The IMF Research Bulletin, a quarterly publication, selectively summarizes research and analytical work done by various departments at the IMF, and also provides a listing of research documents and other research-related activities, including conferences and seminars. The Bulletin is intended to serve as a summary guide to research done at the IMF on various topics, and to provide a better perspective on the analytical underpinnings of the IMF’s operational work.

Abstract

The IMF Research Bulletin, a quarterly publication, selectively summarizes research and analytical work done by various departments at the IMF, and also provides a listing of research documents and other research-related activities, including conferences and seminars. The Bulletin is intended to serve as a summary guide to research done at the IMF on various topics, and to provide a better perspective on the analytical underpinnings of the IMF’s operational work.

Abdul Abiad

Following the crises of the 1990s, the IMF has placed greater emphasis on strengthening its crisis prevention capabilities. One aspect of this work involves developing early warning systems that assess crisis vulnerability, which supplement in-depth country analyses and permit cross-country comparisons. Recent IMF research on the subject has focused on identifying potential indicators of vulnerability, finding the most objective way to synthesize the information from these indicators, and assessing their predictive performance.

Crisis prevention is one of the core responsibilities of the IMF. As such, detecting incipient vulnerabilities plays a central role in IMF work, and much research has been devoted to this task in recent years.

This article focuses on the development of early warning systems to detect currency crises.1

The building blocks of an early warning system consist of a set of vulnerability indicators. Although a large number of indicators are suggested by economic theory—any variable that influences the trade-off between defending the currency or letting it float is a potential candidate—the set of variables is usually constrained by data availability; it is from this set that indicators are selected, based on their predictive ability. Following the Mexican crisis in December 1994, several studies, including Frankel and Rose (1996), Kaminsky, Lizondo, and Reinhart (1998), and Sachs, Tornell, and Velasco (1996), have sought to identify the macroeconomic variables that tend to behave “abnormally” in the run-up to a crisis.2 The variables identified as useful have varied somewhat from study to study, owing to differences in crisis definitions, data sets, and methodologies. However, a small set of indicators—real-exchange-rate overvaluation, reserve adequacy (relative to short-term debt or broad money), domestic credit growth, current account, export growth, and reserves growth—perform well in most studies.

The Asian crisis and the ensuing interest in balance sheets and institutional factors have led researchers to look beyond the standard macroeconomic indicators for signs of vulnerability. Using firm-level data on corporate balance sheets, Mulder, Perrelli, and Rocha (2002) suggest that high leverage and short maturity structures increase both the likelihood and the depth of crises, and that shareholder rights’ protection also significantly affects crisis probabilities. Bussière and Mulder (1999b) document potential interactions between economic and political factors: when reserves are low and fundamentals are weak, political instability worsens crisis severity. Rossi (1999) suggests that countries with a less open capital account, stronger bank supervision, and weaker depositor safety are less prone to currency crashes.

Several studies (Goldfajn and Valdés, 1997; Berg and others, 1999; Berg, Borensztein, and Pattillo, forthcoming) have documented the poor performance of exchange rate expectations, credit ratings, bond spreads, and interest differentials in anticipating currency crises. However, recent research has also found that less conventional measures from financial markets might eventually prove useful as early warning indicators. Examining the period preceding the Mexican devaluation of 1994, Becker, Gelos, and Richards (2000) find that despite an alleged peso overvaluation, shares of firms with high net exports outperformed the market. This pattern is consistent with forward-looking stock prices that assigned an increasing probability to a devaluation benefiting exporting firms. Gelos and Borensztein (2003) find that emerging market mutual funds withdrew their money one month prior to the onset of a crisis. Finally, Prati and Sbracia (2002) show that uncertainty regarding fundamentals—as measured by the dispersion of GDP growth forecasts—is associated with speculative attacks and the effect depends on whether the expected fundamentals are “good” or “bad.”

After a set of useful indicators has been identified, the information contained in the indicators needs to be combined in an objective manner. The two workhorse models in the literature are the indicators model of Kaminsky, Lizondo, and Reinhart (1998) and Kaminsky (1999), and limited dependent variable probit/logit models. Berg and Pattillo (1999a) ask whether these models would have predicted the Asian crisis, had they been in place in late 1996. They find that the best models have real, but limited, predictive power.

Subsequent studies have developed alternative strategies in an attempt to improve predictive performance. These include embedding binary indicators, and a piecewise-linear generalization, in a probit model (Berg and Pattillo, 1999b); refining the crisis variable from a binary to a ternary variable (Mathieson and Yao, forthcoming); focusing only on successful speculative attacks (Kumar, Moorthy, and Perraudin, 2002); and addressing model uncertainty issues (Stone and Weeks, 2001). Other studies explore more sophisticated econometric techniques, such as auto regressive conditional hazard models (Zhang, 2001) and regime switching with time-varying probabilities (Abiad, 2003). Some studies depart from a regression-based setting altogether: Osband and Van Rijckeghem (2000) use classification rules to identify safety zones for fundamentals under which currency crashes are unlikely to occur, and Ghosh and Ghosh (2002) use a binary recursive tree technique to explore complex interactions among structural and economic variables. Many of the proposed approaches look promising, and most report some improvement over the benchmark models. Differing crisis definitions, sample coverage, and forecasting horizons make direct comparison impossible, and a true “horse race” on a single data set is still needed to help identify the strengths and weaknesses of these recent models.

