- Richard Hemming, Axel Schimmelpfennig, and Michael Kell
- Published Date:
- May 2003
Financial crises in emerging market economies have been the subject of considerable analysis. This section draws on work to date to provide essential background for the subsequent sections.
Most studies of financial crises distinguish between currency, debt, and banking crises.3 These three types of crisis are discussed in Box 2.1. Empirical studies—including this one—typically make such a distinction because it is helpful to use specific, and therefore relatively narrow, crisis definitions.4 But many of the recent financial crises in emerging market economies, such as the Mexican crisis of 1994–95 and the Asian crisis of 1997–98, have been characterized by currency, debt, and banking crises occurring at the same time or in rapid succession. Indeed, as international capital and domestic financial markets have become increasingly integrated, the distinction between different types of financial crisis has become blurred.
For this reason, recent analyses of emerging market crises have focused on the linkages between the corporate, banking, and public sectors during times of external financing pressure. In particular, attention is paid to crisis dynamics and spillover effects propagated through sectoral and aggregate balance sheets, as opposed to through more traditional flow variables such as the current account and fiscal deficits.5 From this perspective, financial crises occur when there is a plunge in demand for financial assets of one or more sectors (Allen and others, 2002). Thus creditors may lose confidence in the government’s ability to service its debt, in the banking system’s ability to finance deposit outflows, or in the corporate sector’s ability to service its loans. Problems in one sector are then liable to spread to other sectors: for example, concerns about the government’s balance sheet could undermine confidence in banks holding government debt, and could spark a run on deposits; or banking sector problems could expose large contingent liabilities that could lead to difficulties for the government in servicing its debt, and could even give rise to solvency concerns that could cause a run on the currency.6
Fiscal vulnerability is related to vulnerability to crises more generally.7Furman and Stiglitz (1998, p. 6) equate vulnerability with the causes of crises in a stochastic sense:8 “an increase in vulnerability means an increase in the probability that . . . shocks, rather than being absorbed by the economy, will be translated into a . . . currency or financial crisis.” IMF (1998) argues that vulnerability depends on: the magnitude of economic imbalances, in terms of stocks and flows; misalignments of asset prices, including the exchange rate; the extent of financial sector distortions and structural rigidities; and the credibility of policies. Fiscal vulnerability can then be seen as a component of overall vulnerability that derives from the design and implementation of fiscal policy. Allen and others (2002) adopt a more precise definition, namely the risk that liquidity or solvency conditions are violated and a crisis results.
Fiscal vulnerability can also be defined in its own right. At a fairly general level, Hemming and Petrie (2002, p. 161) see fiscal vulnerability as reflecting a situation in which a government is exposed to the possibility of failing to achieve its aggregate fiscal policy (or macrofiscal) objectives. The latter are characterized as: avoiding excessive deficits and debt; ensuring that fiscal policy contributes to effective demand management; and raising revenue in a manner consistent with maintaining reasonable and stable tax rates (and a stable tax system more generally). Brixi and Schick (2002) define a closely related concept—fiscal risk—which is “a source of financial stress that could face a government in the future,” and focus, in particular, on contingent liabilities and off-budget fiscal activities. This paper is concerned very much with public sector liquidity and solvency risks, although there is increasing recognition of the difficulties involved in distinguishing between the two in emerging market economies.9 However, the paper also pays attention to sources of fiscal vulnerability more broadly defined.
Box 2.1.Currency, Debt, and Banking Crises1
A currency crisis occurs when a speculative attack on the exchange rate results in a devaluation or sharp depreciation, or forces country authorities to defend the currency by expending large volumes of reserves or by significantly raising interest rates. Dornbusch (2001) distinguishes between “old style” or “slow motion” currency crises—where a cycle of overspending and real appreciation weakens the current account, often in a context of extensive capital controls, and ends in devaluation—and “new style” crises—where investor concerns about the creditworthiness of the balance sheet of a significant part of the economy (public or private) lead to a rapid buildup of pressure on the exchange rate in an environment of more liberal and integrated capital and financial markets.
A debt crisis occurs either when the borrower defaults or lenders perceive this as a significant risk and therefore withhold new loans and try to liquidate existing loans. Debt crises can apply to commercial (private) and/or sovereign (public) debt. If there is a perceived risk that the public sector will cease to honor its repayment obligations, this is likely to lead to a sharp curtailment of private capital inflows, in part because it casts doubt on the government’s commitment to allowing private sector debt repayment. By contrast, if (part of) the private sector is unable to discharge its external obligations, this need not lead to a wider crisis; but, in practice, if private sector default is on a significant scale, commercial debt often becomes sovereign debt, through guarantees, bank bailouts, and so on. In an emerging market context, sovereign debt crises will primarily concern debt held by nonresidents and/or denominated in foreign currency. The term debt crisis clearly applies to cases of outright default (such as Russia in 1998), but may also include situations where there is a negotiated “voluntary” restructuring against the background of a real prospect of default (such as Ukraine in 2000), and/or cases where default is only avoided as a result of large and exceptional financing from official donors (such as Mexico in 1994–95).
