passive spotlight · source: db research. all figures as of 31st july 2016. table 2: ecb holdings...

For Professional Investors Only – Marketing Material EUR Fixed Income: Navigating an upside down asset class… l | October 2016 1

Passive Spotlight October 2016

MARKETING MATERIAL – FOR

PROFESSIONAL INVESTORS ONLY

(as defined in MiFID Directive 2004/39/EC

Annex II)

EUR Fixed Income: Navigating an upside down asset class using the risk barbell

Deutsche Bank AG Deutsche Asset Management Mainzer Landstrasse 11-17 Frankfurt am Main, Germany Hotline: +49 (69) 910 30549 [email protected] etf.deutscheam.com Editor Vincent Denoiseux Passive Asset Management [email protected] Authors Olivier Souliac Passive Asset Management [email protected] Shuchang Sun Passive Asset Management [email protected] Content

1. Introduction 1 2. EUR Fixed Income: an upside

down asset class 2 3. Maximising Yield/risk benefits

using the Risk Barbell strategy 11 4. Applications of the barbell for

institutional investors: what makes sense, what does not 17

5. Conclusion 27 6. Annex - Measuring risks in the

fixed income assets 28 References 31

“Копейка рубль бережет”

“Penny and penny laid up will be many”

Russian Proverb

1. Introduction The EUR fixed income asset class has been turned upside down by

the action of central banks, and partly by investors themselves:

yields have turned negative and sovereign risks increased to such

extent that broad corporate bond indices started showing higher

levels of resilience than broad sovereign bond indices!

Also, the adequacy between fundamental risks and market implied

risks has been largely destroyed by large imbalances between

supply and demand. Not only central banks started buying bonds,

but also overall uncertainty combined with sinking yields forced

investors to rush into the asset class as a whole, from AAA

sovereigns to B corporates (cf. Table 1: ETF Fixed Income Flows by

Asset-Classes YTD and Table 2: ECB holdings).

As a consequence of this, not only have the yield spreads

between corporates and sovereigns decreased since the end of

the financial crisis but also the correlation between those two

asset classes increased substantially. This depicts a bleak

outlook for “risk barbell” investors, interested in the diversification

benefits of concentrating their portfolio towards (i) conservative

bonds at acceptable yield costs and (ii) higher yielding bonds at

reasonable costs in terms of yield.

This paper revisits the benefits of traditional barbell and risk barbell

strategies in this context, while pointing out, when relevant, the

added value of using Strategic Beta indices for implementation

purposes. It concludes with several use cases taken from practical

situations faced by fixed income investors: asset managers,

corporate treasurers, investment consultants, pension funds.

The outcome of the paper goes further than what we have seen from

traditional literature on the subject: While traditional barbell

strategies indeed became less and less attractive as interest rates

flattened, a “risk barbelling” approach can still be implemented

http://etf.deutscheam.com/

mailto:[email protected]




For Professional Investors Only – Marketing Material EUR Fixed Income: Navigating an upside down asset class… | October 2016 2

EUR Corporate bonds now

seem to be a better diversifier

versus Equity risk than

sovereign bonds!

as a way to moderately enhance yield and risk-adjusted returns.

However, the traditional use of asset class split (sovereigns vs

corporates) has its weaknesses. Investors therefore need to

reconsider what is risky and what is virtually “risk-free”, both

fundamentally and quantitatively, to ensure proper diversification

benefits between the two components of a barbell.

So as to confirm general common sense, focusing specifically on

fundamental risks for sovereigns helps to avoid the “credit trap” that

characterise peripheral economies and obtain a cleaner “interest

rate” type of signal. Also, looking for yield is a good proxy to increase

risks while maximising returns per unit of risk. Combining these two

approaches presents interesting barbell-friendly features for those

investors willing to carefully increase risks.

2. EUR Fixed Income: an upside down asset class

2.1 Clear Signs of dislocation Triggered by the European debt crisis starting from 2010, the ECB

implemented several unconventional asset purchase and liquidity

programmes due to the emergence of sovereign credit risk in

European peripheral countries. These measures have turned the

market upside down, first impacting periphery government bonds,

later the broader EUR fixed income market. We will however show

that both diversification and investability remain available features in

certain targeted segments of the asset class.

For most European sovereign bonds, diversification benefits versus

equities have been declining since the ECB implemented several

unconventional asset purchase programmes. This can be illustrated

by the moving 5Y correlations among European countries sovereign

bond indices vs. Euro Stoxx 50, which has strongly increased for

almost the whole sovereign bond asset class. While first only

periphery government bonds were affected, risk contagion and

subsequent ECB announcements have also affected near-core

countries such as France and Belgium (Figure 1)1.

1 Note that correlations were calculated on the basis of weekly returns and the picture is the same when using monthly returns



Figure 1: ECB’s unconventional monetary policies have turned the market upside down

Source: Markit Ltd., Bloomberg Finance LP. Weekly data from 01/03/2005 to 31/07/2016.

This development has resulted in a higher correlation for EUR

sovereigns with Euro Stoxx 50 than for EUR corporates now,

bringing us to the qualification of “upside down” (Figure 2, first

chart)2. As a confirmation of this trend, corporate yields became

somewhat more correlated with interest rates than European

sovereign yields (Figure 2, second chart)3.

Figure 2: Historical relationships between sovereigns and corporates have inversed

2 Note that correlations were calculated on the basis of weekly returns and the picture is the same when using monthly returns 3 The relatively high correlation between corporate bond yields and swap rates can seem counter-intuitive, however possible interpretations are the fact that most corporate bonds are priced off a swap curve plus a variable spread, mechanically increasing correlation with swap rates (Elton et al. (2003))



This situation may have been

caused by heavy inflows into

the asset class, both from

investors and from the ECB

German sovereigns have

been the most heavily

purchased bonds by the ECB

Source: Markit iBoxx Ltd., Bloomberg Finance LP. Weekly data from 01/03/2005 to 31/07/2016.

2.2 Supply/demand imbalances: Is there a Fixed

Income Bubble? The unprecedented low interest rate policies as well as asset

financing and purchasing programmes have led to unprecedented

price rallies in the fixed income space, therefore it is at least

questionable whether such prices follow sustainable valuations.

From a supply/demand perspective we see oversubscriptions in

many areas of the asset class. On the one hand, we have double-

digit billion ECB purchases, while on the other hand we see ongoing

strong demand for fixed income. The ETF market for European fixed

income products for example saw sizeable inflows (€21bn, +18.7%)

in the course of 2016 year to date, with main flows stemming from

corporate (€12.7bn) and sovereign bond ETFs (€6.9bn) (Table 1).

Taking a look at the supply side, it is worth mentioning that for

sovereign bonds the ECB is currently purchasing more than what is

issued on a net basis4. Italy for example is expected to have net

issuances of around €45.3bn this year compared to ECB purchases

of around €133.3bn just for Italian government bonds. In Germany

the situation is even more significant with net issuances of -€8.2bn

and expected ECB purchases of €191.1bn in 20165. However, for

corporate bonds we see a slightly different picture. In the corporate

space we see a strong supply pickup, driven by the CSPP, with

expected net issuances of €100bn this year6. This will cover the ECB

purchases of about €7.25bn7 per month but it is at least questionable

4 Net issuance is calculated as gross issuance minus the amount of redemptions. 5 See DB Research (2016): Eurozone Sovereign Issuance: 05th Sep - 09th Sep 6 See DB Research (2016): Credit Bites – EUR and GBP Bond Issuance 7 See DB Research (2016): Special Report – Credit Outlook: Balancing Late Cycle with Overwhelming Technicals



The opportunity costs of

holding AAA sovereign assets

could still seem relatively

high versus sovereign

benchmarks: almost 60 bps

as at 31/07/2016

whether this can cover the current strong demand for corporate

bonds.

Overall there has been strong yield tightening as well as liquidity

deterioration in Europe8, but as we will show in the next section, there

are however interesting opportunities in the AAA sovereign,

Financials and Subordinated space, and not only in terms of

diversification.

Table 1: ETF Fixed Income Flows by Asset-Classes YTD

€ million Net Flows YTD % AuM

Corporates +12,691 +29.3%

Sovereigns +6,937 +12.6%

Overall +1,293 +16.2%

Other +107 1.7%

Total FI +21,029 +18.7% Source: DB Research. All figures as of 31st July 2016.

Table 2: ECB holdings

€ million Sovereigns Corporates IG

ECB Holdings € 1,053 € 13

Approx. Public Market Size € 5,729 € 836

% of Market 18.39% 1.58% Source: European Central Bank (website). All figures as of 31st July 2016.

2.3 The asset allocation perspective: Looking for yield and for diversification

In the current low-yield set-up driven by massive ECB liquidity and

supply/demand imbalances, not all asset classes are rewarding

investors equally for the risks taken. However, as described on Table

3, certain market segments seem to have kept their original virtues,

namely (i) attractive yields (or at least credit premia) for lower quality

or seniority corporate bonds (of course still at the price of higher

risks) and (ii) perceived safety for AAA sovereigns (still at the price

of lower yields, although such price may not necessarily seem

excessive from an historical standpoint).

