Digging deeper

Emerging market debt choices

March 2012

In brIeF
The case for emerging markets debt (EMD) has convinced many investors. Now, they are asking which type of EMDUSD-denominated sovereign debt, USD-denominated corporate debt or localcurrency government debtmakes most sense. Each of the three EMD asset classes has delivered strong returns over time and deserves consideration. If the global economy grows strongly in coming years, local-currency government debt will likely outperform, thanks to its foreign-exchange component. Hard-currency sovereign debt, a lower-beta investment, is likely to outperform in a softer global growth environment. EM corporate debt, which is currently attractively valued, offers more direct exposure to EM Asia than the two government-debt EMD categories.

Interest in emerging markets debt (EMD) has soared in the past few years, with investors responding to sustained rapid growth and solid fiscal metrics in this group of economies. This paper does not rehash the well-understood general arguments in favor of EMD. Instead, it focuses on a question that frequently arises after an investor grows comfortable with the overall story: Which type of EMD makes most sense?
EMD, after all, consists of not one but three separate asset classes: dollar-denominated government debt, typically referred to as sovereign debt; dollar-denominated EM corporate debt; and local-currency government debt. During the past decade or so, each has done well, generating fairly high returns and producing attractive Sharpe ratios. Their track records suggest

For InstItutIonal and ProFessIonal Investor use only | not For retaIl use or dIstrIbutIon

Digging deeper

that any of the three should fit comfortably within most investors portfolios. The characteristics of the three asset classes do differ, however, and each will likely suit particular objectives. The EMD asset classes vary not so much in terms of their geographical composition or their credit quality but rather in the type of exposure to global markets that they imply. Buying localcurrency debt means taking exposure to EM foreign-exchange (FX) trends, while the USD-based assetsas with all unhedged spread productinclude a U.S. Treasury component. The two types of debt are likely to perform relatively well under different states of the world. When global growth is strong, EM FX will tend to appreciate, and Treasury yields will tend to rise. These factors favor local-currency debt. During less boomy periods, Treasury yields will likely decline and EM FX will remain stable or sell off, tilting the scales toward dollar-based assets. Indeed, between 2003 and 2007, when the first set of conditions applied, local-currency debt outperformed USD-denominated EMD. Between 2008 and 2011, when the second type of environment dominated, USD-based sovereign debt outperformed. Over a lengthy time horizon of, say, five years, local-currency debt seems likely to outperform again. Global growth will likely
exhIbIt 1: eMd asset class characterIstIcs

accelerate gradually in coming years after the disappointments of the early part of the recovery; EM FX is coming off a significant sell-off in 2011 and remains well below its pre-crisis peak; and the very low level of Treasury yields makes a rise over a five-year period more likely than a decline. Investors willing to add beta to their portfolios and contemplate a long holding period might thus favor investments in local-currency EMD. The EM FX appreciation path, though, will not necessarily proceed smoothly, and any particular year could bring a sell-off. Moreover, the current global monetary policy atmosphere means that a Treasury sell-off may lie some distance down the road. Investors with a shorter horizon, and those less enthusiastic about adding beta to their portfolios, might thus favor USD-denominated sovereign debt. While USD-denominated EM corporate debt trailed the other two EMD categories in returns during both 2003-2007 and 2008-2011 (though it underperformed local-currency debt only very slightly during the latter period), it possesses two virtues that make it an interesting option as at least part of an EMD portfolio. First, it is the most Asian of the three EMD asset classes, providing more exposure to the fastest-growing part of the global economy and one in which many investors still have only modest holdings.

Type

eMbIG div External government debt ("sovereigns") Currency USD Yield (%) 5.69 Spread over Treasuries (bps) 395 Average life (years) 11 Duration (years) 7.1 Credit quality Baa3/BBBMarket value (USD bn) 276 Return components Coupons and movements in U.S. Treasuries and credit spreads Correlation with risky assets Lowest Correlation with likely funding assets Middle Advantages Relatively low beta; high current yield; no FX risk; likely improvement in EM relative creditworthiness Disadvantages

ceMbI br div External corporate debt USD 5.8 443 8 5.2 Baa2/BBB 213 Coupons and movements in U.S. Treasuries and credit spreads Middle Highest Relatively cheap to its own history as of early 2012; greatest exposure to fast-growing EM Asia; high current yield; no FX risk

GbI-eM G div Local government debt Local EM FX 6.23 N/A 6.7 4.6 Baa2/BBB+ 864 Local yields and movements in EM FX Highest Lowest Exposure to likely EM FX appreciation (estimated at 1% per annum on average); diversification; high current yield; high historical Sharpe ratio; limited interest-rate risk

High degree of exposure to Treasury Modest pickup over U.S. corporates, espe- Volatility of EM FX; relatively high beta yields; lowest yield cially in high-yield component; volatile default rate; likely lowest liquidity; limited exposure to tech and consumer sectors Sluggish but positive global growth Sluggish but positive global growth and absence of financial-market shocks Faster global growth with rising Treasury yields

Backdrop for outperformance

Source: J.P. Morgan Securities, Inc., J.P. Morgan Asset Management. Data as of January 31, 2012. Indices do not include fees or operating expenses and are not available for actual investment.

