Sovereign borrowing can generate risks for essentially two reasons. The first relates to the link between sovereign debt and the probability of financial and debt crises. The second relates to the constraints that sovereign debt places on the ordinary conduct of monetary and fiscal policies.[1] Chapter 11 made the point that, among the determinants of country risk, the structure of debt may be more important than the level of debt. The same is true for the risks of sovereign finance, for which the quality of debt matters more than the quantity of debt. In this context, debt quality refers to the degree of risk associated with any given level of debt. Debt quality depends critically on two dimensions: denomination and maturity. The premise here is that the currency and maturity composition of the debt stock largely determines the debt burden relative to the country’s repayment capacity at each point in time. In particular, currency and maturity composition determine the probability of a debt default and, as a result, the sovereign’s borrowing costs and their sensitivity to both domestic and external factors—as well as to self-fulfilling runs.

Hence, any analysis that aims at offering diagnostic and policy advice on how to limit the risk of sovereign finance should focus on policies with the objective of improving debt structure, so that sovereign borrowing becomes an instrument of growth rather than an impediment to it. Devising such policies (the subject of the next chapter) requires a thorough understanding of the sources of risk.

Before looking into its underlying sources, it is necessary to define risk in an operational way that makes it possible to assess quantitatively the policy responses to particular sources of risk. Unfortunately, this is easier said than done. In the context of efficient financial markets, the most natural measure of sovereign risk (understood as the probability that the issuer does not comply with the terms of the debt contract) is provided by the yield paid by sovereign debt instruments in a continuously trading secondary market. This measure, however, has conceptual and practical drawbacks. The conceptual problem is that it refers to a narrow definition of risk that overlaps only partially with the risk of sovereign finance. More precisely, it measures the risk from the bondholder’s perspective, which is narrower than the risk for the sovereign issuer as defined above. The practical problem is data availability: secondary market prices are limited to the subsample of developing countries that have issued a large stock of liquid global bonds (the emerging economies).[2] An alternative measure of the same concept is provided by the sovereign ratings assigned by credit agencies. While still subject to the first criticism, ratings have the advantage of being available for a larger set of countries. The caveat here is that ratings are highly influenced by outcomes and therefore may overstate risk in the upswing and understate it when the situation deteriorates (see Chapter 5).

Conceptually, the most accurate way of assessing the risk to the sovereign entails an evaluation of the influence of debt size and composition on economic performance (output growth and volatility). This chapter presents some evidence in this direction, but the reader should keep in mind that the causal relationship between debt and economic outcomes is bound to run in both directions, and a quantitative identification of these mutual influences is extremely difficult.

CURRENCY AND MATURITY RISK

There is a large literature showing that currency denomination can substantially increase the risk of sovereign finance. This literature has shown that, in the presence of foreign currency debt, net debtor countries have an aggregate currency mismatch, and a depreciation in the real exchange rate will increase the stock of net liabilities in terms of the national product, whereas a real appreciation will reduce it, creating a potentially perilous exposure to currency fluctuations.

Debtor countries can adopt policies aimed at eliminating the mismatch or preventing it from arising in the first place, but often at considerable expense. For example, they can try to change the denomination of their debt from foreign to domestic currency; however, in the short run, such a policy may sometimes turn out to be exceedingly expensive or unfeasible. Alternatively, countries can eliminate the mismatch by deciding to borrow only to the extent allowed by the supply provided by domestic currency markets, which would in most cases entail a significant reduction in net borrowing. Such self-restraint does not eliminate the problem, however, as countries still need to manage their outstanding stock of debt. Furthermore, countries that decide to become financially autarchic will not be able to benefit from the opportunities for risk diversification and access to resources offered by foreign borrowing. As an intermediate strategy, countries may decide to eliminate the mismatch implicit in short-run flows by accumulating foreign reserves. In this way, they ensure that they will not face a foreign currency liquidity shortage in the near future, at the cost of financing low-return reserves with high-cost sovereign debt.

