The concept of fiscal sustainability asks whether current debt levels in a country can be serviced given the government’s current fiscal position. Thus, it makes no judgment as to whether the country’s debt should be different from its current level based on some optimization criterion (see Chapter 10 for a discussion of these issues).

Several of the now-standard approaches to fiscal sustainability were initially devised with developed countries in mind, stressing mostly long-run solvency matters, but ignoring the characteristics of debt and the macroeconomic environment that are typical of developing countries and key for assessing the sustainability of their debt. For example, issues of currency composition of public debt and exposure to large external shocks are fundamental elements that need to be considered in the case of developing countries.

This chapter introduces different fiscal sustainability tools, taking standard sustainability analysis as a starting point, and then moving on to issues that are specific to developing countries. Those issues include fiscal sustainability in the face of uncertainty from various sources, and in particular, the real exchange rate and public revenue fluctuations. As sovereign credit ratings are important benchmarks for investors, the correlation of the fiscal indicators examined here with credit ratings provides a useful measure of the extent to which credit risk analysts focus on these emerging market issues. In other words, what do rating agencies have in mind, directly or indirectly, when assessing solvency risk?

THE STANDARD APPROACH

The meaning of fiscal sustainability is often blurred. A first point to clarify is the difference between solvency and sustainability. Insolvency refers to a situation in which the future paths of spending and revenue do not generate sufficient net financial resources to service the existing government debt. A policy stance is sustainable if “a borrower is expected to be able to continue servicing its debt without an unrealistically large future correction to the balance of income and expenditure” (IMF, 2002a, 4). Thus, solvency is only a necessary condition for sustainability, because solvency can be achieved with very large and costly future adjustments. In other words, sustainability requires achieving solvency without major policy changes.

The starting point for virtually all standard methods of calculating debt sustainability is a government’s current period budget constraint, which states that the portion of debt payments falling due (inclusive of interest) that cannot be covered by the primary surplus is financed with new debt.[1] This assumes away any direct financing of the deficit through money printing by the central bank. In the long run, for debt to be sustainable, the government must be able to meet the following condition:

where d is the level of debt as a share of GDP, r is the real interest rate, g is the growth rate of the economy, and s is the government’s primary surplus as a share of GDP. This condition has an intuitive interpretation: the level of the sustainable debt-to-GDP ratio is such that the primary surplus is enough to cover the “effective” interest cost of servicing it. The effective interest rate is the real interest rate net of the growth rate of GDP. As all the variables are assumed to be constant over time (or to reflect an appropriately computed average), this equation states that the stream of future long-run primary surpluses—appropriately discounted—has to fully cover existing debt levels.[2]

While sustainable debt thus measured is a useful benchmark, debt sustainability does not necessarily require a government to constantly satisfy a debt target calculated in this way. For example, a government could be faced with a temporary shock and might want to use debt markets to smooth it out. To account for situations in which there are short-run deviations from the long-run debt level but debt will eventually return to sustainability, some studies focus on a more dynamic interpretation.[3] For example, some methods evaluate whether a country’s primary surplus tends to increase when public debt increases.[4] Another, more recent approach makes a short-run evaluation of a country’s current fiscal stance by assessing its current debt stock, its current primary surplus, and ongoing interest and growth rates to determine whether they are consistent with convergence to a targeted long-run debt-to-GDP ratio (see Croce and Juan-Ramón, 2003). The advantage of this approach is that it allows for departures from long-run values in the current debt stock and the primary balance. It analyzes whether deviations of the primary surplus from its steady state value, and/or deviations of current debt levels from target debt levels, are sufficiently large to put convergence in peril, thus rendering a policy stance unsustainable.

SUSTAINABILITY IN LATIN AMERICA
USING THE STANDARD APPROACH

Despite some episodes of fiscal stress in some Latin American countries in the late 1990s and early 2000s, the region has become more fiscally sound than in the 1980s. The sharp decline in inflation rates is an indirect indicator of fiscal improvement, showing that governments have not needed to resort to monetary financing of their deficits. Another commonly used indicator of ability-to-pay performance through noninflationary sources is the debt-to-fiscal revenue ratio, which had declined by 2004 relative to the early 1990s. Although on average Latin American countries have improved their ability to pay as measured by this indicator (which fell from about 460 percent in 1991 to 320 percent in 2004), the indicator is still much higher than in developed countries, where the average debt-to-revenue ratio for the general government was slightly over 130 percent by the end of 2004.[5] Looking within the region, this improved performance as measured by lower debt-to-revenue ratios is confirmed for most countries when information for 1991 is compared to that for 2004 (Figure 11.1).

