A Reconsideration of the Meese-Rogoff Puzzle: An Alternative Approach to Model Estimation and Forecast Evaluation

This study revisits the Meese-Rogoff puzzle by estimating the traditional monetary models of exchange rate determination in state-space form and comparing the accuracy of these forecasts against the naïve random walk model using a wide range of conventional and alternative measures of forecasting accuracy. The results demonstrate that incorporating stochastic movements in the parameters of exchange rate models does not enable the Meese-Rogoff puzzle to be overturned. However, estimating these models in state-space form substantially improves forecasting accuracy to the extent that the model and random walk produce an equivalent magnitude of error. Furthermore, the results prove that the Meese-Rogoff puzzle can be overturned if the forecasts are evaluated by alternative criteria. These criteria include direction accuracy, profitability, and measures that jointly take into account both magnitude and direction accuracy.