Modelling & Forecasting of Re/$ Exchange rate – An empirical analysis

To model the factors affecting rupee dollar exchange rateTheoreticallly justifying use of all right hand side variables affecting rupee dollar exchange rate. Using FGLS,VAR,Bounds Approach and ARIMA for estimating the relationships between rupee dollar exchangeThis study covers two main topics: first modelling, where we discussed about the importance of variables like capital inflows, order flows, central bank intervention in modelling exchange rate. We empirically estimated the coefficients of explanatory variables after overcoming autocorrelation and unit root problems. It was found that only variables order flow, forward premium, trade balance, money supply and output has significant effect on exchange rate. We also checked for any long term relationship between variables, found that only money supply and output has long run relationship with exchange rate. All significant variables shows very small impact on exchange rate except forward premium. Empirical relations between variables and exchange rate support the theoretical relations. Second forecasting of exchange rate, we forecasted exchange rate using VAR, OLS, ARIMA model for three different sample period. Evaluated the forecasting performance of models by graphs and by error statistics. We found that VAR model yield more accurate forecasts than the OLS and ARIMA as it has very low Theil’s U and RMSE for all periods. Using VAR model we also forecasted out of sample for periods from January 2014- June 2014.