A comparative study of bankruptcy prediction models: An application of Altman's Z-ccore, Ohlson's O-score, and Zmijewxki's probit model using Philippines firms

• Main Authors: Ng Uy, Danica Mae, Salazar, Hosea Lejlan L., Sy, Hosea Lejla, Zulueta, Mariel Anne
• Format: text
• Language: English
• Published: Animo Repository 2015
• Subjects: Bankruptcy > Philippines; Business failures; Finance and Financial Management
• Online Access: https://animorepository.dlsu.edu.ph/etd_bachelors/6287
• Institution: De La Salle University

Fundamentally the effect of bankruptcy from an economic standpoint is often quite large. The failure of a business to operate poses a grave threat to numerous stakeholders, depending on how large the company is. Signs of bankruptcy are often elusive, and would only appear upon the failure of the business. Prediction of this certain type of financial phenomena is very important in order to void investment on such companies that are deemed to be sinking ships. Therefore analysis of companies, which the investor might invest upon, is imperative in order to save losses. Fortunately there are many mathematical models to assess the risk of a company. However there are number of these tools and their applications vary to the data available to the investor. There is also the question of the accuracy of these models, which of the variables within the models most significant variables that affect the risk of bankruptcy, as well as the compatibility of its use to a certain economic environment. To address these concerns regarding bankruptcy, this study aims to assess the likelihood of the bankruptcy phenomena within the Philippine setting, the relevant mathematical models that measure the probability of bankruptcy, as well as the adjusting the models to the local environment to see if there is significance in doing so. This study used three models that predict the bankruptcy phenomena these are: Altman's Z-score, Ohlson's O-score, and Zmijewski's probit model. These were applied to two sample sizes, which are the: non-bankrupt, and bankrupt sample size. The same bankruptcy prediction models were then applied with logistic regression in order to calibrate its coefficients to solve the concern of localizing the model for the Philippine setting. The results show that indeed the calibrated models were in fact very effective accuracy wise in the prediction of bankruptcy with the calibrated Ohlson's O-score being the best predictor in both the calibrated and uncalibrated batch of models. The most significant variables for the prediction of bankruptcy are classified for each model, for the Altman Z-score the most significant were earnings before interest and taxes over total assets , and the book value of equity over total assets. The Ohlson's O-score's most significant variable is it's Size variable, which is characterized as the (Log (Total Assets/GNP price-level index)). Likewise, for Zmijewski's model it's the current assets over current liabilities variable. These ratios are the most significant in determining the probability of bankruptcy for the companies based solely upon their financial statements. However investors should take into consideration extraordinary factors as well as qualitative factors in making a sound investment decision as these findings are focused solely on the financial statements of the company.