jp morgan - managing fx hedge ratios - a framework for strategic and tactical decisions

5/14/2018 JP Morgan - Managing FX Hedge Ratios - A Framework for Strategic and Tactical De...

http://slidepdf.com/reader/full/jp-morgan-managing-fx-hedge-ratios-a-framework-for-strategic-and-ta

Managing FX hedge ratiosA framework for strategic and tactical decisions

John NormandAC

(44-20) 7325-5222

[email protected]

Gabriel de Kock

(1-212) 834 4254

[email protected]

Matthew Franklin-Lyons

(1-212) 834 4565

[email protected]

Arindam Sandilya

(1-212) 834 2304

[email protected]

www.morganmarkets.com/GlobalFXStrategy J.P. Morgan Securities

The certifying analyst is indicated by an AC. See page 38 for analyst certification and important legal and regulatory disclosures.

• Unprecedented volatility over the past two years has heightened investor and corporate attention towards managing FX hedge ratios.

• For investors, four issues predominate: (1) how to determine the long-term, optimal hedge ratio; (2) how to time entry into a hedging program; (3) how to deviate from the strategic hedge ratio to generate profits or manage cash flows; and (4) how to choose between forwards

and options in implementing a hedging program.

• For corporates, the most frequent concerns are variants of the first,second and fourth points: how to set a baseline range for hedge ratios;which currencies are most likely to post large moves higher or lower;

and which instruments are best for executing a hedging program.

• Previous J.P. Morgan research has discussed optimal hedge ratios and proposed currency models based on long-term valuation and short-termmomentum. This paper updates those studies, adapts the models for dynamic hedging over various horizons, and extends the original analysisto four base currencies (USD, EUR, GBP, AUD).

• The conventional wisdom that hedging currency risk minimizes portfoliovolatility was proven false at some points during the credit crisis, whenhedged portfolios in some currencies became more volatile thanunhedged ones. A more dynamic strategy is therefore required to avoidonerous cash flow requirements.

• Dynamic hedging should begin with signals from long-term fair value

models, to indicate which currencies merit the most attention over a 6 to18-month horizon. Investors can use these signals to time entry intohedging programs, while corporates can use them to focus on currenciesmost vulnerable to large moves in either direction.

• Short-term trading models commonly used by currency overlaymanagers and global macro funds can be adapted to drive deviationsaround the benchmark hedge ratio over a one to three-month horizon. Aprice momentum model which adjusts dynamically around a 50/50hedge ratio outperforms the benchmark by about 100bp annually,depending on the base currency. Information ratios on the strategy range

from 0.2 to 0.5. A rate momentum (forward carry) model generatescomparable outperformance but with more consistency across sample periods.

• Data limitations constrain long-term analysis of the efficiency of forwards versus options in a hedging program. But since 2003, options

have outperformed forwards for AUD, EUR and GBP-based investorswith USD exposure. A signaling model that switches between forwardsand options is inconclusive.

Contents

I. Overview: Four common questions on h

policy

II. The merits of long-term FX exposure

III. Using fair value models to focus strate

hedge ratios

IV. Using alpha models to adjust hedge ra

tactically

V. Hedging with forwards versus options

Global FX StrategyMay 26, 2010



Global FX StrategyManaging FX hedge ratiosMay 26, 2010

John Normand (44-20) 7325-5222 [email protected]. Morgan Securities Ltd.

2

I. Overview: Four common

questions on hedging policy

Unprecedented volatility over the past two years hasheightened investor and corporate attention towardsmanaging FX hedge ratios. For investors, four issues

predominate:

• how to determine the long-term, optimal hedge ratio for

global stock and bond portfolios;

• how to time entry into a hedging program, to focus on

the most expensive currencies and avoid covering

exposure in undervalued ones;

• how to deviate from the strategic hedge ratio to

generate profits or manage cash flows;

• how to choose between forwards and options in

implementing a hedging program.

For corporates, the most frequent concerns are variants of the first, second and fourth points: how to set a baselinerange for hedge ratios; which currencies are most likely to

post large moves (higher where they have expenses, lower where they have earnings); and which instruments are bestfor executing a hedging program.

Academic studies and previous J.P. Morgan research over the past two decades have discussed optimal hedge ratios indetail.1 More recent J.P. Morgan studies have proposedlong-term valuation and short-term momentum modelswhich can be adapted to answer the second, third and fourthquestions (see blue box on next page). This paper updatesthose studies, adapts the models for dynamic hedging over various horizons, and extends the original analysis to four

base currencies (USD, EUR, GBP, AUD).

Section II reviews and critiques the conventional wisdomon optimal hedge ratios for those unfamiliar with theframework. Section III applies J.P. Morgan’s long-term fair

value model to answer the investor question of how to timeentry into a hedging program, and the corporate question of which currencies merit hedging focus over a 6 to 18-monthhorizon. Section IV modifies short-term FX trading models(alpha models) based on price momentum and interest rate

1 Early J.P. Morgan publications include Currency-hedged

international fixed income investment , Peter Rappoport (1990); Managing currency risk in global portfolios, Jan Loeys (1999);and Introduction to Portfolio Management , John Normand (2002and subsequent revisions).

momentum to drive tactical deviations around a 50/50hedged benchmark over one to three-month horizons.

Section V concludes with a discussion of how these signalscan be used to inform the choice of forwards versus optionsin a hedging program. We find that:

• The conventional wisdom that hedging currency risk

minimises portfolio volatility was proven false at some

points during the credit crisis, when hedged portfolios

in some currencies became more volatile than unhedged

ones. The cash flow implications for some funds was

disastrous, hence the necessity to consider more dynamic

frameworks for managing hedge ratios over time and

across currencies.

• Long-term fair value models can be the starting pointfor dynamic hedges, since they can indicate the most

vulnerable currencies over the medium term. J.P.

Morgan’s original fair value model (2008) has provided

reliable signals for identifying and trading misaligned

currencies versus the dollar over 6 to 18-month horizons.

That analysis is extended to base currencies such as

EUR, GBP and AUD. Investors can use the model’s

signals – reported quarterly in flagship publication World

Financial Markets and FX Markets Weekly – to time the

entry into currency overlay programs by raising hedge

ratios on the most overvalued currencies, lowering them

on the cheapest currencies and remaining at benchmark

on others. For corporates, the signals highlight currencieson which hedge ratios should be raised, because the

company has expenses in undervalued currencies or

earnings in overvalued ones.

• Over shorter horizons (one to three month

rebalancing), tactical trading models commonly used

by overlay managers and global macro funds can

inform deviations from the benchmark hedge ratio. A

price momentum model which adjusts dynamically

around a 50/50 hedge ratio outperforms the benchmark

by about 100bp annually, depending on the base

currency. Information ratios on the strategy range from

0.2 to 0.5. A rate momentum (forward carry) model

generates comparable outperformance but with more

consistency across sample periods.

• Data limitations constrain long-term backtesting to

compare the efficiency of forwards versus options. Still,

since 2003 hedging with options has outperformed

forwards for AUD, EUR and GBP based investors

with USD exposure. A signalling model which switches

between forward and options is inconclusive and will be

addressed in more detail in forthcoming research.





3

Previous J.P. Morgan publications on currency overlay, indices and models available on

www.morganmarkets.com/GlobalFXStrategy, or by clicking the hyperlinks below.

Introduction to Portfolio Management , John Normand, 2002 and subsequent revisions.

J.P. Morgan effective exchange rates: revised and modernized , Derek Hargreaves and Carl Strong, May 30, 2003.

JPMorgan’s FX Barometer , Normand, Mustafa Caglayan, Dan Ko, Nikolaos Panigirtzoglou and Lei Shen, September 22, 2004.

Introducing the JPMorgan VXY & EM-VXY , Normand and Arindam Sandilya, December 11, 2006.

JPMorgan Tradeable Currency Indices, Normand, July 2, 2007.

Rotating Between G-10 and Emerging Markets Carry , Normand, Jul 9, 2007.

A new fair-value model for G10 currencies , Gabriel de Kock, September 6, 2008.

Trading and Hedging Long-Term FX Fundamentals with J.P. Morgan’s Fair-Value Model , de Kock, April 24, 2009.

Alternatives to standard carry and momentum in FX , Normand and Ghia, August 8, 2008.

The month-end effect in FX: small but predictable, Normand, October 23, 2009.

https://mm.jpmorgan.com/servlet/PubServlet?skey=TU1SQy0xNDU4MDctMQA=&Name=MMRC-145807-1.pdf


https://mm.jpmorgan.com/servlet/PubServlet?skey=TU1SQy0xMzEzOTUtMQA=&Name=MMRC-131395-1.pdf


https://mm.jpmorgan.com/servlet/PubServlet?skey=TU1SQy00Njk5ODMtMQA=&Name=MMRC-469983-1.pdf


https://mm.jpmorgan.com/servlet/PubServlet?skey=TU1SQy0zNTUxNzEtMQA=&Name=MMRC-355171-1.pdf


https://mm.jpmorgan.com/servlet/PubServlet?skey=TU1SQy0zODk5ODQtMQA=&Name=MMRC-389984-1.pdf


https://mm.jpmorgan.com/servlet/PubServlet?skey=TU1SQy00Njk5ODItMQA=&Name=MMRC-469982-1.pdf


https://mm.jpmorgan.com/servlet/PubServlet?skey=TU1SQy00NzAwMDUtMQA=&Name=MMRC-470005-1.pdf


https://mm.jpmorgan.com/servlet/PubServlet?skey=R1BTLTI4NTQ1My0w&Name=GPS-285453-0.pdf


https://mm.jpmorgan.com/servlet/PubServlet?skey=TU1SQy00NTkxMzQtMQA=&Name=MMRC-459134-1.pdf


https://mm.jpmorgan.com/servlet/PubServlet?skey=R1BTLTMzNjQyMS0w&Name=GPS-336421-0.pdf
















4

II. The merits of long-term FX

exposure2

• The conventional wisdom is that unhedged FX

exposure raises volatility more than returns over the

long run. This conclusion tends to hold regardless of

base currency.

• There are three exceptions to this conclusion:

emerging markets exposure for long-term FX

appreciation; G-10 forex exposure as catastrophe

insurance; and G-10 forex exposure to diversify

portfolio risk.

• The more-risk-than-return view usually motivates

high hedge ratios on fixed income investments. FX

hedge ratios on equities are more variable, however.

Some investors will hedge to minimise volatility.

Others will not due to FX’s small marginal

contribution to an asset class which is already quite

volatile. Cash flow implications can also deter

hedging.

• The credit crisis has forced a rethink of the

conventional wisdom and standard practices. USD-

based investors suffered massive losses on unhedged

portfolios as the dollar strengthened/foreign

currencies weakened. Non-US investors who hedgedUSD exposure incurred significant cash flow

obligations from being short the dollar as it

appreciated. These developments now motivate

many to consider a more dynamic approach to FX

risk management.

Strategic FX exposure: high risk and lowreturn, with three exceptions

The conventional wisdom on FX makes two claims: thatFX exposure delivers more risk than return over the long-run; and that FX offers abundant short-term profit

opportunities due to structural inefficiencies. If these pointsare correct, then investors should fully hedge FX exposure

but run active overlay programs to capture short-term profitopportunities. Likewise, corporates should hedge as amatter of policy but alter target ratios over shorter horizonswhen they hold strong directional views.

2 This section updates and extends analysis on currencymanagement originally published in Introduction to Portfolio

Management , John Normand (2002 and subsequent revisions).

The notion that passive long-only currency exposure offersno long-term return stems from two conditions in

international finance known as covered and uncovered interest parity. Covered interest parity (CIP) states thatthe interest rate differential between two countries shouldequal the forward premium (discount), expressedalgebraically as

rforeign - rdomestic = (FXforward - FXspot)/ FXspot [Equation 1]

This relationship is a no-arbitrage condition rather than atheory. Consider the following example. Assume 12-moJapanese rates are 1%, US rates are 4%, and USD/JPY spotis 110. The no-arbitrage condition in equation 1 implies thatthe 12-mo forward rate must equal 106.70. If it were 110,an investor could lock in a guaranteed profit by borrowingJPY at 1%, buying a USD deposit at 4% and selling USDforward at 110 for no-risk returns.

Uncovered interest rate parity (UIP) is an extension of CIP. If markets are efficient, the forward rate will be anunbiased estimator of the future spot rate.

rforeign - rdomestic = (FXexpected - FXspot)/ FXspot [Equation 2]

If this condition holds, then there should be no return totaking currency risk, since high (low) interest ratecurrencies should depreciate (appreciate) to the level of theforward rate.

CIP tends to hold in practice, as free capital mobility amongthe major markets enforces the no-arbitrage condition. The

evidence on UIP is mixed, however: the passive return fromG-10 currency exposure tends to be negligible over thelong-term, but sometimes significant over the short term.

One way to illustrate this no-return concept is to consider asset market returns measured from three perspectives: localcurrency, unhedged into a given base currency and hedgedinto a base currency. The difference between local currencyand unhedged returns is the foreign currency’s contributionto total returns. A zero difference would support the viewthat currency exposure does not augment returns, so perhapsshould be hedged if volatility differentials are high.

USD-based investors

For USD-based investors, the return differential betweenlocal currency and unhedged stocks and bonds is modestover the long-term (1988 - 2009).3 Currency moves haveaugmented asset returns by roughly 1% for investments inthe Euro area, Canada, Australia and Japan over the pasttwenty years, due to those currencies’ appreciation versusthe dollar (charts 1 and 4). Currency has subtracted roughly

3 Sample period chosen by the availability of hedged and

unhedged return indices across G-10 markets.





