Time-varying devaluation risk, interest rate differentials and exchange rates in target zones : empirical evidence from the EMS

Stylized empirical facts about the behaviour of exchange rates and interest rate differentials in real world target zone arrangements are at odds with the predictions of the simple (fully credible) target zone model. Incorporating time-varying devaluation risk in target zone models results in much richer data-generating structures and provides an interesting interpretation for the variability which standard target zone models leave unexplained. By using Bayesian time-varying parameter regression, the present paper shows that stochastic devaluation risk actually explains the EMS data quite well. Three key findings should be stressed: first, estimates of expected devaluation rates have recently declined significantly, but devaluation risks are not yet completely eliminated. Second, expected devaluation rates display "hysteresis". This contaminates with noise many of the relationships postulated by target zone models, but adjusting for expected devaluation rates frequently reveals almost noise-free relationships, which strongly supports the prediction from the theory. Finally, the estimates of expected devaluation rates suggest that some of the early EMS realignments were largely expected by the market.