jpm fixed in

8/3/2019 Jpm Fixed In

http://slidepdf.com/reader/full/jpm-fixed-in 1/236

Global Fixed Income Markets2012 Outlook

Pavan WadhwaAC

(44-20) 7777-3370

[email protected]

Fabio Bassi

(44-20) 7325-8615

[email protected]

www.morganmarkets.com J.P. Morgan Securities Ltd.

See page 233 for analyst certification and important disclosures.

Contents

Overview 3

Economics 31

Euro Cash 38

European Derivatives 57

United Kingdom 77

US Cross Sector 90

US Treasuries 107

US Interest Rate Derivatives 131

Japan 158

Inflation-linked Markets 176

Australia

New Zealand 219

…………………………………………………….

Interest rate forecasts 226

Recent curve movements 227

Recent sov cash spread

movements 228

Recent sov CDS spreadmovements 229

Sov & bank redemptions 230

Euro area sov ratings / SMPpurchases / Election calendar 231

Euro area fact sheet 232

European Rates StrategyJ.P. Morgan Securities Ltd.November 24, 2011



European Rates StrategyGlobal Fixed Income Markets 2012 OutlookNovember 24, 2011

Pavan WadhwaAC

(44-20) [email protected]. Morgan Securities Ltd.

2

Key 2012 Global Fixed Income Markets Trading Themes

Overview Pavan Wadhwa, Kedran Panageas The negative feedback loop between sovereigns, banks, and the real economy is exacerbating the peripheral debt crisis.

Technicals are poor for peripheral sovereign debt; the existing investor base has been impaired and distressed credit funds

are keeping their distance due to the highly politicised nature of sovereign debt restructurings. We expect further rounds

of forced capital increases for European banks as the Euro zone slips into recession. The ECB is unlikely to step up its

pace of buying significantly, absent a catastrophe. Position for lower German yields and a flatter curve. Underweight

intra-EMU debt relative to Germany. Australian and Chilean duration is attractive. We highlight trades that are positively

convex in peripheral spreads, and others that are uncorrelated to the sovereign debt crisis.

Euro Pavan Wadhwa, Fabio Bassi, Gianluca Salford We see the refi rate at 0.50% and 10Y Bund yields at 1.25% by mid-2012; go long duration and initiate 2s/10s flatteners

with a 100bp target. Underweight peripherals and position for flatter credit curves. EONIA fixings will fall to 30bp as

excess liquidity stays elevated; go long 6Mx6M EONIA. Position for 1s/5s outright and conditional bull flatteners, and

buy receiver structures on 6Mx5Y swaptions. Swap spreads are likely to trade wider than their Lehman peak; we target

2Y and 10Y swap spreads at 145bp and 90bp, respectively, and favour 2Y wideners. Conditional swap spread wideners

offer better risk/reward than outright wideners. In EUR and GBP gamma, buy intermediate-tails and sell short-tails; we

target EUR 3Mx10Y at 9.4bp/day. Favour longs in Bund volatility vs. swaption volatility.

UK Francis Diamond We expect QE gilt purchases to be upsized to a whopping £425bn in 2012. QE, low non-domestic ownership of gilts, and

a flexible currency should limit any increase in gilt yields should the UK fall under the sovereign risk spotlight. 2Y gilts

should trade in a 50-60bp range. 10Y and 30Y gilt yields are likely to decline: we target 1.50% in 10Y gilts by 2Q12 and

expect the 2s/10s gilt curve to flatten to 100bp. Position for wider 5Y and 10Y swap spreads in 1H12.

US Terry Belton, Srini Ramaswamy

Stay long duration in early-2012, targeting 1.70% in 10Y Treasuries. Avoid consensus trades. Synthetic Treasuriescreated by asset swapping cheap foreign bonds should outperform in 2012. FRA/OIS and intermediate swap spreads will

widen initially before narrowing back. Initiate “synthetic” conditional curve trades by replacing swaptions at the front end

with YCSOs. Approach 2012 with a long gamma bias, but look for 3Yx10Y to decline to 6bp/day by 1Q12 end.

Japan Takafumi Yamawaki, Yuya Yamashita We expect 10Y JGBs to trade in a 0.8-1.1% range in 1H12 and 0.9-1.3% in 2H12, with risks biased towards lower JGB

yields. We have a flattening bias on the JPY swap curve. Large negative USD/JPY cross currency basis makes it

attractive to buy JGBs and swap them into USD. 3s/6s basis is likely to widen.

Inflation Jorge Garayo, Francis Diamond, Kimberly Harano

We expect significant declines in headline and core inflation across the board in 2012, although inflation expectations will

remain anchored. Real yields will test new lows, turning negative in Europe and staying close to 0% in 10Y TIPS.

Breakevens will be biased narrower. Expect cash breakevens to underperform relative to inflation swaps in early 2012.

Euro area: Position for flatter breakeven curve and higher peripheral linker yields. UK: Be long intermediate real yields,

and short inflation breakevens in the 10Y sector. US: Declining inflation and nominal yields, along with fiscal tightening,

should cause breakevens to narrow.

Australia / New Zealand Sally Auld Buy AUD duration on dips as offshore demand is likely to stay elevated, the RBA will ease policy, and domestic investors

are short their benchmarks. Position for curve steepening. We are biased towards wider swap spreads. In NZD, we

expect yields to make new lows in 2012 and the curve to be biased steeper.




Pavan WadhwaAC

(44-20) 7777-3370

Kedran Panageas (44-20) 7777-0326J.P. Morgan Securities Ltd

3

Overview

• The negative feedback loop between sovereigns,

banks and the real economy is exacerbating the

peripheral crisis in the Euro zone

Sovereigns

• Fallen angels: peripheral sovereign debt has too

much credit risk to remain in DM portfolios, but

is too large to be easily absorbed in EM

portfolios…

• …while the ‘benchmarking benefit’ that typically

accrues to corporate ‘fallen angels’ will not accrue

to distressed sovereign debt…

• …and, given the highly politicized nature of

sovereign debt restructurings, distressed credit

funds have been slow to invest in this space

• Non-domestic investors, who own nearly half of

all peripheral sovereign debt, have been steadily

selling and are likely to have limited appetite to

add risk in the near term

• France is at risk of losing its AAA rating by mid-

2012, which would cause EFSF capacity to fall by

one-third

Banks

• Spreads on senior bank debt and covered bonds

are near, if not above, Lehman highs, while Euro

zone bank debt issuance has plummeted to below

Lehman levels

• The exposure of core country banks to all

peripheral assets tops €1tn, or 90% of bank

capital and reserves…

• …and the EBA’s €104bn estimate of bank capital

shortfall for Euro zone banks is €150bn below

J.P.Morgan estimates. We expect further rounds

of forced capital increases for Euro zone banks

Economy

• J.P.Morgan economists expect Euro area growth

to significantly undershoot official forecasts…

• …as further fiscal consolidation by Euro area

sovereigns pressures GDP growth lower…

• …and forced bank deleveraging puts 1-2%-pts of

downward pressure on GDP growth

Recommended Trades

• Position for lower yields and a flatter German

curve

• We recommend underweighting peripheral debt

in the first half of 2012; however, we expect

peripheral yields to fall in 2H12 in response to a

concerted ECB/EU/IMF response to the crisis

• Global markets that offer high real yields,

potential policy rate cuts, and high sensitivity of

GDP growth to Europe are attractive; go long

10Y duration in Australia and Chile

• We highlight trades that are positively convex

when peripheral spreads widen/narrow…

• …as well as trades that offer diversification from

the peripheral debt crisis

The sovereign pandemic

If 2010 was the year that the EU sovereign crisis erupted,

2011 was the year that it morphed into a dangerous new

phase. No longer did it limit itself to small nations such

as Greece, Ireland, and Portugal. Rather, the contagion

spread to Italy and Spain and to a certain extent, core

countries such as France. Italy and Spain constitute

nearly one-third of European debt markets, with acombined €2tn of marketable debt outstanding.

This dangerous new phase began around mid-year, when

Greece started to slip on its fiscal targets and the threat of

private-sector burden-sharing became real (Exhibit 1).

Despite making good progress on fiscal targets and

structural reforms in 2010, Greece hit a wall in 2011. By

mid-May, it was massively behind target and the IMF

needed assurance of future funding in order to continue

disbursing promised monies. This led to a second Greek

bailout in July that stipulated a bond exchange on private

holdings that would extend maturities to 15–30 years and

lower coupon rates modestly. This broke the taboo thatsovereign debt was riskless and reinforced the notion

that peripheral debt was now a credit product.

This debt restructuring, although never actually

consummated, as well as the anticipated need for wide-

scale bank recapitalisation and unhelpful political

rhetoric in Italy, were the three main factors that allowed

the crisis to jump containment. By August Italian and

Spanish 10Y bond yields had breached 6% and the ECB

was forced to start buying Italian and Spanish bonds in




Pavan WadhwaAC

(44-20) 7777-3370


4

their Securities Market Programme (SMP). The crisis

took another turn for the worse in September and

October when Greece’s backsliding on reforms

deepened, the EU negotiated a new and more severe

Greek debt restructuring, and Greek Prime Minister

George Papandreou called for a public referendum on the

bailout package.1 In response, Angela Merkel and

Nicolas Sarkozy raised the prospect of a Greek exit

from the Euro zone which reinforced the risk of a

hard default and/or a currency switch in investors’

minds.

All five peripheral countries have now been forced to go

to early elections or technocratic governments since the

onset of the crisis, with Greece and Italy being the most

recent victims. This indicates the difficulty in enacting

needed reforms in the face of economic slowdown, as

well as the high political and implementation risk of fiscal austerity.

Going forward, we think the crisis will continue to

escalate in 1H12. Our baseline view for 2012 includes

the following developments:

1) The market faces a bimodal distribution of

outcomes going forward: either the ECB will step

1 For reference, Appendix-1 presents a detailed timeline of key crisisevents since 2009.

up its bond-buying pace significantly, or another

country may eventually lose access to capital

markets. Our basecase view is that the ECB (and

Germany) will remain deeply reluctant to greatly

increase the size of the its purchases at this time,

which will contribute to an escalation of the crisis in

1H12. This is due to the ECB’s desire to 1) maintain

oversight and aid conditionality, in order to mitigate

moral hazard, and 2) keep monetary and fiscal policy

distinct. However, we expect a concerted

ECB/EU/IMF policy response in 2H12 which will

help to ultimately stabilize yields.

2) Greece may continue to fail to meet its fiscal and

structural reform targets. We think Greece will

eventually need a more severe debt restructuring

than the 50% haircut agreed to at the 26 October

EU summit, which does not go far enough to restoredebt sustainability.2

3) The Euro area will dip into recession, with the

biggest contraction seen in the periphery (see

Economics). A recession dynamic will make it more

difficult for sovereigns to meet fiscal and structural

reform targets.

2 Please see Overview, Global Fixed Income Markets Weekly, 28October 2011.

Exhibit 1: The sovereign debt crisis engulfed Italy and Spain in 2011, morphing into a dangerous new phase. This was spurred by 1) political risk inGreece and Italy, 2) Greek debt restructuring, and 3) concerns around bank capital needsTimeline of the EU sovereign debt crisis; 10Y weighted* peripheral spreads vs. key events; bp

0

100

200

300

400

500

600

700

800

Jan-10 Apr-10 Jul-10 Oct-10 Jan-11 Apr-11 Jul-11 Oct-11

Greekcrisis

EFSFconcerns

EBA 1st stress test

Irishconcerns

EU plans mandatoryPSI with ESM aid;

Irish bailout

Discussions onEFSF2 / ESM

Portuguese bailout;discussions of Greek PSI

EBA 2nd stress test;2nd Greek bailout

Italian crisis and ECBbuys IT/ES

Greek backsliding /renegotiation of PSI; referendum and Italian politicalgridlock

0

100

200

300

400

500

600

700

800

Jan-10 Apr-10 Jul-10 Oct-10 Jan-11 Apr-11 Jul-11 Oct-11

Greekcrisis

EFSFconcerns

EBA 1st stress test

Irishconcerns

EU plans mandatoryPSI with ESM aid;

Irish bailout

Discussions onEFSF2 / ESM

Portuguese bailout;discussions of Greek PSI

EBA 2nd stress test;2nd Greek bailout

Italian crisis and ECBbuys IT/ES

Greek backsliding /renegotiation of PSI; referendum and Italian politicalgridlock

* Spreads to Germany of Greece, Ireland, Portugal, Italy, and Spain. Weighted by proportional contribution to the J.P. Morgan EMU Bond Index.Note: For reference, Appendix-1 presents a detailed timeline of key crisis events since 2009.




Pavan WadhwaAC

(44-20) 7777-3370


5

4) Faced with recession and disinflation, the ECB will

cut the refi rate to 50bp by June 2012, in

successive 25bp moves occurring at its December,

March, and June meetings. We expect the deposit

rate to be reduced to 25bp at the December meeting

(see Economics).

5) Portugal and Ireland will need to renegotiate

funding packages in mid-to-late 2012, which will

likely require private sector involvement (PSI).3

6) France is at risk of losing its AAA rating given

contingent commitments to other sovereigns and

French banks, especially if economic growth

disappoints. This could reduce EFSF capacity by up

to one-third (see Euro Cash).

Why is this crisis so pernicious?

Exiting this crisis has proven extremely difficult for

several reasons. First, we are in a negative feedback loop

that is proving difficult to stabilize (Exhibit 2). The

banking crisis and recession of 2008/2009 damaged

sovereign balance sheets, and today’s sovereign crisis

(and the policy response) are in turn damaging bank

balance sheets and stunting economic growth. As with

3 Please see Overview, Global Fixed Income Markets Weekly, 8 July2011.

most negative feedback loops, a circuit-breaker is needed

in the form of some external factor such as outside

money, central bank intervention, or a positive growth

shock from abroad. In the sections below, we discuss

each of the three legs of the crisis in more detail –

sovereigns, banks, and the economy – with regard to howthey will drive outcomes in the periphery in 2012.

Second, and critically, the scale of the problem is much

larger than anything experienced before. The Russian

and Argentine defaults in 1998 and 2001, respectively,

only affected between €100–€125bn of marketable debt

each. This was considered to be large at the time, but it

pales in comparison to the nearly €2.5tn of marketable

debt outstanding today across Italy, Spain, Greece,

Ireland, and Portugal.4 Given the scale of the problem, it

is difficult to find a credible and/or willing lender of last

resort.

Third, EU policy response to date has been fractured and

behind the curve. The financially sounder core countries

and the ECB have been reluctant to support at-risk

countries to the extent necessary, for fear of socialising

debt and exacerbating moral hazard. This problem will

4 The marketable debt figure includes T-bills, zero-coupon notes, and

bonds, but does not include official sector loans, private sector loans,commercial credits, etc. We estimate total debt outstanding of the five

peripherals is between €3.0–3.5tn, based on current GDP and debt/GDPratios.

Exhibit 2: The negative feedback loop between sovereigns, banks, and the real economy is exacerbating the peripheral crisisTransmission mechanism between the three pillars of the EU sovereign debt crisis

Banks

Sovereign

Economydeleveraging

credit losses

PROBLEM:

• Overleverage

• Loss of capital market access

• Solvency crisis

PROBLEM:

• Exposure to peripheral assets

• Overleverage

• Loss of capital market access

PROBLEM:

• Slow growth

• Recession risk

• Social unrest

f i s c a l c

o n s o l i d

a t i o

n w r i t e

d o w

n s

b a i l o

u t s

r e c e s s i o n /

s o c i a l u

n r e

s

t




Pavan WadhwaAC

(44-20) 7777-3370


6

be difficult to overcome until there is more control, more

trust, and more homogenisation of fiscal policies across

Europe. The fiscal integration process will take years if not decades to achieve, and will require profound EU

treaty changes.5 We now discuss each of the three pillars

of the EU crisis in more detail.

1. The sovereign debt and funding crisis

First, since the onset of the financial crisis in 2008,

sovereign balance sheets have been severely damaged: it

is, in the words of German Chancellor Angela Merkel, at

least a 10-year project to restore fiscal health. Debt-to-

GDP ratios amongst the peripherals are up 37%-pts on

average (Exhibit 3). Even stronger countries now appear

stretched.

Second, rating agencies have been relentless in

downgrading peripheral sovereign debt. The average

peripheral rating has fallen from A+/Aa2 two years ago

to BBB-/Ba1 today (Exhibit 4). This stands in stark

contrast to the rest of the developed market (DM)

sovereign universe, the vast majority of which has

minimal credit risk and is rated AA or higher (more on

this later).6

Third, the crisis has severely damaged market liquidity.

This is evident in wide bid/offers, thin trading volumes,

and skyrocketing market volatility for peripheralsovereigns (Exhibit 5).

Together, the heightened credit risk and market

volatility have severely impaired the investor base for

peripheral sovereign debt. The natural investor base

consists mainly of rates-based investors, including

central banks, commercial banks, insurance companies,

pension funds, etc. These investors begin to shy away at

the first sign of credit risk or investment volatility.

Further, these investors also face heightened regulatory

scrutiny. This consists of additional disclosure

requirements, capital stresses on assets held in banks’

held-to-maturity books, a possible move to non-zero risk-weightings, and the need for deleveraging or asset sales.

This treatment makes them less able to withstand the

5 See Overview, Global Fixed Income Markets Weekly, 19 August 2011for a discussion on the lengthy process required to change EU treaties.6 In fact, Moody’s classifies only three DM issuers (outside of the

peripherals) as single-A (Israel, the Czech Republic, and South Korea).

These are stronger single-A credits, all of them rated A1 at Moody’sand two split-rated at S&P. Specifically, Israel is rated AA- at S&P,

the Czech Republic is rated AA at S&P, and South Korea is rated A+ atS&P.

Exhibit 3: Peripheral sovereign balance sheets have been severelyimpaired…Debt-to-GDP ratios across selected Euro zone countries, end-2011E* vs. end-2008; %

2011E* 2008 Change

Greece 158 111 47Ireland 112 44 68

Portugal 102 72 30

Italy 120 106 14

Spain 68 40 28

Average**, peripherals 112 75 37

Belgium 97 90 7

France 85 68 17

Germany 82 66 16

Euro area 88 70 18 * Official estimate as provided by Eurostat.** Simple average.

Source: Eurostat

Exhibit 4: …and rating agencies have been relentless in downgradingperipheral sovereign debt…Number of notches that peripheral countries have been downgraded by Moody’s andS&P over the past two years

18-Nov Dec 2009

Downgrade

(notc hes ) 18-Nov Dec 2009

Downgrade

(notches)

Greece CC NEG BBB+ (WN) 12 Ca DEV A2 14

Ireland BBB+ AA 5 Ba1 NEG Aa1 9

Portugal BBB- NEG A+ 5 Ba2 NEG Aa2 9

Italy A NEG A+ 1 A2 NEG Aa2 3

Spain AA- NEG AA+ 2 A1 NEG Aaa 4

Av erage* BBB- A+ 5 Ba1 Aa2 8

S&P Moody's

* Simple average across the peripherals based on a unit ratings scale (AAA=1, AA+ = 2, AA = 3, etc.).Note: NEG, POS, DEV indicates the ratings outlook is negative, positive, or developing, respectively. WN (WP) indicates the rating is on negative (positive) watch.Source: Bloomberg, Moody’s, S&P

Exhibit 5: …while market liquidity has been severely impaired, leadingto wide bid-offers and skyrocketing volatility in peripheral yields6M standard deviation of daily yield changes in Italian 10Y b/m vs. average monthlybid-offer spread on Italian 10Y bonds*bp/day bp of yield*

2

4

6

8

10

12

14

16

Nov 08 May 09 Nov 09 May 10 Nov 10 May 11 Nov 11

1

2

3

4

5

67

8

9

6M yield vol

bid-offer spread

* Estimated from the average bid-offer spread reported in price terms by MTS andprice and duration statistics of the J.P.Morgan Global Bond Index for Italy 7-10Ybonds.Source: MTS, J.P.Morgan




Pavan WadhwaAC

(44-20) 7777-3370


7

MTM volatility that they previously would have been

able to weather through fair value accounting in held-to-

maturity portfolios. As a result, the traditional peripheralsovereign investor base has grown significantly more

risk-averse over the past year, leading to a constant

stream of media reports of large institutional investors

paring their peripheral debt holdings.

This selling has pushed yields significantly higher:

peripheral sovereign debt now has too much credit

risk to remain in DM portfolios, but is too large to be

easily absorbed in EM portfolios. First, outstanding

debt from EU peripherals comprises over half of

outstanding EM sovereign debt, making it a huge

challenge to absorb (Exhibit 6). Second, downgraded

sovereign debt will most likely not be added to EMsovereign indices (which have fairly low country wealth

and income thresholds) or to high-yield or EM corporate

credit indices. Thus, the ‘benchmarking benefit’ that

typically accrues to corporate fallen angels when they

migrate from investment-grade to high-yield territory is

unlikely to accrue to distressed sovereign debt. For

instance, we estimate that the yield on the typical auto

industry ‘fallen angel’ fell an average of 150bp after

crossing into high-yield territory, due to demand from

benchmarked investors in high-yield space (Exhibit 7).

Third, even though credit-focused hedge funds or

distressed investors may theoretically be interested,

the highly politicised nature of sovereign

restructurings makes them more difficult to analyse

as compared to the better-defined corporate

restructuring protocol.7 For instance, GM is routinely

pointed to as an example of a politicised bankruptcy in

corporate space, where the government’s involvement

led to the typical recovery priority being discarded, to the

detriment of bond investors and to the benefit of

unionized employees. Although several investors

explored legal options, they ultimately settled for the

government-negotiated outcome as it was too

difficult/expensive to fight in court. Such treatment is

likely to be amplified many times over in the sovereign

debt space, where there is no defined priority of

payments in a debt restructuring. This is ultimately to

the detriment of the EU as it is likely to dissuade new

bondholder investment in peripheral sovereigns. For

instance, the market reacted quite negatively when

Merkel and Sarkozy first announced in late 2010 that any

official loans granted out of the ESM after 2013 would

7 See Overview, Global Fixed Income Markets Weekly, 10 December 2010, for sovereign debt legal issues.

have preferred creditor status over private debt

investors8. Partly due to these concerns, distressed

credit funds have been slow to invest in the peripheral

sovereign space and will likely remain wary.

8 For instance, see “Cross-country variations in capital structures: therole of bankruptcy codes”, Viral Acharya et. al., Journal of Financial

Intermediation, 2011, and “Creditor rights and Corporate risk-taking”,Viral Acharya et. al., Journal of Financial Economics, forthcoming.

Exhibit 6: Peripheral sovereign debt has too much credit risk to remainin DM portfolios but is too large to be easily absorbed in EMportfolios…Outstanding developed and emerging markets debt universe* by Moody’s rating**; %of total

Rating DM ex peripherals Peripheral EM Total

Aaa 48% - - 48%

Aa 26%^ - 4% 30%

A 1%^^ 7% 1% 9%

Baa - - 6% 6%

Ba - 1% 3% 4%

B - - 2% 2%

<= Caa - 1% - 1%

Total 75% 9% 16% 100%

* Sample consists of 105 countries of which 27 (78) are developed (emerging) debtmarkets. Total gross debt outstanding is $55tn of which $46tn ($9tn) is from developed(emerging) markets. Debt includes all outstanding bonds, loans, and other obligations.** Local currency long-term debt rating. Note: the US is rated AAA by Moody’s.

^ Consists of three countries: Belgium, Hong Kong, and Japan.^^ Consists of three countries: Israel, South Korea, and the Czech Republic.Source: Moody’s, IMF

Exhibit 7: …and the ‘benchmarking benefit’ that typically accrues tocorporate ‘fallen angels’ will not accrue to distressed sovereign debt.Moreover, given the highly politicised nature of sovereignrestructurings, distressed credit funds have been slow to invest in thisspace

Avg. yield-to-maturity of 10Y GM, Ally Financial, Ford, and Ford Motor Credit* bondsaround the 5 May 2005 downgrade to HY**; %

6.5

7.0

7.5

8.0

8.5

9.0

9.5

10.0

10.5

-3 -2 -1 0 1 2 3# months around downgrade to HY**

* Specific bonds used are: GM 7.125% Jul13, Ally Financial 6.5% Feb16, Ford 7.45%Jul31, and Ford Motor Credit 7% Oct13. We use the Jul31 bond for Ford Motor Co. asit did not have a large/liquid 10Y note outstanding at the time.** S&P downgraded Ford, GM, and their subsidiaries on the same date, 5 May 2005.




Pavan WadhwaAC

(44-20) 7777-3370


8

Additionally, non-domestic investors are likely to

have limited appetite to add risk in the near term.

Non-domestic investors own nearly half of all peripheral

sovereign debt; their share ranges from nearly 40% in

Spain to over 80% in Ireland (Exhibit 8). These

investors are typically amongst the first to sell and, once

having exited the asset class, will likely be reluctant to

return in the near future. We expect them to provide a

steady drip of selling pressure going forward.

Given these poor demand technicals, sovereign marketaccess remains tenuous and there is no firm ceiling for

spreads. Indeed, European sovereign spreads are

currently close to their lifetime wides, despite the steady

stream of bailouts, crisis measures, and backstop ECB

buying (Exhibit 9). Even EFSF paper has widened to

120bp over Libor, indicating an aversion to credit risk,

ratings downgrade risk, and structural complexity

(Exhibit 10).

Exhibit 9: Given these poor demand technicals, sovereign spreadscontinue to trade close to their lifetime wides, despite ECB buying10Y spreads to Germany; 18 November 2011 vs. 2Y maximum; bp

18-Nov-11 Max 18-Nov / Max

Greece 2425 2471 98%

Ireland 625 1142 55%

Portugal 910 1161 78%

Italy 494 575 86%

Spain 469 486 97%

Belgium 282 309 91%

France 147 188 79%

Av erage* 765 905 85% * Simple average.

Exhibit 10: Even EFSF paper has widened to 120bp over Libor,

indicating an aversion to credit risk, ratings downgrade risk, andstructural complexity10Y ASW spreads for EFSF* and France; bp

-20

0

20

40

60

80

100

120

140

Aug Sep Oct Nov

France EFSF

* EFSF 3 3/8 July 2021

Exhibit 8: Additionally, non-domestic investors own nearly half of allperipheral sovereign debt. These investors have been steadily sellingand may be reluctant to return to this asset class in the near future

% of marketable debt outstanding held by non-domestic investors; %% non-domestic

Greece 67%

Ireland 83%

Portugal 80%

Italy 45%

Spain 38%

Wtd av g., peripherals* 49%

France 66%

Germany 81%

* Weighted by marketable debt outstanding.Source: National central banks. Data as of latest available update for each sovereign;

generally 2Q11/3Q11.

Exhibit 11: Although core-country banks have been cutting their exposure to peripherals, their total exposurestill tops €1tn, or almost 90% of bank capital and reservesExposure of core-country European banks to peripheral country assets as % of bank capital and reserves; 2Q11 vs. 4Q09; €bn

Claims on Austria Belgium France Germany Netherlands 2Q11 4Q09 % chg

Greece 2 1 38 15 3 60 101 -41%

Ireland 2 18 22 76 12 129 214 -40%

Italy 17 17 287 111 36 468 576 -19%

Portugal 1 2 18 25 5 50 81 -37%

Spain 6 16 104 122 53 302 420 -28%

Total 28 54 469 349 109 1009 1392 -28%

Capital and reserves (C&R) 92 58 492 391 102 1135 1081 5%

Exposure / C&R 30% 93% 95% 89% 107% 89% 129% --

Origin country Total exposure

Source: BIS, ECB. Data as-of 2Q11.




Pavan WadhwaAC

(44-20) 7777-3370


9

2. Bank funding and equity capital crisis

Sovereign fiscal issues and slowing economic growth

have direct repercussions on banking sectors. Due to

cross-border capital flows, losses are transmitted quickly

from borrower to lender countries. Although core-

country banks have been cutting their exposure to

peripherals, their total peripheral exposure still tops €1tn,

or almost 90% of bank capital and reserves (Exhibit 11).

A sizable portion of this total exposure is peripheral

sovereign debt; German and French banks alone hold

nearly €180bn, which is now imperilled given the PSI

precedent set by Greece (Exhibit 12). In addition,

anticipated recessions in the peripherals will increase the

losses on other assets such as consumer and corporateloans. Non-performing loan ratios (NPLs) are up sharply

across the periphery since the recession of 2009 and will

likely continue rising as the periphery slides back into

recession (discussed further below; Exhibit 13).

Not surprisingly, these exposures have pummeled bank

stock prices and impaired funding access. Bank equity

prices are down 35–40% on the year and spreads on

senior bank debt and covered bonds are at multi-year

highs (Exhibit 14). Additionally, debt issuance has

ground to a halt; rolling quarterly long-term debt

issuance is at a 5-year low, lower than even Lehman

levels (Exhibit 15). Net, we estimate that core-country

banks have between 0-2 quarters of funding coverage at

this time.9 ECB reliance has steadily increased and the

ECB’s balance sheet likely will have to expand

massively if it is to contain the incipient bank funding

crisis (discussed further below). For instance, the Wall

Street Journal recently reported on the growing volume

9 See Overview, Global Fixed Income Markets Weekly, 30 September 2011.

Exhibit 13: The recession and its aftermath have led to sharply higher NPLs in the periphery, even on non-sovereign lendingNon-performing loans as a % of total gross domestic loans, quarterly data for Greece*and Spain**; %

0

2

4

6

8

10

12

2006 2007 2008 2009 2010 2011

Greece

Spain

* For 2006-2007, quarterly data interpolated from year-end data.** Total of loans to household and corporate sector.Source: Bank of Greece, Bank of Spain

Exhibit 14: As a result of large exposures to peripheral assets, spreadson senior bank debt and covered bonds are near, if not above, Lehmanhighs…Senior bank debt and covered bond ASW spreads; current as of 18 Nov 2011; bp

-50

0

50

100

150

200

250

2007 2008 2009 2010 2011

Senior

Covered bonds

Current : 216bp

Current : 198bp

Source: J.P. Morgan Maggie Credit Index

Exhibit 15: …and Euro zone bank debt issuance has plummeted tobelow Lehman levelsRolling 3M sum of gross bank debt and covered bond issuance for Euro zone banks;all currencies/all domiciles; €bn/quarter

0

50

100

150

200

250

300

2007 2008 2009 2010 2011

Bank debt + CB

Bank debt only Avg Total = €130bn

(€90+40bn)

Total = €65bn

(€53+€12bn)

Total = €50bn

(€33+€17bn)

Source: Dealogic

Exhibit 12: German and French banks alone hold nearly €180bn of peripheral public sector debt, which is now imperilled given the PSIprecedent set in Greece

Exposure of core-country European banks to peripheral public sector debt (includingsovereign debt) as of 2Q11; €bn

Claims on Belgium France Germany 2Q11 4Q10 % chg

Greece 1 7 9 19 32 -41%

Ireland 0 2 2 5 6 -18%

Italy 11 74 33 125 131 -4%

Portugal 1 4 6 17 20 -16%

Spain 3 21 20 48 50 -4%

Total 17 108 70 215 240 -11%

Origin country Total exposure

Source: BIS. Data as-of 2Q11.




Pavan WadhwaAC

(44-20) 7777-3370


10

of collateral swaps designed to boost banks’ access to

ECB facilities.10

On the equity front, European banks are overleveraged,

especially given growing recession risk and market-

implied losses on sovereign debt. The recently

announced capital shortfall of €104bn11 for Euro zone

banks is roughly €150bn below J.P.Morgan estimates

(Exhibit 16).12 To be sure, the 9% core Tier 1 hurdle that

the EBA used is aggressive, but in our opinion the rapid

deterioration in economic fundamentals is a more

significant driver. We therefore believe that further

capital increases (above and beyond the €104bn estimate)

will likely be required of banks. Capital needs are likely

to be higher still if other peripheral countries such as

Ireland and Portugal eventually undergo debtrestructuring, which is our basecase expectation.13

Bank stress feeds back into the sovereign crisis by

increasing sovereign contingent liabilities. Particularly at

risk are Spain, Belgium, and France. Spain has large

unrecognised capital shortfalls at its cajas14, while the

prospect of bank guarantee schemes and capital

injections have already been cited as risk factors for

France’s AAA rating.

3. Spillover to the real economy

Sovereign fiscal austerity and bank deleveraging arespilling over to the real economy, stunting economic

growth. Our economists expect Euro area real GDP to

grow at a rate of -0.6% in 2012, with the contraction

concentrated in the periphery (we forecast 2012 GDP-

weighted growth of -1.8% in the periphery) (Exhibit 17).

This is driven in large part by the large and sustained

fiscal consolidation that must be done. Based on prior

episodes of fiscal consolidation, we estimate that each

1%-pt of fiscal tightening in the primary deficit lowers

10

See ‘ Banks face funding stress’, 17 November 2011, Wall Street Journal . In these collateral swaps, a bank pays a spread to swap

collateral which is not eligible for ECB repo in order to receive higher-quality collateral which is eligible. This enables the bank to boost ECB

access and is a sign of the increasing scarcity/cost of market funding.11 Note that the €104bn shortfall estimate is for Euro zone banks only.

The total estimated shortfall of bank capital across the entire Euro area

is €106bn.12 We estimate that the largest discrepancy stems from removing the

original macroeconomic stresses applied in July; a modest portion

(roughly €20bn) stems from leaving out the cajas in Spain13 Please see Overview, Global Fixed Income Markets Weekly, 8 July2011.14 See EBA Stress Test Results Analysis: Must do much better , RobertoHenriques, 8 July 2011.

GDP growth by 0.7%-pts (Exhibit 18).15 Additionally,

we estimate that each 1%-pt of bank deleveraging

depresses economic growth by 0.6–0.8%-pts (Exhibit

19). Extrapolating to the Euro area suggests 1–2%-pts of

downward pressure on GDP growth from bank

deleveraging. Thus, in our opinion, economic risk is

skewed to the downside.

15 See Growth in the Euro area periphery to underwhelm, 16 September 2011, Nicola Mai.

Exhibit 16: The EBA’s €104bn estimate* of bank capital shortfall for Euro zone banks is €150bn below J.P. Morgan estimates. This isunlikely to be the last round of forced capital increases for European

banks, especially if economic growth disappoints and sovereigns suchas Portugal and Ireland eventually undergo debt restructuringCapital shortfalls announced by the European Banking Authority (EBA) for variousEuro zone countries’ banking sectors vs. J.P. Morgan estimates**; €bn

EBA target

capital

J.P.Morgan

estimated capital

shortfall

J.P.Morgan -

EBA estimate

104 254 151

France 9 48 39

Germany 5 43 38

Spain 26 53 27

Italy 15 35 20

Greece 30 40 10

Portugal 8 17 9

Others 11 18 7

Country

Total

* Note that the €104bn shortfall estimate is for Euro zone banks only. The totalestimated shortfall of bank capital across the entire Euro area is €106bn.** See ‘The Great Bank Deleveraging: Banking Sector Outlook 2012 ’, 4 November 2011. Assumptions underlying J.P. Morgan estimates are 1) 9% Tier 1 capitalthreshold; 2) original Basel 2.5 Tier 1 capital ratios as provided by the EBA; 3) sovereign MTM on entire holdings of sovereign debt as of end-October; 4) macrostresses as provided by the EBA in June 2011; and 5) cajas included in Spain.Source: EBA, J.P. Morgan

Exhibit 17: J.P.Morgan economists expect Euro area growth in 2012 tosignificantly undershoot official forecasts…2012 GDP growth forecasts; % YoY

JPM Official EC JPM - Official JPM - EC

Greece -6.6 0.8 -2.8 -7.4 -3.8

Ireland 0.0 1.6 1.1 -1.6 -1.1

Portugal -3.9 -2.8 -3.0 -1.1 -0.9

Spain -1.1 2.3 0.7 -3.4 -1.8

Italy -1.5 0.6 0.1 -2.1 -1.6

Wtd avg., peripherals* -1.8 1.1 0.0 -2.8 -1.8

Belgium -0.4 2.3 0.9 -2.7 -1.3

France -0.1 1.0 0.6 -1.1 -0.7

Germany 0.3 1.0 0.8 -0.7 -0.5

Euro area* -0.6 1.1 0.5 -1.7 -1.1 * Weighted by 2011 GDP.Source: National treasury ministries, European Commission and J.P.Morgan forecasts




Pavan WadhwaAC

(44-20) 7777-3370


11

We note that J.P.Morgan economic projections are

significantly more bearish than official forecasts, either

by national treasury ministries or by the EuropeanCommission (EC).16 For example, the J.P.Morgan GDP

forecast for the Euro area is over 1%-pt more bearish

than the EC forecast, and the forecast for the periphery is

nearly 2%-pts more bearish (Exhibit 17).

Slower growth feeds back into the sovereign crisis by

increasing the fiscal deficit, forcing the sovereign into a

vicious cycle of more cutbacks which, in turn, negatively

impact growth and damage political and popular support.

For instance, the realised GDP contraction in Greece has

been several %-pts worse than forecast by the EU/IMF,

coming in close to -5.5% for 2011 vs. an estimated -3%

originally. Similarly, unemployment has surged, leadingto widescale social protests and political backsliding

(Exhibit 20). Across the periphery, all five ruling

coalitions have changed since the onset of the crisis.

However, any benefits from a technocratic government

will likely accrue only over the long term. In the short

term, the effect can be more market-negative if it

increases uncertainty or makes it more difficult to

implement reforms. Net, the result of continued GDP

contraction in 2012 is likely to be further slippage on

fiscal targets, higher deficits, increases in debt/GDP

trajectories, and heightened risk of PSI in other

peripherals besides Greece (despite official declarations

that Greece is a one-off).

Analysing the risk posed by peripherals

The catalogue of risks facing the Euro zone is daunting,

especially where Greece and Italy are concerned. As

discussed previously, Italy’s size and cross-border

linkages make it systemically important. If Italy were to

lose market access, it would greatly increase the

contingent liabilities and funding costs of other

sovereigns, and would most likely result in Spain losing

market access as well. We thus view Italy as the most

challenging issue for 2012.

Why is Italy at risk? First, Italy’s debt burden is large.

For instance, we estimate that Italian debt/GDP will peak

near 125% in 2013, and slowly decline towards 114% by

2020 (Exhibit 21). This is based on our economists’

projections of economic growth and takes into account

recently announced austerity measures (and anticipated

countermeasures when growth disappoints). These

projections envision, for instance, that Italy will run a

16 The latter is typically more up-to-date for latest economic trends.

Exhibit 18: …as urther fiscal consolidation by Euro area sovereignsputs significant downward pressure on GDP growthFiscal consolidation targets and their impact on GDP growth*; % of GDP

2011 2012 2013 2011 2012 2013

Greece 6.0 5.1 0.8 -4.2 -3.6 -0.6

Ireland 2.2 1.6 2.8 -1.5 -1.1 -2.0

Portugal 5.7 6.1 2.2 -4.0 -4.3 -1.5

Spain 2.7 2.6 2.5 -1.9 -1.8 -1.8

Italy 0.8 3.5 2.3 -0.6 -2.5 -1.6

Wtd avg., peripherals* 2.1 3.4 2.3 -1.5 -2.3 -1.6

Belgium 0.2 1.2 1.4 -0.1 -0.8 -1.0

France 1.4 1.6 2.0 -1.0 -1.1 -1.4

Germany 0.0 1.0 0.0 0.0 -0.7 0.0

Euro area 1.2 2.1 1.4 -0.8 -1.5 -1.0

Fiscal consolidation target Impact on GDP growth*

* Fiscal tightening defined as the change in the cyclically-adjusted primary balance.We estimate the follow-on change in GDP growth to be -0.7 times the annual fiscaltightening. Please see Growth in the Euro area periphery to underwhelm, 16September 2011, Nicola Mai.Source: National treasury ministries, EC and J.P.Morgan estimates.

Exhibit 19: Further, the US experience suggests that annual GDPgrowth declines by 0.7%-pts for each 1%-pt increase in bank capitalratio; extrapolating to the Euro area suggests 1–2%-pts of downwardpressure on GDP growth from bank deleveragingResidual of US vs. global real GDP growth (YoY) regressed against US banks’ equitycapital ratio*; January 1999–March 2011; quarterly data; %

y = -70.0x + 6.7

R2 = 31%

-3

-2

-1

0

1

2

3

4

8.0% 8.5% 9.0% 9.5% 10.0% 10.5% 11.0% 11.5%

Bank equity capital ratio*

* Bank equity capital ratio defined as the bank equity capital of FDIC-insuredcommercial banks and savings institutions divided by total assets. Tier-1 capital ratiohas roughly a 1:1 beta to this definition of equity capital ratio over the last 20 years.Source: FDIC, J.P. Morgan

Exhibit 20: The unemployment rate has climbed over 2.5%-pts onaverage since the start of the peripheral debt crisis, leading to socialunrest, political backsliding, and slippage on structural reformsUnemployment rate; current vs. end-2009; %

Current Dec-09 Chg.

Greece 17.6 10.2 7.4

Ireland 14.2 12.9 1.3

Portugal 12.5 11.3 1.2

Italy 8.3 8.3 0.0

Spain 22.6 19.1 3.5

Av g., peripherals* 15.0 12.4 2.7

Euro area 10.2 10.1 0.1

* Simple average.Source: Eurostat




Pavan WadhwaAC

(44-20) 7777-3370


12

sustained primary surplus equivalent to 5% of GDP, and

that real growth will climb to 0.8% per annum, with

annual inflation at 2%.

Second, given that the Italy’s debt burden is second only

to Greece’s, any slippage on fiscal consolidation or debt

costs could quickly turn the liquidity crisis into a

solvency crisis. To illustrate this, we calculate the

sensitivity of debt/GDP ratios projected in 2020 and

2030 to various economic assumptions. Thesesensitivities are shown in Exhibit 22. We find that:

• A -1%-pt shock to estimated nominal GDP growth

rates (vs. J.P. Morgan base case trajectory shown in

Exhibit 21) will boost debt/GDP in 2020 (2030) by

13% (33%)17…

• A -1%-pt shock to the primary surplus each year

(again vs. J.P. Morgan basecase trajectory) will

boost debt/GDP in 2020 (2030) by 10% (26%)

• A 1%-pt higher borrowing rate on new Italian debt

(vs. J.P. Morgan assumption of 6.5%) will boost

debt/GDP in 2020 (2030) by 7% (22%)

Third, while any one of these shocks on its own would

be negative and leave the debt burden at a very high

level, the shocks are likely to be correlated. That is, if

growth surprises to the downside, it will be more difficult

to run the targeted primary surplus, and borrowing costs

will likely rise. Thus, even mild underperformance can

17 Note that a shock to real growth and a shock to inflation have the

same impact on nominal GDP growth. Thus, they would have the sameimpact on debt sustainability.

rapidly spiral out of control. To be sure, a mild

outperformance will have the opposite impact, with

debt/GDP ratios falling more than is reflected in our

baseline scenario.18

How likely is Italy to deviate to the downside from our

basecase scenario? Unfortunately, our assumptions may

already by generous. On the economic front, risk is

skewed to the downside. In the short term, this is driven

18 However, we note that the sensitivity of the debt/GDP trajectory to

interest rate and growth rate shocks is non-linear due the compounding

of financing costs over time. Therefore, the betas are somewhat larger

(in absolute terms) when the situation is deteriorating, and somewhat

smaller when the situation is improving. For instance, for a 2020horizon, the beta for a positive shock (+1%pt to growth, or -1%-pt to

the borrowing rate) is roughly 5-10% smaller than the beta for a

negative shock (in absolute value terms). For the 2030 horizon, the

beta for a positive shock is roughly 15-20% smaller than the beta for anegative shock. On the other hand, the beta to primary surplus

assumption is linear since it is not multiplied by the outstanding debtstock each year but rather is added to it.

Exhibit 21: Fiscal sustainability in Italy and Spain is uncertain…J.P.Morgan’s basecase forecast of debt/GDP trajectory* and key macro assumptions underlying the trajectory for Italy and Spain**;blue selection indicates peak debt/GDP

2011 2012 2013 2014 2015 2016 ...2020 ...2030

Debt/GDP (% of GDP)* 121 124 125 125 123 121 114 100

Nominal grow th rate (% YoY) 2.0 -0.1 -0.3 1.2 2.7 3.0 2.8 2.8

Primary surplus (% of GDP) 0.5 3.2 4.8 5.0 5.0 5.0 5.0 5.0

Wtd av g. coupon rate*** (%) 4.1 4.6 4.9 5.1 5.3 5.4 5.9 6.3

Debt/GDP (% of GDP)*,** 70 75 81 85 86 87 84 77

Nominal grow th rate (% YoY) 1.9 -0.1 0.6 1.6 2.2 2.4 3.4 3.4

Primary surplus (% of GDP) -4.9 -3.2 -1.5 -0.6 0.8 2.0 3.0 3.0

Wtd av g. coupon rate*** (%) 3.7 4.5 4.9 5.2 5.4 5.6 6.0 6.4

Spain

Italy

* Note that J.P.Morgan estimate for 2011 debt/GDP is slightly more bearish than the latest available official forecast (shown in Exhibit 3)which is based on Eurostat data.** Spanish debt/GDP trajectory does not include expected bank recapitalisation needs of roughly €50bn, covering both banks and cajas(J.P. Morgan estimate). 2011 Spanish GDP is estimated near €1.08tn, so this equates to roughly 5% of current GDP. Rating agencyestimates of Spanish bank recap needs run from around €20bn to as high as €120bn in stressed cases.*** Based upon current weighted average coupon rate and average marginal borrowing rate of 6.50% on new issuance.

Exhibit 22: …and highly sensitive to adverse shocks in GDP growth,primary surplus and funding costsChange in projected 2020 and 2030 debt/GDP ratio for a -1%-pt shock to nominalgrowth and primary surplus assumptions*, and for a +1%-pt shock to borrowing rateassumption*; ratios

Year

Nominal

growth rate

Primary

surplus

Marginal borrow ing

rate on debt

2020 13 10 7

2030 33 26 22

2020 9 10 52030 22 25 16

Spain

Variable Shocked

Italy

* Measured around the J.P.Morgan baseline forecast, holding other variables constant.J.P. Morgan base case assumptions shown in Exhibit 21.




Pavan WadhwaAC

(44-20) 7777-3370


13

by the possibility of European policy dysfunction, and in

the long term, by Italy’s deep-seated structural

challenges. For instance, we assume a long-term 0.8% per annum growth rate even though Italy has only

managed a 0.3% annualised growth rate over the past

decade. On the financing front, Italian yields have spiked

past 7% in recent weeks, so that a 6.5% borrowing cost

may also be ambitious.

Fiscal pressures on Spain are also significant. Spain has a

lower debt burden than Italy does, but is running a higher

deficit. Thus, we expect Spanish debt/GDP to peak later,

somewhat shy of 90% of GDP in 2016 (Exhibit 21). This

assumes that the incoming administration in Spain will

pursue additional fiscal tightening relative to what has

been budgeted so far, eventually achieving a primarysurplus of 3% per annum. However, this trajectory does

not include expected bank recapitalisation costs that

could add between 2-12% debt/GDP, depending on the

estimate.19

The sensitivity of Spain’s debt/GDP trajectory

to shocks in macro assumptions is similar to Italy’s, but

somewhat smaller given a lower starting debt burden

(Exhibit 22).

Thus, there is significant slippage potential, especially

for countries whose debt levels are already high. It may

be challenging for Italy to achieve the necessary fiscal

consolidation amidst a recession and a high level of

social resistance. The recently appointed technocraticgovernment may prove more effective than the previous

administration, but the impact of any structural changes

that it implements is not likely to be felt on the economy

for a long time.20 Given these pressure points, Italy, and

possibly Spain, risk losing market access in 2012.

Possible policy response if another

peripheral country loses market access

Given their size, it would be challenging to fund Italy

and Spain in the event that they lose market access.

We estimate total funding needs for Italy and Spain of

nearly €900bn over the next three years (Exhibit 23).

This includes bond redemptions (conventionals and non-

conventionals) and anticipated deficits, but assumes that

both continue to roll T-bills in the private market. Also,

19 J.P. Morgan estimates roughly €50bn of Spanish bank recapitalisation

needs, covering both banks and cajas. Rating agency estimates run from

around €20bn to as high as €120bn in stressed cases.20 See Euro Cash, Global Fixed Income Markets Weekly, 11 November

2011, for an analysis of Italy’s last experience with a technocraticgovernment.

note that funding needs are front-loaded (€350bn in

2012).

Unfortunately, the official resources available to deal

with this outcome are limited. Euro area sovereigns are

already stretched financially, and non-Euro area

sovereigns see little reason to support fundamentally

wealthy countries. We estimate that in extremis, there is a

maximum of roughly €725bn available to support Italy

(€525bn) and Spain (€200bn), as shown in Exhibit 24.

This arises from the following sources:

1) We believe that Italy can raise at most €125–150bn

from one-off measures such as wealth taxes,

privatisations, and sales of gold and FX reserves

over the next 12 months.21 Please see Appendix-2

for our calculations. We assume that anything Spain

21 Because we give Italy the benefit of the doubt by achieving a high

rate of privatisation in the first year, we assume potential privatisationrevenues in the subsequent 1-2 years are negligible.

Exhibit 23: Not counting bills, Italy and Spain have gross funding needsover the next three years of just under €900bn…Yearly Italian and Spanish funding needs; includes bond redemptions and general

government budget deficits*; €bn2012 2013 2014 Total

Italy 233 171 134 539

Spain 113 111 103 327

Italy + Spain 347 282 237 866 * We assume T-bills continue to be rolled. Bond redemptions include both conventionaland non-conventional bond redemptions. Budget deficits based on J.P. Morgan basecase expectations.

Exhibit 24: …while the maximum funding available from varioussources is just over €700bn, or about €150bn short of their 3Y fundingneedsMaximum funding available to Italy and Spain from private and official sources; €bn

One-off wealthtaxes, etc.*

SWFs** IMF*** EFSF^

Italy 135 40 210 140 525

Spain -- 20 110 70 200

Italy + Spain 135 60 320 210 725

Private sources

Total

Official sources

* Includes one-off wealth taxes, privatisation, and sale of gold and FX reserves. Pleasesee Appendix-2 for details.** We estimate sovereign wealth funds would be willing to contribute no more than

€60bn to purchases of peripheral sovereign debt. Please see Appendix-3 for details.*** Please see “IMF 101”, Global Fixed Income Markets Weekly , 21 October 2011 for discussion of IMF lending capacity.^ Please see Overview, Global Fixed Income Markets Weekly , 14 October 2011 and28 October 2011 for discussion of EFSF spare capacity and leverage prospects. Weassume that €123bn of the EFSF’s current unallocated capacity is allotted to Greece

for its revised bailout package and that €61bn will be used for a second bailout of Ireland and Portugal in 2H12.




Pavan WadhwaAC

(44-20) 7777-3370


14

manages to raise from one-off measures will be

small and likely absorbed by further bank

recapitalisation costs.

22

2) We believe that sovereign wealth funds (SWFs) will

be willing to contribute no more than €60bn, or

roughly one-third of the amount currently allocated

to European fixed income in their portfolios. Please

see Appendix-3 for our calculations. Of this, we

split the contribution roughly 2:1 between Italy and

Spain.

3) We believe that the IMF can allocate no more than

€320bn to Italy and Spain combined. Even that

would require upsizing IMF capacity through

exercising New Arrangements to Borrow (NABs)and allocating a huge percentage of available

capacity to the two countries.23 It would also likely

be unpopular politically and might require quid pro

quos such as a reshuffling of voting rights. Of this,

we split the contribution roughly 2:1 between Italy

and Spain.

4) Finally, we do not think that the EFSF will be able to

meet its leverage targets, for reasons discussed

previously.24 After allocating funds for Greece’s

second bailout package and assuming another €60bn

is required to support Ireland and Portugal in

2H12,25 only €210bn of spare EFSF capacity would be left. This assumes France manages to keep its

AAA rating (France provides roughly one-third of

EFSF AAA guarantees and hence one-third of EFSF

issuance capacity). We also think it is unlikely that

the European Stability Mechanism (ESM), the

€500bn planned successor to the EFSF, will be

brought forward early and allowed to run

concurrently, as has been mooted in the media. This

would require greatly increasing the German bailout

contribution, which goes against Chancellor

Merkel’s promises to the German parliament and

would likely be challenged by the German

constitutional court.26

22 Also, note that Spain owns considerably less gold and FX reserves

than Italy does. 23 Please see “IMF 101”, Global Fixed Income Markets Weekly, 21

October 2011.24 See Overview, Global Fixed Income Markets Weekly, 14 October

2011 and 28 October 2011.25 See Overview, Global Fixed Income Markets Weekly, 8 July 2011. 26 See Overview, Global Fixed Income Markets Weekly, 9 September 2011.

Under these assumptions, total available funding of

€725bn is sufficient to fund Italy and Spain for a little

over two years. This may be enough to buy some time,

but is unlikely to keep yields in check. In the long run,

we believe that the ECB will eventually need to step in to

stabilize markets by monetizing EMU debt.

The ECB’s role – How much would the ECB need to

buy to stabilise markets?

Most observers, including ourselves, believe that the

crisis has progressed to such an extent that the ECB is

now the only institution remaining with enough

firepower to short-circuit the contagion and stabilise

markets. However, it is reluctant to do so as it would

mean socialising a large amount of debt, exacerbating

moral hazard, and potentially damaging its credibility by

blurring the line between fiscal and monetary policy.

Additionally, debt monetisation is not a long-termsolution: there is no guarantee that debt capital markets

will reopen for peripheral sovereigns once the ECB stops

buying.

In this section, we assess just how active the ECB would

need to be in order to deliver ‘market stability’, and

whether they would be able to intervene to the extent

required. We focus on the immediate pressures facing the

market, specifically in the next one year. As mentioned

in the previous section, Italy and Spain have combined

Exhibit 25: Selling pressure on Italian and Spanish bonds could be aslarge as €650-700bn…Estimated secondary market selling pressure in Italian and Spanish bonds that mature

in 2013 or after; €bn

Investor type

Fraction

held (%)

Par value

held (€bn)

Fraction

(%)

Par value

(€bn)

Domestic 55% 675 20% 135

Non-domestic 45% 552 66% 364

All -- 1227 -- 499

Domestic 62% 268 20% 54

Non-domestic 38% 165 66% 109

All -- 433 -- 162

Domestic 57% 943 20% 189

Non-domestic 43% 717 66% 473

All -- 1660 -- 662

Debt stock that

matures post-2012*

Amount of debt

sold

Italy + Spain

Spain

Italy

* We include all conventionals, international bonds, linkers, and zero-coupon andfloating rate notes that mature in 2013 or later. We assume that 1) 2/3rd of all non-domestic investors and 1/5th of all domestic investors wish to sell their holdings of Italian/Spanish paper and 2) T-bill markets remain open to Italy and Spain. Domesticand non-domestic ownership shares taken from Exhibit 8 above.




Pavan WadhwaAC

(44-20) 7777-3370


15

primary market funding needs of €350bn over the next

year, and we believe that demand for new Italian and

Spanish bonds will be limited. In addition, we estimate

between €650-700bn of secondary market selling

pressure over the next 12 months (Exhibit 25), based on

the following assumptions:

1) €1.6tn of Italian and Spanish bonds outstanding that

mature in 2013 or later,…

2) …total weighted-average non-domestic ownership

of 43%, and…

3) …an estimated sell-rate of 2/3rds for non-domestic

investors and 1/5th for domestic investors.27

Combined, this nets to over €1tn of funding demand/

selling pressure over the next year. Given the private

sector resources available to Italy and Spain of roughly

€135bn, we estimate that the supply/demand imbalance

would be nearly €900bn in Italian and Spanish paper

over the next year (Exhibit 26). This equates to around

€75bn per month of supply/demand imbalance.

Even if Italian and Spanish gross issuance needs were

funded from other sources, the supply/demand imbalance

would still be around €450-500bn of Italian and Spanish

27 The estimate of investors looking to sell is purely a J.P.Morgan

expectation, but we believe that our assumed sell rates are reasonablegiven anecdotal investor discussions and the empirical pattern of the

past few months: Italian yields have risen over 200bp since August,despite sizable ECB buying.

paper over the next year, or roughly €40bn per month

pre-SMP buying.

To put these numbers in context, we note that the ECB

has been purchasing about €35bn of peripheral paper per

month since August. While the ECB does not identify

individual countries, we assume the split has been

roughly 3:1 between Italy and Spain. Over this same time

period, net issuance has actually been negative for Italy

(Exhibit 27). Despite the negative net issuance and

Exhibit 26: …suggesting that the supply/demand imbalance in Italian/Spanish paper over the next 12 months would be around €75bn/month, or animbalance of around €40bn/month assuming that the SMP continues its current rate of purchase of €35bn/monthEstimated supply/demand imbalance* in Italian and Spanish bonds over the next one year, with and without IMF/EU/SWF commitments**; €bn

Without

IMF/EU/SWF

With

IMF/EU/SWF

Without

IMF/EU/SWF

With

IMF/EU/SWF

Without

IMF/EU/SWF

With

IMF/EU/SWF

Gross issuance / funding needs 233 233 113 113 347 347

Domestic inv estor selling 135 135 54 54 189 189

Non-domestic inv estor selling 364 364 109 109 473 473

One-off w ealth tax es, asset sales 135 135 -- -- 135 135

SWFs -- 40 -- 20 -- 60

IMF/EFSF -- 233** -- 113** -- 347**

€bn total 598 324 276 142 873 467

€bn/month 50 27 23 12 73 39Supply/demand imbalance*

Italy Spain Italy + Spain

Selling pressure on bonds*

Funds available**

* Estimates of funding needs and secondary market selling pressure taken from Exhibits 23 and 25 above. We assume T-bills continue to be rolled.

** Estimates of amounts raised from one-off wealth taxes and SWFs are taken from Exhibit 24 above. However, we assume that the IMF and EFSF contribute just enough tocover gross issuance and funding needs over the next 1Y, rather than providing three years funding in advance.

Exhibit 27: Despite negative net supply in Italy, and significant ECBbuying since August,…Gross bond issuance*, redemptions*, and SMP purchases in Italy and Spain** since 8

August 2011; €bn

52

21

61

14

92

31

0

10

20

30

4050

60

70

80

90

100

Italy Spain

Gross issuance

Redemptions

SMP purchases**

* Excluding T-bills. Bond redemptions include conventional and non-conventional bondredemptions.** The SMP has purchased a total of €123bn of peripheral paper in the roughly 3 1/2months since they initiated purchases on 8 August 2011. We assume that these SMPpurchases are split 3:1 between Italy and Spain.Source: Bloomberg, ECB, J.P.Morgan




Pavan WadhwaAC

(44-20) 7777-3370


16

sizable ECB buying, Italian yields have risen over 200bp

since August (Exhibit 28). This indicates the strength of

the selling pressure currently going through thesecondary markets.

Will the ECB be willing to double its current rate of

purchases to eliminate the supply/demand imbalance?

We make several points here. On the one hand, there are

no hard capacity constraints that would prevent it from

doing so.28 On the other hand, while the ECB has

increased its balance sheet less than other central banks,

it has arguably taken on a good deal more credit risk. The

lion’s share of the growth in the ECB’s balance sheet

stems from increased support for peripheral country

banks and peripheral sovereigns (please see Appendix-

4). Finally, based on recent official commentary, thereappears to be growing reluctance to expand the size of

SMP purchases. The German contingent remains firmly

opposed to this and insists that monetary financing to

sovereigns is against the ECB’s mandate. Further, recent

media reports suggest that this view is spreading. For

instance, there is reported to be a limit on purchases of

no more than €20bn per week. This is said to be a

reduction from the original limit, and a limit that the

ECB Council discussed lowering further, as late as last

week.29 To be sure, such limits are self-imposed and can

always be changed if market conditions change.

However, it appears that at least over the near term,

the ECB remains reluctant to step up the size of its

buying.

Going forward, our baseline view is that the ECB will

not be willing to purchase debt in large quantities,

unless the market forces its hand. This is due to 1) the

need to maintain oversight and aid conditionality in order

to mitigate moral hazard; 2) the desire to avoid blurring

monetary and fiscal policy; and 3) a reluctance to take

any steps that might damage credibility or independence.

The 2012 outlook

The long-term solution to the EU sovereign crisis isconsistent fiscal and structural reform, better governance

in the Euro zone, more fiscal integration, and potentially

jointly guaranteed debt issuance. While some small steps

28 The only hard restriction is that the ECB cannot purchase bonds

directly from a sovereign (i.e. on the primary market) or set up credit

facilities for sovereigns.29 See ‘Upper limit for bond purchases/Obergrenze fur Anleihekaufe’,

Frankfurter Allgemeine, 17 November 2011; and ‘ ECB hits back at

calls for intervention’, FT, 18 November 2011

have been taken, these efforts will take far too long to

bear fruit in order to deliver market stability in 2012.30

In the near term, as in all negative feedback loops or

crises of confidence, outside intervention is required

in the form of a lender of last resort. With the ECB

reluctant to play that role, the path of least resistance

is for the crisis to worsen. Thus, our baseline view is

that the crisis stays on a slow boil over the next fewmonths, possibly resulting in another country losing

market access. In the event that such a risk plays out, we

believe that the ECB will be compelled to start

monetizing debt since the alternative would be

significantly worse.

There are a large number of risks that could go wrong in

other countries as well (Exhibit 29). These include, for

instance, the prospect of

• another round of PSI in Greece,…

• …a hard default in Greece,…

• …potential PSI in Ireland and Portugal,…

• …another country needing to tap the EFSF,…

• …and a French downgrade.

30 See Overview, Global Fixed Income Markets Weekly, 19 August2011, for discussion of the process required to institute Eurobonds.

Exhibit 28: …peripheral yields have continued to rise after initiation of the SMP programPortuguese and Italian 10Y yields around the s tart of ECB bond purchases for eachcountry*; %

4.00

5.00

6.00

7.00

8.00

-3 -2 -1 0 1 2 3 4 5

Months around first day of SMP purchases*

Italy

Portugal

* The SMP started buying Portuguese and Italian bonds on 10 May 2010 and 8 August2011, respectively. Portugal was eventually bailed out by the IMF/EFSF on 3 May2011.




Pavan WadhwaAC

(44-20) 7777-3370


17

Given the large number of risks on the horizon, one or

more of these threats are likely to play out in 2012. The

combination of a steady stream of negative news and

market disappointments will keep liquidity poor, spreads

wide, and investors skittish.

Trading strategies

In our view, the crisis is likely to worsen in 1H12,

before the ECB feels compelled to step in and stabilize

markets by monetizing debt. We therefore expect

peripheral yields to continue to climb going into 1H12,

especially if another country loses market access. Note

that bond yields of a country that loses market access do

not immediately stabilise; as demonstrated in Exhibit 30,

official intervention has thus far provided only a

temporary respite. Thus, short positions in peripheral

debt have been profitable even when initiated post- bailout.

Additionally, in case of loss of market access, the debt of

a country is likely to be downgraded which would

exacerbate the selling pressure. Recent rating agency

commentary around Italy and Spain, and the pattern of

Greece, Ireland, and Portugal when they lost market

access, suggests that Italy would be downgraded if it lost

market access (Exhibit 31). Indeed, sovereign CDS

spreads are trading significantly wider than their

similarly-rated corporate brethren, suggesting that the

risk of further sovereign downgrades is high (Exhibit

32).

Of course, the ECB will be quick to realize the severity

of the situation if a large country loses market access,

suggesting that this will be the tipping point where it will

be compelled to start monetizing sovereign debt in large

Exhibit 29: The number of risks on the horizon is extremely large; we expect the peripheral crisis to escalate in 1H12Risks in 2012, and possible policy responses / positive surprises

Effect Risk

Greece fails to get further funding, leading to a hard default and potential ex it from the Euro zone

Italy and/or Spain lose access to capital markets

Italian coalition collapses, leading to early elections

More sev ere PSI in Greece, triggering contagion fears in Italy and Spain

French downgrade

Ireland / Portugal require PSI

Another country requires EFSF assistance (such as Cyprus)

Italy drags its feet on fiscal consolidation / structural reforms

Bank deleveraging depresses economic grow th

Euro area in deep recession prev ents planned fiscal c onsolidation

LCH continues to boost margin requirements, leading to further delev eraging in peripheral paper

ECB w illing to monetise debt in large quantities

Germany agrees to fund ESM alongside EFSF

Significant new money commitments from IMF/BRICs/SWFs

Upside surprise on economic grow th

Upside surprise on implementation of austerity measures

Potential policy

responses / positive

surprises

Risks

Severe

Significant

Long term

Exhibit 30: Bond yields in bailed-out European countries havecontinued to soar even after the bailout request

Average of Greek, Irish, and Portuguese 10Y benchmark yields around the date whenbailout was first requested*; %

months around bailout request*

6.00

7.00

8.00

9.00

10.00

11.00

-3 -2 -1 0 1 2 3

* Greece, Ireland, and Portugal formally requested a bailout on 2 May 2010, 21

November 2010 and 6 April 2011, respectively.




Pavan WadhwaAC

(44-20) 7777-3370


18

quantities. We expect that the eventual response from

the ECB, IMF and EU will be on a scale sufficient to

ultimately stabilise peripheral yields in 2H12,unlike

the limited attempts made thus far with Greece, Ireland,

and Portugal (Exhibit 33). We also see 10Y Bund yields

falling to 1.25% by 2Q12 in a flight-to-quality, before

rebounding to 1.75% by end-2012. Thus, we

recommend underweighting peripheral duration and

buying German duration going into 1H12.

Below, we also present three different types of trading

strategies that allow investors to take advantage of our bearish view on both the European economy and the

peripheral situation. These strategies, in turn, offer 1)

significant upside in a global recession, 2) positive

convexity to peripheral risk, and 3) diversification away

from the peripheral sovereign debt crisis.

1. Identifying the best long-duration trades in a low-

yield world

Valuations for fixed income are clearly stretched in the

major markets. For instance, US, UK, and European 10Y

real yields are currently below zero, and central bank

policy rates are near lower bounds. Thus, we look for global markets where long duration positions offer more

upside, without compromising too much on credit risk. In

Exhibit 34, we highlight sovereigns that are rated A+ or

above and, at the same time, offer positive real yields,

attractive policy rate downside, and strong economic

linkages to Europe (based on 10Y regression of annual

nominal GDP growth). As a simple metric of

attractiveness, we average each country’s cross-sectional

z-scores across the three measures listed above. This

captures countries that are most likely to be impacted by

the slowdown in growth in Europe and where

policymakers have the most leeway to respond by

slashing rates. Based on this metric, long Australian

and Chilean 10Y duration appear to be the most

attractive trades. We particularly like longs in

Australia: the RBA is expected to cut rates and domestic

technicals are strong (see Australia).

Exhibit 31: Rating agencies have stated that Italy will be downgradedfurther if it loses access to debt capital markets. We expect another three-notch downgrade (from A to BBB) in that event

Number of notches that bailed-out sovereigns were downgraded by Moody’s in the 3months prior and 6 months after the bailout request*

-3M to aid

request

Aid request

to +6M

Greece 1 4 14

Ireland 0 7 9

Portugal 3 4 9

Average 1 5 11

Italy 3 - 3

Downgrade (notches) Total downgrade

since Jan-10

(notches)

* Italy is currently rated A2 by Moody’s and A with a negative outlook by S&P.Source: Bloomberg, Moody’s, S&P

Exhibit 32: Indeed, sovereign CDS spreads are trading significantlywider than their corporate brethren, suggesting that the risk of further sovereign downgrades is high

CDS spreads of iTraxx components regressed against corporate Moody’s rating vs.CDS spreads of sovereigns regressed against Moody’s rating*; triangles showsovereigns and circles show corporates; bp

0

500

1000

1500

2000

Moody's rating

IE

PT

IT

ES

Aaa Aa2 A2 Baa2 Ba2 B2 Caa2

* Using a unit scale for Moody’s rating (AAA=1, Aa1 = 2, Aa = 3, etc.), we model thefollowing relationships based on current cross-sectional data:

5Y Sov. CDS = 76.6*(Moody's) + 38.7; R2 = 88%Sample based on 16 developed market sovereigns (US, UK, Japan, Australia,Sweden, Denmark, and 10 Euro area countries excluding Greece).

5Y Corp. CDS = 7.2*(Moody's)^2 - 73.8*(Moody's) + 358.9; R2 = 66%Sample based on nearly 200 corporates from the iTraxx High-grade, High-yield, andCross-over indices.

Source: Moody’s, Bloomberg, Markit

Exhibit 33: Spreads and yield forecast: We expect the crisis to getworse initially, before concerted policy action pushes peripheral yieldslower in 2H12J.P. Morgan 2012 forecast for 10Y yields and benchmark spreads to Germany

18-Nov 2Q12 4Q12 18-Nov 2Q12 4Q12

Germany -- -- -- 1.97 1.25 1.75

France 147 225 190 3.46 3.50 3.65

Greece 2425 2400 2400 26.32 25.25 25.75

Ireland 625 850 900 8.15 9.75 10.75

Italy 494 775 625 6.95 9.00 8.00

Portugal 910 1200 1300 11.04 13.25 14.75

Spain 469 700 600 6.70 8.25 7.75

Wtd av g., peripherals* 699 955 855 -- -- --

Spread to Germany (bp) Yield (%)

* Weighted by proportional contribution to the J.P. Morgan EMU Bond Index.




Pavan WadhwaAC

(44-20) 7777-3370


19

2. Identifying trades that are positively convex in a

risk-off (or risk-on) world

Given the bimodal distribution of risks, trades that offer

positive convexity to the crisis are attractive. We define a

positively convex ‘risk-off trade’ as one which performs

well if the crisis worsens but does not lose much if the

crisis stabilises. Similarly, we define a positively convex

‘risk-on trade’ as one which performs well if the crisis

abates but does not lose much if the crisis worsens.

To identify such trades, we look at a wide selection of global assets. We regress asset levels against 10Y

weighted peripheral spreads since 2010. Assets that have

a sufficiently strong and convex relationship are

presented in Exhibits 35–36. Specifically, these are

trades where the R-squared exceeds 65% and the t-

statistic on the squared peripheral spread term exceeds

+/- 4. Using the regression betas, we project gains and

losses to the recommended position from a +/-100bp

shock to peripheral spreads. The convexity in each

position is defined as the additional gain that one makes

in the targeted scenario due to the convex performance

profile, divided by the average absolute trade P/L for a

+/-100bp move.

For risk-off positioning, we find that Japanese 2Y swap

spread wideners, long Aussie duration, and short France

vs. Germany offer the greatest positive convexity

(Exhibit 35). However, long Aussie duration is our

preferred trade in that it is amongst the very few that

offer positive carry and it is expected to benefit from

policy rate cuts and strong technicals (see Australia).

For risk-on positioning, short Kiwi 2Y duration, UK 10Y

swap spread narrowers, and short gold offer the greatest

positive convexity (Exhibit 36). As before, there are few

assets that simultaneously offer positive carry and

positive convexity. One that does is long US municipal

credit.

Exhibit 34: Look to global markets for higher real yields, potential policy rate cuts, and high sensitivity of GDP growth to Europe. Long 10Yduration in Australia and Chile appear to be the most attractive tradesStatistics on global benchmark nominal and real yields, central bank policy rates, and beta of GDP growth to Europe; %

Continen t C ountr y 18-Nov 2H12 FC – 18-Nov L10Y, % YoY Beta to Europe Corr to Europe

Americas Chile AA 4.8 3.1 1.7 5.25 -1.25 10.0 1.3 66% 1.5

Americas Canada AAA 2.1 2.1 0.1 1.00 0 4.9 1.3 90% 0.0

Americas USA AA+ 2.0 2.5 -0.5 0.13 0 4.2 1.0 91% -0.7

Europe Czech AA 3.9 2.4 1.4 0.75 0 5.1 1.1 84% 0.5

Europe Sw eden AAA 1.7 1.8 -0.05 2.00 -0.75 4.1 1.1 91% -0.1

Europe Germany AAA 2.0 2.0 -0.04 1.25 -0.75 3.2 1.0 100% -0.3

Europe Denmark AAA 2.0 2.1 -0.2 1.20 -0.75** 3.0 1.0 86% -0.3

Europe UK AAA 2.3 2.5 -0.2 0.50 0 4.1 1.0 92% -0.5

Europe Sw itzerland AAA 0.9 0.8 0.1 0.13 0 2.9 0.8 81% -0.6

Asia/Pacific Australia AAA 4.1 2.9 1.2 4.50 -0.75 7.0 0.8 86% 0.6

Asia/Pacific China AA- 3.7 2.7 1.0 6.56 0 14.8 0.2 16% 0.3

Africa/ME Israel AA- 4.1 2.4 1.7 3.00 -0.50 5.7 0.3 26% 0.1

Asia/Pacific New Zealand AA+ 4.0 2.9 1.0 2.50 0 5.3 0.6 80% 0.1

Asia/Pacific Japan AA- 1.0 -0.2 1.2 0.05 0 -0.3 0.9 89% 0.0

Asia/Pacific Taiw an AA- 1.3 1.0 0.3 1.88 0 3.8 0.8 58% -0.2

Asia/Pacific S Korea A+ 3.8 3.3 0.5 3.25 0 7.1 0.4 43% -0.3

Av erage -- 2.7 2.1 0.6 2.12 -0.30 5.3 0.8 74% --

Average Z-

score*

Avg nominal GDP growthPolicy Rate (%)Rating

(S&P)

10Y Yield

(%)

Avg Inflation,

L10Y (% YoY)

Real yield

(%)

* Defined as the average of three cross-sectional z-scores: 1) 10Y real yield 2) level of central bank policy rate and 3) 10Y GDP beta to Europe. Z-score defined as (level – cross-sectional sample average)/cross-sectional sample standard deviation.** J.P.Morgan does not forecast central bank rates for Denmark but here we assume that they cut rates in line with the ECB, as is their typical pattern.Note: Yields shown for benchmark government bonds. Real yield defined as nominal yield – realised annual headline inflation over the past 10 years. Beta and correlation of GDPgrowth to Euro area GDP growth based on a 10Y regression of annual nominal GDP growth rate data.

Source: Bloomberg, J.P.Morgan




Pavan WadhwaAC

(44-20) 7777-3370


20

Exhibit 35: Trades that offer the greatest positive convexity to peripheral spread widening include 1) Japanese 2Y swap spread wideners; 2) longAussie duration; and 3) short France vs. Germany. Long Aussie duration offers positive carry as wellGlobal ‘risk-off’ trades (trades with a positively convex performance profile when peripheral spreads are widening)

Class Region Item

Trade

Positioning Curr R2 * Beta* Beta_sq*

Risk-

off**

Risk-

on**

Difference

(Risk-off - Risk-on)***

%

Convexity^^

Yields Australia/NZ Aussie 2Y Long duration 320 65% 26.1 -6.6 67 -54 13 22%

Yields Australia/NZ Aussie 10Y Long duration 405 71% -9.3 -2.1 39 -35 4 12%

Sw ap spreads UK UK 2Y SS Wideners 96 75% -3.5 1.2 14 -11 2 20%

Sw ap spreads Japan JPY 2Y SS Wideners 26 66% -8.8 0.9 4 -2 2 65%

EMU spreads Europe France - Germany Wideners 147 90% -7.6 3.1 35 -29 6 19%

Credit Europe Itrax x Wideners 189 76% 6.8 1.5 27 -24 3 11%

Credit Europe Maggie Banks Senior Wideners 216 75% 1.5 2.3 34 -29 5 15%

Credit UK UK Fins Senior Wideners 315 71% -6.0 3.9 49 -41 8 17%

Credit Asia Itrax x Asia Wideners 212 65% -11.0 3.8 42 -35 8 20%

Equities Europe EU banks Short 125 82% -4.8 -1.7 29 -26 3 13%

Equities Europe Italy MIB Short 15233 72% -589.2 -109.3 2120 -1901 219 11%

Regression of asset against

wtd per iph. spreads^ P/L for +/-100bp change in wtd per iph. spreads^

* Trades based on polynomial regression vs. 10Y weighted peripheral spreads since January 2010. Trades shown have R-squared >65% and statistically significant convexity(T-stat vs. squared peripheral spread term > +/- 4). Beta = the regression beta vs. 10Y weighted peripheral spreads measured in % and Beta_sq = the regression beta vs. 10Yweighted peripheral spreads^2, again measured in %. A positive (negative) value of beta_sq indicates a convex (concave) asset.** Modelled gain (loss) for a +100bp (-100bp) spread move, assuming peripheral spreads at 650bp. Gains and losses are calculated with respect to the recommended trade (e.g.reversing the betas where appropriate, or where beta_sq<0).*** Defined as the absolute value of gain in the risk-off scenario minus absolute value of gain in the risk-on scenario.^ 10Y weighted peripheral spread computed against Germany for Greece, Ireland, Portugal, Italy, and Spain (weighted by the size of their outstanding bond market).^^ % Convexity defined as the difference in the absolute value of gains divided by the average of the absolute values of gains in both the risk-off and risk-on scenarios.Units: yields, curve, swap spreads, credit spreads shown in bp of yield. Equities and commodities shown in index points. FX are in ratios.

Exhibit 36: Trades that offer the greatest positive convexity to peripheral spread narrowing include 1) short Kiwi 2Y duration 2) UK 10Y swap

spread narrowers; and 3) short gold. Long US muni credit is one of the few that offers positive carry as wellGlobal ‘risk-on’ trades (trades with a positively convex performance profile when peripheral spreads are narrowing)

Class Region Item

Trade

Positioning Curr R2 * Beta* Beta_sq*

Risk-

off**

Risk-

on**

Difference

(Risk-on - Risk-off)***

%

Convexity^^

Yields Australia/NZ Kiw i 2Y Short duration 250 78% -72.5 4.8 -6 15 10 90%

Curv e UK 2s/10s Steepeners 178 75% -30.0 1.3 -12 14 3 20%

Sw ap spreads UK UK 10Y SS Narrow ers 31 83% 12.4 -0.7 -3 4 1 41%

Sw ap spreads Scandis Denmark 10Y SS Narrow ers 65 80% 12.6 -0.7 -3 4 1 36%

Credit USA 10Y Muni Narrow ers 216 82% 7.9 -0.3 -3 4 1 18%

Commodities Global Gold Short 1720 89% 225.0 -13.1 -41 67 26 49%

FX FX CHF/HUF Short CHF 246 89% 20.6 -1.2 -3 6 2 53%

Regression of asset against

wtd periph. spreads^ P/L for +/-100bp change in wtd periph. spreads^

*, **, ^,^^: See footnotes in Exhibit 35 above. *** Defined as the absolute value of gain in the risk-on scenario minus absolute value of gain in the risk-off scenario.




Pavan WadhwaAC

(44-20) 7777-3370


21

3. Identifying trades that are orthogonal to the

peripheral debt crisis and therefore offer

diversification benefits

One of the difficulties of investing in the current market

environment is that return correlation is extreme. Indeed,

one could be forgiven for assuming that we live in a one-

factor world (with that factor being sovereign risk).

Trades that are relatively uncorrelated with peripheral

spreads can add value in a portfolio context, especially if they carry positively. We screen nearly 200 asset classes

globally and find that just a few, in fact around 5% of

those sampled, demonstrate a significant lack of

correlation. Specifically, Exhibit 37 highlights asset

classes for which the R-squared in a polynomial

regression vs. peripheral spreads is both 1) less than 20%

in 2010-2011, and 2) less than 30% in 2011. Of these,

we note that prime RMBS and Auto ABS in the US carry

positively. Moreover, J.P.Morgan strategists expect

spreads in these sectors to be stable to firmer in 2012.

Exhibit 37: The elusive quest for diversification from the peripheral crisis in Europe: US prime RMBS and Auto ABS carry positively and spreadsare expected to be stable to firmer in 2012Trades with low correlation* to EU peripheral spreads

USA:

Prime RMBS; AAA Auto ABS; CDS-cash basis;

UK:

Rate Vega

Japan:

10s/30s;

Canada:

5s/10s; EUR/CAD

Denmark:

2s/10s;

Brazil:

Real/CLP

Korea:

USD/KRW; EUR/KRW;

Notes:* Trades based on polynomial regression vs. 10Y weighted peripheral spreads since January 2010, as described in footnotes in Exhibit 35 above. Trades shown here have R2 <20% over entire sample period and R-squared < 30% since January 2011. Around 5% of the nearly 200 trades sampled meet these criteria.




Pavan WadhwaAC

(44-20) 7777-3370


22

Trading themes

• Be long German and UK duration; enter 2s/10s

curve flatteners

We expect the peripheral crisis to escalate in 1H12,

driving peripheral yields higher and German and UK

yields lower in a flight-to-quality dynamic. Further, a

Euro area recession, slow growth in the UK, and

disinflation will add downwards pressure to

intermediate yields, flattening the 2s/10s curve.

• Underweight intra-EMU paper vs. Germany

We expect intra-EMU spreads to widen as the

peripheral crisis escalates in 1H12. Also, ratings

downgrades and poor technicals will contribute to

upwards pressure on peripheral yields.

• Go long 10Y duration in Australia and Chile

Global markets that offer high real yields, potential

policy rate cuts, and high sensitivity of GDP growth to

Europe are attractive. Long Australian and Chilean

10Y duration appear the most attractive; we

particularly like longs in Australia given advantageous

technicals and expected central bank rate cuts.

• Position in trades that are positively convex to

peripheral spread widening

Trades that perform well when peripheral spreads arewidening but give up little when they are narrowing

are attractive ‘risk-off’ trades. We find that Japanese

2Y swap spread wideners, long Aussie duration, and

short France vs. Germany offer the greatest positive

convexity to peripheral spread widening . Long Aussie

duration is also attractive as it is one of the few trades

which offer positive carry as well. On the other hand,

short Kiwi 2Y duration, UK 10Y swap spread

narrowers, and short gold offer the greatest positive

convexity to peripheral spread narrowing .

• US Prime RMBS and Auto ABS offer

diversification away from the peripheral crisis

We screen nearly 200 asset classes globally and find

that just a few, around 5%, demonstrate a significant

lack of correlation to the sovereign debt crisis. Of

these, we note that US prime RMBS and Auto ABS

carry positively; J.P.Morgan strategists expect spreads

in these sectors to be stable to firmer in 2012.

Exhibit 38: Bird’s eye view of our major trade recommendations by currencyEuro area UK US Japan

Duration Long 10Y Long 10Y Long 10Y Long bias

Long 6Mx6M EONIA

Curve 2s/10s flatteners 2s/10s flattener Long end steepening bias Swap curve flattening bias

1s/5s bull flatteners 10s/30s flattening bias

Swap spreads 2Y wideners 10Y wideners 10Y wideners Short end wideners

FRA/OIS w ideners Long end narrow ers

TIBOR/LIBOR narrowers

3s/6s w ideners

Swap spread curve Flattening bias Neutral Flattening bias

Gamma Long gamma in 10Y tails Long gamma in 10Y tails Long Neutral

Short gamma in 2Y tails Short gamma in 2Y tails

Vega Neutral Neutral Short Neutral

Inflation 1s/10s HIC P sw ap curv e flattener Long intermediate real yields Short breakev ens N eutral

Long French CPI-linked vs. Euro HICP linked Short 10Y breakevens

Cross-market Wider intra-EMU spreads

Underweight Italy and France vs. Germany

Credit curve flatteners

Short Euro vs. US breakevens




Pavan WadhwaAC

(44-20) 7777-3370


23

Appendix-1: Timeline of key events in the EU sovereign debt crisis2009

04-Oct-09 George Papandreou leads Socialist Pasok Party to landslide victory in Greek elections

20-Oct-09 New Greek Finance Minister Papaconstantinou says deficit will rise to 12.5 percent of GDP this year, triple the prior forecast

2010

13-Jan-10 European Commission report accuses Greece of statistics fraud; new Finance Minister says: "There is no skeleton in the closet"

14-Jan-10 Greece adopts three-year plan to bring the European Union’s biggest budget deficit to w ithin the EU limit in 2012

02-Feb-10 Greek government announces austerity package to get deficit to 3 percent of GDP in 2012

11-Feb-10EU leaders hold first emergency summ it on Greece. EU agrees to take “determined and c oordinated action” to protect financial stability of euro area, w ithout

giv ing further details

08-Mar-10 Portuguese government announces new budget cuts, more asset sales and a freeze on public wages

16-Mar-10 Euro region finance ministers lay groundwork for making emergency loans av ailable to aid Greece

18-Mar-10Papandreou calls on EU partners to come up w ith specific aid measures w ithin a week to help Greece, hints he m ight seek support from IMF if EU partners

don’t act

24-Mar-10 Fitch cuts Portugal’s credit rating to AA-

25-Mar-10Trichet say s that the ECB w ill continue to accept bonds rated as low as BBB- as collateral. An EU summit agrees to general principles behind a Greek bailout,

including IMF inv olvement, but again provides no details30-Mar-10 Ireland says country ’s banks need to raise an additional €31.8bn of capital

12-Apr-10 Euro area finance ministers agree to provide up to €30bn of loans to Greece over the next y ear with the IMF agreeing to put up another €15bn in funds

21-Apr-10 The EU says Greece’s 2009 budget deficit was w orse than it previously forecast and could top 14 percent of GDP as “off- market swaps” c loud its estimates

23-Apr-10 Papandreau asks the EU for aid

27-Apr-10 S&P cuts Greece rating to junk (BB+), downgrades Portugal to A-

02-May-10 Euro area agree to upsize its rescue package to Greece to €110bn. Greece agrees to €30bn in austerity cuts over the next 3 years in ex change for the aid

03-May-10 The ECB says it will indefinitely accept Greek collateral regardless of the country’s credit rating

05-May-10Protests in Athens against the gov ernment’s austerity plans turn violent and 3 people are killed w hen they bec ome trapped in a bank set ablaze by

demonstrators

06-May-10 Flash crash in US stocks

07-May-10European leaders agree in principle to set up a €440bn emergency temporary fund to stem the sovereign crisis, w hich will com e to be know n as European

Financial Stability Facility (EFSF)

07-May-10 In Spain, the Government and the main opposition party reach agreement over mergers betw een cajas and the need of sav ings-bank regulation reform

10-May-10

EU finance chiefs agree to establish a €750bn permanent aid facility to succ eed the EFSF (the European Stabilisation Mechanism , or ESM). Also, the ECB 1)

extends unlimited € liquidity for banks, 2) re-opens $ sw ap lines w ith the Fed, and 3) announces a gov ernment bond purchase program to focus on peripheral

sov ereign debt (the SMP). This is the biggest attempt y et to solv e the peripheral crisis

12-May-10 Spain announces public-wage cuts and a pension freeze

18-May-10 Germany surprises the markets with a ban on naked short selling of debt securities, CDS and shares in 10 financial institutions.

21-May-10 German parliament approves its share of the EFSF commitments

26-May-10 Italian Prime Minister Silvio Berlusconi's ministers approve €24bn in budget cuts for 2011-2012

28-May-10 Fitch cuts Spain’s AAA rating one level to AA+

14-Jun-10 Moody’s cuts Greek rating to junk (Ba1)

15-Jun-10 Chinese sign multibillion euro contracts with Greece hours after Moody 's debt downgrade

16-Jun-10 The French government Wednesday unv eiled a controversial plan to gradually raise the age at which workers can retire with pension benefits

23-Jul-10 Europe publishes the results of the first round of bank stress tests. 7 banks failed with a total capital shortfall of €3.5bn

18-Oct-10

German Chancellor Angela Merke l and French President Nicolas Sarkozy agree that priv ate investors may be required to contribute to future EU bailouts v ia a

mandatory solv ency rev iew of countries requesting aid; they als o declare that official loans granted out of the ESM after 2013 should hav e preferred creditor

status over private debt investors

01-Nov-10 Portugal’s gov ernment and biggest opposition party agree to let nex t year’s budget pass, aiming to stem the euro region’s fourth-biggest fiscal shortfall

01-Nov-10Greek minister Pangalos sugges ts to Greek media that: "Demonizing debt restructuring is wrong. Debt exis ts to be restructured. We may pursue it ourselv es or

it may be proposed to us and it may be too adv antageous to turn it dow n."

10-Nov-10 LCH.Clearnet increases haircuts for Irish bonds

16-Nov-10 December aid tranche for Greece is delayed until January, after the Austrian and Finnish ministers say Greece hasn't met EU criteria

21-Nov-10 Ireland says it will apply for a bailout

28-Nov-10 Ireland gets €85bn bailout. European leaders scale back proposals to inflict losses on bondholders in future EU bailouts

23-Dec-10 Fitch cuts Portugal to A+ rating Source: Bloomberg, Reuters, WSJ, FT, NY Times, Telegraph




Pavan WadhwaAC

(44-20) 7777-3370


24

Timeline of key events in the EU sovereign debt crisis, continuedDate Event

2011

24-Jan-11 Spain announces new capital requirements for banks

11-Feb-11 Axel Weber resigns from Bundesbank after opposing the ECB’s crisis policy

25-Feb-11Fine Gael wins elec tions and knocks out the current party from the Irish government, establishing a coalition with Labour. Backtracks from demands for bank

debt burden-sharing but continues its demands for lower interest rates on official loans

10-Mar-11 Bank of Spain announces €15bn capital shortfalls in national bank stress tests

11-Mar-11 EU summit agrees to expand powers of EFSF by 1 ) allowing it to buy debt in primary m arkets and 2) upsizing guarantees to tap its full €440bn in firepow er

15-Mar-11 EcoFin meeting agrees on new Stability & Growth competitiveness pact

21-Mar-11EU finance ministers decide on mec hanisms for allowing the region’s permanent bailout mechanis m, the ESM, to lend €500bn in 2013, Greece's loan rates are

low ered by 100bp and loan maturity is ex tended to 7.5y rs; Ireland's loan terms are unchanged

23-Mar-11 Portuguese prime minister resigns over loss of support stemming from austerity measures

25-Mar-11 Agreement on capital structure and lending terms of the ESM at the EC heads of state meeting

27-Mar-11 Merkel's coalition loses in state elections in Baden-Wurttemberg, narrowing her support in the upper house of Parliament.

31-Mar-11 Irish bank stress tests announce €24bn shortfall across the remaining 4 un-nationalized banks; €5bn will come from sub debt writedowns

01-Apr-11 ECB agrees to drop its ratings requirement on Irish-backed paper at its tender operations06-Apr-11 Portugal requests aid from the EU

15-Apr-11 Papandreou announces €76bn of austerity measures, later increased to €78bn, running through the end of 2015

17-Apr-11 In the Finnish national election, True Finns, a Euro-skeptic party, polls third in a tight 4-way race, taking roughly 20% of seats in the legislature

03-May-11 Portugal reaches agreement on a €78bn aid package

16-May-11 Eurogroup meeting endorses Portuguese bailout package

16-May-11 Eurogroup meeting endorses Draghi as the next ECB chief

17-May-11 European finance ministers propose talks w ith bondholders around extending Greece’s debt-repay ment schedule

24-May-11 Greece releases a new medium-term plan with more aggressiv e fiscal consolidation and privatization targets

01-Jun-11IMF holds off releasing the nex t tranche of Greece's aid package until a new bailout package, w hich meets necc essary funding needs for at least 1 y ear, is

approved

02-Jun-11 The ECB and EU to come to truce around Greek PSI v ia a "Vienna initiative" approach

05-Jun-11Centre-right party w ins conv incingly in Portuguese parliamentary election; mos t Portuguese parties have already signaled their support for required austerity

measures

15-Jun-11 A parliamentary c risis in Greece sends markets into a tailspin. Fac ed w ith party defections, the Greek prime minister calls for a no-confidence vote to be held

on 21 June. EU em phasises that the disbursement of the nex t tranche of aid package requires approv al of all new austerity meas ures

16-Jun-11 The IMF agrees to release the next tranche of Greek aid w ithout formal agreement on a new package in place

17-Jun-11 Merkel and Sarkozy agree in principle to a "Vienna-initiative" debt rollover for Greece

21-Jun-11 Greek prime minister wins no-confidence vote

24-Jun-11 EU finance ministers agree on details around EFSF upsizing and remove preferred creditor status from ESM loans to Greece, Ireland and Portugal

29-Jun-11 Greek parliament approves austerity measures (principles)

29-Jun-11 French banks propose a Greek debt rollover w ith modest haircuts

30-Jun-11 Greek parliament approves austerity measures (implementation measures)

02-Jul-11 EU releases next tranche of Greek aid package

04-Jul-11 S&P announces that the French proposal for the Greek debt rollover would garner an SD rating

06-Jul-11Follow ing rating agency comments that Vienna initiative approach w ould most likely garner an SD rating, German and Dutch officials return to calls for a formal

debt restructuring in Greece. The ECB remains adamantly opposed

07-Jul-11 ECB agrees to drop the ratings requirement on Portuguese-backed paper at its tender operations

08-Jul-11 IMF releases next tranche of Greek aid package

08-Jul-11 Contagion spreads to Italy . Negative headlines around finance minister and banks under stress (Unicredit trading halted)

11-Jul-11Munitions blast at Cy prus nav al base leads to sev eral multi-notch rating downgrades and fears of a potential Cy priot bailout. Sev eral ministers resign and

protests call for the dissolution of the ruling gov ernment coalition

15-Jul-11 2nd round of EBA bank stress test results released; net €2.5bn capital shortfall estimated across 8 banks

21-Jul-11 Details of 2nd Greek aid package announced; IIF banking federation launches a voluntary Greek debt rollover

02-Aug-11 Italian 10Y yield closes above 6%, at highest levels since joining the EMU

05-Aug-11Italian Prime Minis ter Berlusconi announces new reforms and promises to front-load budget cuts in an apparent quid pro quo; the ECB begins buy ing Italian and

Spanish debt the week after

05-Aug-11 S&P downgrades the US to AA+ Source: Bloomberg, Reuters, WSJ, FT, NY Times, Telegraph




Pavan WadhwaAC

(44-20) 7777-3370


25

Timeline of key events in the EU sovereign debt crisis, continuedDate Event

08-Aug-11 ECB starts buy ing Italian and Spanish debt in its SMP purchase program

16-Aug-11Finland's deal to get collateral in exchange for loans granted to Greece in Greece's 2nd aid package becomes public and sparks controversy amongst other

Euro area countries

02-Sep-11IMF/EU/ECB officials end Greece's 5th quarterly rev iew early , unhappy w ith the pace of structural reforms. They delay decis ion on granting the September aid

tranche to Greece

6-8 Sep 2011 A wide sw athe of German politicians declare that Greece will not get its next aid tranche w ithout a positive IMF rev iew and sufficient progress on reforms

06-Sep-11 Swiss c entral bank announces it will target a 1.2 floor for Swiss/ Euro exchange rate

07-Sep-11German constitutional court rules that the EFSF structure is constitutional but requires m ore strict ov ersight form Parliament (mandatory Parliamentary Budget

committee approval before aid is distributed). The ruling has bearish implications for Eurobonds

11-Sep-11 Greece announces it will speed up reforms and collect a new property tax to help bridge the budget gap

14-Sep-11 Italian parliament gives final approval in a confidence vote to a €54bn austerity package to balance the budget by 2013

19-Sep-11 S&P downgrades Italy 1 notch to A

20-Sep-11 Slovenian government loses a confidence v ote w hich threatens to delay EFSF ratification

29-Sep-11 Bundestag votes to approve EFSF reforms

02-Oct-11 Greece’s gov ernment approves the draft budget for 2012

3-4 Oct 2011EU finance ministers w ork out a rev amped deal on collateral for Greek loans that satisfies Finnish demands and those of other euro-region gov ernments

opposed to a bilateral deal for Finland

07-Oct-11 Fitch cuts Spain rating to AA- and Italy rating to A+

11-Oct-11 Troika releases statement on the fifth review of Greek economy and suggests that the sixth tranche of the bailout pay ments, w orth €8bn, w ill be paid

13-Oct-11Slovak ia becomes the final country to approve the upsized EFSF facility. The ruling government lost a confidence vote tied to the initial EFSF v ote which w ill

lead to new parliamentary elections

21-Oct-11 Papandreou wins parliamentary approval of the latest austerity bill in Greece, w hich includes w age and pensions cuts and plans to lay off 30,000 state workers

23-Oct-11 European leaders say a summit on the euro crisis w on’t produce decisions and set another meeting for 26 October

26-27 Oct 2011EU leaders agree to 1) leverage the EFSF to boost its firepower to €1tn, 2) force private inv estors to accept a 50 percent haircut on Greek bonds, 3) push Euro

zone banks to raise €104bn in new capital, and 4) ex tend a new aid package w orth €130bn for Greece

31-Oct-11 Surprising his ow n party , EU officials, and markets, Papandreou calls for a public referendum on the second bailout agreement in Greece

02-Nov -11 European leaders suspend aid payments to Greece and say Greece must decide w hether it w ants to stay in the euro

03-Nov-11 Papandreou backs down on referendum

06-Nov -11 Papandreou agrees to step aside to make way for a government of national unity in Greece

08-Nov-11Berlusconi fails to secure a majority in a v ote on public finance in the low er house of the Italian parliament. He later announces that he will res ign once he has

secured passage of the upcoming budget bill. Italian 10Y yield c rosses the 7% mark

08-Nov -11 LCH.clearnet increases margin requirements on Italian paper by 3.5-5.0% points across the curv e

09-Nov -11 Greek political leaders agree on a new government, Papandreou steps down

10-Nov -11 Former ECB v ice-president Lucas Papademos is tapped to head Greece's new coalition government

12-Nov -11 The Italian parliament passes the budget bill and Berlusconi resigns

13-Nov -11 President Napolitano nominates former European Commissioner Mario Monti to become prime minister of Italy , heading a new technocratic gov ernment

16-Nov-11 The Papademos-led Greek government wins a v ote of confidence

16-Nov-11 Mario Monti is sw orn in as Italian Prime Minister

17-18 Nov 2011 Mario Monti wins c onfidence v ote in both houses of the Italian Parliament

Source: Bloomberg, Reuters, WSJ, FT, NY Times, Telegraph




Pavan WadhwaAC

(44-20) 7777-3370


26

Appendix-2: What is the maximum

amount of funds that Italy can raise via

one-off wealth taxes and asset sales?

There has been speculation in the media around the

extent of funding that Italy can raise via one-off wealth

taxes and asset sales. In this appendix, we attempt to

quantify the maximum amount that Italy can potentially

raise via three sources of funding:

1. A one-off tax on bank deposits, securities, and

property,

2. Privatisation of financial and non-financial assets,

and

3. Partial sale of gold + FX reserves held at the central bank

Exhibit A1 shows the wealth in Italy as of 2009, as

estimated by the Bank of Italy. Around half of the total

wealth of around €9.4tn is in housing, and another almost

one-third is held in cash deposits, Italian government

bonds, and other securities. We assume that, in extremis,

Italy will be able to impose a one-off tax of 100bp on

housing and Italian government bonds,31 and 60bp on

cash deposits/other securities and mutual funds.32 Such a

move would raise €60–70bn, or around 3 months of

Italian funding needs.

On the privatisation front, it is difficult to get accurate

data on holdings of financial and non-financial assets

held by the Italian government.33 Additionally, it may be

hard to make assumptions on the possibility of large-

scale privatisations given the difficult economic

environment. We therefore look at historical precedent

to estimate how much money can be raised via

privatisations, since Italy has been privatising assets over

the past several years. Exhibit A2 shows the amount of

financial and non-financial assets that Italy has privatised

on average over the past several years, as a percentage of

GDP.34 Also shown is the amount of financial assets

privatised in the top 3 years. Although data on peak non-

31 In the US, an annual tax on property is levied by the local (town)

authority to fund the local school system. This tax ranges from 0.2% to

4.0% (average 1%) of the value of the property. We therefore assume

that Italy can impose such a one-off tax. For more information, see:http://www.taxfoundation.org/publications/show/1913.html.32 There is historical precedent for a 60bp tax on cash deposits. In 1992,

Italy imposed such a tax under Prime Minister Giuliano Amato.33 For example, a recent presentation on the Italian Treasury’s websitehas data on Italian government holdings going no further than 2005. 34 Note that since we use two different data sources, the data periods for

privatisation of financial and non-financial assets do not match.

financial asset privatisation is not available, we estimate

it from the ratio of ‘Top 3 years’ to ‘average’ financial

assets privatised. We give Italy the benefit of the doubt

and assume that it can privatise assets at the maximum

rate achieved over the past few years. Since Italy has

historically achieved peak annual privatisation of 1.9%

of GDP for financial assets, and around 0.9% of GDP for

non-financial assets, we conclude that a similar high rate

Exhibit A1: Even assuming onerous one-off wealth taxes can beimposed on Italian households, the maximum that the government canraise is €60–70bn, or 3 months of Italian funding needs

J.P.Morgan expectation of maximum one-off taxes that may be raised by ItalyAssumed one-off tax

€bn % of total bp €bn

Housing 4830 51% 100 48

Land 241 3%

Non-residential buildings 335 4%

Valuable items 124 1%

Other 347 4%

Deposits, cash, post office 1062 11% 60 6

Italian gov ernment bonds 189 2% 100 2

Other securities and mutual funds 1576 17% 60 9

Insurance technical reserv es 631 7%

Commercial credits 107 1%

Total 9442 100% 66

Wealth

Source: Household wealth in Italy 2009, No. 67, Supplements to the Statistical Bulletin,Banca D'Italia, Eurosystem

Exhibit A2: Historically, Italy has been able to achieve peak annualprivatisation of 1.9% of GDP for financial assets, and an estimated 0.9%of GDP for non-financial assets. Achieving a similar high rate of privatisations in 2012 will likely raise €40–45bn of funding*, or 2 monthsof Italian funding needs

Average and maximum level of privatisation receipts achieved by Italy in past years

Privatisation receipts;

% of GDP

Financial assets (1987-09)

Av erage 0.7%

Top 3 y ears** 1.9%

Non-financial assets (2000-06)

Av erage 0.3%

Top 3 y ears*** 0.9% * Italian GDP is around €1.5tn, so 2.8% of GDP works out to a round €40–45bn.** Top 3 years for financial asset privatisation were 1997, 1999, and 2005 when

€19bn, €27bn, and €26bn were privatised, respectively. Source: Privatisationbarometer.*** Italy privatised non-financial assets worth 0.3% of GDP between 2000-2006.

Although data on peak non-financial asset privatisation is not available, we estimate itfrom the ratio of ‘Top 3 years’ to ‘average’ financial assets privatised. Source: Should Italy Sell Its Nonfinancial Assets to Reduce the Debt ?, Stefania Fabrizio, IMF PolicyDiscussion Paper, April 2008.




Pavan WadhwaAC

(44-20) 7777-3370


27

of privatisations in 2012 will likely raise €40–45bn of

funding,35 or around 2 months of Italian funding needs.

Additionally, the Bank of Italy holds gold + FX reserves

in excess of 8% of GDP, compared to a Euro area

average of just under 6% (Exhibit A3). Bringing them in

line with the average would generate funding of around

2% of GDP (€30bn), or around 1-1/2 months of funding

needs.36,37,38 Note that the Bank of Italy also owns around

€44bn of European sovereign debt (excluding BTPs), a

part of which may also be saleable.

Overall, we find that, in extremis, Italy may be able to

raise around €125–150bn over the next year via a

wealth tax and aggressive asset sales, allowing it to

avoid tapping capital markets for as long as 6–8months. In the long run, however, it is going to be

difficult for Italy to raise taxes significantly in order to

achieve a primary surplus as it is already one of the most

heavily taxed nations in the OECD world (Exhibit A4).

Instead, it will need to rely on increasing the tax base and

boosting growth.

35 Italian GDP is around €1.5tn, so 2.8% of GDP works out to around

€40–45bn.36 Technically, governments cannot use gold for funding needs since

central banks are legally independent and free from governmentinfluence. Indeed, according to an FT article entitled Italy to use gold

reserves to cut national debt , 1 August 2007, “Italy’s governmentcannot unilaterally infringe the central bank’s independence and

autonomy, although the bank could freely agree to contribute some of its reserves”. Given the severity of this crisis, we believe that the

central bank will not hesitate in making such a contribution, especially

since it would contribute to financial stability.37 The ECB will need to approve the sale of gold and FX reserves by

the Bank of Italy. Since Italy will be selling such assets merely to bringthem in line with the Euro area average, we believe that such approval

will be forthcoming.38 In 2009, central banks in the Eurosystem (together with Sweden and

Switzerland) agreed to not sell more than 400 tonnes of gold per year intotal, or around €17bn at current market prices. Given the severity of

the crisis, however, we believe that such restrictions may be relaxed inorder to help Italy.

Exhibit A3: Italy’s holdings of gold + FX reserves are significantlyhigher than the Euro area average; bringing these assets in line withthe average should yield around 2% of GDP (€30bn), or around 1-1/2

months of funding needs*Gold and FX reserves of various Euro zone countries as a % of GDP

Tonnes €bn €bn %GDP

Austria 280 12 5 301 17 5.7%

Belgium 228 10 6 370 16 4.2%

Finland 49 2 4 190 6 3.3%

France 2,435 103 22 1,988 125 6.3%

Germany 3,401 143 29 2,567 172 6.7%

Greece 112 5 0 218 5 2.2%

Ireland 6 0 0 156 1 0.4%

Italy 2,452 103 26 1,586 129 8.1%

Netherlands 613 26 7 607 33 5.5%

Portugal 383 16 1 172 17 9.7%Spain 282 12 12 1,075 23 2.2%

Total 10,239 431 112 9,231 543 5.9%

GDP (€bn)Gold + FX res.Gold** FX reserves

(€bn)

* Note that the Bank of Italy also owns around €44bn of European sovereign debt(excluding BTPs), a part of which may also be saleable.** Gold spot level used: €1310/oz.Source: World Gold Council, ECB

Exhibit A4: It may be difficult for Italy to raise taxes significantly on anongoing basis, given that it is already one of the highest taxedcountries in the OECD worldTotal tax revenue as % of GDP*; %

24

2829

31

34 3537 39

42 43 43 43 4445

20

25

30

35

40

45

50

U S

I r e l a n d

G r e e c e

S p a i n

U K

P o r t u g a l

G e r m a n y

N e t h e r l a n d s

F r a n c e

A u s t r i a

F i n l a n d

B e l g i u m

I t a l y

* * S c a n d i

* Data as of 2008 for Netherlands and Portugal, and as of 2009 for the rest of thecountries.** Scandi average includes Norway, Denmark, and Sweden.

Source: OECD




Pavan WadhwaAC

(44-20) 7777-3370


28

Appendix-3: What is the maximum

amount of funds that sovereign wealth

funds (SWFs) might be willing to invest

in EU peripherals?

EU officials are targeting emerging market countries

such as Brazil, China, and Russia, countries with large

amounts of FX reserves such as Japan, and countries

with significant natural resources such as Norway and

Qatar to help peripheral EU sovereigns. In particular, the

plan to leverage EFSF capacity increases the importance

of finding equity or debt investors from amongst this

pool. Despite stepped-up entreaties from EU officials,

we do not think these nations will be willing to

contribute a large amount of funds.

We base our view on several factors.

1) While sovereign wealth funds do control a large

amount of resources, nearly €2tn in total, the pool of

resources available to EU sovereigns is much smaller

than widely believed. This is because of several reasons.

First, SWFs are typically highly concentrated in equity.

Based on 1H12 data from Monitor, a consulting group,

we estimate that less than a quarter of SWF assets are

allocated to fixed-income. Second, SWF assets are

diversified across currencies; we estimate that a little

over one-third is allocated to European currencies. Thisimplies that in total, just around €170bn is currently

allocated to European fixed income (€170bn = €2tn *

24% * 36%, as shown in Exhibit A5). Third, this

amount includes non-Euro area currencies such as

sterling and Swiss franc, and non-sovereign exposures

such as corporate credit, structured finance, covered

bonds, etc. Although few SWFs break out portfolio

composition in much greater detail than this, it is likely

that the pool of resources available to euro area sovereign

debt is smaller still.

