Emerging market debt is called emerging

Transcription

1 Assessing Value across Emerging Debt Markets Tina Vandersteel, CFA Partner, Fixed-Income Portfolio Manager GMO Boston Emerging market debt offers multiple opportunities for investors across credit categories, interest rates, and currencies. From a credit perspective, it is often the case that dollar sovereign bonds and, in some cases, quasi-sovereign bonds offer the best value. When considering interest rates and currencies, cross-sectional approaches are useful because they mitigate the effects of duration and dollar timing. Emerging market debt is called emerging because it deals with default-sensitive countries that is, countries where an investor lends a dollar but may not get a full dollar back because the country defaults and writes down the debt, the currency devalues, and/or there are capital outflow restrictions for some period of time. Regardless of the reason, it is risky debt. When thought of this way, it becomes important to determine the relevant risk factor exposures in the space. Table 1 shows risk factor exposures for US dollar sovereign debt, local currency sovereign debt, and emerging corporate debt. All three are exposed to some degree of sovereign default risk. Now, there are many names for this particular risk exposure; when I started at Morgan Guaranty Trust a quarter-century ago, we called this crossborder risk. Table 1. Risk Factor Exposures for Emerging Market Debt Subclasses Risk Factors External Local Corporate Sovereign default Corporate default USD rates Local rates Foreign exchange In essence, an investor is going to cross the border, and something bad may happen. Today, this risk is referred to as emerging market risk or, in the language of our funds prospectuses, risk This presentation comes from the Fixed-Income Management 2014 Conference held in Huntington Beach, California, October 2014, in partnership with CFA Society Orange County. of non-us investments. They state that investment in non-us issuers or securities may involve special risk because of non-us economic, political, and legal developments. They go on to highlight the following risk factors: favorable or unfavorable changes in exchange rates (such as the rout in some emerging market currencies in the last 12 months); exchange control regulations, including currency blockage (as in Argentina and Ukraine recently and in Iceland and other countries during 2008); expropriation (as in Cypriot bank deposits in 2012), nationalization, or confiscatory taxation of assets; imposition of withholding or other taxes (as in Brazil s IOF [financial transactions] tax); adverse changes in investment capital or exchange control regulations; political changes; and diplomatic developments, including the imposition of economic sanctions (as in Russia). As investors, we assess these risk factors against the rewards offered in terms of higher yields, and each type of emerging debt has a different combination of these factors. Before I get into assessing emerging market countries in terms of value, I want to contrast the product view of emerging market debt with the investment view. Emerging Debt: The Product View From a product view perspective, the emerging market debt space can be described in terms of dedicated US dollar sovereign debt, sovereign local currency debt, and corporate debt and then the more recent aggregate and unconstrained varieties, referred to as blended debt and emerging unconstrained debt. This product view roughly parallels the development of the emerging fixed-incomerelated benchmarks in the investing universe. 62 First Quarter GMO cfapubs.org

