White Paper September 2015

Transcription

1 White Paper September 2015 How to invest in Emerging Market Debt Combining segments for an efficient allocation Authored by: Xavier Baraton, Global CIO Fixed Income, Regional CIO North America Maria-Laura Hartpence, Head of Fixed Income Quantitative Research For professional clients only

2 How to invest in Emerging Market Debt Combining segments for an efficient allocation Executive summary EMD: from niche to mainstream, but allocating hasn t necessarily become easier Emerging markets remain volatile, with positives and negatives Understanding the factors behind EMD returns Using economic and financial factors to evaluate prospective risks and returns Factors driving EMD Finding the optimal EMD combination: Combining EMD segments for more effective allocations Which segments and in what proportions? Scenario 1: base scenario Scenario 2: adding FX risk and higher expected premiums on hard currency corporate debt Implications of allocating to EMD within a broader context Conclusion: finding the balance Authors Important information Appendices References Endnotes Page 3 Page 4 Page 6 Page 9 Page 9 Page 10 Page 14 Page 14 Page 15 Page 16 Page 17 Page 18 Page 20 Page 21 Page 22 Page 25 Page 28 Page 29 Our study shows that today, an investor should combine debt segments to achieve more robust allocations, to benefit from diversification and the variety of the fixed income components. These have very diverse dynamics made possible by the advent of more complex markets, in particular on EMD, which can no longer be considered as a single homogeneous asset class. Maria-Laura Hartpence Head of Fixed Income Quantitative Research 2

3 Executive summary The facts Key findings EMD is no longer an asset class investors can afford to neglect. As a whole, it has grown by 500% over the last decade. It now represents over 6% of the global fixed income market, with an estimated market capitalisation of USD3 to USD6 trillion, depending on the degree of liquidity of the indices used. Several segments now exist within EMD, making EMD risk premiums and beta more complex to define, but also giving investors the opportunity to build more effective and resilient exposures. Corporate hard currency debt has experienced the most dramatic growth, with a market value 15 times higher in 2014 than it was in 2003, and with more than 500 issuers and 1000 issues. In the most liquid indices, it is now larger than sovereign hard currency debt. Sovereign local currency debt has increased by around 600% in terms of market value over the same period and has also surpassed EM hard currency sovereigns. Investment grade bonds make up over 75% of the EMD universe today, against less than 30% ten years ago. EM hard and local currency sovereign investment grade debt respectively yield over 4% and 5%, compared to less than 2% for developed markets investment grade debt. Illiquidity is a common issue during periods of global turmoil, resulting in sell-offs of the EMD asset class, but these are usually followed by fast recoveries thanks to healthy fundamentals. Overall, EMD investments provide quality and yield, but also come with risks (variable liquidity, idiosyncratic risk) and volatility that allocations should aim to mitigate. Our analysis shows that EMD sectors (hard currency, local currency, corporate) are highly exposed to different systematic (common) risk factors which are not diversifiable and thus earn a long-term premium. The high exposure of EM markets to global interest rate risk, equity risk and credit risk ensures solid prospective premiums and Sharpe ratios over the long run. The fact that the different EMD sectors have different levels of exposure to the above factors, as well as the presence of idiosyncratic risk within each sector, suggests that a combined portfolio should deliver a higher Sharpe ratio than each of the sectors taken individually. In particular, we found that for a portfolio purely invested in EM sovereign hard currency debt, the benefits of adding EM corporate hard currency debt are twofold: - Both assets are highly but not perfectly correlated, so adding corporate bonds has some diversification benefits. - The prospective Sharpe ratio of EM corporate hard currency debt should be higher than that of EM sovereign hard currency debt given the corporate sector s high exposure not only to the systematic interest rate factor but also and principally to the systematic credit factor. The expected continuation of the long-term economic development trend in EM economies suggests EM FX will deliver a positive premium going forward. Therefore, unhedged local currency debt exposure can add EM FX risk and its potential expected premium over the long run. Investors with relatively low exposure to EMD would benefit from increased allocations to this asset class. For instance, under our assumptions, a portfolio invested 55% in equity and 45% in fixed income with an expected Sharpe ratio of 0.30 could raise its risk adjusted returns by nearly 10% by substituting EM bonds for equities. 3

4 EMD: from niche to mainstream But allocating hasn t necessarily become easier This section aims to demonstrate that emerging market fundamentals have improved dramatically since the turn of the century, to the extent that investors can no longer afford to miss out on this asset class. Yet, taking advantage of it presents challenges ( concentration, idiosyncratic, evolving market structure) that require investors to pay particular attention to how exposures are built. Birth of a new asset class EMD was once seen as a niche space in the larger fixed income universe. Ten years ago investors typically only had a very small if any allocation to the asset class. However, as the market grew and matured, more and more investors started to acknowledge the benefits of diversifying into EMD within their strategic allocations and by 2013 EMD accounted for 5 to 6% of the fixed income allocation of the global investment industry. 1 (see end notes) The increasing interest in EMD is not surprising considering that EM sovereign hard currency debt and EM sovereign local currency debt deliver an average yield of around 6%, with even their investment grade segments yielding more than 4% and 5% respectively. This compares to less than 2% for developed market sovereign investment grade bonds. 2 Types of EMD EMD is generally defined by the currency of issue and type of issuer. In other words, EMD can be either hard currency (generally USD denominated) or local currency and is issued by sovereigns, quasisovereigns or corporates. Compared to the USD47.1 trillion global developed nominal bond market 3, EMD is still relatively small at just USD3.4 trillion when considering the most liquid indices. This is split between USD1.5 trillion in hard currency and USD1.9 trillion in local currency. Issuer breakdown for hard currency is USD671 billion for sovereigns and quasi-sovereigns and USD806 billion for corporates. For local currency debt, issuer breakdown is USD1.7 trillion for sovereigns and USD210 billion for corporates. 4 Exhibit 1: EMD outstanding in USD billion EMERGING MARKET DEBT: 3361 Hard EMD: 1477 EM local currency debt: 1884 Sovereign: 671 Corporate: 806 Sovereign: 1674 Corporate: 210 Source: JP Morgan, BofA Merrill Lynch, as of 31 August All segments have grown, but local currency and corporate debt more particularly Emerging sovereign local currency and emerging corporate hard currency debt have exploded over the last decade. Over the past ten years the market value of EMD as a whole has grown by close to 500% against 132% for developed market bonds. 5 A large portion of EMD growth has come from local currency debt, which has increased by over 650% since 2003 to USD1.9 trillion today. In comparison, hard currency debt grew by 380% - from USD305 billion to USD1.5 trillion - over the same period. The corporate market as a whole is still smaller than the sovereign market, but hard currency corporate debt outstanding is now 15 times higher than in 2003 and larger than hard currency sovereign debt, the latter having grown by around 160% over the period. Exhibit 2 Trend in EMD since December 2003 (in USD billion) Hard Currency - Sovereign & Quasi Local Currency - Sovereign Hard Currency - Corporate Local Currency - Corporate Source: JP Morgan, BofA Merrill Lynch, as of 31 December The enhancement of fundamentals has significantly improved the credit quality of EMD While credit fundamentals have certainly deteriorated in the U.S., euro-zone, and Japan over the last decade (reflected by the multiple credit downgrades issued by the leading rating agencies), the improvement in emerging markets fundamentals has led to a significant enhancement of their credit quality. More than 65% of EM hard currency sovereign debt is now rated investment grade, versus around 30% in This figure jumps to over 80% for local currency debt. 6 4

