What Is a Loss Given Default?

Loss default default (LGD, loss given default). The default loss rate refers to the proportion of asset losses that would be caused to creditors by a debtor, that is, the severity of losses. The LGD is also an important parameter in the international banking supervision system.

Loss of Default

LGD means that the debtor will

The two basic elements that constitute a complete risk concept are the probability of loss and the scale of the loss once it has occurred, that is, the severity of the loss. Therefore, LGD is another important parameter that reflects the credit risk level in addition to PD PD. The combination of the two can fully reflect the credit risk level. Obviously, in the case of PD, the higher the LGD, the greater the credit risk.

LGD at the bank

Since the size of LGD is not only affected by the factors of the borrowing company, but also closely related to the specific design of the loan project, the factors affecting LGD are more and more complex than those affecting PD. Specifically, the factors affecting LGD include the following four main aspects:

In view of the limitations of the historical data average method, people began to study more methods to estimate LGD more accurately. These methods mainly include the following three categories:

1.Regression analysis of historical data

This method is based on default

Study on Loss of Default in Foreign Countries and Hong Kong and Taiwan

The main channels for enterprises to borrow funds to obtain funds are direct financing and indirect financing. Each corporate bond that is directly financed has a secondary market price. After a default, the default loss rate can be estimated based on the market price at a certain point in time after the default of the debt instrument. For indirect financing, it is necessary to rely on the default loan data accumulated by the bank to estimate the default loss rate. Public market information is easier to obtain, so research on LGD has also been developed on this basis.

The article "on the Pricing of Corporate Debt: the Risk Structure of Interest Rates" published by Robert C. Merton in 1974 is the theoretical basis of modern credit default probability and recovery analysis. Its shortcoming is that it does not solve the actual observation of credit asset quality, and its application in empirical research is limited. This is also the focus of a lot of subsequent work after the model was born.

In view of the difficulties of Merton's (1974) model in the field of empirical application, several literatures have attempted to provide flexible solutions. Crouhy and Galai (1997) expressed the Merton (1974) model, which cannot be directly observed, as a function of credit default probability and recovery rate, thereby simplifying the core of credit risk management to the observation and analysis of PD and LGD, which had a greater impact.

There are three main ways to observe and measure the LGD of financial instruments (Liu Hongfeng, Yang Xiaoguang, 2003): Market LGD (based on the market price of defaulted bonds or tradable loans after the actual default event); Workout LGD (cleared LGD, liquidation) And the ratio of the present value of a series of cash flow estimates generated by the recovery process to the risk exposure); Implied Market LGD (the market implied LGD, using the asset valuation model and calculated by the spread and price of similar non-default bonds). In fact, there are many empirical studies based on the secondary bond market or the secondary loan market (such as securitized personal home mortgage loans), while there are few empirical studies on ordinary bank loans. The first reason is the complexity of the research method. Second, the non-public nature of the data.

1. Research on the US market

Due to the availability of data, the current literature focuses on the US market.

Asarnow and Edwards (1995) use all economic losses incurred after a default event to measure the expected loss of bank loans. Based on Citibank's 831 default samples of general industrial and commercial loans and Structured loans between 1970 and 1993, the calculated LIEDs were 34.79% and 12.75%, respectively. An important finding of the research is that its distribution is a "bi-model", with samples concentrated at the high and low ends.

Carty and Lieberman (1996) used Moody's to give priority to guaranteeing 58 defaulting bank loans from 1989 to 1996, and conducted an empirical study based on their secondary market transaction prices. The results show that the average recovery rate is 71% and the median is 77% The standard deviation is 32%. The study did not observe a "bi-model", but found that the recovery rate deviated significantly from the high end.

Hamilton and Carty (1999) used the market method to calculate the repayment rate of 159 bankruptcy cases. The average repayment rate was 56.7%, the median repayment rate was 56%, and the standard deviation was 29.3%.

