MODELING INTERNATIONAL LOAN PORTFOLIOS THROUGH LINEAR-PROGRAMMING DISCRIMINANT-ANALYSIS

This paper presents an application of a recently developed decision-mo deling technique known as Linear Programming Discriminant Analysis (LP DA) to the formulation of a commercial bank's international lending po licy. This study uses data from fifty-eight countries who borrowed in international financial markets during the period 1983 through 1987. T hese data were used to demonstrate the decision-making capability of t he LPDA technique as an early warning signal for sovereign debt-repaym ent interruptions. The results of this LPDA study compare favorably wi th those of other decision-modeling techniques that have been used to support international-loan portfolio management.