Demand for Money in Ghana: An Empirical Study of the Pre-Inflation Targeting Period (1960-2006)

Abstract

This paper estimates the demand for real money balances using the Ghanaian data for the pre-inflation targeting period. It applies the Hassler-Berlin’s modified two-step Engel-Granger approach to test for cointegration with different forms of structural shifts. The paper first estimated the long-run demand for real cedi balances for the pre- and post-1983 reforms. The results suggest that, although there was a stable and well-defined real demand for the cedi for both narrow and broad definitions, the stability had broken down around 1983. The paper therefore divided the sample into pre- and post-1983 reforms and estimated the long-run demand for real cedi corresponding to the two sub-samples. This has enabled the paper to undertake an analysis of the effects of the 1983 reforms on the real demand for cedi, as well as on financial intermediation in Ghana (which had both nearly collapsed by the end of 1983). Using the General-to-Specific methodology, the parsimonious dynamic short-run models for the narrow and broad demand for real cedi are estimated. Battery of tests, including amongst others, tests for parameter stability, weak exogeneity tests of income, interest rate and exchange rate, the long-run price homogeneity tests, were conducted. Results show that there was indeed a stable, well-defined relationship between the cedi and a few variables that were identifiable as the real demand for money corresponding to the two sub-samples in Ghana. Secondly, there was a significant increase in the holding of cedi following the reforms, suggesting that the reforms have succeeded in reversing the currency collapse that appeared eminent in the early 1980s. Thirdly, the finding that the elasticity of the inverse velocity with respect to income is higher in the post-reform period (for both broad and narrow money) suggested that: (i) the rate of monetisation in the post-reform period was higher, hence the economic agents exhibited greater confidence in the banking system as they increasingly hold greater fraction of their incomes in bank balances; and (ii) given the higher income level achieved in the post-1983 reforms, the higher income elasticity of the velocity means that the banking sector, or financial intermediation has significantly increased following the reforms. The paper concludes that given the empirical evidence that a stable demand for both real M1 and M2 exists for Ghana, which implies that their respective velocities are predictable, these monetary aggregates can be useful in the conduct of monetary policy.