Most `target zone' models assume perfectly flexible goods prices and predict that an exchange rate usually lies near the edges of the band and domestic interest rates are high (low) when it is near the strong (weak) edge. These predictions are inconsistent with observed data on actual target zone systems, but two extensions to allow random realignments and stabilizing intervention within the band enjoy some empirical support. In Discussion Paper No. 698, Research Fellow Alan Sutherland investigates the importance of the flex-price assumption by developing a target zone model in which goods prices adjust slowly to bring demand into line with capacity output. When random shocks are large, this sticky-price model also predicts that the exchange rate will usually be near the edges of the band, but when shocks are small, the nominal exchange rate will more often lie in the centre, which is consistent with the data. The flex-price model produces a similar prediction once allowance is made for intra-marginal intervention, however, so the two models are observationally equivalent.

Sutherland stresses that the major difference between the two models' structures lies in their determination of goods prices, so they are much more likely to differ in their predictions of the real exchange rate and interest rates. The assumption of continuous purchasing power parity in the flex-price model implies that these are both invariant, while price inertia implies that changes in nominal rates entail changes in the corresponding real rates, so the sticky-price model has testable implications that can be derived from the distribution of fundamentals. Both distributions are symmetric, with a single mode at the centre, and they decline towards zero for extreme values, with no significant difference between the overshooting and undershooting cases. Sutherland concludes that there are major differences with which to distinguish the models empirically: the sticky-price model explains the behaviour of nominal variables no better than the basic flex-price model, but it is likely to improve upon its empirical performance for real variables.

Target Zone Models with Price Inertia: Some Testable Implications
Alan Sutherland

Discussion Paper No. 698, August 1992 (IM)