In the context of political networks, we investigated the relationship between the government and the media. First, we documented two-way favors between these actors: advertising favors from the government to connected media, and content favors from connected media to the government. Estimating a model to match these data, we found that much of the observed patterns were driven by favor exchange. Then, we turned to media and political content. We developed a theory in which politicians supply coherent but false alternative realities, featuring elite conspiracies, which enable them to discredit the news media and deflect criticism. We developed implications of this framework, including predictions about socially harmful government policy (e.g. anti-vaccine) and the emergence of conspiracy-spreading media outlets. Taken together, our analysis identifies important barriers to democratic accountability, and suggests policies that help protect press freedom and reinstate accountability.
In the context of business networks, in a set of studies we have documented the large effects of business networks on firm and industry performance. In one study we showed that business associations substantially improve the performance of participating firms. In another study we showed that firm-to-firm referrals create new supplier-client relationships, lead to a rewiring of the transaction network, and generate firm growth. These results identify a new growth barrier based on firm-to-firm search frictions. Turning to industry-level outcomes, in a third study we documented that access to finance had large positive direct effects on firms, but similarly large negative indirect effects on their competitors due to business-stealing. However, access to finance was still socially beneficial due to positive effects on the firms’ consumers. Taken together, our results demonstrate the importance of accounting for network effects in evaluating industrial policies, and identify concrete policies that improve business and industry performance by leveraging these network effects.