Determining a bankruptcy prediction model: Empirical evidence from the Philippine manufacturing industry from the years 1991-2010

• Main Authors: Demdam, Abigael S., Francisco, Kimberlyn N., Llanto, Brendon T., Tan, Alyssa Sheena P.
• Format: text
• Language: English
• Published: Animo Repository 2011
• Subjects: Bankruptcy > Forecasting > Mathematical models; Manufacturing industries > Philippines
• Online Access: https://animorepository.dlsu.edu.ph/etd_bachelors/14695
• Institution: De La Salle University

Acknowledging the fact that bankruptcy has been an issue which concerns most, if not all business organizations as well as other stakeholders, this study sought to address the bankruptcy concern through the development of a mathematical equation that could foresee possible indications of a company’s susceptibility to bankruptcy. The equation would take into account three variables, as adapted from the study of William H. Beaver, namely return on assets (ROA), cash flow to total liabilities ratio (ETL) and total liabilities to total assets ratio (LTA). Furthermore, the study focuses on Philippine companies from the Manufacturing Industry as this industry accounts for approximately 66% of its whole totality. Data relative to the variables in the developed bankruptcy prediction model were taken from the financial statements of the study’s sample companies. These financial statements were retrieved from the Philippine Securities and Exchange Commission (SEC) through its website application called i-Report. From the retrieved data, the variables ROA, ETL and LTA for all the years prior the bankruptcy are computed for each sample which would then be summarized into panel data. As a result, descriptive statistics of each variable showing the trend will be used to validate expected signs. The data is then subjected to Logistics Regression which would result to the coefficients for each variable. Then, the computed coefficients from the logistic regression model are run using the odds ratio formula to transform the independent variables, resulting to the final bankruptcy prediction model. In addition, in order to validate the overall model, an out of sample validation methodology was done by dividing the total 233 firms into two groups through random sampling. Each group was run using the logistic regression model and odds ratio formula. The coefficients obtained from the group logistic regression model equation shall be used in multiplying the ratios of the companies from the other group in determining the bankruptcy probabilities of companies. Hence, the research study shall comprise a total of three (3) bankruptcy prediction models consisting of the overall model. Group A model and Group B model, in which all utilized the data from years 1991-2010. Comparison of these bankruptcy prediction models were made to validate whether the overall model developed transcends time and is useful in any economic condition. The study is seen significant as the developed bankruptcy prediction model, coined at the PSL (Profitability-Solvency-Leverage) model could aid in decreasing future shutdowns, increase stakeholders’ awareness and improve future decision making.