Applying International Experience in Financial Risk Assessment of Small and Medium Business in Russia

The paper is aimed to build the prediction model of bankruptcy for companies operating in the agriculture sector in Russia for the period from 2011 to 2015. The research is conducted using the sample of companies from Small and Medium field (SME). Every selected financial coefficient is analyzed individually. Coefficients are chosen for a final sample using the CAP – curve analysis and Accuracy Ratio, Area Under Curve indexes. Conducted research has shown a high importance of three groups of indicators. These groups (liquidity ratios, turnover ratios and coefficients of financial stability) are key indicators to predict the probability of default. The research is also conducted using linear methods of prediction, including logit regression analysis, the Fisher’s functions and the canonical function ( Linear Discriminant Analysis ). Built logit regression has not shown expected results. Its predictive power is less than the arithmetic mean of Accuracy Ratios among included coefficients. At the same time, Linear Discriminant Analysis has shown more significant results. The predictive power of the Fisher’s functions and the canonical function is about 80 per cent. That means that built models can predict the probability of default in 80 cases of 100. One of the main conclusions is the identification of some coefficients’ importance. Both models using forward selection have identified the same groups of indicators - liquidity ratios, turnover ratios and coefficients of financial stability, what proves their importance. In general, built models can be used by commercial banks for the credit rating of enterprises.