|
Abstract:
|
As we witness a significant impact of chief executive officers, or CEOs, on the performance and reputation of corporations, we analyze the online networks of 136 renowned CEOs using Twitter data that contain a rich set of information on their follow-networks and tweets. The data allows us to identify who is the most influential among the CEOs on Twitter using well-known measures of degree centrality. We distinguish between directed and undirected networks on Twitter to find distinct characteristics in each type of networks. Furthermore, we estimate a network formation model where degree heterogeneity is unobserved. Since observed homophily or heterophily could be correlated with the unobserved degree heterogeneity, one needs to control for the endogeneity issue to identify the model. We adopt a tetrad-level logit estimator proposed in Graham (2017) and control for the unobserved degree heterogeneity. The estimation results show that observed homophily such as account popularity, interests, physical location, and the industry is widespread in CEO’s Twitter networks. On the other hand, we find observed heterophily in account activeness.
|
