Summary, Conclusions, and Implications

1.  Improved policies for debt and liquidity management are increasingly seen as an important element of crisis prevention. Better guidance in this area would be useful to policymakers in managing reserves and debt, to the Fund in the exercise of its surveillance responsibilities, and to market participants in assessing risks. To that effect, this paper identifies the key debt- and reserve-related indicators of a country’s external vulnerability and, where possible, simple benchmarks for external sector soundness. Key economy-wide indicators, and their uses, are listed in Table 1. A number of sectoral indicators are also discussed.

Assessing the adequacy of reserves

2.  Although import- and money-based measures of reserve adequacy have not been very successful as predictors of recent crises in emerging markets, there is a role for them in certain situations:

• The ratio of reserves to imports, as a current account based measure, is useful for countries with no or limited access to capital markets.
   
• The ratio of reserves to monetary aggregates can be useful as an indicator of the potential impact of capital flight, especially in countries with weak banking systems, but needs to be supplemented with an analysis of other possible sources of capital flight (including short-term government liabilities).

3.  For policymakers in countries with significant (but uncertain) access to capital markets, the ratio of reserves to short-term external debt (by remaining maturity) appears to be the single most important indicator of reserve adequacy:

• In using this indicator, it is important for it to reflect a comprehensive measure of external debt, regardless of instrument or currency denomination. No instrument is entirely free of rollover risk; and although domestic currency external debt—in contrast to foreign-currency-denominated or -indexed debt—shrinks in real terms when the exchange rate depreciates, this does not eliminate the risk of liquidity crisis.
   
• A universal indicator should include the short-term external debt of both public and private sectors.

4.  There is some empirical support for a benchmark with a value of one for such a ratio of reserves to economy-wide short-term external debt, across a range of emerging market economies. But such a benchmark would only serve as a starting point for further analysis, based on a country’s macroeconomic situation (including its “fundamentals” and its exchange rate regime) and on the microeconomic conditions that affect the functioning of the private sector (e.g., taxes, implicit and explicit guarantees, banking supervision, the bankruptcy regime). These microeconomic conditions can result in moral hazard, distort institutions’ financing structure, and make sectors more vulnerable to external shocks. On the other hand, to the extent the private sector has unqualified access to international capital markets (as part of it typically does in industrialized countries), its short-term external debt can be discounted in the overall ratio.

5.  In any measure of reserve adequacy, gross reserves need to be carefully measured in line with the internationally accepted definition. To reflect available liquidity, it is reasonable in the indicators to augment gross reserves with contingent lines of credit that are truly usable, and to subtract from them obligations of the monetary authorities that can materialize immediately, such as derivatives subject to margin calls.

6.  The impact of other potential drains on reserves needs to be carefully analyzed and monitored. For example, the reserves to short-term external debt indicator recommended above does not include derivatives that are not subject to margin calls, nor government foreign currency debt to residents. Nevertheless, these can lead to additional drains on reserves. Short-term domestic currency liabilities of the government to residents can also become drains on reserves. It is especially important to consider short-term government liabilities to residents if there are no capital controls, or other means that create a captive market, and the exchange rate is fixed.

7.  Stress tests are closely related to indicators (a benchmark of one for the ratio of short-term external debt to reserves represents a simplified stress test in which the current account balance and access to capital markets are nil for a year), and can be used to supplement the analysis. By allowing consideration of shorter or longer periods of limited access to capital markets, capital flight, FDI flows and the projected current account deficit, stress tests are useful in identifying major liquidity risks, as well as strategies to mitigate them.

Debt indicators

8. Debt sustainability should be assessed in the context of medium-term scenarios: “snapshot” indicators do not take account of the prospects for the growth of output and exports, or for fiscal performance. Differing macroeconomic prospects, as well as differences in microeconomic conditions across countries, limit the scope for cross-country benchmarks.

9.  Among debt-related indicators:

• The paper identifies the ratio of external debt to exports as an especially useful indicator of trends in debt and repayment capacity.
   
• The ratio of debt to GDP is a useful supplementary indicator that takes account of the potential for switching production to exports.
   
• Where public debt is predominant, the ratios of debt to GDP and to tax revenue are especially useful.
   
• The paper cautions against putting excessive weight on debt service ratios, except as regards public sector debt.

10.  The maturity structure of debt is a critical element in crisis prevention. It is to some extent captured in the reserves to short-term external debt ratio discussed above. It is recommended that this ratio be projected forward several years (making simple, but explicit, assumptions on maturity of new debt) to analyze the impact of the maturity structure. Indicators summarizing the maturity structure should focus on relatively homogeneous categories, such as nonconcessional public debt, rather than total debt.

11.  The effective currency composition and interest rate structure of debt (after taking derivative positions into account) are important ingredients in determining external vulnerability. Sharp swings in exchange and interest rates can have severe cash flow and balance sheet effects. However, the importance of dispersion of debt across debtors, the impact of natural or explicit hedges, and a dearth of data on derivatives at a national level, mean that analysis of such vulnerability may need to be focused primarily at the level of sectors or even individual institutions.

12.  Financial institutions are particularly vulnerable to changing exchange rates and to changes in cash flow, such as withdrawal of foreign currency deposits or credit lines by foreign banks. Important indicators of the external vulnerability of the financial sector are gross external liabilities, open foreign currency positions (if significant), and indicators of the maturity and quality mismatch in the foreign currency position that include off balance sheet items such as derivatives.

13.  Directly relevant for the external vulnerability of the corporate sector are imbalances in companies’ balance sheets and cash flow that make them vulnerable to changes in exchange rates. Especially important indicators of such imbalances are the difference between foreign currency liabilities and liquid foreign currency assets as a ratio to equity, and the mismatch in foreign currency cash flow as a ratio to overall cash flow. In the absence of data on such ratios, indicators such as leverage, the ratio of short-term external debt to overall debt, and the coverage of interest payments by operational cash flow can provide useful information, in conjunction with information on the composition and distribution of corporate external debt.

Follow up and data requirements

14.  The staff plans to pursue the issues raised in this paper in its discussions with country authorities. Special attention will need to be devoted to advice on targets for reserves and the size and structure of debt. The staff plans to report the relevant indicators of vulnerability more systematically in its reports. Moreover, the staff plans to strengthen its dialogue with country authorities on the ways in which they create microeconomic conditions conducive to sound risk management by the financial and the corporate sectors (for example, by avoiding tax or regulatory regimes that create a bias toward short-term external borrowing, or by ensuring that institutional structures—including bankruptcy regimes—support contract enforcement). The principles set forth in this paper also have an important bearing on public debt management and could be used in the development of guidelines.

15.  Data availability is critical for the analysis of vulnerability. At present, the availability of the indicators proposed in this paper varies considerably. Some series can be derived from existing statistical methodologies, and some are commonly used in debt monitoring systems. Important steps have been taken and set in train to improve such data:

• within the SDDS, initiatives to improve data compilation and reporting on reserves, external debt, and the International Investment Position more generally;
   
• new international standards for the compilation of external debt and other pertinent data—to be reflected in a new international debt guide which is being prepared by the Inter-Agency Task Force on Finance Statistics; and
   
• improvements in the availability of creditor side external debt data.¹

16.  Other series needed to construct some of the indicators in this paper go beyond either present statistical frameworks or current common debt monitoring practice, notably indicators that are useful for monitoring individual banks or firms. Such information, often basic, is published by individual firms in the context of listings of bonds and stocks on public stock exchanges and frequently redisseminated by specialized private data providers. A first priority could be to encourage disclosure of key information and its ready availability, for example from public exchanges. Further, statistical agencies could be encouraged to (re)disseminate relevant aggregates. Supervisory regulation could also be used to ensure that the necessary information, especially on the less common currency breakdowns, is collected and disseminated by individual institutions, as a means to foster market discipline and as a basis for adequate monitoring by the appropriate supervisory authorities.
 