5

Stock and bond returns and volatility with USD as base currency

Chart 1: Equity returns hedged vs unhedged into USD, 1988-2009

-3%

0%

3%

6%

9%

12%

USD JPY EUR GBP AUD CAD MSCI

ex-US

local ccy unhedged hedged

Source: J.P. Morgan

Chart 2: Equity volatility hedged vs unhedged into USD, 1988-2009

0%

10%

20%

30%

USD JPY EUR GBP AUD CAD MSCI ex-

US


Source: J.P. Morgan

Chart 3: Equities for USD-based investors: return vs risk, 1988 - 2009

-

5

10

15

5 10 15 20 25

Risk, %

Return, %

US equities

World ex US,

hedged

World ex US,

unhedged

Source: J.P. Morgan

Chart 4: Bond returns hedged vs unhedged into USD, 1988-2009

0%

3%

6%

9%

12%

USD JPY EUR GBP AUD CAD GBI ex

US


Source: J.P. Morgan

Chart 5: Bond volatility hedged vs unhedged into USD, 1988-2009

0%

3%

6%

9%

12%

15%

USD JPY EUR GBP AUD CAD GBI ex

US


Source: J.P. Morgan

Chart 6: Bonds for USD-based investors: return vs risk, 1988–2009

-

2

4

6

8

- 5 10 15Risk, %

Return, %

US bonds

World ex US, unhedged

World ex US, hedged

Source: J.P. Morgan





6

1% from UK investments due to sterling’s depreciationsince the late 1980s.

Though FX’s return impact may be modest, its volatilityimpact can be significant. For stocks, the difference inannualised volatility between local currency returns andunhedged returns can be massive: 7% - 8% for Australianand Canadian exposure, 3% for UK exposure and 1% for Japanese exposure. Only Euro area stocks exhibitcomparable return volatility, whether measured in localcurrency or unhedged terms (chart 2).

Volatility differentials on bonds are also substantial.

Unhedged JGB returns in USD terms are 7 percentage points more volatile than local currency returns, and Euroarea bonds are 6 points more volatile. The volatility of unhedged returns in UK, Canadian and Australian bondmarkets are roughly 4 points higher than local currencyreturns (chart 5). This volatility differential is much moremeaningful in bonds than in stocks, since government

bonds are already a relatively low-volatility asset class.

EUR, GBP and AUD-based investors

Similar conclusions hold for investors with other basecurrencies. For Euro area investors, unhedged overseasexposure has reduced returns by roughly 2% per annum for stocks in the UK, 1.4% for US and less than 1% for Australian and Canadian stocks. Currency moves have hadalmost no long-term impact on Japanese stocks holdingsgiven that EUR/JPY was close to unchanged between 1989

and 2009 (appendix 1, chart 1). Volatility differences aresubstantial, however. Unhedged foreign stocks are 3 to 10

percentage points more volatile (appendix 1, chart 2), whileforeign bonds are about two times more volatile (appendix1, charts 2 and 4).

For UK investors, unhedged foreign currency exposure hasadded 1% - 1.9% per annum to stock and bond returns over the past twenty years due to sterling’s trend decline(appendix 2, charts 1 and 4). In many cases, however,higher return volatility offsets this return advantage.Unhedged Canadian and Australian stock returns are 5%more volatile, and those in Japan 1.7% more. US stock volatility is only 0.8% higher, but Euro area equity exposureis 0.8% less volatile (appendix 2, chart 2). As a group,unhedged non-UK equities (MSCI ex-UK) add 1.1% per annum to returns but are 2.4% more volatile. For most bondmarkets, unhedged exposure is two or three times morevolatile than local currency returns (appendix 2, chart 5).

For Australian investors, the Australian dollar’s trendappreciation over the past two decades has resulted in acurrency loss from unhedged foreign exposure ranging from-0.3% per annum on Euro area investments to -1.7% per annum on UK investments (appendix 3 charts 1 and 4).Exposure to Canadian stocks and bonds has resulted in

almost no currency loss given the steadiness of theAUD/CAD cross, and Japanese exposure has delivered

small currency gains (roughly 0.2%). Volatility differenceshave been significant, however. Unhedged equity exposurein the US, UK and Canada are roughly 2.5% more volatilethan local currency returns, but Euro area and Japaneseexposure is 2%-3% less volatile (appendix 3, chart 2). As inother countries, unhedged bond market returns are two tothree times more volatile than local currency returns(appendix 3, chart 5).

Three exceptions: emerging markets,catastrophe insurance and portfoliodiversifier

Within the G-10, it is generally accepted that FX

exposure brings uncompensated volatility over the longrun. Since unhedged bond market returns are two to threetimes more volatile than local currency returns, a strategic – or benchmark – policy of fully hedging currency risk issensible. In equities, the hedging decision is less clear-cut.Some investors primarily concerned with risk minimizationwould hedge. Others would consider a hedging programcumbersome and expensive since FX’s marginalcontribution to equity volatility over the long run is small;the underlying asset class is already highly volatile. But

before delving further into equity investors’ hedgingdecision, consider three exceptions to the conventionalwisdom that passive FX exposure represents

uncompensated risks. These exceptions concern emergingmarkets exposure, FX as catastrophe insurance and FX as a portfolio diversifier.

Exception 1: emerging markets as asource of long-term returns

Unlike G-10 currencies which tend to mean-revert,

emerging market currencies can offer trend positive

returns. This excess return stems from two sources — real

appreciation and carry. Emerging markets currencies oftenexperience long-run real appreciation due to a faster rate of

productivity growth (the convergence process). At the sametime, interest rates also tend to be above those in G-10

markets, reflecting a higher marginal productivity of capital,a risk premium for convertibility and a policy tool to

promote disinflation. This combination of real exchangerate appreciation and interest rate differentials can persistfor years (chart 7), and thus generates a meaningful returndifferential over time to justify their higher risk.

We track these excess returns through JPMorgan’sEmerging Local Markets Index (ELMI), which measuresthe return in dollars of cash instruments (Libor, T-bills, FXforwards) in 24 emerging markets. Chart 8 shows theSharpe ratio on the index and its regional sub-components





7

since 1994. EM FX has outperformed 6-mo dollar cash by4% annually since the early 1990s. Outperformance has

been much higher for convergence countries: CentralEuropean currencies have outperformed USD libor by over 7.5% annually, while Latin currencies have outperformed

by 1.5%. Risk-adjusted returns on the index are decent(0.42), which is comparable to that on US equities and high-grade credit.

Chart 7: ELMI+ carry and spot returns, 1994-2009Based on returns from J.P. Morgan ELMI+ index

13% 13%10%

17%15%

4% 6%

14%

7% 6% 4% 2% 1% 2% 3%

-10% -9%

-24%

-5% -7%-10% -11%

-6%

7% 8%

-3%

6% 9%

-12%

5%

-30%

-25%

-20%

-15%

-10%

-5%

0%

5%

10%

15%

20%

94 96 98 00 02 04 06 08

return from FX appreciation/depreciation

return from carry

Source: J.P. Morgan

Chart 8: Emerging markets FX Sharpe ratios, 1994-2009Based on returns from J.P. Morgan ELMI+ index

0.42

0.61

0.120.14

-0.02

-0.10

0.10

0.30

0.50

0.70

E L M I +

A s

i a

E u r o p e

L a

t a m

M E / A f r i c a

Source: J.P. Morgan

Exception 2: catastrophe insurance andasset/liability matching require unhedgedFX exposure

Strategic FX exposure is also useful as a hedge against localevent risks, such as political uncertainty and naturaldisasters. For example, many insurance companies oftenhave significant, unhedged USD holdings. In the event of anatural disaster, the hit to their balance sheets would be

partially offset by gains on the local currency value of foreign holdings, assuming the dollar appreciated inresponse to those developments. Similarly, many investorsin emerging markets hold their foreign exposure (in the US,Europe or Japan) unhedged. In the event of political

uncertainty or financial crises, local currency would likelyweaken versus the G-3, providing some offset to the losses

on domestic assets. Finally, investors who have significantforeign currency liabilities — such as global pension funds,multinational corporations or central banks — likewisematch that exposure with unhedged foreign currency assets.

Exception 3: FX exposure may reduceportfolio risk

The third exception which justifies holding unhedged

benchmark exposure is for risk diversification: currencyand asset risk may be sufficiently negatively correlated soas to lower overall portfolio volatility. Since the variance of a portfolio is the sum of individual variances plus a measureof their comovement, the variance of a portfolio combining

a domestic and foreign asset will be a function of the (1)domestic asset’s variance; (2) foreign asset’s variance; (3)currency’s variance; (4) covariance between the domesticand foreign assets; and (5) covariance between the assetsand the currency.

This relationship is expressed algebraically as

σportfolio = w2domestic σ

2domestic + w2

foreignσ2

foreign + 2 wdomestic

wforeign σdomestic, foreign + H2 σ

2fx + 2H( wdomestic σdomestic, fx +

wforeign σforeign, fx) [Equation 3]

where H is the proportion of the portfolio with foreigncurrency exposure.

The last two terms capture the impact of currency exposureon total portfolio risk. For the fully hedged portfolio, thevariance of the currency exposure will be zero (since H =0). The last two terms of Equation 3 drop out and theexpression reduces to

σ fully hedged = w2domestic σ

2domestic + w2

foreign σ2foreign +

2 wdomestic wforeign σdomestic, foreign [Equation 4]

Hedging out the currency risk leaves the variance of the portfolio as a function of the individual asset variances andthe covariances between the asset returns, the same as if thiswere a domestic two-asset portfolio.

Under certain conditions, unhedged exposure can reduceoverall portfolio volatility. Equation 3 represents thevariance of the unhedged (or partially hedged) portfolio,and equation 4 that of the fully hedged portfolio.

Rearranging terms (subtracting equation 4 from equation 3)illustrates that currency exposure will reduce total portfoliorisk if the following condition holds:

w2

foreign σ2

fx + 2 wforeign (wdomesticσdomestic, fx +

wforeign σforeign, fx ) < 0 [Equation 5]


http://slidepdf.com/reader/full/jp-morgan-managing-fx-hedge-ratios-a-framework-for-strategic-and-t



8

In general, the covariance of currency returns with asset

returns must be sufficiently negative to overcome the

volatility of the currency itself. More specifically,• A positive correlation between the assets and the

currency increases total portfolio risk;

• If sufficiently large, a negative correlation between assets

and the currency can offset currency volatility;

• If there is zero correlation between the currency and the

foreign asset, the currency’s correlation with the

domestic asset must be more negative in order to reduce

total portfolio volatility;

• If currency returns are less volatile than asset returns, the

negative correlation can be somewhat closer to zero and

still reduce portfolio volatility.

How negative must the correlation be for currency exposure

to lower portfolio variance? Consider an example in which

the volatility of the domestic and foreign assets are equal,

and the domestic and foreign asset correlations with the

currency are equal. In this case, Equation 3 reduces to

ρ foreign, fx < -wforeign σfx / 2σforeign [Equation 6]

Plugging in various volatilities for the assets and currencyand different allocations to foreign assets generates

breakeven correlations which reduce total portfolio risk.

For example, if the currency and asset returns are equallyvolatile (vol ratio = 1, as in chart 9) and 10% of the

portfolio is allocated to foreign assets, then the correlation between currency and asset returns would need to be morenegative than -0.05 in order to lower overall risk. If 50% of exposure were foreign, the correlation would need to bemore negative than -0.25. If the currency were more volatilethan the asset market (vol ratio = 1.2), then the breakevencorrelations would need to be even more negative for agiven foreign allocation in order to reduce portfoliovolatility.

In practice asset/FX correlations are neither sufficientlynor consistently negative to reduce portfolio volatility by

leaving currency exposure unhedged. For USD-based

investors over the past two decades, only European stockshave correlated negatively with their currency performanceversus the dollar, implying that their currencies depreciateversus the dollar when equities rise (chart 10 and table 1).

Chart 9: Threshold correlation for portfol io risk reduction

-0.70

-0.60

-0.50

-0.40

-0.30

-0.20

-0.10

0.00

0 %

1 0 %

2 0 %

3 0 %

4 0 %

5 0 %

6 0 %

7 0 %

8 0 %

9 0 %

1 0 0 %

percent allocation to foreign equities

b r e a k e v e n F X / a s s e t c o r r e l a t i o n vol ratio = 0.5

vol ratio = 1.2

vol ratio = 1.0

Source: J.P. Morgan

Chart 10: Correlation of foreign stocks and FX, 1990-2009 for USD-based investorBased on monthly returns over 3 year rolling window

-0.8

-0.6

-0.4

-0.2

0.0

0.2

0.4

90 93 96 99 02 05 08

EUR

GBP

Source: J.P. Morgan

Chart 11: Correlation of foreign stocks and FX, 1990-2009 for USD-based investorBased on monthly returns over 3 year rolling window

-0.6

-0.4

-0.2

0.0

0.2

0.4

0.6

0.8

90 93 96 99 02 05 08

CAD AUD JPY

Source: J.P. Morgan





9

But that apparent diversification benefit disappears over aten-year sample period (correlations close to zero) and over

a five-year sample (the long-term correlation has flipped to positive). Correlations between Australian/Canadianequities and their currencies versus the dollar have always

been positive (chart 11 and table 1), which thus makesunhedged equities more volatile for US investors. Thecorrelation between the Nikkei and the yen’s performanceversus the dollar has been negative, but the strength of thisrelationship varies over time.

For euro and sterling-based investors, the dollar’s performance versus the euro has correlated negatively withUS stocks over the past five years (higher stocks, lower dollar), but not over the past twenty years. Commoditycurrency moves versus the euro and sterling correlate

positively with their equities, which makes unhedge

exposure riskier. Yen correlations are negative butsometimes not strong enough to diversify risk. The atypical

market is Australia. The combination of yield andcommodity exposure leads AUD to rise (and foreigncurrencies to fall) when equities rise. Viewed in isolationthis feature lends diversification benefit to unhedged foreignexposure, but given the AUD’s long-term appreciationtrend, unhedged exposure also subjects local investors to atranslation loss.

For bonds, many currencies correlate negatively with fixedincome performance, which suggests diversification

benefits from leaving exposure unhedged. Correlation isonly half the picture, however. Since currency volatility isso much higher than bond market volatility, even high andconsistently-negative correlations are insufficient to reduce

portfolio volatility. This point is illustrated by the appendixcharts at the end of Section II.

Table 1. Correlations between currencies, stocks and bonds from the perspective of four based currencies (USD, EUR, GBP and AUD)Correlations between asset market performance and foreign currency performance versus base currency. For example, a positive correlation of 0.18 for Euro area stockswith USD as base currency implies that the euro tends to rise versus the dollar when European stocks rise. Correlations based on monthly return data.

