An Empirical Study Of The Pricing Of Chinese Collateralized Debt Obligations Based On M-Copula

In recent years,CDO is one of the fastest growing product in financial markets.It has the function of transferring and dispersing the credit risk and can meet the demand of different risk preference of investors.So it is very popular in financial institutions,especially in commercial banks.In the United States subprime mortgage crisis in 2008,the risk effect of CDO was amplified,one of the reasons was its improper pricing.CDO of China started relatively late.The amount and type of CDO are not many in the current financial market.It has strong theoretical significance to study in-depth discussion on Chinese CDO pricing referring to foreign existing research results.And it also has very important practical significance for the healthy development of Chinese CDO and the economy.In this article,we build the pricing model of CDO based on M-Copula,with the default correlation as a starting point.We do empirical analysis according to the first credit asset backed securities issued by the ICBC in 2014.We carry out the following problems in this article:(1)Building the pricing model of CDO based on M-Copula function.We derive the expression of reasonable credit spread under the no-arbitrage principle that continuous breach of contract equal to the premium paid expenses.Then introducing M-Copula function to the pricing problem of CDO and presenting theory about measuring the correlation of asset's default time on M-Copula function.(2)According to KMV model,computing the probability of default on the assets using financial data and stock data.Then calculating the probability of default of assets and solving the default time marginal distribution using the reduced model.(3)Constructing M-Copula model and estimating its parameters.This process requires three steps.Firstly,we solve the marginal distribution of asset returns using estimation of kernel density function that the non-parametric estimation method commonly used.Secondly,we estimate the parameters of a single Copula function with a two-stage maximum likelihood estimation method.And finally,we estimate the parameters in M-Copula function with help of fitting experience Copula function and the minimum distance method.(4)We describe the step of default time simulation in detail by Monte Carlo.(5)We calculate the pricing of CDOs combined with the bond and stock market data.we judge the effectiveness of the model constructed through empirical analysis.In the empirical analysis,we analyze the pricing model of CDO and verify that the M-Copula of CDO pricing model is effective.For different coupons of CDO,different Copula functions have different pricing results.Comparing to asset-backed securities of priority B,the pricing result of asset-backed securities of priority A is closer to the actuaral coupon rate.We do a sensitivity analysis of the recovery of default for each voucher layer in CDO and find that asset-backed securities of priorityB are more sensitive to changes in the recovery of default.When the recovery rate changes from 0.4 into 0.6,each the pricing result calculating by different Copula functions are closer to the coupon rate.In particular,for asset-backed securities of high-yield profile,when the default recovery rate changes from 0.6 to 0.8,its credit spread becomes zero.After analyzing the results,we give related recommendations for the development of Chinese CDO combining with the development of Chinese CDO market.