I.  Introduction

17.  The crises in financial markets in recent years have brought to the fore the importance of sound debt and liquidity management in helping to prevent external crises or to cushion the pace of necessary adjustment. There has been considerable discussion of these issues in recent years, both within the Fund,² and in other fora.³

18.  The increased focus on these topics reflects several motivations. Most importantly, it is incumbent on policymakers to contain the risk of external crisis and to have in place the tools to ensure policies are appropriate. It is apparent by now that there is no simple way of resolving large-scale external crises; the process is costly, uncertain, and complex. Moreover, a crisis in one country can bring large costs in others. Hence the importance of crisis prevention through sound policies, including in the vital areas of debt and reserve management. The clearer the criteria for judging the soundness of debt and reserve management, the clearer the signals that alert policymakers to the need for adjustment.

19.  The present paper attempts to take stock of the current state of knowledge in this area and, in particular, to answer the questions:

What sort of indicators should be used to assess the degree to which a country’s debt and reserve situation makes it vulnerable to shocks? Can these indicators be compared against simple benchmarks to provide a useful test of the soundness of debt and reserve management policies?

20.  Bearing in mind that no indicator is perfect, and qualifications apply to any, the aim is to search for those indicators that are robust and relatively easy to consider. Although the overall framework of analysis of this paper applies to all countries, the primary focus of the search for indicators lies on economies with relatively variable access to capital markets, notably emerging market economies. Indicators, because they capture only a few factors, work better for comparative analysis within groups of countries that are relatively homogenous.

21.  While keeping data availability in mind, the paper does not systematically examine whether data are universally available for the recommended indicators to be calculated. It is evident that many countries still suffer from serious data deficiencies in the area of debt and reserves. To improve this situation, support at the national level for statistical work is essential. At the international level, a number of relevant statistical initiatives have been or are being undertaken: the inclusion of the reserves template and introduction of a new external debt category in the SDDS; the work of the Inter-Agency Task Force on Finance Statistics in updating and harmonizing international guidelines for the measurement of external debt data; improvements in the availability of creditor data; and research by Fund staff into the availability of macroprudential indicators.

22.  The paper is structured as follows. The next section lays out some key concepts, while the two subsequent chapters examine reserve adequacy and debt-related indicators, respectively. The paper concludes with issues for discussion. Several annexes containing data and figures, and elaborating on technical aspects of the discussion, are attached.
 

II.  Some Key Concepts

23.  This paper focuses primarily on the financial relations a country has with the rest of the world—that is, its overall “balance sheet.” The prime focus is on external assets and liabilities (namely, relations between residents and nonresidents), encompassing both the public and private sector. Within this overall balance sheet, the focus is on debt-related liabilities (in contrast to equity) and on reserves.

24.  External debt and reserves affect a country’s external vulnerability through their impact on the country’s ability to discharge external obligations. Inability to discharge obligations may result either from a solvency or a liquidity problem.

25.  Solvency can be defined as the country’s ability to meet the present value of its external obligations.⁴ A perceived lack of solvency leads inevitably to an external crisis, as foreign creditors withdraw and domestic residents seek refuge abroad for their assets. It is relatively easy, but not very helpful, to define a country’s theoretical ability to pay: in theory, countries are solvent as long as the present value of net interest payments does not exceed the present value of current inflows (primarily exports) net of imported inputs.⁵ In practice, countries stop servicing their debt long before this constraint is reached, at the point where servicing that debt is perceived to be too costly in terms of the country’s economic and social objectives. Thus, the relevant constraint is generally the willingness to pay, rather than the theoretical macroeconomic ability to pay.

26.  While a solvency problem almost always leads to a liquidity problem, as capital flees the country, it is also possible for a liquidity problem to arise when countries appear to be solvent. Although a (technically) solvent country typically can service its foreign debt under normal circumstances, it can suffer a “run” on its liquidity as uncoordinated creditors rush for the exit.⁶ In any case, it is difficult in practice to determine whether the cause of illiquidity is insolvency or a coordination failure because “solvency is very much like honesty: it can never be fully certified, and proofs are slow to materialize.”⁷ Hence the importance of foreign exchange reserves and of the maturity structure of external assets and liabilities.

27.  In analyzing the impact of external obligations, a distinction should be made between the public and private sectors. If there is a risk that the public sector will cease to discharge its external obligations, this is likely in itself to lead to a sharp curtailment of capital inflows, in part because it also casts severe doubt on the government’s commitment to allowing private sector debt repayment. By contrast, if (a part of) the private sector is unable to discharge its external obligations, this does not necessarily indicate an external crisis, and, if the problem is contained, should merely lead to lenders, whether domestic or external, carrying the burden of failing loans. However, if private defaults take place on a significant scale, this too is likely to lead to a sharp reduction in capital inflows, and government intervention may follow—in the form of exchange restrictions, a general debt moratorium, or bailouts.

28.  An important implication is that not all external obligations are equal. Relative to public obligations, the weight to be attached to private obligations in risk analysis will depend in large measure on the microeconomic conditions that determine market participants’ risk taking behavior. Such conditions may include implicit or explicit taxes and guarantees on various forms of external borrowing, and weaknesses in institutional structures such as contract enforcement and bankruptcy procedures. These conditions can result in moral hazard, distort the financing structure, and make sectors more vulnerable.⁸ For example, taxes that favor short-term debt financing or implicit guarantees that promote foreign currency funding (be it implicit exchange rate guarantees or expected bail-outs) contribute to financing structures that reduce the ability of individual institutions to withstand shocks, including wholesale withdrawal by creditors.

29.  Uncertainty about such micro conditions can also contribute to external vulnerability. Expectations about the functioning of institutions, such as the bankruptcy regime, and government interventions (bail-outs or exchange restrictions) during crisis circumstances may be volatile and impact foreign creditors’ willingness to provide capital to a country. Given conducive and stable microeconomic conditions, the overall private sector debt burden may well be high, but still not a cause for concern—as is illustrated by the history of a number of industrialized debtor countries.⁹

30.  A final point relates to the importance of the exchange rate regime. A floating exchange rate helps both ensure greater caution in lending and borrowing decisions (as the risk of exchange rate fluctuations must explicitly be taken into account), and contain the cost of external crisis (as, under a fixed exchange rate regime, a crisis may bring an additional cost in the form of a disorderly breakdown of the regime). However, a floating exchange rate regime does not do away with the concern about external obligations. In practice, the response of inflows and outflows to an exchange rate depreciation may be weak—especially in the very short term or if confidence factors dominate—and thus the exchange rate may depreciate substantially in the event of a loss of market confidence. Such a depreciation may bring enormous costs, in terms of both import compression and potentially ruinous balance sheet effects, and would impair the country’s debt-servicing capacity. Thus, even a country with a floating exchange rate should take care not to expose itself to excessive debt service obligations.
 

III.  Reserve Adequacy

31.  Reserves are held for a variety of reasons. Chief among these are maintaining liquidity and allowing time to absorb shocks in situations where access to borrowing is curtailed or very costly. In addition, reserves provide confidence in the authorities’ commitment to the timely discharge of external obligations and to supporting the value of the domestic currency.

