In this paper we test the 'news' model for the peseta-German mark and peseta-US dollar cases, using monthly data for the period January 1986-June 1996. We consider 'news' on the money supply, the level of activity, the nominal interest rate, the inflation rate, the trade balance and the public sector budget balance, for Spain, Germany and the United States, and from four alternative methods: the ARIMA methodology, the Kalman filter, the Hodrick-Prescott filter and the Bean 'surprise' model. The results show a partial effect from 'news' with respect to these variables on the fluctuations of the exchange rates under study.

Key words: exchange rates, news, expectations.
JEL classification: F30, F31, G15.