This thesis reviews five recent methods of calculating the implied cost of equity capital, and then attempts to apply the Korean firms’ ex-post financial and accounting data to those models and verify which of those proved to be the most reliable pr...
This thesis reviews five recent methods of calculating the implied cost of equity capital, and then attempts to apply the Korean firms’ ex-post financial and accounting data to those models and verify which of those proved to be the most reliable proxy for the realized returns of the Korean stock market between 1981 and 2010. Korean firm data between 1981 and 2010 are applied to each model to calculate the implied cost of equity capital estimates between 1981 and 2010. Then the estimates are regrouped into quintile portfolios according to their magnitude. Finally, the average estimate of each portfolio is compared with the average realized return of each quintile portfolio created by order of realized returns of the same sample firms during the same period. The model with the least difference is declared to be the one producing the best estimates as a proxy for the realized returns in the Korean stock market.
The Capital Asset Pricing Model still seems to produce the best estimates as a proxy for the realized returns of the Korean firms. The other four methods, depending on what equity valuation model they are based on, exhibit different level of performance at different quintiles; the ones based on the residual income valuation model perform better at higher quintiles while the ones based on the abnormal earnings growth valuation model perform better at lower quintiles. Among the four models Gode and Mohanram (2003), in general, is the best when tested with the extreme estimates included. Hou et al. (2010), however, is the best at all of the 2nd, 3rd, and 4th quintiles when considered without the highest and the lowest quintiles and with a longer time horizon. Both models exhibited almost identical performance when the extreme values are removed and the time horizon is short.