The Efficiency of Bankruptcy Forecast Models in the Hungarian SME Sector

The paper examines the efficiency of bankruptcy forecast models in the Hungarian SME sector. We also try to construct own models using discriminant-analysis, logistical regression's, and neural network methods, based on a random sample, what we try to validate on a second sample. It has been proved that our own model can only be applied on the first sample with an outstanding result. It has also been proved that complicated statistical solutions themselves are not always applicable; there is a need for the expertise of an experienced economist. The research, of course, does not say that the bankruptcy-forecast methods used in literature have lost their trustworthiness completely. It just draws the attention to the fact that the economic circumstances of Hungarian SMEs can't be compared to that of big foreign companies. Therefore, the results of the indexes developed to the large enterprise sector cannot help accurate decision making in case of SMEs. Huge narrowing of the complexity of economic characteristics may lead to false results.