Estimating Money Demand Function by a Smooth Transition Regression Model: An Evidence for Turkey

The money supply process is assumed to be fixed in economic literature or at least there is a central bank trying to control the liquidity in the economy. On the other hand, the demand side is more volatile and more uncertain. This situation hinders the homogenous and symmetric information assumptions of the monetary models. The amount of money demanded is a dynamic process and changes depending on the transition variable in concern. The money demand increases in the boom periods of the economy but may diminish in the recessions gradually. Therefore the money demand function indicates an asymmetric behavior and nonlinearity. This paper estimates the money demand function by including the inflation uncertainty, that is assumed to be a transition variable for a small-open economy, Turkey by using the monthly data spanning from January, 1990 to May, 2012. The parameters of the money demand function are estimated by the Smooth Transition Regression (STR) models. While modelling the nonlinearity, an appropriate logistic function is determined. The dependent variables that are used to estimate the money function are gold, interest rate, inflation uncertainty, share prices, exchange rate and income. The inflation uncertainty data is gathered from the conditional variances of a specified EGARCH model. The results of the paper have several policy implications for the monetary authorities. First, the behavior of the money demand and its determinants are crucial at the times of adopting the inflation targeting regime. The stability of money demand is also related to the stability of inflation. So the results of the paper may be beneficial for the policy makers and monetary authorities during their decision making process.