Foreign currency

Base currency Foreign asset Sample period USD EUR GBP AUD CAD JPY

USD bonds 2005 - 09 (5 years) NA -0.29 -0.48 -0.56 -0.46 0.35

stocks NA 0.18 0.04 0.40 0.29 -0.34

bonds 2000 - 09 (10 years) NA -0.04 -0.29 -0.34 -0.27 0.14

stocks NA -0.02 -0.04 0.25 0.25 -0.20

bonds 1990 - 2009 (20 years) NA -0.09 -0.21 -0.17 -0.04 0.04

stocks NA -0.12 -0.12 0.19 0.26 -0.08

EUR bonds 2005 - 09 (5 years) 0.04 NA -0.58 -0.58 -0.52 0.19

stocks -0.33 NA -0.16 0.47 0.01 -0.29

bonds 2000 - 09 (10 years) -0.13 NA -0.49 -0.50 -0.44 -0.03

stocks -0.09 NA -0.09 0.39 0.14 -0.12

bonds 1990- 09 (20 years) -0.16 NA -0.23 -0.19 -0.11 -0.04

stocks 0.00 NA 0.00 0.24 0.21 -0.03

GBP bonds 2005 - 09 (5 years) 0.49 0.25 NA -0.12 -0.05 0.42

stocks -0.24 0.05 NA 0.39 0.01 -0.32

bonds 2000 - 09 (10 years) 0.19 0.24 NA -0.14 -0.06 0.15

stocks -0.09 -0.02 NA 0.33 0.13 -0.19

bonds 1990 - 2009 (20 years) 0.02 0.18 NA -0.07 -0.02 0.05

stocks -0.04 -0.12 NA 0.17 0.14 -0.10

AUD bonds 2005 - 09 (5 years) 0.26 0.28 -0.37 NA -0.29 0.25

stocks -0.49 -0.44 -0.48 NA -0.30 -0.36

bonds 2000 - 09 (10 years) 0.12 0.28 -0.20 NA -0.23 0.11

stocks -0.33 -0.36 -0.39 NA -0.19 -0.27

bonds 1990 - 2009 (20 years) 0.11 0.07 -0.12 NA -0.04 0.08

stocks -0.25 -0.22 -0.26 NA -0.13 -0.17Source: J.P. Morgan


http://slidepdf.com/reader/full/jp-morgan-managing-fx-hedge-ratios-a-framework-for-strategic-and-tac


John Normand (44-20) 7325-5222 [email protected]. Morgan Securities Ltd

10

The optimal hedge ratio: no size fits all

In the final balance, what hedge ratio makes sensestrategically? There is no uniform, strategic hedge ratio. The appropriate policy is particular to the investor givenfour variables: (1) the allocation between domestic andinternational assets; (2) the currency allocation of foreignassets; (3) the consistency of historical volatilities andcorrelations in the future; and (4) the investor’s risk

preference. The optimal hedge ratio therefore will vary byinvestor and over time.

While there are no absolutes, several guidelines apply to the

merits of fully hedged, unhedged and semi-hedged

benchmarks.

•

For investors in G-10 bonds, FX exposure should bemostly hedged given how currency volatility often

drives overall portfolio risk. The exceptions are for

catastrophe insurance and asset/liability matching

discussed on page 7.

• FX hedge ratios for equity portfolios are more

debatable. For investors concerned mainly with

minimising volatility over the long term, 100% hedged

benchmarks are lower variance than unhedged ones for

USD, EUR and GBP-based investors given the generally

positive correlation between foreign currency

performance and foreign equity markets. The opposite

holds for AUD-based investors: unhedged portfolios areless volatile over the long term.

Even for those investors seeking to minimise volatility,

100% hedged benchmarks entail important limitations.

• Full hedging eliminates potential short-term gains

from tactical trading. Given the profitability of active

currency managers and rule-based trading strategies,

100% hedging precludes a potential alpha opportunity.

Section III discusses this issue in more detail. Even a

100% hedge ratio with the flexibility to deviate from the

policy is quite constrained. If the benchmark is 100%

hedged against the base currency, the manager may

hedge less than the benchmark but not more. Thus the

manager can only outperform the benchmark in

environments when the foreign currency is rising.

Chart 12: Volatility of hedged and unhedged S&P500 returns forGBP-based investors, 1988-2010Based on monthly returns over rolling 12 month window

0%

5%

10%

15%

20%

25%

30%

35%

88 90 92 94 96 98 00 02 04 06 08

unhedged hedged

Source: J.P. Morgan

Chart 13: Volatility of hedged and unhedged S&P500 returns forAUD-based investors, 1988-2010Based on monthly returns over rolling 12 month window

0%

5%

10%

15%

20%

25%

30%

35%

88 90 92 94 96 98 00 02 04 06 08

unhedged hedged

Source: J.P. Morgan

Chart 14: Volatility of hedged and unhedged Nikkei returns for USD-based investors, 1988-2010Based on monthly returns over rolling 12 month window

0%

10%

20%

30%

40%

50%

89 92 95 98 01 04 07 10

unhedged hedged

Source: J.P. Morgan





11

• Full hedging can also impose substantial cash flow

requirements during period of extreme market stress,

such as the 2008-09 credit crisis. Ironically in 2008,some investors found their fully-hedged benchmarks to

be even more volatile than unhedged ones because they

were short a foreign currency which was appreciating as

the foreign equity market fell. In some cases investors

were forced to liquidate underlying assets to fund the

cash flow obligations of a currency hedging program,

thus reinforcing the decline in equity and currency

markets. Semi-hedged benchmarks (50/50) faced a

similar dilemma, although to a lesser degree Such would

have been the predicament of UK and Australian

investors who hedged the currency risk on S&P500

exposure (so were short USD vs GBP and AUD as both

currencies collapsed), or US hedgers who were short theyen as an overlay to their Nikkei investments. Charts 12

– 14 highlight how, for the first time in decades,

unhedged equity returns were more volatile than hedged

ones.

In those case where a zero hedge ratio seems sensible

long-term due to a negative correlation between foreign

currency and equity returns – the case in foreign equities

from an AUD perspective – such a policy could expose

investors to substantial short-term volatility. And an

unhedged benchmark limits flexibility as much as a 100%

one does. If the benchmark is unhedged against the base

currency, the manager may hedge more than the benchmark

but never less. Thus the manager can only outperform the

benchmark in an environment when the foreign currency is

depreciating.

50/50 hedge ratio as policy of least regret

Given the particulars of investor or corporate exposure andthe tradeoffs inherent in various policy options, risk managers have two options for selecting the optimal hedgeratio. They can run a mean-variance optimization to

determine the minimum variance hedge ratio, thenoverlay some discretion in the final policy decision.

Alternatively they can adopt a symmetric benchmark hedge ratio of 50%. Symmetric benchmarks appeal

because they address the risk inherent in overseas investing;allow flexibility to manage cash flow; and minimize the risk of overfitting inherent in mean-variance optimization.

They also allow investors to capture profit opportunities.With symmetry, if managers believe the base currency willappreciate, they can hedge the foreign currency. If theyexpect the base to depreciate, they can buy additionalforeign currency exposure over the benchmark. The payoff structure resembles a zero cost call option on the currency

Chart 15: Performance with and without overlay

-12%

-9%

-6%

-3%

0%

3%

6%9%

12%

-12% -9% -6% -3% 0% 3% 6% 9% 12%

currency return

p a s s - t h r o u g h

t o i n

v e s t o r

unhedged

symmetric hedge + overlay

Source: J.P. Morgan

which allows the investor to participate in currency gainswhile avoiding currency losses (chart 15). It is thus the

policy of least regret for many investors.

Even if the 50/50 policy appeals intuitively as a benchmark,many situations require a more dynamic approach,

particularly for those investors concerned about entrylevels, cash flows and profit opportunities. Sections III andIV of this paper address these issues in more detail byapplying fair value and momentum models.





12

Appendix 1: Stock and bond returns and volatility with EUR as base currency

Chart 1: Equity returns hedged vs unhedged into EUR, 1989-2009

-6%

-3%

0%

3%

6%

9%

12%

EUR JPY GBP USD AUD CAD MSCI

ex-EU


Source: J.P. Morgan

Chart 2: Equity volatility hedged vs unhedged into EUR, 1989-2009

0%

10%

20%

30%

40%

EUR JPY GBP USD AUD CAD MSCI

ex-EU


Source: J.P. Morgan

Chart 3:Equities for EUR-based investors: return vs risk, 1988 – 2009

-

5

10

15

20

10 15 20 25 30Risk, %

Return, %

Euro area

equitiesWorld ex Euro area,

hedged

World ex Euro area, unhedged

Source: J.P. Morgan

Chart 3: Bond returns hedged vs unhedged into EUR, 1993-2009

0%

3%

6%

9%

EUR JPY GBP USD AUD CAD GBI


Source: J.P. Morgan

Chart 4: Bond volatility hedged vs unhedged into EUR, 1993-2009

0%

5%

10%

15%

EUR JPY GBP USD AUD CAD GBI


Source: J.P. Morgan

Chart 6: Bonds for EUR-based investors: return vs risk, 1993 - 2009

-

5

10

15

20

- 5 10 15Risk, %

Return, %

Euro area bonds

World ex Euro area,

unhedgedWorld ex Euro

area, hedged





13

Appendix 2: Stock and bond returns and volatility with GBP as base currency

Chart 1: Equity returns hedged vs unhedged into GBP, 1988-2009

-3%

0%

3%

6%

9%

12%

GBP JPY EUR USD AUD CAD MSCI

ex-UK


Source: J.P. Morgan

Chart 2: Equity volatility hedged vs unhedged into GBP, 1988-2009

0%

5%

10%

15%

20%

25%

30%

GBP JPY EUR USD AUD CAD MSCI ex-

UK


Source: J.P. Morgan

Chart 3: Equities for GBP-based investors: return vs risk, 1988 - 2009

-

5

10

15

20

25

15 20 25

Risk

Return

UK equities

World ex UK, hedged

World ex UK, unhedged

Source: J.P. Morgan

Chart 4: Bonds hedged vs unhedged into GBP, 1993-2009

0%

3%

6%

9%

12%

15%

18%

GBP JPY EUR USD AUD CAD GBI ex

UK


Source: J.P. Morgan

Chart 5: Bond volatility hedged vs unhedged into GBP, 1993-2009

0%

5%

10%

15%

20%

25%

GBP JPY EUR USD AUD CAD GBI ex

UK


Source: J.P. Morgan

Chart 6: Bonds for GBP-based investors: return vs risk, 1993 - 2009

-

5

10

15

20

- 5 10 15Risk

Return

UK bonds

World ex UK, unhedgedWorld ex UK, hedged

Source: J.P. Morgan





14

Appendix 3: Stock and bond returns and volatilit y with AUD as base currency

Chart 1: Equity returns hedged vs unhedged into AUD, 1988-2009

-3%

0%

3%

6%

9%

12%

AUD JPY EUR USD GBP CAD MSCI

world


Source: J.P. Morgan

Chart 2: Equity volatility hedged vs unhedged into AUD, 1988-2009

0%

10%

20%

30%

AUD JPY EUR USD GBP CAD MSCI

world


Source: J.P. Morgan

Chart 3: Equities for AUD-based investors: return vs risk, 1988 - 2009

-

5

10

15

20

15 20 25Risk

Return

AustraliaWorld,

hedged

World, unhedged

Source: J.P. Morgan

Chart 4: Bonds hedged vs unhedged into AUD, 1993-2009

0%

3%

6%

9%

AUD JPY EUR USD GBP CAD GBI ex

AUD


Source: J.P. Morgan

Chart 5: Bond volatility hedged vs unhedged into AUD, 1993-2009

0%

5%

10%

15%

20%

25%

AUD JPY EUR USD GBP CAD GBI ex

AUD


Source: J.P. Morgan

Chart 6: Bonds for AUD-based investors: return vs risk, 1988 - 2009

-

2

4

6

8

10

- 5 10 15 20Risk

Return

Australia bonds

World, unhedged

World, hedged

Source: J.P. Morgan





15

Appendix 4: Emerging market bond returns and volatili ty with USD as base currency Chart 1: Emerging market bond returns, 2002-2009 Annual %, USD terms, based on J.P. Morgan EM-GBI

0%

5%

10%

15%

20%

EM-GBI GBI-EM

Europe

GBI-EM Asia GBI-EM

Latam

GBI-EM

Mideast & Africa


Source: J.P. Morgan

Chart 2: Emerging market bond volatility, 2002-2009 Annual %, USD terms, based on J.P. Morgan EM-GBI

0%

5%

10%

15%

20%

25%

30%

EM-GBI GBI-EM

Europe

GBI-EM Asia GBI-EM

Latam

GBI-EM

Mideast &

Africa


Source: J.P. Morgan




Gabriel de Kock (1-212) [email protected] Chase Bank NA

16

III. Using fair value models to

focus strategic hedge ratios4

• 100% hedging can create onerous cash flow

obligations, 0% hedging leaves investors and

corporates too exposed to exchange rate swings, and

a 50% hedge ratio strikes some as arbitrary.

• An alternative strategy would focus hedging on the

most misaligned currencies: hedge expensive pairs,

not cheap ones.

• J.P. Morgan’s fair long-term fair value model can

inform this process. The model derives fair valueestimates for G-10 currency cross rates from the

long-term relationship between real trade-weighted

exchange rates and fundamental economic drivers.

• FX misalignments are statistically and economically

significant in predicting currency moves out of

sample over 6-, 12- and 18-month horizons.

• A back test over the 2004-08 period shows that

trading rules exploiting misalignments consistently

outperform carry.

• By implication, adjusting medium-term hedge ratios

in response to misalignments should improve

portfolio performance or reduce hedging costs.

If 100% hedging can create cash flow problems in high-volatility environments, and zero hedging leaves investorsand corporates too exposed to exchange rate swings, whatare the alternatives? From a strategic perspective, investorsand corporates could focus hedging activity on thosecurrencies which are most misaligned. Very expensivecurrencies would seem most vulnerable to a fall, so should

be fully or mostly hedged over a medium-term horizon suchas one year. Very cheap currencies would be biased toappreciate over time, so should be left unhedged or partiallyhedged over the medium term.

