Modeling Oil Price, Exchange Rate, and Interest Rate Volatility in Pakistan

Resumen

The purpose of this study is to model and forecast volatility of Brent and West Texas Intermediate crude oil prices, nominal and real exchange rate, and money market rate and Treasury bill rate in the context of Pakistan. For this purpose, five linear and four non-linear models are used. The estimation period is January 1985 to December 2013.  The in-sample estimation results show that the asymmetric GARCH family models well captured the volatility dynamics as compared to the symmetric GARCH models. We show that for NERR, RERR, MMR, and TBR linear models forecasted well as compared to the non-linear models. However, for both CBNR and WTIR, non-linear models forecasted well. Overall, this study concluded that linear models forecasted well as compared to the non-linear models on the basis of minimum value of RMSE. These findings are important for the policy makers, investors, and financial market participants for making appropriate policies, investment, and asset performance evaluation.