5 Banking Energy Basic Industry Real Estate Utility Telecomms Services Non-Cyclical Capital Goods Financials Cyclical Techno & Electro Media Automotive Healthcare Insurance Brazil China Russia Mexico South Korea Turkey Arab Emirats Indonesia Hong Kong India Venezuela Qatar Colombia Chile Philipines Peru South Africa Singapore Cayman Islands Argentina China has substantially increased its market share on EMD Another prominent trend in EMD over the last ten years has had to do with the country composition of the investment universe. Within EM hard currency debt, one of the most notable country allocation increases has been China, which represented just 1.3% in 2003 and became the second largest country constituent after Brazil, at 9.4%, in Similarly, when looking specifically at the corporate universe over the same period, we can see the share of Chinese issuers progressing from just 0.15% to one of the largest at 14.5%. 7 Exhibit 3 EM Hard Currency Debt Sector Market Capitalisation Shares in % 16% 12% 8% 4% 0% as have the banking and real estate sectors With regards to economic sector shifts in EM hard currency corporate debt between 2003 and 2014, we can observe a significant rise of the Banking (from 14% to almost 25%) and Real Estate sectors (from 1% to 7%). These are trends that we would expect to see in countries that are moving along the path of economic development. EMD indices are developing but tend to track segments... and their flaws The EMD investable universe including corporate bonds is covered by three main index providers today: JP Morgan, Barclays and Bank of America Merrill Lynch. While the three providers index methodologies are relatively similar, there are differences which affect the characteristics of each index. For example, the market capitalisation and effective duration vary between the different index providers, because of the different inclusion requirements set by each firm, such as the minimum face amount outstanding or the minimum maturity. These factors can explain some of the differences in effective duration from one index to the other, since indices with higher maturity requirements will have longer dated bonds and thus a higher duration. For more details on the indices and their characteristics refer to appendix Source: BofA Merrill Lynch, data as of December 2003 and September Exhibit 4 EM Corporate Hard Currency Debt Sector Market Capitalisation Shares in % 30% 15% Implications for investors: The evolution of emerging markets over the last 15 years, both in terms of fundamentals and credit quality has made the asset class an almost compulsory allocation for investors. Nevertheless, market indices remain concentrated, with a modest number of issuers. In addition, the rapid growth and the dynamics of the asset class are a double-edged sword, as they make it difficult to capture its beta. 0% Source: BofA Merrill Lynch, data as of December 2003 and September

6 Emerging markets remain volatile With positives and negatives While the fundamentals of emerging markets are undeniably improving, it does not follow that all risks and volatility have disappeared. However, these risks are also good news because of their associated risk premiums, meaning there are still opportunities for investors to take. Liquidity issues remain and can cause significant drawdowns While the size and diversity of EMD has grown over the last 10 years, liquidity has not necessarily kept pace. Exhibit 5 compares the market capitalisation of EMD according to selected indices and annual trading volumes reported by the Emerging Markets Trade Association (EMTA) in their annual surveys. 8 We can see that EMD has been growing steadily while trading volumes have been stagnating since For instance, while the market capitalisation of EMD more than doubled from USD1.3 trillion in 2007 to USD2.7 trillion in 2013, trading volumes decreased slightly over the same period, to an annual level of USD5.6 trillion. The Institute of International Finance (IIF) argues that the near-zero rates policy has led to the advent of a very liquid primary market, which has thus been able to absorb the record amount of emerging bond issues, while recent regulatory changes and requirements negatively impacted secondary markets. Therefore, in spite of the improvement in EM fundamentals, periods of financial tension have resulted in significant drawdowns of this asset class due to the lack of liquidity, but thanks to the same improved fundamentals these drawdowns have been followed by rapid recoveries. Exhibit 5 - Emerging Market Debt - Market Capitalisation and Trading Volumes in USD billion EMD outstanding in June (lhs) Annual Debt Trading Volume (rhs) Source: JP Morgan, BofA Merrill Lynch, EMTA, as of 31 December Over the past decade, the financial depth of emerging markets has also increased, albeit unevenly, although it remains thinner than in advanced economies. 9 At the same time, the share of emerging market financial assets held by developed investors has increased, exposing more emerging markets to global financial shocks and producing liquidity stress in the asset class. However, it has been argued that those countries with better fundamentals are more resilient to the negative effects of global factors. 10 But improved fundamentals have reduced volatility As mentioned above, the improvement in emerging economies fundamentals has led to a significant enhancement of credit quality, and EM financial markets have developed dramatically. EM corporate debt default rates have shown to be similar if not lower than those observed in developed markets: for instance, the average annual corporate speculative default rate in emerging markets for the period covering reached 3.5% against 4.6% for the US and 3.8% for Europe, while recovery rates were comparable across regions. 11 Exhibit 6 shows the 52-week rolling annualised volatility 12 of the JP Morgan EMBI Global index, one of the most widely followed indices for EM sovereign hard currency debt. Volatility went from 15% in January 1994 to 6% in December Over the same period, the effective duration of the index increased steadily, from around 4 to 7. When calculating the rolling volatility of the index with an equivalent duration of 7, results show the volatility went from 26% to 6% over the period.