2) Recent commentary suggests that these countries are

reluctant to provide large amounts of aid. This is

especially true given that EU officials have indicated

they do not plan to make political concessions in order toobtain commitments. Exhibit A6 highlights recent

statements by high-ranking officials in various countries.

Exhibit A5: Despite having nearly €2tn of resources, very little of sovereign wealth fund resources are allocated to European fixedincome (€170bn). We assume that this allocation to European fixed

income could be increased at most by one-third, or €55–60bnSWF holdings and amount allocated by region and sector as of 2Q11; €bn

Total Europe Americas Asia/EM

Sector Sector w eight (%) 100% 36% 38% 26%

Total 100% € 2,012 € 720 € 771 € 520

--Fix ed Income 24% € 476 € 170 € 183 € 123

--Equity 67% € 1,340 € 479 € 514 € 347

--Other 8% € 165 € 59 € 63 € 43

--Cash 1% € 30 € 11 € 12 € 8

Region w eight (%) and allocation (€bn)

Note: Converted from USD using 1.35 EUR/USD FX rate.Source: Monitor Company Group, J.P. Morgan estimates

Exhibit A6: Recent commentary suggests that most non-Euro area countries are reluctant to provide aid, particularly countries that are lesswealthy than the EU peripheralsSelection of recent comments regarding SWF or FX reserve investments in peripheral sovereign debt

Country Speaker Date Comment Pos/Neg

RussiaSergey Ignatiev, Central bank

chairman18-Nov

Russia’s central bank isn’t planning to purchase debt issued by the EFSF until there’s more clarity

regarding the rescue fund.Negative

02-Nov It’s now “too soon” for his country to contribute. Negativ e

28-Oct"We, of course, mus t wait until its structure is ex tremely c lear. And moreover, this inv estment must be

decided on after serious, technical discussions."Negative

Oeystein Olsen, Central BankGovernor

28-Oct "It's not on the agenda to contribute any ex tra investments in the special purpose inv estment vehic le(SPIV) that have elements of help or aid. That's outside of the mandate."

Negative

Yngve Sly ngstad, Chief

Executive of the Norway oil fund28-Oct

"We do not have sufficient detail really to comm ent on what it would do for us in terms of investment

opportunity. .. Of course w e are committed to Europe although we have signalled that over time the fund

w ill have less of its investment in Europe."

Negative

Jens Stoltenberg, Prime Minister 26-Oct“N orway w ill not participate in any aid package...That is not our task. I believ e it’s w rong that Norw ay

should take part in such support measures.”Negative

Yoshihiko Noda, Prime Minister 02-NovWe welcome the [EU summit] package, and... "from the standpoint of supporting such efforts, w e w ill

continue to study the purchase of EFSF bonds."Positive

Senior Japanese finance minister 01-Nov "I told [Mr. Regling] that w e w ill continue to purchase EFSF bonds." Positiv e

Norway

Japan

ChinaZhu Guangyao, Vice Finance

Minister

Source: WSJ, Bloomberg, Reuters




Pavan WadhwaAC

(44-20) 7777-3370


29

For instance, Norway looks opposed to making any

investments on concessionary terms, and China and

Russia have also emphasised that any investment willdepend on the structure of the investment, how attractive

the investment is, and how extensive EU co-investment

is. It is especially unlikely that countries which are less

wealthy will see a compelling reason to support

relatively wealthy EU countries.

Thus, we believe that the €170bn allocated to

European fixed income will at most be increased by

one-third (or €55–60bn) as a contribution towards

peripheral sovereign debt purchases. Even this would

represent a large portfolio concentration of risk and

would likely require considerable arm-twisting and/or

political quid pro quos.

Appendix-4: The extent of ECB aid to

peripheral countries

Since the onset of the global financial crisis in mid-2007,

the ECB has grown its balance sheet by a substantial

amount, over €1.1tn. This is about half the rate of

increase seen in other central bank balance sheets

(Exhibit A7). However, the lion’s share of the increase

(nearly €700bn) comes from increased support to

peripherals, via

• repo operations to banks in peripheral countries,…

• …emergency liquidity assistance (ELA39) to banks

in peripheral countries,…

• …purchases of covered bonds (largely from Spanish

banks), and…

• …purchases of peripheral sovereign debt (Exhibit

A8).

The rest of the increase comes from items such as gold

reserves (up around €250bn), other securities and assets

excluding ELA (up €250bn), and claims on non-euroarea residents (up €100bn). Thus, while the ECB has

expanded its balance sheet less than its counterparts

have, it has arguably taken on a good deal more

credit risk in doing so.

Given this substantial increase in balance sheet risk, the

ECB is reluctant to further expand the size of its

39 See Overview, Global Fixed Income Markets Weekly, 1 April 2011for more description of ELA.

peripheral sovereign debt purchases. This is confirmed

by recent comments made by EU and ECB officials. For

example, while some German advisors have begun to

speak out in favour of a lender-of-last resort

commitment, key German leaders such as Merkel and

Schäuble, as well as the German Bundesbank, appear toremain firmly opposed (Exhibit A9).

Exhibit A7: The ECB has grown the size of its balance sheet by asmaller percentage than the Fed and BoE over the past few years...Total assets held on central bank balance sheets, current vs. 4Q06

3Q11 4Q06

bn %

ECB 2,289 1,134 1,154 102%

Fed 2,871 863 2,008 233%

BoE 243 81 162 199%

Change

bn of ntl ccy

Source: ECB, Federal Reserve, Bank of England

Exhibit A8: …but asset growth is heavily skewed towards riskier exposures such as peripheral bank repo and ELA* and SMP purchasesof peripheral sovereign debtECB balance sheet composition since 2007**; €bn

600

800

1000

1200

1400

1600

1800

2000

2200

2400

2007 2008 2009 2010 2011

repo to core banks

repo to periphery

banks/ELA*

SMP

Cov ered bond purchases

All other balance-

sheet items**

* ELA, or emergency liquidity assistance, is typically included in “Other assets” or “Other claims” on central banks’ balance sheets. We estimate the current amount of ELA outstanding based on Irish and Greek central bank data. Since the onset of thefinancial crisis in 2009, we estimate ELA reliance has risen to nearly €90bn in thesetwo countries. See Overview , Global Fixed Income Markets Weekly , 1 April 2011 for more description of ELA.** Includes items such as gold and FX reserves, claims on non-Euro area residents,and other assets and securities. We subtract the estimated ELA outstanding from thisfigure.Source: ECB, Central Bank of Ireland, Bank of Greece. Data as of Sep 2011.




Pavan WadhwaAC

(44-20) 7777-3370


30

Exhibit A9: Indeed, comments from senior officials in Germany and at the ECB suggest that they are not ready to monetise peripheral debt in largequantities just yetSelection of recent policymaker comments regarding ECB purchases of peripheral sovereign debt

Country Speaker Date Comment

18-Nov

“Losing credibility c an happen quickly -- and history show s that regaining it has huge economic and social costs." Keeping prices s table

“is the major contribution we can make in support of sustainable growth, employ ment creation and financial stability . And w e are making

this contribution in full independence.”

03-Nov"Our securities market programme has three characteristics: it is temporary; it is limited; it is justified in restoring the functioning of

monetary transmission channels."

03-NovThe [ECB] w ill "not be forced by anybody " to buy bonds. .. it w as pointless to think sov ereign bond yields c ould be brought down for a

protracted period by outside intervention.

18-Nov

“T he economic cos ts of any form of monetary financing of public debts and deficits outweigh its benefits so c learly that it will not help to

stabilize the current situation in any sustainable way . The lack of success i n containing the crisis does not justify overstretching the

mandate of the central bank and making it responsible for solv ing the crisis. "

12-Nov"We have a mandate and w e have to stick to our mandate. Fix ing an interest rate for a country is c ertainly not compatible with our

mandate... y ou could not argue that this w as not monetary financing."

12-Nov"This w ould violate Article 123 of the EU treaty . I cannot see how y ou can ensure the stability of a monetary union by v iolating its legal

provisions."

Jürgen Stark, ECB Ex ecutive Board

Member 11-Nov "The ECB will never be lender of last resort... The ECB can't compensate for deficits in crisis nations."

11-Nov“T he ECB is last resort lender for banks because that is its function, but not for governments because the union’s treaty explicitly forbids

it."

Angela Merkel, Prime Minister 16-Nov "The way w e see the (EU) treaties is that the ECB does not have the possibility of solving these problems here."

Wolfgang Schäuble, German Finance

minister 18-Nov

“If w e did that, the consequence w ould be that for some m onths we’d hav e a certain quiet. But in the longer term, the financial markets

would assume that the euro isn’t a stable currency, w hether financial markets w ould then think that it’s not that bad, I doubt it. That’s w hy

i don’t think it’s the right solution.”

Mario Draghi, ECB President

ECB/

Eurosystem

Jose Manuel Gonzalez-Paramo, ECB

Executive Board Member

Jens Weidman, head of the

Bundesbank

Germany

Source: Bloomberg, Reuters, Telegraph, Businessweek



Economic ResearchGlobal Fixed Income Markets 2012 OutlookNovember 24, 2011

David MackieAC

(44-20) 7325-5040

[email protected] Chase Bank N.A., London Branch

31

Economics

• Sovereign stress is expected to drive Euro area

into a mild recession,…

• …but there is significant risk that the downturn

could be deep

• The ECB main policy rate is expected to be cut to

0.50%

• We do not forecast ECB QE, but SMP purchases

will push the agenda for a fiscal union

• The rest of the world escapes recession, but

growth will be lackluster• Fiscal tightening will be a significant headwind in

the US

• The UK MPC will continue to expand its gilt

purchase programme, possibly extending that to

other assets

• EM growth forecasts have been cut in line with

the historical relationships between DM and EM

The outlook for the global economy in

2012

Since the global recovery began in mid-2009, there has

been a fair amount of tension between the forces of

cyclical lift (easy monetary policy, low levels of

spending on durables, low inventories and healthy non-

financial corporates) and the structural drags

(deleveraging by households, banks and governments,

elevated uncertainty about fiscal and regulatory issues,

and a decline in underlying growth potential).

As the recovery began to unfold, we expected the forces

of cyclical lift to dominate and ensure above-trend global

growth. Even so, the structural drags were always

expected to ensure that the return to full employmentwould be a long journey. Indeed, this was the experience

of the global economy in the first 18 months of the

recovery, from mid-2009 to the end of 2010, when GDP

growth averaged 3.8%ar. However, 2011 has not

followed this script at all (Exhibit 1). It now looks like

the pace of global growth in 2011 will be around two-

thirds of what we expected earlier in the year (a quarterly

average growth rate of 2.3%ar compared with our earlier

expectations of 3.9%ar). Further, our expectations for

the global economy in 2012 show a much more

moderate pace of growth than we expected earlier: an

average quarterly annualised rate of 2.1%ar. Essentially,

after the above-potential performance of the first 18

months of the recovery, the global economy is now

expected to experience below-potential growth through

2011 and 2012.

There are a number of reasons why the global economy

is now performing at a sub-par pace. There were clearly

two huge unexpected shocks in 1H11: the surge in

commodity prices, which dramatically reduced

purchasing power for commodity consumers, and theJapanese earthquake and tsunami, which not only dealt a

huge blow to the Japanese economy but also disrupted

the global manufacturing supply chain. Although the

impact of these shocks started to fade in the summer,

other drags entered the picture. Essentially, the structural

drags have become bigger and are now dominating over

the cyclical lift. The debt ceiling debate in the US has led

to more fiscal tightening next year and beyond.

Meanwhile, the Euro area sovereign crisis has pushed the

region back into recession, reflecting the direct effects of

fiscal consolidation and the indirect effects of a

tightening of financial conditions. The tipping point for

the Euro area was the spread of sovereign stress to Italy,which led to a sharp fall in confidence and asset prices,

and a significant tightening in bank funding conditions.

Even though monetary policy has eased further in both

the US and the Euro area, and emerging market growth is

holding up relatively well, the structural drags in the

US and the Euro area appear to be sufficient to

restrain global growth to a sub-par pace through the

end of 2012.

Exhibit 1: Global GDP growthRealised real GDP growth and J.P. Morgan forecast; %oya

-10

-5

0

5

10

2007 2008 2009 2010 2011 2012 2013

Global

Emerging markets

Developed markets

forecast




David MackieAC

(44-20) 7325-5040


32

Deleveraging shifts from households to sovereignsand banks

Over the past few years, there has been much focus onhousehold deleveraging. However, as far as the broader

economy is concerned, the drag from household

deleveraging is now very limited. The sharp move in

household financial positions ended some time ago, and

we are currently in a situation where households are

generating free cash flow and using that to repay debt

(Exhibit 2). The big deleveraging stories of this year

and the next concern sovereigns and banks. After

easing fiscal policy significantly during the crisis, fiscal

authorities in the US and Europe are turning towards

tightening. Having already begun this year, it is due to

get much more intense in 2012 (Exhibit 3). Meanwhile,

the sovereign crisis in the Euro area and the policyresponse encouraging recapitalisation looks to be

accelerating the deleveraging process in Euro area banks.

It is hard to know exactly where equilibrium

configurations of primary positions and debt levels stand,

but the sovereign deleveraging is likely to be a long and

hard journey, most probably longer and harder than the

journey for households. The key difference is that

sovereigns hope to take more time to complete this

journey, although across the periphery of the Euro area,

financial market impatience is accelerating the

adjustment with inevitable consequences for growth.

Fiscal tightening is a key reason why global growth is

expected to be sub-par in 2012, and this basic picture is

unlikely to change in 2013 and beyond. Sovereign

deleveraging is likely to dominate the macro landscape

for years to come.

Euro area slides into recession

Recent data confirm that the Euro area moved into a

recession in the autumn. When gauging whether an

economy is transitioning into a recession, we look at both

changes and levels of key variables. Over the past 6

months, the area-wide composite PMI has fallen by more

than during any other 6-month period except in theimmediate aftermath of the Lehman bankruptcy (Exhibit

4). Further, the level of the PMI is now well below where

it stood at the start of the 2008-09 recession (April 2008)

(Exhibit 5). The labour market data also are sending a

compelling message of recession. Over the past three

months, unemployment has risen by an average of

125,000, well ahead of the average in the three months

up to the start of the 2008/09 recession (Exhibit 6).

Exhibit 2: US household financial position* and stock of debtoutstanding% of GDP % of GDP

-4

-2

0

2

4

6

8

1980 1985 1990 1995 2000 2005 2010

40

50

60

70

80

90

100Financial position Debt stock

* US households financial position: ((disposable income - personal outlays)-( grossprivate investment -consumption of fixed capital))/GDP

Exhibit 3: US fiscal policy changes for 2012$bn

Current law Net change

Payroll tax -110 +65

Other taxes -25 +45

Infrastructure, S&L support -125 +15

Unemployment relief -50 +12

Other spending -40 -40

Total -350 +97

% of GDP -2.3 +0.6

Imapct on GDP (%pt) -1.8 0.5 Note: Obama plan incorporates changes that will not be fully implemented in 2012

Exhibit 4: Change in Euro area composite PMI6M change in Euro area composite PMI; points

-15

-10

-5

0

5

10

15

1998 2000 2002 2004 2006 2008 2010 2012




David MackieAC

(44-20) 7325-5040


33

At the moment, our forecast is of a mild recession in

the Euro area. The contraction is expected to last

through next autumn and to involve a peak-to-troughdecline in the level of GDP of just over 1%. In the

recessions of the mid-1970s, early 1980s, and early

1990s, the Euro area saw peak-to-trough declines in the

level of GDP of 2.5%, 0.5%, and 1.9%, respectively.

And in the 2008/09 recession, GDP fell by a staggering

5.5%.

Given the relatively unusual nature of the current

environment, and the difficulty in gauging the behaviour

of policymakers, we would emphasise that there is a lot

of uncertainty around our central projection for the Euro

area real economy. Given the stress in the financial

system, the recession could easily be deeper than ourcentral projection, especially if policymakers do not

manage the situation well.

What usually causes a recession?

Recessions occur when one or more negative shocks hit a

vulnerable economy. These shocks can take a variety of

forms, but they involve either a squeeze in current real

disposable income (interest rates, commodity prices,

fiscal policy) or a shift in the forces that influence current

spending relative to current real disposable income

(interest rates, asset prices, wealth, credit availability,

uncertainty). Meanwhile, an economy becomesvulnerable to such shocks for a number of reasons:

households or non-financial corporates have

overextended their spending on durable goods and

inventory relative to their income (through excessive

borrowing, partly driven by elevated asset prices); or

banks have over-levered their balance sheets relative to

their capital positions (again partly driven by elevated

asset prices); or corporate profits have been squeezed to

such an extent that drastic action is required to restore

profitability.

The Euro area recessions of the 1970s, 1980s, and 1990s

were driven by monetary policy shocks that hit a privatesector with overextended levels of spending and a

corporate sector where profitability had been depressed

by an overheated economy. The 2008/09 recession too

involved a monetary policy shock, but there was also a

huge asset-price shock; the overextension was in the

financial sector as well as in the non-financial sector. In

2008/09, the classic recessionary adjustment in the non-

financial private sector was amplified by a severe

adjustment in the financial sector.

Exhibit 5: Euro area composite PMIIndex

35

40

45

50

55

60

2007 2008 2009 2010 2011 2012

Lev el at the start of 2008/09

recession: 51.9

Exhibit 6: Euro area unemploymentMonthly change; 3M average; 000s

-200

-100

0

100

200

300

400

500

2007 2008 2009 2010 2011 2012

Lev el at the start

of 2008/09

recession: 29

Exhibit 7: Euro area bank lending standardsNet tightening of loans to non-financial corporations and households; %

-20

0

20

40

60

80

2003 2005 2007 2009 2011

Net tightening of loans to non-fin.corporates

Net tightening of loans to households

Source: ECB




David MackieAC

(44-20) 7325-5040


34

What does the current Euro area recession look like?

At first blush, the Euro area recession that we are

sliding into now looks more like the 2008/09experience. Instead of a monetary policy shock, the

region is experiencing a confidence shock – confidence

in sovereign solvency – which is being amplified by a

still-vulnerable financial system. As asset prices have

fallen, bank funding stress has risen dramatically.

Although banks have access to unlimited funding from

the central bank, funding stress does lead to higher retail

borrowing rates, tighter lending standards (Exhibit 7),

and a desire on the part of banks to de-lever their balance

sheets. This suggests that the Euro area will experience

something of a credit crunch in the coming quarters. As

we saw in 2008/09, tight credit availability can take a

significant toll on the real economy.

However, a key difference between now and early

2008 is that levels of spending on durable goods and

inventory are currently low, and financial positions in

both the household and corporate sectors are

currently good. There is still a deleveraging process

under way in some of the peripheral economies – as

households generate free cash flow to reduce the size of

their balance sheets – but conditions are not nearly as

stretched as they were at the start of the 2008/09

recession. Moreover, the significant fiscal consolidations

are concentrated in the peripheral economies, and it ishard for the Euro area to have a deep recession

without a significant fall in Germany and France.

Moreover, the global backdrop looks better now than it

did in 2008/09: what we are experiencing at the moment

is an asymmetric Euro area shock, rather than a common

global shock. The differences between now and early

2008 – and the uneven spread of fiscal consolidation –

are why we think the present recession will be relatively

mild.

The distinguishing feature of recessions tends to be depth

rather than duration. The 2008/09 recession in the Euro

area was almost three times deeper than the early 1990s’recession, but it was only one-quarter longer in duration.

This suggests that policy plays a key role in arresting the

negative dynamic in recessions. However, in contrast to

the experience of previous recessions, when

policymakers were adding stimulus to the economy, the

policy action that we anticipate in the coming months –

asset purchases by the EFSF and the ECB, loans to

sovereigns and bank recapitalisation – is more about

crisis containment. This leads to the question of how

effective policy will be in containing the negative

dynamic in the real economy and creating a base for the

ultimate recovery.

All of this indicates that it is difficult to have a lot of

conviction about what the present Euro area recession

will look like. The conditions are unusual and it is

unclear how confidence, asset prices, and bank funding

stress will interact, and how that will impact the real

economy. There is also uncertainty about policy. This is

not simply about whether policymakers will respond in a

timely and effective manner. It is also about the limited

scope for the addition of outright stimulus. Our

inclination is to think that downside risks prevail

around our central projection.

Recession makes fiscal and structural adjustmentsharder to achieve

As the Euro area slides into a recession that looks likely

to be particularly deep in parts of the periphery,

budgetary positions will slip due to the automatic

stabilisers. Given the need for peripheral sovereigns to

keep moving towards more sustainable fiscal positions,

there will be enormous pressure to offset some of the

budgetary slippage with additional tightening measures,

even though these will weigh additionally on demand.

Our forecast does incorporate some additional fiscal

Exhibit 8: Government budget balance projectionsOfficial* and J.P. Morgan budget balance projections**; % of GDP

2011 2012 2013 2014

Greece Official -7.3 -5.6 -4.4 -2.2JPM, no extra aust. -9.1 -10.3 -9.0 -8.2

JPM, more austerity -9.1 -8.3 -6.9 -6.7

Ireland Official -10.0 -8.6 -7.2 -4.7

JPM, no extra aust. -9.6 -9.3 -9.3 -7.9


Portugal Official -5.9 -4.5 -3.0 -1.8

JPM, no extra aust. -6.8 -5.9 -5.4 -5.3


Spain Official -6.0 -4.4 -3.0 -2.1

JPM, no extra aust. -7.1 -6.6 -6.0 -5.9


Italy Official -3.9 -1.6 -0.1 0.2

JPM, no extra aust. -4.3 -2.8 -2.1 -2.3


Belgium Official -3.6 -2.8 -1.8 -0.8JPM, no extra aust. -3.8 -4.1 -3.6 -3.1

JPM, more austerity -3.8 -3.5 -2.6 -1.9 * Official projections are from the Stability and Growth Programmes for Greece,Ireland, Spain and Belgium (the Greek one was published in August rather than in thespring). Portuguese projections are from their fiscal strategy and Budget 2012documents, published in October. Italian projections are from their updated strategydocument published in September.** “JPM, no extra austerity” assumes that governments do not correct the cyclical fiscalslippage due to growth undershooting the official forecasts“JPM, more austerity” assumes additional fiscal tightening worth 50% of the cyclicalslippage




David MackieAC

(44-20) 7325-5040


35

tightening next year, beyond the plans already

announced. Essentially, we assume that governments

seek to offset half of the cyclical slippage (Exhibit 8).

In addition to putting significant pressure on individual

sovereigns, the recession dynamic is also affecting the

rescue mechanism put in place by the region. As

France’s AAA rating looks less secure, due to both the

discussion about leveraging the EFSF and the evidence

that France itself may be falling into a meaningful

recession, the EFSF’s own situation looks less robust.

The response of the ECB to the macro landscape

In early November, the ECB cut the main policy rate

from 1.50% to 1.25%. Although this move was not

widely expected, it is not hard to understand given the

deterioration in the macro landscape over the past few

months. Indeed, the new ECB President, Mario Draghi,

spoke of the region heading toward a mild recession,

which would exert downward pressure on wage and price

inflation. We expect another rate cut to follow in

December, and for the main policy rate to ultimately

reach 0.50%. If the recession turns out to be a deep one,

there will be talk about whether the ECB should engage

in quantitative easing, which we would interpret as a

targeted amount of bond purchases spread across the

region in order to ease the monetary stance further by

lowering longer-term interest rates. We would contrastthis with the Securities Markets Programme (SMP),

which is intended to improve the transmission of the

prevailing monetary stance rather than add more

stimulus. At the moment, we do not anticipate the

ECB engaging in QE, although the central bank’s

balance sheet will still expand further due to SMP

purchases.

The response of the ECB to sovereign stress

As the stress in the sovereign bond markets has spread to

core countries such as France and Austria, it has become

clear that something more dramatic from policymakers is

needed to stabilise markets and ensure financial stabilityacross the region. Not surprisingly, all eyes are on the

ECB as the only institution able to step in quickly and

decisively.

The obvious next step would be for the ECB to announce

its intention to use the SMP aggressively to ensure

financial stability in the region. In our view, there is no

legal or technical limit to the amount of explicit

support that the ECB can give secondary markets

and the amount of implicit support that the central

bank can give to primary markets (by purchasing

around the primary auctions to help dealers get short and

shed unwanted positions, but not at the auction itself).

This more-aggressive stance could be adopted with

explicit yield targets, although we doubt that the ECB

will go that far.

Whether this stabilises the situation remains unclear.What lies at the heart of the crisis is a fear that the

process of sovereigns returning to solvency via austerity

alone, i.e. without local currency depreciation and a

locally orientated monetary stance, is dynamically

unstable. Adding to the prospect of a dynamically

unstable process is the widely held perception that

peripheral economies are profoundly uncompetitive and

that some of the structural reforms needed to improve

competitiveness will depress growth and damage social

cohesion before they deliver any macroeconomic

Exhibit 9: Italian debt-to-GDP under different scenarios% of GDP*

90

100

110

120

130

140

2000 2005 2010 2015 2020

J.P. Morgan

baseline

Official plan

Mild shock case

* See Exhibit 10 for assumptions under different scenarios

Exhibit 10: Key variables for Italian debt dynamicsEvolution of macro variables under alternative scenarios*; % of GDP, unless otherwisespecified

Official

plan

J.P. Morgan

baseline

Mild shock

case

Marginal borr. rate (from now), % 5.8 6.5 6.5

Real long run growth pot., % 1.2 0.8 0.6

Maximum sustained prim. balance 5.7 5.0 3.5

Average borrowing rate, % 5.0 5.4 5.4

Nominal GDP growth, % 3.0 2.0 1.7

Government debt 104.8 120.8 130.2Primary balance 5.4 4.8 3.3

Yearly change in debt -3.4 -0.7 1.4

Key assumptions

Key variable averages over the period 2012-20

* Official projections are partly the result of our assumptions, as they are not availablebeyond 2014.




David MackieAC

(44-20) 7325-5040


36

benefits. This fear of dynamic instability has increased

with the growing evidence that the region is sliding into

recession, possibly a deep one. Greece provides a clear example of how badly things can turn amid political and

social instability and sovereign debt restructuring.

At some point, sizable ECB interventions would stabilize

the system. But, the situation for Italy remains a key

uncertainty (Exhibit 9 and 10). In our view, the SMP is

not an appropriate mechanism for funding Italy’s

gross financing need over the medium term. Thus, the

region does need to create some kind of l iquidity

hospital to accommodate Italy. The EFSF as currently

constructed cannot do this. Thus, it is not surprising that

France has put the idea of turning the EFSF into a bank

back onto the agenda and that there are ongoingdiscussions about how to involve the IMF more

substantially.

Alongside increased ECB intervention and the creation

of a liquidity hospital for Italy, the Germans will

continue their campaign for further fiscal integration.

As this crisis has continued to unfold, it has become

increasingly evident that the current blueprint for the

future – more peer surveillance and pressure but no de

jure loss of fiscal sovereignty, and with a permanent

liquidity hospital based on the idea of shared fiscal

capacity – is unlikely to work, unless sovereign debt

levels are much lower. The risk of sovereign liquiditycrises is just too great. But, in terms of dealing with this

fundamental issue in the functioning of the Euro area,

policymakers are moving incrementally. In the near term,

the Germans want limited treaty change to introduce

automatic sanctions in the workings of the stability and

growth pact, the possibility of referring sovereigns to the

European Court of Justice, and the possibility of greater

interventions in national budgets. It seems that they

continue to want an orderly sovereign default mechanism

as part of the permanent crisis resolution framework. It is

not clear, however, that this will be sufficient.

Impact of the Euro area recession on the rest of theworld

In the middle of 2011, there was a serious growth scare

in the US, but the economy has settled at a close-to-

potential growth rate in the second half of the year. Even

so, two risks loom over the 2012 outlook for the US.

The first risk relates to spill-overs from the economic and

financial crisis in Europe. Around 20% of US exports are

destined for Europe, and the economic contraction will

surely impact demand from that market. If that were the

only transmission channel, the risks would appear to be

contained. However, the potential financial spill-overs

could be larger. Thus far, most measures show only a

modest impact of the European crisis on domestic credit

availability. Nonetheless, if conditions in Europe worsen,

the US economy would not be immune from a seizing up

of global financial markets.

The second risk is a tightening of domestic fiscal policy.

The recession produced an unprecedented fiscal policy

response as the Federal deficit increased to 10% of GDP.

The political tolerance for such large deficits has

lessened recently, and several temporary fiscal stimulus

measures are set to fade in 2012, which could subtract

1%-2%-points from GDP growth next year. There has

been some talk about extending the temporary measures,

but nothing has been agreed upon so far. The

anticipated drag from fiscal policy is the main reason

behind our view that growth will be below potential in

2012, when we anticipate US growth will average

1.75%.

This expected tightening of fiscal policy is likely to hit

the economy where it is weakest: the household sector.

Slack conditions in labour markets have eroded labour’s

bargaining position, resulting in very tepid wage gains

(Exhibit 11). This outcome has been positive for

corporate profits – which have received a significant lift

from restrained unit labour costs – but has held back

consumers’ purchasing power.

Exhibit 11: US labour market3M moving average of monthly change in non-farm payrolls and unemployment ratefor 25 years and older

% %

-1.0

-0.5

0.0

0.5

2006 2007 2008 2009 2010 2011 2012

3

4

5

6

7

8

9

Montlhly

payrolls

Unemployment rate




David MackieAC

(44-20) 7325-5040


37

The low rate of inflation has given the Federal Reserve

leeway to be creative in supporting the recovery. The two

experiments this year were the explicit mid-2013 fundsrate guidance, and Operation Twist. We think the most

likely next step for the Fed is to replace the mid-2013

rate guidance with language that is more explicitly

dependent on a limited number of observable economic

variables, particularly the unemployment rate. While we

are not forecasting more asset purchases in 2012, the risk

of this is high and rising (subjectively, we would put it at

around 40%). Moreover, Fed officials have shown an

increasing openness to considering returning to MBS

purchases. A significant setback in the Euro area, or

an unwelcome fall in inflation expectations, could be

all that is required for QE3 in the US.

The UK has conspicuously chosen to front-load its fiscal

adjustment. While that has shielded it from direct attack

on the gilt market, public sector job losses, tax increases,

and spending restraint have weighed on growth. Through

noisy data, growth has run at a near-1% pace since mid-

2010. As we look forward, fiscal consolidation is

ongoing, while confidence and bank funding conditions

are caught in the back wash of Euro area turmoil.

Leaning against this is the fact that the direct drags on

household purchasing power from rising VAT, energy,

and food bills are dissipating. And as inflation begins to

ease, the MPC will continue to expand its gilt

purchase programme, plausibly extending that to

other assets (such as bank term debt) as 2012

progresses. All told, the forecast shows the UK seeing

near stagnation in output through what will remain

volatile GDP data, with unemployment continuing to

rise.

Given expected sluggish developed market growth,

the emerging market outlook has also markedly

weakened. The drag from DM to EM operates through

two channels. The first is through trade linkages. A

slowing in EM growth from weaker exports to DM is

unavoidable. We estimate that EM export growth will

slow to just 4% in 1H12, which is weaker than at any

time during the 2000s’ expansion.

The second channel of transmission is through financial

markets. As risk appetite wanes, capital flows to the EM

slow or even reverse, equity markets sell off, credit

conditions tighten, currencies fall, and the prices of

commodities – a source of revenue for many EM

countries – decline. In turn, domestic demand in the EM

is hit. This time around, the financial transmission

channel could be magnified by the rapid rise in credit

since the start of the recovery in 2009. While domestic

funding sources have increased, deleveraging of

European banks could still pose a risk as their foreignclaims in EM are sizable – particularly in EMEA and

Latin America.

The interaction of these complex forces has generated

a unit-beta response in the EM to shifts in US/Euro

area growth over the past decade. In other words, a

1%-pt decline in US/Euro area growth has translated, on

average, into a 1%-pt decline in EM growth. EM Asia

has a smaller beta of 0.6 (owing to very low betas in

China and India), Latin America has a beta of 1.2, and

EMEA has the highest beta of 1.4.

Currently, the channels of transmission are operating asusual. Sluggish DM import demand has produced a sharp

slowdown in EM export growth. Risky assets have sold

off across the world and commodity prices have fallen.

In turn, economic activity has decelerated in most EM

countries. J.P. Morgan’s recently lowered GDP growth

forecasts for the next five quarters align well with the

implied beta relationships for the three EM regions, with

the average projected at a sub-trend pace of 5% in 2012.

Should the US/Euro area slide into a deeper downturn

than currently forecast, the greater flexibility of policy in

the EM could lead to a relative outperformance in the

region. This outcome, however, would require a changein EM policymakers’ reaction function. On the downside,

increased leverage in the region could exacerbate

financial stresses, leading to a worse-than-unit beta

outcome.




Gianluca SalfordAC

(44-20) 7325-4334

Aditya Chordia (44-20) 7777-9841J.P. Morgan Securities Ltd

38

Euro Cash

• 2011 saw a further escalation of the Euro area

sovereign crisis as stress reached Italy and core

countries

• In our central scenario the crisis will worsen in

the first part of 2012 on the back of Euro-wide

recession and further market pressure on

sovereigns…

• … forcing policymakers to react aggressively to

stabilise the situation in the second part of the

year

• The ECB will be forced to cut the refi rate to0.50% and expand its balance sheet even further

• Although our fair value model suggests that 10Y

Bund yields could drop below 1.00%, as the

PMIs/core inflation drop and peripheral spreads

widen, we forecast a floor in 10Y Bunds at 1.25%

• We recommend long 10Y duration exposure; but

this view is predicated on policy makers being

able to ultimately contain the crisis; if they fail

Germany’s role as a safe haven will be challenged

• The short end of the curve has little scope to rally

under our assumption that the depo rate will not

fall below 0.25%: we therefore recommend 2s/10s

flatteners targeting around 100bp by 1H12

• With regard to intra-EMU spreads, we forecast

generalized widening vs. Germany

• We also recommend credit curve flatteners: in

line with the experience of Greece, Ireland and

Portugal, investors will switch from focussing on

yields to prices in case of severe sovereign stress

• The empirical price relationship between short

and long dated bonds is convex: we recommend

flatteners weighted by the empirical beta to get

protection from tighter spreads and steepercurves

• We present a framework to analyse the yield

spread of high- and low-coupon bonds of similar

maturity under sovereign stress: Bonos 2032 are

rich vs. Bonos 2037

• We analyse the investor base of Euro area,

German, French, Italian and Spanish bonds. The

reliance on foreign banks and foreign central

banks is high for all of them, except Spain

• On the supply side, higher redemptions will be

generally offset by lower deficits. We expect

€710bn of conventional bond issuance in 2012,

slightly below 2011 levels

• Italy faces heavy bond redemptions in 2012,

especially in the February-April period, whereas

Spain’s redemption calendar is more manageable

2011 recap: Shattered hopes

“It is going to get worse before it gets better” was

investors’ mantra into 2011, a view that we shared. The

beginning of the year caught many investors on the

wrong foot as policymakers seemed to calm fears with

little effort and macro data continually surprised to

the upside, leading to stronger equities and higher

yields, with 10Y bund yields peaking at 3.50% in mid-

April (Exhibit 1). In addition, higher inflation prints

prompted the ECB to surprise market participants and

increase interest rates twice after being on hold for

almost two years. Declining excess cash in the systemadded to the perception of normalisation. It was not

meant to last.

The sovereign crisis raised its ugly head again in mid-

April. Portugal’s request for aid was met with a yawn,

but once again Greece was the catalyst: amid clear signs

of fiscal slippage, the request for private sector

involvement in exchange for further official support

broke a taboo and hit investors’ confidence hard. In less

than 6 months, Bund yields halved on the back of

Exhibit 1: 2011 recap: markets and data peaked in the spring and havebeen on a downward trend ever since

10Y Bund yield and Euro area composite PMI% index

1.50

1.75

2.00

2.25

2.50

2.75

3.00

3.25

3.50

3.75

Jan-11 M ar-11 M ay -11 Jul-11 Sep-11 Nov -11

46

48

50

52

54

56

58

60PMI

10Y Bund




Gianluca SalfordAC

(44-20) 7325-4334


39

pressure from peripherals, a collapse in the PMIs and the

August downgrade of the US by S&P. Since October,

yields have been range-bound and hostage to headline

news related to the sovereign crisis.

We are ending the year with 2s/5s significantly flatter YTD, 5s/10s almost unchanged and 10s/30s steeper. The

2s/5s and 5s/10s curves exhibited typical directionality in

the first part of the year, flattening in the bearish move,

but they lost directionality in the second part. On the

contrary, the 10s/30s Bund curve showed poor

directionality in the first part of the year refusing to

flatten further from already stretched levels, but bull-

steepened with decent directionality in the second half of

the year (Exhibit 2).

Sovereign risk escalated to yet new highs in 2011. In

the first part of the year, the problem remained isolated to

the smaller peripheral countries (Greece, Ireland, andPortugal). However, policy makers’ decision to include

private sector involvement (PSI) in the new package for

Greece spooked investors, spreading contagion to the

larger peripheral countries (Italy and Spain) and beyond.

It is interesting to highlight that in relative terms, France

or supranational issuers such as the EFSF

underperformed more than Italy and Spain did in the

latter part of the year (Exhibit 3), making the distinction

between core and peripheral debt markets somewhat

obsolete.

Although all intra-EMU spreads to Germany reached

their all-time highs since the Euro inception, France and

Italy were the worst performers in relative terms,

whereas Ireland was the best performer (Exhibit 4).

Rating agencies were very busy with downgrades

(Exhibit 5), adding to the pressure.

Once again, policy makers were forced to take

unprecedented actions: 1) ECB purchases: In August,

the ECB was forced to intervene by buying Italian and

Spanish paper through the SMP after a 4 month hiatus

(Exhibit 6). We estimate that the ECB has purchased

Exhibit 2: The curve directionality to the level of the market shifted fromthe short and medium end to the ultra-long end in the second part of the year 2s/5s, 5s/10s, and 10s/30s regressed against 2Y German government par rates; 1 Jan

2011 to 11 Apr 2011 and 12 Apr 2011 to 18 Nov 2011;2s/5s 5s/10s 10s/30s

Jan-April -0.27 -0.30 -0.01

Apr-Nov 0.09 -0.14 -0.19

Jan-April 75% 93% 1%

Apr-Nov 42% 65% 68%

Beta

R-squared

* Peak in 10Y Bund yields.

Exhibit 3: Sovereign risk escalated to yet new highs in 2011YTD % change in 10Y spreads to Germany for France, Greece, and Italy; %

-50%

0%

50%

100%

150%

200%

250%

300%

350%

Jan-11 M ar-11 M ay -11 Jul-11 Sep-11 Nov -11

France Greece Italy

Exhibit 4: Intra-EMU spreads widened aggressively after mid-April 201110Y benchmark spreads to Germany; bp

18-Nov-11 11-Apr-11* 30-Dec-1030-Dec to

Current

11-Apr to

Current

30-Dec to

11-Apr

Austria 147 31 46 216% 372% -33% 189 31

Belgium 282 72 104 170% 291% -31% 309 72

Finland 67 25 28 138% 168% -11% 75 24

France 147 30 42 249% 394% -29% 188 30

Netherlands 56 22 24 131% 152% -8% 65 19

Greece 2425 933 948 156% 160% -2% 2471 747

Ireland 625 561 611 2% 11% -8% 1142 529

Italy 494 125 186 166% 297% -33% 575 125

Portugal 910 521 374 144% 75% 39% 1 161 349

Spain 469 170 248 89% 177% -32% 486 170

% change from

2011

high

2011

low

* Peak in 10Y Bund yields.

Exhibit 5: Rating agencies had a busy year cutting peripheral sovereigndebt ratings further

Average* long-term local currency debt rating for peripheral countries at the end of 2009, 2010 and 18 Nov 2011;

Greece Ireland Italy Portugal Spain

2009 2010 2011

AAA

AA

BBB

BB

CCC

A

B

* Average of Moody’s, S&P and Fitch. AAA/Aaa=1, AA+/Aa1=2, etc.




Gianluca SalfordAC

(44-20) 7325-4334


40

more than €120bn worth of Italian and Spanish paper

since it restarted the SMP (Exhibit 7); 2) Expansion of

EFSF and ESM lending capacity to €440bn and €500bn,respectively, with media speculation of combined fire

power 1 and still pending projects to leverage the EFSF

firepower through different mechanisms;2 3) A second

bailout package for Greece with PSI in July, which is

being renegotiated; and 4) After a mild stress test in the

summer, authorities pushed for a €100bn plan to

recapitalise banks in October.

What has been the impact of this flurry of

announcements? Since the beginning of the crisis,

policymakers’ promises of future action and surprise

ECB purchases have had a large impact both on

German yields and on peripheral spreads, buttypically the impact has been short-lived as evidenced

by sovereign spreads trading close to their widest

levels ever (Exhibit 8).

2012: Flirting with danger

We expect that the outlook for the Euro area fixed

income market will be dominated by a combination of 1)

macro developments, and 2) evolution of the sovereign

crisis and policy making response.

On the macro front, our economists are more negative on

growth than either the latest set of European Commissionforecasts or consensus, forecasting an area wide

recession that only Germany will avoid. Even though

we are far from forecasting deflation, we believe that in

2012 the ECB will be forced to do all it takes to support

the region, cutting the refi to 0.50% (Exhibit 9) in

quarterly 25bp steps.

As discussed at length in the Overview, we fear that, as

was the case in 2011, in the first part of 2012 domestic

political constraints will ensure that policymakers will

continue to be reactive rather than proactive. Italy and

Spain may lose access to the primary market, in which

case our view is that policy makers will realise themistake, step up the effort and manage to contain the

1We think it is unlikely to have the EFSF and the ESM in place

together (with combined capacity) as the German Supreme Court ruledthat Germany’s exposure should be limited and that the German

government must now obtain the approval of the Parliamentary Budget

committee before giving any guarantees. See Overview, Global Fixed Income Markets Weekly, 9 September 2011. 2 See Overview,Global Fixed Income Markets Weekly, 14 October 2011

and 28 October 2011.

Exhibit 6: The ECB started to buy Italian and Spanish bonds on8 August 2011 and currently holds almost €200bn of peripheral countrydebt in the SMP on a cash basisWeekly and cumulative ECB bond purchases through SMP*; official purchases data on

a cash basis; €bn

0

25

50

75

100

125

150

175

200

May-10 Aug-10 Nov-10 Mar-11 Jun-11 Sep-11

0

5

10

15

20

25

Weekly bond purchases (rhs)

Cumulativ e bond purchases (lhs)

* Excludes roughly €6.7bn of matured bonds.Source: ECB

Exhibit 7: We estimate that the ECB’s SMP purchases account for roughly 20% of the bond market in Greece, Ireland, and Portugal, and7% in Italy and SpainJ.P.Morgan estimate of notional amounts of peripheral bonds purchased through theECB SMP* and their current bond market size; €bn

Greece Ireland Portugal I taly Spain

Notional bought** 50 18 20 95 32

Bond market size 247 85 104 1359 474

% 20% 21% 19% 7% 7% * We estimate notional amounts using official total weekly purchases, cash prices andour estimate of country and maturity split.** Includes roughly €6.7bn of matured bonds, mostly Greek ones.Source: J.P.Morgan, ECB

Exhibit 8: Although policymakers’ announcements had a positiveimpact on markets, the effect was short-lived, as evidenced bysovereign spreads currently trading close to their widest levels ever Behaviour of the 10Y Bund and 10Y weighted peripheral spread* after significantpolicy announcements or actions; detailed announcements in grey

Date EventDays of

sell off

Peak sell

off (bp)

Days of

tightening

Peak %

tightening

03-May-10 First bailout plan for Greece 0 n/a 0 n/a

10-May-10 EFSF announcement, SMP 7 16 20+ -46%

23-Jul-10 EBA stress test 4 6 12 -8%

28-Nov-10 Irish bailout 20+ 35 0 n/a

10 -Jan-11 Discussion of comprehensive plan 20+ 42 20+ -27%

25-Mar-11 Comprehensive plan announcement 15 24 14 -8%

04-May-11 Portuguese bailout 1 2 1 -1%

15-Jul-11 EBA second stress test 0 n/a 0 n/a

21-Jul-11 EFSF 2, second Greek bailout 2 11 6 -12%

08-Aug-11 SMP 2 0 n/a 19 -17%

04-Oct-11 Discussion of comprehensive plan 2 8 46 4 -6%

26-Oct-11 Announcement of comprehensive plan 2 2 20 2 -5%

Max 20+ 46 20+ -46%

10Y wtd. peri. spreads*10Y Bund

* 10Y weighted peripheral spread computed against Germany for Greece, Ireland,Italy, Portugal, and Spain (weighted by the size of their outstanding bond market).




Gianluca SalfordAC

(44-20) 7325-4334


41

crisis. If everything plays out as expected, 2H12 should

see a marginal improvement in financial conditions.

Based on these views, we recommend investors:

1) Go long duration. All the drivers of our 10Y Bund

model point to lower yields in 2012 (see below). We

target 10Y Bunds yields at 1.25% (Exhibit 10).

2) Enter 2s/10s flatteners. We see limited scope for

much lower short-term yields and recommend flatteningexposure in 2s/10s, targeting around 100bp by mid-2012.

3) Position for further sovereign stress,

underweighting all intra-EMU countries vs.

Germany. We expect the credit curve to flatten with

wider spreads, but we favour weighting trades for the

empirical convexity of the price curve to minimise risk.

Our forecast for intra-EMU spreads is shown in

Exhibit 11.

Short-end yields: Not far from the bottom

Even with our downbeat view, we struggle to find

value at the short end of the German curve. Short-

term German yields are close to historical lows, already

pricing in 1) an ECB deposit rate at 0.25% for a

prolonged period of time, 2) extra liquidity in the system

to be abundant (Exhibit 12), and 3) a significant

premium for the perceived safety of German government

bonds.

It is unlikely that German 2Y yields will move much in

either direction in the first part of the year if our macro

forecast of declining GDP for four consecutive quarters

Exhibit 10: We forecast a trough in 10Y Bund yields at 1.25%J.P.Morgan interest rate forecast and spread vs. forwards; German benchmarksunless otherwise stated; %

18-Nov-11 1Q12 2Q12 3Q12 4Q12

2Q12 vs.

fwd (bp)

4Q12 v s.

fwd (bp)

ECB refi 1.25 0.75 0.50 0.50 0.50 n/a n/a

1M EONIA 0.65 0.38 0.32 0.32 0.32 -13 -30

2Y 0.46 0.35 0.30 0.40 0.50 -37 -44

5Y 1.10 0.75 0.70 0.85 1.05 -69 -57

10Y 1.97 1.55 1.25 1.50 1.75 -86 -48

30Y 2.61 2.15 1.95 2.25 2.50 -70 -18

2s/10s (bp) 151 120 95 110 125 -49 -4

10s/30s (bp) 64 60 70 75 75 16 30

Exhibit 11: Intra-EMU spreads will reach new highs by mid-2012Forecast of 10Y curve-adjusted spread to Germany; bp

18-Nov -11 Q2 Q4 18-Nov -11 Q2 Q4

Germany - - - 1.97 1.25 1.75 Austria 147 200 175 3.50 3.25 3.50

Belgium 282 400 350 4.80 5.25 5.25

Finland 67 100 80 2.61 2.25 2.55

France 147 225 190 3.46 3.50 3.65

Greece 2425 2400 2400 26.32 25.25 25.75

Ireland 625 850 900 8.15 9.75 10.75

Italy 494 775 625 6.95 9 .00 8.00

Netherlands 56 85 70 2.52 2.10 2.45

Portugal 910 1200 1300 11.04 13.25 14.75

Spain 469 700 600 6.70 8.25 7.75

Weighted spread* 461 625 555 - - -

Weighted peripheral spread** 699 955 855 - - -

Spread to Germ any (bp) Yield (%)

* Weighted spread computed against Germany for the 10 largest Euro area countries(weighted by the size of their outstanding bond market).

** Weighted peripheral spread computed against Germany for Greece, Ireland, Italy,Portugal and Spain (weighted by the size of their outstanding bond market).

Exhibit 12: High liquidity and significant premium for German bondsleave limited scope for a further decline in short-dated German yields

Average monthly seasonally adjusted excess liquidity* and 3M German GC-3M EONIAOIS; monthly data

€bn

0

50

100

150

200

Jan-11 Apr-11 Jul-11 Oct-11

-30

-25

-20

-15

-103M GC-OIS

Excess liquidity ; sa

* Seasonally adjusted excess liquidity = OMOs – Autonomous factors – Reserverequirements.

Exhibit 9: Our economists forecast a mild Euro area recession,declining inflation and official rates troughing at 0.50%Forecasts for 2012 real GDP growth, HICP inflation, and end-of-2012 ECB refi rate; %

-1.00

-0.50

0.00

0.50

1.00

1.50

2.00

Grow th Inflation ECB refi

J.P. Morgan Consensus EC

n/a

Source: J.P.Morgan, European Commission, Consensus Economics, Reuters




Gianluca SalfordAC

(44-20) 7325-4334


42

plays out. However, an improvement in 2H12 should

push yields higher as market participants might start to

see some light at the end of the tunnel. We forecast 2Yyields at 30-35bp until 2Q12 and at 50bp by the end of

2012. The 50bp year-end yield target for 2Y is consistent

with our expectations of 3M GC rates around 10bp and a

2Y benchmark/3M GC rate spread in line with average

since the beginning of 2010 of around 40bp (Exhibit 13).

10Y Bunds to drop to record low: Go long

In 2011, quite a few investors complained that low

German yields did not reflect fundamentals in their view.

Indeed, a decelerating but still-growing economy and

core inflation not far from the ECB target are difficult to

reconcile with 10Y Bund yields almost 200bp lower than

the 10-year average. We find that incorporating intra-

EMU spreads in a macro model explains the dramatic

decline in Bund yields over the past 2 years quite well

(Exhibits 14 and 15).

Looking ahead, all the explanatory variables point to

falling 10Y Bund yields over the next 6-12 months. On

the macro side, to be consistent with our growth forecast,

the composite PMI should fall further by 1H12, before

rebounding above 50 by the end of the year, whereas

core inflation should decline gently in 1H12, and more

aggressively in 2H12. We expect peripheral spreads to

widen further in 1H12 before retracing a bit. Pluggingthe inputs into the model would give a fair value for the

10Y Bund around 1.00% by the middle of the year

(Exhibit 16). We expect non-linearities to eventually

emerge, preventing the Bund from reaching such an

extreme level, but 10Y Bund yields at 1.25% certainly

seems possible. As discussed in detail in the demand

section below, we believe that Germany’s position as

safe haven is at risk in case of an escalation of the crisis

if policy makers are not able to contain the fallout from

the crisis.

Given our views on the short and long end of the curve,

we recommend 2s/10s flatteners, targeting around100bp mid-year. However, we have repeatedly found

that bouts of risk appetite can hurt flatteners, as was the

case in 4Q10 and in September-October 2011. When the

market is pricing in low-for-long, any short-term

improvement in sentiment is felt first at the long end of

the curve, creating a convex behaviour as the 2Y rate

catches up with a lag. Investors who seek protection from

bearish steepening should consider 1s/5s conditional bull

flatteners in swaps (see Euro Derivatives).

Exhibit 13: Under the assumption of 3M GC repo at 10bp, we forecast2Y rates at around 50bp by the end of 2012, 40bp above 3M GC repo, inline with the average of past two yearsEvolution of 2Y German par govt rate minus 3M general collateral repo rate; %

-0.40

-0.20

0.00

0.20

0.40

0.60

0.80

1.00

Jan10 Apr10 Jul10 Oct10 Jan11 Apr11 Jul11 Oct11

Average: 41bp

Exhibit 14: In our fair value model, 10Y Bunds yields are explained by acombination of macro variables and sovereign stress…10Y Bund yields regressed against composite PMI, HICP ex food and energy and 10Yweighted peripheral spread*; daily interpolated data over past 10Y; %

Beta T-stat

Intercept 0.716 9.3

Composite PMI (units) 0.031 25.0

Core inflation (%oya) 0.901 70.1

10Y weighted peripheral spread (bp) -0.003 -47.7

R-squared 83%

Standard error 0.28 * Weighted peripheral spread computed against Germany for Greece, Ireland, Italy,Portugal and Spain (weighted by the size of their outstanding bond market).

Note: In the presence of the above factors, addition of front-end yields as another explanatory variable does not materially improve the goodness of the fit.

Exhibit 15: … with intra-EMU spread widening accounting for around175bp of decline in 10Y Bund yields over the past two yearsMarginal contribution to 10Y Bund yields*: deviation from average; %

-2.0

-1.5

-1.0

-0.5

0.0

0.5

1.0

2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

PMIInflation10Y wtd. peripheral spread

* See Exhibit 14 for 10Y Bund yield model details.




Gianluca SalfordAC

(44-20) 7325-4334


43

Given our forecast for record-low yields, we look at the

example of Japan to analyse the behaviour of the

2s/5s/10s fly under extreme conditions. Typically, in a

scenario of unchanged monetary policy stance, the fly is

fairly directional, allowing investors to choose between

outright longs or barbell trades based on RV and relativecarry and slide considerations. However, even with

unchanged monetary policy, when interest rates drop

significantly the directionality of the 2s/5s/10s can flip,

with the 5Y underperforming a combination of 2Y

and 10Y amid falling yields; this was the case in Japan

(Exhibit 17), and for a few days in the Euro area during

mid-September. We therefore do not recommend

buying 5Y in a 50:50 2s/5s/10s fly, despite our bullish

outlook.

At the ultra-long end of the curve, we anecdotally

witnessed a reduction in client interest in 10Y+ core

bonds, with an impact on relative liquidity comparedwith shorter segments of the curve. This forced primary

dealers to increasingly use the Buxl futures as a hedging

instrument (Exhibit 18). We recommend neutral

exposure on 10s/30s curve.

Finally, on RV, we highlight that the 5s/10s curve is

trading close to the cheapest level of the past 10 years vs.

2s/30s (Exhibit 19). We recommend 5s/10s flatteners

against weighted 2s/30s steepeners.

Intra-EMU spreads: Wider in 1H12

Given the view highlighted in the Overview and above,

we believe investors should underweight all intra-

EMU countries vs. Germany in the first part of the

year. However, stronger policy action should promote a

mild reversal in most countries in 2H12 (Exhibit 11).

To refine our views, we analyse the bond market along

two axes: 1) Country selection, and 2) Maturity/coupon

selection.

Exhibit 16: Under our baseline scenario, a further drop in the PMI, lower core inflation and wider peripheral spreads should pressure Bundyields below 1.00%; however, given non-linearities, we expect Bundyields to bottom around 1.25% in 1H1210Y Bund fair value yield based on the model described in Exhibit 14 and J.P.Morganforecasts for the explanatory variables; 1H12

Level Beta

Composite PMI (units) 43.75 0.031

Core inflation (%oya) 1.50 0.901

10Y weighted peripheral spread (bp) 955 -0.003

Fair value 10Y Bund yield (%) 0.85

Exhibit 17: We do not recommend buying 5Y in 2s/5s/10s fly as abullish duration trade because the directionality can change whenyields drop substantially6M beta of 50:50 2s/5s/10s vs. 5Y JGB yield; 1 Jan 2001 – 18 Nov 2011;

-1.2

-1.0

-0.8

-0.6

-0.4

-0.2

0.0

0.2

0.4

0.6

0.8

0.00 0.50 1.00 1.50 2.00

5Y JGB yield; %

Exhibit 18: Buxl futures volumes have increased amid decliningliquidity in the cash market3M rolling average of Bund and Buxl futures volumes; ‘000s of contracts

400

600

800

1,000

1,200

1,400

1,600

2006 2007 2008 2009 2010 2011

2

3

4

5

6

7

8

9

Bund

Buxl

Exhibit 19: 5s/10s remains close to the cheapest level of the past 10years vs. 2s/30sResidual of 5s/10s regressed against 2s/30s*; German government par rates; past 10years; bp

-20

-15

-10

-5

0

5

10

15

20

2001 2003 2005 2007 2009 2011

Current: 15bp

* 5s/10s = 0.35*2s/30s – 2.94; R-squared: 96%




Gianluca SalfordAC

(44-20) 7325-4334


44

In our view, country selection hinges less than in

previous years on fundamental, medium-term

analysis and more on technical factors, due to theprogressive deterioration of market liquidity. We

define technical factors as events that change market

dynamics in a non-linear fashion, such as rating

downgrades and loss of market access among the

negatives, and explicit or implicit forms of market

support on the positive side.

We expect France to continue to underperform in relative

terms on the back of rating-sensitive selling by foreign

accounts. The importance of the rating is not confined to

France. If Italy loses market access and gets downgraded

to BBB space, as stated by rating agencies under such an

occurrence, there is likely to be further pressure on theinflation-linked segment due to potential index-inclusion

rules (see Inflation-linked Markets section), in addition to

further erosion of its investor base. As discussed in the

Overview, we expect any request for support to be

market-negative: in Greece, Ireland, and Portugal, selling

pressure accelerated after bailouts (see Exhibit 30 in the

Overview). On the positive side, positive technical

pressure can come from aggressive SMP-buying, which

could offset the market signal, and from idiosyncraticissues such as the Irish re-investment of the promissory

note proceeds.3

In terms of valuations, our updated sovereign risk index

(Exhibits 20 and 21) shows that:

1) Austria, Portugal and Spain are cheap vs. fair

value. We agree with Austria, but not on Portugal

and Spain. We believe the case for an Austrian

downgrade is dubious if the authorities take appropriate

action (see below). On Portugal, we believe that there is

a high probability that investors will be at least asked to

extend maturities of existing Portuguese bonds later inthe year, making current valuations (especially at the

short end) unattractive. Despite a low debt/GDP ratio, we

believe that Spanish bonds can come under significant

pressure if the country loses market access.

2) France, Greece and Netherlands trade on the

3 See Euro cash, Global Fixed Income Markets Weekly, 14 October 2011 for details.

Exhibit 20: Sovereign risk index: nobody is perfectJ.P.Morgan sovereign risk index* and its components; lower-than-average risk in grey;

2012

debt/GDP

(%)

2012

deficit/

GDP (%)

Total

private

debt/GDP

2012 current

account

balance (% of

GDP)

Global

competitiveness

index

Political

risk index

Industrial

action index

2012

GDPpc

(EU=100)

Data Austria 73 -3.1 150 2.8 5.14 2 0 124

Belgium 99 -4.6 131 2.1 5.20 7 79 116

Finland 52 -0.7 153 0.0 5.47 2 73 123

France 89 -5.3 138 -3.3 5.14 2 132 104

Germany 81 -1.0 111 4.4 5.41 3 6 108

Greece 198 -7.0 122 -7.9 3.92 4 132 63

Ireland 118 -8.6 341 1.5 4.77 2 38 119

Italy 121 -2.3 126 -3.0 4.43 4 35 90

Netherlands 65 -3.1 219 7.0 5.41 4 6 126

Portugal 111 -4.5 223 -5.0 4.40 1 11 54

Spain 74 -5.9 211 -3.0 4.54 0 60 80

Cross sectional z-score Sov. risk index

Austria -0.6 -0.4 -0.4 -0.7 -0.5 -0.4 -1.1 -0.9 -0.6

Belgium 0.0 0.2 -0.6 -0.6 -0.6 2.2 0.6 -0.6 0.1

Finland -1.2 -1.4 -0.3 -0.1 -1.1 -0.4 0.4 -0.9 -0.6

France -0.2 0.5 -0.5 0.7 -0.5 -0.4 1.7 -0.1 0.1

Germany -0.4 -1.3 -0.9 -1.1 -1.0 0.1 -1.0 -0.3 -0.7

Greece 2.5 1.1 -0.8 1.7 1.9 0.6 1.7 1.5 1.3

Ireland 0.5 1.8 2.4 -0.4 0.2 -0.4 -0.3 -0.7 0.4

Italy 0.6 -0.8 -0.7 0.6 0.9 0.6 -0.4 0.4 0.2

Netherlands -0.8 -0.4 0.6 -1.7 -1.0 0.6 -1.0 -1.0 -0.6

Portugal 0.3 0.1 0.7 1.0 1.0 -1.0 -0.9 1.8 0.4

Spain -0.6 0.7 0.5 0.6 0.7 -1.5 0.2 0.8 0.2 * Calculated as weighted average of cross-sectional z-spreads; debt/GDP ratio accounts for 25% of the weight, other variables have equal weight.Source: Eurostat, EC, ECB, National Central Banks, World Economic Forum, European Industrial Relation Online




Gianluca SalfordAC

(44-20) 7325-4334


45

expensive side. We agree on all accounts. We have

been highlighting for a while that France’s fundamentals

are weaker than its AAA peers. 10Y Greek bonds arefairly priced for the second PSI (with 50% haircut), but

we expect another debt restructuring in the not too distant

future. With regards to the Netherlands, we believe that

domestic support is likely to help only at the ultralong

end of the curve.

We cover the most interesting issues for the major Euro

area countries in the section below.

Austria

We believe most macro variables for Austria look

very good, and certainly consistent with a AAA

rating. However, we see two sources of weakness for Austria that need to be addressed: 1) The bond market is

small, with limited domestic participation (less than 25%

according to financial accounts data) and therefore

potentially vulnerable to the vagaries of the market; and

2) Austrian banks’ net exposure to countries outside

the Euro area, in particular to Eastern Europe,

remains the country’s Achilles’ heel (Exhibit 22). Our

working assumption is that the government has the

resources and the will to provide a stronger support

framework in place: S&P recently expressed the desire to

see non-voting participation capital that some banks

raised during the crisis transformed into more solid

ordinary capital.

Belgium

Despite investors’ focus on higher-than-average

debt/GDP ratio, most of Belgium’s economic variables

are pretty good, especially the net international

investment position. Political issues have been in the

spotlight for more than a year now, and despite the

country working well on ‘autopilot’ for quite some time,

there is a need for a consensus on roughly 3% of GDP

worth of fiscal consolidation in 2012. Discussions are

ongoing but have been so far unsuccessful. The fragility

of financial institutions is the second issue that concerns

us, especially given the amount of support that was

needed in the 2008-09 crisis, and more recently, with

Dexia (Exhibit 23).

Finland

Finland is a very solid AAA country from a macro point

of view, but is likely to be penalised vs. Germany by the

relative lack of liquidity. However, low borrowing

requirements and the potential support coming from its

extensive general government assets should limit

excessive underperformance. Finland is in a unique

Exhibit 21: We find that Austria, Portugal, and Spain trade on the cheapside vs. fundamentals, whereas Greece, Netherlands and France are onthe expensive side

10Y spread to Germany vs. fair value implied by sovereign risk index*; bp;Index Current Model Difference % difference

Austria -0.6 147 81 66 81%

Belgium 0.1 282 299 -18 -6%

Finland -0.7 67 71 -4 -6%

France 0.1 147 304 -156 -51%

Greece 1.5 2,425 4,128 -1703 -41%

Ireland 0.4 625 565 60 11%

Italy 0.2 494 399 95 24%

Netherlands -0.6 56 84 -28 -33%

Portugal 0.4 910 550 360 65%

Spain 0.1 469 298 172 58% * 10Y spread = exp (1.88*sov risk index + 5.6); R-squared: 86%; see Exhibit 20 for thecomputation of the sovereign risk index.

Exhibit 22: Austria’s main weakness lies in the net exposure of itsbanking sector to extra Euro area countries, second only to IrelandNet external assets of the banking sector; as of September 2011; %GDP

-20%

0%

20%

40%

60%

80%

A u s t r i a

B e l g i u m

F i n l a n d

F r a n c e

G e r m a n y

G r e e c e

I r e l a n d

I t a l y


P o r t u g a l

S p a i n

Source: ECB

Exhibit 23: Belgian financial institutions have received considerablestate support since the beginning of the crisisDetails of state support to major financial institutions with significant activities inBelgium; €bn

Capital Guarantees*

Asset

protection

scheme Capital Guarantees

Asset

protection

schemeDexia Belgium 3.0 91 - 4.0 54 -

Others 3.4 59 - - 36 -

KBC Belgium 7.0 - 20 - - -

Fortis** Belgium 4.7 - - - - -

Others 6.5 - - - - -

Ethias Belgium 1.5 - - - - -

2008-09 2011

* Total guarantees agreed for Dexia were €150bn in 2008-09, peak was <€100bn,currently <€30bn.** The original plan for Fortis (capital injection) was eventually changed, withnationalisation of the Dutch arm and sale of the Belgian arm to BNB Paribas.




Gianluca SalfordAC

(44-20) 7325-4334


46

position among the major Euro area countries, with

general government assets exceeding liabilities by a

wide margin (Exhibit 24).

France

We believe that the main issue that investors have with

France relates to the country’s ability to hold to its AAA

rating.

Moody’s and S&P have laid the ground for an action on

France with their recent comments. Moody’s focused on

“the deterioration in debt metrics and the potential for

further contingent liabilities to emerge are exerting

pressure on the stable outlook of the government’s Aaa

debt rating”. S&P instead focused on macro

performance, suggesting that in case of a double-dip in

the Euro area, France would likely be downgraded.

We review rating actions on AAA countries over the past

five years (Exhibits 25 and 26). We find that:

1) Negative outlooks and negative watches are unlikely

to be reversed.

2) In all instances, a downgrade has been preceded by

either a negative outlook or a negative watch.

3) Downgrades have occurred with an average lag of

roughly 3 months from a negative outlook.

4) Fiscal problems are (unsurprisingly) the most quoted

reason for action, followed by deteriorating economic

outlook, problems in the financial sector which lead to

Exhibit 24: The Finnish government balance sheet is very strong asgovernment assets are significantly higher than liabilitiesNet general government assets (financial assets minus financial liabilities) of Euro areacountries; % of GDP; latest data available

-125%

-100%

-75%

-50%

-25%

0%

25%

50%

75%

A u s t r i a

B e l g i u m

F i n l a n d

F r a n c e

G e r m a n y

G r e e c e

I r e l a n d

I t a l y

N

e t h e r l a n d s

P o r t u g a l

S p a i n

Source: National financial accounts

Exhibit 25: Downgrades of a AAA country have always been precededby either a negative outlook or a negative watchNegative outlook/watch and rating actions on AAA countries over the past 5 years;

Country Agency Negativeoutlook

Negativewatch

Downgrade

Outlook -

downgrade

(months)

Watch -

downgrade

(months)

Currentrating

Iceland Moody 's 05-Mar-08 20-May -08 2.5 Baa3

Ireland Moody ' s 30-Jan-09 17-Apr-09 02-Jul-09 5.0 2.5 Ba1

Ireland S&P 09-Jan-09 30-Mar-09 2.7 BBB+

Spain Moody 's 30-Jun-10 30-Sep-10 3.0 A1

Spain S&P 12-Jan-09 19-Jan-09 0.2 AA-

UK* S&P 21-May -09 AAA

US Moody 's 02-Aug-11 13-Jul-11 Aaa

U S S&P 1 8-Apr-11 14-J ul-11 05-Aug-11 3. 6 0. 7 AA+

Av erage - - - - 3.5 1.6 * The UK outlook was changed back to stable on 26 October 2010.Source: Bloomberg, Moody’s, and S&P

Exhibit 26: Fiscal problems have been the most quoted reason for areview/watch/rating action on a AAA rated countryReasons quoted by rating agencies for review/watch/rating action on AAA countriesover the past 5 years;

Country Agency Economic FiscalFinancial

sector

Funding

costsPolitical

External

position

Technical

issues

Iceland Moody's X X

Ireland Moody's X X X X

Ireland S&P X X X

Spain Moody's X X X

Spain S&P X X X

UK S&P X X

US Moody's X X

US S&P X X X

4 6 4 2 3 1 2

X X X X X

Problems

Grand total

Potential issues for

France

Source: Moody’s and S&P

Exhibit 27: In the event of a French downgrade, we expect the 10-15Ypart of the French curve to find better support than other sectors,thanks to the large domestic insurance sector Total assets (financial and non-financial) of insurance corporations of the Euro areacountries; 2010;

€bn % GDP

Austria 119 40%

Belgium 250 67%

Finland 52 27%France 1910 96%

Germany 1549 60%

Greece 16 7%

Ireland - -

Italy 525 33%

Netherlands 414 68%

Portugal - -

Spain 274 25%

Total 5108 100%

Insurance corp. assets

Source: ECB




Gianluca SalfordAC

(44-20) 7325-4334


47

contingent liabilities, in addition to question marks about

political resolve.

5) Only the UK has been able to return to a neutral

outlook thanks to a combination of lower-than-expected

cost of contingent liabilities, strong political resolve, and

improving economic outlook.

In terms of macro and fiscal variables, France starts from

a weaker position than other AAA countries (Exhibit 20).

A further deterioration in the macro outlook and in

financial stress is likely to add to the pressure on France.

Worries about contingent liabilities, increasing cost of

funding, and political uncertainty associated with

presidential elections are likely to prompt a rating

action in 2012.

We expect pressure to come from foreign central banks’

diminished appetite for short-end paper. We are more

sanguine about the prospects for the 10-15Y part of

the curve, given the massive size of the domestic

insurance industry (Exhibit 27) and the likelihood of

asset rotation into long-dated French bonds. Official data

showed that domestic insurance companies hold around

€220bn in French government securities, out of €1.9tn of

assets.

Germany

At the moment of writing, Germany is the only Euro areacountry that has always traded like a safe-haven asset,

thanks to solid macro fundamentals but also helped by

the liquidity that its size and bond futures provide. In our

central scenario this will continue, with yields expected

to drop even further on the back of intra-Euro area

reallocation flows. However, policy makers’ inability to

provide an effective backstop and an escalation of the

crisis beyond our central scenario would likely challenge

our view: at that point, a run on the currency by extra-

Euro area investors would dent Germany’s safe-

haven status.

Greece

The appointment of a technocratic government in Greece

should pave the way for successful implementation of the

second attempted PSI even though the NPV of the new

structure might be below 30c. We expect the voluntary

bond exchange to go through, but we believe that

bond holders will eventually be asked for a further

contribution, although it is difficult to predict the

timing. The bond price curve has flattened aggressively

since the announcement of the first PSI (Exhibit 28). We

expect further selling pressure into the end of the year as

European financial institutions continue to clean their

balance sheets.

Ireland

Ireland is the success story of 2011, with 10Y spreads to

Germany close to unchanged on the year despite a selloff

at the short end. We believe the reasons for this relative

outperformance are both fundamental and technical: 1)

For the first time in 4 years, there appears to be some

stabilisation in the debt/GDP ratio trajectory; 2) The peak

in Irish outperformance coincided with heavy domestic

purchases on the back of the capital injections in the

Exhibit 28: The Greek bond price curve is very flat as investors attach ahigh probability to a debt restructuringClean-price GGB curve after announcement of first and second PSI and current*;

points

20

30

40

50

60

70

80

90

2012 2016 2020 2024 2028 2032 2036 2040

Bond maturity date

PSI 1

PSI 2

Current

* Dates used: First PSI: 22 July 2011, Second PSI: 27 October 2011 and current:18 November 2011.

Exhibit 29: We remain sceptical about Ireland’s ability to deliver theneeded fiscal consolidation from a deficit/GDP above 10% in 2011General government budget balance for Ireland and average of peripheral countries exIreland; realised numbers for 2007-2010 and forecasts for 2011-2012; % of GDP

0.1

-7.3

-14.2

-11.2-10.3

-8.6

-2.3

-4.9-6.3

-8.6

-10.6

-5.2

-16

-14

-12

-10

-8-6

-4

-2

0

2

2007 2008 2009 2010* 2011 2012

Ireland Avg. peripheral ex Ireland

*Irish deficit number for 2010 exclude the 20.1% one-off banking sector supportmeasures adjustment.Source: 2012 EC autumn forecasts




Gianluca SalfordAC

(44-20) 7325-4334


48

summer. Irish banks increased their bond portfolio by

more than 20% in August and September alone. Going

forward, we remain sceptical about Ireland’s ability todeliver the required fiscal consolidation from a

deficit/GDP above 10% (Exhibit 29), still-enormous

net external liabilities, and still-falling house prices.

We are medium-term negative on short-dated bonds as

we believe they do not discount a high enough

probability of restructuring, even in the milder form of

maturity extensions.

Italy

We cover Italy at length in the Overview and in the

demand/supply section below. The bottom line is that we

see a risk that access to the market may become

impaired, prompting significantly higher yields anddowngrades, with technicals especially negative for

inflation-linked bonds.

The Netherlands

Dutch bonds have remained, until recently, remarkably

stable compared with Germany, thanks to their strong

perceived creditworthiness. Amid overall sound macro

fundamentals, the ongoing decline in house prices raises

some concerns in the context of a heavily leveraged

household sector, as highlighted by the Dutch National

Bank in the Financial Stability Report. Going forward,

we see the emergence of further market segmentation

in the DSL market (Exhibit 30), with domesticinsurance companies and pension funds likely to keep the

long and ultra-long ends of the curve supported, and

more pressure on the short end of the curve due to lower

structural domestic demand.

Portugal

Our assessment of the latest troika’s Portugal review is

mixed: the 5.9% 2011 deficit/GDP target will only be

met through an account expedient due to expenses

overruns, but the troika expressed confidence about

hitting the 4.5% 2012 target. The need for structural

reforms was also stressed. We believe that a long period

of fiscal tightening and internal devaluation to restore

competitiveness in a highly leveraged economy is

unlikely to prove successful.

An ex-official Treasury official recently stated that

Portugal might need additional €20-25bn in addition to

the €78bn packaged agreed in the spring as public

companies are unlikely to be able to roll their debt, an

event not contemplated when the package was approved.

Given the risks of PSI in the latter part of the year

when a new package will be negotiated, we are

generally negative on Portuguese bonds, especially in

the 2-3Y part of the curve.4

Spain

Spain’s attempt at rebalancing its economy is impressive,

especially in terms of current account balance, which

moved from -9.6% of GDP in 2008 to a more

manageable -3.4% in 2011. The fiscal journey has also

been laudable, with more than 4% points of deficit

reduction between 2009 and 2011, but Spain continues to

4See Overview, Global Fixed Income Markets Weekly, 4 July 2011 for

details.

Exhibit 30: We expect domestic insurance companies to keep the longand ultra-long ends of the Dutch curve well supported, whereas theshort/intermediate part of the curve will remain under pressure due to

lack of structural domestic demandDutch bonds spread to interpolated German curve; bp

0

20

40

60

80

2012 2018 2024 2030 2036 2042

Bond maturity date

Exhibit 31: Spain’s strong domestic investor base provides support toits bonds markets% of domestic investors in central government securities in selected Euro areacountries; %

34%

19%

33%

17%

55%

20%

62%

0%

10%

20%

30%

40%

50%

60%

70%

F r a n c e

G e r m a n y

G r e e c e

I r e l a n d

I t a l y

P o r t u g a l

S p a i n

Source: J.P.Morgan estimates, national central banks, debt management agencies




Gianluca SalfordAC

(44-20) 7325-4334


49

suffer from 1) poor fiscal discipline at the local level; and

2) uncertainty about the final cost of measures to clean

banks’ balance sheets of non-performing assets. Amongthe positives for Spain, it is worth highlighting the high

percentage of domestic investors (Exhibit 31). However,

we see a risk that Spain may lose market access at

some point during the course of 2012, pushing yields

higher.

Peripheral duration neutral curve flatteners and

weighted flatteners are attractive

After a few years of a sovereign debt crisis, it is

empirically well established that credit spread curves

are directional to the level of spreads: the wider the

spread to Germany, the flatter or inverted the credit curve

(Exhibit 32). The process is very common in credit

markets when bonds transition from high-grade to high-

yield. Investors can analyse the relative merits of curve

positions (boxed vs. Germany) vs. outright spread

positions looking at relative value and/or carry and slide.

Based on our view that spreads are going to widen across

the board and our flattening view on the German curve,

we recommend duration-neutral curve flatteners in

France, Italy, and the Netherlands. Selling high-price,

short-dated bonds and buying low-price, long-dated ones

can make the strategy more attractive (see below).

Price convexity between short and long dated bondsFor investors who want to be protected in case of a

tightening of spreads to Germany that would result in a

steepening of the credit curves, a source of relative

value can be found in the empirically convex

relationship between prices at the short and long ends

of the curve. As an example, we look at 3Y and 30Y

Portuguese bonds (Exhibit 33) to highlight that 1)

initially, 30Y bond prices fall more forcefully than prices

at the short end of the curve, but 2) eventually, the short

end of the curve catches up as the probability of a short-

dated credit event increases. Based on the empirical

regression, when 3Y Portuguese prices were at 100, 90,

and 80, the sensitivity of 30Y prices were 1.7, 1.2, and

0.6, respectively. At an extreme, under most debt

restructurings, bond prices across maturities tend to

converge to similar levels.

The Italian and Spanish price curves have, until now,

exhibited limited convexity, but we expect this to

change in case of further stress, in line with the

experience of Greece, Ireland, and Portugal. Empirically,

30Y Italian bond prices are exhibiting roughly twice as

much price volatility as 3Y ones (Exhibit 34). Selling

Exhibit 32: Credit spread curves are directional to the level of spreads:the wider the spread to Germany, the flatter or inverted the credit curve.Given our view that spreads are going to widen across the board, werecommend duration-neutral curve flatteners in all countries3s/10s spread curve to Germany for the Euro area countries ex Greece, regressedagainst 3Y spread to Germany; bp

y = -0.0002x 2 - 0.014x + 7.4

R2 = 98%

-500

-400

-300

-200

-100

0

100

0 500 1000 15003Y spread to Germany; bp

Ireland

PortugalItaly

Spain

Belgium

Exhibit 33: Under stress, the price curve exhibits convex behaviour asthe short end finally catches up with the long end in terms of losses30Y Portuguese bond price vs. 3Y Portuguese bond price; s ince May 2010*; cleanprices; points

y = 0.021x 2 - 2.7x + 134.5

R2 = 94%

40

50

60

70

80

90

100

60 70 80 90 100 1103Y Portugal bond price; points

Slope: 1.7

Slope: 1.2

Slope: 0.6

* The ECB started the SMP on 10 May 2010.

Exhibit 34: We recommend selling €100mn of BTPs Jun14 vs. €50mn of BTPs Feb37 to exploit the convexity of the price curve relationshipRecommended trade weight on short BTP Jun14-long BTP Feb37 based on empiricalprice relationship since the beginning of the first ECB SMP*;

BTP Jun14 BTP Feb37 Ratio

Dirty price 95.6 69.9 0.73Empirical price beta 1 2.01

Nominal weight 100.0 49.7

Cash weight 100.0 36.3

Yield (%) 6.23 6.67

Modified duration 2.25 12.94

PVBP 2.15 9.04 * BTP Feb37 = 235.3 - 5.64*(BTP Jun14) + 0.041*(BTP Jun14)^2; R-squared: 95%;calculated using clean prices.




Gianluca SalfordAC

(44-20) 7325-4334


50

€100mn of BTPs Jun14 into €50mn of BTPs Feb37

would result in a trade that is locally hedged for small

price changes and would start to be profitable in case of further sovereign stress.

How to price high- low-coupon bond yield spreads

under sovereign stress

As the sovereign crisis evolved from liquidity to

potential solvency issues, investors have been switching

out of high-coupon, high-price bonds into low-coupon,

low-price bonds with similar maturity. Exhibit 35 shows

how the yield spread of the pair with the highest coupon

mismatch (BTP Aug23-Nov23) has moved from virtually

zero to around 60bp.

Although it seems intuitive to recommend switches

between bonds with similar maturity and yield, and

different coupon and price under sovereign stress, it

is not straightforward to estimate the fair value once

yield spreads have already moved as in the example

above. In the current environment of relatively flat yield

curves like the Italian or Spanish ones, traditional

cheap/dear analysis based on bootstrapping would

suggest that wide yield spreads between high- low-

coupon bonds with similar maturities are not justified

and have a source of relative value.5

One needs to incorporate in the analysis risk free rates,the probability of a credit event throughout the bond life,

and the recovery rate in order to justify higher yields for

higher coupon bonds.

We analyse how bond maturity and coupon rates

impact yield spreads for same-maturity bonds in a

simplified world (with constant conditional probability

of default per year, constant recovery rate and constant

zero risk-free rate). We calculate NPVs for theoretical

bonds with different characteristics and compare their

fair value IRRs. Typically, for bonds with the same

maturity and fixed coupon differential: 1) the longer the

maturity of the bond the wider the yield spread (Exhibit36), 2) the lower the coupons the wider the yield spread,

3) the higher the conditional probability of debt

restructuring, the higher the impact on the yield spread

(Exhibit 37).

5In the typical textbook world of risk free government bonds and

upward sloping yield curves, high-coupon, high-price bonds should

trade with a higher yield than low-coupon, low-price bonds with similar maturities.

Exhibit 35: The yield spread between high- and low-coupon bonds hasrisen with higher Italian yieldsYield spread between BTP 9% Nov23 (current clean price 111.6) and BTP 4.75%

Aug23 (current clean price 82.6) regressed against yield of BTP 4.75% Aug23; last sixmonths; %

0

10

20

30

40

50

60

70

4.00 5.00 6.00 7.00 8.00 9.00BTP 4.75% Aug23 yield; %

y =20.1x + -87.5

R 2 = 81%

Exhibit 36: The high- low-coupon impact on yield spread is positivelycorrelated with bond maturity but negatively with the coupon rateTheoretical yield spread between two bonds with the same maturity and coupon of thefirst bond 2% lower than the coupon of the second bond under different maturities andcoupon assumptions; NPV and IRR calculations are based on 0% risk free rate,conditional default rate of 5% per year, 40% recovery rate; %

0.00

0.10

0.20

0.30

0.40

0.50

0.60

0.70

0.80

0.00 2.00 4.00 6.00 8.00 10.00

Coupon rate of low er-coupon bond; %

30Y

10Y

Exhibit 37: The high- low-coupon impact on yield spread is positivelycorrelated with the expected probability of restructuringTheoretical yield spread between two bonds with the same maturity and coupon of thefirst bond 2% lower than the coupon of the second bond under different couponassumptions and conditional default probabilities per year; NPV and IRR calculations

are based on 30Y bonds, 0% risk free rate, 40% recovery rate; %

0.00

0.20

0.40

0.60

0.80

1.00

1.20

1.40

1.60

1.80

0.00 2.00 4.00 6.00 8.00 10.00

10% conditional prob. of default

5% conditional prob. of default

Coupon rate of low er-coupon bond; %




Gianluca SalfordAC

(44-20) 7325-4334


51

We can now apply this stylised framework to two real

cases: 12Y BTPs and 20Y-25Y Bonos (Exhibit 38).

What is the fair value spread between 12Y BTPs withcoupons of 4.75% and 9.00% when the low coupon bond

is yielding around 7.00%? We assume a fixed recovery

rate of 40% and fixed risk free rate of 2%, and we

calculate the annual conditional probability of default

that is consistent with the market yield of the low coupon

bond. We then use the same parameters to calculate the

theoretical yield on the high coupon bonds6. Based on

our model, the spread should be around 75bp vs. current

market level of 60bp, making the market spread broadly

fair. What is the fair value spread between the 20Y

5.75% Bono and the 25Y 4.20% Bono when the latter is

yielding around 7.00%? Based on our model, the spread

should be around -80bp vs. current market level of -7bp,suggesting a significant mispricing. We therefore

recommend investors sell the Bono Jul32 and buy the

Bono Jan37.

In conclusion, under sovereign stress, the standard

cheap/dear analysis is not useful to deal with bonds

issued by the same issuer, with similar maturities but

with significantly different coupons. The fair value

yield spread between bonds with different coupons is

positively influenced by maturity and default probability,

and negatively influenced by the coupon rate for a

constant coupon difference. We believe the BTP Aug23-

Nov23 yield spread is not far from fair price, but theBono Jul32-Jan37 spread is too tight.

Bond demand: In whose hands is the

hot potato?

Given the increasingly technical nature of the Euro area

government bond market, we look at its investor base,

first in aggregate (Euro area) and then for the four major

countries. We conclude that despite different investor

bases, no country is in the position of Japan, where

95% of the country’s domestic investor base acts as a

potent dampener of sovereign stress (see Japan).

Euro area

According to ECB data, in mid-2011, Euro area banks

(monetary and financial institutions [MFIs], to be

precise) held roughly 23% of Euro area (general)

government securities, whereas the Euro system held 7%

of the total. It is important to highlight that SMP holdings

account for less than 50% of that amount. Euro area

6 Using inputs that are extracted from market prices is computationally

more complex, but does not change the results significantly.

insurance companies and pension funds had sizeable

holdings (18%), whereas investors outside the Euro area

held almost 1/3 of the total (Exhibit 39). We estimatethat central banks constitute the lion’s share of external

holdings. For more specific cross-border exposure in the

banking sector, please refer to Exhibits 11 and 12 in the

Overview.

Flow-wise (Exhibit 40), Euro area banks were decent

buyers of bonds in 1H11, but consistent with

anecdotal evidence, they became heavy sellers in

3Q11 (and likely in 4Q11, based on individual bank

updates). In the recent words of the president of the

Exhibit 38: We believe that the BTP 4.75% Aug23-9% Nov23 spread isclose to fair value, but the Bono 5.75% Jul32 is too expensive vs. theBono 4.2% Jan37

Market and theoretical* yields on BTPs 4.75% Aug23 and 9% Nov23 and Bonos4.20% Jan37 vs. Bono 5.75% Jul32; %

Maturity Coupon Dirty price Mkt yield Theoretical yield

BTP Aug-23 4.75 84.1 7.06 7.06

BTP Nov-23 9.00 112.2 7.65 7.81

Spread (bp) 59 75

Bono Jan-37 4.2 71.3 7.01 7.01

Bono Jul-32 5.75 88.7 7.08 7.80

Spread (bp) -7 -79 * We adjust a flat conditional probability of default under 40% recovery rate and 2%risk free rate to match the market yield of the first bond and calculate the theoreticalyield of the second bond under the same parameters.

Exhibit 39: Euro area banks and extra-Euro area investors are thelargest holders of Euro area government bonds…Split of Euro area general government bond holdings; as of 2Q11

bn %

MFIs 1,579 23%

Eurosystem (ECB, NCBs)* 469 7%

Investment funds 671 10%

Insurance companies + pension funds 1,228 18%

Others 633 9%

Rest of the World 2,183 32%

Total 6,763 100%

2Q11

* The number includes holdings other than SMP.Source: ECB

Exhibit 40: … but banks turned net sellers of bonds in 3Q11Change in Euro area general government bond holdings*; €bn

2009 2010 1H11 3Q11 4Q11*MFIs 228 41 53 -43 n/a

Eurosystem (ECB, NCBs)** 42 104 15 81 33

Insurance co. + pension funds 100 101 13 n/a n/a

Others -13 199 47 n/a n/a

Rest of the World 248 154 150 n/a n/a

Total 604 598 278 n/a n/a * SMP purchases from 1 October 2011 till 18 November 2011.** Including revaluation/devaluation changes up to 3Q11.Source: ECB




Gianluca SalfordAC

(44-20) 7325-4334


52

European Banking Federation: “The banks are doing

exactly what they should be doing: they are reducing

their risk … We can see that clearly as now Italian bondsare being sold off.” The country wise split shows that in

3Q11, French banks were the most aggressive in

reducing their government bond portfolios (-€16bn),

followed by Spanish and German ones, while Italian

banks were still adding to their government bond

portfolios.7

After a relatively quiet 1H11, the ECB had to

significantly increase purchases: the central bank releases

on a weekly basis their SMP holdings, providing the

timeliest information. We estimate that between the end

of 3Q11 and mid-November, another €33bn was bought

by the ECB in addition to €81bn in 3Q11. For the other investors, data is available only until 2Q11. It is worth

highlighting the strong buying from extra-Euro area

investors and subdued purchases by insurance companies

and pension funds in 1H11.

Germany

Given the relatively large amount of local debt in

Germany (almost 40% of the total, between securities

and loans), we focus on central government securities. In

trying to estimate central government holdings, we faced

the problem that, based on different sources, there do not

seem to be enough German government bonds and T-

bills around to satisfy all the reported demand. Weestimate that domestic holdings of central government

securities are a small percentage of the total, and foreign

central banks constitute more than one third of the

investor base (Exhibit 41).

Germany’s role of an international safe haven is

confirmed by the 6% jump in foreign holdings of general

government debt in 2010 to 60%, with a further increase

in 1Q11. However, as discussed previously, we believe

that strong international participation, in particular

from overseas investors, makes German bonds

vulnerable to an escalation of the crisis. If the fabric of

the Euro zone is perceived to be under threat, even

German bonds might lose their safe-haven status and be

sold by foreign central banks.

France

Exhibit 42 shows our best estimate of the investor

holding of French government securities.8 Nearly two

7 ECB data shows changes due to outright buying/selling and to

revaluations/devaluations. 8

See Euro cash, Global Fixed Income Markets Weekly, 21 October

2011 for more details.

thirds of French government securities are held by

foreign investors, of which the major chunk is owned by

foreign central banks, based on our estimates (28% of the

total). We believe that, from a technical point of view,

France is fragile due to the large percentage of

foreign investors who are likely to be sensitive to its

AAA rating. Official data shows that foreign holdings

have been declining since mid-2010 after steady

increases since the introduction of the Euro.

Exhibit 41: We estimate that foreign central banks alone hold more thanone third of German bonds and T-bills, making even the German bondmarket vulnerable to a severe escalation of the sovereign crisis

J.P.Morgan estimate of German central government securities holdings; as of 1Q11; €bn

Domestic Foreign Total Domestic Foreign Total

MFIs 107 210 317 10% 19% 29%

Insurance companies 40 150 190 4% 14% 17%

Mutual funds 39 116 155 4% 11% 14%

Central banks 4 395 399 0% 36% 37%

Others 20 10 30 2% 1% 3%

Total 210 881 1091 19% 81% 100%

€bn %

Source: Bundesbank, ECB, BIS, EFAMA, IMF

Exhibit 42: We estimate that almost 30% of French bonds and T-billsare held by foreign central banks that are likely to be sensitive to a

rating downgradeJ.P.Morgan estimate of French central government securities holdings; as of 1H11; €bn


MFIs* 147 141 288 11% 11% 22%


Mutual funds 21 133 154 2% 10% 12%

Central banks n/a 363 363 n/a 28% 28%

Others 53 129 181 4% 10% 14%

Total 442 866 1308 34% 66% 100%

€bn %

* Domestic central bank might be under MFIsSource: AFT, ECB, BIS, EFAMA, IMF

Exhibit 43: We estimate that foreign banks are by far the largest foreign

players in Italian bonds, putting pressure on prices as they de-lever their portfoliosJ.P.Morgan estimate of Italian central government securities holdings; as of May 2011;

€bn

Domestic Foreign Total Domestic Fore ign Total

MFIs 188 189 377 13% 13% 25%


Mutual funds 64 77 141 4% 5% 9%

Central banks 67 66 133 4% 4% 9%

Others* 335 258 593 22% 17% 40%

Total 820 675 1494 55% 45% 100%

€bn %

* 5% of general government securities are held abroad, but is attributable to Italiansavers (4.3% of total debt).

Source: Bank of Italy, ECB, BIS, EFAMA, IMF




Gianluca SalfordAC

(44-20) 7325-4334


53

Italy

Banca d’Italia publishes lagged data on the split between

various domestic investors (MFIs, central banks, mutualfunds, insurance companies, pension funds, and others)

and foreign investors in government securities. They do

not provide the holding split among foreign investors,

who hold nearly 45% (€675bn, Exhibit 43) of

government securities as per the most recent data.

However, they do estimate that 5% of general

government securities held abroad are attributable to

Italian savers9.

We collect information from various sources to get an

idea of the foreign investor split in Italy. BIS data shows

that non-Italian banks have exposure of around €200bn

to the Italian general government, and we assume thataround 95% of that is in Italian government securities.

Euro area investment funds hold nearly €670bn of Euro

area government securities, and we estimate that

holdings of Italian bonds are around €140bn (20% of

total), based on Italy’s contribution to the JPM EMU

government bond index. IMF (CPIS) data shows that

foreign central banks held nearly €50bn of Italian

securities on their books at the end of 2009. Given that

the IMF survey captures only a small sub-sample of

reporting entities, and reserves have increased since

2009, we estimate that total numbers might be 30%

higher (around €65bn). Lastly, based on our insurance

companies’ equity analysts’ database, we estimate thatnearly €85bn of Italian government securities is held by

non-Italian insurance companies.

Flow-wise, available data on government bond securities

shows stable foreign holdings for the first 5 months of

the year, but timelier Balance of Payments data shows a

sharp decline in Italy’s foreign liabilities between July

and August (Exhibit 44). The €11bn increase in

domestic banks’ holding of government bonds in 3Q11

and aggressive purchases of Italian bonds through the

ECB SMP (we estimate €95bn of purchases of BTPs) are

also consistent with domestics and the ECB offsetting

large international selling. As discussed in the Overview,

we expect selling pressure from international investors to

continue in coming months.

9 In the latest Financial and Stability report, Banca d’Italia analysts

stated, “An estimated 4.3% (of total debt) is held by individually

managed portfolios and investment funds administered by foreignintermediaries but attributable to Italian savers.”

Spain

Tesoro Publico provides detailed domestic and foreign

investor holding data for government securities (Exhibit

45), but we use our own estimates for the split of foreign

investor types based on other sources.10 Spain is the

country with the lowest share of foreign investors (latest

data, 38%). Domestic investors, especially the domestic

banking sector, are the most important players in the

Spanish government bond market and based on availableinformation, they have continued to support the segment

in 2011. Domestic banks increased their exposure by

nearly 6%-pts, from around 28% at the end of 2010 to

33% at the end of August 2011. From a technical point

of view, Spain looks less vulnerable than Italy.

10The foreign split by investor type published by the Spanish Treasury

is based on non-resident withholding tax refunds.

Exhibit 44: Foreign investors have reduced exposure to Italian financialassets by €40bn over July-AugustCumulative changes in foreign holdings of Italian assets*; €bn

-20

0

20

40

60

80

100

120

Aug-09 Dec-09 Apr-10 Aug-10 Dec-10 Apr-11 Aug-11

* Italian government securities account for roughly 2/3rds of foreign liabilities.Source: Bank of Italy Balance of Payments data

Exhibit 45: Spanish banks are the key players in the Bono market,making Spain less vulnerable to selling pressure from internationalinvestorsJ.P.Morgan estimate of Spanish central government securities holdings; as of August2011; €bn


MFIs 175 51 227 33% 10% 42%


Mutual funds 23 45 67 4% 8% 13%

Central banks 63 73 136 12% 14% 25%

Others 32 11 43 6% 2% 8%

Total 328 205 534 62% 38% 100%

% €bn

Source: Bank of Spain, ECB, BIS, EFAMA, IMF




Gianluca SalfordAC

(44-20) 7325-4334


54

Bond supply: Expected to decline

Under the assumption that no other country loses market

access, we expect around €710bn of conventional

supply in 2012, around €20bn lower than in 2011

(Exhibit 46), whereas net conventional supply would

decline by almost €75bn due to higher redemptions. Our

economists forecast Euro area central government

deficits to decline by €50bn to around €275bn in 2012

(Exhibit 47), despite their view that non-core countries

will fail to hit the targets.

The most important conclusions on conventional

supply are:

1) In core countries, we expect gross supply to remainstable or decline marginally

2) We project Italian gross conventional issuance to

decline by €5bn to around €145bn. On the other hand, we

expect Spanish gross supply to remain stable as the

increase in redemption is offset by a decline in central

government deficit.

3) Even though Irish officials have expressed their

intention to return to the issuance market in 2012, we

believe this is unlikely and expect them to remain out of

the bond market, along with Greece and Portugal, in

2012.

Other supply

We estimate that floaters, inflation-linked, and zero-

coupon bonds add another €60-65bn to Italy’s supply;

the numbers for the other countries is significantly

smaller. Italy may also increase its T-bill issuance to

meet its funding needs. For the rest of the Euro area

countries, we forecast zero net T-bill issuance due to

governments’ reluctance to further shorten the maturity

of the debt after cutting the T-bill outstanding by nearly

€50bn in 2011. €-denominated, non-conventional net

issuance would likely remain close to zero, but we expecta decline in non-€-denominated issuance, given the

declining appetite of international investors.

Most countries will announce issuance guidelines in

December, and we will publish the Euro area 2012

supply update in early January.

Italy and Spain funding in 2012

Italian redemptions in 2012 are very heavy, with around

€200bn of marketable bonds maturing (Exhibit 48).

Redemptions concentration is particularly high between

February and April (around €90bn). Based on the last

official release, Italy had around €14.5bn in Treasury’s

payment account (conto disponibilità) at the end of

September 2011. We estimate the number to be in the

range of €40-50bn by the end of 2011, based on Italy’s

net issuance since the end of September and our estimate

of Italian deficit seasonal pattern. This amount alone will

not be sufficient to cover their funding requirements in

1H12, but see Overview Appendix-2 for other sources of

funding for Italy.

Exhibit 46: Government conventional bond supply will declinemarginally in 2012Gross conventional bond issuance history and J .P.Morgan 2012 forecast*; €bn

512

584

535 533506

487

562

820 839

733711

400

450

500

550

600

650

700

750

800

850900

2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012f

* Under the assumption that no other country loses market access.

Exhibit 47: 2012 net conventional supply will decline by around €75bndue to an increase in conventional bond redemptionsGross and net conventional bond issuance* and deficit for 2011 and J.P.Morganforecasts for 2012; €bn

Gross bond

issuance

Net bond

issuanceDeficit

Gross

bond

issuance

Net bond

issuanceDeficit

Austria 20 10 10 2 0 -1

Belgium 40 16 13 2 -3 -2

Finland 13 7 7 1 0 -1

France 180 95 82 -4 -10 -10

Germany 175 18 26 -6 -16 4Greece 0 -30 18 0 -8 2

Ireland 0 -6 14 0 -1 -2

Italy 145 35 37 -5 -27 -26

Netherlands 48 18 12 -5 -6 -7

Portugal 0 -10 9 -7 -8 -1

Spain 90 49 45 1 4 -3

Total 711 202 274 -22 -75 -47

2012 v s. 2011

* Gross conventional issuance is not equal to the sum of net conventional issuanceand deficit as a part of the deficit is met by non-conventional supply and other sourcesof funding.




Gianluca SalfordAC

(44-20) 7325-4334


55

In contrast, Spain has a very light redemption schedule in

1H12, giving it some time to front-load. Further, Spain’s

most liquid cash balances have remained in the €50-60bn

range since 2009, with the latest published balance of

€59bn at the end of 2Q11.

EFSF/EFSM supply

The EFSF and EFSM have issued around €45bn of bonds

so far in 2011 to fund the EU share of the Portuguese and

Irish aid packages (Exhibit 49). The bulk of the issuance

came from the EFSM as the EFSF remained out of the

markets for most of 2H11 as the new version of the EFSF

awaited parliamentary approvals. We expect the EFSM

to tap the market one last time before the year ends.

EFSF/EFSM funding requirements for 2012 to

support Ireland and Portugal are quite manageable,

at around €32bn, out of which €10bn would be covered

by the EFSM and the remaining €22bn would be

provided by the EFSF. However, the EFSF issuance

numbers may get revised significantly higher once thesecond bailout package for Greece gets finalised.

Given the EFSF’s several guarantee structure (no joint

guarantees) and significant uncertainty around Greece

PSI and the leveraged structures, EFSF bonds have

recently come under market pressure. EFSF funding

costs, which used to be around 5-10bp above LIBOR,

increased dramatically to 104bp when they last issued in

early November. In contrast, EFSM bonds, which have a

joint and several guarantee structure, have outperformed

their peers (Exhibit 50).

Trading themes

• Enter long 10Y duration trades in Germany

Outright recession in the region, falling core inflation

and a worsening of the sovereign crisis will push

German yields down in the first part of the year. We

target a trough in 10Y Bunds at 1.25% but the failure

to contain the crisis might eventually hurt even

German bonds, the last safe haven asset in the region.

Exhibit 48: Italy has a highly concentrated redemption schedule for 2012, while Spain has a light redemption schedule, which gives it anopportunity to front-loadMonthly redemption data for marketable bonds* issued by the Euro area countries;

€bn

A TS B EF F IM F RF D EM NL G GR D I EP I TL PTE ESP

Jan12 2 0 0 15 25 14 0 0 0 0 0

Feb12 0 0 0 0 0 0 0 0 36 0 1

Mar12 1 4 0 0 19 0 14 6 27 0 1

Apr12 0 0 0 18 16 0 0 0 28 0 12

May 12 0 0 0 0 0 0 9 0 1 0 0

Jun12 0 0 0 0 19 0 0 0 2 10 0

Jul12 10 0 0 28 27 15 0 0 17 1 13

Aug12 0 1 1 0 0 0 8 0 12 0 0

Sep12 0 12 6 12 21 0 0 0 11 0 2

Oct12 0 0 0 19 16 0 0 0 20 0 20

Nov 12 1 0 0 0 0 0 0 0 13 0 0Dec12 0 8 0 5 17 0 2 0 30 1 0

Total 15 25 7 98 160 30 33 6 198 13 49

Core Peripheral

* Marketable bonds include conventionals, floaters, zero-coupons, linkers, andinternational bonds.

Exhibit 49: The EFSF/EFSM has issued around €45bn of bonds so far i2011, with the bulk of the issuance coming from EFSMDetails of the EFSF and the EFSM issuances in 2011; EFSF supply highlighted ingrey; €bn

Date Issuer Maturity Size

Issuance level

(spread to mid-

swap; bp)

EMU beneficiary

05-Jan-11 EFSM 04-Dec-15 5.0 12 Ireland

25-Jan-11 EFSF 18-Jul-16 5.0 6 Ireland

17-Mar-11 EFSM 04-Apr-18 4.6* 8 Ireland

24-May-11 EFSM 04-Jun-21 4.8 14 Portugal/Ireland

25-May-11 EFSM 03-Jun-16 4.8 0 Portugal

15-Jun-11 EFSF 05-Jul-21 5.0 17 Portugal

22-Jun-11 EFSF 05-Dec-16 3.0 6 Portugal

14-Sep-11 EFSM 21-Sep-21 5.0 20 Portugal

22-Sep-11 EFSM 04-Sep-26 4.0 40 Ireland/Portugal

29-Sep-11 EFSM 04-Oct-18 1.1 15 Ireland/Portugal

07-Nov-11 EFSF 04-Feb-22 3.0 104 Ireland

Total EFSM 29.2

Total EFSF 16.0

Grand total 45.2 * Out of €4.6bn issued, €1.2bn was used for BoP and the remaining €3.4bn was givento Ireland.

Exhibit 50: EFSM bonds have outperformed their peers whereas EFSFbonds have been under market pressure, given the high level of uncertainty and the several but not joint guarantee structure

ASW spread for EFSF Jul16, EIB Jul16 and EFSM Dec15 bonds; bp

-20

0

20

40

60

80

100

May -11 J un-11 Jul-11 Sep-11 Oct-11 Nov -11

EIB Jul16

EFSF Jul16

EFSM Dec15




Gianluca SalfordAC

(44-20) 7325-4334


56

• Enter 2s/10s flatteners

The short end of the curve has little scope to rally

even if the ECB cuts the refi to 0.50%: we thereforerecommend 2s/10s flatteners, with a target of around

100bp by mid-year.

• Enter intra-EMU spread wideners

We forecast generalised spread widening vs. Germany

in 1H12.

• Enter flatteners in non-German curves

We also recommend credit curve flatteners: in line

with the experience of Greece, Ireland and Portugal,

in case of sovereign stress, we expect investors to

move from yield to price considerations. Technicals

point to flatter curves also in core countries such as

France and Netherlands, as we expect domesticsupport to the ultralong end of the curve.

• RV on peripherals: weighted flatteners and high-

coupon low coupon

We also recommend flatteners weighted by the

empirical price beta to exploit the convexity between

short-dated and long-dated bond prices. We present a

framework to analyse the yield spread of high- and

low-coupon bonds with similar maturity under

sovereign stress: We recommend selling Bonos 2032

vs. Bonos 2037 as the current yield spread is not

consistent with the coupon differential.




Pavan WadhwaAC

(44-20) 7777-3370

Fabio Bassi (44-20) 7325-8615

Khagendra Gupta (44-20) 7777-1980J.P. Morgan Securities Ltd

57

European Derivatives

• The EONIA curve was driven by ECB policy

rates and liquidity conditions in 2011

• We expect excess liquidity to remain elevated in

2012…

• …as peripheral banks continue to rely on the

ECB for funding, given limited market access and

large bank debt redemptions in 1H12

• The EONIA/refi bias is likely to remain around

70-75% of the corridor…

•

…and, with the ECB expected to cut refi anddeposit rates to 0.50% and 0.25% respectively,

we expect EONIA fixings to fall to 30bp

• We recommend longs in 6Mx6M EONIA with a

target of around 30bp…

• …and bullish option structures on Dec12 Euribor

• With peripheral sovereign stress expected to

escalate in 1H12, we expect the swap curve to bull

flatten…

• …and recommend carry efficient flatteners which

are a proxy for bullish positions but a with better

risk profile,…• …and receiver structures on 6Mx5Y swaptions

• We present a framework to analyse the

attractiveness of long-dated forward steepeners…

• …but refrain from recommending these trades

due to the prevalence of technical factors in this

part of the curve

• The 10s/30s swap curve will continue to be driven

by technical factors and is expected to flatten as

peripheral spreads widen

• Implied curve directionality has frequently

underestimated delivered directionality during

2011…

• …making conditional structures attractive at the

very front end of the curve; we recommend 1s/5s

bull flatteners

• Swap spreads will continue to be driven by

peripheral risk and are likely to eventually trade

wider than their Lehman peak

• We target 2Y and 10Y swap spreads at 145bp

and 90bp, respectively, by mid-2012

• Delivered directionality of swap spreads hashistorically exceeded implied directionality…

• …as options markets have failed to price in

greater delivered counter directionality in risk-off

environments,…

• …suggesting that conditional swap spread

wideners offer better risk/reward than outright

swap spread wideners

• Although spread markets are being driven

largely by a single-factor, i.e. sovereign risk,…

• …some spread markets such as OIS swap

spreads and 3s/6s basis are still trading far belowtheir local/Lehman peaks,…

• …suggesting that investors should express risk-

off views in these markets

• EUR volatility in the belly of the curve will

increase as the peripheral crisis escalates; we

target 3Mx10Y at 9.4bp/day, up from its current

level of 8bp/day

• Front-end volatility, however, will decline as the

ECB eventually goes on hold and excess liquidity

stays high…

• …and we recommend fading flare-ups in 2Y tail

volatility

• Favour long positions in Bund volatility vs.

swaption volatility

• On a cross-market basis, we prefer buying EUR

gamma to GBP gamma

EONIA curve

The EONIA curve in 2011 was driven primarily by

ECB activity and liquidity conditions. Early in the

year, EONIA yields drifted higher on declining excessliquidity and a hawkish ECB. The ECB hiked the refi

rate in April and July, and phased out some of the

extraordinary liquidity measures to reduce European

banks’ reliance on ECB funding. In the summer,

however, further escalation of the peripheral crisis,

declining inflationary pressures, and a deteriorating Euro

area macro outlook led the ECB to initially extend

liquidity measures and later cut refi by 25bp at its

November meeting (Exhibit 1).




Pavan WadhwaAC

(44-20) 7777-3370

Fabio Bassi (44-20) 7325-8615


58

The dynamic of excess liquidity in 1H11 was broadly in

line with our 2011 outlook 1 of declining excess liquidity.

Indeed, non-seasonally adjusted excess liquidity

averaged €40bn and €20bn in 1Q and 2Q, respectively,

close to our forecast of €50bn in 1Q and €20bn in 2Q.

Our outlook for a further decline in excess liquidity in

2H11 was severely challenged, however. At its August

meeting, the ECB, faced with increasing concerns

surrounding peripheral sovereign debt and slowing

economic growth, provided unlimited liquidity until

1Q12. At its October meeting, it extended liquidity

measures even further and re-introduced two new 1Y

LTROs, covering liquidity funding until the beginning of

2013.