32.  The recent financial crises have made clear the very high costs that countries bear when they run short of liquidity, and the question of what is an adequate level of reserves has acquired new prominence. There are two separate issues here. The first is whether indicators of reserve adequacy can be formulated that would be a useful predictor of the incidence and severity of crises. The second is whether a rule or benchmark level for such an indicator can be defined, which countries (or at least some set of countries) could use as a policy guide in this area.

33.  Following a discussion of the definition of reserve assets (gross and net) to be used for indicators, this chapter discusses three types of indicators of reserve adequacy:   ratios of reserves to imports, to monetary aggregates, and to measures of external debt. It then examines whether a benchmark can be established for the reserves to short-term debt ratio¹⁰ and how stress testing can be used to complement an analysis based on indicators. It concludes by reviewing the impact of derivative liabilities of the monetary authorities.
 

A.  Defining Reserves

Assets

34.  Reserves are defined as “external assets that are readily available to and controlled by monetary authorities for direct financing of external payments imbalances, for indirectly regulating the magnitudes of such imbalances through intervention in exchange markets to affect the currency exchange rate, and/or for other purposes” (Balance of Payments Manual, 5^th edition—BPM5). The reserves template for the Special Data Dissemination Standard (SDDS) provides more detailed guidance on what should be included in reserves.¹¹ As detailed in the operational guidelines for the reserves template, if the authorities are to use reserve assets for the financing of payments imbalances and to support the exchange rate, the reserve assets must be foreign currency assets. The guidelines also call for assets to be marked-to-market in line with BPM5, and for gold to be included at market prices.

35.  Contingent lines of credit are not part of reserves,¹² but, if they are truly usable on demand, they can be used to augment the reserve number used in reserve indicators. Such lines of credit must allow immediate drawings without material conditionality, and spreads should not effectively prohibit drawings. Very few countries, however, have such lines, and it should be borne in mind (for instance, in performing stress tests) that drawing on such lines could reduce other inflows, by inducing the providers of credit lines to reduce other exposures.

Netting drains on reserves?

36.  The crises of the last few years have shown the importance of taking account of future drains on reserves, and the question arises how to treat such drains in the “headline” reserves number for use in indicators on which policymakers should focus.¹³ The idea of netting such drains from the headline reserves number raises questions related to time horizon. Just as, on the asset side, the headline number used in reserve indicators should include only liquid assets and assets immediately available, so only drains on reserves that might materialize immediately should be netted out of the headline number.¹⁴ (This would include derivatives subject to margin calls.)¹⁵ Among the drains on reserves that may materialize over a longer time horizon, the short-term external debt of the monetary authorities will be included in one of the indicators discussed below (the ratio of reserves to short-term external debt).
 

B.  Import-Based Measures of Reserve Adequacy

37.  The traditional indicator of reserve adequacy is reserves in months of imports, and this ratio is likely to remain relevant as a simple way of scaling the level of reserves by the size and openness of the economy. Reserves in months of imports of goods and services has a straightforward interpretation: the number of months a country can continue to support its current level of imports if all other inflows and outflows cease. This measure—focused on the current account—is of use especially in judging reserve needs for countries that have limited access and vulnerabilities to capital markets.¹⁶ In addition, import data generally suffer from relatively few measurement problems.¹⁷
 

38.  It is increasingly clear, however, that the appropriate level of reserves cannot be determined without reference to the capital account. Box 1, for instance, illustrates the weak relation between reserves over imports and the occurrence or depth of crises in a recent period of financial turmoil.¹⁸ Reflecting the main sources of possible pressure on the capital account, it is generally considered useful to measure reserves against some monetary aggregate and/or some measure of external debt.
 

C.  Money-Based Measures of Reserve Adequacy

39. Money-based indicators of reserves provide a measure of the potential for resident-based capital flight from the currency.¹⁹ An unstable demand for money or the presence of a weak banking system indicates a greater probability of such capital flight. In these circumstances the ratio of reserves to broad money is thus a potentially useful indicator. In addition, the ratio of reserves to base money may be especially useful for assessing reserve adequacy under fixed exchange rate regimes whose credibility needs to be established (a focus on this measure is institutionalized in the context of currency boards).

40.  Money-based indicators, however suffer from several drawbacks. In countries where money demand is stable and confidence in the domestic currency high, domestic money demand tends to be larger and the reserves over money ratio, ceteris paribus, relatively small. Therefore, while a sizable money stock in relation to reserves suggests a large potential for capital flight out of money, it is not necessarily a good predictor of actual capital flight. Money-based measures also do not capture comprehensively the potential for domestic capital flight. They do not, for example, take account either of the potential for residents to shift out of short-term debt of the public sector into foreign assets, or of the potential for residents to take derivative positions in exchange rates vis-à-vis the monetary authorities. Box 2 illustrates the weak relation between reserves over broad money and the occurrence and depth of crises in a recent period of international crisis.

D.  Debt-Based Measures of Reserve Adequacy

41.  In recent years there has been increasing interest in comparing the level of reserves to a measure of external debt, in particular to short-term external debt by remaining maturity.²⁰ A measure comparing reserves and short-term external debt is useful to gauge risks associated with adverse developments in international capital markets. Short-term debt by remaining maturity provides a measure of all debt repayments to nonresidents over the coming year and, as such, constitutes a useful measure of how quickly a country would be forced to adjust if it were cut off from external borrowing.

42.  Empirical work both in the Fund and elsewhere²¹ suggests that the ratio of reserves to short-term external debt (R/STD), as further defined below, is indeed the single most important indicator of reserve adequacy in countries with significant but uncertain access to capital markets. A smaller reserves to short-term debt ratio is associated with a greater incidence and depth of crises (see also Box 3). Whether the resulting indicator is sufficiently robust that a benchmark level can be identified is a separate issue, which is discussed below.²²

Defining the relevant concept of short-term external debt

43.  A general reserves to short-term debt ratio should use a comprehensive measure of external debt (i.e., the non-equity elements of external liabilities),²³ regardless of instrument or currency denomination (Box 4). All categories of instruments should be included: loans, debt securities (bonds and notes, and money market instruments), currency and deposits, trade credits, and the debt component of foreign direct investment (Box 4). Debt to nonresidents is included, and debt to residents excluded, regardless of the currency of denomination. In determining maturity, debt with embedded put options should be taken to be payable on the earliest date when the creditor can demand repayment.²⁴

44.  On balance it seems best to include both public and private debt for the purpose of defining a general reserves to short-term debt indicator. Clearly, debt issued or guaranteed by the public sector should be included, and this will often encompass many of the liabilities of the banking system, which typically enjoy explicit or implicit guarantees. Beyond this, as noted above (¶28-29), the degree to which (unguaranteed) private short-term external debt is of concern varies across countries—with microeconomic conditions and with the volatility of investor perceptions. However, it is possible for the private sector’s access to international capital markets to be curtailed for reasons related not to its own solvency, but because of increased transfer risk (the risk that exchange rate restrictions are imposed that prevent the repayment of obligations).²⁵ Moreover, the liquidity position of particular sectors (see also chapter IV.C) could contribute to self-fulfilling expectations that affect the sector or the economy as a whole, if not compensated by liquidity at a national level. It is thus appropriate for the authorities to monitor the liquidity position of the private sector and to take private short-term debt into account in their macroeconomic policies, including their reserve policies. In parallel, arguments can be put forward for including in the numerator of the “reserves” measure the liquid foreign assets of the private sector, but on balance it seems best to exclude these (Box 5). ²⁶

Augmenting short-term debt with the current account deficit

45.  There is a case for comparing reserves to a measure of short-term external debt augmented with other borrowing needs. Analysts often use an indicator that includes the current account deficit, and Mr. Guidotti, former Deputy Finance Minister of Argentina, and others have argued that reserve adequacy should be judged based on how long a country can go without foreign borrowing. This implies that other borrowing needs—in particular the current account deficit—would be included alongside short-term debt.

