Prescriptions for network strategy: Does evidence of network effects in cross-section support them?

Although intuitively appealing (and common), drawing network strategy implications from empirical evidence of network performance effects in pooled cross-section is not necessarily warranted. This is because network positions can influence both the mean and variance of firm performance. Strategic prescriptions are warranted if empirically observed network effects reflect increases in mean firm performance. If network effects reflect increases in firm performance variance, however, such prescriptions are warranted only if the increase in the odds of achieving high performance is sufficient to compensate for the concomitant increase in the odds of realizing poor performance. Our simulation study, designed to examine network performance effects in both pooled cross-section and within-firm over time across a wide range of conditions, counsels caution in drawing implications for network strategies. We discuss the implications of our findings for research on network effects, and more broadly for drawing strategic inferences from studies of firm performance in pooled cross-section.