Determining which currencies qualify as cheap or expensiveis a valuation exercise, for which we employ a fair valueframework developed in 2008 (see footnote). This sectionsummarizes J.P. Morgan’s fair-value model for the G-10

4 This section updates and extends to other base currencies the fair value model outlined in two previous publications: A new fair

value model for G10 currencies (September 6, 2008) and Hedging

and trading long-term FX with J.P. Morgan’s fair value model

(April 24, 2009), both by Gabriel de Kock.

exchange rates. It explains our approach to modelling fair values, the drivers of fair value exchange rates and their

empirical importance, and the model’s performance inguiding currency investment and hedging decisions.

The fair-value exchange rate model is designed as anobjective method for estimating exchange rate benchmarksthat would be sustainable over longer time horizons

between one and a half and five years. It takes a multi-currency approach to fair value exchange rates, defining fair values in terms of trade-weighted exchange rates. Startingfrom the presumption that real trade-weighted exchangerates vary systematically with long-run economicfundamentals, the model defines the relationship betweenreal trade-weighted exchange rates and their fundamentaldrivers as the long-run equilibrium (or sustainable) real

exchange rates (ERER). Estimated ERER, together withtrade weights and historical bilateral nominal exchangerates are used to calculate mutually consistent fair-value

bilateral nominal exchange rates.

Long-term FX fundamentals

The fundamental drivers of long-term equilibrium realexchange rates in JP Morgan's fair-value model are thosethat economic theory suggests should affect a wide range of currencies over a wide range of time periods andcircumstances. These variables and their expected impactson the ERER are:

Terms of Trade: A rise in the terms of trade (export pricesrelative to import prices) increases profits and incomes andhence national wealth, attracting capital inflows and

boosting capex and consumer demand. Rising demandtightens capacity constraints, driving up real interest ratesand thereby attracting foreign capital. As a result, theERER impact of the terms of trade is proportional to theopenness of the economy and the duration of the terms of trade change.

Current account balance and external debt: We interpreta current account deficit as a capital account surplus,reflecting the attractiveness of domestic assets in the globalmarketplace and the fact that capital can cross borders

instantaneously, while trade and investment flows changeslowly over time. In short, a rising current account deficit isassociated with exchange rate appreciation. However, acountry cannot run a current account deficit – spendingmore than it earns – forever. Persistent external deficitsimply an accumulation of external debt and rising debtservice payments to foreign investors. Thus, higher foreignliabilities require a weaker currency to generate the increasein export earning to service the debt.

Government debt: High levels of government debtdepreciate the currency over the long run, because





17

government borrowing crowds out private borrowing andinvestment, lowering potential economic growth and long-

term asset returns and raises expected future tax rates andinflation rates, both of which tend to undercut long-termasset market performance.

Inflation: High inflation typically reflects bad economic policies and is associated with high inflation volatility, bothof which distort resource allocation lowering the return oncapita and the ERER.

Estimating fair values

We estimate the fair-value exchange rates in three steps.First, we use a panel cointegration analysis to estimate thestable long-term relationship between the real exchange rateand its fundamental drivers. This step combines data from

all of the G-10 economies and by assuming that thefundamental drivers affect real exchange rates similarlyacross countries, allows us to estimate a long-termrelationship with relatively short data spans.

Next, we use the parameters of the estimated long-termrelationship and observations on the fundamental drivers tofind the equilibrium or fair-value real exchange rates.However, period-to-period swings of the drivers containtransitory noise that distorts the estimates when pluggedinto a relationship designed to capture long-term shifts. Tocompensate, we scale down the elasticities on thefundamental drivers in proportion to the relative importanceof transitory noise in their period to period fluctuations.And, reflecting the forward-looking nature of financialmarkets, to the extent that real observed real exchange ratechanges predict future shifts in the permanent component of the macroeconomic drivers, we incorporate a fraction of

period-to period changes in the observed real exchange inthe estimated long-run equilibrium real exchange rate. Todetermine the appropriate scale factor, we use time-seriestechniques to decompose the data into permanent andtemporary components.

Finally, we use the our estimated long-run equilibrium realexchange rates obtained from this decomposition, alongwith trade weights and historical nominal exchange rates, to

calculate fair-value bilateral nominal exchange rates.Intuitively, since price levels at each point in history aregiven, all of the real misalignment can be attributed tonominal misalignments. So, we use the trade weights todistribute nominal misalignments across the G-10 to match,simultaneously, the real misalignments of all of the G10countries.

Table 1 shows the equation used to estimate the G10equilibrium real exchange rates, with the scaled-downcoefficients denoted as “permanent.” The elasticitiesmeasuring the equilibrium real exchange rate impact of shift

in the economic drivers are of the expected sign andmagnitude. Moreover, the analysis attributes about one-

quarter of the swings in observed real exchange rate to shiftin the underlying fair value.

Charts 1 and 2 illustrate historical real equilibriumexchange rate estimates for the USD and the AUD.Strikingly, the estimated real fair value of the USD appearsquite stable over time compared to the swings in the marketreal exchange rate, while most of the swings in the AUDreal exchange rate are attributed to swings in its fair value.The key difference between the two currencies is that as acommodity price-driven swings in Australia’s terms of tradeaccount for much of the swings in the real AUD and hencefor much of the swings in its fair value. By contrast, the USwith its well-diversified import and export baskets

experiences mild terms of trade swings over time and thoseswings, given that the US is a relatively closed economyhave only a small impact on the real exchange rate.

Our latest fair value estimates for G-10 real trade-weightedexchange rates and bilateral exchange rates vs. the USD areshown in Tables 2 and 3. Table 2 highlights that, as of March, the USD and EUR were close to fairly valued inreal-trade-weighted terms, while GBP was about 15%undervalued and the JPY 11% overvalued. However, onaverage, deviations from fair value had declined notablysince 2009 Q3, led by NOK, SEK, CAD, and EUR.

Table 1: Estimated long-run equilibrium real exchange rate model,1980-1Q08 Driver Estimate Permanent

Log(RER) 0.2637

Log(Terms of Trade) x Openness 1.3429 0.9869

(26.86)

C urrent Ac count Balance (% GDP) -1.2971 -0.9550

(-10.09)

Net Inv estment Income (% GDP) 0.4755 0.3501

(2.02)

Gov ernment Debt (% GDP) -0.2331 -0.1716

(-12.27)

Inflation (%) -0.6427 -0.4732

(-1.92)

R2

= 0.88

Note: Standard errors in parentheses.Source: J.P. Morgan





18

The EUR, CHF, CAD and AUD were less than 3% mis- priced against the USD in 1Q10, while GBP, NOK and

SEK remained cheap and JPY the most persistentlyovervalued G10 currency.

Currency misalignments as signals forhedging/investment decisions

In most quarters, the JP Morgan fair value model identifiessignificant misalignments among the G-10. At the mostrecent update of the fair value estimates in mid-March 2010the average misalignment against the USD stood at 5.2%.But, as Table 4 shows, the misalignments are much moreextreme for some G-10 crosses, particularly for GBP andJPY. For example, JPY was 13% overvalued against CHF,15% against SEK and 19% against NOK, while GBP was

more than 20% undervalued against AUD, CAD and NZD.Misalignments relative to fair value predict exchange-ratechanges out of sample, particularly over 12-18-monthhorizons, suggesting that using deviations from fair valueinto account in hedging decisions could improve portfolio

performance. Below we demonstrate that a simple buy-and-hold trading rule – acting when misalignments exceeded thecost of carry – generally outperformed the forwards over 6-,12- and 18-month horizons, typically performing best over 6-month investment horizons. Of course the usual caveatapplies to these results: they reflect FX developments in a

particular time period, and cannot prove that using the fair-value model in this manner would necessarily produce

profits in the future.

A back test of fair value signals

To assess whether the fair-value model provides usefultrading/hedging signals, we back-test a simple buy-and-holdtrading rule. For a signal to go long or short FX against a

base currency, we compare its misalignment with itsforward premium or discount as follows:

(a) go short foreign currency, buying 1 unit of the basecurrency forward if the percentage overvaluation of theforeign currency exceeds its forward discount (and hencethe cost of carry) by more than a predetermined critical or threshold value. Reverse the position at the spot rate when

the forward contract matures. Similarly, (b) go long FX,selling $1 forward if the percentage undervaluation of foreign currency exceeds its forward premium by more thanthe predetermined threshold and reverse the position at thespot rate when the forward contract matures.

Chart 1: Equilibrium and actual real trade-weighted USD, 1980-1Q10

100

110

120

130

140

150

160

80 85 90 95 00 05 10

Actual Fair Value

Source: J.P. Morgan

Chart 2: Equilibrium and actual real trade-weighted AUD, 1980-1Q10

55

65

75

85

95

105

80 85 90 95 00 05 10

Actual Fair Value

Source: J.P. Morgan

Table 2: G-10 actual and fair-value real trade-weighted exchange rateindexes (2Q09=100), 2Q09-1Q108

2Q09 3Q09

Actual FV Actual FV Actual FV Actual FV

USD 100 99.6 96.2 98.3 93.1 96.4 94.9 96.8

EUR 100 96.6 101.6 97.3 102.9 97.2 96.1 95.8

JPY 100 92.3 100.7 92.5 101.9 92.0 102.3 92.1

GBP 100 120.2 102.4 120.4 100.1 118.5 101.3 118.8

CHF 100 103.4 98.9 102.7 101.2 104.2 102.0 104.4

CAD 100 108.2 103.9 111.2 107.7 111.9 109.5 112.3

AUD 100 105.3 106.3 107.3 110.9 109.4 109.5 109.1

NZD 100 102.2 108.5 101.2 114.0 106.4 113.2 106.2

NOK 100 116.9 104.6 118.1 109.2 119.0 110.6 119.3

SEK 100 108.3 103.0 111.6 102.7 111.4 106.7 112.3

4Q09 1Q10

Note: Exchange rates are quarterly averages.Source: J.P. Morgan.





19

The objective of the back test is to determine whether thereare values of the threshold for which this trading strategy

would have yielded profits in excess of carry.The back-test covers the period from 1Q04 to 4Q08 and is

based on 21 sequential estimates of the fair value model ondata samples from 1Q80 to 4Q03 through 1Q80 to 4Q08.To accurately reflect the time lags in the availability of thedata used in estimating fair values, we measuremisalignments relative to the quarterly fair value estimatelagged one quarter. And to mimic a realistic tradingenvironment we assume that transactions occur in the firsttrading week of every quarter and at the averages of theforward and spot rates over the first five trading days of thequarter. And the signal is derived by comparing the averageforward rates over the first five trading days of the quarter

to the quarterly fair-value estimate.

Finally, for ease of interpretation and to explore therobustness of the results we report back-tests for USD-,EUR-, GBP-, and AUD-based investors and corporates.

USD-based investors and corporates

Table 5 shows the results of the strategy for a USD-basedinvestor or corporate over 6-, 12- and 18-month horizons(two, four and six quarters) and decision threshold values at0%, 5% and 10%. The strategy outperformed carry, onaverage, performing best over 6-month investment horizonsirrespective of the size of the decision threshold. Asexpected, the strategy is more profitable for greater initial

misalignments, that is, for higher threshold values. Indeed,in our sample, acting on initial carry-adjustedmisalignments greater than 10% outperforms carry by about3.6% over a 6-month horizon, about one-and-a-half timesthe average return for a 0% threshold.

A 10% misalignment threshold might seem large. In Table4, only the JPY and GBP were misaligned by more than10% against the USD in mid-March. But for the G10currencies 19 out of 45 crosses were misaligned by morethan 10%. Finally, irrespective of the decision threshold,

the profitability of the buy-and-hold strategy

deteriorated over longer holding periods.

Table 3: G-10 actual and fair-value bilateral exchange rates vd USD,2Q09-1Q10

2Q09 3Q09

Actual FV Actual FV Actual FV Actual FV

EUR/USD 1.36 1.36 1.43 1.39 1.48 1.41 1.39 1.41

USD/JPY 97 105 94 103 90 101 90 100

GBP/USD 1.55 1.84 1.64 1.86 1.63 1.85 1.57 1.83

USD/CHF 1.11 1.07 1.06 1.05 1.02 1.03 1.06 1.03

USD/CAD 1.17 1.10 1.10 1.05 1.06 1.03 1.05 1.04

AUD/USD 0.76 0.79 0.83 0.81 0.91 0.86 0.90 0.88

NZD/USD 0.60 0.61 0.67 0.60 0.73 0.65 0.71 0.65

USD/NOK 6.50 5.47 6.11 5.47 5.69 5.31 5.84 5.35

USD/SEK 7.91 7.19 7.28 6.81 7.01 6.61 7.19 6.75

4Q09 1Q10

Note: Exchange rates are quarterly averages.

Source: J.P. Morgan.

Table 4: G-10 bilateral exchange rate deviations from 1Q10 fairvalues (%), 12-18 Mar 10

EUR JPY GBP CHF CAD AUD NZD NOK SEK

USD -3.2 10.4 -18.6 -2.9 2.0 4.5 9.0 -8.8 -5.0

EUR 13.5 -15.5 0.2 5.2 7.7 12.2 -5.7 -1.8

JPY -29.0 -13.3 -8.4 -5.8 -1.4 -19.2 -15.3

GBP 15.7 20.6 23.2 27.6 9.8 13.7

CHF 4.9 7.5 12.0 -5.9 -2.0

CAD 2.5 7.0 -10.8 -6.9

AUD 4.5 -13.4 -9.5

NZD -17.8 -14.0

NOK 3.9

Misalignments measured as average market rates for 12-18 Mar vs 1Q10 fair values. Anegative number denotes undervaluation of the column currency vs the row currency.

Source: J.P. Morgan.

.





