Using the asset market approach to estimate future exchange rates in a developing country:the case of Kenya

Abstract

The study investigates the short term behaviour of exchange rates in Kenya with the aim of making relatively accurate short term forecasts using the asset market approach to exchange rate modelling. A sample of six different currencies is used to bring out the relationship between the short term movement of exchange rates and the nominal interest rates. A regression of the current exchange rate against the previous one and the differential interest rates reveals that the time series of exchange rates is predominantly a first order autoregressive process. The inclusion of interest rates does not significantly improve the data fitting performance of the model,though it appears to improve the predictive performance. To investigate this unexpected finding,the first order difference in exchange rates is regressed against the the interest differential. The result gives poor data fitting and the F-test confirms that the regression is not significant. A close examination of the time series of the first order difference of exchange rates using the autocorrelation function shows that it is very close to white noise,thus suggesting that exchange rate movement in Kenya have closely .followed a random walk process with a drift.