This study tests Krugman's model on exchange rate target zones using four old EMS exchange rates (those of the currencies of Belgium, France, Italy and the Netherlands against the Deutsche mark). Vast theoretical literature is available but there are very few empirical studies testing the main results of Krugman's model, namely the 'honeymoon effect', 'smooth pasting' and the overall S-curved relationship between fundamental and exchange rate. The present study supports the honeymoon effect but does not find any evidence in support of 'smooth pasting' or of an S-curved relationship between the exchange rate and fundamentals for the Belgian franc, French franc and Italian lira against the Deutsche mark. The Dutch guilder very closely follows movements in the German mark throughout the period 1981-90. End-of-month exchange rates and Euro-currency interest rates data are used.