Nous analysons, à l’aide d’un modèle théorique à deux pays, si la reprise précoce de l’économie américaine implique une forte hausse des taux d’intérêt en Europe, qui mettrait en péril la reprise européenne. Nous examinons les effets des politiques monétaires, du degré de mobilité internationale du capital, de l’intensité des effets de richesse, et enfin de l’accumulation de dette extérieure aux États-Unis.

This paper asks how EMU will affect foreign-exchange markets, the demands for various key currencies, the behaviour of exchanges rates among them, and policy coordination. The advent of the Ecu as the single currency of the EC may reduce the vehicle role of the dollar in foreign-exchange markets. Yet the share of the Ecu in global reserves may be lower initially than the sum of the shares of the present EC currencies. The private demand for Ecu, domestic and foreign, may also be smaller initially than the demands for the present EC currencies. By helping to unify EC capital markets, however, advent of the Ecu may raise the flow demand for Ecu-denominated claims, causing the Ecu to appreciate in the early years of EMU. Finally, EMU may not help Europe to speak with one voice in international negotiations and may thus interfere with policy coordination.

In this paper we compare the interregional and international degree of real exchange rate variability of a number of EMS-countries. This allows us to gain insights into the costs of a monetary union. Our major conclusion is that a monetary union among all EMS members would be costly for some of them if it were instituted today. We also find that a monetary union among a «hard-core» of EMS-countries (Benelux, Germany, Denmark) would not involve major costs.

Exchange rate theory has recently been concerned with versions of the uncovered interest rate parity (UIP) condition, as an alternative to purchasing power parity. The UIP condition is examined, for the U.S. Dollar, the Deutschemark, the Yen, and the Swiss Franc all against the Pound Sterling, using dynamic specification tests based on residual from the LUS and OLS classes, and selected error orthogonality tests. Market errors for one currency may have an information role for other currencies, this also is tested. It is suggested that spot rates are determined by the same underlying process which requires a considerable amount of past information on market errors for efficient spot rate determination. A role for time varying risk premia is suggested.