As a result of the extension of these extraordinary

liquidity measures, excess liquidity increased to around

€200bn (this is the peak 1M average of non-seasonally

adjusted excess liquidity). The increase in excess

liquidity has widened the bias between EONIA fixings

and refi once again, broadly following the dynamic

captured in our non-linear model of the EONIA/refibias.2 However, segmentation in the EONIA market has

reduced the impact of excess liquidity on the EONIA/refi

bias in recent months.

Over the past couple of years, whenever excess liquidity

crossed €120bn, the EONIA/refi bias tended to widen to

about 85% of the corridor between refi and deposit

facility (which would currently be equivalent to -65bp).

1 See Global Fixed Income Markets 2011 Outlook , 26 November 2010.2 See Global Fixed Income Markets 2011 Outlook , 26 November 2010.

More recently, however, despite high excess liquidity,the EONIA/refi bias has averaged around -55bp. For

instance, the EONIA/refi bias averaged around -45bp in

the September maintenance period and around -55bp in

the October and November maintenance periods

(Exhibit 2). In our view, this decline in the EONIA/refi

bias is driven by segmentation in the EONIA market,

resulting from peripheral sovereign stress. Indeed, the

residual from our long-term model of EONIA/refi bias

increases (EONIA/refi bias becomes less negative) as

peripheral spreads widen (Exhibit 3). Going forward,

Exhibit 2: As excess liquidity increased to over €200bn, EONIAfixings declined but remained above the level implied by our modelof EONIA/refi bias…

EONIA/refi bias* regressed against non-seasonally adjusted excess liquidity; past2Y; bp

-80

-60

-40

-20

0

20

40

60

0 100 200 300 400

Oct & Nov11 Maint period

Excess liquidity (NSA); €bn

y = -9E-06x 3 + 0.006x 2 - 1.1x

R2 = 60%

* EONIA/refi bias defined as EONIA fixings – refi rate.

Exhibit 3: …likely because peripheral stress resulted in weaker banks finding it more difficult to access funding marketsResidual* from Exhibit 2 regressed against 10Y weighted peripheral spread**; past3M; bp

y = -0.0004x 2 + 0.47x - 142.1

R2 = 37%

-30

-20

-10

0

10

20

30

300 400 500 600 700 800

10Y weighted peripheral spread; bp

18-Nov -11

* Residual obtained by regressing EONIA/refi bias against non-seasonallyadjusted excess liquidity over the past 2Y.** Weighted peripheral spread computed against Germany for Greece, Ireland,Portugal, Italy, and Spain (weighted by the size of their outstanding bond markets).

Exhibit 1: EONIA yields rose in the first half of 2011 on the back of ECB rate hikes and declining excess liquidity. Yields have fallensince early August as the ECB cut rates and re-introduced

extraordinary liquidity measures3M EONIA and ECB refi rate; 1 January 2011 – 18 November 2011; %

0.40

0.60

0.80

1.00

1.20

1.40

1.60

04-Jan 08-Mar 11-May 13-Jul 15-Sep 18-Nov

3M EONIA

Refi rate




Pavan WadhwaAC

(44-20) 7777-3370

Fabio Bassi (44-20) 7325-8615


59

we expect this bias to stay at 70-75% of the deposit/refi

corridor (currently equivalent to -55bp) in 2012.

We believe excess liquidity will remain elevated in

2012. Peripheral banks have funded an increasingly

higher proportion of their assets at the ECB over the past

few months, and we do not expect this to change next

year. On the contrary, we expect an escalation of the

peripheral crisis, which will make wholesale funding

even more difficult for European banks, compelling them

to further increase their reliance on the ECB for funding.

The borrowing needs of peripheral European banks

for 2012 are quite heavy as there are significant

senior unsecured redemptions, especially in 1H12

(Exhibit 4). Absent a full re-opening of the unsecured bond market, we expect that most of these redemptions

will be covered by the ECB’s MRO/LTRO, and only

partially by alternative sources of funding such as retail

bonds, private placement and covered bonds.

The last 1Y tender will take place in December. With

about €140bn expiring at the 3M tender, we expect

€75bn to be rolled and about €100-125bn to be tendered

at the 1Y (13M tender), which will extend over both

2011 and 2012 year-ends. Overall, we expect an increase

in excess liquidity from these tenders of about €30-60bn

and a significant increase in the average duration of the

existing tenders.

Where does that leave the ECB in terms of policy rate?

We expect the ECB to cut the refi rate to 1.00% at the

December meeting and then to cut it a further 25bp

each quarter, down to 0.50% by June 2012, while

leaving the deposit facility rate at 0.25% after the

December cut. In that scenario, we expect EONIA to

trade as low as 32bp, with the EONIA/refi bias at 70-

75% of the deposit/refi corridor. Therefore, we see value

in long positions in mid-to-late 2012 such as 6Mx6M

EONIA (Exhibit 5). Although there is a risk that the

ECB will not cut rates below 1%, we believe long

positions at the front end are likely to offer an option-like

profile, with significant downside only if rates are left

unchanged at 1.25%.

Recent widening of the FRA/OIS basis at the front-end

of the curve has worsened the carry in long Euribor

positions, which is now negative for the first year of the

Euribor curve. Additionally, we expect funding

pressures to increase, driving FRA/OIS close to its

Lehman highs by mid-2012, before it starts to decline in

2H12. From both a carry and FRA/OIS perspective, we

see value in bullish structures only further out theEuribor curve.

In Exhibit 6, we plot our projected level of EONIA,

under the assumption that the refi rate will be cut to

0.50% by June, and the projected Euribor levels obtained

by adding to the projected EONIA curve the maximum

and minimum levels of the FRA/OIS basis since

Exhibit 4: With bank funding markets unlikely to open fully inearly 2012, we expect a reasonable portion of European banks’borrowing needs to be funded by the ECB

Cumulative total redemptions of senior unsecured bonds for peripheral*country banks in 2012; €bn

59

114

135

153

40

60

80

100

120

140

160

1Q12 2Q12 3Q12 4Q12

* Greece, Ireland, Italy, Portugal, and Spain.Source: Dealogic

Exhibit 5: We are bullish on front-end EONIA as ECB refi anddeposit rate cuts are expected to drive EONIA yields into the low30s; receive 6Mx6M forward EONIAProjected upside from outright longs across the EONIA curve; %

Market

Date ECB OIS Refi Depo EOIN A/ refi bias ECB OIS

Dec11 0.55 1.00 0.25 -0.55 0.45 0.10

Jan12 0.47 1.00 0.25 -0.55 0.45 0.02

Feb12 0.45 1.00 0.25 -0.55 0.45 0.00

Mar12 0.48 0.75 0.25 -0.37 0.38 0.10

Apr12 0.47 0.75 0.25 -0.37 0.38 0.08

May 12 0.47 0.75 0.25 -0.37 0.38 0.08

Jun12 0.51 0.50 0.25 -0.18 0.32 0.19

Jul12 0.50 0.50 0.25 -0.18 0.32 0.18

Aug12 0.51 0.50 0.25 -0.18 0.32 0.19

Sep12 0.53 0.50 0.25 -0.18 0.32 0.22

Oct12 0.55 0.50 0.25 -0.18 0.32 0.24

Nov 12 0.57 0.50 0.25 -0.18 0.32 0.26

Dec12 0.61 0.50 0.25 -0.18 0.32 0.29

J.P.Morgan Upside from

longs




Pavan WadhwaAC

(44-20) 7777-3370

Fabio Bassi (44-20) 7325-8615


60

1 January 2009.3 We then show the current and

projected Euribor curves consistent with our FRA/OIS

assumptions. We draw the following conclusions from

this analysis. First, although the FRA/OIS is trading

close to its widest levels since 1 January 2009, we do not

see value in outright longs at the very front end of the

curve, as we expect FRA/OIS to widen further. Second,

Euribor fixings are expected to decline towards the end

of 2012, with our forecast for Dec12 Euribor at 1.00%.

Third, we see more value in bullish option structures

than outright Euribor longs.

We analyse the projected P&L of various bullish option

structures on Dec12 Euribor under various yield

projections in Exhibit 7. We draw the followingconclusions. First, call spreads are relatively expensive

and offer only modest upside in our central scenario of

Dec12 Euribor at 1%. Second, symmetric structures

such as call flies are relatively cheap and offer upside for

a wide range of Euribor fixings (0.785% to 1.215%).

Third, 1x2s are the most attractive structures as they can

be implemented at a credit and are expected to be

profitable as long as Dec12 Euribor remains above

0.68%, or a mere 3bp above the minimum Euribor fixing

level reached between June 2009 and March 2011. We

therefore recommend 1x2s on Dec12 Euribor.

Swap curve

The swap curve was driven by ECB activity and the

peripheral crisis in 2011. The ECB hiked the refi rate

in April and July, driving the curve flatter in a selloff.

Monetary policy turned at the August meeting when it

3 The period since 1 January 2009 does not include the FRA/OIS peak

witnessed during the Lehman crisis, when front FRA/OIS widened to150bp.

became clear that the ECB had gone on-hold. The 2s/10s

swap curve steepened aggressively as the market started

pricing in ECB easing. Since then, it has mostly bull–

flattened (Exhibit 8) going into the ECB cut in

November. The 10s/30s swap curve remained

uncorrelated with front-end yields, driven mostly by

technical factors. Since the beginning of the year, the

2s/10s swap curve flattened 55bp, about 35bp more

than was implied in the forward curve early in the

year, whereas 10s/30s flattened only 8bp, 12bp less

than the forward.

Exhibit 6: We expect bank funding pressures to increase,driving the FRA/OIS basis close to its Lehman highs in 1H12before declining in 2H12. On balance, we expect December 2012Euribor to trade below its current level…1) Projected EONIA levels under the assumption that the refi rate is cut to50bp, 2 ) min/max Euribor threshold based on the min/max* FRA/OIS since1 January 2009, and 3) 18 November 2011 vs. forecast y ield on front Euribor strip; %

0.20

0.40

0.60

0.80

1.00

1.20

1.40

1.60

Dec-11 Mar-12 Jun-12 Sep-12 Dec-12

Projected EONIA

EONIA + Min FRA/OIS

EONIA + Max F RA/OIS

Euribor 18 Nov

Euribor forecast

* Min/max since 1 January 2009. See footnote 3 further discussion.

Exhibit 7: …and therefore favour bullish option structures in December 2012 Euribor Upfront cost and projected P&L under different Euribor projections for various bullish option structures on Dec12 Euribor futures; bp

Projected P&L (bp) at m aturity when Euribor is

70 80 90 100 110 120 130 65* 130**

Call spread 98.750/99.000 +1/-1 14.0 11.0 11.0 11.0 11.0 1.0 -9.0 -14.0 11.0 -14.0

Call spread 98.750/99.250 +1/-1 24.5 25.5 20.5 10.5 0.5 -9.5 -19.5 -24.5 25.5 -24.51x 2 98.750/99.000 +1/-2 -7.0 2.0 12.0 22.0 32.0 22.0 12.0 7.0 -3.0 7.0

1x 2 98.750/99.250 +1/-2 14.0 31.0 31.0 21.0 11.0 1.0 -9.0 -14.0 26.0 -14.0

Ladder 98.750/99.000/99.250 +1/-1/-1 3.5 16.5 21.5 21.5 21.5 11.5 1.5 -3.5 11.5 -3.5

Fly 98.750/99.000/99.250 +1/-2/+1 3.5 -3.5 1.5 11.5 21.5 11.5 1.5 -3.5 -3.5 -3.5

Condor 98.750/99.000/99.250/99.500 +1/-1/-1/+1 7.5 12.5 17.5 17.5 17.5 7.5 -2.5 -7.5 7.5 -7.5

Dec 2012 Euribor futures; Forward = 126bp;

Cost (bp)Ty pe Strike Contracts

* Minimum Euribor fixing between June 2009 to March 2011, a period when the refi rate was 1%.** Maximum Euribor fixing between June 2009 to March 2011, a per iod when the refi rate was 1%.




Pavan WadhwaAC

(44-20) 7777-3370

Fabio Bassi (44-20) 7325-8615


61

A notable characteristic of the swap curve during

2011 has been frequent changes in curve

directionality observed on the back of central bank

activity and the peripheral crisis. The change in swap

curve directionality may be seen in the evolution of PCA

factor loadings (Exhibit 9). Early in the year, when the

market was expecting the ECB to remain on hold, the

curve was driven by the intermediate sector, flattening in

rallies and steepening in selloffs. This resulted in higher first-factor loading on 10Y yields relative to 2Y yields.

This dynamic changed in 2Q and 3Q as the ECB started

to hike rates and the curve became driven by the front

end. Consequently, the first factor loading on 2Y yields

increased relative to 10Y. This directionality has flipped

once again over the past three months, when

intermediates started driving the curve.

Going forward, we expect the ECB to eventually cut the

refi rate to 0.50% in 1H12, while keeping the deposit

facility rate at 0.25%. With the market already pricing

much of the expected easing (see EONIA curve), we

expect the front end to remain tethered and the curveto exhibit positive directionality in 2012, flattening in

a rally and steepening in a selloff .

2s/10s flatteners have been profitable on average during

2011 and their P&L performance has been correlated

with the carry at inception. Exhibit 10 shows the

success ratio and the average 3M P&L on 2s/10s

flatteners initiated at different levels of ex ante carry

during 2011. The analysis shows that flatteners have

performed better when initiated at positive carry, and

Exhibit 8: A year of flattening: the swap curve bear flattenedearly in the year on the back of ECB rate hikes, and bull flattenedlater due to escalation of the peripheral crisis

2Y swap yields and 2s/10s EUR swap curve; 1 January 2011 – 18 November 2011;% bp

1.20

1.40

1.60

1.80

2.00

2.20

2.40

2.60

Jan 11 Mar 11 May 11 Jul 11 Sep 11 Nov 11

80

100

120

140

160

1802Y sw ap rate

2s/10s swap curve

Exhibit 9: Changes in swap curve directionality may be seen in theevolution of PCA factor loadings. Going forward, with the front endexpected to remain tethered, we expect the swap curve to continueto flatten in a rally and steepen in a selloff 3M rolling 2Y and 10Y first factor loading obtained from PCA*; %

0%

10%

20%

30%

40%

50%


10Y

2YBear

flattening

Bull

flattening

* Principal component analysis (PCA) is a tool that recombines observed variablesinto factors such that a smaller number of these factors explain a large percentageof variation observed in the original data set. We run PCA on demeaned levels of 1Y, 2Y, 3Y, 5Y, 7Y, 10Y, 12Y, 15Y, 20Y, 25Y, and 30Y swap yields.

Exhibit 10: Over the past year, flatteners have been more profitablewhen initiated at high levels of carry; by this metric, flattenerscurrently appear attractiveSuccess ratio and statistics on P&L of 2s/10s flatteners held for three monthsrelative to the 3M carry at inception*; 1 January 2011 – 18 November 2011; bp

3M Carry <-8bp >-8bp and <0bp >8bp

# trades 88 48 25

# profitable trades 34 32 25

Success Ratio 39% 67% 100%

Av erage -3 7 29

Min -38 -28 8

Max 24 30 53

SD 17 17 15

* We split the sample of trades into three categories based on the top, the central,and the bottom third of carry level at trade inception.

Exhibit 11: Further escalation of the peripheral crisis is expected toput flattening pressure on the swap curve…2s/10s EUR swap curve regressed against 10Y weighted peripheral spreads*;1 January 2011 – 18 November 2011; bp

y = -0.08x + 156.5

R2 = 50%

80

100

120

140

160

180

200 300 400 500 600 700 800

18-Nov-11


* Weighted peripheral spread computed against Germany for Greece,Ireland, Portugal, Italy, and Spain (weighted by the size of their outstanding

bond markets).




Pavan WadhwaAC

(44-20) 7777-3370

Fabio Bassi (44-20) 7325-8615


62

have averaged negative returns if implemented when the

carry was in the bottom third of the sample.

Another significant driver of the curve has been the

peripheral sovereign crisis. We believe that this crisis

will likely worsen in 1H12 (see Overview and Euro

Cash), resulting in wider peripheral spreads. This

will likely put flattening pressure on the Euro swap

curve (Exhibit 11). Empirical analysis over the past

year indicates that the 2s/10s swap curve tends to flatten

about 8bp for every 100bp of widening in 10Y weighted

peripheral spreads. We expect about 250bp of widening

in 10Y weighted peripheral spreads (see Euro Cash),

which will put around 25bp of flattening pressure on the

2s/10s swap curve. Additionally, 3M carry on 2s/10sflatteners is +5bp; we therefore recommend that

investors implement swap curve flatteners.

Next, we further analyse carry in the swap curve to find

the most attractive sectors to implement flatteners. Since

early 2011, carry in long positions at the front end of the

curve has flipped from positive to negative as the

fronts/reds curve has inverted. This has turned carry in

flattening trades positive, especially when anchored at

the very front end of the curve (Exhibit 12). On a

carry-to-risk basis, we believe 1s/3s and 1s/5s are the

most attractive flatteners on the swap curve. These

trades have a strong bullish bias and sport a better risk profile than outright longs.

Investors wishing to express a bullish view on the

intermediate sector may also consider call structures.

With our 5Y German benchmark yield target at 0.65%

(see Euro Cash), we see value in receiver structures on

5Y swaps despite our view that 5Y swap spreads will

widen to 130bp (see Swap spreads below). In Exhibit

13, we analyse the projected upside for various bullish

structures implemented with 6M receivers. Among

them, we prefer symmetric structures such as receiver

butterflies or condors, given their limited downside.

However, investors less concerned about 5Y swap rates

declining below 1.50%-1.60% may consider asymmetric

Exhibit 12: …and investors wishing to position for this shouldconsider flatteners that offer attractive risk-adjusted carry…

Risk-adjusted carry in outright long positions in swaps and swap curve flatteners;bp

18-Nov-11 18-Nov-11 30-Jun-11 04-Jan-11

1Y 1.62 -8 14 9 43 -0.7

2Y 1.60 -1 12 12 71 -0.1

3Y 1.71 4 13 14 82 0.2

5Y 2.09 6 11 12 93 0.3

10Y 2.68 4 7 8 102 0.2

15Y 2.96 3 6 6 107 0.1

30Y 2.82 1 3 2 116 0.0

1Y/2Y -2 7 -2 4 32 0.9

1Y/3Y 10 12 -2 5 43 1.1

1Y/5Y 47 14 -4 3 59 0.9

2Y/5Y 49 7 -1 0 39 0.7

2Y/10Y 108 5 -5 -4 60 0.4

2Y/30Y 122 2 -9 -10 78 0.1

5Y/10Y 59 -2 -3 -4 29 -0.2

Current ann. risk

adj. carry**

3M Carry (bp)Trade

Ann Risk*

(bp)

* Annualised risk is defined as sqrt(251)*3M standard deviation of daily changes.** Annualised risk-adjusted carry is defined as 4*3M Carry/Annualised risk.

Exhibit 13: …or bullish option structures on intermediate tailsCost and projected payoff of various bullish option structures on 5Y EUR swap rates; bp

Ty pe Strike Contracts Low er Upper Upside Dow nside

Call spread ATM/ATM-1C +1/-1 6.1 13.2 - 216.1 7.1 6.1

Call spread ATM/ATM-2C +1/-1 11.2 13.2 - 211.0 15.2 11.2

1x 2 ATM/ATM-1C +1/-2 -16.4 13.2 179.3 238.5 29.6 -1x 2 ATM/ATM-2C +1/-2 -6.2 13.2 163.2 228.3 32.6 -

Ladder ATM/ATM-1C/ATM-2C +1/-1/-1 -11.3 13.2 171.3 233.4 24.5 -



Ladder ATM/ATM-2C/ATM-4C +1/-1/-1 1.4 13.2 144.3 220.7 25.0 -

Butterfly ATM/ATM-1C/ATM-2C +1/-2/+1 1.0 13.2 196.7 221.2 12.2 1.0

Butterfly ATM/ATM-2C/ATM-4C +1/-2/+1 3.6 13.2 172.9 218.5 22.8 3.6

Condor ATM/ATM-1C/ATM-2C/ATM-3C +1/-1/-1/+1 1.9 13.2 184.4 220.2 11.3 1.9

Condor ATM/ATM-1C/ATM-3C/ATM-4C +1/-1/-1/+1 2.7 13.2 172.0 219.4 10.5 2.7

Max P&L (bp)

6Mx5Y Swaptions; Carry = 13.2bp; ATMF = 222.1bp; Spot = 208.9bp; Spot 1Y Min = 184.1bp

Yield bounds (bp)**Cost (bp)

Projected payoff

at maturity (bp)*

Note: ATM-1C refers to the swaption strike that is 1-carry (13bp) away from ATM swap rate.* As the forward rate is expected to slide to spot rate, the projected payoff at maturity is equal to the current slide.** Yield levels above and below which the trade becomes unprofitable.




Pavan WadhwaAC

(44-20) 7777-3370

Fabio Bassi (44-20) 7325-8615


63

structures such as receiver ladders, which may be

implemented at a credit.

Another set of trades that have captured investor interest

over the past year has been long-dated forward

steepeners on the swap curve. We analyse the

attractiveness of such trades by building a ‘carry

efficient frontier’ of steepener trades that: 1) offer

high risk-adjusted carry, and 2) are trading flat from

a historical perspective (Exhibit 14). In addition, given

our view that the spot 2s/10s curve will flatten, we

eliminate trades that have large positive correlation to the

2s/10s curve. The best trades are in the top right

quadrant of Exhibit 14 which shows forward curves that

are relatively flat and offer high risk-adjusted carry.

Unfortunately, however, we found that most trades havetheir long leg anchored in the 20Y+ area of the curve.

Since long dated forward curves are positively correlated

with 10s/30s, which is itself driven by technical factors

(see below), we remain cautious on long-dated

forward steepeners and refrain from recommending

trades in this sector.

Long-dated swap curve

The 10s/30s swap curve has been in a 25bp range since

the beginning of the year and is now trading close to the

bottom of the range. Various technical factors have, over

the past five years, overwhelmed the typical relationshipof the 10s/30s curve with front end rates. For example,

in mid- and late-2008, the flattening in 10s/30s was

driven by the hedging of non-inversion notes from the

dealer community.4 Additionally, hedging flows from

pension funds have frequently been an important driver

of the 10s/30s curve. Over the past couple of years,

however, 1-way CSA hedging has dominated long end

flows due to the peripheral debt crisis.5

We develop a relative value model that captures the

sensitivity of the 10s/30s curve to both macro and

technical factors using the following variables: 1) 10Y

weighted peripheral spread, 2) 2Y swap rate, 3) pension

fund coverage ratio, and 4) dummy variables for hedging

of non-inversion notes in mid-2008 and early 2009

(Exhibit 15). The betas of the regression model are

significant and have the correct signs. The model

indicates that 10s/30s tends to flatten:

4 See Global Fixed Income Markets Weekly, 25 April 2008 and 06 June

2008.5 See Global Fixed Income Markets Weekly, 30 July 2010.

Exhibit 14: Although flatteners appear attractive on the spotcurve, steepeners appear attractive on the forward curve based onvaluations and carry, especially when they have little correlation

with the spot 2s/10s curve; however, we remain cautious…Risk-adjusted carry* of various forward steepeners** vs. its 5Y z-score of relativesteepness*** (reverse axis); unitless

0.0

0.2

0.4

0.6

0.8

1.0

-1.5-1.0-0.50.00.5

3Yx5s/20s2Yx5s/20s

2Yx10s/20s

3Yx10s/20s

4Yx5s/20s

5Yx5s/20s

5Yx10s/ 20s

Relative steepness; z-score

* Risk-adjusted carry defined as 1Y carry divided by 2*SD of quarterly changeson the forward curve.** Black dots indicate curve segments with correlation lower than 20% to 2s/10s;grey dots indicate curve segments with correlation higher than 20% to 2s/10s.*** 5Y z-score; negative numbers (right side of X-axis) indicate a flat curve.We include 1Y to 5Y forward steepeners on 5s/10s, 5s/20s, 5s/30s, 10s/20s,and 10s/30s.

Exhibit 15: …as technical factors such as 1-way CSA hedgingmay drive the 10s/30s swap curve flatter if the crisis gets worse10s/30s EUR swap curve regressed against 1) 10Y weighted peripheral spread*,2 ) 2Y EUR swap yield, 3) pension fund coverage ratio, and 4) two dummy

variables for the 31 May 2008 to 15 August 2008 and the 15 September 2008 to15 January 2009 periods; past 5Y; bp

y = -0.03*(Peri sprd)-13.0*(2Y)+42.6*(PF ratio)

- 6.8(I1) - 28.3*(I2)+15.7

R2 = 74%

-40

-20

0

20

40

60

0 200 400 600 80010Y we ighted peripheral spread; bp

18-Nov -11

* Weighted peripheral spread computed against Germany for Ireland, Portugal,Italy, Spain and Greece (weighted by the size of their outstanding bondmarkets).




Pavan WadhwaAC

(44-20) 7777-3370

Fabio Bassi (44-20) 7325-8615


64

• 3bp for every 100bp widening in the 10Y weighted

peripheral spread,

• 13bp for every 1% increase in 2Y yields,

• 4.2bp for every 10% decline in pension fund

coverage ratio, and

• 7bp and 28bp in the two episodes of hedging of non-

inversion notes of mid-2008 and early 2009,

respectively.

Anecdotal evidence suggests pension fund and exotic

desk hedging has declined recently, leaving 1-way CSA

hedging as the main driver of this segment of the curve.

With peripheral spreads expected to widen significantly

from current levels, we see risk of further flattening in

10s/30s. Specifically, we expect the 10Y weighted

peripheral spread to widen 250bp, putting around 6-

8bp of flattening pressure on the 10s/30s swap curve.

Indeed, the correlation between weighted peripheral

spreads and 30Y swap yields has become more negative

in 2H11, as the peripheral crisis has spread to Italy(Exhibit 16). Thus, our negative view on peripherals

leads us to a 10s/30s swap curve flattening bias.

1-way CSA hedging on the swap curve has richened the

30Y sector. Indeed, 30Y is trading rich on a level neutral

10s/30s/50s fly. However, historical analysis indicates

that 30Y richness is a function of peripheral stress, with

the residual of the 10s/30s/50s level neutral fly being

strongly correlated with peripheral spreads (Exhibit 17).

Given our negative view on the crisis, we believe that the

30Y sector could richen further and we recommend that

investors do not fade the move.

Implied and delivered curve directionality

With the ECB unlikely to cut rates below 0.50%, we

believe the curve will remain driven by the intermediate

sector, flattening in rallies and steepening in selloffs.

During 2011, this directionality was not fully priced in

the curve and offered several opportunities for

Exhibit 17: …since we believe that 30Y swap yields could richenfurther if peripheral sovereign spreads continue wideningResidual of 10s/30s/50s EUR swap fly regressed against 30Y EUR swap yields

vs. 10Y weighted peripheral spreads*; past 2Y; bp

y = -0.02x + 6.26

R2 = 54%

-15

-10

-5

0

5

10

0 200 400 600 800


18-Nov-11

* See Exhibit 15 for definition of weighted peripheral spread.

Exhibit 18: Implied curve directionality has frequentlyunderestimated delivered directionality in 1s/10s, makingconditional structures attractive in the past…Implied* and 3M forward looking delivered directionality** of 1s/10s EUR swapcurve; %

0%

20%

40%

60%

80%

100%

Jan 10 Apr 10 Jul 10 Oct 10 Jan 11 Apr 11 Jul 11 Nov 11

Implied direc tionality

Delivered directionality

* Implied directionality for swap curve calculated as ratio of implied volatility of 3MxLong leg/3MxShort leg -1.** Delivered directionality for swap curve calculated as 1 - 3M beta of dailychanges in Short leg regressed against daily changes in Long leg.

Exhibit 16: With the crisis spreading to Italy, the correlationbetween peripheral spreads and 30Y swap yields has becomemore negative on the back of technical flows such as 1-way CSA

hedging. We have a 10s/30s swap curve flattening bias…2M correlation of daily changes in 10Y weighted peripheral spreads to dailychanges in 30Y EUR swap yields; %

-100%

-80%

-60%

-40%

-20%

0%

J an 10 Apr 10 J ul 10 Oc t 10 J an 11 Apr 11 J ul 11 N ov 11




Pavan WadhwaAC

(44-20) 7777-3370

Fabio Bassi (44-20) 7325-8615


65

conditional curve trades. For example, implied

directionality in the 1s/10s swap curve has been, until

recently, in a 20-50% range whereas delivered

directionality has rarely fallen below 50% (Exhibit 18).

Moreover, since delivered directionality is autocorrelated

(Exhibit 19), investors may use historical delivereddirectionality as a measure of future expected delivered

directionality, thereby providing them with a way to

judge the attractiveness of conditional curve trades at any

point in time.

For example, implied directionality is significantly below

delivered directionality for curve trades anchored at the

very front end (1Y), making such conditional curve

trades attractive. However, trades anchored further out

(2Y+) are unattractive due to recent richening of implied

directionality in such structures (Exhibit 20). Therefore,

we favour bull flatteners in 1s/2s and 1s/5s, while

avoiding other structures.

Swap spreads

Like most asset classes, swap spreads were held hostage

to market risk aversion stemming from the burgeoning

peripheral debt crisis (Exhibit 21). After staying in a

narrow 20-35bp range in the first half of the year, 10Y

swap spreads embarked on a widening trend in the

second half in response to a sharp deterioration in risk

markets, reaching a peak of 75bp. Subsequently, spreads

declined below 50bp in October as equity markets

rallied, before widening sharply in response to the threat

of a Greek referendum and significant political/market

stress in Italy. Since the end of last year, 2Y, 10Y and30Y swap spreads have widened 40, 45 and 15bp,

respectively.

Historically, we have used 5 factors to explain 10Y swap

spread movements:

• German government bond issuance,

• Amount of liquidity in the banking system, proxied

by EONIA rates,

• Directionality to 10Y Bund yields,

Exhibit 20: Implied directionality is below delivered directionalityfor curve trades anchored at the very front end, makingconditional curve trades attractive

Current implied and delivered directionality* for various curve trades; %Trades Implied Deliv ered Implied-Deliv ered

1s/2s 1% 42% -41%

1s/5s 21% 59% -38%

1s/10s 47% 67% -19%

1s/30s 65% 73% -8%

2s/5s 20% 28% -8%

2s/10s 46% 41% 5%

2s/30s 63% 51% 12%

5s/10s 22% 12% 10%

5s/30s 36% 26% 9%

10s/30s 12% 14% -2%

* See Exhibit 18 for definition of implied and delivered directionality.

Exhibit 21: Although the primary driver of swap spreads in 2011was risk aversion,*…10Y German b/m swap spreads vs. DJ Euro Stoxx 600 index (inverted); since1 January 2011 – 18 November 2011;bp points (inverted axis)

20

30

40

50

60

70

80


200

220

240

260

280

300

10Y ASW

Euro Stoxx 600(inverted axis)

* The correlation between DJ Euro Stoxx 600 index and 10Y swap spreads sincethe beginning of the year is around 95%.

Exhibit 19: …especially since delivered directionality isautocorrelatedSpread of ex post 3M delivered and implied 1s/10s curve directionality*

regressed against the spread of ex ante 3M delivered and implied 1s/10scurve directionality; past 2Y; %

y = 0.76x + 0.08

R2 = 57%

-20%

0%

20%

40%

60%

80%

-20% 0% 20% 40% 60% 80%3M lagged (delivered - implied) directionality ; %

18-Nov-11

* See Exhibit 18 for definition of implied and delivered directionality.Y variable: Delivered directionality at time t minus implied directionality at timet – 3M.X variable: Delivered directionality at time t – 3M minus implied directionality attime t – 3M.




Pavan WadhwaAC

(44-20) 7777-3370

Fabio Bassi (44-20) 7325-8615


66

• Risk aversion, proxied by swaption implied volatility,

and

• Swapped issuance.

Much of the widening this year, however, may be

explained by sovereign spreads. Indeed, a regression of

10Y swap spreads against 10Y weighted peripheral

spreads over the past year shows an R-squared of 80%,

up from just over 40% in 2010. This suggests that, much

like every other asset class, swap spreads are living in a

single-risk factor world.

Although swap spreads appear to be driven solely by

sovereign risk, in actuality it is impossible to distinguish

between this (relatively) new driver of swap spreads

versus traditional drivers. This is because sovereign risk is highly correlated with typical high frequency drivers

of swap spreads. Indeed, as peripheral spreads have

widened, Bund yields have fallen, implied volatility has

shot up, and swapped issuance has plummeted. Since the

beginning of the year, the correlation between peripheral

spreads and these three drivers is 70-90% (Exhibit 22).

Given these high correlations, the change in 10Y swap

spreads since end-2010 can be explained reasonably well

with traditional drivers. Between December 2010 and

November 2011, for example, 10Y German b/m swap

spreads (averaged over a 1M period) widened 26bp, of

which 25bp can be explained by high frequency factors(Exhibit 23).

Going forward, we expect swap spreads to continue to

be driven by the peripheral debt crisis which is likely

to get worse before it gets better (see Overview). Over

the past 2 years, 10Y swap spreads have widened an

average of 9bp for each 100bp widening in 10Y weighted

peripheral spreads (Exhibit 24). Given our expectation

that 10Y weighted peripheral spreads are likely to widen

from their current level of 700bp to a peak of 955bp in

1H12, we believe that 10Y swap spreads will widen to a

Exhibit 22: …high cross correlations make it is difficult todisentangle the effect on swap spreads of peripheral sovereign

spreads versus more traditional drivers…Correlations between traditional high frequency drivers of swap spreads and10Y weighted peripheral spreads; data since 1 January 2011; %

10Y swap

spreads

10Y Bund

yield

3Mx10Y

implied

v ol

Swapped

issuance*

10Y

peripheral

spread**

10Y swap spreads 100%

10Y Bund y ield -95% 100%

3Mx 10Y implied v ol 91% -96% 100%

Sw apped issuance* -77% 72% -71% 100%

10Y peripheral spread** 90% -93% 86% -68% 100%

* We use fixed-rate, €-denominated issuance by financial institutions (includingcovered bonds) and supras/agencies, plus one-half of corporate bond issuance,as an indicator of potential swapped issuance.

** Weighted peripheral spread computed against Germany for Greece, Ireland,Portugal, Italy, and Spain (weighted by the size of their outstanding bondmarkets).

Exhibit 23: …and, indeed, much of the swap spread widening in 2011 may be explained by lower Bund yields and higher implied volatility,which themselves reflected an escalation of the peripheral debt crisisReturn attribution of 10Y German b/m swap spreads between 30 December 2010 and 18 November 2011*; bp

20-day MA as of:

18-Nov -11 30-Dec-10 Chg Beta Ex p. Chg

10Y sw ap spread; bp 62 36 26

Low freq regression; 1 Jan 00 - 1 Jan 08

IMM1 EONIA; % 0.56 0.73 -0.17 11.0 -2

1Y sum of German govie issuance**; €bn 170 194 -24 -0.12 3

Exp. chg from low freq factors 1

High freq regression; 1 Jan 08 - Present

Yields (10Y Bund); % 1.91 2.97 -1.07 -19.0 20

3Mx10Y sw aption v olatility ; bp/day 7.5 6.0 1.5 4.7 7

20-day MA of sw apped issuance*** (€bn/day ) 1.3 0.5 0.8 -2.6 -2

Exp. chg from high freq factors; bp 25

Exp. chg from all factors; bp 26

* J.P.Morgan’s low frequency model regresses 10Y swap spreads against first IMM date EONIA and gross annual issuance of German government bonds over the 8Yperiod ending 1 January 2008. Residual from this low-frequency regression is computed over the period 1 January 2008 – present and regressed against the three highfrequency factors.** We use an exponential model of govie issuance as an explanatory variable because the sensitivity of swap spreads declines with increasing govie issuance. For simplicity, however, we show the overall (linear) sensitivity of spreads to issuance in the table, at current issuance levels.*** We use fixed-rate, €-denominated issuance by financial institutions (including covered bonds) and supras/agencies, plus one-half of corporate bond issuance, as anindicator of potential swapped issuance.




Pavan WadhwaAC

(44-20) 7777-3370

Fabio Bassi (44-20) 7325-8615


67

peak of 90bp. Thereafter, as the peripheral debt crisis

starts to wane, we expect 10Y swap spreads to narrow to

around 80bp by end-2012.

We also expect 2Y, 5Y and 30Y swap spreads to widen

concurrently with 10Y spreads, in line with their

historical pattern. For example, over the past two years,

2Y spreads have widened around 60bp while 10Y

spreads have widened 40bp. This suggests a ratio of 1.5

between 2Y and 10Y spreads. We use this ratio to

project 2Y spreads. Exhibit 25 shows our swap spread

forecast. Given that we believe the sovereign debt crisis

will continue to worsen in 1H12, swap spreads across the

curve are likely to widen, with the front end being the

most susceptible to spread widening. We therefore

recommend that investors position for wider swap

spreads in anticipation of further peripheral stress,

especially in the front part of the swap spread curve.

Correlations between various spread markets have

skyrocketed as the peripheral crisis has intensified.

Exhibit 26 shows 10Y weighted peripheral spreads

versus the % of variability explained by the first factor in

a rolling 6M PCA6 performed on various spread

6 Principal component analysis (PCA) is a tool that recombinesobserved variables into factors such that a smaller number of these

factors explain a large percentage of variation observed in the originaldata set.

markets.7 The % of variability explained by the first

factor has jumped from a low of 50% to its current levelof around 90%, suggesting that spreads have increasingly

been driven by a single factor – sovereign risk. Given

this dynamic, investors who are bearish on the

peripheral crisis should consider positioning for

wider spreads in markets which have lagged the

recent widening in peripheral spreads, and are trading

far from their peak levels. Such spreads have more

7 We run PCA on the correlation matrix of spread levels shown in

Exhibit 27 below (not including bank CDS and Euro Stoxx bank index).

Exhibit 25: …and, given our view that the peripheral debt crisiswill get worse before it gets better, we believe that swap spread

widening has more to goJ.P.Morgan swap spread forecast*; bp

18 Nov 11 1Q12 2Q12 3Q12 4Q12

10Y w td peripheral** spread 699 835 955 900 855

2Y Sw ap spreads 110 130 145 135 130

5Y Sw ap spreads 96 115 130 120 115

10Y Sw ap spreads 67 80 90 85 80

30Y Sw ap spreads 19 25 30 25 25

* We use the model in Exhibit 24 above and our projections on 10Y weightedperipheral spreads (see Euro Cash) to forecast 10Y swap spreads. Swapspreads at other maturities (such as 2Y) are projected by assuming that theywiden proportionally based on their widening relative to 10Y swap spreads over the past 2 years.** See Exhibit 22 for definition of weighted peripheral spreads.

Exhibit 26: Although correlations between various spreads haveskyrocketed as the peripheral debt crisis has intensified…10Y weighted peripheral spreads* versus % of variability explained by the firstfactor in a rolling 6M PCA** on various spreads***;bp %

200

300

400

500

600

700

800


40

50

60

70

80

90

100% v ariability explained

by 1st PCA factor

Weighted peripheral

spread

* See Exhibit 22 for definition of weighted peripheral spreads.** Principal component analysis (PCA) is a tool that recombines observedvariables into factors such that a smaller number of these factors explain a largepercentage of variation observed in the original data set.*** We run PCA on the correlation matrix of spread levels shown in Exhibit 27below (not including bank CDS and Euro Stoxx bank index).

Exhibit 24: Going forward, we expect swap spreads to continue tobe driven by the peripheral debt crisis, which has caused 10Y

swap spreads to widen an average of 9bp for each 100bp wideningin peripheral sovereign spreads…10Y German b/m swap spreads regressed against 10Y weighted peripheralspreads*; past 2Y; bp

y = 0.09x + 7.7

R2 = 82%

0

20

40

60

80

0 200 400 600 800

10Y weighted peripheral spread*; bp

* See Exhibit 22 for definition of weighted peripheral spreads.




Pavan WadhwaAC

(44-20) 7777-3370

Fabio Bassi (44-20) 7325-8615


68

potential to widen than those that are trading at, or close

to, their wides.

Exhibit 27 shows how far various market variables are,

on average, from their local and Lehman peaks.8 We

examine FRA/OIS basis, Libor and OIS swap spreads,

interest rate bases, FX bases, sovereign spreads,

European bank CDS, and European bank equities.

Markets that appear to be farthest away from their

local/Lehman peaks are OIS swap spreads and 3s/6s

basis. We therefore recommend that investors consider

positioning for a widening of these spreads. On the other

hand, FRA/OIS, sovereign spreads and European bank

CDS spreads are close to their wides; investors who

believe that the sovereign debt crisis will subside should

consider positioning in these markets.

As the sovereign debt crisis has progressed, the optionsmarket has frequently underpriced swap spread

directionality. As Exhibit 28 shows, the implied

directionality of 10Y swap spreads has consistently

stayed above delivered directionality. This is because

markets have failed to account for the fact that swap

spread counter directionality increases significantly

during stressed environments as investors rush into the

8 Distance from peak is defined as (Peak minus current level) / (Peak

minus 1 June 2011 level). 1 June 2011 was roughly the start of this legof the sovereign debt crisis.

Bund contract (Exhibit 29). Since market stress has

increased almost monotonically in the past few months,

swap spread directionality has increased concurrently

(that is, become more negative). Given that we expect

market stress to increase in 2012, we recommend

Exhibit 28: As the sovereign debt crisis has progressed, theoptions market has consistently underpriced swap spreaddirectionality…2-week MA of implied directionality of front Bund swap spreads versus 3Mdelivered directionality*; past 1Y; %

-40

-30

-20

-10

0

10


Implied direc tionality


* Implied directionality computed as (implied vol on maturity-matched swaption /implied vol on Bund) – 1. Delivered directionality computed as rolling 3M beta of daily change in swap spread regressed against daily change in CTD spot yield.Contract is rolled 10-days before option expiration.

Exhibit 27: …some spread markets remain further away from their local and Lehman peaks than others. Investors who are bearish on theperipheral crisis, should consider risk-aversion trades where levels are far from peak levels achieved during periods of stress. On this basis,

OIS swap spread wideners appear attractive, followed by 3s/6s basis wideners Average distance of various spreads from local and Lehman peaks; %

Peak level* Distance from peak**

18 Nov 11 01 Jun 11 Since 1-Jun-11 LEH times Since 1-Jun-11 LEH times Av erage

1 IMM2 FRA/OIS 77 24 77 93 0% 24% 12%

2 IMM6 FRA/OIS 60 28 60 62 0% 8% 4%

3 2Y Libor sw ap spreads 110 53 112 118 2% 11% 7%

4 10Y Libor Sw ap spreads 67 33 78 82 24% 30% 27%

5 2Y OIS sw ap spreads 18 10 25 41 46% 74% 60%

6 10Y OIS sw ap spreads 13 -2 35 40 59% 63% 61%

7 1Y 1s/3s basis 33 14 33 54 1% 53% 27%

8 1Y 3s/6s basis 29 15 35 44 26% 51% 39%

9 1Y EUR/USD Fx basis -87 -26 -88 -123 2% 38% 20%

10 2Y EUR/USD Fx basis -72 -26 -73 -87 3% 25% 14%

Average 16% 38% 27%

11 10Y w td peri spread 699 353 739 - 10% - 10%

12 Bank CDS*** 312 123 322 322 5% 5% 5%

13 Euro Stox x Bank Index 125 193 119 90 8% 34% 21%

* Peak level between 1 June 2011 and 18 November 2011 and past 5Y to capture Lehman peaks.** Distance from peak defined as (Peak minus current level) / (Peak minus 1 June 2011 level).*** Average of 5Y CDS on Deutsche, UBS, BNP, Barclays, Santander, and Unicredit.




Pavan WadhwaAC

(44-20) 7777-3370

Fabio Bassi (44-20) 7325-8615


69

conditional bull wideners as a less risky alternative to

outright swap spread widening trades.

Euro volatility

Much like 2010, 2011 was not a year to buy swaption

gamma since such a strategy would have yielded flat

returns with large P&L swings (Exhibit 30). On

average, losses from buying volatility during risk-on

episodes offset gains collected during risk-off

episodes, resulting in high P&L volatility (Exhibit 31).

We analyse returns from buying 3Mx10Y swaption

gamma during various risk-on and risk-off episodes from

January 2010, a period during which the sovereign debt

crisis dominated market sentiment. Exhibit 32 lists the

date ranges used for this analysis. We find that, on

average, buying gamma during risk-off periods is indeed profitable.9 For instance, buying 3Mx10Y gamma has

returned an average of 40bp of notional during these risk-

9 We note that there are episodes when volatility markets initially do

not respond in an expected manner at the onset of a risk-off regime. Nevertheless, on the back of continued stress, they eventually do catch

up, resulting in positive returns to long volatility positions. For

example, during the risk-off episode in May 2011, volatility markets

were unmoved by discussions surrounding the Greek PSI even thoughequity markets were trended lower. Long gamma positions resulted in

losses during this period. However, volatility shot up as soon asconcerns surrounding Italy came to the fore in July.

off episodes. On the other hand, buying volatility during

risk-on episodes has resulted in an average loss of 75bp

of notional (Exhibit 31).

Drivers of volatility in 2011

The sovereign debt crisis and the ECB were the

dominant drivers of the volatility surface in 2011

(Exhibit 33). Volatility in the belly of the curve headed

lower in 1Q11 as markets expected the peripheral crisis

Exhibit 30: Despite heightened uncertainty and peripheral stress,2011 turned out to be a mediocre year for outright long EURgamma positions…

Average 1M returns* and annualised information ratio (IR) from buying EUR3Mx10Y swaption straddles;bp of notional information ratios

-60

-40

-20

0

20

40

2001 2003 2005 2007 2009 2011

-6

-4

-2

0

2

4 Av erage (left)

Information ratio (right)

* Trades entered daily and held for 1M. Returns calculated using J.P.Morganvolatility indices, which assume daily delta-hedging and zero transaction cost.Options are re-struck at the beginning of each month.

Exhibit 31: …as positive returns from buying gamma during risk-off periods was offset by losses during risk-on episodes

Average* cumulative return** from buying 3Mx10Y EUR swaption straddles; bpof notional

-80

-60

-40

-20

0

20

40

60

0% 20% 40% 60% 80% 100%

% days into risk on/off period; %

Risk-off

Risk-on

* Averaged over risk-on/off episodes (see Exhibit 32 below for definition of dates).** See Exhibit 30 for definition of returns.

Exhibit 29: …as markets have failed to take into accountincreasing counter directionality of swap spreads in stressedenvironments. Going forward, we recommend conditional swap

spread trades as a less risky alternative to outright trades3M delivered directionality* of 10Y swap spreads versus 20D MA of 10Yweighted peripheral spreads (inverted axis);% bp (inverted axis)

-35

-30

-25

-20

-15

-10

-5


200

300

400

500

600

700


Weighted peripheral

spread (inv erted axis)

* Delivered directionality computed as rolling 3M beta of daily change in swapspread regressed against daily change in CTD spot yield. Contract is rolled 10-days before option expiration.




Pavan WadhwaAC

(44-20) 7777-3370

Fabio Bassi (44-20) 7325-8615


70

to be contained to Greece, Ireland and Portugal.

However, as discussions around Greek PSI gained

momentum, implied volatility inched higher and finally

spiked up as contagion spread to Italy and Spain.

The front end of the volatility curve was, however,

equally impacted by both economic data and the

peripheral crisis. An improving outlook for globalgrowth and rising inflation expectations led the ECB to

start hiking policy rates in April. Higher rates and the

escalating peripheral crisis drove front end volatility

higher. Later on, as the Euro area economy started to

deteriorate, the ECB reacted by extending extraordinary

liquidity measures and eventually cutting policy rates in

November. At that point, front-end volatility decoupled

from peripheral stress as it became clear that the ECB

was likely to remain in a low-for-long mode.

Trade the belly of the volatility curve from the long

side

Our baseline view for 2012 is that the peripheral crisis

will worsen leading to higher implied volatility in the

belly of the curve (see 2012 targets below). While

implied volatility will likely increase in tandem with

peripheral spreads, delivered volatility in the belly of the

curve is also expected to remain high for two reasons.

First, event risk remains high. Even if policy makers

come up with a resolution to the crisis, implementation

risks of such measures will remain high. Second, bank

funding pressures and thin markets will support delivered

volatility. Thus, we recommend trading the belly of

the volatility curve from the long side in 2012.

The ex-ante level of 1) implied volatility, 2) weighted

peripheral spread, and 3) 2s/10s swap curve were the

best predictors of volatility returns in 2011. Buying

gamma at lower levels of implieds yielded better returns.

Similarly, buying gamma when peripheral spreads are

high and the 2s/10s curve is steep resulted in better

Exhibit 32: Risk-on/off episodes from January 2010 to presentStart and end dates for various risk-on/off episodes since 1 January 2010, and% change in Euro Stoxx50 between start and end dates (%);

Risk-on

Start End # months Start End % change*

05-Feb-10 15-Apr-10 2.3 2632 3013 14%

30-Nov -10 18-Feb-11 2.7 2651 3068 16%

16-Mar-11 02-May -11 1.6 2721 3009 11%

12-Sep-11 28-Oct-11 1.6 1995 2462 23%

Av erage 2.1 16%

Risk-off

04-Jan-10 05-Feb-10 1.1 3018 2632 -13%

15-Apr-10 24-May -10 1.3 3013 2558 -15%

18-Feb-11 16-Mar-11 0.9 3068 2721 -11%

02-May -11 12-Sep-11 4.5 3009 1995 -34%

28-Oct-11 18-Nov -11 0.7 2462 2237 -9%

Av erage 1.7 -16%

Euro Stoxx 50Date

* % Change of Euro Stoxx 50 defined as (End level – Start level) / Start level.

Exhibit 33: Implieds on longer tails were boosted by the sovereigndebt crisis, whereas implieds on shorter tails were negativelyimpacted by high excess liquidity

EUR 3Mx10Y and 3Mx2Y implied volatility versus 10Y weighted peripheralspreads*; 1 January 2011 – 18 November 2011;bp/day bp

4

5

6

7

8


200

300

400

500

600

700

800

3Mx2Y

3Mx 10Y Peripheral spread

* 10Y weighted peripheral spread computed against Germany for Greece,Ireland, Portugal, Italy, and Spain (weighted by the size of their outstanding bondmarkets).

Exhibit 34: Long EUR volatility returns have been driven by thelevel of implieds and peripheral spreads at trade initiation…Rolling 1M return* from buying EUR 3Mx10Y volatility regressed against 1Mlagged levels of 1) EUR 3Mx10Y implied volatility, 2 )10Y weighted peripheralspread**, and 3) 2s/10s EUR swap curve; 1 February 2011– 18 November 2011;bp of notional

y = -52.3*(3Mx10) + 0.67*(Peri sprd) + 2.1*(2s/10s) - 207

R2 = 68%

-700

-650

-600

-550

-500

-450

-400

4 5 6 7 8

1M lagged 3Mx10Y implied vol; bp/day

* See Exhibit 30 for definition of returns.

** See Exhibit 33 for definition of weighted peripheral spreads.




Pavan WadhwaAC

(44-20) 7777-3370

Fabio Bassi (44-20) 7325-8615


71

returns. The intuition behind this is that wide peripheral

spreads and a steep curve results in high subsequent

delivered volatility, leading to improved returns.Together these three factors explain about 2/3rds of the

total variability in returns in 2011 (Exhibit 34).

Using the above relationship we develop a framework for

identifying trading opportunities in 2012 (Exhibit 35).

Over the past few months, the 2s/10s swap curve has

exhibited a strong relationship to peripheral spreads (see

Exhibit 11). Thus, in our model, we replace the 2s/10s

curve with peripheral spreads. We then use the modified

2-factor model to calculate the level of implied volatility

that generates flat returns over the next month given

current level of peripheral spreads. We find that buying

gamma when the combination of implied volatility and peripheral spreads is below the -1 standard deviation line

is generally profitable. Strong sell signals are rarely

generated because of the overall bias towards higher

volatility and wider peripheral spreads over the past few

months. For 2012, we therefore recommend initiating

gamma trades if the combination of implied volatility

and peripheral spread is outside the 1 standard

deviation line.

With the peripheral debt crisis expected to escalate, and

10Y weighted peripheral spreads targeted to widen by

250bp, we target implied volatility on 3Mx10Y at

9.4bp/day (Exhibit 36).10

Fade volatility flare-ups at the front end in 2012

With excess liquidity high and the ECB expected to be

on perma-hold starting in 2H12 (see EONIA curve),

front-end yields are likely to stay rangebound, leading to

a low volatility environment (Exhibit 37). We,

therefore, target 3Mx2Y implied volatility at 4bp/day

and recommend short gamma positions on 2Y tails.

A risk to this outlook comes from FRA/OIS volatility

which may stay elevated due to bank funding pressures

(see Swap spreads). Front-end volatility is likely to

flare-up during episodes of FRA/OIS widening. We

recommend fading such moves.

10 We use the following model to arrive at our target for 10Y implied

volatility: 3Mx10Y = 0.0061*(10Y weighted peripheral spread) + 3.56;R-sqr = 73%; 1 Jan 2010 – 18 November 2011.

Exhibit 35: …leading us to recommend buying volatility when thecombination of implied volatility and peripheral spreads is 1SDbelow the zero iso-return line

3Mx10Y EUR volatility versus 10Y weighted peripheral spreads* at initiation of long gamma trades in 2011; solid line is t he zero iso-return line and light lines

are ±1SD lines**; bp/day

4

5

6

7

8

9

10

300 400 500 600 700

Iso-return line

<0 returns

>0 returns

Current

Sell volatility

Buy v olatilityStay

neutral

Weighted peripheral spread; bp

18-Nov-11

* See Exhibit 33 for definition of weighted peripheral spreads.** Based on regression beta and standard errors in Exhibit 34.Note: we reduced our 3-factor model to a 2-factor model as follows. The 2s/10sswap curve has exhibited strong directionality with peripheral spreads over thepast few weeks (2s/10s swap curve = -0.0014*(10Y wtd peripheral spread) +1.94; 8 August 2011–18 November 2011; R2 = 71%). We use this relationshipto replace the 2s/10s curve with peripheral spreads. We then generate the zeroiso-return line based on implied volatility and peripheral spreads alone.

Exhibit 36: EUR implied volatility targets for 2012EUR implied volatility targets for 1Q12 and 2Q12*; bp/day

18 Nov 11 1Q target 2Q target 2Q target - current

3Mx 1Y 5.4 4.0 3.5 -1.9

3Mx 2Y 5.5 4.5 4.0 -1.5

3Mx 5Y 6.6 6.6 6.7 0.1

3Mx 10Y 8.0 8.7 9.4 1.4

3Mx 30Y 8.9 9.8 10.6 1.7

2Yx 2Y 5.9 5.5 5.2 -0.7

2Yx 5Y 6.1 6.0 6.1 0.0

2Yx 10Y 6.4 6.6 6.7 0.4

5Yx 5Y 5.6 5.5 5.6 0.0

5Yx 10Y 5.6 5.7 5.7 0.0

10Yx 10Y 5.0 5.0 5.0 0.0

* Target for 3Mx10Y is based on its relationship with the 10Y weightedperipheral spread and our forecast of this spread in 2012 (see Euro Cash). For 30Y, we assume that the beta between 30Y and 10Y remains at 120%. 2Yimplieds in 2Q is the average implied of 3Mx2Y for 1H10 – a period when theECB was firmly on hold and excess liquidity was high. 5Y is computed as anaverage of 2Y and 10Y.




Pavan WadhwaAC

(44-20) 7777-3370

Fabio Bassi (44-20) 7325-8615


72

With front end swap yields range-bound (Exhibit 38),

mean reversion may be an attractive source of return as

delta-hedging short volatility positions infrequently could

boost performance. Indeed, 2Y swap yields have been

about 30% less volatile measured on a periodic basis than

on a daily basis.11 Exhibit 39 shows statistics on 1M

rolling returns from selling 3Mx2Y EUR gamma using

time-based and delta-based hedging strategies. We draw

the following conclusions from this analysis. First,delta-hedging less frequently provides higher risk-

adjusted returns than daily delta-hedging. Second, re-

setting hedges based on a delta threshold is better than

using a time-based rule. Thus, going into 2012, we

recommend infrequent delta hedging of short

volatility positions at the front end of the volatility

curve.

Stay neutral on ultra long tails

The 30Y sector remains the ‘forbidden fruit’ of European

volatility markets because volatility on 30Y tails has

historically been impacted by technical factors such asnon-inversion note hedging, pension fund receiving, and

1-way CSA hedging. Due to these technical factors,

implied volatility on 30Y tail tends to increase in excess

of 10Y tails whenever we encounter a crisis

(Exhibit 40).

11 For instance, 1M delivered volatility of 2Y swaps measured on 5D

changes is currently at 3.5bp/day (after adjusting by sqrt(5)) while thatmeasured on daily changes is at 5bp/day.

Exhibit 37: Look for delivered volatility at the front end to declinedue to exceedingly high excess liquidity; fade volatility flare-upsin 2Y tails

1M delivered volatility on 3Mx2Y EUR swaps regressed against 1M deliveredvolatility of 3Mx3M EONIA; 1 January 2009 – 18 November 2011; bp/day

y = 0.71x + 1.78

R2 = 64%

1

2

3

4

5

6

7

8

0 1 2 3 4 5 6 71M deliv v ol of 3Mx 3M EONIA; bp/day

18-Nov-11

Exhibit 38: High excess liquidity is also expected to keep frontend rates in a range,…EUR 2Y swap yield; %

1.35

1.40

1.45

1.50

1.55

1.60

1.65

05-Aug 26-Aug 16-Sep 07-Oct 28-Oct 18-Nov

Exhibit 39: …and we recommend enhancing returns of a shortvolatility position in 2Y tails by delta-hedging infrequentlyStatistics on 1M rolling returns* from selling 3Mx2Y EUR swaptions usingvarious time-based and delta-based hedging strategies; 8 August 2011– 18November 2011; bp of notional

Av erage SD Information Ratio

Days based hedging rule

Daily 9.3 8.8 3.7

Weekly 16.3 8.6 6.5

Fortnightly 18.0 10.8 5.8

Delta threshold based hedging rule

10% 9.5 8.8 3.7

30% 15.0 9.9 5.2

50% 20.2 9.6 7.3


Exhibit 40: Implieds on 30Y tails have lost some of their sensitivity to 10Y tails as technical factors have declined; weexpect technicals to remain a driver of 30Y tails, however, andsuggest a neutral view in this sector Rolling 3M beta from regressing daily changes in EUR 3Mx30Y implied volatilityagainst daily changes in EUR 3Mx10Y implied volatility; unitless

0.6

0.8

1.0

1.2

1.4

1.6

1.8

2.0

Jan 08 Oct 08 Jul 09 Apr 10 Jan 11 Nov 11

1-way CSA

hedging

Non-inversion

note hedging

Current




Pavan WadhwaAC

(44-20) 7777-3370

Fabio Bassi (44-20) 7325-8615


73

The sensitivity of 30Y tail volatility to 10Y tail volatility

appears to have declined somewhat over the past few

months. This may be because some of the technicalflows supporting 30Y tails have declined recently (for

example, pension funds refuse to receive at these low

levels of yields). Unfortunately, however, other hedging

flows such as 1-way CSA flows have continued to drive

30Y yields lower (and volatility higher). Going forward,

it is difficult to judge the evolution of these technical

flows and therefore we recommend staying away from

30Y tails.

Buy Bund volatility vs. swaption volatility as a source

of relative value

2011 has been a year of govie volatility. Regular boutsof flight-to-quality have helped long gamma positions in

govie options produce high risk-adjusted return,

especially relative to swaption gamma. Exhibit 41

shows the average monthly returns and annualised

information ratio from several such trades. We draw

three main conclusions from this analysis. First, buying

govie gamma outright was profitable across the curve,

with Bunds outperforming Bobl and Schatz. Second,

buying govie gamma vs. maturity-matched swaption

gamma was also profitable. Here again a strategy

involving Bunds was the clear winner. Finally, returns

from buying govie gamma vs. swaption gamma were

more stable than buying outright govie gamma, i.e., theyyielded superior information ratios.

Swap spread directionality was consistently negative in

2011 and generally exceeded implied directionality (see

Swap spreads). This disparity, which was most

prominent in Bunds, was the primary reason for the

outperformance of govie vs. swaption gamma. We

expect this disparity to continue in 2012 and recommend

buying Bund gamma vs. swaption gamma as an

efficient way to position for an escalation of the

peripheral crisis.

To help generate a trading signal, we develop a

framework akin to the one in Exhibit 35. As a first step,

we estimate that 60% of the variability in returns is

explained by peripheral spreads and implied volatility

differential (defined as govie implied minus swaption

implied) at trade initiation (Exhibit 42). The regression

signs are intuitive; buying govie vs. swaption gamma at

low implied volatility differential and wide peripheral

spreads is attractive. This is because wide peripheral

spreads indicate stress in the system which typically

results in high subsequent delivered volatility differential

as asset swaps become increasingly counter directional.

Exhibit 41: Buying govie volatility vs. maturity matched swaptionvolatility was a profitable strategy in 2011. Implementing thestrategy in Bunds would have resulted in higher P&L and

information ratio than buying Bobl or Schatz volatility outright Average 1M returns* (bp of notional) and annualised information ratio (unitless)from buying govie volatility and selling a gamma-equivalent amount of maturitymatched swaption volatility, vs. buying volatility outright; all trades scaled tohave the same gamma as Bund; 1 January 2011– 18 November 2011;bp of notional Information ratios

0

3

6

9

12

15

18

Bund Bobl Schatz Bund Bobl Schatz

Long gov ie vs . mat matched

swpt

outright

0.0

0.5

1.0

1.5

2.0

2.5

3.0


Exhibit 42: Outperformance of govie vs. swaption volatility can beexplained by the implied volatility differential and peripheralspreads at trade initiation…Rolling 1M returns* from buying Bund volatility vs. selling a gamma equivalentamount of maturity matched swaptions regressed against 1M prior levels of 1)10Y weighted peripheral spreads, and 2 ) implied volatility differential; 1 January

2011– 18 November 2011; bp of notional

y = 0.23*(Peripheral sprd) - 24.4*(Imp vol s pread) - 57.2

R2 = 63%

-20

0

20

40

60

80

100

120

200 250 300 350 400 450 500 550 600 650

1M ago 10Y w td peripheral spreads; bp





Pavan WadhwaAC

(44-20) 7777-3370

Fabio Bassi (44-20) 7325-8615


74

In Exhibit 43, we present the zero iso-return line and ±1

standard deviation lines from buying Bund gamma vs.maturity-matched swaption gamma. Buying Bund

volatility vs. swaptions when the combination of implied

volatility differential and peripheral spreads is below the

zero iso-return line is generally profitable. The table

within Exhibit 43 shows statistics on trades initiated

following the signal from this model in 2011. In addition

to a higher average P&L (overall average is shown in

Exhibit 41), this strategy has a success ratio of 85%. Of

course, these statistics are based on in-sample data and so

need to be viewed accordingly. However, given that the

fundamental drivers are not likely to change in 2012, we

recommend buying govie vs. swaption gamma when

the combination of implied volatility differential and

peripheral spreads is below the -1 standard deviation

iso-return line.

Stay neutral on vega

Unlike gamma, buying vega would have been

profitable in 2011 (Exhibit 44). Outright long vega

positions would have resulted in higher returns with

smaller swings. For instance, buying 5Yx10Y swaption

straddles would have produced an average P&L of 21bp

of notional per month compared to flat returns for

3Mx10Y. This outperformance of vega is mainly due to

the increase in implied volatility and reduced impact of

lower delivered volatility. For instance, 5Yx10Y implied

volatility has increased from 5.0bp/day to 5.6bp/day in2011 while 3Mx10Y implied volatility has increased

from 5.7bp/day to 8bp/day. However, subsequent

delivered volatility on 10Y swaps has been relatively low

compared to implied volatility at inception. For

example, average 3Mx10Y implied volatility at inception

during 2011 is 5.8bp/day, while subsequent 1M delivered

volatility has been around 5.1bp/day. Thus, for short

dated options, vega gains (from increasing implieds)

have been largely offset by gamma losses.

Exhibit 43: …suggesting that investors should buy govie vs.swaption volatility when the combination of implied volatility

differential and peripheral spreads is 1SD below the zero iso-return lineImplied volatility spread* vs. 10Y weighted peripheral spread; solid line shows

the zero iso-return line and light lines are ±1SD** lines; table shows statistics of trades initiated in 2011***; 1 January 2011– 18 November 2011; bp/day

-1

0

1

2

3

4

5

200 300 400 500 600 700

Iso-return line

<0 returns

>0 returns

Current

1SD

line


18-Nov-11

Av erage 22

IR 3

% profitable 85%

Light (dark) dots represent trades that were profitable (unprofitable) in 2011 over the next month, without considering the trading rule discussed here.* Bund implied – maturity-matched swaption implied volatility; futures rolled 20Dbefore option expiry.** Based on regression beta and standard errors in Exhibit 42.*** 1M statistics of trades entered when the combination of implied volatilitydifferential and peripheral spreads was below the zero iso-return line.

Exhibit 44: Vega outperformed gamma in 2011 as low deliveredvolatility was a drag on the latter

Cumulative returns from buying €100mn notional EUR 3Mx10Y vs. 5Yx10Yswaption straddles; 1 January 2011 – 18 November 2011; bp of notional

-200

-100

0

100

200

300


5Yx10Y

3Mx10Y

Exhibit 45: Vega has exhibited a concave profile to gamma. Thus,despite our bullish view on gamma, we stay neutral on vega5Yx5Y EUR implied volatility regressed against 3Mx5Y EUR implied volatility;past 2Y; bp/day

y = -0.08x 2 + 1.20x + 0.99

R2 = 73%

4.0

4.5

5.0

5.5

6.0

3 4 5 6 7 8

18-Nov-11

3Mx 5Y implied vol; bp/day




Pavan WadhwaAC

(44-20) 7777-3370

Fabio Bassi (44-20) 7325-8615


75

Additionally, vega has exhibited a concave profile to

its gamma counterparts (Exhibit 45). Thus, despite

our bullish view on gamma, we stay neutral on vega.

We present our targets for various vega points in Exhibit

36.

Sterling volatility

The story of Sterling volatility in 2011 has been eerily

similar to its EUR counterpart. Despite changing

inflation expectations and BoE policy reaction, GBP

volatility in the belly of the curve has been largely driven

by peripheral concerns. Thus, it is not surprising to see

that buying GBP swaption gamma in 2011 would have

also yielded mixed results, i.e., low monthly average

returns with large swings (Exhibit 46).

We do not expect this to change in 2012 and

recommend buying GBP swaption gamma in 10Y

tails. At the front end of the curve, we recommend

fading flare-ups by selling 2Y tail volatility and

enhancing returns in a range-bound environment by

hedging infrequently. Indeed, the argument for selling

volatility in GBP is even stronger despite the recent

widening of FRA/OIS at the front end. Our economists

forecast the BoE to engage in further QE purchases (see

United Kingdom) and the base rate to remain at 0.50%

throughout 2012, given the gloomy economic backdrop

and declining inflation expectations. A flat SONIA

curve will likely keep delivered volatility at the front-end

low (Exhibit 47). We present our target for various

points on the GBP volatility surface in Exhibit 48.

Given a choice between buying EUR and GBP gamma,

what should an investor do? Investors facing this

Exhibit 46: Similar to EUR volatility, buying GBP volatility in 2011would have generated mixed returns, since positive returns during

risk-off periods were offset by losses during risk-on periods Average 1M returns* and annualised information ratio achieved by buying GBP3Mx10Y swaption straddles;bp of notional ratios

-40

-30

-20

-10

0

10

20

30

40

2008 2009 2010 2011

-3

-2

-1

0

1

2

3

Av erage (left)

Information ratio (right)

* Trades entered daily and held for 1M. See Exhibit 30 for definition of returns.

Exhibit 47: We expect front-end volatility in GBP to declinemarginally as BoE remains on hold in 2012 and the policy

expectations curve remains flat…3Mx2Y GBP implied volatility regressed against 3M/15Mx3M GBP OIS curve;past 1Y; bp/day

y = 0.02x + 4.04

R2 = 76%

3

4

5

6

7

-20 0 20 40 60 80 100 120 140

18-Nov -11

3M/15Mx 3M OIS curve; bp

Exhibit 48: …even as volatility on 10Y tails increases in responseto an escalation in the peripheral debt crisisGBP implied volatility targets for 2012 vs. 18 November 2011; bp/day

18 Nov 11 1Q target 2Q target 2Q target - current

3Mx 2Y 4.8 4.5 4.0 -0.8

3Mx 5Y 5.9 6.0 6.1 0.2

3Mx 10Y 7.2 7.5 8.2 0.9

Exhibit 49: Given our view that the peripheral crisis will get worse,

we prefer buying EUR gamma over GBP gamma… Average* cumulative return** from buying 3Mx10Y EUR and GBP swaptionstraddles; bp of notional

-80

-60

-40

-20

0

20

40

60

0% 20% 40% 60% 80% 100%

EUR

GBP

EUR

GBP

Risk-off

Risk-on

% days into risk on/off period; %

* Averaged over risk-on/off episodes (see Exhibit 32 for risk-on/off dates).** See Exhibit 30 for definition of returns.




Pavan WadhwaAC

(44-20) 7777-3370

Fabio Bassi (44-20) 7325-8615


76

choice should consider buying EUR gamma over GBP

gamma. Although GBP volatility has been driven by

peripheral spreads, EUR gamma remains at the fore front

of the crisis and is more likely to outperform GBP.

Indeed, long gamma positions in EUR have, on

average, outperformed long GBP gamma positions

during various risk-off episodes in the past two years

(Exhibit 49). We also note that selling EUR gamma

during risk-on episodes is better than selling GBP

gamma. One reason for superior returns of EUR gamma

is that GBP implieds tend to decline with respect to EUR

implieds during periods of stress and vice versa (Exhibit

50). We use this relationship to arrive at our targets for

GBP gamma for 2012, listed in Exhibit 48.

Trading themes

• We are bullish on EONIA and high excess

liquidity; receive 6Mx6M EONIA on the back of

ECB policy rate cuts; implement bullish option

structures on Dec12 Euribor

− With the ECB expected to cut refi and deposit rates to0.50% and 0.25% respectively, we expect EONIA

fixings to fall to 30bp by mid-2012. We recommend

longs in 6Mx6M EONIA with a target of around 30bp,

and bullish option structures on Dec12 Euribor.

• We see value in bullish option structures on 5Y

swaps and recommend 1s/5s outright or

conditional swap curve flatteners

− With peripheral sovereign stress expected to escalate

in 2012, we expect the swap curve to flatten and

recommend carry efficient flatteners such as 1s/5s,which are a proxy for bullish positions but with better

risk profile, along with receiver structures on 6Mx5Y.

• 10s/30s is likely to remain driven by technical

flows; we have a flattening bias

− The 10s/30s swap curve is likely to remain driven bytechnical factors and is expected to flatten as

peripheral spreads widen. We present a framework toanalyse the attractiveness of long-dated forward

steepeners but refrain from recommending this trade

due to the prevalence of technical factors.

• We favour 2Y wideners and have a swap spread

curve flattening bias

− Swap spreads will likely widen beyond their Lehman peak in 1H12. We target 2Y and 10Y swap spreads at

145bp and 90bp, respectively, by mid-2012.

Conditional swap spread wideners offer better risk/reward than outright swap spread wideners.

• Position for further widening in spread markets,

such as OIS swap spreads and 3s/6s basis, that

have lagged the recent widening in swap spread

− Although spread markets are being largely driven by a

single-factor, i.e. sovereign risk, some spread markets

such as OIS swap spreads and 3s/6s basis are stilltrading far below their local/Lehman peaks,

suggesting that investors should express risk-off views

in these markets.

• Trade EUR gamma from the long side in the belly

of the curve, while fading spikes in volatility at the

front end; buy Bund volatility vs. swaption

volatility

− EUR volatility in the belly of the curve will increaseas the peripheral crisis escalates; we target 3Mx10Y at

9.4bp/day by 2Q12. Buy gamma in the belly of the

volatility curve and fade flare-ups in volatility at the

front end. Favour long positions in govie volatility vs.

swaption volatility.

− On a cross-market basis, buy EUR gamma vs. GBP

gamma.

Exhibit 50: …as GBP implieds will likely underperform EURimplieds during such episodes

3Mx10Y GBP-EUR implied volatility spread regressed against 3Mx10Y EURimplied spread; risk-off periods* from 1 January 2010 – 18 November 2011;bp/day

y = -0.44x + 2.57

R2 = 55%

-2

-1

0

1

2

3.4 4.0 4.6 5.2 5.8 6.4 7.0 7.6 8.2

EUR 3Mx 10Y implied vol; bp/day

18-Nov -11

* See Exhibit 32 for dates of risk-off episodes.




Francis DiamondAC

(44-20) 7325-3541

[email protected]. Morgan Securities Ltd

77

United Kingdom

• Gilt yields fell during 2011 on the back of

peripheral spread widening and reached all-time

lows across the curve. The BoE restarted QE gilt

purchases despite headline CPI breaking the 5%

level

• The macro picture for the UK is worrying, with

2012 growth expected to be just 0.5%. The BoE

will likely keep rates on hold and we expect QE

gilt purchases to increase to £425bn in 2012. As a

result we estimate the BoE will own just over

40% of the total secondary market of outstanding

gilts

• UK banks have relatively small exposures to

peripheral sovereign debt but they could be

vulnerable if financing markets freeze up given

their exposure to wholesale funding markets and

2012 refinancing needs

• We expect 2Y gilts to trade in a 50-60bp range

and recommend fading any richening move below

the 50bp level

• Wider peripheral spreads will drive 10Y and 30Y

gilt yields lower – we target 1.50% in 10Y gilts

and 2.25% in 30Y gilts by 2Q12. Should the UK

fall under the sovereign risk spotlight, we expect

ongoing QE purchases, a relatively low

proportion of non-domestic ownership and an

independent currency to limit any fiscal-driven

rise in gilt yields

• The 2s/10s curve will remain highly directional

with 10Y yields. We position for flattening and

expect 2s/10s to reach the 100bp level by mid-

2012

• We have a bias for a flatter 10s/30s curve as

directionality with 10Y gilts should weaken and

increasing QE purchases should drive the curve

flatter. We target the 75bp level by 2Q, 15bpflatter than currently implied by the forwards

• Position for wider 5Y and 10Y swap spreads in

1H12, driven by wider peripheral spreads, lower

gilt yields and a £150bn increase in QE gilt

purchases

2011 – The year of the long bond

2011 saw the European sovereign risk crisis move up

another notch. The year started off relatively benignly,

with the market pricing in rate-tightening expectations as

inflation moved higher and surprised to the upside. The

SONIA curve was pricing 100bp of tightening over a 1Yforward horizon in the first couple of months of the year.

10Y gilt yields briefly touched the 4% level in February

and traded in a broad range until the sovereign risk crisisreignited in the spring. From then on, gilt yields steadily

fell almost in lockstep with Bunds to make new all-time

historic lows, with 10Y real yields moving into negativeterritory.

Domestic data took a backseat as peripheral spreads

were the main driver of gilt yields and of the 2s/10s

Exhibit 1: The story of 2011 – it’s a one-factor worldCorrelation of UK par yields and curve with 10Y weighted peripheral spreads*, last 6M;%

74% 72%78% 79%

73%

62%

37%

66%

36%

45%

0%

20%

40%

60%

80%

100%

2 Y

5 Y

1 0 Y

3 0 Y

5 0 Y

2 s / 5 s

5 s / 1 0 s

2 s / 1 0 s

1 0 s / 3 0 s

3 0 s / 5 0 s

* 10Y weighted peripheral spread computed against Germany for Greece, Ireland,Portugal, Italy, and Spain (weighted by the size of their outstanding bond markets).

Exhibit 2: 2011 was the year of the long bondTotal return performance, 1 January – 18 November 2011, government bonds;%

0%

5%

10%

15%

20%

25%

30%

35%

40%

2Y 5Y 10Y 30Y

UK US Germany




Francis DiamondAC

(44-20) 7325-3541


78

curve (Exhibit 1). The large fall in 30Y yields resulted

in long-end bonds outperforming on the curve in total

return terms, a theme that was common to both USTreasuries and Bunds (Exhibit 2). The BoE increasedQE by £75bn in October on the back of heightened

downside growth risks despite headline CPI inflation

reaching the highest level since mid-2008 (5.2% oya),

although the market impact was more limited than whenQE1 first started. The 2s/10s curve was highly directional

and flattened almost 70bp whilst 10s/30s steepened from

April to September only to flatten back following the QEextension in October.

Sterling funding markets reflected the various facets of

the sovereign risk crisis as Libor/OIS spreads reached

their widest levels since May 2009 whilst the flight-to-

quality demand for gilts and the increase in QE gilt purchases drove gilt GC below SONIA for short dates1.

2012 looks bleak

The macro outlook for the UK in 2012 is predominantly

a function of the global outlook and the ongoing

sovereign risk issues in Europe. Our economists expect

0.8% growth in the UK for this year and 0.5% oya in

2012, with the likelihood of two consecutive quarters

of mild contraction estimated at close to 50%. We believe that that the UK will avoid a deep recession

provided growth outside of Europe holds up.

Empirically, the UK PMI has demonstrated larger sensitivity to the global ex-Euro area PMI, but other

factors such as bank funding concerns and the ongoing

fiscal contraction present downside risks to UK growth.

There are worrying signs that UK output couldremain at anemic levels for several years. Corporates

have been running financial surpluses (i.e. have been net

lenders) for several years but since 2007, households

have shifted from being net borrowers to net lenders possibly indicating signs of a balance sheet recession

(Exhibit 3). As a result, the government stepped in with

fiscal supports in order to prevent the economy from

shrinking, with total government net borrowing moving

from 2.5% of GDP in 2007 to 10% of GDP in 2010.Fiscal contraction has prevented this pace of government

support from continuing, but the private sector is in debt-

reduction mode and appears unlikely to take up themantle to drive the economy forward for some time (we

estimate consumption will not begin recovering until at

least 2H12). The growth rate of loans to the householdsector has recently fallen to negative levels, and both

1 This has been the case in Euro cash markets for some time.

Exhibit 3: There are signs of a balance sheet recession in the UKFinancial balance by sector (negative is net borrower, positive is net lender); % GDP

-12%

-8%

-4%

0%

4%

8%

1990 1995 2000 2005 2010

Financial Corporates

Gov t Household

Exhibit 4: Private sector loan growth in the UK is in the doldrumsGrowth rate in MFI lending to private sector; %oya

-20%

-10%

0%

10%

20%

30%

40%

50%

60%

1998 2000 2002 2004 2006 2008 2010

Financial

Corporates

Household

Exhibit 5: Several more years of fiscal correction are needed to bringthe deficit downFiscal thrust* and Public Sector Net Borrowing (PSNB); % of GDP

-6

-4

-2

0

2

4

6

8

10

1972 1982 1992 2002 2012

Fis cal thrus t N et borrow ing

forecast

2% fiscal drag is max imum

likely achiev able in one year

* Fiscal thrust/drag is defined as year on year change in PSNB.




Francis DiamondAC

(44-20) 7325-3541


79

corporate and financial loan growth remain at very

depressed levels historically (Exhibit 4).

We expect the base rate to be kept on hold at 50bp

during 2012, with QE purchases increasing by

another £150bn by the end of the next year, taking

total QE purchases up to £425bn (around 30% of

GDP and 40% of total outstanding of gilts). Headlineinflation is forecast to fall to 1.9% oya by the end of

2012, but we expect this to have little impact on either

gilt yields or inflation breakevens (see Inflation Linked Markets), as markets will remain focused on the risks

around peripheral Europe and the growth outlook.

The UK budget deficit remains historically high, and we

expect this to come in at 5.6% of GDP for FY11/12, but

several more years of fiscal correction of around 1.5–2%of GDP per year are needed to reduce the deficit to more

appropriate levels (Exhibit 5). The Chancellor has statedthat the government’s self-imposed rules are to achieve a

balanced, cyclically adjusted deficit and have a falling

trajectory of debt/GDP by 2015/16. The recent downshiftin UK growth will likely mean that the 2012 Budget may

just show these rules being missed and that the

Chancellor will have to pencil in some additional modest

tightening in the next few years.

Summary themes

The evolution of peripheral spreads will likely be thekey driver of gilt yields for 2012. Macro, fiscal and

technical factors will all be important, but will likely

have a larger impact on the evolution of the curve andswap spreads than the outright level of yields. We

describe our trading themes and gilt market views indetail in the following pages, but in summary we think:

1) 2Y gilt yields will remain anchored and will likely

trade within a tight range. We believe the BoE isunlikely to cut base rates below the current level and

think 2Y gilt yields are floored at close to 50bp.

Exhibit 6: We expect 10Y and 30Y yields to fall substantially in 1H12, with the 2s/10s curve flatteningCurrent and J.P. Morgan gilt forecasts for 2012; % unless stated

18-Nov -11 1Q12 2Q12 3Q12 4Q12Q2 vs. fwd

(bp)

Q4 vs. fwd

(bp)

BoE base 0.50 0.50 0.50 0.50 0.50 n/a n/a

1M SONIA 0.52 0.50 0.50 0.50 0.50 4 0

2Y 0.48 0.50 0.50 0.55 0.65 4 -1

5Y 1.13 1.00 1.00 1.05 1.15 -22 -17

10Y 2.26 1.80 1.50 1.75 1.95 -89 -61

30Y 3.19 2.70 2.25 2.40 2.60 -103 -75

2s/10s (bp) 178 130 100 120 130 -93 -60

10s/30s (bp) 93 90 75 65 65 -14 -14

Exhibit 7: Gross gilt sales likely to rise to close to £200bn in FY12/13,but net sales will be lowGross gilt issuance and net gilt issuance adjusted for redemptions and QE*, fiscal year

basis; £bn

0

50

100

150

200

250

0 4 / 0 5

0 5 / 0 6

0 6 / 0 7

0 7 / 0 8

0 8 / 0 9

0 9 / 1 0

1 0 / 1 1

1 1 / 1 2

1 2 / 1 3 ( J P M )

Net of QE and redemptions Gross gilt sales

* We expect £150bn of additional QE taking total gilt purchases to £425bn by the endof 2012.

Exhibit 8: Gilt issuance in the belly of the curve is likely to increase inFY12/13Gilt issuance by sector as proportion of total issuance, FY11/12 and J.P.Morganforecast for FY 12/13; %

0%

5%

10%

15%20%

25%

30%

35%

40%

Short nom inal M edium

nominal

Long nom inal* Linkers *

FY 11/ 12 F Y 12/13

* Includes syndicated issuance.




Francis DiamondAC

(44-20) 7325-3541


80

2) Wider peripheral spreads will likely drive 10Y and

30Y gilt yields lower – we target 1.50% in 10Y gilts

and 2.25% in 30Y gilts by 2Q12 (Exhibit 6). We seelimited risk from sovereign concerns in the UK. Indeed,if the UK were to fall under the sovereign risk spotlight,

we expect ongoing QE purchases, a relatively low

proportion of non-domestic ownership and an

independent currency to limit any fiscal-driven rise ingilt yields.

3) The 2s/10s curve will remain highly directional

with 10Y yields. Position for flattening as we expect

2s/10s to reach the 100bp level by mid-2012.

4) Bias for a flatter 10s/30s curve. We expect

directionality with 10Y gilts to weaken as yields fall

further. We think 30Y gilts will modestly outperform asQE gilt purchases increase to the £425bn level.

5) Position for wider swap spreads in the 5Y and 10Y

sectors.

Gilt supply, QE purchases and bank

funding

We expect gross gilt issuance of around £190bn in

FY12/13, some £25bn above planned gilt sales for the

current fiscal year. Net borrowing for FY12/13 willlikely be some £10bn higher than forecast by the OBR

(£122bn) given a small degree of fiscal slippage, and weexpect any increase in the current year net borrowingrequirement to be reflected in higher borrowing needs for

FY12/13. Gilt redemptions are also higher next fiscal

year at just under £50bn.

Issuance net of redemptions and QE purchases is

expected to be just below £25bn in the current fiscal

year, which will then rise to just under £60bn in FY12/13

(Exhibit 7). In terms of maturity split, we expect a £17bnincrease in issuance in the 5-15Y sector, to partially

counter the expected increase in QE gilt purchases

(Exhibit 8). We expect long nominal gilt issuance to

increase around £5bn, but remain roughly unchanged as a proportion of total gilt issuance. On a calendar year basis

we expect £185bn of gilt supply in 2012 to be fully offset

by £43bn of redemptions and £150bn of QE gilt purchases.

We think that the demand impact through QE gilt

purchases could be a significant driver of gilt yields.

Our central view is for the BoE to announce QE

increases of £75bn in both February 20122 and May2012, taking total QE asset purchases to £425bn. We

expect these purchases to be entirely in gilts as no other sterling fixed income markets are large enough to

accommodate the scale of QE purchases, although we

don’t rule out that the BoE could extend purchases into

bank debt on a much smaller scale. The BoE currently

2 Note: The £75bn of purchases which began in October 2011 will be

completed in early February 2012, taking total QE purchases to£275bn.

Exhibit 9: Total QE BoE gilt holdings are expected to peak at around40% of the secondary market next year, but the BoE could own almost60% of the 10-25Y sector

Evolution of BoE holdings as % of gilt secondary market outstanding* based onadditional QE** and gilt issuance profile for 2012; %

10%

20%

30%

40%

50%

60%

Nov 11 Mar 12 Jun12 Dec12

3-10Y 10-25Y 25Y+ Total

* Excludes DMO collateral holdings.** We assume a further £150bn of QE gilt purchases taking total QE gilt holdings to£425bn.

Exhibit 10: If BoE QE gilt holdings were to reach £450-500bn, then theliquidity of the gilt market could be severely impacted as the BoE wouldown large segments of the gilt marketEstimates for distribution of BoE QE gilt ownership by sector by end 2012incorporating our issuance forecast; £bn

37%40%

52%59%

66%

60%

45%50%

33%

43%

52%

40%

0%

10%

20%

30%

40%

50%

60%70%

80%

400 450 500

Hypothetical total BoE QE gilt holdings; £bn

3-10Y 10-25Y 25Y+ Total




Francis DiamondAC

(44-20) 7325-3541


81

owns around 25% of the gilt market and around 38% of

the secondary market outstanding of bonds in the 10-25Y

sector.

Given our QE forecast and supply estimates for the

next fiscal year, we estimate that the BoE could own

just over 40% of the entire gilt market and close to

60% of the 10-25Y sector (Exhibit 9). In our view, thisis close to the limit at which QE gilt purchases will affect

gilt market liquidity, and we think that the BoE will be

unable to increase QE gilt purchases above the £450-500bn level as holdings in some sectors of the curve

would become prohibitively large (Exhibit 10).

Concerns around the impact of the sovereign risk crisis

on banks’ increased in the past month. UK banks have

relatively small exposures to peripheral governmentbonds (Exhibit 11), but they could be vulnerable if

financing markets freeze up given their exposure to

wholesale funding markets and 2012 refinancing

needs. UK banks have around €100bn of refinancing

needs in 2012, which is large in absolute terms comparedwith other European countries, but is relatively small in

terms of total banking sector assets (Exhibit 12).

Nonetheless, should conditions in financing markets

deteriorate if sovereign risk stresses increase, then this

could well pose some problems. We would expect the

BoE to respond by increasing the amount of fundingavailable via the 3M and 6M LTROs, which are currently

small in size at £5bn and £2.5bn, respectively, as wasdone in 2009.

The BoE is unlikely to reopen the SLS3, which will close

in 2012 as other mechanisms such as the discountwindow facility are now in place. UK banks have raised

capital and reduced leverage, compared with European

banks, but capital remains below the 10% Basel III

requirement which needs to be implemented by 2019.

Our equity analysts expect them to meet this targetorganically, but in the event of a severe recession, they

are uncertain as to whether regulators would intervene

and force capital-raising at some point over the next few

years.

Front end duration view

2011 was almost a re-run of 2010 in SONIA space, asrate-tightening expectations priced in during the first few

months of the year were steadily unwound from April

onwards, although the magnitude of the re-pricing was

3 Special Liquidity Scheme (SLS) was launched in April 2008 andallowed institutions to lend a wide range asset-backed bonds to the BoE

in exchange for 9M T-bills.

Exhibit 11: UK bank exposures to peripheral government bonds arerelatively small compared with total bank reservesUK bank exposures to Euro zone sovereigns, public sector claims*; $bn

$bn %

% bank

reserves

Austria 2 1% 0%

Belgium 6 3% 1%

Finland 4 2% 1%

France 56 31% 10%

Germany 62 34% 11%

Greece 4 2% 1%

Ireland 4 2% 1%

Italy 17 10% 3%

N etherlands 20 11% 3%

Portugal 2 1% 0%

Spain 8 4% 1%Peripheral 34 19% 6%

Ita + Spa 25 14% 4%

Public sector

* Public sector claims as proportion of all Euro zone public sector claims. Derived fromBIS data 2Q11.

Exhibit 12: UK bank refunding needs for 2012 are large in absolute termsbut are comparatively small as a proportion of banking sector assetsTotal securitised, secured and unsecured bank refinancing requirements for 2012 €bn proportion of total banking sector assets; %

0

20

40

60

80

100

120

140

A u s t r i a

B e l g i u m

F i n l a n d

F r a n c e

G e r m a n y

G r e e c e

I r e l a n d

I t a l y


P o r t u g a l

S p a i n

U K

0.0

1.0

2.0

3.0

4.0 Amount

Proportion of bank assets

Exhibit 13: We don’t expect the spread between O/N SONIA and the baserate to fall to the levels seen during QE1, as access to the BoE reservefacility was expanded in late 2009O/N SONIA – base rate spread and periods of QE purchases; bp

-15

-10

-5

0

5

10

Mar 09 Sep 09 Mar 10 Sep 10 Mar 11 Sep 11

QE1

QE2




Francis DiamondAC

(44-20) 7325-3541


82

larger than that seen in 2010. We expect the BoE to

keep base rates on hold at 50bp, with additional

monetary easing coming from further QE rather thana lower base rate. The BoE views 50bp as an effectivefloor and has previously stated that its preferred tool for

further easing monetary conditions is more QE. To this

extent, we think that SONIA is effectively floored at

50bp, and we do not expect it to trade much below thislevel. Following on from the start of QE in March 2009,

O/N SONIA fell around 15bp below the base rate, but

this was due to fragmentation in the SONIA marketgiven the limited number of participants who had access

to the BoE reserve facilities. In the autumn of 2009, the

BoE lowered the requirements for institutions to become

reserve scheme members and access to the BoE reserve

accounts was expanded; the move saw the O/N SONIA

vs. base rate spread gradually increase (Exhibit 13).Given that access to the reserve facilities remains

unchanged, we don’t think that any additional QE in

itself should push O/N SONIA below the level of base

rates.

Given this, we expect SONIA rates up to 12M

forward to hover around 50bp, and we would look to

fade any cheapening on a forward basis as we think

the probability of rate hikes over the next 12 months is

very small. Libor rates have widened to reflect increased

funding pressure concerns stemming from the sovereignrisk crisis, and we think that Libor/OIS spreads can

widen further in the UK as peripheral spreads widen as per our forecast.

Historically, the level of 2Y par yields can be well

explained by the slope of the SONIA curve (we use18Mx1M – 6Mx1M), and this has been the case over

most of the 2011 (Exhibit 14). Given our view that base

rates will be on hold for 2012 as QE is increased further,

we expect the SONIA curve to remain at very flat levels,

with a limited amount of term premia priced in. The6Mx1M/18Mx1M curve has traded in a range around the

10bp level over the past few months, and we expect this

dynamic to continue in 1Q12 as we see little chance of

the curve pricing in any BoE rate tightening expectations.We think the forward SONIA curve should not trade

below 0bp (as the BoE will not lower base rates further in our view) which effectively means 2Y gilt yields

should not fall much below the 50bp level. We expect

2Y gilt yields to range-trade around the 50-60bp level

for much of 2012.

Be long 10Y and 30Y duration….

In the belly of the curve, 5Y and 10Y gilt yields made

all-time lows in 2011, and are far below the levels that

Exhibit 14: 2Y gilts are driven by the slope of the SONIA curve andshould not trade much below 50bp in our view given that the base rateis floored at 50bp

2Y par gilt yield regressed against level of SONIA curve, November 2010 – November 2011; %

y = 0.0004x 2 + 0.005x + 0.58

R2 =98%

0.40

0.60

0.80

1.00

1.20

1.40

1.60

-20 0 20 40 60 80 100 120SONIA curve*; bp

current

* 6Mx1M/18Mx1M SONIA curve.

Exhibit 15: Peripheral spreads are more significant in explaining 10Ygilt yields than macro factors…10Y par gilt yield regressed against 10Y peripheral spread*, UK composite PMI, CPIand QE dummy; monthly data March 2007–October 2011 Factor Coefficient T-Stat

Peripheral Spread, bp -0.01 -12.7

Composite PMI 0.03 3.7

CPI, %oy a 0.24 3.7

QE dummy -0.36 -3.0

R-squared 83%

Std. error, bp 32

* 10Y weighted peripheral spread computed against Germany for Greece, Ireland,Portugal, Italy, and Spain (weighted by the size of their outstanding bond markets).

Exhibit 16: …and the model suggests 10Y yields are some 60bp abovefair value

Actual and model predicted 10Y par yields, monthly data March 2007–November 2011; %

1.00

2.00

3.00

4.00

5.00

6.00

Mar 07 Mar 08 Mar 09 Mar 10 Mar 11

Actual Predicted




Francis DiamondAC

(44-20) 7325-3541


83

macro-economic models would suggest. However, the

low level of yields is less a function of the UK macro

landscape and more a function of the sovereign risk crisisin Europe. We present a simple regression model tohighlight this point (Exhibit 15) and whilst the level of

the composite PMI and level of headline CPI are

important drivers, their statistical significance is much

lower than 10Y weighted peripheral sovereign spreads.We also include a QE dummy to control for periods of

BoE gilt-buying. This model suggests that 10Y yields

are some 60bp above fair value level (Exhibit 16). We

expect 10Y yields to gradually fall to the 1.50% level

over the first half of 2012 as peripheral spreads widen

substantially, with the weighted peripheral spread

expected to reach the 955bp level, from the current

700bp (see Euro Cash for our forecasts). A tightening

back in peripheral spreads should result in a rise in 10Ygilt yields in 2H12.

We note that over recent weeks gilt yields have not fallen

as much as the model suggests given the move in

peripheral spreads, but we think this is reflection of alack of investor appetite into year end to enter fresh

duration positions. We expect the usual linear

relationship between gilt yields and spreads to resume in

the New Year. We think that 5Y yields will become

sticky in a rally and will fall less than 10Y yields. We

expect 30Y yields to outperform, driven by the

ongoing QE gilt purchases and weakening

directionality (See curve section below).

Gilts have acted as a safe haven in recent months as

European sovereign risks have increased, and we expect

this to continue into 2012. In the event that the situationin Europe escalates far enough so that Germany looses its

safe-haven status then Bund yields could rise. In that

environment, we think gilt yields will also rise but may

well outperform bunds due to flight-to-quality flows.

…as we don’t expect fiscal pressures to

drive gilt yields higher

In presenting our duration view, we are assuming that thecurrent sovereign risks do not envelop the UK in the

coming months. Although the UK is one of the worst

AAA countries in terms of its debt dynamics and still hassome way to travel on its fiscal journey, our base case

scenario is that UK gilt yields do not rise materially in

response to fiscal concerns. Through the large-scale

purchases of gilts the BoE appears to be sacrificing long-

term inflation certainty for near-term fiscal stability. QEgilt purchases are expected to rise to £425bn by the end

of 2012, which should limit the potential for substantial

fiscally-driven increases in gilt yields. If fiscal stresses

were to become severe enough then the currency could

also act as a pressure valve (Exhibit 17), limiting the

magnitude of any rise in gilt yields (note our central viewis for relatively stable trade weighted sterling in 2012)

4.

The impact of weaker growth on the UK’s debt/GDP

dynamic is important. Exhibit 18 shows our estimates of the evolution of this metric under several different

economic scenarios. In our central case of 0.5% GDP in

2012 and a slow recovery to the 2% level over thesubsequent few years, debt/GDP (excluding the impact

of financial interventions) will peak just under 77% on a

4 See Global FX Strategy 2012

Exhibit 17: The currency can provide an escape valve when UK fiscalpressures buildTrade-weighted sterling index regressed against PSNB*, yearly 1976–2011

y = -2.7x + 106.3

R2 = 41%

70

80

90

100

110

120

-2.0 0.0 2.0 4.0 6.0 8.0 10.0

PSNB, % GDP

2011 est

* Public Sector Net Borrowing.

Exhibit 18: We expect debt/GDP to reach the 75%–85% level in fiveyears’ timeJ.P. Morgan’s debt/GDP forecasts under various economic scenarios* on a fiscal year basis, dashed line is end of government forecast period; %

60

70

80

90

100

110

2011/12 2012/13 2013/14 2014/15 2015/16 2016/17

Baseline

Severe Recession

Mild Recession

95% lev el

* Baseline is our central GDP forecast; mild recession assumes mildly negative GDP in2012 followed by a modest recovery back to the 2% level; severe recession assumesa repeat of 2008/09, with 2012 GDP at -1.5% and 2013 GDP at -3.5%.




Francis DiamondAC

(44-20) 7325-3541


84

five-year horizon, some 10% higher than the estimate for

FY11/12. In a mild recession scenario where output is

mildly negative for 2012 this peak is predicted to bearound 85% with a steeper trajectory. It would take a

recession of the severity of 2008/09 to push debt/GDP

to close to 100% in five years time, something we

think is not likely at the moment.

If either the baseline or mild recession scenarios play

out then gilt yields will not rise materially due to

fiscal concerns as long as the government maintainsits fiscal commitments and credibility. However, if the

severe recession scenario plays out then we expect rating

agencies to act, not only because of the weaker growth

outlook but also because of an expected increase in

contingent liabilities as the UK government may have to

re-introduce its guarantee programmes, and financialmarkets will likely price an increase in UK sovereign

risk. In this environment safe-haven purchases of giltswould abate and non-domestic holdings of gilts would

probably fall.

In addition, as non-domestic holdings are just less than

one third of the total gilt market which should limit an

impact from foreign investors taking flight from the UK

gilt market. This compares favourably with the main

Euro zone countries where non-domestics own between

40 and 65% of the total bond markets and domesticholdings in the UK are not highly concentrated in the

banking sector (Exhibit 19). It is worth noting thatforeign central bank ownership has been very stable ataround 6-8% of the market and that the increase in non-

residents holdings has come from other investors. In fact,

the increase in other foreign investors (ex central banks)holdings of gilts has been well correlated with gilt

supply, suggesting that non-domestic investor

participation has increased as the size of the gilt market

has expanded (Exhibit 20).

In our view, market pressure is only likely to come to

bear if the government’s fiscal plans lose credibility.

So far the coalition has shown no signs in wavering on its

fiscal tightening commitment but the Chancellor’s self-imposed rules are close to being breached. If output

weakened to the point where fiscal contraction needed to be slowed then we think that the Government can back

load tightening further into the current parliament term.

However, we think that the maximum tightening

achievable in any one year would be 2% of GDP.

Attempting to do any more than this amount would risk

straining the government’s fiscal credibility in our view.

Our base case view is that the UK retains its AAA

rating and stable outlook but rating agency action

Exhibit 19: Non-domestic investors’ holdings of gilts are lower thannon-domestic holdings in the main European bond marketsForeign and domestic bank* holdings of government bonds as % of total marketoutstanding; %

66

81

4538

30

46

611 10 13

33

12

4

44

0

10

20

30

40

50

60

70

80

90

F r a n c e

G

e r m a n y

I t a l y

S p a i n

U K

U S

J a p a n

Non-domestic Domestic bank*

* Excludes Central Bank holdings.

Exhibit 20: Non-domestic investors’ gilt holdings have increased as giltissuance and market liquidity have increasedForeign investor ex-Central Banks gilt holdings regressed against gross gilt issuance,annual data 1976–2011; %

y = -0.0001x 2 + 0.002x + 0.10

R2 = 74%

10%

15%

20%

25%

30%

0 50 100 150 200 250

Gross gilt issuance, £bn

Exhibit 21: Our central view is that the UK’s sovereign rating will not bedowngraded, but should rating agencies put the UK on negativeoutlook, we expect only a modest 20bp rise in 10Y gilt yields

Adjusted 10Y gilt yield* around the time when S&P placed the UK on negative outlookin May 2009; bp

-25

-20

-15

-10

-5

0

5

1015

20

Mar 09 Apr 09 May 09 Jun 09 Jul 09

S&P places UK on negative outlook

* 10Y par gilt = 0.7* 10Y par Bund yield + 0.76 * 12Mx3M (SONIA – EONIA) + 1.2. R-

squared: 82%, std. error: 9bp.




Francis DiamondAC

(44-20) 7325-3541


85

cannot be ruled out if output weakens substantially.

However, if the UK were to be put on review for a

potential downgrade then we think the gilt marketreaction will be muted. When the UK was placed onnegative outlook by S&P on 21 May 2009 the adjusted

10Y gilt yield rose by less than 20bp (Exhibit 21) and

the impact dissipated quite quickly. Downgrades of AAA

sovereigns have typically had a limited impact on yieldshistorically with the Japanese downgrades in November

1998 and February 2001 resulting in less than a 15bp rise

in yields in the week following the move. The USdowngrade earlier this year had a similarly limited

reaction on the day of the announcement with Treasury

yields falling some 30bp in the following week on flight

to quality flows.

Curve view – position for 2s/10s

flattening

The 2s/5s curve is flatter since the start of the year but

has been trading in a 30bp range since mid-August and

has shown little correlation with European peripheral

spreads over this period. We can model the par 2s/5scurve as a function of three variables: 1) the forward

SONIA curve (24Mx1M – 12Mx1M), to capture

changing base rate expectations; 2) 5Y RPI to capture

changing inflation expectations; and 3) a QE dummy

variable to take into account QE gilt purchases (whichhave been in the 3Y+ sector of the gilt curve). The model

fit has been good over the past couple of years (Exhibit22), and currently the 2s/5s curve is fairly valued.

Going forward, we expect this range-trading dynamic

to continue. Under the assumption that QE gilt purchases continue for the next 6 months, we outline

how the curve could change under different assumptions

for the slope of the SONIA curve and the level of 5Y RPI

swaps (Exhibit 23). In our view, the SONIA curve is

unlikely to trade below zero (as we think the chances of base rate cuts are low), and we don’t expect this curve to

trade much above 35bp, given our low-for-long view.

We think 5Y RPI swaps could modestly fall to the 2.50%

level at worst, and with the SONIA curve likely toremain close to current levels, we expect very modest

flattening of the 2s/5s curve. Any steepening is likely to

be limited, and we think the 2s/5 curve can trade in a

50-70bp range over 1H12.

The 2s/10s curve has been highly directional with 10Y

yields when the 10Y rate is below 3.20% (Exhibit 24).

We expect this to remain the case over 2012 as 2Y yieldsare anchored, and we find it hard to envisage a scenario

in which 10Y gilt yields will rise above this 3.20% level

in the next couple of quarters. We don’t expect the

Exhibit 22: The 2s/5s curve can be well explained by the slope of theSONIA curve, the level of 5Y RPI, and QE purchases

Actual 2s/5s par gilt curve vs. model curve*; March 2009–November 2011; %

0.60

0.80

1.00

1.20

1.40

1.60

1.80

Mar 09 Sep 09 Mar 10 Sep 10 Mar 11 Sep 11

Actual Model

* 2s/5s par = 0.25 * 5Y RPI inflation swap + 0.49*OIS curve – 0.18 * QE dummy +0.028. R-squared: 79%, std. error: 10bp.

Exhibit 23: We think the 2s/5s curve cannot flatten much more, and anysteepening will struggle to beat the forwards – we recommend playingthe rangeProjected changes* in 2s/5s curve over the next 6 months, given various levels of 5YRPI swaps and OIS curve**; shaded indicates our expected outcomes; bp

OIS curve**; bp 2.50 2.75 3.00 3.25 3.50

75 15 25 30 35 40

50 5 10 15 25 30

25 -10 0 5 10 150 -20 -15 -10 0 5

-25 -30 -25 -20 -15 -10

5Y RPI swap, %

* Calculated using regression equation in Exhibit 22. Assumes QE buying is in placeover the next 6 months. Current OIS curve = 17bp, 5Y RPI swap = 3.04%.**OIS curve = SONIA OIS 24Mx1m – SONIA OIS 12Mx1M.

Exhibit 24: With the front end anchored, the 2s/10s curve is likely toremain highly directional with 10Y yields during 20122s/10s par curve regressed against 10Y par yields, last 12M, %

y = 0.77x - 0.03

R2 =93%

y = -0.04x + 2.6

R2 = 0%

1.60

1.80

2.00

2.20

2.40

2.60

2.80

2.00 2.20 2.40 2.60 2.80 3.00 3.20 3.40 3.60 3.80 4. 00 4. 20

10Y par rate, %

10Y par > 3.2%10Y par < 3.2%

current




Francis DiamondAC

(44-20) 7325-3541


86

increase in QE to generate an increase in forward

inflation expectations and we think it is too early for this

dynamic to be reflected via a steeper curve. Inflationexpectations are only likely to become unanchored to theupside when excess demand has recovered and the

excess central bank liquidity is being reflected in strong

credit and loan growth. This scenario is several years

away yet in our view. We forecast the 2s/10s curve to

flatten to the 100bp level by the end of 2Q,

substantially more than priced into the forwards.

The 2s/5s/10s fly is typically viewed as a directional fly,

and this has been the case for most of the past 12 months,

although we note that the directionality has weakened

substantially when 5Y par yields have been below 1.5%

(Exhibit 25). This convexity phenomenon is likely due

to investors attributing some degree of risk premia tolonger maturity rates when the base rate is anchored (i.e.

some upward slope) and has been observed in Japan (see Euro Cash) in the past when yields get to very low

levels. We think this convexity phenomenon will remain

in place as 5Y and 10Y yields fall further.

The 10s/30s curve has also been highly directional with

10Y yields (Exhibit 26), although the curve has

persistently traded too flat vs. the level of yields over the

past couple of months. We think this is a reflection of the

resumption of QE gilt purchases in early October as the10s/30s curve, adjusted for the level of 10Y yields, has

been broadly flattening as the BoE has been buyingrelatively more 25Y+ gilts compared with 10-25Y gilts,taking into account DMO supply (Exhibit 27). Any

implied steepening from lower 10Y gilt yields will likely

be gradually offset by the BoE purchasing relativelymore 25Y gilts compared with 10-25Y gilts (in 10Y

equivalent terms).

We note that since QE restarted in October that the BoE

has had to pay relatively more compared with pre-auction levels for 25Y+ gilts and that cover ratios for

buybacks in this sector have been much lower than in the

other two sectors (Exhibit 28). We expect this dynamic

to continue as QE purchases increase. The directionalityof the 10s/30s curve with the level of yields is likely to

become weaker as 10Y yields fall further as investors rollalong the curve in search of higher yields. Overall, these

factors should result in modest flattening in the

10s/30s curve and we target the 75bp level by 2Q,

some 15bp flatter than currently implied by the

forwards.