The currency denomination problem is particularly important in Latin America and the Caribbean, a region characterized by narrow domestic markets and a dependence on external funds. Because the economies of the region are also relatively closed in terms of international trade, balance sheet effects arising from adjustments of the real exchange rate become magnified (Calvo, Izquierdo, and Talvi, 2005). There is evidence that a large share of foreign-currency-denominated external debt is associated with lower credit ratings, higher volatility of both GDP growth and capital flows, and limited ability to conduct an independent monetary policy (Eichengreen, Hausmann, and Panizza, 2005b). Additionally, foreign currency external debt increases the sensitivity of spreads to real exchange rate fluctuations (Berganza and García-Herrero, 2004) and leads to contractionary devaluations (Bebczuk, Galindo, and Panizza, 2006).

Furthermore, although there is incomplete cross-country data on the structure of domestic public debt, there seems to be a close correlation between the dollarization of government debt and bank loans (Cowan et al., 2006), and dollarization of the domestic bank sector also presents macroeconomic risks. Dollarization of bank loans increases a country’s propensity to suffer disruptive sudden stop episodes (Calvo, Izquierdo, and Mejía, 2004) and leads to high relative price volatility and thus macroeconomic instability (Calvo, Izquierdo, and Loo-Kung, 2005). Along similar lines, the dollarization of bank deposits increases financial fragility and leads to higher output volatility and lower economic growth (Levy Yeyati, 2006a).

Another major factor that affects the risk of sovereign borrowing is the maturity structure of debt. Short-term debt, by bunching debt payments together (specifically, by increasing the size of the obligations maturing at each point in time) deepens rollover risk and paves the way to possible debt crisis. This is especially true when debt is denominated in foreign currency, but even if all debt is denominated in domestic currency and the government is running a primary budget surplus, the bunching of payments induced by short-term debt creates a financing gap that opens the door to self-fulfilling liquidity runs (Obstfeld, 1994).

Most of the literature on debt crises has focused on the combination of these two risks and shows that short-term external (hence, mostly foreign-currency-denominated) debt is a strong predictor of debt crises. In particular, studies have found that the higher the ratio of a country’s short-term dollar debt to international reserves, the higher the probability of a crisis in that country (Manasse, Roubini, and Schimmelpfennig, 2003). This provided a rationale for what became known as the Guidotti-Greenspan rule of reserve adequacy, which states that countries should always hold enough reserves to cover at least their external liabilities falling due within one year.[3]

Not surprisingly, there is much less evidence of a link between debt crises and short-term domestic debt denominated in domestic currency. This is because, in the event that the sovereign’s capacity to pay is severely affected, domestic currency debt offers an alternative to outright default: the government can keep servicing debt by “printing money” at a seemingly negligible immediate cost. The result, of course, is inflation and also higher interest rates in anticipation of the expected inflation over the life of the bond. If inflation wins the race against nominal interest rates—which is always possible but may require accelerating inflation until levels of hyperinflation are reached—the real value of the debt is effectively diluted, and the government regains solvency. In fact, in most cases, governments faced by a large amount of domestic currency debt have preferred to recur to inflation (sometimes to hyperinflation) rather than defaulting (the Russian default on ruble-denominated GKOs is a notable exception).[4]

Does this mean that short-term domestic debt denominated in domestic currency is less risky than foreign currency debt? Not necessarily. First of all, high inflation is costly economically and socially. It creates uncertainty and reduces growth, and stabilizing from high inflation has been a long and costly process in Latin America. Moreover, inflation hits especially hard the poorest segments of the population, who have more limited ways of protecting their savings from sudden increases in prices and for whom basic necessities can become suddenly unaffordable with a price increase. Second, there have been cases in which the presence of a substantial amount of short-term domestic currency debt was one of the fundamental causes of a financial crisis even though, by the time the crisis came, most of this debt had been swapped into foreign currency.

