State-owned Enterprises In Shenyang City Financial Distress Early Warning Research

Financial distress is a world problem. Since 1960s, more and more researchers try to predict bankruptcy through quantitative analysis. In the recent 50 years, many models such as Multivariate Discriminate Analysis and Neural Network come out. However, in our country, financial distress prediction just begins. The main reasons are as follows: one reason is the lack of uniform Accountant Rule before July 1, 1993, the other reason is the difficult of having the financial dates of unlisted companies. However, unlisted companies are many times bigger than listed companies in our country, it is urgent for us to work on financial distress prediction under status quo.As one of old industry bases of P.R. China, there are many state-owned enterprises in Shenyang City. Now, insolvency and deficit is a critical phenomenon in state-owned enterprises of Shenyang City. From 2003, the state-owned enterprises of Shenyang City use a united and normative accounting reporting soft which make the study of financial distress prediction a solid foundation. So, if we can found a model with the dates we have and predict the financial distress with the model beforehand, it is significant not only for State-owned Assets Supervision and Administration Commission of Shenyang and banks, but also for employee of these enterprises.This study regards that deficit surpass depreciation as a signal of financial distress, and tries to predict financial failure of state-owned enterprises of Shenyang City. At first, we pose hypotheses before we begin the study, select enterprise of financial distress and non-financial distress and classify them to sample part and test part. We compare the financial rates between the enterprise of financial distress and non-financial distress and use Logistic Regression and BP Neural Network to found models of financial distress. We also predict the financial distress of test part with the models that we just found and compare accurate rates. We find that current asset turnover, debt ratio, revenue growth rate, profit margin before taxes and investing gains, return before taxes on assets et al. can predict financial distress accurately. We also find that BP Neural Network model overwhelms Logistic Regression model in prediction accuracy in sample part and has a accuracy of 95% one year before financial distress. But BP Neural Network model have the similarity accuracy in test part with Logistic Regression model.