In the end, however, the true test of these models lies in operationalizing them within a comprehensive framework that includes in-depth country vulnerability analyses. This has been done for the indicators and probit models, and their performance has been evaluated by Berg, Borensztein, and Pattillo (forthcoming), who emphasize the importance of out-of-sample forecasting in model evaluation. They find that the IMF’s probit model and the Kaminsky, Lizondo, and Reinhart models have done reasonably well since their operational adoption about four years ago. The March 2002 issue of the IMF’s Global Financial Stability Report (Chapter 4) also assesses the performance of early warning systems and lays out possible directions for future work in this area.

Books from the IMF

Japan’s Lost Decade: Policies for Economic Revival

Edited by Tim Callen and Jonathan D. Ostry

It has been 13 years since the bursting of the asset price bubble in Japan, yet the economy remains weighed down by excess capacity and debt, anemic growth, and entrenched deflationary pressures. If Japan’s weak growth continues, this would affect not only its own people but also the world economy generally, given Japan’s importance as a trading partner and supplier of capital.

Japan’s Lost Decade: Policies for Economic Revival evaluates the key issues facing the Japanese economy and policymakers. The book argues that an integrated policy strategy—combining decisive bank and corporate restructuring with supportive macroeconomic policies—is needed to lift the economy from its prolonged period of lackluster performance. While welcoming the progress that the Koizumi government has made in addressing the weaknesses in the Japanese economy, the book stresses that the reform strategy is not sufficiently comprehensive and that implementation has been slower and less complete than hoped for.

The book reflects the analysis prepared by the IMF staff working on Japan. Contributors include Tim Callen, Jonathan Ostry, Taimur Baig, Giovanni Dell’Ariccia, Hamid Faruqee, Ben Hunt, Sanjay Kalra, Kenneth Kang, Douglas Laxton, Martin Muhleisen, Takashi Nagaoka, Ramana Ramaswamy, Hossein Samiei, and Warwick McKibbin.

References

  • Abiad, Abdul, 2003, “Early Warning Systems: A Survey and a Regime-Switching Approach,” IMF Working Paper 03/32.

  • Aziz, Jahangir, Francesco Caramazza, and Ranil Salgado, 2000, “Currency Crises: In Search of Common Elements,” IMF Working Paper 00/67.

    • Search Google Scholar
    • Export Citation
  • Becker, Torbjorn, Gaston Gelos, and Anthony Richards, 2000, “Devaluation Expectations and the Stock Market: A New Measure and an Application to Mexico, 1994–95,” IMF Working Paper 00/28; also published in International Journal of Finance and Economics, 2002, Vol. 7, No. 3 (July), pp. 195214.

    • Search Google Scholar
    • Export Citation
  • Berg, Andrew, and Catherine Pattillo, 1999a, “Are Currency Crises Predictable? A Test,” IMF Staff Papers, Vol. 46, No. 2 (June), pp. 10738.

    • Search Google Scholar
    • Export Citation
  • Berg, Andrew, and Catherine Pattillo, 1999b, “Predicting Currency Crises: The Indicators Approach and an Alternative,” Journal of International Money and Finance, Vol. 18, No. 4 (August), pp. 56186.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Berg, Andrew, Eduardo Borensztein, and Catherine Pattillo, “Assessing Early Warning Systems: How Have They Worked in Practice?” IMF Working Paper, forthcoming.

    • Search Google Scholar
    • Export Citation
  • Berg, Andrew, Eduardo Borensztein, Gian Maria Milesi-Ferretti, and Catherine Pattillo 1999, Anticipating Balance of Payments Crises: The Role of Early Warning Systems, IMF Occasional Paper No. 186.

    • Search Google Scholar
    • Export Citation
  • Bussière, Matthieu, and Christian Mulder, 1999a, “External Vulnerability in Emerging Market Economies: How High Liquidity Can Offset Weak Fundamentals and the Effects of Contagion,” IMF Working Paper 99/88.

    • Search Google Scholar
    • Export Citation
  • Bussière, Matthieu, and Christian Mulder, 1999b, “Political Instability and Economic Vulnerability,” IMF Working Paper 99/46; also published in International Journal of Finance and Economics, 2000, Vol. 5, No. 4 (October), pp. 30930.

    • Search Google Scholar
    • Export Citation
  • Detragiache, Enrica, 2001, “Systemic Banking Crises: Causes, Consequences, and Policy Lessons,” IMF Research Bulletin, Vol. 2, No. 1 (March).

    • Search Google Scholar
    • Export Citation
  • Detragiache, Enrica, and Antonio Spilimbergo, 2001, “Crises and Liquidity: Evidence and Interpretation,” IMF Working Paper 01/2.

