The comparison of the predictive ability of different cost-of-equity capital models for the lodging industry

by Lee, Seoki.

The purpose of this study is to evaluate six different ICE estimates and two traditional asset pricing model estimates with regard to their predictive ability for the lodging industry. Six implied cost of equity capital estimates use six different versions of equity valuation models. They are: two versions of residual income valuation models (RIV1 and RIV2), two versions of dividend models (DIV1 and DIV2), Ohlson-Juettner (OJ) model, and Price-Forward-Earnings (PFE) model. The two traditional asset pricing models are Capital Asset Pricing Model (CAPM) and Fama and French Three-Factor (FF) model. The study collected lodging firm data for the entire sample period of 1976 to 2005. EPS forecasting data and historical financial data of both firm and market level are collected from I/B/E/S, Compustat and CRSP. The sample period is limited to 1976 to 2005 due to the limited forecasting data availability from I/B/E/S. The sample data is analyzed by performing a relation test between the cost of equity capital estimates and the future realized returns from quarter-1 to year-3. The findings of the main analysis strongly suggest that the PFE model provides the cost of equity capital estimate with the higher predictive ability than the other eight methods for the lodging industry in the period of 1976 to 2005. The PFE estimate consistently presents statistically significant positive correlations with the future realized returns for the entire sample period; none of the other methods show such relationships.