7 Exhibit 6-52-week rolling volatility of the JP Morgan EMBI Global index Exhibit 7: weekly returns, empirical and theoretical frequency of the JP Morgan EMBI Global index 50% 9 40% 40% 30% 20% 10% % 30% 25% 20% 15% 10% 5% 0% 12/ / / / % -7%-6%-5%-4%-3%-2%-1% 0% 1% 2% 3% 4% 5% 6% 7% 1-y ann. rolling index vol. (lhs) 1-y ann. rolling index vol. in constant effective duration (lhs) Effective duration of the index (rhs) Source: JP Morgan, HSBC Global Asset Management calculations, data as of 31/12/1993 to 12/12/2014. Emprical frequency Theoretical frequency for a Normal distribution Source: JP Morgan, HSBC Global Asset Management calculations, data as of 31/12/1993 to 12/12/2014 And although some volatility clearly remains Emerging markets are still subject to sell-offs and significant drawdowns during periods when investor risk aversion rises. Looking at the JP Morgan EMBI Global weekly returns from January 1994 to November 2014 (Exhibit 7), we can see that the frequency of returns lower or higher than 3 standard deviations (5%) was 1.5% against a theoretical value of 0 for a normal distribution. 13 Exhibit 7 also shows the empirical frequency of these weekly returns are quite different from the theoretical frequency implied by a normal distribution. Drawdowns have been less severe Exhibits 8.a and 8.b on the next page highlight the most significant drawdowns in EM sovereign hard currency debt since 1994 as measured by the JP Morgan EMBI Global index. These show that, overall, recent drawdowns have been less severe than previous ones. It is interesting to compare the behaviour of the JP Morgan EMBI Global total return index against the MSCI World index. Since the beginning of the century the latter experienced two major drawdowns of -46% and - 27% (the tech-bubble and the credit crisis), with return to peak lengths of 71 and 67 months respectively. Meanwhile the average annualised returns of the JPM EMBI Global between December of 1993 and November 2014 reached over 9% per annum against 7.6% for the MSCI World s total returns (in USD, unhedged). Exhibit 9 presents the major drawdowns in hard currency corporate debt as measured by the JP Morgan CEMBI Broad diversified index. The most severe drawdown occurred during the Credit Crisis, and still the recovery was relatively rapid. Using the CDS Bond Spread basis for EM sovereign debt as a liquidity indicator, Darolles et al. (2013) noted that most periods with a significant fall in EM prices since 2007 have been characterised by a simultaneous tensing of liquidity indicators, and not because of any particular correlation of economic fundamentals. Significant impact of idiosyncratic risk in EMD Exhibit 8b also shows that country risk is an important source of volatility in EMD. This risk is amplified by the concentrated nature of indices (compared to credit indices in developed markets for instance), making default loss or contagion more detrimental. This is a particular challenge for investors aiming at building an allocation that delivers robust beta. 7

8 Exhibit 8.a Major drawdowns in emerging markets JP Morgan EMBI Global total return index Drawdown periods Year drawdown starts Peak to trough index return Source: JP Morgan, HSBC Global Asset Management calculations, December Number of months from through to recovery date Number of months from peak at the beginning of drawdown until recovery date Mexican crisis % 9 23 Asian crisis % 5 6 Russia default % Turkey crisis % 2 5 Argentine crisis % 3 8 Argentine crisis, Brazilian elections % 4 7 US interest rate rise % 3 5 Monetary policy tightening % 2 6 Concerns about Bear Sterns and subprimes % 2 5 Great Financial Crisis % 8 13 Sovereign Debt Crisis % 4 7 Start of Quantitative Easing tapering talks % 9 13 Russian crisis, Oil price collapse % Drawdown not ended Dec 2014 Exhibit 8.b Major drawdowns in emerging markets JP Morgan EMBI Global total return index 800 0% 700-5% % % % % % 0-35% 12/ / / / /2013 Implications for investors Emerging market debt can still provide liquidity, credit and volatility risk premiums but the challenge for investors is in how to capture them. In practical terms, what investors should aim for is to define and implement an appropriate allocation allowing them to access emerging market growth while limiting the attendant idiosyncratic risk. Major drawdowns (rhs) JP Morgan EMBI Global - Total return index (lhs) Major drawdowns: beginning of drawdown until recovery date Source: JP Morgan, HSBC Global Asset Management calculations, December Exhibit 9 Major drawdowns in EM Hard Currency Corporate debt, as measured by the JP Morgan CEMBI Broad diversified index Drawdown periods Year drawdown starts Peak to trough index return Source: JP Morgan, HSBC Global Asset Management calculations, December Number of months from trough to recovery date Number of months from the peak at the beginning of drawdown until recovery date US interest rate rise % 3 5 Great Financial Crisis % 9 14 Sovereign Debt Crisis % 5 7 Start of Quantitative Easing tapering talks % 7 7 8

9 Understanding the factors behind EMD returns Using economic and financial factors It is difficult to evaluate prospective risks and returns on a historical basis In the previous section, we observed how, over the last ten years, emerging market fixed income has been garnering interest among investors. In this chapter, we will analyse how best to allocate these assets across sectors. The challenge is that the history of these asset classes is relatively short. Market indices for EM sovereign local currency debt and EM corporate hard currency debt have only been in existence since 2002: in terms of the number of observable economic cycles, 14 years is too short a time to infer robust conclusions. Inevitably, with an insufficient track record and an asset class still developing, this raises questions about prospective risks and returns. In addition, due to the diversity of risks attached to the EMD market, it becomes difficult to efficiently assess its beta. And fundamentals have been constantly evolving More importantly, we have noted that during the structural evolution of the market discussed in the first section of this paper, emerging market indices were continuously subject to changes, in everything from the quality of indices to issuer numbers and diversification, and sector and country weights. Additionally, as a result of all these developments Emerging Market Debt has seen variations in its volatility levels though these have generally decreased while liquidity has not necessarily improved, particularly during periods of turmoil. One way to approach this issue is to use economic and financial factors to evaluate prospective risks and returns Therefore, in order to address our questions on prospective risks and returns, we must make certain assumptions, which we will base on modern economics and financial theory. In order to limit the potential impact of current and changing valuations, we use a factor-based approach, which does not take into account expected returns. This approach makes our analysis less time-sensitive but investors will of course have to take market valuations into consideration when building or reviewing their allocation. We will start by shedding some light on the economic and financial factors that influence emerging fixed income markets. This will enable us to evaluate the risks to which these assets are exposed and hence what we should expect in terms of premiums. Indeed, several economic and financial factors affect the entire market, resulting in non-diversifiable risk, and financial theory tells us that a financial asset delivers a positive premium over an equivalent riskfree asset when exposed to risk factors that cannot be diversified. EM assets are exposed to a variety of factors including economic growth, inflation, exchange rates and several others. These factors fluctuate with the economic and financial cycle and explain asset return movements. The impact of investors risk aversion The other element that influences returns is investors risk aversion. Risk-averse investors, who already require a premium on average expected losses and fear the added uncertainty, may require even higher risk premiums in difficult times, resulting in strong return fluctuations. Lack of market liquidity could compound these negative effects as investors could be forced to sell the bond if they needed liquidity, even those not especially fearful of future economic uncertainty. Shocks or periods of financial turmoil can thus translate into abnormally high market volatility with extremely low asset returns (tail risk) even though the resulting fall in return can be relatively short-lived. Investors require a premium for risk factors that cannot be diversified away The economic and financial factors mentioned above underpin well-known global risk factor categories such as interest rate risk, credit risk, liquidity risk, equity risk, foreign exchange (FX) risk or volatility risk (and the list is not exhaustive). We here use the word global to express the notion that these factors lie at the origin of correlations between assets. 9