The 1994 Mexican crisis provides an example of this situation. At the beginning of 1994, Mexico had basically no domestic debt in foreign currency, but it had about 60 percent of its domestic debt denominated in short-term peso notes (called CETES, for Certificados de la Tesoreria de la Federación). During the year, pre-electoral political turmoil, amplified by the assassination of presidential candidate Luis Donaldo Colosio and an insurgency in the state of Chiapas, led to expectations of a currency devaluation and a surge in the interest rate on CETES (which, given their short maturity, needed to be rolled over during the year). In fact, in the month of the Colosio assassination, the rate on CETES jumped from 10 to 16 percent. Deeming a devaluation unlikely to become necessary, Mexican authorities decided to substitute dollar-denominated Tesobonos for some of its CETES holdings. The result was a significant leveraging of risks: if the exchange regime survived the attack on the Mexican currency, the cost of defending the country’s exchange rate peg would have been much lower, but if a currency devaluation became unavoidable (as happened), the government’s losses would be much higher. With the benefits of hindsight, the swap of CETES for Tesobonos was probably a bad decision, but the alternatives (either pay a high real interest rate or accommodate the inflationary expectations by abandoning the peg) were extremely costly from both a political and an economic perspective. These alternatives were determined to no small degree by the presence of short-term domestic debt denominated in domestic currency and by the fact that the Mexican authorities knew well that the arithmetic of diluting short-term debt with inflation can be unforgiving, as the path to high inflation can be gradual, unplanned, and hard to reverse.[5]

IS THERE A CURRENCY-MATURITY TRADE-OFF?

Inflation and currency depreciation risks have been pervasive elements in the structure of government debt in Latin America. Seeking protection against those risks, investors have gravitated towards foreign-currency-denominated debt. Two notable exceptions to the prevailing dollarization cases are Chile, where financial instruments are widely indexed to inflation, and Brazil, where a large fraction of domestic currency debt is indexed to the overnight interest rate. These cases suggest, at an anecdotal level at least, that there is a trade-off between currency and maturity, namely, that countries can avoid foreign currency debt, but only by going heavily to the short end of the maturity spectrum.

At a more systematic level, the evidence on a trade-off between maturity and denomination is more mixed. Firm-level data on liabilities show that in highly dollarized countries (Uruguay, Argentina, Costa Rica, and Peru) there is a strong positive relationship between the degree of dollarization and maturity of debt, suggesting that firms that issue dollar debt are able to extend the maturity of their obligations. Interestingly, this pattern is not present in economies characterized by low levels of dollarization (such as Brazil and Chile), probably because these countries can extend maturity by using other forms of indexation (to prices or interest rates rather than foreign currency). But the sign of the relationship between dollarization and long-termism reverses when a cross-section of countries is examined. Countries that have on average higher degrees of firm liability dollarization have on average a lower share of long-term liabilities in total debt (Kamil, 2004).[6] One possible interpretation of the above results is that country-specific factors like policies and the credibility of policymakers affect in the same direction both dollarization and short-termism, leading to the positive cross-country relationship between these two variables. There is some debate, however, on the extent to which economic policies affect the structure of countries’ debt. Some studies have failed to find a strong correlation between policies and the structure of external debt (see Eichengreen, Hausmann, and Panizza, 2005a). However, Hausmann and Panizza (2003), Jeanne and Guscina (2006), and Mehl and Reynaud (2005) find that policies may matter for the structure of domestic debt.

Cross-country government debt data show less evidence of a trade-off between currency and maturity. Although data for 19 emerging market countries suggest that a movement towards lower levels of dollarization is associated with an increase in the share of short-term debt, the effect is completely driven by the behavior of Russia (Figure 13.1).[7] If Russia is excluded from the sample, there is little evidence of a correlation between changes in dollarization and changes in maturity structure. Furthermore, there seems to be no shortening of debt maturities associated with recent dedollarization in Latin America and the Caribbean (Cowan et al., 2006).

From a more conceptual point of view, it can be argued that the currency-maturity trade-off should be analyzed in terms of cost. Namely, governments can always issue long-term, domestic currency debt that pays the currency premium that the market demands. Of course, this premium may be too high, and hence the sovereign will decide not to pursue this option. But what does “too high” mean? If the difference between the interest rate paid on long-term domestic currency debt and that paid on foreign currency debt correctly reflects inflation or devaluation expectations (that is, if uncovered interest parity holds), then the costs of all instruments will be equal ex ante. Why, then, would a sovereign choose to borrow in foreign currency or in short-term domestic instruments?

One reason has to do with the fact that the sovereign cares about the level of inflation, which carries heavy economic and political costs. If the structure of debt affects expected inflation, the government can use dollar debt or short-term debt to provide a signal that it is committed to maintaining low inflation. In this way, it will end up paying the same interest rate as with long-term debt denominated in the domestic currency, but the economy will have a lower inflation rate.