46.  The ratio of reserves to short-term external debt augmented with the projected current account deficit (or another measure of expected borrowing) is useful as an additional indicator, but it is not recommended that it supersede the narrower reserves to short-term debt indicator. Empirical studies suggest, and Annex I illustrates, that the augmented ratio does not in general perform better than the basic reserves to short-term debt ratio.²⁷They also generally find the current account deficit in percent of GDP to be a crucial additional variable in predicting external crisis,²⁸ so that this variable needs to be examined separately anyway (Footnote 32 contains a simple rule of thumb for taking the current account deficit, as well as the real exchange rate, when assessing the reserves to short-term debt ratio). In any case, the narrower measure is easier to judge because it does not depend on a projected current account deficit of uncertain magnitude—all the more uncertain because it may also depend directly on the availability of new funding.

E.  A Simple Benchmark for Reserve Adequacy?

47.  The ratio of reserves to short-term external debt has recently received attention in attempts to find a simple benchmark that countries should satisfy in order to minimize their vulnerability. Mr. Greenspan (Chairman of the Federal Reserve Board of the United States) has elaborated on Mr. Guidotti’s proposal with suggestions that reserves should exceed official and officially-guaranteed short-term debt.²⁹

48.  The Fund staff has found some empirical support for a benchmark of one for the ratio of reserves to economy-wide short-term external debt.³⁰ The staff would recommend using a ratio of unity as a starting point for analysis of reserve adequacy for countries with significant but uncertain access to capital markets. In addition, however, the above discussion points to the need to take into account various factors that serve to enhance or mitigate the need for reserves in a particular country compared to such a benchmark:³¹

• the exchange rate regime (a flexible regime reduces the likelihood and costs of crisis);
   
• the currency denomination of external debt (ideally after hedging);
   
• other macro fundamentals, especially the current account deficit and the real exchange rate;³²
   
• the microeconomic conditions that impact the soundness of the private sector debt position. For example, to the extent a country’s private sector has unqualified access to international capital markets, its short-term debt can be discounted in the overall ratio; and
   
• the possibility of capital flight by residents or the possibility for residents and nonresidents to take derivative positions or otherwise short the currency. Thus, close attention should also be paid to monetary aggregates (especially when the banking system is weak), and short-term public debt held by residents (especially in the absence of capital controls or captive markets, and the public sector’s financial position is weak) (Box 4).
   

F.  Derivative Liabilities of the Monetary Authorities

49.  The comparison of reserves to short-term debt, as discussed above, takes account of the short-term external debt of the monetary authorities (which is included in the denominator), but not of their notional derivative liabilities (unless subject to margin calls). It is important to consider the impact of such liabilities separately.³³ In the recent crises, several central banks took large forward positions in support of the domestic currency. The potential for running up such positions is in general very great, since central banks may face the full force of international capital markets on the other side of the market, and since there are no natural limits on the amount of forward intervention. As a result, forwards can be a dangerous instrument.³⁴ It is thus particularly important that the full implications of forward liabilities be understood by policy makers.

50.  As discussed in Annex III, when the central bank seeks to maintain the value of the currency, forward contracts generally presage a drain on reserves in the amount of the full notional value of the contracts. In the absence of exchange restrictions, there is no substantial difference in this respect between outright and nondeliverable forward contracts.³⁵ These drains may be less likely to materialize to the extent that the ultimate counterparties to the forwards are residents, but this outcome depends largely on the presence of effective capital controls and the effectiveness of tax regimes. Thus, any substantial forward commitments to sell foreign currency should be taken into account in assessing reserve adequacy, and if they are of shorter maturity and no effective capital controls exist, would generally be taken to increase the need for reserves one-for-one.
 

G.  Stress tests

51.  A straightforward method for taking risk enhancing or mitigating factors into account in judging the level of reserves is to conduct stress tests. Stress tests are in essence a simple and relatively uniform method of scenario analysis, and generally limit the behavioral reactions to a bare minimum to enhance comparability and increase transparency of the test. Thus a ratio of reserves to short-term debt of one means that a country can withstand a simple stress test in which the current account and access to capital markets are nil for a year: the reserve loss will be equal to short-term debt (by remaining maturity). Stress tests are particularly useful to take into account the impact of additional pressures such as capital flight or a current account deficit. If the magnitude of these pressures is expressed in percent of reserves, they can easily be taken into account in a judgment on the reserves to short-term debt ratio. For instance, a country could withstand a simplified stress test in which it faced a current account deficit equivalent to 30 percent of reserves, in addition to a complete cutoff from capital markets, only if its reserve to short-term debt ratio was at least 130 percent. (Annex IV discusses “liquidity at risk,” a special form of stress testing.)
 

52.  Stress tests can be used to test a variety of scenarios, to take account, for instance, of:

• the likelihood that access to different types of inflows (FDI, trade credit, other credits) would be affected differently in a period of turbulence;
   
• shorter or longer periods of limited access to capital markets than the one year implicit in the ratio of reserves to short-term debt;
   
• the extent to which different parts of the private sector might retain access to financial markets in a crisis;
   
• possible reserve losses as a result of changes in the exchange rate through the derivative positions of the monetary authorities;
   
• the potential for capital flight both from deposits with the domestic banking system and short-term government debt; and
   
• additional risks such as changes in the price level of major commodities, in the interest rates charged for floating rate loans, or large cross-currency exchange rate movements that affect the debt and current account deficit in disparate ways.

53.  This kind of analysis can usefully complement evaluations of sustainability through medium-term scenarios. Such stress tests could prove useful in identifying the major risks facing countries, as well as strategies to mitigate such risks, such as exchange rate flexibility, entering into long-term export contracts, or enhancing the liquidity buffer by increasing reserves or by establishing private contingent credit lines.

IV.  Debt-Related Indicators

54.  Appropriately defined debt indicators are a useful tool to support sound debt management, but care is needed—especially in attempts to formulate benchmarks—to make adequate distinctions between different types of debt and their impact on external vulnerability. As emphasized in Chapter II, private sector debt that is not based on distorted incentive structures is of less concern than private sector debt that reflects severe distortions, for example, in favor of short-term or foreign currency financing. These differences limit the usefulness of macroeconomic debt indicators as benchmarks to comparisons within fairly homogeneous groups of countries and categories of debt. This also underscores the importance of sectoral indicators.

55.  This chapter is structured as follows. It first addresses traditional debt-related indicators and distills conclusions regarding their formulation and use as indicators and benchmarks. Implications of the structure of debt—maturity structure and exchange rate and interest rate composition—are discussed next. The chapter concludes with a discussion of sectoral indicators.³⁶

A.  The Use of Debt Indicators for Sustainability Analysis

56.  External debt sustainability analysis is generally conducted in the context of medium-term scenarios. These scenarios allow for analysis of diverse flows such as new borrowing driven by current account deficits, of the terms at which new borrowing is conducted, and of the interaction with policies such as demand management policies affecting imports and exports. Judgment in these numerical evaluations tends to focus on the conditions under which debt and other indicators would stabilize, on major risks, and on the need and scope for policy adjustments. Macroeconomic uncertainties (such as the outlook for exports and imports) and policy uncertainties (fiscal expansion or consolidation) tend to dominate the medium term balance of payments outlook, and feature prominently in the scenarios conducted by staff in the context of Article IV consultations and program design.

