MODELLING THE DEMAND FOR MONEY : THE CASE OF INDONESIA

Abstract

The apparent breakdown of the partial adjustment model as the conventional approach to the demand for money has cast serious doubt on the adequacy of its economics and econometrics specifications. Two competing research strategies, the buffer stock money and the error correction approaches, have been proposed to surmount the difficulties with the partial adjustment model. This thesis is an attempt to investigate the most appropriate modelling strategy of the demand for money in the case of Indonesia by evaluating the empirical performances of partial adjustment, buffer stock, and error-correction models. The period of the study, 1978.1 to 1992.4, covers the era of financial and monetary reform which caused financial innovation. Therefore, this thesis also examines the effects of financial innovation on the demand for money in Indonesia. Chapter 2 presents the outstanding problems and current developments on the demand for money. Here we start by providing the rationale of the partial adjustment model and listing of some its empirical and theoretical difficulties. Next we present the idea of the buffer stock and error-correction models as the current developments in the demand for money. Since the error-correction model intimates with the relatively new concept of cointegration, we also provide a brief on the rationale of cointegration and its testing procedures at the end of this chapter. In Chapter 3, we provide the developments of monetary aggregates in Indonesia as useful grounds in constructing the models and interpreting the empirical results. In general, the results of this thesis suggest that the demand for money has not yet been as seriously distorted by financial innovation as might have been expected. In fact a new stable relationship may now exist after the transitional period of financial innovation. In assessing the demand for money, we suggest that the partial adjustment approach might not be the best strategy for modelling the demand for money when money supply is treated as exogeneous. The Chow tests for parameter instability show that partial adjustment performed poorly during the period of exogeneous money supply shocks. The buffer stock money approach appears successful in providing a plausible explanation for the long lag structure in the partial adjustment model and for the inflation characteristic in Indonesia post the reform. This model, however, fails to reject parameter instability tests. This unsatisfactory result might be because of the restrictive specification and the financial innovation that drives the longer term idea of buffer stock. The consequences of longer term buffer stock may be vital for monetary control. If buffer balances are induced by high real interest rates, then they will be spent when interest rates fall with consequent inflanationary effect. Conversely at times of monetary restraint, balances that have amassed for investment characteristics may be used to sustain inflationary levels of spending. The buffer stock model then provides the explanation of inflation mechanism in Indonesia. The combination of cointegration analysis and error correction are found to be succesful in providing satisfactory results. Evidence of cointegration increases confidence about the presence of equilibrium relationship and the stabilizing characteristics of monetary policy. The short-run disequilibrium is modelled dynamically by using error correction. Two approaches are used, the Engle-Granger and the Bardsen. The two-stage Engle-Granger works well on narrow money, while the Bardsen gives satisfactory results for broad money.