8 See Passive Insights (2016): ‘ETFs: Catalyst or Mirage for Market Liquidity?’



But AAA sovereigns are the

only asset class exhibiting

satisfying capital preservation

features for buy and hold

investors.

Table 3: Breaking down Fixed Income into its European sub-components9

Source: Markit Ltd., Bloomberg Finance LP. All figures as of 31st July 2016. (*)Rating is

calculated as average of S&P and Moody’s. Spread volatility is calculated as the standard

deviation of: weekly spread differences multiplied with the corresponding modified duration.

Although AAA sovereign bonds seem unattractive from a yield

seeking perspective, within a diversified portfolio context, investors

may want to keep considering the lowest (and negative) yielding

sovereign bonds. This asset class has consistently helped generate

diversification benefits (Table 4 and Figure 3) versus the widest

range of riskier assets. It is also important to consider that AAA

sovereigns yields may be negative because interest rates are

negative (as an effect of central bank monetary policy), not

necessarily because the asset class would have been particularly

oversubscribed. Although this is arguably the case that especially

German sovereigns have been heavily subscribed as per Table 1

and footnote 5, AAA sovereign yields seem to be a broadly fair

reflection of their fundamental quality and quantitative stability, as

shown in Figure 4.

The point of keeping a very conservative anchor in portfolio

construction may become all the more relevant for yield seeking

investors, who will per definition need to increase risks within their

portfolio.

Table 4: 5Y Correlations showing diversification benefits within the Fixed Income space

Source: Markit Ltd. All figures as of 31st July 2016. Correlation of spreads versus riskless

rates, except for AAA sovereigns where prices were taken into account. As AAA sovereigns

are considered „riskless“ for spread calculation purposes, spreads of AAA sovereigns are

insignificant and would only entail noise. Prior to 01/08/2016 iBoxx Italy was taken for 5Y

calculations as BBB Eurozone has no continuous operating history before 01/03/2011.

9 We excluded covered bonds as they hardly offer any yield pickup to EUR sovereigns and liquidity deteriorated sharply due to three covered bond purchase programmes

undertaken by the ECB.



Diversification benefits from

holding corporate bonds

alongside sovereigns started

diminishing since the end of

the 2011 sovereigns crisis

Diversification benefits are

however partially recovered

by investors concentrating

their portfolio into the

extremes of the EUR IG asset

class: e.g. AAA sovereigns

and Subordinated/BBB

corporates

As a supplement to the correlation table above, Figure 3 illustrates

the moving 5Y correlation versus equities which confirms the point

that the FI market was turned upside down, with the noticeable

exception of the AAA Eurozone sovereign asset class. Two

noticeable examples:

- price correlation between AAA sovereigns on one side

and both (i) BBB sovereigns (not so intuitive result 5

years ago!) and (ii) BBB corporates on the other is clearly

negative,

- spread correlation (i.e. adjusted for interest rate

variations) between corporates and sovereigns is largely

positive with 57%, indicating that European sovereigns

risk does have only limited diversification features versus

Corporate credit risk

Figure 3: Moving 5Y correlations indicating strong stock/bond co-movements

Source: Markit Ltd., Bloomberg Finance LP. Weekly data as of 01/03/2005 – 31/07/2016.

On the “higher yielding” side of the spectrum, BBB sovereigns may

now look attractive, the question remaining whether such yields are

a fair reward for the fundamental risks taken. When looking at a

qualitative analysis of Italy and Spain, their fundamentals 10 are

slightly worse than during the European debt crisis, while their bonds

are trading at substantially lower yields (cf. Figure 4) as a result of

ECB intervention. The reminiscent fundamental weakness of

southern European countries can also be perceived quantitatively

when measuring markets nervousness around the pricing of

10 See Passive Insights (2014): ‘Fundamental Scoring for Fixed Income’



BBB sovereign bonds may

remain attractive for long-

term or liquidity affine

investors, however offer a

poor “yield vs fundamental

risks” profile for income

investors

periphery countries, as measured by spread volatility. Such spread

volatility is substantially higher than that of safer countries such as

France or Germany indicating that the market is aware of its

fundamental risks.

Figure 4: Qualitatively speaking Italy’s and Spain’s situation is not

better than during the European debt crisis while the quantitative

picture below shows that the market recognizes this risk.

Source: Deutsche Bank, Markit Ltd., Bloomberg Finance LP. As of 31st July 2016. Spread

volatility is calculated as the volatility of: yield minus 10Y EONIA swap as risk-free rate

(100% correlation with 10Y bund yields).



When comparing the spread volatilities of Italy and of BBB

corporates, the results are similar. Both segments have the same

rating, similar yields (whereas corporate bonds exhibit higher yields

than Italian sovereigns for a given maturity) whereas Italy shows

substantially higher spread volatilities (Figure 5). Note that, however,

southern European sovereigns have both the advantage of

exhibiting high liquidity versus corporate bonds and providing longer

duration exposure, which can be of interest for institutional investors,

especially those looking to increase duration and not worrying about

short term volatility.

Figure 5: Italy has substantially higher spread volatility than BBB

corporates.

Source: Markit Ltd. As of 31st July 2016.

Generating additional income in bonds while only taking marginally

higher risks can generally be obtained via the credit asset class,

although its issuance volume is lower than sovereigns and its

transaction costs are expected to be higher. The chart below (Figure

6) confirms this assumption, showing that corporate credit premia

have not reached pre-crisis levels (even after taking into account the

slight fundamental degradation of the asset class 10 years ago).

Arguably, since corporate bond prices already started incorporating

CSPP effects 11 and ECBs willingness to provide markets with

liquidity, corporate bonds are still trading at comparably attractive

levels of credit risk premia (Figure 6). Therefore, the corporate bond

asset class presents the doubled benefits of diversification with still

11 See DB Research (2016): Credit Outlook - Balancing Late Cycle with Overwhelming Technicals



EUR corporate bonds credit

premia have remained broadly

stable (although recently

slightly declining) in the past

few years

More specifically, at the

border of the IG Corporate

Bond asset class, certain

subordinated bonds retain

attractive risk/yield features

are they are not eligible for

CSPP and investors tend to

avoid more complex assets

reasonably attractive valuations, despite recent corporate rally in

2016.

Figure 6: Corporate benchmark spread evolution from 2004 to 2016

and average ratings before the global financial crisis and after the

European debt crisis.

Source: Markit Ltd., Factset. Average Ratings as of 31/01/2005 and 31/01/2015

Yield seeking investors may also consider financials (as banks are

excluded from CSPP) as well as subordinated bonds (mostly issued

by banks and insurances) to capture additional risk premia. As an

example, lower tier 2 subordinated bonds issued by banks share the

same default probability as their related senior bonds while

additional risks can be expected to stem from additional complexity

of the bond covenants (driving lower demand and lower liquidity) and

lower recovery rates in case of default. The exclusion of such bonds

in the CSPP made them less sensitive to ECB’s QE, but the

possibility of spill-overs cannot be ruled out. The first effects on

eligible CSPP debt and bank bonds can already be observed when

looking at benchmark spreads12: while benchmark spreads of bank

senior unsecured bonds were trading at lower levels compared to

eligible bonds before the CSPP announcement, this relationship has

now inversed13. However, one should keep in mind that the two

TLTRO 14 programmes undertaken by the ECB had CSPP-like

effects on banks financing, because it allowed banks to access

cheap long-term refinancing while incentivizing lending 15 .

Considering this argument, bank bonds as well as subordinated

bonds may be currently trading at attractive credit risk premia.

12 Markit calculates the benchmark spread as the difference between the yield of the bond and a government bond as proxy of a ‘default-free bond’ (e.g. German

government bonds). For further information see Markit iBoxx Spread Analytics:

https://products.markit.com/indices/download/products/guides/Markit_iBoxx_Spread_Analytics.pdf 13 See DB Research (2016) p. 14, Figure 20 14 Targeted longer-term refinancing operations 15 See DB Research: Credit Bites - Will the ECB Buy Bank Senior Unsecured Bonds?



The compression of yields

and the re-correlation

between sovereigns and

corporates create an

unfavourable environment for

barbell type investors

This holds particularly true for

maturity barbell investors,

aiming to increase yields from

additional premia offered by

greater durations

3. Maximising Yield/risk benefits using the Risk Barbell strategy

3.1 Revisiting Risk-Barbelling: Is this the end of the bond barbell strategy?

This question has been asked by a senior active asset manager16 in

respect of USD assets recently and deserves equally to be asked

when looking at EUR bonds. Despite the EUR aggregate fixed

income markets have been turned “upside down”, the extremes of

the risk spectrum still exhibit orderly correlations to equities and rates

(See Figure 3). Traditional fixed income investors may want to

reconsider the usual sovereigns/corporates allocation in order to

cope with such central-bank driven market. Therefore, it is worth

revisiting the risk-barbelling concept which is looking at the extremes

to recoup diversification benefits. In this section, we will first review

the theoretical foundation of the barbell strategy and then follow by

recapping the risk-barbelling strategy. Then, we will look at concrete

application examples of risk-barbelling for investors in the EUR

market while specifically also looking at the use of strategic beta

indices by risk-barbelling investors.

