Exchange rates in the modern floating era: what do we really know? created by Kenneth Rogoff

Forecasting nominal exchange rates remains a remarkably difficult task, despite the proliferation of new floating currencies, the maturation of the floating rate period, the deepening of financial markets, and the development of more sophisticated econometric tests that make use of today’s more powerful computing possibilities. Despite these advances, the basic results of Meese and Rogoff in the 1980s stand up remarkably well—it is still extremely difficult to forecast exchange rates. To the extent that there is any forecasting power, the most promising models are those based on purchasing power parity or the current account, although it must be noted that these mainly predict the real exchange rate, rather than the nominal exchange rate. Thus, some of the adjustment takes place in prices. Finally, it should be noted that panel methods help in exchange rate forecasting, albeit mainly by allowing better estimation of nonstructural factors such as shift parameters.