These measures hint at improvements in comparison with the 1990s, but how has the region performed in terms of the standard sustainability approach? To answer this question, a more comprehensive indicator based on the standard sustainability approach was constructed based on equation (11.1). This is the real interest rate that would have been required to make prevailing debt levels sustainable given prevailing primary surpluses, current debt stocks, and average GDP growth rates. It is equivalent to a “slackness” interest rate to the extent that it represents the maximum interest rate consistent with fiscal sustainability. This indicator is convenient, because accurate data on average real interest rates are very hard to come by. Although there are some available measures of effective real interest rates—obtained by computing interest payments as a share of beginning-of-period debt levels, net of inflation—these are subject to substantial measurement problems (and/or lack of information for the early 1990s).6 Figure 11.2 shows values of this variable for each country in the region in 1991 vis-à-vis 2004. The lower interest rates that would have been required, on average, in the early 1990s to make debt sustainable—particularly in a context of higher international interest rates than in 2004—suggest that the countries’ fiscal position was more precarious at that time relative to 2004.

Focusing on more recent periods for which data on effective real interest rates are available, a standard debt sustainability exercise is to compute the “required” vis-à-vis the observed levels of the primary surplus. The required primary surplus is defined as the one that would make current debt sustainable, as indicated by equation (11.1). As a first approximation, required primary surpluses were calculated on the basis of effective real interest rates paid in 2004, average growth for the 10-year interval covering the period 1994–2004, and current debt-to-GDP levels.[7] Results are shown in panel (a) of Figure 11.3, and at first glance they suggest a very favorable scenario, as most countries’ observed primary surpluses are higher than their required levels.[8]

A more conservative estimate of required primary surplus would be based on a longer-term average of interest rates. Effective real interest rates for 2004 are probably far from being long-run real interest rates for several reasons. International real interest rates in recent times have been quite low and may not remain at current levels in the future. Additionally, countries that have recently restructured their debt may also be making lower-than-usual interest payments. For these reasons, and in order to stress-test fiscal performance with respect to interest rate fluctuations, an alternative exercise was performed, this time assuming that countries would eventually roll over their debt stocks at the median real interest rate that prevailed for the period 1992–2004, based on behavior of the aggregate Latin Eurobond Index yield.[9] Recomputed required primary surpluses are contrasted with observed surpluses in panel (b) of Figure 11.3. This time, the results are significantly different, with several more countries now facing required primary surpluses above actual levels, highlighting the fact that current fiscal positions may look much more comfortable than they would if interest rates returned to the higher levels that prevailed in the past. The large fluctuations in real interest rates that Latin American countries have faced in the past—and their profound impact on fiscal standing—suggest that, although fiscal positions have improved, vulnerability to real interest rate increases still remains. It is also worth noting that, at least in the experience of the 1990s, much of the increase in interest rates was due to increases in spreads rather than increases in the risk-free interest rate. Thus, several countries could be as exposed as they were in the 1990s to shocks resulting from increases in real interest rates, even if risk-free interest rates were to remain low.

Furthermore, the current high primary surpluses reflect in part the current favorable conditions for export demand and high commodity prices that increase government revenue. Again, this favorable phase may pass. It is worth noting, however, that under similar favorable conditions in the early 1990s, primary surpluses were lower on average, which indicates that the region has expended significant policy effort toward achieving stronger fiscal positions.

INCORPORATING VOLATILITY

Developing countries differ in many respects from developed countries. To begin with, developing countries face substantially greater volatility in real exchange rates, real interest rates, and terms of trade. Developing countries also frequently have to cope with sudden loss of access to international credit markets, which in many cases is completely unexpected. Additionally, the magnitude of shocks faced by developing countries can be so large and these shocks so persistent that, especially in the presence of relatively low credibility in fiscal policy and budget institutions, countercyclical policies may not be an option at all.

This higher volatility and persistence in shocks introduces various additional sources of vulnerability that must be considered when debt sustainability is analyzed, suggesting that until developing countries are able to reduce volatility, it may be relevant to consider debt levels to be “safe” when they can both be sustained in the long run and withstand the pressure of large shocks. In particular, to the extent that shocks are large and persistent, as, for example, was the case of the capital flow standstill experienced by Latin America in the late 1990s, solvency issues may quickly come into play. Thus, the concept of sustainability should be expanded to ensure solvency not only against realistic adjustment scenarios, but also against realistic external volatility.