A banking crisis occurs when actual or potential bank runs or failures induce banks to suspend the internal convertibility of their liabilities or compel the government to intervene to prevent this by extending assistance on a large scale. Banking crises tend to be protracted, and have severe effects on economic activity, through their impact on financial intermediation, confidence, capital flight, currency substitution, and public finances. Banking crises were relatively rare during the Bretton Woods era, due to capital and financial controls, but have become increasingly common since the 1970s, often in tandem with currency crises (Kaminsky and Reinhart, 1999).1 These definitions are taken from IMF (1998).
Assessing Fiscal Vulnerability
In principle, fiscal vulnerability can be assessed using financial market indicators. For example, spreads on foreign currency long-term government debt should reflect all information relevant to the risk of a sharp depreciation or debt default, including information about fiscal vulnerability. However, the evidence suggests that financial market indicators have often failed to signal impending crises in a timely fashion.10 Moreover, to the extent that financial market indicators do contain useful information for predicting and explaining crises, there is the separate issue concerning how to identify the specific contribution of fiscal factors.
A second approach would be to apply the “Value at Risk” (VaR) methodology to the public sector. The VaR methodology, which was developed as a means for assessing risks in private sector financial institutions, provides a single measure of the multiple risks applying to a portfolio or balance sheet.11 The growing emphasis on balance sheet problems as a major cause of financial crises led Dornbusch (2001, p. 5) to suggest that “vulnerability can, at least conceptually, be determined through a value at risk exercise: What are the relevant shocks? What are the exposure areas? How large a deterioration of the balance sheet would result?” In principle, this could be applied to the public sector balance sheet, to derive a measure of solvency—and hence vulnerability—explicitly accounting for the expected volatility of the macroeconomic and financial environment.12
There are, however, significant conceptual and practical problems with implementing the VaR methodology. An obvious problem is the lack of comprehensive data on public sector balance sheets for even the most advanced economies. Variances and covariances drawn from past periods of relative calm have also been shown to be unreliable in periods of market turbulence; the same drawback might well apply in the context of the macroeconomy and public finances. Furthermore, whereas the asset prices that are typically used in VaR models tend to follow a random walk with movements dominated by a probability distribution, this is less likely to be true for macroeconomic and public finance aggregates, which depend on behavioral relationships, including policy responses.
A more comprehensive approach to assessing country-specific fiscal vulnerability has recently been proposed. Hemming and Petrie (2002) outline a framework based on assessments of four sources of fiscal vulnerability: the initial fiscal position (i.e, deficit and debt levels), which should relate to the full range of government operations, including quasi-fiscal activities and contingent liabilities; short-term fiscal risks, including the sensitivity of fiscal outcomes two or three years ahead to variations in key underlying assumptions about the economic environment, the calling of contingent liabilities, and unanticipated spending needs; longer-term fiscal sustainability, based on the standard analysis of debt dynamics, supplemented by alternative scenarios and stress testing; and structural fiscal weaknesses, such as sizable nondiscretionary expenditure, a volatile revenue base, and limited institutional capacity for fiscal management.
This paper is informed by these approaches to assessing fiscal vulnerability, and focuses on the evidence from previous financial crises about the relationships between crises and fiscal variables. As such, the study is backward-looking in nature, but it hopefully leads to conclusions that can inform and improve forward-looking assessments of vulnerability and the likelihood of crises.
Conclusions from the Literature
This section summarizes that part of the vast literature on financial crises dealing with fiscal aspects of crises. Only the main conclusions from the literature relating to the fiscal causes of crises and fiscal consequences of crises are reported. A full literature review is provided in Appendix I.
Fiscal Causes of Crises
The theoretical literature suggests the following fiscal causes of crises.
An overly expansionary fiscal policy—the traditional flow variable channel—can lead to crises in several ways, including through a depletion of reserves eventually leading to a currency crisis; by contributing to a lending boom that leaves the banking sector vulnerable to shocks; or through a consumer boom and consequent current account deficit that becomes unsustainable. Such a policy could show up in the conventional fiscal deficit, or it may only be apparent from some alternative measure of fiscal imbalance; but, in either case, a crisis is likely to manifest itself gradually.
