Skip to main content

LUP Student Papers

LUND UNIVERSITY LIBRARIES

Predicting Corporate Defaults: Evaluating Moody's Credit Rating Institute

Baltaev, Alexander LU and Chavdarov, Ivaylo LU (2014) NEKP02 20141
Department of Economics
Abstract (Swedish)
The ability of the Merton model and the logistic regression to accurately forecast corporate defaults is evaluated. Additionally, the probability-of-default (PD) estimates obtained from these two models are compared with the corresponding rating class historic default rates presented by Moody’s. Data for 56 defaulted and 272 healthy US publicly traded organizations serves as the basis for this study. Results reveal that: (i) the logistic regression is more accurate in distinguishing between defaulted and healthy companies, but provides overly conservative PD estimates; (ii) the Merton model struggles to correctly identify true defaults and true non-defaults, while providing default probabilities that are in line with historic default rates... (More)
The ability of the Merton model and the logistic regression to accurately forecast corporate defaults is evaluated. Additionally, the probability-of-default (PD) estimates obtained from these two models are compared with the corresponding rating class historic default rates presented by Moody’s. Data for 56 defaulted and 272 healthy US publicly traded organizations serves as the basis for this study. Results reveal that: (i) the logistic regression is more accurate in distinguishing between defaulted and healthy companies, but provides overly conservative PD estimates; (ii) the Merton model struggles to correctly identify true defaults and true non-defaults, while providing default probabilities that are in line with historic default rates (iii) no framework was deemed superior in this context, ascertaining the difficulty associated with identifying the precise timing of a corporate default. (Less)
Please use this url to cite or link to this publication:
author
Baltaev, Alexander LU and Chavdarov, Ivaylo LU
supervisor
organization
course
NEKP02 20141
year
type
H2 - Master's Degree (Two Years)
subject
keywords
Credit Ratings., Probability of Default, Logistic Regression, Merton Model, Moody's
language
English
id
4499681
date added to LUP
2014-06-26 15:00:22
date last changed
2014-06-26 15:00:22
@misc{4499681,
  abstract     = {{The ability of the Merton model and the logistic regression to accurately forecast corporate defaults is evaluated. Additionally, the probability-of-default (PD) estimates obtained from these two models are compared with the corresponding rating class historic default rates presented by Moody’s. Data for 56 defaulted and 272 healthy US publicly traded organizations serves as the basis for this study. Results reveal that: (i) the logistic regression is more accurate in distinguishing between defaulted and healthy companies, but provides overly conservative PD estimates; (ii) the Merton model struggles to correctly identify true defaults and true non-defaults, while providing default probabilities that are in line with historic default rates (iii) no framework was deemed superior in this context, ascertaining the difficulty associated with identifying the precise timing of a corporate default.}},
  author       = {{Baltaev, Alexander and Chavdarov, Ivaylo}},
  language     = {{eng}},
  note         = {{Student Paper}},
  title        = {{Predicting Corporate Defaults: Evaluating Moody's Credit Rating Institute}},
  year         = {{2014}},
}