57.  Within this context, the most common debt indicators scale the external debt stock by either exports, GDP, or government revenue. In each case total outstanding debt at the end of a 12-month period is scaled by the relevant flow over that period.³⁷ A high ratio indicates a greater burden of servicing the debt, with the magnitude of the burden depending also on the difference between the interest rate and the growth rate of exports, GDP or revenue (Annex V). And a growing ratio, especially if the level of debt is already high, may suggest that a country is on an unsustainable path.

58.  Among the various indicators based on the debt stock, debt to exports is the most useful, but other ratios may also be relevant:

• Debt to exports has the advantage that it is less volatile than debt to GDP in the presence of changes in the real exchange rate, and allows a more meaningful inference of trends.³⁸
   
• Debt to exports also has the advantage that exports provide the basis for debt repayments.³⁹ However, the debt to GDP indicator may also provide some indication of the potential to service debt by switching resources from production of domestic goods to the production of exports
   
• In cases where public debt is predominant, the ratios over GDP or tax revenue may also be useful in that they relate debt to the underlying source of repayment, the country’s tax base.

59.  At the same time, all indicators based on the stock of debt suffer from certain drawbacks, particularly if viewed only in the form of a single snapshot in time. First, the interest burden of servicing the debt can vary widely, depending in particular on whether the debt is concessional or not. In general, it is useful to report and analyze the average interest rate on the debt alongside the debt to exports ratio, as an indicator of concessionality of debt and to analyze the impact of changing interest rates on the real debt burden.⁴⁰ Second, the pace of output and export growth and the contribution to it of debt creating inflows differ across countries.⁴¹ These drawbacks could to some extent be addressed by focusing on the present value of debt payments rather than debt itself, by using debt over prospective exports or GDP, or by multiplying the ratios of debt to export or GDP by the nominal rate of interest minus the rate of nominal export growth (defined in the same currency)—see Annex V. These alternative constructs, however, are significantly more complicated. The difficulty of identifying “snapshot” indicators underscores the importance of examining the evolution of debt in medium-term scenarios, and of reviewing the debt ratios in conjunction with expected interest and growth rates.

60.  For the larger emerging market economies, the ratio of debt to exports does seem to positively impact interest spreads on sovereign bonds. Since spreads are a useful proxy for the loss creditors expect, this suggests that repayment expectations are affected by the overall level of debt (illustrated in Figure 6 for a relatively tranquil period). Nevertheless, the relationship varies considerably from one year to another, implying that other factors are also important. In particular, spreads appear to be significantly affected by variations in market liquidity and by fluctuations in the markets’ assessment of the government’s willingness to pay.

61.  The ratio of debt service to exports is another possible indicator of economy-wide debt sustainability, but care should be taken in its use.

• In practice, an indicator that is different from the core concept is generally used. Although debt service is in principle defined including principal payments and interest paid on both long- and short-term debt, in practice amortization payments on short-term debt are usually excluded from debt service.⁴² In addition, the coverage of private sector data is often limited, either because the indicator is intentionally focused on the public sector or because data on private debt service are not available.
   
• The ratio of short-term debt to reserves already provides an indication of potential liquidity problems, while the ratio of interest payments to exports provides an insight into solvency problems.

These factors may contribute to the limited correlation between the debt service ratio and bond spreads for the larger emerging market economies (as illustrated in Figure 7).⁴³

62.  Nonetheless, a narrow version of this ratio, focused on government and government guaranteed debt service, can be a particularly useful indicator of government debt sustainability and transfer risk, for the following reasons.⁴⁴ If the official sector wants to offset amortization payments through new borrowing from official sources, it will in most cases need to take some action entailing financial and non-financial costs, e.g., new loans may require imports, the implementation of projects, or other conditionality, including of a political nature. In this case, interest payments and amortization are more comparable, and including both in a single measure may provide some broad indication of the overall cost to service the debt.⁴⁵

63.  The scope for identifying critical ranges for debt indicators is rather limited. The absolute value of overall debt ratios is not very useful across heterogeneous countries, without additional information, notably on factors that affect the soundness of borrowing and lending decisions. Difficulties in comparing debt indicators across countries are also reflected in the absence of a relation between debt and crisis indicators (illustrated in Annex I), and in the unstable relation with spreads.

• For more homogeneous country groupings and for debt definitions focused on the public sector, there is more potential to identify ranges for debt-related indicators that suggest that debt or debt service is approaching dangerous levels, i.e., that these indicators are approaching levels which in other countries have resulted in suspension or renegotiations of debt service payments, or have caused official creditors to consider whether the debt burden may have reached levels that are too costly to support. For example, assistance under the HIPC Initiative is determined based on a target for the public debt to exports ratio (150 percent) or debt to tax revenue ratio (250 percent). In these ratios the net present value of debt is used, and only a subset of external debt is taken into consideration, namely medium- and long-term public and publicly guaranteed debt.⁴⁶
   
• Empirical work on macro debt indicators is limited and does not provide much guidance. Based on whether or not a country has been subject to rescheduling, Cohen estimates critical points for debt to GDP of 50 percent, debt to exports of 200 percent and debt to tax of 290 percent.⁴⁷ Above these points the risk of rescheduling exceeds 60 percent.⁴⁸ However, the estimations do not control for variables such as the composition of debt (private or public), export growth rates, international interest rates, the level of per capita income, or institutional variables, and capture only rescheduling, which is but one manifestation of external crisis.

B.  The Impact of the Composition of Debt

The maturity structure

64.  The maturity structure of debt has a profound impact on liquidity, and can be tracked using the ratio of reserves to short-term debt and through stress testing. An emphasis on the reserves over short-term debt ratio as recommended above will in itself focus attention on the maturity structure of debt, and will create, to some extent, incentives to lengthen maturities and avoid a lumpy repayment structure. Moreover, stress testing or scenario analysis, as discussed above, can be used to analyze the impact of limited access to financial flows for a period longer than a year, thus taking into account debt falling due over longer time horizons.

65.  Such stress testing can be supplemented with rolling liquidity analysis to analyze the impact of the maturity structure on future, rather than present, liquidity, by predicting the reserves over short-term debt ratio forward in the context of medium-term scenarios. Simple and explicit assumptions need to be made to support a meaningful interpretation of such projections. For example, short-term debt by original maturity could be assumed to be rolled over (or grow in line with trade), amortization on medium- and long-term loans and bonds could be assumed to be rolled over with a specific maturity profile (such as the current maturity structure), and similar assumptions can be made for additional debt. The implications of achieving a desired reserves over short-term debt ratio for the needed maturity structure of newly contracted or rolled over debt could be highlighted. Useful maturity indicators to provide shorthand information on assumptions and prospects are the levels of medium- and long-term debt falling due over specified periods (e.g., two, three, and five years).⁴⁹

66.  While the maturity structure is important, it is difficult to capture this in a single number—such as average maturity of overall debt—that is useful, especially for comparison across countries.⁵⁰ Short average maturity is not necessarily indicative of external vulnerability:

• average maturity masks important differences in the sectoral composition of debt. For example, countries with limited debt mostly in the form of private trade credit will show a relatively short average maturity, but will not necessarily be particularly vulnerable;
   
• average maturity also masks important differences in the dispersion of maturities, when it is really the whole spectrum (as analyzed in rolling liquidity analysis) that matters. For example, countries with a small fraction of very long-term concessional debt and a significant amount of short-term government debt may show a high average maturity.⁵¹

67.  Therefore, only adequately focused measures, such as average maturity by sector and by debt category, can usefully be compared over time.52 Such measures can alert policy makers and market participants to maturity structures that are potentially problematic and warrant closer consideration of sectoral vulnerabilities.

