Font Size: a A A

A Study On Bond Rating Method Applied Of China Corporation

Posted on:2013-07-28Degree:MasterType:Thesis
Country:ChinaCandidate:Y B TianFull Text:PDF
GTID:2249330395982357Subject:Financial engineering
Abstract/Summary:PDF Full Text Request
The corporate bond market is an important part of the bond market, the development and improvement of corporate bond market is directly related to operating efficiency of the bond market and even capital market. Therefore, it is fundamental for the healthy development of capital market to improve China corporate bond market. China bond market, especially the corporate bond market, However, has been lagging far behind in the stock market. Except for the regime factors, corporate bond rating method backward is the main reason that China corporate bond market is lagging behind.Based on the researches about bond rating at home and abroad, the paper chooses four types methods including MDA, Logistic Model, Probit model and neural network, and according to the risk features of China list corporations, such methods are optimized by the angle of the variable selection. Beside the part of the widely recognized financial indicators, another four indicators such as the corporate controlling attribution, Tobin q, beta and the ratio of EBIT to current liabilities are selected, and simultaneously, the data of China list corporation is used to conduct an empirical test for the application capabilities of these rating methods. The conclusions show that the rating variables selected in this paper could capture the risk features of China list corporations better than the rating variables which were often chosen in the literatures abroad, and all of Logistic Model, Probit model and neural network have much more capability of rating classification to the bonds of China list corporations. To the training sample, the three rating models could correctly classified more than95%of bonds, especially the Probit model classified correctly all bonds in the training sample, and to the holdout sample, such three models predicted93%of bonds correctly. Consider the training sample and the holdout sample together, Probit model and neural network method are very precise, the error classification rates of which are almost0.
Keywords/Search Tags:bond rating, MDA, Logistic, Probit, neural network, variableselection
PDF Full Text Request
Related items