2 | Digging deeper: Emerging market debt choices

Second, in contrast with sovereign debt, its spread over Treasuries currently stands wide to its medium-term average. Investors with a focus on value thus may find EM corporates more appealing than the two government-based EMD assets.

exhIbIt 2: GeoGraPhIc coMPosItIon oF eMd IndIces

LatAm Asia Europe 19 27 28 29 30 13 6 3 5 16 5 10 15 20 25 Market cap (%) 30 35 40 45 10 34 EMBIG Div GBI-EM Div CEMBI Br-Div 40

40

Snapshot of the EMD asset classes

Exhibit 1 (on the previous page) presents the main characteristics of each asset class within the EMD category. In each case, we use the relevant JPMorgan index as a proxy for the underlying asset class. Among the various JPMorgan indices, we chose the ones that we believe to be most used as benchmarks by asset managers: the EMBI Global Diversified for dollar-based external debt; the CEMBI Broad Diversified for EM corporates; and the GBI-EM Global Diversified for local-currency government debt. All of the exhibits and numbers in this report refer to these indices, to which we will refer as EMBIG, CEMBI, and GBI-EM for simplicitys sake. Each of these indices comes in other forms, including an undiversified version that does not constrain the weights that individual countries can take. At the moment (using figures as of January 31, 2012), the GBIEM offers, by a slim margin, the highest yield among the three EMD asset classes. This has not always been the case, but the very low level of Treasury yields has dragged down the overall yield on the two spread product classes, even though spread levels are either in line with (EMBIG) or wider than (CEMBI) medium-term averages. Its higher yield comes despite a shorter duration and a marginally higher average credit rating than are the case for the USD-based indices.

Africa Middle East 0 0

Source: J.P. Morgan Securities, Inc., J.P. Morgan Asset Management. Data as of January 31, 2012.

What does distinguish EMBIG from GBI-EM is the greater number of individual country exposures, currently 44 in the former against 14 in the latter. This difference reflects the considerably larger number of countries that have issued dollar-denominated debt in international markets, compared with the relatively small number of countries whose local bond markets are easily accessible to foreign investors. The smallest 30 country positions in the EMBIG represent a non-trivial 32% of the total index. The EMBIG thus provides greater diversification across countries. At the same time, the EMBIG also implies a somewhat higher degree of exposure to commodities. About 56% of this index consists of countries whose exports are dominated by such primary goods, compared with 45% for the GBI-EM (Exhibit 3). This contrast owes less to the broad regional compositions and more to differing weights attached to individual countries. The manufacturing economy Poland, for example, makes up 10% of the GBI-EM versus 3.6% of the EMBIG. Among the three indices, the corporate one (CEMBI) gives the highest weight to Emerging Asia40%, compared with 29% and 19% for the GBI-EM and EMBIG, respectively. Its Latin America
exhIbIt 3: coMPosItIon oF eMd IndIces by tyPe oF econoMy
60 50 40 Percent 30 20 10 0 Commodity Manufacturing GBI-EM Div EMBIG Div

Composition of the asset classes

A first step in discussing the three main asset types within EMD is to figure out what exactly an investor is buying in each case, in terms of geography, type of economy and, for EM corporate bonds, sector exposure. Do the different EMD categories expose investors to vastly different parts of the world? Not really. The two government bond indicesEMBIG and GBI-EMdo not differ greatly in terms of regional breakdown (Exhibit 2). Relative to the dollar debt, the local debt index contains a little more Emerging Asia and a bit less Latin America, but not dramatically so. EMEA represents a broadly similar component of each. Of the seven largest country weightings in each of the diversified indices, four match: Brazil, Indonesia, Mexico, and Turkey.

Source: J.P. Morgan Securities, Inc., J.P. Morgan Asset Management. Data as of January 31, 2012.

J.P. Morgan Asset Management | 3

Digging deeper

component basically matches the GBI-EM. One factor that sets the CEMBI apart from the other two is the high representation of Middle Eastern issuers, which make up 16% of the total index, at the expense of Emerging Europe, which is much less represented in the CEMBI than in the government debt classes. In terms of sectors, the CEMBI gives investors greater exposure to high-yield banks than is the case in U.S. corporate indices. It also assigns a somewhat larger weight to industrials in both the investment grade and high yield components and contains quite a bit more oil in investment grade and metals in both, while underweighting consumer products relative to its U.S. counterparts (Exhibit 4). In general, banks aside, the CEMBI appears to possess something of a bias toward export-oriented companies, an inclination that makes sense given investors presumably greater willingness to lend dollars to firms that generate hard currency through their regular operations.
exhIbIt 4: ceMbI coMPosItIon by sector and ratInG cateGory

Performance history
returns
All three EM debt indices have delivered strongly positive returns since the early 2000s (Exhibit 5 and 6, on the following page). The newest of the three, the GBI-EM, goes back to 2003. Since its inception, it has outperformed the other two on average. Local debt, though, did experience two down years during that period, 2008 and 2011, both times when EM FX suffered sharp depreciation. The dollar-based indices each lost money once, in 2008. The largest full-year drop and increase occurred in the CEMBI, which lost 15.9% in 2008 and jumped 34.9% the following year. Going back further, the EMBIG Div started in 1994, and its average annual return for that longer period almost exactly matches the 2003-2011 era although two of the first nine years were negative ones.

volatility
IG (%) 37 0 16 6 16 4 11 11 hy (%) 23 1 36 9 5 6 13 7

Banks Consumer Products Industrials Metals Oil Retail Telecom Utilities

Unsurprisingly, considering that 2008-2009 experience, the CEMBI has shown the highest volatility of annual returns, with a 13.4% standard deviation for the 2003-2011 period. Despite its two negative years, the GBI-EM has displayed the lowest annual return volatility, with a standard deviation of 10.1%. The EMBIG sits marginally higher at 11.5%. These figures put the Sharpe ratio for the GBI-EM at 1.16, above both the EMBIG at 0.90 and the CEMBI at 0.59. In its early days, the EMBIG behaved in somewhat more volatile fashion than it has more recently, so the Sharpe ratio for its entire period of existence since 1994 drops to 0.73.