20

Table 5: USD as base currency - profitability of fair value-based trading rules (% annual, net of carry), 2004-08

Horizon Avg. AUD CAD EUR JPY NOK NZD SEK CHF GBP

2 2.3% 2.9% 2.9% -0.4% 2.9% 2.1% 3.5% 2.9% 0.5% 2.9%

4 0.4% -0.4% 3.4% -2.1% 1.3% -1.5% 2.8% 0.0% -0.6% 0.4%

6 0.2% -1.5% 1.8% -1.9% 2.0% -0.6% 2.7% -0.6% -0.3% -0.1%


2 2.6% 2.4% -0.9% -0.4% 1.1% 3.0% 7.2% 4.3% 0.5% 6.5%

4 0.5% -1.1% 2.1% -2.1% 0.5% -0.5% 2.9% 0.3% -0.6% 3.2%

6 0.3% -1.4% 0.8% -1.9% 2.8% -0.4% 1.8% -0.5% -0.3% 1.3%


2 3.6% 4.5% NA 2.5% -0.9% 3.3% 5.5% 6.4% 1.2% 6.1%

4 1.1% 2.3% NA -0.4% 1.6% 0.7% 0.4% 1.7% -0.8% 3.2%6 0.8% 1.7% NA -1.1% 1.6% 0.1% 1.3% 0.9% -0.3% 2.1%

Threshold = 5%

Threshold = 10%

Threshold = 0%

Note: NA reflects situations when there were no trades for a particular currency and time horizon.Source: J.P. Morgan

EUR-based investors and corporates

Among the four base currencies considered here, the G-10fair value model is the least successful in generating excessreturns for EUR-based investors or corporate hedgers. Asfor USD-based investors, positions taken on the basis of fair value model signals generate excess returns over carry, butthe excess returns are modest – about one-fifth on averagethe excess returns enjoyed by USD-based decision-makers.And more clearly than elsewhere the signals are morevaluable at higher decision thresholds, albeit by a relativelysmall return margin.

Consistent with the patterns observed for USD-basedinvestors, EUR/USD positions always underperformedinterest rate spreads, irrespective of the decision horizon or the reaction threshold. This outcome appears to reflect thespecifics of the sample period used. The euro trendedabove and away from its fair value for most of our back-testsample – the powerful trend reversal in the second half of 2008 simply came too late to affect the overall results, butwould have done so if we had extended the sample to allowmore positions to mature.

Table 6: EUR as base currency - profitability of fair value-based trading rules (% annual, net of carry), 2004-08

Horizon Avg. AUD CAD USD JPY NOK NZD SEK CHF GBP

2 0.2% -0.9% NA -0.4% 2.0% -1.3% 4.2% -2.0% 3.4% -3.4%

4 0.0% 0.9% 0.7% -2.1% -2.0% 1.9% 0.1% 0.0% 3.0% -2.5%

6 0.2% 1.2% NA -1.9% 0.7% 0.3% 0.1% 1.0% 2.1% -1.6%


2 0.5% 0.7% NA -0.4% 2.9% 0.5% 1.4% -2.2% 1.4% -0.7%4 0.4% 2.1% 0.7% -2.1% 0.4% 0.0% 1.3% 0.2% 1.0% 0.1%

6 0.3% 1.5% NA -1.9% 0.7% -0.1% 0.0% NA 0.8% 0.7%


2 0.6% NA NA 2.5% 5.6% -0.1% -2.6% NA 0.2% -1.7%

4 0.6% 0.5% 1.5% -0.4% 2.5% -0.1% 0.1% NA 0.5% NA

6 0.1% 0.3% NA -1.1% 0.8% NA 0.7% NA NA NA

Threshold = 10%

Threshold = 0%

Threshold = 5%

Note: NA reflects situations when there were no trades for a particular currency and time horizon.Source: J.P. Morgan





21

GBP-based investors and corporates

The results of the misalignment-driven buy-and hold

strategy for a GBP-based decision-maker are shown inTable 7. In comparison with the results using the USD asthe base currency, the performance of the model is quiteconsistent over decision thresholds and investment

horizons. In general, an investment strategy based onthe fair value signals outperformed carry by more than

2% annually on average and the deterioration in performance over longer decision horizons appears quite

modest. But, by the same token, waiting for large initialmisalignments does not consistently improve performance. Also, before late 2008, GBPmisalignments tended to be modest. As a result, thedecisions based on a 10% action threshold resulted in avery small number of positions and, therefore, the resultare potentially less reliable.

Table 7: GBP as base currency - profitability of fair value-based trading rules (% annual, net of carry), 2004-08

Horizon Avg. AUD CAD EUR JPY NOK NZD SEK CHF US

2 2.3% 4.5% 4.4% -3.4% 2.5% -0.7% 6.2% 2.0% 2.2% 2.9%

4 2.5% 3.5% 3.2% -2.5% 11.3% -2.1% 4.1% -0.3% 5.2% 0.4%

6 1.5% 1.7% 2.9% -1.6% 9.9% -0.9% 1.8% 0.1% 0.1% -0.1%


2 1.9% 0.5% 5.3% -0.7% 2.0% -3.2% 1.5% 2.4% 2.4% 6.5%

4 2.6% NA 4.4% 0.1% 7.6% -1.6% 2.6% 0.4% 4.1% 3.2%

6 1.7% 1.3% 3.2% 0.7% 6.0% -0.8% 2.0% -0.1% 1.7% 1.3%


2 2.2% 0.5% 6.9% -1.7% 2.1% 0.2% NA NA 1.4% 6.1%

4 2.0% NA 3.1% NA 0.9% NA NA NA 0.9% 3.2%

6 1.7% NA 1.5% NA NA NA NA NA 1.4% 2.1%

Threshold = 0%

Threshold = 5%

Threshold = 10%

Note: NA reflects situations when there were no trades for a particular currency and time horizon. Source: J.P. Morgan

AUD-based investors and corporates

The fair-value model also proved a reliable guide for AUD- based decision-makers. Table 8 shows that the signals aremost reliable for a 6-month investment horizon, as in thecases of decisions where the USD and GBP are the basecurrencies. The strategy outperformed carry by over 2.5%

annualized, on average, over a 6-month investment horizon.Perhaps counter-intuitively and unlike the case of the USD-

based decisions, higher action thresholds did not improve

portfolio performance. And similar to the results for GBP-based investors and corporate hedgers, thedeterioration of trading strategy’s performance at longer horizons is relatively small.

Table 8: AUD as base currency - profitability of fair value-based trading rules (% annual, net of carry), 2004-08

Horizon Avg. CAD EUR JPY NOK NZD SEK CHF GBP USD

2 3.7% 1.9% -0.9% 8.4% -0.3% 2.9% 5.6% 8.1% 4.5% 2.9%

4 2.4% -0.1% 0.9% 6.4% 0.9% 0.3% 3.2% 7.4% 3.5% -0.4%

6 1.9% 0.1% 1.2% 10.8% 0.1% -0.3% 1.9% 2.9% 1.7% -1.5%


2 2.4% 1.9% 0.7% 6.5% -0.4% 0.3% 3.2% 6.4% 0.5% 2.4%

4 2.0% 0.6% 2.1% 7.1% -0.1% 0.3% 2.3% 5.0% NA -1.1%

6 1.5% 0.4% 1.5% 5.8% 0.0% 0.4% 1.6% 3.8% 1.3% -1.4%


2 2.3% 3.4% NA 6.4% 0.2% 0.0% 0.7% 2.5% 0.5% 4.5%

4 1.9% 0.4% 0.5% 3.8% NA NA 1.1% 3.0% NA 2.3%

6 NA 0.7% 0.3% 3.0% NA NA 0.7% 2.7% NA 1.7%

Threshold = 0%

Threshold = 5%

Threshold = 10%

Note: NA reflects situations when there were no trades for a particular currency and time horizon. Source: J.P. Morgan




Matthew Franklin-Lyons (1-212) 834 [email protected] Chase Bank NA

22

IV. Using alpha models to

adjust hedge ratios tactically5

• Dynamic hedging around a medium-term

benchmark requires discretion or trading rules. The

latter is tested in this section, building on two

momentum models commonly used by currency

overlay managers and global macro funds.

• Price momentum is the most basic trading strategy;

it simply buys (sells) the best (worst) performing

pairs. A rate momentum strategy, or forward carry,

would buy currencies in whose favor interest rate

expectations are moving.

• Standard industry models generate signals for two

to four week horizons. Minor modifications,

however, can generate signals of one to three

months, which are more appropriate for hedge

rebalancing.

• A price momentum model which adjusts

dynamically around a 50/50 hedge ratio

outperforms the benchmark by about 100bp

annually, depending on the base currency.

Information ratios on the strategy range from 0.2 to

0.5.

• A rate momentum (forward carry) model generates

comparable outperformance but with more

consistency across sample periods.

• Strategies are robust to various sample periods and

model specifications.

Alpha strategy basics

Short-term trading models, or alpha strategies, areemployed extensively in currency overlay and global macrohedge fund communities to replicate the decision making

process of investors. The approaches go by several names,such as systematic, tactical, short-term or alpha. We use the

term alpha strategy since it conveys the underlyingobjective of outperforming a benchmark. Over the pastdecade J.P. Morgan has developed a suite of simple,intuitive trading rules to replicate common investmentstrategies across all major asset classes. The series of methodology notes – entitled Investment Strategies – isarchived on www.morganmarkets.com.

5 This section amplifies models originally detailed in Alternatives

to Standard Carry and Momentum in FX , Normand and Ghia,(2008).

Carry and momentum are the most commonly followedtrading strategies across asset markets and in currencies.

Momentum strategies are based on the empirical tendencyof strong-performing assets to outperform again in thefuture, usually due to behavioural biases such asunderreaction and overreaction.6 In FX, the standardmomentum model trades in the direction of previous spotmovements, as determined by a filter, moving average, or lookback rule. Despite the tendency to dismiss theseframeworks as overly simplistic, the profits from such astrategy have been decent, at roughly 3.7% over the past tenyears (table 1). Momentum strategies perform better inhigh-volatility environments but incur sizable drawdowns,and perform poorly in range bound environments or thoseexhibiting frequent reversals. Hence the poor performanceof a basic momentum rule over the past year in which thedollar has experienced several sharp reversals (chart 1).

Chart 1: Returns on a simple price momentum rule since 2000Indexed returns based on trading G-10 currencies versus USD. Tradingrule uses 1-mo lookback for price momentum and daily rebalancing

90100

110

120

130

140

150

160

170

180

00 02 04 06 08 10

Momentum only (9 USD pairs)

Source: J.P. Morgan

The other standard alpha model is carry, which buys highyield currencies versus low-yield ones. This strategy hasreturned 4.2% per annum over the past decade, but given itshigh volatility (12%), its risk adjusted returns are the lowestof all alpha strategies we track (table 1). The very lowfrequency of signal changes – a carry model would almostalways be long AUD and short JPY – also renders it inappropriate for dynamic hedging.

An alternative to standard carry, which derives signals fromthe level of interest rate differentials, is forward carry,

6 Underreaction reflects investors’ inability or unwillingness toadjust views and positions quickly; either because they await fuller information to make a decision, or because they are reluctant toappear non-consensus. Accordingly, prices adjust slowly towards amarket’s fundamental value, in the process producing short-termtrends. Overreaction is also based on cognitive biases. Althoughmost investors adjust their expectations fully, some extrapolatethis positive news into the future, thus leading prices to overshootfundamental value.





23

which derives signals from changes in interest ratedifferentials. This strategy is based on the observation that

currencies respond as much to spread changes as they do tospread levels. Low-yielding currencies can appreciateversus high-yielding currencies when rate spreads narrow intheir favor, either because rates are falling more quickly inthe high-yield country or rising more quickly in the low-yield country. For this reason, the forward carry framework

proposes to position on changes in expected cash ratespreads rather than current cash rate differentials.

This pattern is evident in both EUR/USD and USD/JPY inrecent years (charts 2 and 3). In the case of EUR/USD, ratedifferentials as measured by the difference in Euro and USlibor rates reliably signaled currency moves in the late1990s and early 2000s. The euro depreciated when it

yielded less than the dollar (1998 – 2000) and appreciatedwhen it yielded more (2002 – 2004). This relationship

between carry and EUR/USD has broken down since 2005-2008 when the euro rallied almost continuously despite arate disadvantage to the US. In this case, the moresignificant driver has been the change in rate spreads: theeuro rallied as the rate gap between the US and Europenarrowed, even though the dollar was the higher-yieldingcurrency (chart 2). In USD/JPY as well, the mere existenceof a US rate advantage over the yen has brought onlylimited spot appreciation (2000 – 02 and 2005 - 06). Inmany other periods, the change in rates has been moresignificant than the level of rates, as was the case in 2002 –

04 and in 2007 (chart 3).A simple strategy to exploit this phenomenon would buythe currency in whose favour interest rates are moving,regardless of whether it is a low or high-yielder. In essence,this strategy uses rate momentum as a signal for tradingcurrencies. Over the past decade this strategy has returns6% per annum on a volatility of 5% (table 1), so delivers

performance with much less drawdown than most other alpha models (chart 4).

Table 1 displays the historical performance of various alphastrategies and compares to those of currency managers andglobal macro hedge funds. Price momentum and forward carry (rate momentum) are the two we modify in thissection for the purpose of a dynamic hedging model. Pricemomentum is tested since it is the simplest to construct; andforward carry because it has offered the best combination of

performance and low volatility/drawdown. The originalmodels were constructed to issue signals ever two to four weeks, consistent with the investment horizon of manytactical trading accounts. Minor modifications to the

parameters generate signals for one to three monthhorizons, which is more consistent with the FX hedgerebalancing frequency of long-term investors and

Chart 2: EUR/USD versus Euro – US rate expectationsrate expectations based on 1-mo rates 3-mos forward

-300

-200

-100

0

100

200

300

99 01 03 05 07 09

0.80

0.90

1.00

1.10

1.20

1.30

1.40

1.50

1.601.70

EU -US libor differential, bpEUR/USD

Source: J.P. Morgan

Chart 3: USD/JPY versus US – Japan rate expectationsrate expectations based on 1-mo rates 3-mos forward

0

100

200

300

400

500

600

700

99 01 03 05 07 09

80

90

100

110

120

130

140US-JA libor differential, bp

USD/JPY

Source: J.P. Morgan

Chart 4: Forward carry performance since 2000Based on G-10 currencies vs USD. Trading rule used 1mo rates 3mosforward, 1-mo lookback for rates and daily rebalancing

100

120

140

160

180

200

00 01 02 03 04 05 06 07 08 09 10

Forward Carry (9

USD pairs)

Source: J.P. Morgan





24

corporates. As in Section III, all models are backtestedagainst four base currencies: USD, EUR, GBP and AUD.

