Policy as a Private Good: Firm-Lobbyist-Politician Networks in the Legislative Process

Massachusetts Institute Of Technology, Cambridge MA

Investigators

Abstract

General Abstract This project investigates a key question in political science, "Who governs?". It seeks to identify the set of political actors exert who greater levels of influence over policy-making. Representation is one of the fundamental pillars of democracy, ensuring citizens the right to petition elected government officials in order to influence policies. Yet, a vast literature has demonstrated that the policy-making process in the U.S. is dominated by individuals with substantial economic resources. Especially important players include the firms owned by those elites who make campaign contributions, lobby, and make political endorsements to increase their leverage in influencing government policies. Despite the significance of this "unequal representation," empirical studies of political representation have been limited because it is difficult to observe a direct link between political actors and politicians, let alone the true underlying preferences that drive their interactions. The proposed project will investigate the direct representation of specific firms' interests by politicians in the form of highly complex policy outputs. The investigator will develop a theoretical framework to identify the conditions under which firms invest in political networks. To overcome the empirical challenge in identifying the links between firms and politicians, the investigator will then construct a large database of records of firms' lobbying on congressional bills and their political connections to the legislators who sponsor lobbied bills. The project will not only provide valuable insight into the influence of interest groups, it will also produce an easily accessible website that will illustrate the important players and their positions in the policy process. This will be of great value to scholars, and will serve as a valuable teaching resource as well. Technical Abstract This project will develop a set of machine learning tools used to analyze political connections. Specifically, the project will combine the methodologies of ideal point estimation widely used in political science with probabilistic models of latent networks originating in computer science. The primary goal of the proposed methodology is to identify the location of firms, lobbyists, and politicians in a common 'ideological' space in which the relative proximity of various participants implies a closer political connection or alignment of interests. Groups of firms, lobbyists, and politicians who share similar political interests will be clustered together. This methodological framework will allow researchers to identify political networks that are prevalent in various policy dimensions. The open-source software that implements the proposed methodology and the entire database with its web-interface will be made publicly available.