engleski

Interest Rate Modelling Of Croatia: From Kitchen Sink Model, And Granger Causality Testing To VAR Models

This paper presents empirical tests of the hypothesis that higher budget deficits in Republic of Croatia raise real interest rates. The interest-rate equation used for the tests in this paper is based upon the type of equation in Feldstein and Eckstein (1970). Such an equation can be developed from a simple modified LM curve giving the condition for equality of the supply of and demand for the monetary supply. The basic model is estimated using quarterly data for the sample period 1995 QI - 2004 QIII. In modelling the interest rate we used the kitchen sink model, satisfactory model is obtained within the reduced model framework which left out insignificant variable ; because the budged deficit is left out as insignificant variable we didn’ t proof the main hypothesis, but we find some interesting implication examined by the Granger causality test: the first of all is evidence of money nonneutrality because increase of money supply does Granger-cause interest rate to fall (as real variable). In the last part of paper is a VAR model with three endogenous variables and a fourth-order autoregressive specification. We use the quarterly data for a three-equation model relating real money supply (M), interest rate (r), and real output (x), because, previous results indicated that the money supply and real output remain significant in the reduced model of interest rate. The main results of VAR models is tested by the MAPE value which is less than 1 percent for the real output series, whereas other two series they are much higher beside for the real money supply in the case of the final model. The final model of real money supply and real output have a strong edge over unrestricted model in terms of predictability outside the sample.

LM curve; interest rate; kitchen-sink model; Granger-causality; VAR-models

nije evidentirano