This article applies multivariate analysis to measure the exchange market pressures in 20 countries, in the context of the Asian crisis of 1997/98. We develop an index of pressure on the exchange market as a weighted average of exchange rate changes, reserve changes and interest rates changes. Changes in the exchange rate enter with positive weights, changes in reserves have negative weights and changes in interest rates have positive weights. These weights are derived from a measure of the communality of each variable used in the analysis. Using a exchange market pressure index we classify the sample, identifying by Ward's method of cluster analysis, the countries most affected by the crisis initiated in mid 1997. The classification is validated with the use of discriminant analysis.
exchange market pressure; Asian crisis; cluster analysis; discriminant analysis