These issues can be considered from three different perspectives. The first focuses on financial frictions that have played a major role in recent emerging market crises and their effect on fiscal sustainability. The second addresses the issue that emerging market governments are typically confronted with considerable sources of aggregate uncertainty as they try to assess the pattern of government revenue and expenditure, as well as the level of debt they can afford.[10] A third perspective extends the uncertainty analysis to the evaluation of net worth performance, focusing on novel techniques that account for the valuation of assets.

Currency Mismatches, Sudden Stops, and Valuation Effects

Liability dollarization is a major source of vulnerability to sudden stops in capital flows, and the combination of the two can have devastating effects on fiscal sustainability.[11] For instance, loss of access to credit markets need not be the result of overindebtedness in the context of a good equilibrium, but can instead be the result of an economy’s having fallen into a bad equilibrium triggered by a sudden stop in capital flows (see Calvo, Izquierdo, and Mejía, 2004). This inverse fiscal view—in the sense that there is little that is fiscal in the origin of the crisis—finds support in the fact that sudden stop episodes tend to occur around the same time, and for countries exhibiting a variety of fiscal situations before the shock.

Sudden stops in capital flows force abrupt adjustments in the current account deficit that may require sizable currency depreciation in real terms when the capital flow standstill is highly persistent—otherwise, real exchange rate fluctuations would be short lived, and solvency would not necessarily be at stake.[12] This adjustment may have large valuation effects that multiply the cost of servicing foreign currency debt because of excessive liability dollarization, thus pushing a country’s debt out of the sustainable range.

The best example of the effect of external conditions on fiscal sustainability can be found in the Russian debt default of August 1998 and the spread of this shock to global capital markets, an event that would lead to generalized capital flow reversals in emerging markets. In many ways, the Russian crisis worked as a liquidity shock to international investors, who spread it across different countries as they sold assets in their portfolios to restore liquidity and cover margin calls resulting from collapsing Russian bond prices (Calvo, 2005b). This shock turned into a full-fledged crisis in countries that had two key domestic vulnerabilities: a small supply of internationally tradable goods (such as exports) and domestic liability dollarization (Calvo, Izquierdo, and Mejía, 2004). These vulnerabilities can be summarized in the following “mismatch” ratio (see Calvo, Izquierdo, and Talvi, 2005):

where e is the real exchange rate, B is domestic currency debt, B* is foreign currency debt, Y is output of nontradables, and Y* is output of tradables. Mismatches between debt and output composition can lead to substantial differences in valuation of the debt-to-GDP ratio following depreciation. For example, consider a limit case in which all valuation effects take place on debt only, because debt is fully denominated in foreign currency, and output is fully nontradable (that is, when m = 0). This is the worst-case scenario, in which real exchange rate depreciation hits fully on sustainability. Another case that is particularly relevant is that in which the composition of debt (in terms of tradables vis-à-vis nontradables) matches that of output (that is, when m = 1). When this condition holds, real exchange rate depreciation has no effect on the debt-to-GDP ratio and thus on fiscal sustainability. The effect can be compounded by the fact that contingent liabilities, such as those emanating from banking sector bailouts, can materialize simultaneously. When firms in nontradable sectors are heavily indebted in foreign currency to the banking system, substantial depreciation brings along bankruptcies and an urge for the government to bail out the banking system in an effort to preserve the payments system. Thus, public sector debt can skyrocket once the direct and indirect effects of depreciation add up.

Figure 11.4 shows how Latin American countries ranked in terms of the mismatch ratio defined in equation (11.2) in 1998 (at the time of the Russian crisis) and in 2004.[13] As the figure shows, from a liability perspective, Latin American countries were more exposed to exchange rate fluctuations in 1998 than in 2004. Following the sudden stop episode that plagued the region after the Russian crisis, many countries reacted by reducing mismatches and expanding issuance of debt in domestic currency, in many cases indexed to the CPI.