Currency and Interest Rate Composition

68.  The currency and interest rate composition of debt can also have significant implications for external vulnerability, but analysis and indicators of such vulnerability may similarly need to be focused at a sectoral level or even at the level of individual institutions.

69.  As discussed above (Box 4), there is a significant difference in implications between domestic and foreign currency external debt. The burden imposed by the former in terms of foreign currency is eroded if the exchange rate depreciates, and the balance sheet effects of a depreciating exchange rate vary with the extent to which the debt is denominated in foreign currency. Information on the currency composition of debt at the sectoral level is particularly important, because the balance sheet effects also depend on foreign currency relations between residents (such as government foreign currency-denominated or -indexed liabilities to residents).

70.  The specific foreign currency and interest rate composition of external debt may also have significant implications. If sharp changes occur in the exchange rates in which debt is denominated, but these are not offset by similar changes on the inflow side (for example, in exports), significant income effects can result. Similarly, sharp increases in short-term interest rates, such as those experienced in the early 1980s, can have profound implications for the real cost of debt, especially if a significant share of debt is floating rate.

71.  While the currency and interest composition of debt can be important in analyzing external vulnerability, it is necessary to take account of derivatives, as these may significantly change the effective composition of debt. Interest derivatives can be used to swap floating rate obligations into fixed rate liabilities. Similarly, currency derivatives can be used to change the exposure from foreign to domestic currency or to a different foreign currency. In the presence of large scale derivative activities, on which available data are often limited, it is doubly difficult to assess the overall implications of private sector exposure to (cross) currency or interest fluctuations. In such a situation it is difficult to build up a picture of overall vulnerability without information at the sectoral level or at the level of individual banks and firms.

C.  The Role of Sectoral Indicators

72.  A sectoral analysis of debt is important because the sustainability of public and private sector debt is governed by different considerations. Analysis at the level of sectors can shed light on the degree to which private debt should be of concern.⁵³

73.  The present paper has a narrow focus, and it will merely note the extensive work underway elsewhere (see below) on the general issue of risk management at the institutional or sectoral level. The question this paper seeks to answer is whether there are simple debt-related indicators that provide insights into the external vulnerability of sectors—i.e., their vulnerability to changes in exchange rates and, consequently or in expectation thereof, to a cutoff from foreign currency lending.⁵⁴

The public sector

74.  Public debt management must seek to ensure that the public sector can service its debt while minimizing costs in the long run—which implies also minimizing the costs to the economy of crises resulting from imprudent debt management. Debt management should thus contribute to limiting the risk of recourse to a moratorium on public debt service, and minimizing transfer risk. Careful management of all public debt is required both for the sake of effective stewardship of fiscal resources, and also for the sake of external stability, since external crises can have their roots in fiscal sustainability problems. These wider issues, however, are beyond the scope of this paper.⁵⁵

75.  A number of indicators can be identified that help capture the solvency and liquidity risks associated with external public debt:

• the public sector debt service ratio, ratios of public debt to GDP and to tax revenue, the average interest rate, various maturity indicators, and indicators of the composition of debt—these have all been noted above;
   
• the share of foreign-currency-denominated or indexed debt, which provides a handle on the extent to which exchange rate depreciation may help erode away a debt problem; ⁵⁶
   
• the share of overall short-term and longer-term floating rate debt in domestic government debt—short-term debt may be a source of capital flight in the absence of effective capital controls, and the size of interest repayments on floating rate debt may be affected by measures to defend the currency and to stem capital flight.

76.  In addition, the public sector has a special role to play in ensuring that it creates or maintains conditions for sound risk management in other sectors (for instance, by avoiding policies that create a bias toward short-term foreign currency borrowing). The discussion that follows will provide some insight as to whether useful indicators can be developed of the extent to which the government creates conditions for sound risk management.

The financial sector

77.  Financial sector vulnerability is a particular cause for concern as regards external vulnerability.⁵⁷ Financial institutions are especially vulnerable because they are usually highly leveraged, exposed to maturity mismatches, operate in markets where asymmetric information is important, and are subject to moral hazard through explicit or implicit deposit insurance and limited liability.58 Reflecting these characteristics, financial institutions are vulnerable to the balance sheet effects of changing exchange rates (or interest rates), and to changes in cash flow such as withdrawal of foreign currency deposits or credit lines by foreign banks.

78.  Among aggregated microprudential indicators, two have a foreign currency dimension—“sensitivity to foreign exchange risk” and “foreign-currency-denominated lending:”

• The open foreign exchange position is a key indicator of sensitivity to foreign exchange risk.⁵⁹ Because of the obvious direct risk attached to this factor, the Basel Core Principles of Banking Supervision include limits on the open foreign currency positions of banks. In the presence of such regulatory constraints (as long as they are abided by), the open foreign currency position is typically not a prime concern.
   
• Also important are foreign currency maturity mismatches—i.e., the extent to which foreign currency liabilities are matched by foreign assets that can be liquidated in the same time frame. In the presence of capital controls, the indicator could focus on a subset, namely, the mismatch in foreign assets and liabilities (i.e., claims on and liabilities to nonresidents). In the presence of such mismatches, a cut-off from external funding can put considerable pressure on the international reserves of the central bank.
   
• Foreign currency quality mismatches (and in particular the quality of foreign currency assets) matter, for the same reasons, although here data can be very difficult to obtain. A particular concern is foreign currency-denominated credit to local borrowers without significant foreign currency cash flow, or who display other foreign currency related weaknesses discussed under the corporate sector indicators below.
   
• Information on gross foreign currency liabilities is useful to the extent assets are not usable by banks to offset withdrawals in liquidity, for example because of fiduciary or liquidity requirements.

79.  Assessment of financial sector vulnerabilities needs to take account not only of the numerical values of these indicators, but also of the quality of supervisory policies, institutions’ own systems, and data availability, and efforts are underway to improve these. Supervisory policies create rules and incentives for individual financial institutions to limit their own vulnerability, and institutions’ own systems should enhance their capacity to do so. Data availability is necessary for markets to price risk adequately and for supervisors to perform adequate oversight. To support the assessment of vulnerabilities, oversight and market discipline, the staff is surveying the availability of data to devise indicators as part of the review of Macro Prudential Indicators, while the Basel Committee on Banking Supervision is reviewing the guidelines on bank liquidity.

The corporate sector

80.  In the wake of the recent crises it has become obvious that vulnerability analysis cannot ignore the corporate sector. Although individual firm failures can and should be addressed through bankruptcy and resolution systems, it is increasingly recognized that the organization of the corporate sector, and especially its financial structure, can impact external vulnerability (including indirectly through its impact on the credit risk of the financial sector).⁶⁰ Overextension of foreign currency financing to the corporate sector, combined with limited profit margins, and/or a financing structure that is highly leveraged or geared toward short-term financing, can lead to widespread corporate nonpayment. By overwhelming the legal system for contract enforcement and bankruptcy, and undermining asset prices and the value of collateral,⁶¹ this can have a domino effect. The ensuing corporate failures may contribute to a recall of foreign loans for trade and working capital and a dramatic fall-off in new capital inflows, which can precipitate or worsen an external crisis.