Exhibit 25: 2s/5s/10s fly directionality has weakened as yields havefallen and is likely to remain low2s/5s/10s par yield fly (50:50) regressed against 5Y par yield, last 24M; %

y = 0.16x - 0.4

R2 =6%

y = 0.26x - 0.61

R2 = 90%

-0.40

-0.30

-0.20

-0.10

0.00

0.10

0.20

0.30

0.50 1.00 1.50 2.00 2.50 3.00 3.50

5Y par yield; %

5Y par < 1.5%

5Y par > 1.5%

Exhibit 26: The 10s/30s gilt curve has been driven by 10Y yields…10s/30s par gilt curve regressed against 10Y par yields, January 2011–November 2011; bp

y = -35.2x + 196.6

R2 = 80%

40

60

80

100

120

140

2.00 2.50 3.00 3.50 4.00 4.50

10Y par yield; %

current

Exhibit 27: …and the excess flattening since October can be explainedby BoE gilt purchases in the 25Y+ sector 10s/30s curve adjusted for level of 10Y gilt yields* vs. cumulative excess of QE giltpurchases in the 25Y+ sector compared with the 10-25Y sector adjusted for DMOsupply **; weekly data 10 October 2011–18 November 2011bp £bn, 10Y equivalents

-24

-22

-20

-18

-16

-14

-12

-10

10 Oct 17 Oct 24 Oct 31 Oct 07 Nov 14 Nov

-4

-2

0

2

4

6

8

10

Cum. ex cess of 25Y+ purchases (£bn**)

Adjusted 10s/30s

* Residual from linear regression of 10s/30s curve vs. 10Y yields over the past 12months, weekly average.** Expressed in £bn of 10Y equivalents, cumulative purchases from 6 October 2011when QE restarted minus DMO supply. Scale inverted.




Francis DiamondAC

(44-20) 7325-3541


87

Market technicals – curve-trading rule

and trading short-dated gilts

The 5s/10s curve has generally been non-directional with

the level of yields for most of the past year, although thisdirectionality has increased in the past couple of months.

We think this sector of the curve is the most likely to be

affected by momentum dynamics as regular DMOsupply, QE purchases and reasonable liquidity can result

in opportunities for investors wishing to express short-

term tactical views.

To this extent, we have developed a momentum-trading

rule model for the 5s/10s par gilt curve, based on a

moving average indicator 5. The intuition is that when the

signal is large and positive, then the curve should flattenin subsequent days, and when the signal is large and

negative, the curve should steepen in subsequent days.

We test a trading rule based on different signal thresholds

for initiating and exiting trades since the start of 2010

and since the start of 2011. A rule that initiates curve

trades when the signal threshold is +/- 2.5 and exitstrades when the threshold is +/- 0.25 has performed the

best over the two periods, generating a total P/L since

2010 and 18bp P/L since 2011 (Exhibit 29), with a

success ratio around 75% in both periods. The

performance of this rule has been lower in 2011compared with 2010 as general risk appetite has fallen as

a result of the worsening sovereign risk crisis. In anenvironment in which both regular supply and BoE

buying continue to take place in the 5Y and 10Y

sectors, we expect the 5s/10s curve momentum model

to continue delivering positive results.

We also look at the behaviour of very short-end gilts

from a technical perspective. As bonds fall below 1Y

remaining maturity, they become less liquid and cease to

become part of most investors’ portfolios. We wouldexpect gilts to start to cheapen vs. surrounding lines at

some point before they fall below the 1Y point. Looking

at examples over the past couple of years and using a

yield spread adjusted for the level of 2Y gilts we find thatthis starts to occur, on average, 50 days before the bonds

become sub 1Y in maturity (Exhibit 30). Gilt 4.5%

Mar13 falls below the 1Y remaining maturity point

5 We first generate an MACD (Moving Average Convergence

Divergence) measure. This is calculated as the expected moving

average (EMA) of the 5s/10s curve over a 5-day period minus the EMA

of the 5s/10s curve over a 25-day period. We then generate a signal

measure which is the 10-day EMA of the derived MACD measure. Thetrading signal is defined as the MACD measure minus the signal

measure.

Exhibit 28: The BoE has had to pay more vs. market levels to buy 25Y+bonds – we think this will continue

Average coverage ratio and average yield of bonds purchased vs. market levels for

QE gilt buybacks conducted from 6 October 2011 onwardsRatio bp

1.5

2.0

2.5

3.0

3.5

3-10Y 10-25Y 25Y+

0.0

0.1

0.2

0.3

0.4 Avg. cover ratio

Yield vs. market

Exhibit 29: A momentum trading rule for the 5s/10s gilt curve hasgenerated positive P/L since 2010Total P/L and number of signals generated by 5s/10s curve momentum model *trading rule ** since 2010 and 2011; bp

18

46

4

11

0

10

20

30

40

50

Since 2011 Since 2010

P/L Total no. s ignals

* We first generate a MACD (Moving Average Convergence Divergence) measure.This is calculated as the expected moving average (EMA) of 10Y yields over a 5-dayperiod minus the EMA of 10Y y ields over a 25-day period. We then generate a signalmeasure which is the 10-day EMA of the MACD measure. The trading signal is definedas the MACD measure minus the signal measure.** The trading rule is calibrated based upon a threshold for the signal at which toinstigate and unwind trades. We use a signal of + (-) 2.5 as the threshold to enter

trades and a signal of + (-) 0.25 as the threshold to close open trades.




Francis DiamondAC

(44-20) 7325-3541


88

towards the end of 1Q12 and we recommend

underweighting this line vs. gilts 8% Sep14 and 2.25%

Mar14.

Position for wider 5Y and 10Y swap

spreads

Sterling swap spreads have widened over 2011, with the bulk of this occurring in the past few months of the year.

10Y swap spreads were relatively stable during 1H11,

trading in an 8bp range from January to July, beforewidening out during 2H11 to reach levels not seen since

mid-2009 (Exhibit 31). The move wider occurred across

the swap spread curve, with 2Y swap spreads widening

the most (53bp up to 18 November 2011), driven by

increased European sovereign risk fears and renewed bank funding concerns. The resumption of QE in October

resulted in 30Y swap spreads moving out of the trading

range that had been in place for most of 2011.

Swap spreads in the 2Y sector have broadly mirrored

the moves in European peripheral spreads, as Libor/OIS

spreads have trebled with the increase in bank fundingconcerns, and we expect this dynamic to continue in

2012.

Further out, 5Y and 10Y swap spreads have generally

been directional with the level of yields, although thisrelationship has started to become non-linear as yields

have reached extremely low levels. The resumption of BoE gilt purchases has also affected swap spreads. Over the past three years, 10Y swap spreads can be reasonably

well explained by the level of 10Y yields, monthly QE

purchases in the 10Y sector net of DMO supply, andimplied volatility (Exhibit 32).

Given our views of lower 10Y gilt yields and an increase

in QE to £425bn, we expect 10Y swap spreads to widen

to around the 55bp level in 1H12 (Exhibit 33) beforeslowly tightening back in 2H12 as yields rise modestly.

Our expected increases in net gilt supply for FY12/13 is

relatively small, and we think that the market is already

expecting a slight worsening in public finances.However, the scope for the government to increase

borrowing is limited, and we expect any tightening

impact on 10Y swap spreads from increased borrowingforecasts from the OBR this year or increased gilt

issuance in FY12/13 to be limited.

30Y swap spreads can be explained by the level of 10Y

swap spreads and net QE buying of 30Y gilts (Exhibit

34). As discussed, we don’t expect the BoE to alter the

maturity split of gilt purchases if QE is extended, and we

expect 30Y swap spreads to gradually widen over 1H12,

Exhibit 30: As bonds approach 1Y remaining maturity they tend tounderperform vs. 18M and 2Y gilts

Adjusted yield curve* for selected 1Y vs. 18M and 2Y bond pairs** in the businessdays around gilts falling below 1Y remaining maturity; bp

-20

-15

-10

-5

0

-90 -60 -30 0 30

Days around bond falling sub 1Y

* Yield spread adjusted for 2Y par gilt yield.** Yield spreads used: Mar11/Mar12 yield spread around 7 March 2010; Dec11/Jun12yield spread around 7 December 2010; Mar12/Mar13 spread around 7 March 2011;Jun12/Mar13 spread around 6 June 2011.

Exhibit 31: Following a period of relative stability, 2Y and 10Y swapspreads have substantially widened back to mid-2009 levels2Y and 10Y maturity matched swap spreads; bp

-50

0

50

100

150

Nov 08 May 09 Nov 09 May 10 Nov 10 May 11 Nov 11

2Y 10Y

QE1 QE2

Exhibit 32: 10Y swap spreads can be modeled as a function of giltyields and net QE purchases10Y swap spread regression model*, monthly data November 2008–October 2011 Variable Coefficient T-stat

10Y y ield, % -9.7 -2.2

Monthly Net QE purchases, £bn 2.3 5.2

3Mx 10Y v ol, bp 3.4 2.2

R-squared 59%

Std error 11

* Maturity matched benchmark gilt swap spread.

Exhibit 33: We expect wider swap spreads across the curve, with 30Yspreads breaking the 0bp levelCurrent maturity matched swap spreads and J.P. Morgan forecasts for 2012

Sw ap spreads 18-Nov -11 1Q12 2Q12 3Q12 4Q12

2Y 96 110 120 110 100

5Y 71 85 95 85 65

10Y 31 45 55 45 25

30Y -9 0 5 0 -10




Francis DiamondAC

(44-20) 7325-3541


89

reaching and then modestly breaking through the 0bp

level. However, we note that 30Y swap spreads globally

are much tighter than their pre-crisis levels (Exhibit 35),and whilst we expect 30Y UK swap spreads to widen, wedon’t expect a move back to the levels seen in 2007,

given the current global backdrop. In addition, 30Y swap

spreads are more affected by investor flows and range-

trading behaviour, and the relative cheapness of linkerson ASW compared with nominal gilts (see Inflation

Linked Markets) could encourage asset swap investors to

shift out of holding nominal gilts into inflation-linkedgilts on ASW. Given these technical factors, we think

that swap spread widening to 0bp will be a slow

process interspersed by bouts of tightening.

Trading themes

• Be long duration in 10Y and 30Y gilts in 1H12

We expect wider peripheral spreads to drive gilt

yields lower and we forecast 10Y gilts to reach

1.50% by 2Q. The fiscal environment is challenging

but we expect ongoing QE gilt purchases and a

relatively low proportion of non-domestic investors

to prevent UK yields from rising due to a spill over

of sovereign risk fears into the UK

• Enter 2s/10s curve flatteners

The curve will likely remain highly directional with

10Y yields. We forecast almost 90bp of flatteningcompared to what is priced by the forwards by the

middle of 2012

• Bias for a flatter 10s/30s curve

We think the 10s/30s curve can flatten despite our

view of lower nominal yields. As gilt yields fall

further we expect the directionality of the 10s/30s

curve to weaken as investors move out the curve in

search of yield. In addition, further QE gilt purchases should support 30Y gilts relative to 10Y

gilts

•

Position for wider 5Y and 10Y swap spreadsSwap spreads are expected to widen further driven

by wider peripheral spreads, lower nominal yieldsand a total of £425bn of QE gilt purchases by the

end of 2012

Exhibit 34: 30Y swap spreads can be broadly explained by the level of 10Y swap spreads and net QE buying in the 25Y+ sector…

Actual 30Y maturity matched swap spread and model predicted 30Y swap spread*;

weekly data, past 12 months; bp

-30

-25

-20

-15

-10

-5

Nov 10 Feb 11 May 11 Aug 11 Nov 11

Actual Predicted

* 30Y ASW = 0.45 * 10Y ASW + 0.62 * net QE buying of 25Y+ gilts – 25.3. R-squared:58%, std error: 4bp.

Exhibit 35: …and we expect further widening but not back to the levelsseen prior to 2008Maturity matched 30Y swap spread for the UK and weighted global 30Y swap spread*;bp

-75

-50

-25

0

25

50

75

Nov 06 Nov 07 Nov 08 Nov 09 Nov 10 Nov 11

UK Global ex UK

* Calculated as the average of US, German and Japanese 30Y maturity matched swapspreads, weighted by total government bond market value.



US Fixed Income Strategy

US Fixed Income Markets 2012 Outlook

November 24, 2011

Srini RamaswamyAC

(1-212) 834-4573

Kimberly L. Harano (1-212) 834-4956J.P. Morgan Securities LLC

90

US Cross Sector

• We look for growth to slow in 1H12 as past fiscal

stimulus programs roll off, but accelerate

thereafter. With the unemployment rate

remaining elevated, core inflation should slow

• Although our baseline forecast does not yet call

for an imminent QE3, we think there is a

significant chance that the Fed resumes balance

sheet expansion in 2012, most likely by buying

MBS

• The European crisis will likely worsen before it

gets better. Key developments in 2012 are likely

to include further funding pressures on Italy, a

Euro-area recession, and increased risks of a

downgrade of France

• The crisis should pressure Treasury yields lower

into 1Q12, but a deteriorating supply/demand

balance, rich valuations, and a decoupling from

Europe will limit the rally

• Despite near-term risks from a worsening crisis,

spread products are likely to benefit from

significant offsets such as solid credit

fundamentals, negative net supply (in securitized

products), and attractive valuations—cautiously

overweight ‘B’ rated high yield bonds, new issue

‘AAA’ and subordinate CMBS relative to legacy

paper, select ABS, and Agencies; stay overweight

EMBIG versus CEMBI

• Overweight mortgages given significant risks of

QE3 and attractive valuations

• Tail risk is not what it used to be: the European

crisis is far from the only “tail” scenario investors

will need to consider in 2012. We characterize

four distinct risks—the European crisis, an EM

Asia hard landing, and deflationary and

inflationary scenarios—and explore the best

hedging strategies

• We find that synthetic strategies designed to

match the duration of liabilities or hedge the

funding gap for defined-benefit pension plans

significantly lower the tracking error and

volatility

The year that was

The past year was an unforgettable one, even if we might

wish we could forget it. It is impossible to forget the

human tragedy in Japan from the earthquake, tsunami

and the nuclear reactor crisis in the first quarter. The

earth appeared to shift under market participants’ feet as

well, with a number of crises coming out of left field.

The year started off strongly, with forecasts for robust

growth pushing equities and bond yields higher and

Exhibit 1: Another wild ride in markets11/17/11 level, quarter-to-date change, quarterly changes over 2011, and changes over 2010, 2009, and 2008 for various market variables

CurrentQTDchg

3Q11chg

2Q11chg

1Q11chg

2010chg

2009chg

2008chg

Global Equities (level)

S&P 500 1216.1 84.7 -189.2 -5.2 68.2 143 212 -565

E-STOXX 2242.8 63.1 -668.9 -62.4 118.1 -172 517 -1952

FTSE 100 5423.1 294.7 -817.2 36.9 8.8 487 979 -2023

Nikkei 225 8479.6 -220.7 -1115.8 61.0 -473.8 -318 1687 -6448

Sovereign par r ates (%)

2Y US Treasury 0.290 -0.006 -0.175 -0.301 0.197 -0.58 0.57 -2.43

10Y US Treasury 2.015 -0.007 -1.238 -0.277 0.148 -0.59 1.09 -1.32

2Y Germany 0.373 -0.135 -1.058 -0.162 0.939 -0.55 -0.46 -2.21

10Y G ermany 1.963 0.024 -1.111 -0.346 0.391 -0.44 0.33 -1.23

2Y JGB 0.132 -0.012 -0.026 -0.023 0.034 -0.02 -0.19 -0.33

10Y JGB 0.895 -0.100 -0.166 -0.143 0.138 -0.17 0.14 -0.28

5Y Sovereign CDS (bp)

Greece 4964 -975 3827 1019 86 721 61 205

Spain 478 94 122 29 -116 235 4 92

Portugal 1173 -42 417 206 101 403 -7 77

Italy 570 96 303 24 -86 126 -58 145

Ireland 763 22 -50 128 57 441 -5 156

Funding spreads (bp)

2Y EUR par swap - par gov't spd 119.6 20.4 38.0 -0.7 -18.3 28.5 -42.3 39.6

2Y USD par swap - par gov't spd 50.9 22.9 5.7 3.8 -3.8 -5.4 -61.2 8.4

EUR FRA-OIS spd 81.4 18.0 34.0 3.1 -7.2 3.3 -39.9 27.0

USD FRA-OIS spd 67.7 24.3 22.6 -0.2 1.7 4.6 -60.9 27.6

1Y EUR-USD xccy basis -88.2 -15.6 -46.5 0.1 22.7 -22.0 22.6 -36.0

Currencies

EUR/USD 1.352 0.009 -0.107 0.031 0.092 -0.103 0.016 -0.056

USD/CHF 0.917 0.011 0.065 -0.073 -0.023 -0.103 -0.012 -0.077

USD/JPY 76.96 0.21 -4.29 -1.73 1.30 -10.81 1.49 -23.00

JPM Trade-weighted USD 81.27 -0.74 4.41 -1.46 -1.71 -3.20 -4.66 6.38

Spreads (bp)

30Y CC MBS L-OAS 47.7 -7.0 17.0 9.6 -2.4 35 -66 59

10Y AAA CMBS spd to swaps 315.0 -50.0 145.0 30.0 -50.0 -260 -375 793

JULI (ex-EM) Z-spd to Tsy 235.6 -13.6 96.7 14.9 -17.4 -11 -363 333

JPM US HY index spd to worst 741.4 -66.6 241.0 50.1 -66.0 -74 -1068 1128

EMBIGLOBAL spd to Tsy 416.6 -48.4 176.5 -10.4 10.3 -6 -430 470

MAGGIE (Euro HG spd to govies) 76.2 9.8 18.1 6.3 -6.8 14 -27 38

US Financials spd to Tsy 325.6 -8.5 153.0 21.1 -19.2 -11 -359 336

Euro Financials spd to gov ies 316.3 18.5 127.9 20.1 -29.4 47 -164 208

10Y AAA muni/Tsy ratio (lev el) 114% 3.8% 25.4% -5.8% -3.0% 18% -47% 36%

30Y AAA muni/Tsy ratio (lev el) 126% 6.0% 21.5% -7.4% -0.9% 18% -104% 97%

Commodities

Gold futures ($/t oz) 1774.30 157.00 106.90 86.60 17.90 313 223 27

Oil futures ($/bbl) 98.82 19.49 -16.09 -11.30 15.34 12.0 34.8 -51.4





November 24, 2011

Srini RamaswamyAC

(1-212) 834-4573


91

causing credit spreads to narrow (Exhibit 1). By the

middle of the year, however, expectations for US growth

plunged sharply as the run-up in commodity prices posed

a threat to growth and the supply shocks from Japan

appeared more severe than anticipated. At the same time,

the sovereign crisis in Europe ballooned from impacting

only smaller, peripheral countries into a much larger

issue, spreading to Spain and Italy, and threatening the

banking system. Third, political deadlock in the USnearly produced a crisis in markets as Congress played a

game of chicken with the vote to increase the debt

ceiling. In part as a result of that political maneuvering,

S&P downgraded the US sovereign rating from AAA to

AA+ with a negative outlook. All of these events led to a

sharp underperformance of risky assets in the second and

third quarter.

With the economy looking vulnerable, the Fed was

forced to ease again, this time launching “Operation

Twist” instead of another round of balance sheet

expansion. Finally, in the fourth quarter, economic data

turned more upbeat, and with some positivedevelopments coming out of Europe after the October 26

summit, US risky asset markets have managed to rally

this quarter. All in all, given crises erupting left and right,

2011 was a year with trading ranges in yields and credit

spreads that were only surpassed by the financial crisis

periods in the US (Exhibit 2).

Given the high volatility, it is unsurprising that investors

were generally not rewarded for taking on additional risk

this year, particularly in high-beta sectors. Unlike years

past, when the riskiest/most illiquid sectors such as

emerging markets and CMBS produced the best returns,

this year CMBS, high yield, and especially equities

sharply underperformed, while longer-maturity US and

German government bonds posted the best returns

(Exhibit 3).

Exhibit 3: In 2011, investors were generally not rewarded for

taking on more risk in high-beta sectorsRolling 3-month total returns (not duration adjusted), averaged over 3/31-10/31/11 versus the standard deviation of those returns for various assetclasses; %

-4

-2

0

2

4

6

0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0

Standard deviat ion of 3M returns; %

EM

HY

JULIMBS

Agy

US 1-3Y

DEM 7-10Y

ABS

CMBS

US 7-10Y

DEM 1-3Y

S&P

Exhibit 2: This crisis-filled year produced trading ranges in yields

and credit spreads that were only surpassed by the 2008-09 crisisand recoveryYearly range in 10-year Treasury yields and JULI ex-EM I-spread to Treasury;bp

0

100

200

300

400

20112010200920082007200620052004200320022001

10Y UST yield JULI

Exhibit 4: Risky asset spreads have been very well correlated tomeasures of European exogenous riskCross-sector spread index* versus average CDS for French banks**;Level bp

50

100

150

200

250

300

350

400

-1.0

-0.8

-0.6

-0.4

-0.2

0.0

Nov 10 Jan 11 F eb 11 Apr 11 Jun 11 Jul 11 Sep 11 Nov 11

Cross-sector spread index

French bank CDS

* Average of 3-year z-scores for 5-year swap spreads, JULI spread to Tsy, JPMHY index spread to worst, 10Y AAA CMBS spread to swaps, 2Y AAA card ABSspread to swaps, and EMBIGLOBAL strip spread to Tsy.** Average of 5-year CDS spreads for Societe Generale, Credit Agricole andBNP Paribas.





November 24, 2011

Srini RamaswamyAC

(1-212) 834-4573


92

Lifting the covers off 2012: What lies

beneath

Although it has been nearly two years since Greece’s

fiscal problems entered the spotlight, the European crisis

continues to impact US markets in highly significant

ways. On a fundamental basis, Europe now risks falling

back into recession, posing a threat to global growth as

well. In addition, the flight-to-quality premium has been

a factor keeping US Treasury yields well below fair

value. Indeed, we have often noted that yields, adjusted

for their usual long-term drivers as well as short-term

technical factors such as positions, have been very well

correlated to measures of European exogenous risk. This

has been true for risky assets as well: as Exhibit 4

shows, our cross-sector spread index has been very wellcorrelated to French bank CDS spreads. As a result,

proxy variables for the European crisis have become part

of our frameworks for thinking about valuations in

various sectors.

Amidst this backdrop of weak fundamentals and high

geopolitical risk, regulatory and litigation risks have also

been growing. In general, the pace of macro-prudential

regulation has been slower than we anticipated, causing

regulatory uncertainty to persist. For example, the

implementation of Dodd-Frank has been slow, and the

rulemaking process continues to suffer from considerable

delays, perpetuating uncertainty. However, severalcritical changes have come to pass: US bank ratings have

been downgraded as a result of lower systemic support,

and the FDIC changed its deposit insurance assessment

to shift more of the burden onto larger banks.

One area where some progress has been made is money

fund reform. Recently, SEC Chairman Mary Shapiro

indicated that the Financial Stability Oversight Council is

close to proposing rules for money fund reform, which

we expect to be implemented by the end of 2012. At this

point, we think the preferred route for regulation will be

a capital buffer requirement, likely in combination with

gating restrictions. However, we caution that in the

current environment of very low yields it will likely be

difficult for money funds to raise capital (see Short-Term

Fixed Income).

Litigation risk around banks has also taken a toll on asset

markets, particularly in securitized product markets

where banks have a large footprint. To see this, we note

that in an environment where banks are very well

capitalized, litigation risk should negatively impact bank

equity prices more than bank credit. Thus, we define our

legal risk metric as the residual of the log of bank equity

prices regressed against the average CDS level for the

USD Libor bank panel. (We multiply the residual by -1

such that higher values of our index indicate heightened

legal risk.) As Exhibit 5 shows, our index of legal risk

has been well-correlated to MBS spreads, with MBS

cheapening when legal risk rises.

All of these factors will continue to drive markets in

2012. In particular, Europe will very much continue to be

a factor. To help gauge where we are in the crisis, we

find it is interesting to compare the European crisis with

the 2007-08 financial crisis in the US and in Japan in the

early 1990’s. Although the events that may have sparked

the crises are not exactly the same (the bursting of real

estate bubbles in the case of Japan and the US, and the

bursting of a sovereign credit bubble in the case of

Europe), the subsequent propagation and contagion, as

well as the eventual measures that will be necessary to

end the crisis, are similar. In Exhibit 6, we present a

combined crisis timeline, which compares events in theJapan banking crisis to similar events in the US financial

crisis and European sovereign crisis.

As in the US and Japan, the European crisis began with

“localized” losses (on subprime mortgages in the US and

on peripheral sovereign bonds in Europe). As bank

capital positions weakened in an environment of

declining risky asset prices, banks began deleveraging.

These actions now appear likely in Europe, as several

banks have indicated that they expect to delever. In fact,

Exhibit 5: Litigation risk for banks has taken a toll on securitizedproductsLegal risk index* versus 1-week average of 30-year MBS Libor OAS;

bp

10

20

30

40

50

60

70

80

-0.20

-0.15

-0.10

-0.05

0.00

0.05

0.10

0.15

May 10 Aug 10 Nov 10 Mar 11 Jun 11 Sep 11

Legal risk index

MBS OAS

* Legal risk index calculated as the residual of the log of the KBW bank stockindex regressed against the average 5-year CDS spread for USD Libor panelbanks** over the past two years (1-week averages are used). The residual ismultiplied by -1, such that increasing values of the index indicate heightened risk.** Excludes Bank of Nova Scotia.





November 24, 2011

Srini RamaswamyAC

(1-212) 834-4573


93

our European bank credit analysts estimate that the 28

large banks they cover will reduce assets by anywhere

from around €830bn to almost €2tn over the next twelve

months in order to raise capital ratios (see “The Great

Bank Deleveraging: Banking Sector Outlook 2012,”

Roberto Henriques et al, 11/4/11).

Eventually, forced recapitalization through public capital

was employed in the US and Japan. Although the EBA

has already mandated that some banks raise €106bn of

capital (rather than de-lever) to improve their capital

ratios, we think another round of capital raising may beneeded given that the EBA’s estimate of capital needs is

nearly €200bn below our estimates. Moreover, definitive

public programs to enforce recapitalization (rather than

deleveraging) have yet to be announced in Europe.

However, in the US case, bank recapitalization alone was

not enough to reopen broken primary markets and short-

term funding markets. The Fed established the CPFF to

provide liquidity for CP issuers, and supported

securitized product markets through TALF. In addition,

Treasury established a guarantee program for money

market funds, and the FDIC established the TLGP to

temporarily guarantee the senior debt of FDIC-insured

institutions. Although some guarantee and deposit

insurance schemes from the 2007-08 financial crisis are

still in play in Europe, comprehensive programs

spanning the Euro area have yet to emerge, and

policymakers are currently focused on other mechanisms

(such as providing monoline insurance) to allow for

primary issuance by affected sovereigns.

Of course, the rest is history—ZIRP and quantitativeeasing followed suit in the US as well as in Japan. It is as

yet unclear if and when this might occur in Europe, but

the ECB did cut rates this month to 1.25%, and we look

for the ECB to cut rates to 0.5% by June 2012.

One last thing to note from our three-crisis comparison is

the timescales. The pace with which policy measures are

unfolding in Europe is clearly slower than was the case

in the US, but hopefully faster than in the Japanese

experience.

Exhibit 6: Side-by-side comparison of Japan’s banking crisis, the US financial crisis, and the European sovereign crisisTimeline of key events in Japan’s banking cr isis and comparable events in the US financial crisis and ongoing European sovereign c risis

US Euro area

1991-92 Peak in land prices Jul-06 Peak in housing prices Aug-09 Trough in sovereign CDS

1996

Diet passes six laws and establishes Housing Loan

Administration Corporation and Resolution Collection Bank to

cope with liquidation and recovery of assets of failed jusen

and cooperatives. Deposit Insurance Scheme strengthened 10/03/08

Congress passes TARP/EESA to buy troubled assets from

financial institutions 05/10/10

EU and IMF launch a €750bn stabilization scheme with

government-backed loan guarantees and a commitment to

buy European sovereign bonds

Nov-97

Government to guarantee full amount of deposits in yen and

other currencies, bank debentures and certain trusts offered

by trust banks 10/03/08

Under TARP/EESA, deposit limit increased to $250K for

individuals; government still working on plan for unlimited

guarantee for small businesses

(Guarantee schemes from the 2008 financial crisis are still in

place)

Feb-98

Two new laws passed including a provision of ¥30tn to

recapitalize banks 10/14/08 TARP modified to prov ide up to $250bn of capital for banks 10/26/11

EU officials call for a €106bn recapitalization of banks and

plan to leverage the EFSF v ia a monoline insurance scheme

or via a leveraged CDO-style investment vehicle

M ar-98 Government injec ts ¥1. 8tn into large banks 10/ 14/ 08

Government injects $125bn into 9 banks; offers another

$125bn to others Jun-12 Recapitalization of banks to be completed

Late

1998 and

Sep-99

Government offers ¥20tn for credit guarantee schemes, and

then introduces scheme to guarantee bonds of small and

medium-sized compnaies Oct-08

FDIC to temporarily guarantee the senior debt of all FDIC-insured institutions and their holding companies, as well as

deposits in non-interest bearing deposit transaction accounts;

Treasury establishes guarantee program for money market

funds; CPFF created to provide a liquidity backstop for CP

issuers 10/26/11

EU officials discussed exploring a coordinated guarantee

scheme, but nothing concrete has been proposed

Feb-99-

Aug-00 BoJ employs zero interest rate pol icy 12/16/08 The Fed introduces a zero interest rate pol icy

EC B cut the refi rate by 25bp to 1.25% on 11/3/11, and we

expect the refi rate to be cut to 0.5% by June 2012

Mar-01 BoJ introduces quantitative easing 03/18/09

The Fed announces Treasury purchases and increased

purchases of Agency debt and MBS (QE1)

ECB has been purchasing sovereign debt, but it is not

"quantitative easing" since they have been sterilizing

purchases

Jun-03 BoJ starts puchasing ABS (including ABCP )

Oct-Nov

08

MMIFF created to purchase assets from money market

funds; TALF created to provide liquidity to ABS investors

JAPAN

Source: Akihiro Kanaya and David Woo, “The Japanese Banking Crisis of the 1990s: Sources and Lessons,” IMF Working Paper, January 2000. Masaaki Kanno, “How Japan’s bankingcrisis ended in a lost decade,” J.P. Morgan Economic Research note, 10/31/08.





November 24, 2011

Srini RamaswamyAC

(1-212) 834-4573


94

Look for further deterioration in Europe over the

near term

Our global strategists look for the followingdevelopments in Europe in 2012. First, we expect

funding spreads on the semi-peripherals to worsen early

next year. Second, Greece is likely to continue to fail to

meet austerity targets, which could eventually lead to a

more severe debt restructuring. Third, we think the Euro

area will dip into recession, with the peripheral area hit

hardest. This could make it difficult for EU sovereigns to

meet their fiscal and structural reform targets. Fourth, we

expect increasing risks of Portugal and Ireland needing to

renegotiate their funding packages in mid- to late-2012,

which will likely require private sector involvement.

Finally, France is at risk of losing its AAA rating,

especially if the Euro area economy weakens more than

expected.

Now that sovereign debt of many European countries is

too low-rated to remain in portfolios seeking “risk-free”

assets, these bonds will slowly but surely need to find

new homes in risk-seeking investors’ portfolios. The US

experience with GSE debt in 2008-09 (when foreign

investors sharply scaled back on holdings of this asset

class) makes for a useful comparison, and highlights the

staying power of bearish technicals when such shifts are

underway. Moreover, although the Fed was willing to

intermediate the transfer, by stepping in to buy Agencydebt, the ECB has thus far been reluctant to intermediate

in the size required to stabilize spreads. Thus, we remain

bearish on intra-EMU spreads and expect them to widen

going into 1H12, before concerted policy action pushes

them narrower in 2H12.

Separate from our baseline view, it is interesting to note

that our J.P. Morgan Fixed Income Markets Investor

Survey indicates that 33% of investors believe Greece

will exit the EMU by the end of 2012, while 25% of

investors believe a country other than Greece will exit by

the end of 2013 (see Appendix at end of piece for

detailed results). Thus, with all these risks on the

horizon, we expect European developments to remain a

major driver for markets in 2012.

The outlook for the economy and Fed policy

Beyond European factors, the US growth outlook will

also be an important driver. Although the super

committee failed to agree on a deficit reduction proposal,

given that the mandatory cuts will not begin until 2013,

there is considerable uncertainty around the nature of

fiscal policy adjustments (both potential stimulus

packages and potential deficit reductions) that may occur

over the next year.

In our baseline economic forecast, we assume no

additional fiscal stimulus will be passed by the end of

2011. This is roughly in line with the results of our

investor survey, which shows that 49% of investors

believe no additional stimulus will be passed. However,if any fiscal stimulus is passed, we think it is unlikely to

be large enough to completely offset the drag from past

stimulus programs rolling off, which we estimate will

reduce GDP by 1.7% in 2012 on a year-over-year basis.

As a result, we look for growth to slow substantially in

the first half of 2012 and then recover in the second half

(Exhibit 7). With growth likely to weaken, we expect the

unemployment rate to remain elevated, and high resource

slack should push core and headline inflation lower.

Thus, with the economy likely to weaken further, what

will the Fed do? The results of our investor survey show

that 80% of investors believe the Fed will embark on

another round of balance sheet expansion via asset

purchases (QE3) in 2012. We think the next policy action

the Fed will undertake will most likely be the so-called

“Evans plan,” which calls for specifying the economic

conditions that need to be met for the Fed to tighten. That

said, we also think there is a significant chance that QE3

will be deployed, especially in the form of MBS

purchases, if inflation expectations fall enough. Thus,

QE3 remains on the horizon, even if not yet a force in

markets.

Implications for US fixed incomemarkets

As mentioned above, Europe is likely to remain the

biggest driver of US markets in 2012. In Treasuries, we

think the worsening crisis will sustain the flight-to-

quality bid, pushing yields lower in early 2012 despite

already-rich valuations. Some evidence of these flows

can already be seen in data on foreign purchases:

Europeans have increased purchases of Treasuries, likely

reflecting asset allocation shifts out of riskier EU

Exhibit 7: We expect growth to slow substantially in 1H12 due tofiscal drags and then recover in 2H12J.P. Morgan forecast for real GDP growth, the unemployment rate and core CPI

4Q11 1Q12 2Q12 3Q12 4Q12Real GDP (% q/q, saar) 3.0 0.5 1.5 2.5 2.5

Unemployment rate (% ) 9.0 9.0 9.0 9.0 9.0

Core CPI (% q/q, saar) 1.2 1.2 1.0 1.2 1.2





November 24, 2011

Srini RamaswamyAC

(1-212) 834-4573


95

sovereign debt. However, such effects are unlikely to be

permanent. Correlations between US rates and EU

sovereign spreads will decline eventually, and such portfolio rebalancing will eventually wane, likely later in

1H12. As this European demand subsides, the supply-

demand imbalance in US Treasuries is likely to shift in

an unfavorable direction, and we expect poor technicals

and rich valuations to eventually cause yields to rise over

the course of 2012 (see Treasuries).

At the other end of the risk spectrum, emerging markets

will also remain hostage to European developments

thanks to heightened correlations. However, we think a

number of factors make EM a relatively defensive asset

class. In contrast to 2008, EMBIG yields and spreads are

well above those of—say—high grade corporates,making relative valuations attractive. In addition, stable

growth, greater policy flexibility, low refinancing needs,

a diversified investor base and an upward ratings

trajectory are all supportive factors. Thus, we look for 5-

10% total returns in 2012.

High yield investors face a similar backdrop. On one

hand, Europe remains a risk, but on the other hand, we

think default rates will remain low over the next few

years due to strong liquidity and the health of corporate

balance sheets. We project default rates of 1.5% and

2.0% for high yield bonds and loans, respectively, in

2012; even in a US recessionary scenario, we see defaultrates peaking near 5-6%, versus 10-15% peaks in prior

recessions. High carry and modest spread tightening

should produce attractive total return—we expect returns

to reach 9.6% in 2012. Within the sector, however, we

remain biased towards greater tiering, and recommend

underweighting ‘CCC’s versus higher rated credits. In

particular, we find ‘B’s to be the most attractive rating

category as they offer the best balance between

heightened macro risks, sound credit fundamentals and

low default risk.

In higher quality spread product, we recommend

overweights for various reasons. In CMBS, we are

cautiously optimistic given cheap fundamental credit risk

and negative net issuance of private label CMBS. We

favor new issue ‘AAA’s and subordinates relative to

legacy CMBS. Although early 2012 will very likely

experience ongoing spread volatility, over a 12-month

horizon, we look for legacy benchmark ‘AAA’s to

outperform corporates, tightening by 70bp by mid-year.

Similarly, in ABS, we think credit fundamentals will

remain solid despite stubbornly-high unemployment and

look for technicals to remain favorable due to a persistentsupply shortage. Although we believe ABS spreads have

limited incremental tightening potential from here, we do

think some sectors offer value. Specifically, our top picks

in ABS heading into next year in the ‘AAA’ space are

non-prime auto, retail cards and cross-border ABS. In

addition, we like subordinate credit card and auto ABS.

In Agencies, we are overweight—albeit cautiously so.

Agency spreads remain vulnerable to a worsening of the

European crisis, and our investor survey points to a

desire to reduce exposure to Agencies; however, supply

technicals are strong, as net issuance in 2012 will once

again be significantly negative.

In MBS, the prospect of the Fed buying MBS under a

QE3 program is a powerful wildcard, and should limit

the downside in the asset class. Given attractive spreads

currently, we recommend heading into 2012 with an

overweight.

In high grade, the broad outlines of our view are similar

to our view on rates; prospects for a near-term worsening

in Europe keep us underweight in the near term, but the

significant divergence between strong corporate credit

metrics and wide spread valuations will eventually cause

a decoupling, and likely cause spreads to narrow stronglyat some point in the future. We forecast high grade bond

spreads at 175bp at year-end 2012, versus spreads at

235bp today. Within the high grade market, we

recommend underweighting Financials versus Non-

Financials and underweighting high beta sectors and

those most closely tied to Europe.

Contemplating risks in 2012 …

Although we are generally positive on risky assets over a

medium-term horizon in 2012, given that markets face

unusually high risks, we think it is useful to quantify how

much various spreads might move in an environment of weakening growth and a worsening European crisis. To

do this, we present a stylized model for spreads in

various sectors. We use four factors to model spreads

over the past two years. First, we use 4-week average

jobless claims as a proxy for economic conditions.

Second, we use the size of the Fed’s securities holdings

as a proxy for QE. Third, we use the average CDS spread

for French banks to account for the impact of the

European crisis. Finally, we use an interaction term

defined as the product of French bank CDS and our





November 24, 2011

Srini RamaswamyAC

(1-212) 834-4573


96

flight-to-liquidity index to account for the negative

feedback loop between crisis-induced de-risking and

declining liquidity. Exhibit 8 presents the statistics for

our model.

Although our model does not account for the

idiosyncratic drivers of each market, it does allow us tomake broad projections for spreads based on projections

for the underlying drivers, and—even more

importantly—assess the risk exposure of each sector to

the underlying macroeconomic undercurrents. Exhibit 8

shows our model’s fair value estimate, based on current

levels of the independent variables and also assuming the

Fed’s security holdings increase by $500bn under QE3.

In addition, we use the betas from our model to

characterize the risk exposure in each sector to economic

deterioration (proxied by jobless claims worsening by,

say, 25K), the exposure to a worsening crisis in Europe

(proxied by a 100bp rise in French bank CDS spreads),

and the exposure to QE3 (assuming a $500bn increase inthe Fed’s security holdings). To put the different asset

classes on a comparable footing, we divide the projected

changes by the 2-year standard deviation of 3-month

changes in spreads.

Based on these risk exposure metrics, we find that

CMBS, high yield, and equities are most exposed to a

weakening of the economy. In contrast, CMBS seems to

be least exposed to European risks, whereas CDX.IG, US

Financials, and equities are most exposed. Finally,

CMBS and equities appear poised to benefit the most if

QE3 occurs.

… and tail risks

As investors contemplate the tail risks to their portfolios

in the year ahead, the European crisis perhaps looms

largest in their minds; however, we think there are three

other key risk scenarios that also merit focus. These

include a “hard landing” in China and emerging Asia, a

return of deflation fears in the US, and (in the opposite

direction) the potential for high future inflation (or

inflation expectations) as a result of easy monetary

policy now.

In order to assess the impact of these four scenarios on

markets, we first pick proxy variables that would be

expected to move significantly in these crises—variables

that we could view as almost definitional in

characterizing each of the risk scenarios that are to behedged. For example, we use an average of French bank

CDS as a metric for the EU crisis and define our tail risk

scenario with respect to this metric. We use a basket of

EM currencies to proxy the health of EM Asian

economies and in specifying the risk associated with that

scenario. Third, we use 10-year Treasury yields as a

reference variable for our deflation and high inflation

scenarios.

Exhibit 8: Quantifying the risks for spreads

Current level of spreads; our strategists’ targets for mid-year 2012*; statistics for spreads regressed against 4-week average of jobless claims (000s), size of the Fed’ssecurities holdings ($bn), average French bank CDS** (bp), and average French bank CDS multiplied by flight-to-liquidity index*** using weekly data over past 2 years;projection for spreads if $500bn of QE3 occurs; 2-year standard deviation of 3-month changes, and normalized risks****

Beta T-stat Beta T-stat Beta T-stat Beta T-stat R-sq

10Y AAA CMBS spd to swaps 320 250 3.09 8.8 -0.103 -2.4 0.24 1.2 0.365 3.5 71% 317 81 0.95 0.30 -0.64

JULI (ex-EM) spd to Tsy 235 213 0.46 6.4 -0.025 -3.0 0.35 8.1 0.085 3.9 80% 237 31 0.37 1.11 -0.40

JPM US HY index spd to worst 743 700 2.29 10.7 -0.085 -3.3 1.02 7.9 0.270 3.9 82% 789 95 0.60 1.07 -0.45

EMBIGLOBAL spd to Tsy 415 375 0.69 5.0 0.025 1.5 0.54 6.4 0.074 1.8 74% 446 48 0.36 1.12 0.26

US Financials spd to Tsy 346 300 0.77 7.3 -0.018 -1.5 0.63 10.0 0.139 4.3 86% 357 50 0.39 1.26 -0.18

5Y CDX.IG spd 136 118 0.28 6.0 0.004 0.6 0.28 9.6 -0.005 -0.3 76% 142 18 0.38 1.49 0.10

5Y CDX.HY spd 772 650 2.92 9.6 -0.008 -0.2 1.52 8.3 0.228 2.3 78% 811 150 0.49 1.01 -0.03

S&P 500 (points) 1216 1475 -2.13 -8.6 0.182 6.0 -1.16 -7.6 0.044 0.6 81% 1245 86 -0.62 -1.35 1.06

Projection

with QE3 (bp)

Mid-year

target (bp)

Jobless claims Fed sec holdings French bank CDS FTLIxFrench bank CDSCurrent

level (bp)

Econ

risk

QE3

risk

EU

risk

SD

(bp)

* For JULI, CDX.IG, and CDX.HY, mid-year targets are interpolated based on year-end 2011 targets or current levels and year-end 2012 targets. For US Financials, targetis based on the JULI target and the historical relationship between the two spreads. For the S&P 500, this is the year-end 2011 target.** Average 5-year CDS spread for Societe Generale, Credit Agricole, and BNP Paribas.*** Average of the 2-year Z-scores of (i) spread between 3-month Fed funds strip and 3-month T-bill, (ii) market depth, which is calculated as the half the size of the sum of the top three bids and offers for the 5-year hot run Treasury note, averaged between 8:30am and 10:30am daily, and (iii) 5-year muni pre-refunded bond yields spread to 5-year on-the-run Treasury yield.**** Econ risk assumes a 25K rise in the 4-week average of jobless claims. EU risk assumes a 100bp widening in French bank CDS. QE3 risk assumes the Fed’s balancesheet increases by $500bn. Normalized risks are calculated as the projected change in each of these scenarios is divided by the 2-year standard deviation of 3-monthchanges.





November 24, 2011

Srini RamaswamyAC

(1-212) 834-4573


97

Next, we define our tail risk scenarios based on changes

in these variables. For the Europe and EM scenarios, we

define the tail risk as a two standard deviation move in

the reference variables, which works out to a 125bp

widening of French bank CDS and a 2.6 point rise in our

EM currency index. For the deflation risk scenario, we

looked at the beta between 10-year yields and 5Yx5Y

inflation swap rates during periods when both declined

sharply. Then, if we assume that forward inflation swap

rates fall 1.25%, we project that 10-year yields fall 80bp;

this is our specification for the deflationary scenario.

Finally, for the high inflation scenario, we looked at the

relationship between 10-year yields and headline CPI in

1971-74 and 1977-80, two periods when CPI surged over

8%-points. We then estimate that if headline inflation

doubles from current levels, 10-year Treasury yields

should rise 110bp.

In general, the best ways to position for tail risk scenarios

involve taking advantage of heightened correlation in

these crisis periods. Therefore, in order to identify the

best hedges for these four tail risk scenarios, we

calculated empirical betas between a number of market

variables and our proxy variables, both in normal periods

and “stress” periods, which are periods when our proxy

variables moved significantly. Exhibit 9 presents the

results of our analysis, as well as the projected moves for

the various market variables in each risk scenario. We

also show the ratio of the stress move to the normal move

for the market variables that show substantial increases

in correlation in crises.

We can make several observations based on this table.

First, we find that in general, currencies are poor hedges

for crises since “stress betas” were not meaningfully

higher than normal period betas, perhaps reflecting the

impact of FX intervention. Second, for the European

crisis scenario, we find that longs in 10-year Treasuries

and 2s/10s curve flatteners offer attractive “beta pick-

up.” Equities, commodities, and high yield also show

increased correlation in an intensifying European crisis,

but interestingly, high grade credit does not.

Exhibit 9: The tail wagging the dog? Projected market moves in four tail risk scenariosStatistics for various market variables regressed against our proxy variables, in normal periods and stress periods*; projected moves in our risk scenarios**, and ratio of stress move

to normal move

Stress

beta

Normal

beta

Stress

move

Normal

move Ratio

Stress

beta

Normal

beta

Stress

move

Normal

mov e Ratio

Stress

beta

Normal

beta

Stress

move

Normal

mov e Ratio

Stress

beta

Normal

beta

Stress

move

Normal

mov e Ratio

10Y Tsy yields (bp) -0.847 -0.501 -105.8 -62.5 1.69 -20.3 -15.4 -52.2 -39.6 1.32 110 -80

2s/10s Tsy curve (bp) -0.644 -0.171 -80.5 -21.3 3.78 -6.22 -6.02 -16.0 -15.5 46.2 41.2 50.8 45.3 1.12 74.4 41.2 -59.5 -33.0 1.81

10s/30s Tsy curve (bp) 0.021 0.070 2.6 8.8 -1.64 -2.07 -4.2 -5.3 -14.3 -14.9 -15.8 -16.4 4.2 -14.9 -3.4 11.9

S&P 500 (points) -2.22 -1.19 -278 -148 1.87 -99.7 -83.6 -256 -215 1.19 46.9 96.1 51.6 105.7 241 96 -193 -77 2.51

MSCI G7 Index (points) -1.95 -1.15 -244 -144 1.70 -96.4 -82.0 -248 -211 1.18 42.3 83.2 46.6 91.5 206 83 -164 -67 2.47

MSCI EM Index (points) -1.93 -1.30 -241 -163 1.48 -127.5 -109.6 -328 -282 1.16 48.2 88.1 53.0 96.9 188 88 -150 -70 2.13

Nikkei Index (points) -18.1 -10.3 -2264 -1288 1.76 -1024.7 -827.7 -2636 -2129 1.24 337.7 894.2 371.4 983.6 1235 894 -988 -715 1.38

EUR/USD 0.00015 -0.00044 0.019 -0.055 -0.042 -0.046 -0.107 -0.118 0.0157 0.0036 0.017 0.004 0.0563 0.0036 -0.045 -0.003

USD/JPY -0.01176 -0.01156 -1.47 -1.44 -2.487 -1.349 -6.40 -3.47 1.84 2.72 4.08 3.00 4.49 1.48 4.08 -1.19 -3.26

AUD/USD -0.00062 -0.00038 -0.077 -0.047 -0.062 -0.054 -0.160 -0.139 1.15 0.0286 0.0303 0.031 0.033 0.0899 0.0303 -0.072 -0.024

Gold futures ($/t. oz) 1.959 0.933 244.69 116.52 2.10 -49.40 -37.87 -127.1 -97.4 1.30 25.9 -12.5 28.4 -13.8 -2.07 -113.0 -12.5 90.4 10.0

WTI oil futures ($/bbl) -0.149 -0.053 -18.55 -6.57 2.83 -9.59 -7.02 -24.7 -18.1 1.37 9.47 8.01 10.4 8.8 1.18 19.01 8.01 -15.2 -6.4 2.373M fw d copper price ($/m ton) -15.51 -4.12 -1937.45 -514.84 3.76 -714.5 -566.0 -1838.1 -1456.1 1.26 378.8 614.0 416.7 675.4 1131.9 614.0 -905.5 -491.2 1.84

5Y CDX.IG spread (bp) 0.32 0.49 40.27 60.72 31.6 25.2 81.2 64.9 1.25 -19.1 -26.3 -21.0 -28.9 -53.3 -26.3 42.6 21.0 2.03

5Y iTraxx Main spread (bp) 0.44 0.57 54.87 70.70 29.2 23.4 75.2 60.2 1.25 -20.9 -28.4 -22.9 -31.3 -43.6 -28.4 34.9 22.7 1.53

5Y CDX.HY price (points) -0.098 -0.069 -12.26 -8.65 1.42 -2.70 -4.23 -6.95 -10.89 4.21 5.30 4.63 5.83 4.05 5.30 -3.24 -4.24

European crisis EM Asia hard landing High inflation Deflation/low y ields

* For the European crisis, the stress beta is calculated as monthly changes in market variables regressed against monthly changes in average French bank CDS*** over 7/1/11-9/10/11, while the normal beta is calculated over the past 5 years. For the EM Asia hard landing scenario, the stress beta is calculated as weekly changes in market variablesregressed against weekly changes in an EM currency basket**** over 8/1/08-11/30/08 and 8/1/11-10/31/11, while the normal beta is calculated over the past 5 years. For the highinflation scenario, the stress beta is calculated as week ly changes in market variables regressed against weekly changes in 10-year Treasury yields (%) over 5/16/07-6/22/07,5/5/09-8/7/09, and 6/10/11-6/29/11, while the normal beta is calculated over the past 5 years. For the deflation/low yield scenario, the stress beta is calculated as weekly changes inmarket variables regressed against weekly changes in 10-year Treasury yields (%) over 4/29/10-9/1/10 and 7/25/11-10/3/11, while the normal beta is calculated over the past 5years.** For the European crisis scenario, we assume the French bank CDS average rises 125bp. For the EM As ia hard landing scenario, we assume the EM FX basket cheapens 2.6points. For the high inflation scenario, we assume 10-year Treasury yields rise 110bp, and for the deflation scenario, we assume 10-year yields fall 80bp. Then we calculateprojected moves using both the stress beta and normal beta.*** Average 5-year CDS spread for Societe Generale, Credit Agricole, and BNP Paribas; in bp.

**** Average of USD/BRL, USD/INR and USD/MXN.





November 24, 2011

Srini RamaswamyAC

(1-212) 834-4573


98

Third, in the EM Asia hard landing scenario, we find

increased correlation for 10-year Treasuries, equities and

commodities, but not for high yield. This scenario is the

only one in which currencies serve as effective hedges:

the yen typically strengthens as Japanese investors

unwind long-EM trades, and the Aussie dollar typically

weakens along with other EM/commodity currencies.

Fourth, we find that hedging against high inflation is

difficult, with very few market variables displaying

increased correlation in this scenario. Only 2s/10s curve

steepeners and longs in gold or oil show up as attractive

hedge trades. In contrast, several market variables show

increased correlation in a deflation scenario, with selling

equities and commodities among the most attractive

trades. It is also interesting to note that the beta between

equities and yields in a deflation scenario is higher than

the beta in inflation scenarios. This is likely because a

weak growth/deflation scenario is clearly negative for

equities, whereas the effect of high inflation on equities

is mixed—on one hand, equities are real assets, so equity

prices should rise with inflation, but on the other hand,

increased risks of Fed tightening should constrain prices.

Now that we have defined our scenarios and our

projected market moves, in order to evaluate which

hedges are currently most attractive, we look at digital

and dual digital option prices for these market variables.

Digital options, which give a payoff of 1 if a certain

condition is met, are insightful because they may be

interpreted as the market-implied probability of a given

scenario. In Exhibit 10, we show indicative prices for

several digital options and dual digital options that are

potential hedge trades in our four scenarios. We have

chosen the strikes for the single digital options based on

the expected “stress moves” of market variables, while

the strikes for the dual digital options are based on the

“normal moves” as implied by Exhibit 9 and reflect

moves that are highly likely to occur in a given risk

scenario. The rationale for choosing less out-of-the-

money strikes in dual-digitals is this: although each move

may be more probable individually, as well as jointly in a

given tail risk scenario, the dual digital can often trade

Exhibit 10: Indicative prices for digital and dual digital options that we expect to pay out under our stress scenariosInstrument, strike*, payoff if market is below or above the strike, and indicative premium for 1-year digital or dual digital option, with USD payouts (where applicable); the

most attractive trades for a given panel are highlightedEUROPEAN CRISIS

Instrument Strike Payoff if Premium

10Y Tsy yields (bp) -105 Below 17.6%

2s/10s Tsy curve (bp) -80 Below 11.4%

S&P 500 (pts) -275 Below 22.0%

Gold futures ($/t. oz) 245 Above 30.0%

WTI futures ($/bbl) -18.5 Below 35.3%

Copper fut. ($/m ton) -1950 Below 33.5%

Instrument #1 Strike Pay off if Instrument #2 Strike Pay off if Premium

Gold futures ($/t. oz) 115 Above S&P 500 (pts) -145 Below 15.6%

WTI futures ($/bbl) -6.5 Below S&P 500 (pts) -145 Below 23.8%

Copper fut. ($/m ton) -515 Below S&P 500 (pts) -145 Below 23.5%

EM ASIA HARD LANDING



Nikkei Index (points) -2600 Below 10.7%

USD/JPY -6.5 Below 19.0%

Gold futures ($/t. oz) -125 Below 50.0%

WTI futures ($/bbl) -25 Below 29.1%



Nikkei Index (points) -2100 Below USD/JPY -3.5 Below 12.5%

Nikkei Index (points) -2100 Below Gold futures ($/t. oz) -95 Below 11.3%

Gold futures ($/t. oz) -95 Below 10Y Tsy y ields (bp) -40 Below 25.0%

Copper price ($/m ton) -1450 Below USD/JPY -3.5 Below 23.0%

HIGH INFLATION


Gold futures ($/t. oz) 30 Above 46.0%

WTI oil futures ($/bbl) 10.5 Above 44.7%

10Y Tsy yields (bp) 110 Above 13.3%


Gold futures ($/t. oz) 15 Abov e WTI futures ($/bbl) 9 Abov e 34.6%

DEFLATION/LOW YIELDS


2s/10s Tsy curve (bp) -60 B elow 15.4%

S&P 500 (pts) -190 Below 28.1%

WTI futures ($/bbl) -15 Below 37.5%


CDX.IG spd (bp) 40 Above 42.0%



S&P 500 (pts) -75 Below WTI futures ($/bbl) -6.5 Below 27.8%

S&P 500 (pts) -75 Below Copper fut. ($/m ton) -490 Below 28.9%

S&P 500 (pts) -75 Below CDX.IG spd (bp) 20 Above 42.0%

* Strike is ATM spot + indicated value for equity indices, and ATM forward + indicated value for all others.

Note: Swaptions are used instead of Treasury options.





November 24, 2011

Srini RamaswamyAC

(1-212) 834-4573


99

cheap because the correlation is underpriced in normal

markets. As such, cheap dual digitals—based on less

extreme strikes—will give an indication of which pair-

wise correlations are underpriced currently in markets,

relative to what can be expected in a tail risk scenario.

For example, our analysis projects that the S&P 500 will

fall 278 points if the European crisis worsens

significantly, compared to a 148-point fall using normal

betas. One way to hedge for such a move in the S&P 500

is to buy a 1-year digital option that pays out if the S&P

500 falls 275 points. We estimate that the premium on

this option is approximately 22.0%. Alternatively, one

can buy a dual digital option, which pays out only if two

conditions are met. In this case, we use two weaker

conditions, both of which would be very likely to occur

in the risk scenario. For example, one can buy an option

that pays out only if the S&P 500 falls to a certain level

and gold futures rise to a certain level. Since we projectthat the S&P 500 will fall 148 points and that gold

futures will rise $116.5 based on typical betas, we

consider a dual digital option that pays out only if the

S&P 500 falls at least 145 points and gold futures rise as

least $115. Interestingly, even though the dual digital has

strikes that are closer to at-the-money, the premium on

the dual digital option is cheaper—15.6% versus 22.0%

for the single S&P 500 option. This suggests that the

correlation between S&P 500 and gold in a European

crisis scenario as predicted by our analysis is not priced

into markets. Put differently, tail risk is priced in to a

greater extent in the S&P implied distribution, but less

priced into the S&P/gold correlation market.

Indeed, we find that the correlations between

commodities and other variables appear underpriced

versus what our model predicts for all four scenarios. In

each case, a dual digital structure that pairs a commodity

with another variable is cheaper than a single digital

option on that commodity.

We can draw several other conclusions from the table.

First, we find that floors on the 2s/10s Treasury curve

appear to be the most cost-efficient way to position for a

deterioration of the European crisis. The dual digital with

gold and S&P 500 is the second cheapest way to hedge

the European crisis, and is cheaper than all the other

single digitals we considered.

Second, the best ways to position for economic weaknessin EM Asia tend to involve exploiting correlations with

other liquid markets, particularly the Nikkei Index. We

find that puts on the Nikkei are the most cost-effective

way to hedge for EM Asia weakness, followed closely by

dual digitals involving the Nikkei and gold futures or the

Nikkei and USD/JPY.

Third, for the two inflation scenarios, rate hedges are the

most attractive—for the high inflation scenario, outright

payer swaptions are most attractive, and for the deflation

scenario, floors on the 2s/10s Treasury curve are

cheapest. This is likely because changes in inflation

expectations have the greatest impact on rates as opposedto other asset classes.

Exhibit 11 presents a summary of the best trading

themes for each risk scenario.

Trends in supply and demand in fixed

income markets

Net issuance across fixed income markets averaged

around $1.5tn over 2010 and 2011, down from the nearly

Exhibit 11: Summary of the best tail risk hedgesSummary of the best trading themes for each risk scenario based on Exhibit 10Scenario Best trading themes

European crisis 2s/10s curve floors, yield floors, S&P puts

EM A sia hard landing Nikkei puts, USD puts/JPY calls

H igh inflation P ay er sw aptions

Deflation/low y ields 2s/10s curve floors, yield floors, S&P and commodity puts

Exhibit 12: We expect total net issuance across fixed incomemarkets to decline in 2012Historical data and J.P. Morgan forecast for long-term (>1Y) net issuance in2011 and 2012; $bn

Net ($ bn) 2004 2005 2006 2007 2008 2009 2010 2011 E 2012 E

IG corporates 215 169 326 408 103 95 247 353 303

HY corporates -17 -25 15 6 -25 76 151 100 85

EM corporates 54 76 102 117 17 80 150 141 129

Municipals 166 156 173 38 29 108 34 -71 3

Non-Agency MBS 423 616 539 201 -359 -379 -274 -218 -186

Agency MBS -24 97 269 510 519 524 -80 0 -55

CMBS 79 97 165 175 -40 -47 -57 -46 -17

ABS 1 50 38 47 -50 -25 -96 -50 -50

CLOs 25 43 86 76 5 -17 -15 -8 -13

Agency Debt 116 54 86 -64 -39 -19 -119 -162 -185

Treasuries 296 263 177 135 396 1,549 1,611 1,364 1,036

Total 1,334 1,596 1,976 1,649 556 1,945 1,552 1,403 1,050





November 24, 2011

Srini RamaswamyAC

(1-212) 834-4573


100

$2tn peak reached in 2009. Next year, we expect net

issuance to take another leg down, driven primarily by a

sharp decline in net issuance of long-term Treasuries(Exhibit 12). Although we expect gross issuance of

coupon Treasuries to be virtually unchanged in 2012, the

net number will likely fall about $328bn due to increased

redemptions. Muni issuance, on the other hand, is

projected to rebound from -$71 to $3bn. Finally, we

expect net issuance across corporates and securitized

products to remain roughly the same at $520bn and

-$320bn, respectively.

How does our supply projection compare to expected

demand? In Exhibit 13, we present our projections for

the sources of “organic” demand for fixed income assets

in 2012 from six major classes of investors: mutualfunds, foreigners, insurance companies, commercial

banks, pension and retirement funds, and the Federal

Reserve. Clearly, prices in markets will adjust so that

demand will meet supply, but we think it is useful to

estimate the demand flows that might be expected given

the current state of markets.

First, we can think of mutual fund purchases of fixed

income assets as being driven by two factors: on one

hand, as the economy weakens, mutual funds will likely

increase their purchases of bonds, and do the opposite if

economic prospects improve. Second, as front-end rates

fall, investors will likely shift out of money market fundsand into bond mutual funds in order to pick up yield.

Thus, we find that using initial jobless claims as a

barometer for the economy and 3-month T-bill yields as

a proxy for money market rates does a good job of

explaining changes in mutual fund holdings of fixed

income over the past 25 years. If we assume that initial

jobless claims rise slightly as the economy weakens in

1H12 and that T-bill yields remain near current levels,

then we project that mutual funds will add about $230bn

of fixed income assets in 2012, close to the average

annual purchase over 2006-10.

Second, we model changes in foreign holdings of fixed

income assets by regressing annual changes in holdings

versus the annual trade deficit of the US. Given that we

expect the trade deficit to widen modestly next year, to

about $582bn, we project $585bn of net purchases by

foreigners in 2012. This is slightly below the net

purchases we have seen over 2006-10.

Third, we model flows from insurance companies using

the slope of the 10s/30s curve (with a steeper curve

leading to greater purchases) and the backward-looking

performance of the S&P 500 (with positive past

performance leading to lower fixed income purchases).

Given that the 10s/30s curve has averaged around 1.15%

over the past year, and given that quarterly changes in the

S&P have averaged around 140 points, we project that

insurance companies will purchase around $165bn fixed

income assets in 2012. This is substantially higher than

the average over 2006-10, but slightly below the $200bn

Exhibit 13: Expected 2012 fixed income purchases by variousinvestor classesStatistics for regression models for quarterly changes (1-year average) or

annual changes in fixed income holdings for various investor classes, estimatedpurchases based on forecast for variables, expected purchases in 2012, andaverage actual annual purchase over 2006-10 ; $bnMUTUAL FUNDS

Period 3Q86 -2Q11

R-sq 58%

Factor Beta T-stat Forecast

Initial jobless claims (1Y avg; 000s) 0.12 3.7 420

3-month T-bill y ield (1Y avg; % ) -5.82 -6.5 0.11

Intercept 8.22 0.6

Estimated 2012 purchases ($bn) 228

Actual avg purchase over 2006-10 255

FOREIGNERS

Period 1992-2010

R-sq 66%


Annual trade balance ($bn) -0.94 -5.7 -582

Intercept 39.67 0.6

Estimated 2012 purchases ($bn) 585


INSURANCE COMPANIES

Period 3Q86-2Q11

R-sq 55%


1Y lag 10s/30s Tsy curve (1Y avg; % ) 23.46 5.2 1.10

1Y lag 1Y chg in S&P 500 (1Y avg; pts) -0.06 -6.8 140

Intercept 23.35 11.1

Estimated 2012 purchases ($bn) 163 Actual avg purchase over 2006-10 86

COMMERCIAL BANKS

Ex pected 2012 purchases ($bn) 80


PENSION AND RETIREMENT FUNDS

Ex pected 2012 purchases ($bn) 135


FEDERAL RESERVE

Ex pected 2012 purchases ($bn) 0 But some chance of QE3


T OT AL EXPECT ED PURCHASES ($bn) 1191

Note: Insurance companies consist of life insurance and property & casualtyinsurance companies. Pension and retirement funds include private pension

funds, as well as state, local, and federal government retirement funds.Source: Federal Reserve Flow of Funds





November 24, 2011

Srini RamaswamyAC

(1-212) 834-4573


101

and $188bn of purchases seen in 2009 and 2010,

respectively.

For banks and pension funds, we assume that they

maintain their pace of purchases over recent years. Banks

added $83bn of fixed income securities in 2010 and are

on track to add a similar amount in 2011, so we estimate

that banks will purchase around $80bn of fixed income

assets in 2012. Similarly, pension funds have been

steadily adding around $135bn of fixed income assets

each year, so that is our estimate for 2012 purchases.

Finally, our baseline view is that the Federal Reserve will

not (yet) embark on QE3, but there is a significant risk

that it will sometime in 2012, particularly involving

mortgages. All in all, we estimate net organic demand for fixed-income assets based on current conditions to be

about $1200bn in 2012, with additional upside risk due

to the Fed. This is comfortably above our net supply

forecast.

Aggregate measures of net demand and net supply do not

tell the whole story, however. The majority of fixed

income investor classes discussed in our table are natural

buyers of fixed income spread product ; in contrast, net

supply in 2012 will come almost entirely from

Treasuries. With Fed purchases of Treasuries likely to be

zero on a net basis in 2012 (based on our assumption that

if QE3 occurs, it will be in MBS), the Treasury marketwill likely witness a deteriorating supply/demand

imbalance as we move through 2012 (see Treasuries).

Investors’ asset allocation plans for

2012

In our J.P. Morgan US Fixed Income Investor Survey,

we also asked investors about the changes they plan to

make to their positions over the next year. In Exhibit 14,

we show the weighted average response for expected

changes in portfolio currency allocations. Unsurprisingly,

investors generally plan to reduce exposure to European

assets and add exposure elsewhere. In particular, most

investors indicated that they planned to add exposure to

emerging market assets (denominated in local currencies)

and US assets.

Exhibit 15 shows how investors plan to change their

asset class exposures over the next year. According to

our survey, investors generally plan to reduce exposure

to Agency debt and cash, with a smaller fraction

planning to reduce duration risk on net. At the opposite

end of the spectrum, investors generally plan to increase

exposure to high yield and investment grade corporates.

They also plan to add exposure to Agency MBS, perhaps

in anticipation of outperformance due to additional Fed

purchases of mortgages. It is interesting to note that only

a small fraction of investors plan to add exposure to

equities and commodities this year, since real assets like

equities and TIPS were investors’ top picks last year (see

US Fixed Income Markets 2011 Outlook , 11/24/10).

Exhibit 14: Investors plan to reduce exposure to European assetsand add exposure to US and EM local currency assetsWeighted expected change in exposure for assets in various currencies

according to our J.P. Morgan US Fixed Income Markets Investor Survey

-0.07

0.07 0.08

0.16 0.17

-0.10

-0.05

0.00

0.05

0.10

0.15

0.20

EUR ass ets Other DM EM USDassets USD assets EM local FXassets * To calculate the weighted expected change in exposure for each asset class,we assigned values of 1, 0, and -1 to responses of add, maintain, and reduceexposure, respectively, and then we summed the values and divided by thenumber of responses.

Exhibit 15: Investors plan to add exposure to corporate credit riskand reduce exposure to Agency debt and cashWeighted expected change in exposure* for various asset classes according toour J.P. Morgan US Fixed Income Markets Investor Survey

-0.19 -0.19

-0.04

0.040.07 0.09 0.09

0.160.19 0.20 0.20

0.24

0.30

-0.2

-0.1

0.0

0.1

0.2

0.3

A g y

d e

b t

C a s

h

D u r ' n r i s

k

C o m m o

d i t i e s

E q u

i t i e s

N o n - A

g y

M B S

A B S

T I P S

C M B S

E M

A g y

M B S

I G H Y

* To calculate the weighted expected change in exposure for each asset class,we assigned values of 1, 0, and -1 to responses of add, maintain, and reduceexposure, respectively, and then we summed the values and divided by the

number of responses.





November 24, 2011

Srini RamaswamyAC

(1-212) 834-4573


102

ALM strategies for pension funds

Over the first half of 2011, the funding gap for defined

benefit pension funds fluctuated around -$200bn, then it

deteriorated sharply as equities and rates both tumbled in

the second half of the year. As of the end of October, the

funding gap stood at -$398bn, nearly double its value at

the start of the year (Exhibit 16). This dramatic

deterioration highlights the risks of running a significant

asset-liability duration mismatch.

Indeed, the most recent Flow of Funds data on defined

benefit pension fund assets showed that funds had 49%

of their assets invested in equities and mutual funds and

only 38% invested in fixed income securities

(Exhibit 17). If we assume that fixed income assets

mirror the composition of an aggregate index like the J.P.

Morgan Global Aggregate Index (GABI), then given the

38% fixed income allocation, the estimated duration of

assets would be only 1.9 years as of the end of October.

On the other hand, if we regress monthly changes in D.B.

pension fund liabilities against monthly changes in 30-year swap rates and AA rated corporate bond spreads, we

estimate an empirical duration of $128bn per 1%-pt

change in 30-year swap rates (Exhibit 18). Given

aggregate liabilities of about $1615bn currently, this

relationship implies the duration of liabilities is around

7.9 years, substantially greater than the estimated asset

duration of 1.9 years.

One way to address this asset-liability duration gap is to

increase allocations to fixed income, and specifically

long-duration assets, since the duration of pension fund

liabilities is longer than the aggregate duration of the US

fixed income universe. However, the prospects for better

ALM-matching using cash assets may be difficult given

our expectation of low long-end fixed income supply. As

Exhibit 19 shows, long-end gross issuance fell sharply in

2011 due to declines in municipal and high grade

corporate supply. Although we expect long-end gross

issuance to increase in 2012, it will likely remain below

Exhibit 18: Estimated empirical duration of pension fund liabilitiesMonthly changes in top 100 D.B. pension fund liabilities regressed against monthlychanges in 30-year swap rates (%) and monthly changes in JULI AA spread to

Libor (bp); monthly data over 11/08-10/11; $bn

-100

-50

0

50

100

150

-1.2 -1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8

1-mo chg in 30Y swap rate; %

Y = -127.8(swap rate chg) - 1.07(JULI spdchg) + 0.59R-sq = 75%

Source: Milliman 100 monthly Pension Funding Index

Exhibit 19: Issuance of long-term debt will likely increase in 2012,but remain below historical averagesGross issuance of long-term (>10Y) debt, actual and forecast for 2011 and 2012, anddebt currently outstanding with >10Y to maturity*; $bn

2005 2006 2007 2008 2009 2010 2011 E 2012 E

Treasuries 14 41 42 49 143 183 191 194 830

Corporates 67 104 162 95 118 120 85 115 1080

Municipals 334 320 373 332 303 301 135 169 1678

TOTAL 415 465 578 477 564 605 411 478 3588

Current

outstanding

Gross issuance of long-term debt

* For Treasuries, we exclude the amounts held by the Fed. Corporate issuance isfixed-rate only.

Exhibit 16: The defined benefit pension funding gap has nearlydoubled during 2011

Assets minus liabilities for top 100 defined benefit pension plans; $bn

-500

-400

-300

-200

-100

0

Oct 08 May 09 Nov 09 Jun 10 Dec 10 Jun 11


Exhibit 17: Pension fund asset breakdown Assets of private defined benefit pension funds as of 2Q 2011; $bn

Fixed income 848

Tsys 353

GSE-backed securities 121

Corporates/foreign bonds 374

Equities 1095

Other 272

Total assets 2215

Source: Federal Reserve Flow of Funds Table L.118.b. Note “Equities” includesequities and mutual fund holdings.





November 24, 2011

Srini RamaswamyAC

(1-212) 834-4573


103

the average levels seen over 2006-10. Furthermore, the

Fed’s Operation Twist will likely take another $100bn of

long-end Treasuries out of the market in December

through June.

In such an environment of constrained long-term debt,

we think synthetic overlay strategies to increase duration

and spread exposure are attractive. We consider two sets

of strategies. The first set of strategies is designed to

match the duration exposure of the liabilities. This can be

thought of as the “end game” position that pension funds

undertaking ALM strategies aim to reach, and represents

the benchmark for any “new” investments that a pension

fund manager might use. The second set of strategies is

designed to hedge the funding gap itself, and could be

useful to a fund manager seeking to hedge all or part of a pension fund’s current funding gap.

For the liability-matching strategies, we examine three

variations of overlay strategies: 1) buying 30-year

(classic bond) Treasury futures, 2) receiving fixed in 30-

year swaps, and 3) receiving fixed in 30-year swaps and

buying risk (selling protection) in 5-year CDX.IG. We

calculated the monthly returns on these three strategies

assuming monthly rebalancing, whereby the duration and

credit exposures were given by the rolling 3-year betas of

changes in liabilities regressed against changes in 30-

year swap rates and AA rated corporate bond spreads (as

shown in Exhibit 17). We then compared the returns on

these strategies to the actual changes in liabilities, and in

Exhibit 20, we show the monthly tracking error over

January 2010-October 2011. As the exhibit shows, all

three synthetic strategies showed substantially lower

magnitudes of tracking error. Unsurprisingly, the strategy

Exhibit 20: Strategies to match the duration of liabilities hadsignificantly lower tracking error than the actual assetsStandard deviation of the net P/L of three strategies* and of the actual change inassets minus liabilities**; 1/10 – 10/11; $bn

29.5

26.524.5

50.1

20

25

30

35

40

45

50

55

Futures Swaps Swaps + CDX Actual assets

* Net P/L defined as monthly P/L on assets plus any overlays, net of monthlychange in value of liabilities. Strategies match the duration and spread exposureas given by rolling 3-year regressions of monthly changes in liabilities versusmonthly changes in 30-year swap rates and AA corporate spreads to Libor.They involve 1) buying 30-year (classic bond) Treasury futures, 2) receivingfixed in 2Mx30Y swaps, and 3) receiving fixed in 2Mx30Y and buying risk(selling protection) in 5-year CDX.IG. We assume monthly rebalancing.** Change in assets minus change in liabilities for top 100 defined benefit plansas given by Milliman 100 monthly Pension Funding Index.

Exhibit 21: The pension funding gap is exposed to rates, creditspreads, and equitiesMonthly changes in top 100 D.B. pension fund funding gap regressed againstmonthly changes in 30-year swap rates (%),S&P 500 (points), and JULI AA spreadto Libor (bp); monthly data over 11/08-10/11; $bn

-150

-100

-50

0

50

100

-1.5 -1.0 -0.5 0.0 0.5 1.01-mo chg in 30Y swap rate; %

Y = 134.9(swap chg) + 0.61*(S&P chg) +1.36(JULI spd chg)+3.95R-sq = 85%


Exhibit 22: Strategies to hedge the funding gap substantiallylowered the volatility of the gapStandard deviation of monthly changes in gap plus P/L on hedge strategy* and of actual monthly change in funding gap**; 1/10-10/11; $bn

25.1

28.1

50.1

20

25

30

35

40

45

50

55

Swaps + C DX+ S&P Fut ures + C DX+ S&P C hg in gap

* Strategies hedge the duration, equity and spread exposure as given by rolling3-year regressions of monthly changes in the funding gap versus monthlychanges in 30-year swap rates, the S&P 500, and AA corporate spreads toLibor. They involve 1) buying 30-year (classic bond) Treasury futures or receiving fixed in 2Mx30Y swaps, 2) buying risk (selling protection) in 5-year

CDX.IG, and 3) selling S&P 500 futures. We assume monthly rebalancing.** Change in assets minus change in liabilities for top 100 defined benefit plansas given by Milliman 100 monthly Pension Funding Index.





November 24, 2011

Srini RamaswamyAC

(1-212) 834-4573


104

involving swaps and CDX.IG had the lowest tracking

error, since it hedged both the duration and credit

exposure of the liabilities. The outperformance of thestrategy using swaps versus the strategy using futures

may be in part explained by the fact that we used 30-year

swap rates in our models to determine the duration

exposure, while the bond futures contract (currently the

most liquid long-end Treasury futures instrument) creates

exposure to the 15-year point.

To hedge the funding gap, we first have to identify the

drivers of the gap. Exhibit 21 shows that changes in the

funding gap have been well explained by changes in 30-

year swap rates, AA rated corporate bond spreads, and

the S&P 500, with increases in these factors leading to an

improvement in the gap, and decreases leading to aworsening of the gap. Thus, the strategies to hedge the

funding gap involve positions with the opposite

exposures, i.e., a long duration position (either via 30-

year swaps or bond futures), a long risk position in

CDX.IG, and a short in S&P futures. We calculated the

returns on these strategies assuming monthly rebalancing

based on the betas from rolling 3-year regressions, and

we combined them with the actual changes in funding

gap. As Exhibit 22 shows, using these hedge strategies

significantly lowered the monthly volatility of the

funding gap. As with the first set of strategies, the

outperformance of the strategy using swaps is likely

partly because 30-swap rates are used in our empiricalestimation of the duration exposures.

Cross sector trading themes

• Stay long duration early in 2012 targeting 10-year

yields to reach 1.70% in 1Q12

A worsening of the European sovereign debt crisis

should accelerate the flight-to-quality bid into

Treasuries early next year pushing yields lower; rich

valuations, a deteriorating supply/demand imbalance,

and poor technicals will limit the upside, however, and

keep rates range bound later in 2012.

• Look for TIPS breakevens to narrow sharply over

the next few months and then widen modestly over

2012

Breakevens will hit a number of headwinds in early

2012: we expect nominal yields to plummet, fiscal

policy to tighten, and headline inflation to fall. Thus,

our fair value model projects that breakevens should

narrow over the next few months, though breakevens

should widen over the remainder of the year as

nominal rates rise. QE3 also poses an upside risk to

our targets for breakevens.

• Position for wider 10-year maturity matched swap

spreads in the near term, but look to initiate

narrowers towards the end of 1Q12

In the near term, FRA-OIS spreads will likely

continue to widen, and banking stock valuations will

likely remain under pressure, both of which should

pressure 10-year swap spreads wider. We expect 10-

year swap spreads to hover near 27bp towards the end

of this year and in early 1Q12. Beyond that point,

issuance-related swapping is likely to pick up, and the

growing likelihood of concerted action by central

banks should help spreads narrow. Look for narrowing

to 18bp by mid-year.

• Stay long gamma going into year end and in 1H12

A persistent crisis in Europe and poor risk appetite

should cause market depth to stay depressed for much

of 4Q11 and 1H12, creating conditions favorable for

long gamma positions.

• Modestly overweight Agency bullets tactically

hedged with swap spread wideners

With 5-year Agency spreads wider than our T+33bp

fair value target for 2Q12, we recommend an

overweight position. However, given the significant

impact of the European crisis on Agency spreads, werecommend tactically overlaying swap spread

wideners as a hedge. In particular, we like adding to

front-end bullets (2- to 3-year) on dislocations. We

also recommend that investors use short-dated

callables to enhance excess returns versus lockout-

matched bullets. Given greater clarity around short-

term rates due to the Fed’s low-for-long policy, we

would look at structures with lockouts as far out as 12-

15 months.

• Modestly overweight MBS going into the new year

We recommend beginning the new year with an

overweight given solid fundamentals and the risk of

QE3 in mortgages. Technicals should also be

supportive given negative net supply. However, we

only look for 10bp of tightening in the basis due to

lower leverage and higher capital requirements.

• Add CMBS versus corporates

We position relatively defensively and favor new issue

‘AAA’s and subordinates relative to legacy CMBS.

For investors looking for yield in legacy paper, add

exposure to higher-quality 2005/2006 AJs but avoid





November 24, 2011

Srini RamaswamyAC

(1-212) 834-4573


105

2007 AJs given higher downside in worse-than-

expected economic scenarios. While early 2012 will

experience ongoing spread volatility, over a 12-monthhorizon, look for legacy benchmark ‘AAA’s to

outperform corporates, tightening by 70bp by mid-

year.

• We are neutral on ABS

ABS spreads likely have limited incremental

tightening from here and tiering will be pronounced.

Within ABS, our top picks in ABS heading into next

year in the ‘AAA’ space are non-prime auto, retail

cards and cross-border ABS. In addition, we like

subordinate credit card and auto ABS.

• Stay underweight high grade corporates but look for spreads to tighten later in 2012

Although we are currently underweight, we look for

spreads to tighten sharply at some point in the future

given the divergence between strong corporate credit

metrics and wide spreads. We forecast high grade

bond spreads at 175bp at year-end 2012, versus

spreads at 235bp today. On a sector basis, we

recommend underweighting Financials versus Non-

Financials and underweighting high beta sectors and

those most closely tied to Europe.

• Remain bearish on CDX in the near term but look

for spreads to tighten over the longer term

Similar to high grade cash markets, we expect strong

credit fundamentals and technical forces to drive

spreads tighter. At year-end 2012, we expect CDX.IG

to trade at 100bp (versus 136bp today) and CDX.HY

at 550bp/$98.0 (versus 772bp/$90.0 today), in line

with our 175bp year-end HG bonds forecast. In

addition, we believe that investors should take

advantage of opportunities where CDS-bond basis is

more negative than -75bp in HG and -100bp in HY.

• Overweight ‘B’s, be neutral on ‘BB’s, and

underweight ‘CCC’s in high yield

We think default risk is likely to be negligible over the

next few years due to strong liquidity and the health of

corporate balance sheets. Thus, we forecast high-yield

bond and loan spreads will tighten to T+705bp and

L+660bp by year-end 2012. In particular, we find ‘B’s

to be the most attractive rating category as they offer

the best balance between heightened macro risks,

sound credit fundamentals and low default risk.

• Stay overweight Senior CLOs (AAA/AA) but

neutral CLO Mezzanine to Subordinates

(A/BBB/BB)Given the macro risks and lack of liquidity, we stay

neutral CLO Mezzanine to Subordinates (A/BBB/BB)

but overweight Seniors (AAA/AA). Trading themes

include contraction upside as reinvestment periods

end, spread tiering in mezzanine bonds, and some

opportunities to take advantage of the high current

carry in equity

• Look for muni ratios to fall and for spreads to

tighten in early 2012

We look for ratios to fall in early 2012, but remain

elevated versus historic norms due to strong

reinvestment capital, high volatility, modest inflows,and some flight-to-quality bid. Over the first half,

spreads may also tighten as yields remain ultra low

and both market returns and the credit environment

remain unremarkable. As a relative value trade, we

like overweighting ‘A’ and ‘BBB’ rated spreads given

investors’ grab for yield and wide ‘A’ and ‘BBB’

spreads.

• Remain marketweight EMBIG but underweight

CEMBI

EMBIG remains the more defensive asset class with

cash flows of $64bn next year versus projected

sovereign issuance of $60bn. Within the EMBIG,underweight EMEA EM sovereigns (through

underweights in Croatia and Ukraine) as we see the

much lower growth environment due to a European

recession contributing to a higher level of sovereign

stress in this region. We are also underweight

idiosyncratic downside risks in the Middle East, where

we see continued political instability and worsening

fundamentals as underpriced; we are underweight

Egypt (through quasi-sovereign Nile Finance) and

Lebanon.





November 24, 2011

Srini RamaswamyAC

(1-212) 834-4573


106

Appendix: Results of the J.P. Morgan Fixed Income Markets Investor Survey

Yes 80%

No 20%

None 49%

$1-100bn 19%

$100-200bn 19%

>$200bn 14%

Yes 52%

No 48%

Yes 33%

No 67%

Yes 25%

No 75%

<1.5% 6%

1.5% - 2.0% 22%

2.0% - 2.5% 48%

2.5% -3.0% 20%

3.0% -3.5% 5%

>3.5% 0%

1) Do you think the Fed will expand its balance sheet via

large-scale asset purchases (QE3) in 2012?

2) How much new fiscal stimulus do you think Congress will

pass by the end of the 2011?

3) Do you think any ratings agencies will lower the US

sovereign rating by one or more notches by the end of 2012?

4) Do you think Greece will exit the EMU by the end of

2012?

5) Do you think any country other than Greece will exit the

EMU by the end of 2013?

6) Where do you expect 10-year yields to be at the end of 2Q

2012?

Reduce M aintain Add N /A

USD assets 2% 51% 18% 29%

EUR assets 16% 27% 9% 47%

Other developed markets 2% 35% 9% 54%

EM local currency assets 4% 17% 20% 59%

EM USD denominated assets 2% 21% 9% 68%

Reduce Maintain Add N/A

Duration risk 27% 38% 23% 13%

TIPS 7% 27% 24% 42%

Agency debt 21% 30% 2% 47%

Agency MBS 11% 17% 31% 41%

Non-agency MBS 7% 30% 17% 46%

CMBS 6% 24% 24% 46%

ABS 7% 28% 17% 48%

IG Corporates 9% 19% 33% 39%

High Yield 2% 17% 31% 50%

EM 4% 17% 24% 56%

Equities 7% 18% 15% 60%

Commodities 4% 13% 7% 76%Cash 22% 31% 4% 43%

I already have 11%

Yes 4%

No 37%

N/A 48%

7) How do you expect to change your portfolio currency

allocations over the next year?

8) Are you looking to reduce, maintain or add exposure to the

following asset classes over the next year?

9) In 2012, are you planning to modify your benchmark to

exclude Fed holdings of Treasuries or MBS?





November 24, 2011

Terry BeltonAC

(1-212) 834-4650

Meera Chandan (1-212) 834-4924

Kimberly L. Harano (1-212) 834-4956

J.P. Morgan Securities LLC

107

US Treasuries

• A worsening of the European sovereign debt crisis

will accelerate the flight-to-quality bid into

Treasuries early next year; rich valuations, a

deteriorating supply/demand imbalance, and poor

technicals will limit the upside, however, and keep

Treasury rates broadly range bound in 2012

• Foreign investors will need to double their

Treasury purchases next year to clear the market;

while the bid from the EU has been strong,

demand from EM has weakened as they have

slowed their accumulation of USD reserves• Disintermediation of money funds, limited supply,

and Fed policy support lower front-end yields

• With poor liquidity, whipsaw risk in the Treasury

market will be elevated next year; we discuss the

importance of avoiding consensus trades in 2012

and highlight a measure of investor positioning

that helps identify turns in Treasury rates

• Positive carry yield curve trades in the front end

are likely to perform poorly given negative curve

convexity; we favor negative carry curve trades

anchored in the front end and positive carry

trades anchored in intermediates

• Synthetic Treasuries created by asset swapping

cheap foreign bonds should outperform in 2012

• Reconstitutions of STRIPS should be strong in

2012, causing short maturity Cs richening versus

similar maturity Ps

• The US fiscal outlook improved in 2011, although

debt levels are still on an upward trajectory. We

expect ratings agencies to maintain their existing

ratings through 2012; another $1.8tn of deficit

reduction is still needed for the US to achieve fiscal

sustainability

Fall from grace

Treasury rates fell sharply in 2011 with 10-year yields

reaching an all-time low of 1.70% in 3Q11. With great

irony, the record low occurred in the aftermath of a major

fall from grace; for the first time ever, the US lost its

‘AAA’ sovereign rating as S&P downgraded the US in

recognition of the sharp deterioration in US debt metrics

over the last few years. Despite a worsening sovereign

credit outlook, Treasury rates moved sharply lower,

with 10-year yields falling 135bp since the start of the

year to their current level of 1.96%. With front-end

yields approaching the zero-rate boundary, the rally

also caused the curve to flatten. Year to date, 2-year

yields have declined 30bp, 5-year yields have declined

110bp, and 30-year yields have declined 140bp.

Exhibit 2: The Fed’s biggest impact this year was arguably on 30-year rates; the long end flattened 50bp after the Fed began signalingOperation Twist in August10s/30s Treasury yield curve; %

0.9

1.0

1.1

1.2

1.3

1.4

1.5

Feb 11 May 11 Aug 11 Nov 11

Exhibit 1: Yields tracked US economic data in 1H11 beforedecoupling in 2H11 as the European debt crisis moved to the core

Average Italy and France CDS spreads (bp) versus 10-year Treasury yields (%)and 3-month rolling average of J.P. Morgan EASI*bp %

50

100

150

200

250

300

350

400

450 1.5

2.0

2.5

3.0

3.5

4.0


10-year Treasury yieldsItaly / France CDS spreads (inverted)EASI

* EASI is the J.P. Morgan Economic Activity Surprise Index and measures thenumber of positive surprises minus the number of negative surprises divided bythe total number of data releases over the past six weeks. This index has beenscaled to fit on the chart as 3.2+J.P. Morgan Economic Surprise Index/10.





November 24, 2011

Terry BeltonAC

(1-212) 834-4650




108

The downtrend in rates in 2011 masks two very different

dynamics driving the Treasury market this year. In 1H11,

rates were comparatively range-bound, with yields largely

driven by changing perceptions around the US economic

outlook. Increased optimism on better growth following

the payroll tax cut helped push 10-year yields to the high

of the year (3.72%) in 1Q11; these moves were reversed

in 2Q11 as economic data disappointed to the downside.

During both quarters, volatility was comparatively low,

with rates mostly trading within 30bp of their 1H11

average of 3.30%.

The dynamic in 2H11 was quite different, as the European

debt crisis spread from the smaller peripherals to the much

larger bond markets of Italy and core Europe. In response,volatility spiked, Treasury rates plunged, and correlations

with European sovereign spreads moved to extreme levels

(Exhibit 1). Despite some analysts arguing that the US

was in recession, the 2H rally actually occurred as

economic data surprised to the upside, highlighting that

the rally primarily reflected a strong flight-to-quality bid

into Treasuries, as well as unwinds of bad positions.

The Treasury market was also supported by Fed policy

this year, with stimulus coming from both $640bn of asset

purchases under QE2 and another $175bn (10-year

equivalents) of duration buying in 2H11 as part of

Operation Twist. The overall impact on yields from allthat buying was small, however; most estimates including

our own suggest that QE2 lowered intermediate rates by

15-30bp (see US Fixed Income Markets Weekly, July 15,

2011). Surprisingly, Fed policy this year likely had its

biggest impact on the long end of the curve; after

steepening nearly 50bp during the first seven months of

the year, the 10s/30s yield curve flattened all that and

more after the Fed began signaling Operation Twist in

August (Exhibit 2).

Fed policy was also modestly effective in helping reduce

term premium in the front end of the curve. Two-year

rates fell sharply after the August FOMC meeting as theFed adopted a new communication strategy and clarified

its intention to keep the funds rate near zero until mid-

2013. Since then, 1-year forward 3-month OIS has

averaged 11bp compared to a 24bp average in the 1-month

period leading up to the August FOMC meeting

(Exhibit 3). To be sure, low front-end yields are also

being supported by a collateral shortage in the Treasury

bill market, and by the European-led flight to quality into

Treasuries.

Finally, we note that the large decline in Treasury rates in

2011 occurred in an environment where long-term

inflation expectations remained relatively stable. For the

first time since 1997, 10-year Treasury yields moved

significantly below long-term inflation expectations, thus

implying negative 10-year real returns. With the exceptionof the UK, these real yields compare unfavorably to

virtually every other developed bond market in the world,

including Japan, where real rates are currently 140bp

above those in the US (Exhibit 4). Such low real rates

highlight both the strength of demand for Treasuries

during this year’s debt crisis as well as the significant

challenges Treasury investors face to earn reasonable

returns going forward.

Exhibit 4: The 2011 rally has pushed US 10-year real rates 140 bpbelow those in Japan10-year government bond yield minus 10-year inflation swap rate; %

-1.0

-0.5

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

2009 2010 2011

USJapan

Exhibit 3: Term premium in the front end declined sharply this year in part due to the Fed’s new guidance on policy rates2-year Treasury yields versus 1Yx3M OIS rate; %

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9


2-year Treasury yields1yx3m OIS rate





November 24, 2011

Terry BeltonAC

(1-212) 834-4650




109

The outlook for Treasury yields

Looking into 2012, our outlook is for Treasury volatilityto remain high but with rates ultimately remaining range-

bound as further deterioration in Europe competes with

rich valuations. Within this range, we look for rates to

move lower early in the year as the crisis in Europe

deteriorates further and investors add to the risk-off trade.

At the same time, we see the flight-to-quality bid

eventually fading, as rich valuations, crowded investor

positions, and still-heavy supply allow the Treasury

market to decouple from the ongoing stress in Europe. In

the discussion below, we assess the major factors likely to

drive Treasury yields in 2012, present our interest rate

forecast, and highlight the many risks around that forecast.

The case for lower Treasury yields in 2012 really rests on

only two factors, but they are important ones that, at least

early next year, are likely to cause Treasury yields to trade

at even richer levels than those reached in 2011. The first

is Europe. Our baseline view is that the sovereign debt

crisis in Europe is now set to take another turn for the

worse and is likely to fuel a renewed flight-to-quality bid

into Treasuries early next year. While the details are more

fully discussed in our Global Fixed Income Markets

Outlook publication (see Global Fixed Income Markets

2012 Outlook , November 24, 2011), we would highlight

three of the most worrisome aspects of that outlook withimplications for the Treasury market.

• We look for intra-EMU spreads to widen

significantly early next year. The widening will

likely be driven by increased selling from

unlevered investors who thought they owned

risk-free government bonds but are now

overweight high yielding risky assets. With

yields elevated, Italy’s access to private market

funding may become less secure, creating some

risk that they are forced to seek external funding

support. Italy is the world’s third largest bond

market, however, and with €250bn of funding

needs in 2012 alone (including €200bn of

redemptions), they will stretch the limits of the

EU’s and IMF’s ability to meet their funding

needs.

• Policymakers will likely continue to move too

slowly to get ahead of the deepening crisis. The

ECB will likely remain a reluctant lender of last

resort and, with heavy investor selling likely to

continue, is unlikely to buy enough to stabilize

spreads near current levels. While we do not

expect a EU breakup, statements from officials in

Germany, the Netherlands, and other creditor

countries about EU exit strategies are likely to

increase and provide ongoing support for

investors to short semi-peripheral bond markets.

• The crisis is likely to broaden further beyondItaly, especially if the Euro area recession proves

severe. In that case, risks would rise of a France

sovereign downgrade; this would lower the EFSF

funding capacity and further increase the

resources required from the creditor countries.

We note that these views represent our base case outlook;the headline risk around Europe is extraordinarily highand a permanent solution to the crisis could certainly befound just when the stress reaches its deepest point.

How low could intermediate Treasury yields go if Europe

deteriorates further? Correlations between Treasury yieldsand European sovereign spreads remain high and shouldstay that way into the early part of next year. Using our fair value model (see box), we estimate that for everyadditional widening of 100bp in 10-year Italy spreads toGermany, 10-year Treasury yields should fall by 16bp, allelse equal. If 10-year Italy yields reach 9% next year, wewould expect 10-year Treasury yields to move below therecord 1.70% reached last year. To be sure, US rates willeventually decouple from Europe as rich valuations bring

Exhibit 5: Correlations between 10-year Treasuries and European

sovereign spreads remain elevated and were especially high when10-year yields hit record lows1-month rolling average of 10-year Treasury yields (%) versus the 1-month rollingcorrelation between weekly changes in 10-year Treasury yields and semi-peripheral* Europe sovereign CDS spreads% correlation; inverted

1.9

2.0

2.12.2

2.3

2.4

2.5

2.6

2.7

2.8

2.9

3.0 -1.0

-0.8

-0.6

-0.4

-0.2

0.0

0.2

Sep 11 Oct 11 Nov 11

Treasury yieldsCorrelation; inverted

* Average 5-year CDS spreads of France, Spain and Italy.





November 24, 2011

Terry BeltonAC

(1-212) 834-4650




110

sellers. But for now, correlations (and betas) remain quiteelevated and were especially high when rates made their

record lows back in September (Exhibit 5).

After Europe, the second most important factor supportinglower Treasury yields in 2012 is likely to be Fedsponsorship. With growth likely to remain sluggish in1H12 as fiscal drags come into play and unemploymentremains high, we look for the Fed to announce another round of asset purchases in MBS. While the eventualimpact on Treasury yields may be modest, the initialsignal by the Fed that more QE is coming is likely to leadto an outsized rally as investors rebalance their portfolios.In each of the previous four instances where the Fedannounced large scale asset purchases, Fed signaling inadvance of the formal announcement caused 10-year

Treasury yields to fall sharply (by an average of 50bp) aslevered investors significantly increased duration longs(Exhibit 6).

Beyond Europe and the Fed, other factors supportive of lower yields in 2012 are small by comparison. Theyinclude a worsening of the front-end collateral shortagehelping push GC rates lower (see GC repo discussion below), a modest weakening in growth and inflationexpectations early next year, and further changes in the

Flight to quality and fair value of 10-year TreasuriesIn February 2011, we introduced a new short-term fair value

model for 10-year Treasury yields that combines both economicand technical factors likely to drive yields in the near term.Since then, we have modified the model somewhat byintroducing better measures of investor positioning, and addedEuropean sovereign debt spreads as a proxy for flight-to-qualityflows from Europe. The model now explains 10-year Treasuryyields as a function of five variables including 5-year forwardinflation expectations derived from the inflation swap market,the 1-year ahead J.P. Morgan forecast for real GDP growth,near-term expectations on policy rates measured by the 3-monthforward on 3-month OIS, the average 5-year CDS spreads of France, Italy, and Spain, and a measure of levered investor

positions (see box below on measuring investor positions). Themodel is estimated using weekly averages for data since June2008. All five variables are highly significant and explain about

80% of the variability in 10-year rates during this period. Thecoefficient on the flight-to-quality variable indicates that every100bp increase in semi-peripheral spreads lowers 10-year yields

by 33bp on average. Alternatively, because average semi- peripheral CDS spreads have been moving only half as much asspreads on Italian government bonds recently, we estimate 10-year yields should fall by roughly 16bp for every 100bp increasein spreads on Italian government bonds. This estimate holdsother variables constant; we expect the actual change in 10-year rates to be larger early next year as increased sovereign stresswill also be accompanied by an increase in levered investor duration longs.

10-year Treasury yield model parameters:*

Variable Current level Beta T-stat

Intercept - 1.983 14.2

Inflation expectations; % 2.72 0.328 7.3

GDP growth forecast; % 1.75 0.184 24.9

3Mx3M OIS; % 0.10 0.476 24.9

Sovereign CDS; % 4.01 -0.326 -26.5

Levered investor positions -0.64 -0.296 -19.1

*Estimated from 6/1/08-11/10/11.N=864, SER=0.229, R2=0.824

10-year Treasury yields versus model fair value; %

1.5

2.0

2.5

3.0

3.5

4.0

2009 2010 2011

ActualModel

Exhibit 6: Buy the rumor, sell the fact; 10-year Treasuries haverallied an average of 50bp leading into the last four Fed QE

announcements10-year Treasury yields (%) and J.P. Morgan index of levered investor bondpositions*% positions index; inverted axis

2.3

2.4

2.5

2.6

2.7

2.8

2.9

3.0 -0.6

-0.5

-0.4

-0.3

-0.2

-0.1

0.0

0.1

0.2

0.3

0.4

-30 -20 -10 0 10 20 30Business days around quantitative easing

10-year Treasury yields

Positions index; inverted

* Levered investor positions index is the weighted average of aggregate net longsheld by non-commercial investors as provided by CFTC and the partial beta of hedge fund returns versus 10-year Treasury yields, converted to a z-score. QEannouncement dates are 11/25/08, 3/18/09, 9/21/10, and 9/21/11.





November 24, 2011

Terry BeltonAC

(1-212) 834-4650




111

Fed’s policy rate guidance that should help lower term premium in the yield curve next year.

Against these bullish factors for the Treasury market,there are also some important offsets that we believe willultimately push Treasury yields higher in 2012. The firstis valuations. Ignoring the impact of the (presumably

temporary) flight-to-quality bid into Treasuries, our fair value model for 10-year yields suggests Treasuries are80bp expensive (Exhibit 7). While this richness can persist when Europe remains under stress, the magnitudeof the mispricing highlights the significant cheapeningthat is likely to occur in the Treasury market as soon as thesituation in Europe stabilizes.

Second, we view the supply/demand outlook in theTreasury market as unfavorable next year, with durationsupply rising and Fed buying likely to be significantlyweaker than in 2011. As discussed in more detail below,we expect overall duration supply (net of Fed purchases)across US fixed income markets to increase 11% nextyear, assuming a moderately sized QE3 program in MBS.Treasury supply net of Fed purchases is projected to morethan double to $1tn, requiring a significant increase in buying from foreign investors. This increased supply isoccurring at the same time that the largest foreign holdersof Treasuries have slowed their accumulation of USDreserves with Fed custody holdings of Treasuriesessentially flat since June (Exhibit 8).

Third, with Treasury rates very rich and most investorsalready extremely pessimistic about Europe, we would

expect Treasury yields to begin decoupling from Europeanheadlines at some point next year. In part, this reflectsinvestor positioning: correlations tend to rise early on inthe crisis as the news forces investors to rebalance but fallonce those rebalancing trades have occurred. Indeed, prior to the broadening of the crisis in August to include Italyand France, US rates had already begun decoupling from peripheral spreads; correlations between 10-year Treasuryyields and the average spreads of Greece, Portugal andSpain had fallen to zero this summer, even though

Exhibit 9: Correlations between US rates and spreads inGreece/Portugal/Spain had fallen to zero this summer despitecontinued spread widening

Average Greece, Portugal and Spain CDS spreads versus the 3-month rollingcorrelation between 1-month changes in 10-year rates and average Greece,Portugal and Spain CDS spreadsbp correlation; inverted axis

400

500

600

700

800

900

1000

1100

1200 -0.6

-0.4

-0.2

0.0

0.2

0.4

0.6

Feb 11 Apr 11 Jun 11 Aug 11

Average CDS spreadsCorrelation; inverted

Exhibit 8: The largest foreign holders of Treasuries slowed their accumulation of USD reserves this year with Fed custody holdingsof Treasuries flat since JuneTreasuries held at custody at the Fed excluding repo versus foreign exchangereserves of the largest Treasury holders*;$bn $bn

5100

5200

5300

5400

5500

5600

5700

2560

2580

2600

2620

2640

2660

2680

Feb 11 Apr 11 Jun 11 Aug 11 Oct 11

Treasuries held incustody at the Fed

FX reserves

* Includes FX reserves of South Korea, Brazil, Japan, China, Russia and Taiwan.Dollar is held constant at March 2009 levels.

Exhibit 7: 10-year Treasury rates are rich by 80bp10-year fair value* excluding the impact of the flight to quality bid into Treasuriesversus actual 10-year Treasury yields; %

1.5

2.0

2.5

3.0

3.5

4.0

4.5

5.0

2009 2010 2011

Model Actual

* 10-year Treasury model yield = 0.34 + 0.66 * 5yx5y inflation swap rates (%) +0.85 * 3mx3m OIS rate (%) + 0.2 * 1-year ahead J.P. Morgan real GDP growthforecast (%) – 0.33 * J.P. Morgan investor positions index in bonds; Estimatedover 6/1/08-11/11/11 period.





November 24, 2011

Terry BeltonAC

(1-212) 834-4650




112

peripheral spreads continued to move significantly wider (Exhibit 9). To be sure, rebalancing flows out of Italy are

far from done, suggesting the Treasury rally is not over.But at some point next year, we expect the short risk tradeto become crowded again limiting the upside in Treasurieseven if Italy continues to widen.

The offsetting factors of further deterioration in Europeand already rich valuations in the Treasury market willmake it difficult for Treasuries to trend for an extended period in either direction next year. Along with our outlook for near-trend growth and for policy rates to stayanchored near zero at least until 2014, we expect range- bound Treasury rates in 2012. Within that range, we look for rates to move lower early next year as Europe

deteriorates and investors reset short risk positions. Our 1Q12 target for 10-year rates is 1.70% (Exhibit 10),which we view as consistent with another 100-200bpincrease in Italian 10-year yields combined with anincrease in duration buying from levered investors earlynext year as they reset the risk-off trade. Our year-endtarget for 10-year yields is 2.5%. This assumes only a partial retracement back to fair value as we expect Europeto remain unsettled, resulting in some lingering flight toquality-related demand.

At the front end of the curve, our forecast tracks theoutlook for 10-year rates (adjusted for the lower volatility

of short rates) but incorporates some additional richeningdriven by an anticipated decline in GC repo and OIS (see“The front end and the outlook for GC” below). Weexpect the Fed to further strengthen its forward guidanceon policy rates early in 2012 helping further flatten theOIS curve. In addition, front-end Treasuries are cheap toOIS and we expect 3-5bp of richening in 2-year yieldseven if OIS remains unchanged.

We are biased to a modestly steeper 10s/30s yield curve in2012 targeting 110bp by mid-year. Following theannouncement of Operation Twist in September, the longend of the curve had significantly flattened relative to fair value before retracing much of the mispricing over the past month. Our 10s/30s yield curve model, which is based on the level of front-end rates, inflationexpectations, variable annuity (VA) hedging needs, andlong-end supply (see U.S Fixed Income Markets Weekly,September 23, 2011) currently shows the curve as 7 bp tooflat (Exhibit 11). While we could see some modestflattening in 1Q12 as stresses in Europe increase long-endVA-related duration buying, increased prospects that theFed will expand its balance sheet as we move through theyear provides an offset. On balance, with Treasury rateslargely range bound into mid-year, we look for a modestly

steeper 10s/30s curve and target 30-year yields to reach3.60 by mid-year.

Finally, we would note that this is a year where volatilityis likely to be extraordinarily high even if rates ultimately prove range-bound. With poor liquidity and balance sheetsof many financial institutions becoming more constrainednext year, the market is likely to continue to experiencefrequent whipsaws in Treasury rates driven as much bytechnicals and investor rebalancing flows as byfundamentals. In the next section, we discuss why

technicals are likely to be so important next year and makethe case that contrarian duration trading strategies arelikely to be especially profitable in 2012. We also discussimproved metrics we have developed this year for trackinginvestor technicals that have proven to be a good leadingindicator of turns in the Treasury market.

Exhibit 11: After a significant overshoot following Operation Twist,the 10s/30s curve has moved back to fair value

Actual versus model* for the 10s/30s Treasury yield curve; %

0.7

0.8

0.9

1.0

1.1

1.2

1.3

1.4

1.5

1.6

May 10 Nov 10 May 11 Nov 11

ActualModel

* 10s/30s curve modeled as 1.147 - 0.003292 * index of variable annuity hedging($bn 20-year equivalents) + 0.1169 * 5yx5y inflation swap rates (%) – 0.2446 * 3-

year yields (%) + 0.001028 * Amount of Treasuries outstanding in the 10- to 30-year sector net of Fed purchases via Operation Twist ($bn)

Exhibit 10: J.P. Morgan interest rate forecast

% Current 4Q11 1Q12 2Q12 4Q12

21 Nov 11 31 Dec 11 31 Mar 12 30 Jun 12 31 Dec 12

2-yr Treasury 0.28 0.22 0.17 0.30 0.30

5-yr Treasury 0.91 0.90 0.75 1.25 1.25

10-yr Treasury 1.96 2.00 1.70 2.50 2.50

30-yr Treasury 2.94 3.00 2.70 3.60 3.60





November 24, 2011

Terry BeltonAC

(1-212) 834-4650




113

Avoiding crowds in 2012

While our baseline forecast is for yields to move lower early in 2012 as contagion spreads in Europe, markets are

very unlikely to move in a straight line; we expect

volatility to remain extremely high with frequent turns in

US interest rates during the course of the year driven by

poor liquidity and technicals around investor positioning.

While technicals have always been an important factor

driving yields, a combination of cyclical and structural

factors have amplified their impact since the end of the

2008 financial crisis, causing intermediate Treasury rates

to be extraordinarily volatile despite stable policy rates

(Exhibit 12). These factors include both declining market

depth and a growing disparity between the size of dealer

and end-user balance sheets.