Source: J.P. Morgan Securities, Inc., J.P. Morgan Asset Management. Data as of January 31, 2012.

the treasury component

Given that both of the dollar debt indices contain a significant U.S. Treasury component, movements in Treasuries have played a role in driving returns in these spread products. Treasuries generally represented a negative in the 2003-2006 period, but the subsequent plunge in yields boosted returns. The EMBIG returned a cumulative 46.1% during the 2007-2011 years, with Treasuries accounting for approximately 19.8 percentage points of that return (or a little less than half the total). A similar result is seen for the CEMBI, which returned a cumulative 36.4% during those same years, with half (18.4 percentage points) owing to the Treasury rally. The shorter duration of CEMBI compared with EMBIG helps account for its mild underperformance during the period of falling Treasury yields.

Composition differences alone do not appear to present solid arguments for favoring one type of EM debt versus another. Investors who prefer to gain exposure to fast-growing EM Asia should probably prefer local debt to dollar debt, but the difference between the two indices is not compelling, and the greater diversification within EMBIG Div represents a counterargument. CEMBI is the most Asian of the three, but it does not contain much within the consumer or technology areas that tend to attract investors to the region. At the moment, its very small Emerging European component probably looks like a plus, though over time investors might want to obtain exposure in that region to play a structural re-convergence with the rest of the world. CEMBI does, naturally, contain a far larger number of individual issuers336 corporatesthan either EMBIG or GBIEM, providing a greater degree of internal diversification than is available in the government indices.

the eM Fx component
Local-currency debt obviously does not include a U.S. Treasury component. Unlike the two dollar-based indices, however, it

4 | Digging deeper: Emerging market debt choices

exhIbIt 5: eMd annual returns by asset class

2003 eMbIG 22.2% GbI 16.9% ceMbI 16.2%

2004 GbI 23.0% eMbIG 11.6% ceMbI 10.3%

2005 eMbIG 10.2% GbI 6.3% ceMbI 6.1%

2006 GbI 15.2% eMbIG 9.9% ceMbI 6.5%

2007 GbI 18.1% eMbIG 6.2% ceMbI 3.9%

2008 GbI -5.2% eMbIG -12.0% ceMbI -15.9%

2009 ceMbI 34.9% eMbIG 29.8% GbI 22.0%

2010 GbI 15.7% ceMbI 13.1% eMbIG 12.2%

2011 eMbIG 7.3% ceMbI 2.3% GbI -1.8%

Source: J.P. Morgan Securities, Inc. Data as of December 30, 2011.

exhIbIt 6: eMd Index PerForMance sInce 2003 exhIbIt 7: cuMulatIve GbI-eM return decoMPosItIon

Average return (%) Standard deviation (%) Max (%) Min (%) Sharpe ratio

eMbIG 10.3 11.5 29.8 -12.0 0.90

GbI-eM 11.8 10.1 23.0 -5.2 1.16

ceMbI 7.9 13.4 34.9 -15.9 0.59

140 120 100 Percent 80 60 40 20 0 -20 2003

Coupon and price return Currency return

Source: J.P. Morgan Securities, Inc., J.P. Morgan Asset Management. Data as of December 30, 2011.

contains FX risk. EM FX appreciation (at least in the countries represented in the index) boosts returns, while FX depreciation eats into performance. We can decompose the returns over time on the overall index into those stemming from spot FX movement and those coming from coupons and changes in bond prices in local-currency terms. For the first several years of the indexs history2003 until mid-2008overall returns reflected currency appreciation and the coupon/price effect in roughly equal size. During that time, major EM currencies rose strongly against the dollar. In the most spectacular case, the Brazilian real appreciated from 3.54/USD at the end of 2002 to 1.60 in June 2008. At that point, though, EM FX went violently into reverse. The sharp selloff in the second half of 2008 wiped out most of the previous years FX gains. Although EM FX then appreciated steadily until mid-2011, returning to within shouting distance of its previous high, another wobble in the second half of last year unwound a significant portion of those renewed gains. Even the Brazilian real, one of the better performers, currently stands a bit weaker than its mid-2008 level. For the 2003-2011 period as a whole, the FX component of the GBI-EM Div returned a cumulative 20.1%, about 2.1% per annum, compared with 127.4% for the coupon and price parts, a 9.6% annual average (Exhibit 7). The annual return table presented above shows that, unsurprisingly, the GBI-EM Div outperformed its dollar-based counterparts in aggregate during the FX appreciation period (though not in each individual year) and has generally underperformed since. To the extent that EM FX appreciation is likely to occur during periods of strong global

2004

2005

2006

2007

2008

2009

2010

2011

Source: J.P. Morgan Securities, Inc., J.P. Morgan Asset Management. Data as of December 30, 2011.

growth, when U.S. bond yields might be expected to rise, the EM FX component of the GBI-EM will serve as a mirror image of the Treasury component of the EMBIG and CEMBI.

Correlations with other asset classes

Relative to the asset classes out of which investors might wish to fund EMD positions, local-currency debt appears to offer somewhat greater diversification benefits than the two dollardenominated categories. On average, over the past eight years, returns on the GBI-EM have displayed less correlation with the WGBI (its DM equivalent) than returns on the EMBIG and the CEMBI have with U.S. corporate spread product. The differences, though, are not enormous, especially between the GBI-EM and the EMBIG. The correlation of the GBI-EM with the WGBI has also risen since the start of the financial crisis, while remaining a bit lower than the correlations of dollar-based indices with U.S. corporates (Exhibit 8, on the following page). The latter correlations appear to have changed character somewhat in the past two years: Before 2010 both the CEMBI Div and the EMBIG Div correlated more with the U.S. investment-grade index than with U.S. high yield, but more recently the connection with high yield has strengthened, possibly because of the relatively low