10 Factors driving EMD Current and/or expected changes in economic activity typically lag behind interest rate, credit and equity risk factor movements. Inflation on the other hand should be more in step with the interest rate and FX factors. Higher volatility tends to be linked to the credit and equity factors, although we will aim to identify a volatility factor per-se to detect periods of significant increases in market volatility. What needs to be taken into account is that economic activity and inflation tend to be correlated across markets. Correlation between default rates rises in periods of poor economic activity while at the same time companies recovery rates tend to fall. Liquidity crises have often showed to be systemic. Therefore, to the extent that these factors cannot be diversified, they generate correlation across returns, a phenomenon that is amplified by common investors perception of risk. We retained these variables as proxies of the economic and financial factors for our analysis. Some of the variables, typically asset indices, can be considered as direct representatives of the global risk factors (for example the Global Government bond index is a very good representative for the global interest rate factor). Other variables constitute a partial explanation for a factor. For instance, oil prices partially explain inflation, but in order to understand inflation fully we also need to look at other variables such as other commodity prices, wage movements, market competition, etc. In all cases these variables, which we will now call factor proxies, are expected to be highly correlated with the economic and financial factors they represent. Most common global risk factors: Indeed, the majority of investors are risk-averse and require compensation above that of risk-free assets to accept being exposed through their investments to risk factors that cannot be diversified. This results in a remuneration of these factors in the long run. 14 All these global risk factors can therefore be explained by economic fundamentals and by investors perception of risk. Crucially, in most instances, the latter is the more important determinant in global factor movements. Interest rate risk factor Credit risk factor Liquidity risk factor Equity risk factor FX risk factor Volatility risk factor Evaluating the sensitivity of EMD assets to these factors To estimate the sensitivity of EMD assets to these economic and financial factors, we performed a linear regression of the excess returns (relative to US T-Bill returns) of the following EM indices on the variables described on the next page, in Exhibit 10: EMBI Global (EM sovereign + quasi-sovereign debt in USD hard currency) CEMBI Broad Diversified (EM corporate debt in USD hard currency) GBI-EM Global Diversified in USD unhedged (EM sovereign debt in local currency unhedged in USD) GBI-EM Global Diversified hedged in USD (EM sovereign debt in local currency hedged in USD) Implications for investors In constantly changing emerging market conditions, a backward-looking approach cannot be sufficient, particularly given the lack of available historical data. Implementing a forward-looking approach therefore seems to be the only viable option to determine optimal allocations. A thorough understanding of the different underlying factors and their dynamic thus becomes a crucial factor for success. 10

11 Exhibit 10: Description of the variables taken as proxies of economic/financial risks Variables: Proxies of the economic/financial factors Economic/financial factor the proxy represents and/or explains Global Factor to which the economic/financial factor is mainly associated Source Oil Price, monthly relative change Inflation Interest Rate risk factor Crude Oil-WTI Spot Cushing U$/BBL Commodities Index, monthly excess return Inflation, economic activity in commodity producers markets Interest Rate risk factor Credit risk factor for EMs? S&P GSCI Commodity total return index OECD CPI, monthly relative change Inflation Interest Rate risk factor OECD CPI All Items index OECD Industrial Production, monthly relative change Global economic activity Interest Rate risk factor Credit risk factor OECD Production Total Industry excluding Construction index Emerging Markets Industrial Production, monthly Economic activity in Emerging relative change Markets US dollar, monthly relative change Foreign Exchange (FX) FX risk factor Government Bonds in Developed Markets, monthly excess return Interest Rates Based on the average of industrial Interest rate risk factor (in EMs?) production growth in Brazil, India and Russian Federation, source OECD Interest Rate risk factor US Trade Weighted Value of US Dollar against Major Currencies index BofA Merrill Lynch Global Government total return index hedged in USD Illiquidity indicator: interbank rate minus T-Bill rate Liquidity Liquidity risk factor US Overnight Interbank rate T-Bill rate Market Volatility, tailed returns. Implied volatility of S&P, monthly relative change Considered as an indicator of general market risk aversion Volatility risk factor CBOE SPX Volatility VIX index Global Equities, monthly excess return Equity risk Equity risk factor MSCI total return index hedged in USD US Corporate BBB relative to AAA bonds, difference of monthly excess returns Spread changes Downgrades Credit risk factor Liquidity risk factor The BofA Merrill Lynch AAA and BBB US Corporate total return Indices Equity risk Equity risk factor US High Yield, monthly excess return Defaults Spread changes Credit risk factor Liquidity risk factor BofA Merrill Lynch US High Yield Master II total return Index Downgrades EM Money Market local returns x FX exchange rate change relative to the US dollars, monthly excess return Emerging Market risk EM FX risk EM risk factor JP Morgan ELMI total return index in USD Note: Index excess returns are index returns minus US T-Bill returns. Source: HSBC Global Asset Management, for illustrative purposes only, December Disentangling the impact of each factor It is not an easy task to disentangle exactly to what extent assets are exposed to each of the economic and financial risk factors. Take corporate bonds for example: during an economic or financial crisis spreads will probably widen, but to what extent will this be due to changes in expected bond defaults and/or downgrades or to liquidity issues? Will the widening of spreads result from an objective perception of the default risks implied by the crisis, or rather from a heightened level of risk aversion causing an exaggerated fluctuation in price? Similarly, if we consider a fall in equity markets during a period of uncertainty such as the inflexion of the economic cycle, would prices be falling as a result of the changes of the current and expected business conditions, or would they be stemming from a high level of risk aversion (as seen by fluctuations on the VIX)? 11