Another reason has to do with the fact that long-term domestic currency debt may indeed be more costly in term of the ex post real cost. This can happen if the maturity or currency premium overstates inflation or devaluation expectations, possibly as a result of information asymmetry between the government and investors. Suppose, for instance, that there are two types of governments: a good one (which desires to avoid inflation and currency depreciation) and a bad one (which has a tendency to use inflation and currency depreciation). Also suppose that the current government is of the good type, but investors do not know this for sure and hence assign a 50 percent probability that the government will inflate and/or devalue the currency. As a consequence, investors will ask a currency/maturity premium that the good-type government knows to be excessive, and hence the government will decide to issue either short-term or foreign currency debt.[8]

A similar situation arises from the absence of a mechanism to commit a government to maintaining low inflation. If the government cannot make a credible commitment to this objective in the eyes of investors, they will demand a premium on local currency debt, and then the best option for the government will be to reduce its borrowing costs by resorting to inflation (Calvo, 1988). This will lead to a situation characterized by two equilibria: in the good equilibrium, credible governments will be able to borrow cheaply long term in local currency and will fulfill lenders’ expectations by maintaining low inflation; in the bad equilibrium, however, noncredible governments will be required to pay a premium and will end up fulfilling lenders’ high inflation expectations. Faced with such a situation, a government characterized by low credibility can do better by borrowing either short term or in an indexed unit, be it inflation indexed or in foreign currency, because it will end up paying the same interest rate but without needing to resort to inflation. In other words, a risky form of debt (such as foreign currency or short-term debt) is a commitment mechanism that can solve the time inconsistency problem faced by governments with low credibility.[9] Tirole (2003) focuses on the currency denomination of private debt and also highlights that risky debt may serve as a commitment device. Similarly, Alfaro and Kanczuk (2006) conduct a welfare analysis that suggests that, under certain conditions, risky debt is welfare improving, and nominal debt may not be sustainable in volatile emerging economies. An additional reason why foreign currency debt may be relatively cheap is that foreign currency debt instruments may be perceived as implicitly senior to local currency debt instruments, as defaults usually are associated with a highly depreciated real exchange rate, which increases the value of dollar claims relative to local currency claims (Chamon, 2001).

Summing up, when a government has low credibility in regard to inflation, short-term domestic currency debt often turns out to be a more feasible option than long-term domestic currency debt. From the point of view of investors, it is reasonable to assume that the risk of an inflation outbreak is increasing over time, simply as a result of increasing uncertainty. There is in fact a tendency for countries with lingering inflation fears to shorten the maturities on their debt instruments or to employ frequently adjustable interest rates. From the point of view of the government, there is also a tendency to prefer short tenors if the authorities expect inflation to stay low and credibility to increase gradually. In this type of situation, governments would not want to lock in high interest rates that incorporate high-inflation fears in long-term instruments and would prefer instead to issue short-term debt until credibility improves and maturity can be lengthened at lower interest rates.

It is critical to note, however, that even if local currency borrowing at long maturities is indeed excessively costly for some of the reasons noted above, there are cases in which foreign currency borrowing still might not be a better alternative. The higher expected cost of long-term local currency debt may be worth paying in exchange for the insurance benefits it provides. The final decision will hinge, more generally, on the price that a sovereign is willing to pay to increase debt quality and limit the risks of sovereign finance.

This suggests a less benign view of why a government may choose not to issue long-term debt in the local currency. Even in the absence of distortions, a fair currency premium would typically incorporate the incidence of a possible sharp devaluation in the future—much in the same way as an insurance fee against sudden real exchange rate adjustments.[10] Finding that the premium exactly compensates for currency risk, a forward-looking policymaker may opt for the safer local currency debt factor because of its insurance benefits. By contrast, a myopic policymaker who cares only about the present will disregard negative events that may materialize when he is no longer in office. If, as usual, the probability of a currency adjustment increases with the time horizon, such a policymaker will find the premium expensive relative to short-term risk and will opt for either foreign currency debt or short-term debt, which will command a lower premium but leave the next government exposed to the risk of sovereign finance.

