Analyzing Effects of Missing Data, Financial Data, and Corporate Information in Credit Risk Assessment of Japanese Small and Medium Enterprises

2015 Fiscal Year Final Research Report

• Project/Area Number: 24530355
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Research Field: Public finance/Monetary economics
• Research Institution: Akita Prefectural University
• Principal Investigator: Miyamoto Michiko 秋田県立大学, システム科学技術学部, 教授 (30469598)
• Project Period (FY): 2012-04-01 – 2016-03-31
• Keywords: 中小零細企業 / 信用リスク / ロジスティック回帰分析 / 多重代入法 / neglog変換 / ロバストロジスティック回帰分析

Outline of Final Research Achievements

Missing data is found in nearly every use of financial data. In the study of credit risk assessment of small banks, many prior researches have led to information loss by deleting cases with missing values, are left with complete data for all subjects, which do not take advantage of the original dataset. Financial variables may also be non-Gaussian and skewed. The neglog transformation has been introduced to overcome this difficulty. This study analyzes the nature and effects of missingness and skewness of data in credit risk modeling. This study also investigates indicators needed for credit risk measurement for Japanese SMEs, using financial information, as well as corporate information of a large SMEs database and those of a small bank. Lastly, a robust logistic regression extends the conventional logistic model by taking outlier into account, to implement forecast of defaulted firms.

Free Research Field

財政学・金融論