Study On Integrated Risk Measurement With Market And Credit Risk For Defautable Bonds Portfolio

Defaultable assets, including defaulted bonds and credit derivatives, have higherincome, thus financial institutions hold as important positions. Domestic defaultableassets market continues to make breakthroughs, short-term financing bills introduced in2005is in true sense the first debenture in China. Domestic SME private debt (so calledjunk bonds) which is more market-oriented and obvious default risky, pushs defaultedbonds development to a higher level. However, because of being exposed to market riskand credit risk for these defaulted assets, financial institutions need to have a strongcapabilities of risk management.As risk management transfers focus from managing risk in single loan or singlebond to a portfolio perspective. Credit correlation between each asset in the portfolioshould be taken into account. Therefore, this paper study credit correlation betweeneach individual credit risk in a portfolio, and correlation between credit risk and marketrisk will be studied in details under a joint framework to integrated measure defaultedbonds portfolio. Little research about defaultable assest portfolio is made in domestic,so this paper hope to discuss in the following areas.To construct the joint correlation framework, two correlation is introduced tocapture risk. To credit correlation, details about related technologies is given. Forcorrelation between credit and market risk, introduce risk factor. Especially defaultintensity in simple model is taken as credit risk factor in this paper. The above twocorrelations are connected in affine process.Secondly, in the credit risk measurement of reduced-form model, we use theintensity pricing process, and assume that the two basic state process are independentone-dimensional CIR process under the actual probability measure, further shows creditrisk factor h (default intensity)and market risk factor r (risk-free rate) as affine form ofthe basic state variables, so that we can build the correlation between credit risk andmarket risk. To market risk and credit risk for portfolio, more work should be done.Intensity h should be further showed affine intensity process as system credit riskfactors and traits credit risk factors in linear representation, so that the individual creditcorrelations is taken into account. In order to obtain the measurement results,establishing defaulted bond credit risk market risk integration measurement framework,we use the Monte Carlo simulation technique to simulate the portfolio value of thedefaulted bonds, obtain portfolio loss distribution in special horizon and VaR value under a certain level of confidence.Then we respectively select a group of defaulted bonds for portfolio, Shibor interestrates and SSE Corporate Bond yield and SSE Government Bond yield as sample tomake parameter estimation and integration risk estimates. From parameter estimationresult, risk-free rate and default intensity are negatively correlated and have a trend. Thenegative correlation from the given empirical result is consistent with research ofdomestic and foreign scholars. Correlation coefficient of the default intensity withsubstract of SSE Corporate Bond yield and SSE Government Bond yield is possitive, sothe default intensity and credit spreads has positive relationship. It is consistent with theresult of U.S. economic cycle and credit spread fluctuations correlation which is givenby Jorion (2005).For integrated risk estimation, at first we just make comparison and analysis aboutdifferent VaR in three situation, that is, credit-market risk comprehensive measurecases, only considering credit risk, only considering the interest rate risk a. The resultsillustrating directly total add up two types of risk by traditional method may neglect thecomplicating effects among different risks, it will underestimate total risk. Directly totaladd up two types of risk by traditional method may have dispersion effects among risks,it will overestimate total risk; due to the complication effects, it will underestimate thetotal risk. Then portfolio risk integration measurement results are compared under notconsidering credit correlation and considering credit correlation from two aspects: valueat risk and occurrence numbers of extreme default events from simulation, illustratingthat without of consideration of the credit correlation, the probability of default isseriously underestimated, which also seriously affects risk estimation for defaultableassets portfolio. The conclusion we get is consistent with domestic and foreignresearch.