Testing an Augmented Fisher Hypothesis for Money Market Interest Rates in Finland

In this paper we augment the famous Fisher hypothesis by introducing foreign interest rate and exchange rate variables to a tradional Fisherian test equation for the Finnish money market interest rates. Theoretically this augmentation is based on the use of uncovered interest rate parity for nominal rates and certain assumptions concerning the formation of exchange rate expectations. The empirical data covers monthly observations from the beginning of 1987 until the time of Finnish ERM-connection in October 1996. For the purposes of this analysis we have previously generated a time series of adaptive inflation expectations in Finland utilising a univariate time series model and a special structural change procedure in Junttila (1997). The use of cointegration techniques enables to include this generated regressor to the estimation equation, because the possible error-in-variable problem connected to the use of generated regressor in the empirical analysis asymptotically vanishes when unit root methods are applied. In this work the primary method in the cointegration analysis is the Johansen procedure, which is based on maximum likelihood estimation. Furthermore, we also apply Kalman filtering to the recursive estimation of error-correction models for Finnish interest rates. The foreign interest rates and exchange rates are chosen in view of obtaining some empirical results to be used as a background for the vivid discussion in Finland at the moment concerning the possible joining to the European Monetary Union (EMU) at its first phase in 1999. The foreign interest rates corresponding to the analysed domestic money market rates are from Germany, the United Kingdom and Sweden. The effects of each of them were analysed separately in a context, where also the world, i.e. U.S. market effects were examined. The main empirical result from this study is that the decision to float the Finnish markka in September 1992 might have caused remarkable structural changes in the economic relationships. However, applying recursive estimation to error-correction models indicated that at least for the monetary variables analysed here the structural instabilities were more connected to the short-run dynamics than the long-run (stationary) economic relationships. We were able to identify the Fisherian one-to-one positive relationship between the domestic nominal interest rate and domestic inflation expectations in all the analysed settings involving different foreign variables, but at varying maturities in different settings. We also found that the Finnish money market is very much dependent on the development both in the German and Swedish markets, but also the role of British market would seem to be important, because at least both at shortest and longest maturities the Fisherian one-to-one long-run relationship between domestic interest rates and inflation expectations with corresponding expectation horizons was accepted when the effects of British and world markets were controlled for