This paper examines the performance of the different versions of the generalized comparable-multiples-based valuation method by reference to their ability to predict future returns (i.e. to detect mis-pricing). In contrast to the previous studies, which usually look at the market price or the adjusted market price in the evaluation of the different valuation methods, this research focuses on the post-valuation-date returns from the stocks. In particular, this study measures the raw and abnormal returns from quintile portfolios constructed by reference to the value-to-price ratio, where value is calculated by a range of valuation methods, including the generalized comparable-multiples-based valuation methods (with the conventional price multiples as special cases), the weighted-average models based on earnings and book value, as well as the Residual Income Valuation Model (RIVM) and its equivalent Abnormal Earnings Growth Model (AEGM). If the valuation methods deliver better estimates for the intrinsic value of equity than the current stock price, then a lower (higher) value-to-price ratio would be associated with lower (higher) post-valuation-date returns. I evaluate the performance of the different valuation methods by comparing the returns from the hedge portfolios, defined as the portfolios with a long position in the high value-to-price firms and a short position in the low value-to-price firms. I measure the raw returns and risk-adjusted abnormal returns from the portfolios. The comparison of post-valuation-date returns is done at both the aggregate pooled sample level and a finer nine-industry sub-samples level. It is found that capitalized forward earnings as the valuation anchor has higher return predictability than other stand-alone methods, i.e. the book value of equity, capitalized trailing earnings valuation anchor and RIVM (AEGM). The weighted-average models based on book value and earnings produce average performance. Among the comparables-based valuation methods, it is shown that value estimates produced by the generalized comparable-multiples-based valuation methods using two different value drivers have higher return predictive power than that of the conventional price multiples, especially over longer horizons. The generalized comparable-multiples-based valuation methods with the combination of book value of equity and earnings, especially forward earnings, are found to perform the best in detecting mis-pricing.

This paper examines the performance of the different versions of the generalized comparable-multiples-based valuation method by reference to their ability to predict future returns (i.e. to detect mis-pricing). In contrast to the previous studies, which usually look at the market price or the adjusted market price in the evaluation of the different valuation methods, this research focuses on the post-valuation-date returns from the stocks. In particular, this study measures the raw and abnormal returns from quintile portfolios constructed by reference to the value-to-price ratio, where value is calculated by a range of valuation methods, including the generalized comparable-multiples-based valuation methods (with the conventional price multiples as special cases), the weighted-average models based on earnings and book value, as well as the Residual Income Valuation Model (RIVM) and its equivalent Abnormal Earnings Growth Model (AEGM). If the valuation methods deliver better estimates for the intrinsic value of equity than the current stock price, then a lower (higher) value-to-price ratio would be associated with lower (higher) post-valuation-date returns. I evaluate the performance of the different valuation methods by comparing the returns from the hedge portfolios, defined as the portfolios with a long position in the high value-to-price firms and a short position in the low value-to-price firms. I measure the raw returns and risk-adjusted abnormal returns from the portfolios. The comparison of post-valuation-date returns is done at both the aggregate pooled sample level and a finer nine-industry sub-samples level. It is found that capitalized forward earnings as the valuation anchor has higher return predictability than other stand-alone methods, i.e. the book value of equity, capitalized trailing earnings valuation anchor and RIVM (AEGM). The weighted-average models based on book value and earnings produce average performance. Among the comparables-based valuation methods, it is shown that value estimates produced by the generalized comparable-multiples-based valuation methods using two different value drivers have higher return predictive power than that of the conventional price multiples, especially over longer horizons. The generalized comparable-multiples-based valuation methods with the combination of book value of equity and earnings, especially forward earnings, are found to perform the best in detecting mis-pricing.