2 Assessing Value across Emerging Debt Markets The emerging market debt asset class started in the early 1990s with US dollar sovereign debt. But soon enough, US dollar sovereign debt spreads became fairly narrow, and investment managers began to cheat against the index by buying higheryielding local currency debt. Back then, many of the emerging currencies were pegged to the US dollar, so unless the peg broke, the higher-yielding local currency debt could outperform comparable dollar debt. Of course, investors caught on, so in 1996, J.P. Morgan produced an index called ELMI (Emerging Local Markets Index) to track the performance of passively investing in emerging currencies. This was not local currency debt per se because these countries had very nascent capital markets; it was just a currency index. Some 10 years after the Asian financial crisis of , the sell side came back and talked about local currency debt again. Around 2007, J.P. Morgan came out with an emerging market corporate index and tried to get people seriously interested in the idea of dedicated emerging market corporate bonds. Although the sell side may tout emerging corporate bonds as a diversification play, in terms of fundamentals, it does not work that way. First, there is a lot of sovereign risk embedded in these corporate bonds: think of the corporations in Ukraine these days, which may be financially sound but cannot buy foreign exchange to service their foreign debts. Second, these corporations have a lot of overlap with emerging market equities. Third, the US interest rate exposure is identical to (but merely shorter duration than) US dollar sovereign debt hardly diversifying, in my opinion. It does not mean that corporate emerging bonds are never worth owning, but the compensation for owning them needs to rise (because they are less liquid and carry idiosyncratic default risks relative to the sovereign) rather than fall (which would happen if they were truly diversifying). Finally, there are the blended and global unconstrained product categories. I like to say that in terms of an investment opportunity set, we at GMO have always thought in blended and unconstrained ways. From a performance measurement standpoint, we insist on stating our products benchmarks. We pride ourselves on the fact that a 50/50 blend of our dedicated US dollar and local currency strategies has outperformed the new breed of blended (with 50/50 benchmarks) and unconstrained (with cash benchmarks) strategies. What leaves us confused is the idea of emerging unconstrained. Is it a strategy that is unconstrained except for the constraint of being in emerging? Panel A of Figure 1 presents benchmark returns from a product view for the time period. Again, I take the view that investment results should be examined at a risk factor level and not at a broad index level, as is common from a product perspective. Decomposing total returns into risk factor returns relating to such factors as sovereign credit risk, corporate credit risk, foreign exchange risk, and local interest rate risk allows for more meaningful assessment of investment managers. Panel B shows the performance of emerging country debt over the same time period but from a risk factor perspective. The Investment View: Value Opportunities The investment view of emerging market country debt considers the risk factors and yield spreads across the space. Again, in Figure 1, it is important to note that the common risk factor across the three emerging market debt subclasses is sovereign default risk credit risk of these emerging market sovereigns. Sovereign Credit Risk. To assess emerging market country debt from an investment perspective with respect to credit risk, consider Mexico, a solid investment-grade sovereign. Figure 2 illustrates a typical representation of risk and reward for emerging market debt. On the x-axis is spread duration, the sensitivity of a bond price to small changes in sovereign spreads. On the y-axis, we plot spreads over the US dollar LIBOR curve that is, what investors are getting paid for taking on credit risk. Figure 2 plots the Mexican sovereign curve in USD, EUR, GBP, and MXN as well as quasi-sovereign Pemex issues and the issues of two corporations (América Móvil and Televisa). Pay close attention to the USD-denominated Mexican 100-year bond. As a portfolio manager who wants to be invested in Mexico over the long run, I care about default risk and loss given default and not some local volatility estimate, such as spread duration. Figure 2 shows the credit curves to be upward sloping as cumulative default probability grows with time. It is important to note that the instantaneous default probability is the same for all of these bonds on the curve. That is, if Mexico unexpectedly defaults tomorrow, then all of the bonds cross-default and a par claim results. So, if default risk is the only risk I am worried about, then I want the bond with the lowest price. This explains why, as we watch Venezuelan bond prices fall every day, the bonds that started with higher prices are converging down to the bonds with lower prices. cfapubs.org First Quarter

3 CFA Institute Conference Proceedings Quarterly From a risk reward standpoint, in the no-default case and having no more instantaneous default probability than any of these other bonds, the century bond offers a better reward because it pays 80 Figure 1. Benchmark and Risk Factor Returns, Total Return (%) 50 A. Benchmark Returns USD Sovereign Local Corporate Total Return (%) 50 B. Risk Factor Returns Sovereign Credit Corporate Credit Foreign Exchange USD Rates (EMBIG) USD Rates (CEMBIG) Local Rates Notes: In Panel A, the USD sovereign benchmark is the EMBIG (Emerging Market Bond Index Global); the local benchmark is the GBI-EMGD (Government Bond Index Emerging Markets Global Diversified); and the corporate benchmark is the CEMBIG (Corporate Emerging Markets Bond Index Global). Source: Based on data from J.P. Morgan. 64 First Quarter 2015 cfapubs.org

