Comment on Lars Svensson's Assessing Target Zone Credibility: Mean Reversion and Devaluation Expectations in the ERM, 1979-1992