Need for Mean Reversion in Forecasting Emerging Market Exchange Rates

Abstract

A lot of research has been done over the past few decades for arriving at the models which can accurately estimate the exchange rates for currencies. Various approaches including time series analysis, macro-economic factors analysis, analysis using vector algebra and stochastic processes have been considered to estimate the exchange rates. Estimation of exchange rates assumes huge prominence in the present era, where most of the countries have moved to the floating exchange rate regime from the fixed rate regime of the past. Countries across the world consider that the risk due to fluctuations in the exchange rate needs to be minimized because of the huge positions assumed by them. This paper initially provides a brief on some of the major models available in the literature for forecasting exchange rates. It further analyzes the findings available in the literature for the accuracy of the various models and summarizes findings of the various researchers for those models. Further, it tests the model developed using GBM for forecasting USD~GBP and USD~INR exchange rates to assess the need for mean reversion. From the above study, it is concluded that GBM model has some shortcomings in terms of accuracy of the forecasting for the emerging market currencies. Based on the above study, it is proposed that there is a need for mean reversion that needs to be taken into consideration for estimation of the exchange rates for the emerging market currencies.