Financial Distress Prediction Variables & Industry Effects Analysis
Hong, Ye (2013) Financial Distress Prediction Variables & Industry Effects Analysis. [Dissertation (University of Nottingham only)] (Unpublished)
In light of the speedy development in the economics market, corporate bankruptcy problems have become more and more serious. In this reason, I began this research to get a set of my model, which can predict the probability of bankrupt at least three year prior the occurrence of the event. I study a sample of 100 delisted US companies from 2002 to 2006, for the investigation of the time interval and industry effects use regression models. I find evidence that the logistic regression model is more useful than the ordinary least square (OLS) regression model in the non-linear relationship problems. Secondly, working capital divided by total assets (WCTA) is the least contributor to discrimination between the financial distress group and the health group. On the contrary, the size of the company (LNCAP), solvency (CR) and working performance (EBITTA) are the most contributors to the financial distress prediction model. Thirdly, the model with five-year horizon backward from two-year prior bankruptcy event performs better than the model with five-year horizon as well, but from three-year prior bankruptcy event. Finally, the group sectors do explain the determinants of financial distress. The researches implemented in the industry effects illustrate that period of 2002 to 2004 with the highly proportion of bankruptcy and industry of industrials and consumer service have the most serious problems of financial distress. Both suspects have been confirmed by the out of sample test, which is used to assess the efficiency of the prediction model.
Actions (Archive Staff Only)