In this study we employ a multilevel analysis to investigate the relative importance of industry, business group, corporation, and business unit effects on the performance of business units that are nested within business group affiliates. Empirical results demonstrate that business unit effects explain (1) about four times as much variance as does the corporate effects, (2) over four times as much variance as the industry effects, and (3) over eighteen times as much variance as the business group effects in business unit performance. Moreover, results provide evidence that (4) corporate effects explain almost five times as much variance as the business group effects in determining business unit performance. These findings extend our theoretical and empirical understanding of performance variability in the context of business groups