Order Information

O C C A S I O N A L   P A P E R      
Fiscal Vulnerability and Financial Crises in Emerging Market Economies

Richard Hemming, Michael Kell, and Axel Schimmelpfennig

©2003 International Monetary Fund
May 23, 2003

I Overview
II Background
  Assessing Fiscal Vulnerability
  Conclusions from the Literature
III Data and Event Studies
Event Studies
Summary of Findings
IV EWS Models and the Severity of Currency Crises
Using Fiscal Variables to Predict Crises
Fiscal Variables and the Severity of Currency Crises
Summary of Findings
V Case Studies
Fiscal Causes of Crises
Fiscal Indicators
Fiscal Consequences of Crises
VI Conclusions
What Are the Fiscal Causes of Crises?
Which Fiscal Vulnerability Indicators Help to Predict Crises?
Can Fiscal Variables Explain the Severity of Currency Crises?
What Are the Fiscal Consequences of Crises?
I Literature Review
II Country and Area Coverage and Data Availability
III Event Studies
IV EWS Models and the Severity of Currency Crises
V Case Studies
2.1. Currency, Debt, and Banking Crises
3.1. Fiscal Variables
Text Tables
4.1. Summary Results from the Signals Approach
4.2. Summary of Probit EWS Results
4.3. Explaining the Severity of Currency Crises (STV Approach)
4.4. Explaining Changes in the FMP Index (Panel Approach, Fixed Effects)
5.1. Causes of Crises
5.2. Fiscal Consequences of Crises
Appendix Tables
A1.1. Literature Review: Empirical Studies of Financial Crises
A2.1. Countries and Areas Included in EWS and Event Studies
A2.2. Dates of Crisis Episodes
A2.3. Description and Sources of Fiscal Variables
A2.4. Data Availability
A4.1. Performance of Indicators of Currency Crises Using the Signals Approach
A4.2. Performance of Indicators of Debt Crises Using the Signals Approach
A4.3. Performance of Indicators of Banking Crises Using theSignals Approach
A4.4. Country-Specific Thresholds for Currency Crises
A4.5. Country-Specific Thresholds for Debt Crises
A4.6. Country-Specific Thresholds for Banking Crises
A4.7. Regression Results for the DCSD Specification (Maximum Sample)
A4.8. Regression Results for Deficit and Financing Variables (Joint Sample)
A4.9. Regression Results for Reduced Set of Deficit and Financing Variables (Maximum Sample)
A4.10. Regression Results for Selected Fiscal Variables
A4.11. Explaining Crisis Depth (Benchmark Specification)
A4.12. Best Performers Among Fiscal Variables
A4.13. Benchmark Specification for Panel Approach (Fixed Effects)
A4.14. Including Deficit and Financing Variables in the Benchmark Specification (Panel Approach, Fixed Effects)
A4.15. Including Public Debt Variables Individually in the Benchmark Specification (Panel Approach, Fixed Effects)
A4.16. Including Government Expenditure Variables in the Benchmark Specification (Panel Approach, Fixed Effects)
A4.17. Best Performers Among Fiscal Variables (Panel Approach, Fixed Effects)
A5.1. Mexico: Selected Fiscal Vulnerability Indicators (Annual)
A5.2. Mexico: Selected Fiscal Vulnerability Indicators (Quarterly)
A5.3. Argentina: Selected Fiscal Vulnerability Indicators (Annual)
A5.4. Argentina: Selected Fiscal Vulnerability Indicators (Quarterly)
A5.5. Bulgaria: Selected Fiscal Vulnerability Indicators
A5.6. Czech Republic: Selected Fiscal Vulnerability Indicators
A5.7. Thailand: Selected Fiscal Vulnerability Indicators (Annual)
A5.8. Thailand: Selected Fiscal Vulnerability Indicators (Quarterly)
A5.9. Korea: Selected Fiscal Vulnerability Indicators (Annual)
A5.10. Korea: Selected Fiscal Vulnerability Indicators (Quarterly)
A5.11. Pakistan: Selected Fiscal Vulnerability Indicators
A5.12. Russia: Selected Fiscal Vulnerability Indicators (Annual)
A5.13. Russia: Selected Fiscal Vulnerability Indicators (Quarterly)
A5.14. Ukraine: Selected Fiscal Vulnerability Indicators
A5.15. Brazil: Selected Fiscal Vulnerability Indicators
A5.16. Ecuador: Selected Fiscal Vulnerability Indicators
3.1. Number of Crises in the Sample
3.2. Overall Balance
3.3. Public External Debt
3.4. Short-Term Debt
Appendix Figures
A3.1. Actuarial Deficit
A3.2. Total Financing
A3.3. Change in Net Claims on Government
A3.4. Foreign Debt
A3.5. Total Expenditure
A3.6. Interest Expenditure
A3.7. Social Expenditure
A3.8. International Trade Taxes

