Herrmann et al. (2001) provide evidence in support of a pricing anomaly whereby the Japanese market underestimates the persistence of current changes in subsidiary earnings for future changes in consolidated earnings. They find a positive relation between changes in subsidiary earnings and year‐ahead stock returns (i.e., the Japanese subsidiary earnings anomaly) and are able to use this information to form zero‐investment hedge portfolios producing positive abnormal returns for 12 out of 13 years and an average annual abnormal return over 7 percent. The purpose of this study is to investigate further whether this relation between subsidiary earnings and year‐ahead returns is attributable to the market behaving naively or failure of the research design to control adequately for cross‐sectional differences in risk. If the anomaly is attributable to naive behavior, then the anomaly should be more pronounced for firms with less informed investors and for firms with higher persistence in subsidiary earnings. If the anomaly occurs due to failure to control adequately for cross‐sectional differences in risk, then the anomaly should exist independent of investor informativeness or subsidiary earnings persistence. Our evidence is consistent with investors behaving naively in further support of the Japanese subsidiary earnings anomaly, although as in all earnings anomaly studies, it is impossible to ever completely rule out risk as an alternative explanation.
