Commercial Bank Loans Default Risk And Recovery Risk Impact Factor Empirical Research

How to measure the credit risk is one of the greatest challenges faced by Chinese banks. Therefore, it is very necessary and urgent for Chinese commercial banks to use international experience for reference, study and explore both theoretically and empirically the measurement and management methodolog.The paper analyzes the key composition elements of credit risk in terms of the basic definition of risk through literature review. It places emphases on the key composition elements of credit risk-default risk and recovery risk. It systemically combs, analyzes and sums up both domestic and international relative theory and empirical results about default risk and recovery risk. It is stated that the affecting factors of default risk include the firm property, firm characteristics, financial situation, industrial sectors , geographical characteristics , market factors, macroeconomic factors and other non-economic factors. The affecting factors of recovery risk include the firm property, firm characteristics, financial situation, contract characteristics, industrial characteristics , business relationship, macroeconomic factors and other non-economic factors.By the review of existing research, this paper confirms the research focal points of three respects: First, study the affecting factors of the default risk of the loan; second, study the sensitiveness of default risk model of the loan to the default definitions; third, study the influence factors of the recovery risk of the loan.According to the focal point, this paper constructs the affecting factor model of the default risk and the affecting factor model of the recovery risk and carries on theempirical research respectively.This paper selects 3742 loan enterprises of one commercial bank in 2004 as samples, separately carries on the empirical research on the affecting factors of the default risk according to the three default definitions. On the foundation of analysis of the sample characteristics, the assumptions of the research model of the default risk are tested through describing statistics, relevant analysis and Logistic regression through SPSS.The empirical result shows that, the ownership property, asset scale, financial situation, industrial attribute of loan enterprises have significant effects on probability at default, but the financial factor is still the most important affecting factors. The probability of default is positive relevant to cash assets rate which represents liquidity, retained earning rate which represents profitability and total assets turnover rate which represents service efficiency of assets. Because big enterprises have advantages in fund, technology, personnel's quality, status of market, financing, can disperse risk in pluralism, the greater the scale of the enterprise is, the lower the probability of default is. Opposite to the research conclusion of Evelyn Hayden (2003), in our country, because the governance structure advantage of limited company is obvious, PD of limited company is lower than that of unlimited company. The industrial attribute of loan enterprises has significant effects on PD, but because the phased characteristic of industrial structure and influence of the macro economic policy in our country, there exists bigger difference between its conclusion and foreign research results(Sjur Westgaard and Nic van der Wijst, 2001; Evelyn Hayden, 2003). The PD of real estate and service enterprises is higher than that of other enterprises. Under the three default definitions, a lot of variables selected by Logistic default models are the same, and the structure of default risk model is still similar, but some variables and their significant levels are different.Therefore the default risk models are sensitive to the default definitions.This paper selects 147 charged-off loan enterprises at default of one commercial bank between 2002 and 2005 as samples, Carries on the empirical research to the influence factors of the recovery rate with the whole 147 samples and the 83 sub-samples respectively. On the foundation of analyzing the sample characteristics, it carries on descriptive statistical analysis, relevant analysis, analysis of variance and linear regression analysis through SPSS, to test the assumptions of the research model of the recovery risk.The empirical conclusion shows that , opposite to the research conclusion of Jens Grunert and Martin Weber (2005), because of the advantage of the governance structure,clear-cut property right and capital, the recovery rate of limited company is higher than that of unlimited company; the recovery rate is positive relevant to registered capital; opposite to the research conclusion of J. Dermine and C.Neto De Carvalho (2005), the longer the enterprise age, the lower the recovery rate; the recovery rate is positive relevant to asset-liabilities rate; opposite to the research conclusion of Viral V. Acharya et al.