It is conceived that the percentage of completion method make periodical earnings smoother than the completed contract method and that it can reflect construction activities of the firm on earnings more timely. Based on this intuition, some people believe that earnings recognized by the percentage of completion method is more value-relevant than that by the completed contract method. Though various measures of value relevance have been adopted in prior studies, this paper adopts the earnings capitalization model and investigates whether the capitalization multiple of earnings by the percentage of completion method is higher than the other. This research focuses on not only the revenue recognition method but also the ratio of revenues recognized by the percentage of completion method. By controlling the differences in firm characteristics, the difference in value relevance between revenue recognition methods is examined by means of cross-sectional comparison. The results show that the capitalization multiple for firms adopting the percentage completion method is not higher than that of the completed contract method. The capitalization multiple for subgroup of lower ratio is significantly lower than other subgroups. Those results are opposed to the common belief that adoption of percentage of completion method improves the usefulness of earnings information. Though unification into the percentage of completion method is took up for discussion in the process of conversion of international accounting standards, empirical results in this paper suggest that the grounds for the unification should be re-examined.