재무비율의 CP등급 예측능력에 관한 연구-상장기업총감에 수록된 재무비율을 중심으로-

A Multiple Discriminant Analysis of Commercial Paper Ratings in Korea

Only a few studies have been conducted with attempt to determine the financial information that the rating agencies employ to rate bank holding company and public utility commercial paper. By observing various combinations of pertinent statistical variables, these studies have been moderately successful in classifying commercial paper issues into their proper agency rating group. However, no such study exist for replicating commercial paper ratings in Korea. The purpose of this study is to determine the financial variables that are most important in explaining the Korea Investors Service's rating assigned to the commercial paper. Multiple discriminant analysis is used to predict the Korea Investors Service's rating attributed to the issuers. Financial data are gathered from Korea Listed Companies Associations's Annual Report of Listed Companies(1987) and the Korea Investor Service's rating on July 15, 1988. A total of 157 companies meet these criteria in the first year of prediction and 142 companies in the second year of prediction. To reduce the multicollinearity problems which could hamper the combined predictive ability of financial variables to discriminate among paper issuers in different rating groups, a stepwise inclusion procedure is employed. Based on an 'F-to-enter/F-to-remove' criterion, the original 26 variables are reduced to a smaller set of explanatory variables. Although this procedure does not necessarily eliminate all interrelated variables, it tends to discard those variables which explain similar amounts of data variability, emphasizing the relative uniqueness of the selected variables. By sequentially selecting the variables with the highest F-value, the resulting Wilks' lambda is minimized, thereby maximizing discrimination among rating groups. Two discriminant functions account for 95.51% in the first year and for 97.80% in the second year. As a result, two functions are employed for classification purposes in order to avoid discarding a potentially meaningful amount of discriminating information. The model performs at a low degree of predictive accuracy, correctly classifying 49.68% in the first year and 57.75% in the second year of all commercial paper issuers. These predictive results do not conform to those obtained in the previous commercial paper rating studies conducted in the U. S. This study suggests that the rating agencies do, in fact, observe different parameters in assessing the quality of commercial paper. Accordingly, it appears that only financial ratios do not explain commercial paper ratings adequately.