Return to search

The Prediction of Industrial Bond Rating Changes: a Multiple Discriminant Model Versus a Statistical Decomposition Model

The purpose of this study is to investigate the usefulness of statistical decomposition measures in the prediction of industrial bond rating changes. Further, the predictive ability of decomposition measures is compared with multiple discriminant analysis on the same sample. The problem of this study is twofold. It stems in general from the statistical problems associated with current techniques employed in the study of bond ratings and in particular from the lack of attention to the study of bond rating changes. Two main hypotheses are tested in this study. The first is that bond rating changes can be predicted through the use of financial statement data. The second is that decomposition analysis can achieve the same performance as multiple discriminant analysis in duplicating and predicting industrial bond rating changes. To explain and predict industrial bond rating changes, statistical decomposition measures were computed for each company in the sample. Based on these decomposition measures, the two types of analyses performed were (a) a univariate analysis where each decomposition measure was compared with an industry average decomposition measure, and (b) a multivariate analysis where decomposition measures were used as independent variables in a probability linear model. In addition to statistical decomposition analysis, multiple discriminant analysis was used in duplicating and predicting bond rating changes. Finally, a comparison was made between the predictive abilities of decomposition analysis and discriminant analysis.

Identiferoai:union.ndltd.org:unt.edu/info:ark/67531/metadc332370
Date12 1900
CreatorsMetawe, Saad Abdel-Hamid
ContributorsWalker, Michael C., Coda, Bernard A., McDonald, James L.
PublisherNorth Texas State University
Source SetsUniversity of North Texas
LanguageEnglish
Detected LanguageEnglish
TypeThesis or Dissertation
Formatvii, 161 leaves, Text
RightsPublic, Metawe, Saad Abdel-Hamid, Copyright, Copyright is held by the author, unless otherwise noted. All rights reserved.

Page generated in 0.002 seconds