Financial Early Warning Of Listed Companies Difficulties Panel Logit Model

Listed companies which trapped into financial distress can not only be athreat to the companies’ development and survival, but also inflict huge losseson creditors and investors. In order to reduce the economic losses of creditorsand investors and promote the healthy development of the listed company, it isnecessary to build up a financial early-warning system to reflect all risks of thelisted companies which they may come across, and to predict enterprisefinancial situation effectively and prevent the financial distress. It has theimportant meaning in improving quality of operating management andprotecting the interests of stakeholders and promoting the benign developmentof the capital market in China.Domestic and foreign scholars have conducted lots of academic researchwork about financial early warning of listed companies. But previous studiesmostly used a single phase of the cross-section data to predict financial distress,it belongs to static prediction research. The financial condition of enterprises hasthe feature of continuity and cumulative effects, companies in financial distressis a process of gradually evolved, and corporate balance sheets temporarydeviation from the normal should not be classified as trouble. In order torespond financial situation of listed enterprises as much as possible, in this paper,considered two dimensions of financial and non-financial for early warningindex selection, build Panel Logit model which can treat Panel data effectivelyto research the financial distress of listed companies in our country. This methodmake up for the deficiency that the traditional cross-section data research cannotdynamic response the evolution of the enterprise financial situation, andobjectively reflect the company’s financial situation of the development ofdynamic facts accurately, and make the early warning model having morepractical significance.Taking manufacturing enterprises in Shanghai and Shenzhen A shares asthe research object in this paper. Firstly, with the proportion of1:2, this paperhas selected30companies which are first announced into ST from2011to2013 as samples of troubled companies and60companies which have never been inST as the samples of normal companies,28financial indicators and10non-financial indicators preliminary were selected for early warning indicators.In order to have a better warning effect, this paper not only selected the28financial indicators; Secondly, this paper analyses the primary38indexes of thefinancial data by entropy method, and according to the different amount ofinformation that influence of the financial early-warning model provided byeach of the primary indicators, eventually chose21financial indicators and8non-financial indicators; Thirdly, factor analysis is carried out on the29variableindex, and eliminate the effect of multicollinearity of variable indicators forestimate in early warning model, extract7financial public factors and3non-financial public factors that has important influence to the enterprisefinancial situation, and7financial public factors as explanatory variables of thePanel Logit financial early-warning model based on financial indicators, and the7financial public factors and3non-financial public factors as explanatoryvariables of the Panel Logit of financial early-warning model that combinedfinancial indicators and non-financial indicators; Fourthly, the paper builds therandom effects regression model with Panel Logit by Housman test. Theempirical results of the Panel Logit financial early-warning model based onfinancial indicators show that repayment factor, profit factor, capital utilizationfactor, capital structure factor are the important factors influencing companies infinancial distress. The empirical results of the Panel Logit of financialearly-warning model that combined financial indicators and non-financialindicators show that repayment factor, profit factor, capital utilization factor,capital structure factor, ownership concentration factor, evaluation factor are theimportant factors influencing companies in financial distress; Finally, this paperestablished the financial early-warning model made outside samples’ inspection,and the correct rate of predicted results are86.67%and86.67%,and the result ofPanel Logit financial early-warning model based on financial indicatorscontrasts to the result of Panel Logit of financial early-warning model thatcombined financial indicators and non-financial indicators, the results show that the non-financial indicators in the financial early-warning model is conducive toimprove the prediction ability of model.