This paper is a theoretical and empirical investigation into the duration of exchange-rate pegs. The theoretical model considers a policy-maker who must trade off the economic costs of real exchange- rate misalignment against the political cost of realignment. The optimal time to spend on a peg is derived and factors that influence peg duration are identified. The predictions of the model are tested using logit analysis with a data set of exchange-rate pegs for sixteen Latin American countries and Jamaica during the 1957-1991 period. We find that the real exchange rate is a significant determinant of the likelihood of a devaluation. Structural variables, such as the openness of the economy and its geographical trade concentration, also significantly affect the likelihood of a devaluation. Finally, political events that change the political cost of realignment, such as regular and irregular executive transfers, are empirically important determinants of the likelihood of a devaluation.