In addition, one should factor in the positive externalities of issuing domestic currency debt—an aspect ignored in the static cost-benefit framework discussed above—in terms of the development of a new market and the creation of an investor base for domestic currency instruments, which may flourish once credibility is finally built. In the presence of start-up costs, a sovereign may then decide to issue part of its debt in the domestic currency even if costs are excessive, simply to keep open the option of resorting to domestic currency borrowing in the future.

INFLATION INDEXATION

The above discussion has addressed the conventional view that governments with credibility problems face a trade-off between short-term domestic currency debt and long-term foreign currency debt. But neither time inconsistency nor asymmetric information necessarily implies that the only viable option is foreign-currency-denominated government debt. More precisely, the government may have another method for committing to low inflation: by issuing local currency long-term bonds indexed to inflation. Indexed debt is not as safe as nominal debt, because a crisis that causes inflation to accelerate can undermine fiscal solvency, potentially with self-fulfilling implications. But inflation is a slowly moving variable, and from the sovereign’s point of view, inflation-indexed bonds are clearly safer than short-term local currency bonds or long-term foreign currency bonds because they reduce rollover risk and do not generate negative balance sheet effects in the presence of devaluation of the real exchange rate. It is therefore puzzling that Latin American countries (with the exception of Chile) have made such a limited use of price indexation in recent years.[11]

There are three possible answers to this puzzle. The first has to do with negative past experience with indexation. In some countries, past indexation of financial contracts spilled over to the whole economy, generating a situation in which everything was indexed (including wages, pensions, and subsidies). Such a pervasive indexation system amplified inflationary cycles, as increases in prices led to rapid increases in wages, which would then immediately feed back to prices (Bernanke, 2005). As a consequence, the widespread conventional wisdom was that it was best to steer clear of any form of indexation, although that view seems to have been reconsidered more recently.[12]

A second aspect that explains the unpopularity of price indexation has to do with the fact that, while more difficult to dilute than nominal long-term bonds, inflation-indexed bonds are easier to dilute than dollar-indexed bonds. To the extent that the price index is measured with a lag—or averaged over a longer period—it will not fully protect investors from an acceleration of the inflation rate, and hence it will not protect investors from hyperinflation. Whereas this could have been a problem in the early 1990s, when most Latin American countries were emerging from a period with recurrent hyperinflation episodes, it should be less of a problem in the current context, when most central banks in the region have made important gains in monetary credibility.

A third, more practical reason has to do with the fact that inflation-indexed bonds may simply be too expensive given inflation expectations, an anomaly that may be related to a number of distortions. The premium may reflect the incipient nature of the markets for such bonds—associated with the lack of both trading volume and specialized traders—the mistrust of government-produced inflation statistics, or simply lack of familiarity with this new instrument. It should be clear that none of these reasons constitutes a fundamental objection to the use of inflation-indexed bonds; rather, they are transient obstacles to the use of such bonds that should be taken into account in the design of policies to promote them. The small-market problem is certainly transitory and can be mitigated with the help of international financial institutions, which can fund themselves using CPI-indexed bonds on the borrowing members’ currencies. Investors’ mistrust can also be attenuated by international financial institutions, which can serve as auditors to increase the credibility of official statistics. In turn, the contractionary devaluations at the core of the implicit seniority problem should be gradually eliminated by dedollarization. It is thus not surprising that several Latin American countries have started to rely more on domestic debt indexed to prices as a mechanism for lengthening maturities while avoiding currency risk.[13]

OTHER SOURCES OF RISK

So far, the discussion in this chapter has focused on rollover and exchange rate risks. However, developing and Latin American and Caribbean countries are also subject to several other sources of risk. The most important ones are terms-of-trade shocks; catastrophic events (such as earthquakes and hurricanes), which can have a very high cost in poorly diversified small countries; and, more generally, high output volatility.

