Comparative Value Relevance Studies: Country Differences Versus Specification Effects

This paper sheds light on the sensitivity of findings in comparative international value relevance studies regarding two fundamental methodological choices. We hypothesize and find that, first, using the regression vs. the portfolio returns specification and, second, the choice of the return window, is not arbitrary. Both choices will have an impact on country rankings and the significance of cross-country differences in comparative designs. This makes us conclude that findings in previous comparative international value relevance studies are partly driven by differences in market characteristics across countries. Extending the findings of Francis and Schipper (1999) and Collins and Kothari (1989), our results suggest that previous comparative studies might thus have overstated value relevance differences and institutional variables' power to explain these differences across countries. Findings are based on a treatment sample of 56,000 firm-year observations from 12 countries and from 12 matched U.S. control samples, with observations from 1988 to 2007.