Modelling Exchange Rates Returns Using a Nested Design Model

This paper examines cross-country currency variations regarding using a “nested design model“. It attempts to address the following three fundamental questions frequently asked by market practitioners. Is the market efficientŒ How significant is the variation of currencies volatility across countriesŒ How significant is the risk premium in the currency marketŒ The logic of these questions is that, if countries have similar trade policies, then they are likely to share common market risk factors. These factors could include, among others, a common risk shared by the various currencies. The study here compares the prediction power of two models. The first is by Levich (1978) and Frankel (1982). Wilkinson and Rogers (1973) motivated the second in the experimental design framework. In the model due to Wilkinson and Rogers (!973), we incorporated a time component into the model structure to yield what we term the nested design model (or nested exchange rate model). This model is then compared with a traditional exchange rate model. The daily spot and forward exchange rate data (in US dollar equivalence) from six developed countries were used in the study. The speculative returns and forward premium of both models were calculated. Evidence here suggests that the nested exchange rate model appears to be a better model.