3.2 Traditional Barbell and Risk-Barbell strategies The barbell strategy is widely used among asset managers. It

traditionally exists as a yield curve strategy seeking to invest the

portfolio in two extreme maturity. This is in fact relative to a fixed-

maturity strategy which is a portfolio of only concentrating in one

maturity. As an example, a fixed-maturity portfolio which has only 10-

year maturity bonds and a barbell portfolio which has only 5-year and

20-year maturity bonds might have similar overall portfolio duration.

However, the portfolio yield and convexity characteristics can differ

substantially from each other. Given these differences, the barbell

portfolio helps to achieve diversification between both ends of yield

curves especially during periods of large or non-parallel yield curve

shifts. Specifically, the strategy allows to lock-in higher interest rate

due to exposure to longer-term bonds and the short-term bonds

investment provides an income cushion against sharply falling rates.

Note that a barbell strategy may incur an element of turnover when

rolling the short term component and is extremely reliant on the

shape of the yield curve. Due to its flatness, the barbell strategy

recently lost some attractiveness.

16 Is this the end of the bond barbell strategy? By Arif Husain, Online article in Portfolio Adviser, April 2016: http://www.portfolio-adviser.com/viewpoint/1028390/bond-barbell-strategy



3.3 Risk-Barbell Strategy Applying the traditional yield curve barbell strategy to a portfolio risk

context, risk-barbelling seeks to achieve diversification benefits by

investing at two extreme risk spectrums. It is more and more widely

applied17, as the risky parts of the barbell can be equities or other

diversified sources of risks (and hopefully performance). In the light

of Modern Portfolio Theory (Markowitz, 1952), individual asset risk

can be reduced by holding a diversified portfolio of assets which are

not perfectly positively correlated. Looking from an investment grade

investor’s perspective, this means the risk-barbell portfolio combines

“risky” securities such as corporate bonds with rather “safe”

securities such as sovereign bonds. The corporate bonds have

higher risk as they are significantly more present to idiosyncratic risk

than sovereigns, contrarily sovereign bonds have lower risk and

more sensitive to interest rates. Their sensitivity to different sources

of risks and the fact that interest rates and credit spreads should not

be perfectly correlated unveils the potential portfolio diversification

that can be exploited.

Figure 7 Traditional Barbell Strategy (above) and Risk-barbelling strategy (below)

17 For further reference, see Investors Reach for a New Kind of 'Barbell' in Wall Street Journal, 2014, http://www.wsj.com/articles/more-investors-use-a-barbell-strategy-

to-structure-their-portfolios-1410120110



As an alternative to the

traditional allocation between

sovereigns and corporates,

an allocation between safer

sovereigns and more yield-

oriented corporates seems to

maintain the original

properties of a barbell

strategy

3.4 Long Live Risk-Barbelling: Applications for EUR Investors

Looking at the current EUR overall fixed income market from a risk-

barbelling perspective, Figure 2 demonstrates the evolution of the

yield curve per rating bucket of the current EUR overall market and

the yield difference between the BBB- and AAA-rated segments. The

current market shows more steepness in the yield curve than the

pre-crisis market in 2005 (cf. Figure 8). This manifests the potential

benefits that can be exploited in the current market situation for a

risk-barbell investor at the two ends of the yield curve. The large yield

difference between AAA-rated segment and BBB-rated segment in

2013 was mainly caused by the sharp increase of Italian Sovereigns

spread followed by its downgrade to BBB in 2012.

Figure 8: Yield per Rating Evolution

Yield Difference BBB-AAA

2005 2009 2013 2016

0.43% 1.59% 2.69% 1.37% Source: Markit Group Limited, Deutsche Bank, from 28.02.2005 to 30.06.2016.



Using traditional index carve

outs such as AAA Sovereigns

and BBB Corporates to

implement a barbell can be a

simple and transparent

solution, however at the price

of more idiosyncratic risks

Source: Markit Group Limited, Deutsche Bank, from 28.02.2005 to 30.06.2016.

Table 5 Current EUR Overall Segment by Market Cap

Table 5 provides an overview of the current EUR overall market,

sovereigns and corporates together currently represent

approximately 78% of the EUR overall market. Furthermore, most of

the rating buckets are dominated by the sovereign issues. However,

the high yielding issues in the BBB sovereigns segment are mainly

longer maturity Italian/Spanish sovereign issues. The BBB

sovereigns yields currently 1.19% and has duration of 6.99, whereas

the BBB corporates yields 1.26% and has duration of 5.13 18 .

Therefore, the yield difference is higher between the BBB corporates

and AAA sovereigns than the BBB sovereigns. In light of this, one

would construct a risk-barbelling portfolio by combining AAA

sovereigns with BBB corporates.

Figure 9 shows that the evolution of their yield and the current

convexity between AAA sovereigns and BBB corporates is 1.26%,

which is favourable for a risk-diversified approach. Main issues with

looking into the “extremes” to maximise diversification benefits are

using carve outs might limit the gain of idiosyncratic risks.

Additionally, rating carve outs do address risk reduction but not

necessarily yield increase and it has less representativeness of the

asset class.

EUR Overall AAA AA A BBB

Collateralized 5.39% 2.39% 0.64% 0.30%

Corporates 0.11% 2.04% 7.32% 7.83%

Sovereigns 14.80% 21.84% 1.59% 22.64%

Sub-Sovereigns 5.86% 5.78% 0.50% 0.97%

Source: Markit Group Limited, Deutsche Bank, data as of 30.06.2016.

18 Data as of 03.10.2016.

0%

1%

2%

3%

4%

5%

6%

AAA AA A BBB

Yiel

d

2005 2009 2013 2016



Figure 9: Yield Development of EUR Fixed Income Market Indices

Source: Markit Group Limited, Deutsche Bank AG, from 31.08.2009 to 01.09.2016.

Figure 10: 1-year Average Yield Difference of EUR Fixed Income Market Indices

Source: Markit Group Limited, Deutsche Bank AG, from 31.08.2010 to 01.09.2016.

-1

0

1

2

3

4

5

6

7

8

Aug-09 Aug-10 Aug-11 Aug-12 Aug-13 Aug-14 Aug-15 Aug-16

iBoxx EUR BBB Corporates

iBoxx EUR AAA Sovereigns

0

0.5

1

1.5

2

2.5

3

3.5

Aug-10 Aug-11 Aug-12 Aug-13 Aug-14 Aug-15 Aug-16

iBoxx EUR BBB Corporates-AAA Sovereigns YieldDifference

iBoxx EUR Corporates Yield Plus-EUR SovereignsQuality Weighted Sovereigns Yield Difference



Furthermore, both the price of

holding conservative assets

and the additional yield from

holding BBB corporates may

seem unattractive.

As an alternative to such

carve out indices, strategic

beta indices have been

developed that address such

limitations while maintaining

barbell-friendly features, at

the price of slightly increased

risks but lower tracking error

Figure 11: 5-year window moving correlation with equity

Source: Markit Group Limited, Deutsche Bank AG, 31.03.2010 to 25.07.2016.

3.5 Interesting building blocks for a Risk-barbell: Strategic Beta indices

In addition to the rating carve-out indices, Figure 3 also exhibits the

yield evolution of the strategic beta indices. Interestingly, the Markit

iBoxx EUR Corporates Yield Plus index provides slightly higher yield

than the BBB-corporates although it invests in all rating buckets

across the EUR Corporates market. Similarly, the 1-year average

yield difference of the strategic beta indices exhibits comparably

similar historical trend to the yield difference of AAA-rated

sovereigns and BBB-rated Corporates (Figure 10).

On the corporates side, the Corporates Yield Plus strategy 19

repositions the traditional EUR Corporates benchmark to be focused

only on corporate risk premium. The higher risk premium that Yield

Plus strategy provides result in higher positive correlation to equity

instead of interest rate and to the extent that its yield evolution is very

similar to BBB Corporates (5-year correlation: Yield Plus – 36.25% /

BBB Corporates – 33.28% vs. Euro Stoxx 50)20. The traditional BBB

Corporates carve-out maximises the yield at the cost of

diversification, but Yield Plus strategy is able to maintain relatively

similar yield but not losing diversification (the yield plus strategy has

allocation in different rating bucket from AA to BB).

On the Sovereign side, Looking at the historical yield evolution of

AAA Sovereigns vs. that of the Markit iBoxx Eurozone Sovereigns

Quality Weighted index, the quality oriented approach applied to a

19 For further information, please refer to the relevant index guide from HIS Markit on this index:

http://content.markitcdn.com/corporate/Company/Files/DownloadDocument?CMSID=a07bde73403e495fbbb073e184dc4f89 20 Correlation timeframe: 22/08/2011 to 25/07/2016.



broad sovereign benchmark21 is effective in achieving risk reduction

and to an extent that is comparable to (although not as good as) AAA

Sovereigns while maintaining broad market access (Figure 11). The

broad diversification that the quality weighted approach offers is of

great help in avoiding portfolio concentration risks (at the cost of

lower yield than the EUR Sovereigns index). As a result, the Quality

Weighted index has lower correlation to equities than the traditional

EUR Sovereign index (5-year correlation: Quality Weighted - 4.98%

/ EUR Sovereigns - 18.97% vs. Euro Stoxx 50)22. See Figure 5 for

historical correlation evolution.

