Abstract
This paper estimates the long-run demand for money in Romania using monthly data from January 1994 to August 2003. The Johansen–Juselius cointegration procedure provides evidence of one cointegrating vector for both narrowly and broadly defined money demand functions. Real money balances are not only sensitive to real income, but also the domestic interest rate and the depreciation of the domestic currency.
Notes
In the transition to a market-oriented economy in 1989, Romanian authorities adopted a policy of gradual transition with regard to financial markets. The National Bank established a fully free foreign exchange market in March 1997 when it introduced full convertibility of current account transactions. This time also coincides with the removal of state control over a wide range of prices. The impact of this change on financial markets did not occur as rapidly as it was expected. The domestic currency remained overvalued during 1997 and 1998 due in part to the large foreign capital inflows.
Payne (Citation2003) estimated money demand for Croatia and found that the interest rate, inflation and exchange rates have a significant negative effect on money demand. Buch (Citation2001) found evidence that only inflation has a negative impact on money demand in Poland and Hungary while changes in the exchange rate appear to have an insignificant influence.
Preliminary money demand specification incorporating inflation as an opportunity cost variable yielded a positive coefficient on the inflation variable and corresponding instability in the money demand function.
For previous applications of this procedure to money demand see Hafer and Jansen (Citation1991), Mohammadi and Smith (Citation1993) and references therein.
The X11 procedure was developed by the US Bureau of Census.
The reports may be found on the National Bank of Romania's website, www.nbr.ro.
In light of the possibility that the exchange rate may have been controlled prior to 1998, real foreign deposits were used to account for currency substitution instead of the exchange rate. The results yielded a negative and significant coefficient on the real foreign deposits variable in the normalized long-run estimates of money demand. These results are available from the authors upon request.