The Monte Carlo Algorithm Optimization And Application Of CDO Credit Risk Assessment

Extensive growth gradually turned to intensive economic development in China, and the past extensive lending mechanism is bound to affect the non-performing loans of financial institutions. To reduce the pressure on financial institutions management, rich financial products market, CDO products welcomed the vigorous development of the open policy in the new period, which not only provide diversified business and investment varieties for Banks and financial markets, but bring into problems about CDOs credit pricing. So the research of CDOs credit risk quantification is of great practical significance.This paper mainly base on general Credit Metrics and KMV credit evaluation model and use the Monte Carlo simulation method to measure the CDO portfolio credit risk. Considering the structure of the portfolio correlation, we introduce the latent variables to construct copulas model and carry on the MC simulation. According to the low efficiency and high error of the simple Monte Carlo in measuring rare event probability, we introduce the important sampling Monte Carlo method based on cross entropy in order to improve the simulation efficiency and precision. Experiment prove that cross entropy method not only solves the key problem that it is difficult to determine unbiased operator for sampling Monte Carlo, but also greatly improves the accuracy of the Monte Carlo simulation. At the end we can get a more accuracy result about the credit risk of CDOs.