The thesis of Bryan Caplan’s The Myth of the Rational Voter, in the briefest terms and paraphrasing Mencken, is that the common people know what they want, and democracy gives it to them good and hard. As Caplan explains it, in the realm of trade policy, steel tariffs are imposed not because clever lobbyists fight to reap concentrated benefits at the (dispersed) expense of the general public, but because average people (and the median voter) like steel tariffs. That’s why the president holds a press conference to announce new protectionism, rather than quietly erecting barriers that favor special interests.
Caplan’s story is very different from the most popular political economy model of trade policy — Grossman & Helpman’s “Protection for Sale.” That model abstracts from much of the political process and applies contract theory (see “Protection for Sale Made Easy“). In short, the lobbies present the government with contribution schedules that induce the government to do what the lobbies want the government to do. The government’s objective function is some linear combination of social welfare and political contributions. The result of lobbying is that governments care too much about producer profits and thus impose tariffs. The model’s predictions are borne out empirically to some degree. Moreover, investigations also find a positive correlation between the size of contributions and tariff levels.
So which story better explains reality? Do lobbies or public opinion trigger protectionism? Are there data that would allow us to distinguish between the two hypotheses, given that they both predict protection?
A synthesis of the two would be Jagdish Bhagwati’s distinction between downstream directly unproductive profit-seeking (DUP) activities – rent-seeking taking place in response to an existing distortion, such as lobbying for a quota allocation – and upstream DUP activities that aim to create distortions. The broad contours of trade policy – a country’s aggregate openness – may be determined by public opinion or the median voter, while the specific manifestations of protection are sufficiently insulated from public perception that they are largely determined by an inside-the-Beltway lobbying process.
Empirically testing this suggested synthesis would be difficult, however. If firms lobby rationally and know that they are constrained by public opinion, then we should expect them to largely engage downstream DUP activities, foregoing upstream DUP activities rendered impotent by the democratic process. But classifying real-world lobbying activities as upstream or downstream may be impossible, making the test hopeless.
Other suggestions for tackling this topic?