Most Speculative Attacks Do Not Succeed: Currency Crises and Currency Crashes

Abstract

While there has been considerable research on currency crises, relatively little attention has been given to whether they are successful or not. We investigate this question for a set of 32 emerging market economies for the period 1980–2014. In the literature, many different measures of currency crises have been used, but almost all use some variants of exchange market pressure indices that look at changes in exchange rates, international reserves, and often also interest rates. These vary mainly in their specific specifications such as how to weigh the different variables. Therefore, to check the robustness of our results we use six different specifications. A second type of measure is also sometimes used. These focus only on large depreciations of exchange rates. While often called measures of currency crises they are really only measures of currency crashes. We thus take this approach as a measure of successful attacks. Using a wider range of thresholds than studies such as Lavean and Valencia’s, a well-known dataset of different types of financial crises, we still find that the vast majority of speculative attacks are not successful.