4 Assessing Value across Emerging Debt Markets Figure 2. Mexican Credit Spread over the US Dollar LIBOR Curve Spread (bps) Hurdle Mexico 100 Year Spread Duration Mexico (USD) Mexico (EUR/GBP) Pemex (USD) Pemex (EUR) América Móvil (A2/A ) Televisa (Baa1/BBB+) Mexico (MXN) Sources: Based on data from J.P. Morgan and Deutsche Bank. bps more than the next nearest USD bond. It turns out that the century bond has the lowest price of all the nearby USD bonds priced slightly above par value whereas the others have prices of 120 and higher. So, if default is the primary risk to be concerned with, the century bond offers the best value. In essence, this century bond becomes the hurdle; every other bond with direct or embedded Mexican sovereign risk local, external, corporate should have to be compared with this hurdle. What are some alternatives at the Mexican sovereign level? I could buy euro- or sterlingdenominated Mexican bonds denoted Mexico (EUR/GBP) in Figure 2 and because of liquidity, they do pay more than their local-dollar peers, but they do not meet the hurdle. Corporate Credit Risk. I could buy corporates, such as bonds issued by Pemex, the 100% state-owned Mexican oil company. Pemex delivers about 40% of its revenues to the Mexican state. From a strategic standpoint, Pemex is to Mexico what Fannie Mae turned out to be to the United States. The United States could not afford to let Fannie Mae go, even though it never guaranteed Fannie Mae s bonds. The same is true, we believe, of Mexico and Pemex. If Mexico defaults, Pemex will default, but Pemex is unlikely to default on its own. Therefore, we believe the default risk for Pemex bonds is effectively the same as it is for sovereign debt. Now, Figure 2 shows that a couple of the longer-dated Pemex bonds denoted Pemex (USD) and Pemex (EUR) trade up to the level of the century bond. The Pemex bonds require a little room for liquidity because they are less liquid than sovereign debt. Furthermore, some additional spread is required of the Pemex bonds to compensate investors for the possibility of idiosyncratic default risk. With that in mind, I think the century bond is the bond that offers better value. What about other pure corporate alternatives? Consider bonds issued by América Móvil, the telecom company owned by Carlos Slim, the world s richest man. One would think that the bonds issued by the company trade at a premium. They do; the bonds do not even pay as much as the Pemex bonds, which does not make sense. Why bear idiosyncratic default risk in América Móvil bonds when a higher spread can be earned on Pemex bonds with virtually no idiosyncratic default risk? Even a move down in credit quality to BBB, as indicated by the Televisa bonds in Figure 2, does not offer a better spread than the century bond. From a practitioner s standpoint, if a manager wants to buy credit in Mexico, he should buy the century bond and go home. Local Currency Bonds. How do local currency bonds stack up to the century bond? Of course, it is important to remember that, relative to the dollar century bond, US investors take on cfapubs.org First Quarter

5 CFA Institute Conference Proceedings Quarterly several additional risks, such as capital control risk, when buying local currency bonds. Figure 2 shows that the local bonds, when measured in US dollar terms denoted Mexico (MXN), offer very small spreads relative to the alternatives. The reason for the low spreads is very simple; in Mexico, the banks and insurance companies consider them to be riskfree assets. Just as US Treasuries are risk free to US investors by fiat, the local currency government bonds are risk-free assets for Mexican investors. Of course, the local currency bonds are definitely not risk-free assets for US investors. So, again, the century bond is a better value. At GMO, our portfolios are positioned such that the hard-currency portfolio owns the century bond and the local-currency portfolio owns the century bond swapped back into Mexican pesos. From the perspective of the local currency portfolio, there are a couple of other things that are attractive. For example, peso-denominated Pemex bonds offer a slightly inferior spread to the century bond. But after considering the effort and leverage involved in swapping the century bond back into pesos, it may be better from a risk reward perspective to just buy one of the peso-denominated Pemex bonds. Liquidity, which is also a common risk across all categories, has been receiving a lot of attention lately. The October 2014 International Monetary Fund Global Financial Stability Report focused on the current market liquidity being poor. Liquidity has a role as a key feature of the emerging debt asset class. Figure 3, which presents the weighted average price bid offer spread across time for emerging market country sovereign debt and corporate debt, illustrates that emerging market sovereign bonds are typically more liquid than emerging market corporate bonds. Of course, this evidence makes the century bond even more attractive. It is important to note that investors interested in local currency debt face a couple of bid offer spreads first in the currency exchange and then in the actual local currency bond purchase. Figure 3 reveals several periods of lower liquidity over time, including the periods around the tequila crisis, the Asian financial crisis, the Long-Term Capital Management blowup, the Lehman Brothers crisis, and the recent European sovereign debt crisis. Exchange Rates. Identifying opportunities in the foreign exchange market is difficult because there is a directional component. At GMO, we start with valuation; absent clear valuation signals, we look at interest rate differentials and price momentum. In terms of valuation, Figure 4 shows the relative position of several currencies based on real exchange rates and current account positions. Moving from left to right, the real exchange rate is getting more expensive; either the currency is appreciating relative to everything else or is compounding in higher inflation. Similarly, moving from top to bottom, the current account position is deteriorating; in other words, the real exchange rate has become so rich now that the current account surplus is diminishing or, as in the case of China (renminbi), the current account deficit is getting worse. At GMO, we pay attention to those currencies that are considered outright cheap or outright expensive based on these factors and disregard the rest. Figure 3. Index Bid Offer Spreads, December 1993 December 2014 Spread Corporate Local Foreign Exchange External Dec/93 Dec/96 Dec/99 Dec/02 Dec/05 Dec/08 Dec/11 Dec/14 Note: Spread is in terms of price points for bonds, the implied yield for foreign exchange. Source: Based on data from J.P. Morgan. 66 First Quarter 2015 cfapubs.org