  • Frankel, Jeffrey, and Andrew Rose, 1996, “Currency Crashes in Emerging Markets: An Empirical Treatment,” Journal of International Economics, Vol. 41 (November), pp. 35166.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Gelos, Gaston, and Eduardo Borensztein, 2000, “A Panic-Prone Pack? The Behavior of Emerging Market Mutual Funds,” IMF Working Paper 00/198; also forthcoming in IMF Staff Papers.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Ghosh, Swati, and Atish Ghosh, 2002, “Structural Vulnerability and Currency Crises,” IMF Working Paper 02/9; also forthcoming in IMF Staff Papers.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Goldfajn, Ilan, and Rodrigo Valdés, 1997, “Are Currency Crises Predictable?” IMF Working Paper 97/159; also published in European Economic Review, 1998, Vol. 42, Nos. 3–5 (May), pp. 87385.

    • Search Google Scholar
    • Export Citation
  • Huang, Haizhou, 2000, “Financial Contagion: ABC Channels” IMF Research Bulletin, Vol. 1, No. 2 (September).

  • Kaminsky, Graciela, 1999, “Currency and Banking Crises: The Early Warnings of Distress,” IMF Working Paper 99/178.

  • Kaminsky, Graciela, Saul Lizondo, and Carmen Reinhart, 1998, “Leading Indicators of Currency Crises,” IMF Staff Papers, Vol. 45, No. 1 (March), pp. 148.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kumar, Manmohan, Uma Moorthy, and William Perraudin, 2002, “Predicting Emerging Market Currency Crashes,” IMF Working Paper 02/7.

  • Mathieson, Don, and James Yao, “Toward a New Early Warning System for Currency Crises: An Ordered Logit Approach,” IMF Working Paper, forthcoming.

    • Search Google Scholar
    • Export Citation
  • Milesi-Ferretti, Gian Maria, and Assaf Razin, 1998, “Current Account Reversals and Currency Crisis: Empirical Regularities,” IMF Working Paper 98/130; also in Currency Crises, ed. by Paul Krugman (Chicago: University of Chicago Press, 2000).

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Mulder, Christian, Roberto Perrelli, and Manuel Rocha, 2002, “The Role of Corporate, Legal and Macroeconomic Balance Sheet Indicators in Crisis Detection and Prevention” IMF Working Paper 02/59.

    • Search Google Scholar
    • Export Citation
  • Oka, Chikako, 2003, “Anticipating Arrears to the IMF: Early Warning Systems,” IMF Working Paper 03/18.

  • Osband, Kent, and Caroline Van Rijckeghem, 2000, “Safety from Currency Crashes,” IMF Staff Papers, Vol. 47, No. 2 (December), pp. 23858.

    • Search Google Scholar
    • Export Citation
  • Prati, Alessandro, and Massimo Sbracia, 2002, “Currency Crises and Uncertainty About Fundamentals,” IMF Working Paper 02/3.

  • Rossi, Marco, 1999, “Financial Fragility and Economic Performance in Developing Economies: Do Capital Controls, Prudential Regulation and Supervision Matter?” IMF Working Paper 99/66.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Sachs, Jeffrey, Aaron Tornell, and Andrés Velasco, 1996, “Financial Crises in Emerging Markets: The Lessons from 1995,” Brookings Papers on Economic Activity: 1, Brookings Institution.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Stone, Mark, and Melvyn Weeks, 2001, “Systemic Financial Crises, Balance Sheets and Model Uncertainty,” IMF Working Paper 01/162.

  • Zhang, Zhiwei, 2001, “Speculative Attacks in the Asian Crisis,” IMF Working Paper 01/189.

Visit the Research at the IMF website at www.imf.org/research

IMF Staff Papers

Volume 50, Number 2

Sources of Economic Growth in East Asia: A Nonparametric Assessment

Shigeru Iwata, Mohsin S. Khan, and Hiroshi Murao

Can Inheritances Alleviate the Fiscal Burden of an Aging Population?

Erik Lueth

Population Aging and Global Capital Flows in a Parallel Universe

Robin Brooks

The Properties of the Equity Premium and the Risk-Free Rate: An Investigation Across Time and Countries

Fabio Canova and Gianni De Nicolo

Resources and Incentives for Reform

Alberto Dalmazzo and Guido de Blasio

A Dynamic General Equilibrium Framework of Investment with Financing Constraint

Danyang Xie and Chi-Wa Yuen

Refocushig the Fund: A Review of James M. Boughton’s Silent Revolution: The International Monetary Fund, 1979–1989

Peter B. Kenen

1

Previous research summaries in the IMF Research Bulletin have covered banking crises (Detragiache, 2001) and contagion (Huang, 2000). Some of the models described in this article have also been applied to debt crises (Detragiache and Spilimbergo, 2001) and arrears to the IMF (Oka, 2003).

2

Other IMF studies identifying macroeconomic indicators include Bussière and Mulder (1999a); Aziz, Caramazza, and Salgado (2000); and Milesi-Ferretti and Razin (2000).

IMF Research Bulletin, June 2003
Author: International Monetary Fund. Research Dept.