Modelamiento del capital por riesgo operativo en entidades financieras de mercados emergentes Entre 1990 y 2013 Usando Cópulas Multivariadas

Abstract

This thesis seeks to provide a more efficient estimation of capital requirement for operational risk in financial institutions across emergent markets by applying the LDA (Loss Distribution Approach) proposed in Basel II Capital Accord (2004) and modelling multivariate dependence between severities using Gaussian and t copulas. Three models were built to draw conclusions. Model (i) estimates capital requirements for operational risk using the standard model. Model (ii) incorporates into the LDA the multivariate analysis of dependences between operational losses (severities) using the Gaussian multivariate symmetric elliptical copula. Model (iii) incorporates the multivariate symmetric elliptical t copula into the standard model LDA. This research analyses an updated operational loss data set, SAS OpRisk Global Data, to model operational risk at international financial institutions from emerging markets between 1990 and 2013. The impact of models (ii) and (iii) was evaluated on the estimates of the total regulatory capital for operational risk and compare the resulting estimates with the one predicted by the standard model LDA. The results confirm the existence of diversification benefit up to 33% with Gaussian copula (model ii) and up to 34% with t copula (model iii). These results are consistent with previous researches and evidence that the incorporation of the multivariate Gaussian and t copulas provide a more efficient quantification of capital requirement for operational risk.