Lobbying shapes what bills get passed---best data

In this article, I have presented a unique approach to the empirical analysis of political influence by interest groups based on the specification and estimation of an all-pay contest with heterogeneous interest groups over policies considered in the US Congress. One of the main contributions of this article is that I debut a novel unit of analysis: policies, which are parts of bills, rather than bills themselves as in previous works. This is particularly relevant for the study of lobbying behaviours because the content of a bill can and often does change throughout the entirety of the legislative process. I show that bill-level analyses which take the content of a bill as exogenously given can generate biased estimates of the effects of lobbying expenditures on policy changes.

Using a newly-constructed dataset that contains information on policies and lobbying activities, I have quantified the effect of lobbying expenditures on the probability that a policy is enacted, and estimated the average returns to lobbying expenditures for or against a policy. I find that the effect of lobbying expenditures on a policy's equilibrium enactment probability is very small. Nonetheless, the average returns from lobbying expenditures are estimated to be over 130%. In this study, I focus on energy policies and lobbying activities targeting these policies by energy firms. Given that lobbying expenditures by the energy sector comprise 12% of all lobbying expenditures, these findings are interesting in their own right, though it remains to be seen whether the results for lobbying in the energy sector will extend to lobbying in other domains. The approach developed in this article can, however, be applied to study the effects of lobbying in other policy domains.

The findings are closely related to the puzzle that the total amount of lobbying expenditures is relatively small when compared to the value of the government policies they are intended to influence. A similar observation regarding campaign contributions was made by Tullock (1972) and Ansolabehere et al. (2003) . If lobbying is a part of the economic activities of interest groups, one potential explanation for the puzzle is that the average returns to lobbying are small. However, I find that the average returns are much larger than normal market returns. Furthermore, articles that look at lobbying expenditures and stock returns, such as Hill et al. (2013) , find that shareholders value lobbying activities. This implies that lobbyists could charge the interest groups much more, but they do not.

This suggests that significant frictions may exist in the market for policy influence. One such friction is limited access to the market. Granted, in this article, I impose very minimal market frictions on the four energy lobbying coalitions. The only friction in the model is that the coalitions are supposed to incur the minimum initial lobbying costs. However, this almost unrestricted access to the market may be available only to certain firms and trade associations. Another friction in the market is related to political organization as described by Olson (1965) . A further study on these potential frictions in the market can be very important to our understanding of the policy-making process and the welfare implications of the regulation of lobbying.