81.  The present paper will not attempt to summarize extensive research work underway in this area, but will focus on indicators of the financial structure to the extent they provide insight into the impact of the corporate sector on external vulnerability. Some general background information is provided in Box 8.

82.  Directly relevant for external vulnerability are imbalances in companies’ balance sheets and cash flow that make them vulnerable to changes in exchange rates. The exchange rate can have a balance sheet effect by increasing the value of a firm’s debts more than that of its assets, which may reduce its ability to obtain new loans. (Leverage is an important indirect factor here: leveraged companies are especially vulnerable to such imbalances as their cushions are more limited.) The exchange rate will also impact a company’s cash flow. Corporations that borrow in foreign currency but produce few exportable goods or services (as was common, for instance, in Indonesia) are exposed to sharp cash flow (and profit) reductions as the exchange rate depreciates.⁶² Moreover, the extent to which corporations are exposed to short-term foreign debt impacts, in conjunction with their cash flows, the extent to which they are exposed to cut-offs from foreign financing. Thus, key financial indicators for the corporate sector from the standpoint of the present paper would be:

• the mismatch in foreign currency cash flow as a ratio to overall cash flow taking account of hedging and other foreign currency revenues; and
   
• the difference between foreign currency liabilities and liquid foreign currency assets as a ratio to equity.

83.  More traditional corporate indicators are useful as well. Indicators to assess general profitability and vulnerability include the coverage of interest payments by operational cashflow, leverage, the ratio of short-term to overall debt (for domestic and foreign currency), and the ratio of domestic currency versus foreign currency debt. Indicators aggregated for all corporations (Box 8 includes several that are derived from a large data base on corporate data), or by type of sector (traded or nontraded sector), combined with information on dispersion, provide useful information on the overall state of the corporate sector and its contribution to external vulnerability. In addition, they provide information on interest sensitivity, which is important, as the prevention of an exchange rate crisis could call for sharply increased interest rates, while a failed defense may result in at least temporarily high interest rates.

84.  Data availability with regard to corporate indicators focused on exchange rate vulnerability—especially indicators that would require breakdowns in domestic and foreign currency—leaves much to be desired, but could be promoted using present dissemination practices as a starting point. For large corporations, publication standards that make the relevant data available are often established as part of the conditions for stock exchange listings of equity and bonds. A first priority could be to ensure that information that individual corporations are required to publish is readily available, for example from stock exchanges. Further, statistical agencies could be encouraged to (re)disseminate relevant aggregates. Collection and publication of additional pertinent information could be promoted, as part of supervisory regulation and listing arrangements, for companies that are allowed direct access to foreign capital markets (bank or bond funding), or whose stocks and bonds are listed on the domestic exchanges. This would both foster market discipline and provide a basis for adequate monitoring by the appropriate supervisory authorities.

85.  An overview of indicators that can be used at a sectoral level or at the level of an individual firms or banks is contained in Table 2.

1/ No standard indicator is suggested. Lending criteria can provide important information. For bank claims on the corporate sector indicators could be formulated based on a weighted average of corporate indicators.
 

V.  Issues for Discussion

86.  The paper has reviewed a variety of reserve and debt related indicators of external vulnerability. The results are summarized in the “Summary, Conclusions and Implications.” The conclusions with regard to indicators are reflected in Tables 1 and 2, on pages 6 and 39.

• Directors may wish to comment on the recommended indicators, the characterization of the different circumstances under which these indicators would apply, and the data and conceptual issues that would arise.
   
• What potential would Directors see for member countries to draw upon these recommendations in the context of their efforts to monitor and limit sources of external vulnerability? Does this paper provide useful operational guidance?
   
• The staff intends to use the results of the paper and Executive Board discussion to further refine the presentation of vulnerability indicators in staff reports and other operational papers. The staff would also propose to draw on these results in its response (in collaboration with the World Bank staff) to the request by the IMFC to develop guidance on best practices in sovereign debt management. Do Directors agree with this approach?

ANNEX I

Illustrative Charts of Debt- and Reserve-Related Indicators

The figures in this annex show, across a range of emerging market economies, the relationship between a number of reserve and debt indicators before the onset of crises, and the degree to which these countries were subsequently affected by currency crises (as measured by a crisis index). Footnotes 18 and 24 in the main text contain key information for interpreting the charts.

Panel 2. Reserve and Debt Indicators of SElected Emerging Market Economies in 1997

Panel 3. Reserve and Debt Indicators of SElected Emerging Market Economies in 1998

ANNEX II

Illustrative Table on Reserve and Debt Indicators for Industrial Countries

The table below describes key debt- and reserve- related indicators for a number of industrialized countries. The data are derived from International Investment Position (IIP) data reported to the Fund. Certain conventions had to be used to distinguish between short-term debt and medium-and long-term debt in the reported data (e.g., currency and deposits and the data on the value of financial derivatives have been included under short-term), and these data should be considered to be approximations. National sources may contain better estimates.

The availability of IIP data for these countries, as well as the importance of assets, both allow and suggest the presentation of a number of additional indicators such as the net IIP position, and the ratio of short-term debt over total assets. Other IIP based indicators could provide useful information as well, such as the ratio of currency and deposits to short-term debt, and a sectoral breakdown of the ratios.

ANNEX III

Interpreting Central Bank Forward Positions

Several central banks have taken or currently have large forward positions in support of the domestic currency. This annex asks how such positions should be interpreted. Is forward intervention equivalent to spot intervention? Will it eventually cause reserve drains, and to what extent?

1.  Equivalence

In response to shifts in portfolio preferences by market participants against the local currency (e.g., because of changed exchange rate expectations), the central bank may choose to relieve the pressure on the exchange rate by intervening in either the spot or the forward market. Abstracting from monetary policy changes affecting the interest rate, the spot and forward exchange rates will move in concert, since the difference between them will remain equal to the interest rate differential plus a risk premium/cost margin reflecting counterparty, legal risks, etc. (otherwise there would be arbitrage opportunities).

Given the equivalence between spot and forward market intervention of the central bank, it is important that the central bank reveal both—changes in reserves and in its forward book—in order for market participants to realize the full extent of pressure on the exchange rate. Transparency about the size of the forward book is also important if market participants are accurately to gauge the quasi-fiscal risks involved. Therefore, merely revealing a forward position that was not previously known to markets can create or relieve pressure on the exchange rate, as market participants reassess their expectations and adjust their portfolio positions.

2.  Eventual drain on reserves

In analyzing the eventual drain on gross reserves from unwinding forward contracts, we posit that the unwinding takes place by decision of the central bank (and is thus a given), and we assume that the exchange rate is fixed by the central bank.⁶³

The key factor that will determine the drain on reserves is whether the shift in desired portfolio allocations described above is temporary or permanent—that is, whether it will have been reversed by the time the contract matures.

•          If the shifts in desired portfolio allocations are temporary and are reversed by the time the contract matures, the demand for forward cover subsides and the central bank can concomitantly reduce its forward position without losing reserves (compared to the point in time when the desired portfolio allocation first shifted).