J.P. Morgan Asset Management | 5

Digging deeper

yield now prevailing on investment grade and correspondingly lower volatility (Exhibits 9, 10). On the flip side of the GBI-EMs diversification benefit relative to its funding category, this asset class shows higher correlation with non-fixed-income risk assets than do the USD-based instruments, largely because of the Treasury component of the latter. The negative correlation of Treasuries with, for example,
exhIbIt 8: GbI-eM rollInG 2-year correlatIon wIth wGbI
0.90 0.80 2-year rolling correlation 0.70 0.60 0.50 0.40 0.30 0.20 0.10 0.00 2005 2006 2007 2008 2009 2010 2011

the S&P 500 dampens the overall correlations of the USD indices with such risk assets. The GBI-EM lacks that Treasury cushion, leaving it more exposed to swings in risk sentiment. During the financial crisis, all three EMD assets showed extremely high correlations with the S&P 500 and, to a lesser extent, commodities (Exhibits 11, 12). During more normal times, even including the second half of 2011, the GBI-EM moves significantly more closely with risk assets than do the EMBIG or the CEMBI.
exhIbIt 11: eMd rollInG 2-year correlatIon wIth s&P 500
1.00 0.90 2-year rolling correlation 0.80 0.70 0.60 0.50 0.40 0.30 0.20 0.10 0.00 -0.10 2005 2006 2007 2008 2009 2010 2011 CEMBI EMBIG GBI-EM

Source: J.P. Morgan Securities, Inc., J.P. Morgan Asset Management. Data as of December 30, 2011.
exhIbIt 9: eMbIG rollInG 2-year correlatIon wIth u.s. corPorates
1.00 0.90 2-year rolling correlation 0.80 0.70 0.60 0.50 0.40 0.30 0.20 0.10 0.00 2004 2005 2006 2007 2008 2009 2010 2011 IG HY

Source: J.P. Morgan Securities, Inc., J.P. Morgan Asset Management. Data as of December 30, 2011.
exhIbIt 12: eMd rollInG 2-year correlatIon wIth coMModIty Index
1.00 0.80 2-year rolling correlation 0.60 0.40 0.20 0.00 -0.20 -0.40 2005 2006 2007 2008 2009 2010 2011 CEMBI EMBIG GBI-EM

Source: J.P. Morgan Securities, Inc., J.P. Morgan Asset Management. Data as of December 30, 2011.
exhIbIt 10: ceMbI rollInG 2-year correlatIon wIth u.s. corPorates
1.00 0.90 2-year rolling correlation 0.80 0.70 0.60 0.50 0.40 0.30 0.20 0.10 0.00 2004 2005 2006 2007 2008 2009 2010 2011 IG HY

Source: J.P. Morgan Securities, Inc., Bloomberg, J.P. Morgan Asset Management. Data as of December 30, 2011.

The future of EM FX
As discussed earlier, FX exposure represents the main difference between GBI-EM and the two dollar-based EMD asset classes. We saw that over the 2003-2011 period, nominal FX appreciation added about 2% per annum to the GBI-EMs returns, but that this boost came entirely between 2003 and mid-2008, with negative FX returns thereafter. How should investors think about prospects for EM FX in coming years? We can derive an estimated path for EM FX by answering a series of questions.

Source: J.P. Morgan Securities, Inc., J.P. Morgan Asset Management. Data as of December 30, 2011.

6 | Digging deeper: Emerging market debt choices

First, should EM real (inflation-adjusted) exchange rates appreciate over time? Almost certainly they should. Economists think in terms of real exchange rates because these capture the effects of both nominal FX moves and domestic price changes and therefore help measure trends in competitiveness. Emerging economies grow more quickly than their DM counterparts, in large part because of more rapid productivity increases. In turn, their equilibrium real exchange rates should appreciate over time: As an economy becomes more productive, its exchange rate should become stronger to keep trade flows broadly unchanged. Another perspective on upward EM FX pressure comes from the International Monetary Fund (IMF), which calculates exchange rates based on purchasing power parity (PPP). These PPP exchange rates for 2011 stood 54% above their nominal counterparts, again suggesting strong likelihood of EM FX appreciation. Second, how rapidly should EM real exchange rates appreciate? Emerging economies excluding China (which is not part of the GBIEM Div) grow, in real terms, about 3% per annum faster than DM economies. That outperformance has been fairly stable in recent years and is likely to persist as the still-large income gap between the two sets of economies narrows gradually. A growth differential of that size probably warrants a roughly similar trend rate of EM real FX appreciation against DM. This equivalence has broadly worked in recent years: the aggregate EM real exchange rate rose 3.5% per annum, on average, between 2004 and 2011 (Exhibit 13). Third, what will EM inflation rates look like in coming years? Changes in real exchange rates come from nominal FX moves and price changes, and we therefore need to determine the likely breakdown of expected real FX appreciation between these two factors. Since 2001, EM CPI inflation has averaged a touch over 5%. EM inflation slowed gradually during the first half of that period and dropped to 4.4% in 2006; the secular rise in commodity prices pushed EM inflation higher in 2007-2008 and
exhIbIt 13: eMerGInG econoMIes real eFFectIve exchanGe rate
140 135 Index (2000 = 100) 125 120 115 110 105 100 95 90 1994 1997 2000 2003 2006 2009 Flat +3.5% p.a. 130

