Моделирование вероятности дефолта в строительном секторе: факторы корпоративного построения

• Main Author: Alexey Rybalka
• Format: Article
• Language: English
• Published: National Research University Higher School of Economics 2017-09-01
• Series: Корпоративные финансы
• Subjects: probability of default; corporate governance; logit model; regularization; construction industry; russia
• Online Access: https://cfjournal.hse.ru/article/view/7212/8126

In this paper, we have estimated the probability for default in large construction companies in Russia using the classic method for this purpose – logistic regression. Our task incorporates testing corporate governance factors and analyzing the predictive power of the model with regularization (Lasso and Ridge). For the dependent variable, we tested four definitions of default and then compared them. The model was conducted on the basis of information from the SPARK, Rosstat and the Bank of Russia database for the period of 2007-2015 – the final sample, after eliminating the outlying ob-servations, consists of 4761 construction companies. The added value of the corporate governance factors is verified on the basis of comparison of the ROC-curves (AUC) and the I and II errors. Seven hypotheses were formed, some of which were statistically significant, using an assessment based on both international and domestic experience dealing with the influence of corporate structure on the company's stability. In particular, everything else being equal, the default prob-ability of the company will be lower if the CEO is also a co-owner; whereas the default probability of the company will be higher if the company is a subsidiary. Note also, that, in fact, companies with a small board of directors overcome fi-nancial distress better (with a negative return on assets) in the Russian construction business. There was no confirmation of the hypothesis that older companies are less likely to default. Confirmed hypotheses give a new perspective to look at with a comprehensive risk assessment of large construction companies in the country. According to our estimates, corporate governance factors really improved the predictive ability of the models, and regularization methods confirmed the stability of these models. Using cross-validation, the robustness of the coefficients of the final specification was con-firmed. This result may be of interest to a greater extent for banks, commercial investors and partners-contractors.