Financial Crises: New Data, New Evidence

Financial crises are one of the dominant features of the 1990s. The fact that so many episodes of financial turbulence and distress have been packed into fewer than ten years points to what is new and troubling about our current economic and financial environment. Or does it? The premise implicit in recent analyses - that the crisis problem is growing more severe - is just that: implicit. Virtually the entire recent literature on financial crises uses the same crisis dates for the same sample of developing countries for the same 25 years. We lack systematic, rigorous, quantitative comparisons of the last quarter century with earlier times and places. The goal of this project is therefore to construct a new database encompassing not just the 1970s, 1980s and 1990s but more than a century of historical experience. These historical data, for a substantial number of developed and emerging markets, allow this project to formulate more precise answers to the question of whether crises are growing more frequent and severe, and to construct more powerful tests of their determinants. It sheds new light on the causes and consequences of financial crises. It provides a unique resource for future investigators.

This project asks and answers questions like the following: Have there been changes in the frequency of currency and banking crises? How do they compare in terms of severity? What is the role of macroeconomic, financial and political variables in their incidence? How long has output taken to recover? What was the impact on the current account, money supply and interest rates? The project addresses these questions with a combination of statistical and historical techniques. The statistical analysis uses a newly generated set of crisis dates, together with macroeconomic variables and econometric techniques designed to disentangle the two-way causation between recessions and crises. The historical analysis looks deeply at particular episodes where there are parallels with today.

This research has immediate policy implications. If crises are indeed becoming more frequent and/or severe, then there exist important weaknesses in our current financial environment requiring urgent reform. If history shows that crises are especially pervasive and/or devastating in countries with particular exchange rate systems, monetary regimes, regulatory arrangements, and capital account regimes, then the results will have obvious implications for the choice of exchange rate regime, the desirability of capital-account liberalization, and the emphasis that should be placed on financial deregulation. The data developed for this project and the institutional variation provided by history should enable us to uncover these patterns. Even if not everyone agrees with our conclusions, our database will provide a new resource with which others can also pursue such questions.