Search results

This paper proposes a method that estimates credit ratings by mapping empirical probability of default (PD) and standardized historical financial ratios. Unlike standard approaches such as the parametric logit model. discriminant analysis. neural network. and survival function model. the proposed approach has an advantage of offering a multiple credit rating categories. as opposed to either default or not default. of obligors. It would provide an useful information to practitioners because the probability of default for each credit rating category is a critical input under New Basel Capital Accord. Emoirical results based upon the historical PD and financial ratios of Korean savings bank industry from 2000 and 2003 suggest that the industry’s average credit rating belong to a speculative grade. that is BB and below.

In addition, the computed transition matrix indicates that volatility of transition matrix fluctuates substantially each year and the orobability of staying in the same rating category at the end of the year tended to be much smaller than the average reported by the rating agencies for the overall Korean companies. The proposed method can easily be applied to industries other than savings bank industry.

This paper tries to estimate multivariate latent factor model with jump in order to find common factor and jump risk of KOSDAQ markets. Using five major daily KOSDAQ indexes such as construction, wholesale, transportation, finance, and IT/SW/SVC from January 2 2003 to August 29 2008, this study finds the evidence of significant systematic jump risk in addition to industry-specific idiosyncratic risk. According to the main estimated results of this paper, jump risk comes every 31 trading days in KOSDAQ markets and approximately twenty percent of the common factor of the KOSDAQ market can be explained by the KOSPI market risk.

This paper investigates whether stock‐price indexes of emerging markets can be characterized as random walk (unit root) or mean reversion processes. We implement a panelbased test that exploits cross‐sectional information from seventeen emerging equity markets during the period January 1985 to April 2002. The gain in power allows us to reject the null hypothesis of random walk in favor of mean reversion at the 5 percent significance level. We find a positive speed of reversion with a half‐life of about 30 months. These results are similar to those documented for developed markets. Our findings provide an interesting comparison to existing studies on more matured markets and reduce the likelihood of earlier mean reversion findings as attributable to data mining.