ABSTRACT
The behavior of the dollar exchange rates of the last few years has been perplexing in many ways. Not only has it turned out to be almost impossible to predict the movements of the dollar (something which is not really surprising); more importantly, the link between the dollar exchange rates and the "fundamental" variables (the money stock, the interest rates, the business cycle, etc.) has been very tenuous. For example, those who thought that there was some predictable relationship between economic activity and the dollar, or between the interest rate movements and the movements of the dollar rates, have found that this relationship is very unstable. In 1994, when the Federal Reserve started to raise the domestic interest rates and the Bundesbank initiated a policy of lower domestic rates, most analysts predicted that the dollar would increase in value. Exactly the opposite happened. Put differently, although many economists today have lost their ambition to forecast exchange rates, most of them would like to be confident in making conditional forecasts — i.e., to be able to predict what the dollar will do when, say, the money stock is reduced in the USA. It turns out that these conditional forecasts are as difficult to make as the unconditional ones. In this paper we try to explain this puzzling phenomenon of the weak link between exchange rates and the fundamental variables.