In a nutshell, both strategic beta strategies seek to reposition

the traditional benchmark indices to recover the diversification

benefits by either via risk reduction or yield enhancement.

Consequently, the correlation benefits in a portfolio context

provided by the strategic beta indices are greater than their

traditional benchmarks and to some extent comparable to that

of the traditional rating carve-out indices (Table 6: Sovereigns

and Corporates CorrelationTable 6).

Table 6: Sovereigns and Corporates Correlation

Correlation Comparison

EUR Sovereigns Quality Weighted

vs. Corporates Yield Plus

EUR Sovereigns vs.

EUR Corporates

EUR AAA Sovereigns vs.

EUR BBB Corporates

45% 59% 44% Source: Markit Group Limited, Deutsche Bank, 07.03.2005-25.07.2016.

4. Applications of the barbell for institutional

investors: what makes sense, what does not23 In this section, we aim to examine the practical added value of the

risk-barbelling strategy by simulating several topical portfolio

strategies. Each portfolio strategy seeks to illustrate the utility of a

risk-barbell portfolio from different investment perspective. For this

purpose, we first studied the benchmark replication approach to

assess the existence of risk-barbelling benefits in the current EUR

overall market, then we looked at the risk-barbelling portfolio strategy

from a total return perspective, then about the positive aspects of

adding sovereign bond exposure for single corporate bond investors

looking to increase yield and last about the positive aspects for broad

EUR sovereign bond investors to slightly reposition their exposure

so as to increase capital preservation and risk-adjusted returns. The

portfolio analysis of this section is based on either daily or weekly

returns of the respective indices from 2005 to 2016. The rebalancing

frequency and procedure is described in details in each following

section.

21 For further information, please refer to the relevant index guide from HIS Markit on this index:

http://content.markitcdn.com/corporate/Company/Files/DownloadDocument?CMSID=a9aa579b36274cf7a78724cfe69da023 22 Correlation timeframe: 22/08/2011 to 25/07/2016. 23 Please note for this section: past performance, actual or simulated, is not a reliable indicator of future performance.



Strategic fixed income

investors may find it

interesting to implement a

risk barbell when looking to

enhance risk-adjusted returns

However, due to higher

idiosyncratic risks, a barbell

based on carve-out indices

implied higher tracking error

versus an aggregate

benchmark than a barbell

based on Strategic Beta

indices in the simulation

Also, lower yields obtained

from carve-out indices

indicate that the performance

benefits of using such

components in this use case

may be limited going forward

4.1 Case 1: SAA: enhancing FI Aggregated returns To validate the theoretical benefits of risk-barbelling, we have re-

constructed the iBoxx EUR Overall Index which represents a

benchmark for the EUR denominated investment grade fixed income

asset class. It is to be noted that this case does not aim to serve as

an investable portfolio strategy but rather as an illustration of

theoretical implications. The EUR Overall on average consists of

61% sovereigns, 16% corporates, 11% sub-sovereigns and 12%

collateralized. We have substituted the sovereigns segment of the

EUR Overall with AAA sovereigns and the corporates segment with

BBB corporates. In a similar vein to demonstrate the added value of

risk-barbelling with strategic beta indices, we have replaced the

sovereigns segment with the iBoxx EUR Eurozone Quality Weighted

Index and the corporates segment with the iBoxx EUR Corporates

Yield Plus Index. The reconstructed portfolios are calculated based

on the daily return and rebalanced annually to comply with the EUR

Overall market-cap weights of the respective segment. Table 7

documents the results from this empirical analysis. Both

reconstructed strategies have outperformed the traditional EUR

Overall benchmark. This validates the empirical added value of the

risk-barbelling strategy. Furthermore, the reconstruction with

strategic beta indices have further outperformed the benchmark

without affecting its risk profile. This demonstrates the further added

value of using index strategies instead of traditional index rating

carve-outs. The outperformance of the strategic beta indices did not

come at the cost of diversification or lower yield. Specifically, the

materially lower tracking error obtained when using strategic beta

indices indicates that the broad market access features of the

Strategic Beta indices helped reduce deviations versus benchmark

compared to the traditional portfolio. It is also worth noting that the

traditional portfolio combining AAA Sovereigns and BBB Corporates

is currently only yielding 20 basis points whereas the strategic beta

portfolio has an annual yield of 49 basis points, in line with that of the

benchmark.



Table 7: Use Case 1 Simulation Results

Use Case 1

Benchmark EUR Overall Reconstruction

EUR Overall

Using AAA Sovs and BBB Corps

Using Strategic Beta Indices

Annualized Return 4.78% 4.93% 5.04%

Volatility (Duration Hedged)

2.61% 2.31% 2.46%

Sharpe Ratio (Duration Hedged)

1.83 2.13 2.05

Turnover N/A 3.67% 3.83%

Max. Drawdown -5.11% -5.21% -5.10%

Drawdown Date 03/18/2009 03/18/2009 03/18/2009

Tracking Error N/A 1.31% 0.44%

Avg. Yield 3.18% 3.00% 3.20%

Avg. Duration 5.56 5.60 5.54

Stress Test: Lehman Default - 200824

-0.79% -0.62% -0.89%

5-year Correlation with EUROSTOXX 50

22% -10% 17%

Annual Yield 0.48% 0.20% 0.49%

Duration 6.74 6.85 6.78

Time frame of analysis is from 28.02.2005 to 31.07.2016. Annual yield and duration data

are as of 29.07.2016.

4.2 Case 2: Replicating aggregate bond returns with only corporates and sovereigns

In accordance with popular practice in the asset management space,

the EUR Overall index can be reconstructed using only sovereigns

and corporates, as such approach requires low implementation

sizes. First, the “Duration + Yield Matching” approach seeks to obtain

the weekly optimal portfolio allocation that minimizes the yield and

duration differences between the reconstructed portfolio and the

EUR Overall index (cf. Bolder, 2015).

The portfolio is rebalanced when the daily “Duration + Yield

matching” target weight of the corporates/sovereigns component

differ by more than 10% from the current corporates/sovereigns

weight in the portfolio (“10% buffer” so as to avoid turnover). The

“Duration + Yield matching” objective function is minimization of ex

ante portfolio yield and duration distance to the benchmark index

(𝐷𝑝)

* 3-Y weekly price correlations between corporates and sovereigns benchmark went up from 46% in September 2012, indicating that sovereign credit risk correlated

little with corporate credit risks during the 2011 crisis, to 75% in August 2016. 24 Please refer to the Annex for further information about this stress test scenario. Source: Bloomberg LLP.



In our historical simulation,

the heuristic way of

replicating aggregate returns

(combining sovereigns and

corporates while trying to

match duration and yield)

recently suffered from an

increase in price correlations*

between corporate and

sovereign bond benchmarks

Using a traditional barbell

allocation helped restore

diversification benefits while

matching benchmark returns

Using a barbell allocation

between Strategic Beta

indices currently helps

increase yields, at lower

tracking error and risk

features in line with the

benchmark in the simulation

𝐷𝑝 = (𝜔′𝑌 − 𝑌𝐵𝑀)2 + (𝜔′𝐷𝑈𝑅 − 𝐷𝑈𝑅𝐵𝑀)2

where ω is a 2-by-1 vector of index weights, Y is an 2-by-1 vector of

asset annual yield, 𝑌𝐵𝑀 is the annual yield of the benchmark index

(i.e. iBoxx EUR Overall Index), DUR is an 2-by-1 vector of index

annual modified durations and 𝐷𝑈𝑅𝐵𝑀 is the annual modified

duration of the benchmark index.

Second, as a simplistic proxy, the 25/75 replication approach

matches broadly the historical market cap allocation of the EUR

Overall index, of which the reweighted (i.e. without sub-sovereigns,

allocated to sovereigns and collateralized, allocated to corporates)

sovereigns and corporates segments represent 75% and 25%

respectively. Such 25/75 allocation is maintained on an annual basis

and the portfolio is calculated based on daily returns. For comparison

purposes, both strategies have been studied based on standard

benchmark corporate/sovereign indices and strategic beta indices.

The standard portfolio uses iBoxx EUR Sovereigns index and iBoxx

EUR Corporates index and the strategic beta portfolio uses iBoxx

EUR Sovereigns Quality Weighted index and iBoxx EUR Corporates

Yield Plus index.

The strategic beta based portfolios of both strategies offered an

improvement of the risk return profile while exhibiting similar levels

of tracking error versus overall benchmark to the use of standard

indices. The strategies have especially outperformed the EUR

Overall benchmark (Table 8) and the simplistic 75/25 strategy

reaches similar effects (at the price of slightly higher duration and

credit risks than the more dynamic approach). See Figure 12 for

the historical weight evolution of the Duration + Yield Matching

strategy. It is interesting to note that the distance minimisation

method did not diverge and remained within a constrained range.

Other portfolio characteristics such as yield and duration have

remained comparably similar to the EUR Overall benchmark.

Broadly similar results can be achieved at level of the AAA Sovs /

BBB Corps based strategy, confirming the added value of looking

into the extremes, however at the price of materially higher tracking

error and currently lower yields.




