Abstract
We investigate the long held but empirically untested assumption that diverse networks drive performance by providingaccess to novel information. We build and validate an analytical model of information diversity, develop theory linkingnetwork structure to the distribution of novel information among actors and their performance, and test our theory usinga unique ten month panel of email communication patterns, message content and performance data from a mediumsized executive recruiting firm. While our theory and results demonstrate that network structures predict performancedue to their impact on access to information, we also find important theoretically driven non-linearities in these relationships.Novel and diverse information are increasing in network size and network diversity, but with diminishingmarginal returns. There are also diminishing marginal productivity returns to novel information, consistent with theoriesof cognitive capacity, bounded rationality, and information overload. Network diversity contributes to performanceeven when controlling for the performance effects of novel information, suggesting additional benefits to diverse networksbeyond those conferred through information advantage. Our theory and results suggest subtle nuances in relationshipsbetween networks, information and economic performance, and the methods and tools developed are replicable,opening a new line of inquiry into these relationships.