6 Assessing Value across Emerging Debt Markets Figure 4. Real Exchange Rate vs. Current Account Positions for Currencies 4 3 Cheap Currencies 2 CZK EUR RON Current Account Position HUF INR CAD TRY GBI-EM Global BRL COP SGD IDR JPY ILS CHF USD EGP CNY NGN GBP 3 Expensive Currencies Developed Country Real Exchange Rate Out-of-Benchmark Country Member of J.P. Morgan GBI-EM Global Notes: The units represent Z-scores. CAD is the Canadian dollar; CNY is the Chinese renminbi; INR is the Indian rupee; TRY is the Turkish lira; COP is the Colombian peso; SGD is the Singapore dollar; BRL is the Brazilian real; EGP is the Egyptian pound; NGN is the Nigerian naira; RON is the Romanian leu; CHF is the Swiss franc; ILS is the new Israeli shekel; GBP is the British pound; JPY is the Japanese yen; EUR is the euro; HUF is the Hungarian forint; CZK is the Czech koruna. Source: Based on data from GMO. How do the major currencies fare? According to Figure 4, the sterling rate looks quite expensive, and the dollar has been moving left as the real exchange rate appreciates. The yen and euro have been moving left as well; the euro has also been moving up because of an improved current account position. In some ways, this is merely a statistical artifact; the current account measurement in Figure 4 is relative to the rest of the world. Of course, the current account problem in the eurozone is not relative to the rest of the world but rather relative to the countries within the eurozone for example, Spain and Italy versus Germany. Germany s export machine has simply redirected the collapse in domestic demand in the Portugal, Italy, Greece, and Spain (PIGS) region to the rest of the world. Consequently, the eurozone current account surplus is going up. Therefore, I do not think that necessarily makes the euro cheap. With respect to major emerging market currencies, defined as those currencies that are in the local currency debt index, the pegged Nigerian naira is veering toward expensive and the Hungarian forint is starting to look a little bit cheap partially because of the effect of the euro moving left. The Czech koruna also looks a bit cheap. Figure 5 shows the various currencies ranked on a net signal based on interest rate differentials and momentum. Those currencies whose light blue bars are above the line reflect better-than-average interest rate differentials, and those below the line reflect worse than average. It is not surprising to see that the currencies to the left are those from countries that were forced to raise rates last year after the taper tantrum, when the Federal Reserve signaled that quantitative easing of asset purchases might be reduced earlier than investors expected, causing a rise in US bond cfapubs.org First Quarter

