Tests of currency substitution, capital mobility and nonlinearity of Hungary's money demand function

Abstract

The demand for M2 in Hungary is positively associated with real output and the nominal effective exchange rate and negatively influenced by the deposit rate, the euro interest rate, and expected inflation rate. The coefficient of the euro interest rate for the demand for M1 is insignificant. Hence, depreciation of the forint or a higher euro interest rate would help raise Hungary's real output. The Box–Cox transformation shows that the log-linear form for M1 or M2 demand cannot be rejected at the 5% level while the linear form for M1 or M2 demand can be rejected at the 5% level. However, the log-linear form for M2 demand can be rejected at the 10% level. The CUSUM and CUSUMSQ tests show that M2 demand based on the Box–Cox model and M1 demand are relatively stable. Based on the mean absolute percent error and the Theil inequality coefficient, M2 demand based on the Box–Cox model yields better forecasting outcomes.