Use Case 2

Bench

mark Duration+ Yield Matching 25/75

EUR

Overall

Stand

ard

AAA Sovs &

BBB Corps

Strategic

Beta

Stand

ard

AAA Sovs &

BBB Corps

Strategic

Beta

Annualized Return

4.78% 4.77% 4.98% 5.02% 4.90% 5.09% 5.17%

Volatility

(Duration Hedged)

2.61% 2.46% 1.91% 2.41% 2.95% 2.52% 2.73%

Sharpe Ratio

(dur. Hedged) 1.83 1.94 2.61 2.08 1.66 2.02 1.89

Turnover N/A

21.01%

20.95% 22.95% 1.41% 2.39% 2.27%

Max. Drawdown -5.11%

-4.99%

-5.12% -4.87% -5.75% -5.65% -5.52%

Drawdown Date 03/18 /2009

03/09/2009

06/22/ 2015

03/16/ 2009

03/18/2009

06/10/ 2015

03/18/ 2009

Tracking Error N/A 0.92% 1.74% 0.67% 0.39% 1.45% 0.49%

Avg. Yield 3.18% 3.33% 3.13% 3.36% 3.29% 3.07% 3.34%

Avg. Duration 5.56 5.62 5.77 5.71 5.80 5.84 5.78

5-year

Correlation with EUROSTOXX 50

22% 24% -7% 24% 26% -10% 21%

Annual Yield 0.48% 0.67% 0.36% 0.73% 0.58% 0.26% 0.62%

Duration 6.74 6.71 7.00 6.85 7.03 7.15 7.07

Time frame of analysis is from 28.02.2005 to 31.07.2016 unless specified otherwise. The

Standard portfolio is constructed based on iBoxx EUR Eurozone Sovereigns and iBoxx EUR

Corporates indices. AAA Sovs & BBB Corps portfolio is constructed based on iBoxx EUR AAA

Sovereigns Index and iBoxx EUR BBB Corporates Index. The Strategic Beta portfolio is

constructed based on iBoxx EUR Eurozone Sovereigns Quality Weighted and iBoxx EUR

Corporates Yield Plus indices. Sharpe Ratios are calculated on the basis of zero rates and

rate-adjusted volatility, so as to provide a measure of credit risk adjusted returns

Figure 12: Historical Weight Allocation-Duration+Yield Matching Risk-Barbelling Portfolio

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Feb-05 Feb-07 Feb-09 Feb-11 Feb-13 Feb-15

Markit iBoxx EUR Corporates Yield Plus Index Markit iBoxx EUR Eurozone Quality Weighted



While following a mean-

variance strategy may

generate additional turnover,

it is a way to illustrate the

added diversification benefits

of using a risk barbell

4.3 Case 3: Simulating an active allocation strategy - Applying minimum variance-based risk barbell to reap diversification benefits

The minimum variance-based strategy capitalises especially on

diversification benefits by minimizing portfolio spread volatility

(weekly basis) especially in case of volatile events to increase the

risk-adjusted returns. We have implemented a 20% rebalancing

buffer for this portfolio to mitigate frequent rebalancing due to the

strategy’s sensitive nature. This means that when the weekly

minimum variance optimization target weight of a

corporates/sovereigns component differs by more than 20% from the

current corporates/sovereigns weight in the portfolio, a rebalancing

is triggered. 20% buffer has been chosen given this strategy’s large

turnover characteristic and targeting an approximate 50%25 turnover

p.a. while remaining dynamic (as opposed to rebalancing regularly

but more seldom, which may be less reactive and more dependent

on the timing of rebalancing). The minimum-variance objective

function is minimization of ex ante (i.e., estimated) portfolio spread

variation (𝜎𝑝2) (Clarke et al., 2013)

𝜎𝑝2 = 𝜔′Ω𝜔

where ω is an 2-by-1 vector of asset weights, and Ω is an 2-by-2

asset benchmark spread variation covariance matrix. The equation

is solved analytically for convenience purposes.

While sovereigns typically offer the lowest spread volatility, it is

interesting to note that a portfolio combination with corporates

started increasing diversification benefits from the start of the

Eurozone crisis in H2 2011. The lower drawdown confirms the

volatility reduction feature of the strategy; meanwhile it generated

attractive outperformance. Such strategy, if implemented, would

suffer relatively high turnover costs26 (cf. Figure 13, Table 9).

This case should be considered a proxy of an active/dynamic

strategy and its results have important implications as the strategy

exploits the aforementioned upside down market situation: The

underperformance of the Standard portfolio (vs. EUR Overall)

confirms the existence of lack of correlations in the traditional

benchmark indices, while the 76 basis points performance pickup

from the strategic beta strategies demonstrate that such

unconventional indices can be helpful in repositioning the asset

class’s correlation and recovering portfolio correlation benefits.

Results are similar if not even more attractive when using the AAA

Sovereigns & BBB Corporates portfolio components, which further

proves our point. The attractive outperformance in the simulation can

25 50% per annum turnover is already an upper limit in terms of portfolio stability. 26 Assuming 50 bps bid-offer costs on corporates index/ETF and 15 bps bid-offer on sovereigns index/ETF, average trading costs would be approx. 19 bps p.a., however

being applied very irregularly over the period.



Applying this dynamic

strategy using Strategic Beta

or carve out indices

contributed to material

improvements in the risk

return profile in the

simulation, especially

lowering risks. However this

occurred at the price of high

tracking error.

however seem slightly less attractive as of now, given the currently

lower yields and slightly higher duration available. This case also

means that the use of strategic beta building blocks lends itself

particularly well to dynamic active allocation strategies aiming at

varying allocation between conservative (risk off) and higher yielding

(risk-on) components.


Use Case 3

Benchmark Minimum Variance

EUR Overall Standard

AAA Sovs & BBB Corps

Strategic Beta

Annualized Return 4.78% 4.61% 5.88% 5.55%

Volatility

(Duration Hedged) 2.61% 2.33% 2.07% 2.55%

Sharpe Ratio (Duration Hedged)

1.83 1.98 2.84 2.18

Turnover (2-way) N/A 49.71% 48.74% 54.47%

Max. Drawdown (Duration Hedged)

-6.59% -6.50% -5.15% -4.94%

Drawdown Date 03/18/2009 11/28/2011 03/09/2009 03/09/2009

Tracking Error N/A 3.68% 3.48% 3.50%

Avg. Yield 3.18% 3.32% 3.09% 3.06%

Avg. Duration 5.56 5.35 5.82 5.98

5-year Correlation with EUROSTOXX 50

22% 20% -7% 20%

Annual Yield 0.48% 0.61% 0.17% 0.42%

Duration 6.74 6.85 7.34 7.54

Time frame of analysis is from 28.02.2005 to 31.07.2016 unless specified otherwise. The

Standard portfolio is constructed based on iBoxx EUR Eurozone Sovereigns and iBoxx EUR

Corporates indices. The Strategic Beta portfolio is constructed based on iBoxx EUR Eurozone

Sovereigns Quality Weighted and iBoxx EUR Corporates Yield Plus indices.

Figure 13: Historical Weight Allocation-MVP Risk-Barbelling Portfolio

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Mar-06 Mar-08 Mar-10 Mar-12 Mar-14 Mar-16

Markit iBoxx EUR Eurozone Sovereigns Quality Weighted

Markit iBoxx EUR Corporates Yield Plus



Interestingly, adding

corporate bond exposure to a

sovereigns portfolio helped

reduce risks from the end of

2011 in the simulation

Adding conservative

sovereign bond exposure to a

corporate bond helped yield

seeking investors to mitigate

their risk increase in the

simulation.

4.4 Case 4: Corporate bond investors: Yield enhancement by mixing in quality sovereigns

This case simulates the behaviour of a single-line corporate bond

investor who focuses on relatively safe and liquid, large corporates.

As an alternative to their investment behaviour, we look at a yield

enhancement approach which aims to allocate a diversified portfolio

by matching the yield of the Markit iBoxx EUR Liquid Corporates 100

index plus 30 basis points and similarly plus 70 basis points27.

Results can be seen on Table 10. For 30 basis points yield

enhancement, this approach leads to an allocation of 55% in EUR

Corporates Yield Plus and 45% in EUR Sovereigns Quality Weighted

1-15Y. Note that we have used the 1-15Y maturity carve-out of the

Quality Weighted Index to match the duration objective of an

equivalent corporate bond investor. Note the results are similar (if

not better) when using the standard index, at the price of almost 1Y

additional duration. An additional 70 basis points in yield requires an

allocation of 78% in EUR Corporates Yield Plus and 22% in EUR

Sovereigns Quality Weighted (main index as proportion of

sovereigns is limited and their effect on portfolio risk mitigation is

increased as duration is increased).

Adding the sovereigns segment to the portfolio does not only provide

additional safety when looking to increase yields, it may also help

mitigate immediate portfolio liquidity risks, given the current EUR

corporates market has recently been heavily subscribed. Besides

the yield enhancement, this approach aims to maintain similar/

reasonably increase risk levels thanks to better diversification (1136

bonds instead of 100) than a single-line corporate bond investment.