12 To obtain a clearer idea of the sensitivity of assets with regard to the various factors considered, we calculated "adjusted" or uncorrelated proxies, 15 whose returns represent pure factors". We used the following steps: 1. We took the oil-price return itself as the first adjusted proxy. 2. We regressed 16 the excess return of the commodities index on the oil return. The residual of this regression defined the commodity proxy ex-oil, constituting the adjusted commodity proxy. 3. We regressed OECD monthly inflation on the oil factor (the oil return) and the adjusted commodity proxy (the commodity index return ex-oil). The residual of this second regression produced an adjusted global inflation proxy, representing global inflation explained by effects other than oil and commodities ex-oil. 4. We continued this process following the order of factors as presented in the first column of Exhibit 10. It is interesting to note that the adjusted global bond proxy i.e. the developed government bond market index excess returns adjusted to remove inflation and activity proxies can give a fair representation of the portion of interest rate movements due to fluctuations in the required risk premium on government bond markets. In the same vein, the adjusted high yield proxy is a good representation of pure credit risk (assuming the US High Yield market risk is a good representation of the Global High Yield market risk) adjusted to remove equity risk. Exhibit 11 presents the estimated sensitivities of the EMD indices relative to these adjusted factor proxies. The analysis runs from January 2002 to August 2014 for Hard Currency debt, while it starts in January 2003 for Local Currency debt. 17 For Sovereign hard currency debt, we also ran estimates from January 1994 to December 2001, as per the available data for this asset class. The results presented below are based on a sample excluding those months that could be considered outliers, since the regression analysis can be very sensitive to such events and our goal was to obtain sensitivities illustrating the exposure of EMD markets to the factors over the long-run Exhibit 11 Estimated Sensitivities of EMD indices to Selected Factors - Explained variables (Indices' Excess Returns) EM Hard Currency Sovereign Debt in USD EM Hard Currency Sovereign Debt in USD Source: HSBC Global Asset Management, December T-stats are between brackets. *Significant at a 5% level, **Significant at a 1% level 12 EM Hard Currency Corporate Debt in USD EM Local Currency Sovereign Debt in USD Unhedged EM Local Currency Sovereign Debt hedged in USD Adjusted factor proxies (Intercept) (2.1)** (7.3)** (9.3)** (10.9)** (5.9)** Oil (0.2) (0.8) (5.8)** (6.8)** (-1.1) Commodities (1.5) (2.6)** (1.7) (3.1)** (0.3) OECD Inflation (1.7)* (-1.9)* (-3.8)** (-0.7) (-2.9)** OECD Eco Activitiy (-0.3) (-0.5) (1.6) (-0.7) (-0.4) EM Eco Activity (1.1) (2.4)** (4.4)** (5.2)** (1.9) US dollar (-0.2) (-3.3)** (-6.6)** (-11.9)** (-3.6)** Global Gvt Bonds (4.2)** (6.3)** (10.6)** (6.2)** (7.1)** Illiquidity (2.7)** (-2.4)** (-4.4)** (-2.4)* (-2.9)** VIX (-5.2)** (-8.8)** (-9.9)** (-11.2)** (-4.4)** Global Equities (3.4)** (7.9)** (9.3)** (10.9)** (4.5)** IG BBB minus AAA (-0.4) (4.0)** (10.1)** (4.6)** (1.9) High Yield (1.4) (5.4)** (9.3)** (3.0)** (1.4) EM Money Market (4.4)** (3.4)** (2.1)* (19.4)** (5.7)** Adjusted R 2 45% 66% 82% 88% 55% R 2 53% 69% 84% 90% 60%

13 Analysis of the main sensitivity factors Global inflation and a strong dollar appear to have a negative influence on EMD returns, while the impact of EM economic activity is positive Adjusted inflation on the other hand appears to have a significant negative impact on returns: a 1% change in prices translates to a fall of both the hard currency and USD hedged sovereign local currency indices of nearly 1%. While the relationship between OECD economic activity and EMD index returns is not statistically significant, we do see a positive and statistically significant relationship between our proxy of EM economic activity and the EMD index returns. However, as we will we see further on, emerging indices are sensitive to the MSCI equity index fluctuations, which include fluctuations in expectations about future global activity. The effect is relatively strong as a 1% rise of our EM industrial production proxy translates into positive excess returns of between 0.3% and 0.5%. The impact of a strengthening dollar is systematically negative for all indices and for all periods analysed. Could this be due to the fact that periods of poor liquidity stemming from financial crises and a flight to quality have always translated into a rise of the dollar and a fall of EM markets? Global interest rate, credit, equity and liquidity risks are important determinants of EMD returns The estimated coefficients of index excess returns on the adjusted Global Government Bond index excess returns are positive, strong and statistically significant. Noting that this adjusted factor proxy is a perfect representative of global interest rate risk, these results show that EM indices are definitely sensitive to that factor. This was to be expected for EM hard currency debt by construction, but comes as more of a surprise for EM local currency debt. Illiquidity has a statistically significant negative impact on all indices (except for EM sovereign hard currency debt before 2002). For example a differential of the interbank rate/t-bill of 1% translates in a fall of the indices ranging from 8 to 12%. The VIX variable is widely known as being a proxy for global risk aversion and indeed we see that indices are significantly negative in correlation with this variable. EM fixed income markets are exposed to global equity and credit risk as shown by the statistically significant EM indices positive sensitivities estimated relative to the adjusted excess returns of the MSCI World index, the US BBB relative to triple AAA and the US High Yield indices. These indices adjusted to remove the economic variables effects can be considered not only as excellent representatives of equity and credit market perceptions of risk, but also reflect expectations of future global economic activity, something not accounted for by the economic variables used in our regressions. The question of causality for commodities Overall for the period running from 2002 (2003 for EM sovereign local currency debt) to 2014, the estimated coefficients are significant in most cases and with the sign we were expecting. As is usual in this kind of exercise there is no way to tell if the estimated coefficients result from a correlation between EMD indices and the adjusted proxy factors for the period, or from a true relationship of causality between the factors and the EMD indices. For instance there is a small positive impact of the price of oil on the indices (whereby if oil prices double, corporate debt rises by 5%), the source of which is not completely clear. While some emerging countries are major oil exporters (e.g. Russia, Venezuela or Mexico) and thus benefit from a rise in the price of oil, other emerging economies would suffer from this movement. Oil may in fact be a coincident indicator: when oil prices were rising (or falling), markets were rising (or falling) too. As a matter of fact, we observed this for the entire period of analysis except for the 2008 Credit Crisis. The estimated coefficient on the adjusted commodity proxy is also positive. We could argue that, as exporters of commodities, some emerging economies benefit from rises in commodity prices, but at the same time we know that the growth in emerging economies was itself at the origin of the rise in commodity prices including oil. Regardless, the positive impact measured is relatively small. 13

14 Finding the optimal EMD combination Combining segments for more effective allocations Exhibit 12 shows the variance breakdown by factor across the four indices examined in this analysis. The interest is that it takes into account the volatility of the factors, as well as their contribution to the overall volatility of the indices. 20 Exhibit 12 Risk Breakdown by Factor across EM Indices 100% 80% 60% 40% 20% 0% 47% 11% 1% 7% 16% 1% 4% 10% 10% 3% 4% 4% EM HC Sovereign Debt in USD % 3% 11% 16% 20% EM HC Sovereign Debt in USD Economic Activity and Inflation Gvt Bonds Risk Volatility Risk Credit Risk Unexplained 16% 25% 11% 13% 1% 15% 4% 10% 8% Source: HSBC Global Asset Management, December We also ran regressions for the three sub-periods of January 2002 (2003 for EM local currency debt)- September 2008, July 2007-March 2009 and April August The general pattern of risk exposures remains broadly the same except for interest rate exposure, which rises for both EM sovereign hard and local currency hedged in USD, while it falls for EM corporate hard currency debt after It is important to remember that, in contrast to a market index where the underlying countries are relatively stable advanced economies, EM indices represent an evolving world: the sensitivities observed here could continue to exhibit significant changes in the future. 3% 6% EM HC Corporate Debt in USD % 37% 3% 11% 12% 14% EM LC Sovereign Debt in USD Unhedged % 12% 2% 8% 7% 3% 18% 5% 5% EM LC Sovereign Debt hedged in USD US dollar Illiquidity Equity Risk Emerging Market Risk (inc. FX) Combining segments can help investors take advantage of this As stated earlier, the economic/financial factors discussed above are at the origin of non-diversifiable risk, and should thus pay a positive premium over the risk-free asset. A portfolio looking to recreate any of the EM indices presented above should therefore deliver a Sharpe ratio reflecting their exposure to these non-diversifiable risk factors, in line with the sensitivities we have estimated. Such a portfolio will also exhibit additional risks (corresponding to the unexplained variance in our regression analysis above), although these could potentially be diversified away through adequate investments. As a result they would not be remunerated over the long run but would nonetheless add volatility to the portfolio. The indices analysed here all have varying levels of exposure to the different factors and all of them present some index-specific risk. Portfolios allocated so as to recreate combinations of indices would therefore not only improve risk diversification but also reduce the volatility stemming from non-remunerated specific risk, thus improving the Sharpe ratio of the portfolio. The next question is finding how to best balance the allocation between the EMD segments. From a broader perspective, we also decided to evaluate the contribution of EMD to a global bond/equity portfolio. Starting with a portfolio 100% invested in EM sovereign hard debt, we looked at how the introduction of other EM fixed income asset classes, namely EM corporate hard currency debt and EM sovereign local currency debt, would impact the portfolio s performance. We also investigated portfolios invested in EM local currency corporate debt, though this asset class is extremely young and would not be adequate for those investors who need a minimum level of liquidity. 14

