Abstract. Since the large number of parties involved in corporate failure or ‘business failure’, the
avoidance of failure has always been an important issue in the field of corporate finance and business
management. In this paper, the model was developed to predict business failure in Thailand particular
in technology industry by using four variables from Altman’s model and adding one variable to the
model. Descriptive statistics, correlation, and independent T-test are used for testing to see the
characteristics of each variable on both failed and non-failed companies. The model was developed by
using the stepwise logistic regression. Samples were developed by using financial information from
private limited companies based on technology industry in Bangkok. The result from this empirical
study can conclude that financial ratios are useful analytical techniques for forecasting financial health
of companies in technology industry. The result of independent T-test has pointed out sales to total
assets ratio is the only significant independent variable indicating significant differences between
failed and non-failed group. The Nagelkerke R2 indicated 42.4% of the variation in the outcome
variable. The predictability accuracy of the model is 77.8% which is under 95% confidence level
Research Framework
The financial status of a business firm can be produced information to creditors, investors, stockholders and
others for making decision. There are many researchers have been used financial information for predicting financial health of companies. The literature of Edward I. Altman developed the popular model called Z-score for a predictor of business failure. Therefore, the model would serve to reduce such losses by providing warning to these interested parties and that model predicts business failure to assess financial status as
early as possible. Hence, this study tends to focus on modified Altman’s model to develop the business failure prediction which is suitable for Bangkok based companies in technology industry.
Method Used
Descriptive statistics, correlation, and independent Ttest are used for testing to see the characteristics of each
variable both failed and non-failed companies. Stepwise logistic regression is the method used to develop the model which derived the variables from modified Altman’s model. Samples are developed by using financial information from private limited companies based on Bangkok in technology industry. Logistic regression produces all predictions, residuals, influence statistics, and goodness-of-fit tests using data at the individual case level. This empirical finding will provide warning signs to both the internal and external users of financial statements in planning, controlling, and decision-making. The warning signs and stepwise logistic regression model have the ability to assist management for predicting corporate problems early
enough to avoid financial difficulties. Moreover, the evidence from analysis of warning signs and the model can signal going concern problems early before eventually enters bankruptcy. Financial analysts could improve the development of this model since there are limitations associated that affected the ability of warning signs on financial status and predictive ability. For example, the failed firms selected in this study
were dissolved not bankruptcy. To improve the significance and predictive ability of the model, it is suggested to select failed firms as bankruptcy. The reason is that dissolved companies may need to stop doing business without financial problem while as bankruptcy companies have to stop doing business because of financial problem. Moreover, since financial information in this paper is limited to only income statement and balance sheet, this leads to the limited number of ratios in the model. One way to improve the model is to add
more ratios so that the predictability will be more accurate. It would be suggested that other statements such as statement of owner’s equity and statement of cash flow should be put into consideration. Lastly, besides the important financial ratios that measure the internal state of the firm, the existing macroeconomic conditions need to be included to help properly model the external environment of the firm. In the past, corporate failure prediction has been based on traditional methods of financial ratio analysis with multivariate discriminant analysis (MDA). This paper is limited to logistic regression. However, there are methods can be applied for further company failure prediction. For example, neural networks [K. D. Gunawardana, 2002] would
be applied to predict corporate failure. Artificial Neural Networks (ANN) are an attempt to imitate the human
reasoning and have been used in various applications of financial modeling that can be an effective forecasting alternative when compared to traditional techniques.