The argument that any exchange rate regimes other than firmly fixed and freely floating rates were infeasible — the so-called bipolarity thesis — acquired great popularity in the wake of the Asian crisis of a decade ago, but it has almost vanished today. One reason is surely the unkind empirical evidence, which shows that intermediate regimes — measured as those where both reserve and exchange rate changes lie in an intermediate range — are not in fact tending to disappear (Levy Yeyati and Sturzenegger, 2002). Another reason is the recognition that exchange rate policy should have other objectives besides avoiding crises, and that in the world we live in today it is reasonable to give these other objectives a significant priority. And perhaps a third factor is growing recognition that it is possible to design or operate intermediate regimes in ways that avoid exposing them to the dangers that were focused on by the disciples of bipolarity. This article starts by distinguishing the options that countries face in choosing an exchange rate regime. It examines the advantages and disadvantages of each of them, finally suggesting that for most countries the real choice lies between freely floating rates, floating rates disciplined by a reference rate system, and an ill-defined managed floating with the management undefined. Three issues may influence the choice between those alternatives: transparency; perceived consistency with that pillar of current macroeconomic thinking, inflation targeting; and the theory of what determines exchange rates. In the latter context, it is argued that the current conventional wisdom of the economics profession is wrong, and that a more convincing diagnosis of the process of exchange rate determination lends support to the proposal for a reference rate system.
