Estimating Dependence Structure and Risk of Financial Market Crash

Abstract

Dependence structure of nancial market is crucial in determining investment po- sitions and strategies to reduce nancial market risk. Linear correlation model is not suitable to capture asymmetries and dependence structure of nancial market as the only capture the degree of correlation. In order to address the problem, the study estimates dependence structure between nancial markets using the copula concept. Di erent relationships that exist in normal and extreme periods were estimated using copula. The Inference Functions for Margins method was used in estimating copula parameter thereby obtaining dependence estimates. The study show analytically how dependence estimates are imputed into Value-at-Risk. The Inference Function for Margin estimator was found to be consistent and asymp- totically normal. From the empirical ndings, to diversify market risk during the crisis period (2007-2009) the market pairs with the highest maximum possible loss is evident in the stock market and followed by the stock market-Tbills pairs. However, the less risk portfolio is the stock-bond. As bonds are nancially con- sidered as safer investments over time, implies that investment in a stock-bond portfolio is less risky during the crisis period.