The role of credibility in exchange rate-based stabilization: The case of Mexico

My dissertation argues that high credibility causes the demand boom during exchange rate based stabilization. I present a theoretical model with risk aversion where increased credibility of a currency peg lowers real interest rates, causing demand and real exchange rate appreciation. I suggest that a gradual increase in credibility over time is responsible for the stylized facts of most exchange rate based stabilizations.; I review the specific circumstances of the Mexican stabilization episode from 1988 to 1994, and claim that it is consistent with my view of how credibility shapes the macroeconomic dynamics during stabilization. I estimate a variety of models for the Mexican episode, using the Kalman filter algorithm to estimate FIML systems with an unobserved interest premium that reflects credibility. I argue that a measure of the perceived currency risk, based on capital flow and interest rate levels, reflects the credibility of the exchange rate policy. The Kalman filter allows me to measure this unobserved “credibility premium.”; In addition to extracting a measure of credibility, I test whether interest rates affect demand. I also explore whether credibility affects demand via channels other than the interest rate. There is little robust statistical evidence to prove that credibility has the hypothesized effects, but the pattern of inferred credibility and the results are largely consistent with the theory.; The other main purpose of the empirical analysis is to test which factors cause credibility movements. A measure of fiscal responsibility and a debt reduction agreement are two of the main candidates. In addition, I test whether credibility improves over time. The effects of fiscal policy and debt reduction are contingent on the particular model specification. I argue for a most plausible specification that supports the view that credibility improves over time. The time effect is statistically significant and economically large. This specification does not have significant causality from the fiscal and debt reduction variables.