Applying recurrent event analysis to understand the causes of changes in firm credit ratings

Yan Shing Chen, Po Hsin Ho, Chih Yung Lin, Wei Che Tsai

Research output: Contribution to journalArticlepeer-review

1 Scopus citations

Abstract

This study applies recurrent event analysis to examine the determinants of changes in firm credit ratings. This study uses two extended Cox proportional hazard models to examine upgrade and downgrade data separately. Explanatory variables are taken from financial ratios in Z-score (Altman, 1968) and AR-score (Altman and Rijken, 2004) models. The empirical results first suggest that sales to asset ratio and market equity to book debt ratio are the key explanatory variables for the sample comprising credit rating upgrade firms examined using Z-scores specification. Next, the sample of credit rating upgrade firms examined using AR-score variables reveals that the first rating of young firms is generally underestimated. Additionally, analysis of sample comprising credit downgrade firms examined using Z-score specification identifies working capital to asset ratio and market equity to book debt ratio as the key explicative variables. Furthermore, analysis of sample of credit downgrade firms examined using AR-score model reveals that larger firms are not easily downgraded, and old firms are more likely to be downgraded because of their ratings typically having initially been overestimated. Finally, high q firms with high retained earnings may suffer from underinvestment problem. Consequently, credit agencies may be reluctant to upgrade such firms.

Original languageEnglish
Pages (from-to)977-988
Number of pages12
JournalApplied Financial Economics
Volume22
Issue number12
DOIs
StatePublished - Jun 2012

Keywords

  • AR-score
  • credit rating change
  • recurrent event analysis
  • underinvestment problem
  • Z-score

Fingerprint

Dive into the research topics of 'Applying recurrent event analysis to understand the causes of changes in firm credit ratings'. Together they form a unique fingerprint.

Cite this