In terms of ranking within the region, in the healthier range of the spectrum are countries like Chile, with a high level of domestic debt issuance and a large supply of tradable goods. Countries with past inflationary experiences and a tradition of liability dollarization are typically more mismatched (Argentina, Peru, Uruguay).[14] For example, Argentina before debt restructuring had a much lower matching ratio, which has substantially improved since debt restructuring (from 0.15 at the end of 2004 to 0.40 right after debt restructuring in June 2005). Although this index captures the main determinants of exposure to sudden stops, it does not comprise all sources of vulnerability to events of this type. Where domestic currency debt is short term, expectations of currency depreciation can make interest rates soar and have a significant impact on public finances to the extent that devaluation is resisted (see Chapter 13).

The mismatch ratio can be expanded to reflect the interaction of mismatches and debt levels. For example, a highly mismatched country with very low debt levels would not be seriously affected by the aggregate effects of debt valuation following real currency depreciation. For this reason, a second measure was calculated, this one evaluating the debt increase in terms of GDP that would take place were a country subject to real depreciation of 100 percent (see Figure 11.5).[15] Although in general, the ordering is similar to that obtained using equation (11.2), some heavily indebted countries, like Nicaragua, jump up in the vulnerability rankings, while other less-indebted countries, like Guatemala, are on safer ground.

THE UNCERTAINTY APPROACH

Standard sustainability analysis was originally conceived for developed countries, in which volatility issues are generally less important. For emerging markets, it becomes essential to incorporate key sources of uncertainty into fiscal analysis and redefine what should be considered “safe” debt levels, taking into account possible changes in economic conditions. The key question that emerging market governments face is whether their debt levels will still be sustainable given the range of possible changes in the international and domestic economic environments.

Most of the debt sustainability literature has explored the use of stochastic methods to obtain representations of the process that drives the dynamics of public debt or net worth.[16] For example, Barnhill and Kopits (2003) adapt the concept of value at risk used in the finance industry to the analysis of government net worth by computing measures of dispersion relative to present values of a government’s assets and liabilities to determine the value at risk or exposure to negative net worth. In an application to the case of Ecuador, they find that net worth valuation could not resist large shocks without turning negative. Xu and Ghezzi (2003) instead follow a liquidity approach, in which a government may be exposed to depletion of treasury reserves. Through estimation of the processes followed by variables that influence treasury reserves (such as exchange rates, interest rates, and the primary fiscal balance), the probability of default, that is, a depletion of treasury reserves, can be estimated at any point in time.[17] Garcia and Rigobón (2004) estimate the joint behavior of key variables affecting the evolution of government debt and perform simulations of the joint paths of these variables (the real interest rate, GDP growth, the primary deficit, the real exchange rate, inflation, and shocks to debt or “skeletons”), which amount to repeated simulations of the path of government debt. Based on this information, the probability that debt will reach a level considered unsustainable can be computed.

Celasun, Debrun, and Ostry (2006) emphasize that the high-frequency data used in Garcia and Rigobón (2004) for the primary balance may not be a good indicator of a government’s policy stance, because in many cases high-frequency data are quite noisy, as they reflect cash management operations which may differ substantially from true fiscal policy response to changes in the environment. For this reason, Celasun, Debrun, and Ostry separately estimate, using annual data, how the fisc reacts to a key set of variables—a fiscal reaction function—and they combine this with an estimation, based on quarterly data, of the joint behavior of nonfiscal determinants of public debt dynamics (real foreign and domestic interest rates, GDP growth, and the real exchange rate), very much in the vein of Garcia and Rigobón (2004). Using these two pieces of information, they produce “fan charts” indicating the associated potential paths of the debt-to-GDP ratio, which allow for evaluation of the probability that debt as a share of GDP will lie below a particular threshold at any point in time.

Mendoza and Oviedo (2004, 2006) provide a framework that rationalizes why governments may want to impose a debt threshold on themselves. In this framework, a government may want to provide insurance to society by keeping government outlays as smooth as possible (except for inevitable adjustments in times of crises), given uncertainty in regard to public revenue and an environment in which such insurance cannot be bought from financial markets. The framework determines sustainable debt ratios based on the ability of a government to credibly commit to debt repayment, that is, the ability to repay debt even after fiscal revenue hits very low levels for a prolonged period of time. Under these conditions, the government will determine the maximum liability position that it can sustain—or “debt limit”—and will set a contingent plan for adjusting expenditures so as to smooth outlays as much as possible while abiding by the debt limit.

This concept of debt limit is similar to that introduced in equation (11.1) for the standard sustainability approach, except that it considers the primary balance that can be achieved in a fiscal crisis (when revenue is at its minimum and expenditures are adjusted as much as possible in times of crisis). Thus, the debt limit is not the same as the sustainable debt, except in times of crisis.