The importance of technicals on Treasury rates in the last

couple of years is highlighted in Exhibits 13 and 14

which compare 10-year Treasury yields to a measure of

investor duration positioning we have developed (see grey

box). The exhibits highlight four major turns (whipsaws)

in the Treasury market since 2Q10 that were triggered or

at least amplified by investor unwinds of crowded

consensus trades. This includes: 1) the Treasury rally in

2Q10 following a period where improving growth and Fed

exit strategy discussions resulted in significant duration

shorts by levered investors; 2) the sell-off in Treasuries

following the start of QE2 in October 2010; 3) the 3Q11rally in Treasuries following the end of QE2; and, 4) the

4Q11 sell-off in Treasuries following the EU driven flight-

to-quality bid in Treasuries. While the initial catalyst for

the unwinds differed, the whipsaw in Treasury rates

occurred in each case just after our investor position

measure reached a fairly extreme level with unwinds

amplifying the market move. Over the four episodes, the

change in 10-year Treasury yields (with Treasury yields

moving in the opposite direction of the consensus view in

all four cases) averaged 70bp. The largest move (97bp

sell-off) followed the start of QE2 when our position

measure indicated levered investors held their largest long

duration position since 2007.

Exhibit 13: The importance of avoiding crowdsConsensus View

Duration Rationale

Short Growth/Fed exit 5/18/10 10/21/10 5.0 -85

Long QE2 coming 10/21/10 4/12/11 5.6 97

Short QE2 ending 4/12/11 6/23/11 2.3 -59

Long EU crisis 9/23/11 10/18/11 0.8 34

Average move (abs value) 3.4 69

10y yield

change; bp

Time

(months)

Positions

reversed by

Peak

positons by

Exhibit 12: Going nowhere fast: a structural decline in liquidity has

caused Treasury volatility to spike despite stable policy ratesRolling 1-year standard deviation of daily changes in 10-year Treasury yieldsdivided by the rolling 1-year standard deviation of daily changes in the 3Mx3M OISrate

0

2

4

6

8

10

12

1995 1999 2003 2007 2011

Exhibit 14: Crowded trades have led to four major rate whipsawssince 2Q1010-year Treasury yields (%) versus the J.P. Morgan levered investor position indexin bonds*

1.5

2.0

2.5

3.0

3.5

4.0 -1.5

-1.0

-0.5

0.0

0.5

1.0

1.5


10-year Treasury yields; invertedPositions index

* Positions index is the weighted average of aggregate net longs held by non-commercial investors as provided by CFTC and the partial beta of hedge fundreturns versus 10-year Treasury yields, converted to a z-score.





November 24, 2011

Terry BeltonAC

(1-212) 834-4650




114

Measuring investor positionsOne of the more active areas of research for us in 2011 has been developing alternative measures of investor positioning that better explain

Treasury yield movements. While survey data (e.g., the J.P. Morgan Treasury Client Survey) are often used as a barometer of investor positioning, we have found it to be less useful for explaining and predicting Treasury yields than actual position measures available fromfutures markets. In addition, position measures calculated by reverse engineering return data (i.e., return betas) appear more reliable to us thansurvey measures.The table below provides a comparison of alternative measures of investor positioning based on how well they explain yield movementswhen added to our short-term fair value model for 10-year Treasury yields. The investor position measures we compare include (durationweighted) CFTC data on net speculative positions in interest rate futures, survey data from our weekly duration survey, and return betascalculated using daily returns published by the largest core bond funds as well as hedge funds. The fair value model (see grey box above),which is estimated using weekly data since June 2008, explains 10-year Treasury yields as a function of inflation expectations, the J.P.Morgan 1-year ahead growth forecast, the level of 3-month forward OIS, a flight-to-quality measure proxied by European semi-peripheralspreads, as well as a measure of investor positions. While most position indicators are statistically significant and improve the overall fit of the model, we find levered investor position measures explain Treasury yield movements much better than unlevered investor

positions. The best-fitting models are the ones that use CFTC data on speculative positions followed by a position measure based on macrohedge fund return betas. These models have the lowest regression standard error, highest t-statistic on the position variable, and highest R-squared (see Table). By contrast, the J.P. Morgan duration survey measure and the return betas for core bond fund managers have the lowest

explanatory power with the bond fund return beta statistically insignificant.We also find that the 10-year model can be improved by combining some of these position measures. In particular, we have created a leveredinvestor index that converts the CFTC and hedge fund return betas to a z-score, and takes a weighted average based on regression weights.This index has the best fit in our 10-year yield model when compared to other position measures and has been stable through time; the R-squared increases to 0.82, and the standard error of the regression falls to 0.23. This index is currently at -0.5 suggesting levered investors arenow modestly short duration.Beyond helping understand the factors driving yield changes, we also find position indicators to be useful contrarian indicators that help usforecast future yield changes. Reflecting the fact that levered investor positions tend to be mean-reverting, we find our position index to be auseful leading indicator of yield changes. This is highlighted in the last exhibit below which compares monthly changes in 10-year yields tothe ex ante level of our J.P. Morgan positions index.10-year Treasury yield model summary statistics*Model includes: Current value Coefficient T-stat RSQ SER

CFTC spec longs -1.1 -0.175 -11.8 78.3 0.25

Hedge fund beta -0.4 -0.090 -11.3 78.0 0.26

JPM duration survey 0.6 -0.059 -6.2 75.9 0.27

Mutual fund beta 0.3 -0.006 -0.5 74.8 0.27

CFTC/HF Index -0.8 -0.295 -21.5 82.3 0.23 *All models estimated from 6/1/08-11/8/11. In addition to a measure of investor positions, the model includes 5Yx5Y forward inflation expectations, the JPM 1-year aheadgrowth forecast, the level of 3-month forward OIS, and an average of France/Italy/Spain 5-year CDS.

JPM index of levered investor positions:* 1M change in 10Y UST yields vs. one-month ago JPM index of leveredpositions; 5/10-11/11

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5


-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

-1.0

-0.8

-0.6

-0.4

-0.2

0.0

0.2

0.4

0.6

0.8

Nov 10 May 11 Nov 11

1m ago Positions Index1m change in yields; %

* Defined as 0.66 * CFTC net spec longs z score +.33 * hedge fund return beta z-score. Hedge fund betas are calculated using daily returns on the IQ Global Macro Beta Index.





November 24, 2011

Terry BeltonAC

(1-212) 834-4650




115

While markets often zigzag, the magnitude of the recent

whipsaws appear unusually large, given that the economic

outlook has been comparatively stable (slow but positive

growth) and the Fed is expected to keep rates near zero for

a prolonged period of time. In our view, the volatilityreflects a secular decline in market liquidity as a declining

pool of dealers face increasingly larger capital

requirements and stricter leverage ratios. Some evidence

of this can be seen in the decline of dealer reported VAR

levels since 2010 and the corresponding decline in market

depth (Exhibit 15). We would note that this reduced

ability of dealers to provide liquidity is occurring at the

same time that the balance sheets of the largest asset

managers are growing, creating an imbalance in the

supply and demand for liquidity (Exhibit 16).

Because poor market depth is unlikely to improve next

year as US and especially European banks de-lever further, technicals are likely to again be a dominant driver

of Treasury yields with investor rebalancing flows having

an outsized impact on yields. Thus, despite our 1Q12

outlook for lower yields driven by renewed stress in

Europe, our bias in 2012 will be to trade duration

tactically with an emphasis on fading crowded trades

and positioning for mean reversion. This strategy

proved especially profitable in 2011; a simple rule of

going long duration each day our levered investor

positions index was below -0.75 and short duration each

day our index was above 0.75 had a success rate of over

95% (based on 71 trades) and average P/L for 1-month

holding periods of 25bp (Exhibit 17). With the index

currently at -0.5, we are currently biased to holding long

duration trades expecting investors to build long duration

positions as the risk-off trade gains momentum early in

2012.

The front end and the outlook for GCThe past year has been characterized by increased demand

for front-end Treasuries from money market funds. As the

European crisis escalated and risk aversion increased, the

assets under management (AUM) in prime money market

funds fell dramatically and investors instead favored the

relative safety of government money market funds. Since

the end of May, the AUM in prime money market funds

have fallen $200bn, while the AUM in government money

market funds have increased by $130bn. As a result,

Treasury holdings of money market funds (including repo)

Exhibit 15: Market depth has declined as dealers continue to lower VAR and shrink their balance sheets…Reported average quarterly VAR for the 9 largest investment banks* versusmarket depth**; $mn$mn $mn

90

95

100

105

110

115

120

125

80

100

120

140

160

180

200

220

240


VARMarket depth

* Average VAR reported by JPM, GS, MS, BAC, C, UBS, CS, Soc Gen and DB.3Q11 VAR is an estimate, projected from the results reported to date (whichexcludes C).** Market depth is calculated as the 3-month moving average of average size of the top three bids and offers, in $mn, for the on-the-run 10-year Treasury note,averaged between 8:30 a.m. and 10:30 a.m. daily.

Exhibit 16: …while end-user demand for liquidity has increased AUM for the largest asset managers*; $tn

14.5

15.0

15.5

16.0

16.5

17.0

17.5

18.0

1Q10 2Q10 3Q10 4Q10 1Q11 2Q11 3Q11 *Includes BlackRock, Allianz Group, State Street Global, Deutsche Bank,Vanguard Group, Fidelity Investments, J.P. Morgan Chase, BNP Paribas, AXAGroup and BNY Mellon.Source: Bloomberg, Company websites, P&I/Towers Watson Global 2010 ranking

Exhibit 17: A simple contrarian trading strategy based on our indexof investor positioning was highly profitable in 2011Number of trades, success rate and average P/L from back-testing a trading rule*based on the J.P. Morgan positions index

TradeNumber of

tradesSuccess rate (%) Average P/L (bp)

Long 10s 39 95 22.9

Short 10s 32 100 28.3

Total 71 97 25.3 * The rule involved going long duration each day our levered investor positionsindex was below -0.75 and short duration each day our index was above 0.75.Positions were held for a one month period.





November 24, 2011

Terry BeltonAC

(1-212) 834-4650




116

have increased dramatically (Exhibit 18). This increased

demand has come in a period of falling T-bill supply. The

combination of these two factors proved to be very

supportive of valuations: T-bills across the curve richened

to historical levels, with yields at or near-zero levels.

Supportive technicals notwithstanding, GC repo rates have

stayed stubbornly high, with 1-month term GC currently

at around 11bp.

Looking ahead, we expect GC repo rates to head lower

targeting 1-week GC to average 5bp in 1H12 for the

following reasons. First, supply technicals in the front-end

are very supportive. Our short-term fixed income

strategists expect issuance in most front-end markets to be

negative in 2012, with Non-Financial CP being a notable

exception (Exhibit 19). Moreover, net T-bill issuance for

FY2012 will be close to zero or negative, depending on

the magnitude of fiscal stimulus that is passed. We expect

the budget deficit to total $1.1tn in FY2012, assuming that

a $100bn stimulus is passed. With coupon net issuance

expected to be $1.09tn, the net issuance of T-bills will

total a mere $10bn over that period. However, if nostimulus is approved, then T-bill issuance will total

-$90bn. Second, demand for T-bills in particular is likely

to stay elevated despite shrinking supply. Investors will

likely stay risk-averse since headline risk from Europe

will stay elevated in 1H12, and demand for government

paper will persist as assets continue to migrate out of

prime and into government money market funds.

Moreover, unattractive net yields in prime funds (a mere

3bp) relative to government funds (1bp) will also be

supportive of this trend. Finally, possible regulatory

changes in money market fund capital requirements and

the constant NAV model in late 2012 could further

accelerate this trend (see Short-term Fixed Income).

The combination of these factors is likely to put

downward pressure on GC repo rates as we head through

the year. Historically, the outstanding balance of T-bills

net of money market holdings of Treasuries has been a

significant driver of repo rates, with every $100bn of

additional supply biasing 1-week GC repo rates 4bp

higher (Exhibit 20). Given our expectation of close-to-

zero net bill issuance in 1H12, combined with our

expectation that money market funds will purchase $20bn

of Treasuries per quarter, we expect that 1-week GC repo

Exhibit 19: We expect net issuance in the front end to be negative

in 2012Projected 2012 net issuance in short-term fixed income markets by sector; $bn

-200

-150

-100

-50

0

50

Non-fin CP T-bills ABCP Agencydiscos

FinancialCP

YankeeCDs

Exhibit 20: We look for 1-week GC repo rates to average near 5bpin 1H given our outlook for supply and money market funddemand…1-week Treasury GC repo rate (bp) modelled as a function of outstanding T-bills

net of money market fund holdings of Treasuries ($bn)

0

5

10

15

20

25

950 1000 1050 1100 1150 1200 1250 1300 1350 1400

Effective T-bills outstanding; $bn

Y = 0.0421 X1 - 36.6723R² = 65.92%standard error = 3.9300period = Nov 19,09 - Nov 19,11

Exhibit 18: The perfect storm: high money market demand in a

period of falling supply resulted in lower bill yieldsTreasury bills outstanding ($bn) versus Treasuries held by money market funds($bn)

1450

1500

1550

1600

1650

1700

1750

1800

430

440

450

460

470

480

490

500

510

520


T-bills outstandingTreasuries held by money market funds

Source: Treasury, J.P. Morgan, iMoneyNet





November 24, 2011

Terry BeltonAC

(1-212) 834-4650




117

rates will eventually average around 5bp in 1H12, or 6bp

lower than current levels.

This environment will be broadly supportive of lower

2-year yields as well, and will push them closer to 17bp

in 1Q. In addition, to lower financing rates, our positive

outlook on 2-year Treasuries is driven by the following

factors.

First, the expansion of the Fed’s balance sheet under QE3

will increase the supply of reserves and help lower the

effective funds rate. Historically, every $100bn of

additional reserves in the system has lowered 3-month

OIS rates by 1.5bp, as shown in Exhibit 21. Thus, if the

Fed announces quantitative easing of $500bn, it couldsignificantly lower front-end OIS rates, bringing them

close to zero.

Second, changes in the Fed’s communication policy will

be supportive as well, since it will further reduce term

premium from the front end of the curve. We expect the

Fed to modify its communications on policy rates during

1Q12 by linking future interest rate hikes to specific

economic conditions along the lines suggested by Chicago

Fed President Evans. This should lengthen the market’s

expectation of the Fed-on-hold horizon beyond mid-2013.

Assuming that this outcome causes forward OIS rates to

narrow back to levels reached when the Fed firstintroduced the mid-2013 language on August 9 suggests

1Yx1Y OIS rates could fall by almost 10bp (Exhibit 22).

We estimate that this would lower 2-year Treasury yields

by another 8bp.

Finally, we expect 2-year Treasuries to richen relative to

OIS as we approach mid-year when the Fed is scheduled

to complete Operation Twist. Fed selling of the front end

has helped Treasuries cheapen relative to OIS with the

spread between 2-year Treasury yields and 2-year OIS

currently at 13bp, or 6bp wider than its 1-year average.

In sum, given our outlook for GC, term premium, and anexpansion of the Fed’s balance sheet, we expect 2-year

Treasuries to richen on an outright basis and relative to

OIS. We look for 2-year rates to trade back to 17bp, or

around 10bp richer than current levels.

Finally, we also note that changes to the Fed’s

communication policy are also likely to remove term

premium even further out the curve, such as in the 3-year

sector. As shown in Exhibit 22, at close to 35bp, the

1Yx1Y/2Yx1Y OIS curve is quite steep and has the

potential to flatten significantly further if the market starts

to price in a longer period of Fed-on-hold.

Exhibit 21: The expansion of the Fed’s balance sheet under QE3will increase the supply of reserves and help lower the effective

funds rate1-month rolling average of 3-month OIS rate (bp) regressed against US bankreserves ($bn)

6

8

10

12

14

16

18

20

22

24

1000 1100 1200 1300 1400 1500 1600 1700

Reserves; $bn

Y = -0.0145 X1 + 33.8150R² = 78.67%standard error = 1.9268period = Nov 21,09 - Nov 21,11

Exhibit 22: OIS forwards in the front end of the curve havesubstantial room to fall if the Fed adopts the Evans planMinimum 1-year spot, 1Yx1Y and 2Yx1Y OIS rates since the FOMC meeting on

August 9 versus current levels; %

0.12

0.19

0.55

0.06 0.08

0.46

0.0

0.1

0.2

0.3

0.4

0.5

0.6

1y spot 1yx1y 2yx1y

Current

Min since 8/9





November 24, 2011

Terry BeltonAC

(1-212) 834-4650




118

Supply Outlook

Supply technicals improved during 2011 as gross duration

supply across fixed-income markets fell 10% and duration

supply net of Fed purchases fell 24% (Exhibit 23). The

decline in gross issuance was fairly broad-based and came

primarily from spread products led by Municipals (down

29%), Agency debt (down 28%), MBS (down 14%), and

high-grade credit (down 9%).

Issuance within the Treasury market also fell in 2011

driven by a combination of lower deficits and the

suspension of SFP bills issuance. For calendar year (CY)

2011, we project net Treasury issuance to equal $1.09tn

versus $1.59tn in CY2010 (Exhibit 24). Notably, even

though nominal coupon sizes were left unchanged during

the year, net issuance of nominal coupon Treasuries fell

by $318bn as redemptions were higher (Exhibit 24). 2011

also became the third consecutive year of negative bill

issuance, causing the share of T-bills in the Treasurymarket to fall further to 15.2% currently from 19.6% at the

beginning of the year. As a result, the weighted average

maturity of outstanding debt increased to 63 months, its

longest since August 2002, and up from 55 months at the

end of 2009 (Exhibit 25).

Looking ahead, we expect overall duration supply in

fixed-income markets to increase 8% in 2012 approaching

the record levels set in 2009-2010 (Exhibit 23). This

increase is primarily driven by higher supply in the MBS

and Municipal markets with mortgage refinancing

Exhibit 24: Look for net issuance of coupon Treasuries to fall

further in 2012 Net issuance in Treasuries by sector and by calendar year; $bn

Sector CY 2010 CY 2011* CY 2012*

Bills -19 -274 -8

Nominal coupons 1563 1245 936

TIPS 48 120 100

Total 1592 1090 1028

Total ex-SFP bills 1397 1290 1028

Nominal coupons+TIPS 1611 1364 1036 * 2011 and 2012 are J.P. Morgan projections.

Exhibit 25: The share of T-bills in the Treasury market fell againthis year, further lengthening the weighted average maturity of theTreasury market Weighted average maturity of outstanding Treasuries (months) versus the shareof T-bills in the Treasury market (%); dashed lines are projections*Months %

10%

15%

20%

25%

30%

35%

45

50

55

60

65

70

75

Jan 00 Aug 02 Feb 05 Aug 07 Feb 10 Sep 12

Weighted average

maturity

Share of T-bills

* Assumes unchanged coupon sizes and FY budget deficit of $1.1tn

Exhibit 26: We look for the budget deficit to fall 15% in FY2012Budget deficit by fiscal year; $bn

-237-127

158

377 413

319 248162

455

141612941298

1100

-500

0

500

1000

1500

2000 2003 2006 2009 2012*

* 2012 is J.P. Morgan forecast.

Exhibit 23: Gross duration supply fell by 10% in 2011, but is poised

to increase by 8% in 2012 Annual duration supply via MBS, Municipal, investment grade corporates,Treasuries and Agency debt markets, gross and net of Fed purchases; $bn 10-year equivalents

1500

2000

2500

3000

3500

2007 2008 2009 2010 2011 2012

Gross duration supply

Supply ex-Fed

* 2011 and 2012 are J.P. Morgan forecasts. We assume that the Fed purchasesa total of $500bn MBS 2012 via QE3.





November 24, 2011

Terry BeltonAC

(1-212) 834-4650




119

projected to increase and municipal finances staying weak.

Within the Treasury market, the outlook is more

favorable, with the deficit expected to be on a declining

trajectory. Our forecast assumes a FY2012 budget deficit

of $1.1tn, lower than the $1.3tn in FY 2011 (Exhibit 26).

Thus, net issuance will be lower, too, totaling $1.04tn

(Exhibit 24).

The composition of issuance will likely be similar to thatin recent years. We look for nominal coupon sizes to stay

unchanged in 2012, but look for TIPS issuance to increase

modestly (Exhibit 27). Given our deficit projection, this

implies that Treasury bill net issuance will be close-to-

zero in 2012, with the share of bills in the Treasury market

falling modestly to 14.6% (Exhibit 25). This projection

assumes Congress passes $100 bn of fiscal stimulus; if no

stimulus is passed, net issuance of bills is projected to fall

to -$90bn, bringing the share of bills to 13.7% of the

market by the end of FY2012. In this case, there is some

possibility that Treasury will cut coupon sizes later in the

year in order to stabilize the size of the bill market.

Demand Outlook

At the same time that duration supply is set to increase

across fixed income markets, demand technicals are likely

to be more challenging for the Treasury market in 2012.

The primary reason is a decline in Fed sponsorship.

While the Fed purchased 60% of net Treasury issuance in

2011, we expect limited net buying of Treasuries by the

Fed next year with any increase in its balance sheet likely

to be targeted in MBS. As a result, private investors and

foreign central banks will need to more than double (from

$454bn in 2011 to $1,000bn in 2012) the amount of

Treasuries they buy in order for the Treasury market to

clear in 2012.

After the Fed, the single largest buyer of Treasuries in

2011 was foreign investors. Net purchases in 2011 are on

track to reach $290bn with most of the buying occurring

in 3Q11 after the end of QE2. While this is significantly

lower than the magnitude of purchases in the last three

years of $640bn per year, it is higher as a percentage of

effective supply. Foreigners took down 64% of effective

Exhibit 27: Nominal coupon sizes should be stable in 2012 while

TIPS issuance increases modestly J.P. Morgan gross issuance forecast; $bn

2s 3s 5s 7s 10s 30s

5y

TIPS

10y

TIPS

30y

TIPS Subtota l

Jan 12 35 32 35 29 21 13 14

Feb 12 35 32 35 29 24 16 10

Mar 12 35 32 35 29 21 13 12 537

Apr 12 35 32 35 29 21 13 16

May 12 35 32 35 29 24 16 12

Jun 12 35 32 35 29 21 13 8 537

Jul 12 35 32 35 29 21 13 14

Aug 12 35 32 35 29 24 16 14

Sep 12 35 32 35 29 21 13 12 541

Oct 12 35 32 35 29 21 13 8

Nov 12 35 32 35 29 24 16 12Dec 12 35 32 35 29 21 13 14 535

Total 420 384 420 348 264 168 44 76 26 2150

Chg from 2011 0 0 0 0 0 0 6 6 3 15

Exhibit 28: While the magnitude of buying by foreign investors was

lower in 2011 than that in recent years, they took down a larger portion of effective supplyForeign purchases of Treasuries ($bn) versus foreign purchases as a percentageof Treasury net issuance excluding Fed purchases (%)% $bn

0

200

400

600

800

20%

40%

60%

80%

100%

120%

140%

2001 2003 2005 2007 2009 2011*

%ge of netissuancebought byforeigners

Amount of foreignTreasury purchases

Source: Federal Reserve Z.1, TICNote: 2011 is an estimate of the total foreign purchases reported by the FederalReserve in Flow of Funds data. 1H11 uses actual Flow of Funds data, Q3 modelsthe Flow of funds purchases as -3.7 + 0.975 * Monthly net purchases reported byTreasury in the TIC report, while Q4 is estimated by a combination of: (a) Flow of funds modeled as 3.8 + 1.34 * Monthly change in Treasury custody holdings and(b) the YTD monthly average purchases of Treasuries by foreigners.

Exhibit 29: A flight-to-quality bid: EU countries were net buyers of Treasuries in 3Q11 as the crisis in Europe escalated, more thanoffsetting selling by EM countries

Average monthly purchases of coupon Treasuries by foreigners in August andSeptember according to TIC data versus its prior 6-month average; $bn

Region Aug-Sep av erage 6m av erage

EU 44.4 10.9

UK 33.7 15.1

France 10.9 -1.6

Japan 21.8 4.2

China -9.8 4.4

EM ex China* -9.4 2.0

* Includes Brazil, Russia, African countries, Taiwan, South Korea and Mexico.Countries in grey are also included in the EU total.Source: TIC





November 24, 2011

Terry BeltonAC

(1-212) 834-4650




120

supply (i.e., net issuance excluding Fed purchases), higher

than the 3-year average of 51% (Exhibit 28). The

composition of demand was favorable and foreigners

heavily favored coupon securities to bills.

While overall foreign demand for Treasuries has been

strong since the end of QE2, Treasury data for the third

quarter shows a significant shift in the composition of foreign buying. Demand from EM reserve managers

weakened during the quarter as they slowed their building

of USD FX reserves, while demand from the EU increased

in response to the sovereign debt crisis (Exhibit 29).

While we view the recent slowing in EM demand as

mostly a cyclical phenomenon, it is also likely a sneak

preview of a longer-term trend underway for many EM

countries to gradually slow their reserve accumulation.

Officials in China have become increasingly vocal about

the need to slow FX reserve building and diversify away

from the US dollar (see US Fixed Income Markets Weekly,

September 19, 2011). At the same time, this

diversification has been slow-moving and is likely to

remain so (Exhibit 30). On balance, we view weaker

sponsorship by EM reserve managers as a risk to the

Treasury market rather than our base case. In the

meantime, softer demand from the EM is likely to be

offset by the flight-to-quality bid from the EU.

Exhibit 31 shows our 2012 projections for Treasury

demand by investor type. The estimates are based on a set

of simple empirical models discussed in more detail below

(see box and Exhibit 32). For foreign investors, we

project total purchases of $540bn or 53% of net Treasury

issuance; by comparison, foreign purchases amounted to

64% of net issuance (net of Fed purchases) in 2011 and

48% in 2009-2010 (Exhibit 31). With the curve flat and

rates low, we expect demand from pension funds and

insurance companies to be modestly below average at

around $150bn (Exhibit 32). We also expect commercial

Exhibit 32: Models for estimating Treasury buying by pensionfunds, insurance companies and commercial banksEmpirical models for quarterly purchases by pension funds/ insurance companiesand commercial banks; $bn

Pension funds/ insurance companies

Variable Coefficient T-stat Projection

Intercept 35.9 1.7 1

30-year Treasury yields; % -7.8 -2.1 3.25

10s/30s curve; % 18.5 2.7 1

Quarterly effective net issuance of Treasuries; $bn 0.04 2.6 260

Projected purchases in 2012: 150

Quarterly data regressed over 15-years; R2=54%; Standard error = 62

Commercial banks

Variable Coefficient T-stat Projection

Intercept -5.4 -1.23

30-year MBS CC Libor OAS; bp -0.6 -4.7 43

2s/5s Treasury curve; % 7.0 1.4 0.8

Quarterly effective net issuance of

Treasuries; $bn 0.06 5.4 260

Projected purchases in 2012: -35

Quarterly data regressed over 5-years; R2=74%; Standard error = 10. Projections

are for the full year 2012.

Source: J.P. Morgan, Federal Reserve H.Z.1

Exhibit 31: Despite our expectation of increased foreign demand,

the overall supply/demand imbalance will worsen next year giventhe lack of Fed buyingNet issuance of Treasuries per calendar year versus purchases by differentinvestor types; $bn

Treasury net issuance; $bn 1476 1592 1093 1028

Investor

Foreign Investors 555 654 290 540

Commercial banks 92 113 -75 -35

Pension funds/ Insurance companies 211 224 130 150

Broker dealers -68 -29 100 -20

Mutual Funds 68 42 45 60

Federal Reserve 300 244 639 0Other 318 344 -36 333

2009 2010 2011* 2012*

Source: J.P. Morgan, Federal Reserve* 2011 and 2012 are J.P. Morgan estimates.

Exhibit 30: The share of USD in FX reserves has stabilized at

around 65%Estimated* share of USD in foreign exchange reserves; $bn

64%

66%

68%

70%

72%

74%

Jun 00 Jun 02 Jun 04 Jun 06 Jun 08 Jun 10

Source: IMF, Bloomberg* Assumes that the allocation of Japan’s and China’s FX reserves mirror thedistribution of FX reserve currency allocation of advanced economies andEmerging economies, respectively.





November 24, 2011

Terry BeltonAC

(1-212) 834-4650




121

banks to be net sellers of coupon Treasuries next year,

likely preferring MBS given its cheapness and the

relatively flat front end of the Treasury yield curve.

Finally, we expect mutual fund demand and demand from

broker-dealers to be in line with their recent averages.

In total, demand across the six largest investor types sums

to $695bn. This suggests that a comparatively large

amount must be funded by other investors, highlighting a

significantly worse supply/demand imbalance than in

2011.

One potential offset is the proceeds of maturing FDIC

guaranteed bank debt: $45bn is due to mature in

December 2011, and another $165bn is due to mature in

2012 (Exhibit 33). While little data is available on the buyers of this product, our best guess is that they were

primarily Treasury/Agency debt buyers and that the

proceeds will likely make their way back into the front

end of the either the Treasury or the Agency debt market.

Exhibit 33: Redemptions of FDIC guaranteed bank debt will beheavy in 2012 providing some support for Treasuries

FDIC guaranteed bank debt redemptions; $bn

0

10

20

30

40

50

60

Nov 11 Jan 12 Mar 12 May 12 Jul 12 Sep 12 Nov 12

Source: Bloomberg

Forecasting foreign demand for TreasuriesIn order to forecast foreign demand for Treasuries, we model average monthly foreign Treasury purchases (available from the Federal ReserveFlow of Funds release) as a function of the US trade deficit with the rest of the world and net issuance of Treasuries (i.e., Treasury net issuance

minus Fed Treasury purchases via quantitative easing). The trade deficit of the US with the rest of the world is a significant driver of foreign purchases of US securities, since it leads to an accumulation of USD reserves abroad. In addition, the magnitude of foreign purchases of Treasuries also depends on available supply as foreign investors have shown a willingness to increase their buying in response to larger USdeficits.

Exhibit A1 shows the details of this model, and Exhibit A2 shows that the model has tracked actual purchases reasonably well since 2002. Themodel shows that each additional $10 bn increase in the monthly trade deficit results in an additional $2bn of foreign buying. Similarly, eachadditional $10 bn increase in monthly Treasury supply results in an additional $4.3 bn of foreign buying. Given our expectation that the tradedeficit and effective net issuance will average $48bn and $85bn per month in 2012, respectively, we look for foreign buying of Treasuries toaverage $45bn per month or $540bn for the year, representing 53% of Treasury net issuance next year.

Exhibit A1: A model for forecasting foreign purchases of Treasuries Exhibit A2: Actual versus predicted monthly average of Treasurypurchases by foreigners; $bn

Net purchases by foreigners modeled as:

Variable Coeff T-stats Proj. 2012

Intercept -0.3 -0.1

Trade surplus; $bn -0.20 -2.5 -48.0

Net issuance of Treasuries, net of Fed

purchases; $bn 0.43 19.1 85.0

Projected foreign purchases of Treasuries; $bn/ month 45

R Square: 91%

Std. Error: 5.6

Regression uses rolling 1-year monthly averages and is fitted on quarterly

data over 10-years.

Source: Federal Reserve Z.1, US Census Bureau, Treasury

0

10

20

30

40

50

60

70

2002 2005 2008 2011

ActualModel





November 24, 2011

Terry BeltonAC

(1-212) 834-4650




122

With the Agency debt market shrinking every year,

Treasuries are likely to be the beneficiary of this demand.

Ratings Risk and the US Fiscal Outlook

One of the great ironies in the Treasury market this year is

the fact that 10-year yields reached all-time lows in the

immediate aftermath of S&P’s downgrade of the US

sovereign rating. The loss of the US ‘AAA’ rating

followed three consecutive years of record deficits that

caused net federal debt to increase to 73% in 2011 from

40% of GDP in 2008. In August, with a divided Congress

unable to reach a consensus on how to stabilize US fiscal

metrics and debt levels projected to continue to grow,

S&P fired a warning shot to the US about the need to get

its fiscal house in order.

While a lot more work lies ahead for the US to stabilize its

finances, the good news is that the debate changed

dramatically in 2011, and a down payment on spending

cuts was made with the passage of the Budget Control Act

(BCA) of 2011. These cuts include $900bn agreed to in

August and another $1.2tn in automatic spending cuts

required by the bill, with the Deficit Reduction Committee

failing to reach agreement. With these cuts, federal net

debt is projected to grow from 73% of GDP currently to

84% by 2021; while still on an uptrend, the trajectory is

much flatter than estimates prior to the passage of BCA,

with July CBO projections for 2021 equal to 95% of GDP(Exhibit 34). In order to stabilize debt levels, we estimate

Congress still needs to reduce the structural deficit by an

additional 1%-pt of GDP per year, or $1.8tn over the 9-

year period. These additional cuts are not likely in an

election year but are certainly possible in 2013.

Exhibit 35 presents a short menu of alternatives that could

be used to reduce the structural deficit further; any

combination that totaled an additional 1% (beyond the

BCA) would stabilize the ratio of US debt to GDP.

To be sure, these projections are sensitive to a number of

assumptions including economic growth. CBO’s baseline

projections, upon which Exhibit 34 is based, assume real

GDP growth will average 2.8% from 2012-2021. If

growth averaged 1% per year lower, we estimate debt

levels would deteriorate by 5%-pts, reaching 89% of GDP

by 2021 (Exhibit 36). In order to stabilize debt-to-GDP

under this weaker growth outlook, Congress would need

to find an additional $2.3tn in savings beyond the $1.2tn

in automatic cuts triggered by the Budget Control Act.

With some, albeit modest, progress made on budget

reform this year, we expect S&P and Moody’s to maintain

their existing US sovereign ratings in 2012, although Fitch

Exhibit 36: 2021 debt levels will deteriorate 5%-pts if growth slipsby 1% per year Projected Net Federal debt to GDP ratio in different growth scenarios; %

70

75

80

85

90

2012 2013 2014 2015 2016 2017 2018 2019 2020 2021

Baseline 2.8% growth

1.8% real GDP growth

Exhibit 34: Half way home: Debt to GDP in the US is still on anupward trajectory; another $1.8tn in cuts are needed to stabilize debtlevelsProjected Net Federal debt to GDP ratio*; %

70

75

80

85

90

95

100

2012 2013 2014 2015 2016 2017 2018 2019 2020 2021

BCA baseline

BCA+ 1.8 tn

Pre BCA baseline

* CBO’s August alternate baseline and J.P. Morgan estimates

Exhibit 35: Alternatives to reduce the structural deficit: anycombination that totals 1% of GDP will stabilize the ratio of US debtto GDPVarious alternatives to reduce the structural deficit; as percentage of GDP and $bn

Cumulative

deficit reduction*

Sunset all Bush tax cuts 1.4 2459

Sunset tax cuts on wealthy (per Obama**) 0.4 700

Feldstein 2% itemized deduction limit 0.2 350

Cap itemized deductions at 28% 0.2 293

Asset and spectrum sales; user fees 0.1 200

Tax international income 0.1 133

% of GDPOptions

*Cumulative deficit reduction relative to BCA baseline for 2013-2021** Extend cuts for families with income under $250,000





November 24, 2011

Terry BeltonAC

(1-212) 834-4650




123

is likely to move to a negative outlook. In their recent

statements, S&P and Moody’s both noted that the US

rating would not be lowered in response to the failure of

the Joint Select Committee, as long as the automatic

spending cuts of $1.2tn are implemented (Exhibit 37).Fitch has been more negative in its comments around the

failure of the Super Committee and we expect them to

place the US on negative outlook while maintaining the

‘AAA’ rating.

Finally, we note that derivative markets also appear

sanguine about the US fiscal outlook and the near-term

risk of further downgrades. CDS on US Treasuries have

been remarkably stable since August, outperforming most

other AAA-rated sovereigns. With the exception of

Norway, US CDS is now tighter than every other AAA-

rated sovereign; spreads are currently 45bp tighter than

Germany and 180bp tighter than France (Exhibit 38).

Trading themes for 2012

While Treasury yields are expected to remain low in 2012,

volatility will still be high, creating regular opportunities

for active investors to enhance the return on their

investment portfolios. In the discussion below, we discuss

four broad areas of opportunities for Treasury investors in

2012 including yield curve carry trading, supply cyclicals,

relative value opportunities around Fed purchase

operations, and cross-currency arbitrage that involves

creating synthetic Treasuries with cheap foreign bonds.

Yield curve trading

While yield curve carry trades typically perform well

when the Fed is on hold and rates are range-bound, the

current environment will likely prove more challengingwith front-end rates so low. In 2012, our approach to yield

curve trading will generally be driven by carry

considerations. But, in contrast to conventional wisdom,

our bias will be to favor negative carry curve trades in the

front end of the curve where convexity is attractive. In

intermediates, however, we expect positive carry trades to

outperform and will generally concentrate long carry

exposure in the intermediate to long end of the curve.

Some support in favor of avoiding positive carry yield

curve trades in the front end of the curve is highlighted in

Exhibit 38: US sovereign CDS spreads are below those of countries

with ‘AAA’ ratings5-year sovereign CDS spreads; bp

Country Rating Outlook Rating Outlook Rating Outlook

Norway AAA Stable Aaa Stable AAA Stable 45

United States AA+ Neg Aaa Neg AAA Stable 52

Sweden AAA Stable Aaa Stable AAA Stable 66

Finland AAA Stable Aaa Stable AAA Stable 70

Australia AAA Stable Aaa Stable AAA Stable 90

Britain AAA Stable Aaa Stable AAA Stable 92

Germany AAA Stable Aaa Stable AAA Stable 97

Netherlands AAA Stable Aaa Stable AAA Stable 117

Denmark AAA Stable Aaa Stable AAA Stable 123

Austria AAA Stable Aaa Stable AAA Stable 215

France AAA Stable Aaa Stable AAA Stable 230

Fitch

5y CDS

S&P Moody's

Source: Bloomberg

Exhibit 37: Comments by rating agencies on further downgrades

Rating

Agency

CurrentUS

sovereign

rating

Comment on Super Committee

S&P

AA+;

Negative

outlook

"The Fiscal Committee's inability to agree on fiscal measures

that would stabilize U.S. government debt as a share of GDP is

consistent with our Aug. 5 decision to lower our rating to 'AA+'.

However, we expect the caps on discretionary spending as laid

out in the Budget Control Act of 2011 to remain in force. If these

limits are eased, downward pressure on the ratings could

build."

Moody's

Aaa;

Negativeoutlook

"The Aaa government bond rating for the United States is

unaffected by the lack of a deficit reduction agreement by the

Joint Select Committee on Deficit Reduction...While a change in

the composition of the spending cuts would not be a major ratingconsideration, a reduction in the total amount that would increase

the projected increase in federal debt over the coming decade

could have negative rating implications"

Fitch

AAA;

Stable

outlook

"Failure by the Super Committee to reach agreement would likely

result in a negative rating action -- most likely a revision of the

rating Outlook to Negative...Less likely would be a one-notch

downgrade. Fitch now expects to conclude its review of the US

sovereign rating by the end of November."

Exhibit 39: Location, location, location: carry was a contrarianindicator of performance for front-end curve trades in 2011 3-month holding period P/L from going long the belly (5s) versus a level andcurve neutral combination of wings (2s/10s and 2s/7s) in 2011 vs. the 3-monthcarry and roll in the trade at time of initiation between 1/1/11 and 8/12/11; bp

-30

-20

-10

0

10

20

30

-6 -4 -2 0 2 4 6

2s5s10s

2s5s7s

3m carry and roll at initiation; bp





November 24, 2011

Terry BeltonAC

(1-212) 834-4650




124

Exhibit 39, which compares the performance of 2s/5s/7s,

and 2s/5s/10s weighted butterfly trades this year to ex ante

carry and roll on the trade. The weights on each trade were

designed to make the butterfly curve and level neutral and

returns are calculated over 3-month holding periods. The

Exhibit highlights that while carry is a good leading

indicator of P/L on these trades, it is a contrarian

indicator; the more negative the carry, the larger the

subsequent P/L. Initiating these weighted curve tradeswhen carry was positive in 2011 has resulted in a loss

80% of the time, with an average 3-month P/L of -8bp.

The poor performance of front-end butterflies during 2011

arises from the negative curve convexity on these trades.

As shown in Exhibit 40, positive carry 2s/5s/10s weighted

butterflies that are long the belly tend to become

steepeners as the curve flattens and become flatteners as

the curve steepens. This behavior reflects the changing

volatility of 2-year yields with 2s becoming sticky, as

rates fall and the curve flattens, and more volatile as rates

rise and the curve steepens. To offset these changing

volatilities, rebalancing the position would involve addingflattening risk (i.e., selling more 2s versus 5s) when the

curve flattens to offset the falling volatility of 2s.

Similarly, rebalancing involves adding steepening risk

(i.e., scaling back the short in 2s versus 5s) when the curve

steepens to offset the increasing volatility of 2s.

In contrast to carry trades anchored in the front end, we

favor positive carry yield curve trades anchored in the

intermediate sector where convexity is less of an issue.

Exhibit 41 shows that the 3-month holding period returns

from being long 7s versus a risk-weighted 5s/10s barbell

have been well correlated with the carry and roll on the

trade.

In addition to carry, we note that butterflies anchored in

the long and intermediate sector of the curve have been

more mean reverting than front-end butterflies. To

determine how mean reversion of various butterfly trades

performed in 2011, we tested a simple trading rule that

involved initiating a long position in the belly versus the

wings when the belly appeared cheap by more than 1.5

standard deviations on a curve and level neutral basis. The

reverse was done when the belly appeared rich. The trade

was then held for a 1-month period. Exhibit 42 shows the

results of such a trading rule and shows that intermediate

and long-end butterflies had higher risk-adjusted returns

by following such a strategy. By contrast, front-end curve

Exhibit 40: A long position in 5s versus 2s and 10s increasingly

becomes a steepener in a rally and a flattener in a sell-off 3-month holding period P/L from initiating a long position in a level and curveneutral 2s/5s/10s butterfly when it carries positively (bp), regressed against thechange in the 2s/10s curve during the holding period (%)

-35

-30

-25

-20

-15

-10

-5

0

5

10

15

-1.2 -1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8

Change in 2s/10s curve over the holding period; %

Y = -4.03 X1 - 19.43 X1̂ 2 - 5.34R² = 29.0%standard error = 8.26

period = Oct 12,10 - Aug 12,11

Exhibit 41: The performance of intermediate butterflies is well

correlated with carry at inception 3-month holding period P/L of a long position in 7s versus 5s and 10s versus 3-month carry and roll in the trade at inception; bp

-6

-4

-2

0

2

4

6

8

10

-2.5 -2.0 -1.5 -1.0 -0.5 0.0 0.5 1.03m carry and roll in weighted trade at time of initiation; bp

Y = 3.0385 X1 + 4.1926R² = 63.35%standard error = 2.0730period = Jan 01,11 - Aug 12,11

Exhibit 42: Butterflies in the long end of the curve are more likely tobe mean-reverting than front-end butterflies 1-month holding period returns from initiating level and curve neutral butterflieswhen the belly is mispriced by at least 1.5 standard deviations in 2011; bp unlessotherwise specified

Bfly Avg P/L Min Max # Trades Hit rate Std Dev Risk adj returns

10y15y20y 1.5 1 3 20 100% 0.5 3.00

10y12y15y 0.5 0 1 21 100% 0.2 2.88

5y10y15y 5.0 -7 8 20 95% 3.3 1.50

5y7y10y 3.8 -2 7 8 88% 2.8 1.36

7y15y20y 1.8 0 4 34 85% 1.4 1.285y10y20y 6.2 -10 12 34 91% 5.8 1.06

3y5y7y 2.1 -4 7 14 71% 3.8 0.55

2y5y10y 1.5 -11 12 17 65% 6.7 0.22

2y5y7y 0.7 -9 6 16 63% 5.0 0.15

1y3y5y 0.2 -5 28 16 13% 8.0 0.03

10y20y30y -0.2 -4 9 5 20% 5.3 -0.04

2y3y5y -0.3 -4 6 19 37% 2.4 -0.13

1y2y3y -0.6 -3 1 2 50% 2.8 -0.20





November 24, 2011

Terry BeltonAC

(1-212) 834-4650




125

trades exhibited weak mean reversion, with the simple

trading rule producing poor risk-adjusted returns.

In sum, our yield curve bias in 2012 will be to favor

negative carry curve trades in the front end of the

curve where convexity is attractive. In intermediates,

we expect positive carry trades to outperform and will

look to add to long carry positions as high carry points

become cheap on the curve.

Supply cyclicals

Trading strategies around Treasury auction concessions

had a mixed track record in 2011, with returns weakest intrades around 2-, 3-, and 5-year auctions. The largest and

the most consistent supply concessions have been in the

very long end of the curve. Exhibit 43 shows the

performance of various trades in the days before and after

auctions in different sectors. As shown in the exhibit,

hedged for the level of rates, the 7s/30s curve has

steepened going into the bond auction in 9 out of 11

months this year by an average of 4bp. Notably, this

steepening has persisted despite Fed purchases in the very

long end via Operation Twist. Exhibit 44 shows that even

though the magnitude of the supply concession has

been smaller since Operation Twist began, it is still

sizable.

With liquidity poor, we expect these concessions to persist

in 2012, and we favor curve steepeners going into bond

auctions. Our outlook is motivated in large part by our

expectation that market depth will likely remain poor in

1H12, as bank deleveraging continues. Historically,

supply concessions have increased in periods of poor

market liquidity, as shown in Exhibit 45. Thus, we

Exhibit 43: Supply concessions have been the largest and most

consistent going into the bond auction, a trend we expect willcontinue in 2012 Average P/L in select weighted curves* from being short the sector beingauctioned in the days prior to its auction versus being long the sector in the dayspost auction calculated between January and October 2011; bp unless otherwisespecified

Avg Chg Hit rate Avg Chg Hit rate

2s3s 1.5 80% 1.0 60% 2y

2s10s 2.4 50% 1.7 60% 2y

2s3s 1.0 80% -0.6 50% 3y

3s5s 1.0 40% 1.6 60% 3y

3s7s 1.4 40% 0.7 80% 3y

3s5s -0.8 50% -0.3 60% 5y

5s10s 0.9 40% 0.7 60% 5y

5s30s 3.0 50% 2.5 70% 5y

3s7s 2.5 70% 4.6 70% 7y

7s10s 0.3 50% 0.5 60% 7y

7s30s 2.7 60% -1.2 50% 7y

2s10s 0.4 70% 1.1 50% 10y

5s10s 1.1 70% 0.6 30% 10y

7s10s 1.6 70% 0.0 40% 10y

2s30s 1.6 80% 1.1 60% 30y

5s30s 4.1 80% 1.1 40% 30y

7s30s 4.0 80% 3.6 50% 30y

Wtd.

Trade

Pre auction Post auctionCentered

around

auction in

* Weighted trades include 0.57*3s-2s, 0.31*10s-2s, 0.57*5s-3s, 0.58*7s-3s,0.85*10s-5s, 0.64*30s-5s, 0.95*10s-7s, and 30s-0.85*7s. Curves are weighted to

be level neutral.

Exhibit 45: Poor liquidity tends to amplify auction cyclicals30-year – 0.85 * 7-year Treasury rates averaged in the business days around the30-year auction in periods of high market depth* and low market depth (Data over last 3-years has been used); %

1.65

1.70

1.75

1.80

1.85

1.96

2.01

2.06

2.11

2.16

-12 -10 -8 -6 -4 -2 0 2 4 6 8 10 12

Business days around 30-year auction

High marketdepth

Low marketdepth

* Market depth is calculated as the 10-day moving average of average size of thetop three bids and offers, in $mn, for the on-the-run 10-year Treasury note,averaged between 8:30 a.m. and 10:30 a.m. daily. Periods with market depth of greater than $100mn are classified as high market depth periods.

Exhibit 44: While supply concessions ahead of the bond auction

have fallen since Operation Twist began, they are still sizable Average change in 30-year - 0.85 *7-year curve in eight business days before thebond auction during QE2 (January to June 2011), post QE2 (July and August)and Operation Twist (October and November); bp

0

2

4

6

8

10

QE2 Jul-Aug Operation Twist

Average in different periods

2011 YTD average





November 24, 2011

Terry BeltonAC

(1-212) 834-4650




126

recommend that investors actively position for a steeper

curve ahead of bond auctions, particularly in periods in

which risk aversion is elevated and balance sheet is

constrained.

Trading around the Fed

The Fed’s purchase operations still continue to provideshort-term trading opportunities in the Treasury market.

Its preference to consistently purchase securities that are

cheap to a fitted curve has resulted in further compression

in the dispersion of valuations along the Treasury curve,

as measured by the root mean square error of a fitted

Treasury curve (RMSE). Even though this has been a

well-identified trading theme for some time (through QE1

and QE2), the strategy of owning issues cheap on the

curve in anticipation of Fed buying has remained

profitable, even under Operation Twist. Exhibit 46

presents the results from following a simple daily strategy

of buying issues that are cheap on the curve and selling

similar maturity issues that are trading rich; the strategyhas been most profitable in the 6- to 8-year sector, and 25-

to 30-year sector. As a result, the dispersion of valuations

in these sectors has compressed to their lowest level in a

year (Exhibit 47). On the other hand, as also shown in the

exhibit, dispersion in the 8- to 10-year sector still has

room to compress further relative to its recent history.

Typically, the richest issue in this sector is the on-the-run

10s, and these have historically tended to cheapen relative

to the older securities as they get reopened in the weeks

Exhibit 48: New-issue 10s have cheapened relative to old 10s inthe weeks after they have been auctioned

Adjusted* and unadjusted on-the-run / triple olds yield spread in the days after anew 10-year is auctioned; %

-0.070

-0.065

-0.060

-0.055

-0.050

-0.045

-0.040

-0.035

-0.030

0.120

0.125

0.130

0.135

0.140

0.145

0.150

0.155

0.160

0 10 20 30 40 50 60

# of business days after a new 10-year auction

New/ olds spread adj. for 7s10s curveNew/ olds spread

* Adjusted for the 7s/10s yield curve. Equals on-the-run 10-year yield – tripleolds yield – 0.25 * 7s/10s curve

Exhibit 49: The yield pickup of synthetic assets constructed bybuying JGBs/ Bunds and swapping back to USD is substantial…Yield on synthetic asset* minus Treasury yield for various government bondsand sectors; bp

JGB Bunds JGB Bunds UST

2y 117 30 2.1 0.2 0.3

3y 124 23 1.6 0.2 0.2

5y 133 15 1.1 0.4 0.4

Sector

Spread pick of synthetic

versus US Treasuries; bp

Annualized risk-

adjusted return**

* 2-year synthetic yield equals JGB 2-year yield – basis swap spread + (USDswap yield – Yen swap yield) + 6M/3M Libor basis.** 3-month return/ risk. 3-month return is one-fourth the yield plus the roll. Risk is(5-year standard deviation of 3-month changes in yield)*(modified duration).

Exhibit 46: A strategy of buying issues with the highest yield errors

versus issues with the lowest yield errors has been profitable duringOperation Twist Results from back-testing a simple trading rule*; table shows average change**,standard deviation of changes, minimum change, average change/ standarddeviation and hit rate*** for par asset swap spread difference and yield differencefor each bond pair

Avg chg Min chg Avg/Std.Dev Hit Rate Avg chg Min chg Avg/Std.Dev Hit Rate

6-7 1.3 -1.3 0.8 80% -1.0 -4.3 -0.6 40%

7-8 2.1 0.3 1.9 100% 0.9 -2.1 0.6 67%

8-9 -0.7 -2.2 -0.6 20% 0.8 -1.0 0.7 73%

9-10 -2.0 -4.9 -1.0 13% 0.9 0.4 2.6 100%

10-15 1.3 -4.1 0.5 80% -1.9 -5.2 -0.9 13%

15-20 -1.6 -3.2 -1.5 7% -0.4 -2.8 -0.3 53%

24-27 1.2 0.6 3.2 100% 0.8 0.1 1.6 100%

27-30 0.3 -0.6 0.6 80% 0.5 -0.3 0.9 80%

Yrs to

mat

Par asset swap Yields

* Trading rule involves buying the issue with the highest yield error in each maturitybucket versus issue with the lowest yield error in each bucket on a daily basisstarting September 21, 2011. Holding period is one month.** Positive number for average change implies that bond with highest yield error outperformed the bond with the lowest yield error *** Hit rate equals number of trades in which spread widened/ total number of trades in each maturity bucket

Exhibit 47: Mispricings are largest in the 8- to 10-year sector of the

curve RMSE of Treasuries in select parts of the Treasury curve; bp

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

May 11 Aug 11 Nov 11

6y-8y8y-10y20y-30y

Latest observation





November 24, 2011

Terry BeltonAC

(1-212) 834-4650




127

after they are first auctioned (Exhibit 48). In general, we

would favor overweighting off-the-run 10s versus newly

issued on-the-runs to benefit from supply-related

cheapening as well as the Fed’s preference for purchasing

cheap off-the-run issues.

Creating front-end synthetic Treasuries

The cross currency basis of USD with JPY and EUR was pushed further into negative territory this year, as the

pullback by US money funds lending to European banks

resulted in increased demand for USD funding. At its

current level of -80bp, the 2-year JPY/USD cross-currency

basis is near the lows of the last decade, while the

EUR/USD basis is at its lowest since December 2008.

This has made it very attractive for investors to create

synthetic USD assets constructed by buying foreign

government bonds that are swapped back to USD (see

grey box).

Synthetic Treasuries and cross currency basis swapsA cross-currency basis swap is an agreement to exchange floating rates in a foreign currency against USD Libor. The transaction involvesan exchange of foreign currency up front with floating payments exchanged over time. The floating payments equal USD Libor versusforeign Libor plus a spread where, by convention, the spread on the foreign Libor is referred to as the basis swap spread ( Exhibit A).Currently, spreads for 2-year USD/JPY basis swaps equal -80bp meaning an investor can pay 3-month Yen Libor minus 80bp versusreceiving 3-month USD Libor flat for the next eight quarters.The large negative basis spread currently makes it attractive for a USD floating-rate investor to create synthetic floating rate USD assets by

buying floating rate assets in Yen and swapping them to USD floating-rate assets with a cross-currency basis swap. For example, a USinvestor can synthetically create a 2-year floating rate USD asset that earns USD Libor + 80 bp by 1) investing in a Yen floating-rate assetthat pays Yen Libor flat and 2) entering in a cross-currency basis swap at a spread of -80.The yield pick-up embedded in the negative basis swap also creates opportunities for fixed-rate investors. An investor benchmarked to 2-

year Treasuries, for example, can create a synthetic 2-year Treasury note with a much higher yield by instead buying 2-year JGBs and usingthe basis swap and interest rate swap markets to convert the cash flows back into fixed-rate USD flows. To convert the JGB cash flows intofixed-rate USD flows, the investor 1) buys the JGB and asset swaps it into a semiannual Yen floater, 2) uses a 3s/6s basis swaps to convertJPY semiannual cash flows into quarterly cash flows, 3) enters a cross currency basis swap (paying Yen Libor + spread versus receivingUSD Libor) to convert it to a USD floater, and 4) converts the floating rate USD cash flows into fixed semiannual cash flows using plainvanilla interest rate swaps. Of course, these trades can be collapsed into a single trade to effectively convert the 2-year JGB cash flows intoa 2-year semiannual fixed-rate USD asset. The cash flows on this trade have the investor paying a Yen coupon equal to the yield on theJGB (offsetting JGB cash flows) and receiving a USD coupon (Exhibit B). The yield on the synthetic 2-year Treasury can be calculated as:Synthetic yield = JGB asset swap spread + 6M/3M Libor basis – basis swap spread + USD fixed swap rateAt current levels, this implies

Synthetic 2-year yield = -25 + 12 - (-80) + 77 = 1.44% or 117bp higher than 2-year Treasury notes.

Exhibit A: Basic cross currency swap Exhibit B: 2-year fixed semiannual currency swap

JPY 7,716 mm initial exchange

Investor Dealer Yen Libor + spread

USD Libor

US $ 100 mm initial exchange

JPY 7,716 mm initial exchange

Investor Dealer JPY 0.12%

USD 1.44%

US $ 100 mm initial exchange





November 24, 2011

Terry BeltonAC

(1-212) 834-4650




128

JGBs in particular look very attractive when swapped

back into US dollars. A trade constructed in this fashion in

the 2-year sector currently results in a pick-up of 117bp

versus similar maturity Treasuries (Exhibit 49). Not only

is this incremental yield high relative to recent history

(Exhibit 50), but the unlimited USD swap lines provided

by the Fed limit the potential widening in the EUR/JPY

basis. Currently, the 3-month JPY/USD FX basis is

trading at -70bp. Converting this basis (which is quoted

versus Libor) to an OIS basis (OIS basis = Libor basis +

JPY 3M Libor/OIS spread – USD 3M Libor/OIS spread)

translates to an OIS basis of -100. Because this spread is

equivalent to the penalty rate the Fed charges banks to use

its currency swap line (OIS+100), we should not expect

any further widening in the 3-month basis. This limits the

downside risks in buying JGBs swapped to US dollars.

Outlook on STRIPS

The size of the P-STRIPS market is unchanged this year,

with $197bn outstanding. However, the intra-year pattern

exhibited more volatility: net stripping activity picked up

in Q2 bringing the market to its largest size in a decade,

but fell in 2H11. The quarter ending October 2011 saw the

largest 3-month decline in the size of the market since

January 2009, as the very long end of the curve flattened

massively and Operation Twist started taking long-end

supply out of the market (Exhibit 51). Investor demand

for STRIPS remained largely in the very long end of the

curve. P-STRIPS with more than 17 years to maturity now

comprise 66% of the market versus 60% at the beginning

of the year (Exhibit 52).

Exhibit 53: Low yields, a flat curve, and Operation Twist are likelyto keep net stripping activity muted in 1H12

A model for quarterly changes in P-STRIPS outstanding; $bn

Variable Coefficient T-statistics

Intercept -55.3 -5.7

10-year yields; % 10.5 5.4

10s/30s curve; % 24.9 5.8

Change in S&P 500; points 0.01 1.7Quarterly Treasury duration purchases

by the Fed; $bn of 10-year equivalents -0.0209 -1.5

R2 = 50%; Std. error= 5

Model fitted for quarterly data over a 5-year period. 10-year Treasury rates andthe 10s/30s curve are quarterly averages.

Exhibit 52: STRIPS in the very long-end of the curve continued togain market share in 2011P-STRIPS outstanding by sector as a percentage of total; %

0

10

20

30

40

50

60

70

2002 2005 2008 2011

<5.5-yrs5.5-to 10-yrs10- to 17-yrs17+ yrs

Exhibit 50: …and at multi-year highsSynthetic 2-year Treasury yield* minus 2-year US Treasury yield; %

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4


* Synthetic 2-year Treasury yield equals JGB 2-year yield – basis swap spread +(USD swap yield – Yen swap yield) + 6M/3M Libor basis.

Exhibit 51: The size of the STRIPS market fell in 2H11P-STRIPS outstanding; $bn

170

175

180

185

190

195

200

205

210

Nov 09 May 10 Nov 10 May 11 Nov 11





November 24, 2011

Terry BeltonAC

(1-212) 834-4650




129

Looking ahead, we expect net stripping activity to stay

muted and for the market to continue to shrink in 1H12.

This view is in part motivated by our outlook that the

market environment will be characterized by high risk

aversion as the crisis in Europe escalates, which will keep

yields low and the very long end of the curve flat causing

risky assets to underperform. Historically, a flat curve,

low yields, and weak equities have all caused the STRIPS

market to shrink (Exhibit 53).

Fed purchases of Treasuries via Operation Twist will also

have a significant impact and is likely to result in

increased reconstitution activity for the purpose of

delivering whole bonds into the Fed. As a rule of thumb,

quarterly Treasury duration purchases of $100bn of 10-

year equivalents by the Fed have lowered P-STRIPS

outstanding by around $2bn per quarter. In 1H12, we

expect duration buying by the Fed to average $180bn of

10-year equivalents per quarter, suggesting that the market

should shrink by $4bn per quarter due to this factor alone.

With 32% of Fed purchases in maturities greater than 20-

years and the bulk of the STRIPS market concentrated in

that sector of the curve, we expect reconstitutions to be

particularly high. Indeed, there was already evidence of

this in October (the first month of Operation Twist).

Exhibit 54 shows that issues that were purchased in size

by the Fed in October were also the ones that were

reconstituted the most.

In sum, we look for the STRIPS market to shrink

further in 1H12. Whole bonds that are trading cheap

relative to the fitted Treasury curve will likely lead the

way in reconstitution activity (Exhibit 55), since the

Fed concentrates its purchases in such issues (almost

75% of its purchases have occurred in positive yield

error bonds thus far in Operation Twist). This has two

implications for the STRIPS market. First, shorter

maturity Cs should richen versus similar maturity

Ps, since reconstitution of longer maturity bonds will

reduce the relative supply of shorter maturity Cs.

Indeed, this has been one driver of the C-P spread

recently (Exhibit 56). This richening of Cs will likely

occur most in Cs that correspond to coupon dates of

highly stripped whole bonds that have positive yield

Exhibit 55: Issues with high yield errors like Aug-40s and May-40s

will likely lead reconstitution activity given the Fed’s preference for issues trading cheap on the curve Amount stripped by issue ($bn) versus current yield error (bp)

F-36

F-37

M-37F-38

M-38

F-39

M-39 A-39

N-39F-40

M-40

A-40

N-40F-41

M-41 A-41

N-41

0

5

10

15

20

25

30

-2.5 -2.0 -1.5 -1.0 -0.5 0.0 0.5

Yield error; bp

Exhibit 54: Operation Twist Impact: Issues that were purchased in

size by the Fed in October were also reconstituted the most duringthe monthChange in stripped amount for the top 10 issues purchased by the Fed in the 10-to 30-year sector in October versus Fed purchases of the bond in October; $bn

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

0.0 0.5 1.0 1.5 2.0 2.5 3.0

Fed purchases in Oct; $bn

Exhibit 56: Increased reconstitutions are likely to drive C-P spreadsnarrower in the coming monthsCoupon minus Principal STRIPS yield spread in the 2022-23 sector* (%) versusthe total amount of P-STRIPS outstanding ($bn)% $bn

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.08

0.09

170

175

180

185

190

195

200

205

210

2010 2011

Average C-P spreads; %P-STRIPS outstanding; $bn

* Average C-P spread between issues maturing on 8/15/22, 11/15/22, 2/15/23and 8/15/23.





November 24, 2011

Terry BeltonAC

(1-212) 834-4650




130

errors in the very long end of the curve. Second, Ps

corresponding to bonds that are likely to be reconstituted

should richen relative to other surrounding Ps.

In sum, we expect the size of the STRIPS market to

decline further in 2012. Short maturity Cs should richen

versus similar maturity Ps, as should Ps on long-dated

bonds that are currently cheap to the par curve.

Trading themes

• Stay long duration in early-2012 targeting 10-year

yields to reach 1.70% in 1Q12

A worsening of the European sovereign debt crisis

should accelerate the flight-to-quality bid into

Treasuries early next year pushing yields lower; rich

valuations, a deteriorating supply/demand imbalance,

and poor technicals will limit the upside, however, and

keep rates range bound later in 2012.

• Position for lower front-end yields in 2012

Front-end rates should trend lower in 2012 driven by

declining front-end supply, further policy rate guidance

by the Fed that flattens term premium, an increase in

the supply of excess reserves following QE3, and the

end of Operation Twist; look for GC repo to average

near 5bp, and 2-year yields to fall to 17bp.

•

Avoid crowds in 2012A secular decline in liquidity will lead to heightened

whipsaw risk in 2012 and cause Treasury yields to be

volatile within a broad trading range; trading strategies

based on our measure of investor positions that fade

crowded consensus trades are likely to perform well in

2012 as they did this year.

• Create synthetic 2-year Treasury notes by buying

JGBs and currency swapping them back to USD

Synthetic Treasuries constructed by buying foreign

bonds and currency swapping them back to US dollars

are at the most attractive levels of the last decade. The

yield on a synthetic 2-year note created by assetswapping a 2-year JGBs currently equals 1.44% or 117

bp higher than 2-year Treasury yields.

• We expect positive carry yield curve trades in the

front end to perform poorly in 2012; favor convexity

over carry in the front end of the curve and carry in

the intermediate sector of the curve

Positive carry yield curve trades in the front end are

likely to perform poorly reflecting the extreme negative

convexity in these trades. We recommend negative

carry curve trades in the front end of the curve where

convexity is attractive. In intermediates, we expect

positive carry trades to outperform and will look to add

to long carry positions as high carry points become

cheap on the curve.

• The long end of the curve is biased slightly steeper;

our mid-year target for the 10s/30s Treasury curve

is 110bp

The 10s/30s curve is slightly flat to fair value; we favor

positioning for a steeper curve in advance of long-end

supply to take advantage of the large supply concession

in the long end.

•

An increase in reconstitutions in the STRIPS marketshould cause shorter maturity Cs to outperform Ps

We expect the size of the STRIPS market to decline

further in 2012. Short maturity Cs should richen versus

similar maturity Ps, as should Ps on long-dated bonds

that are currently cheap to the par curve.





November 24, 2011

Srini RamaswamyAC

(1-212) 834-4573

Praveen Korapaty (1-212) 834-3092

Alberto Iglesias (1-212) 834-5116


131

US Interest Rate Derivatives

• There is a new normal in the swaps market—

convexity-hedging flows and issuance-related

swapping matter less, but European markets and

Fed purchases matter more

• FRA-OIS spreads should remain under pressure

into 1Q12, but concerted central bank policies

should cause it to narrow thereafter—look for

FRA-OIS to widen in the short term, but

eventually narrow to 50bp by mid-year.

Intermediate maturity swap spreads are similarly

biased wider in the near term, but should end

1H12 close to current levels

• Thanks to sticky front-end Treasury yields,

short-expiry swaptions on short tails are

effectively like options on front-end spreads. We

recommend utilizing 1Mx1Y payer / receiver

swaptions to initiate asymmetric exposure to

front-end spread widening / narrowing

• Fed purchases of longer-end Treasuries will offer

opportunities in 1H12—trade the impact of intra-

month swings in purchase pace on swap spreads,

and look to trade maturity-matched swap spread

switches based on the Fed’s metric of value

• Initiate steepeners between intermediates and the

long end hedged with Eurodollars, to position for

yield curve normalization as well as carry

• Look to initiate positive carry, belly richening

butterflies as attractive ways to gain empirically

convex exposure to higher front-end yields

• Yield curves between the front end and

intermediates will remain directional in 2012, but

conditional curve trades are not cheap—initiate

“synthetic” conditional curve trades, created by

replacing swaptions at the front end with YCSOs

• Approach 2012 with a long gamma bias, but look for 3Yx10Y to decline to 6bp/day by 1Q12 end

• Bermudan receiver swaptions offer directional

vega exposure, as well as exposure to cheap

implied correlation and forward volatility. We

introduce a novel approach to measure relative

value in Bermudan swaptions, based on using

approximating Canaries

Swaps

Swap spreads across the curve traded in significantly

wide ranges over the past year, and have recentlywidened to the upper end of those ranges (Exhibit 1).

The move wider late in the year has been mainly due to

the steadily deteriorating sovereign debt crisis in Europe,

which has now spread well beyond its origins in Greece.

The resulting moves in swap spreads have been

reminiscent of 2008, although trading ranges this year

remain small in comparison; 10-year spreads, for

instance, traded in a 75bp range in 2008.

The volatility in swap spreads is a reflection of the fact

that spreads have been influenced by numerous factors in

significant ways. Intermediate spreads narrowed going

into the second quarter as markets priced in the end of

QE2, but worsening conditions in Europe steadily have

pressured spreads wider since then; more recently, the

insufficient nature of policy measures announced after

the recent EU summit has sparked renewed widening in

spreads, bringing them closer to the wides of the year in

the 2- to 10-year sector of the curve.

In short, spreads have been largely driven by the Fed’s

Treasury purchases and by the European banking system

Exhibit 1: Swap spreads are near the upper end of a relatively wide

trading range in 2011Statistics regarding maturity matched swap spreads in various benchmark sectors;bp

* As of 11/17/2011

Exhibit 2: Collateralization rates have been on the rise across all OTCderivative marketsPercent of trade volume subject to credit support agreements

Source: ISDA Margin Survey

Current Start Average High Low Range

2Y 53.2 18.1 23.7 53.2 12.6 40.6

3Y 50.1 23.9 28.2 50.1 20.4 29.7

5Y 44.3 15.9 25.0 44.3 16.2 28.1

7Y 34.1 9.6 20.0 34.1 10.5 23.5

10Y 19.3 5.5 11.5 22.7 3.9 18.8

30Y -27.3 -21.6 -26.3 -19.8 -39.0 19.2

2011 YTD statistics

2003 2004 2005 2006 2007 2008 2009 2010 2011

All OTC derivatives 30 51 56 59 59 63 65 70 70

Fixed Income derivatives 53 58 58 57 62 68 63 79 79

Credit derivatives 30 45 59 70 66 74 71 93 93

FX derivatives 21 24 32 37 36 44 36 57 58

Equity derivatives 27 45 45 46 51 52 52 71 72

Commodities 17 25 33 44 40 39 39 62 60





November 24, 2011

Srini RamaswamyAC

(1-212) 834-4573




132

crisis—factors that swap market participants would not

normally view as endogenous. Indeed, endogenous

factors that have historically exerted considerable

influence on swap spreads—mortgage market hedging

flows and the swapping of corporate issuance, as well as

longer-term drivers such as budget deficit expectations

and funds rate expectations—have played a less

significant role this year. With deficit expectations

remaining largely stable and with the Fed on hold for

several years, longer-term drivers have not experienced

enough volatility to influence swap spreads in a

significant way. Also, the numerous well-understood

hurdles to refinancing mortgages have made mortgages

ever less negatively convex; this fact, together with

falling portfolios at the two major GSEs, has largelyhelped to explain the decline of mortgage-hedging flows

as a driver of swap spreads. Somewhat less permanently,

low and sticky yield levels appear to have lessened the

imperative to swap fixed rate debt issuance on the part of

many high grade corporate issuers, including those who

traditionally swap their issuance.

Such behavioral shifts on the part of market participants

are only one among numerous changes impacting the

swaps market. Another key transformative development

is of course the ongoing implementation of the Dodd-

Frank reforms, as they pertain to the derivatives markets.

Much uncertainty remains in this regard, and we discussthe current status of regulatory implementation; however,

one key change that has already taken place in the swaps

market is a move towards OIS discounting as a quasi-

standard for valuing off-market swaps. To be sure, this is

not exactly a new development, and fixed income

derivatives markets have led the broader trend towards

higher levels of collateralization in recent years

(Exhibit 2). Given this trend, the move towards OIS

discounting (which is arguably closer to the funding rates

for commonly used collateral) has been in the making for

sometime now. Nonetheless, there is growing evidence

that markets have gravitated towards value measures that

are more consistent with OIS discounting. For instance, par asset swap spreads for premium/discount bonds

appear to serve as better signals of value when computed

using OIS discounting, as opposed to Libor discounting.

In other words, swap market participants will need to

adjust to a world where the reference curve (with respect

to which asset valuations are measured) and the discount

curve are different; we discuss this further in a later

section.

Swap spread drivers: the new normal

As we consider the likely outlook for intermediate

maturity swap spreads over the course of next year, it is just as important to recognize what will not be driving

swap spreads. These include several traditional and long-

standing factors that influenced swap spreads in the past,

such as (most notably) budget deficit expectations. To be

sure, should the path of fiscal policy deviate from

expectations in a sufficiently large way, budget deficit

expectations will undoubtedly re-emerge as a key driver

Exhibit 3: No convexity to hedge—the magnitude of convexity in current

coupon mortgages is near historical lows Absolute value of FNMA 30-year current coupon TBA Libor OA-convexity;

Exhibit 4: The declining footprint of mortgage hedgers in the swapsmarket is also due to mandated reductions in the GSEs’ retainedportfoliosTotal outstanding swaps notional amount held by Fannie Mae and Freddie Mac,versus their total retained portfolio size;$bn $bn

0.0

0.5

1.0

1.5

2.0

2.5

3.0

Feb 00 Oct 02 Jul 05 Apr 08 Jan 11

Current

1400

1450

1500

1550

1600

1650

1700

800

1000

1200

1400

1600

1800

2000

Dec 06 May 08 Sep 09 Feb 11

Retainedportfolio

Swapnotional





November 24, 2011

Srini RamaswamyAC

(1-212) 834-4573




133

of swap spreads. In this regard, it is still possible (but

perhaps not likely) that fiscal policy adjustments will be

enacted sometime next year; despite the initial failure of

the Congressional super committee, mandatory cuts do

not take effect until 2013, and thus significant time

remains for Congressional action on fiscal policy.

A tactical model for benchmark 10-year maturity matched swap spreadsIt is no exaggeration to say that the world of swap spreads is experiencing its own “new normal.” Traditional, long-term drivers have simply become lessvolatile and more range-bound—Fed funds expectations are virtually assured to remain stable for a long period of time, and budget deficit expectationshave become quite stable. Traditional tactical factors have become less relevant too; declining portfolio sizes at the GSEs, and structural impediments torefinancing mortgages have made mortgage-hedging flows relatively insignificant in terms of driving swap spreads. In addition, anecdotal evidence

points to reduced swapping of debt issuance, thanks to low outright yield levels.

Long-term structural factors remain relevant, and budget deficit expectations could still prove significant in 1H12 if fiscal policy were to deviatematerially from current expectations. That said, trading swap spreads in the 10-year sector now requires a different, more tactical model to assess fair value, which is outlined here. This model is more short term in nature out of necessity, given the recent nature of the developments discussed above. Thefactors incorporated in our model are (i) 6-month FRA-OIS spreads (to capture bank funding pressures and Libor expectations), (ii) the bank stock index(a broader measure of financial system health), (iii) weekly open market gross purchases (i.e., not net of sales at the front end) of Treasuries by the Fed(which matters for swap spreads even under a balance sheet neutral construct as with Operation Twist), and (iv) weekly swapped corporate issuance.

The first three factors are relatively straightforward; the last—swapped issuance—has been more interesting, because low-yield levels have likelydeterred the amount of fixed rate issuance that has been swapped this year. Understanding the aggregate swapping activity of high grade debt issuers isdifficult given sketchy data, but necessary nonetheless since moves in yields could alter swapping behavior by material amounts. Therefore, here we

attempt to estimate the fraction of high grade issuance that gets swapped as a function of rates. Our approach is as follows. For a representative sample of non-financial companies over a 3-year period, we examined data on the aggregate amount of fixed-rate issuance (as a percentage of overall fixed rateissuance) that was swapped to floating, using information from financial statements. As seen in Exhibit A1, the fraction of overall non-financialcorporate debt issuance that gets swapped to floating has indeed been sensitive to yield levels, falling sharply as 10-year swap yields fall below 3%. Alsointeresting is that the data suggests that swapping activity may rise again in very low yield environments. While the data in these ultra-low yield regimesis admittedly limited, it stands to reason that swapping activity might begin to rise again as yields fall below a second threshold, perhaps reflectingexpectations of the Fed being on hold for a very long time.

Unfortunately, such data cannot be compiled for financial issuers; it is nonetheless important to account for potential changes in their swapping behavior too, given that these institutions are both large issuers and also tend to swap considerable fractions of their issuance. Moreover, anecdotal evidencesuggests that these issuers too have curtailed their swapping activity as yields have fallen. To estimate their swapping behavior as a function of yields, wetake the leap of faith that the fraction of financial issuance that is swapped may be modeled as a linear function of the corporate issuance swap-fraction.That is, we assume that financial issuance fractions are a similar non-linear function of yields, but could differ in magnitude since financials typicallytend to swap a greater portion of their fixed-rate debt to floating. We estimate this function to be consistent with broad anecdotal evidence regarding the

percentage of financial issuance that gets swapped. To be specific, our linear transformation causes financial issuance swapping percentages to fall to30% if corporate swapping fractions go to zero, and financial issuance swapping fractions go to 100% when corporate swapping fractions go to 100%.

Exhibit A2 presents a time series of our estimate for total swapped monthly high grade issuance, calculated by multiplying actual monthly financial andnonfinancial issuance by our estimate for swapping fractions (based on yield levels at the time of issuance).

Armed with an approach to estimate issuers’ debt swapping percentages, we can now present our model for benchmark 10-year maturity matched swapspreads. This new, high-frequency model for swap spreads is detailed in Exhibit A3. As a rule of thumb, each 10bp of FRA-OIS widening implies a2.5bp widening in 10-year spreads; $10bn in weekly Treasury gross purchases is worth 1bp of widening in spreads; $10bn in swapped high gradeissuance (estimated per our framework) would narrow swap spreads by 2.5bp; and a 10 point rise in the BIX bank stock index would cause swap spreadsto narrow 1.3bp.

Exhibit A1: Issuance swapping fraction vs. yields Exhibit A2: Estimated swapped issuance Exhibit A3: A model for 10-year swap spreads

0%

5%

10%

15%

20%

25%

2.0 2.5 3.0 3.5 4.0

Low yieldperiods

High yieldperiods

Overall fit

0

10

20

30

40

50

Nov 08 May 09 Dec 09 Jul 10 Jan 11 Aug 11

Factor Coefficient T-stat Current

Intercept 20.4 6.3

FRA-OIS 0.26 10.2 72.3

Wkly Treasury purchases by Fed 0.10 4.6 10Swapped HG issuance -0.25 -4.1 1.7

BIX -0.13 -6.5 119.0

R^2

Projected 10Y swap spread

Actual

73%

23.9

19.3





November 24, 2011

Srini RamaswamyAC

(1-212) 834-4573




134

However, barring such a surprise, the fact remains that 1-

year ahead budget deficit expectations have remained

(and are likely to remain) very stable, moving in a

$200bn range. As a result, assuming there are no

unforeseen developments on the fiscal policy front,

deficit expectations will likely not particularly impact 10-

year swap spreads in 1H12. Similarly, given a

historically low level of negative convexity in mortgages

(Exhibit 3), mortgage hedging flows have not been, and

will likely not be, a significant driver of spreads in 2012.

The footprint of mortgage hedgers in the swaps market is

also declining for a much more obvious reason—

mandated caps on the retained portfolios of Fannie Mae

and Freddie Mac point to a decline in the swap hedging

activity of these hedgers (Exhibit 4), who havetraditionally been sizeable players in the swaps market.

In order to reflect this altered market environment that

has now become the “new normal,” we have augmented

our more traditional longer-term benchmark 10-year

swap spread model with a considerably revised tactical

model. Not only is this model more short term in nature

(out of necessity given the recent nature of the

developments discussed above), but it also addresses

newly emergent tactical forces impacting swap spreads—

Fed purchases of Treasuries at the longer end, behavioral

shifts with respect to swapping of debt issuance, and of

course the European crisis. A detailed description of thistactical model is included in the grey box on benchmark

10-year swap spreads.

Swap spread trading themes in 2012

Such a model has several implications for swap spreads

looking ahead into 1H12. First and foremost, our outlook

on the various drivers points to widening in intermediate

swap spreads heading into year-end and in 1Q12, and an

eventual narrowing (to slightly below current levels) by

mid-year 2012; this is laid out in our swap spread

forecast in a later section.

Second, swap spreads—even in the 10-year sector—willlikely remain quite vulnerable to exogenous risk from the

European crisis. Exhibit 5 presents an estimate of the

partial impact on 10-year swap spreads, if each of the

drivers moves by an amount equal to one standard

deviation of rolling 3-month changes over the past year.

Not only are FRA-OIS spreads the single most important

driver of intermediate spreads—as noted in Exhibit 5, but

they are also significantly inversely correlated to the BIX

(whose partial beta is negative), implying that the two

most significant drivers of swap spreads are likely to

move in correlated fashion, amplifying the effect of

European developments on swap spreads. Hedging swapspread positions with 35% risk (rather than the 25%

partial beta) in FRA-OIS spread positions would help

mitigate the effective risk exposure to the European

crisis.

Third, although the Fed’s Treasury purchases are likely

to be stable on average (with Operation Twist scheduled

to continue through the end of June 2012, and with QE3

unlikely to include Treasuries until after Operation Twist

is concluded), the actual schedule of Treasury purchases

Exhibit 5: Ranking swap spread drivers—FRA-OIS spreads and bank

equity prices are likely to be the dominant drivers of swap spreadsPartial impact of a 1 standard deviation move* in each of the drivers in our model for benchmark 10-year maturity matched swap spreads; bp

* Estimated as the standard deviation of rolling 3-month changes over the past year.Impact estimated as the 1 SD move in each driver multiplied by its beta in our model.

Exhibit 6: The weekly pace of Treasury purchases by the Fed can exhibitsignificant tactical variability, even as it remains stable on averageduring Operation Twist …Weekly gross purchase of Treasuries by the Fed; $bn

-3

-2

-1

0

1

2

3

FRA-OIS Fed purchases Corp. issuance BIX

0

5

10

15





November 24, 2011

Srini RamaswamyAC

(1-212) 834-4573




135

by the Fed will likely create attractive tactical trading

opportunities in swap spreads. It is interesting to note

that there can be considerable variation in the pace of

gross Treasury purchases on a week-to-week basis; for

instance, the weekly pace of gross purchases declined

from a mid-October high of $14bn to under $5bn in just

two weeks (Exhibit 6). Such tactical swings in the

weekly pace of Fed purchases tend to impact swap

spreads (adjusted for FRA-OIS spreads). To illustrate

this tactical effect, we identified the three dates that saw

the biggest 2-week upward change in rolling weekly

purchases, as well as the three dates that saw the biggest

downward change. As seen in Exhibit 7, spreads

(adjusted for FRA-OIS) narrowed in the periods that saw

sharp declines in the pace of buying, while periods that

saw a sharp pickup in the pace of purchases were

characterized by a widening in swap spreads. Given that

the Fed will be publishing schedules at the end of each

month for its purchases in the following month, shifts in pace will be known ahead of time, and this tactical effect

can therefore be traded.

Fourth, barring considerable disruptions in primary

issuance markets due to the European crisis, typical intra-

year seasonal patterns in issuance will likely offer

attractive opportunities as well. Over the past five years,

high grade swapped issuance has tended to be far below

average in the months of February and July, while

reaching highs in the months of May and September

(Exhibit 8). The best way to position for this intra-year

seasonal pattern is to initiate swap spread narrowers inearly March. Benchmark 10-year swap spreads, adjusted

for FRA-OIS, have narrowed significantly in the

March/April period in each of the past four years, and

narrowed by a modest amount in 2007. Given the

significant narrowing in spreads in this period on average

(Exhibit 9), we would look to initiate spread narrowers

in early March.

Exhibit 8: Our estimate of swapped high grade issuance also exhibitsconsiderable intra-year seasonal patterns, with the Feb–May period

seeing the biggest rise in pace …Weekly swapped high grade issuance*, averaged by month over the past five years;$bn

* J.P. Morgan estimate, based on a model for the fraction of financial andnonfinancial debt issuance that is swapped

Exhibit 9: … causing swap spreads to be biased narrower in that periodon averageCumulative change in 10-year maturity matched swap spread adjusted for 6-monthforward FRA-OIS in the 2-month period from the first business day of March,averaged over the past four years; bp

Business days after the beginning of March

Exhibit 7: … which should make for tactical trading opportunities, sincespreads have tended to narrow in periods when the Fed’s purchase

pace declines sharply, and widen when the opposite is trueCumulative change in 10-year maturity matched swap spread, net of a 35% risk-weighted 6-month forward FRA-OIS hedge, averaged in periods when the weeklypace of Treasury purchases by the Fed fell sharply* and rose sharply**; bp

# business days around selected date* Dates used are 1/5/11, 7/15/11 and 11/2/11. ** Dates used are 11/18/10, 1/20/11and 5/11/11.

3

4

5

6

7

8

9

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

-15

-10

-5

0

5

0 5 10 15 20 25 30 35 40

-4

-3

-2

-1

0

1

2

3

-10 -8 -6 -4 -2 0 2 4

Rise in pace

Decline inpace





November 24, 2011

Srini RamaswamyAC

(1-212) 834-4573




136

Modeling the duration exposure in variable annuities and its impact on 30-year swap spreads

Since the start of QE2 in 4Q10, long-end swap spreads have been mainly driven by three factors—the slope of the 2s/30s Treasury curve, the Fed’s gross purchases of Treasuries, and the shifting duration of the variable annuity universe.

Estimating the last of these three factors is itself a challenge. Pricing variable annuities (henceforth referred to as VAs) is an exceedingly complex undertaking, requiringthe joint modeling of long-term interest rates as well as equities; in addition, actuarial risks stemming from life insurance related guarantees, and other features make the

product path dependent, adding to the complexity. It is nonetheless important to capture the nonlinear dependencies of VA duration exposure with respect to equities andlong-term swap yields, since VAs have been a popular product for insurance companies for many years with over $1.5tn estimated to be outstanding, and the interest ratein these products can be highly nonlinear.

A common product is to pay policy holders a variable return based on the performance of the S&P 500 (or another benchmark) but provide downside protection byguaranteeing a minimum income stream for a specified time period. In rising equity markets, the product creates little risk to insurance companies, since equity returns aremerely passed through to policyholders. However, in falling equity markets, the minimum guaranteed income stream becomes more binding. Effectively, as the“moneyness” of the embedded equity put increases, insurance companies increasingly become short a fixed income annuity, requiring them to add long duration hedges asa result.

Recognizing that our objective is not to price variable annuities accurately, but merely to capture the trend in their duration exposures as well as their nonlinear

relationship with equities/yields, we devise a simpler approach. Our approximation approach is based on three principles. First, we start with the assumption that theduration of the VA universe may be approximated by a weighted combination of the durations of a set of much simpler “VA lite” instruments, which we refer to as “VAkernels.” This is not unlike “series approximation” techniques commonly used in mathematics to solve difficult problems. One example of a VA kernel is a simple producwhere a policyholder pays $100 on (say) January 1, 2007, intended to be invested in the equity market for a 10-year (fixed) horizon, with a guaranteed withdrawal amountof $5 per year for the subsequent 20 years. Several VA kernels may be created by varying the start date, the length of time of the intended equity investment, and theminimum guarantee amount. Effectively, we price the complex VA universe as a linear combination of simpler VA kernels, each of which is priced in a manner thatcaptures the nonlinearities with rates and equities. Second, we price the present value of this instrument by ignoring actuarial risks and using an option pricing framework.We use implied distributions from the swaptions market as well as long-term S&P vols and correlation estimates to calculate the price. We also use numerical tweaks tocalculate the partial exposure with respect to long-term swap rates (i.e., duration). Third, we use a calibration approach to solve for the appropriate weights on the variousVA kernels. Our calibration relies on the anecdotally known fact that VA risk exposures were significant influences on long-end swap spreads in certain periods of time,such as 4Q08; we may thus solve for non-negative coefficients that maximally explain the portion of long-end swap spread behavior not explained by other factors inthose select periods of time. In order to mitigate circularity (since we plan to use VA duration estimates to model long-end swap spreads), no data after 2008 has been usein calibration, and out-of-sample performance has been tested and found to be reasonable. The aggregate duration of the VA universe is shown in Exhibit B1; as can beseen this remains near historical highs and could worsen if yields and equities were to decline in 1Q12 on the back of continuing deterioration in Europe.

Armed with this estimate for VA duration, we may now model 30-year swap spreads as a function of the three factors outlined above. Exhibit B2 presents the statisticsfrom regressing 30-year benchmark maturity matched swap spreads versus the 2s/30s curve, cumulative gross purchases of Treasuries by the Fed (i.e., not net of front-en

sales, and since the start of QE1), and the duration of the VA universe in 20-year equivalents, over the past 15 months (i.e., since QE2 expectations became a significantfactor, resulting in altered dynamics at the long end of the Treasury curve). As a rule of thumb, each 10bp flattening in the 2s/30s Treasury curve would widen 30-year spreads by 2.8bp, a fall in VA duration by $10bn 20-year equivalents would cause a 5bp widening in spreads, and the cumulative effect of the Fed’s long-end purchasesunder Operation Twist ($400bn in all) should be about 3.6bp (not including indirect effects due to the impact of such purchases on the curve). As seen in Exhibit B3, thismodel has been successful in fitting the observed behavior of 30-year swap spreads.

Exhibit B1: Estimated duration of VA universe Exhibit B2: 30-year swap spread model Exhibit B3: Actual spreads versus model fair value

100

125

150

175

200

225

250

275

Jan 07 May 08 Sep 09 Feb 11

Factor Coefficient T-stat Current


2s/30s Treasury curve, % -27.7 -20.7 2.897

VA aggregate duration; $bn 20s -0.5 -24.1 240.4

Cum gross Tsy purchases by Fed; $bn 0.009 17.0 1200.5

R^2

Projected 30Y swap spread

Actual

-25.2

-27.1

74%-45

-40

-35

-30

-25

-20

-15

Nov 10 May 11 Nov 11

30-year maturity matched swap spread; bpModel





November 24, 2011

Srini RamaswamyAC

(1-212) 834-4573




137

Long end swap spreads: does VA hedging still

matter?

Swap spreads at the long end of the curve are broadly at

the mercy of two large technical forces—Fed purchases

of Treasuries (particularly at the long end of the curve),

and the shifting duration hedging needs of life insurance

companies’ variable annuity portfolios (VA duration, for

short). Although the significance of the latter has been

diminishing, as insurance companies broaden their use of

Treasury-based derivatives and/or cash Treasuries, we

believe it is too soon to dismiss VA duration needs as a

driver of long-end swap spreads.

Estimating the duration of the VA universe is a complex

undertaking, and we have outlined an approximate

scheme for doing this in previous research (see Interest

Rate Risk in Variable Annuities, J.P. Morgan Research

Note, September 28, 2011). A short summary of our approach to estimating VA duration, as well as its use in

our revised model for 30-year benchmark maturity

matched swap spreads, is presented in the grey box on

long-end swap spreads.

Several points are interesting to note with respect to our

long-end spread model. First, the partial sensitivity of

long-end spreads to VA duration has indeed been

declining in magnitude (Exhibit 10). That said, reports

of its demise as a factor impacting long-end spreads are

exaggerated, at least for now, and it is equally

important not to overstate the extent of this decline.

To all appearances, it would seem that VA hedging

has become insignificant, or even a contrarian

indicator of long-end swap spreads. Indeed, in recent

months, long-end spreads have widened as VA

duration has risen (Exhibit 11). However,

Exhibit 12: … masks a more subtle message—the recent strong equity-rate correlation has caused the 2s/30s curve to become negativelycorrelated with VA duration, producing offsetting impacts on long-endspreadsRolling 3-month correlation between weekly changes in the S&P and 30-year swapyields, versus the effective beta* of 30-year swap spreads versus VA duration;

bp per $bn 20-year equivalents

* Calculated as partial beta of 30-year spreads with respect to VA duration (asdetailed in the footnote to exhibit 10), plus the partial beta of long end spreads withrespect to the 2s/30s curve (from the J.P. Morgan 30-year spread model) times therecent beta of the 2s/30s Treasury curve with respect to VA duration. Recent betacalculated over rolling 3-month periods.

Exhibit 11: Appearances can be deceptive—the seemingly wrong-way

correlation between VA duration and long end swap spreads …30-year maturity matched swap spread versus estimated VA duration*$bn 20-year equivalents bp

* J.P. Morgan estimate

Exhibit 10: The partial sensitivity of long-end swap spreads to VA

duration has indeed been declining, but reports of its demise (as afactor impacting spreads) are as yet exaggeratedRolling partial beta of 30-year maturity matched swap spreads versus VA duration*;bp per $bn 20-year equivalents

* Based on rolling 15-month regressions of 30-year benchmark maturity matchedswap spreads versus VA duration, the 2s/30s Treasury curve and cumulative grossTreasury purchases by the Fed. Regressions are over 15 months because our current model for 30-year swap spreads is designed to cover the period from

August 2010, when QE2 became a significant factor affecting long end spreads.

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

-0.50

-0.45

-0.40

-0.35

-0.30

-0.25

-0.20

-0.15

-0.10

-0.05

0.00

0.05

May 11 Jul 11 Sep 11 Nov 11

S&P/30Y yield correlation

Effective VA beta

220

225

230

235

240

245

250

255

260

-38

-36

-34

-32

-30

-28

-26

-24

-22

-20

-18


VA duration30-year swap spread

-0.75

-0.70

-0.65

-0.60

-0.55

-0.50

-0.45


VA beta





November 24, 2011

Srini RamaswamyAC

(1-212) 834-4573




138

appearances can be deceiving, and this masks a more

subtle message, which is that in recent months VA

duration and the 2s/30s curve have become much more

negatively correlated than before, causing their impacts

to offset each other; moreover, this negative correlation

is itself a consequence of the elevated positive correlation

between long-end yields and the S&P. In other words,

increased equity-rate correlation has caused VA duration

to rise sharply as long-end yields fall and the 2s/30s

curve flattens. As a result, the partial impact of increased

VA duration needs has been mitigated by the counter-

directional exposure of long-end spreads to the curve

(Exhibit 12).

Such correlations could persist for some time, producinglimited net impact on long-end swap spreads. However,

this should be seen as a happy coincidence, and is

unlikely to persist in the medium term. Should equities

and rates decouple in 1H12, as indeed we expect them to,

such a canceling effect will be less pronounced, and the

impact of VA duration on long-end spreads will likely be

apparent once again. Therefore, we continue to rely on a

fair-value framework for long end spreads that includes

VA duration as well as the 2s/30s curve.

Front-end spreads

At the front end of the curve, swap spreads have become pure plays on the European crisis, exhibiting strong

correlation to FRA-OIS spreads but little else. This will

likely remain the case in 1H12, and opportunities in this

sector will likely revolve around trading the mispricing

in front-end spreads relative to FRA-OIS on a hedged

basis.

Thus, a view on FRA-OIS is essential to determine the

likely evolution on front-end swap spreads in 2012. To

this end, we present a simple model that attempts to

explain the behavior of FRA-OIS differentials in recent

months. USD FRA-OIS spreads have been driven by two

factors recently—the EUR/USD OIS FX basis, and semi- peripheral European sovereign CDS spreads. (Although

the commonly quoted EUR/USD FX basis is with respect

to a Euribor/Libor basis swap, it is more useful to adjust

this quoted spread to be with respect to an EONIA/OIS

basis swap. We do this because (i) EONIA and OIS rates

are more reflective of the marginal cost of funds for

banks in the Eurozone and the US, and (ii) the existence

of the Fed’s dollar swap lines at a 100bp penalty bounds

this EUR/USD OIS FX basis at -100bp. For a more

detailed discussion of cross-currency basis swaps, see Decoding the FX basis market for signs of $ funding

stress, Pavan Wadhwa et al , August 23, 2011.

A more negative basis implies a higher USD borrowing

rate at which a European borrower would be indifferent

to borrowing in USD versus borrowing in euros and

swapping to USD, and may thus be thought of as a

fundamental funding market factor. We also use semi-

peripheral spreads as a second factor to capture the broad

correlations between credit spreads and metrics of

European stress.

Exhibit 13 presents statistics regarding such a model,

and also uses the model parameters to project 3-month

forward FRA-OIS spreads by year end and by mid-year

2012. To make these projections, we are assuming that in

the near term, conditions deteriorate further into year end

and in 1Q12, and the EUR/USD OIS FX basis reaches -

100bp, which should act as a near-term floor thanks to

unlimited central bank funding. We also make

projections for European stress metrics based on

estimates from our European rates strategists, and look

for a near-term widening in our measure of semi-

Exhibit 13: Our model for FRA-OIS spreads suggests that FRA-OIS

could widen into year-end, but narrow going into mid-year Statistics from regressing 3-month forward constant maturity FRA-OIS differential (bp)versus the 3-month EUR/USD FX OIS basis (bp) and semi-peripheral sovereign CDSspreads* (bp), and projections for FRA-OIS at end of 4Q11 and 1H12 based onassumptions for the drivers.

* Average of Italy, France and Spain 5Y CDS spreads. Regression based on 6 monthsof history, and are as of COB 11/18/2011. Shaded values represent projections.

Exhibit 14: Swap spread forecastJ.P. Morgan projections for maturity matched swap spreads in various benchmarksectors; bp

* Current values as of COB 11/18/2011. Projections assume 3-month forward FRA-OIS spreads evolve according to the trajectory shown in the previous exhibit.

Factor Coefficient T-stat Current 1Q12 1H12

Intercept 4.6 3.8

EUR/USD FX OIS basis -0.43 -11.9 -73 -100 -40

Semiperipheral spreads 0.06 7.7 406 450 475

R^2

FRA-OIS fair value 60 74 50

Actual 63

95%

Current 1Q12 1H12

2Y swap spread 49 53 38



30Y swap spread -29 -23 -23





November 24, 2011

Srini RamaswamyAC

(1-212) 834-4573




139

peripheral spreads. Finally, we assume that policy actions

to be enacted sometime in 1H12 will include concerted

actions by central banks (alluded to by Rosengren lastweek) such as dropping the penalty on the Fed’s dollar

swap lines to 50bp from 100bp. Under such a scenario,

FRA-OIS spreads may be expected to widen to nearly

75bp in the near term, but narrow back to 50bp by mid-

year 2012 despite a steady widening in semi-peripheral

spreads.

Given these projections for FRA-OIS spreads, which is a

key determinant of spreads in the 2- to 10-year sector,

and based on our projections for other inputs, our

estimates for maturity matched swap spreads in various

sectors as of the end of 1Q12 and 1H12 are shown in

Exhibit 14. Broadly speaking, we see the potential for swap spreads to widen going into 1Q12; however, since

front-end spreads are generally wide to fair value

currently, this upside is greatest in the 10-year sector.

Heading into 2Q12, we would look for swap spreads to

narrow, a view that is premised upon favorable policy

actions in Europe as discussed earlier. In the 30-year

sector, we expect spreads to widen initially and then

remain largely range-bound.

Front-end spread… options?

The correlation between front-end swap spreads and

FRA-OIS spreads is a reflection of a broader

phenomenon at the front end of the yield curve.

Subsequent to the Fed announcing its commitment to low

rates until at least mid-2013, OIS rates and Treasury

yields have fallen and the OIS curve has become both

flat and sticky; as a result, the volatility in front-end swap

spreads is now almost entirely due to the volatility of the

swap rate itself. Put differently, bearish positions at the

front end are effectively the same as front-end swap

spread wideners. This is strikingly evident in Exhibit 15,

which shows that 2-, 3- and 5-year swap spreads have all

become highly correlated to front-end swap rates, with

little differences between the three sectors.

This is quite useful—options on front-end yields do exist,and are effectively options on front-end swap spreads.

Short expiry payer swaptions and/or payer swaption

spreads on front-end swap yields (such as 3Mx1Y payer

swaptions, for instance) can be used to create asymmetric

exposure to wider front-end swap spreads. As we have

noted often in recent weeks, given their limited-risk

nature, such trades can offer better risk-reward than

outright swap spread wideners or narrowers. Indeed, in

the past month, we have recommended payer swaption

spreads as proxies for front-end swap spread wideners, as

well as receiver swaption spreads in place of spreadnarrowers (see US Fixed Income Markets Weekly,

Interest Rate Derivatives, dated September 23, 2011 and

November 18, 2011, respectively).

This state of affairs is likely to persist into next year.

Going into 2012, a key part of our strategy with respect

to front-end swap spreads will be to trade them using

payer and receiver swaption spreads as proxies.

Exhibit 16: Sectors where the Fed is actively buying Treasuries are likelyto see a convergence in yields between issues of similar durationDispersion* of bond yields in the 3- to 6-year sector (where the Fed is not buying)and the 20- to 30-year sector (where the Fed is buying); bp

* Calculated as the root mean squared y ield error of bonds in that sector relative toan overall par Treasury fitted curve

Exhibit 15: Swap spreads across much of the front end have become

highly correlated to front-end swap yields with little distinction acrosssectorsMaturity-matched swap spreads in 2-, 3- and 5-year sectors versus the 3Mx1Yforward swap yield; past three months; bp

3Mx1Y forward swap yield; %

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6


3- to 6-year 20- to 30-year

20

25

30

35

40

45

0.4 0.45 0.5 0.55 0.6 0.65 0.7

2Y 5Y 3Y





November 24, 2011

Srini RamaswamyAC

(1-212) 834-4573




140

O-spreads and spread curve relative value

With Operation Twist already underway and slated tocontinue until the end of 1H12, relative value spread

switch trades will likely prove to be a profitable trading

strategy as we head into 2012. The logic underlying this

is quite simple; experience from the Fed’s Treasury

purchases under QE1 as well as QE2 suggest that the

Fed’s issue selections are based on yield differentials of

various bonds with respect to a fitted curve—i.e., yield

errors (see Treasuries). Thus, it is reasonable to expect

convergence between rich and cheap issues, measured

through metrics that mirror the Fed’s, particularly in

sectors where the Fed is actively purchasing Treasuries.

Such convergence is indeed already evident since the

commencing of Operation Twist; yield errors are

converging at the longer end of the curve, producing

little dispersion across bonds in those sectors, while a

similar fall in dispersion is not evident in the shorter end

(Exhibit 16).

Of course, yield errors are not directly tradable, but

switch trades on a maturity matched swap spread basis

are a close proxy. Such trades should prove attractive in

periods and sectors characterized by active open market

Treasury purchases by the Fed. But what relative value

metrics are suitable for sectors of the curve or periods of

time where the Fed is not actively purchasing Treasuries?

Historically, asset swap spreads (some times called a

proceeds asset swap spread) have served as a useful

metric of value. The logic behind this measure is this:

suppose that an investor puts up $100 to purchase a

Treasury bond regardless of its actual full price, finances

(or lends) the difference between the bond’s full price

and par via a swap where the investor pays the bond’s

coupon cash flows versus receiving Libor plus a spread

on the floating leg. This spread is called the proceeds

asset swap spread (or simply, asset swap spread), and is

a clean metric for comparing bonds with different

coupons and prices (see Exhibit 17 for an illustration).

Traditionally, these asset swap spreads were solved for

by requiring that cash flows from the associated swap,

discounted using swap curve zero rates, should produce

an NPV that offsets the amount being financed (or lent).That is, the swap curve served as both the reference

curve as well as the discount curve. In principle,

however, the discount curve should depend on the

funding curve for the type of assets that would be used to

collateralize the mark-to-market variations in the swap.

Since US dollar interest rate swaps are commonly

Exhibit 17: An illustration of the roceeds asset swap spread and O-

spread A schematic illustration of the proceeds asset swap spread for a given bond

Exhibit 18: Might is right—the Fed’s preferred metric displaces the market’s value metric in sectors and times where the Fed is actively buying Treasuries,but O-spreads are likely to be the best value metric otherwiseStatistics regarding the back-testing of trading strategies* using three metrics of value: yield error, asset swap spreads (based on swap curve discounting) and o-spreads (or assetswap spreads to the swap curve, calculated using OIS discounting)

* Trading strategy is as follows: every day, the cheapest and richest bonds in each sector are identified, based on each of the three value metrics. The rich bond is then sold and thecheap bond is bought, on a maturity matched swap spread switch basis. Trades are held for a month and unwound thereafter. We back-tested these strategies from the start of QE2through mid-November.

Dirty Bond

price

For bond: 8% Nov 2021

Spread: -12.9bp under Libor discounting

-17.4bp under OIS discounting

Investor Swap counterparty

Bond coupons

3M Libor + spread

Dirty price minus par

During Jul - Oper. During Jul - Oper. During Jul - Oper. During Jul - Oper. During Jul - Oper. During Jul - Oper.

Overall QE2 Sep Twist Overall QE2 Sep Twist Overall QE2 Sep Twist Overall QE2 Sep Twist Overall QE2 Sep Twist Overall QE2 Sep Twist

# trades 230 146 64 20 230 146 64 20 230 146 64 20 230 146 64 20 230 146 64 20 230 146 64 20

# wins 202 135 53 14 196 119 60 17 208 125 63 20 139 103 18 18 137 86 43 8 139 87 42 10

# losers 28 11 11 6 34 27 4 3 22 21 1 0 91 43 46 2 93 60 21 12 91 59 22 10

Hit ratio 88% 92% 83% 70% 85% 82% 94% 85% 90% 86% 98% 100% 60% 71% 28% 90% 60% 59% 67% 40% 60% 60% 66% 50%

Avg Gain 4.2 4.2 4.7 2 4 3.7 4.9 3.1 4.7 3.9 6.6 3.1 2.5 2.9 1.5 1.5 3.7 3.2 5.1 1.2 4 3.8 5.2 0.9

Avg Loss -2.4 -0.7 -4.7 -1.3 -1.2 -1.4 -0.7 -0.5 -1.1 -1.1 -0.8 N/A -1.6 -1.6 -1.6 -0.4 -3.1 -3.2 -3.6 -1.9 -3.5 -3.4 -4.2 -2.6

Overall avg 3.4 3.8 3.1 1 3.2 2.7 4.5 2.5 4.1 3.2 6.5 3.1 0.9 1.5 -0.7 1.3 0.9 0.6 2.2 -0.7 1 0.9 1.9 -0.8

7- to 15-year sector

Value metric: ASW Value metric: OSWValue metric: Yield error Value metric: Yield error Value metric: ASW Value metric: OSW

2- to 5-year sector





November 24, 2011

Srini RamaswamyAC

(1-212) 834-4573




141

collateralized by Treasuries, it would be more correct to

use OIS-curve based zero rates to discount future cash

flows (since Treasuries fund at GC rates, which arerather close to OIS rates, and since a well developed OIS

curve exists). With regulatory reform forcing almost all

swaps to be collateralized going forward, a more correct

way to define a proceeds asset swap spread would be to

still define a spread versus the Libor curve (as in the

previous case), but using the OIS curve for discounting

cash flows. We introduce new terminology to reference

this new measure of value, and will refer to it as

proceeds asset swap o-spread (or simply, o-spread )

heretofore in our research. O-spreads will be understood

as a reference to spreads on the floating leg of a swap,

where the spread has been solved for using OIS zero

rates for discounting purposes.

Which of these metrics works well in practice? To

answer this question, we back-tested trading strategies

based on three different value metrics, in two sectors of

the curve, over the past year. Every day, we identified the

richest and cheapest bonds in each of the two sectors,

based on each of three value metrics (yield error, asset

swap spread and o-spread). We then initiated a long in

the cheap bond and a short in the rich bond, but on a

maturity-matched swap spread basis (for the purpose of

lessening P/L noise due to curve movements). All trades

were held for a month and unwound thereafter. The twocurve sectors we chose are the 7- to 15-year sector (a

beneficiary of Fed purchases during QE2 as well as

Operation Twist), and the 2- to 5-year sector (which

benefited from Fed purchases during QE2, but not during

Operation Twist). Statistics regarding our results are

presented in Exhibit 18.

Three points are interesting to note from our results.

First, yield errors are the best metric for sectors/periods

where Fed purchases are a factor. Put differently, Fed

purchases distort normal value metrics, and essentially

impose the Fed’s value metric in its place. This is evident

from the fact that hit ratios were highest in both curvesectors during QE2 when using yield errors as the value

metric. More recently under Operation Twist, using yield

errors once again results in the highest hit ratios in the 7-

to 15-year sector.

Second, when Fed purchases cease to be a factor, the

resulting distortion effect on value metrics is quick to

dissipate. This is seen in the fact that during the brief 3-

month hiatus in Fed purchases in the 7- to 15-year

sectors, yield errors were a poor indicator for

convergence trades, underperforming asset swap spreads

as well as o-spreads as a value metric. This is also

evident in the fact that since the ending of Fed purchases

in the 2- to 5-year sector of the curve, yield error has

underperformed the other two value metrics.

Third, there is some evidence to suggest that o-spreads—

rather than traditional asset swap spreads—are a better

value metric when Fed purchases are not a factor. This is

seen by looking at the performance of the three differentvalue metrics in the 2- to 5-year sector, which has been

free of Fed purchase distortions since the end of QE2.

