Theoretical And Empirical Study Of Credit Risk In Commercial Banks Based On Herding

Herd behaviors in financial market have been studied for several decades in foreign,and there are also some studies about herd behavior in banks.Herd behaviors of China securities market also have been studied for several years,but there are not theoretical study and empirical analysis about herding in China banks yet.Because there are still no literate reviews of reputational herding,informational herding,and running herding,literates about these three kinds of herding are summarized in my paper.Herd behaviors in our commercial banks are classified as follow:reputational herding, informational herding,and running herding.Theoretical study,empirical analysis and risk prevention against these three kinds of herding in our commercial banks are discussed respectively.A model of reputational herding in commercial banks is set up basis on the model of Scharfstein and Stein(1990).Under certain circumstance,when concerning about their reputation,managers of commercial banks simply mimic the investment of others,ignoring private information.But different from Scharfstein and Stein(1990),this paper assume the public reliefλ∈[0,1]and the informative signal correlationρ∈[0,1],these make my model more closed to realistic commercial banks.And how initial reputation,public relief and ability affect the follower's incentive to herd is analyzed.By statistical analysis of the loans for real estate and the real estate load accounts for ratio of total load,it is manifested that there are reputational herdings in commercial banks of China.In the discussion of informational herding,a model of two managers is set up.Because of the costliness to obtain information,managers of commercial banks often give up mining information,but simply mimic the investment of others.To interpret the formation of informational herding,an exogenous model of multiple managers is set up patterned after the model of Banerjee(1992).Then an endogenous model is set up patterned after Teraji(2003). The formation of "spurious herding" and "intentional herding" is interpreted.Different from Teraji(2003),which assumed all participants are homogenous,in this model commercial banks are classified into two categories:banks of large-scale,and banks of middle and small scales.The actions of these two types are more complicated.By statistical analysis of loads to the first six industries and its proportion to total loan and the biggest customer ratio of different commercial banks,it is manifested that there are also informational herding in commercial banks of China.Patterned after the models of Bilthchandani,Hirshleifer and Welch(1992),Diamond and Dybvig(1983),Allen and Gale(1998),Yorulmazer(2003),a model of bank run is set up. Since the private signals about the bank are noisy,information revealed by the actions of depositors have value for other depositors.A few numbers of withdrawals can lead depositors to withdraw.Different from the above models,it's assumed that the public reliefα∈[0,1] and the informative signal correlationρ∈[0,1]in this paper,these make my model more closed to realistic commercial banks.By an empirical example of bank run,it's showed that bank run can be prevented by increasing information quality.By statistical analysis of deposit and lending ratio,medium and long-term load ratio of different commercial banks,it is manifested that there are also runing herding in commercial banks of China.The models of herding in commercial banks are the most important innovations in this paper.