Modifying alpha strategies for dynamichedging

The alpha signals have proven to be effective tools for positioning among currencies, and this section incorporatesthe strategies within dynamic hedging systems. Hedgingdecisions result in increased or decreased long exposure toone or more currencies, meaning that even default decisionsto ascribe to a fixed coverage level implicitly results inforeign currency exposure. If systematic signals generated

positive returns over time, then they should provide a usefulguide to timing hedging decisions by signalling when, andin what direction to adjust coverage ratios.

Before testing the efficacy of the signals, severaladjustments to the basic strategies are necessary so as tomake them relevant for medium-term hedgers. The originalstrategy backtests revealed that shorter lookback windowstend to generate the best results for forward carry signals.Since hedgers are unlikely to adjust their coverage at highfrequencies, it is first necessary to extend the trading rulesin such a way as to lengthen the holding period of the

trades. Both strategies, price momentum and forward carry,have two parameters in common which can be used to

calibrate the signals to hedging behaviour: (1) the lookback period for measuring changes in price or rates; and (2)rebalancing frequency (daily, monthly).

Various lookback windows impact returns and holding periods. A wide range of lookback windows are tested toidentify strategies with appropriate holding periods, and

positive returns. J.P. Morgan’s original models arerebalanced daily; but longer holding periods can also beimposed by restricting the rebalancing frequency.

Signals are evaluated relative to a 50% hedged benchmark

index. As discussed in Section II, no uniform hedge ratioapplies to all investors or corporates, but a semi-hedged

benchmark brings some advantages. Symmetric mandatesallow the manager to be over or underhedged relative to the

benchmark. Most select a 50% hedged benchmark for practical considerations, and anecdotal evidence confirmsthat this describes a fairly large cross-section of typicalhedging behavior. For this reason, the following backtestsuse a 50% benchmark hedging index.

Table 1: Long-term performance of FX alpha strategies, currency managers and global macro funds

2010 YTD

YTD return -11.0% -7.0% -1.3% -5.6% 7.7% 2.1% 1.1% 4.1% 1.9% -0.8%

Std dev 18.7% 10.4% 19.2% 7.3% 7.0% 4.4% NA 3.9% 3.1% 5.7%

IR -0.59 -0.67 -0.07 -0.77 1.11 0.48 NA 1.07 0.61 -0.13

2009

Avg annual return 57.9% 20.6% 10.2% -11.0% 7.9% 3.0% 0.9% -2.9% -1.2% 4.3%

Std dev 27.7% 13.6% 7.7% 12.0% 8.6% 5.6% 1.5% 3.8% 2.8% 6.3%

IR 2.09 1.51 1.34 -0.92 0.92 0.53 0.61 -0.76 -0.41 0.69

2005 - 2009 (5 years)

Avg annual return 4.8% -0.1% 7.3% -1.5% 9.4% 8.1% 1.1% 5.0% 3.2% 7.0%

Std dev 31.6% 16.0% 14.4% 8.7% 3.6% 6.1% 1.9% 12.4% 3.0% 2.7%

IR 0.15 -0.01 0.51 -0.18 2.61 1.33 0.58 0.40 1.09 2.56

2000 - 2009 (10 years)

Avg annual return 9.0% 4.2% 10.2% 3.7% 6.2% 5.7% 3.2% -7.0% -1.3% 7.6%

Std dev 22.1% 12.1% 16.2% 8.9% 5.2% 5.1% 3.5% 10.4% 19.2% 6.1%

IR 0.41 0.35 0.63 0.41 1.18 1.12 0.92 -0.67 -0.07 1.26

G-10 carry

(IncomeFX)

G-10 carry

(unlevered)

Alpha strategies Manager performance

Forward Carry

(22 Major pairs)

Barclay

Currency

Traders Index

Barclay Group

BTOP FX

Parker

Blacktree CMI

HFR global

macro hedge

fundsPrice momentum

Emerging Markets

carry (IncomeEM)

Forward Carry

(9 USD Pairs)

Source: J.P. Morgan





25

From the perspective of a USD-based investor, alphamodels are used to adjust hedge ratios as follows.

• Benchmark: 50% hedged, 50% unhedged

• Signal based strategy: Long USD signal: 75% hedged,

25% unhedged

• Short USD signal: 25% hedged, 75% unhedged

If the model issues a buy signal on the dollar, the hedgeratio is increased to 75% to capture the expecteddepreciation of the foreign currency. If a sell USD signal isissued, the hedge ratio is reduced to 25% to profit fromexpected foreign currency strength. These +/-25%

fluctuation bands are chosen for simplicity but areconsistent with anecdotally-acceptable hedge ratios.

Backtests are conducted using J.P. Morgan IndexResearch’s hedged and unhedged bond indices7 whichhedge monthly via forward contracts and rebalance at theclose of the first business day of the month. The percentageof the portfolio is chosen according to the benchmark specification (50%), or as indicated by a given signal.Results are generated across G-10 pairs, from the standpointof USD based investors, GBP based investors, and AUD

based investors hedging G-10 pairs back to their respectivehome currencies in each case. 8 The full set of resultsinclude backtests across a wide range of strategyspecifications, across G-10 currencies, and two sample

periods (past five and ten years). Given these permutations,complete results are confined to the Appendix.

The backtests focus on the two most simple strategy building blocks, simple momentum and forward carry.Forward carry is tested using lookback windows of 20, 40,

60, 100 and 120 business days. Price momentum is testedon lookback windows of 20, 40, 100, 200 and 252 business

days. Both strategies are also tested with and withoutrestrictions on a fixed rebalancing day at month-end.

All strategies are assessed for the excess return

(outperformance) versus the 50/50 benchmark, the volatility

of that outperformance (tracking error) and their information ratio (excess return divided by volatility of excess return). Performance tables also display averages

7 The hedging simulation relies on the JPMorgan hedged and

unhedged bond indices across G10 currencies with the exceptionof NOK and CHF, which are Bloomberg government bondindices. Please see JPMorgan Government Bond Indices, FrancisDiamond (2002).8 These base currencies were chosen to highlight performanceacross three economically diverse regions, with highly differentexchange rate behavior.

across all of the individual currency results for a givenstrategy. This provides an overview which helps to gauge

the general performance of a given strategy across all of thecurrencies.

Results: Price momentum

The simple momentum results are displayed within tables 2and 3. This strategy generates significant and consistently

positive excess returns across currencies, sample periodsand model specifications. For example, in the case of a 100or 40-day lookback windows, the strategy generated anaverage excess performance of around 2% over 10 yearsresulting in an average information ratio of around 0.7, andwas profitable across 100% of the 9 major G10 pairs (table2).

Not surprisingly, the average holding periods for thestrategies are dependent on the lookback window chosen.For example, 20 to 40-day lookback windows generateholding periods of around one month, 100-day lookbacksgenerate holding periods of about two months, and longer (200 to 252-day) lookback windows generate holding

periods of three to four months. In the aggregate, it appearsthat price momentum tends to favor shorter lookback windows over the past decade.

These cross currency averages mask fairly significant intra-currency details. Table 3 displays the results for only the40-day lookback window. As already indicated in table 2,the strategy is profitable across all G-10 pairs, over bothtime periods. The best performances over a ten-year sampleare AUD, NZD, CAD, and EUR.

The appendix tables replicate the analysis across threeadditional base currencies (GBP, EUR, and AUD) and thefull range of lookback windows. These results are broadlyconsistent with that of the USD based strategies as mostlookback windows are profitable across most G10 pairs.The EUR strategies are the worst performing, with averageannual performance of only 0.5% above that of the

benchmark strategy in most cases (appendix table 1). GBP based backtests are more successful, with average annualexcess returns of around 1% for 40 to 100-day lookback

windows using ten years of data (appendix table 3).Finally, the AUD based results have positive excess returns,with the best performances for shorter lookback windowswhere average annual returns are 1-3% above that of the

benchmark index, resulting in information ratios of around0.4 to 0.8 over the past decade (appendix table 5).





26

Table 2: Dynamic hedging with a price momentum signal: performance of G-10 currencies with USD as the baseExcess returns over the 50/50 benchmark. Table entries are averages of results across all G-10 USD pairs

2000 - 2009 (10 years) 252 Day 200 Day 100 Day 40 Day 20 Day

Avg annual return 1.3% 1.7% 2.0% 2.2% 2.7%

Std dev 3.0% 3.1% 3.2% 3.1% 2.7%

IR 0.44 0.57 0.65 0.73 1.01

%ccy with profitable strategies 100% 86% 100% 100% 100%


Avg annual return -0.1% 0.9% 2.2% 1.7% 2.1%

Std dev 2.7% 2.9% 3.0% 3.6% 3.0%

IR -0.12 0.29 0.87 0.51 0.84


# of trades* 52 63 97 172 256

average months per trade* 4 3 2 1 1*Using 10y or maximum available history

Source: J.P. Morgan

Table 3: Individual currency excess returns of pri ce momentum rule-based index versus benchmark index: 9 USD pairsExcess returns over the 50/50 benchmark.

USD/CAD USD/EUR USD/NOK USD/GBP USD/SEK USD/CHF USD/AUD USD/JPY USD/NZD

2000 - 2009 (10 years)

Avg annual return 2.4% 2.1% NA 1.2% 2.0% NA 3.3% 1.1% 3.7%

Std dev 2% 2% NA 2.5% 3.2% NA 5.0% 2.9% 3.8%

IR 1.18 0.83 NA 0.49 0.62 NA 0.66 0.36 0.96

2005 - 2009 (5 years)

Avg annual return 2.4% 1.2% 1.9% 1.0% 1.3% 0.2% 2.9% 2.0% 2.8%

Std dev 2% 3% 4% 3.5% 4.0% 2% 6.7% 3.2% 3.8%

IR 1.25 0.42 0.43 0.28 0.32 0.10 0.43 0.60 0.74

# of trades* 170 140 196 189 192 140 163 180 181

average months per trade* 1 1 1 1 1 1 1 1 1

*Using 10y or maximum available history

40 bus. day lookback window

Source: J.P. Morgan

Results: Forward carry

Forward carry results are displayed within Table 4 andTable 5. As before, the strategies generate decent excess

returns across currencies, sample periods and modelspecifications. Forward carry tends to favor 60 to 100-daylookback windows, delivering excess returns of 1-2% per annum. For example, in the case of 60-day lookback, thestrategy generated an average excess performance of 1.5%over ten years, 2.2% over five years, and was profitablyover all currencies over both periods (table 4). For medium-term lookbacks (40 to 100-day), information ratiosranged from 0.3 to 0.9. As before, the average holding

period is longer for longer lookback windows, with average

trades of one month for shorter windows, and as long asthree months for the longest lookback (120-day).

Table 5 displays the individual currency results for a 60-day

lookback. This strategy performs well across mostcurrencies, and has been best for AUD, NZD, CAD andGBP. Holding periods are in the one to two month rangedepending on the lookback. All of these results appear wellsuited for dynamic hedging where investors or corporatesrebalance on a one to three month horizon.

The appendix also presents results for the other basecurrencies of EUR, GBP and AUD. The results areconsistent with that of the USD-based strategies. The EUR results are good, with average annual excess returns of 0.6%to 0.8% (appendix table 2). GBP-based strategies result inexcess performance of around 1% and IRs of 0.3 to 0.4 over





27

shorter lookbacks (20 to 60-day), as shown in appendixtable 4. Finally, the AUD based results have positive annual

excess returns, with the best performance for shorter lookback windows where average annual returns are around1%, resulting in information ratios of around 0.3 to 0.5(appendix Table 6).

Cash flow analysis

Excess returns from these signaling models can also be

illustrated through cash flows. Charts 5 through 7 displaythe monthly returns of the signal based index, and thecumulative monthly returns of the signal based index andthe benchmark index from a USD perspective against CAD,from a GBP perspective against USD, and from an AUD

perspective against USD. In cases where the alpha strategy

signals generate excess returns over the benchmark, thestrategy’s cumulative cash flows will be greater than that of the benchmark strategy. In each of these cases, the signal

based strategy is able to outperform the benchmark indexconsistently.

Robustness tests: month-end restrictions

Tables 6 and 7 display the results of the simple momentumand forward carry strategies respectively, both withrebalancing restricted to the month-end. The month-endrestriction has a significant impact on performance in mostcases, but performance remains positive across G-10 pairs.Additionally, adding month-end restrictions greatly

increases the holding period of the trades across allstrategies. This is of course not surprising, since themonthly restriction is essentially a lower bound on thenumber of trades that can be taken.

In the case of price momentum, (table 6) it appears that profitability is preserved under the restricted strategy. Ashorter lookback window provides consistently positive

performance across the G10. The dramatic effect of themonth-end restriction on holding periods is also apparent:even in the case of shorter lookback windows, most holding

periods are in the range of three or four months.

The month-end restriction results in profitable averageannual returns in the case of forward carry as well (table 7).Medium (40-100d lookback windows) are able to generate1% of excess returns versus the benchmark on average.Interestingly, for certain lookback windows, the monthlyrebalancing restriction actually increases returns, althoughthis is likely an aberration.