again in 2010-2011, with a sharp recession-related fall to about 2% in between. Several factors suggest a somewhat slower pace of inflation in the future. First, a few large EM economies, like Russia and Turkey, made the transition to single-digit inflation during the past 10 years; this convergence helped the overall EM inflation deceleration observed through 2006. Second, much of the upward pressure on EM inflation since 2007 has come from food and energy commodity prices, which may continue to rise but which seem unlikely to duplicate their previous increases in percentage terms. Third, inflation targeting has spread among EM central banks as an organizing principle for monetary policy. Of the 14 countries included in the GBI-EM Div, 12 maintain explicit inflation targets. While debate persists about the long-term appropriateness of inflation targeting, for now this system appears to have helped anchor inflation expectations in many EM countries and made it easier for EM central banks to formulate coherent policies. The midpoint of the inflation targets for the GBI-EM countries stands at 3.5%. Even assuming a small miss, a 4% expectation for EM inflation looks reasonable. Assuming DM inflation of 2%very close to its average since 2001, and around most DM central banks targetsthe inflation differential between EM and DM should run at roughly 2% (Exhibit 14). The answers to these questions point to a trend rate of EM nominal FX appreciation of roughly 1% per yeara number suggested by a 3% rate of real appreciation, of which 2% would come from the inflation differential between EM and DM, leaving 1% for nominal FX moves. This overall appreciation rate could translate into a higher or lower pace of appreciation against the U.S. dollar alone, if the dollar were to fall or rise on a sustained basis against other DM currencies. At the moment, though, the overall dollar index does not appear particularly misaligned relative to its own history. And while the U.S. is running a nontrivial current account deficit (around 3% of GDP), the gap looks
exhIbIt 14: consuMer PrIces
10 8 Percentage change (y/y) 6 4 2 0 -2 2003 Emerging Markets Developed Markets

2004

2005

2006

2007

2008

2009

2010

2011

Source: J.P. Morgan Securities, Inc. Data as of December 30, 2011.

Source: J.P. Morgan Asset Management. Data as of December 30, 2011.

J.P. Morgan Asset Management | 7

Digging deeper

stable. One risk factor is the euro, which might fall against the dollar in order to stimulate euro area growth. Other things equal, significant euro weakness against the dollar would chip away at the expected rate of EM FX appreciation against USD. As a base case assumption, though, EM nominal FX rates seem likely to rise against the dollar by about 1% per year, on average.

The future of EM creditworthiness

Recent investor interest in EMD stems in large part from EM economies evident improvement in creditworthiness relative to DM economies, especially in terms of fiscal metrics. As a group, EM economies run smaller fiscal deficits than DM and carry considerably lighter public debt burdens. The considerable attention given to this phenomenon, however, suggests that markets have likely priced in the relative creditworthiness improvement that has occurred thus far. From here, the question is whether the trend will continue. Several metrics suggest that it will, but probably at a slower pace than in recent years.

expects a narrowing in the difference between the two to 1.4 p.p. by 2015.1 This trend makes sense. Many DM countries (such as peripheral euro area economies) are tightening fiscal policy under pressure from markets, and others, like the U.S. and U.K., are implementing medium-term improvements to public finances in efforts to avoid future financing difficulties. EM countries generally are not bumping against borrowing constraints and face little near-term incentive to improve fiscal positions. Moreover, the IMFs medium-term projections reveal another source of risk. The IMF assumes that EM government expenditure declines marginally as a share of GDP between now and 2016. Such a trend, set against the ongoing improvement in per capita GDP, would conflict with the well-known tendency for governments to spend more as incomes rise (essentially, demand for public goods tends to increase as populations become richer). EM fiscal balances may feel some spending pressure from this source in coming years. EM countries will likely continue to run narrower fiscal deficits than DM economies for the foreseeable future, but the improvement in perceived relative creditworthiness in this area likely has already occurred.

Per capita GdP

Often overlooked in this context, per capita GDP actually correlates better with creditworthiness than almost any other single economic indicator. The projections outlined earlier suggest that the gap in per capita GDP between EM and DM economies will continue to narrow. Taken together, real GDP growth about 3% per year faster than DM; an inflation differential around 2% per annum also in favor of EM; and 1% nominal EM FX appreciation against DM currencies, point to 6% faster GDP growth for EM compared with DM, measured in USD terms. With population growing about 1% per year more rapidly in EM than DM, EM per capita GDP should rise about 5% per annum more quickly than the DM measure. Relative creditworthiness on this basis should thus continue to improve. Markets likely understand the growth outperformance story well, though, and may have priced in at least some of this future convergence.

Public debt
In contrast with public finance flow data, stock numbers should continue to favor emerging economies in relative terms. While EM countries will likely run aggregate fiscal deficits only slightly smaller than DM shortfalls, their higher growth rates and lighter initial debt burdens will push public debt ratios downward. DM public debt, as a share of GDP, will probably continue to rise. The IMF Fiscal Monitor calculations from September 2011 show the gap between EM and DM public debt ratios widening to 78.5 p.p. in 2016, versus 65.1 p.p. in 2011, helped by GDP growth that exceeds real interest rates by 4.5 p.p. between 2012-2016 (compared with 0.4 p.p. in DM). Even if these numbers suffer some revision, the divergence between falling and rising debt ratios, even if already understood, should contribute to improvement in perceived relative EM creditworthiness during the next several years.

Fiscal balances
The much smaller collapse in public finances during the Great Recession and its aftermath boosted perceived EM creditworthiness greatly. At least in terms of flow measures, however, this story may have played itself out. The IMFs September 2011 Fiscal Monitor contained projections for general government balances through 2016, and these forecasts show the gap between EM and DM fiscal deficits shrinking significantly. From 4.0 percentage points (p.p.) of GDP in 2011, the IMF

current account balances

The creditworthiness of issuers included in the EMBIG depends not only on growth and fiscal performance but also external accounts: Borrowers need to generate hard-currency receipts in order to repay hard-currency debt. In the mid-2000s, EM economies as a
1

Moreover, the January 2012 interim update to the Fiscal Monitor, which includes revised projections through 2013 only, worsens the near-term outlook for both sets of countries, but slightly more for EM than DM (putting next years difference at 2.1 p.p. of GDP, compared with 2.0 p.p. in the September edition).