15 Which segments and in what proportions? Finally, we studied the impact of including these assets classes in a broader portfolio, taking as a reference a global portfolio with a strong weight on US fixed income assets. We used well-known global indices 22 to represent each asset class. To assess future asset risks and returns of these portfolios, we employed a proprietary risk model that provides an estimate of future market correlations between the main asset categories and of their volatility. The model statistically identified three systematic risk factors that explained most of the covariance of returns; an interest rate risk factor, a credit/liquidity/equity risk factor and a US dollar vs. euro risk factor. These findings were perfectly consistent with the analysis in section 2 on factor returns. In other words, our risk model statistically identified three main systematic risk factors which actually encompass the adjusted economic and financial factors discussed in section We then evaluated expected long-term asset premiums. Based on modern financial theory, we made the assumption that fixed income long-term asset premiums are determined by the assets exposure to risk factors that cannot be diversified away namely the main systematic risk factors identified by our risk model and by the expected premiums on these factors. We made the further assumption that the interest rate and credit/equity/liquidity factors would be equally remunerated while the dollar/euro factor would not. To assess the expected market correlations between the main asset categories as well as their volatilities, we used a proprietary risk model which uses a factorial approach. 24 This has the advantage of reducing the number of estimated parameters and is intuitively appealing from an economic and financial point of view. It is based on the idea that asset excess returns can be defined as a combination (given by the estimated Betas of the assets relative to the factors) of the excess returns of common or systematic risk factors (which explain the covariance between excess returns) and specific or idiosyncratic risk factors, i.e. factors that explain excess returns specifically for each asset. The systematic risk factors can be represented by asset portfolios, or proxies that have unit sensitivity to these factors. 25 The model identifies three main common or systematic risk factors that explain the correlation between asset returns. The three common or systematic factors of risk identified in this analysis are: Interest rate risk factor Credit/liquidity/equity risk factor Dollar/euro bloc risk factor If we assume that the long run Sharpe ratios of the interest rate and credit risk factors are equal, and close to a value of 0.40, the resulting long-run equilibrium Sharpe ratios for different asset classes are consistent with historic Sharpe ratios computed over long time periods. 26 Looking forward, our central scenario is more cautious in terms of the absolute expected level of the factor Sharpe ratios, based on the assumption of more mature economies and financial markets. We thus adopt a value of 0.25 for both risk factors. 15

16 Scenario 1: Base Scenario Expected EMD Sharpe ratios come from the exposure of assets to interest rate and credit / equity / liquidity risk factors only The expected Sharpe ratios for the assets classes in our analysis are shown in Appendix 2. Exhibit 13 shows how a portfolio invested in EM sovereign hard currency debt might perform if diversified across a broader EM space. 29 In Portfolio 1, we can see that reducing EM sovereign hard currency debt exposure in favour of some unhedged EM sovereign local currency debt raises the expected portfolio premium but also increases volatility, leaving the Sharpe ratio unchanged. The benefits of diversification in this portfolio are lost given the higher volatility of unhedged EM sovereign local currency debt, a risk that we assume not to be remunerated in this scenario. Portfolio 2, of which a portion is invested not only in unhedged EM sovereign local currency debt but also in unhedged EM corporate local currency debt, should in theory benefit from its exposure to credit risk, but in fact does only slightly better than Portfolio 1 in terms of the Sharpe ratio. Increasing the corporate hard currency bond allocation further as in portfolios 4 and 5 does not improve the portfolio Sharpe ratio, but does slightly improve the expected return. Finally, Portfolio 6 is highly diversified in terms of both hard and local currency as well as sovereign and corporate debt, but given our assumptions concerning unhedged EM sovereign local currency debt no remuneration for the EM FX risk this does not suffice to improve the Sharpe ratio relative to the three former portfolios, though Portfolio 6 does perform better than the reference portfolio. The conclusion we can draw from this scenario is that looking forward, the right long-run strategic allocation for investors looking for EM debt exposure without EM currency risk would be a mix of between 60%/40% and 50%/50% EM sovereign hard currency debt/em corporate hard currency debt. Portfolio 3 invests in a 60/40% mix of EM sovereign hard currency debt and EM corporate hard currency debt which reduces the overall portfolio volatility and, given our scenario on the expected Sharpe ratio for EM corporate hard currency debt, enables it to deliver a significantly higher Sharpe ratio. Exhibit 13: Impact of diversification on the reference portfolio 100% EM sovereign hard currency debt EM Fixed Income Reference Portfolio Portfolio 1 Portfolio 2 Portfolio 3 Portfolio 4 Portfolio 5 Portfolio 6 EM Sovereign Hard Currency 100% 60% 60% 60% 50% 40% 30% EM Sovereign Local Currency 0% 40% 30% 0% 0% 0% 30% EM Corp Hard Currency 0% 0% 0% 40% 50% 60% 35% EM Corp Local Currency 0% 0% 10% 0% 0% 0% 5% EM Fixed Income Total 100% 100% 100% 100% 100% 100% 100% Weighted Effective Duration Annualised Expected Excess return 1.79% 1.91% 1.87% 1.86% 1.87% 1.89% 1.93% Annualised Expected Volatility 7.5% 8.0% 7.5% 6.7% 6.7% 6.7% 7.2% Sharpe ratio Source: HSBC Global Asset Management calculations, December Excess returns in USD. See Appendix 1 for a full list of the indices used to represent the investment universe. For additional important information, including index descriptions and disclosures, please refer to pages 25 and 26. Past performance is no guarantee of future results. For illustrative purposes only and is not intended as an offer or solicitation for the purchase or sale of any financial instrument, nor should this be considered a recommendation for any investment or trading strategy. 16