Gupton, Daniel Gates, and Carty used a sample of 121 defaulted loans in 2000. The results showed that the average value of preferentially guaranteed and preferentially unsecured bank loans defaulted was 69.5% and 52.1%, respectively. The deviation is also significant.

Gupton and Stein (2002) introduced for the first time a market value forecast based on the LGD forecasting model LossCalc, a multi-factor statistical model of LGD for US bonds, bank loans and preferred stocks.

Til Schuermann (2004) introduced the recovery rate distribution of all bonds and loans of Moody's 1970-2003, and explained the reason for the formation of the bimodal distribution.

Michel A., M. Jocobs Jr., P. Varshey (2004) used JP Morgan Chase s loan loss historical data from 1982 to 1999 (a total of 3,761 defaulting customers) to study LGD. The average accounting LGD and economic LGD are respectively 27.0% and 39.8%. The study also analyzed mortgage LGD. Through a study of a total of 1705 samples from the first quarter of 1982 to the fourth quarter of 1999, the average LGD of mortgage loans (1279 samples) was 27.7%, the standard deviation was 35.3%, the average LGD of unsecured loans was 40.3%, and the standard deviation was 42.5% The study released the mean and standard deviation of LGD for different types of collateral.

2. Empirical research on LGD in other markets

Citi's Hurt and Felsovalyi (1998) research on 1,149 bank loans in 27 countries in Latin America from 1970 to 1996 showed that the average default recovery rate was 68.2%, LGD showed a skewed distribution, and the macroeconomic and loan amounts were the recovery rate. One of the influencing factors, the larger the amount, the lower the recovery rate;

La Porta et al. (2003) studied the PD and LGD of related loans in Mexico. The average recovery rate of unlinked loans in 1995-1999 was 46%, while related loans were 27%. The distribution shows that LGD deviates from the high end.

Taiwan's Xu Zhongmin (2004) conducted LGD empirical research on the information of bank loan corporate defaults of Taiwan Lianzheng Center from 1996 to 2002. The annual operating income is 5 million euros, which is smaller than this standard (16,454 small companies). The average LGD is 75% and the median is 88%, which is larger than the standard LGD (84 samples). The average LGD is 84% and the median is 92%.

Standard & Poor's Franks et al. (2004) used about 8000 original data from the United Kingdom, France, and Germany for research. Data show that the recovery rate in the UK is significantly higher than in France and slightly higher than in Germany. The distribution of recovery rate in France is obviously "double model distribution", and the distribution in England and Germany is skewed.

Grunert and Weber (2005) studied the LGD data of 120 German companies from 1992 to 2003. The data shows that the average recovery rate is 72.45%, and the variance is 35.46%. The distribution of the recovery rate deviates significantly from the high end. The report also studies the impact of macroeconomics, industries, loan conditions and tax policies.

The above research reports only published intensive data after intensive processing. The original data, model parameters, etc. were not published, and no LGD research report on specific mortgage loans was found.

Domestic data on LGD

Due to the underdeveloped domestic corporate bond market, the research time for a bank default loan recovery data system has not been studied for a long time, and there are many domestic research theories on the default loss rate, and influential empirical data are scarce. There are:

1. Relevant data of the four major asset management companies. The asset recovery data released by the four major asset management companies, including China Huarong, can be used as indirect data to study the domestic loan default loss rate. In 2004, the asset disposal results of the four major financial asset management companies in China were an asset recovery rate of 26.60% and a cash recovery rate of 20.16%.

2. Other research. Zhang Haining (2004) An empirical study using 191 credit projects of large Chinese commercial banks as a sample (as of 1998) (involving a loan principal of 26.629 billion yuan and interest of 7.708 billion yuan) shows that the average recovery rate is 33%, the largest The value is 80% and the minimum value is 0.

On May 28, 2004, CCB conducted an auction of non-performing real estate assets with a book value of RMB 4 billion in mortgages through international bidding. Citibank, Deutsche Bank, Lehman Brothers, JP Morgan Chase, Morgan Stanley and other 15 institutions Participating in the bidding, the comprehensive fund recovery rate of the final bid was 34.75%.