8 | Digging deeper: Emerging market debt choices

group ran large current account surpluses, with DM economies in deficit. The aggregate EM surplus has since shrunk significantly (to 2.4% of GDP in 2011, according to IMF estimates, compared with a peak of 5.0%/GDP in 2006), suggesting somewhat less (though still perfectly sufficient) dollar-earning capacity. The outlook seems mixed. Global rebalancing, if it occurs, will likely involve some improvement in DM balances and worsening in EM (albeit led, in the latter case, by China, which represents only 2.5% of the EMBIG). Other things equal, one might expect that faster EM growth alongside currency appreciation, even if warranted by productivity gains, would put downward pressure on EM surpluses. On the other hand, rising commodity prices would improve aggregate EM terms of trade (especially for the relatively commodity-focused EMBIG group), blunting the growth and FX effects. Moreover, EM countries and regions currently present a hodgepodge of external-accounts stories, with Central and Eastern Europe running large deficits, Latin America moderate shortfalls, and Emerging Asia sizeable surpluses. No common story seems likely to prevail in the coming years, though IMF projections show only a slightly smaller total EM surplus (as a share of GDP) in 2016 than in 2011. In relative terms, the evolution of current account balances probably represents a minor negative for perceived EM creditworthiness compared with DM during the next several years. What does all this mean for EM credit spreads? As Exhibit 15 shows, sovereign spreads since 2003 have not necessarily tracked perceived relative creditworthiness. Instead, they have behaved broadly similarly to other spread products (including EM corporate), tightening gradually but significantly through early 2007, widening thereafter and eventually blowing out during the financial crisis, rallying sharply as the recovery took hold, then selling off again during the second half of last year
exhIbIt 15: eMbIG sPread over u.s. treasurIes and eM credIt ratInGs
1,600 1,400 Spread (basis points) 1,200 1,000 800 600 400 200 0 1994 1997 2000 2003 2006 2009 BBBB BB+ BBB2012 EMBIG spread over Treasury EMBIG average credit rating CCC+ Credit rating (inverted scale) BB B+

alongside other risky assets. For the entire period, they have averaged 360 basis points, very similar to their current level, but with a standard deviation of 144 basis points. Spreads have traded within 50 basis points of that 360 basis points mean only 23% of days since 2003. The fact that spreads currently stand slightly wide of their medium-term average, despite what almost certainly is high perceived creditworthiness compared with the period as a whole, likely reflects several factors. Spread product in general appears to incorporate a volatility discount at the moment and, in some cases, an illiquidity discount as well (for example, U.S. high yield spreads are trading considerably north of their long-term average despite a low default rate). Very tight spreads in 2006 and early 2007 owed in part to a credit and leveraged-trading boom that may not return. And the very low level of Treasury yields may be playing a role, too, as some investors focus on all-in yield rather than spreads. Can spreads tighten from here, consistent with the likely, if small, further improvement in relative creditworthiness? A move to the 250-300 basis points range certainly seems quite plausible. After all, spreads traded in that area as recently as the first half of 2011. Getting back to those levels, though, would likely require a lengthy extension of the recent drop in overall market volatility. Under those circumstances, probably associated with an improved global growth outlook, Treasury yields would likely increase, chipping away at the gains from spread tightening (a move to, say, 275 basis points from todays roughly 375 basis points would, if Treasuries did not move, generate about a 7% gain for the EMBIGbut this increase would be countered by a 100 basis points rise in Treasury yields).

CEMBI, the default rate, and U.S. corporate spreads

The relatively small number of issuers included in CEMBI336 total, with about one-third of that figure in the high-yield bucket, compared with 959 in the JPMorgan Domestic High-Yield Index creates a significantly lumpier default rate than is the case for U.S. high yield (Exhibit 16, on the following page). The average CEMBI HY default rate since 2002, 3.7%, is just slightly above that for U.S. HY during the same era (3.2%). Twice in the past 10 years, though, no members of CEMBI HY defaulted, whereas on two other occasions more than 10% of the high-yield part of the index did. Last years default rate of 0.5% significantly bettered its U.S. counterpart (1.7%). In 2012, by contrast, the CEMBI HY default rate seems likely to top U.S. HY by some margin, at around 5.8% versus 2.0%. That spike owes mostly to two large

Source: J.P. Morgan Securities, Inc., J.P. Morgan Asset Management. Data as of January 31, 2012.

J.P. Morgan Asset Management | 9

Digging deeper

single-name defaults, an illustration of the volatility in the default rate created by the lower number of issuers. Do CEMBI spreads trade with the default rate? Exhibit 17 suggests a genuine if tenuous relationship. The connection, though, falls short of the close tie between defaults and spreads in U.S. HY, where the default rate is less discontinuous and more predictable. The expected CEMBI HY default rate seems best used to identify large deviations from fair value, rather than serving as a tactical trading tool. The EM corporate default history suggests that the high yield component of CEMBI should probably trade slightly wide to U.S. HYas has, indeed, generally been true in recent years. The marginally higher EM default rate over time, along with the greater concentration of risk within the EM index, both point to a small discount. Recent global market trends, with liquidity dropping in asset classes less closely linked to core broker-dealer activity, also imply a somewhat wider spread for EM corporates compared with their U.S. counterparts. A possible counterargument might be that EM corporate credit quality is structurally