A key determinant of the debt limit is the volatility of government revenue.[18] As Mendoza and Oviedo show, in general, higher revenue volatility will imply lower debt limits. Figure 11.[6] plots revenue volatility for Latin American and Caribbean countries by computing the volatility of the cyclical component of the revenue-to-GDP ratio for the period 1990–2004.[19] The mean volatility for the sample of developed countries used here is 3.3 percent. A one standard deviation interval around that mean yields volatility coefficients ranging roughly between 2 and 4 percent. Only 6 out of 24 countries in the Latin American and Caribbean sample fall within that range or lower, providing an indication of the significance of the volatility problem facing the region.

The relevance of revenue volatility highlighted by the Mendoza-Oviedo framework becomes apparent in the computation of the minimum tolerable level of government expenditure in a time of crisis that would make current debt levels sustainable (and barely below the debt limit) given historical revenue volatility.[20] This exercise provides an idea of the level of insurance that a government would be able to provide given its current fiscal position (as it represents the level of expenditure that could be sustained during a crisis assuming that current debt levels are barely below the debt limit). Results for the Latin American and Caribbean sample are shown in Figure 11.7, in which the minimum level of primary expenditure during a fiscal crisis is expressed as a share of prevailing primary expenditure levels in 2004. As is clear from an examination of the figure, very few countries in the sample could sustain current expenditure levels in a time of fiscal crisis. On average, current expenditure levels would need to be adjusted by 22.3 percent in times of crisis in order to make current debt levels sustainable at the debt limit. This points to relatively low insurance levels for expenditure in times of crisis, even when current debt levels are used as threshold levels.[21]

Changes in revenue volatility have a strong impact on debt limits. To illustrate this point, Latin American and Caribbean countries were ranked according to Moody’s credit rating and split into two categories: lower-risk and higher-risk countries.[22] Figure 11.8 shows that the average higher-risk country could benefit substantially from reducing its revenue volatility to avoid a fiscal crisis: if it could bring down revenue volatility from 9.5 percent to that of the average lower-risk country (6 percent), it could increase its debt limit from 42 to 108 percent of GDP.[23] Thus, its chances of hitting a fiscal crisis could be greatly reduced.

THE NET WORTH APPROACH REVISITED

While most of the approaches presented in this chapter have focused on vulnerability to shocks to government liabilities, government assets are equally affected by economic shocks. For example, the results of exercises considering the effect of real exchange rate depreciation could change substantially once assets such as oil or copper reserves are added to the equation. A typical representation of the government balance sheet is displayed in Table 11.1. A key component of the balance sheet is the net present value of the stream of future revenues, which can be highly susceptible to changes in the macroeconomic environment and is a key determinant of sustainability. Building upon the framework established by Barnhill and Kopits (2003), Levy Yeyati and Sturzenegger (2006) explicitly estimate the behavior of different types of revenue (income tax, value-added tax, etc.) as a function of GDP growth and the real exchange rate (they also do this for different types of government expenditures).

In this framework, the joint behavior of GDP growth, the real exchange rate, and international interest rates is estimated to produce simulations of future joint paths for these three variables that are replicated several times. Results from each simulation are fed into the estimated equations of revenues and expenditures in order to come up with a stream of revenues and expenditures whose net present value can be calculated and included in net worth estimations. Thus, for each replication of GDP growth, real exchange rate, and interest rate paths, a net worth position can be obtained. After a sufficient number of replications, a distribution for net worth can be constructed. This distribution is quite useful, as it allows for the estimation of the probability that net worth may fall into negative territory, that is, the probability that the government may become insolvent. In contrast to the approach pursued in other studies that model the primary surplus as a function of the three key variables previously mentioned, this approach has as an additional benefit that, through the estimation of separate revenue and expenditure elasticities, simulations can be estimated relative to the actual value of taxes and expenditures, which is tantamount to estimating the elasticity of the primary surplus for the current fiscal policy mix.

