The exchange rate is such an important macroeconomic policy variable that the debate is still raging about fixed versus flexible exchange rates. Recent developments in Europe and the move towards the European Monetary Union and a single currency point to the belief in the harm inflicted on the economy by sharp fluctuations in the exchange rate. Under a system of flexible exchange rates, central banks intervene in the market on a regular basis to smooth' and iron out' fluctuations in exchange rates. Exchange rate forecasting may take one of three forms.
The first is the forecasting of event timing, that is forecasting the timing of an event of a once-and-for-all nature. This form of forecasting is mostly relevant to fixed exchange rates, more precisely to devaluation and revaluation of currencies under a system of fixed exchange rates. Realignment of exchange rates within the European Monetary System (EMS) is an example of such an event. Forecasting the timing of these realignment is an example of the forecasting of event timing. The second form is forecasting event outcome. Sometimes a model that is based on a univariate time series of the exchange rate is called a time series model while a model that is based on a multivariate time series is called an econometric model. A distinction may also be made between structural forecasting, which is based on an explicit theory, and non-structural forecasting, which is not.
Structural forecasting is normally associated with econometric models while non-structural forecasting is associated with univariate time series models. While microeconomic forecasting pertains to the forecasting of variables specific to one sector, macroeconomic variables, one of which is the exchange rate. Therefore, exchange rate forecasting is macroeconomic forecasting. One problem with macroeconomic forecasting arises from the effect of structural breaks resulting, for example, from significant changes in the price of crude oil, inflation surges and policy shifts. This problem may make it necessary to combine the forecasts generated by a forecasting model with judgment that is to modify the raw forecasts using the forecaster's judgment.
The two most commonly used methods for exchange rate forecasting is Fundamental approach and technical approach. In fundamental approach factors like GDP, Inflation rate, productivity indices, balance of trade and unemployment is taken into consideration. In technical approach investor sentiments determines changes in the exchange rate and makes predictions by charting out patterns.