improving over time and that, as a result, EM issuers should trade tighter to their current ratings than U.S. borrowers. While appealing given the faster trend rate of economic growth in EM, this view conflicts with the extremely sound corporate fundamentals in the U.S. at the moment and the resulting prevalence of upgrades over downgrades in U.S. HY. CEMBI typically trades at a small discount to its U.S. counterparts. As seen in Exhibit 18, the spread on the high-yield component of CEMBI has averaged 114 basis points over the U.S. high-yield index since 2003, though with considerable volatility around that mean (including extended periods in which CEMBI HY has traded through U.S. HY). The investment-grade portion of CEMBI has averaged 65 basis points over U.S. IG corporate during that same period (Exhibit 19). At the moment, the gap between the high-yield component of CEMBI and U.S. HY is a little narrower than average, though not unusually so, especially considering that U.S. HY itself has recently been trading wider than its currently low default rate would generally imply. The discount attached to CEMBI IG is a bit larger than usual but fairly typical of the last three years. This persistently wider spread may reflect a liquidity premium accruing to U.S. issuers; if so, it will likely narrow only over the medium term.
exhIbIt 18: ceMbI hy and u.s. hy sPread over u.s. treasurIes

exhIbIt 16: ceMbI hy and u.s. hy deFault rates

18 16 14 12 Percent Basis points 10 8 6 4 2 0 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 500 0 2003 CEMBI U.S. High Yield 2,500 2,000 1,500 1,000 CEMBI High Yield U.S. High Yield

Source: J.P. Morgan Securities, Inc. Data as of December 30, 2011.

exhIbIt 17: averaGe annual ceMbI sPread over u.s. treasurIes versus annual deFault rate
16 Percentage points spread (%) 14 12 Basis points 10 8 6 4 2 0 0.00 5.00 10.00 Annual default rate (%) 15.00 20.00 y = -420.1x + 128.22x + 5.0708 R = 0.8577

2004

2005

2006

2007

2008

2009

2010

2011

2012

Source: J.P. Morgan Securities, Inc. Data as of January 31, 2012.

exhIbIt 19: ceMbI IG and u.s. IG sPread over u.s. treasurIes
800 700 600 500 400 300 200 100 0 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 CEMBI IG U.S. IG

Source: J.P. Morgan Securities, Inc., J.P. Morgan Asset Management. Data as of December 30, 2011.

Source: J.P. Morgan Securities, Inc. Data as of January 31, 2012.

10 | Digging deeper: Emerging market debt choices

Conclusion: Putting money to work

Emerging markets debt has performed strongly over the past decade, and the underlying characteristics of its economies and companiesfast growth and good fiscal indicators suggest a reasonable probability of strong returns in the years to come. The particular EMD asset class purchased should depend on an investors overall view of the market, holding period and risk appetite.

exhIbIt 20: duratIon oF eMd asset classes

8.0 7.5 7.0 6.5 6.0 Years 5.5 5.0 4.5 4.0 3.5 3.0 2003 EMBIG CEMBI GBI-EM

Market view
Investors who believe that, over the next five years or so, global growth will likely accelerate back to a reasonably strong pace should favor EMD exposure through the GBI-EM. Local-currency debt tends to outperform when EM FX is appreciating, and this outcome is more likely when global growth is fairly solid. At the same time, satisfactory global growth will tend to put upward pressure on what are currently extremely low U.S. Treasury yields, dampening returns on USD-denominated debt. True, such a growth environment will tend to reduce credit spreads, a plus for USDbased debt, while pushing up local-currency yields, a negative for the GBI-EM. But EMBIG credit spreads are not particularly high by medium-term standards. Indeed, they stand about 300 basis points below their level at the start of 2003, the beginning of the previous high-growth period. During that 2003-2007 period, the GBI-EM outperformed the EMBIG despite a more than 400 basis points spread tightening on the latter. The GBI-EMs outperformance occurred even though its own yield climbed about 100 basis points, in part because the relatively short duration of the GBI-EM (currently 4.6 years) makes its returns less sensitive to yield changes than is the case for the EMBIG (7.1 years) (Exhibit 20). (Note that CEMBI also has a shorter duration than EMBIG, making it somewhat less exposed to rising Treasury yields.) Conversely, an investor who believes the global backdrop will not change much with relatively soft growth and little upward pressure on Treasury yieldsshould probably prefer USD-based EMD, because the FX component of local-currency debt is unlikely to perform well.

2004

2005

2006

2007

2008

2009

2010

2011

Source: J.P. Morgan Securities, Inc. Data as of December 30, 2011.

instead of EMBIG or CEMBI results in a higher-beta portfolio, given GBI-EMs higher correlation with risky assets like the S&P 500 or commodities (mostly because of the dampening effect of the Treasury component in the dollar-denominated instruments).

holding period
Even assuming a relatively upbeat global outlook, local-currency debt outperformance of GBI-EM seems more likely over a lengthy holding period of at least a few years than in any particular year, thanks to the unpredictable nature of EM FX moves. True, the volatility of annual returns on the EMBIG is slightly higher than that of the GBI-EM. But even in the 2003-2007 period in which GBI-EM significantly outperformed on average, the EMBIG did better in two of the five years. And during the 2003-2011 timeframe, GBI-EM experienced two down years, versus just one for EMBIG. Given the uncertainty that surrounds the global view, and the likelihood that U.S. Treasury yields remain fairly steady at least for the next year or two, investors with short-term horizons of a year or so should probably favor USD-based debt.

special interests
CEMBI represents an appealing investment for portfolio managers with particular interests in assets that are currently trading cheap to medium-term averages. Even following a significant rally at the start of 2012, the CEMBI spread stands roughly 80 basis points wide to its 2003-2011 average (as is also the case for U.S. high yield). CEMBI also should appeal to investors who especially wish to gain exposure to Emerging Asia, the planets fastest-growing region. Asian issuers make up 40% of CEMBI, versus 29% of GBIEM and just 19% of EMBIG. While CEMBI contains a relatively small number of companies in the technology and consumer fields that represent a significant part of EM Asias appeal, it nonetheless provides reasonable sector diversification, and its members should benefit from the favorable overall EM Asian outlook.

