The Study Of Asset Securitizations’ Credit Risk Economic Capital Based On Correlation Of LGD And PD

Since the 1970s, asset-backed securitizations developed very quickly, and they were important to lower financing costs, spread and transfer risk, abundant financial market investment products. However, asset securitizations’structure is rather complex, and therefore its risk is difficult to measure and manage. The root of the subprime crisis in 2007 is to securitization and re-securitization of subprime mortgages, and the outbreak of the subprime crisis let us re-examine the reasonableness of asset securitizations’risk and economic capital provision. Since 2011 Chinese asset securitization faced explosive development. In 2012, more than 200 million asset securitizations were issued and the number was soared to 300 billion in 2014. Asset securitizations are indispensable financing tools to China’s financial market but bitter historical experience tells us asset securitization is a double-edged sword that when we use them we must take into account its risks.For financial institutions, it is essential to keep enough economic capital to deal with significant uncertainty and keep themselves from bankruptcy. The literature of asset securitizations’economic capital is not very much. Based on a single risk factor and Vasicek model, Pykthin and Dev et al build a progressive model, which is analytic method of measurement of economic capital asset securitizations. However the approach assuming LGD (Loss Given Default) is a constant, many scholars verified this hypothesis inconsistent with the actual. So we use stochastic LGD to model, specifically considering LGD and default rates are affected by the same systemic factors. By Monte Carlo simulation we calculated the economic capital of asset securitization layers the case of constant and stochastic LGD. By comparing the result, we find that under constant LGD economic capital is seriously underestimate, especially in the mezzanine seeuritization products. We also find that on the hypothesis of constant LGD, the economic capital of re-securitizations is underestimated more severely.This paper consists of five parts:The first part is the introduction, mainly introducing the main document of this writing background, significance and major domestic and foreign research portfolio credit risk economic capital and economic capital of the securitization. Then we lay the framework for analysis theoretical basis.The second part focuses on the basics of asset securitization. Asset-backed securitization has a unique cash flow structure-low-level hierarchy to provide security for high-level hierarchy. Asset-backed Securitization products have unique features:bankruptcy isolation and credit enhancement, which make the asset-backed securitization product various financial products with different risk and return, but which is also the root of its risks. This section also analyzes the risks faced by asset securitization, including credit risk, market risk, repayment risk. However, credit risk is the main risk to asset-backed securitization.The third part analyzes main methods of VaR and ES model and the application of modern credit risk and economic capital calculation method. The two methods, VaR and ES model, have different advantages. VaR is more widely used by financial institutes, but which has a fatal flaw:not satisfied subadditivity. Gordy proposed asymptotic single risk factor model in 2003, whose hypothesis are assets pooling being fully dispersed and driven by only single risk factor. Based on such hypothesis, he have proved unexpected loss of assets pooling is equal to the combined system in sufficient factor at q quantile loss expectations. This model greatly simplifies the calculation of economic capital for the theoretical foundation. Under this assumption, we can not only prove that VaR meet subadditivity, also get that the economic capital of the whole asset pool equals to the sum of each part’s economic capital and each part’s economic capital equals the size of its own risk.In Part Ⅳ, on the basis of a single risk factor model of progressive, we can deduce the asset securitization model of economic capital. We assume that LGD have certain correlation with systemic risk factors, on the basis of random LGD and the calculation method of asset securitization economic capital.The fifth part is simulation analysis. According to Moody’s Investor Report, we first evaluate the parameters involved in the model estimated with the maximum likelihood estimation method. After obtaining the model parameters, we do the Monte Carlo simulation with 500,000 times to calculate the economic capital of securitization and re-securitization on the condition of constant LGD and statistic LGD.The sixth part is the conclusion of this article. Through this article we can get that constant LGD underestimate economic capital securities, particularly mezzanine economic capital, and this effect is magnified in calculating the re-securitizations’ economic capital. Through this study, we know that a random LGD obviously more reasonable, which is essential for financial institutions to identify and manage critical asset securitization risk. Further, for asset-backed securitized products, there is a higher model risk because of their complex structure. Any small model setting incorrectly may pose a risk of multiplied, so we have to be more cautious in dealing with re-securitization products.