Looking ahead to 2012, we anticipate that the general

trend in these results will continue to hold—convergence

trades at the long end of the curve will continue to be

attractive when chosen using yield errors as a measure of

value; for other sectors of the curve, convergence trades

indicated by o-spreads as a value metric will likely prove

profitable. With Fed purchases likely to be a mainstay of

Exhibit 19: The yield curve has flattened significantly this year in all

sectorsYear-to-date statistics regarding the slope of the curve in various sectors; bp

* As of 11/17/2011 COB

Exhibit 20: Investors reaching for carry will only find it further out on thecurve

3-month carry and slide in various sectors of the yield curve; bp

* As of 11/17/2011 COB

Current Start Average High Low YTD chg

1s/2s 4.6 35.8 25.5 56.4 0.4 -31.2

2s/3s 9.9 48.3 36.1 59.8 9.9 -38.4

3s/5s 42.7 90.8 76.8 98.0 42.7 -48.1

5s/7s 42.6 65.5 59.8 68.2 40.8 -22.9

7s/10s 40.1 57.0 54.7 63.2 36.9 -17.0

10s/30s 55.1 76.7 77.0 93.5 51.3 -21.7

2011 YTD statistics

0

2

4

6

8

10

12

14

ED2 1Y ED6 2Y ED10 3Y 5Y 7Y 10Y 30Y





November 24, 2011

Srini RamaswamyAC

(1-212) 834-4573




142

markets in 2012, such relative value switch trades will be

a key part of our market strategy in the year ahead.

Swap yield curve

The yield curve has been on the move in 2011, flattening

significantly in all sectors of the curve (Exhibit 19). The

flattening mostly occurred in 3Q, leading into and after

the August FOMC meeting, as a number of factorsconspired to flatten the curve. At the very front end of

the curve, the biggest driver has been the Fed’s

conditional commitment, announced at the August

FOMC meeting, to hold the funds rate near current levels

until mid-2013. Although falling short of a firm

commitment, the Fed’s indication that the current

economic outlook would likely be consistent with a low

funds rate for at least two years had a significant

impact—front-end term premium collapsed, OIS rates

fell, the OIS curve flattened sharply, and yield levels

declined sharply across much of the front end of the

curve, out to the 5-year sector. The very front end of theswaps curve has been additionally biased flatter by the

widening in Libor/OIS spreads. Further out the curve, the

key driver has been the Fed’s Operation Twist.

Much of this backdrop is likely to stay unchanged as we

head into next year. Operation Twist will continue to

dominate rates markets through the end of 1H12, and the

Fed’s mid-2013 conditional commitment is likely to

remain valid for quite some time, which should keep the

very front end of the curve very flat for the foreseeable

future. In addition, the overhang from Europe will likely

persist, weighing on funding spreads and thus the front

end of the swap curve.

The implications for yield curve trading strategies in

2012 are three-fold. First, investors will need to reach

further along the curve for carry, as the flatness of the

very front end has erased carry in the front and red

Eurodollar sectors (Exhibit 20). Second, the very frontend of the swap curve—1-year swaps and front

Eurodollars—will likely be the least attractive sector to

earn carry, since longs in these sectors are vulnerable to a

worsening of the European crisis. Carry in this sector is

more optical than real, as spot Libor rates continue to

rise, making forwards more likely to be realized than

spot. Ironically, wider FRA-OIS spreads can lead to

better carry/slide optically, making front-end longs look

more attractive, even as rising spot Libor rates means

that carry in this sector is merely a mirage.

Empirical evidence supports this view. Exhibit 21 shows

that the usefulness of carry as an indicator (measured asthe beta between the 1-month P/L on a cross section of

Eurodollar sector curve and butterfly trades and the ex-

ante carry/slide on those trades) has declined as spot

Libor has risen; in other words, slide on the Eurodollar

curve ceases to predict future spot yields when funding

pressures are at work in pressuring Libor higher.

Thus, going into 1H12, we will not look to earn carry at

the front end of the Libor swap curve (unless of course if

policy actions were to cause a more sanguine outlook on

Exhibit 22: Yield curves are now highly directional with yield levels

thanks to ultra low—and sticky—front-end yieldsRolling 3-month beta between weekly changes in various yield curves and weeklychanges in the corresponding longer end y ield;

Exhibit 21: At the front end of the swap curve, carry is now a mirage—

rising spot Libor rates at the front end has decreased the usefulnessof carry as a measure of the attractiveness of Eurodollar trades1-week moving average of carry-beta* versus spot 3-month Libor;

%, inverted axis

* Beta between the 1-month P/L on a c ross section of Eurodollar sector curve andbutterflies versus the 1-month ago slide on those trades

-0.2

0.0

0.2

0.4

0.6

0.8

May 11 Jul 11 Aug 11 Oct 11

2s/5s

5s/10s

0.28

0.30

0.32

0.34

0.36

0.38

0.40

0.42-1

0

1

2

3

09 Aug 19 Aug 29 Aug 08 Sep 18 Sep 28 Sep 08 Oct

3-month Libor (inverted)

Carry beta





November 24, 2011

Srini RamaswamyAC

(1-212) 834-4573




143

the European crisis at some point). Indeed, as we noted

in an earlier section, rather than owning the front end of

the Eurodollar curve for the sake of carry, being short thesector via short-expiry payer swaptions offers the most

asymmetric way of positioning for deterioration in

Europe.

Third, the yield curve is likely to exhibit a very high

degree of directional exposure to yield levels, an artifact

of ultra-low yields in the less-than-5-year maturity

sector. This is seen in the elevated positive betas between

various yield curves and yield levels (Exhibit 22).

Hedge curve steepeners with Eurodollars

Based on these observations, three types of tradingthemes are likely to prove attractive in 2012. First,

forward curve steepeners (or spot curve steepeners with

the front end being anchored in sectors that offer the best

carry) hedged with shorts in the Eurodollar sector are

likely to be attractive ways to earn carry while also

protecting against exogenous risk from Europe. To

identify the best trades, we identified the potential reward

and risk in numerous such combinations. We estimated

the risk weight on the Eurodollar hedge as the 1-year

average of the rolling 3-month beta between weekly

changes in each curve and the selected Eurodollar yield.

Based on this hedge ratio, we then estimate the potential

gain as the 3-month carry/slide, plus 50% of the potentialupside, which in turn is defined as the average level of

the weighted yield spread since the start of 2010 minus

its current level. In order to properly assess risk, we

focus on a downside-vol measure (in addition to a more

typical standard deviation based volatility estimate).

Specifically, we estimate the downside risk by assuming

that semi-peripheral spreads (an average of Spain, France

and Italian sovereign CDS spreads) widen by 100bp, and

estimate the likely move in the weighted spread, and we

additionally assume a 1-standard error downward move

from this relationship. Thus, our downside risk measure

reflects an attempt to capture the effect of deteriorating

conditions in Europe. Finally, we calculate an efficiency

ratio, which is the ratio of potential gain to the downside

risk.

Exhibit 23 presents a visual snapshot of results, and

charts the potential gain versus the magnitude of the

downside risk, while Exhibit 24 presents more detailed

statistics regarding numerous such trades, ranked

according to their efficiency ratios. As can be seen,

10s/30s and 5s/30s steepeners, hedged with shorts in red

Exhibit 24: Detailed statistics regarding the potential reward anddownside risk in yield curve steepeners hedged with shorts in red

EurodollarsBeta of various yield curves against Eurodollar yields*, the current weighted yieldspread corresponding to each trade**, 3-month carry and slide, potential upside***, 3-month standard deviation of weekly changes, downside risk**** and efficiencyratio***** for various yield curve s teepener trades hedged with red Eurodollar shorts.

* 1-year average of 3-month beta of weekly changes in the curve versus weeklychange in ED yield.** Current value as of 11/14/2011. *** See footnote in previous exhibit. **** Seefootnote in previous exhibit.***** Efficiency ratio defined as the ratio of 3-month carry plus 50% of upside todownside volatility.

Exhibit 23: Steepeners such as 10s/30s and 5s/30s, hedged with shorts

in red Eurodollars, are attractive ways to earn carry and position for yield curve normalization, while hedging European exogenous risksPotential gain* versus the magnitude of downside risk** for various curve steepenershedged with short Eurodollars; bp

Downside risk; bp* Upside defined as average of the weighted yield spread corresponding to eachtrade since 1/1/2010, minus its current value as of 11/14/2011. Potential gain definedas carry/slide plus one half of the upside.** Downside risk estimated as the expected decline in the weighted yield spread for a100bp rise in semi-peripheral spreads, plus a 1-standard error downward move inaddition. Beta with respect to semi-peripheral spreads estimated based on aregression of monthly changes since 1/1/2010.

Curve ED Beta

Spread

(%)

Carry

(bp)

Upside

(bp)

Vol

(bp)

Downside

vol (bp)

Efficiency

ratio

3M fwd 10s/30s 6th -0.24 0.73 4.4 26.5 11.0 20.8 0.85

1Y fwd 10s/30s 6th -0.25 0.61 4.1 22.2 9.2 18.4 0.83

6M fwd 10s/30s 6th -0.24 0.69 4.0 25.0 10.3 20.0 0.82

2Y fwd 10s/30s 6th -0.24 0.42 4.2 17.3 7.5 15.9 0.81

2Y fwd 5s/30s 6th -0.47 1.16 9.0 32.3 16.7 32.1 0.78

3M fwd 5s/30s 6th -0.37 1.71 7.7 51.6 25.5 43.2 0.78

1Y fwd 5s/30s 6th -0.46 1.53 8.0 42.1 21.5 38.0 0.77

3Y fwd 5s/30s 6th -0.39 0.75 7.5 25.8 12.9 26.9 0.76

3Y fwd 5s/10s 6th -0.20 0.53 4.0 12.1 7.9 14.4 0.702Y fwd 5s/10s 6th -0.23 0.74 4.8 15.0 10.4 17.6 0.69

3Y fwd 3s/10s 6th -0.32 0.97 7.9 18.2 13.8 25.5 0.67

3Y fwd 2s/10s 6th -0.40 1.30 9.6 22.9 17.8 32.5 0.65

2Y fwd 3s/10s 6th -0.36 1.34 7.4 24.4 18.4 30.6 0.64

1Y fwd 5s/10s 6th -0.21 0.92 3.9 19.9 13.3 21.6 0.64

2Y fwd 5s/10s 5th -0.26 0.76 5.0 11.2 10.0 16.8 0.63

2Y fwd 2s/10s 6th -0.42 1.71 7.2 32.2 24.0 37.6 0.62

5

10

15

20

25

30

35

10 20 30 40 50

10s/30s

5s/30s

2s/10s and 3s/10s

5s/10s

1s/5s, 2s/5s and 3s/7s





November 24, 2011

Srini RamaswamyAC

(1-212) 834-4573




144

Eurodollars are among the best such trades; variations

involving 3- and 2-year forward 2s/10s steepeners (again

hedged with red Eurodollars) are also attractive, but

slightly less so. As we head towards 2012, we would

look to a framework such as this to identify the best ways

to position for a normalizing in yield curves.

Take advantage of non-linearities in the yield curve

due to sticky front-end yields

Our second yield curve trading theme takes advantage of

yield curve nonlinearity in low yield regimes. One

consequence of yields being very close to the zero bound

at the front end of the curve is that yield curve butterflies

will likely exhibit nonlinear behavior with respect to

yield levels. This is strikingly evident by looking at—for

instance—a curve neutral 2s/10s/30s butterfly (25:75

weights on the wings) versus 2-year yields (Exhibit 25).

This suggests that curve-neutral belly-richening

butterflies locally hedged for exposure to front-end yields

should experience empirically convex upside, in either a

front-end rally or a sell-off. Moreover, since the

underlying reason for this nonlinearity—the stickiness of

yields when near the zero boundary—is also affecting

swaption implied volatility skews, we may use

information from swaption skews to infer which

butterflies are likely to exhibit the most nonlinear

behavior (see Implied Weights in Yield Curve Relative

Value Trades, J.P. Morgan Research note, May 25, 2011,

for a detailed discussion of our framework). Exhibit 26

outlines several butterflies that are likely to exhibit the

most asymmetric behavior with respect to rates under

different rate scenarios, thanks to such nonlinearities.

Some of these trades have positive (or very slightly

negative) carry, which also makes them attractive ways

of positioning for a large, unanticipated, move higher infront-end yields.

Use payer swaptions to position for an eventual

normalization in rates

Third, with yields extremely low at the front end, and

with steep implied volatility skews, payer swaption

spreads (1:1 weighted) now offer attractive reward-to-

premium ratios (Exhibit 27), and are attractive as

limited-risk ways of positioning for an eventual

normalization in rates towards higher levels, should that

occur in 2Q12 as our rates strategists expect. Although

we expect yield levels to fall in the early part of 2012, we

do expect such a move to higher rates in the middle of

next year (see Treasuries). Given the highly uncertain

nature of policy developments in Europe, and the

resulting risks in outright short-duration positions, we

view 1:1 weighted limited-risk payer swaption spreads as

the preferred way to position for the normalization in

yield levels that we expect as we head towards the

middle of next year. Moreover, with yields expected to

drift initially lower in 1Q12, we would expect to find

even better entry levels for such trades in the early part of

next year, and would look to initiate payer swaption

Exhibit 25: Yield curve butterflies begin to exhibit non-linear behavior as front end yields become sticky near the zero bound

A curve neutral (25:75 weighted) 2s/10s/30s swap yield butterfly spread, versus 2-year swap yield; %

2-year swap yield; %

Exhibit 26: The nonlinearity implied by the skew is pronounced for various benchmark butterfliesProjected evolution of left- and right-leg weights for several butterflies under various left leg swap yield shifts* and projected upside** in each scenario and carry (bp).

* Our methodology is as follows. For each left leg rate shift, we calculate the right leg rate shift based on the current ratio of ATM vols. We then combine the left and right rateshifts with the current implied weights to calculate a belly rate shift for that scenario. Third, we use the current implied vol skews to estimate the change in then-ATM implied volsfor those rate shifts in each leg. Last, armed with estimates for then-ATM implieds in all three legs, we calculate new implied left and right weights.** The projected upside is estimated as sum over left and right leg values of half the change in weight from the un-shifted scenario times the corresponding shift in rate.

-0.3

-0.2

-0.1

0.0

0.1

0.2

0.4 0.5 0.6 0.7 0.8 0.9 1 1.1

Projected Projected Projected

upside from upside from upside from

Butterfly -25bp 0bp +25bp +75bp -25bp 0bp +25bp +75bp -25bp scenario +25bp scenario +75bp scenario 3M Carry

3s/7s/15s 1.32 1.03 0.85 0.69 0.34 0.36 0.34 0.30 3.3 2.1 10.4 1.22s/7s/15s 0.86 0.60 0.46 0.35 0.62 0.64 0.62 0.54 2.7 1.3 4.8 -3.7

3s/10s/30s 0.78 0.60 0.48 0.37 0.63 0.66 0.63 0.55 1.5 0.8 3.1 -0.4

5s/7s/15s 1.07 1.02 0.99 0.96 0.15 0.15 0.15 0.13 0.6 0.3 1.7 2.6

Left weight given a shift in left rate of Right weight given a shift in left rate of





November 24, 2011

Srini RamaswamyAC

(1-212) 834-4573




145

spreads on the back of declines in yield levels.

Conditional trades … with a twist

The fourth yield curve trading theme that we will attempt

to take advantage of in 1H12 involves exploiting the high

degree of directionality between yield curves and yield

levels. With yield curves anchored at the front end

expected to be highly directional with yield levels,

conditional curve trades (i.e., conditional flatteners in a

rally and conditional steepeners in a sell-off) will likely be an ongoing trading theme in 1H12.

Options markets are pricing in this directional nature of

yield curves to some extent, but not fully. Although

swaption implied volatility on short tails (such as 2s) is

well below implied volatility on longer tails (such as

10s), the implied beta between the curve and long end

yields is well below empirical betas (Exhibit 28). This

allows for the construction of premium-neutral weighted

conditional curve trades at the current time; however,

with yield curves expected to remain highly directional

for quite some time, we would not expect the options

markets to persistently allow for attractive entry levels on

such trades.

How else can one exploit yield curve directionality oncethe swaptions markets have priced in the relative vol

differentials fairly? To answer this question, we explore a

variant of the “usual” conditional curve trades. Our

variant is based on two observations: first, the 10s/30s

curve is highly correlated (in magnitude, but with a

negative beta) to front-end yields, and second, implied

volatility on the 10s/30s curve in the YCSO market has

tended to trade rich. This suggests that YCSOs on the

Exhibit 28: Yield curves have become highly directional with rates, and

options markets are pricing this in to some extent but not yet fullyImplied* and empirical** betas between the 3-month forward 2s/10s curve and3Mx10Y forward swap yield;

*Implied beta calculated as 1-(3Mx2Y / 3Mx10Y) implied bp vol ratio. ** Empiricalbeta calculated as the 3-month beta between weekly changes in the 3M fwd 2s/10scurve and 3Mx10Y forward swap yield.

Exhibit 27: Payer swap spreads (1:1 weighted) are attractive as limited-

risk ways of positioning for an upward drift in rates by mid-year Statistics regarding 1:1 weighted payer swaption spreads (buy the ATMF strike, sellthe A+50 strike) with a 6/30/2012 expiry on various underlying tails

* As of COB 11/17/2011

Exhibit 29: Conditional curve trades with a twist—since the 10s/30s curve is well correlated to front-end yields, and since YCSOs on 10s/30s are rich,2-year tail swaptions can be replaced with options on the 10s/30s curve to create “synthetic” conditional curve trades at better entry levelsImplied volatilities (bp/day), notionals ($mn) and premium($) for conditional curve trades and corresponding values for trades in which the left leg option is replaced by a YCSOcurve cap.

0.4

0.5

0.6

0.7

0.8

0.9

Aug 11 Sep 11 Oct 11 Oct 11 Nov 11

Empirical beta Implied beta

Slide Premium Max

Tail Spot Forward bp bp yield gain Ratio

1Y 0.752 0.846 9.3 12.1 37.9 3.1

18M 0.780 0.859 7.9 12.1 37.9 3.1

2Y 0.799 0.897 9.8 12.2 37.8 3.1

3Y 0.897 1.080 18.3 13.3 36.7 2.8

5Y 1.324 1.570 24.6 16.1 33.9 2.1

7Y 1.750 1.963 21.3 17.6 32.4 1.8

10Y 2.151 2.317 16.6 18.6 31.4 1.7

15Y 2.503 2.615 11.2 19.2 30.8 1.6

30Y 2.701 2.764 6.3 20.1 29.9 1.5

Yields, %

Net prem

Net Beta wrt on modified

Expiry Left Right Type Left Right Left Right Premium Expiry YCSO Imp. Vol Left leg Notional Type Premium trade

3m 2Y 5Y Rec 3.46 5.26 -100 40.8 111,655$ 3m 10s/30s 4.6 -0.41 -485.9 Cap 14.4 (373,100)$

3m 2Y 7Y Rec 3.46 6.61 -100 29.8 195,317$ 3m 10s/30s 4.6 -0.41 -485.9 Cap 14.4 (289,438)$

3m 2Y 10Y Rec 3.46 7.86 -100 21.7 272,424$ 3m 10s/30s 4.6 -0.41 -485.9 Cap 14.4 (212,331)$

3m 2Y 30Y Rec 3.46 8.96 -100 9.6 339,559$ 3m 10s/30s 4.6 -0.41 -485.9 Cap 14.4 (145,196)$

6m 2Y 5Y Rec 3.36 5.28 -100 40.7 167,224$ 6m 10s/30s 4.0 -0.34 -582.5 Cap 17.7 (569,328)$

6m 2Y 7Y Rec 3.36 6.41 -100 29.7 265,495$ 6m 10s/30s 4.0 -0.34 -582.5 Cap 17.7 (471,057)$

6m 2Y 10Y Rec 3.36 7.46 -100 21.6 355,936$ 6m 10s/30s 4.0 -0.34 -582.5 Cap 17.7 (380,616)$

6m 2Y 30Y Rec 3.36 8.26 -100 9.6 423,282$ 6m 10s/30s 4.0 -0.34 -582.5 Cap 17.7 (313,270)$

Trade Notionals

Desired Trade

Implied vol

Replacement for left leg





November 24, 2011

Srini RamaswamyAC

(1-212) 834-4573




146

10s/30s curve may be used in place of swaptions on 2-

year tails; specifically, receiver swaptions on 2-year tails

may be replicated via similar expiry caps on the 10s/30s

curve, and payer swaptions on 2-year tails may be

replaced with floors on the 10s/30s curve.

Exhibit 29 presents details for a sample set of

conditional curve trades, with the 2-year leg being

replaced by options on the 10s/30s curve; for instance,

the short receiver swaption position in a conditional

bullish flattener would be replaced with short position in

1-look caps on the 10s/30s curve with the same expiry.

The short payer swaption leg in a conditional bearish

steepener would similarly be replaced with a short floor

position on the 10s/30s curve. As can be seen in the

table, such modified “synthetic” conditional curve trades

are interesting because they can mimic an unweighted

conditional curve trade and also take in premium

currently, which cannot be done right now in the

swaptions market. A visual confirmation that such a trade

would indeed track the yield curve (which, after all, isthe ultimate intent of a conditional curve trade) is seen in

Exhibit 30. The yield spread corresponding to a 6-month

forward 2s/10s curve trade, where the 2-year leg has

been replaced with a beta weighted position in the

10s/30s curve, is shown in this exhibit; as can be seen,

this spread does indeed closely track the 2s/10s curve,

which was the original desired trade in this example.

Exhibit 31: A tale of two different vol markets in 2011—being short

volatility was profitable until July, and long volatility positions haveprevailed afterwards1-month Gamma and Vega P/L* from long straddles in 3Mx10Y swaptions ; bp of notional

* Returns calculated using the J.P. Morgan Volatility Index, which assumes dailydelta rebalancing and zero transaction costs. Options are re-struck at the start of each month.** November-to-date P/L.

Exhibit 32: Expect liquidity to remain poor: risk appetite continues todecline as the European crisis spreads into semi-peripheralcountries …Reported average quarterly VAR for the 9 largest investment banks* versus 3-monthmoving average of semi peripheral spreads**; $mn

* Average VAR reported by JPM, GS, MS, BAC, C, UBS, CS, Soc Gen and DB.3Q11 VAR is an estimate, projected from the results reported to date (excluding C).** Average of 5-year CDS spreads on France, Italy and Spain.

Exhibit 30: “Synthetic” curve trades—where the front end leg has been

replaced with a beta-weighted position in the 10s/30s curve—are likelyto track the desired original trade6-month forward 2s/10s swap curve, versus the yield spread corresponding toreplacing the 2-year leg with a beta weighted 6-month forward 10s/30s curve position

*Calculated as 6Mx10Y minus (6Mx10s/30s curve divided by -0.34)

-60

-40

-20

0

20

40

60

J a n

F e

b

M a r

A p r

M a y

J u n

J u

l

A u g

S e p

O c

t

N o v

* *

G amma PL Vega PL

50

100

150

200

250

30090

100

110

120

130

140

Dec 09 Apr 10 Jul 10 Oct 10 Feb 11 May 11 Aug 11

Semi-peripheralspreads

VAR

3.63.8

4.0

4.2

4.4

4.6

4.8

5.0

1.41.5

1.6

1.7

1.8

1.9

2.0

2.1


Yield spread for synthetic trade; %6M fwd 2s/10s; %





November 24, 2011

Srini RamaswamyAC

(1-212) 834-4573




147

Options

Short volatility strategies have thus far managed to

remain profitable in 2011 year-to-date, but it has been a

clear case of two different market environments. From

the start of the year through the end of July, selling

gamma proved considerably profitable overall. Incontrast, since August, the opposite has been the case,

and long gamma strategies have steadily gained ground

(Exhibit 31).

While the outperformance of long gamma strategies in

August is a typical seasonal phenomenon related to

declines in market depth, heightened macro uncertainty

stemming from Europe has thus far forestalled a recovery

in market depth, helping preserve conditions that are

favorable to long gamma positions.

Looking ahead towards the remainder of this year and

1Q12, our best guess is that risk appetite will remain low,and liquidity conditions will remain poor. Declining risk

appetite is perhaps most evident in the daily VARs

reported by major investment banks in their quarterly

financials. As seen in Exhibit 32, VARs remain on the

downtrend, reflecting growing caution as the European

crisis has spread into semi-peripheral countries.

In the near term as we head into year end, seasonals are

unfavorable as well and could make for worsening risk

appetite and market depth. As seen in Exhibit 33, market

depth has exhibited a tendency to decline to annual lows

in December. This should prove unhelpful to marketliquidity and could support elevated realized volatility.

One way to see this is to estimate the width of the market

at which a given fixed size can be transacted by a price-

taker. For instance, an investor needing to (say) buy

$150mn 10s might simply lift the highest offer in a

period when market depth is high, but might need to lift

(say) the top three or four offers in periods of lower

market depth, assuming a static order book. Examining

the implied width of the market for a given fixed size (we

Exhibit 35: A model for seasonally-adjusted market depthModel* for market depth in the 10-year sector, and forecast for 2012

* Model based on regression of seasonally-adjusted market depth against semi-peripheral spreads and cross asset correlation, which we proxy using the rolling 3-month correlation between weekly changes in the S&P and 30-year swap yields.Semi-peripheral spreads are an average of 5-year CDS spreads on France, Spainand Italy, and are used to capture broader risk aversion. Regression estimated over two years of history.

Exhibit 33: Intra-year seasonals are likely to be unhelpful to market

liquidity, as evidenced by the tendency of market depth to reach itslows in December Percentage deviation of market depth in the 10-year sector from average over the pastfive years

* Market depth is calculated as the average size of the top three bids and offers, in$mn, for the on-the-run 10-year Treasury note, averaged between 8:30 a.m. and10:30 a.m. daily.

Exhibit 34: Larger market moves become likely in periods of low

market depth, as a price-taker will need go deeper down the order stack to trade a given amountRegression of market width* spread against market depth** in the 10-year sector; bp

Market depth; $mn* Market width calculated as the spread between mid-market and the level at whicha fixed size ( $150mn) of the on-the-run 10-year Treasury note can be traded,assuming a static order book. Bid-side and offer-side widths are averaged here.** see footnote in previous exhibit

Factor Coefficient T-stat 1Q12 2Q12 4Q12


Semi-peripheral spreads -0.47 -27.8 70 125 70

Cross-asset correlation -76.1 -19.2 -0.1 -0.3 -0.3

R-squared

Std error

Impact on market depth -26 -36 -10

Projected market depth 70 60 86

77%

26.4

Chg in driver, from current

-60%

-40%

-20%

0%

20%

40%

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

0.5

1.0

1.5

2.0

2.5

3.0

50 100 150 200 250





November 24, 2011

Srini RamaswamyAC

(1-212) 834-4573




148

use $150mn) is insightful. As Exhibit 34 shows, this

implied market width grows nonlinearly with depth,

highlighting the fact that the market is much more prone

to violent moves in periods of low market depth—as

even small-sized trades produce sizeable moves.

Looking past year end, the key question then is whether

market depth will begin to improve with a renewal of

risk appetite next year. While this has historically been

the case seasonally, there are good reasons not to expect

such a pick-up after year-end 2011. In order to better

understand the drivers of market depth, which has

become a key determinant of realized volatility, we

analyze the drivers of market depth on a seasonally-

adjusted basis (given significant intra-year variation dueto seasonals). As seen in Exhibit 35, market depth

(adjusted for intra-year seasonals) has been well

explained by two factors over the past two years. The

first factor is a measure of cross asset correlations,

proxied here by the rolling equity-rate correlation. All

else equal, rising cross-asset correlations effectively

increase the risk in an investor’s portfolio, in turn leading

to cuts in notional amounts of risk-taking in individual

markets. As short-term cross-asset correlations have risen

in markets, measured VARs in portfolios have likely

risen as well, prompting risk reduction; market depth

would likely suffer in such periods.

A second factor that has also biased market depth lower

has been the steady contagion from the European crisis,

which we measure by using an average of semi-

peripheral spreads. Finally, although not formally in our

model, we note that some of this decline in market depth

is also structural, reflecting stricter capital and leverage

requirements on banks.

Looking ahead, cross-asset correlations are likely to

remain elevated or even increase further in the near term

going into year end, but decouple thereafter in 2012.

Semi-peripheral spreads, however, are likely biased

wider, in the view of our European rates strategists.Incorporating their views into the framework of our

model for market depth suggests that the typical new-

year rebound in market depth may be unlikely; we

project that market depth is likely to average similar or

lower levels for much of 1H12, as also seen in

Exhibit 35.

What does this mean for delivered volatility? To answer

this, we turn to a model for realized volatility we have

used in recent months. This model explains realized

volatility in the 10-year sector using yield levels, the

curve, market depth and an index of hedge fund leverage

as independent variables. It is presented in Exhibit 36

together with the details of our projections for 2012. As

can be seen, we generally project elevated realized

volatility over much of 2012 (i.e., levels comparable to

current implieds), and would thus look for an

Exhibit 36: Look for realized volatility to decline only modestlyafter year end, and remain elevated in 1H12Statistics regarding a model* for 1-month ahead realized volatility on 3Mx10Yforward swap yields, and forecast for 2012

* Model based on 1-year regress ion of 1-month 3Mx10Y realized volatilityagainst ex-ante levels of the 2s/10s swap yield curve, market depth in the 10-

year sector**, 3M forward 10-year swap yields and the hedge fund leverageindex***.** Market depth is calculated as the average size of the top three bids andoffers, in $mn, for the on-the-run 10-year Treasury notes, respectively,averaged between 8:30 a.m. and 10:30 a.m. daily.*** Leverage defined as the sum of the absolute value of each of the six betacoefficients in the following multiple regression: daily excess returns on the IQHedge Global Macro beta index regressed against daily excess returns on 1)J.P. Morgan global bond index ex-US, 2) J.P. Morgan US 7-10Y bond index, 3)J.P. Morgan US 7-10-year minus 1-3-year bond index, 4) MSCI G7 index, 5)USD (J.P. Morgan USD cash index minus J.P. Morgan global cash index), and6) GS Commodity index. Higher values imply more leverage.

Exhibit 37: Mortgage market convexity needs are near historiclowsConvexity exposure in Fannie Mae and Freddie Mac’s portfolios*;$bn 10s per 10bp

* Estimated from their reported interest rate sensitivities.

Factor Coefficient T-stat Current 1Q12 2Q12 4Q12

Intercept -0.6 -0.6

3M fwd 2s/10s curve (%) 4.0 5.0 1.49 -0.3 0.4 0.5

Market depth ($mn) -0.017 -11.1 98 -26 -36 -10

3Mx10Y yield (%) -1.5 -2.9 2.29 -0.35 0.45 0.65

H.F. leverage index 5.85 15.6 1.4 -0.1 -0.4 -0.5

R-squared

Std error

Impact on del. Vol (bp/day) -0.8 -0.8 -1.7

Projected realized vol 8.4 7.6 7.6 6.7

72%

0.97

Chg in driver, from current

-10

-9

-8

-7

-6

-5

-4

-3

-2

2009 2010 2011





November 24, 2011

Srini RamaswamyAC

(1-212) 834-4573




149

environment that is largely favorable to long gamma

positions. In the early part of the year, the main driver of

this is likely to be heightened risk aversion and low

market depth that combine to create illiquid conditions

that support high realized volatility. Later in the year,

however, the driver is likely to be higher rates—we

expect US yields to decouple from the European crisis

later in 2012, and rising yields will likely emerge as themain driver sustaining elevated realized volatility.

Long-dated volatility

As we contemplate the outlook for longer-dated swaption

implied volatility in the year ahead, it is useful to revisit

the key drivers of implieds in this sector in recent years.

In addition to the overall level of volatility, which would

of course drive implieds all across the vol surface, there

have been two important drivers. The first is mortgage

market convexity, which has of course steadily declined

in magnitude, resulting in declining demand for

intermediate expiry swaption volatility. We use theimputed convexity of the two GSEs’ portfolios net of

hedges (which they report monthly) as our preferred

variable to capture mortgage market convexity needs in

the aggregate (Exhibit 37).

A second factor that has been relevant in recent years is

long-term inflation expectations. As we note in the Cross

Sector Overview, there are not many attractive

alternatives to position for a sharp rise in long-term

Exhibit 38: Longer-dated swaption implied volatility has been wellexplained recently by three factors—the level of volatility overall,mortgage market convexity, and inflation expectations3Yx10Y implied volatility versus in-sample model fair value *

* Model based on 3-year regression of 3Yx10Y implied vol against GSEportfolio convexity (see Exhibit 37), 10-year inflation swap yields (%) and3Mx10Y implied vol (bp/day).

Exhibit 40: Implied correlation has plunged in anticipation of Operation Twist …

Implied correlation between 6M forward swap yields from YCSO market. Averagevalue from 8/18/2010 to 8/18/2011, value as of 8/18/2011 and drop during the period8/18/2011-11/18/2011; %

.

Exhibit 39: Look for 3Yx10Y swaption volatility to decline to 6bp/day in1Q12

Statistics regarding a model* for 3Yx10Y implied volatility and forecast for 2012

* See Exhibit 38

Exhibit 41: … and that, coupled with an inverted volatility curve, hasled to cheap forward volatilityBeta weighted differential between forward and spot volatility, current level and 2-year Z-score

* Beta weighted.

5

6

7

8

9

10

Nov 08 Jun 09 Dec 09 Jul 10 Jan 11 Aug 11

3Yx10Y implied vol

Model

-100%

-50%

0%

50%

100%

6 M x

5 Y

6 M x

1 0 Y

6 M x

3 0 Y

1 Y x

5 Y

1 Y x

1 0 Y

1 Y x

3 0 Y

2 Y x

5 Y

2 Y x

1 0 Y

2 Y x

3 0 Y

Avg 8/18/2011 Chg 8/18/2011-11/18/2011

Factor Coefficient T-stat Current 1Q12 2Q12 4Q12

Intercept 1.4 9.7

GSE portfolio cvx ($bn 10s/10bp) -0.17 -17.1 -2.6 -2.1 -3.3 -3.6

10Y inflation swap rate; % 1.0 21.3 2.26 1.81 1.96 1.96

3Mx10Y implied vol; bp/day 0.30 37.9 7.65 7.6 7.6 6.7

R-squared

Std error

Actual implied vol; bp/day 6.9

Projected implied vol, bp/day 6.4 5.9 6.2 6.0

80%

0.335

Projections

2-year average

Forward Spot (bp/day) 2-year Z-score

6Mx5Yx5Y 5Yx5Y 1.02 -0.6 -0.3 -3.46Mx5Yx10Y 5Yx10Y 1.00 -0.4 -0.2 -3.11Yx5Yx5Y 5Yx5Y 0.93 -0.2 0.2 -2.9

1Yx5Yx5Y 5Yx5Y 0.93 -0.2 0.2 -2.91Yx5Yx30Y 5Yx30Y 1.03 -0.9 -0.5 -2.76Mx1Yx10Y 1Yx10Y 1.12 -1.4 -0.9 -2.21Yx1Yx10Y 1Yx10Y 1.10 -1.4 -0.7 -2.11Yx1Yx10Y 1Yx10Y 1.10 -1.4 -0.7 -2.16Mx6Mx10Y 6Mx10Y 0.96 -0.7 0.4 -1.91Yx6Mx10Y 6Mx10Y 0.90 -0.2 0.7 -1.91Yx6Mx10Y 6Mx10Y 0.90 -0.2 0.7 -1.9

Richness/cheapness of forward vol

Forward minus

spot vol spread*

(bp/day)

2-year beta

of forward

vs spot vol





November 24, 2011

Srini RamaswamyAC

(1-212) 834-4573




150

inflation expectations, other than outright bearish trades

or payer swaptions. Indeed, recent years have seen

considerable demand for trades designed to protect

against a rise in long-term inflation, via payer swaptions/

payer swaption spreads (or CMS caps or cap spreads),

although this demand has ebbed and flowed with market

sentiment in this regard.

These two factors, together with 3Mx10Y swaption

implied volatility (as a proxy for the overall level of

volatility) do indeed successfully explain most of the

variation in implied volatility levels in the intermediate

expiry sector, as seen in Exhibit 38. We also project the

likely evolution of the drivers in Exhibit 39. In

particular, we project the convexity of the GSEs’

portfolios based on the most recent known data point, the

recent relationship of this variable with yield levels, and

our projections for yield levels in 2012; we project 10-year inflation swap yields based on our projections for

TIPS breakevens (see TIPS); and we draw upon our

forecast for short expiry swaption implied volatility

detailed earlier.

Putting it all together, our projections point to declines in

3Yx10Y implied volatility to 6bp/day by the end of

1Q12, with range-bound levels thereafter.

Bermudan Swaptions

As investors look ahead to devising their volatility

strategies in 2012, several underlying principles will

likely be important.

First, given extremely low rates across much of the

curve, implied volatility across much of the volatility

surface has become highly directional with rate levels.

Exhibit 42: A Bermudan swaption may be thought of as spanning

numerous European swaptions, corresponding to each of thevarious possible exercise dates …Graphic representation of a 1Yx5Y Bermudan swaption with semi-annual exercisepoints, and its spanned European swaptions.

Exhibit 43: … also, it can be thought of as spanning numerous

Canary swaptionsGraphic representation of a 1Yx5Y Bermudan with annual exercise points, and its 10spanned Canary swaptions

Exhibit 44: The BC basis is consistently narrower than the BE basisunder different yield curve scenarios1Yx5Y ATM receiver Bermudan/Canary basis and Bermudan/European basis for the CTD divided by the Bermudan vega* for various parallel shifts of the yieldcurve**; %

Shift in yields; %* Bermudan vega in bp of notional per bp shift in daily vol** COB 11/10/2011

t=0 t=1Y t=6Y

─ ─ ─►

t=1Y (first exercise) ─ ─ ─► 5Y swap

t=1.5Y ─ ─ ─ ─ ─►

─ ─ 2Y expiry ─► 4Y swap

─ ─ ─ ─ ─ ─ ─ ─ ─►

─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─►t=3.5Y

─ ─ ─ ─ 3.5Y expiry ─ ─ ─ ─ ─► 2.5Y swap

─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─►

─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─►t=5Y

─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─►

─ ─ ─ ─ ─ ─ 5.5Y expiry (last exercise) ─ ─ ─ ─ ─► 6M

Spanned European swaptions

1Yx5Y Bermudan swaption

t=0 t=1Y t=6Y

─ ─ ─►

t=1Y t=2Y ─ ─ ─► 1Y(2Y)x5Y Canary

t=3Y ─ ─ ─► 1Y(3Y)x5Y Canary

t=4Y ─ ─ ─►

t=5Y ─ ─ ─► 1Y(5Y)x5Y Canary

t=2Y ─ ─ ─ ─ ─ ─ ─► 2Y(3Y)x4Y Canary

─ ─ ─ ─ ─ ─ ─► t=4Y ─ ─ ─ ─ ─ ─ ─►

t=3Y ─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─► 3Y(4Y)x3Y Canary

─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─►t=4Y

─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─►

1Yx5Y Bermudan swaption

Spanned Canary swaptions

20%

40%

60%

80%

100%

120%

-150 -75 -25 0 25 75 150

BC basis/Berm. vega BE basis/Berm. vega





November 24, 2011

Srini RamaswamyAC

(1-212) 834-4573




151

All else equal, this argues for structuring portfolios to

have directional vega exposure—i.e., positions that

become long vega in a sell-off are attractive.

Second, one consequence of the Fed’s recent initiation of

Operation Twist has been the plunge in correlation

(Exhibit 40). While this is most noticeably true for

correlations between front-end and back-end rates, it is

also true more broadly. Last, with liquidity likely to

remain poor, realized volatility is likely to remain high as

well and preserve an inverted implied volatility curve

across expiries—i.e., short expiries are likely to remain

elevated relative to longer expiries. One corollary of the

last two points—cheap implied correlation and an

inverted volatility curve—is that forward volatility

appears cheap (Exhibit 41).

Thus, vega strategies for 2012 should seek to leverage

the current cheapness of implied correlations and forward

volatility, while creating positions that gain vega in aselloff. Interestingly, Bermudan receiver swaptions offer

a way to do all of those in one package. Since Bermudan

swaptions can be exercised at several points in time, the

value of near-term exercise grows with the moneyness of

the option. Thus, Bermudan receivers, for instance,

should increasingly track short-expiry receiver swaptions

in a rally, but effectively behave more like longer-dated

vega instruments in a sell-off (see Introduction to

Bermudan Swaptions and a Framework for Analysis,

Exhibit 45: The Canary CTD mimics the Bermudan riskcharacteristics better than individual European counterparts,

both for delta …Deviation of the delta profile of selected European and Canary swaptions fromthe Bermudan delta profile plotted against parallel shifts of the yield curve; %

The delta profile is obtained as a % of the original delta (that is, for no changein the yield curve). Each curve is obtained by subtracting the European or Canary delta profile from the Bermudan delta profile. The 1Y(2.5Y)x5Y Canaryand the 1Yx5Y European are the CTDs.

Exhibit 46: … and vegaDeviation of the vega profile of selected European and Canary swaptions from the

Bermudan vega profile plotted against parallel shifts of the yield curve; %

The vega profile is obtained as a % of the original vega (that is, for no change in theyield curve). Each curve is obtained by subtracting the European or Canary vegaprofile from the Bermudan vega profile. The 1Y(2.5Y)x5Y Canary and the 1Yx5YEuropean are the CTDs.

Exhibit 47: Sources of value in the 5Yx30Y Bermudan receiver swaptionProperties* of Bermudan swaption, Canary CTD and volatility curve as of COB11/01/2011

* Premia quoted in bp of notional and vega in bp of notional by 1/10 bp shifts in dailyvolatility.

-0.3

-0.2

-0.1

0.0

0.1

0.2

0.3

-150 -100 -50 0 50 100 150

1Y(2.5Y)x5Y Canary

1Yx5Y European

2.5Yx3.5Y European

-3%

-2%

-1%

0%

1%

2%

3%

-150 -100 -50 0 50 100 150

1Y(2.5Y)x5Y Canary

1Yx5Y European

2.5Yx3.5Y European

Premiums from a Bermudan model

5x30 Berm receiver 1891.5 bp

5Y(14Y6M)x30Y Canary 1741.1 bp

Vega 33.1 bp of notl per 0.1 bp/day

Alternate fair values for Canary based on:

Implied correlation 1848.6 bp

Cheapness of Canary (& Berm)

In price terms 107.5 bp of notional

In vol units 0.32 bp/day

Mispricing in correlation markets

Canary price using implied correlation 1848.6 bp

Canary price using realized correlation 1851.7 bp

Correlation driven cheapness of Berm

In price units 3.1 bp of notional

In vol units 0.009 bp/day

Mispricing of 5x30/14.5x20.5 vol curve 0.001 bp/day

Total mispricing 0.335 bp/day

plus

plus

equals





November 24, 2011

Srini RamaswamyAC

(1-212) 834-4573




152

J.P. Morgan Research Note, July 1, 2003 for more

details). In addition, Bermudan swaptions are also

naturally long exposure to correlation and forward

volatility.

However, the complexity of Bermudan swaptions (due to

several possible exercise dates) means that they have

exposure to a number of different correlations and

forward volatilities. It is therefore unclear exactly which

implied correlations and which forward volatilities a

given Bermudan structure is chiefly exposed to, if such

simplification is even possible. More broadly, while

Bermudan receiver swaptions may offer a slate of

desirable characteristics, it is unclear if they offer cheap

means to add such exposures since developing a relativevalue view on Bermudan swaptions is challenging.

In this section, we build on our previous research on

Bermudan swaptions to develop exactly such a relative

value metric for each Bermudan structure. Our approach

is as follows.

First, we identify a simpler instrument that is “spanned”

by a Bermudan swaption that most closely mirrors the

Bermudan itself. In our earlier work, we noted that each

Bermudan swaption may be thought of as spanning a

basket of European swaptions—one for each possible

exercise date. Moreover, the premium of the Bermudanswaption must be higher than each of the European

swaptions spanned by it, and the European swaption with

the highest premium in that basket may be thought of as

the best European approximation to the more complex

Bermudan, at least locally (Exhibit 42). (We label that

particular European swaption as the Bermudan’s “CTD,”

motivated by the Treasury futures analogy and

notwithstanding the fact that it is the most expensive

European swaption in the basket, not the cheapest).

While that observation is useful in developing some

intuition regarding the behavior of Bermudan swaptions,

the relative instability of the Bermudan swaption’sEuropean CTD has in practice meant that it does not

easily lead to a value metric. In this piece, we examine a

natural generalization of that idea that does lead to a

value metric. Much like the Bermudan swaption may be

thought of as spanning a basket of one-exercise date

European swaptions (one for each possible date on which

the Bermudan can be exercised), we may also think of

the Bermudan swaption as spanning a basket of two-

exercise date Canary swaptions, with one Canary for

each possible pair of dates on which the Bermudan

swaption can be exercised (Exhibit 43). We recall here

that a Canary option is one that can be exercised on

exactly two dates; for instance, a 1Y(2Y)x5Y has

possible exercises one and two years from now into

swaps that end in six years. Moreover, just as was the

case earlier, the price of the Bermudan swaption must be

at least as large as the highest premium canary that is

spanned by the Bermudan—we refer to that particular

Exhibit 48: The relative value metric shows several sectors ascurrently mispricedComponents* of our relative value metric for Bermudan receiver swaptions andtotal in various sectors expressed in price terms (bp of notional) and Bermudanvolatility terms (bp/day) as of COB 11/01/2011

* Canary CTD premium using implied correlation minus Premium of the CanaryCTD implied by Bermudan price.** Canary CTD premium using realized correlation minus premium using impliedcorrelation.*** Canary CTD premium using the second forward rate volatility shifted by volcurve mispricing (from a 2-year regression of vol curve versus second vol level)minus premium using current implied volatilities and realized correlation.

Exhibit 49: Total mispricing of the 1Yx10Y Bermudan receiver swaptionover the past four yearsTotal mispricing in the 1Yx10Y Bermudan receiver swaption; bp/day

* Negative values indicate cheapness of Bermudan swaptions

Sector

Price Vol Price Vol Price Vol Price Vol

5Yx30Y 107.5 0.32 3.1 0.01 19.3 0.06 129.8 0.39

6Mx30Y 93.5 0.33 -1.1 0.00 13.3 0.05 105.7 0.37

1Yx30Y 91.0 0.31 -0.9 0.00 13.7 0.05 103.9 0.36

3Yx30Y 93.2 0.30 0.7 0.00 15.9 0.05 109.7 0.35

2Yx30Y 86.1 0.28 -0.4 0.00 14.6 0.05 100.4 0.33

6Mx2Y 2.6 0.24 -0.2 -0.02 0.3 0.03 2.8 0.25

6Mx10Y 7.7 0.09 7.3 0.09 1.7 0.02 16.6 0.20

3Yx10Y 13.8 0.12 5.1 0.04 0.3 0.00 19.2 0.17

5Yx10Y 12.0 0.09 8.3 0.06 1.3 0.01 21.5 0.17

3Yx7Y 4.5 0.06 6.0 0.08 1.0 0.01 11.5 0.14

1Yx10Y 7.2 0.08 3.5 0.04 0.9 0.01 11.6 0.13

2Yx10Y 8.2 0.08 4.5 0.04 0.4 0.00 13.1 0.13

5Yx5Y 6.4 0.10 -0.3 -0.01 1.6 0.02 7.7 0.12

6Mx3Y 1.6 0.09 1.9 0.11 -1.6 -0.09 1.9 0.11

6Mx5Y 1.6 0.05 3.8 0.11 -3.3 -0.10 2.0 0.06

Bermudan / Canary

basis cheapness*

Correlation

cheapness** Total

Vol curve

cheapness***

-0.6

-0.4

-0.2

0.0

0.2

Nov 07 May 08 Dec 08 Jun 09 Jan 10 Aug 10 F eb 11 Sep 11





November 24, 2011

Srini RamaswamyAC

(1-212) 834-4573




153

Canary as the Canary CTD, in keeping with our earlier

terminology.

But why is it useful to think of a Bermudan swaption as

spanning a set of Canaries? Because the Canary CTD is

much more stable than the European CTD, across a range

of rate, curve and volatility moves, making it a useful

simplification (Exhibits 44-46). Moreover, canaries also

are exposed to exactly one correlation and forward

volatility, which will be useful to us. For instance, the

5Yx30Y Bermudan currently has the 5Y(14.5Y)x30Y as

its Canary CTD, with resulting exposure to 5Y forward

9.5Yx20.5Y correlation and forward volatility.

We may now develop a relative value metric. In doing

so, we note that there are three possible sources of value.First, the price of the canary CTD in the Bermudan

swaptions market (or from a suitably calibrated

Bermudan swaption pricing model) may differ from the

price of a canary that can independently be calculated

from vanilla implied volatilities and implied correlations

from the YCSO market. Second, using an empirical

framework, we may conclude that implied correlations in

the YCSO market are themselves rich or cheap, based on

comparisons to realized correlations over time. Third,

again using an empirical framework, we may take a view

on the volatility curve corresponding to a given canary.

Last, we may total up the mispricing from all threesources, and divide by the vega of the Bermudan

swaption to compute the richness or cheapness of the

Bermudan swaption, expressed in basis points per day.

This is illustrated in detail below for the case of the

5Yx30Y Bermudan receiver swaption in Exhibit 47. The

sources of value are also illustrated for the currently most

mispriced Bermudan receiver swaptions in Exhibit 48,

while Exhibit 49 shows a time series of the total

mispricing in the past four years for the particular case of

the 1Yx10Y Bermudan receiver. Large negative values

of our total mispricing should be taken as a signal to buy

the Bermudan.

As a way to analyze the usefulness of our metric we

evaluated the performance of buying several Bermudan

receiver swaptions at previous points in time in 2011

when our metric suggested cheapness. Three strategies

were tested: (i) hedging only the delta, (ii) hedging the

delta as well as the vega by selling the European “CTD”

swaption, and (iii) hedging the delta as well as vega by

selling a portfolio of replicating European swaptions.

Our results in this regard are preliminary, and much more

analysis is needed. However, our initial results are

promising, and suggest that such a relative value metric

is in fact useful in signaling cheapness of Bermudan

swaptions (Exhibit 50); our (limited) results currently

appear to suggest that buying the Bermudan versus the

European CTD swaption (and dynamically delta-hedging

the switch) is the best way to monetize the cheapness of

Berms.

Regulatory updateUnder the Dodd-Frank Act, several government agencies

such as the CFTC, SEC, FDIC, and the Federal Reserve

were tasked with rule-making responsibilities. We review

developments this year on this front, as well as some

potential changes in markets that could result from the

broader regulatory reform effort.

Proposed and finalized rules

The CFTC, the key rulemaking body as it pertains to

derivatives markets, was tasked under the Dodd-Frank

Act to regulate Swap Dealers (SD) on capital and margin

requirements, as well as on more stringent recordkeeping

and reporting. The act also required the CFTC to specify

rules related to the migration of trading standardized

derivatives to regulated exchanges or Swap Execution

Facilities (SEF).

The process of arriving at final rules has proved fairly

time intensive. In July, the CFTC issued an order

clarifying the effective date of the provisions in the swap

regulatory regime established by Title VII of the Dodd-

Exhibit 50: More empirical analysis is needed, but our results currentlysuggest that our metric does indeed identify value in Bermudanswaptions, and also that the best way to monetize this metric is to buythe Bermudan swaption against the European CTD on a delta hedgedbasis3-month return of long Bermudan positions delta hedged*, vega hedged with respect tthe European CTD, and vega hedged to a replicating portfolio**;

* Options are delta hedged daily and rolled every month. Returns assume zerotransaction costs.

** Replicating portfolio refers to a weighted combination of European swaptions that ar“spanned” by the Bermudan swaption. The weights are solved for using a scheme toreplicate the Bermudan swaption’s characteristics under a range of scenarios. SeeIntroduction to Bermudan Swaptions and a Framework for Analysis, J.P. MorganResearch Note, 7/1/2003 for details.

bp of notl. bp/day bp of not l. bp/day notl. bp/da

1/3/2011 -0.44 88.5 0.4 25.0 0.1 8.5 0.0

5Yx30Y 3/1/2011 -0.44 -105.4 -0.4 311.3 1.3 28.5 0.1

4/1/2011 -0.43 -125.8 -0.5 61.9 0.3 -25.9 -0.1

5/2/2011 -0.38 -217.4 -0.9 185.8 0.8 -152.6 -0.6

Trade start

dateBermudan

Delta hedged returns over 3-months

Vega hedged to

European CTD

Vega hedged to

replicating portfo

Mispricng at

inception

Bermudan

vol terms

(bp/day)





November 24, 2011

Srini RamaswamyAC

(1-212) 834-4573




154

Frank Act. The order provides temporary relief from

provisions that that would have become effective July 16,

2011 up to the effective date of final rules or December

31, 2011. Specifically, the CFTC provided relief from

certain provisions in the Commodities and Exchange Act

(CEA) that do not require rule-making but reference one

or more terms regarding swap entities and instruments

(such as the terms “swap,” “swap dealer,” “major swap

participant,” or “eligible swap participant”) that the act

requires be further defined. The CFTC has, jointly with

the SEC, issued notices of rulemaking to further define

these terms. In its latest amendment to its July order, the

Commission is proposing to extend the latest date of

expiry of relief from December to July 16, 2012 as it has

not yet finalized its definition of these terms. As such,several aspects of the rule-making are still in various

stages of either being open to comments or under

discussion by the commission. Rules in this stage

include, but are not restricted to, key items affecting

OTC derivative markets as relating to margin

requirements for uncleared swaps, capital requirements

for Swap Dealers (SD) and Major Swap Participants

(MSP), and collateral type and custody rules. Other items

yet to be finalized include rules on Derivatives Clearing

Organizations (DCOs) with regards to financial resources

and the resource waterfall in case of a default by a

clearing member, as well as end-user exceptions from the

mandatory clearing requirement. Several rules were infact finalized, such as those related to swap data

repositories, on particular products such as agricultural

swaps, and certain enforcement-related items, but as

noted earlier, key rules remain outstanding.

The Dodd-Frank Act also required the Federal Reserve to

implement several provisions, sometimes in conjunction

with other agencies. As part of this inter-agency effort

(between the Fed, FDIC, FCA, FHFA and OCC), a

proposal seeking comment was issued in April that

would require swap entities regulated by the five

agencies to collect minimum amounts of initial margin

and variation margin from counterparties to non-clearedswaps and non-cleared, security-based swaps (with a

commercial end-user exception as long as the entities’

margin exposure is below an appropriate credit exposure

limit). The amount of margin required under the

proposed rule would vary based on relative risk of the

counterparty and of the swap. The proposal would apply

to new, non-cleared swaps entered into after the proposed

rule’s effective date. Existing capital standards that apply

to swap entities as part of their prudential regulation

already address non-cleared swaps, and therefore are not

separately addressed in the rule.

The Fed also requested public comment on the “Volcker

Rule” requirements, developed jointly with the FDIC,

OCC, SEC and CFTC. Briefly, the Volcker Rule prohibits insured depository institutions, BHCs, and their

subsidiaries or affiliates from engaging in short-term

proprietary trading of any security, derivative, and

certain other financial instruments for the entity’s own

account, subject to certain exemptions. The act also

prohibits exempt transactions that may result in a conflict

of interest with customers or in a material exposure to

high-risk assets or trading strategies as defined by the

rule. The proposal clarifies the scope, as well as provides

Exhibit 51: Cleared interest rate swap volumes have risen significantlyover the last few yearsPercent of notional outstanding cleared via LCH

Source: ISDA, BIS, LCH

Exhibit 52: Cleared volumes of other interest rate derivatives have risenover the year Percent of notional outstanding centrally cleared in 2011

Source: Trioptima

20%

30%

40%

50%

60%

2007 2008 2009 2010

10%

13%

16%

19%

22%

25%

28%

31%

58%

60%

62%

64%

66%

68%

70%

72%

Dec 10 Feb 11 Apr 11 Jun 11 Aug 11 Oct 11

IRSOISIR Basis Swap (rhs)





November 24, 2011

Srini RamaswamyAC

(1-212) 834-4573




155

certain exemptions to prohibitions in the act.

Transactions in certain instruments, such as obligations

of the US government or a US government agency, theGSEs, and state and local governments are exempt from

the statute’s provisions. As expected, it also exempts

market-making, underwriting, and risk-mitigation

hedging. It also provides commentary to assist entities in

distinguishing market making-related activities from

proprietary trading.

Finally, we note that next year brings the observation

period for the Liquidity Coverage Ratio (LCR) proposed

under Basel III. Relevant regulators will start collecting

data to assess LCRs, with mid-2013 being the deadline

before which any changes to the current LCR proposal

may be announced. Our short-term analysts note that inits existing form, LCR is likely to be fairly onerous on

banks. In its current form, this requirement would result

in increased demand for scarce high-quality collateral.

Market developments and outlook

Despite extensions in the rule-making process, markets

have moved ahead towards a central clearing model. This

trend towards a clearing model predates the new

regulations (Exhibit 51). As of October 2011, around

half of the $511tn interest rate derivatives outstanding

were cleared. A majority of the cleared derivatives were

interest rate swaps (87%), which are about 60% of notional of interest rate derivatives outstanding. Over the

last year, roughly two-thirds of interest rate swaps

outstanding (based on USD-equivalent notionals) have

been centrally cleared, and the percentages of other

products currently cleared (OIS and basis swaps) have

grown substantially (Exhibit 52).

Another trend that predates post-crisis regulations is the

rise in the level of OTC derivative collateralization. As

seen in Exhibit 2, margin coverage in fixed income

derivatives is high, and covered over half the trade

volume well before the crisis.

One outcome of the move to central clearing and more

stringent collateral requirements has been the move to

OIS discounting as discussed earlier. The move would

result in a one-time accounting gain (loss) depending on

whether the counterparty was long (short) an in-the-

money swap position. More interesting is the question of

whether there would be a significant switch from Libor-

based hedges to OIS-based ones. Given the much higher

levels of liquidity in Libor swaps, we think this is

unlikely in the near term.

Another potential consequence of regulatory changes is a

change in issuer behavior. Corporate issuers that swap

their fixed-rate debt issuance may incur higher costs

irrespective of whether they may avail of the end user

exemption from mandatory clearing because their dealer

counterparties will incur higher capital costs on

uncleared swaps. The proposals could also impact

callable issuance. Many issuers of callables routinely sell

the embedded option on issuance to monetize favorable

pricing and achieve a lower funding cost. While

swaptions are not yet a cleared product, they will

nevertheless require their own collateral. Even though

this should not be an issue for some larger issuers like

FNMA and FHLMC that carry a significant amount of

eligible collateral, other issuers like FHLB that do not

(and non-US issuers such as KFW that do not typically

fully collateralize their derivative positions) would have

to evaluate the cost savings from callable issuance

against additional collateral costs. On the margin, this

could result in lower callable issuance and lower vol

supply. We note, however, that the rules are not finalized

with regards to which of these entities will be eventually

subject to mandatory clearing and margin requirements,

even as entities seek exemptions from various provisions

of the act.

Trading themes

• Look for FRA-OIS spreads to stay wide or widen

further into 1Q12, but begin to narrow thereafter

In the near term, worsening conditions in Europe are

likely to bias the EUR/USD cross currency basis

narrower, bias the semi-peripheral European sovereign

spreads wider, and preserve upward pressure on FRA-

OIS spreads. However, looking past this likely near-

term intensification, policy actions (a potential

reduction in the penalty on the Fed’s USD swap lines

to 50bp from 100bp) are likely to eventually dominate, pushing FRA-OIS spreads narrower to 50bp by the

end of 1H12.

• Swaptions on short tails are effectively options on

front-end swap spreads—use receiver/payer

swaptions in place of front-end spread

narrowers/wideners

With the Fed’s conditional commitment to low rates

until mid-2013 helping to keep front-end Treasury

yields low and sticky, front-end swap spreads have





November 24, 2011

Srini RamaswamyAC

(1-212) 834-4573




156

become highly correlated to front-end swap yields.

Thus, options on front-end yields can effectively serve

as options on front-end swap spreads. Buy payer

swaptions or 1:1 weighted payer swaption spreads in

place of front-end swap spread wideners, and buy

receiver swaptions or receiver swaption spreads as

proxies for front-end spread narrowers.

• Position for wider 10-year maturity matched swap

spreads in the near term …

In the near term, FRA-OIS spreads will likely

continue to widen, and banking stock valuations will

likely remain under pressure, both of which should

pressure 10-year swap spreads wider. We expect 10-

year swap spreads to hover near 27bp towards the endof this year and in early 1Q12.

• … but look to initiate narrowers towards the end of

1Q12

High grade corporate issuance tends to exhibit strong

intra-year seasonal patterns, and issuance-related

swapping is likely to pick up over the March-May

period. Intermediate maturity swap spreads have

consequently tended to narrow in this period. In

addition, as we look into next year, the growing

likelihood of concerted action by central banks

(including perhaps lower penalties on USD swap lines

with the Fed) should also help swap spreads to turn thecorner. Look for narrowing to 18bp by mid-year.

• Trade intra-month swings in the pace of Fed

purchases of longer-end Treasuries

While the Fed tends to purchase similar amounts each

month in total, the actual schedule of purchases

(which is known ahead of the month) can produce

significant intra-month variations in pace.

Intermediate maturity swap spreads have tended to

narrow in periods characterized by a sharp slowing in

pace, while widening when the opposite is true.

• Take advantage of relative value opportunities insectors subject to Fed purchases

A simple strategy of buying the cheapest issue

(measured using yield error as a metric) and selling

the richest issue on a maturity matched swap spread

switch basis has been attractive in periods and sectors

where the Fed is actively purchasing Treasuries. With

Operation Twist slated to continue until the end of

1H12, we would look to make such switches a key

part of our relative value trading strategy in 1H12.

• Position for a steeper yield curve between

intermediates and the long end, hedged with red

Eurodollar shorts

Steepeners anchored in the 5-year sector and beyond

offer attractive carry and slide in addition to being

considerably mispriced relative to front-end yields.

Also, thanks to the flattening of the front end of the

curve, shorts in red Eurodollars have small carry costs,

and also help mitigate exogenous risks.

• Take advantage of yield curve nonlinearities

induced by the stickiness of front-end yields near

the zero bound

As front-end yields become highly sticky near the zero

bound, their relative volatility versus points further outthe curve falls, causing locally level and curve neutral

belly-richening butterflies to exhibit empirical

convexity versus front end yields. Butterflies that are

likely to exhibit the greatest asymmetry can be

identified by examining implied volatilities and skew

information. In addition, some of these butterflies

have positive carry and slide, making them

particularly attractive ways to position for an

unanticipated front end selloff.

• The yield curve should remain highly directional

with rates in 1H12; combine YCSOs with

swaptions to create “synthetic” conditional curvetrades that offer better entry levels than regular

conditional curve trades

Sticky front-end yields are likely to cause the yield

curve between the front end and intermediate yields to

remain highly directional with yield levels for the

foreseeable future. However, swaptions markets have

priced such directionality in (albeit to varying

degrees), and entry levels are often unattractive. By

exploiting the correlation of the 10s/30s swap curve

with front-end yields, and given the richness of

YCSOs on the 10s/30s curve, replacing swaptions at

the front end of the curve with suitably weighted

YCSOs results in a “synthetic” conditional curve tradethat can be constructed at much better entry levels.

• Stay long gamma going into year end and in 1H12

A persistent crisis in Europe and poor risk appetite

should cause market depth to stay depressed for much

of 4Q11 and 1H12, creating conditions favorable for

long gamma positions.





November 24, 2011

Srini RamaswamyAC

(1-212) 834-4573




157

• Look for intermediate-expiry swaption implied

volatility to drop into 1Q12, but stay range-bound

thereafterIntermediate expiry swaption implied volatility

currently looks rich; look for 3Yx10Y swaption

implied volatility to fall to 6bp/day by the end of

1Q12, and remain range-bound thereafter.

• Look to add exposure to Bermudan receiver

swaptions

Cheap implied correlations and an inverted implied

volatility curve have made forward volatility cheap. In

addition, the strong relationship between implied

volatility and rates that emerges in low yield regimes

makes it desirable to hold positions that become

longer vega in a sell-off. One attractive way to initiatedirectional vega exposure, while also positioning for a

rise in implied correlations and forward volatility, is to

replace European receiver swaptions with Bermudan

receiver swaptions.



Japan Rates ResearchGlobal Fixed Income Markets 2012 OutlookNovember 24, 2011

Takafumi YamawakiAC

(81-3) 6736-1748

Yuya Yamashita, CFA (81-3) 6736-1493J.P. Morgan Securities Japan Co., Ltd

158

Japan

• 2011 was a noteworthy year for the JGB market

in three respects: 1) The Great East Japan

Earthquake of 11 March 2011 2) JGB-UST

decoupling, and 3) A weakening of the ultra-long

sector. However, JGB stayed within a narrow

range of 0.95-1.35% in 2011

• Our main scenario points to a continuation of

range-bound trading while our risk scenarios

suggest that JGB yields are more likely to fall

than rise. The downside in yields could be tested

if the global economy loses momentum in 1H2012

while a gradual cyclical recovery in 2H12 should

trigger only moderate rises in yen interest rates.We are therefore anticipating ranges of 10Y JGB

yields as 0.8-1.1% for 1H12 and 0.9–1.3% for

2H12

• This report considers why Japanese sovereign

debt can be considered relatively ‘safe’ by

international standards while also identifying

some cause for concern in the longer term

• We would consider any rise in JGB yields that

might be triggered by the FY12 issuance plan to

be little more than extremely short-lived noise

• We expect short-end swap rates to remain sticky

under the BoJ’s interest rate policy. We also have

in mind a risk scenario where LIBOR rises on a

possible worsening of the crisis in Europe. We

have a flattening bias on JPY swap curve as a

whole

• We have a widening bias on short-end swap

spreads. We think the cheapness of JGB vs. swap

rates in the super-long sectors will be sustained

into 2012

• Large negative cross-currency basis spreads

make it attractive to create synthetic USD bonds

by buying JGBs and swapping them to USD.

With higher FX hedge cost, Japanese investorsmay accelerate the repatriation of USD funds

back to JGBs

• We have a tightening bias on the TIBOR/LIBOR

fixing spread on positive convexity. We expect

more tightening if LIBOR increases than

widening if LIBOR falls

• We think the short-to-medium-term 3s/6s basis

spreads are biased wider

The JGB market in 2011

Three key points – What sort of year has 2011

been for the JGB market?

US Treasury yields rose sharply in January and February

as equity prices were driven higher by the Fed's late-2010

QE2 announcement and the Obama administration's

decision to extend the so-called ‘Bush tax cuts’. The 10Y

JGB yield hit its year-to-date high of 1.35% as yen bonds

were caught up in this global selloff, but the domestic

market then turned unusually volatile in the wake of the11 March 2011 Great East Japan Earthquake, with losses

on stock positions apparently triggering a certain amount

of profit taking in JGBs towards the 31 March end of the

Japanese fiscal year (Exhibit 1)

Yen interest rates followed UST yields lower from mid-

April through July, but then failed to keep pace with

further declines from August onwards. The year as a

whole has thus been characterised by a rather gradual

downtrend in JGB yields since May.

2011 was a noteworthy year for the JGB market in three

respects.

1: The Great East Japan Earthquake of 11 March 2011

The Great East Japan Earthquake of 11 March 2011

appears set to be one of the defining events of Japan's

already long history.

While the initial earthquake ended up causing relatively

little direct damage, the subsequent tsunami destroyed a

number of towns and villages and triggered a major crisis

at the Fukushima Daiichi nuclear power plant. Japan is

Exhibit 1: JGB stayed in a narrow range of 0.95-1.35% in 201110YJGB yield history; %

0.80

0.90

1.00

1.10

1.20

1.30

1.40

1.50

Jan-10 Jul-10 Jan-11 Jul-11

QE2

Earthquake

Decoupling

from UST

high correlation

with UST

1.35%

0.95%




Takafumi YamawakiAC

(81-3) 6736-1748


159

by no means unaccustomed to earthquakes and tsunami,

but the virtually unprecedented size of this tsunami

nevertheless shocked the nation, with significant negativeramifications for consumer spending, manufacturing

supply chains, and electricity supply capacity. The

collapse and rebound of industrial production provides a

particularly clear picture of the extent to which economic

activity was impacted by the disaster (Exhibit 2).

The following couple of months saw JGB market

participants focus on: 1) the potential for post-quake

reconstruction to drive a V-shaped economic recovery;

and 2) the prospect of an increase in JGB issuance to pay

for post-quake rebuilding. The subsequent economic

recovery fell somewhat short of initial expectations,

however, and with the JGB market asked to absorb onlya limited amount of additional issuance, a significant rise

in yen interest rates was averted.

2: The Euro zone crisis and JGB-UST decoupling

UST-JGB correlations declined markedly as the

European sovereign debt crisis dominated headlines from

August onwards. UST and German Bund yields fell

sharply as the US economy too began to lose momentum,

but JGB yields dropped only slightly during this period.

Exhibit 3 illustrates the extent to which yen interest rates

were ‘left behind’ by the fall in UST yields.

3: A weakening of the ultra-long sector

The performance of ultra-long JGBs was a key difference

between 2010 and 2011. The 30s/40s curve steepened

sharply this year (Exhibit 4), as demand for these

longest-dated bonds dropped dramatically while demand

for other parts of the super-long sector remained strong.

The 40s have a very thin investor base compared with

that of the 20s, however, and we will be monitoring this

recent slackening of the supply/demand in the 40s to see

if it continues to put steepening pressure to the JGB yield

curve in 2012.

Exhibit 3: UST-JGB correlations declined markedly from Augustonwards.UST 10Y vs. JGB 10Y since October 2010; %

0.70

0.80

0.90

1.00

1.10

1.20

1.30

1.40

1.50

1.50 2.00 2.50 3.00 3.50

JGB 10y r

UST 10y r

Oct10 - July11

Aug11-Nov11

Exhibit 2: The collapse and rebound of industrial production dataprovide a particularly clear picture of the extent to whicheconomic activity was impacted by the disaster Industrial production seasonal adjusted; CY 2005 = 100

75

80

85

90

95

100

Jul-10 Oct-10 Jan-11 Apr-11 Jul-11

Earthquake

Exhibit 4: The 30s/40s curve steepened sharply this year JGB 30s/40s curve history; bp

0

5

10

15

20

25

30

Oct-10 Jan-11 Apr-11 Jul-11 Oct-11




Takafumi YamawakiAC

(81-3) 6736-1748


160

Outlook for the JGB market in 2012

Six factors support a relatively narrow trading

range

While we do not expect volatility to remain as low as it

has been in 2H11, we nevertheless expect to see a

continuation of relatively range-bound trading in 2012.

Looking at the key factors summarised in Exhibit 5, we

see little prospect of significant yield movements in

either direction. Below, we discuss our main scenario

and risk scenarios for each of these factors.

1: Trade balance

Main scenario: Large current account surplus despite

near-zero trade balance (neutral bias on yields: 75%)Exhibit 6 shows Japan’s trade balance, which remained

in deficit this year as exports fell in the wake of the

earthquake, and import value was driven up by high

commodity prices. The trade deficit exceeded ¥400bn

(seasonally adjusted) in both April and May, but had

narrowed to ¥21.8bn as of September. Japan may find it

difficult to book a consistent trade balance surplus in

2012, given the prospect of exports being held back by a

global economic slowdown, but we expect it to remain

roughly in equilibrium (i.e. near zero) over the year as a

whole. This would see Japan's massive income surplus

translate into a similarly large current account surplus,

implying that the market should have little problem

absorbing even relatively high levels of JGB issuance.

Risk scenario: Severe overseas economic slowdown or

rise in commodity prices (upward bias on yields: 20%)

Japan’s trade balance would likely plunge deep into

deficit under a severe global recession, thereby fueling

speculation that a current account surplus might be

difficult to sustain in the longer term and potentially

triggering a selloff of JGBs by foreign investors.

Exhibit 6: Japan's trade balance remained in deficit this year and thecurrent account surplus has been gradually shrinking recentlyCurrent account surplus in Japan; ¥tn

-10

-5

0

5

10

15

20

25

30

35

96 97 98 99 00 01 02 03 04 05 06 07 08 09 10 11

Income Trade Balance

C urrent T rans fer Serv ic e

Exhibit 5: Six factors support for the JGB market moving in a relatively narrow range, but with a downward bias in yieldsSix factors for the JGB market, their bias towards yield move and our subjective probabilities assigned to each scenario

ProbabilityYield

Down

Yield

Neutral

Yield

UpComments

1 Trade Balance 5% 75% 20%If trade balance stays in the negative for a long time, some investors

may have concerns about JGB sustainability.

2 BoJ Additional Easing 35% 60% 5%If the BoJ indroduce further easing such as increasing Rinban,

extending the maturity of the Asset Purchase Program, and lowering the IOER

3 Domestic Politics 10% 65% 25%If Japanese economy goes into a recession, tax increase might

be extended.

4 Lifer's Investment Activity 40% 50% 10%If US rates stay around the current levels for a long time, lifers will

shift their investment money from USTs to JGBs.

5 Bank Activity 30% 60% 10%If banks face difficulties in non-performing loans, they will buy

JGBs more aggressively.

6 UST Rates 30% 50% 20% If UST stay below 2%, JGB yields will follow USTs.




Takafumi YamawakiAC

(81-3) 6736-1748


161

2: Additional BOJ easing

Main scenario: Additional easing limited to further

expansion of Asset Purchase Program (APP) (neutral bias on yields: 60%)

The Bank of Japan eased monetary policy on three

occasions following the 11 March disaster.

14 March 2011: Increase in the amount of APP: + ¥5tn

4 August 2011: Increase in the amount of APP: + ¥10tn

27 October 2011: Increase in the amount of APP + ¥5tn

However, each of these actions was limited to an

expansion of the central bank’s APP, with the BoJ

refraining from exercising any of its other options (2-4

below):

1) Further expansion of APP

2) Lengthening of duration of APP JGB purchases

(from 2 years to 5 years)

3) Lowering of the interest rate paid on excess

reserve balances

4) Increase in Rinban outright JGB purchases

With overseas central banks seemingly reluctant to lower

IOER (interest on excess reserves) rates or increase their

outright bond purchases, we see little prospect of the BoJ

opting for any of the options (2-4) unless the domestic

economy were to slow dramatically. Central banks are

increasingly prone to criticism for both direct

underwriting of government debt and excessive

purchases via the secondary market. With the BoJ

understandably keen to avoid any negative surprises on

the monetary policy front, we do not expect to see any

announcements with sufficient impact to knock JGB

yields out of relatively narrow trading ranges.

Risk scenario: Yen driven higher by aggressive Fed

easing (downward bias on yields: 35%)

There is at least some chance that the BoJ will bow to

pressure from markets and politicians by announcing

slightly more aggressive measures, with (2) perhaps themost likely option (other than a further APP expansion)

by virtue of its relative ease of implementation. If

USD/JPY were to drop below 70 following a QE3

announcement by the Fed, then we would expect the BoJ

to respond by lengthening the duration of its APP JGB

purchases, thereby generating downward pressure on the

curve as a whole and the 5Y yield in particular.

3: Domestic political situation

Main scenario: Decision on consumption tax hike helps

to maintain a minimum level of fiscal discipline (neutral bias on yields: 65%)

Fiscal discipline became a major policy priority after

Yoshihiko Noda took over from Naoto Kan as Prime

Minister, with the need for sharply higher spending in the

wake of the 11 March disaster seemingly

counterbalanced by a strong commitment to responsible

fiscal management. The Noda administration also

appears determined to hike the consumption tax rate. In

fact, the administration has set its guidline to hike the

consumption tax rate from 5% to 10% in two stages from

October 2013. Fiscal policy thus appears unlikely to be

an Achilles heel for the JGB market provided that the Noda government can remain in office.

Risk scenario: Political squabbling sees fiscal discipline

thrown out the window (upward bias on yields: 25%)

A severe global economic downturn could potentially see

the fiscally hawkish Noda administration overthrown, in

which case the new government could be expected to

loosen the fiscal reins and relegate a consumption tax

hike.

4: Investment by life insurers

Main scenario: Shift out of foreign bonds into JGBs(neutral bias on yields: 50%)

Japan has also seen an increase in USD funding costs

(Exhibit 7), with 3M roll cost rising to around 0.8%

annualised. This means that less than 1.5% of carry is

now earned by buying 10Y USTs, as a result of which

domestic investors have started to shift back into JGBs.

The focus is life insurance companies as their activities

are the key for the demand and supply balance of the

long end of the curve. Their buying has been already

modestly increasing over the past few years (Exhibit 8).

Some investors may also have replaced a portion of their

FX-hedged foreign bond holdings with unhedged

positions, but full substitution appears unlikely given the

risk it would entail. We therefore expect to see a gradual

shift into (super-long) JGBs that should help to limit any

rise in yields towards the far end of the curve.

Risk scenario 1: Renewed investment in foreign bonds

following a resolution of the Euro zone crisis (upward

bias on yields: 10%)

Confidence in USD funding may be boosted if a

significant improvement in the Euro zone situation drives

10Y UST and German bund yields substantially higher,




Takafumi YamawakiAC

(81-3) 6736-1748


162

in which case European bonds and USTs could start to

look more attractive to domestic lifers from a relative

value perspective (versus JGBs).Risk scenario 2: Increased investment in JGBs as the

Euro zone situation continues to deteriorate (downward


A further escalation of the Euro zone crisis would

probably make lifers increasingly risk averse, with a

selloff in domestic equities – and the corresponding need

to reduce risk exposure – likely to accelerate purchases

of super-long JGBs.

5: Investment by banks

Main scenario: Little change in earnings, little need for

aggressive trading (neutral bias on yields: 60%)

Recent financial results indicate that bad-debt disposal

costs remain well under control, suggesting that there is

little need for banks to chase relatively risky returns in

the JGB or the broader securities markets. We therefore

expect to see only moderate trading activity, with banks

likely to take profits when yields fall and buy on (price)

dips when yields rise. This should help limit overall

market volatility.

Risk scenario: Economic downturn increases reliance on

trading profits (downward bias on yields: 30%)

It is possible that a further escalation of the Euro zonecrisis could have negative ramifications for Japan and the

rest of Asia, in which case domestic banks may be forced

to step up their bond trading with a view to covering

higher bad loan costs. This could result in a slightly more

volatile market, with JGB yields likely to decline at least

temporarily.

6: Correlation with US Treasuries

Main scenario: JGB yields unlikely to rise sharply if the

10Y UST yield remains somewhere around 2.5% (neutral


As should be evident from Exhibit 3, the JGB marketwas much more sensitive to movements in UST yields

prior to August 2011. A return to pre-August correlations

would imply that a 10Y UST yield of around 3.0%

should be commensurate with a 10Y JGB yield of around

1.20%. (Our UST research team is forecasting that the

10Y UST will stay below 3.0% throughout 2012.)

Our main scenario for 2012

Our main scenario for each of the above six factors

points to the continuation of range-bound trading, while

our risk scenarios suggest (on balance) that JGB yields

are more likely to fall than rise. The downside could be

tested if the global economy loses momentum in 1H12,

while a gradual cyclical recovery in 2H12 should trigger

only moderate rises in yen interest rates.

We are therefore anticipating ranges of 10Y JGB

yields as 0.80%–1.10% for 1H12 and 0.90%–1.30%

for 2H12.

Exhibit 8: Lifers’ activities are the key for the demand and supplybalance of the long end of the curve. Their buying has been modestlyincreasing over the past few years3M moving average of monthly purchase amount in super long end from lifers; ¥bn

0

200

400

600

800

1,000

1,200

1,400

Apr-08 Apr-09 Apr-10 Apr-11

3M moving average

Exhibit 7: The increase in USD funding cost has reduced theattractiveness of USTs. We therefore expect to see a gradual shift into(super-long) JGBs that should help to limit any rise in yields towards

the far end of the curve3M FX hedge cost in USD/JPY: % annualised

0.00

0.20

0.40

0.60

0.80

1.00

Oct-10 Jan-11 Apr-11 Jul-11 Oct-11

* 3M FX hedge cost is calculated as annualized 3M forward discount vs. spot FX level.Source: Bloomberg




Takafumi YamawakiAC

(81-3) 6736-1748


163

Japanese sovereign risk

Still relatively safe, but some longer-termconcerns remain

A continued escalation of the European debt crisis has

focused substantial attention on Japanese sovereign risk,

but the stability of domestic interest rates over the past

few years suggests that a ‘sovereign shock’ is still quite a

remote prospect. This report considers why Japanese

sovereign debt can be considered relatively ‘safe’ by

international standards, while also identifying some

cause for concern in the longer term.

Why the JGB market is still relatively safe?

1: Massive domestic savings: A healthy balance sheetfor the nation as a whole

Exhibit 9 shows (inward and outward) financial assets

and liabilities for the various sectors of the Japanese

economy, with equities counted separately owing to their

significant price volatility. These figures should help

give an overall picture of Japanese financial flows and

leverage.

The household sector is seen to hold net financial assets

totaling about ¥1,048tn, while the social security

system’s pension reserves amount to about ¥175tn in net

assets. The (central + local) government sector has netliabilities totaling ¥840tn, while non-financial private

corporations have around ¥39tn in net liabilities.

Foreigners’ net liabilities total ¥348tn, implying an

equivalent amount of Japanese claims against foreign

entities.

Overall, we believe that Japan's balance sheet is healthy,

with the existence of massive domestic asset holdings a

key difference from the current situation in Europe.

2: High levels of issuance digested with little difficulty

The past behaviour of these various sectors should help

explain why the market as a whole has had little

difficulty in absorbing continued increases in JGB

issuance.

• Foreigners’ net liabilities (= Japan’s net assets) tend to

decline when the yen is strengthening and increasewhen the domestic currency is weakening. In the

absence of particularly dramatic exchange rate

fluctuations, however, it seems likely that Japan will

continue to gradually accumulate net assets by virtue

of its consistent current account surplus.

• The existence of a current account surplus means that

the government sector’s fiscal deficit is being more

than offset by increases in the net assets of the private

sector (households + corporate). The corporate

sector’s net financial liabilities have declined sharply

in recent years owing to a diminished need for

borrowing, while the household sector has continuedto build up its net assets. Some analysts have

suggested that Japan’s adverse demographics – an

Exhibit 9: Given massive domestic asset holdings, we believe that Japan's balance sheet is healthyFinancial assets and liabilities by sectors; ¥tn

Inward Asset Inward Liability Inward Asset Inward Liability Inward Asset Inward Liability

2,430 2,642 607 480 1,392 443

Outward Asset Outward Liability Outward Asset Outward Liability Outward Asset Outward Liability

244 30 106 3 9 0

Net Asset Net Asset Net Asset

10 -39 1,048

Equity Asset Equity Liability Equity Asset Equity Liability Equity Asset Equity Liability113 105 138 407 90

Inward Asset Inward Liability Inward Asset Inward Liability Inward Asset Inward Liability

128 1,109 147 24 212 217

Outward Asset Outward Liability Outward Asset Outward Liability Outward Asset Outward Liability

88 2 31 0 34 472

Net Asset Net Asset Net Asset

-840 175 -348

Equity Asset Equity Liability Equity Asset Equity Liability Equity Asset Equity Liability

73 18 21 95

Financial Institution Non Fin Corporate Household

Government Social Security Foreigners




Takafumi YamawakiAC

(81-3) 6736-1748


164

aging and shrinking population – will eventually push

the household sector’s net assets into a downtrend, but

this has thus far been prevented by significanttransfers of income from the government sector to

households (Exhibits 10&11).

The resulting decline in corporate demand for funds and

increase in deposits have left banks awash with surplus

funds to invest in JGBs, thereby helping to keep yen

interest rates low and relatively stable. We expect that

this mechanism will continue to function for as long as

Japan’s current account balance remains in surplus.

3: Current account surplus has declined slightly but is

unlikely to disappear

It is of course necessary to recognise that Japan’s current

account surplus has been gradually shrinking recently as

shown in Exhibit 6. The trade balance deteriorated

sharply in the wake of the Great East Japan Earthquake

of 11 March, owing to supply chain disruptions and a

surge in oil imports. We expect to see only a moderate

surplus going forward given the likelihood that Japan

will continue to import fossil fuels while exporters are hit

by a global economic downturn.

Japan continues to book a large income surplus,

however, and we expect this to remain somewhere in the

order of ¥10tn per annum unless there is a significantdecline in net claims against foreigners. This should also

provide substantial support for the JGB market.

4: Difficult to envisage a default scenario

While there has indeed been some discussion of the

possibility that Japan might default on its debt, we

consider this an extremely unlikely scenario for the

following reasons.

Shifting out of JGBs during a ‘flight to quality’ will

entail foreign currency risk, making this a very difficult

option to contemplate for most market participants.(Shifting from Greek bonds into German bunds involves

no such risk). We expect support from domestic

investors not willing to take currency risk.

• Japan has its own central bank, which should be able

to purchase JGBs in the event of a crisis. This makes it

theoretically possible for Japan to avoid a default

indefinitely, i.e. the BoJ is theoretically able to

underwrite JGBs in the primary market with necessary

law changes.

• Even if the yen were to weaken as a consequence of

JGB purchases by the BoJ, Japan’s ample foreign

reserves mean that this would actually help to reduce

the state’s net debt. (In the case of Argentina,

depreciation of the home currency triggered a sharp

increase in USD-denominated liabilities.)

• A depreciation of the yen would also be likely to boost

exports and thereby bring about a significant increase

in the current account surplus, with positive

ramifications for the JGB market.

We therefore see very little risk of Japan facing a Greece-

like sovereign debt crisis.

Exhibit 10: Despite adverse demographics, the household sector hascontinued to build up its net assets…History of net asset in household sector; ¥tn

700

750

800

850

900

950

1,000

1,050

1,100

1997 1999 2001 2003 2005 2007 2009

Exhibit 11: … on the back of significant transfers of income from thegovernment sector to householdsHistory of net liability by sectors; ¥tn

0

100

200

300

400

500

600

700

800

900

1997 1999 2001 2003 2005 2007 2009

Government

Foreigners

Corporate




Takafumi YamawakiAC

(81-3) 6736-1748


165

5: Some gradual progress towards more stable fiscal

footing

There were concerns that the three supplementary budgets made necessary by the Great East Japan

Earthquake of 11 March would require a sharp increase

in JGB issuance, but the government has been able to

deplete frontloaded JGB issuance to the point where the

market will only be asked to absorb an additional ¥2tn

(or thereabouts), which suggests that any impact on the

overall supply/demand balance should be minimal.

Moreover, the government appears to be making slow-

but-steady progress in its efforts to return Japan to a

more sustainable fiscal footing.

• The government’s medium-term fiscal framework

covering the period through FY2014 (endorsed by the

Cabinet in August) seeks to limit annual general

account spending (excluding debt-servicing costs) to

¥71tn and new JGB issuance to ¥44tn. Cabinet Office

projections indicate that meeting these targets would

still leave Japan unable to achieve its longer-term goal

of wiping out the primary fiscal deficit by FY2020,

but it is nevertheless impressive that Japan has clearly

announced its intention to maintain a certain level of

fiscal discipline despite the massive shock of the 11

March disaster.

• The government has also indicated that the post-quake

rebuilding process is to be managed outside thegeneral budget, with the necessary funds to be repaid

via temporary tax hikes – including higher personal

income and residential taxes – and non-tax revenues.

• The Noda administration has set its guidline to hike

the consumption tax rate from 5% to 10% in two

stages from October 2013.

Cause for concern in the longer term

6: Massive imbalance between the government and

private sectors

Japan faces a severe imbalance where the governmentsector’s massive net liabilities (currently exceeding

¥800tn) are covered in their entirety by the domestic

private sector (households + non-financial private

corporations). It is also somewhat troubling that the

banking sector holds such a large proportion of the

private sector’s net assets. This lack of diversification has

the potential to exacerbate any upward movement in yen

interest rates.

As should be evident from Exhibit 12, there have been

two particularly sharp rises in JGB yields over the past

decade:

• June–August 2003, when a surge in VaR for the

banking sector triggered an across-the-board reduction

of interest rate risk, and

• 2005–2006, when banks sought to reduce their

exposure to medium-term JGBs ahead of a BoJ rate

hike.

Insufficient diversification on the part of investors was a

contributing factor in both cases.

7: Super-long issuance outstanding continues to rise

A sharp increase in outstanding issuance of super-long

JGBs is also worthy of some attention. Total outstanding

issuance beyond the 11Y sector has risen by around

¥40tn over the past two years and is projected to exceed

¥140tn by end-FY2011 (Exhibit 13). This extremely

rapid pace of supply reflects large and regular offerings

via 20Y, 30Y, and 40Y tenders as well as Auctions for

Enhanced Liquidity.

Japanese life insurers are already preparing for the

launch of a domestic version of the Solvency II

framework (currently being implemented in Europe)

around 3 – 4 years from now, and their associated ALM

measures appear likely to ensure relatively strong

demand for super-long JGBs over the next 2 – 3years.

Beyond that point, however, it may be necessary to find a

new investor base for super-long offerings. The current

debate over a possible lengthening of the average

Exhibit 12: The lack of diversification in the banking sector has beenthe main driver of the sharp rises in JGB yields over the past decadeLong history of JGB 10Y yield; %

0.20

0.40

0.60

0.80

1.00

1.20

1.40

1.60

1.80

2.00

00 01 02 03 04 05 06 07 08 09 10 11

Banks reduced

their rates exposure

in a rate hike cycle

Bank

TradingBank

Trading




Takafumi YamawakiAC

(81-3) 6736-1748


166

duration of JGB holdings for public pension portfolios

could be a step in the right direction.

Long-term outlook for the JGB market

As indicated in Exhibit 14, we expect the 10Y JGB yield

to remain below 1.5% over the coming 3 – 4 years before

shifting into a 1.5%–2.0% range for the next few years,

with fiscal risk unlikely to become a major market factor

until at least a decade from now.

Exhibit 13: Total outstanding issuance beyond the 11Y sector isprojected to exceed ¥140tn by end-FY2011, increasing the need tofind a new investor base for super-long offerings.

Total market outstanding of 10Y+ JGB excluding 15Y CMT, nominal bonds only; ¥tn

0

50

100

150

200

03 04 05 06 07 08 09 10 11 12 13 14

Forecast

Exhibit 14: We expect the 10Y JGB yield to remain below 1.5% over the coming 3–4 years before shifting into a 1.5%–2.0% range for thenext few years, with fiscal risk unlikely to become a major marketfactor over the next 10 yearsLong-term forecast of 10Y JGB yield; %

0.00

0.50

1.00

1.50

2.00

2.50

3.00

11 12 13 14 15 16 17 18 19 20 21 22 23 24

Short Term

Rangebound

10Y=0.8-1.5%

Medium Term

High Volatility

10Y=1.5-2.0%

Long Term

Gradual Yield Rise

10Y>2.0%




Takafumi YamawakiAC

(81-3) 6736-1748


167

Forecasts for FY2012 calendar basemarket issuance

The Ministry of Finance tends to announce its JGB

issuance plan for the following fiscal year around

Christmas. As indicated in Exhibit 15, we expect the

FY2012 issuance plan (which is also likely to be

announced before the end of 2011) to show an increase

of ¥300-400bn per month in calendar base market

issuance (supply at scheduled auctions from April

through March) relative to the offering amounts that will

be seen from December 2011 onwards. We would expect

this to break down into a ¥100bn/month increase in

issuance of 1Y, 2Y and 5T JGBs and a further

¥100bn/month increase in the 20Y sector (Exhibit 16).

Increases of this magnitude are unlikely to have a

significant impact on overall market sentiment, and we

would therefore consider any rise in JGB yields that

might be triggered by the announcement to be little more

than extremely short-lived noise.

Our key assumptions are as follows.

We currently expect the FY2012 issuance plan to show:

• No more than ¥44tn in new financial resource bonds

(thereby meeting the target set out in the government’s

medium-term fiscal framework)

We expect this to break down as follows:

• Around ¥150tn in calendar base market issuance

(offerings to regular auction participants)

• ¥4.2tn in Non-price Competitive Auction II issuance

• ¥2tn in adjustments by depleting front-loaded issuance

• Around ¥3.5tn in issues to individual (retail) investors,

which would be on a par with the third FY2011

supplementary budget

• Around ¥16tn in long-term JGBs held by the BoJ are

due for redemption during FY2012

Exhibit 16: … and we would expect this to break down into a¥100bn/month increase in issuance of either 2Y or 5TY JGBs anda further ¥100bn/month increase in the 20Y sector Calendar base JGB issuance in FY2012; JPYtn

Source: J.P. Morgan

Exhibit 15: FY2012 issuance plan (which is also likely to be announced before the end of 2011) to show an increase of ¥300-400bn per month incalendar base market issuance…JGB issuance plan (JPM estimate) in FY2012; JPYtn

FY2011 FY2012

3rd Suple Lowest Forecast Highest

Legal Grounds 181.7 173.9 ～ 176.1 ～ 179.2

New Resource JGBs 44.3 43.9 ～ 44.0 ～ 44.1 Government Target

Refunding JGBs 109.3 113.5 ～ 114.0 ～ 114.5 Around 114 trJPY

FILP JGBs 16.5 12.0 ～ 13.0 ～ 14.5 Redemption of FILP bonds is smaller than FY2011

Reconstruction JGBs 11.6 2.0 ～ 2.5 ～ 3.5 3.5trn at a maximum JGBs for govt pension 0.0 2.5 ～ 2.5 ～ 2.6 Temp issuance for Govt pens ion

FY2011 FY2012

3rd Suple Lowest Forecast Highest

Issuance Methods 181.7 176.1

Calendar Base 145.7 148.0 ～ 149.7 ～ 152.0 Calendar base will be around 150trn

Non-Price Comp 5.9 4.2 ～ 4.2 ～ 5.0 3.75% of total amount of 2Y + JGBs

Depleting Front Loaded 14.8 1.0 ～ 2.5 ～ 4.0

For Households 3.5 2.5 ～ 3.5 ～ 4.0 Same size as 3rd supplementary budget of FY11

Bank of Japan 11.8 15.0 ～ 16.2 ～ 18.0 Same size as the redemption in BoJ portfolio in FY12.

Comment

Comment

FY11 FY2012

Dec～ Amt Total

6mTB 0.9 0.9 1 0.9

1Y TB 2.5 2.5 12 30.0

2Y JGB 2.7 2.7 12 32.4

5Y JGB 2.5 2.6 12 31.2

10Y JGB 2.2 2.2 12 26.4

20Y JGB 1.1 1.2 12 14.4

30Y JGB 0.7 0.7 8 5.6

40Y JGB 0.4 0.4 4 1.6Liquidity 0.6 0.6 12 7.2

149.7

Calendar

Base




Takafumi YamawakiAC

(81-3) 6736-1748


168

Interest Rate Derivatives

Swap Curve

Over the past year, the swap curve bull-flattened up to

10Y sectors while the curve in the super-long sectors was

shifted down in a parallel-to-modest bull-steepening

fashion (Exhibit 17). The direction and magnitude were

generally in line with the movement of the JGB yield

curve. Even after markets recovered from the post-

earthquake turmoil, the BoJ continued to provide

abundant liquidity in response to the deepening debt

crisis in Europe and global economic slowdown. In

addition, the Fed introduced an explicit time frame of

interest rate policy until mid-2013 in August and then the

so-called Operation Twist in September. The BoJ is

expected to hold the policy rate at a low level even

longer than the Fed. The bull-flattening of the JPY swap

curve up to 10Y sectors this year is consistent with such

global low-for-long environment.

6M LIBOR effectively lays a floor to the swap rates up

to 5Y. Therefore, in a bullish environment, the level of

the shorter-term swap rate tends to be well supported

while longer end rates have more room for downshift

given a slow fall in 6M LIBOR. Therefore, in the course

of a downshift of the curve, short-end flatteners begin to

show convexity from a certain point. And butterflies,

after the belly richens as rates decline, begin to cheapenfrom a certain point. In fact, as the curve flattens further

out, the 2s/5s/10s fly did not richen in a rally this year.

On the other hand, the 5s/10s/20s spread has been on a

richening trend for about a year (Exhibit 18) and may

start to show convexity going forward. The long end of

the curve remained flat. The inversion in the long end of

the forward swap curve, which had first taken place after

Lehman shock in 2008, was not corrected either this

year, although the extent of inversion declined. The

10Yx10Y / 20Yx10Ycurve is still in negative territory at

-36bp (Exhibit 19).

The supply and demand balance in 30Y sector remainedone-sided in 2011 with strong receiving of the long end

of the curve. Firstly, declining FX rates and risky assets

put a flattening pressure on the long end of the curve.

While the issuance sum of structured notes remained

small, hedging activities for existing positions continued.

Secondly, asset swap investors stayed away from the

30Y sector. Domestic investors buy ASW up to 20Y.

Overseas investors, who had once invested in 30Y JGBs

as a LIBOR+ floating asset by using ASW, did not come

back to the market. Lastly, hedge funds sporadically paid

Exhibit 17: Swap curve bull-flattened up to 10Y sectors while thecurve in the super-long sectors was shifted down in a parallel tomodest bull-steepening fashion over the past year Change over the past year;

% bp

0.00

0.50

1.00

1.50

2.00

0 5 10 15 20 25 30

-30

-25

-20

-15

-10

-5

0

5Change (RHS)

Term structure (LHS)

Exhibit 18: As the curve flattens further out, the 2s/5s/10s fly did notrichen in a rally this year whereas the 5s/10s/20s has richenedconsistentlySwap 2s/5s/10s and 5s/10s/20s 50;50 butterfly since 2011; bp

-16

-14

-12

-10

-8

-6

-4

-2

0


-50

-45

-40

-35

-30

-25

-20

2s/5s/ 10s Butterfly (RHS)

5s/10s/20s Butterfly (LHS)

Exhibit 19: The inversion in the long end of the forward swap curveremained, though the extent of inversion declined. The10Yx10Y/20Yx10Y curve is at -36bp10Yx10Y / 20Yx10Y curve since 2011; bp

-60

-50

-40

-30

-20

-10

0





Takafumi YamawakiAC

(81-3) 6736-1748


169

super-long swap rate and bought long-dated payer

swaptions as a so-called ‘Japan fiscal premium trade’

with the anticipation of an increased Japanese fiscaldeficit leading to higher long-term yields. However, they

generally preferred benchmark 10Y tail to express the

view compared with 20Yx10Y for a liquidity reason

because the trade was essentially more macro-driven.

Going forward, for short-end rates, we expect them to

remain sticky under the BoJ’s interest rate policy as a

main scenario. At the same time, we also have in mind a

risk scenario that LIBOR rises on a possible worsening

of the crisis in Europe, which we think is getting to be

more of a reality. Short-term swap rates will likely rise if

LIBOR rises. We have a flat-to-upward bias on short-

term swap rates. On the other side of the curve, weexpect the flatness of the long-end swap curve to

continue into 2012. The crisis in Europe still has a long

way to go before it approaches a resolution. The crisis

puts a downward pressure on long-term swap rates via

two routes. The supply and demand balance is unlikely to

see a large change. Risk appetite of foreign (European)

investors is unlikely to recover enough for them to come

back to 30Y ASW in the near term. Also, a worsening of

the crisis would lead to a selloff in risky assets and a JPY

appreciation, which would flatten the long end of the

curve through dealers’ hedging activities. In sum, we

have a flattening bias on JPY swap curve as a whole.

With short-end rates held at a low level, we focus on

convexity trades such as weighted flatteners and

weighted butterflies. In addition, if the markets price in a

longer time frame of a low interest-rate policy, the effect

will likely spill over from shorter ends to longer ends, as

investors seek carry and move along the curve. In that

scenario, we expect a cheapening of the belly on shorter

dated butterflies such as 2s/5s/10s and then possibly

5s/10s/20s spreads as a medium-term trend. For example,

the 5s/10s/20s 50:50 fly has a convex profile vs. 5s/20s

curve, therefore a curve neutral 5s/10s/20s fly is expected

to show a convex P&L (Exhibit 20). In the long end,

long-dated steepeners are attractive in carry

considertations..

Swap spreads

Swap spreads tightened over the past year (Exhibit 21).

For one, we think this was related to a rise in fiscal

premium for JGBs. As mentioned earlier, the increase in

JGB issuance associated with the aftermath of the 11

March earthquake has continued to be in focus. On the

other hand, from a swap perspective, the swap curve

Exhibit 20: The 50:50 5s/10s/20s butterfly has historically showedpositive convexity to the 5s/20s swap curveSwap 5s/10s/20s butterfly vs. swap 5s/20s spread since 2010; bp

y = 0.02 x 2 - 3.52 x + 188.77

R2 = 0.46

-30

-25

-20

-15

-10

-5

0

5

100 110 120 130 140 150

5s/20s sw ap curv e; bp

Exhibit 21: Swap spreads have broadly tightened across the curvewith the exception of below 2Y over the past year Change of swap spreads over the past year; bp

-30

-20

-10

0

10

20

30

1 2 3 4 5 6 7 8 910 12 15 20 25 30

Maturity

-30

-20

-10

0

10

20

30

Change (RHS)

Term structure (LHS)

Exhibit 22: A PCA* indicates that swap spreads tended to move inthe same direction, but the move tended to be larger in the longer end. There was also another factor that impacted the 7Y sector and super-long sectors in the opposite directionJanuary 2011– November 2011; first two PCA factors, first factor explains 73%,second factor explains 15% unit

-0.60

-0.40

-0.20

0.00

0.20

0.40

0.60

2Y 5Y 7Y 10Y 20Y 30Y

1st factor

2nd factor

* The PCA is conducted on the daily changes in swap spreads across the curve.The data period is from January – November 2011.




Takafumi YamawakiAC

(81-3) 6736-1748


170

faced a bull-flattening pressure, led by super-long sectors

on the back of a selloff of risky assets and JPY

appreciation.

To get a better understanding of this year’s trend, we

conducted a PCA. Exhibit 22 shows the result of the

PCA analysis of the daily changes in swap spreads over

this year. The first and second PCA factors explain 73%

and 15% of the total variance, respectively. The result

indicates that swap spreads generally moved in the same

direction but the move was larger in the longer end. This

is in line with the above explanation that a rise in fiscal

premium caused this year’s swap spread changes. The

second factor, which is independent of the first one and

smaller in its impact, had an effect on the 7Y sector and

super-long sectors in the opposite direction. We can think of this factor as somehow related to ‘flight-to-quality’

movement with richening of 7Y on JGB futures buying

and cheapening of 30 on swap curve receiving.

In general, JPY swap spreads are largely affected by 1)

repo/LIBOR spread as a determinant of baseline values,

and 2) JGB yield levels. Since December 2010, the

repo/LIBOR spread has stayed within the 21-23bp range

(Exhibit 23), therefore had only a limited impact on

swap spreads. In fact, in a simple multiple regression

analysis over this year, the repo/LIBOR spread does not

have a meaningful explanatory power and JGB yield

levels, in turn, have higher t-values.

Partly mentioned in the Swap Curve section, we have a

flat-to-upward bias on LIBOR. From a monetary policy

perspective, we do not expect the BoJ to hike a rate

ahead of the Fed. Neither do we expect the BoJ to step in

to an IOER cut. However, a cut in IOER is still possible.

In this scenario, it is likely that with LIBOR falling faster

than repo rates, the repo/LIBOR spread will tighten. A

more likely risk scenario is that markets face stress and

LIBOR rises. In that scenario, LIBOR rises while repo

rates are kept low and the repo/LIBOR spread widens.

All in all, we have a widening bias on short-end swap

spreads. For the super-long sectors, we think the

cheapness of JGB vs. swap rates will remain into

2012, given our view that the flatness in the long end of

the JPY swap curve will continue.

From a relative value and shorter-term perspective, swap

spread curve trades are worth considering. One way to

see such relative values is to run simple multiple

regression. For example, Exhibit 24 shows the residuals

of swap spreads from the regression analysis where

swap spreads are regressed on underlying JGB yields and

1Y LIBOR/OIS basis (which we think better reflects

day-to-day moves than the repo/LIBOR spread) as

explanatory variables. It indicates that front end spreads

are too wide while long end spreads are too narrow. We

would use this model to assess the relative

richness/cheapness on the swap curve, however given

our view of swap curve flattening we recommend

investors not to chase this relative value.

Exhibit 23: Since December 2010, repo/LIBOR spreads havestayed within the 21-23bp range, therefore had only a limitedimpact on swap spreads.

Spread between the S/N repo rate and 6M JPY LIBOR since 2010; bp

20

22

24

26

28

30

32

34

36

38


Source: Bloomberg, J.P. Morgan

Exhibit 24: Our JGB swap spreads model on underlying JGByields and 1Y LIBOR/OIS basis* indicates that front end spreadsare too wide while long end spreads are too narrow. However, wewould not fade this relative value.Residuals of swap spreads as of 17 November 2011; bp

-6

-4

-2

0

2

4

6

8

2 3 4 5 6 7 8 910 12 15 20 25 30

Maturity

JGB cheap

JGB rich

* The swap spreads are regressed on the underlying JGB yields and 1YLIBOR/OIS basis. The data period is from January – November 2011.The regression equation for 10Y swap spread is, for example,10Y swap spreads = 8.9 – 19.0 x 10Y JGB Yield - 0.48 x 1Y LIBOR / OIS




Takafumi YamawakiAC

(81-3) 6736-1748


171

Cross-Currency Basis Swap

USD/JPY cross-currency basis spreads gradually

narrowed in 1H11. However, they widened back (moved

–ve) in 2H11. In addition, we had two sharp widening in

August and November (Exhibit 25).

A medium-to-long-term driver is dollar funding pressure.

Concerns about European banks have heightened. Some

banks that found it difficult to raise dollar funding via

standard funding markets probably headed for cross-

currency basis markets. USD/EUR cross-currency basis

spreads widened, and USD/JPY cross-currency spreads

followed. On a flow side, coupon swap (forward USD

buying) flows from Japanese investors, which had once

counteracted a widening of the spreads in the past, were

substantially reduced as the USD/JPY FX rate continued

to break record lows this year. Meanwhile, periodic

samurai issuance put pressure in a negative direction to

the spreads.

Short-term and sharp widening in August and November

is explained by the FX intervention by the MoF and BoJ,

which created a significant number of short USD

positions in markets. Our FX team estimates the amount

of intervention to be around ¥4.5tn in August and ¥7.4tn

in October. Some market participants covered those

positions in the cross-currency basis market as well as

FX markets. August, in particular, saw some investors

spread tightening positions as a carry trade. The FX

intervention triggered stop-loss activities on such

positions, which accelerated further widening. In fact, the

magnitude of widening was milder over October –

November. Probably a large part of speculative positions

had been already cleared after the intervention in August.

It is possible to estimate the break-even level of short-

term FX basis spreads. If a bank funds JPY cash at

overnight rate and swaps it into USD by FX OIS basis

swap, it effectively funds USD cash at USD OIS + X

(Exhibit 26, left). Moreover, short-term FX LIBOR

basis should move interactively with 3M LIBOR/OIS

basis in each currency and FX OIS basis (Exhibit 26,

right). Even if the overnight call market is not available,

banks can resort to the BoJ’s liquidity providing

operation which lends JPY cash at 0.1% p.a. for 3M or

6M terms in exchange for certain eligible collaterals.