diversification versus beta

An investors relative interest in diversification benefits compared with keeping portfolio beta low should also play a role in the EMD choice. The GBI-EM shows significantly lower correlation with its closest comparable in a typical portfolio international fixed-incomethan the EMBIG or CEMBI do with theirsU.S. corporate debt. If an investor would fund a GBI-EM purchase by selling international fixed-income and an EMBIG or CEMBI purchase by selling U.S. corporates, the diversification benefit is larger for the GBI-EM. At the same time, buying GBI-EM

J.P. Morgan Asset Management | 11

Digging deeper

recent PaPers
November 1, 2011. Price/earnings investing: One picture requires a thousand words, P. Rappoport, A. Barrett November 9, 2011. Risky businessdont de-risk, right risk, P. Sweeting Michael Hood Market Strategist, J.P. Morgan Asset Management
http://www.jpmorgan.com/institutional/marketviews

November 13, 2011. Twelve in 2012: What Politics Can Mean for Markets in the Year Ahead, R. Patterson December 1, 2011. Long-term capital market return assumptions: 2012 estimates and the thinking behind the numbers, D. Shairp, A. Werley, M. Feser, E. Efeyini, G. Koo, A. Sheikh, J. Warner. January 2, 2012. Top investment questions for 2012, R. Patterson, M. Hood, K. Franceries, P. Sweeting, P. Rappoport, A. Valluri February 8, 2012. Twelve in 2012 Update: How political events are shaping markets, R. Patterson

Contributor Annapurna Valluri, Vice President

Important Disclaimer FOR PROFESSIONAL INVESTORS ONLYNOT FOR RETAIL USE OR DISTRIBUTION. `Professional investor means the definition ascribed to it within the european union directive 2004/39/eu on markets in financial instruments (mifid). Opinions, estimates, forecasts, and statements of financial market trends that are based on current market conditions constitute our judgment and are subject to change without notice. We believe the information provided here is reliable but should not be assumed to be accurate or complete. The views and strategies described may not be suitable for all investors. References to specific securities, asset classes and financial markets are for illustrative purposes only and are not intended to be, and should not be interpreted as, recommendations. Indices do not include fees or operating expenses and are not available for actual investment. The information contained herein employs proprietary projections of expected return as well as estimates of their future volatility. The relative relationships and forecasts contained herein are based upon proprietary research and are developed through analysis of historical data and capital markets theory. These estimates have certain inherent limitations, and unlike an actual performance record, they do not reflect actual trading, liquidity constraints, fees or other costs. References to future net returns are not promises or even estimates of actual returns a client portfolio may achieve. The forecasts contained herein are for illustrative purposes only and are not to be relied upon as advice or interpreted as a recommendation. International investing involves a greater degree of risk and increased volatility. Changes in currency exchange rates and differences in accounting and taxation policies outside the U.S. can raise or lower returns. Also, some overseas markets may not be as politically and economically stable as the United States and other nations. The risk of investing in foreign countries is heightened when investing in emerging markets. In addition, the small size of securities markets and the low trading volume may lead to a lack of liquidity, which leads to increased volatility. Also, emerging markets may not provide adequate legal protection for private or foreign investment or private property. Fixed income securities will increase or decrease in value based on changes in interest rates. If rates increase, the value of the investments generally declines. On the other hand, if rates fall, the value of the investments generally increases. Your investment will decline in value if the value of the investments decreases. Securities with greater interest rate sensitivity and longer maturities tend to produce higher yields, but are subject to greater fluctuations in value. Usually, the changes in the value of fixed income securities will not affect cash income generated, but may affect the value of your investment. Investing in high yield securities involves high risk securities (also known as junk bonds) which are considered to be speculative. These investments may be issued by companies which are highly leveraged, less creditworthy or financially distressed. Although non-investment grade debt securities tend to be less sensitive to interest rate changes than investment grade debt securities, non-investment grade debt securities can be more sensitive to short-term corporate, economic and market developments. During periods of economic uncertainty and change, the market price of high yield securities may be volatile. Furthermore, though these investments generally provide a higher yield than higher-rated debt securities, the high degree of risk involved in these investments can result in substantial or total losses. J.P. Morgan Asset Management is the brand for the asset management business of JPMorgan Chase & Co. and its affiliates worldwide. This communication is issued by the following entities: in the United Kingdom by JPMorgan Asset Management (UK) Limited which is regulated by the Financial Services Authority; in other EU jurisdictions by JPMorgan Asset Management (Europe) S. r.l., Issued in Switzerland by J.P. Morgan (Suisse) SA, which is regulated by the Swiss Financial Market Supervisory Authority FINMA; in Hong Kong by JF Asset Management Limited, or JPMorgan Funds (Asia) Limited, or JPMorgan Asset Management Real Assets (Asia) Limited, all of which are regulated by the Securities and Futures Commission; J.P. Morgan Asset management (Canada) Inc. is registered as an Exempt Market Dealer, Portfolio Manager and Investment Fund Manager in all Canadian provinces and territories, except the Yukon; in Singapore by JPMorgan Asset Management (Singapore) Limited which is regulated by the Monetary Authority of Singapore; in Japan by JPMorgan Securities Japan Limited which is regulated by the Financial Services Agency, in Australia by JPMorgan Asset Management (Australia) Limited which is regulated by the Australian Securities and Investments Commission and in the United States by J.P. Morgan Investment Management Inc. which is regulated by the Securities and Exchange Commission. Accordingly this document should not be circulated or presented to persons other than to professional, institutional or wholesale investors as defined in the relevant local regulations. The value of investments and the income from them may fall as well as rise and investors may not get back the full amount invested.
270 Park Avenue, New York, NY 10017 2012 JPMorgan Chase & Co. | Mini WP_Digging deeper_EMD

jpmorgan.com/institutional