An Empirical Analysis of a Scandinavian Exchange Rate Model

Abstract

The Scandinavian inflation model (SIM) is augmented by a monetary sector in order to allow for the application of the SIM to floating currency regimes. The structural equations of the resulting generalized SIM are estimated using Swiss data. Two major results are obtained: (1) Systematic lags must be incorporated into the real sector equations of the orthodox SIM, even if estimations are performed on annual data. (2) Exchange rate expectations apparently render the empirical model unstable. The latter result may explain why Swiss monetary authorities felt compelled to terminate the free float of the Swiss franc in late 1978. I. Introduction Ever since currency prices were granted flexibility in early 1973, economists have been struggling hard to find a theoretically consistent and empirically workable explanation for what goes on in the foreign exchange markets. Few will deny that the subsequent erratic behavior of major currencies caught the majority of the profession by surprise. In general, goods arbitrage apparently takes considerably more time than is assumed in the bulk of theoretical studies. Also, and probably more importantly, exchange rate expectations seem to include an extrapolative component which is so strong that it allows for periods of sustained currency over- and undervaluation, thus putting considerable pressure on the exporting or import-competing sector of the domestic economy. It appears that early proponents of flexible exchange rates had scarcely anticipated problems such as these, which seem to indicate that the world is not always as frictionless and expections not quite as rational as mainstream economic reasoning postulates. In an attempt to cope with the challenge posed by unexpectedly volatile exchange rates, the profession soon set out to construct analytically tractable models which go beyond good old purchasingpower-parity doctrine, and attempt to model short-run disequilibrium exchange