L. Forastiere, V. Leone Sciabolazza, C. Tortú (2020), Dyadic Treatment Effect on Network Formation using Multi-level Propensity Score Matching: Lobbying Activities and Legislative Collaborations.

Firms and corporate companies often work to sway a legislator agenda. For instance, they financially support the electoral campaign of political candidates running for a seat, trying to influence their political activity once elected. It is thus natural to expect that politicians funded by the same firms will collaborate in Congress to achieve the goals of their funders.

In this work, given the bipartite network of financial support from firms to politicians, we define a network of strong ties between congress members where a strong tie is present if the two politicians have a substantial number of common supporters. We then use this network of support ties and the network of collaborations to evaluate the effect for two elected politicians of being supported by common firms on their legislative collaboration.

To conduct our analysis, we develop an estimator for causal effects of the formation of links on a ‘treatment network’ on the formation of links on an ‘outcome network’, with both networks being directed. The estimator is based on an extension of the propensity score matching approach to handle multi-valued treatments, network data and conditional effects.

Using data from the US House of Representatives (111-113 Congress), our results show that sharing common supporters encourages collaborations among politicians.