A Research On Early-Warning Financial Crisis Of Manufacturing Listed Companies In China

With the rapid development of China's economy and securities markets, the number and the scale of listed companies increased gradually. The growing international and domestic markets bring unlimited business opportunities, but it also brings a significant financial risk. If the enterprises can't take effective measures when facing financial crisis, the failure of capital allocation and management confusion will be inevitable, or even insolvency. How to predict and find the crisis early and achieve the rapid and healthy development is not only the focus of listed companies, but also the urgent needs of stakeholders. Therefore, building an effective model for early warning of financial crisis is of great practical significance to listed companies and its stakeholders. The purpose of this paper is to analyze the early warning indicators about China's manufacturing companies, extract the financial indicators which are able to distinguish financial crisis and non-financial crisis, add non-financial indicators into early warning indicators system, build financial distress models, and ultimately provide a reference of the views of stakeholders and decision-making basis.This paper selects 77 crisis companies that were all first rendered special treatment from 2006 to 2009 and the paired 231 healthy companies as samples, and they are divided into training samples and testing samples. Besides, this paper selects 25 financial indicators and 12 non-financial indicators. After the significance test and principal component extraction, the ultimate indicators are selected as the initial model variables. Then two T-3 early warning Logistic models, before and after the introduction of non-financial indicators, are constructed. The result shows that consolidated forecast accuracy increased by 3.90% and 4.68% for training samples and testing samples after the introduction of non-financial indicators, which confirms that the introduction of non-financial indicators of early warning models have a better warning effect than the simple financial indicators models. On this basis, the paper constructs the hybrid financial crisis early warning models, and the result shows that compared with Logistic models consolidated forecast accuracy of the hybrid models increased by 2.35%, which confirms that the hybrid models have higher accuracy and stability than the single early warning models. The characteristic of this paper is that non-financial factors are introduced into early warning models, and data of T-3 years rather than T-2 years or T-1 years are used, and finally this paper constructs hybrid models to improve early warning capacity and accuracy.