I. Overview

Fiscal problems have long been considered a central feature of financial (i.e., currency, debt, and banking) crises. However, fiscal problems received less attention at the time of the Asian crisis, because financial and corporate sector vulnerabilities were seen to be more important than fiscal vulnerability.1 But the recent crises in Argentina and Turkey illustrate the continuing importance of fiscal problems in precipitating crises. Moreover, whatever their cause, financial crises always have important fiscal dimensions.

This paper focuses on the fiscal aspects of financial crises in emerging market economies. As such, it is intended to complement other IMF work in progress on vulnerability and crises.2 In particular, the paper seeks to provide answers to the following four questions:

  • What are the fiscal causes of crises?

  • Which fiscal vulnerability indicators help to predict crises?

  • Can fiscal variables explain the severity of crises?

  • What are the fiscal consequences of crises?

Two empirical approaches are followed. First, a large set of fiscal variables for 29 emerging market economies over the period 1970–2000 is used to examine whether: there are systematic patterns in fiscal variables in the periods before and after crises; fiscal variables can improve the performance of existing early warning system (EWS) models in predicting crises; and fiscal variables help to explain the severity of currency crises. Second, detailed case studies of 11 recent crises in emerging market economies focus on some of the structural and institutional dimensions of fiscal vulnerability.

The main conclusions of the paper are the following:

  • Fiscal policy has contributed significantly to past crises. While previous cross-country empirical studies have not found a significant role for fiscal variables in explaining currency crises, this paper concludes that fiscal variables are correlated with crises. The change in net claims of the banking sector on government, together with public debt level and composition variables, are consistently important. Some structural variables, reflecting expenditure and revenue rigidities, also matter. Country case studies reveal the importance of fiscal variables even more clearly, with fiscal problems being the direct cause of crises in a number of countries, and a contributing factor elsewhere. But in a few countries, fiscal problems did not play a catalytic role.

  • Fiscal variables add relatively little to the predictive capabilities of existing EWS models. The main contribution of fiscal variables to such models is that their inclusion reduces the chances of sending false alarms. However, closer analysis of crisis events suggests that fiscal deficits are higher than normal in the run-up to crises. This is not the case for debt as whole, but short-term debt turns out to be one of the best predictors of currency, debt, and banking crises. Fiscal variables are also found to contribute to pressure on the foreign exchange market in both crisis and noncrisis periods, although their impact is dominated by other variables.

  • Crises have significant fiscal consequences. Fiscal deficits tend to decline following crises, the adverse impact of slower growth on revenue and expenditure being more than offset by the discretionary fiscal adjustment normally called for in response to a crisis. Debt tends to increase, driving forces being the recognition and securitization of contingent liabilities associated with bank restructuring and often sizable exchange rate effects. But the share of short-term debt in the total is reduced. Expenditure and revenue changes are not particularly marked.

The paper is structured as follows. Section II provides background: it defines financial crises and fiscal vulnerability; discusses the different ways in which fiscal vulnerability can be assessed; and reports the main conclusions from the theoretical and empirical literature on financial crises. This literature is more extensively reviewed in Appendix I. Section III describes the dataset constructed for this study, and presents an event study analysis that summarizes the univariate correlations between crises and fiscal variables. More details about the dataset and the event studies are provided in Appendixes II and III. Section IV reports the main results of econometric analysis that uses fiscal variables to predict the timing of financial crises and investigates the role of fiscal variables in explaining the severity of currency crises. Further details of the methodology and results are contained in Appendix IV. Section V presents the results from 11 case studies of emerging market crises in the 1990s; the case studies themselves are in Appendix V. Section VI concludes by providing answers to the four questions listed above.

1For example, in several recent overviews of the emerging market crises of the 1990s-Beim and Calomiris (2001); Glick, Moreno, and Spiegel (2001); and Feldstein (2002)-the coverage of fiscal issues is marginal at best.
2As reported in IMF (2002a); Allen and others (2002); and Berg, Borensztein, and Pattillo (2003).