Corporate Credit Rating Models for Industrial Companies in Emerging Market

The main purpose of this work was to predict corporate credit ratings using a different models (RF, DT,SMV,NN, nearest Nearest neighbor and ordered probit model) based of publicly available information for industrial companies in emerging market. The paper compares the predictive strength of models and tests the hypothesis of the importance of belonging to the industry or to the country. The hypothesis of the influence of belonging to the industry and the country is confirmed. Also identified key financial, market and macro-economic indicators which have the greatest influence on the rating. As the result, the best models became RF and DT, after to compare the variables was built probit ordered model. And it was found that country and macroeconomic indicators play an important role for all 3 models. In probit ordered model, GDP and CPI have a positive impact. However, inflation has a greater p-value in the DT and RF models, while GDP and CPI are more significant in the ordered model. From the ordered model we got: significant variables with a positive effect - GDP, CPI, marcap, income and EBITDA margin. Also significant variables, but with a negative effect, are ROE, debt / total assets, debt/EBITDA.