Chart 5: Cash flows on CAD hedged into USD based on 60-dayforward carry signalCumulative and monthly cash flows

-6%

-4%

-2%

0%

2%

4%

6%

8%

10%

00 01 02 03 04 05 06 07 08 09 10

-30%

-20%

-10%

0%

10%

20%

30%Monthly Signal CF, lhs

Cumulative Benchmark CF, rhsCumulative Signal CF, rhs

Source: J.P. Morgan

Chart 6: Cash flows on USD hedged into GBP based on 60-dayforward carry signalCumulative and monthly cash flows

-10%

-5%

0%

5%

10%

00 01 02 03 04 05 06 07 08 09 10

-10%

0%

10%

20%


Cumulative Benchmark CF, rhs

Cumulative Signal CF, rhs

Source: J.P. Morgan

Chart 7: Cash flows on USD hedged into AUD based on 60-dayforward carry signalCumulative and monthly cash flows

-10%

-5%

0%

5%

10%

00 01 02 03 04 05 06 07 08 09 10

-30%

-20%

-10%

0%

10%

20%

30%


Cumulative Benchmark CF, rhs

Cumulative Signal CF, rhs

Source: J.P. Morgan





28

Table 4: Dynamic hedging with forward carry signal: performance of G-10 pairs with USD as the base currencyExcess returns over the 50/50 benchmark. Table entries are averages of results across all G-10 USD pairs



Std dev 2.9% 3.1% 3.3% 2.9% 2.6%

IR 0.29 0.32 0.50 0.60 0.55




Std dev 2.8% 3.1% 3.3% 2.6% 2.2%

IR 0.56 0.64 0.82 0.89 1.13


# of trades* 55 60 98 136 221


Source: J.P. Morgan

Table 5: Individual currency performance with forward carry signal with USD as the base currency

USD/CAD USD/EUR USD/NOK USD/GBP USD/SEK USD/CHF USD/AUD USD/JPY USD/NZD

2000 - 2009 (10 years)

Avg annual return 1.7% 0.6% NA 1.3% 0.7% NA 1.3% 0.6% 4.3%

Std dev 1.9% 1.8% NA 4.2% 3.4% NA 4.9% 3.5% 3.2%

IR 0.86 0.32 NA 0.32 0.19 NA 0.27 0.17 1.37

2005 - 2009 (5 years)

Avg annual return 1.7% 1.7% 2.4% 1.1% 1.6% 2.1% 2.2% 2.6% 4.3%

Std dev 2.7% 1.4% 1.7% 5.5% 3.6% 2.3% 5.4% 3.7% 3.0%

IR 0.62 1.25 1.37 0.19 0.44 0.94 0.41 0.69 1.44

# of trades* 88 130 82 80 93 101 144 74 90

average months per trade* 2 1 2 2 2 2 1 2 2

*Uses 10y or maximum available data period

60 bus. day lookback window

Source: J.P. Morgan





29

Table 6: Performance of price momentum signal with month-end restriction and USD as base currencyExcess returns over 50/50 benchmark



Std dev 3.3% 3.2% 3.0% 3.1% 2.8%

IR 0.35 0.40 0.51 0.36 0.38




Std dev 3.0% 2.9% 2.8% 3.4% 3.2%

IR -0.21 0.10 0.82 0.12 0.03


# of trades* 11 14 22 41 57

average months per trade* 19 14 9 4 3*Using 10y or maximum available history Source: J.P. Morgan

Table 7: Performance of forward carry signal with month-end restriction and USD as base currencyExcess returns over 50/50 benchmark



Std dev 3.1% 2.9% 3.1% 2.9% 2.8%

IR 0.26 0.31 0.27 0.45 0.26




Std dev 2.9% 3.1% 3.0% 2.6% 2.7%

IR 0.60 0.64 0.63 0.74 0.36


# of trades* 12 13 20 31 22

average months per trade* 15 13 9 5 9

*Using 10y or maximum available history Source: J.P. Morgan





30

Appendix charts: Excess returns for price momentum and forward carry signal with EUR, GBP and AUD as the base currency.

Table 1: EUR as base currency – performance of price momentum signal

Excess returns over the 50/50 benchmark. Table entries are averages of results across all G-10 EUR crosses.



Std dev 2.5% 2.5% 2.3% 2.5% 2.4%

IR 0.08 0.20 0.21 0.13 0.35




Std dev 2.7% 2.6% 2.3% 2.8% 2.6%

IR -0.06 0.03 0.26 0.04 -0.04


# of trades* 91 93 128 204 275


*Using 10y or maximum available history

Source: J.P. Morgan

Table 2: EUR as base currency – performance of forward carry signalExcess returns over the 50/50 benchmark. Table entries are averages of results across all G-10 EUR crosses.



Std dev 2.5% 2.6% 2.5% 2.4% 2.4%

IR 0.17 0.24 0.28 0.31 0.32




Std dev 2.5% 2.6% 2.3% 2.2% 2.4%

IR 0.42 0.50 0.46 0.28 0.52


# of trades* 77 87 131 172 236







31

Table 3: GBP as base currency – performance of price momentum signalExcess returns over the 50/50 benchmark. Table entries are averages of results across all G-10 GBP crosses.



Std dev 3.2% 3.1% 3.1% 3.3% 2.8%

IR 0.18 0.18 0.38 0.14 0.52




Std dev 3.6% 3.9% 3.4% 4.1% 2.9%

IR 0.33 0.23 0.52 0.06 0.47


# of trades* 68 94 130 204 272

average months per trade* 3 2 1 1 1*Using 10y or maximum available history Source: J.P. Morgan

Table 4: GBP as base currency – performance of forward carry signalExcess returns over the 50/50 benchmark. Table entries are averages of results across all G-10 GBP crosses.



Std dev 2.6% 3.1% 3.5% 3.2% 2.8%

IR 0.07 0.25 0.27 0.32 0.38




Std dev 3.1% 3.9% 4.1% 3.2% 2.7%

IR 0.29 0.34 0.31 0.28 0.64


# of trades* 79 94 122 156 244







32

Table 5: AUD as base currency – performance of price momentum signalExcess returns over the 50/50 benchmark. Table entries are averages of results across all G-10 AUD crosses.



Std dev 2.4% 2.9% 2.8% 3.5% 3.4%

IR 0.10 0.00 0.37 0.50 0.79




Std dev 2.4% 2.7% 3.1% 4.7% 4.0%

IR -0.11 0.13 0.65 0.43 0.84


# of trades* 70 75 103 174 249


Source: J.P. Morgan

Table 6: AUD as base currency – performance of forward carry signalExcess returns over the 50/50 benchmark. Table entries are averages of results across all G-10 AUD crosses.



Std dev 3.6% 3.6% 3.6% 3.3% 3.3%

IR 0.27 0.36 0.30 0.38 0.45


2005 - 2009 (5 years) 120 Day 100 Day 60 Day 40 Day 20 Day Avg annual return 1.7% 2.4% 1.8% 1.2% 1.4%

Std dev 3.8% 4.0% 4.1% 3.7% 3.7%

IR 0.26 0.67 0.43 0.30 0.60


# of trades** 60 65 113 140 226

average months per trade** 3 3 1 1 1






33

This page left intentionally blank




Arindam Sandilya (1-212) [email protected] Chase Bank NA

34

V. Hedging with forwards

versus options• Compared to extensive studies on forwards-based

hedging, work on options approaches is limited.

• This section compares the returns of a forward

versus option strategy in hedging a portfolio

employing a 50/50 hedge ratio.

• For AUD, EUR and GBP based investors with USD

exposure, hedging with options has outperformed

forwards since 2003. From a corporate perspective,

options outperformed for long USD/short AUD,

EUR and GBP exposure.• A signalling model which switches between

forwards and options does not provide substantial

advantage over an outright option based hedge. This

conclusion is sensitive to methodology, however.

FX hedging analysis is typically conducted based onforwards rather than options. For completeness, optionsshould also be considered. Although there exist manyempirical studies on forward contracts to hedge currencyrisk, the literature on derivatives-based hedging is relativelylight. The primary difference between the two instrumentsis options’ non-linear payoff which leads to complex,

distinctly asymmetric return distributions. Of course, thenon-linearity of payoffs comes at a cost ─ the option

premium ─ which can be a disincentive for some hedgersgiven the zero upfront cost of forwards. Option premia can,however, be higher or lower than the subsequent delivered

performance; it is not immediately clear a priori if thenecessity of an upfront cash payment in itself impairs theeffectiveness of these instruments. This section backtests aseries of option-based alternatives to forward hedges since2003. We also test the efficacy of a switching rule which

positions in either forwards or options depending upon theconviction of the signal in forecasting a spot trend.

Base case: replicating a 50/50 hedge ratioThis analysis assumes a base case forward hedge for

benchmarking the performance of option hedges. As before,this is taken to be 50% of the total currency risk hedgedusing 1-mo forwards: the benchmark hedge for 100 units of foreign currency exposure is to sell 50 units of it forward. Itis trickier to compute the equivalent notional of optionsrequired to perform an apples-to-apples study. Atheoretically appropriate way to proceed would back out anotional number that leaves the option-hedged foreigncurrency risk at 50% of the original exposure, exactly

mimicking the base case forward hedge. For ATM optionswith delta of approximately 50, this means buying 100 units

notional of foreign currency puts. The 50% hedge ratioholds true only at inception, however. Unlike forwards thatoffer a constant coverage ratio, option deltas fluctuate withmoves in spot, interest rates, volatility and the passage of time, causing the hedge ratio to vary between the upper andlower delta thresholds of 0% and 100%. Thus market movesalone can leave investors either under- or over-hedgedrelative to the forward hedge benchmark, an unavoidableconsequence of this approach.

An alternative uses the same notional of options as thenotional of the benchmark forward hedge ─ in this example,50 units of the foreign currency. But given that option deltasvary between 0 and 1, the directional exposure from an

ATM option sized according to this scheme can range between 0 to 50 foreign currency units, or a hedge ratiorange of 0% - 50% on the total FX risk of 100 units. Thisapproach renders a portfolio with a (foreign currency) putoption overlay systematically under-hedged vis-à-vis onewith a short (foreign currency) forward overlay, which maynot be desirable. For consistency with Section IV of this

paper we prefer the former scheme, but the second optionusing equivalent notionals will be worth exploring inforthcoming research.

The sample period for the study is 2003 to the present,reflecting the availability of options data. The universe of option hedges examined is limited to 1-mo vanillas, and to

only the two structures most commonly used for hedging purposes: straight ATM puts on the foreign currency, and10D, 25D and 40D risk reversals (long foreign currency

puts versus short foreign currency calls). The analysiscontains one important simplification relative to the

previous section. The notional size of the forward hedgesconsidered in Section IV is an estimate of the expectedvalue of the core asset holdings (e.g. JPMorgan bondindices). Because the purpose of this section is to comparethe efficacy of two hedging instruments, we hold thenotional of foreign currency exposure constant here, therebyavoiding the influence of moves in core holdings on theresults.

The study assumes that non-US investors hold $100 of USassets, whose value fluctuates solely due to variations in thelocal currency/USD exchange rate. Chart 1 depicts thehedged and unhedged returns from the perspective of AUD,EUR and GBP- based investors, and Table 1 details the

performance of the various hedging strategies. From acorporate perspective, the company is long USD and shortAUD, EUR and GBP. Note that the results in the table holdonly for US investments by foreign investors, unlikeSections II – IV of this paper which presented results acrossa broader range of base currencies and related crossrates.





35

The narrower sample here reflects the microstructure of FXoption markets – options on crosses are less liquid than

those in the USD majors. This means that while notionalsize is not an issue for AUD-based investors hedgingEuropean or UK assets through forwards, deployingEUR/AUD or GBP/AUD puts as hedges runs into notionalconstraints. Given this liquidity issue, options performanceis shown for USD-pairs only.

Results: forwards versus options

The charts and table highlight the following:

• Given the dollar’s general downtrend against most

currencies over the sample, period, hedging currency

exposure on US assets has tended to be profitable.

• ATM options sized to provide the same hedge ratio as

the benchmark forward hedge at inception generally

dominate forwards as hedge instruments, more in

AUD than in EUR or GBP. This pattern is largely

attributable to the direction of the underlying spot rates

over the backtest period. With dollar weakness prevailing

for most of the last seven years, USD put/local currency

call options spent the majority of their lives in the

money, and effective hedge ratios offered by options

were generally greater than the 50% provided by the

benchmark forward hedge. In other words, options

generally overhedged investors relative to the

benchmark, which turned out to be quite profitable in adeclining dollar environment. In option parlance, ATM

foreign currency puts possess favorable

dHedgeRatio/dSpot properties: the effective notional of

the hedge increases precisely when needed ─ during

periods of a declining foreign currency. This convexity

of course comes at a price in the form of paying away

premium for put options that expire worthless in periods

of strengthening foreign currency. EUR and GBP have

encountered steeper and more extended spot declines

against USD than AUD over the past six months, and

ATM USD puts/currency calls therefore have given back

a large part of their outperformance relative to forwards

over this period.

• Risk reversals perform better in EUR and GBP than

in AUD. Intuitively, risk reversals should track the

forward hedge, given their similarity to a forward in

being exposed to both appreciation and depreciation of in

the underlying exchange rate. (Thus symmetry becomes

clearer if one thinks of a forward as the combination of a

call and a put, both struck at the forward; a risk reversal

can be created by simply pulling the call and strikes away

from the forward strike in opposite directions). The more

out-of-the-money the two strikes, the more obvious the

delinkling from the performance of forwards. 10D risk

reversals exhibit interesting behavior in that selling wingUSD calls for a leveraged notional was a painful

experience during the credit crisis as vols and skews

blew out, thus detracting from hedge performance

relative to forwards. This phenomenon was especially

pronounced in AUD, where the penalty was substantially

higher than in most currencies.

Table 1. Option-based hedges (ATM puts or risk reversals) haveoutperformed forwards and a rule-based switch ing st rategy betweenforwards and options.

Assumes USD exposure of $100 for all local currencies. Benchmark hedge is a50% short foreign currency forward. Option hedges are sized to provide the samehedge ratio as the benchmark forward hedge at inception. No transaction costs

Local

CurrencyStrategy

Average

ReturnVo lati li ty IR

Unhedged -6.5% 15.2% -0.43

Hedged v ia Forw ard -1.8% 6.3% -0.28

Hedged v ia Option 0.1% 7.8% 0.01

Hedged v ia 40dRiskRev ersal -1.5% 5.9% -0.26


Hedged via 10dRiskRev ersal -6.4% 13.4% -0.48

Hedged v ia Signal (option and fw d) -0. 6% 8.2% -0.08

Hedged v ia Signal (25d rr and fw d) -1.8% 5.9% -0.30

Unhedged -3.1% 10.1% -0.31

Hedged v ia Forw ard -1.6% 4.6% -0.35

Hedged v ia Option -1.5% 5.5% -0.28




Hedged v ia Signal (option and fw d) -1. 5% 5.6% -0.28

Hedged v ia Signal (25d rr and fw d) -1.6% 4.8% -0.33

Unhedged 0.7% 10.5% 0.07

Hedged v ia Forw ard 0.9% 5.0% 0.18

Hedged v ia Option 1.0% 6.2% 0.15

Hedged v ia 40dRiskRev ersal 1.0% 5.1% 0.19



Hedged v ia Signal (option and fw d) 0.8% 5.9% 0.13

Hedged v ia Signal (25d rr and fw d) 1.1% 5.5% 0.19

AUD

EUR

GBP

Source: J.P. Morgan

Results: a signalling model for switching

between forwards and options

Table 1 also lists the performance of a signal-based

hedging strategy that switches between a forward andan equivalent delta option or a risk reversal dependingon the conviction from the alpha models described in





36

Section IV. A buy or sell signal from the models representsa high-conviction view on spot’s direction, so would justify

hedging with forwards. A neutral reading from the modelsrepresents a low-conviction view, so would justify hedgingwith options instead.