17 Scenario 2: adding FX risk and higher expected premiums on hard currency corporate debt Expected EMD Sharpe ratios come from EMD assets exposure to interest rate, credit / equity / liquidity and EM FX risk factors; and higher expected premiums on EM hard currency corporate debt. In this alternative scenario, we used the same logic as in Scenario 1 but relaxing our earlier assumption on FX risk remuneration for local currency emerging markets assets so that in Scenario 2 FX risk can earn a premium over the long run. 30 There is indeed some evidence that investors require a premium to invest in unhedged local currency EM assets, not only in relation to interest rate risk but also to FX risk, which ultimately results in a remuneration of this risk factor. Moreover, we do see opportunity in emerging markets local currency debt over the longer term. This is philosophically consistent with the Balassa- Samuelson effect, which underscores how prices in developing countries converge with advanced countries through adjusted foreign exchange rates. 31 In this second scenario, we also raised the expected premiums on EM corporate hard currency debt. Indeed, we believe that EM corporate hard currency debt will continue to benefit from the ongoing economic development of EM markets with strong returns, a greater number of issuers resulting in lower volatility of the asset class and hence a higher Sharpe ratio. For instance, the CEMBI Broad Index has more than 500 names today, compared to around 200 issuers in July There is an additional reason to expect a higher Sharpe ratio in EM corporate hard currency debt, which is that the number of analysts for each issuer is quite low compared to corporates in developed markets. This is due to the youth of the EMD corporate market and the very rapid growth in its issuer numbers. The age of a company and the number of analysts researching it have an impact on the quality of the information relative to that company, both in terms of distribution and general level of understanding, which in turn affects the required risk premium. 32 Investors demand higher premiums for less transparent companies with an asymmetrical distribution of information (which goes on to translate into higher returns). This is probably one of the main factors behind the systematic positive spread of USD hard currency emerging corporates relative to US corporates for an equivalent rating, the latter being older and followed by a large number of analysts. In addition, going forward we could see a slow downward trend of these required spreads, which might temporarily inflate EM corporate hard currency debt returns. Exhibit 14 shows that shifting our allocation from 100% EM sovereign hard currency to a 60/40% hardlocal sovereign debt mix as in Portfolio 7 raises the portfolio s volatility but also increases its annualised returns by almost 50 basis points, raising the Sharpe ratio from 0.24 to Portfolio 8, with some additional allocation to EM corporate local currency debt, does slightly better in terms of its Sharpe ratio. Portfolios 9, 10, and 11, which are only invested in EM hard currency debt, deliver higher Sharpe ratios than portfolios 3, 4 and 5 as a result of our new Sharpe ratio scenario for EM corporate hard currency debt. The best result however is achieved by Portfolio 12, which is highly diversified in all four sectors: this portfolio yields almost 60 additional basis points on the return compared to the reference portfolio, as well as a reduction in volatility, resulting in a Sharpe ratio of Exhibit 14: Impact of diversification on the reference portfolio 100% EM sovereign hard currency debt according to Scenario 2 EM Fixed Income Reference Portfolio Portfolio 7 Portfolio 8 Portfolio 9 Portfolio 10 Portfolio 11 Portfolio 12 EM Sovereign Hard Currency 100% 60% 60% 60% 50% 40% 30% EM Sovereign Local Currency 0% 40% 30% 0% 0% 0% 30% EM Corp Hard Currency 0% 0% 0% 40% 50% 60% 35% EM Corp Local Currency 0% 0% 10% 0% 0% 0% 5% EM Fixed Income Total 100% 100% 100% 100% 100% 100% 100% Weighted Effective Duration Annualised Expected Excess return 1.79% 2.28% 2.22% 2.03% 2.09% 2.16% 2.40% Annualised Expected Volatility 7.5% 8.0% 7.5% 6.7% 6.7% 6.7% 7.2% Sharpe ratio Source: HSBC Global Asset Management calculations, December Excess returns in USD. See Appendix 1 for a full list of the indices used to represent the investment universe. For additional important information, including index descriptions and disclosures, please refer to pages 25 and 26. Past performance is no guarantee of future results. For illustrative purposes only and is not intended as an offer or solicitation for the purchase or sale of any financial instrument, nor should this be considered a recommendation for any investment or trading strategy. 17

18 Implications of allocating to EMD within a broader context So far we have explored portfolios exclusively invested in emerging market fixed income, but we also wanted to explore the potential implications of our findings in the broader context of a portfolio invested in other asset classes. To find out, we took the example of a US portfolio similar in profile to what a US pension plan portfolio could look like. We applied the approach developed above to show how an increased allocation to EM bonds might affect a balanced portfolio of 45% mostly domestic US fixed income and 55% equity, the equity segment having a relatively larger exposure to foreign markets. 33 Foreign assets are assumed to be hedged in USD, with the exception of local currency EM assets. Expected premiums are based on the assumptions of Scenario 2. Exhibit 15 presents the expected premiums (relative to the cash rate) and Sharpe ratios resulting from the range of various asset mixes. It is worth noting that these results using a US portfolio are comparable to what we would expect to see for benchmarks from other regions, allowing for certain conditions: in particular, for another given region, we would need to assume that the expected Sharpe ratios and volatilities of the domestic assets were close to those of the US market and that the impact on premiums resulting from hedging the US dollar-denominated assets in the applicable currency was negligible. In addition, we would need to consider the specific EM FX risk for this other region when dealing with EM local currency debt. However, we do not believe that the impact of these assumptions is significant when considering highly diversified markets. 18

