A Generalized Model for Bankruptcy Prediction of the Electricity Industrial Firms: Empirical Evidence for the Restructured Iranian Distribution Companies

In today's world, security of capital investment is one of the most important concerns forexisting economic enviro-nment. Insurance of productive capital investment and reducing economical riskcauses more fundraising and therefore the greater economic boom cycle. One way to arrive capital investmentsecurity is to predict bankruptcy of a business unit. Predi-cting possibility of a company’s bankruptcynot only can prevent losing the principle and capital interest of investing, but also faci-litate the mostimportant decision makings. Considering the imp-ortance of this subject, many researches have been donein this area. Since the Iranian electricity stock exchange is going to start working in 2012, it would beappropriate to propose a new model for bankruptcy perdition of Electrical and energy industries which willparticipate in the new stock exchange. This paper pre-sents a generalized model for bankruptcy predictionof an electr-ical business unit. Prevalent linear models are changed into a nonlinear model and all unknowncoefficients are determined through optimization process using a Learning Automata based algorithm. Thedeveloped model is conducted for bankruptcy prediction of electrical and energy industrial companies listedin Tehran Stock Exchange (TSE). In addition obtained results are compared to those of multiplediscriminant analysis (MDA), Logit, Altman model and results of the prevalent linear model. Moreover,linear Altman model is reformed into a nonlinear model relevant to the Iranian firms. Detailed numericalstudies and comparisons presented in the paper show that proposed model could improve noticeably thequality of prediction and can be used as an effective model for bankruptcy prediction of an electrical firm inthe new stock market.

کلیدواژه ها:

Bankruptcy Prediction ، Financial Ratios ، Lear-ning Automata ، Iranian Electrical and Energy Companies ، Nonli-near Model