From the Dept. of Academics Confirming Something You Already Suspected comes a new study concluding that rich people and organizations representing business interests have a powerful grip on U.S. government policy. After examining differences in public opinion across income groups on a wide variety of issues, the political scientists Martin Gilens, of Princeton, and Benjamin Page, of Northwestern, found that the preferences of rich people had a much bigger impact on subsequent policy decisions than the views of middle-income and poor Americans. Indeed, the opinions of lower-income groups, and the interest groups that represent them, appear to have little or no independent impact on policy.
“Our analyses suggest that majorities of the American public actually have little influence over the policies our government adopts,” Gilens and Page write:
That’s a big claim. In their conclusion, Gilens and Page go even further, asserting that “In the United States, our findings indicate, the majority does not rule—at least not in the causal sense of actually determining policy outcomes. When a majority of citizens disagrees with economic elites and/or with organized interests, they generally lose. Moreover … even when fairly large majorities of Americans favor policy change, they generally do not get it.”
It is hardly surprising that the new study is generating alarmist headlines, such as “STUDY: US IS AN OLIGARCHY, NOT A DEMOCRACY,” from, of all places, the BBC. Gilens and Page do not use the term “oligarchy” in describing their conclusions, which would imply that a small ruling class dominates the political system to the exclusion of all others. They prefer the phrase “economic élite domination,” which is a bit less pejorative.
The evidence that Gilens and Page present needs careful intepretation. For example, the opinion surveys they rely on suggest that, on many issues, people of different incomes share similar opinions. To quote the paper: “Rather often, average citizens and affluent citizens (our proxy for economic elites) want the same things from government.” This does get reflected in policy outcomes. Proposals that are supported up and down the income spectrum have a better chance of being enacted than policies that do not have such support. To that extent, democracy is working.
The issue is what happens when some income groups, particularly the rich, support or oppose certain things, and other groups in society don’t share their views. To tackle this issue, Gilens and Page constructed a multivariate statistical model, which includes three causal variables: the views of Americans in the ninetieth percentile of the income distribution (the rich), the views of Americans in the fiftieth percentile (the middle class), and the opinions of various interest groups, such as business lobbies and trade unions. In setting up their analysis this way, the two political scientists were able to measure the impact that the groups have independent of each other.
This is what the data shows: when the economic élites support a given policy change, it has about a one-in-two chance of being enacted. (The exact estimated probability is forty-five per cent.) When the élites oppose a given measure, its chances of becoming law are less than one in five. (The exact estimate is eighteen per cent.) The fact that both figures are both below fifty per cent reflects a status-quo bias: in the divided American system of government, getting anything at all passed is tricky.
The study suggests that, on many issues, the rich exercise an effective veto. If they are against something, it is unlikely to happen. This is obviously inconsistent with the median-voter theorem—which holds that policy outcomes reflect the preferences of voters who represent the ideological center—but I don’t think that it is a particularly controversial claim. A recent example is the failure to eliminate the “carried interest” deduction, which allows hedge-fund managers and leveraged-buyout tycoons to pay an artificially low tax rate on much of their income. In 2012, there was widespread outrage at the revelation that Mitt Romney, who made his fortune at the leveraged-buyout firm Bain Capital, paid less than fifteen per cent in federal income taxes. But the deduction hasn’t been eliminated.
One of the study’s other interesting findings is that, beyond a certain level, the opinions of the public at large have little impact on the chances a proposal has of being enacted. As I said, policy proposals that have the support of the majority fare better than proposals which are favored only by a minority. But, in the words of Gilens and Page, “The probability of policy change is nearly the same (around 0.3) whether a tiny minority or a large majority of average citizens favor a proposed policy change.”
The paper is a provocative one, and there’s sure to be a lot of debate among political scientists about whether it wholly supports the authors’ claims. One issue is that their survey data is pretty old: it covers the period from 1982 to 2002. (On the other hand, it hardly seems likely that the influence of the affluent has declined in the past decade.) Another issue is that, in a statistical sense, the explanatory power of some of the equations that Gilens and Page use is weak. For example, the three-variable probability model that I referred to above explains less than ten per cent of the variation in the data. (For you statistical wonks, R-squared = 0.074.)
Even in this sort of study, that’s a pretty low figure. Gilens and Page, to their credit, draw attention to it in their discussion, and suggest various reasons for why it’s not a big issue. They also acknowledge another possible objection to their conclusions:
Me, too. There can be no doubt that economic élites have a disproportionate influence in Washington, or that their views and interests distort policy in ways that don’t necessarily benefit the majority: the politicians all know this, and we know it, too. The only debate is about how far this process has gone, and whether we should refer to it as oligarchy or as something else.
Photograph: Ryan Heffernan
