Analyzing Industry‐level Vulnerability by Predicting Financial Bankruptcy

Abstract

This study introduces a novel framework for building company bankruptcy models and a methodology for assessing the vulnerability of industrial economic activities. We consider the identification of bankruptcy as a classification problem and assume that bankruptcy criteria differ across industries. We build highly accurate industry bankruptcy models by constructing separate models for each industry. We also propose a method of analyzing the vulnerability of industrial economic activities in various countries and industries using new indicators we call “expected potential loss,” which we obtain using the predicted likelihood of bankruptcy and company information. (JEL G0, C0)

New Dataset for Forecasting Realized Volatility: Is the Tokyo Stock Exchange Co-Location Dataset Helpful for Expansion of the Heterogeneous Autoregressive Model in the Japanese Stock Market?

“…The random forest method excels in prediction accuracy compared with the logistic model on average, as we expected, corresponding to the model accuracy in the previous section. Regarding this point, we consider that it is natural and consistent with previous research (Tanaka et al 2018a;Tanaka et al 2019), which have similar frameworks. As noted in the previous section, linear models do not work in the large explanatory variable space.…”

“…This is beneficial to both researchers and practitioners because it allows them to make a better selection toward the financial problem in advances. Third, we found that the random forest method framework works effectively and can be superior to the linear model in the framework of RV forecast, which is in line with the previous studies that used the random forest method for building bankruptcy models of companies (Tanaka et al 2016(Tanaka et al , 2018a(Tanaka et al , 2018b(Tanaka et al , 2019. Our study uses a sufficiently long observation period (2012 to 2019) to consider the change in market quality affected by the HFT system and participants.…”

The predictive strength of MBS yield spreads during asset bubbles

“…It is unlikely that tranches within a specific deal are independent of each other; for instance, the ratings on multiple tranches tend to be modified around the same time (Adelino 2009). Therefore, the reported standard errors are clustered at the deal level to mitigate the correlation of errors within cross-sectional clusters (Cuchra 2004 Tanaka et al 2016Tanaka et al , 2019. Our aim here is to apply these innovative machine-driven techniques to assess the reliability of our findings.…”

“…Feature engineering is a data science procedure of data transformations to achieve higher classification accuracy power with ML methods. This procedure is common in the relevant investigations (Beutel et al 2019;Tanaka et al 2016Tanaka et al , 2019.…”

“…For example, Chen (2011) combine regression trees with logistic regression to analyse and predict corporate financial distress. Tanaka et al (2016) develop an early warning RF framework for bank failures and Tanaka et al (2019) also use RF to predict industry level financial bankruptcy. Mercadier and Lardy (2019) employ RF to approximate credit default swaps to assess default risk of companies when these financial derivatives are not available.…”