Differences in Financial Distress Prediction Models for Small and Medium-Sized Enterprises

Abstract

     Financial problems are one of the biggest issues affecting the survival of small and medium-sized enterprises (SMEs). Consequently, providing a warning before a company fails should be an effective method to help the survival of SMEs. There are many models that are used as early warning tools, and each model performs differently. Therefore, the primary aim of this article was to compare the principles of financial distress prediction models. The methods studied consisted of: Logit, Probit, Multivariate Discriminant Analysis (MDA) and Artificial Neural Network (ANN) models. In addition, the strengths and weaknesses including the nature of prediction of each method were summarized. The forecasting efficiency of these methods was compared by reference to relevant research studies. It was found that the Logit and Probit models are flexible in application and they are also easy to understand and explain. For more complex research studies, which require more complex techniques to identify several multivariate groups, the appropriate tool is MDA. For even more complicated research requiring more sophisticated techniques or nonlinear equations, ANN modeling is the most effective tool. The variables contributing the highest opportunity to identify financial distress were also identified.