7 CFA Institute Conference Proceedings Quarterly Figure 5. Emerging Currencies Ranked on Net Signal Emerging Currency Relative Value BRL RUB TRY IDR ZAR INR PEN COP THB CLP MYR MXN KRW RON PLN HUF PHP SGD TWD CZK ILS Implied Yields Price Momentum Net Signal Note: PEN is the Peruvian nuevo sol; THB is the Thai baht; CLP is the Chilean peso; MYR is the Malaysian ringgit; MXN is the Mexican peso; KRW is the South Korean won; PHP is the Philippine peso; and TWD is the new Taiwanese dollar. Sources: GMO calculations are based on data from Bloomberg and J.P. Morgan. yields. These currencies are from the Fragile Five countries Brazil, Turkey, Indonesia (IDR), South Africa (ZAR), and India as well as Russia (RUB). Interestingly, Russia finds itself in this awkward situation in which it has been trying to set an inflation target for five years but cannot do it. Russia used to manage the ruble, but now the ruble is going down precipitously, resulting in the threat of inflation. So, it raised rates. Toward the right side of Figure 5 are some unattractive interest rate currencies, most of which are Central European countries. A majority of these countries have compressed their policy interest rates down as low as they can toward the very low European Central Bank rates. The Czech Republic does not even have an interest rate policy anymore; instead, it has created a floor in the 27 koruny/1 euro exchange rate. I believe the next candidate for a floor is the new Israeli shekel. The only reason that the shekel has not gone to a floor yet is that former Bank of Israel governor Stanley Fischer left to join the US Federal Reserve, but I think it is coming. Hungary and Poland (PLN) both score unattractively based on interest rate differentials. But if the euro goes down and they follow, it would be very helpful. Of course, the dollar s recent momentum is a global phenomenon, so all the momentum signals (the darker blue bars) are negative. Thus, from an active strategy perspective, we like the higher yielders and the US dollar relative to the lower yielders. In addition to valuations, interest rate differentials, and momentum, it is important to monitor the electoral calendar in emerging market countries. A number of key countries have just gone through key elections. India and Indonesia ended up with fairly marketfriendly outcomes, but unfortunately, the results in South Africa and Turkey maintain the status quo. South Africa s President Jacob Zuma and Turkey s President Recep Erdoğan are not particularly attractive candidates, but their policies are at least predictable. Local Interest Rates. As might be expected, interest rates in emerging market countries are almost uniformly higher than the zero rates currently practiced in most developed markets. Similar to exchange rates, we think about interest rates in emerging market countries in the cross section. To rank them, we use the same factors that we use in our developed market strategies, such as real policy interest rates and term premiums. Figure 6 plots the real policy rates in the G 3 against real policy rates in various emerging market countries. The highest real policy rate in the world is Brazil. Actually, the rate is somewhat artificially high; almost nobody pays that rate in Brazil. Practically speaking, there is a subsidized lending rate that is paid by most of the important people. 68 First Quarter 2015 cfapubs.org

8 Assessing Value across Emerging Debt Markets Figure 6. Interest Rates of G 3 Countries vs. Emerging Markets Interest Rate Market Relative Value EUR JPY IDR RUB ZAR INR TRY MXN MRY THB HUF SGD ILS USD BRL COP RON PHP PEN PLN CNY TWD CLP KRW CZK HKD Real Policy Five-Year Rate Less Floating Rate Net Signal Sources: GMO calculations are based on data from Bloomberg and J.P. Morgan. But foreign investors who show up and receive a Brazilian interest rate contract get the higher rate. It is not surprising that everybody who is short Japan is also thinking about being long Brazil. From a foreign investor s point of view, Brazil looks great, and in the cross section within emerging markets, it also looks great. With respect to term premiums, we are typically interested in the 10-year term premium. But in a number of emerging market countries, this market is immature and barely goes beyond five years in duration, so then we look at a five-year term premium over the local floating rate. Figure 6 shows Mexico having negative real policy rates, but it is positive in nominal terms and, therefore, has room to cut; it is not at the zero floor yet. Mexico s term premium is quite steep; I like to think of it as a levered play on the United States. Similarly, Romania, which has slightly positive real policy rates, has a very steep term premium that makes it more attractive than, for example, the eurozone. US Interest Rates. The final risk factor directly relevant to sovereign and corporate US dollar debt but also indirectly relevant to local debt is US interest rates. When there is a lot of cushion in US rates, the rates are generally benign forces on risky assets of the kinds we have just discussed. But when rates are very low as they were in mid-2013 investors cannot have the same expectation. Instead, interest rates are a source of continuous danger hanging over the market. No wonder many in the market parse Federal Open Market Committee members every word. Let us consider the 10-year US dollar swap rate over time. Rather than placing an emphasis on the current levels, I think it is more important to think about what the market is pricing in for the future. In the pretaper tantrum period, there was not much cushion at all. The 10-year rate was under 2%, and out 3 years, the market was pricing a level of merely 2.25%. In contrast, in December 2013, the level was 3%, but more importantly, the slope had increased and thus the market was pricing in a rate of over 4% on the 10-year rate out 3 years. So, in that case, there was quite a bit of cushion. Unfortunately, that cushion has come down a bit this year; as of September 2014, the 10-year rate was 2.25%, and 3 years out, it was just over 3%. Conclusion In emerging market debt, it is important not to confuse the opportunity set with performance measurement. Although we are in favor of blended and unconstrained universes, it is important to be able to assess a manager s normal portfolio against a correct benchmark. cfapubs.org First Quarter