The strategies’ information ratios – defined asoutperformance versus benchmark divided by the volatilityof outperformance – indicate that while the switching ruleapplied either to an ATM put or a risk-reversal outperformsforward hedges in most cases, the magnitude is small. Anoutright option-based hedging strategy has delivered better

performance for all currencies, although this conclusioncould be sensitive to sample period and baseline hedge ratiochosen (50/50 benchmark versus a neutral of either zero or 100% hedging).

For instance, AUD-based investors actually increase thevolatility of their hedged portfolio by switching betweenforwards and options (8.2%) as opposed to hedging purelywith options (7.8%), even as the returns from doing so arelower (-0.6% with forwards plus options versus 0.1% withoptions alone). This pattern probably owes to the filter rule.While correct more than 50% of the time in identifying theappropriate spot environments for switching into options,the rule does not incorporate rich/cheap considerations inoption prices as an input, which is a critical considerationfor most investors.

The delta-adjusted methodology probably also shapes theoutcome. By allowing the option to have a notional twice aslarge as the forward contract, a simple directional signalwhich does not take into account the expected extent of spot’s upward or downward movement becomes difficult toapply. A more neutral framework to assess the signal’svalue would maintain the same notional of the two hedginginstruments. This issue can be addressed in subsequentresearch.





37

Chart 1. Hedging USD exposure has been prudent AUD, EUR, and GBP- based investors since 2003, and ATM options have performed at paror better than forwards.

Assumes core USD exposure of $100 for all local currencies. Benchmark hedge is a 50% short foreign currency forward. Option hedges are sized to provide the same

hedge ratio as the benchmark forward hedge at inception. No transaction costs

AUD/USD

100

150

200

250

Jan-03 Jun-04 Dec-05 Jun-07 Nov -08 May -10

UnhedgedFw d Hedged

ATM Option Hedged

100

120

140

160

180

200


40D RR Hedged25D RR Hedged

10D RR Hedged

Source: J.P. Morgan Source: J.P. Morgan

EUR/USD

60

70

80

90

100


UnhedgedFwd Hedged ATM Option Hedged

65

70

75

80

85

90

95

100


40D RR Hedged25D RR Hedged10D RR Hedged


GBP/USD

45

50

55

60

65

70

75

80


UnhedgedFw d Hedged ATM Option Hedged

45

50

55

60

65

70

75


40D RR Hedged25D RR Hedged10D RR Hedged





John Normand (44-20) 7325-5222

[email protected]. Morgan Securities Ltd

38

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 theresearch analyst(s) in this report.

Disclosures: J.P. Morgan is the global brand name for J.P. Morgan Securities Inc. (JPMSI) and its non-US affiliates worldwide. J.P.Morgan Cazenove is a brand name for equity research produced by J.P. Morgan Securities Ltd.; J.P. Morgan Equities Limited; JPMorganChase Bank, N.A., Dubai Branch; and J.P. Morgan Bank International LLC. Legal Entities Disclosures: U.S.: JPMSI is a member of

NYSE, FINRA and SIPC. J.P. Morgan Futures Inc. is a member of the NFA. JPMorgan Chase Bank, N.A. is a member of FDIC and isauthorized and regulated in the UK by the Financial Services Authority. U.K.: J.P. Morgan Securities Ltd. (JPMSL) is a member of theLondon Stock Exchange and is authorised and regulated by the Financial Services Authority. Registered in England & Wales No.2711006. Registered Office 125 London Wall, London EC2Y 5AJ. South Africa: J.P. Morgan Equities Limited is a member of theJohannesburg Securities Exchange and is regulated by the FSB. Hong Kong: J.P. Morgan Securities (Asia Pacific) Limited (CE number AAJ321) is regulated by the Hong Kong Monetary Authority and the Securities and Futures Commission in Hong Kong. Korea: J.P.Morgan Securities (Far East) Ltd, Seoul Branch, is regulated by the Korea Financial Supervisory Service. Australia: J.P. Morgan

Australia Limited (ABN 52 002 888 011/AFS Licence No: 238188) is regulated by ASIC and J.P. Morgan Securities Australia Limited(ABN 61 003 245 234/AFS Licence No: 238066) is a Market Participant with the ASX and regulated by ASIC. Taiwan: J.P. MorganSecurities (Taiwan) Limited is a participant of the Taiwan Stock Exchange (company-type) and regulated by the Taiwan Securities andFutures Bureau. India: J.P. Morgan India Private Limited is a member of the National Stock Exchange of India Limited and BombayStock Exchange Limited and is regulated by the 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 ExchangeCommission. Indonesia: PT J.P. Morgan Securities Indonesia is a member of the Indonesia Stock Exchange and is regulated by theBAPEPAM LK. Philippines: J.P. Morgan Securities Philippines Inc. is a member of the Philippine Stock Exchange and is regulated bythe Securities and Exchange Commission. Brazil: Banco J.P. Morgan S.A. is regulated by the Comissao de Valores Mobiliarios (CVM)and by the Central Bank of Brazil. Japan: This material is distributed in Japan by JPMorgan Securities Japan Co., Ltd., which isregulated by the Japan Financial Services Agency (FSA). Singapore: This material is issued and distributed in Singapore by J.P. MorganSecurities Singapore Private Limited (JPMSS) [MICA (P) 020/01/2010 and Co. Reg. No.: 199405335R] which is a member of theSingapore Exchange Securities Trading Limited and is regulated by the Monetary Authority of Singapore (MAS) and/or JPMorgan ChaseBank, N.A., Singapore branch (JPMCB Singapore) which is regulated by the MAS. Malaysia: This material is issued and distributed inMalaysia by JPMorgan Securities (Malaysia) Sdn Bhd (18146-X) which is a Participating Organization of Bursa Malaysia Berhad and aholder 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 authorised by the Capital Market Authority of the Kingdom of Saudi Arabia (CMA) to carryout dealing as an agent, arranging, advising and custody, with respect to securities business under licence number 35-07079 and itsregistered 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., Dubai Branch is regulated by the Dubai Financial Services Authority (DFSA) and its registeredaddress is Dubai International Financial Centre - Building 3, Level 7, PO Box 506551, Dubai, UAE. Country and Region Specific

Disclosures: U.K. and European Economic Area (EEA): Unless specified to the contrary, issued and approved for distribution in theU.K. and the EEA by JPMSL. Investment research issued by JPMSL has been prepared in accordance with JPMSL's policies for managing conflicts of interest arising as a result of publication and distribution of investment research. Many European regulators requirethat a firm to establish, implement and maintain such a policy. This report has been issued in the U.K. only to persons of a kind describedin Article 19 (5), 38, 47 and 49 of the Financial Services and Markets Act 2000 (Financial Promotion) Order 2005 (all such persons beingreferred to as "relevant persons"). This document must not be acted on or relied on by persons who are not relevant persons. Anyinvestment or investment activity to which this document relates is only available to relevant persons and will be engaged in only with

relevant persons. In other EEA countries, the report has been issued to persons regarded as professional investors (or equivalent) in their home jurisdiction. Australia: This material is issued and distributed by JPMSAL in Australia to “wholesale clients” only. JPMSAL doesnot issue or distribute this material to “retail clients.” The recipient of this material must not distribute it to any third party or outsideAustralia without the prior written consent of JPMSAL. For the purposes of this paragraph the terms “wholesale client” and “retailclient” have the meanings given to them in section 761G of the Corporations Act 2001. Germany: This material is distributed inGermany by J.P. Morgan Securities Ltd., Frankfurt Branch and J.P. Morgan Chase Bank, N.A., Frankfurt Branch which are regulated bythe Bundesanstalt für Finanzdienstleistungsaufsicht. Hong Kong: The 1% ownership disclosure as of the previous month end satisfiesthe requirements under Paragraph 16.5(a) of the Hong Kong Code of Conduct for Persons Licensed by or Registered with the Securitiesand Futures Commission. (For research published within the first ten days of the month, the disclosure may be based on the month enddata from two months’ prior.) J.P. Morgan Broking (Hong Kong) Limited is the liquidity provider for derivative warrants issued by J.P.Morgan Structured Products B.V. and listed on the Stock Exchange of Hong Kong Limited. An updated list can be found on HKExwebsite: http://www.hkex.com.hk/prod/dw/Lp.htm. Japan: There is a risk that a loss may occur due to a change in the price of the sharesin the case of share trading, and that a loss may occur due to the exchange rate in the case of foreign share trading. In the case of share




John Normand (44-20) 7325-5222


39

trading, JPMorgan Securities Japan Co., Ltd., will be receiving a brokerage fee and consumption tax (shouhizei) calculated bymultiplying the executed price by the commission rate which was individually agreed between JPMorgan Securities Japan Co., Ltd., andthe customer in advance. Financial Instruments Firms: JPMorgan Securities Japan Co., Ltd., Kanto Local Finance Bureau (kinsho) No. 82

Participating Association / Japan Securities Dealers Association, The Financial Futures Association of Japan. Korea: This report mayhave been edited or contributed to from time to time by affiliates of J.P. Morgan Securities (Far East) Ltd, SeoulBranch. Singapore: JPMSS and/or its affiliates may have a holding in any of the securities discussed in this report; for securities wherethe holding is 1% or greater, the specific holding is disclosed in the Important Disclosures section above. India: For private circulationonly, not for sale. Pakistan: For private circulation only, not for sale. New Zealand: This material is issued and distributed byJPMSAL in New Zealand only to persons whose principal business is the investment of money or who, in the course of and for the

purposes of their business, habitually invest money. JPMSAL does not issue or distribute this material to members of "the public" asdetermined in accordance with section 3 of the Securities Act 1978. The recipient of this material must not distribute it to any third partyor outside New Zealand without the prior written consent of JPMSAL. Canada: The information contained herein is not, and under nocircumstances is to be construed as, a prospectus, an advertisement, a public offering, an offer to sell securities described herein, or solicitation of an offer to buy securities described herein, in Canada or any province or territory thereof. Any offer or sale of the securitiesdescribed herein in Canada will be made only under an exemption from the requirements to file a prospectus with the relevant Canadiansecurities regulators and only by a dealer properly registered under applicable securities laws or, alternatively, pursuant to an exemptionfrom the dealer registration requirement in the relevant province or territory of Canada in which such offer or sale is made. The

information contained herein is under no circumstances to be construed as investment advice in any province or territory of Canada and isnot tailored to the needs of the recipient. To the extent that the information contained herein references securities of an issuer incorporated,formed or created under the laws of Canada or a province or territory of Canada, any trades in such securities must be conducted througha dealer registered in Canada. No securities commission or similar regulatory authority in Canada has reviewed or in any way passed

judgment upon these materials, the information contained herein or the merits of the securities described herein, and any representation tothe contrary is an offence. Dubai: This report has been issued to persons regarded as professional clients as defined under the DFSArules.General: Additional information is available upon request. Information has been obtained from sources believed to be reliable butJPMorgan Chase & Co. or its affiliates and/or subsidiaries (collectively J.P. Morgan) do not warrant its completeness or accuracy exceptwith respect to any disclosures relative to JPMSI and/or its affiliates and the analyst’s involvement with the issuer that is the subject of theresearch. All pricing is as of the close of market for the securities discussed, unless otherwise stated. Opinions and estimates constituteour judgment as of the date of this material and are subject to change without notice. Past performance is not indicative of future results.This material is not intended as an offer or solicitation for the purchase or sale of any financial instrument. The opinions andrecommendations herein do not take into account individual client circumstances, objectives, or needs and are not intended asrecommendations of particular securities, financial instruments or strategies to particular clients. The recipient of this report must make itsown independent decisions regarding any securities or financial instruments mentioned herein. JPMSI distributes in the U.S. research

published by non-U.S. affiliates and accepts responsibility for its contents. Periodic updates may be provided on companies/industries based on company specific developments or announcements, market conditions or any other publicly available information. Clientsshould contact analysts and execute transactions through a J.P. Morgan subsidiary or affiliate in their home jurisdiction unless governinglaw permits otherwise.

“Other Disclosures” last revised March 1, 2010.

Copyright 2010 JPMorgan Chase & Co. All rights reserved. This report or any portion hereof may not be reprinted, sold or

redistributed without the written consent of J.P. Morgan.




John Normand (44-20) 7325-5222


40

J.P. Morgan Global FX Strategy

London

John Normand MD Head, Global FX Strategy (44-20) 7325-5222 [email protected]

Paul Meggyesi MD G-10 FX Strategy (44-20) 7859-6714 [email protected]

Thomas Anthonj ED Technical Strategy (44-20) 7742-7850 [email protected]

Kamal Sharma VP G-10 FX Strategy (44-20) 7777-1729 [email protected]

Talis Bauer VP FX Derivatives Strategy (44-20) 7777-5276 [email protected]

New York

Ken Landon MD G-10 FX Strategy (1-212) 834-2391 [email protected]

Gabriel de Kock ED G-10 FX Strategy (1-212) 834-4254 [email protected]

Niall O’Connor ED Technical Strategy (1-212) 834-5108 [email protected]

Arindam Sandilya ED FX Derivatives Strategy (1-212) 834-2304 arindam.x.sandilya@jpmorg

Matthew Franklin-Lyons Analyst G-10 FX Strategy (1-212) 834-4565 [email protected]

Asia

Tohru Sasaki MD G-10 FX Strategy (81-3) 6736-7717 [email protected]

Junya Tanase ED G-10 FX Strategy (81-3) 6736-7718 [email protected]

Yen Ping Ho VP (Emerging Markets Research) Asia FX Strategy (65) 6882-2216 [email protected]

Yoonyi Kim Analyst G-10 FX Strategy (81-3) 67367729 [email protected]

mailto:[email protected]

mailto:[email protected]

jp morgan - managing fx hedge ratios - a framework for strategic and tactical decisions

Documents