Bad news about contingent liabilities and/or prospective future deficits could be taken to indicate an increase in inflation or default risk, and hence cause sudden pressure on the exchange rate and/or debt rollover problems for the government. Adverse developments in the government’s balance sheet could also undermine confidence in its deposit guarantee and contribute to a bank run. On the asset side, large holdings of government debt can leave banks vulnerable to sovereign default. Again, the underlying fiscal problem may not show up in conventional deficit or debt measures, but in this case a crisis could emerge suddenly.
For emerging market economies, borrowing short term and in foreign currency are ways of resolving time-inconsistency problems related to the lack of credibility of monetary and fiscal policies. However, the currency and maturity composition of public debt can be critical to investor perceptions of government liquidity, and hence lead to self-fulfilling currency and debt crises.
Turning to the empirical literature, it is difficult to reach firm conclusions on the causes of crises. This reflects the lack of agreement on the appropriate theoretical model, which in turn is due to the complexity and heterogeneity of financial crises in emerging market economies. It also reflects problems of measurement and collinearity in the data. Nevertheless, there is a large empirical literature on crises, and on currency crises in particular. Much of the recent empirical work takes the form of EWS models, where the focus is on predicting rather than explaining crises (this literature is summarized in Table A1.1 of Appendix I). But the approach can still shed some light on the causes of crises.
Relatively few empirical studies of financial crises focus on fiscal variables. Indeed, many studies omit fiscal variables entirely. This in part reflects the fact that fiscal data are less widely available, especially at monthly or quarterly frequencies and for a long time period, compared with many monetary, financial, and real sector variables.
In empirical studies that do include fiscal variables, the evidence of direct effects of fiscal variables on crises is not strong. Some studies find the deficit (or credit to government) to be significantly correlated with currency crises, but others do not; the same applies to studies that include public debt variables. Some studies of the determinants of crisis severity that include fiscal variables find evidence that a lax fiscal policy is associated with greater pressure in the exchange market during periods of turbulence and contagion. But this finding is not robust, either to the inclusion of nonfiscal explanatory variables or to changes in data and sample period. Empirical studies of banking crises do not generally find a significant role for fiscal variables. The literature review uncovered only one (recent) cross-section study of debt crises, and this finds the deficit to be uncorrelated with debt crises.
Measurement problems associated with most fiscal variables, and the deficit in particular, could play a role in explaining the absence of a strong correlation with crises. Kharas and Mishra (2001) test this possibility by constructing an alternative measure of the fiscal deficit based on the change in the stock of base money plus government debt (which is viewed as a proxy for the change in government net worth). They refer to this measure as the actuarial deficit, and find that it significantly outperforms conventional deficit measures in explaining the number of currency crises experienced in a sample of industrial countries and emerging market economies.
There is evidence of a range of indirect means by which fiscal policy can contribute to crises. Several currency crisis studies find that a measure of the fiscal deficit is significant when interacted with the real exchange rate. Two causes of banking crises that are generally deemed robust—rapid domestic credit growth and high real interest rates—could reflect an indirect causal effect from loose fiscal policy. Recent empirical studies have also found that the extent of state ownership of banks is correlated with underdeveloped, poorly functioning—and, by implication, more crisis-prone—banking sectors.
As regards debt crises in particular, there is some support for fiscal policy as a cause of crises via the liquidity channel. Thus Detragiache and Spilimbergo (2001) find that the probability of a debt crisis increases with the proportion of short-term debt (public and private) and debt service falling due. Recent studies of debt crises in Latin America identify several key causes related to fiscal policy. First, in most cases, external debt is concentrated in the public sector and is high relative to exports and tax revenue. This means that devaluation provides a limited boost to activity and hence government revenues, but causes a large increase in debt-service costs. Second, a history of macroeconomic volatility exacerbates liquidity problems and increases default risk. Fiscal policy is an important source of volatility, reflecting a ready willingness to implement discretionary measures and the recognized tendency for fiscal policy in Latin America to be procyclical.
Fiscal Consequences of Crises
There are relatively few empirical studies of the effects of crises on fiscal policy. Green and Campos (2001) find that the Asian crisis resulted in a large fall-off in revenues, due to enterprise insolvency and declining trade and personal incomes; devaluation rapidly increased the domestic cost of external debt service, particularly in those countries with a worse precrisis fiscal situation; and concerns about banking and corporate sector collapse led to massive infusions of public funds. As regards the fiscal costs of crises, Burnside, Eichenbaum, and Rebelo (2001) estimate that 30 percent of the costs of the Mexican crisis in 1994–95 had been paid for through 2000 by a combination of debt deflation, fiscal reform, and seigniorage, and that seigniorage could cover the remaining costs. They also predicted that the costs of the Korean crisis in 1997 would be paid for through future fiscal reform. Other studies have reported the direct costs of resolving banking sector problems, which in some cases exceed 40 percent of GDP.