Research On Credit Score Of Internet Finance Listed Companies Based On Super-efficiency DEA Model

Recently,China's financial industry and information technology industry have made continuous progress,and the Internet finance industry generated under this background has seen rapid development.After the continuous development of internet finance,six major business segments focusing on third-party payment,P2P online lending,crowdfunding,big data finance,information-based financial institutions,and internet financial portals have gradually emerged to improve the efficiency of financial services and meet the need for diversified investment.Financing needs,improving the inclusiveness of financial services and coverage have played an active role,but in recent years there have been a number of adverse events in Internet finance,such as Tang Xiaoseng's P2P platform illegally raising more than 80 billion road events,car crowdfunding platform "Jinfu Online" attracted more than 100 investors to raise funds in just a few days.The occurrence of these incidents will have a negative impact on the entire Internet financial industry and even affect the stability of the entire financial industry.Therefore,studying the credit risk of Internet finance has very important academic value and has realistic reference significance for investors.This paper explores an effective credit scoring model for the Internet finance industry based on a two-step empirical analysis.This paper selects 53 internet financial listed companies in the Internet financial concept stocks as the research object.Taking 159 sets of data from 2015 to 2017 as the research sample,based on high frequency indicators and related authoritative literatures,the sea selection index system is constructed,and then combined with the Internet financial industry.Its characteristics include adding R&D expenditures to the proportion of operating revenues and technicians and technical aspects,and finally forming a first-level criteria layer based on financial indicators and non-financial indicators.The financial indicators include profitability,solvency,and growth.There are four secondary criteria levels for competence and operational capabilities,and non-financial indicators include two secondary criteria layers,basic conditions and repayment willingness,with a total of 53 indicators.In the first step,the indicators are screened,and the index selection process is based on the constructed sea-selection index system.Since the indicator system of sea-selection may contain indicators with duplicate information,first the equation-based inflation factor method removes multiple correlations in each criterion layer.The indicators,which include a small number of repetitive indicators in the secondary criteria level of profitability,debt repayment ability and growth ability,are excluded.Secondly,according to the binary logistic model,based on the default sample and non-default model,the indicators are screened in each index layer,and the indicators that have significant contributions to the default samples are selected.Finally,16 indicators are selected.The second step is to score the credit of the Internet financial industry,and bring the selected indicators into the ultra-efficient DEA model to obtain the credit scores of 53 Internet finance industries.Through empirical research,this paper draws the following main conclusions:First,the Logistic regression model has good index screening capabilities.Based on publicly available information,this paper constructs a comprehensive and objective pool of initial credit evaluation system indicators.Based on the classified default companies and non-defaulting companies in the sample,the Logistic regression model is used for index selection,which can effectively screen out non-compliant enterprises and non-default enterprises.Default company's index.The initial indicator pool is effectively reduced,and the final selected indicator pool can be seen to cover multiple levels of information and improve the accuracy of the model.Second,the credit scoring model based on the super-efficient DEA model performs well.By evaluating the efficiency of the DEA model for 53 sample values,it can be seen that the final credit score of the defaulting companies is generally low.The credit scores of the five companies in the six defaulting companies are all below 100 points.The accuracy of the model prediction is 83.33%.Non-default Internet financial listed companies have relatively good ratings,and most companies have a credit score of 100 or more.Third,financial indicators in the credit rating index system of Internet finance listed companies account for the major proportion.The credit score system finally obtained in this paper consists of 16 indicators,including 12 financial indicators and 4 non-financial indicators.The financial indicators make a greater contribution to the Internet financial credit evaluation system.At the same time,based on the characteristics of Internet financial enterprises with high technical requirements and attention to asset status,in the construction of the indicator pool,based on the selection of high-frequency indicators,targeted additions to the total of R&D expenditures and the total amount of guarantees accounted for net assets.The related indicators,such as proportion,total R&D spending as a percentage of operating income,are still retained in the final selected indicator evaluation system,indicating that the technology and guarantee ratios when evaluating the Internet finance industry are indeed indicators worth considering.This article based on Internet financial listed company related credit data to score the Internet financial industry,not only can quantify the credit risk of the Internet financial industry,has a clear understanding of the industry's overall credit risk,but also can provide effective investment for market investors suggestions,as well as providing effective reference to the Internet finance industry regulators.