Study On Financial Crisis Forewarning For Listed Companies Of Manufacturing Based On Cash Flow

The pillar industry of national economy, manufacturing industry, is facing with an unprecedented pressure as a result of the slow domestic economic growing and world economic recovery. More and more enterprises are in danger of financial crisis or collapse because of low economic efficiencies and high debt ratios. The financial crisis can be identified at the early stage and the collapse of enterprise can be avoided by implementing effective options. This could be done based on the research of financial crisis forewarning. The research results can be provided as basis of decision making for investors, creditors and stakeholders. Therefore, the research of financial crisis forewarning is very valuable for theory development and practice.This thesis reviewed the relevant literature all over the world. The financial forewarning theory of the enterprise that based in cash flow was analyzed. This thesis also compared the strengths, weaknesses and applicability of different financial forewarning models. We developed the index selection principle for financial forewarning index system of the manufacturing enterprise’s cash flow. 17 cash flow indexes were selected using this principle from the perspective of obtaining ability, debt paying ability, profit quality, and elastic development capacity of the financial affairs. These selected indexes were used to build a financial forewarning index system of the manufacturing enterprise’s cash flow. We selected all 61 ST manufacturing industries that were treated specially because of financial problems for the first time from 2011 to 2014 and their paired companies as samples. We collected the annual financial report of the recent three years and semi-annual financial report of last year. Significant test and factor analysis has been done to select 10 key indexes and model variables. Financial forewarning model of the manufacturing enterprise’s cash flow was built using these indexes and the BP neural network. To this end, the BP neural network was optimized using adaboost algorithm to build a BP-Adaboost combining forewarning model.The forecasting accuracy of BP-Adaboost combination forewarning model is higher than the BP neural network model from the perspective of empirical analysis. This system has better forecasting ability. The accuracies of financial crisis for training samples and examine samples are 91.46% and 82.50% three years ahead the crisis. Two years ahead the crisis, the accuracies are 95.12% and 92.50%. One year ahead the crisis, the accuracies are 93.90% and 90.00%. Half year ahead the crisis, these accuracies are 91.46% and 80.00%. These results indicated that the cash flow indexes that are selected in this thesis are scientific. The BP-Adaboost combining forewarning method based on adaboost algorithm is applicable and effective.Finally, according to the financial forewarning indexes system based on cash flow and results of BP-Adaboost combination forewarning model, to analyze the causes of the financial crisis of representative companies, and combined with the main risks of the manufacturing, proposed countermeasures to avoid financial crisis.