Since the collapse of the Bretton Woods system in the early 1970s, the choice of the exchange rate regime has been the subject of a lively debate in international finance. In this study, we investigate the determinants of three exchange rate regimes (fixed, flexible and intermediate). Our contribution to the literature is threefold: (i) we examine the unordered choice between the three regimes, with a multinomial logit model and cross-section data from 1982 to 1999. Previous studies are generally based on binomial and/or ordered models and use a short data span; (ii) we cover three different currency zones, the US dollar, the ECU/EMU, and the CFA Franc. The previous literature confines the analysis mainly to the US dollar zone; (iii) we compare two streams of the literature, the optimal currency area (OCA) and the currency crises (CC) models, and examine when each model explains the choice of exchange rate regimes better. Our approach provides a rich framework for the analysis of exchange regimes. Results with the OCA model show that the probability of choosing the flexible or intermediate regimes increases with a rise in fundamental variables (economy size, exchange rate volatility, capital mobility, inflation, low openness). Results with the CC model, however, suggest that policy variables and financial factors also affect the choice of the exchange regime. Adverse shocks to foreign factors (deteriorating current account, rising foreign liabilities, falling reserves) increase the probability of choosing the flexible or intermediate exchange regime over the fixed regime. Furthermore, we find that there is no one-size-fits-all model for all regions. The OCA model performs better in Europe and the Latin America, and the CC model in East Asia, Pacific regions and the CFA area. This result is consistent with the general perception that the currency crises that hit the international markets since the 1980s were fundamentals-driven in Europe and Latin America but financial-driven in Asia.
