Indeed, this result is derived in a model of perfect capital mobility and sticky prices.
The overshooting paper not only was a great piece of research, but also had important policy implications. In the context of flexible exchange rates, not only among major currencies, but also increasingly with emerging market currencies, the excessive volatility is usually mentioned as the main disadvantage of free floating. A policy sequel is that overshooting is often used to justify intervening in foreign exchange markets. This is also a strong reason why policymakers suffer from "fear of floating" (Calvo and Reinhart, 2002).
From the empirical point of view, the evidence has been mixed and there are several dimensions in which the model performs poorly.1 Starting from the "exchange rate disconnect puzzle" from Meese and Rogoff (1983), which shows that no structural model can predict exchange rates, not even monetary ones, there have been many attempts to explain exchange rate fluctuations. Faust and Rogers (2003) and, more recently, Bjjournland (2006) propose new identification restrictions that reduce this delayed overshooting. Although the researcher does not intend to address empirically the overshooting hypothesis, it is useful to review analytically the robustness of overshooting and which type of conditions are required to generate a different behavior of exchange rates.
The researcher plans to examine the conditions under which the exchange rate undershoots instead of overshoots as in the original model. This could help to reconcile the evidence with Dornbush's model (Rogoff, 2002). However, in the basic theoretical framework, the conditions to generate undershooting are rather contrived, namely, that the interest rate rises as a result of a monetary expansion. Therefore, under perfect capital mobility, with the consequent uncovered interest rate parity, overshooting should be a natural outcome. I also show that dropping perfect capital mobility as suggested by Frenkel and Rodriguez (1982) also requires special conditions. In such case it would be necessary for the current account deficit to narrow after a monetary expansion.
Empirical Analysis: Exchange rate shooting during Financial Crises of 1990s
This part of the papers documents main characteristics of the exchange rate movement in the countries that experienced currency crises in the 1990s. First, the researcher introduces the data set that we use and then analyze the exchange rate movements in these countries to examine the existence of any systematic regularity that derives the exchange rate overshooting.
The sample includes currency crises in the 1990s. First, the researcher collects all episodes of speculative attacks in the 1990s based on Glick and Rose (1999). The researcher excludes unsuccessful speculative attacks where countries maintained stable exchange rates even under the pressure of speculative attacks. the researcher also excludes a few recent cases in which the exchange rate is still unstable and the complete exchange rate dynamics during currency crises are not revealed. This selection process reduces the available data set to 24 episodes, which consist of 10 cases in the 1992-3 European crisis, 4 cases in the 1994-5 Mexican crisis, and 10 cases in the 1997-8 Asian crisis and related others. List of countries is reported