Purchasing Power Parity and Exchange Rate Overshooting: An Econometric Investigation

Abstract

<p>After the abandoning of fixed exchange rates by most major countries in 1973, it was observed that exchange rates fluctuated disproportionally to relative prices. "The asset market approach to exchange rate determination" to a large country framework is the subject of this dissertation. The effect of an increase of nominal money supply on the exchange rate is dampened in a two-country world and overshooting is not a necessary feature of the model. The model has the advantage of examining the effect of real shocks, like an increase in government spending. It can also handle shocks due to changes in foreign variables (nominal or real) which are equally important in a large country setting. The assumption that both domestic and foreign variables have the same coefficient but opposite signs, which has been used extensively in the theoretical and empirical literature, is relaxed since its imposition could even reverse the signs of the constraint coefficients. Even within a two-country model, the long-run proportionality between money, prices and exchange rates, that is, purchasing power parity, holds. This is derived as a long-run zero degree homogeneity restriction in money and prices which can be tested. The impact period results, however, are different from those derived in a small country world. The sample consists of 42 quarterly observations (1973I-1983II) and only the large OECD countries, France, Germany, Japan, the UK and the US are considered. The notion of effective exchange rate is applied. The foreign variables are constructed by aggregating over the major trade partners of each country using trade weights. It is shown that the reduced form system is quite successful in tracking the actual data and that the model is indeed stable. There is no indication of overshooting except for France, and the restriction implying purchasing power parity cannot be statistically accepted. Finally, the predictive performance of the purchasing power parity is evaluated.</p>