The validity of the real interest differential model for Turkey with Markov regime switching

Abstract

Factors driving the fluctuations in exchange rates have always been on the agenda of economists. For this purpose, many theories have been developed to find the determinants of exchange rate movements. One of the best known of these theories is Frankel's (1979) Real Interest Differential (RID) Model, which explains exchange rate movements with a monetarist approach. RID model is an approach that tries to explain exchange rate changes among countries by the differences between money supplies, income levels, interest rates and expected inflation rates. Although this monetary model is useful in explaining long-term relationships, it has been considered as inadequate to explain the mutual dynamic interactions between monetary fundamentals and exchange rates that change over time. The nonlinear relationship between exchange rates and monetary fundamentals has led to the analysis of theoretical models to explain exchange rate movements with nonlinear methods. Accordingly, in this study the validity of Frankel's (1979) RID model is tested for Turkey during the 1990:01-2018:12 period by performing the Markov Regime Switching (MS) method. Findings show that there are two regimes reflecting; specifically the periods of decrease in exchange rates (first regime) and the periods of increase (second regime) throughout the analyzed period. According to the MS estimation results, while all the explanatory variables other than the relative interest rate are statistically significant in the first regime, the effect of relative income on the exchange rate is positive and contrary to expectations. In the second regime, all the explanatory variables except the relative income level are statistically significant and the coefficient signs are consistent with the predictions of the RID model. In this context, it is possible to state that the RID model helps to understand the movements in the USD/TL exchange rate in certain periods. © Peter Lang GmbH Internationaler Verlag der Wissenschaften Berlin 2020. All rights reserved.