For the 30 basis point case, the historical average yield pickup was

approximately 46 basis points while maintaining the same Sharpe

ratio (based on credit) and even widely improved Sharpe ratio when

using price volatility measurement, which is interesting as more risky

corporate portfolios can rather be expected to offer lower risk

adjusted returns, see for example Leote De Carvalho (2014). The 70

basis point case has achieved historical average yield improvement

of 86 basis points, while delivering higher credit risk (well

compensated by interest rates variations, hence the lower price

volatility) in the past.

It is interesting to note that a pure BBB corporates allocation would

also offer approx. 70 bps of yield enhancement (as of 31/07/2016),

while AAA sovereigns offer negative yields, making the proportion of

AAA-rated sovereigns insufficient to compensate for the risk

increased by the BBB index within the frame of a yield enhancement

portfolio. It is also interesting to note that the stress test figures for

27 Yield matching is considered as at 30.06.2016.



Increasing yields can result in

lower price volatility

historically, which can be

beneficial for certain clients

looking for price stability.

However credit and tail risks

will be materially increased, at

least in our simulation.

BBB bond exposures as of now are slightly worse than the +70bps

Yield Risk Barbelling portfolio, with 4.0% estimated loss upon

Lehman Default instead of 3.8%. Also the price volatility of the BBB

index would be higher.


Use Case 4

Benchmark Corporate Bonds Case

EUR Liquid Corporates 100

Index

+30 bps Yield Risk Barbelling

55% YP/45% QW 1-15

+70bps Yield Risk Barbelling

78% YP/22% QW


Volatility 2.74% 2.73% 2.57%

Volatility (Duration Hedged)

2.01% 2.35% 2.83%

Sharpe (Duration Hedged) 1.97 1.99 1.69


Drawdown -8.99% -10.10% -15.38%

Drawdown Date 03/18/2009 03/11/2009 03/17/2009


Average Yield 3.24% 3.70% 4.26%

Average Duration 4.16 4.51 4.78

Stress Test: Lehman Default - 2008

-2.01% -2.80% -3.81%

Annual Yield 0.74% 1.04% 1.61%

Duration 4.80 5.26 5.63

Time frame of analysis is from 31.10.2005 to 30.06.2016 unless specified otherwise. QW

refers to iBoxx EUR Eurozone Sovereigns Quality Weighted index and YP refers to iBoxx EUR

Corporates Yield Plus index. Annual Yield and duration data are as of 30.06.2016.

4.5 Case 5: Sovereign bond investors - Looking for additional safety at reasonable opportunity costs, looking to slightly increase yields while maintaining capital preservation features

We have looked at simulating the behaviour of middle term sovereign

bond investors who remained in this little attractive asset class as it

serves diversification or capital preservation features. An example

could be balanced Equity / multi-asset managers invested in

Eurozone sovereigns (as simulated using the iBoxx EUR Eurozone

Sovereigns 1-15Y index carve out) as a means to preserve capital

during equity bear markets situations.

First, an overall “quality improvement” of their sovereigns exposure

can be aimed at by combining a quality oriented sovereigns index (in

this case also limited to 1-15Y maturities so as to avoid duration

biases due to including corporate bonds, to the extent possible) with



In line with the results

suggested in Use Case 3,

historical simulations show

that adding a limited

corporate bond exposure to a

sovereign bond portfolio

while reducing exposure to

BBB sovereigns can help

slightly improve the capital

preservation features of an

intermediate sovereign

benchmark.

higher risk corporates to match original benchmark yields while

avoiding peripheral exposure.

The second case aims to obtain 20 basis points ex-ante “yield

enhancement” on the sovereign bond portfolio while keeping similar

capital preservation features and similar diversification benefits vs.

equities to a benchmark sovereign index, to the extent possible.

The results (cf. Table 11) have shown noticeable improvements in

terms of capital preservation throughout the 2011 crisis, a sign of

reduction of the exposure to Southern Europe. The “quality

improvement” case shows a risk reduction according to almost all

measurements: lower expected drawdowns in an equity crisis

scenario (here it is worth noting that the Bloomberg engine expects

a very high correlation to apply to bond prices if stocks went to go

down by 10%!) while the correlation versus equities also slightly

decreased (iBoxx EUR Eurozone Sovereigns 1-15Y vs. EURO

STOXX 50: -1% and Yield Matching Risk-Barbelling Portfolio vs.

Euro Stoxx 50: -11 %28), a further sign that objective was achieved

historically. Also, interestingly for the same yield on average, the

performance improved.

For the “yield enhancement” case, objective was also achieved in

terms of realised performance (more than 25 bps p.a. of

outperformance) at no additional cost in terms of capital preservation

(as measured by max. drawdown) and at some element of cost in

terms of diversification benefits versus equities (Yield Matching Risk-

Barbelling Portfolio vs. Euro Stoxx 50: 0%), while certain metrics like

Stress Tests indicate higher potential risks on the back of the yield

increase. It is worth noting, as mentioned in the first part of this

document, that long term sovereigns benefit from higher yields,

therefore making an investment in corporates (with typically 5-year

duration) look less attractive for long term sovereign investors.

28 Correlation figures were calculated on a weekly basis. Using monthly returns, correlations vs Euro Stoxx 50 are respectively -9% for the iBoxx EUR Eurozone

Sovereigns 1-15Y index, -9% for the “quality improvement” portfolio and 15% for the “yield enhancement” portfolio.




Use Case 5

Benchmark Sovereign Bonds Case

EUR Eurozone Sovereigns1-15

Yield Matching Risk Barbelling

5% YP /

95% QW 1-15

+20 bps Yield Risk Barbelling

26% YP/

74% QW 1-15


Volatility

(Duration Hedged) 2.87% 2.38% 2.30%

Sharpe Ratio 1.55 1.90 2.03


Max. Drawdown

(Duration Hedged) -6.02% -4.18% -7.86%

Drawdown Date 11/25/2011 03/18/2009 03/18/2009


Average Yield 2.77% 2.67% 3.08%

Average Duration 4.66 4.65 4.59

Stress Test:

Lehman Default - 2008 -0.31% -0.48% -1.34%

Stress Test: Equities

down 10% -0.10% -0.07% -0.34%

Annual Yield 0.16% 0.16% 0.45%

Duration 5.37 5.32 5.23

Time frame of analysis is from 31.10.2005 to 31.07.2016 unless specified otherwise. QW

refers to iBoxx EUR Eurozone Sovereigns Quality Weighted index and YP refers to iBoxx

EUR Corporates Yield Plus index.

Note that clearly looking to further enhance yields on the safe

bucket of such investors does not seem to make sense. It may also

be of interest to keep in mind that the motor of returns in the

balanced portfolio is the risky part of it, the mission of the

conservative part of it being often rather oriented towards capital

preservation, at whatever (reasonable) cost.

5. Conclusion As a lot of investors use yields as a proxy to estimate fixed income

returns, may traditional fixed income indices may currently look very

unattractive, leading investors to go down the quality ladder in the

search for yield. The growing imbalances that characterise the EUR

IG fixed income asset class, as mentioned in the first part of this

paper, mean above all that the relation between risk and yield

becomes more and more unclear in many segments of this

asset class, rendering certain traditional bond strategies or

exposures less and less attractive from a yield vs risk perspective.

This being said, neither Risk Barbelling strategy nor Strategic

Beta indices are an adapted solution to all investors nor to all

investors’ issues. As an example, unconstrained, long-term

institutional investors may not be bothered with strong market

variations nor downgrades, and may very well be better served to

look into less liquid or lower rated asset classes. As mentioned by

our active management colleague16, diversifying sources of income

is also key. A barbell strategy should not be used by broad fixed



income investors from a standalone perspective: the loss in

risk/adjusted returns that can be noticed when replicating an

aggregate component passively with only sovereigns and

corporates is a testament to the importance of diversification,

not only internal (de-correlation) but also external (number of

risk sources) diversification.

Strategies combining broadly diversified tools that are more

explicitly targeted at achieving higher yields on one side and

more quality on the other side, has however become more and

more relevant in light of the current market situation.

Such strategies, when applied to parts of a portfolio, can not only

help investors continue pursuing their usual objectives (e.g.

achieving broad fixed income exposure or actively allocating

between market segments) in a more efficient way, they can also

help less sophisticated investors looking to harness higher yields at

the price of a reasonable risk increase.

Going forward, the strategic beta indices we specifically

analysed can help investors address the scarcity of yields in the

corporates space as well as the high costs of safe assets thanks

to broader market exposure. Although the features of the barbell

using Strategic Beta indices seem slightly less attractive for certain

investors (cf. case 3), those indices offer a good alternative to

standard sovereign/corporate benchmarks on one side (higher

diversification benefits, clearer signals enabling investors to

better implement their views) and to barbell investors on the

other (lower tracking error to overall benchmarks, higher yields

on both sides of the barbell)

6. Annex - Measuring risks in the fixed income assets

In this paper, we mostly measured risk, especially within the frame

of risk/return analysis. While this decision was not made for the lack

of better options, we found that certain measurements may not have

been adequate for this study:

Traditional value at risk measures have been identified as

being weak in predicting credit crises, as identified by

Deventer (2013), especially as in practice higher value at risk

figures in the Investment Grade fixed income market mostly

reflected the higher duration (and especially interest rate

sensitivity) of the bond components analysed. Certain

models dedicated in applying a VAR analysis to credit

spreads are available (see Raunig, 2008), however they

deviate from the natural and intuitive price percentile

approach applied for equity.



