Working Paper: NBER ID: w3320
Authors: Linda S. Goldberg
Abstract: This paper predicts ex-ante the probability of currency crises end size of expected devaluations month by month for Mexico between 1980 and 1986 using a heterodox linear discrete time model of exchange rate crises. The forces contributing to speculative attacks on the Mexican peso include internal money creation, external credit shocks, and relative price shocks. The framework proves highly successful for generating forecasts of the probability of speculative attacks on the peso and for predicting lower bounds for post- collapse exchange rates using a range of assumptions about critical levels of central bank reserve floors. Simulation results suggest that reducing domestic credit growth, increasing the uncertainty surrounding this growth, and reducing the size and perhaps increasing the frequency of currency realignments might have greatly reduced the amount of currency speculation against the peso in some of the crisis periods between 1980 and 1986.
Keywords: exchange rate crises; currency speculation; Mexico; monetary policy
JEL Codes: F31; F32
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| internal money creation (E51) | speculative attacks on the Mexican peso (F31) |
| external credit shocks (F65) | speculative attacks on the Mexican peso (F31) |
| relative price shocks (P22) | speculative attacks on the Mexican peso (F31) |
| excess domestic credit creation (E51) | depletion of foreign exchange reserves (F31) |
| depletion of foreign exchange reserves (F31) | speculative attacks on the Mexican peso (F31) |
| reduction of domestic credit growth (E51) | reduction in currency speculation against the peso (F31) |
| increased uncertainty around domestic credit growth (F65) | reduction in currency speculation against the peso (F31) |