The development of bankruptcy prediction models in modern Russian economy

The paper focuses on problems of bankruptcy prediction models in Russia. This issue gains more and more relevance in recent years, in the light of falling incomes of population. The bankruptcy prediction models are widely used to predict bankruptcy. However, they might have low accuracy. In that regard, there is an active discussion in the foreign literature. In Russia there is a growing number of models developed, however so far no detailed studies to assess their efficiency. There are several objectives of the current study: a detailed review of domestic literature to reveal the flaws of existing models; design of a new set of models accounting to analyzed flaws; suggestions for future research. The estimation of Russian models revealed their inefficiency, due to data instability problem, poor financial reports quality, insufficient sample size, and distortion effect of “black accounting” and criminal bankruptcy practices. The set of models offered was proved to be efficient and robust. As the way of further model improvement it was suggested to distinguish between possible scenarios of failure (including liquidation, merger and “freezing”) using statistical classification methods. The Benfords Law analysis was introduced to separate companies falsifying financial statements. The application showed this method as promising.