Influence of Financial and Non-financial Factors on the Probability of Bankruptcy: Evidence from Russian Construction Industry

Recently in the construction sector of Russia there has been a significant reduction in the number of operating companies, which reinforces the interest in modeling early warning of financial insolvency. The purpose of this study is to identify the most significant financial and non-financial factors in order to predict the likelihood of bankruptcy of Russian medium and small non-public construction companies. In addition, the task is to establish the existence of the business cycle of construction companies. The results of the study showed that financial and non-financial factors are significant in predicting the bankruptcies of Russian non-public construction companies, among them the greatest contribution was made by profitability indicators, turnover ratios, debt load, number of employees, age of the company. It has been empirically determined that there is a business cycle of construction companies. It was also proved that for medium and small construction companies a higher value of the average number of employees leads to a greater likelihood of bankruptcy, which makes it possible to interpret this factor as an indicator of labor inefficiency. The described results were obtained by building statistical models ( Logit ) and machine learning models ( CART , SVM , RF ). In accordance with the prediction accuracy indicators, the advantage of decision tree models for predicting the probability of bankruptcy of Russian construction companies was proved. The results of this study can be used by all groups of economic agents interested in predicting the financial instability of a firm. For owners, the developed models will allow evaluating the efficiency and sustainability of an enterprise, for lenders — the risk of bad debts, for the government — to assess the impact of a particular reform of the construction sector on the financial stability of companies.