engleski

Firm failure prediction: prediction accuracy improvements based on firm size groupings

Firm failure has been an interesting research topic for a long period and many researches try to explain this phenomenon, often trying to improve failure prediction accuracy. In order to explore the issue of firm failure, researchers use different statistical methods and samples. Findings generally indicated that financial ratios are useful for failure prediction. Many early papers dealt with the samples of large listed firms, but later papers put focus on SMEs. Comparison of studies for large firms and SMEs reveal that models are not identical regarding financial ratios and prediction accuracy. Therefore, in order to explore the issue of possible differences in failure profile of different sized firms we developed separated models, not only for the two groups of firms commonly investigated (large or SMEs) but for four different size firm groups. Empirical research was conducted for failed firms from Croatia, while the data was obtained from Amadeus database. Failure models were evaluated with the usage of logistic regression and findings confirmed that failure prediction accuracy for size based subsamples was higher than for all firms sample.

firm failure prediction, firm size groups

nije evidentirano