An econometric investigation on the determination of the monetary base and money multiplier in Korea.

The purpose of this study was to investigate the behavior of the determinants of the Korean money supply and the various factors affecting it. Three major subjects were examined: (1) An investigation was made as to how monetary authorities respond to economy goal variables, such as output, unemployment and the inflation rate by using the monetary policy reaction function. (2) The roles of important variables that determine the money multiplier and each component of the monetary base comprising the level of the money supply were determined. (3) The predictability of the structural and ARIMA model was examined to see which one is a more appropriate model for forecasting the money multiplier; a naive forecasting technique was also introduced to compare with the structural and time series approaches. The empirical findings are as follows: (1) The growth rate of the monetary base responded to the growth rate of output under an accommodation policy while it responded countercyclically to the inflation rate and the unemployment rate. The effect of the balance of payments was almost totally sterilized, and the effect of budget deficit was not monetized. (2) During the period of a M1 monetary target, non-policy variables contributed 74.9 percent to the growth rate of M1 but the portion of the contribution of non-policy variables declined to 58.5 percent during the M2 monetary target period. Policy variables such as the required reserve ratio on both time deposits and demand deposits played a relatively more important role in determining the level of money stock M1 while the non-policy variable of currency ratio had an important role in determining the level of M2. (3) The structural forecasting approach for the multiplier is based on the simulation of three determinants--the currency ratio, the time deposit ratio and the excess reserve ratio. The currency ratio and time deposit ratio are expressed as a function of the average family real income, rate of interest on time deposits and an average return on stock and bonds. The excess reserve ratio is expressed as a function of the interest rate on loans and the discount rate. The structural approach, which uses the actual values for exogenous variables, is the best forecasting model for the money multiplier m1 while the component approach of ARIMA which is dynamic simulation is the best for the multiplier m2