Traditional price variation driven measurements such as

volatility or correlation exhibits the same weakness. This is

also important when comparing diversification benefits of

certain fixed income instruments based on covariances of

prices, where all asset classes seem to be much correlated

because they essentially have similar sensitivity to moves in

interest rates. While interest rates risk is a major contributor

of returns in the asset class,

(i) Most investors are fully aware of that risk, which is the

common denominator of the asset class, whereby

certain institutional clients even need this risk, e.g.

For Liability Driven Investment purposes;

(ii) as interest rates and credit risks have been negatively

correlated historically (as confirmed in this paper),

interest rates risks are expected to present an

element of noise when trying to assess other, market

related risks such as (corporate and sovereign) credit

risks

(iii) Even within the frame of a relative value analysis,

removing this common component in asset class

returns is expected to help clarify the analysis.

We, of course, show modified duration in our analyses as that

number is broadly used by academics to measure sensitivity

to interest rates risks.

Equity price volatility, as clearly identified in the Merton model

(Merton, 1974) as one of the potential factor to determine

credit risks, and any implied probability of default from, it,

does unfortunately not cover sovereign credit risks. However,

traditionally, measuring correlation of a fixed income asset

class against equity prices will usually help distinguish how it

reacts to equity market variations, a negative correlation

being intuitively expected for the least risky asset classes and

a positive to the more risky ones. We used this measurement

in our paper as well, as a complement to other measures.

The duration times spread (DTS) measure, calculated by

multiplying duration with the spread of a bond (or an index)

and introduced by Ben Dor et al. (2007) does take into

account credit and its “risk multiplication factor” which is

spread (or credit) duration, therefore giving, in our view, a

very interesting (because transparent and interest rates

neutral) measurement of credit risk. However, one key

limitation of this figure is that it is based on the spread, which

is an implied measurement of the risks of a bond as

determined by market prices. It therefore assumes full

efficiency of the fixed income markets, which is an

assumption that can definitely be challenged by virtue of the



large offer/demand imbalances in the EUR IG asset class

mentioned in this paper.

Certain low maximum drawdown figures for conservative

fixed income portfolios are mostly driven by interest rates (as

in April 2015) and may not sufficiently be representative of

risk.

We mainly chose to use variation in spreads when measuring risks

and correlations for the following reasons:

Volatility of spreads is an intuitive measure of interest-rates

adjusted market risks that is widely used in the industry (and

specifically by Ben Dor et al. as well).

It being clear that spreads are a premium on top of riskless

rates and therefore encompass all other risks around a bond

price, namely, among others, credit but also liquidity (also

including supply/demand imbalances), “complexity” (due to

convexity, special covenants in the prospectus, non-vanilla

coupon payouts etc.) and long duration, spread volatility has

been observed to also be consistent with fundamental credit

risks in the works of Delianedis et al. (2013).

While spread volatility using daily changes is implied from

price volatility and therefore also assumes market efficiency,

the relative tardiness of the biggest fixed income purchasers

that can be expected to skew the supply demand balance

(such as central banks and banks financed by central banks)

cannot seem to prevent day-to-day spread variations, as

seen in the case of Italian sovereigns in this paper.

We chose to multiply the spread by duration when calculating

volatilities, so as to re-scale the importance of risk by the

maturity of the relevant bond/index29. Note that this approach

comes down (qualitatively at least) to measuring the volatility

of interest rates hedged price variations, which we also used

as a metrics for convenience purposes in this paper.

For the conservative portfolio, we apply a naïve “bullet” interest rate

hedging (using the portfolio duration and simulating a short interest

rates swap exposure as overlay) to calculate maximum drawdown.

It also can help better appreciate risks especially in a situation where

interest rates may not be able to move further down in the negative

territory

As a complement to these measurements, we chose to add the

stress test scenario “Lehman Default - 2008” (a stress event

simulating variation of asset prices, using historical factor returns,

one month following the default of Lehman Brothers) developed by

the Bloomberg Analytics team and the “Equities down 10%” case (a

29 Note that multiplying with duration does not impact correlation figures due to the consistency of duration in indices across time and to the fact that correlations are

normalised.



test mimicking the behaviour of the portfolio should stock prices go

down 10% as of now). Those stress tests are applied to the current

portfolio composition and would be expected to represent the

behaviour of the current portfolio in case a similar situation occurred

again.

Note that this measurement, is far from being comprehensive

especially when applied at level of a single issuer. Despite the ex-

post consistency, at broader levels, in the results obtained by

Delianedis et al., it can only rely on markets perceptions of risks,

which can of course be perceived as limited. This measurement can

and must therefore, in our view, be complemented with fundamental

driven measurements as well as, if possible, with further stress tests

that correspond better to an investors specific perception of risk.



References

Merton, R. C., 1974, On the Pricing of Corporate Debt: The Risk Structure of Interest Rates, Journal of Finance, 29,

Donald R. Van Deventer, Kenji Imai, Mark Mesler, 2013 Advanced

Financial Risk Management: Tools and Techniques for Integrated

credit risk and interest risk management.

Burkhard Raunig, Martin Scheicher, November 2008 - A Value At

Risk Analysis Of Credit Default Swaps – ECB Working Paper

Series No 968 /

Arik Ben Dor, Lev Dynkin, Jay Hyman, Patrick Houweling, Erik van

Leeuwen and Olaf Penninga - DTS (Duration Times Spread) - A

New Measure of Spread Exposure in Credit Portfolios (Journal of

Portfolio Management, Winter 2007)

Delianedis, G., Geske, R., 12/01/2001, The Components of Corporate Credit Spreads: Default, Recovery, Tax, Jumps, Liquidity, and Market Factors, University of California

Bolder, D.J. (2015) Fixed-income portfolio Analytics; A practical guide to implementing, monitoring and understanding fixed-income portfolios. United States: Springer International Publishing AG. Clarke, Roger G and de Silva, Harindra and Thorley, Steven, Risk Parity, Maximum Diversification, and Minimum Variance: An Analytic Perspective (June 1, 2012). Journal of Portfolio Management, Vol. 39, No. 3, pp. 39-53 (Spring 2013). Edwin J. Elton, Martin J. Gruber (15/09/2003), On The Valuation Of Corporate Bonds Using Rating-Based Models, New York University/Nomura

Michal Jezek, Nick Burns, Jim Reid (07/09/2016), Will the ECB Buy Bank Senior Unsecured Bonds?, Deutsche Bank Markets Research

Michal Jezek, Nick Burns, Jim Reid (05/07/2016), EUR and GBP IG Bond Issuance, Deutsche Bank Markets Research Jack Di-Lizia, Abhishek Singhania, Eurozone Sovereign Issuance: 05th Sep-09th Sep, Deutsche Bank Markets Research

Raul Leote De Carvalho, Patrick Dugnolle, Xiao Lu (Spring 2014), Low-Risk Anomalies in Global Fixed Income: Evidence from Major Broad Markets, The Journal Of Fixed Income



Disclaimer © Deutsche Bank AG 2015 As of: 31.10.2016

This document has been issued and approved by Deutsche Bank AG London Branch. Deutsche Bank AG is authorised under German Banking Law (competent authority: European Central Bank) and, in the United Kingdom, by the Prudential Regulation Authority. It is subject

to supervision by the European Central Bank and by BaFin, Germany’s Federal Financial Supervisory Authority, and is subject to limited regulation in the United Kingdom by the Prudential Regulation Authority and Financial Conduct Authority. Deutsche Bank AG is a joint

stock corporation with limited liability incorporated in the Federal Republic of Germany, Local Court of Frankfurt am Main, HRB No. 30 000; Branch Registration in England and Wales BR000005 and Registered Address: Winchester House, 1 Great Winchester Street, London

EC2N 2DB. This document is a “non-retail communication” within the meaning of the FCA's Rules and is

directed only at persons satisfying the FCA’s client categorisation criteria for an eligible counterparty or a professional client. This document is not intended for and should not be relied upon by a retail client.

Deutsche Asset Management is the brand name for the asset management arm of Deutsche Bank AG.

