|
| Country Risk Ratings: Statistical and Combinatorial Nonrecursive Models by Peter L. Hammer of Rutgers University, March 2004 Abstract: The central objective of this paper is to develop transparent, consistent, selfcontained, and stable country risk rating systems, closely approximating the country risk ratings provided by a major rating agency (Standard & Poor). We propose two models that achieve the stated objectives, the first one utilizing the classical econometric technique of multiple linear regression, and the second one using the combinatorial-logical technique of Logical Analysis of Data. The proposed models use economic-financial and political variables, and are nonrecursive (i.e., they do not rely on the previous years' ratings). The accuracy of the proposed models' predictions, measured by their correlation coefficients with Standard and Poor's ratings, and confirmed by k-folding cross-validation, exceeds 95%. The stability of the constructed non-recursive models is shown in three ways: by the correlation of the predictions with those of other agencies (Moody's and The Institutional Investor), by predicting 1999 ratings using the non-recursive models derived from the 1998 dataset applied to the 1999 data, and by successfully predicting the ratings of several previously non-rated countries. The confidence in the results and in the validity of both models is strongly reinforced by the fact that the traditional linear regression model and the qualitatively different combinatorial-logical model produce almost identical results. Keywords: Country Risk Ratings, Multiple Regression Analysis, Cross-validation, Sovereign. Books Referenced in this Paper: (what is this?) |
|
Please contact me with problems or suggestions. |