With a focus on the particular cases of Argentina and Chile, several interesting results emerge from the net worth approach. An examination of Argentina’s balance sheet shows that bonded debt (i.e., explicit liabilities) adds up to just 8 percent of total liabilities, a result that stresses the relevance of including all elements of the balance sheet when assessing solvency. More interesting are the findings regarding the effects of real exchange rate depreciation. As expected, in the case of Argentina—a relatively dollarized country—net worth falls following depreciation (see Table 11.2). But an appealing claim that Levy Yeyati and Sturzenegger make is that most of this effect comes from the fact that, with real depreciation, the tax base shrinks as a share of GDP, thus highlighting the relevance of income effects on fiscal accounts (however, this statement should be weighed against the fact that dollar liabilities themselves could be responsible for the obtained income effect).

This result differs from that for Chile (see Table 11.2). With a high share of its income base linked to the tradables sector (due in part to resources provided by copper production), Chile would experience an improvement in its net worth position following depreciation (despite a countervailing increase in expenditure).

DEBT SUSTAINABILITY AND CREDIT RATINGS IN LATIN AMERICA

Are these different perspectives on debt sustainability reflected in any way in how solvency risk is perceived by rating agencies? Do agencies focus on measures associated with standard sustainability analysis only, or are recent concerns brought up by the sustainability literature also considered when default risk is assessed?

A first pass at the data in regard to these questions suggests that standard sustainability measures may have an impact on credit ratings. Panels (a) and (b) in Figure 11.9 show the relationship between indicators associated with standard sustainability analysis and credit ratings assigned by Moody’s. Panel (a) shows that the difference between required and observed primary surpluses, using prevailing real effective interest rates in 2004, has a low correlation with credit ratings. However, as mentioned previously, real effective interest rates are subject to several sources of measurement error. Given this shortcoming, credit ratings are contrasted next against “slackness” real interest rates, or the maximum level of real interest rates that would make a country’s current fiscal position sustainable. The results are displayed in panel (b), indicating a much tighter relationship (the correlation coefficient in this case is 0.49).

How about measures associated with currency mismatches? Panels (c) and (d) in the figure provide a preliminary answer. Panel (c) plots the mismatch measure of equation (11.2) against credit ratings, while panel (d) does the same using the augmented mismatch measure that also takes into account debt size. Interestingly, both measures indicate a relatively strong association with credit ratings (the correlations are 0.38 and −0.40, respectively).

Finally, panels (e) and (f) in Figure 11.9 show associations with measures of revenue volatility and minimum expenditures derived from the Mendoza-Oviedo framework. This time there is indication of negative correlation between credit ratings and the volatility of the cyclical component of the revenue-to-GDP ratio, and of positive correlation between credit ratings and minimum levels of expenditures that can be “guaranteed” in times of crisis (correlations are −0.49 and 0.51, respectively). These results open up interesting venues for future research exploring further links between revenue volatility and credit ratings.

Many of these partial, cross-section regression plots may in one way or another be capturing a common element taken into account by credit-rating agencies. However—and despite the limitations of the sample—there is some evidence that, directly or indirectly, these new measures may provide pieces of information which could be relevant in and of themselves, as indicated by a simple regression of credit ratings on the matching measure that takes debt size into account and the measure of revenue volatility previously described.[24]

HAS DEBT SUSTAINABILITY IMPROVED IN LATIN AMERICA?

Latin America made some strides in terms of debt sustainability coming out of the 1980s, which has been aptly termed the “debt crisis” decade. However, several indicators suggest that, on average, the region is far from being safe. To begin with, the region remains vulnerable to interest rate fluctuations. Although some progress has been made in reducing liability dollarization, fiscal sustainability in many countries remains susceptible to large shocks in the real exchange rate, given lingering currency mismatches and relatively high debt levels.

Furthermore, volatility is far from over in Latin America. Periods of systemic capital market turmoil could well return to haunt the region. Although many countries in the region have improved their current account balances, partly as a result of the current bonanza in export prices, they are still exposed to potential interruptions in capital flows that could trigger substantial changes in the real exchange rate and in debt sustainability. In such situations, countries would need to resort to large expenditure adjustments. In this regard, several exercises suggest that reducing revenue volatility could be beneficial in terms of the ability to sustain higher debt levels.

Fiscal sustainability analysis is undergoing a quantum leap with the incorporation of the effects of economic and financial volatility into analysis frameworks. Exposure to volatility in government revenues and real exchange rate fluctuations are particularly important in the Latin American context, as underscored by the fact that credit ratings seem to reflect these vulnerabilities as well. Although progress made in modifying the debt structure in Latin American countries has reduced the exposure to exchange rate fluctuations, debt sustainability could be further enhanced by fiscal and institutional reforms that limit the impact of volatility in government revenues.