(2004); Daniel Covitz and Song Han(2004), the recovery rate is negative relevant to the fixed asset-liabilities rate; the recovery rate is positive relevant to loan exposure at default, but its influence is small; the recovery rate is negative relevant to default time and work out time; the recovery rate is positive relevant to secured loan, credit relationship time and negative relevant to lending bank counts; the recovery rate is irrelevance to the credit level, asset scale, the industrial characteristics, profit margin in sale, this bank loan rate.The main contribution and innovation herein are listed: 1. Based on the bankrupt theory, corporation financial theory and default risk theory, the paper defines three kinds of default definitions with criteria for classification of four grades and five grades based on the credit risk classification methods of loan in Chinese commercial bank, includes the influence factors of enterprise property (Limited company or not), enterprise characteristics (Asset scale), 39 financial ratios (Leverage rate, liquidity rate, productivity, efficiency, turnover rate, profit rate), industrial characteristics (Public utilities, real estate industry, service), to construct Logistic default risk influence model. It has offered a new theoretical visual angle and more complete analysis frame for the fact that the Chinese commercial bank measures the default risk of the loan.2. Take the data enterprises loaned by a commercial bank as samples and carry on the empirical analysis of the influence factors of the default risk, it verifies that the default risk of loan is affected by enterprise property, enterprise characteristics, financial factors and industrial characteristics. The result enriches the influence factors of the default risk in the aspects of enterprise and industry. It has offered the theoretical and empirical basis for measuring the default risk of loan for Chinese commercial banks.Past reseaches studied less on the influence factors of PD about the aspect of industry and enterprise's ownership properties. While considering financial variable and variable of scale of the enterprise, the research introduces enterprise's property variable and industrial characteristic variable.Past default risk researches adopted more disclosed bond data, listed company data, the same proportion of default and non-default samples, less enterprise's samples of commercial bank loans; frequently defined default as "bankrupt"," ST of listed company". The research herein adopts loan enterprise's data of a commercial bank in our country, gives default definitions on criteria of credit risk classification methods of loans in Chinese commercial banks. The definitions are generally used in the management practice of commercial bank loans, nearly included all states at default loans, and concord with the ones stipulated by the new capital agreement of Basel. It has real meanings for Chinese commercial banks that will introduce the foreign default risk models and employ their own internal data and default standards to measure the default risk of the loan.3. Carry on empirical analysis on the effect of choice of default definitions on the structure and predictable ability of logisitic default risk models. The empirical results show that the final models are similar in their structure under the three default definitions, but there exists difference with respect to the selected variables and their significant levels. The default models seem to be sensitive to the definitions of default.Past studies on default risk paid little attention to the choice of default definitions. They frequently defined default as bankruptcy, ST of listed company. Other definitions included large bank overdraft, non-payment for preferred stock dividend, debt restructuring, overdue more than ninety days, etc. There are few studies (Evelyn Hayden, 2003) on the effect of choice of default definitions on the structure and predictable ability of default models. The empirical result of the paper that the default risk models are sensitive to the definitions of default further enriches the theory of default risk.4. Based on asset pricing theory, structural theory, industrial theory, this paper regards recovery rate of loan at default as dependent variable, takes the enterprise's properties(limited company), enterprise's characteristics (credit level, asset scale, registered capital, age of enterprise) , financial rates(Asset liability rate, tangible asset-liability rate, profit margin in sale), contract characteristics(exposure at default, default time, workout time, guarantee way), industrial characteristics(The tertiary industry or not), commercial relationship(The lending bank counts, credit relationship time, this bank loan rate) as independent variables. In terms of the 6 respects and 16 factors, the paper constructs the linear regression model of the influence factors of recovery risk. It has offered a new theory visual angle and more complete analysis frame for the fact that the Chinese commercial bank measured the recovery risk of the loan.5. Select the charged-off loan enterprises in one commercial bank as samples, try to measure the recovery rate of default loan with the Cash flow discounted method, carry on empirical analysis of the aspects of enterprise's property, enterprise's characteristics, financial rates, contract characteristics, industrial characteristics and commercial relationship. It verifies that the recovery rate is affected by enterprise's property, enterprise's characteristics, financial rates, contract characteristics and commercial relationship. The study further enriches the theoretical and empirical basis of the affecting factors of recovery risk.