9 CFA Institute Conference Proceedings Quarterly Also, a common risk among the various emerging market debt subclasses is credit risk. For credit risk, usually, dollar sovereign bonds and, in some cases, quasi-sovereign bonds are the best alternatives. Periodically, a few corporations sneak through in terms of being attractive alternatives, but it is rare. In most cases, the US investor is better off in dollar sovereign credit than in local currency credit. Finally, currencies and interest rates should be viewed through a relative, cross-sectional lens. In the current environment, I recommend looking closely at the Fragile Five countries because some of them are looking pretty good at this point. This article qualifies for 0.5 CE credit. 70 First Quarter 2015 cfapubs.org

10 Question and Answer Session Tina Vandersteel, CFA Q&A: Vandersteel Question: In the Mexico example, you said that risk should increase over time. Is that true if emerging market countries evolve and get wealthier? It seems that may not necessarily be the case for all emerging market countries. Vandersteel: It is important to distinguish between a decision made in the time series and one made in the cross section. An investment decision today is based on the shape of the curve today, which is upward sloping, indicating that default probability grows with time. Of course, at some future date, the credit curve could be flatter, higher, and inverted, and with those shapes, there are different decisions to make. But at any given point in time, this is the decision-making process. Question: When investing in emerging markets, what is an appropriate time horizon? Do you attempt to time these markets? Vandersteel: The average life of emerging sovereign debt is 12 years, which is probably about the right time horizon to think about when investing in this space. Of course, at GMO, we are buy-andhold investors. Question: Do you know if anyone has been able to demonstrate being good at timing emerging markets both in terms of performance and in terms of the crossover from local to hard currency and back? Vandersteel: I do not know anyone who has been able to do well in terms of timing. Actually, I do not think there is any alpha in that game. So, to sell the Mexican century bond from one portfolio only to buy it back in another portfolio to hedge the currency makes no sense to me. If there is no alpha in the switch and you want blended, then it is better to just buy one unit of each and go home. Question: As someone who is a dedicated emerging market portfolio manager with a long-term view, are you concerned with the large positions that have been built up by those tourist unconstrained fund managers who are not as knowledgable about the space? Vandersteel: That is a very good question. Generally, I would not buy the stuff these folks are buying because they buy it at prices at which I would not buy it. As a dedicated emerging market debt person, I have set a high hurdle: I buy the cheapest bond, and I pass on all of the stuff that is marketed to me. Question: Within the fixed-income or non-equity space, which asset classes do you expect to outperform emerging market debt over the next three to five years? Vandersteel: I can say it is no longer asset-backed securities. But I do not know enough about the US corporate and high-yield markets to make a meaningful statement. I think there are too many unknowns to say anything with precision. One of those important unknowns is default correlation. For example, Argentina took down Uruguay as a result of some default correlation. Analogously, I do not know the current industry composition of high yield and the level of industry concentration. It is relevant because the higher the industry concentration, the higher the default correlation and the lower the recovery value that can be expected to be received. cfapubs.org First Quarter