A simple exercise can illustrate how different debt structures can affect the evolution of public debt in an economy that faces these different types of risk (details are provided in Box 13.1). Consider the public finances of a government that faces uncertainty in regard to the key economic variables that determine its financial position, including interest rates, the exchange rate, and economic growth (which determines tax revenue and, indirectly, the fiscal surplus before interest payments). To be sure, the volatilities of each of these variables are not independent of one another, but this exercise ignores that complication. The government has choices in terms of the structure of its liabilities. For simplicity, these choices are limited to three instruments: a foreign-currency-denominated bond, a domestic-currency-denominated bond, and a bond whose payments are linked to the evolution of GDP.

Figure 13.2 plots the probable evolution of a country’s debt-to-GDP ratio over time, after the government makes a choice regarding debt structure and the level of spending. The line in the center of the figure is the country’s projected debt-to-GDP ratio if there are no shocks to the economy. In this case, the debt-to-GDP ratio is projected to remain constant (at 0.45). But when economic uncertainty is considered, a “fan” of possible values for the debt-to-GDP ratio opens up. The fan reflects the fact that in the face of uncertainty, the variance is larger at points more distant in the future. The figure shows the range of values of the debt-to-GDP ratio that may be obtained with an 80 percent probability under various mixes of currency and indexation. When all of a country’s debt is denominated in foreign currency, the fan is the widest. When the country’s debt structure is composed of foreign currency and domestic currency instruments in equal parts, the fan is less wide, and when its debt structure comprises equal parts of foreign-currency, domestic-currency, and GDP-linked instruments, the fan is even narrower. This illustrates how a government can reduce the variance of its debt ratio or, equivalently, how it can reduce the fiscal adjustment that would be necessary to preserve debt sustainability if a negative shock were to occur.

This example assumes that the government can issue three different types of instruments at a cost that roughly reflects the expected return to investors. This is not always the case, because numerous factors—like the liquidity of the instrument and concerns that investors may have about possible capital controls, the manipulation of the exchange rate or economic statistics that determine the value of the asset—may affect the cost of instruments other than foreign-currency-denominated bonds issued in global markets. Further, in the case of innovative instruments, there is a “novelty” premium to be paid, as investors may prefer to stay away from less familiar assets unless the expected payoff is sufficiently rewarding. Debt management must balance these costs of more favorable instruments against the gains of a more resilient debt structure and protect the country against the need for costly fiscal adjustment in times of economic malaise.

DOES IT MATTER WHERE TO ISSUE?

Another possible source of risk is the jurisdiction where bonds are issued. There are two ways in which the jurisdiction in which sovereign debt is issued could influence the risk of sovereign finance. The first relates to the fact that, for emerging markets, most financial assets are denominated in the currency of the market in which they are issued, with only a few exceptions. If there are deep-seated reasons for this feature of global financial markets, such as perhaps the absence of deep markets for many emerging economies’ currencies in the major global markets, any dedollarization effort must involve primarily the development of domestic markets, as has been mostly the case recently in Latin America.[14]

A second aspect relates to the differential legal treatment that a sovereign may receive from local and international courts in the event of a debt restructuring. Domestically issued sovereign debt may be thought of as a way of reducing overlending, under the assumption that, unlike in courts abroad, international lenders cannot enforce any of their rights in domestic courts and will therefore be more reluctant to volunteer funds (Bulow, 2002). Furthermore, if domestic courts are more amenable to a debt restructuring that is believed to improve the country’s welfare, they may take into account the social costs of debt service and legitimize a renegotiation under more favorable conditions than a judge in New York or London.[15] While this may make domestic debt less risky than external debt from the perspective of the sovereign issuer, it will also make it more expensive. Figure 13.3 is consistent with the idea that issuing in international markets is marginally cheaper than issuing in domestic markets, but the difference is small and could be due to a variety of factors (like tax treatment, regulatory issues, and market segmentations) other than a difference in risk.[16]

CAN RISKS BE REDUCED?

Most discussion on the risks of sovereign finance has focused on the problems related to currency and maturity risks and on a possible trade-off between these two forms of risk. Countries can try to avoid having to make such a trade-off using forms of debt, such as inflation-indexed long-term bonds or bonds indexed to real variables (like the GDP), that expand the efficiency frontier, but there are several reasons why countries still make only limited use of these (apparently superior) forms of financing. The next chapter will focus more specifically on policies aimed at removing the perceived obstacles to countries’ use of these forms of financing.