Portfolio VaR using coulas

Abstract

With the increased focus on Risk management there is a need for advancement in the techniques to model profits and losses accurately. A non availability of an easily implementable procedure to estimate VaR for a portfolio of assets( in our work, equities) taking into account the individual marginal distributions and using various copulas to model the same has prompted our research. In our effort we have also calibrated the model for ideal window size and through various other sensitivity analysis we have arrived at an optimum VaR estimation methodology. We have conceptualized, implemented and setup a tool to model an n-equity portfolio VaR using Historical Simulation (HS) and studied the impact of window size, marginal distributions, copula used and in a broad sense the advantage of using a copula over the traditional VaR methodology.