Central Bank's Value at Risk and financial crises: An application to the 2001 Argentine crisis

Abstract

Blejer and Schumacher (1999) were the first to suggest that Central Bank's Value at Risk (VaR), a widely used composite measure of potential portfolio losses in the corporate sector, could be used as an early warning indicator of financial crises. We extend their research in two aspects. First, we develop an operational model to calculate Central Bank's VaR and illustrate the methodology using data from the recent financial crisis in Argentina. Second, we compare the predictive performance of diverse measures based on the VaR approach to that of another well known early warning system, the signals approach, and several univariate leading indicators. The results reveal a strong relationship between the measures proposed and the crisis. Furthermore, one of the measures provides higher accuracy and announces the probability of a crisis sooner than the competing indicators.