19 Exhibit 15: Impact of Emerging Debt on a Global Fixed Income-Equity portfolio Global Reference Portfolio 13 Portfolio 14 Portfolio 15 Portfolio 16 Portfolio 17 Portfolio 18 Portfolio 19 Portfolio 20 Portfolio US Fixed Income 40% 30% 30% 30% 30% 30% 30% 30% 20% Government + Quasi-Government 18.7% 13.6% 13.6% 13.6% 13.6% 13.6% 13.6% 13.6% 8.6% Corporate Investment Grade 8.8% 6.4% 6.4% 6.4% 6.4% 6.4% 6.4% 6.4% 4.0% Corporate High Yield 0.6% 0.5% 0.5% 0.5% 0.5% 0.5% 0.5% 0.5% 0.3% Mortgage-Backed Securities 8.9% 6.5% 6.5% 6.5% 6.5% 6.5% 6.5% 6.5% 4.1% Government Inflation-Linked 3.0% 3.0% 3.0% 3.0% 3.0% 3.0% 3.0% 3.0% 3.0% Foreign Fixed Income 5% 15% 15% 15% 15% 25% 25% 25% 35% Government Developed Markets 5.0% 6.6% 3.6% 3.6% 3.6% 12.0% 5.0% 4.0% 4.2% Corporate Investment Grade 6.6% 3.6% 3.6% 3.6% 10.0% 5.0% 4.0% 4.5% Corporate High Yield 0.0% 1.8% 1.8% 1.8% 1.8% 3.0% 3.0% 3.0% 3.0% EM Sovereign Hard Currency 0.0% 0.0% 6.0% 3.0% 1.5% 0.0% 6.0% 4.0% 3.5% EM Sovereign Local Currency 0.0% 0.0% 0.0% 3.0% 0.0% 0.0% 6.0% 7.6% EM Corp Hard Currency 0.0% 0.0% 3.0% 1.5% 0.0% 6.0% 4.0% 9.1% EM Corp Local Currency 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 3.0% US Equities 33% 33% 33% 33% 33% 27% 27% 27% 27% Foreign Equities 22% 22% 22% 22% 22% 18% 18% 18% 18% Developed Markets 19.8% 19.8% 19.8% 19.8% 19.8% 16.2% 16.2% 16.2% 16.2% Emerging Markets 2.2% 2.2% 2.2% 2.2% 2.2% 1.8% 1.8% 1.8% 1.8% TOTAL 100% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% Effective Duration FI Portfolio Weighted Fixed Income Eff. Duration Annualized Expected Excess return 2.06% 2.09% 2.14% 2.16% 2.19% 1.88% 2.04% 2.12% 2.27% Volatility 6.8% 6.9% 7.1% 7.1% 7.1% 6.0% 6.4% 6.6% 6.9% TE vis-à-vis Reference Portfolio 0.4% 0.6% 0.5% 0.6% 0.9% 0.7% 0.9% 5.9% IR vis-à-vis Reference Portfolio Sharpe-ratio Source: HSBC Global Asset Management calculations, December Differences due to rounding. See Appendix 1 for a full list of the indices used to represent the investment universe. For additional important information, including index descriptions and disclosures, please refer to page 25 and 26. Past performance is no guarantee of future results. For illustrative purposes only and is not intended as an offer or solicitation for the purchase or sale of any financial instrument, nor should this be considered a recommendation for any investment or trading strategy. The US market is itself rich and diversified, and beating the US reference benchmark presented in this section is not an easy task. However, our scenario showed that even such a portfolio can benefit from diversification. In Portfolio 13, we can see that reducing domestic bond exposure in favour of foreign bonds, using mostly global investment grade, plus a small amount of foreign high-yield bonds, does not result in any significant benefits, either in terms of yield or reduced volatility. Portfolio 14, with a significant portion invested in EM sovereign hard currency debt, does slightly better than Portfolio 13 in terms of performance, though it adds some volatility and leaves the Sharpe ratio unchanged. Adding some EM corporate hard currency debt as in Portfolio 15 or including a mix of EM hard currency debt and local currency debt as in Portfolio 16 resulted in an increased return over the reference portfolio with only a slight increase in volatility and even a slight increase of the Sharpe ratio, showing the benefits of diversification with EM assets under the assumptions of Scenario 2. Portfolio 17 reallocates more aggressively to foreign bonds from equities, but without diversifying into the Emerging Market fixed income space. 19 The Sharpe ratio increases at the expense of a significant fall in performance relative to the reference portfolio. 34 While keeping the same fixed income/equities split as in Portfolio 17, Portfolio 18 starts investing in the EM Fixed Income market, through sovereign and corporate hard currency debt, improving the Sharpe ratio relative to the reference portfolio. Portfolio 19 invests in EM local currency debt, further enhancing the prospective premiums against the reference portfolio. Finally, Portfolio 20 goes even more aggressively into the foreign fixed income emerging market space. It is in this more diversified portfolio that the highest risk-adjusted returns are to be found. It delivers a premium of 20 basis points over the reference portfolio return, with a Sharpe ratio of 0.33 compared to 0.30 for the reference portfolio. Naturally, Portfolio 20 has a high tracking error with respect to our chosen reference portfolio, due to the 25% allocation to emerging markets (including equities). This figure is not as shocking on a prospective basis, assuming that financial depth will continue to develop in emerging markets, and considering that emerging and developing economies represent 56% of world GDP to date.

20 Conclusion Finding the balance Implications for investors Emerging markets now represent around 50% of the global GDP, with EM bonds accounting for over 6% of the global fixed income market, making it an asset class on which investors can no longer afford to miss out. Although their fundamentals have dramatically improved over the last few years, EMD sectors remain exposed to different systematic risks (interest rate risk, equity risk and credit risk notably) that provide investors with a long-term premium and has the potential to generate solid Sharpe ratios over the long run. The emergence of EM hard currency corporate debt and EM local currency sovereign debt as well-established asset classes makes the overall asset class deeper, but also more complex when identifying the EMD beta and when building allocations aiming to capture it over the long term. What investors should aim for is to find the optimal asset mixes within their EMD allocation and to assess the contribution of EMD in terms of risk-adjusted returns in their global portfolios. Using an approach based on the underlying factors that influence the behaviour of the asset class, our research demonstrates that: - In a pure USD EM hard currency portfolio, the optimal allocation is to a balanced mixed of sovereign and corporate debt. Adding local currency debt to this mix is beneficial if we make the assumption that EM currency risk will be remunerated in the long run. - In a broader portfolio allocated to global bonds and global equities, the inclusion of a balanced portfolio of EMD can be an interesting alternative to equities, bringing a significant enhancement in terms of risk-adjusted returns. The continuous changes in Emerging market indices (ratings, constituents) over the last few years and their still high concentration (modest number of issuers overall) are all the more reasons to be looking to combine the various EMD sub-segments. These factors also stress the need to re-assess our approach regularly and possibly complement it with alternative index construction techniques (such as GDP-weighted) or by factoring in more timely valuation elements, while still favouring the active management of EMD allocations. 20

21 Authors? Xavier Baraton (New York) Global CIO Fixed Income, Regional CIO North America HSBC Global Asset Management (USA) Inc. Maria-Laura Hartpence (Paris) Head of Fixed Income Quantitative Research HSBC Global Asset Management (France) Xavier Baraton joined HSBC in September 2002 to head the Paris based Credit Research team and became Global Head of Credit Research in January From 2006, Xavier managed euro credit strategies before being appointed as Head of European Fixed Income in 2008 and as Global CIO, Fixed Income in Prior to joining HSBC, Xavier spent six years at Credit Agricole Indosuez, including fi ve years as Head of Credit Research. Xavier began his career in 1994 in the CCF Group. Xavier graduated from the Ecole Centrale Paris as an engineer with a degree in Economics and Finance in 1993 and holds a postgraduate degree in Money, Finance and Banking from the Université Paris I Panthéon Sorbonne (France) in Maria-Laura Hartpence has been in the industry since 1984, working extensively with economic and financial forecast models. Prior to joining HSBC in 1995, Maria-Laura worked as a senior economist at Bunge-Born Group in Sao Paulo and at Cecogest, an independent French investment firm. She graduated with a degree in Economics from the University of Brasilia, and holds a Master of Arts in Economics from the University of Miami and a Master of Science in Probability and Statistics from the University of Sao Paulo. 21