This paper tackles the issue of hysteresis in the desired equilibrium exchange rate (DEER) arising from misalignment and changes in debt stocks. The DEER is a “steady state” concept, calculated as that (real) rate of exchange that will ensure external balance when the economy is operating at a full rate of utilization (internal balance). When the actual exchange rate is different--a so-called misalignment--the external balance realization will, in general, be different from that implicit in the DEER. This implies that the economy’s net foreign asset stock will also be different from that implicit in the DEER, as will debt-service obligations. But different debt-service obligations require a different DEER, which must be recomputed. The dependence of the desired equilibrium exchange rate on the actual rate of exchange indicates that it is subject to hysteresis.
The first task of the paper is to formalize the presence of hysteresis effects in the DEER. This process is straightforward and indicates that a misalignment has an effect of opposite sign on the DEER. That is, if the actual exchange rate is depreciated relative to the DEER, the DEER value appreciates. This result suggests that the stability properties of a system in which the actual rate may respond to the desired equilibrium exchange rate should be examined. Such an examination yields some quite intuitive results with respect to the adjustment speed.
The presence of hysteresis effects in the DEER is not in doubt, but particular interest attaches to their empirical significance. Three relevant sets of calculations are performed in this respect. First, the analytical formalization derived below is used to examine the possible significance of hysteresis effects for the Group of Five during 1975-90. For the United States, particularly, a prolonged misalignment over this period indicates a sizable hysteresis effect. Second, rules of thumb are derived for computing the amount by which the desired equilibrium exchange rate will shift for a given initial misalignment and the length of the period of adjustment envisaged for the actual rate to converge on the desired rate. Finally, a 1990 set of DEERs for the Group of Five is used to compute by how much the DEERs would need to be adjusted if convergence were to take place over five or ten years, with account being taken at the same time of the adjustment of utilization rates over the same horizons. The results broadly confirm the rules of thumb and suggest that the extent to which the desired rate shifts, relative to the initial amount of misalignment, is not negligible.