Dekle, Hsiao, and Wang (hereafter DHW) discuss two important issues in their chapter. The first is whether interest rates affected nominal exchange rates in postcrisis Asia, and in what direction; the second is whether “news” and exchange rate movements elsewhere had an impact on the behavior of domestic exchange rates in some of the crisis-stricken Asian countries. Due to space limitation, I will focus my discussion only on the empirical evidence that they attempt to provide on the first issue.
As noted by the authors, the interest rate-exchange rate link has been discussed in a variety of other recent studies, including Ghosh and Phillips (1998), Goldfajn and Baig (1998), Goldfajn and Gupta (1999), Gould and Kamin (Chapter 11, this volume), Kraay (1998), Kaminsky and Schmukler (1998), and Furman and Stiglitz (1998). A review of this literature makes it clear that results are mixed and often not robust. Goldfajn and Baig (1998), for instance, used daily data and found no stable relationship between nominal interest rates and exchange rates. Gould and Kamin (Chapter 11, this volume), using weekly data and Granger-causality tests, were also unable to find any significant effect of interest rates on exchange rates.
Should the lack of robustness be surprising? In my view, not really. The short-run relationship between exchange rates and interest rates is well known to be unstable in “normal” times (Frankel and Rose, 1995); there is no reason to expect to find a stable relationship in the immediate aftermath of crisis episodes. A highly volatile environment is bound to make it more, not less, difficult to isolate a robust link between changes in policy instruments and forward-looking asset prices.