Research summary: We revisit the empirical relationship between multinationality and performance by attempting to replicate the widely cited S‐shape relationship reported in Lu and Beamish (2004). Using a longitudinal and comprehensive database on the population of U.S. MNCs from 1989 to 2007, we find no evidence of an S‐shaped relationship; nor do we see a moderating effect of intangible assets. Although our results do show a marginally significant U‐shaped association between multinationality and performance for a subsample of manufacturing firms, this relationship disappears once we account for the endogeneity of multinationality. Our study contributes to empirical research on the multinationality‐performance relationship, highlighting the need for caution in generalizing results across countries and the importance of controlling for the endogeneity of multinationality when assessing its effect on performance. Managerial summary: Our study examines the relationship between a firm's multinationality and its performance. In a much‐cited study, Lu and Beamish (2004) found evidence of an S‐shaped relationship—with firm performance first decreasing, then increasing, then decreasing again as firms internationalized—in a sample of Japanese firms from 1986 to 1997. We test for the same relationship across all U.S. MNCs from 1989 to 2007, and find no evidence of an S‐shaped pattern, or indeed, of any effect of multinationality at an aggregate level. Our study thus suggests that the effect of multinationality may vary with firm capabilities and home country environments, and that managers and academics alike should focus on understanding these specifics, rather than searching for a universal effect of multinationality on performance. Copyright © 2016 John Wiley & Sons, Ltd.