Unpredictable After All? A short note on exchange rate predictability

Earlier research has shown that it is very hard to outperform the random walk model with respect to forecasting exchange rates. In this paper we propose an extension to the regular regime-switching model in order to capture the exchange rate dynamics. The model is extended by including macro-economic variables, like inflation and interest rates, into both regimes. The regimes not only have different means and volatility's, but also different sensitivities to the macro-economic variables. We will show that this approach doesn't work over the whole sample, although previous research work might indicate otherwise. Furthermore we will elaborate on sub samples, in which the model showed a better performance than the random walk model, and show that this is rather coincidence.