Two political science articles I read recently have surprisingly dissonant conclusions.
Gilens and Page’s study “Testing Theories of American Politics: Elites, Interest Groups, and Average Citizens” is very interesting. You may have spotted it in the news media under any of a host of diverse titles:
The New Yorker: Is America An Oligarchy?
Business Insider: Major Study Finds That The US Is An Oligarchy.
And my favorite, Daily Kos: Too Important For Clever Titles: Scientific Study Says We Are An Oligarchy
(the word “oligarchy” appears in the study only once, at the bottom of page six, as a reference to an alternative theory the authors do not endorse)
But RAMPANT MEDIA PLAGIARISM aside, it’s not a bad summary. The study tries to determine what factors predict whether or not a policy gets implemented in the United States. They compare popular support to elite support, where “elites” are the wealthiest ten percent, and find that elite support is a stronger predictor. I believe the way they put it is that once you know whether elites support a policy, learning whether or not the general public supports it improves your model’s ability to predict whether or not it gets passed only an tiny amount, even though elite opinion and popular opinion are often quite different.
Also recently, Rationalist Conspiracy had a good post on Money Doesn’t Matter In Politics. A lot of anecdotes, but also links to some convincing studies, like the one that shows how “in Congressional races where candidates spent about $250K (1990 dollars), every $100K spent got another 0.3% of the vote, a tiny amount.”
To Alyssa’s list I would add Ansolabehere, Figueiredo and Snyder’s: Why Is There So Little Money In Politics?, recently spotted on Marginal Revolution. The summary (which does not include the word “oligarchy”):
“We show that only one in four studies from the previous literature support the popular notion that contributions buy legislators’ votes. We illustrate that when one controls for unobserved constituent and legislator effects, there is little relationship between money and legislator votes. Thus, the question is not why there is so little money in politics, but rather why organized interests give at all.”
I call these “dissonant” because the simplest explanation for the Gilens and Page finding is that the economic elite are buying elections. But the Ansolabehere et al result says they couldn’t even if they tried. If we take both of these studies at face value, how can we reconcile them?
I can think of a few hypotheses:
1. Legislators vote based on their personal opinions. Most legislators are elite, therefore their opinions correlate with the opinions of other elites.
2. Elites control the media, the universities, et cetera. They affect legislators indirectly, by affecting the entire culture (but how would they do this without influencing commoners? Maybe this is a subset of , in that elites consume elite-produced media?)
3. Legislators would like to think they are elite, and so they vote with elite opinion in the hopes of looking cool and getting elites to like them.
4. Money does not buy elections, but legislators think it does, so they try to satisfy the people with the money in order to win elections.
5. Money does not buy elections, but money can fund think tanks and lobbyists who can persuade legislators through non-election-buying means. This doesn’t take the form of promising financial support or during elections, it just comes from talking and befriending and advising and convincing them. The studies showing money doesn’t affect campaigns miss this effect. Ansolabehere seems to like this one, pointing out that interest groups spend ten times as much as lobbying as on direct campaign contributions. But even here there are economic arguments against. They estimate that one hour of a legislator’s time costs $10,000. This is a high number, but if talking to legislators seriously affected legislation it would be an amazing steal.
6. Elites vote more and are more politically active in terms of volunteering, letter-writing, etc. Legislators try to cultivate their affection to win elections, but it has nothing to do with money. But this effect doesn’t seem strong enough to make up for the small number of elites.
7. The connection between elites and successful policies is a coincidence – not in the sense that the study found a nonsignificant finding, but in the sense that elite opinion and legislative success are both biased in the same direction for different reasons. For example, maybe elites tend to lean conservative, and the conservative party in government is much better organized and able to push more legislation through. Gallup finds there is not a big difference between elites and commoners in terms of basic party labeling. But this study (which does define “elite” somewhat differently) shows that elites are predictably less supportive of welfare and redistribution programs than commoners are (I am enraged that this study doesn’t give good comparative data on social issues). If those programs tend to fail for some reason, that could help produce some of these effects.