Article ID: | iaor19951690 |
Country: | United States |
Volume: | 40 |
Issue: | 10 |
Start Page Number: | 1329 |
End Page Number: | 1342 |
Publication Date: | Oct 1994 |
Journal: | Management Science |
Authors: | Mulvey John M., Zenios Stavros A. |
Keywords: | financial, investment, statistics: multivariate |
This paper develops a framework for managing portfolios of fixed income instruments based on traditional principles from the equities market, i.e., based on diversification. It shows, through an analysis of the high-yeild bond market over the period 1987 to 1991, that fixed-income prices could be highly correlated. These correlations can be quantified and integrated, in a systematic way, in an asset/liability management framework. For vanishing fixed income securities, however, we cannot resort to statistical analysis of historical data in order to quantify correlations. The paper develops a forward-looking simulation procedure for capturing correlations. Applications are illustrated for examples from high-yield bonds and mortgage-backed securities. The superiority of the proposed approach over the traditional