The cost involved in the above method should

theoretically not exceed the cost via other dollar funding

means. Currently, the BoJ’s dollar lending facility is in

place. It lends unlimited USD at approximately USD

overnight rate + 100bps in exchange for certain eligible

collaterals with the additional haircut of about 12.5% for

equal to or less than the 1M term and 20% for more than

1M and equal to or less than the 3M term. Taking the

cost of the haircut into account, the effective cost of USD

funding via the BoJ’s dollar facility is higher than USD

FF + 100bps. Therefore, FX OIS USD/JPY basis spreads

Exhibit 25: USD/JPY FX basis spreads gradually narrowed in 1H11.However, they widened back (moved –ve) in 2H11Cross currency basis spreads since 2011; bp

-100

-90

-80

-70

-60

-50

-40

-30

-20

-10

0


1Y FX LIBOR basis

(bp)

3M FX OIS basis


Exhibit 26: (1) If a bank funds JPY cash at overnight rate and swapsit into USD by FX OIS basis swap, it effectively funds USD cash atUSD OIS + X. (2) Short-term FX LIBOR basis should moveinteractively with 3M LIBOR/OIS basis in each currency and FX OISbasis. The cost involved in the above method should theoreticallynot exceed the cost via other dollar funding means…How to calculate the break-even level of short-term FX basis spreads

Funding JPY from

overnight marketJPY O/N

Funding JPY from overnight

marketJPY O/N

Fixing via JPY OIS JPY OISConverting it to JPY LIBOR

via JPY LIBOR/OIS basis

- O/N +

3M JPY LIBOR- Y

Swapping JPY cash

into USD by FX OIS

basis

- ( JPY OIS – X)

+ USD OIS

Swapping JPY cash into

USD by FX LIBOR basis

- (3M JPY LIBOR - Z)

+ 3M USD LIBOR

Convert it to USD OIS via

USD LIBOR/OIS basis

- 3M USD LIBOR

+ USD OIS + W

Net USD OIS + X Net USD OIS - Y + Z + w

(1) (2)

Exhibit 27: …but 1Y FX OIS spread at the right is already below –100bp. The implied FX OIS USDJPY spreads exceed those seen inthe actual FX OIS USD/JPY basis market.Breakeven USD funding cost; bp

JPY 3M

LIBOR/OIS

basis (Y)

FX LIBOR

USD/JPY

basis (Z)

USD 3M

LIBOR/FF

basis (W)

USD OIS +

(-Y+Z+W)

JPY OIS +

(Implied

FX OIS)

FX OIS

USDJPY

basis

1Y 16.9 66.0 49.3 109 -109 -110

2Y 17.0 75.5 58.9 120 -119 -120

3Y 16.8 82.3 66.2 124 -123 -125 Source: Bloomberg, J.P.Morgan




Takafumi YamawakiAC

(81-3) 6736-1748


172

can go under -100. But the divergence should not be

large, assuming that the two OIS rates are equal to the

overnight rate in each currency and constant. In addition,if the spread charged on the dollar lending facility by the

central banks is cut from the current 100bps, FX OIS

USD/JPY basis spreads are expected to move in a

positive direction accordingly,

Exhibit 27 shows the current level of LIBOR/OIS basis

spreads in each currency, FX LIBOR USD/JPY basis

spreads and the FX OIS spreads implied from those

numbers. 1Y FX OIS spread at the right is already below

–100bp. Further, the implied FX OIS USDJPY basis

spreads exceed those seen in the actual FX OIS USDJPY

basis market. We have to be reminded that it starts from

the assumption that a bank funds JPY at the overnightrate. In addition, it depends on many other variables such

as how long the BoJ’s lending facilities (in JPY and

USD) would be available, how much the collateral and

additional haircut actually costs, and bid-offer. Lastly,

there could be the possibility that markets overshoot at

extreme stress, irrespective of these fundamentals.

There are some implications from large negative

USD/JPY cross-currency basis spreads. First, large

negative basis spreads make it attractive for a USD

investor to create synthetic USD bonds by buying JGBs

and swapping them into USD with a cross-currency basis

swap. The level of yield pick-up appears to at ahistorically attractive level (Exhibit 28). This applies

even when compared with the difference of CDSs of the

two countries (Exhibit 29). Although we need to bear in

mind that CDSs of JGBs and USTs are mainly traded in

USD and EUR, respectively, the comparison still gives

us some guidance on the level of the pickups.

Secondly, the return of FX-hedged foreign bonds has

become less attractive (Exhibit 30). Some investors,

when investing in foreign bonds, hedge their FX risk.

They often use rolling 3M FX forward. As a result of

sharp widening in the 3M term FX OIS basis spread, the

cost of 3M FX hedge has become higher. If short-term

basis spreads are stabilised at a large negative level,

Japanese investors may accelerate the repatriation of

USD funds back to JGBs. The focus is on life insurance

companies because these investors would increase

investment into super-long sectors with extra JPY cash.

The return of FX-hedged 10-year USTs is now even

lower than in August, when we saw ¥550bn net selling of

medium- and long-term foreign bonds from life

insurance companies, according to MoF data.

Exhibit 28: The level of yield pick-up appears to be at a historicallyattractive level on 1YPick-up in yield of USD-asset-swapped 1Y JGB over 1Y USTs since 2010; bp

0

20

40

60

80

100

120



Exhibit 29: The level of yield pick-up appears to be attractive evenwhen compared with the difference of CDSs of the two countriesPick-up in yield of USD-asset-swapped JGBs over USTs; bp

JGB JGB ASWJPY 3s/6s

Basis

FX LIBOR

USD/JPY

basis

3M USD

LIBOR +

Pickup vs

UST

JGB/UST

CDS spread

1Y -25.5 13.5 -60.0 48 107 8

2Y -23.9 14.0 -70.0 61 110 23

3Y -20.6 14.1 -78.0 73 119 31

* For Exhibits 27&29, all the numbers are based on mid levels. For relevant basisspreads in each currency, cross-currency basis spreads, ASWs and CDSs, weused the nearest benchmark points. We calculated these numbers, with theassumption thatUSD/JPY conversion factors are the ratio of modified durations for simplification.

Exhibit 30: The return of FX-hedged foreign bonds has become lessattractive with higher FX hedge cost, which could lead to further demand to the super-long sectors from life insurance companies.20Y JGB yields and 10Y UST yields with FX hedge since 2010; %

0.50

1.00

1.50

2.00

2.50

3.00

3.50


10Y UST hedged by 3M roll

20Y JGB

FX hedge cost is calculated as 3-month forward point divided by spot FX rateSource: Bloomberg, J.P. Morgan




Takafumi YamawakiAC

(81-3) 6736-1748


173

Volatility

The implied volatility surface as a whole was shifted

downward with a steepening bias over the past year

(Exhibit 31). Implied volatilities soared last December,

as QE2 pushed up 10Y UST yields to above 3.50%.

JGBs were sharply sold off and 10Y JGB yields moved

higher from 0.85% in October to 1.3% in December. The

selloff also triggered unwinding of short volatility

positions, and volatilities spiked. Implied volatilities rose

sharply soon after the 11 March earthquake. The BoJ

provided abundant liquidity to ease the post-earthquake

turmoil and maintained liquidity at a high level. Interest

rates gradually declined thereafter. The 10Y swap rate

and 10Yx10Y forward swap rate moved lower to 0.96%

and 2.43%, respectively, from 1.21% and 2.67% at the beginning of this year. Steadily declining yields also

lowered realised volatilities and then implied volatilities

(Exhibit 32).

From an investor flow perspective, first, the fall in the

surface was led by constant selling flows in gamma

sectors such as target buying and covered call from

domestic investors. Second, issuance volume of

structured notes remained small, but domestic banks kept

selling callable structured deposit to mass retail. Banks

generally sell options to hedge long vega positions. Their

selling tends to take place in the middle of the surface as

structured deposits often have a call option after 1–3

years and a maturity of 5–10 years. Such volatility

selling flows are another factor for cheapening of the

vega sector. On the other hand, there were sporadic

buying flows in longer tail vega sectors from

macro/leveraged investors. They bet on a long-term

fiscal deficit deterioration scenario of the Japanese

government. These flows pushed up and supported

longer ends of the surface.

Into 2012, we think that the current shape of the surface

is likely to remain. We are neutral on gamma. Under a

low-for-long regime, short-term interest rates will likely

remain low and move within a narrow range and

volatility selling flows are expected to continue. On theother hand, we cannot discount the possibility of a rise in

gamma sector volatilities under a stressed scenario, given

the situation in Europe. In long-tail vega sectors, we

expect volatilities to remain relatively high. With

absolutely low interest rates and volatilities in the JPY

market, we do not expect selling demand to outweigh

buying demand in these sectors. Rather, since an

issuance increase of JGBs continues to be a major topic

of debate, concerns about the long-term fiscal deficit of

Japan are likely to be priced in and out from time to time.

Exhibit 31: The implied volatility surface as a whole was shifteddownward with a steepening bias over the past year.Implied volatility changes over the past year; bp/day

-1.8

-1.6

-1.4

-1.2

-1.0

-0.8

-0.6

-0.4

-0.2

0.0

3M 1Y 2Y 5Y 10Y

1Y tail 2Y tail 5Y tail 10Y tail 20Y tail

Exhibit 32: Implied volatilities soared last December as QE2pushed up 10Y UST yields to above 3.50% They also rose sharplysoon after the earthquake in March before declining later in theyear.1Yx10Y normal implied volatility since 2010; bp/day

2.0

2.5

3.0

3.5

4.0

4.5

5.0


Earthquake

UST Yield Rise

after QE2

Exhibit 33: TIBOR remained sticky to fall relative to CP rateTIBOR 3M and a-1 rated CP rate; %

0.00

0.10

0.20

0.30

0.40

0.50

0.60

Nov -09 May -10 Nov -10 May -11 Nov -11

CP rate a-1

3M D Tibor

Source: Bloomberg, JASDEC, J.P. Morgan




Takafumi YamawakiAC

(81-3) 6736-1748


174

TIBOR/LIBOR Spread

There are three categories of TIBOR-LIBOR spreads: 1)

spot fixing TIBOR-LIBOR spread; 2) IMM 3M TIBOR-

LIBOR spreads whose liquidity runs up to the 8th

contract; and 3) Term TIBOR/LIBOR spreads which are

the difference between swap rates usually against TIBOR

6M and LIBOR 6M, but for short maturities also against

TIBOR 1M and LIBOR 1M.

TIBOR fixing remained sticky to fall while CP rates

were stable at a low level (Exhibit 33). We think the

reason for the stickiness has remained unchanged over

the past few years, i.e. market practice is that TIBOR is

used as a reference rate to corporate lending. As the

credit spread charged to domestic companies continuedto be tight (Exhibit 34), TIBOR tends to be maintained

at a high level relative to the policy rate. As a result, the

3M TIBOR/LIBOR fixing spread remains positive.

There were two periods in the past when TIBOR/LIBOR

fixing spread was negative (Exhibit 35). One was just

after the end of the BoJ’s quantitative easing in 2006

when LIBOR priced in the future rate hike earlier, and

the other was after the subprime shock in summer 2007

as LIBOR jumped. We think that the former case is not

impossible, but very unlikely. Since the Fed introduced

an explicit time frame of interest-rate policy until mid-

2013 in August, the BoJ is expected to continue thecurrent interest-rate policy even longer than the Fed. The

latter case is more likely given the current situation in

Europe. But even if it happens, the magnitude may be

milder as central banks are committed to avoiding a

liquidity crisis.

With regards to term TIBOR-LIBOR, we think that

negative spreads are unlikely unless the spot fixing

TIBOR-LIBOR spread becomes negative. Short-term

spreads are generally driven by the spot fixing spread.

For longer tenors, the spreads are largely driven by

domestic banks’ hedging activities. They use TIBOR as

their internal pricing and TIBOR swaps fit the purpose.However, as LIBOR swaps are more liquid, they often

use LIBOR swap and hedge the TIBOR/LIBOR risk

from time to time. Their choice of hedge timing is the

source of term TIBOR-LIBOR fluctuation. However,

term TIBOR-LIBOR spreads generally moved in a very

tight range with limited flows and a sticky spot-fixing

spread this year. (Exhibit 36)

Looking ahead, we have a tightening bias on the

TIBOR-LIBOR fixing spread. Historically, LIBOR

Exhibit 34: As the credit spread charged to domestic companiescontinued to be tight, TIBOR tends to be maintained at a high levelrelative to the policy rate.

Average rate charged to corporate in bank lending; %

1.00

1.20

1.40

1.60

1.80

2.00

2004 2005 2006 2007 2008 2009 2010 2011

Source: BoJ, J.P. Morgan

Exhibit 35: LIBOR fixing was higher than TIBOR fixing at the endof QE in 2006 and after the subprime shock in 20073M TIBOR and LIBOR fixing; %

0.00

0.20

0.40

0.60

0.80

1.00

1.20

2005 2006 2007 2008 2009 2010 2011

3M LIBOR

3M Euroyen T IBOR

Source: Bloomberg

Exhibit 36: Term TIBOR-LIBOR spreads generally moved in a verytight range with limited flows and sticky spot-fixing spread thisyear 5Y term TIBOR/LIBOR spread and spot-fixing spread; bp

-4

-2

0

2

4

6

8

10

12


5Y Term T/L

TIBOR-LIBOR fix ing spread

Source: Bloomberg




Takafumi YamawakiAC

(81-3) 6736-1748


175

tended to move faster and be more volatile both in a rise

and in a fall. Therefore, the TIBOR-LIBOR fixing spread

has had a tendency to narrow in a rise and to widen in afall. As mentioned in the previous sections, there are

some risk scenarios. One is a fall of LIBOR on a cut in

IOER. Another scenario is that LIBOR rises under

extreme stress, which we think is more likely. Besides,

the level of LIBOR is absolutely low and lower than

TIBOR. The magnitude of the spread move is likely to be

asymmetric, i.e. we expect more tightening if LIBOR

increases than widening if LIBOR falls.

3s/6s spread

The Term LIBOR 3s/6s spread curve steepened this year

(Exhibit 37). The spreads moved in a parallel fashionuntil 2010. However, although the spreads widened, the

widening was larger at the longer ends. Widening can be

explained by one-sided demand in the 3s/6s market. With

relatively tight credit spreads in Japan, samurai issuance

continued in 2011. When a non-resident entity issues

JPY bonds (e.g. EMTN or samurai) and swaps into USD,

it enters a paying position of 3s/6s basis swap (pay 3M

vs. rec 6M) as well as a receiving position in a cross-

currency swap (Exhibit 38), driving 3s/6s basis higher

and FX basis wider (more –ve). Such flows tended to

take place in 5–10Y sectors. Conversely, if a non-

resident investor buys a JPY bond with a cross-currency

swap as a synthetic USD bond, it has the opposite impacton the spreads. However, there was little such demand

this year. Rather, those investors who had such positions,

particularly in the super-long sectors, might have

unwound them for the purpose of reducing balance

sheets. This could be a reason for the widening on

longer term 3s/6s spreads.

We think the short-to medium-term spreads are

biased wider. Tight credit spreads in Japan are still

attractive for non-resident entities to raise funding, even

after large negative cross-currency basis spreads and

extra premium for non-resident entities are considered.

Issuance by non-residents is expected to continue in 2012and such flows would be a widening factor in the

medium-term spreads. There is also the possibility that

LIBOR will rise if the crisis in Europe deepens. If the

scenario realises, the spot fixing 3s/6s spread will widen

as it had amid the Lehman shock and lead to a widening

of term 3s/6s spreads at the short ends.

Exhibit 37: Term LIBOR 3s/6s spread curve widened and steepenedthis year Spot fixing 3s/6s and 5Y and 20Y term 3s/6s basis; bp

10

15

20

25

30


Spot fixing 3s/6s LIBOR

5Y term 3s/6s

20Y term 3s/6s


Exhibit 38: Flows of non-resident entity issuing JPY bonds andswapping into USD drive 3s/6s basis higher and FX basis wider (more –ve)How 3s/6s is involved in swapping JPY into USD

Issuer Samurai/EMTN

JPY 3M L + 3s/6s basisJPY 6M

USD 3M L

JPY 3M L +FX LIBOR basis

JPY Fixed

JPY FixedJPY 6M

3s / 6s

marktet

Swap marktet

FX basis

marktet




Jorge GarayoAC

(44-20) 7325-4820


176

Inflation-linked Markets

• 2011: Linkers underperformed as nominal yields

declined substantially on the back of an

escalation of the sovereign risk crisis in the Euro

area and a deteriorating global macro outlook

• 2012 outlook for inflation: Substantial falls in

headline and core inflation across the board

• We believe that the market will focus more on

downside risks from a deeper recession than on

medium-term inflation, as inflation expectations

remain anchored

• German and UK 10Y real yields to test new lowsand fall well below zero in Europe but expected

to stay close to 0% in 10Y TIPS…

• …while breakevens should be biased lower as the

sovereign crisis continues in 1H12

• Expect cash breakevens to underperform relative

to inflation swaps in early 2012

• Euro area: Position for flatter breakeven curve,

as inflation should be sticky and longer-dated

breakevens remain highly directional

• UK strategy: Be long intermediate real yields,

and short inflation breakevens in the 10Y sector

• UK tactic: Go short UK real yields into linker

syndications as a tactical trading strategy

• US: Fair-value model suggests falls in inflation

and nominal yields, combined with fiscal

tightening, should cause breakevens to narrow

into early 2012

• Beyond 1Q12 longer maturity TIPS breakevens

should widen as nominal rates rise

• Supply: UK and US to see modest increases in

gross supply, whereas Euro area issuance should

be below average, with Italy being the mainquestion mark

2011: Linkers outpaced by nominals in a

relatively high inflation environment

2011 was dominated by large declines in nominal yields,

as the European sovereign debt crisis accelerated from

the summer. Throughout the year, market-implied

inflation breakevens were initially driven by a spurt of

Exhibit 1: Inflation breakevens were supported by energy price inflationin 1Q11 but started to decline in the summer Inflation breakevens for 10Y benchmarks in US TIPS, Germany and the UK; %

1.00

1.50

2.00

2.50

3.00

3.50

Nov 10 Feb11 May 11 Aug11 Nov 11

UK Germany US

Exhibit 2: Headline inflation rose steadily over the past year, withinflation in 1Q11 being lifted by energy prices…Inflation measures, in oya terms; %

-4.00

-2.00

0.00

2.00

4.00

6.00

Nov -06 Nov -07 Nov -08 Nov -09 Nov -10 Nov -11

Euro HICP US CPI UK RPI

Exhibit 3: …and core inflation also moving up significantly over theperiodCore inflation measures, in oya terms; %

0.00

1.00

2.00

3.00

4.00

Nov 06 Nov 07 Nov 08 Nov 09 Nov 10 Nov 11

HICP core US CPI core UK CPI core




Jorge GarayoAC

(44-20) 7325-4820


177

energy price inflation in 1Q11, but they entered a clear

downward phase from May onwards (Exhibit 1). From

this point, inflation breakevens were driven lower by a prevailing risk-off environment and worries of a

worldwide recession as data printed particularly weak

globally.

The evolution of inflation prints would suggest a

positive return year for inflation-linked markets.

Headline inflation rose steadily since November 2010,

peaking at the end of the year (Exhibit 2). Even though

energy inflation played a significant part in 1Q11, core

inflation also rose throughout the year (Exhibit 3). This

was contrary to our January predictions that a high

degree of slack would push core inflation lower (with the

caveat that the move in UK core CPI was heavilyimpacted by VAT effects).

However, inflation prints were not the main driver

and inflation expectations were pushed lower by the

macro landscape. The downgrading of growth

expectations in the Euro area and globally from the

summer was likely exacerbated by sovereign risk

dynamics in the periphery. Commodities reversed their

upward trend, and equity markets fell over 15% in the

space of two weeks in August. Central banks recognised

a shift in the macro fundamentals and risks around them,

dismissing the contemporaneous ‘relatively high’

inflation prints. The ECB eased rates in November after having tightened policy in April and July, whereas the

Fed committed to keeping rates unchanged up to mid-

2013 and initiated ‘Operation Twist’. The BoE

announced another £75bn of QE purchases and is

expected to announce another round of QE in the coming

months.

Survey-based measures of medium-term inflation

expectations remained well anchored throughout the

year despite the significant deterioration in the macro

outlook . Inflation expectations from 2013 onwards were

broadly unchanged according to economists’ consensus

forecasts (Exhibit 4), supporting the message of

anchored inflation expectations that central banks have

been highlighting thus far.

So how did inflation-linked portfolios fare relative to

conventional government bonds? Their returns

(adjusted for duration differences) were negative in the

Euro area and the UK and just about matched those

of Treasuries in the US (Exhibit 5). In other words, a

strategy of being long inflation breakevens across the

curve would have given investors a return of -3.3% and

Exhibit 4: Medium-term inflation expectations are well anchoredaccording to economists’ surveys, with the Euro area and the US likelyclose to target and UK CPI expected to be 0.8% above target

Long term inflation forecasts from economists*; %

Oct10 Oct11 C hg Oct10 Oct11 C hg Oct10 Oct11 Chg

2011 1.9 3.1 1.2 2.6 4.4 1.8 1.6 2.6 1.0

2012 2.2 2.1 -0.1 2.1 2.7 0.6 1.5 1.8 0.3

2013 2.1 2.1 0.0 2.6 2.5 -0.1 1.7 1.8 0.1

2014 2.2 2.3 0.1 3.0 2.7 -0.3 1.9 1.9 0.0

2015 2.2 2.3 0.1 2.6 2.6 0.0 2.0 2.0 0.0

2016 2.3 2.2 -0.1 2.7 2.8 0.1 2.1 2.0 -0.1

2017-21 2.3 2.2 -0.1 2.7 2.8 0.1 2.1 2.1 0.0

US CPI UK CPI Euro HICP

* Source: Consensus Economics, October 2011 long-term forecasts.

Exhibit 5: Despite the upward trend in inflation in 2011, buyingEuropean linkers vs. conventional bonds would have beenunprofitable, with Italian linkers being the main drag for the Euro area;the same strategy in the US would have posted returns close to zeroTotal return between 15 November 2010 and 18 November 2011; %

Total return

Euro linkers (JPM ELSI) -10.7%

Euro gov comparator* portfolio -5.8%

Long inflation B/E (JPM EBEX) -3.3%

Euro linkers ex Italy ex Greece -2.7%

Italian linkers -23.5%

US TIPS (JPM JUSTINE) 12.5%

USTs comparator* portfolio 10.4%

Long inflation B/E (JPM UBEX) 0.0%UK Linkers (FTSE) 18.7%

Gilts 7Y+ (JPM GBI) 20.4%

Long inflation B/E (est.**) -5.7%

* Comparator portfolios are based on the same weights used by linker indices for comparable nominal bonds, adjusting for the relative duration.** Calculated by replicating a strategy of being long FTSE All-Share index linkedagainst being short the JPM Gilts 7Y+ index in a duration-adjusted amount, re-weighted weekly.

Exhibit 6: The deterioration in liquidity was an important driver of performance. Our estimated bid-offers widened to levels above thoseseen in 4Q08 for Italian linkers, and to similar levels in French linkerswhereas other linker markets and inflation swaps bid-offers held inremarkably wellIndicative bid/offers across different inflation products*; bp of yield

Assumptions

Prior 4Q 08 1H 11 2H 11

Italy linkers 2 15 2.5 20 €50mn size

France linkers 2 15 2.5 15 €25-50mn size

German linkers 2 15 2.5 10 €25-50mn size

Euro HICP sw aps 2 15 2 4 €50mn size

UK linkers 2 7 2 2 £10mn size

UK RPI sw aps 4 18 4 4 £25mn size

US TIPS 2.5 7.5 3 4 $50mn size

US CPI sw aps 5 15 5 5 $25-50mn size

2008 2011

* Source: J.P. Morgan




Jorge GarayoAC

(44-20) 7325-4820


178

-5.7% in the Euro area and the UK, respectively, and a

flat return in the US.

As in 2008, the deterioration in linkers liquidity was

an important driver of performance, particularly in

the Euro area. Bid/offers reached similar or larger

dislocations than those seen in 2008. In particular

Italian BTPeis bid/offers widened on average 20bp vs.

15bp at their worst point in 2008. But even bid-offers for

French linkers widened to levels reached in late 2008

(Exhibit 6). Supply in Euro area linkers was heavily

affected by the sovereign debt dynamics. Germany

cancelled its linker programme in 2Q11 and 3Q11, not

tapping the market between April and November. Italy’s

net issuance of linkers in 2H11 was just around €1.5bn

(vs. usually €2–4bn per quarter). Unstable marketconditions resulted in reduced auctions, and the Treasury

conducted buybacks of heavily dislocated lines. It is

interesting to note that US and UK linkers, as well as

inflation swaps, held up much better in terms of bid-

offers, despite some evident deterioration in liquidity

dynamics.

Inflation outlook for 2012

The inflation outlook for 2012 can be characterised

by an expected decline in both core and headline

inflation (Exhibit 7). In terms of headline inflation, we

expect more significant declines, in great part because of energy and food base effects from substantial price

increases in the past year. In sequential terms (seasonally

adjusted q/q changes in CPI indices), we have already

seen a clear declining trend in Euro area headline

inflation, and we expect significant falls in the UK and

the US in the coming quarters (Exhibit 8). Euro area

inflation should drop to an average of 1.9% oya during

2012, in line with the ECB’s target. In the UK, we expect

headline inflation to fall sharply but average 2.5% (above

the BoE’s target of 2%) whereas we see US inflation

averaging 1.4%.

The assumptions behind our forecasts are as follows:

• Stable oil prices, assuming Brent around

$110/bbl, most likely stuck in a $100–115/bbl

range. We see OPEC supporting prices above

$100 (barring a very significant recession,

which could push it below that figure), while

marginally supportive fundamentals could push

it to $115/bbl.

Exhibit 7: J.P. Morgan 2012 inflation forecasts: core measures todecline across the board; headline inflation to fall but average ‘abovetarget’ in the UK, ‘around target’ in the Euro area, and ‘well below

target’ in the USUS= US CPI-U nsa and core CPI, Euro = Euro HICP all items, core HICP (excludingall food and energy), UK= UK RPI, CPI and core CPI; oya%

CPI Core HICP Core RPI CPI Core

Oct 11 3.5 2.1 3.0 1.6 5.4 5.0 3.4

Nov 11 3.5 2.1 3.0 1.6 5.2 4.8 3.2

Dec 11 3.1 2.2 2.8 1.7 4.9 4.2 3.0

Jan 12 2.7 2.1 2.5 1.7 4.4 3.6 2.6

Feb 12 2.4 1.9 2.4 1.7 3.7 3.2 2.1

Mar 12 1.6 1.9 2.0 1.6 3.5 3.2 2.1

Apr 12 1.2 1.7 1.8 1.5 3.0 2.5 1.5

May 12 1.0 1.5 1.9 1.6 2.9 2.5 1.6Jun 12 1.3 1.3 2.0 1.5 3.1 2.7 2.0

Jul 12 1.2 1.1 1.9 1.6 3.2 2.5 1.9

Aug 12 1.0 1.0 1.9 1.5 2.9 2.3 1.6

Sep 12 0.9 1.1 1.7 1.4 2.6 1.9 1.6

Oct 12 1.1 1.1 1.5 1.4 2.7 2.0 1.6

Nov 12 1.2 1.1 1.5 1.5 2.6 1.9 1.5

Dec 12 1.3 1.2 1.5 1.5 2.6 1.9 1.5

Av e 2012 1.4 1.4 1.9 1.5 3.1 2.5 1.8

US Euro UK

Exhibit 8: Euro area sequential inflation has already initiated adownward trend and should stabilise around 1%; we expect significantfalls in the US and the UK in the next few quartersEuro area HICP inflation Q/Q saar, US CPI inflation Q/Q saar and UK CPI Q/Q saar,actual and J.P. Morgan forecasts; %

-2.00

-1.000.00

1.00

2.00

3.00

4.00

5.00

6.00

7.00

Mar 09 Sep 09 Mar 10 Sep 10 Mar 11 Sep 11 Mar 12 Sep 12

Forecasts

Euro area

UK

US




Jorge GarayoAC

(44-20) 7325-4820


179

• Stable currencies in the major crosses, in line

with their recent behaviour this year

• 50% chance of a technical recession in the UK

(base case is one-quarter of GDP contraction,

annualised 0.5% growth in 2012), a moderate

recession in the Euro area (peak-to-trough

contraction in GDP just over 1%) and the US

growing at sub-par pace of 1.8% in 2012.

One of the main forces that poses significant

downside risks to inflation is the prospect of a more

severe and prolonged recession in the Euro area,

which could significantly impact the region and the rest

of the world. A global spillover could potentially bring

oil prices well below the $100/bbl level and pushheadline inflation two to three tenths lower, purely

through energy components, as well as impact long-term

inflation expectations. This is not our central scenario,

however, and there are also upside risks if core

inflation remains sticky on the way down if we

misjudge the slack in the economy. As always, the

behaviour of currencies could also have a significant

impact on imported inflation. In the Euro area, ongoing

fiscal retrenchment in key countries may prove

inflationary over the near term, as governments are

likely to increase taxes for the consumer (in the form of

VAT, administered prices, and other duties). Our forecast

does not incorporate any of these factors, and thereforewe should be conscious of them.

We show a summary of some of these factors’

sensitivities in Exhibit 9. We also discuss some of these

risks in some detail in the individual market sections.

Are we approaching deflation? Not quite

The quick deterioration in the macro data and a gloomy

outlook for coming years has raised questions about the

possibility of entering a very low inflation world. In our

view, a zero-inflation world would come with a

significant decline in medium term inflation

expectations. We are far from that, as inflation

expectations have remained firmly anchored in the

three main markets. The prospect of a global recession is

pushing expectations lower in the Euro area, but not to a

worrying degree.

Exhibit 9: Our inflation view is sensitive to assumptions aboutvolatile elements in baskets and model risk for core inflation; we seedownside risks coming from a deeper recession (lower oil) and

upside risks from tax increases (VAT) or currency depreciationImpact on headline inflation; % points

10%

move in

oil price^

10%

move in

food

prices^

10% drop

in

currency*^

Core

models

error

State

controlled

prices / VAT

increase**

Euro HICP 0.10 0.15 0.25 +/- 0.7 0.25

UK RPI 0.20 0.10 0.65 +/- 0.7

US CPI 0.50 0.15 0.25 +/- 0.5

J.P.

Morgan

view

$100-120

range -Downside

in a

deeper

recession

Near termdownside

to food

prices

Central

view of

stable

currencies -

EUR at risk

Risk of

deeper recession

/

Misjudge

slack

Most likely

increases in

Euro area

(-) (-) (+) (-) / (+) (+)

* Based on drop in Trade Weighted currency.** Euro area based on combined 1% moves in Italy, Spain and France, assuming a65% pass through on 65% of baskets, for more granular information see Exhibit 26.^ For simplicity we assume broadly equal betas on the downside and upside.(-) Downward impact on inflation (+) Upward impact on inflation.

Exhibit 10: Investors expect central banks to do a good job atavoiding deflation; the Fed and the BoE are seen as giving significant

weight to growth, while the ECB is more inflation-oriented butremains the most credible in terms of keeping inflation close totargetInvestor perception of weights (median response) given to growth and inflation for each central bank*; %

80

35

70

20

65

30

0%

20%

40%

60%

80%

100%

US Euro area UK

Growth / employment Inflation

* See: “J.P. Morgan inflation expectations survey, November 2011” by Garayo &Chordia




Jorge GarayoAC

(44-20) 7325-4820


180

With central banks watching these carefully, there is no

reason to believe in a changing inflation environment,

which would only be consistent with the world economy

sinking into a prolonged depression (not our current

expectation). As we showed in our J.P. Morgan

November Inflation Expectations Survey (11 November

2011), an overwhelming 45% of investors expect

inflation to be close to target in the Euro area over the

medium term. In the case of the BoE and the Fed, they

have both been extremely vocal about the risks of low

inflation and continue to be seen as giving a lot of weight

to growth at the current cross roads (Exhibit 10).

Inflation expectations 1 year out remain remarkably high

according to our latest Inflation Expectations survey

(November 2011), even if there has been a shift lower inthe Euro area (Exhibit 11). Our own expectation is that

Euro area inflation will average 1.9% in 2012, which is

likely to keep expectations relatively high. Furthermore,

the risks of low inflation over the medium term seem to

be pretty much counterbalanced by perceived risks of

very high inflation. This is most clear in the US and the

UK, where investors attach a particularly high

probability of inflation being above 4% in 2 out of the

next 5 years (Exhibit 12). Admittedly, investors’

perceived probability of a deflation in the Euro area has

increased recently (to 15%), but this is counterbalanced

by an equal probability of very high inflation. Looking at

inflation volatility markets, the implied probability of

deflation in 2 or 5 years’ time is not significantlydifferent from that of inflation being above 4% in our

survey (Exhibit 13).

Exhibit 11: Inflation expectations 1 year out do not point toimmediate disinflation; investors were likely much more worriedabout low inflation in July 2010 than in our latest survey

Percentage responses to the question: “Where do you expect core inflation to be ayear from today?” from the J.P. Morgan inflation expectations client survey,November 2011 and previous surveys (from July 2010); % applies to responsesfrom investors across different asset classes

US Euro area UK

Nov 2011 survey

Below 0.5% 4% 6% 4%

0.5-1.5% 24% 42% 13%

1.5-2.5% 58% 38% 37%

2.5%+ 15% 14% 47%

July 2010 survey

Below 0.5% 20% 22% 12%

0.5-1.5% 53% 55% 30%

1.5-2.5% 24% 20% 39%

2.5%+ 4% 4% 18%

Mean (Nov 11) 1.86 1.61 2.34

Chg v s. prev -0.01 -0.16 -0.06

Mean (Jul 11) 1.87 1.77 2.40

Mean (Mar 11) 1.80 1.80 2.40

Mean (Nov 10) 1.40 1.35 1.90

Mean (Jul 10) 1.10 1.00 1.60

48%

84%

77%

57%

Exhibit 12: The Euro area shows the highest probability of deflationamong investors, but it is still counterbalanced by an equalprobability of high inflation; probabilities remain heavily skewed to

the upside in the UK and the US% probability; J.P. Morgan inflation expectations client survey* November 2011

0

5

10

15

20

25

30

35

US Euro UK US Euro UK

Prob of deflation Prob of high inflation

Jul 11

surveyNov 11

survey

* Probability perception of deflation/high inflation as per the question below:“What is, in your view, the likelihood of getting deflation (CPI annual inflation below0%) or high inflation (CPI annual inflation above 4%) for two years out of the next five years in each of these regions?”.

Exhibit 13: The implied probability of deflation in 2 or 5 years’ time isnot significantly different from that of inflation being above 4%; theUK prices in high probability of high inflation in 5 years’ time; over the near term US is pricing a relatively higher probability of deflation% probability of inflation <0% or above 4% in 2 and 5 years’ time*

2Y 5Y 2Y 5Y 0% floor 4% cap

Euro HICP 8.5% 12.8% 5.2% 11.8% 35 44UK** 10.7% 10.9% 8.9% 35.9% 52 67

US 14.3% 14.6% 12.4% 23.5% 39 74

Prob Inf>4% p.a. 5Y ZC (bp of notional)Prob Inf<0% p.a.

* Probability is derived from pricing caps and floors spreads using the J.P. Morganvol surface calibrated with current zero coupon style caps and floors as well as y/ ycaps and floors.** For UK RPI, we report the probability of inflation > 4.8% and <0.8%, roughlyequivalent to UK CPI > 4% or <0% p.a. if we applied the historical spread betweenRPI and CPI of 0.8%.




Jorge GarayoAC

(44-20) 7325-4820


181

Euro area market dynamics

Sovereign risk to dominate inflation

breakevens in 2012

Euro area breakevens have been intrinsically linked to

the sovereign risk dynamics during 2011, with a very

clear correlation between the risk-off/risk-on dynamics

and linker breakeven performance (Exhibit 14). This

dynamic is unlikely to change in 2012 and the

performance of peripheral spreads is likely to continue to

dominate while inflation fundamentals will take the back

seat.

Italian linkers to continue to feel the pressure

The main question is whether the underperformance of

BTPei linkers vs. BTPs will continue, and whether it will

spread to French linkers vs. OATs as we have seen

recently. The large declines in Italian BTPei breakevens

relative to inflation swaps have been significantly larger

than what we saw in the 2008 financial crisis, while

breakevens for French OATeis have already moved by a

similar degree vs. 2008 (Exhibit 15). This has led to

negative breakevens in Italian linkers out to 2019 and a

complete decoupling from inflation expectations.

Our central view is that Italian breakevens willcontinue to underperform as spreads widen further.

There are risks that Italy may be downgraded and if this

is severe enough it could impact inflation indices. Based

on our understanding of current index rules of one of the

most widely followed benchmarks, a downgrade to BBB

will result in investors reallocating BTPei holdings

(Exhibit 16). These numbers are based on our estimate

that 20–30% of investors are benchmarked, holding

around €20bn of the €77bn of current market value. With

such a large percentage of the market potentially being

dropped from indices, we believe other non-

benchmarked investors may also be reluctant to hold

onto their BTPei investments, resulting in

underperformance ahead of the event and exacerbating

the impact.

We note the following could well limit any potential

impact to Italian linkers:

1) The index rules could be changed if a majority

of investors consider them to be very harmful

for the market.

Exhibit 14: Inflation breakevens have been extremely correlated withequity markets over the past year, rising in risk-on and falling in risk-off environments

Euro Stoxx 50 and DBRi 2020 inflation breakevens vs. risk-on and risk-off *;Index points %

1800

2000

2200

2400

2600

2800

3000

3200

Nov10 Jan11 Mar11 May11 Jul11 Sep11 Nov11

1.20

1.40

1.60

1.80

2.00

2.20

2.40

10Y B/E

Euro STOXX 50

o n

o f f

o f f

o n

o f f

o n

* Risk-on and Risk-off periods defined with reference to sustained trends in EuroSTOXX 50 index – See European Derivatives for definition.

Exhibit 15: BTPei linker breakevens have underperformed vs. HICPswaps by more than twice as much as seen in late 2008; Frenchbreakevens have already moved by a similar degree as in 2008BTPei 2019 breakeven minus HICP swaps, OATei 2020 breakevens minus HICPswaps; bp

-250

-200

-150

-100

-50

0

2008 2009 2010 2011

BTPei 19 b/e minus

HICP s w aps

OATei 20 b/e minus HICP swaps

Financial

crisis late

2008

Exhibit 16: In the event that Italian linkers leave linker indices, wewould see a reallocation out of BTPeis by benchmarked investors;

France should be the main beneficiary in terms of inflows, given thatit constitutes 54% of Euro benchmarksRelative market value of constituents of Euro linker indices and expected activityfrom benchmarked investors; €bn

MV % split MV ex Ita Chg 20% 30%

Italy 77.1 28% 0% -28% -15.4 -23.1

France (€) 74.4 27% 37% 10% 5.7 8.5

France (F) 75.3 27% 37% 10% 5.7 8.6

Germany 52.5 19% 26% 7% 4.0 6.0

Total 279 100.0%

Euro l inker index* Index ex Ita ly Indexed activ ity**

* Based on Euro area government linkers with maturity over 1Y excluding Greece** Expected selling/buying under two assumptions for percentage held bybenchmarked investors in €bn.

(€) Linked to Euro HICP ex tobacco (F) Linked to French CPI ex tobacco.




Jorge GarayoAC

(44-20) 7325-4820


182

2) Investors could switch to a new linker

benchmark that includes Italy regardless of the

rating. This is more of an administrative process, and would only be likely adopted by

Italian domestic investors (which we estimate

own around 50% of the outstandings).

Taking all risks into account, we see a risk that Italian

real yields go significantly higher. If nominal yield

spreads were to widen by 250bp (broadly consistent with

our projection of 7.75% for 10Y Italy-Germany spread,

see Euro cash), Italian breakevens could trade very

negative out to the 10Y sector, with real yields between

9.50% and 11% in the sub-10Y sector possible. Our

model uses recent betas of breakevens to peripheral

spreads, which are then adjusted according to dirty pricedynamics of linkers. Linkers whose dirty prices reached

levels of 60–70 have seen buying flows, resulting in a

better performance in breakeven terms (as has been the

case in BTPei 2041 and BTPei 2035).Exhibit 17 shows

the moves in inflation breakevens during the first two

weeks of November vs. the dirty prices of linkers, we

think this behaviour could continue.

The beta of Italian breakevens to peripheral spreads has

been approximately 40% out to the 15Y sector over the

past 6 months (i.e 4bp of breakeven falls for every 10bp

of widening of spreads, Exhibit 18) and around 25% for

the very long end. Two factors drove the BTPei real yieldunderperformance relative to nominal BTPs. First, linker

real yields rose reflecting liquidity concerns and balance

sheet de-leveraging, and, second, the ECB bought

conventional BTPs in its SMP, focused on maturities out

to the 10Y sector of the curve. While the illiquidity factor

is unlikely to result in further rises in real yields, the

SMP has continued to expand, and we expect increased

buying volumes in BTPs. We see a very small

probability that the SMP starts to buy linkers or BTPs at

the very long end of the curve, so we would continue to

expect high betas with nominals in sub-10Y breakevens.

Another factor we must take into account is the deflation

floor embedded in the principal of BTPeis,

particularly relevant in a negative inflation breakeven

environment. The fact that principal payments of linkers

cannot go below par creates a theoretical floor for

breakevens. This floor is relatively close for the BTPei

September 2016, which only has around 3% of accrued

inflation but reached the breakeven lows of around

-0.50% in mid-November. We can calculate the

breakeven level which is consistent with the principal

hitting the deflation floor, which is somewhere between

-0.60% and -0.70% (which on a compounded basis

would completely offset the 3% of inflation accrual).

Exhibit 19 projects the nominal cash flows and future

HICP index levels applied to BTPei 2016 using the

current inflation breakeven expectation, as well as under

the scenario where we reach the threshold at which the

principal floor kicks in. In this scenario, the final HICP

index level expected is at the same level as the base CPI

value of 109.5227.

Putting all the information together, we have modeled the

expected breakeven moves for all BTPeis under a

significant increase in BTP nominal yield spreads to

Bunds (ranging from 220bp to 290bp) in Exhibit 20. Our

Exhibit 17: Italian linkers with a high dirty price saw significantdeclines in breakevens in November, whereas those with low dirtyprices outperformed – We must allow for this when projecting

performance going forwardInflation breakeven change 1 November–18 November vs. cash dirty price of BTPeilinkers on 1 November; bp

14

16 1719

212326

3541

-150

-100

-50

0

50

100

60 70 80 90 100 110 120Dirty price, points

B / E c h a n g e

Exhibit 18: BTPei breakevens have declined 4bp per every 10bpwidening in the nominal yield spread to Germany and are likely todecline further given our view on peripheral spreadsBTPei 2021 breakeven vs. nominal yield spread in 10Y nominal benchmarks over the past 6M; %

y = -0.43x + 2.79

R2 = 79%

0.0

0.5

1.0

1.5

2.0

2.5

1.0 2.0 3.0 4.0 5.0 6.0

Italy -Germany nom y ld spread, %

B T P e i 2 1

b r e a k e v e n




Jorge GarayoAC

(44-20) 7325-4820


183

beta for breakevens is based on the factors highlighted

above. The BTPei 2016 is expected to hit the threshold

breakeven of -0.65%, at which point the deflation floor on the principal starts to kick in, and therefore

outperforms other lines. BTPei 14s and BTPei 17s would

suffer most under our assumption while we would expect

more limited falls in breakevens at the very long end of

the curve.

While we have not modeled alternative widening

scenarios, we should expect broadly proportional

relative performance. Clearly a scenario where Italian

spreads tighten (unlikely in our view) would result in

normalisation in Italian real yields and significant

performance, both in absolute and relative terms.

French linkers should prove more resilient than

Italian to a worsening of the sovereign risk dynamics

Is France likely to follow the footsteps of Italian

linkers? As Italian and Spanish yields spreads widened

significantly, French lines started to underperform in

November as France widened vs. Germany. However,

we believe there are a series of factors that put French

linkers in a better position than Italian ones.

• Sales of Italian linkers (28% of the linker

market) could result in switches into French

(54% of the market if we include both Frenchand Euro HICP-linked lines). With German

linkers only around 20% of the market and

trading at negative real yields, French linkers

should be reasonably attractive alternatives.

Looking at the relative weights of linkers within

indices, the bonds that should attract more

inflows are (in this order) OATei 20, OATi 17,

DBRi 16 & DBRi 20, and OATi 13.

• We expect domestic demand to remain strong

on the back of French CPI hedging (at least in

1Q12) and interest from Libor-based domestic

investors to resurface. Inflows into Livret A

accounts have been particularly strong in 2011

(Exhibit 21), and we predict the Livret A rate

will increase to 2.50% in February 2012 (before

falling back to 1.75% later in the year, see

Exhibit 22). This means that we are likely to

see some French CPI hedging demand in the

early stages of 2012. Domestic Libor-based

demand for linkers should provide some support

to the market in the event of significant

Exhibit 20: Possible scenario for BTPei breakevens if BTP nominal yield spreads widened by an average 250bpCurrent linker metrics for Italian BTPeis and projected levels under nominal yield widening assumptions

Line

Dirty

price

Cum

Infl

Real

Yld

B/E

Infl

Threshold

B/E*

Nom Yld

widening

assume

Beta real

to nom

ylds

Beta to

B/E

Expected

Real Yld

Expected

B/E Infl

Expected

Move (bp)

15/09/2014 103.23 17% 7.12% -0.49% -5.51% 288 1.40 40% 11.1% -1.64% -11515/09/2016 82.58 3% 7.13% -0.34% -0.61% 275 1.10 24% 10.1% -0.61% -27

15/09/2017 85.51 12% 7.20% -0.30% -1.90% 260 1.36 36% 10.7% -1.23% -94

15/09/2019 78.19 6% 6.87% 0.11% -0.80% 250 1.24 24% 10.0% -0.49% -60

15/09/2021 70.96 4% 6.63% 0.31% -0.42% 250 1.24 24% 9.7% -0.29% -60

15/09/2023 73.21 10% 6.81% 0.59% -0.79% 240 1.14 14% 9.5% 0.26% -34

15/09/2026 68.29 2% 6.72% 0.74% -0.13% 230 1.12 12% 9.3% 0.47% -28

15/09/2035 70.20 15% 5.27% 2.36% -0.58% 220 1.10 10% 7.7% 2.14% -22

15/09/2041 63.74 5% 5.16% 2.22% -0.15% 220 1.08 8% 7.5% 2.04% -18 * Threshold breakeven level at which the principal deflation floor would be hit.

Exhibit 19: The breakeven floor-threshold for BTPei is not far fromcurrent levels given current market volatility; breakevens should notgo below -0.66% as this would imply hitting the floor on the principal

Cash flow calculation under current breakeven assumption and thresholdassumption for BTPei Sep 2016 (Base CPI value 109.5227);Nominal cash flows and Index points

Date Current B/E= -0.20%p.a. Threshold= -0.66% p.a.

Cash Flows HICP index Cash Flows HICP index

21/11/2011 -83.988 112.796667 -82.176 112.79667

15/03/2012 1.083 112.999 1.082 112.863

17/09/2012 1.082 112.855 1.078 112.482

15/03/2013 1.081 112.715 1.075 112.111

16/09/2013 1.079 112.574 1.071 111.736

17/03/2014 1.078 112.432 1.068 111.361

15/09/2014 1.077 112.293 1.064 110.996

16/03/2015 1.075 112.152 1.061 110.623

15/09/2015 1.074 112.013 1.057 110.258

15/03/2016 1.073 111.872 1.054 109.890

15/09/2016 103.088 111.732 101.050 109.5227

IRR 6.64% 6.71%




Jorge GarayoAC

(44-20) 7325-4820


184

cheapening of linkers in asset swap terms,

though the timing of such flows is difficult to

predict.

Outlook for real yields, breakevens, and

linker asset swaps

Real yields can go more negative

Real yields for German linkers are negative out to the 7Y

sector and a mere 30bp in the DBRi 2020 line. Real

yields for French linkers are slightly higher given the

current credit risk dynamics. The main driver for real

yields has been a deteriorating macro outlook,

combined with the flight-to-quality resulting from the

sovereign risk crisis. We think these factors are likely tocontinue to drive real yields into more negative territory

while the ECB easing the real refi rate should also

contribute to lower real yields at the front end (with our

view not fully priced in, see European Derivatives). With

a forecast of 10Y Bunds reaching 1.25% by 2Q12 and

Italian spreads widening 250bp (see Euro Cash), we

believe real yields will go negative in the 10Y sector. If

we use the beta of real to nominal yields of the past 12

months, the DBRi 2020 should decline around 40bp, to

negative -0.10% real yield, while 10Y breakevens would

fall by a similar extent, to around 1.10%.

Upside risks to our inflation call

The J.P. Morgan inflation assumption is consistent with

inflation staying a bit stickier than the path implied by

inflation swaps (Exhibit 23). Moreover, it does not

factor in upside risks from rises in state-administered

prices (VAT and duties), which we see as a potential

upside risk for the beginning of 2012. With Italian,

Spanish, and French public finances under close scrutiny,

we believe that any shortfalls in austerity plans are likely

to be met with rises in taxes or state-controlled prices.

Italy and Spain have just formed new governments,

which will be very much focused on providing a deficit-

reduction plan, while France is being closely watched by

markets.

We have already seen a rise in ‘administered-prices’

inflation in Italy and Spain, where these correspond to

around 8% of the respective inflation baskets (Exhibit

24). In Exhibit 25, we set out the impact state-controlled

prices may have on Euro HICP inflation, looking at VAT

rises, VAT reclassifications, general rises in administered

prices (which include services such as refuse collection,

water supply, as well as pharmaceutical products), and

Exhibit 21: Livret A inflows have been significant during 2011, weexpect continued inflows in the first months of 2012…Inflows into Livret A accounts, centralised by CDC; €bn

-2

-1

0

1

2

3

4

5

J a n - 0

9

A p r - 0 9

J u l - 0 9

O c t - 0 9

J a n - 1

0

A p r - 1 0

J u l - 1 0

O c t - 1 0

J a n - 1

1

A p r - 1 1

J u l - 1 1

2 0 0 8 *

2 0 0 7 *

18.3YTD €16.2bn2010 €7.8bn

Exhibit 22: … as we expect the Livret A rate to go up to 2.50% inFebruary before moving down to 1.75% in the following fixing (giventhe expected downward trajectory in inflation)Livret A rate*, French CPI ex tobacco inflation oya, previous month average of 3MEuribor and EONIA; %

-1.00

0.00

1.00

2.00

3.00

4.00

5.00

Jan07 Jan08 Jan09 Jan10 Jan11 Jan12

Inflation

Avg of 3M Euribor & EONIA (1M)

Liv ret A Rate

Current: 2.25%

Forecasts

* Livret A rate is set by Bank of France every 15 January and 15 July with referenceto an average of inflation and rates (also depicted in the Exhibit), floored at inflationplus 25bp.

Exhibit 23: Inflation swaps assume a larger fall in inflation than theJ.P. Morgan forecast, even if this does not include upside risks fromadministered prices, state-controlled duties and VATEuro HICP ex-tobacco J.P. Morgan forecasts and implied by HICP swaps; % oya

1.0

1.5

2.0

2.5

3.0

Oct11 Dec11 Feb12 Apr12 Jun12 Aug12 Oct12 Dec12

JP Morgan EURO HICP sw aps




Jorge GarayoAC

(44-20) 7325-4820


185

duties on energy prices. Our conclusion is that the

potential impact from these measures on Euro HICP

can be quite substantial, somewhere between 0.2%and 0.5%, on inflation.

Inflation breakevens offer value at the front end…

Traditional fundamental models for inflation breakevens

would fail to explain recent performance. Breakevens

were driven by nominal yields, peripheral spreads and

equities more than by inflation fundamentals. As a

measure of value in inflation breakevens, we prefer

looking at the excess nominal returns that sub-5Y linkers

would provide over conventional bonds, using a blend of

our inflation scenario to 2013 and the path implied by

inflation swaps thereafter (Exhibit 26). Linkers offer above-average returns vs. conventionals in the case of

France (except OATei 12), but even the German OBLi

13 offers a substantial 65bp pickup vs. Bunds. Even if

the current illiquidity dynamics justify positive excess

returns in linkers, we think front end breakevens offer

good value at these levels.

Long end breakevens to be driven by nominal yields

and a flight to quality dynamic – Favour breakeven

flatteners

Inflation breakevens further out are much more

likely to be driven by the performance of Bund yieldsand a flight-to-quality dynamic. The inflation

breakevens implied by the French HICP-linked lines

are only a touch lower than German breakevens (Exhibit

27), and we see risks that they underperform in

relative terms, driven by a general widening of

French nominal yields vs. Bunds.

Exhibit 24: “Administered prices” inflation has sky rocketed in Italyand Spain, as one would expect with countries engaging in fiscaltightening, thus we expect more pressure on this front

Administered price inflation as defined by Eurostat*; % oya

-4%

-2%

0%

2%

4%

6%

8%

10%

2002 2004 2006 2008 2010

France Italy Spain

* Items correspond to around 8% of basket in Italy & Spain, 13.5% in France.

Exhibit 25: Fiscal austerity programmes can impact short-term inflation significantly, with numerous challenges putting pressure to reducedeficits: optimistic growth forecasts and increased visibility will force countries to hike VAT, administered prices, and other duties. Italy, Spain,and France could push Euro HICP by up to half a percentage pointImpact on Euro HICP inflation from hypothetical changes in government-controlled measures

Euro area

HICP

weight

VAT rate

(standard

%)

Admin

prices

weight

>>

% duties

on

gasoline

1% higher

VAT*

VAT

reclass**

Admin

prices***

2.5% rise

Rise

duty on

gas^

Total

potential

Germany 26% 19.0 12.9 59.2 0.11 0.06 0.08 0.25

France 21% 19.6 13.6 57.3 0.09 0.06 0.07 0.22

Italy 18% 21.0 8.1 55.8 0.08 0.04 0.04 0.01 0.16

Spain 13% 18.0 8.0 49.0 0.05 0.03 0.03 0.06 0.16

Netherlands 5% 19.0 13.6 60.3 0.02 0.01 0.02 0.05

Greece 4% 23.0 9.8 59.3 0.02 0.01 0.01 0.03

Euro area 19.8 11.3 56.9 0.37 0.21 0.24 0.07 0.88

France + Italy + Spain impacts combined 0.22 0.12 0.13 0.07 0.55

Euro HICP inflation sens itivities

>> Weight of items whose prices are fully or mainly administered by the State, as defined by the European Commission.* Effect from a 1% rise in s tandard VAT rate. Assumes a pass through of 65% on 65% of the total basket, or around a 0.40% increase in domestic inflation.** Assuming a VAT reclassification of 2% of basket items from the reduced rate of VAT to a standar rate of VAT.*** Assuming a 2.5% rise in all administered prices in country baskets (average administered price inflation has been 3.5% over the last 7 years).^ Assuming convergence in duties as a % of overall price of gasoline to the Euro area average, affecting the item “Fuels for personal transport equipment” in the basket.




Jorge GarayoAC

(44-20) 7325-4820


186

Overall, being short breakevens is not necessarily the

best risk-adjusted trade, given the expected volatility

with sovereign risk dynamics. Therefore, we preferexpressing this view as a curve trade through

inflation breakeven flatteners. This is consistent with

our expectation of significant flattening in the nominal

yield curve (in 2s/10s), plus our view that inflation can

surprise on the upside in 1H12. There are different

variants in which we can express this view: 1)1s/10s

flatteners in HICP swaps (more of a play on inflation

surprising on the downside), 2) overweighting OBLi 13

vs. BTPei 19 in breakeven terms (incorporating a

negative view on Italian breakevens), or 3) BTPei 16s vs.

BTPei 21 (incorporating the value of the deflation floor

in BTPei 16).

Inflation swaps are likely to outperform cash

breakevens as peripheral spreads continue to widen,

and in general, we wouldn’t fade the relative historical

cheapness in the first months of 2012. A further

cheapening may provide good medium-term

opportunities to buy linker breakevens vs. swaps.

Regarding Euro HICP-linked vs. French CPI linkers,the latter have underperformed significantly in November and present opportunities for investors to

overweight OATis vs. OATeis over the medium term.Whereas OATis have traded historically at higher

breakevens than OATeis, they are currently tradingaround similar breakeven levels, around 20bp too cheapvs. a regression since January 2010 in the sub-10Y sector (Exhibit 28). While this may reflect the relatively lower liquidity of French CPI-linked lines, we expect a decentamount of Livret A hedging to take place early in 2012,which should favour French CPI lines. Any demand for French linkers from Libor-based investors should alsofavour French CPI-linked lines, given they provide amore attractive pickup.

Supply/demand dynamics

Euro area linker supply: risk of very low supply fromItaly, lower volumes than in 2011

Supply of inflation-linked bonds in 2011 was below our

initial expectations (€41bn vs. our initial forecast of

€51bn), as Italy issued only around €1.5bn of linkers (net

of buybacks) in 2H11. This was clearly a reflection of the

limited demand for Italian linkers during the sovereign

debt crisis that began unfolding in the summer. France

continued to issue at regular auctions, supplying around

Exhibit 26: Sub 5Y linkers will provide reasonable pickup vs.conventionals under our inflation scenario, which suggestsbreakevens offer value

Current levels, Nominal IRR under J.P. Morgan inflation assumption and pickuprelative to conventional bonds; % and bp

Lines Real Yld Inf B/E

Nom Rtn

(JPM*)

Pickup vs.

Noms (bp)

Euro HICP lines

Germany OBLI 2.250 0413 -0.60 0.86 0.97 65

Germany DBRI 1.50 0416 -0.35 1.20 1.31 35

France OAIE 3.000 0712 -1.52 2.43 0.84 7

France OAIE 1.600 0715 1.30 0.91 3.22 88

France CPI lines

France OAIE 2.500 0713 0.28 1.12 2.14 55

France BTNS 0.45 16 IL 1.59 0.87 3.50 91

France OAIE 1.000 0717 1.75 1.04 3.70 86

* Projected nominal return using J.P. Morgan forecast out to December 2012 andinflation forwards implied in Euro HICP swaps thereafter with J.P. Morgan seasonalmodel.

Exhibit 27: Inflation breakevens for French linkers are trading closeto those of German lines, and could under perform, though we do notexpect them to behave like Italian breakevensInflation breakevens for Euro HICP linked lines (x-axis denotes redemption year); %

-1.00

-0.50

0.00

0.501.00

1.50

2.00

2.50

12 14 16 18 20 22 24 26 28 30 32 34 36 38 40

Germany

France

Italy

Exhibit 28: French CPI-linked breakevens are cheap vs. French EuroHICP-linked breakevensResidual from regressing OATi 2019 vs. OATei 2020 breakeven, since January2010; %

-30

-20

-10

0

10

20

Jan10 Apr10 Jul10 Oct10 Jan11 Apr11 Jul11 Oct11




Jorge GarayoAC

(44-20) 7325-4820


187

€3.5bn during this period, while Germany did not hold

any auctions between April and November.

With sovereign risk dynamics likely to continue to

impact markets, we expect supply in 2012 to be a

below-average year. The total supply to the market will

very much depend on Italy’s strategy towards linker

issuance in a dislocated market. While we are calling for

wider BTP spreads, the timing of such widening may

impact the issuance strategy from the Italian Treasury,

which will remain as flexible as possible. We have

therefore forecast Italian linker supply to be of the order

of €12bn, assuming they continue to tap the market in an

opportunistic manner. We expect Germany to issue €9bn

while we have penciled €18bn for France (with a bit

more issued in Euro HICP lines through a new 5Y benchmark in Q1). There are clear downside risks to

Italian linker issuance, which could result in a bit more

issued from France and Germany. Overall, we expect the

supply of linkers to be at best around €39bn, slightly

below what was issued in 2011 (Exhibit 29).

On the demand side, inflation-hedging at the long end

of the curves has been very limited in 2011, and we do

not expect a huge change in this dynamic in 1H12.

The pension reform agenda in the Netherlands can be a

significant development for inflation-hedging, but we do

not expect a quick implementation of the proposed

measures, which give more leeway for Dutch pensionfunds to be more active in inflation hedging programmes.

Cross market: relative flattening vs. the

US inflation curve

Our view for Euro area breakevens contrasts with our

outlook for US TIPS, where we see short dated

breakevens most at risk and expect the breakeven curve

to steepen (around 40bp in 5s/30s by 2Q 12, see US TIPS

section). This is on the back of significant expected falls

in US inflation prints and further impact from fiscal

tightening on short dated breakevens.

Thus we generally recommend overweighting Euro

area breakevens at the front end vs. US TIPS, whilst

doing the opposite in 10Y+ maturities. In swaps space,

the 5s/30s curve in HICP swaps is broadly at the same

level as in US CPI swaps, whereas on average the latter

has traded 25bp steeper (Exhibit 30).

Exhibit 29: Euro area linker supply forecasts for 2012: lower volumesthan an already low 2011Gross supply (net of buybacks); €bn

Q1 Q2 Q3 Q4 Total % vs 2011Euro HICP 10.5 8.0 7.5 6.0 32.0 82% +0.8

France (Euro) 5.3 2.8 1.5 1.5 11.0 28% +1.6

Italy (Euro) 3.3 3.3 3.0 2.5 12.0 31% (1.8 )

Germany (Euro) 2.0 2.0 3.0 2.0 9.0 23% +1.0

France CPI 2.8 2.8 0.8 0.8 7.0 18% (3.6 )

Total 13.3 10.8 8.3 6.8 39.0 (2.8 )

% of total 32% 26% 20% 16% 93%

vs 2011 (3.5 ) (4.6 ) +4.6 +0.7 (2.8 )

By Maturity

Short (<3yrs) 1.3 1.5 1.3 1.0 5.0 11% +5.0

Medium (3-7yrs) 6.8 4.3 0.0 2.0 13.0 28% (5.7 )

Long (7-12yrs) 2.5 2.8 2.5 2.3 10.0 22% (5.6 )

Ultra long (13yrs+) 2.8 2.3 4.5 1.5 11.0 24% +3.6

Exhibit 30: The 5s/30s inflation curve should be steeper in US CPIswaps compared to Euro HICP swaps5s/30s HICP swap and US CPI swap curves; bp

0

20

40

60

80

100

120

140

Nov 09 May 10 Nov 10 May 11 Nov 11

US CPI 5s/30s

Euro HICP 5s /30s




Jorge GarayoAC

(44-20) 7325-4820

Francis Diamond (44-20) 7325-3541J.P. Morgan Securities Ltd

188

UK inflation-linked markets

2011 reviewIndex-linked gilts underperformed nominals during 2011.

Despite the fact that both CPI and RPI inflation both

broke the 5% level (5.2% and 5.6%, respectively), the

linker market was driven by the collapse in nominal

yields as investors sought the safe haven of UK gilts as

the situation in peripheral Europe deteriorated. On the

real yield curve, 5Y real yields fell the most but were

also the most volatile (Exhibit 31), while at the long end

of the curve, 30Y and 50Y real yields have moved

broadly in lockstep. Real yields are barely above 0bp at

the long end, and although this part of the curve is much

less affected by fundamental valuations and macro

drivers than the short end, there has been a high

correlation between 30Y and 10Y real yields.

Breakevens are lower on the year with the movement

driven entirely by the fall in nominal gilt yields. The

resumption of QE purchases in October had little impact

on cash breakevens, despite the fact that linkers are not

included in the QE purchase basket. This was not the

case when QE was first started in March 2009 when 10Y

breakevens initially fell 70bp following the

announcement, only to reverse this over the following

weeks. 2011 saw the issue of new ILG29, ILG34, and

ILG62 gilts via syndication with the latter, attracting anorder book close to £10bn.

2012 inflation view

2011 saw CPI inflation rise to the highest levels since

September 2008, driven by higher food and energy

prices. Core CPI rose above 3.5% oya during 1H11

before subsiding. Looking forward into 2012, we think

headline and core inflation will fall markedly to 2.7%

oya and 2.0% oya, respectively by the end of 2Q12 (see

Exhibit 7 in the earlier section Inflation outlook for

2012). The drivers are lower food and energy prices on

the back of base effects, a downshift in core goods pricesdriven by a weaker growth outlook and the impact of the

VAT hike falling out of oya inflation (Exhibit 32). We

forecast the contribution of core goods inflation to

headline CPI to be slightly negative by the end of 2012.

The main risks to our inflation forecast stem from

downside risks to growth. Our baseline view is that the

UK generates anaemic growth of 0.5% oya in 2012 but

there is clearly the risk that output is flat or negative over

the year, dragging core goods and services inflation

down. Aside from the usual monthly volatilities

surrounding the CPI comments, it is difficult to envisage

a scenario in which inflation does not fall from current

levels. The impact on inflation of QE increasing to

£425bn as per our forecast is difficult to gauge, but the

BoE estimates that the impact of £200bn QE on headline

Exhibit 31: 5Y real yields outperformed in 2011 and correlation withnominal yields was high across the curvePerformance of UK index linked gilts from 1 January 2011–17 November 2011; bp

unless stated Real y ield Current High Low Chg.

Std.

Dev*

R-sq vs.

10Y real

R-squ

vs. nom

ILG16 -140 6 -153 -132 5 93% 96%

ILG22 -30 87 -36 -96 5 100% 95%

ILG32 5 91 3 -63 4 97% 91%

ILG40 5 88 5 -61 4 95% 88%

ILG55 4 72 3 -48 3 88% 73%

Exhibit 32: Our forecast for lower headline inflation in 2012 is drivenby falls in all the main componentContribution to headline CPI by component, %oya

-0.5

0.0

0.5

1.0

1.5

2.0

2.5

Jan-11 May -11 Sep-11 Jan-12 May -12 Sep-12

FoodEnergy

Core goods

Services

Exhibit 33: An increase in QE gilt purchases will push 5Y real yieldslower– we forecast a 50bp fallEstimated change in 5Y real yields based on two factor model* under variousscenarios for additional QE and level of RPI by mid 2012; bp

4.50 4.00 3.50 3.00 2.50 2.00

50 -20 -10 -5 5 10 20

75 -35 -25 -20 -10 -5 5

100 -45 -40 -35 -25 -20 -10

125 -60 -55 -45 -40 -35 -25

150 -75 -70 -60 -55 -45 -40

Mid 2012 RPI, %oya

A d d i t o n a l Q E b y

m i d 2 0 1 2 ,

£ b n

* we model 5Y real y ields as a function of the base rate adjusted for QEpurchases** and the level of RPI.5Y = 0.48 * effective policy rate – 0.15* RPI. R-squ: 81%, std. error: 50bp** we assume £25bn of QE lowers base rate by 0.3%. Current effective rate is -2.8%.




Jorge GarayoAC

(44-20) 7325-4820


189

CPI has been between 0.75% and 1.50%, and we don’t

discount that a further increase in QE could prevent

inflation falling as low as we forecast. The monetaristargument that excess reserves will generate inflation

will only become relevant once economic recovery is

achieved and demand for credit increases – a

situation which is several years away, in our view.

Real yields view – be long 10Y real

yields

Real yields are at historically low levels, but we think

that they can fall further in 2012, particularly in the sub-

10Y sector of the curve. The QE increase in October

2011 and the very dovish set of forecasts in the

November Inflation Report suggest that the BoE isfocused on managing downside risks to growth, with

more QE likely in the near future in our view. Our model

of 5Y real yields as a function of the base rate adjusted

for QE purchases and the level of RPI (to measure

demand for inflation protection) suggests that 5Y real

yields can fall around 40bp-50bp by mid-2012

(Exhibit 33) from current levels. Given the strong

correlation between 5Y and 10Y real yields we expect

ILG22 real yield to fall around 30-40bp by mid 2012.

We note that, long positions in 5Y and 10Y linkers offer

attractive carry and slide vs. the forwards (Exhibit 34).

During 2010, 30Y real yields traded in a very pronounced 60bp–90bp range (ILG40) and many end

investors tailored their investment strategies around

playing this range. This has not been the case this year as

30Y yields have trended downwards (Exhibit 35). One

of the key questions for 2011, in our view, is the

direction of 30Y real yields. Will 30Y yields rise back

up to the 2010 levels or will a new trading range be

established around current levels?

To answer this question, we need to take a look at the

drivers of 30Y real yields, namely nominal 30Y yields

and the supply/demand outlook. We note that that

correlation between 30Y real and nominal yields has not

been constant over the past few years, particularly during

periods of QE gilt purchases when the relationship has

tended to be weaker (Exhibit 36). This correlation also

tends to increase when nominal yields are trending

higher or lower (Exhibit 37).

We can model 30Y real yields as a function of 30Y

nominal yields and the size of pension fund liabilities

(Exhibit 38) where an increase in liabilities tends to

result in lower real yields. The intuition here is that as

Exhibit 34: Long positions in 5Y and 10Y real yields look attractivecompared to the forward real yield curveImplied carry and slide for being long real par rates over 3M, 6M, 9M, and 12M

horizon; bp

7

14

21

28

6

11

17

22

13

45

0

5

10

15

20

25

30

3M 6M 9M 12M

5Y 10Y 30Y 50Y

Exhibit 35: Long-end real yields trended lower in 2011 after range-trading for most of 2010Par 30Y real yield; %

0.00

0.20

0.40

0.60

0.80

1.00

1.20

Feb 10 Aug 10 Feb 11 Aug 11

Exhibit 36: The correlation between 30Y real and nominal yields isweaker when the BoE is conducting QE gilt purchasesBeta and r-squared between 30Y real par and 30Y nominal par yields, January2009–October 2011

0.00

0.10

0.20

0.30

0.40

0.50

0.60

0.70

Beta R-squ

QE Non-QE




Jorge GarayoAC

(44-20) 7325-4820


190

liabilities, which are inflation-linked, increase, then

pension funds are more likely to buy index-linked gilts to

hedge these, although we note that changes in liabilitiesmay not always result in renewed hedging flows as they

can also increase due to increases in longevity and tend

to increase as the point of payment increases. This

model suggests that 30Y index-linked gilt yields

currently look cheap (Exhibit 39), and our forecast for

a 100bp fall in 30Y gilts would imply 30Y real yields

rallying 30bp to around the -25bp level, although we

think that in practice any move to this level would be

very slow and gradual. The 0bp level may represent a

psychological barrier around which investors may

attempt to fade any move to negative 30Y real yields. At

the other extreme, we think any move in 30Y real yields

back up to the 1% level is highly unlikely to occur over 2012 as 30Y nominal gilt yields would need to rise to

around 5% based on the current betas.

Breakeven view

Breakevens have fallen over the past year by 50bp in the

10Y and 30Y sectors, and by nearly 100bp in the 5Y

sector, despite the rise in CPI inflation to 5.2% and RPI

inflation reaching the 5.6% level. In fact, over the past

few years 5Y cash breakevens have show a weak

relationship with spot inflation prints unless RPI is below

the 2% level (Exhibit 40) and over the past couple of

years, nominal yields have been the dominant drivers of breakevens (Exhibit 41).

Our breakeven view for 2012 is as a straightforward one

– lower nominal yields will drive breakevens lower in the

5Y and 10Y sectors of the curve. Our forecast for a

75bp fall in 10Y gilt yields during 1H12 (to the

1.50%) level would broadly translate into a 30bp fall

in ILG22 breakevens, to the 2.30% level. Our

expectations for CPI and RPI inflation may, at the

margin, also contribute to lower breakevens. In the 30Y

sector, we note that non-linearities are starting to appear

in the relationship between breakevens and nominal

yields (Exhibit 42), which is likely a result of demand-specific drivers. Pension funds and ALM-driven

investors tend to have specific trigger levels at which

they look to hedge inflation which may limit the

magnitude of any down move in 30Y breakevens.

The 5s/10s inflation curve has been broadly range-bound,

particularly in the RPI swap space, and we think this can

continue. We think it is far too early to position for

sustained steepening on the back of unanchored inflation

expectations from an increase in QE. The monetarist

Exhibit 37: The relationship between 30Y real yields and nominals isweaker when nominal yields are range-boundStatistics from regressing 30Y par real yields vs. 30Y par nominal yields

Period 30Y y ield Beta Rsq

19/11/09 - 15/2/10 sell-off 0.8 64%

16/2/10 - 16/12/10 range 0.2 37%

10/2/11 - 18/11/11 Rally 0.5 89%

Exhibit 38: 30Y real yields can be explained by nominal yields andthe size of pension fund liabilities…Statistics from regressing 30Y par yield vs. 30Y nominal par yield and the PPFmeasure of total liabilities for DB pension funds, monthly data, March 2003–Oct11

Variable Coefficient T-stat

30Y nominal y ield (%) 0.308 3.6

PPF liabilities (£bn) -0.003 -18.2

R-squ 81%Std. error (bp) 21

Exhibit 39: …and currently look some 30bp cheap Actual and predicted level* of 30Y par real yields, monthly data March 2003–October 2011; %

-0.50

0.00

0.50

1.00

1.50

2.00

2.50

Mar-03 Mar-05 Mar-07 Mar-09 Mar-11

Predicted

Actual

0bp level

* predicted using model in Exhibit 38.

Exhibit 40: 5Y breakevens are generally non-directional with inflationwhen RPI is above 2%5Y par cash breakeven* regressed against RPI print, 2005–2011**; %

y = 0.1x + 2.55

R2 = 3%

y = 0.31x + 2.3

R2 = 29%

1.00

1.50

2.00

2.50

3.00

3.50

4.00

-2.0 -1.0 0.0 1.0 2.0 3.0 4.0 5.0 6.0

RPI, %oya

* Derived from BoE fitted par curve.** Excludes period from December 2008–February 2002.




Jorge GarayoAC

(44-20) 7325-4820


191

arguments that, for additional QE, translates into higher

inflation is not easy to demonstrate empirically, and we

think that inflation expectations run the biggest risk of becoming unanchored when the macro backdrop has

improved. This scenario is several years away, in our

view, and we see little benefit in positioning for this

outcome now.

Cash breakevens vs. RPI swaps

The past few months have seen a marked

underperformance of index-linked cash breakevens

compared with RPI swaps, with the relative spread now

at the widest levels ever for ILG55 and at its widest level

since March 2009 for ILG37 (Exhibit 43). This spread

essentially reflects the relative ASW levels betweenindex-linked gilts and conventional gilts, and indicates

that linkers offer an attractive pick-up vs. Libor

compared with conventional gilts.

Going forward, we think this relative spread can

widen further, and we would not advocate fading this

apparent cheapness. We think Libor-based investors may

look to take advantage of the relative cheapness in

linkers compared with nominals, but over the past 6

months, this relative spread has been driven by nominal

gilt ASW levels and European peripheral spreads,

particularly in the 10Y+ sector (Exhibit 44). We think

cash breakevens will continue to underperform swapbreakevens, given our view that peripheral spreads will

widen further in the coming months and that nominal

ASW levels may fall further (i.e. gilts outperform

swaps), driven by additional QE and lower nominal gilt

yields. For investors who do want to fade this cheapness,

we think the sub-10Y part of the curve is the best place

to position, given the lower correlation with peripheral

spreads and nominal ASW levels.

Market technicals

Issuance

Our fiscal outlook and issuance forecasts for gilts are

described in the United Kingdom section, and for

FY12/13, we expect linker issuance to break through the

£40bn level (Exhibit 45). The low level of spot and

forward real yields, and the continued backdrop of strong

demand from institutional investors warrant an increase

in index-linked gilt supply, in our view. Total gilt supply

is expected to rise to £190bn, and we forecast linker

issuance remaining around 23% of total gilt supply. As is

typically the case, we expect the bulk of issuance to

come in the long end (15Y+) of the real yield curve, with

taps to ILG34, ILG40 and ILG62 likely. We also see a

reasonable chance of a new linker in the 35Y sector of

the curve, as well as the possibility of a short-dated linker

in the 10-12Y sector, although this is less likely now that

the ILG29 has been issued.

Exhibit 41: Breakevens have been driven by nominal yields over thelast couple of years. We expect the strong relationship to continue …Statistics from regressing breakevens vs. nominal yields over the last 12M and 24M

Beta R-squ Beta R-squ

ILG22 0.4 87% 0.4 68%

ILG32 0.5 91% 0.5 77%

ILG40 0.5 92% 0.5 81%

ILG55 0.6 93% 0.7 84%

12M 24M

Exhibit 42: …but the long end of the BE curve is becoming non-linear ILG40 regressed against nominal 30Y yields, last 3M; %

y = 0.64x 2 - 3.9x + 9.0

R2 = 83%

2.90

3.00

3.10

3.20

3.30

3.40

3.50

3.0 3.2 3.4 3.6 3.8 4.0

30Y nominal y ield, %

18 Nov

Exhibit 43: Cash breakevens are at cheap levels vs. RPI swaps

across the curve…Cash breakeven – RPI swap for selected linkers; bp

-50

0

50

100

150

Oct 08 Jun 09 Feb 10 Oct 10 Jun 11

ILG17 ILG37 ILG55




Jorge GarayoAC

(44-20) 7325-4820


192

The results of the consultation process on whether the

DMO should issue CPI-linked gilt should be available in

the coming months. In our view, the decision as towhether CPI-linked gilts will be issued will be very much

driven by anticipated end-investor demand, particularly

pension funds and other investors who will be required to

provide CPI-linked payments. We expect around 15% of

pension schemes with indexation linked to RPI will be

able to or will want to make the switch from RPI to CPI

indexation, as mandated by new legislation. However,

given that pension fund schemes hedge only around 60%

of their inflation risk, they may prefer to continue to use

RPI-linked instruments and run the CPI-RPI basis risk.

If CPI-linkers were to be issued, we think they would

likely be in the 15-30Y maturity sector of the curvewhere demand is likely to be the highest. We would also

expect a flexible issuance programme to enable the DMO

to tailor issuance needs in either the CPI-linked or RPI-

linked space to meet demand, with the allocation of

supply between the two instruments likely to vary during

the fiscal year.

Real yield behaviour into syndications

For the past couple of fiscal years, linker syndications

have risen to around 40%–50% of total index-linked gilt

sales, and we expect a similar proportion of linkers to be

issued via syndications in FY12/13, equating to £18bn– 22bn of syndicated linker issuance, based on our

forecast. Since the DMO introduced syndications as a

permanent feature of the issuance programme in April

2009, there have been 9 index-linked gilt syndications,

raising a total of £38.75bn, with an average size of

£4.3bn.

These syndications have given rise to some interesting

behaviour on the linker curve. Looking at the

performance of the closest maturity bond to the linker

being issued, we note that, on average, real yields have

risen 8bp in the 15 business days prior to the syndication

(Exhibit 46). This is then quickly reversed in the fivedays following the syndicated supply. This phenomenon

has occurred into 7 out of the 9 syndications, with the

largest increase close to 25bp. We think that this can be

explained by the fact that 1) syndications are often

multiples of the size of typical index-linked gilt auctions

and represent large concentrated amounts of linker

duration supply into the market. This creates uncertainty

around the actual potential demand for the bond; and 2)

the actual details of the bond maturity are not announced

until a few weeks before the syndication date.

Exhibit 44: … and peripheral spreads and nominal ASW levelsexplain the bulk of this cheapening for 10Y+ l inkersR-squ from separately regressing* cash BE – RPI spread vs. nominal ASW and

peripheral spreads, last 6M; %

0%

20%

40%

60%

80%

100%

ILG17 ILG22 ILG37 ILG55

Nom ASW Peripheral spread

* Single factor regressions.

Exhibit 45: We expect over £40bn of linker issuance in FY12/13Index linked gilt total issuance and supply as % of total gilt issuance£bn %

0

10

20

30

40

50

2 0 0 0

2 0 0 1

2 0 0 2

2 0 0 3

2 0 0 4

2 0 0 5

2 0 0 6

2 0 0 7

2 0 0 8

2 0 0 9

2 0 1 0

2 0 1 1

2 0 1 2

( f )

0%

10%

20%

30%

40%

Amt, £bn Proportion, %

Exhibit 46: Real yields rise prior to linker syndications on average… Average real yield of closest maturity linker around index linked gilt syndications*,July 2009–October 2011; bp

55

60

65

70

-20 -15 -10 -5 0 5 10 15 20

Business day s around linker sy ndications

* Dates and linkers used: 23/07/09 ILG42, 24/09/09 ILG50, 27/01/10 ILG40,26/05/10 ILG50, 27/07/10 ILG40, 27/01/11 ILG55, 24/05/11 ILG34, 26/07/11 ILG34,25/10/11 ILG62.




Jorge GarayoAC

(44-20) 7325-4820


193

We expect this trend to continue and we think

investors can exploit this dynamic by selling the

closest maturity bond to the linker being issued viasyndicate in 15 business days ahead of linker

syndications. Instigating a trading rule in which the

closest maturity linker to the syndicated bond is sold on

the 15th day prior to the syndication and unwound on the

day of the syndication itself would have generated a total

of 24bp of P/L in yield terms in 2011, and on average,

6bp per syndication over the past three years (Exhibit

47).

Demand for linker syndications is almost entirely from

domestic investors, and the size of the book as well as

the actual amount of the bonds issued can be affected by

large asset allocation orders. Predicting the size of the potential demand for linker syndication in advance is not

easy, but looking at the relative valuations between

equities and real yields may indicate potential demand.

The idea here is that pension funds and asset managers

are more likely to switch from equities into linkers when

equities are relatively rich compared with linkers on a

real-yield basis. This is because pension funds

asset/liability mix is not managed based on specific

funding level triggers but more holistic, long-term

funding plans, and on aggregate, funds are reluctant to

crystallise equity losses.

We can measure the relative valuations by looking at theratio of equity E/P (as a measure of equity real yield) vs.

the real yield on long linkers. Looking at this valuation

metric and its evolution in the few days prior to

syndications there appears to be a loose relationship i.e.

syndication demand is stronger if equities have richened

vs. linkers, although this is skewed by the strong demand

for the ILG62 syndication in October 2011. It is worth

noting that the syndication of ILG62 coincided with the

richest levels of equity/linkers for 2011, and we think

that looking at the performance of this ratio ahead of

future linker syndications may provide some indication

of potential investor demand.

Exhibit 47: …creating attractive trading opportunities for investorsTotal and average P/L from trading rule which goes short closest maturity linker onT-15 and unwinds the trade on T*; bp

0

5

10

15

20

25

30

35

2009 2010 2011

Total P/L Average P/L

* T = date of linker syndication.




Jorge GarayoAC

(44-20) 7325-4820

Kimberly Harano (1-212) 834-4956J.P. Morgan Securities Ltd, J.P. Morgan Securities LLC

194

US inflation-linked markets

Up, down, then back aroundThe path of breakevens over 2011 looks remarkably

similar to the 2010 experience. First, breakevens steadily

widened over the first few months of the year thanks to a

surge in energy prices (Exhibit 48). In late spring,

however, European concerns resurfaced, the US

economy appeared to hit a speed bump, and the resulting

risk-off trade pushed breakevens narrower. Finally, in the

fall, the Fed was forced to step in with Operation Twist,

and this action, along with some positive news out of the

October EU summit, helped breakevens reverse some of

their narrowing. On a total return basis, the sizable

increases in headline CPI helped TIPS outperform

nominals early in the year, but TIPS sharply

underperformed Treasuries in 3Q11 as nominal yields

declined more than real yields (Exhibit 49).

Real yields, in turn, refreshed their all-time lows this year

(Exhibit 50). Five-year yields traded negative for most

of 2011, hitting a low of -1.12% in early November. As a

result, the Apr-16 TIPS became the first Treasury

security issued with the new minimum 0.125% coupon

rate, despite a clearing yield of -0.18%. Even 10-year

real yields turned negative in 3Q11, reaching -0.15% at

the lows.

TIPS supply and demand in 2012

After Treasury announced modifications to the TIPS

calendar at the November 2010 refunding, 2011 became

the first year with a TIPS auction every month. At this

year’s November refunding, Treasury wrote that “TIPS

are an important part of Treasury’s overall debt

management strategy” and indicated that it “expects to

continue to gradually increase gross issuance of TIPS in

2012.” Based on these comments, we look for modest

increases in auction sizes across the curve, for a total

of $146bn of gross issuance in 2012 (Exhibit 51)

compared to $131bn of expected issuance in 2011. With$46.3bn of redemptions in 2012 and net Fed purchases of

zero, effective net issuance in 2012 will likely rise to

$100bn versus $79bn (net of redemptions and Fed

purchases) in 2011.

Alternatively, we can evaluate changes in TIPS supply

on a duration-weighted basis, since the Fed will

remain a source of duration demand in 2012, even if

its holdings of TIPS will likely remain unchanged in par

notional terms. Over two rounds of purchases and sales

operations, the Fed has bought about $2.3bn of 10-year

TIPS equivalents per month. If we assume the Fed

maintains this pace, then it will likely purchase about

$14bn 10-year Treasury equivalents in 2012. However,

these purchases will only partially offset the expected

Exhibit 48: Breakevens widened early in the year thanks to a surgein energy prices, but narrowed sharply in the summer as growthprospects waned and risk aversion rose

5-year, 10-year, and 30-year TIPS breakevens (all left axis) versus rolling frontWTI futures contract price (right axis);bp $/bbl

70

80

90

100

110

120

100

150

200

250

300

Nov 10 Jan 11 Feb 11 Apr 11 Jun 11 Jul 11 Sep 11 Nov 11

30Y

10Y

5Y

Oil prices

Exhibit 49: TIPS outperformed nominals on a total return basis in1H11, but sharply underperformed in 3Q11Quarterly total returns on J.P. Morgan TIPS index (JUSTINE) and GBI-US indexby maturity bucket; duration of the indices, and TIPS excess returns*; %

1-10Y 10Y+ 1-10Y 10Y+ 1-10Y 10Y+ 1-10Y 10Y+ 1-10Y 10Y+

1Q11 -0. 1% -0. 9% 3. 9 13. 0 2. 3% 1. 3% 4.7 14. 4 2. 4% 2. 3%

2Q 11 2.3% 3.3% 3.9 13.2 2.9% 5.0% 4.9 14.3 0.0% 1.4%

3Q11 3.8% 23.5% 4.0 14.7 1.8% 11.7% 5.1 14.8 -3.0% -12.0%

QTD 0.1% -0. 3% 4.1 14.8 1.4% 3.8% 5.1 15.2 1.3% 4.0%

Nominal returns Nominal dur'n TIPS returns TIPS dur'n TIPS-nom returns*

* Calculated as TIPS return-(Tsy duration)/(TIPS duration)*(Tsy return).

Exhibit 50: Real yields refreshed their lows in 20115-year and 10-year TIPS real yields; %

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

Nov 10 Feb 11 Jun 11 Sep 11

10Y

5Y




Jorge GarayoAC

(44-20) 7325-4820


195

rise in TIPS duration supply. As Exhibit 52 illustrates,

even with Fed demand, we expect TIPS duration

supply to rise to $140bn in 2012 from $108bn 10-yearTIPS equivalents in 2011.

Although supply should increase in 2012, we expect

end-user demand to rise to meet it. As Exhibit 53

shows, while the average offering size for TIPS has

steadily increased over recent years, so has end-user

demand, with direct and indirect bidders taking down

even greater fractions of (even larger) issues. The effect

of larger auction sizes can also be seen in primary dealer

transaction data: as Exhibit 54 shows, transaction

volumes in TIPS have steadily increased since Treasury

committed to increase auction sizes and improve

liquidity in late 2009. Although increased transactionvolume by itself is not necessarily indicative of improved

liquidity, anecdotal reports suggest that the increased

frequency and size of auctions have improved the

depth of the TIPS market and attracted a broader

base of investors1.

Outside of auctions, we can gauge investor demand for

TIPS based on flows into inflation-protected mutual

funds. In our 2011 Outlook, we noted that mutual fund

flows have historically been correlated to changes in

energy prices, with higher prices leading to greater flows.

In addition, since the Fed began its QE programs, long-

term inflation concerns have increased—a development

1 See “US Treasury’s Makeover on TIPS Ushers in Boom Era; China

Buys,” Min Zeng, Wall Street Journal, 11/15/11

that has been captured in the rise in the 5Yx10Y payer

swaption skew. As a result, we found that the level of theswaption skew has helped explain flows into inflation-

protected mutual funds over the past three years. With

the skew holding fairly constant this year, however,

energy prices were the dominant driver of mutual fund

flows (Exhibit 55). As our oil strategists look for modest

gains in oil prices next year (see below), we expect to

see continued inflows into inflation-protected mutual

funds.

Exhibit 51: We expect gross issuance of TIPS to increasemodestly in 2012…J.P. Morgan forecast for gross issuance; reopenings shaded in grey; $bn of

real principal5yr TIPS 10yr TIPS 30yr TIPS Subtotal

Jan-12 14

Feb-12 10

Mar-12 12 36

Apr-12 16

May-12 12

Jun-12 8 36

Jul-12 14

Aug-12 14

Sep-12 12 40

Oct-12 8

Nov-12 12

Dec-12 14 34

TOTAL 44 76 26 146

Exhibit 52: …and duration supply will rise substantially evenwith Operation TwistGross issuance of TIPS versus duration supply net of Fed purchases*; $bn of real principal in 10-year TIPS equivalents

41

51

77

61

86

68

41

47

54 54

7268

40

50

60

70

80

90

1H10 2H10 1H11 2H11 1H12 2H12

Gross issuance Dur'n supply net of Fed

* Assumes pace of duration buying over October and November 2011 ($2.3bn 10-year TIPS equivalents per month) is maintained.

Exhibit 53: If you sell it, they will comeOffering size for TIPS auctions averaged by calendar year, versus average of sum of direct and indirect bidder percentage*;$bn %

35%

40%

45%

50%

55%

5

6

7

8

9

10

11

2006 2007 2008 2009 2010 2011

Avg auction sizeEnd user %

* 2011 average is over January-November .Source: US Treasury




Jorge GarayoAC

(44-20) 7325-4820


196

In sum, although we look for TIPS supply to increase

next year, we expect demand at auctions and retail

demand to remain robust. Indeed, our J.P. Morgan USFixed Income Investor Survey indicates that 24% of

respondents plan to increase their exposure to TIPS next

year, versus just 7% that plan to reduce exposure (see US

Cross Sector Overview).

The outlook for US inflation

Last year, heading into 2011, we looked for headline and

core inflation to end the year below 1%. Instead, monthly

gains in headline inflation averaged 0.6% over the first

five months of the year, reaching 3.9% (year-over-year

ago) by September. Similarly, core CPI surged to 2.1%

in October from 0.65% in December 2010.

Looking ahead, we expect both headline and core

inflation to moderate. As Exhibit 56 shows, the upside

surprises in core CPI were driven primarily by strong

gains in apparel and vehicle prices as well as some

firming in owners’ equivalent rent (OER). The gains in

the first two categories were driven by temporary

factors—high cotton prices and supply shocks from

Japan—which we do not expect to recur in 2012. In

addition, with incomes constrained by the weak labor

market, gains in OER should remain muted. Thus, we

look for core CPI to decline to 1.0% and headline CPI to

fall to 1.4% by year end. Exhibit 57 shows our forecastfor the monthly changes in headline CPI. We look for

prices to rise modestly over 1H12 and be roughly flat

over 2H12.

Two factors that pose risks to our inflation forecast are

energy prices and the dollar. Our commodity strategists

currently look for modest declines in Brent and WTI oil

prices early in the year followed by increases in 2H12(Exhibit 58). In addition, our currency strategists look

for the dollar to weaken modestly over the next year.

Inflation expectations and dispersion

How does our forecast for inflation compare to investors’

expectations? To answer this question, we examine the

November 2011 results of our J.P. Morgan Inflation

Expectations Survey, which surveys a wide range of

investors across the globe. As Exhibit 59 shows,

Exhibit 54: Transaction volumes in TIPS have steadily increasedsince Treasury committed to improving liquidityRolling 3-month moving average of primary dealer transaction volume in TIPS

(daily average for each week); $bn

4

6

8

10

12

Oct 06 Mar 08 Jul 09 Nov 10 Source: Federal Reserve Bank of New York.

Exhibit 55: Flows into inflation-protected mutual funds weredriven by energy prices this year Monthly flows into inflation-protected mutual funds* versus 1-year percentagechange in oil futures prices, lagged 1 month;$bn %

-5

0

5

10

15

20

25

30

35

40

-1.0

-0.5

0.0

0.5

1.0

1.5

2.0

2.5

Aug10

Sep10

Oct10

Nov10

Dec10

Jan11

Feb11

Mar 11

Apr 11

May11

Jun11

Jul11

Aug11

Sep11

Mutual fund flows

1Y chg in oil prices

* Includes both weekly and monthly reporters.Source: EPFR Global

Exhibit 56: The upside surprise in core CPI this year was driven bygains in OER, apparel and vehicle prices, which we do not expectto recur in 2012Year-over-year-ago percent change in owners’ equivalent rent, apparel prices,and vehicle prices; %

-4%

-2%

0%

2%

4%

6%

Oct 08 May 09 Nov 09 Jun 10 Dec 10 Jul 11

OER Apparel Vehicles




Jorge GarayoAC

(44-20) 7325-4820


197

investors’ expectations for near-term core inflation are

higher than they were a year ago. Our November 2011

survey shows that over half of investors expect coreinflation to be between 1.5%-2.5% next year, while a

similar number expected inflation to range between 1-2%

in 2011. In addition, the “tails” of the distribution have

shifted: last year, 10% of investors expected core

inflation to be below 0.5% while 8% expected inflation

to exceed 2.5%. The corresponding figures for this year’s

survey are 4% and 15%, respectively, indicating that

investors now appear to be more concerned about upside

risk in inflation than downside risk.

The message is similar for medium-term expectations.

Sixty percent of respondents think headline inflation will

be above target or significantly above target over themedium term, while 8% think it will be below target

(Exhibit 60). In contrast, in last year’s survey 57% of

respondents thought headline inflation would be above

target, while 13% thought it would be below target. It is

interesting to note that the tails of the distribution have

fallen: last year, about 30% of investors expected

inflation to be either below target or significantly above

target, while only 20% of investors held those extreme

views this year.

This decline in the dispersion of medium-term inflation

expectations can be seen in other survey measures. For

example, Blue Chip collects economists’ forecasts for various economic indicators over the long term. As

Exhibit 61 shows, the dispersion of economists’

forecasts for medium-term inflation (over the next five

years) rose sharply around the financial crisis, but has

now declined to levels last seen in early 2008.

The outlook for breakevens

Last year, we modified our long-term model for

breakevens to use an “illiquidity cost metric” as one of

the explanatory variables in the model, instead of trying

to model liquidity-adjusted breakevens. This year, we

make two additional modifications to our model. First,given that the Fed’s quantitative easing programs have

had substantial impacts on inflation expectations and

breakevens, we have added the size of the Fed’s

securities holdings as an additional explanatory variable.

Second, instead of using Nov-27 P-STRIPS asset swap

spreads minus a barbell of 10-year and 30-year Treasury

asset swap spreads as our metric for the cost of

illiquidity, we are now using the spread between 5-year

pre-refunded muni yields (taxable-equivalent) and 5-year

Treasury yields. Pre-refunded munis are fully

collateralized with Treasuries, so they have effectively

the same credit risk as Treasuries, but they trade cheaper

because of lower liquidity. Thus, the pre-refunded

muni/Treasury spread serves as a metric for liquidity.

The other factors in our model are unchanged. We model

3-month forward breakevens to take out the impact

of near-term carry, and we use four additional

Exhibit 57: We look for headline inflation to moderate in 2012J.P. Morgan forecast for monthly % changes in headline CPI and year-over-year ago rate;

% %

0.02

0.200.23

0.260.24

0.16

-0.02

0.09

0.050.08

0.15

-0.06

0.5

1.0

1.5

2.0

2.5

3.0

-0.2

-0.1

0.0

0.1

0.2

0.3

Jan12 Feb12 Mar 12 Apr 12 May12 Jun12 Jul12 Aug12 Sep12 Oct12 Nov12 Dec12

Monthly change

Year-over-year ago rate

Exhibit 58: We expect modest increases in oil prices and aweakening of the dollar over the course of 2012Our forecast for WTI (period average) and the J.P. Morgan USD index (nominalnarrow effective exchange rate)

Current 1Q12 2Q12 3Q12 4Q12

Brent oil futures; $/bbl 107.56 108.00 105.00 115.00 120.00

WTI oil futures; $/bbl 97.41 90.00 90.00 100.00 110.00

JPM Dollar index; level 81.4 82.4 80.5 80.5 80.0

Exhibit 59: Most investors expect higher core inflation than theydid at this time last year Distribution of expectations for US core inflation over the next year from the J.P.

Morgan Inflation Expectations Client Survey as of November 2011 andNovember 2010; %

1 03

5

18

33

24

96

1 1

7

19

31

20

12

4 4

0

10

20

30

40

B e l o w -

0 . 5 %

- 0 . 4

9 -

0 %

0 - 0 .

4 9 %

0 . 5 - 0 .

9 9 %

1 . 0 - 1 .

4 9 %

1 . 5 -

1 . 9 9 %

2 - 2 .

4 9 %

2 . 5 - 2 .

9 9 %

A b o v e

3 %

Nov 11 Nov 10




Jorge GarayoAC

(44-20) 7325-4820


198

factors: 1) the level of nominal yields to account for the

directionality between breakevens and nominal yields; 2)

the unemployment rate, as a measure of slack in theeconomy; 3) the one-year-ahead forecast for the

budget surplus as a percentage of GDP, as a proxy for

fiscal policy-driven inflation pressures, and 4) realized

headline CPI, since inflation expectations tend to be

anchored to realized inflation itself. Exhibit 62 presents

the statistics for our model for 5-, 10-, and 30-year

breakevens.

We can make several observations about breakevens

based on this model. First, as expected, breakevens are

positively correlated with nominal rates, with breakevens

widening as nominal rates rise. Interestingly, long-end

breakevens appear to be most sensitive to nominal rates,widening about 3.6bp for every 10bp rise in long-end

nominal yields, while front-end breakevens are less

sensitive to rates. On the other hand, front-end

breakevens are most sensitive to the current level of the

unemployment rate and realized inflation, as these

factors have a greater impact on near-term inflation

expectations.

Forward-looking budget surplus expectations have a

similar impact across the curve, with every 1% reduction

in the annual deficit (roughly $160bn) causing

breakevens to narrow about 9-10bp. We can use these

betas to estimate the impact of future deficit reduction proposals on breakevens. For example, if market

expectations shift to expect an additional $1tn of cuts

over 10 years (above and beyond baseline expectations),

that would equate to $100bn of cuts per year, or about

0.5% of GDP. Based on the betas in Exhibit 62, these

expected cuts would lower the fair value for breakevens

by about 4.5-5bp.

As for the Fed’s balance sheet, we find that increases

(due to quantitative easing) lead to wider breakevens,

Exhibit 62: Our updated model for breakevens

Statistics for 3-month forward breakevens regressed against nominal yields, the unemployment rate, the 3-month average of 1-year ahead budget surplus expectations asa percentage of GDP, year-over-year ago headline CPI, the size of the Fed’s balance sheet, and our illiquidity cost metric*; monthly data over 12/2002-10/2011

Nominal

yields (% )

Unemployment

rate (%)

Budget surplus as

% of GDP (% )

Headline CPI,

yoy rate (% )

Fed's securities

holdings ($bn)

Illiquidity cost

metric (bp)

Intercept

(bp) R-sq

Beta 20.97 -29.3 -8.7 15.9 0.034 -0.46 237.7 85%

T-stat 4.5 -4.2 -2.4 6.8 4.4 -7.1

Beta 25.67 -21.4 -9.6 11.4 0.029 -0.23 172.7 83%

T-stat 6.1 -5.2 -4.0 7.2 5.6 -5.1

Beta 35.66 -21.6 -8.8 6.8 0.023 -0.16 161.4 85%

T-stat 9.0 -7.1 -4.4 5.0 5.1 -3.830Y

5Y

10Y

* 5-year muni pre-re yield (taxable equivalent) minus 5-year hot-run Treasury yield

Exhibit 60: Fewer investors expect extreme outcomes for medium-term inflation compared to last year…Distribution of expectations for US headline inflation over the next 2-5 years

from the J.P. Morgan Inflation Expectations Client Survey; %

8

33

48

1213

30

41

16

0

10

20

30

40

50

60

below target around target above target s ig. above target

Nov 11 Nov 10

Exhibit 61: …and this decline in the dispersion of inflationexpectations can be seen in other surveys

Average of top 10 responses minus average of bottom 10 responses for long-range consensus projection for year-over-year ago percent change in headlineCPI (average rate for the next five years); %

0.6

0.8

1.0

1.2

1.4

1.6

Mar 03 Jul 04 Dec 05 Apr 07 Sep 08 Jan 10 May 11

Source: Blue Chip Economic Indicators




Jorge GarayoAC

(44-20) 7325-4820


199

with the effect strongest on the front end. Using the betas

from our model, we can also estimate the impact of QE1

and QE2 on breakevens. Given that the Fed’s securityholdings have grown by about $2.14tn since late 2008,

when the Fed first began purchasing MBS, we estimate

that the fair value of TIPS breakevens have risen by

72bp, 62bp, and 49bp, in the 5-, 10-, and 30-year sectors,

respectively. Our model also implies that 5-year

breakevens will benefit the most if the Fed embarks on

QE3, with every $200bn in net purchases increasing the

fair value of breakevens by 6.7bp.

Five-year breakevens also appear likely to benefit the

most if liquidity improves, or underperform the most if

liquidity deteriorates. This suggests that in flight-to-

liquidity periods, the breakeven curve is likely tosteepen, ceteris paribus.

Putting it all together, we project the level of breakevens

in 2012 based on our forecasts for the underlying

variables and our expectation that our illiquidity cost

metric declines to its average level over the past year. We

also assume that the Fed does not expand its balance

sheet via QE3, though there is upside risk to our fair

value projections for breakevens if it does. As Exhibit 63

shows, we expect breakevens to be hit by quite a few

negative factors: the unemployment rate will remain

at elevated levels, budget deficit expectations should

trend lower as fiscal tightening takes hold, and

headline CPI should decline sharply from its current

level. In addition, we look for nominal rates to plunge

in 1Q12 as the European crisis worsens, and then rise

over the rest of the year. Thus, early next year, negative

factors will dominate, and we expect breakevens to

narrow sharply. As the year progresses, however,

breakevens should widen out as rates rise. Thus, based on

this analysis, we recommend underweighting TIPSversus nominals over the next few months. Looking

beyond 1Q12, we would look to initiate breakeven

wideners.

A model for TIPS asset swaps

How are TIPS asset swaps likely to perform in 2012? To

answer this question, we present a simple model for TIPS

asset swaps in Exhibit 64. Since TIPS asset swaps have

no inflation risk due to the embedded inflation swap,

Exhibit 65: Because of seasonality in inflation, the value of theinflation “stub” for TIPS varies over timeValue of the inflation stub at mid-month points over the calendar year for TIPSwith different maturity months; %

For month

starting

Seasonal factor,

with lag applied Jan Feb Apr Jul

15 Jan -0.28% 0 -0.28% -0.25% 0.48%

15 Feb -0.11% 0.27% 0 0.03% 0.76%

15 Mar 0.14% 0.38% 0.10% 0.14% 0.87%

15 Apr 0.26% 0.24% -0.04% 0 0.73%

15 May 0.31% -0.02% -0.30% -0.27% 0.47%

15 Jun 0.16% -0.33% -0.61% -0.58% 0.16%

15 Jul 0.06% -0.49% -0.77% -0.74% 0

15 Aug -0.10% -0.55% -0.82% -0.80% -0.06%

15 Sep -0.16% -0.45% -0.72% -0.70% 0.04%

15 Oct -0.09% -0.29% -0.57% -0.54% 0.19%

15 Nov -0.07% -0.20% -0.48% -0.45% 0.28%

15 Dec -0.13% -0.13% -0.41% -0.38% 0.35%

Inflation stubs for TIPS lines (by maturity month)

Note: We estimate the seasonal factor as the difference between non-seasonally adjustedchanges in headline CPI minus seasonally-adjusted changes, averaged by calendar monthover 2001-10. This series is lagged by two months to account for the lag in TIPS indexationand interpolated to get mid-month values.

Exhibit 63: Our targets for breakevens in 2012J.P. Morgan forecast for the drivers of forward breakevens and projected level of breakevens based on the model in Exhibit 15; bp

Factors Current 1Q12 2Q12 4Q125Y UST y ield, % 0.92 0.75 1.25 1.25

10Y UST y ield, % 2.01 1.70 2.50 2.50

30Y UST y ield, % 3.00 2.70 3.60 3.60

Unemployment rate, % 9.1 9.0 9.0 9.0

Budget surplus as % ge of GDP, % -7.5 -6.4 -6.1 -5.4

Headline CPI, oya % 3.72 1.46 1.10 1.40

Fed's security holdings ($bn) 2625 2625 2625 2625

Illiquidity metric, bp 91.8 80.3 68.7 45.7

Breakeven Targets Current 1Q12 2Q12 4Q12

3Mx5Y breakevens, bp 174.9 120 130 140



Exhibit 64: TIPS asset swaps have been negatively correlated withthe level of breakevens10-year TIPS asset swap spread regressed against 10-year TIPS breakevensand Nov-27 P-STRIPS asset swap spread; all data are 1-week averages;weekly data over past three years; bp

-20

0

20

40

60

80

100

0 50 100 150 200 250 30010-year TIPS breakeven; bp

Y = -0.31(10Y BE)+0.46(20Y P-STRIPS) + 79.0R-sq = 77%




Jorge GarayoAC

(44-20) 7325-4820


200

they should trade similarly to asset swaps on other less-

liquid sectors of the Treasury market. Thus, we use 20-

year P-STRIPS asset swaps as one of the explanatoryvariables in our model.

In addition, although TIPS asset swaps should not have

inflation risk, we find that they have been correlated with

TIPS breakevens over recent years, with asset swaps

narrowing (i.e., TIPS richening versus swaps) when

breakevens widen. This is likely because demand for

TIPS—which will impact asset swaps—has been

correlated to inflation expectations and thus breakevens.

Given our expectation that breakevens will narrow over

the next few months and that liquidity will likely remain

constrained, we look for TIPS asset swaps to widenmodestly, though longer-maturity asset swaps will

likely retrace that widening over the course of next

year.

Trading seasonality in TIPS

Since TIPS are indexed to headline CPI, which is not

seasonally adjusted, TIPS are exposed to seasonality.

Furthermore, for a given TIPS issue, the value of

inflation seasonality varies at different points in the year,

even if we assume the seasonal pattern in inflation

remains the same over time. To see this, we consider a

hypothetical TIPS with a January 15 maturity date. If welook at this TIPS on November 15, we can think of

inflation over the remaining life of the bond as n full

years of November 15-November 15 inflation (with a net

seasonality component of zero), plus an inflation “stub”

for a November 15-January 15 period. Similarly, if we

look at the same bond on March 15, we can think of it

accruing n-1 full years of March 15-March 15 inflation

(with a net seasonality component of zero), plus a March

15-January 15 inflation stub.

In Exhibit 65, we show the value of those inflation stubs

for the different TIPS maturity months over the course of

a calendar year. Here we estimate the seasonal factors bytaking the difference between non-seasonally adjusted

changes in headline CPI minus seasonally-adjusted

changes and averaging the differences by calendar month

over 2001-10. We also lag the series by two months to

account for the lag in TIPS indexation, and then we

interpolate to get mid-month values. As the table shows,

some TIPS lines benefit from seasonality more than

others—in particular, the inflation stubs for July TIPS are

almost always positive, while the stubs for February

TIPS are almost always negative.

