Demand for money in inflation-targeting monetary policy

In order to study the role of money in an inflation-targeting regime for monetary policy, we compare the interest rate and money as monetary policy instruments.The theoretical part of the study builds on a dynamic stochastic general equilibrium model that combines the money-in-the-utility-function approach with sticky prices.Preference and technology shocks are the driving forces of the economy.We show that conditioning the interest rate on the expected future cost change can be used to achieve constant inflation or constant inflation expectations.The assumed adjustment costs in 'money demand' lead to an equilibrium in which inflation can be controlled by money growth without having information on the current state of the economy.The tradeoff between money and the interest rate as a monetary policy instrument depends on the parameter stability of the technology change process relative to that of the 'money demand' function. We experiment with the parameter stability of the demand for money using Finnish monthly data for 1980 - 1995.The steadystate - utility function - parameters of the model of narrow money (M1), estimated with cointegration techniques, are stable; whereas in the model of harmonized M3 (M3H) the parameters are not stable.The theoretical model fits the M1 data.The adjustment cost parameters of the M1 model describing the dynamics of the demand for money could indicate the occurrence of technological improvements in banking and payments during the sample period.These results suggest that from the Finnish viewpoint M1 would be a more appropriate intermediate target for monetary policy than harmonized M3.Due to small sample problems, we compare parameters of the theoretical model estimated using the Generalized Method of Moments and Full Information Maximum Likelihood method.The process driving the forcing variables is approximated with vector autoregression. Both the GMM and FIML parameter estimates are reasonable and the differences are negligible.The cross-equation restrictions implied by the rational expectations hypothesis are clearly rejected.