Exchange Rate Bands
Analysing credibility

After the demise of the Bretton Woods system of `fixed but adjustable' exchange rates in the early 1970s, there followed a period of free floating characterized by a far greater degree of exchange rate variability than had generally been anticipated. The exchange rate mechanism of the European Monetary System established in 1979 was intended in part to avoid this problem, by committing member states to take suitable action to prevent cross-parities from staying outside some given, publicly announced range; and John Williamson's proposal for `target zones' was explicitly aimed at preventing exchange rate misalignments among the principal industrial economies.

In Discussion Paper No. 421, Programme Director Marcus Miller and Research Fellow Paul Weller extend Krugman's formal analysis of the effects of such policy commitments on the exchange rate to allow for imperfect price flexibility, departures from purchasing power parity and output dynamics. They use a stochastic version of the Dornbusch overshooting model.

Miller and Weller examine the effects of imposing a real currency band, or `target zone'. In the absence of a target zone, the real exchange rate follows a path that reverts to the mean, but it can wander far from equilibrium in response to shocks to fundamentals. When a fully credible target zone is imposed, the range of variation of the real exchange rate is contained; but the possibility of stochastically <MI>unstable<D> paths for the exchange rate is introduced. This instability is not inevitable, as some have argued, but depends upon parameter values in the model.

A target zone is defended by `small' interventions designed to limit the range of variation of real balances. In contrast, however, a nominal currency band must be defended by discrete interventions to reduce the interest rate differential to zero. There will be finite periods of time during which the currency is defended at the limits of its permitted range of variation.

Since, in practice, the commitment to defend a nominal currency band will never be fully credible, it is important to provide a framework of analysis for `realignments' of the band. Miller and Weller consider two alternative models: in the first, realignments are expected to occur only when the exchange rate is being held at the edge of the band by suitable intervention; in the second, realignments occur with certainty if the rate ever hits the edge of the band, and market participants realize this. The two cases imply very different behaviour for the exchange rate.

Miller and Weller contrast their results with those of Krugman. They conclude that the solution to the Krugman model, as currently interpreted, is indeterminate and that their own model provides a means of resolving this indeterminacy.

Exchange Rate Bands with Price Inertia
Marcus Miller and Paul Weller

Discussion Paper No. 421, June 1990 (IM)