Although the inflation stubs for individual TIPS vary

over time, the difference between stubs for TIPS with

different maturity months (i.e., January versus April)

Exhibit 66: We estimate that the April-July seasonality differentialis worth about 36bp of yield for 2-year TIPSDifference in value of the inflation stub for various pairs of TIPS (in price terms)

as given in Exhibit 23, and in yield terms for various maturitiesJan-Apr Jan-Jul Apr-Jul

Price impact (% ) 0.25% -0.49% -0.73%

Yield impact, by maturtiy (bp)

2Y 12.2 -23.9 -36.1

5Y 4.8 -9.4 -14.2

10Y 2.5 -4.8 -7.3

20Y 1.3 -2.5 -3.8

Exhibit 67: Projected value of seasonality differentials for TIPSwith different maturities over timeDifference in value of the April-July inflation stub for bonds with different maturityyears; bp of yield

-60

-50

-40

-30

-20

-10

0

May 10 Aug 10 Dec 10 Mar 11 Jun 11 Sep 11

201320152017201920212031

Exhibit 68: The flattening of the Apr-13/Jul-13 TIPS yield curve hasbeen roughly in line with what we would expect given the changein the value of the seasonality differential

Apr-13/Jul-13 TIPS real yield curve versus projected value of the April-Julyseasonality differential; bp of yield

-55

-50

-45

-40

-35

-30

-25

-20

-30

-25

-20

-15

-10

-5

0

5

10

15

May 10 Aug 10 Nov 10 Mar 11 Jun 11 Sep 11

Apr-13/Jul-13 TIPS curve

Value of seasonality differential




Jorge GarayoAC

(44-20) 7325-4820


201

stays constant, assuming that the seasonal pattern is

unchanged. Thus, we can estimate the difference in the

value of seasonality for various pairs of TIPS, as shown

in Exhibit 66. As the table shows, April TIPS are most

adversely affected by seasonality, accruing 0.25%-pt less

inflation than January TIPS, and 0.73%-pt less inflation

than July TIPS. We can also translate this seasonality

differential into yield terms by dividing by duration. As

the table shows, the April-July seasonality differential is

worth 36bp of yield for 2-year TIPS, but only about 4bp

of yield for 20-year TIPS.

To see how the value of this seasonality differential (in

yield terms) changes as bonds age, we project the

differential for April and July TIPS with various maturity

years in Exhibit 67. For longer-maturity TIPS, the

change in the value of the seasonality differential over

time is small: for example, for TIPS maturing in 2019,

the differential was worth 8.4bp of yield 18 months ago,

and is currently worth 9.5bp. In contrast, for TIPS

maturing in 2013, the differential was worth about 25bp

of yield 18 months ago, and is currently worth 51bp. We

can see the impact of this seasonality effect on the Apr-

13/Jul-13 TIPS real yield curve—as Exhibit 68 shows,the flattening of this curve over the past 18 months has

been roughly in line with what we would expect given

the change in the value of the seasonality differential.

In practice, it may be difficult to isolate the value of

seasonal differentials, since the spread between two

issues will also be impacted by the slope of the overall

curve as well as liquidity differentials. For example, even

though April TIPS have a seasonal disadvantage

compared to July TIPS, in the 2016 sector, the April

TIPS is the benchmark 5-year and enjoys greater

liquidity, while the Jul-16 TIPS is an old 10-year issue

that is less liquid. As a result, the fact that the spread between these two issues is almost zero likely indicates

that the liquidity premium for Apr-16 TIPS counters the

seasonality disadvantage. In addition, at the very front

end of the TIPS curve, investors’ near-term inflation

forecasts are likely to have a greater influence on spreads

than backward-looking estimates of seasonality.

With these constraints in mind, we think it is still

possible to find instances where investors can take

advantage of seasonal differentials. For example, we

estimate that Jul-16 TIPS should trade about 11bp rich to

Jan-16 TIPS based on seasonals, but they are currently

about 1bp cheap. In addition, this curve currently lookstoo steep relative to the broader Jan-15/Jan-19 curve, and

liquidity differentials should be a minor issue, since both

Exhibit 69: TIPS breakevens have closely tracked equities over thepast two years10-year TIPS breakevens versus S&P 500;

bp level

1000

1100

1200

1300

1400

140

160

180

200

220

240

260

280

N ov 09 Feb 10 J un 10 Sep 10 Dec 10 Mar 11 J ul 11 Oc t 11

10-year TIPS breakevens

S&P 500

Exhibit 70: Going short breakevens versus the S&P 500 whenbreakevens look wide has been more profitable than going longbreakevens when breakevens look narrowStatistics for long or short breakeven trades initiated when the residual* reached acertain level and hedged with the opposite position in S&P 500**; 11/09-10/11

Long BE if

residual (bp) <

#

trades

Avg P/L

(bp)

#

winners

Hit

ratio

Avg gain

(bp)

Avg loss

(bp)

-10 125 -0.6 64 51% 9.5 -11.2

-15 85 -0.5 45 53% 8.7 -10.8

-20 421.9

2662%

8.8 -9.4

-25 11 6.4 10 91% 8.3 -11.8

Short BE if

residual (bp) >

#

trades

Avg P/L

(bp)

#

winners

Hit

ratio

Avg gain

(bp)

Avg loss

(bp)

5 130 6.8 82 63% 15.8 -8.6

10 70 9.3 49 70% 17.3 -9.4

15 32 16.3 27 84% 20.2 -4.4

20 10 18.0 10 100% 18.0 N/A

* Backward-looking 3-month residual of the level of breakevens versus the level of S&P 500.** The hedge ratio is given by the backward-looking 3-month beta of weekly changes inbreakevens versus weekly changes in the S&P 500.Note: P/L calculated as change in breakevens minus (hedge ratio)*(change in S&P) plus 1-month ex-ante breakeven carry for long breakeven positions, and the inverse of that quantityfor short breakeven positions.

Exhibit 71: Going long breakevens when carry is attractive andbreakevens look narrow has been profitable…Statistics for long 5-year breakeven trades (held for one month) initiated when ex-ante 1-month carry was above a certain level and the rolling 6-month residual of 5-year breakevens regressed against 5-year nominal yields and WTI oil futuresprices was below a certain level; 1/09-10/11

Carry

(bp) >

Residual

(bp) <

#

trades

Avg carry

(bp)

Avg P/L

(bp)

#

winners

Hit

ratio

Avg

gain (bp)

Avg loss

(bp)

3 0 88 8.0 6.9 56 64% 19.1 -14.4

3 -2 77 7.7 6.5 48 62% 19.4 -14.9

5 0 56 10.4 6.2 34 61% 17.7 -11.6

5 -2 47 10.3 5.6 28 60% 17.4 -11.9




Jorge GarayoAC

(44-20) 7325-4820


202

issues are old 10-year TIPS. Thus, we think

overweighting Jul-16 TIPS versus Jan-16 TIPS is

currently an attractive way to take advantage of seasonality differentials. Such trades, based on the theme

of identifying seasonality mispricings, will be a key

element of our TIPS strategy in 2012.

Relative value trading themes for 2012

Trading breakevens versus the S&P 500

Although we expect breakevens to narrow in early 2012

as the European crisis worsens, one way we can protect

against being wrong is to trade breakevens on a hedged

basis—namely, by trading breakevens versus equities.

Over recent years, the correlation between breakevens

and risky assets has been striking, with breakevenswidening in “risk-on” trades and narrowing in “risk-off”

trades. In particular, breakevens have closely tracked the

S&P 500 over the past two years (Exhibit 69). Given this

strong correlation, we would expect trading breakevens

versus a beta-weighted amount of equities when

breakevens look either very narrow or very wide relative

to equities to be an attractive trading theme.

We test this hypothesis by back-testing trades over the

past two years as follows: each day, we calculate the

backward-looking 3-month residual of the level of

breakevens versus the level of S&P 500. If the residual is

below (above) a certain threshold, we initiate a long(short) breakeven position and hedge with the opposite

position in the S&P 500. The hedge ratio is given by the

backward-looking 3-month beta of weekly changes in

breakevens versus weekly changes in the S&P 500.

Exhibit 70 presents the results of our analysis. We would

expect the strategy to work equally well when

breakevens look either too wide or too narrow, but we

find that over this period, going long breakevens hedged

for the S&P when breakevens look narrow has generally

been unattractive, unless breakevens look extremely

narrow. In contrast, going short breakevens on a hedged

basis when breakevens look too wide has been

profitable—these trades have had high hit ratios andattractive P/Ls. This asymmetry could be a result of our

observation that equities and inflation expectations have

been more correlated during deflationary periods than

inflationary periods in recent years (see US Cross Sector

Overview).

Thus, one trading theme we like for 2012 is

opportunistically selling breakevens versus a hedged

amount of S&P 500 futures when breakevens look too

rich versus equities.

Earning carry when breakevens look narrow

A second trading theme we like is opportunistically

going long breakevens when carry is attractive and

breakevens look narrow relative to fair value. To see this,

we back-tested a strategy under which long breakeven

trades were initiated for a 1-month period when 1) 1-

month carry exceeded a certain level and 2) the

backward-looking 6-month residual of breakevens

regressed against nominal yields and oil prices fell below

a certain threshold. In Exhibit 71, we present statistics

for these trades over January 2009-October 2011. As the

exhibit shows, this strategy generated hit ratios around

60% and average 1-month P/L of 5-7bp.

We found that this strategy was even more effective

using inflation swaps—specifically, receiving

inflation/paying fixed when carry was attractive and

inflation swap rates looked too low relative to fair value.

As Exhibit 72 shows, using inflation swaps produced

significantly higher hit ratios and average P/Ls than the

strategy using breakevens. Even if we account for the

wider bid-offer in inflation swaps, the derivatives version

of this strategy still appears more attractive than the cash

version.

Thus, a second trading theme we like is receiving

inflation/paying fixed in swaps when carry is

attractive and inflation swap rates look too low.

Exhibit 72: …and this strategy works even better with inflationswapsStatistics for receiving inflation/paying fixed positions in 5-year or 3-year inflation

swaps (held for one month) when ex-ante 1-month carry was above a certain leveland the rolling 6-month residual of 5-year or 3 -year inflation swaps regressedagainst 5-year or 3-year nominal yields and WTI oil futures prices was below acertain level; 1/09-10/11

Carry

(bp) >

Residual

(bp) <

#

trades

Avg carry

(bp)

Avg P/L

(bp)

#

winners

Hit

ratio

Avg

gain (bp)

Avg loss

(bp)

5-year swaps

3 -6 63 7.6 15.2 54 86% 19.3 -9.4

3 -14 25 7.2 15.7 22 88% 19.9 -14.7

5 -6 45 9.1 12.6 38 84% 16.4 -7.8

7 -6 29 10.9 10.3 23 79% 15.4 -9.1

3-year swaps

7 -6 42 16.5 13.0 32 76% 19.5 -8.1

7 -10 25 17.2 14.8 20 80% 20.4 -7.8

9 -8 27 18.3 14.7 22 81% 20.1 -8.9

11 -8 20 21.2 12.7 15 75% 19.9 -8.9




Jorge GarayoAC

(44-20) 7325-4820


203

Medium term themes for inflation

markets

Euro area

• Favour breakeven flattening trades

Even if inflation should fall, we see upside risks to

prints in 1H12 vs. expectations, while the long end

likely to be driven by sovereign risk dynamics which

should flatten nominal yield curves. Favour 1s/10s

flatteners in HICP swaps, overweight OBLi 13 vs.

BTPei 19 breakevens or BTPei 16s vs. BTPei 21

breakevens.

• Overweight French CPI-linked vs. Euro HICP-

linked lines into early 2012French-CPI linked should perform well in 1Q12 when

we expect further hedging from Livret A accounts and

support from Libor-based domestic investors.

• Italian real yields to be under pressure

Real yields can reach 10% on the back of our

peripheral spread views as well as risks that Italian

linkers are dropped from linker benchmarks.

UK

• Be long real yields in intermediate part of the curve

Our model predicts 5Y real yields to fall 30-40bp fromcurrent levels. Long positions in 5Y and 10Y linkers

offer attractive carry and slide vs. the forwards.

• Short 10Y breakevens

Nominal gilts will continue to be the major driver of

breakevens. Our predicted 75bp fall in 10Y gilt yields

during 1H12 (to the 1.50%) level translates into a

30bp fall in ILG22 breakevens, to the 2.30% level.

• Go short real yields into linker syndications as a

tactical trading strategy

A trading rule selling the closest maturity bond to the

linker being issued via syndicate 15 business daysahead of syndications has been profitable.

US

• Look for TIPS breakevens to narrow sharply over

the next few months and then widen modestly over

2012

Breakevens will hit a number of headwinds in early

2012: we expect nominal yields to plummet, fiscal

policy to tighten, and headline inflation to fall. Thus,

our model projects that breakevens should narrow

over the next few months, though breakevens should

widen over the remainder of the year as nominal rates

rise. QE3 also poses an upside risk to our targets.

• Overweight TIPS with attractive seasonals versus

TIPS with unattractive seasonals when mispricings

existWe recommend taking advantage of seasonality

differentials among TIPS that are close in maturity

when mispricings exist.

• Opportunistically trade breakevens versus the S&P

500 when breakevens look too wide

With correlations expected to remain high as long as

European risks remain significant, trading breakevens

hedged for moves in the S&P 500 is likely to prove

attractive, as it has in recent years.

• Look to tactically trade breakevens and inflation

swaps for carryWe find that strategies involving receiving inflation

when carry is attractive and swap rates look low have

been profitable over recent years and recommend that

investors use these strategies to earn carry.

Cross market themes

• Generally overweight US inflation breakevens vs.

Euro area equivalents in 10Y+, vs. doing the

opposite at the front end

This is based on our relative inflation views, the risk

of sovereign dynamics impacting Euro breakevens

more than US equivalents.




Sally AuldAC

(61-2) 9220-7816

[email protected]. Morgan Securities Australia Limited

204

Australia

• J.P. Morgan economists expect the Australian

economy to grow 3.0% in 2012, but risks are

biased to lower growth outcomes given the fragile

global environment. Against this backdrop, J.P.

Morgan economists expect the RBA to deliver

50bps of easing in Q1 next year

• While the domestic economic outlook might

suggest the potential for modestly higher yields,

the experience of 2011 suggests that domestic

drivers may not be the main influence on term

yields

• As such, the bearish outlook we retain on Europe

means that we recommend extending duration on

dips in AUD fixed income. Look to lighten

duration in 2H12, given the risks that the market

prices in a more optimistic global outlook

• We target the 3s/10s bond curve spread to make

new highs in 2012. Investors trading long

duration should implement cash bond curve

flatteners in order to gain some portfolio

protection

• We think swap spreads will be biased wider in

2012, with the spread curve likely to flatten. We

like owning cash spread product against swap

rather than bond, and prefer NSWTC in the

semi-government space

• Look to buy the NSWTC Nov-2035i against the

ACGB Sep-2030i. The ACGB Aug-2020i looks

cheap against the UST Jul-2021i; real yield

spreads look wide given AUS-US growth

differentials

Economy expected to improve in 2012,

but global headwinds are fierce

2011 was a year of disappointment on the

macroeconomic front, at least relative to expectations.

Consensus forecasts now expect that Australia will likely

record growth of 1.5-2.0% in 2011, well below trend.

This was a lower number than most expected at the

beginning of the year, but the disappointment was largely

due to the impact of severe floods in Queensland and a

slower-than-anticipated recovery from the natural

disaster.

The outlook for Australia in 2012 is more optimistic,

with the ongoing recovery in coal exports expected to see

solid growth momentum in 1H12. J.P. Morgan expects

growth of 3.0% for the calendar year; the RBA expects

4.0% growth, although we would argue that this looks a

little ambitious given risks surrounding the European

(and global) outlook. While Australia’s direct trading

links with Europe are small (around 8% of Australianexports go to Europe), the indirect links via Asia are

reasonably strong. As the RBA noted in its most recent

quarterly Statement on Monetary Policy:

“The possibility of a sharp economic deterioration in

Europe represents a downside risk for the Australian

economy. Given strong trade links with Asia, it is likely

that Australia would be less directly affected than some

other countries by a deterioration in Europe, although

the economy would still be affected through falls in asset

Exhibit 1: Australia’s (export weighted) trading partner growth –forecast to be below trend in 2011 and 2012Calendar year % growth

0.0

1.0

2.0

3.0

4.0

5.0

6.0

7.0

1995 1998 2001 2004 2007 2010

Average

Exhibit 2: Movements in the RBA cash rate generally correlate well withglobal GDP growth%-pts Qtrly. annualised %

-5.5

-4.5-3.5

-2.5

-1.5

-0.5

0.5

1.5

2.5

Mar-98 Dec-00 Sep-03 Jun-06 Mar-09 Dec-11

-10.0

-8.0

-6.0

-4.0

-2.0

0.0

2.0

4.0

6.0

8.0

RBA cash rate, annual

change, lhs

Global GDP, rhs, pushed

fwd 2q




Sally AuldAC

(61-2) 9220-7816


205

prices and weaker household and business confidence.

Commodity markets could be expected to weaken and

growth in domestic incomes would be lower .”

The clear message is that Australia will not be immune to

a severe economic contraction in Europe. Against this

backdrop, J.P. Morgan economists are forecasting the

RBA to deliver modest easing in early 2012 and expect a

cash rate of 3.75% by March next year. The risks to this

view are clearly biased to the downside, given fragility in

the global growth environment and risks around an

escalation in the European crisis. Indeed, our estimate of

Australia’s trading partner growth in 2012 suggests little

change from 2011 (Exhibit 1), with trading partner

growth at 4.2% (below the average of 4.9%). It is also

worth noting that movements in the RBA cash rategenerally correlate well with the global growth outlook

(Exhibit 2).

Of course, the main medium-term influence on the

Australian macro-economic outlook continues to be

China, and while our economists have recently shaved

their growth forecasts for China, they still expect a

reasonably positive outcome in 2012 (growth of 8.3%).

As Exhibit 3 illustrates, Australian and Chinese GDPs

have been well correlated over the past 5-6 years, with

growth in China leading Australian growth by around 3

months. Although our expectation for Chinese economic

growth would be historically consistent, with Australiangrowth around 2.0%, the consensus expects strong capex

growth and a rebound in flood-affected exports to

temporarily drive a divergence between Australian and

Chinese growth outcomes (hence the disconnect between

Australian and Chinese growth forecasts in Exhibit 3).

Overall, our forecasts for Australian growth next year

represent a cautiously optimistic outlook, especially

given global risks and our forecasts for another year of

below-trend trading partner growth. Accordingly, we

believe risks are biased to the downside to our growth

forecasts. This is especially the case given a reasonably

subdued pulse in the domestic dataflow of late.

Can the RBA validate front-end

pricing…another 165bp of rate cuts?

As noted above, J.P. Morgan economists expect the cash

rate to reach 3.75% by March 2012. We still think the

risks to this view are to the downside, given the situation

offshore. And while the RBA may not yet believe it has

begun a traditional easing cycle, we would expect the

RBA’s policy stance to retain a degree of optionality in

the months ahead. But the market is already pricing in

this risk; Exhibit 4 illustrates that the OIS curve is

pricing in 165bp of rate cuts over the next 12 months.

Exhibit 5 illustrates that since early August, the markethas oscillated between pricing in 100bp and 165bp of

rate cuts from the RBA (over a 12-month horizon). Thisis a distinct change from the period January–August2011, when the market priced anywhere from 0-50bp of rate hikes.

Investors may be tempted by the attractive carry in paid

positions in the very front end of the Australian curve;

for example, a paid position in AUD 1Y swap currently

offers 30bp of carry and roll over a 3-month horizon. But

assuming the RBA does not validate current market

pricing anytime soon and with no ‘end-game’ solution in

Exhibit 3: Chinese and Australian GDP growth have been wellcorrelated over the past 6 yearsDotted lines represent J.P. Morgan forecasts

oya % oya %

6.0

8.0

10.0

12.0

14.0

16.0

Dec-04 Apr-06 Aug-07 Dec-08 Apr-10 Aug-11 Dec-12

0.5

1.6

2.7

3.8

4.9

6.0Chinese GDP, lhs, pushed fw d 1q

Australian GDP, rhs

Exhibit 4: The market is pricing in over 150bp of rate cuts from theRBA…

AUD OIS curve, %

2.50

2.75

3.00

3.25

3.50

3.75

4.00

4.25

4.50

4.75

Current Feb-12 Apr-12 Jun-12 Aug-12 Oct-12




Sally AuldAC

(61-2) 9220-7816


206

Europe on the horizon, we see little reason for current

market pricing to change materially into 2012 given:

1) Ongoing concerns surrounding Europe, China

and global growth;

2) Data suggesting that while the Australian

economy is not on the verge of a recession, nor

is it showing any sign of a significant uptick in

momentum;

3) Technical aspects to the OIS market which

currently bias yields lower (domestic banks are

structural receivers of OIS given their need to

hedge deposit books); and

4) The fact that the RBA could – if needed – cut

rates by a lot, relative to other developed market

central banks.

Interestingly, the cycle trough in 3Y yields has often

been a good indicator of the subsequent trough in the

cash rate (2001 and 2008). Exhibit 6 illustrates that with

3Y government bonds currently yielding 3.25%, this

would be consistent with a further 125bp of easing from

the RBA. It also implies that there is probably 40–50bp

of risk premium in the very front end of the AUS yield

curve at present.

The outlook for outright yields – lower-

yield highs and lower-yield lows in 2012

as domestic fundamentals continue to

take a back seat

Broadly speaking, the trend over the course of 2011 was

towards lower yields in Australia, with a rally in global

bond markets, some disappointment on the domestic and

global growth fronts, and a November rate cut from the

RBA being the main drivers of lower yields. A persistent

bid from offshore investors also helped to suppress yields

(see our analysis below for more detail on this theme).

Australian 3Y bond yields have traded a 215bp range

over the course of 2011, and at 3.23% at the time of

writing, sit at the lows of the year (Exhibit 7).

Remarkably, the calendar year yield high for 3Y bonds

has been very similar over the past 3 years, at around

5.35%. Exhibit 7 illustrates the structurally lower trading

range for 3Y yields post the financial crisis; with

offshore risks mounting, we think it is likely that

Australian yields post new lows in 2012.

Exhibit 5: …and has been pricing in cuts since early Augustbp of tightening* priced in over the next 12 months

-175

-125

-75

-25

25

75


* negative number implies easing

Exhibit 6: Even term yields suggest the possibility of more easing – thelevel of 3-year bond yields suggest the risk of rate cuts, if history is anyguide%

2.0

3.0

4.0

5.0

6.0

7.0

8.0

Jan-00 Feb-02 Mar-04 Apr -06 May-08 Jun-10

3-year bond yield

RBA cash rate

Exhibit 7: 3-year government bond yield calendar year trading rangeshighlight a structurally lower trading range post crisis and the risk of new yield lows in 2012Trading range for 3Y government bond yields over the course of 2011; %

2.5

3.0

3.5

4.0

4.5

5.0

5.5

6.06.5

7.0

7.5

2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011




Sally AuldAC

(61-2) 9220-7816


207

Even the 10Y bond has traded within a 175bp range over

the course of the year, with a yield high for the year of

5.75% also close to the yield highs seen in 2009 and2010 (Exhibit 8).

With outright yields at the lows for the year (and not far

from the yield-lows seen since 2009) and with J.P.

Morgan economists expecting a pick-up in growth

momentum next year, it might be intuitive that short-

duration strategies are the correct view on Australian

bonds in 2012.

However, the message from 2011 has been that

domestic fundamentals have not been the main driver

of yields over the course of this year – indeed, Exhibit

9 illustrates the divergence between domesticfundamentals and the level of 3Y bond yields. Against

this backdrop, we think there are a number of reasons

why investors are not likely to be rewarded for short-

duration strategies in the coming year:

1) Our forecast for US 10Y yields does not

envisage a significant selloff, at least for the

first half of calendar year 2012. After an early

rally to 1.7%, we expect the UST 10Y yield to

oscillate around the 2.5% level for much of the

year, limiting the extent to which Australian

bond yields are likely to rise.

2) Domestic real money accounts are already

running short-duration positions and have done

so for much of 2011 (Exhibit 10), and as such,

we believe there will be little appetite to shorten

the duration of portfolios further from current

levels.

3) Our European rates strategy team remains

extremely cautious on the outlook for the

European peripherals in 2012. We think

developed market bond yields will struggle to

selloff significantly in such an environment.

Furthermore, this backdrop will reinforce the

attractive risk/reward in owning AAA-rated

Australian sovereign risk at (relatively)

attractive yield levels.

4) Following on from point 3, we expect offshore

demand for Australian government bonds to

remain solid in 2012. As we note below, 71% of

ACGBs are owned by foreign investors (as of

June 2011).

Exhibit 8: The 10Y bond has traded within a 175bp range over thecourse of the year, with a yield high for the year of 5.75% also close tothe yield highs seen in 2009 and 2010Trading range for 10Y government bond yields over the course of 2011; %

3.5

4.0

4.5

5.0

5.5

6.0

6.5

7.0

2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011

Exhibit 9: Australian bond yields have not traded in line with domesticfundamentals in 2011…% oya %

2.0

3.0

4.0

5.0

6.0

7.0

Mar-99 Dec-01 Sep-04 Jun-07 Mar-10

0.0

3.0

6.0

9.0

3y bond yield, lhs

Real GDP + CPI (2q ave. ), lagged

2 qtrs, rhs

Exhibit 10: …but this has not prevented domestic real money managersfrom trading short duration for most of 2011

Actual portfolio duration relative to benchmark index, %

86

90

94

98

102

106

110

Nov -06 Dec-07 Jan-09 Feb-10 Mar-11




Sally AuldAC

(61-2) 9220-7816


208

Of course, our fair value models for both 3Y and 10Y

government bonds continue to suggest that current yield

levels for AGCBs should be regarded as rich. Exhibit 11 shows our fair value for 10Y bond yields. For 2012, we

see fair value around 4.75%.

But with higher-than-usual levels of uncertainty

persisting offshore and ongoing demand for yields from

foreigners, the valuation richness of the ACGB market

may be a constant theme in 2012. In 2011, investors have

not been paid to position portfolios on domestic

fundamentals alone, and we suspect this will continue to

be the case in 2012 given our global strategy views.

Indeed, movement in a weighted European peripheral

10Y swap spread has accounted for almost 86% of the

moves in AUD 1y/1y swap over the course of 2011.

Our preferred duration strategy for 2012 would be to

use any selloff in bond markets to enter long-duration

portfolio structures. Indeed, we expect the next couple

of months to provide good opportunities for investors to

initiate long-duration positions. Short-risk trades are

likely to be unwound going into the year-end as money

managers, hedge funds, and dealers alike pare back their

balance sheet. Towards 2H12, we would look to lighten

duration given the risk that the market could start to price

in a more optimistic global outlook.

Against the US, we see Australian 10Y bonds as cheap at present. Our medium-term model of the AUS-US 10Y

bond spread suggests that the fair value for the spread is

probably around the 175bp level for 2012 (Exhibit 12).

The spread is currently trading at 205bp, and we suspect

that AUD 10Y bonds will be well bid on any widening of

this spread. Increasingly, relative (and not absolute) yield

levels are becoming the driver of cross-border bond

market flows.

Offshore investment becomes a key

feature of Australian fixed income

markets and is here to stay

One of the features of the Australian government bond

market in recent years has been strong demand for

ACGBs by offshore investors. Indeed, foreign buyers

have accounted for much of the additional supply in

ACGBs (Exhibit 13). Offshore holders now own 71% of

ACGBs, just shy of all-time highs (Exhibit 14). The

solid demand for ACGBs in recent years reflects a

number of factors:

Exhibit 11: Our fair value model suggests current 10Y governmentbond yields are rich. For 2012 we see fair value around 4.75%10Y government bond yield, actual and model*; %

2.0

4.0

6.0

8.0

10.0

12.0

14.0

16.0

Sep-88 Mar-93 Sep-97 Mar-02 Sep-06 Mar-11

Actual

Model

+/- 2 s.d.

* model is 10yr yield = 2.52 + 0.60*90-day rate + 0.31*moving avg. of nominal GDP –4.49*% of ACGBs held offshore; R-squared is 0.92, standard error is 0.72ppts.

Exhibit 12: Australian 10Y bonds still look cheap vs. US 10Y notes onour medium-term modelQuarterly data for spread between 10Y Australian government bonds and USTreasuries, actual and model*; bp

0

50

100

150

200

250

300

Jun-96 Apr-99 Feb-02 Dec-04 Oct-07 Aug-10

Actual

Model

* model is 10yr spread = 58.13 + 5.32*real GDP growth spread (pushed fwd. 4q) +30.98*1y swap yield differential (%-pts); R-squared is 0.83, standard error is 0.29pptsLatest read for the actual spread is 30-Sep-11.

Exhibit 13: Foreign buying of ACGBs has accounted for much of theadditional supply in recent years

AUDbn

0

50

100

150

200

Dec-88 Jun-92 Dec-95 Jun-99 Dec-02 Jun-06 Dec-09

ACGBs (offshore holdings)

Total ACGBs outstanding




Sally AuldAC

(61-2) 9220-7816


209

1) Reserve diversification by central banks into

AUD;

2) A largely non-leveraged investor base with

longer investment horizons (central banks and

offshore pension funds) and a small level of

holdings in absolute terms pre-crisis ($34.4bn as

of December 2007);

3) Wide yield differentials with most developed

market economies and outperformance by the

Australian economy (suggesting potential for

AUD appreciation); and

4) A strong fiscal situation in Australia.

When we look at the relative rankings of AAA-rated

sovereigns1, Australia should definitely be considered a

‘strong’ AAA-rated sovereign. And when one considers

the additional yield pick-up, it is not difficult to see why

offshore capital has found a home amongst ACGBs

(Exhibit 15).

As we have noted above, ongoing foreign demand for

ACGBs is one reason we do not expect a significant sell-

off in AUD rates. Indeed, we can estimate the extent to

which changes in offshore participation can impact the

level of bond yields. As we note in the footnote to

Exhibit 11, we include the proportion of ACGBs held byoffshore investors as an explanatory variable in our

medium-term, fair-value model for Australian 10Y bond

yields.

Our other inputs are a moving average of nominal GDP

growth (to capture the influence of domestic economic

fundamentals on long-end yields) and the 90-day rate (to

capture the influence of short-term interest rates on the

level of term yields). The model statistics are outlined in

Exhibit 16.

The model suggests that a 1-ppt rise in offshore

ownership will see 10Y yields decline 4.5bp. This

suggests that the rise in offshore holdings though 2010

and 2011 has lowered 10Y yields by around 34bp, all

other things being equal.

We think that much of rise in offshore demand for

ACGBs has been driven by central bank reserve

diversification. This would be consistent with recent data

1 For more details on the analysis behind the ranking of AAA-rated

sovereigns see S. Auld, The Antipodean Strategist , 18 August 2011

from the IMF’s composition of foreign exchange

reserves, which suggests that non-G4 currencies now

account for 5% of reserves, up from just over 2% in

Exhibit 14: Offshore holders now own 71% of ACGBs, just shy of all-time highsProportion of ACGBs held by offshore investors; %

0

10

20

30

40

50

60

70

80

Sep-88 Jun-93 Mar-98 Dec-02 Sep-07

Exhibit 15: Australia is one of the stronger AAA-rated sovereigns, yet ithas the highest 10Y bond yields among all AAA-rated peers*10-year sovereign bond yield vs. standard deviations away from AAA median**

0.51.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

-0.6 -0.3 0.0 0.3 0.6 0.9 1.2

Std. Devs . aw ay from AAA median

CHF

AUSNZ

UKGER

Stronger AAAWeaker AAA

* we include New Zealand even though Moody’s is the only rating agency to give thesovereign an AAA credit rating.** based on a model based on real GDP growth and 1Y swap y ield differentials.

Exhibit 16: Our models for 10Y yields suggest a 1% increase in offshoreownership will result in yields declining 4.5bpModel statistics for our 10Y ACGB model

Model Statistics Value

Intercept 2.5

Nominal GDP (moving ave.) 0.6

90-day rate 0.3

% of offshore holdings -4.5

R-squared 0.92

Standard error (ppts) 0.7




Sally AuldAC

(61-2) 9220-7816


210

2009. And even with much of the new allocation to AUD

now out of the way, quarterly net buying of ACGBs still

remains well above average.

But the other driver of offshore demand for ACGBs

continues to be the wide yield differential – both in

nominal and real terms – between Australian government

bonds and bonds in most developed market economies.

Exhibit 17 shows that the 10Y real yield spread between

Australia and the US has a leading relationship with

changes in offshore holdings of ACGBs. If anything, the

chart suggests that current levels of foreign demand look

a touch low relative to rate spreads. And with the outlook

for G4 government bond markets consistent with

relatively low yields for a long time, we expect the ‘real

money carry trade’ to support foreign sponsorship of theACGB market.

Curve spreads: the 3s/10s spread has

scope to steepen further

Generally speaking, the shape of the yield curve over

time in Australia is determined by movements in the

RBA’s cash rate. Exhibit 18 illustrates the point –

investors who think the RBA will validate market pricing

should clearly position for a steeper curve spread.

Indeed, our fair value model for the 3s/10s bond spread

suggests that the curve has traded flat relative to fair value for some time (Exhibit 19). There are possibly two

reasons for the persistent cheapness of the curve – it

could be a reflection of strong demand for AUD duration

from offshore (as discussed above), or the negative carry

inherent in curve steepeners may prevent investors from

positioning for a steeper curve.

Given our expectation of rate cuts from the RBA, we

believe the 3s/10s bond spread will make new highs for

the cycle in Q1 next year and target a move to +105bps.

The carry and roll associated with cash bond steepeners

is very negative. Accordingly, we think there is better

value in entering forward starting steepeners; forward

curve spreads are very flat, suggesting that investors

biased towards steeper curves should easily beat the

forwards.

The least carry and roll negative forward starting

steepener which offers some roll-up to spot is the 3s/7s

swap curve steepener, 1Y forward. The 1Y forward

spread is around 4bps flatter than the spot spread, and the

carry and roll on the trade costs 4.8bps over a 3-month

Exhibit 17: Real yield spreads are a key driver of offshore demand for ACGBs%-pts %-pts

-15

-10

-5

0

5

10

15

20

25

30

Sep-97 Aug-00 J ul-03 J un-06 May -09

-1.0

-0.5

0.0

0.5

1.0

1.5

2.0

2.5

3.0 Annual change in offshore holdings of ACGBs, lhs

10y AUS-US real yield spread, rhs, 6 mths fw d

Exhibit 18: The policy cycle tends to be the medium term driver of movements in the 3s/10s bond curve in Australia%-pts %

-0.8

-0.4

0.0

0.4

0.8

1.2

1.6

Aug-98 J un-01 Apr-04 Feb-07 Dec -09

2.00

3.00

4.00

5.00

6.00

7.00

8.00

AUS 3s10s curve, pushed fwd. 3 months, lhs

RBA cash rate, rhs, inv erted

Exhibit 19: The 3s10s bond curve still looks a touch flat relative to our fair value model3s/10s bond curve actual and model*; %-pts

-1.25

-0.75

-0.25

0.25

0.75

1.25

1.75

Oct-01 Apr-04 Oct-06 Apr-09 Oct-11

Actual

Model

* model is 3s/10s bond spread = 2.20 + 0.07*1s6s AUD OIS curve (%-pts) - 0.41*3ybond yield + 0.18*US2s10s bond spread; R-squared is 0.88, standard error is0.15ppts.




Sally AuldAC

(61-2) 9220-7816


211

horizon. The 1Y forward spread is currently 66bps; we

target a move to +90bps.

Investors already trading long duration

should look to use cash bond flatteners

as a means of portfolio hedging

Given the prospect of difficult trading conditions in

2012, we think it is wise to think about trades which are

1) positive carry, and 2) not so correlated with moves in

European peripheral spreads. In the Australian market,

carry-efficient curve trades offer good opportunities for

investors trading long duration looking for some

portfolio diversification or protection in the event that

peripheral spreads narrow. Exhibit 20 illustrates that the

3s/10s curve has been reasonably invariant to moves in peripheral spreads over the course of the year.

This is probably contrary to expectations – most would

have expected the curve to bull-steepen in a risk-averse

environment and when the RBA is easing policy. As we

note above, we think the most logical explanation for this

is constant demand for AUD duration from offshore.

The positive carry associated with cash bond flatteners

means that curve flattening trades can offer some

protection for investors trading long duration. In the cash

space, the most carry-efficient flatteners involve shorting

April 2012 bonds to buy April 2023 or April 2027 bonds

(around 15bp of carry over a 1-month horizon); based off

the relationship in Exhibit 20, this suggests the flattener

offers protection against around 215bp of 10Y peripheral

spreads narrowing (over 1-month).

Swap spreads should be biased wider in

2012, 3s/10s spread curve likely to

flatten

3Y and 10Y swap spreads have drifted wider over the

course of 2011 (Exhibit 21), and we would expect the

higher trading range to persist into 2012. We expectspreads to make new wides next year, and target 3Y

spreads to trade a 50bp to 85bp range, and 10Y spreads

to trade a 55bp to 90bp range in 2012. We expect spreads

to retain strong directionality next year, with spreads

widening on a rally and narrowing on a selloff.

We still believe that a paid position in AUD 3Y swap

spreads is a good way in which to position for still-wider

peripheral spreads in Europe; Exhibit 22 illustrates the

strong correlation between the two series. Aside from

Exhibit 20: Even though the 3s/10s bond curve has shown somecorrelation to peripheral spreads this has been reasonably limited3s/10s bond curve vs. 10Y weighted peripheral swap spread*, %-pts and bp

y = 0.0007x + 0.2384

R2 = 0.5181

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

200 300 400 500 600 700 800

10y w eighted peripheral swap spread

* A weighted peripheral spread computed against Germany for Ireland, Portugal, Italy,Spain and Greece (weighted by the size of their outstanding bond markets)

Exhibit 21: AUS 3Y and 10Y swap spreads have drifted wider over thecourse of the year Swap spread measures; bp

10

20

30

40

50

60

70

80

Jan-11 May -11 Sep-11

3y swap spread

10y swap spread

Exhibit 22: AUS 3Y swap spreads exhibit strong directionality with 5Yperipheral spreads*bp bp

200

400

600

800

1000


20

35

50

655-year weighted

peripheral spread tobund, lhs3-year sw ap spread,

rhs

* as defined in the footnote in Exhibit 20.




Sally AuldAC

(61-2) 9220-7816


212

reflecting heightened banking system risk, wider swap

spreads should also reflect strong demand for Australia’s

AAA-rated sovereign paper in times of market stress(especially stress that is driven by sovereign debt

concerns). And if peripheral spreads move wider and

markets remain illiquid, then AUD corporate and

Kangaroo issuance will likely be limited in 2012,

implying the absence of issuance-related, receive-side

flow in the swap market. Indeed, year-to-date issuance in

Australia has been the lowest since 2008 for all spread

products (Exhibit 23).

Despite our view of wider swap spreads, there are a

couple of factors which might work to cap the move

wider in spreads next year. The first is further rate cuts

from the RBA. Generally speaking, a lower cash rate has been associated with narrower 3Y swap spreads (less

demand for fixed-rate mortgage products implies less

demand from mortgage hedgers). But we expect this

effect to be limited given the likelihood of an escalation

of the European crisis.

The second is the potential for AUD/JPY-related

receiving in the long end (10Y sector) of the swap curve;

generally speaking, a lower AUD/JPY exchange rate

generates receiving of longer-dated swaps due to the

hedging needs inherent in the issuance of callable power

reverse dual currency notes2.

Taking into account all the varied influences on swap

spreads next year, we are biased towards a flatter 3s/10s

swap spread curve.

Macro-prudential regulation continues

to be the major structural influence on

relative value in AUD cash spread

product

In terms of cash spread product, regulation continues to

be the major structural influence upon the spread product

market in Australia. In 2011, the major development wasgreater clarity on APS210, the regulatory standard which

defines the type of assets that can be held in domestic

bank liquidity books. In addition, the Australian bank

regulator (APRA) and the RBA announced the

establishment of a committed secured liquidity facility

(CLF) in order to facilitate compliance by Australian

2 See S. Auld, Why does a higher AUD/JPY FX rate generate paying in

AUD long end swap yields and AUD/USD cross currency basis

spreads?: Explaining the impact of Callable Power Reverse Dual

Currency Notes, 21 April 2011).

Authorised Deposit-Taking Institutions (ADIs) with the

new macro-prudential framework outlined by Basel III.

The purpose of this facility is to enable an ADI to

establish a committed secured liquidity facility with the

RBA, sufficient in size to cover any shortfall between the

ADI’s holdings of high-quality liquid assets and theliquidity coverage ratio requirement (LCR). APRA has

determined that the only assets that qualify as Level 1

(high quality liquid) assets are cash, balances held with

the Reserve Bank of Australia, and Commonwealth

Government and semi-government securities. APRA has

announced that (at this point in time) there are no assets

that qualify as Level 2 assets. This implies that other

AUD AAA-rated assets, such as supra-national bonds,

covered bonds and RMBS, will only qualify towards the

LCR via the CLF.

Exhibit 23: In 2011, $A issuance volumes are likely to be the lowestsince 2008

AUDbn

0

20

40

60

80

100

120

2008 2009 2010 2011 YTD

Financials SSA Corporate

Exhibit 24: The market has made a clear distinction between Europeanand non-European supra-national paper this year

Asset swap levels (semi/quarterly basis), bp*

-10

20

50

80

110

140

170

200

Apr-2012 Dec-2014 Sep-2017 Jun-2020 Mar-2023

ASIA EIB EUROFIADB IFC KFWRENTEN

*Levels as of 17-November-2011




Sally AuldAC

(61-2) 9220-7816


213

The ADI will pay a 15bp fee to access the CLF. Given

current market spreads and RBA haircuts/repo charges,

we estimate that spread product – particularly Europeansupra-national paper, RMBS and covered bonds – look

cheap to a bank liquidity portfolio. In terms of AUD

AAA-rated paper eligible for the CLF, an optimally

structured liquidity portfolio should overweight RMBS

and covered bonds3.

More generally, we remain cautious on spread products,

given the risk of further spread widening in peripheral

Europe. In the supra-national space, the market is very

clearly making a geographic distinction, as investors and

dealers reduce exposure to European names in favour of

Washington names (Exhibit 24). As long as European

issues remain largely unresolved, we expect thisdivergence to remain with non-European AAA-rated

assets well sought after (and note that European supra-

nationals – EIB and KfW – represent 35% of 2011 YTD

Kangaroo issuance so far this year).

Semi-government paper – we favour

owning NSWTC bonds against swap

and ACGBs

Given our views on swap spreads, we like owning cash

spread product relative to swap rather than government

bonds. When we look at AAA-rated semi-government

paper using an assessment of relative credit metrics, both

SAFA and NSWTC paper look cheap against swap

(Exhibit 25). However, given recent commentary from

the South Australian Premier, we see a significant risk

that the state of South Australia could loose its AAA

rating. As such, we prefer owning NSWTC paper against

swap.

In contrast, WATC and TCV paper looks rich/fair value

against swap. This relative value assessment holds across

both the 5Y and 10Y sectors, but we prefer to exploit this

relative value in the 5Y sector, given our view on the

swap spread curve (likely biased flatter) and like owning

3 For further detail on this issue, see S. Auld, The RBA's committed

liquidity facility and initial margin changes: A low fee means that European supras, covered bonds and RMBS look cheap, 16 November 2011.

the NSWTC April-15 against the TCV November-

164(Exhibit 26).

As we noted above, we remain cautious on spread

product given 1) the likelihood of an escalation in the

European crisis and 2) the prospect of disappointment on

State government budget outcomes (given a weaker domestic growth backdrop). But we note that the latter

point works in favour of NSWTC paper, given the strong

possibility of an announcement concerning the full

privatisation of New South Wales’ electricity assets. One

implication of any such policy announcement is that the

4 More detail on our relative value methodology can be found in S.

Auld, Strategy Update: Ranking the AAA-rated State Government Issuers – Where is the risk/reward? 4 October 2011.

Exhibit 25: 5-year semi-government spread to swap vs. relative creditranking – NSWTC looks cheap relative to TCV

Asset swap vs. average z-score away from AAA mean; bp

-8.0

0.0

8.0

16.0

24.0

-0.5 -0.4 -0.3 -0.2 -0.1 0.0 0.1 0.2 0.3 0.4

Average Z-score aw ay from AAA mean

WATC (Apr-15)

NSWTC (Apr-16)

SAFA (Apr-15)

TCV (Nov-16)

Exhibit 26: We like owning the NSWTC April-15 against the TCVNovember-16 given the cheapness of the asset swap boxSpread in asset swap measures between the NSWTC April 15 and TCV November 16;bp

-2.0

-1.0

0.01.0

2.0

3.0

4.0

5.0

6.0

7.0

8.0

May-11 Jun-11 Jul-11 Aug-11 Sep-11 Oct-11




Sally AuldAC

(61-2) 9220-7816


214

sale of electricity assets could mean significantly reduced

funding requirements for NSWTC; currently, around

30% of the NSWTC net borrowing requirement is usedto fund electricity assets. All other things equal, this

should aid out-performance of NSWTC bonds, especially

in the mid- to longer-dated maturities, and drive

flattening of the NSWTC bond curve5.

Australian inflation-linked bonds: the

outperformer in 2011

Australian inflation-linked bonds have performed very

well over the course of 2011, with real yields declining

considerably over the year (Exhibit 27). The rally has

been most pronounced at the front end of the curve,

helping the real yield curve to bull-steepen. This isconsistent with both downward revisions to the

Australian growth profile over the course of the year and

a global rally in inflation-linked bonds.

Interestingly, Australian 10Y real yields are close to their

all-time lows (since the data series began in the mid-

1980s) (Exhibit 28). Given that real yields are

theoretically meant to encapsulate expectations about

GDP growth over the long run, this performance is

perhaps a little counter-intuitive (especially given the

consensus expectation of strong, medium-term and long-

term growth prospects in Australia thanks to its close ties

with China).

We think one potential explanation for the collapse in

real yields is Australia’s disappointing productivity

performance. In the long run, productivity growth is an

important determinant of an economy’s GDP

performance. Exhibit 29 illustrates that labour

productivity growth and 10Y real yields are reasonably

well correlated. The implication of this is that ongoing

trend declines in productivity might imply a structurally

lower trading range for longer-dated real yields going

forward.

But one important point to remember is that despite the

impressive rally in Australian real yields this year,

spreads to offshore markets continue to remain wide

(Exhibit 30). As we have noted above, this real yield

differential continues to be a driver of offshore demand

for ACGB products in both the nominal and inflation-

linked space.

5 For more detail, see S. Auld, Strategy Update: NSW Electricity

Privatization - implications for the NSWTC bond curve, 2 November 2011.

Exhibit 27: Australian real yields have fallen significantly across thecurve over the course of 2011Real yields across Australian linkers; %

0.75

1.00

1.25

1.50

1.75

2.00

2.25

2.50

2.75

3.00

2015i 2020i 2025i 2030i

3-Jan-11

18-Nov-11

Exhibit 28: The recent rally has pushed Australian 10Y real yields tohistoric lowsMonthly data for 10Y real yields; %

1.5

2.5

3.5

4.5

5.5

6.5

J ul-86 Mar-91 N ov -95 J ul-00 Mar-05 N ov -09

Exhibit 29: The trend decline in productivity may be driving structurallylower real yields in Australiaoya % %-pts

-3.5

-2.5

-1.5

-0.5

0.5

1.5

2.5

3.5

4.5

5.5

6.5

Mar-00 Aug-02 J an-05 J un-07 N ov -09

-1.20

-0.80

-0.40

0.00

0.40

0.80

1.20Labour productivity, lhs

Annual change in 10y real y ield, rhs




Sally AuldAC

(61-2) 9220-7816


215

Indeed, the growth differential between Australia and the

US correlates well with the 10Y real yield spread; based

on J.P. Morgan growth forecasts for Australia and theUS, the 10Y real yield spread looks too wide at present

(Exhibit 31). Accordingly, we think the better AUS-US

spread contraction trade is to own the ACGB Aug-2020i

against the UST Jul-2021i.

Australian breakevens have not performed well over the

course of the year. This was largely contrary to

expectations, where the consensus anticipated rising

inflation in Australia. However, a softer growth outcome

in 2011 and a lower inflation forecast track from the

RBA has meant that market inflation expectations have

struggled to rally. This has been mirrored in J.P.

Morgan’s November 2011 Inflation Expectations Survey,where expectations of the chance of above-target

inflation in Australia have declined consistently over the

course of the year. The survey suggests that it was

changing perceptions of domestic investors (as opposed

to offshore investors) which, in turn, have driven the

move towards more benign inflation expectations.

J.P. Morgan economists now expect core inflation to

remain within the target band until the end of 2012.

Headline inflation should decline markedly in the near

term (as base effects associated with the impact of the

Queensland floods roll off), but will move higher as the

impact of the carbon tax works its way through theeconomy. J.P. Morgan forecasts for inflation are outlined

in Exhibit 32.

We think one of the better opportunities in the AUD

inflation-linked market is to be found in NSWTC

inflation-linked bonds. NSWTC has been the

predominant issuer in the semi-government inflation

market, largely due to its requirement to fund state-

owned utility assets. However, a recent discussion paper

suggests the strong probability that the NSW

Government will consider the full privatisation of its

electricity assets (generation and transmission and

distribution assets).

In our view, the main consequence should the

government decide to proceed with privatisation is that

there is likely to be far less supply in NSWTC inflation-

linked bonds as a large part of longer-dated inflation

linked issuance is used to fund electricity assets. And

given the recent under-performance of the 2035i

maturity, we think there is good value in owning

NSWTC 2035 linkers vs. bonds. For investors able to

Exhibit 30: Despite the rally in Australian real yields, 10Y cross-marketreal yield spreads remain wideMonthly data for 10Y real yield spreads to TIPS, UK linkers and Canadian Real bonds;

%-pts

-1.0

-0.5

0.0

0.5

1.0

1.5

2.0

2.5

3.0

Mar-00 Nov-01 Jul-03 Mar-05 Nov-06 Jul-08 Mar-10

Spread to US

Spread to UK

Spread to Canada

Exhibit 31: The 10Y real yield spread between Australia and the USlooks too wide, given their GDP differential10Y AUS-US real yield and real GDP growth spreads (J.P. Morgan forecast dashedline);%-pts %-pts

-1.0

-0.5

0.0

0.5

1.0

1.5

2.0

2.5

Mar-99 May -02 Jul-05 Sep-08 Nov -11

-2.0

-1.0

0.0

1.02.0

3.0

4.0

5.0

6.0

7.010y AUS-US real yield spread, lhs

Real GDP spread, rhs, pushed fwd 4q

Exhibit 32: We expect core inflation to remain within the target band in2012. Headline inflation should decline markedly in the near term onbase effects, but will move higher as the impact of the carbon taxworks its way through the economy

Australian headline and core inflation, actual and J.P. Morgan forecasts; oya %

0.0

1.0

2.0

3.0

4.0

5.0

6.0

03Q1 05Q3 08Q1 10Q3 13Q1

Headline

Core




Sally AuldAC

(61-2) 9220-7816


216

trade asset swaps, we think that the 2035i also looks

attractive on an asset-swap basis.

Australian bond supply update

The Australian Office of Financial Management

(AOFM) expects that gross Treasury bond issuance will

be around AUD51bn in the current 2011/12 financial

year. In net terms, supply will be around $37bn. So far

this year, the AOFM has issued $44.7bn of ACGBs, with

net issuance at $33.5bn due to the maturity of the Jun-11

benchmark. Four new benchmark bonds have been

introduced, including two new long-end bonds with

maturity dates of 21 April 2023 and 21 April 2027. The

latter bond extended the benchmark ACGB curve by a

further four years. Exhibit 33 shows the ACGB benchmark curve and benchmark issue size as of 31

October 2011.

There is currently $14.4bn of ACGB inflation-linked

bonds on issue across four benchmark lines, and we

expect a further $1.2bn to be issued by June 2012. We

expect gross (and net) issuance of $2.0bn in 2012. As of

October 2011, the AOFM had no plans to extend the

inflation-linked benchmark curve, but in the May budget,

it committed to keeping the size of the inflation-linked

market at 10% to 15% of the total ACGB market.

At present, the outlook for bond supply in 2012 isdifficult to ascertain. One reason is that there is some

uncertainty about whether or not the Federal Government

will be able to meet its commitment to returning the

budget to a small surplus in the 2012/13 fiscal year. J.P.

Morgan economists are doubtful, and do not envisage the

budget returning to surplus until 2014/15. Exhibit 34

highlights the difference between Federal Government

and J.P. Morgan forecasts for the budget in the coming

fiscal years. Over the forecast horizon, the cumulative

difference is $29bn. We should get greater clarity on the

Government’s intentions when it releases the next Mid-

Year Economic and Fiscal Update before the year ends.

Depending on which set of forecasts we use, the outlook

for ACGBs (nominal and inflation-linked) on issue

through to mid-2014 varies from $215bn to $250bn

(Exhibit 35). Note that this estimate does not include T-

notes (usually around $10bn-15bn on issue).

Either way, it is likely that Australia will require the debt

ceiling to be raised at some point in either 2012 or 2013.

Section 5 of the Commonwealth Inscribed Stock Act

(1911) states, “The total face value of stock and

Exhibit 33: The benchmark ACGB curve, maturity and issue sizeOutstanding amount by line in ACGBs; AUDbn

0.0

3.0

6.0

9.0

12.0

15.0

18.0

1 5 - A

p r - 1 2

1 5 - N

o v - 1 2

1 5 - M

a y

- 1 3

1 5 - D

e c - 1

3

1 5 - J

u n - 1 4

2 1 - O

c t - 1 4

1 5 - A

p r - 1

5

2 1 - O

c t - 1 5

1 5 - J

u n - 1

6

1 5 - F

e b - 1 7

2 1 - J

u l - 1 7

2 1 - J

a n - 1

8

1 5 - M

a r - 1

9

1 5 - A

p r - 2

0

1 5 - M

a y

- 2 1

1 5 - J

u l - 2 2

2 1 - A

p r - 2

3

2 1 - A

p r - 2 7

Exhibit 34: Federal and J.P. Morgan government budget forecasts – J.P.Morgan economists are less optimistic on the government’s ability toreturn to surplus in 2012/13Budget balance; AUDbn

-60

-50

-40

-30

-20

-10

0

10

2009-10 (a) 2010-11 (a) 2011-12 (f) 2012-13 (f) 2013-14 (f)

J.P. Morgan

Government

Exhibit 35: Supply of ACGBs depends heavily on budget forecasts for coming fiscal yearsOutstanding amount of ACGBs (nominal and inflation-linked), forecast under FederalGovernment and J.P. Morgan assumptions; AUDmn

0

65,000

130,000

195,000

260,000

Dec-70 Jun-78 Dec-85 Jun-93 Dec-00 Jun-08

Using Federal Governm ent budget forecasts

Using J.P. Morgan budget forecasts




Sally AuldAC

(61-2) 9220-7816


217

securities on issue under this Act and the Loans

Securities Act 1919 at any time must not exceed $250

billion.” As of 11 November 2011, total CommonwealthGovernment Securities on issue amounted to $215.1bn.

Looking to the end of the forecast horizon, there is still

some debate about the future size of the ACGB market.

This is because if the Federal Government realises its

forecasts for a return to surplus by 2012/13, then this

would imply a negative net bond supply at some point. In

the 2011/12 Budget, the Government noted the

recommendation from a panel of financial market

participants and regulators that the ACGB market be kept

at a constant size of nominal GDP (around 12% to 14%)

(Exhibit 36). If the Government decides to accept this

recommendation, then the AOFM will be required tomanage a portfolio in order to invest surplus cash.

Exhibit 36: There is a proposal to keep the size of the ACGB market asa constant proportion of nominal GDPSize of ACGB market as a proportion of nominal GDP; %

0.0

5.0

10.0

15.0

20.0

25.0

30.0

1983 1988 1993 1998 2003 2008 2013

Proposed target lev el:

12%-14% of GDP




Sally AuldAC

(61-2) 9220-7816


218

Trading themes

• Extend duration on dips− The lesson in 2011 for investors in AUD fixed income

was that events in Europe – rather than domesticfundamentals – had a major impact on government bond yields.

− Our European strategists remain bearish on Europe,and for this reason we are biased towards long-duration strategies in AUD fixed income in 2012.

− While this is a somewhat difficult call given thatyields are close to their lows for the year, we note thatdomestic investors are still running short of benchmark duration; US 10Y yields should declinefurther and strong demand from offshore investors

should continue to support the market.− Look to lighten duration in 2H12, given the risk that

markets price in a more optimistic global outlook.

• Curve

− The 3s/10s bond curve should steepen in 2012, givenour economists’ expectation of further rate cuts fromthe RBA. But cash bond steepeners have negativecarry and roll; instead, we prefer to express this viewvia forward starting swap steepeners.

− There is very attractive carry and roll in cash bondcurve flatteners. Accordingly, we think these are aneffective portfolio hedge for investors already running

long-duration portfolios as they provide protectionagainst the peripheral spread narrowing.

• Spread product

− We are biased towards wider swap spreads next year

and believe that paying 3Y swap spreads is a goodinsurance trade to hold into 2012. If risk aversiondominates markets, then we would expect the assetswap curve to flatten.

− In the cash spread product, we remain cautious.Domestic positioning still feels overweight spread product and as such, we think there will be better buying opportunities. We favour owning NSWTC paper to swap and AUD European supras against USDEuropean supras.

• Inflation

− We like owning the ACGB Aug-2020i against theUST Jul-2021i; this spread looks too wide given theexpected growth spread between Australia and the US.This trade should also perform given our view that thenominal AUS-US 10Y bond spread is too wide versusour medium term forecasts.

− Our preferred trade in the inflation space is to own the NSWTC Nov-2035i inflation-linked bond against theACGB Sep-2030i bond. Any decision to privatise NSW electricity assets will be extremely positive for longer-dated NSW inflation-linked paper.

• AUS-US 10Y spread

− Our medium-term model suggests the 10Y spreadshould be biased narrower from current levels.

Exhibit 37: AUD interest rate forecastsPercent

Current Dec-11 Mar-12 Jun-12 Sep-12 Dec-12

RBA cash rate 4.50 4.25 3.75 3.75 3.75 3.75

90-day rate 4.60 4.00 3.65 3.90 4.00 4.00

3Y yield 3.22 3.25 2.75 3.30 3.50 3.70

10Y yield 3.99 4.15 3.80 4.30 4.35 4.40

3s/10s curve (bp) 77.0 90.0 105.0 100.0 85.0 70.00

3Y swap yield 3.84 4.00 3.65 4.05 4.10 4.20

10Y swap yield 4.77 4.95 4.65 5.05 5.05 5.05

3Y swap spread 62.0 75.0 90.0 75.0 60.0 50.0

10Y swap spread 75.0 80.0 85.0 75.0 70.0 65.0

AUS-US 10Y spread (bp) 203.0 210.0 210.0 180.0 175.0 170.0




Sally AuldAC

(61-2) 9220-7816


219

New Zealand

• Despite the fact that both the NZD 2Y and 10Y

swap yields are close to all-time lows, we think

New Zealand yields will likely reach new lows

again in 2012. Accordingly, we prefer to extend

duration on any selloff

• This position should work well for 1H12, and we

would look to lighten duration into the second

half of the year due to the risk of a move towards

higher yields globally

• We think the NZD OIS curve looks cheap relative

to the AUD OIS curve, especially looking at Jun-

12 RBA and RBNZ OIS dates

• Generally, our view is that the 2s/10s swap curve

looks too flat, especially if the RBNZ does not

tighten policy next year. We like entering 2s/5s

swap curve steepeners, 3 months forward. An

alternative expression of this view is to pay the

belly of the 2s/5s/10s swap butterfly, 3 months

forward

• Given our view that 1) New Zealand swap yields

are likely to reach new lows, and 2) that 10Y asset

swap spreads already look too wide, we would

look to underweight the long end of the NZGB

curve against swap.

The economic recovery that wasn’t

New Zealand has faced a plethora of headwinds over the

course of 2011. Indeed, we estimate that the economy

grew by just 2.3% this year, a small improvement on the

1.7% growth recorded in 2010. Indeed, Exhibit 1

illustrates that in level terms, the New Zealand economy

is yet to return to its pre-crisis levels.

The New Zealand economy has been buffeted by

ongoing household sector deleveraging, an elevated NZD, weaker-than-expected trading partner growth

(especially Australia) and the impact of the Canterbury

earthquake (and the subsequent delayed recovery

efforts). Indeed, we expect household deleveraging to be

a permanent feature of the New Zealand economic

backdrop in the years ahead – the experience of the

Scandinavian countries in the early 1990s is quite

illustrative for New Zealand too (Exhibit 2).

Exhibit 1: The New Zealand economy is yet to return to its pre-crisispeak…Real GDP, Index is December 2007 = 100

96

97

98

99

100

Dec-07 Sep-08 Jun-09 Mar-10 Dec-10

Exhibit 2: …and if the Scandinavian experience is any indication,household deleveraging could take a long timeHousehold debt to GDP, %

47

55

63

71

79

87

95

103

-7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7

Years from start of recession

35

40

45

50

55

New Zealand, lhs

Scandinavian average, rhs

Exhibit 3: Low number of building consents suggest that post-earthquake-related activity is yet to have a material impact on theeconomyNumber of building consents issued

500

1000

1500

2000

2500

3000

3500

Jan-90 Oct-93 Jul-97 Apr-01 Jan-05 Oct-08

Seasonally Adjusted

Trend




Sally AuldAC

(61-2) 9220-7816


220

The one bright spot for the New Zealand economy in

2011 was the Rugby World Cup (RWC), although

sentiment data since the RWC suggests that any boostwas very short-lived.

Accordingly, we only expect a modest uptick in growth

in 2012, with calendar growth likely to be 2.8% during

the year. It has become evident that ongoing seismic

shocks are continuing to delay any rebuilding efforts in

the earthquake-affected regions and as such, the boost to

broader economic activity is likely to be more of a 2013

story. Exhibit 3 shows that so far, there has been no

notable uptick in building consents issued.

The RBNZ responded to the severe earthquake in

Canterbury by cutting the OCR by 50bp to 2.5%, back tothe lows seen during the financial crisis (Exhibit 4).

Initially, the RBNZ characterised this as an emergency

setting and was keen to withdraw the extra stimulus as

soon as possible. It has since become clear that this

setting of monetary policy looks increasingly appropriate

given the lack of momentum in the domestic economy

and mounting risks offshore (Exhibit 5). We see a risk

that RBNZ rhetoric will turn more dovish as we move

into 2012.

J.P. Morgan economists do not expect the RBNZ to

begin normalising policy until September 2012. Clearly,

risks are biased to unchanged policy for longer and possibly, a new low in the OCR. Indeed, the RBNZ’s

decision in early November to delay the introduction of

some macro-prudential regulatory changes (the core

funding ratio for New Zealand banks) because of

offshore instability may go some way to reflecting policy

makers’ current mindset.

Front-end pricing: A new low in the NZ

policy rate in 2012?

Front-end pricing in New Zealand has been very volatile

over the course of 2011, with the market pricing in

anywhere from 20bp of rate cuts over a 12-month

horizon to 110bp of rate hikes over the same period

(Exhibit 6). Currently, the market is expecting the

RBNZ to deliver 20bp of rate cuts by June 2012. In

contrast, the local market economists’ consensus strongly

favours the next move in rates as up in 2012.

While we believe that offshore (and domestic)

fundamentals would need to deteriorate markedly from

here before the RBNZ will countenance the need for a

lower policy rate, we are also cognizant that illiquid

Exhibit 4: The RBNZ returned the policy rate back to lows seen duringthe financial crisis …OCR; %

2.0

3.0

4.0

5.0

6.0

7.0

8.0

9.0

Mar-99 Dec-01 Sep-04 Jun-07 Mar-10

Exhibit 5: …but if global conditions deteriorate further, then the casefor further rate cuts will start to mount – there is a strong correlationbetween moves in the OCR and J.P. Morgan’s Global PMI%-pts Index

-6

-5

-4

-3

-2

-1

0

1

2

Sep-99 F eb-02 Jul-04 Dec -06 M ay -09 Oc t-11

32

37

42

47

52

57

62

Annual difference in

cash rate, lhs

Global PMI, pushed

fwd 6 months, rhs

Exhibit 6: Monetary policy expectations have been quite volatile in NewZealand this year Tightening* priced in over a 12-month period; bp

-40

-20

0

20

40

60

80

100

120


* Negative number implies easing




Sally AuldAC

(61-2) 9220-7816


221

global funding markets could also pressure the RBNZ to

ease policy again. Exhibit 7 illustrates that there is a

reasonable disconnect between the level of the OCR and bank lending rates. More expensive bank funding will

only pressure these spreads wider.

In a relative value sense, we think the front end of the NZ

OIS curve looks cheap against the AUD OIS curve. For

example, the June 2012 RBNZ meeting date OIS implies

20bp of rate cuts, while the June 2012 RBA meeting date

OIS implies almost 160bp of rate cuts. If the RBA has

cut 160bp by the middle of 2012, then the RBNZ will

have likely cut by more than 20bp. Similarly, if markets

normalise, then there is scope for a more rapid selloff in

the AUD front end than in the NZD front end.

The outlook for outright yields – new

lows in 2012

New Zealand yields have largely moved lower over the

course of the year, with a new historical yield low in the

2Y swap rate made this month. As Exhibit 8 illustrates,

price action has been consistent with the bullish trend

seen in 2Y rates over the past couple of years. This

theme has also been replicated in the longer end of the

swap curve, with the 10Y yield reaching a new historical

yield low in November.

The move lower in yields has been in sympathy with the broader macroeconomic environment in New Zealand

and offshore (disappointing growth outcomes), and a

growing expectation that central bank policy could turn

more dovish. In addition, we think the recent move lower

has been a function of positioning, with most dealers

looking for higher yields into the year-end and not

prepared for a rally.

In the short term, some of the domestic sentiment

indicators suggest the possibility of a bounce in the NZD

2Y swap yields (Exhibit 9). However, as we noted in our

outlook for Australian yields, domestic fundamentals

have not necessarily been a reliable guide to interest-rate

market movements in 2011. And given the risks to the

global backdrop, we would look to use any short-term

sell-off in New Zealand yields as an opportunity to

extend duration.

In the 10Y part of the curve, a similar theme has

prevailed. The NZD 10Y swap has made new, all-time

lows recently but still looks to be trading largely in line

with our estimate for fair value (Exhibit 10).

Exhibit 7: Lending rates have not fallen to the same extent as the OCRin New Zealand, meaning bank funding stress could also be a keydriver of RBNZ policy moves in 2012Various bank lending rates and OCR; %

2

3

4

5

6

7

8

9

10

11

Mar-99 Oct-01 May -04 Dec -06 Jul-09

OCR

Variable business lending rateFloating mortgage rate

Exhibit 8: NZD 2-year swap yields have made new historical lowsrecently, consistent with the bull trend in place since late 20092Y swap yield, %

2.0

3.0

4.0

5.0

6.0

7.0

8.0

9.0

Jan-07 May -08 Sep-09 Jan-11

Exhibit 9: Similar to Australia, domestic indicators suggest the risk of higher yields – but this can easily be swamped by offshoredevelopmentsIndex %-pts

-55.0

-45.0

-35.0

-25.0

-15.0

-5.0

5.0

15.025.0

35.0

45.0

Jun-04 Oct-06 Feb-09 Jun-11

-5

-4

-3

-2

-1

0

1

2

3Domestic economic activity expected, lhs,

pushed fwd. 1q Annual change in 2y swap, rhs




Sally AuldAC

(61-2) 9220-7816


222

Our view on duration in New Zealand next year looks for

lower yields, at least in 1H12. This is likely to be driven

by:

1) The prospect of lower UST yields in early 2012;

2) The strong likelihood of noticeably more dovish

commentary from the RBNZ;

3) The risk that the domestic data flow in New

Zealand starts to disappoint; and

4) Ongoing flight to quality into New Zealand

fixed income.

Curve spreads should be biased steeper in 2012

The NZD 2s/10s swap curve has traded a wide range

over the course of 2011, reaching a peak of 214bp in late

April and a trough of 127bp in late September. The curve

is currently towards the lower end of this range, at

around 152bp. Generally speaking, our view is that the

curve looks too flat at current levels, especially given the

risks surrounding the OCR in the coming year. Exhibit

11 suggests that the 2s/10s swap curve is already priced

for a modest tightening cycle in 2012.

We think the better curve trade in New Zealand is to position for steeper forward curves in 2012. At present,

we think the forward swap curves look too flat

(especially 2s/5s and 2s/10s) and like paying the 2s/5s

swap curve 3-months forward. Indeed, our models

suggest that the 2s/5s forward curve spread is pricing in a

cash rate close to 3.5% in just six months’ time, an

expectation well above our economists’ view and current

front-end pricing (Exhibit 12).

These trades are attractive because they roll positively up

to spot, especially in 2s/10s (the 1Y forward drop to spot

is around 12bp). Investors who prefer a more defensive

curve steepening position should look to short the front

end against a forward-curve steepener; either the front

end is too rich or the forward curve spreads are too flat.

The shape of the forward curves (in terms of what they

imply for the path of the policy rate) also suggests that

the 2s/5s/10s swap butterfly, 3-months forward, is

starting to look too low. We like paying the belly of the

fly on any move below flat (currently, the spread is

market at 1bp).

Exhibit 10: Fair value model for the NZD 10Y swap yield – according toour forecasts, new yield lows are likely in the new year 10Y swap yield; %

4.0

4.5

5.0

5.5

6.0

6.5

7.0

7.5

8.0

8.5

Mar-98 Oct-00 May-03 Dec-05 Jul-08 Feb-11

Actual

Model

Model is 10-year swap yield = 0.44 + 0.10*90-day rate + 0.65*US 10-year swap yield +0.46*moving average of nominal GDP growth; R-squared = 0.75, standard error is0.33ppts.

Exhibit 11: The 2s/10s curve looks flat, even if the RBNZ executedanother 50bp of rate hikes next year %-pts %

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

2.0

2.5

3.0

Mar-99 Jan-02 Nov -04 Sep-07 Jul-10

0.5

1.5

2.5

3.5

4.5

5.5

6.5

7.5

8.5

9.5

2s10s curve, lhs

NZ OCR, rhs, inverted

Exhibit 12: The cash rate implied by 2s/5s forward curve is too highrelative to our forecast or market pricing. Therefore given the negativecorrelation* between the 2s/5s curve and cash rate, we favour 2s/5sswap curve steepeners 3M forwardCash rate over various horizons; %

2.0

2.2

2.4

2.6

2.8

3.0

3.2

3.4

3.6

3.8

4.0

Current 3m 6m 1y 2y

Cash rate implied by 2s5s forward curv e

Cash rate, JPM forecast

Cash rate implied by market pricing

* Model used for the curve-implied cash rate: 2s5s swap curve = 1.11 - 0.29*real cash

rate; R-squared is 0.86, standard error is 0.22ppts.




Sally AuldAC

(61-2) 9220-7816


223

Swap spreads look close to their wides

in the long end of the curve

New Zealand swap spreads have been quite volatile over

the course of this year, with the 2021 bond trading at

almost 40bp over the swap earlier in the year (Exhibit

13). Since then, NZGBs have generally outperformed

swap, and swap spreads look to be pushing back to the

wides seen earlier in the year.

Generally speaking, NZGB swap spreads tend to exhibit

strong directionality with 2Y swap yields. Indeed, our

model suggests that movements in 2Y swap yields can

explain 86% of the variation in 10Y swap spreads. This

is because the swap market is generally seen as the more

liquid market and will usually outperform bonds in arally and underperform bonds in a selloff, as market

participants look to express duration views through

swaps rather than bonds. Our model of 10Y swap spreads

(Exhibit 14) suggests that swap spreads look wide given

the level of 2Y swap yields.

Given our view that 1) New Zealand swap yields are

likely to reach new lows, and 2) that 10Y swap spreads

already look too wide, we would look to underweight the

long end of the NZGB curve against swap on any out-

performance of NZGBs in Q1 next year. Given the

relative issue size ($6.6bn of Mar-19s and $10.4bn of

May-21s), we would prefer to express this view via the2019 maturity. Look to position for a narrower swap

spread on the May-19 bond on any move to the +30bp

level (currently 24bps).

New Zealand bond supply update

The New Zealand Debt Management Office (NZDMO)

announced that gross bond supply in the 2011/12 fiscal

year will be $13.5bn. In net terms, this amounts to a

supply of $5.9bn, given the redemption of the Nov-2011

bond. Pre-funding in the prior fiscal year (2010/11) has

meant that gross bond supply remains unchanged despite

significant costs associated with the post-earthquakerebuilding. So far this year, the NZDMO has issued

$6.45bn of benchmark bonds and is running ahead of

schedule (on a pro-rata basis) for the fiscal year.

The NZDMO lengthened the maturity of the NZGB

curve this year with the issuance of a 5.5% April 2023

maturity. The NZGB benchmark curve is illustrated in

Exhibit 15; in addition, benchmark line sizes for the

April 2013 and April 2015 bonds were increased to a

maximum of $12.0bn this year. In New Zealand’s

Exhibit 13: NGZBs have generally outperformed swaps since April2011, when they traded 40bp over, and look to be pushing back to their widesSwap spread; bp

-50

-40

-30

-20

-10

0

10

20

30

40

50


NZGB Apr-13

NZGB May-21

Exhibit 14: 10Y swap spread model – spreads look too wide given thelevel of yields10Y swap spread actual and model*; bp

-60

-40

-20

0

20

40

60

80

100

120

140

160

Oct-03 Jul-05 Apr-07 Jan-09 Oct-10

Actual

Model

+/- 2 s.d.

* Model 10Y swap spread = -87.55 + 23.90*2Y swap yield; R-squared = 0.86, standarderror = 12.95bp

Exhibit 15: NZGB benchmark curveOutstanding by line; NZDbn

0.0

2.0

4.0

6.0

8.0

10.0

12.0

1 5 - A p r - 1 3

1 5 - A p r - 1 5

1 5 - D e c - 1 7

1 5 - M a r - 1 9

1 5 - M a y - 2 1

1 5 - A p r - 2 3

.




Sally AuldAC

(61-2) 9220-7816


224

inflation-linked bond space, the NZDMO has postponed

plans for an inflation-linked issue (2023 maturity) until

1H12. There has been no new issuance into the soleinflation-linked February 2016 maturity so far this year.

Total bonds on issue (nominals) in New Zealand now

total $47.4bn (we exclude ‘non-market’ bonds in our

calculations – those held by the RBNZ and EQC). We

expect that gross (and net) issuance in 2012 to likely total

$12.3bn, taking bonds on issue close to $60bn by the end

of 2012 (Exhibit 16).

Exhibit 16: Issuance of NGZBs in 2012 will take outstanding amount toclose to $60bnTotal outstanding of NGZBs; NZDbn

0

20

40

60

80

Jan-00 Apr-03 Jul-06 Oct-09 Jan-13




Sally AuldAC

(61-2) 9220-7816


225

Trading themes

• Extend duration on dips− Despite the fact that both the NZD 2Y and 10Y swap

yields are close to all-time lows, we think NewZealand yields will likely reach new lows again nextyear. Accordingly, we prefer to extend duration on anyselloff.

− This position should work well for 1H12, and wewould look to lighten duration into the second half of the year due to the risk of a move towards higher yields globally.

− We think the NZD OIS curve looks cheap relative tothe AUD OIS curve, especially looking at the June2012 RBA and RBNZ OIS dates.

• Curve

− Generally, our view is that the 2s/10s swap curvelooks too flat, especially if the RBNZ does not tighten policy in 2012. Carry + roll on 2s10s steepeners islargely neutral.

− We think the forward 2s/5s swap curves look too flatrelative to policy expectations. Implement 2s/5ssteepeners, 3-months forward (3-month carry + roll of 2.5bps). A more defensive expression of this tradewould be to hold steepeners against paid 3-month OIS positions.

− We like paying the belly of the 2s/5s/10s swap curve,

3-months forward. This trade has positive carry + rollof 3.0bp over a 3-month period.

•

Spreads− Given our view that 1) New Zealand swap yields are

likely to reach new lows, and 2) that 10-year assetswap spreads already look too wide, we would look tounderweight the long end of the NZGB curve againstswap on any out-performance of NZGBs in Q1 nextyear.

− Given the relative issue size ($6.6bn of Mar-19s and$10.4bn of May-21s), we would prefer to express thisview via the 2019 maturity. Look to position for anarrower asset swap spread on the May-19 bond onany move to the +35bp level.

Exhibit 17: NZD interest rate forecasts%

Current Dec-11 Mar-12 Jun-12 Sep-12 Dec-12

RBNZ OCR 2.50 2.50 2.50 2.50 2.75 3.00

90-day rate 2.66 2.65 2.65 2.75 3.00 3.25

3-year yield 2.82 2.85 2.50 3.25 3.40 3.50

10-year yield 3.91 4.05 3.80 4.40 4.40 4.40

10-year swap spread (bp) 27.5 20.0 25.0 15.0 10.0 10.0

2-year swap yield 2.71 2.65 2.30 2.95 3.00 3.10

10-year swap yield 4.23 4.25 4.05 4.55 4.50 4.50

2s10s swap curve (bp) 152.0 160.0 175.0 160.0 150.0 140.0

NZ-US 10-year spread (bp) 190.0 200.00 210.0 190.0 180.0 170.0




Aditya ChordiaAC

(44-20) 7777-9841


226

Interest rate forecasts

Euro area 18-Nov-11 Mar-12 Jun-12 Sep-12 Dec-12 United States 18-Nov-11 Mar-12 Jun-12 Sep-12 Dec-12

Refi rate 1.25 0.75 0.50 0.50 0.50 Fed funds 0 - 0.25 0 - 0.25 0 - 0.25 0 - 0.25 0 - 0.25

Euribor 3M 1.47 1.40 1.45 1.20 1.00 Libor 3M 0.49 0.60 0.50 0.40 0.35

2Y 0.46 0.35 0.30 0.40 0.50 2Y 0.28 0.17 0.30 0.30 0.30

5Y 1.10 0.75 0.70 0.85 1.05 5Y 0.92 0.75 1.25 1.25 1.25

10Y 1.97 1.55 1.25 1.50 1.75 10Y 2.01 1.70 2.50 2.50 2.50

30Y 2.61 2.15 1.95 2.25 2.50 30Y 3.00 2.70 3.60 3.60 3.60

2s/10s 151 120 95 110 125 2s/10s 173 153 220 220 220

10s/30s 64 60 70 75 75 10s/30s 99 100 110 110 110

2s/30s 215 180 165 185 200 2s/30s 271 253 330 330 330

2Y 110 130 145 135 130 2Y 48 53 38 38 38

5Y 96 115 130 120 115 5Y 40 45 34 34 34

10Y 67 80 90 85 80 10Y 16 27 18 18 18

30Y 19 25 30 25 25 30Y -30 -23 -23 -23 -23

United Kingdom 18-Nov-11 Mar-12 Jun-12 Sep-12 Dec-12 Australia

Base rate 0.50 0.50 0.50 0.50 0.50 Cash rate 4.50 3.75 3.75 3.75 3.75

Libor 3M 1.02 1.05 1.10 1.05 1.00 Govt curve 10Y 3.99 3.80 4.30 4.35 4.40

2Y 0.48 0.50 0.50 0.55 0.65

5Y 1.13 1.00 1.00 1.05 1.15 New Zealand

10Y 2.26 1.80 1.50 1.75 1.95 Cash rate 2.50 2.50 2.50 2.75 3.00

30Y 3.19 2.70 2.25 2.40 2.60 Govt curve 10Y 3.91 3.80 4.40 4.40 4.40

2s/10s 178 130 100 120 130

10s/30s 93 90 75 65 65

2s/30s 271 220 175 185 195

2Y 96 110 120 110 100

5Y 71 85 95 85 65

10Y 31 45 55 45 25

30Y -9 0 5 0 -10

Japan

O/N call rate 0.05 0.05 0.05 0.05 0.05

2Y 0.12 0.15 0.15 0.15 0.15

5Y 0.31 0.30 0.35 0.40 0.45

10Y 0.94 0.90 0.95 1.10 1.15

20Y 1.70 1.75 1.80 1.90 2.00

30Y 1.93 1.95 2.00 2.10 2.15

2s/10s 82 75 80 95 100

10s/30s 99 105 105 100 100

2s/30s 181 180 185 195 200

Govt curve

Swap spreads

Govt curve

Swap spreads

Govt curve

Govt curve

Swap spreads




Aditya ChordiaAC

(44-20) 7777-9841


227

Recent curve movements

Govt. curve 18-Nov-11 3Q end 2Q end 1Q end YTD min YTD max 18-Nov-11 3Q end 2Q end 1Q end YTD min YTD max

2Y 0.46 0.56 1.61 1.78 0.31 1.92 0.48 0.56 0.83 1.34 0.47 1.55

5Y 1.10 1.17 2.28 2.66 0.86 2.82 1.13 1.07 1.84 2.25 0.95 2.55

10Y 1.97 1.89 3.03 3.35 1.69 3.50 2.26 2.26 3.38 3.68 2.10 3.88

30Y 2.61 2.64 3.76 3.79 2.46 3.97 3.19 3.55 4.28 4.35 3.12 4.58

2s/5s 63 61 67 88 44 109 64 50 101 91 41 135

2s/10s 151 133 142 157 128 207 178 169 255 234 152 276

10s/30s 64 75 73 44 39 99 93 129 90 67 62 144

Swap curve 18-Nov-11 3Q end 2Q end 1Q end YTD min YTD max 18-Nov-11 3Q end 2Q end 1Q end YTD min YTD max

2Y 1.60 1.50 2.18 2.35 1.37 2.47 1.54 1.28 1.45 1.86 1.08 5.05

5Y 2.09 1.98 2.82 3.08 1.84 3.23 1.86 1.78 2.51 2.97 1.64 3.22

10Y 2.68 2.55 3.44 3.63 2.40 3.79 2.60 2.60 3.55 3.80 2.48 4.05

30Y 2.82 2.74 3.85 3.86 2.60 4.02 3.11 3.34 4.08 4.13 3.04 4.292s/5s 49 48 64 73 41 88 32 50 106 111 -257 125

2s/10s 108 105 126 128 97 165 105 132 211 194 -162 214

10s/30s 14 19 42 23 14 42 51 74 53 32 20 85

Swap spreads 18-Nov-11 3Q end 2Q end 1Q end YTD min YTD max 18-Nov-11 3Q end 2Q end 1Q end YTD min YTD max

2Y 110 92 55 53 47 112 96 67 48 48 42 96

5Y 96 81 49 39 37 100 71 50 40 36 29 71

10Y 67 65 41 26 21 78 31 21 6 6 2 37

30Y 19 9 9 4 0 22 -9 -21 -20 -22 -30 -6

2s/5s -14 -11 -6 -14 -20 2 -25 -17 -8 -12 -26 -7

2s/10s -43 -27 -14 -27 -44 -3 -65 -46 -43 -41 -65 -27

10s/30s -48 -56 -32 -21 -67 -18 -40 -42 -26 -29 -46 -22

EUR GBP

Govt. curve 18-Nov-11 3Q end 2Q end 1Q end YTD min YTD max 18-Nov-11 3Q end 2Q end 1Q end YTD min YTD max2Y 0.28 0.25 0.46 0.79 0.15 0.85 0.12 0.14 0.17 0.21 0.12 0.25

5Y 0.92 0.97 1.75 2.22 0.77 2.40 0.31 0.37 0.43 0.49 0.29 0.62

10Y 2.01 1.93 3.16 3.45 1.71 3.72 0.94 1.03 1.13 1.25 0.94 1.35

30Y 3.00 2.92 4.38 4.51 2.76 4.76 1.93 1.91 2.01 2.18 1.85 2.26

2s/5s 64 71 130 144 57 155 19 23 26 28 17 38

2s/10s 173 167 270 267 151 289 82 89 96 104 82 114

10s/30s 99 99 123 106 95 148 99 89 88 93 86 101

Swap curve 18-Nov-11 3Q end 2Q end 1Q end YTD min YTD max 18-Nov-11 3Q end 2Q end 1Q end YTD min YTD max

2Y 0.78 0.58 0.69 0.96 0.42 1.05 0.38 0.35 0.38 0.39 0.31 0.46

5Y 1.34 1.26 2.02 2.42 1.07 2.62 0.46 0.49 0.56 0.63 0.42 0.78

10Y 2.18 2.11 3.26 3.56 1.90 3.83 0.95 1.03 1.16 1.31 0.95 1.40

30Y 2.70 2.70 4.07 4.29 2.56 4.54 1.72 1.80 2.01 2.17 1.72 2.25

2s/5s 56 69 133 147 53 158 7 14 17 24 7 322s/10s 140 154 257 260 135 279 57 68 78 91 57 98

10s/30s 52 58 81 73 51 93 77 77 85 86 75 88

Swap spreads 18-Nov-11 3Q end 2Q end 1Q end YTD min YTD max 18-Nov-11 3Q end 2Q end 1Q end YTD min YTD max

2Y 48 32 22 16 13 52 26 21 22 19 15 26

5Y 40 29 25 19 16 44 14 11 13 13 6 17

10Y 16 16 8 8 2 22 -1 0 2 5 -3 9

30Y -30 -27 -31 -23 -38 -20 -22 -13 -4 -1 -22 2

2s/5s -9 -3 3 3 -7 9 -12 -10 -9 -5 -13 -5

2s/10s -32 -15 -14 -8 5 33 -27 -21 -19 -13 -27 -12

10s/30s -46 -43 -39 - 31 27 54 -21 -13 -6 -6 -21 -4

USD JPY




Aditya ChordiaAC

(44-20) 7777-9841


228

Recent sovereign cash spread movements

Spread to Germany; bp

18-Nov-11 1W ago 2W ago 1M ago 3M ago 1Y min 1Y max 1Y avg 1Y SD 1Y z-score

Austria 111 94 45 47 30 6 130 27 19 4.4

Belgium 344 273 219 212 132 32 357 104 62 3.9

Finland 39 23 13 17 12 -4 51 10 7 4.0

France 119 105 63 69 33 4 140 28 24 3.7

Netherlands 29 19 15 14 11 1 48 9 6 3.3

Greece 9131 8414 8379 6499 3321 1092 9131 3209 2330 2.5

Ireland 773 765 801 732 774 361 2164 808 314 -0.1

Italy 572 543 522 408 283 83 684 225 136 2.6

Portugal 1507 1556 1900 1603 1099 299 1901 962 500 1.1

Spain 484 421 370 322 264 102 492 225 83 3.1

Wtd. peri. spread* 1418 1321 1315 1042 632 229 1418 574 336 2.5


Austria 177 157 110 100 54 30 202 55 30 4.1

Belgium 330 294 283 281 196 78 351 158 63 2.7

Finland 82 63 52 52 33 6 91 26 19 3.0

France 151 135 106 104 47 21 187 49 30 3.4

Netherlands 66 50 41 51 35 5 83 25 17 2.5

Greece 4003 3771 3573 2939 1734 937 4059 1714 763 3.0

Ireland 706 707 704 614 780 495 1509 759 211 -0.3

Italy 533 548 526 435 304 102 675 250 132 2.1

Portugal 1322 1351 1459 1375 1085 292 1527 905 392 1.1

Spain 453 409 388 344 302 155 474 266 72 2.6

Wtd. peri. spread* 892 871 840 709 503 241 951 451 179 2.5

18-Nov-11 1W ago 2W ago 1M ago 3M ago 1Y min 1Y max 1Y avg 1Y SD 1Y z-score Austria 147 149 113 105 60 31 189 58 28 3.1

Belgium 282 256 253 244 176 72 309 141 56 2.5

Finland 67 55 51 51 43 24 75 37 11 2.8

France 147 148 120 110 64 30 188 58 30 3.0

Netherlands 56 47 43 45 38 19 65 33 10 2.3

Greece 2425 2357 2407 2266 1352 747 2471 1342 491 2.2

Ireland 625 611 631 637 761 501 1142 679 128 -0.4

Italy 494 492 471 399 285 125 575 244 113 2.2

Portugal 910 907 986 961 930 306 1161 722 269 0.7

Spain 469 395 374 338 291 170 486 267 65 3.1

Wtd. peri. spread* 699 673 666 603 444 229 739 400 135 2.2


Austria 121 124 95 82 41 28 158 45 22 3.5Belgium 238 238 219 210 147 75 280 129 43 2.5

Finland - - - - - - - - - -

France 152 158 123 109 57 32 199 58 30 3.2

Netherlands 11 6 7 9 11 6 15 10 2 0.6

Greece 1523 1487 1448 1303 761 458 1544 781 305 2.4

Ireland - - - - - - - - - -

Italy 450 444 426 392 287 153 515 254 92 2.1

Portugal 648 652 639 628 544 254 682 443 146 1.4

Spain 431 389 365 343 270 185 448 267 53 3.1

Wtd. peri. spread* 563 545 525 484 345 205 595 320 104 2.3

10Y

30Y

2Y

5Y

*Weighted peripheral spread computed against Germany for Greece, Ireland, Portugal, Italy and Spain (weighted by the size of their outstanding bond market). 30Y doesnot contain Ireland.




Aditya ChordiaAC

(44-20) 7777-9841


229

Recent sovereign CDS spread movements

CDS spread*; bp


US 29 29 29 50 50 16 50 33 13 -0.3

UK 51 50 49 52 36 21 58 34 9 1.8

Germany 46 45 43 48 30 10 61 24 12 1.8

Austria 149 127 109 112 48 20 159 54 33 2.9

Belgium 306 279 260 275 202 62 320 156 68 2.2

Finland 34 32 30 38 30 12 45 22 9 1.3

France 158 138 121 131 98 29 167 68 38 2.4

Netherlands 55 49 48 41 26 10 60 26 12 2.5

Greece*** 55 52 54 57 37 15 59 31 14 1.7

Ireland 911 930 917 968 1051 531 1606 858 223 0.2Italy 573 551 463 428 336 71 627 229 150 2.3

Portugal 1507 1500 1471 1527 1196 392 1757 934 440 1.3

Spain 451 405 360 362 351 142 464 274 82 2.2

Wtd. peri. spread** 616 591 521 506 429 148 657 316 145 2.1


US 40 40 40 65 65 37 65 52 9 -1.3

UK 90 89 87 92 79 47 104 70 13 1.6

Germany 95 92 87 95 82 36 120 61 21 1.7

Austria 207 185 159 163 120 50 222 98 39 2.8

Belgium 332 305 286 301 235 115 346 201 58 2.3

Finland 68 63 60 76 60 24 90 45 18 1.3

France 221 202 178 191 148 63 234 116 45 2.3Netherlands 113 100 98 98 67 28 122 60 25 2.2

Greece*** 60 58 59 61 45 25 63 41 12 1.7

Ireland 715 730 720 760 785 510 1199 694 121 0.2

Italy 533 530 490 453 360 125 583 269 130 2.0

Portugal 1060 1054 1034 1119 869 390 1235 750 258 1.2

Spain 465 423 383 385 370 197 479 310 68 2.3

Wtd. peri. spread** 557 545 507 492 415 189 595 331 115 2.0

2Y

5Y

* 25bp running coupon used for Finland, France, Germany, Netherlands and US. 100bp running coupon used for UK, Austria, Belgium, Greece, Ireland, Italy, Portugal andSpain. Spreads for all the countries except US are in $ CDS and for US it is in € CDS.** Weighted peripheral spread computed as CDS spread of Ireland, Portugal, Italy and Spain, weighted by the size of their outstanding bond market.*** For 2Y and 5Y Greece CDS, we show upfront premium instead of flat CDS spread for 100bp running coupon.




Aditya ChordiaAC

(44-20) 7777-9841


230

Euro area marketable bond* and bank debt** redemptions

€bn

Sov. Banks Sov. Banks Sov. Banks Sov. Banks Sov. Banks Sov. Banks Sov. Banks

Dec-11 0 2 0 0 0 0 0 10 18 3 0 1 18 16

Jan-12 2 3 0 5 0 0 15 5 25 28 14 8 56 50

Feb-12 0 2 0 2 0 0 0 7 0 8 0 12 1 31

Mar-12 1 2 4 1 0 0 0 7 19 13 0 3 24 26

Apr-12 0 2 0 0 0 1 18 6 16 3 0 10 34 22

May-12 0 2 0 2 0 0 0 2 0 2 0 4 0 12

Jun-12 0 1 0 1 0 1 0 5 19 9 0 2 19 20

Jul-12 10 1 0 0 0 0 28 13 27 10 15 2 81 25

Aug-12 0 0 1 1 1 1 0 0 0 5 0 3 2 10

Sep-12 0 2 12 0 6 0 12 5 21 11 0 2 51 21

Oct-12 0 1 0 3 0 0 19 9 16 9 0 1 35 23

Nov-12 1 1 0 1 0 0 0 1 0 5 0 1 1 9

Dec-12 0 2 8 0 0 0 5 3 17 3 0 1 30 9

Total 15 18 25 17 7 4 98 74 178 108 30 52 353 272

Sov. Banks Sov. Banks Sov. Banks Sov. Banks Sov. Banks Sov. Banks

Dec-11 3 0 0 0 0 2 0 1 0 6 3 10

Jan-12 0 2 0 1 0 4 0 4 0 6 0 18

Feb-12 0 1 0 2 36 5 0 1 1 21 38 29

Mar-12 14 2 6 2 27 6 0 1 1 14 48 25

Apr-12 0 0 0 3 28 2 0 0 12 15 40 21

May-12 9 0 0 0 1 6 0 1 0 10 9 17

Jun-12 0 1 0 2 2 3 10 2 0 25 12 33

Jul-12 0 1 0 0 17 2 1 2 13 8 31 12

Aug-12 8 1 0 0 12 2 0 0 0 3 20 5

Sep-12 0 2 0 0 11 4 0 0 2 7 13 12

Oct-12 0 1 0 1 20 6 0 0 20 5 40 12

Nov-12 0 1 0 3 13 4 0 0 0 10 13 18

Dec-12 2 0 0 0 30 10 1 0 0 5 34 15

Total 36 11 6 14 198 55 13 14 49 133 301 227

Austria Belgium

Greece Ireland

Netherlands Core Euro-area** *

Peri. Euro-area****

Finland

Portugal Spain

France Germany

Italy

* Marketable bonds include: conventionals, linkers, floaters zero coupons and international bonds.** Maturities in all currencies and jurisdictions and include secured, unsecured and securitised issuance, including MTNs but excluding short-term (maturity of less than oneyear) and self-funded deals (deals where there is only one bookrunner and it is also the issuer). The data also include any government guaranteed issuance by the banksbut no direct issuance by government or government sponsored institutions.*** Austria, Belgium, Finland, France, Germany and Netherlands**** Greece, Ireland, Italy, Portugal, SpainSource: J. P. Morgan, Dealogic




Aditya ChordiaAC

(44-20) 7777-9841


231

Euro area sovereign ratings / SMP purchases /

Election calendar

Sovereign ratings

Austria AAA Aaa AAA

Belgium AA+ NEG Aa1 NEG* AA+ NEG

Cy prus BBB NEG* Baa3 NEG* BBB NEG

Finland AAA Aaa AAA

France AAA Aaa AAA

Germany AAA Aaa AAA

Greece CC NEG Ca DEV CCC

Ireland BBB+ Ba1 NEG BBB+ NEG

Italy A NEG A2 NEG A+ NEG

Netherlands AAA Aaa AAA

Portugal BBB- NEG Ba2 NEG BB+ NEG

Slov akia A+ POS A1 A+

Slov enia AA- Aa3 NEG* AA- NEG

Spain AA- NEG A1 NEG AA- NEG

S&P Moody's Fitch

* represents under watch; grey highlight: below IG;Outlook:NEG - negative outlook, POS – positive outlook, DEV - developing outlookand blank represents stable outlookNotes:1 – For a country’s exclusion, Barclays Capital requires 2 out of 3 credit ratings (S&P,Moody’s and Fitch) to be below IG. Citigroup requires both S&P and Moody’s ratingbelow IG. J.P. Morgan EMU Investment Grade index requires Moody’s, S&P, andFitch each to rate a country above IG to maintain inclusion.

2 – Markit iBoxx uses an average rating methodology (S&P, Moody’s and Fitch) for acountry's exclusion.Source: Bloomberg

SMP purchases

First Last First LastWeekly

(€bn)

Amount

matured

(€bn)

Amount

offered for

sterilization

(€bn)

09-Nov 15-Nov 14-Nov 18-Nov 8.0 0.1 194.5

02-Nov 08-Nov 07-Nov 11-Nov 4.5 0.6 187.0

26-Oct 01-Nov 31-Oct 04-Nov 9.5 0.0 183.0

19-Oct 25-Oct 24-Oct 28-Oct 4.0 0.0 173.5

12-Oct 18-Oct 17-Oct 21-Oct 4.5 0.2 169.5

05-Oct 11-Oct 10-Oct 14-Oct 2.2 0.0 165.0

28-Sep 04-Oct 03-Oct 07-Oct 2.3 0.0 163.0

21-Sep 27-Sep 26-Sep 30-Sep 3.8 0.0 160.5

14-Sep 20-Sep 19-Sep 23-Sep 4.0 0.1 156.5

07-Sep 13-Sep 12-Sep 16-Sep 9.8 0.0 152.5

31-Aug 06-Sep 05-Sep 09-Sep 14.0 0.0 143.0

24-Aug 30-Aug 29-Aug 02-Sep 13.3 0.0 129.0

17-Aug 23-Aug 22-Aug 26-Aug 6.7 1.3 115.5

10-Aug 16-Aug 15-Aug 19-Aug 14.3 0.0 110.5

03-Aug 09-Aug 08-Aug 12-Aug 22.0 0.0 96.0

78.3 4.3 74.0

Trade date Settlement date

Cumulativ e amount till 2 Aug 2011

Note: Every Tuesday ECB sterilizes SMP purchases during the period between the

Wednesday two weeks back (first trade date) and the Tuesday of the previousweek (last trade date). This is equivalent to SMP purchases which settle onMonday (first settlement date) to Friday of the previous week (last settlement date).The ECB started buying Italian and Spanish bonds on Monday, 8 August 2011.Source: ECB

Election calendar Period Date Month Election in

2012 19 February Greece

22 April France

2013 - - Austria

- - Germany

- - Italy

2014 - - Belgium

- - Netherlands

2015 - - Finland

Portugal

- - Spain

2016 - - Ireland




Aditya ChordiaAC

(44-20) 7777-9841


232

Euro area fact sheet

Avg debt maturity Non-dom Bank assets % of bank assets in govt % loans/ % of bank assets

Bonds Tbi lls % Tbi lls (years, JPM) Gross (€bn, JPM) Net (€bn, JPM) investors /GDP ratio securi ties (incl . non-dom.) deposi ts rat io funded at the CB

Oct-11 Oct-11 Oct-11 Oct-11 Tot. 2012 Tot. 2012 (%) Sep-11 Sep-11 Sep-11 Latest

Austria 167 1 1 7.6 20 10 79 3.3 2.9 109 0.7

Belgium 264 32 11 6.7 40 16 59 3.2 8.8 59 1.6

Finland 62 7 10 5.2 13 7 84 3.0 1.7 141 0.0

France 1098 163 13 7.3 180 95 66 4.2 3.9 117 1.0

Germany 1021 63 6 6.2 175 18 81 3.2 3.7 84 0.4

Greece 247 12 5 6.7 0 -30 67 2.3 9.4 134 15.9

Ireland 85 0 0 6.3 0 -6 83 4.1 3.3 156 21.8

Italy 1356 130 9 6.9 145 35 45 2.5 6.5 126 2.8

Netherlands 261 40 13 6.2 48 18 68 4.0 4.0 125 0.7

Portugal 104 10 9 6.1 0 -10 80 3.4 4.8 119 8.0

Spain 469 81 15 6.7 90 49 38 3.3 5.0 107 2.4

GDP wtd. avg. 9 6.7 65 3.3 4.7 108 2.1

US* 8246 1480 15 5.3 2150 936 46 - 3.7 81 -

UK* 944 57 6 14.5 186 142 30 5.4 1.8 91 2.0

Marketable debt outstanding (€bn) Conv. bond issuance

GDP GDP growth Inflation** Budget balance^ Prim. Balance Gross debt Curr. acc. bal. GDPpc Unempl.

(€bn) (oya, %) (oya, %) (% of GDP) (% of GDP) (% of GDP) (% of GDP) (EU=100) rate (%)

2012 2012 2012 2012 2012 2012 2012 2012 Latest

Austria 310 0.9 2.2 -3.1 -0.3 73 2.8 124 3.9

Belgium 382 0.9 2.0 -4.6 -1.3 99 2.1 116 6.7

Finland 198 1.4 2.6 -0.7 0.5 52 0.0 123 7.8

France 2,028 0.6 1.5 -5.3 -2.5 89 -3.3 104 9.9

Germany 2,623 0.8 2.0 -1.0 1.3 81 4.4 108 5.8

Greece 212 -2.8 0.8 -7.0 1.0 198 -7.9 63 17.6

Ireland 159 1.1 0.7 -8.6 -4.3 118 1.5 119 14.2

Italy 1,617 0.1 2.0 -2.3 3.1 121 -3.0 90 8.3

Netherlands 623 0.5 1.9 -3.1 -1.2 65 7.0 126 4.8

Portugal 169 -3.0 3.0 -4.5 0.8 111 -5.0 54 12.5

Spain 1,094 0.7 1.1 -5.9 -3.5 74 -3.0 80 22.6

GDP wtd. avg. 9,415 0.5 1.8 -3.4 -0.2 91 0.0 101 10

US 15,495* 1.5 1.9 -8.5 -5.4 105*** -3.1 - 9.0

UK 1,566* 0.6 2.9 -7.8 -4.6 89 -0.9 - 8.3

* Local currency** HICP; National index if not available*** IMF Estimate^ Net lending (+) or net borrowing (-)Source: EC European Economic Forecast, Autumn 2011, IMF, Eurostat and ILO




Pavan WadhwaAC

(44-20) [email protected]. Morgan Securities Ltd

233

Analyst Certification:

The research analyst(s) denoted by an “AC” on the cover of this report certifies (or, where multiple research analysts are primarilyresponsible for this report, the research analyst denoted by an “AC” on the cover or within the document individually certifies, with

respect to each security or issuer that the research analyst covers in this research) that: (1) all of the views expressed in this reportaccurately reflect his or her personal views about any and all of the subject securities or issuers; and (2) no part of any of the researchanalyst's compensation was, is, or will be directly or indirectly related to the specific recommendations or views expressed by the researchanalyst(s) in this report.

Other Disclosures

J.P. Morgan ("JPM") is the global brand name for J.P. Morgan Securities LLC ("JPMS") and its affiliates worldwide. J.P. Morgan Cazenove is amarketing name for the U.K. investment banking businesses and EMEA cash equities and equity research businesses of JPMorgan Chase & Co.and its subsidiaries.

Options related research: If the information contained herein regards options related research, such information is available only to persons whohave received the proper option risk disclosure documents. For a copy of the Option Clearing Corporation's Characteristics and Risks of Standardized Options, please contact your J.P. Morgan Representative or visit the OCC's website athttp://www.optionsclearing.com/publications/risks/riskstoc.pdf

Legal Entities DisclosuresU.S.: JPMS is a member of NYSE, FINRA, SIPC and the NFA. JPMorgan Chase Bank, N.A. is a member of FDIC and is authorized andregulated in the UK by the Financial Services Authority. U.K.: J.P. Morgan Securities Ltd. (JPMSL) is a member of the London Stock Exchangeand is authorized and regulated by the Financial Services Authority. Registered in England & Wales No. 2711006. Registered Office 125 LondonWall, London EC2Y 5AJ. South Africa: J.P. Morgan Equities Limited is a member of the Johannesburg Securities Exchange and is regulated bythe FSB. Hong Kong: J.P. Morgan Securities (Asia Pacific) Limited (CE number AAJ321) is regulated by the Hong Kong Monetary Authorityand the Securities and Futures Commission in Hong Kong. Korea: J.P. Morgan Securities (Far East) Ltd, Seoul Branch, is regulated by the KoreaFinancial Supervisory Service. Australia: J.P. Morgan Australia Limited (ABN 52 002 888 011/AFS Licence No: 238188) is regulated by ASICand J.P. Morgan Securities Australia Limited (ABN 61 003 245 234/AFS Licence No: 238066) is a Market Participant with the ASX andregulated by ASIC. Taiwan: J.P.Morgan Securities (Taiwan) Limited is a participant of the Taiwan Stock Exchange (company-type) andregulated by the Taiwan Securities and Futures Bureau. India: J.P. Morgan India Private Limited, having its registered office at J.P. MorganTower, Off. C.S.T. Road, Kalina, Santacruz East, Mumbai - 400098, is a member of the National Stock Exchange of India Limited (SEBIRegistration Number - INB 230675231/INF 230675231/INE 230675231) and Bombay Stock Exchange Limited (SEBI Registration Number -INB 010675237/INF 010675237) and is regulated by Securities and Exchange Board of India. Thailand: JPMorgan Securities (Thailand)Limited is a member of the Stock Exchange of Thailand and is regulated by the Ministry of Finance and the Securities and Exchange

Commission. Indonesia: PT J.P. Morgan Securities Indonesia is a member of the Indonesia Stock Exchange and is regulated by the BAPEPAMLK. Philippines: J.P. Morgan Securities Philippines Inc. is a member of the Philippine Stock Exchange and is regulated by the Securities andExchange Commission. Brazil: Banco J.P. Morgan S.A. is regulated by the Comissao de Valores Mobiliarios (CVM) and by the Central Bank of Brazil. Mexico: J.P. Morgan Casa de Bolsa, S.A. de C.V., J.P. Morgan Grupo Financiero is a member of the Mexican Stock Exchange andauthorized to act as a broker dealer by the National Banking and Securities Exchange Commission. Singapore: This material is issued anddistributed in Singapore by J.P. Morgan Securities Singapore Private Limited (JPMSS) [MICA (P) 025/01/2011 and Co. Reg. No.: 199405335R]which is a member of the Singapore Exchange Securities Trading Limited and is regulated by the Monetary Authority of Singapore (MAS) and/or JPMorgan Chase Bank, N.A., Singapore branch (JPMCB Singapore) which is regulated by the MAS. Malaysia: This material is issued anddistributed in Malaysia by JPMorgan Securities (Malaysia) Sdn Bhd (18146-X) which is a Participating Organization of Bursa Malaysia Berhadand a holder of Capital Markets Services License issued by the Securities Commission in Malaysia. Pakistan: J. P. Morgan Pakistan Broking(Pvt.) Ltd is a member of the Karachi Stock Exchange and regulated by the Securities and Exchange Commission of Pakistan. Saudi Arabia: J.P.Morgan Saudi Arabia Ltd. is authorized by the Capital Market Authority of the Kingdom of Saudi Arabia (CMA) to carry out dealing as an agent,arranging, advising and custody, with respect to securities business under licence number 35-07079 and its registered address is at 8th Floor, Al-Faisaliyah Tower, King Fahad Road, P.O. Box 51907, Riyadh 11553, Kingdom of Saudi Arabia. Dubai: JPMorgan Chase Bank, N.A., DubaiBranch is regulated by the Dubai Financial Services Authority (DFSA) and its registered address is Dubai International Financial Centre -Building 3, Level 7, PO Box 506551, Dubai, UAE.

Country and Region Specific DisclosuresU.K. and European Economic Area (EEA): Unless specified to the contrary, issued and approved for distribution in the U.K. and the EEA byJPMSL. Investment research issued by JPMSL has been prepared in accordance with JPMSL's policies for managing conflicts of interest arisingas a result of publication and distribution of investment research. Many European regulators require a firm to establish, implement and maintainsuch a policy. This report has been issued in the U.K. only to persons of a kind described in Article 19 (5), 38, 47 and 49 of the Financial Servicesand Markets Act 2000 (Financial Promotion) Order 2005 (all such persons being referred to as "relevant persons"). This document must not beacted on or relied on by persons who are not relevant persons. Any investment or investment activity to which this document relates is onlyavailable to relevant persons and will be engaged in only with relevant persons. In other EEA countries, the report has been issued to personsregarded as professional investors (or equivalent) in their home jurisdiction. Australia: This material is issued and distributed by JPMSAL inAustralia to "wholesale clients" only. JPMSAL does not issue or distribute this material to "retail clients". The recipient of this material must notdistribute it to any third party or outside Australia without the prior written consent of JPMSAL. For the purposes of this paragraph the terms"wholesale client" and "retail client" have the meanings given to them in section 761G of the Corporations Act 2001. Germany: This material is




Pavan WadhwaAC


235




Pavan WadhwaAC


jpm fixed in

Documents