Research On Credit Rating System Of Rail Transit Company Based On Discrete Choice Model

With the rapid development of urbanization in China,the urban rail transit,such as subway and light rail,has become more and more important in public transport.With the rapid development of rail transportation industry,city rail transportation companies emerge rapidly,and rail transportation projects will attract social capital inflows in a large scale,the related loans,bonds,and non standardized assets of related financial products will also emerge with big volume,which will promote the demand for credit rating.At present,the establishment and use of rating model has become one of the most common methods for credit risk management and investment decision making.But the majority of financial institutions face problems like lacking independence,simplicity of rating methods,lack of risk disclosure,cannot be completed on the technical requirements of the higher rating business and professional pertinence.At first,in view of characteristics of rail transportation industry in China,this paper constructs a complete and flexible index system,which is consisted of four major classes of scale factors,factors of profitability,asset-liability structure factors and potential of growth factors.By using a classification system with three levels,we find measurable indicators for each class of factors.The index system has high coverage.And the classification structure of the index system with three levels makes it possible for the rating agencies to flexibly adjust the reference factors according to the specific situation,and to build the models with specific needs.Secondly,used the principal component analysis to descend dimension for the multidimensional indexes system of four major classes and extract the principle component of each major class,then we build a bridge between the ever extracted principal components and credit rating of target companys by the using the model with discrete dependent variables.The study offered a new idea for credit rating of companys in subdivision industry.Thirdly,regard the current 24 rail transit companys as sample,we applied the principal component analysis to descend dimension for the multidimensional indexes system,then constructed a discrete choice model based on the first principle component of each major class of multidimensional indexes and credit rating of 24 rail transit companys.Several tests had been tried by adjust the number of independent variables,we find a three-independent-variables regression model get the best goodness of fit.Fourthly,we applied the three-independent-variable regression model to evaluate credit rating of a test sample,we find that the calculation result is consistent with its real credit rating.Several conclusions have been presented as follows: scale factors,profitability factors,asset-liability structure factors and potential of growth factors could all affect the credit rating of a rail transit company;simultaneously,the estimation of the discrete choice model with a independent variable vector including scale factors,profitability factors,asset-liability structure factors shows that scale factors andasset-liability structure factors have greater influence on the credit rating of the companys,while the profitability factor's affection on credit rating of rail transit companys is relatively weak.The evaluation results of the other regression test models showed that the above three-independent-variable got the best goodness of fit and potential of growth factors could nothave an impact on the credit rating of a rail transit company.The innovation of this paper is reflected in the creation of a rating model of rail transit companys which avoided the weakness of original expert scoring models,our credit rating system and model could also easily be applied to the analysis of credit rating on other industries.With the further development of rail transportation industry,researchers can use new samples to enrich and improve the credit rating index system,and the multiple choice model could be applied for credit rating,to establish a more functional credit rating system.