Theory and empirics on lobbying friends and enemies

The primary approach to studying lobbying assumes that it allows interests to provide information to politicians. Literature has largely focused on the strategies that lobbyists can use to overcome the problem of credible communication under misalignment of preferences. This appears at odds with apparent empirical regularities showing that lobbying targets friends. Yet this evidence has not convincingly identified politicians' pre-lobbying preferences. Avoiding this problem, I present a formal model of lobbying whose empirical implications hinge on the degree to which business firms are internally hedged against political risk, i.e., politically diversified. Importantly, the firms most desperate to manage this risk through lobbying (that would target enemies) are least able to credibly communicate. I then develop a novel measurement strategy that employs network analysis methods to quantify the degree to which firms are politically diversified. Empirical results show that diversified firms lobby most, implying that lobbying targets friends