This document is intended for discussion purposes only and does not create any legally binding obligations on the part of Deutsche Bank AG and/or its affiliates (“DB”).This material was not produced, reviewed or edited by the Research Department, except where specific documents

produced by the Research Department have been referenced and reproduced above. Without limitation, this document does not constitute an offer, an invitation to offer or a recommendation to enter into any transaction. When making an investment decision, you should rely solely on

the final documentation relating to the transaction and not the summary contained herein. DB is not acting as your financial adviser or in any other fiduciary capacity in relation to this Paper. The transaction(s) or products(s) mentioned herein may not be appropriate for all investors and

before entering into any transaction you should take steps to ensure that you fully understand the transaction and have made an independent assessment of the appropriateness of the transaction in the light of your own objectives and circumstances, including the possible risks

and benefits of entering into such transaction. For general information regarding the nature and risks of the proposed transaction and types of financial instruments please go to https://www.db.com/en/content/Risk-Disclosures.htm. You should also consider seeking

advice from your own advisers in making this assessment. If you decide to enter into a transaction with DB, you do so in reliance on your own judgment. The information contained in this document is based on material we believe to be reliable; however, we do not represent

that it is accurate, current, complete, or error free. Assumptions, estimates and opinions contained in this document constitute our judgment as of the date of the document and are subject to change without notice. Any projections are based on a number of assumptions as to

market conditions and there can be no guarantee that any projected results will be achieved. Past performance is not a guarantee of future results. Any opinions expressed herein may differ from the opinions expressed by other DB departments including the Research Department. DB

may engage in transactions in a manner inconsistent with the views discussed herein. DB trades or may trade as principal in the instruments (or related derivatives), and may have proprietary positions in the instruments (or related derivatives) discussed herein. DB may make

a market in the instruments (or related derivatives) discussed herein. You may not distribute this document, in whole or in part, without our express written permission. DB SPECIFICALLY DISCLAIMS ALL LIABILITY FOR ANY DIRECT, INDIRECT, CONSEQUENTIAL OR OTHER

LOSSES OR DAMAGES INCLUDING LOSS OF PROFITS INCURRED BY YOU OR ANY THIRD PARTY THAT MAY ARISE FROM ANY RELIANCE ON THIS DOCUMENT OR FOR THE RELIABILITY, ACCURACY, COMPLETENESS OR TIMELINESS THEREOF.

Markit and iBoxx are trademarks of Markit and have been licensed by International Index Company, a subsidiary of Markit (“Licensor”). The Licensor of the indices listed here make no

representations or warranties concerning the results obtained by using their index and/or index levels or in any other respects, on any given day.

https://www.db.com/en/content/Risk-Disclosures.htm



Deutsche Asset Management (Deutsche AM) represents the asset management activities

conducted by Deutsche Bank AG or any of its subsidiaries. Clients will be provided Deutsche Asset Management products or services by one or more legal entities that will be identified to clients pursuant to the contracts, agreements, offering materials or other documentation

relevant to such products or services. This document has been prepared without consideration of the investment needs, objectives or financial circumstances of any investor. Before making an investment decision, investors

need to consider, with or without the assistance of an investment adviser, whether the investments and strategies described or provided by Deutsche Bank, are appropriate, in light of their particular investment needs, objectives and financial circumstances. Furthermore, this

document is for information/discussion purposes only and does not constitute an offer, recommendation or solicitation to conclude a transaction and should not be treated as giving investment advice.

Deutsche Bank does not give tax or legal advice. Investors should seek advice from their own tax experts and lawyers, in considering investments and strategies suggested by Deutsche Bank. Investments with Deutsche Bank are not guaranteed, unless specified. Unless notified

to the contrary in a particular case, investment instruments are not insured by the Federal Deposit Insurance Corporation (“FDIC”) or any other governmental entity, and are not guaranteed by or obligations of Deutsche Bank AG or its affiliates.

Although information in this document has been obtained from sources believed to be reliable, we do not guarantee its accuracy, completeness or fairness, and it should not be relied upon as such. All opinions and estimates herein, including forecast returns, reflect our judgment on

the date of this report and are subject to change without notice and involve a number of assumptions which may not prove valid. Investments are subject to various risks, including market fluctuations, regulatory change,

counterparty risk, possible delays in repayment and loss of income and principal invested. The value of investments can fall as well as rise and you may not recover the amount originally invested at any point in time. Furthermore, substantial fluctuations of the value of the

investment are possible even over short periods of time. This publication contains forward looking statements. Forward looking statements include, but are not limited to assumptions, estimates, projections, opinions, models and hypothetical

performance analysis. The forward looking statements expressed constitute the author’s judgment as of the date of this material. Forward looking statements involve significant elements of subjective judgments and analyses and changes thereto and/or consideration of

different or additional factors could have a material impact on the results indicated. Therefore, actual results may vary, perhaps materially, from the results contained herein. No representation or warranty is made by Deutsche Bank as to the reasonableness or

completeness of such forward looking statements or to any other financial information contained herein. The terms of any investment will be exclusively subject to the detailed provisions, including risk considerations, contained in the Offering Documents. When making

an investment decision, you should rely on the final documentation relating to the transaction and not the summary contained herein. This document may not be reproduced or circulated without our written authority. The manner

of circulation and distribution of this document may be restricted by law or regulation in certain countries, including the United States. This document is not directed to, or intended for distribution to or use by, any person or entity who is a citizen or resident of or located in any

locality, state, country or other jurisdiction, including the United States, where such distribution, publication, availability or use would be contrary to law or regulation or which would subject Deutsche Bank to any registration or licensing requirement within such jurisdiction not currently

met within such jurisdiction. Persons into whose possession this document may come are required to inform themselves of, and to observe, such restrictions. Past performance is no guarantee of future results; nothing contained herein shall constitute

any representation or warranty as to future performance. Further information is available upon investor’s request.

This Document may not be distributed in Canada, Japan, the United States of America, or to any U.S. person.

© 2016 Deutsche Bank AG



Risk Warning Investments are subject to investment risk, including market fluctuations, regulatory change,

possible delays in repayment and loss of income and principal invested. The value of investments can fall as well as rise and you might not get back the amount originally invested at any point in time.

Investments in Foreign Countries - Such investments may be in countries that prove to be politically or economically unstable. Furthermore, in the case of investments in foreign securities or other assets, any fluctuations in currency exchange rates will affect the value of

the investments and any restrictions imposed to prevent capital flight may make it difficult or impossible to exchange or repatriate foreign currency. Foreign Exchange/Currency - Such transactions involve multiple risks, including currency risk

and settlement risk. Economic or financial instability, lack of timely or reliable financial information or unfavorable political or legal developments may substantially and permanently alter the conditions, terms, marketability or price of a foreign currency. Profits and losses in

transactions in foreign exchange will also be affected by fluctuations in currency where there is a need to convert the product's denomination(s) to another currency. Time zone differences may cause several hours to elapse between a payment being made in one currency and an

offsetting payment in another currency. Relevant movements in currencies during the settlement period may seriously erode potential profits or significantly increase any losses. High Yield Fixed Income Securities - Investing in high yield bonds, which tend to be more

volatile than investment grade fixed income securities, is speculative. These bonds are affected by interest rate changes and the creditworthiness of the issuers, and investing in high yield bonds poses additional credit risk, as well as greater risk of default.

Hedge Funds - An investment in hedge funds is speculative and involves a high degree of risk, and is suitable only for "Qualified Purchasers" as defined by the US Investment Company Act of 1940 and "Accredited Investors" as defined in Regulation D of the 1933 Securities Act. No

assurance can be given that a hedge fund's investment objective will be achieved, or that investors will receive a return of all or part of their investment. Commodities - The risk of loss in trading commodities can be substantial. The price of

commodities (e.g., raw industrial materials such as gold, copper and aluminium) may be subject to substantial fluctuations over short periods of time and may be affected by unpredicted international monetary and political policies. Additionally, valuations of commodities may be

susceptible to such adverse global economic, political or regulatory developments. Prospective investors must independently assess the appropriateness of an investment in commodities in light of their own financial condition and objectives. Not all affiliates or subsidiaries of Deutsche

Bank Group offer commodities or commodities-related products and services. Investment in private equity funds is speculative and involves significant risks including illiquidity, heightened potential for loss and lack of transparency. The environment for private

equity investments is increasingly volatile and competitive, and an investor should only invest in the fund if the investor can withstand a total loss. In light of the fact that there are restrictions on withdrawals, transfers and redemptions, and the Funds are not registered under the

securities laws of any jurisdictions, an investment in the funds will be illiquid. Investors should be prepared to bear the financial risks of their investments for an indefinite period of time. Investment in real estate may be or become nonperforming after acquisition for a wide variety

of reasons. Nonperforming real estate investment may require substantial workout negotiations and/ or restructuring. Environmental liabilities may pose a risk such that the owner or operator of real property may

become liable for the costs of removal or remediation of certain hazardous substances released on, about, under, or in its property. Additionally, to the extent real estate investments are made in foreign countries, such countries may prove to be politically or economically unstable. Finally,

exposure to fluctuations in currency exchange rates may affect the value of a real estate investment. Structured solutions are not suitable for all investors due to potential illiquidity, optionality, time

to redemption, and the payoff profile of the strategy. We or our affiliates or persons associated with us or such affiliates may: maintain a long or short position in securities referred to herein, or in related futures or options, purchase or sell, make a market in, or engage in any other

transaction involving such securities, and earn brokerage or other compensation. Calculations of returns on the instruments may be linked to a referenced index or interest rate. In such cases, the investments may not be suitable for persons unfamiliar with such index or interest rates, or

unwilling or unable to bear the risks associated with the transaction. Products denominated in a currency, other than the investor’s home currency, will be subject to changes in exchange rates, which may have an adverse effect on the value, price or income return of the products.

These products may not be readily realizable investments and are not traded on any regulated market.



Notes



Notes db X-trackers_2083

Deutsche Asset Management represents the asset management activities conducted by Deutsche Bank AG or any of its subsidiaries. Clients will be provided Deutsche Asset Management products or services by one or more legal entities that will be identified to clients pursuant to the contracts, agreements, offering materials or other documentation relevant to such products or services. © 2016 Deutsche Asset Management. All rights reserved.