Application Of Discriminant Analysis, Factor Analysis, Logistic Regression, And KMV-Merton Model In Credit Risk Analysis
Handbook of Financial Econometrics, Mathematics, Statistics, and Machine Learning, Chapter 126, ed. by Cheng-Few Lee, World Scientific Publishing Co. Ltd. ISBN: 978-981-12-0238-4
Posted: 6 May 2020
Date Written: April 9, 2020
Abstract
The main purposes of this paper are to review and integrate the applications of discriminant analysis, factor analysis, and logistic regression in credit risk management. First, we discuss how the discriminant analysis can be used for credit rating such as calculating financial z-score to determine the chance of bankruptcy of the firm. In addition, we also discuss how discriminant analysis can be used to classify banks into problem banks and non-problem banks. Secondly, we discuss how factor analysis can be combined with discriminant analysis to perform bond rating forecasting. Thirdly, we show how the logistic regression technique can be used to calculate the default risk probability. Fourthly, we will discuss the KMV-Merton model and Merton distance model for calculating default probability. Finally, we will compare all techniques discussed in previous sections and draw conclusions and give suggestions for future research. We propose using the CEV option model to improve the original Merton DD model. In addition, we also propose a modified naïve model to improve Bharath and Shumway's (2008) naïve model.
Keywords: Discriminant Analysis, Factor Analysis, Logistic Regression, KMV-Merton Model, Probit Model, MIDAS Logit Model, Hazard Model, Merton Distance Model
JEL Classification: C40, C50,C53,C58
Suggested Citation: Suggested Citation