Is Long-Run Demand for Money Stable in India? - an Application of the Gregory-Hansen Model[dagger]
Singh, Prakash, Pandey, Manoj K., IUP Journal of Applied Economics
The paper is an attempt to analyze the behavior of money demand function in India using annual data for the period 1953-2008. Gregory and Hansen (1996) cointegration results show the presence of cointegration between demand for money, real GDP and nominal interest rate with structural break in the year 1965. The study also suggests downward shiftin the demand for money by about 0.33% around the year 1965. The analysis puts forward that demand for money is stable except for the period 1975-1998.
(ProQuest: ... denotes formulae omitted.)
The issue of long-run relationship between broad money and its determinants and the stability of the demand for money has always been in the center of conduct of the monetary policy and has gained momentum due to increased financial innovations, reforms in financial sector, shiftin exchange rate policy worldwide and increased financial integration. India has witnessed many ups and downs in macroeconomic policy framework in the past five decades, responding to uncertain political and economic climate. Invoked by this, determinants of demand for money function, e.g., GDP, interest rates, inflation, etc., have seen changes too. For instance, India experienced two consecutive wars with China and Pakistan and famine during 1960-1965 and political instability in the form of two general elections in three years (during 1977-1979).1 In the 1980s, financial deregulation took place, followed by Reserve Bank of India (RBI) moving to monetary targeting regime as recommended by the Chakravarty Committee in 1985. Further, during the 1990s, interest rate structure was leftto the invisible hands of market due to deregulation of deposit rates. The year 1986-87 is marked as the time when India had its first stroke of hesitant reform. Again in the early 1990s, amid the reserve crisis and pressures from external international agencies (IMF and World Bank), India was forced to devaluate its currency and move to a flexible exchange rate regime from the fixed one. In 1994, RBI got some muscle to exercise monetary policy independently in the form of its agreement with Government of India to bear fiscal burden only through issuance of 91 days ad-hoc Treasury bills. Again in 1998, RBI moved to multiple indicator approach for the conduct of monetary policy from a regime of monetary targeting approach (RBI, 1999).2 In the light of these policy framework changes, numerous quantitative research efforts have been made to analyze demand for money in India (see e.g., Vasudevan, 1977; Arif, 1996; Mohanty and Mitra, 1999; Bhanumurthy, 2000 and Ramachandran, 2004). However, theoretical and empirical approaches for these studies have been quite complex and multidirectional. The econometric results are found to be very sensitive to the functional specifications and methodologies used in estimating money demand function suggesting mix of both stable (see e.g., Pradhan and Subramanian, 1997; Ramachandran, 2004; and Rao and Singh, 2006) and unstable scenarios of demand for money in India (Bhanumurthy, 2000). Motivated by these mixed econometric results and increased capital inflow in the economy during the recent years, this study is an attempt to re-look at the issue using annual data for the period 1953-2008 by applying the Gregory and Hansen (1996) cointegration approach with structural break. The study uses Gregory and Hansen method of cointegration against the general practice of using Johansen (1991) cointegration method, because Johansen and other cointegration techniques assume time invariant cointegration coefficient among the variables. The assumption of time invariance of cointegrating coefficient could not estimate demand for money in the Indian context precisely due to the number of macroeconomic policy changes during the time of study. Whereas, Gregory and Hansen cointegration technique estimates long-run relation with time variant cointegrating coefficient. Additionally, it identifies three different sources of break (namely, break in intercept, break in trend and break in intercept and trend both) in the cointegrating relationship. …