Money Demand Function for Pakistan

Abstract

The main objective of this study is to empirically estimate the long run money demand function for Pakistan using time series data. For this purpose we used annual data from 1953 to 2003. The results of unit root analysis have suggested that both log of nominal GDP and log price are possibly I (2) variables. The results of unit root with structural break for real M1 and M2 have suggested that log of real M1 is fractionally integrated where as log of real M2 is trend stationary. Results of I(2) co-integration analysis have suggested that there are some I(2) trends in the model with nominal variables hence in order to avoid complications involved in the analysis of I(2) trends we transformed our model in real variables. We found one co-integration relation for both M1 and M2. The sign of the estimated coefficients for GDP and interest rate in M1 money demand function are according to theory but coefficient of interest rate has wrong sign for M2 money demand relation but it is statistically insignificant. We accepted the hypothesis that both real GDP and interest rate are weakly exogenous in long run money demand relation for both M1 and M2. Demand for money for both M1 and M2 are found to be inelastic with respect to interest rate which is quite obvious for an underdeveloped country like Pakistan where financial markets are underdeveloped, infrastructure is poor and information system is still very slow.