•          If the shifts in desired portfolio allocations are permanent, the central bank can unwind its forward position only by accepting a drain on its reserves. The drain will encompass the full notional value of the forward contract.

The analysis above considers whether desired portfolio allocations have shifted between the time the contract was issued and when it matures. The analysis generalizes to the situation where one views the forward position at any intermediate time, between issuance and maturity, and where one asks what will be the drain on reserves at maturity: the key question then becomes whether desired portfolio allocations will shift between the time of viewing and the time of maturity.

The above analysis does, however, assume that portfolios are continuously in equilibrium and is based implicitly on the absence of capital controls or transactions costs, allowing investors to swiftly adjust their portfolio position in response to changing economic parameters. To the extent that, at the time one views the forward position, portfolios are not in equilibrium, the drain on reserves may be more limited.

3.  Instruments of intervention

Central banks can use outright forward contracts or non-deliverable forward (NDF) contracts for their forward market interventions. Outright forward contracts provide for the sale or purchase of a specified amount of foreign currency, at a specified exchange rate and date with the settlement involving the full exchange of principal in foreign currency. NDFs, in contrast, do not involve the exchange of principal. But they provide for the payment of the difference, usually in domestic currency but sometimes in foreign currency, between the actual exchange rate on the day the contract matures, and the contractual NDF rate, multiplied by the notional value of the contract.

The general analysis above goes through for NDFs as well as for outright forwards. This is because, for the counterparty, a NDF produces exactly the same profit or loss at maturity as an outright forward. Thus, portfolios of NDFs and of outright forwards will adjust similarly in the presence of shifts in preferences as to currency exposure (regardless of whether settlement takes place in domestic or foreign currency and involves the corpus of the contract or not). In particular, in both cases, the counterparties face similar incentives to acquire domestic currency ahead of the unwinding of the contracts.⁶⁵ In the case of an outright forward, that domestic currency is then used to purchase foreign currency at the forward rate from the central bank; in the case of a NDF, the counterparty receives from the central bank, usually in local currency, the difference necessary to enable purchases of the same amount of reserves as would be available with an outright forward contract. Thus, unwinding NDF positions normally has the same impact on reserves as unwinding outright forward contracts.

The equivalence between outright forwards and NDFs breaks down, however, in the presence of exchange restrictions. Under an outright forward, the counterparty receives foreign exchange from the central bank; under a NDF (all the more so if it is settled in domestic currency), the counterparty must purchase the foreign exchange with domestic currency, and may encounter exchange restrictions.

4.  Resident vs. nonresident counterparties

The effect on reserves of unwinding the central bank’s forward book may also depend on whether counterparties are residents or non-residents:

•          Most obviously, if capital outflow controls prevent residents from cross-border transactions, the foreign currency paid out to residents by the central bank as a result of maturing forward contracts will generally find its way back into reserves, probably through the deposit of such foreign currency at domestic banks.

•          In addition, taxes paid on residents’ profits from forward contracts will work to limit the loss of reserves.

•          Moreover, permanent shifts in desired portfolio allocations against the domestic currency may be more limited both in size and frequency among residents, given that residents’ expenditure is usually denominated in domestic currency, and that their portfolios tend to be smaller compared to the amounts managed by nonresidents. Thus, the chances that the portfolio shift that induced the buildup of the forward book will be reversed, so that the book can be unwound without loss of reserves, may be higher if the counterparties are residents than if they are nonresidents.

ANNEX IV

Liquidity at Risk

Mr. Greenspan, Chairman of the Federal Reserve Board of the United States, has recently proposed what he called a “liquidity-at-risk” standard.⁶⁶ Under this standard, a country’s external liquidity position would be calculated for a wide range of possible events, as in stress tests; but events now have to be assigned probability distributions. An appropriate level of reserves is defined as one which provides a high probability (e.g., 95 percent) that external liquidity will be sufficient to (say) avoid new borrowing for one year. This approach is closely related to the value-at-risk methodology, widely used by financial institutions to provide a single yardstick for a multitude of risks in their portfolio, by estimating (usually from past information on prices) the maximum loss an institution can expect over a given period of time, at a given confidence interval.

There are, however, a number of drawbacks to value at risk methodology, especially in a macroeconomic setting. The allure of value-at-risk has declined recently, as variances and covariances over historic periods of relative calm ceased to be reliable guides in a period of turbulence in capital markets, and similar problems would likely beset attempts to apply this methodology to the economy as a whole. In addition, prices of assets, to which value-at-risk techniques are normally applied, tend to follow random walks, and their movement is dominated by the probability distribution. This key assumption cannot be made for most economic time series that influence the balance of payments: exports, imports, or reserves do not follow random walks. Therefore, in a macroeconomic setting, value-at-risk depends less on the probability distributions of unexplained events, and much more on the behavioral relations underlying the movements in economic variables—in other words, on a macroeconomic model, which, if it were to be used to generate not just point estimates but very detailed probability distributions, would have to be highly accurate. Moreover, there is the further complication that the probability distributions, both past and future, of many relevant economic events (e.g., declines in capital inflows) are importantly influenced by the authorities’ policy responses to them.

Nevertheless, in particular country cases random movements of balance of payments variables may play an important role. Examples are countries whose exports are dominated by commodity prices. Even in such cases stress tests may hold more promise, as they are more analytically tractable. The type of stress test, however, can usefully be inspired by value-at-risk or liquidity-at-risk type of arguments. For example, stress tests can be conducted that seek to analyze the effect of a change in the price of a key commodity export of the order of twice the standard deviation (for a 97 percent probability interval if the prices are normally distributed). For countries whose trade balance is geographically tilted (imports from one region, exports to another), such statistical analysis might also be useful. All in all, while the liquidity-at-risk approach may be useful in the evaluation of individual instruments and to inspire stress tests, underlying economic relations are at present too uncertain to allow more general applicability of this technique.

ANNEX V

Numerical Interpretation of Debt Ratios

To obtain an idea of the magnitude of the transfers implied by debt to export ratios, as background for assessing the sustainability of debt, one can conveniently make use of the following formula which describes the constant fraction of exports, b, to be transferred (interest after net debt inflows), if the debt stock is to satisfy the solvency condition that the present value of debt goes to zero:⁶⁷

b = (r-n)D/X₊₁

where r is the (constant) interest rate, n the growth rate of exports X, and D the inherited debt stock measured in net present value terms. X₊₁ is next year’s exports.

A similar ratio can be derived for GDP, replacing X by GDP, and interpreting b as the fraction of GDP to be transferred.

Therefore, the higher the initial stock of debt relative to exports or GDP, and the higher the interest rate on debt relative to the country’s growth outlook, the larger the share of GDP or export earnings that has to be earmarked for debt payments.

The formula can readily be used for quantitative illustrations. For instance:

• For r-n of 5 percentage points, and D/GDP of 50 percent (key ratios and assumptions used by Cohen, 1997, op. cit.), 2.5 percent of GDP needs to be transferred in order to stabilize the debt ratio.
   
• For r-n of 5 percentage points, and D/X of 200 percent, 10 percent of export earnings needs to be transferred in order to stabilize the debt ratio.

The formula can also be used to illustrate the point that the pure ability to pay is not a very useful concept of solvency:

• For r-n of 5 percentage points, if all resources were devoted to repayment (b=100 percent), D/X could amount to 20. Countries rarely reach such levels of indebtedness and normal debt service ceases well before they do.

ANNEX VI

Summary of Aggregated Microprudential Indicators