Predicted probability of policy adoption (dark lines, left axes)  by policy disposition; the distribution of preferences (gray columns, right axes) for average U.S. citizens and elite groups. Date are compiled from roughly 1,800 different policy initiatives in the years between 1981 and 2002, these policy changes are compared with the expressed opinion of the United State public. Graphic: Gilens and Page, 2014 [UPDATE: Objections have been raised about this study, e.g., here. I asked Prof. Gilens for his rebuttal, and he graciously replied. His response is appended at the end of this post. –Des]  By Tom McKay
16 April 2014 (PolicyMic) – A new scientific study, “Testing Theories of American Politics: Elites, Interest Groups, and Average Citizens” [pdf], from Princeton researchers Martin Gilens and Benjamin I. Page has finally put some science behind the recently popular argument that the United States isn’t a democracy any more. And they’ve found that in fact, America is basically an oligarchy. An oligarchy is a system where power is effectively wielded by a small number of individuals defined by their status called oligarchs. Members of the oligarchy are the rich, the well connected and the politically powerful, as well as particularly well placed individuals in institutions like banking and finance or the military. For their study, Gilens and Page compiled data from roughly 1,800 different policy initiatives in the years between 1981 and 2002. They then compared those policy changes with the expressed opinion of the United State public. Comparing the preferences of the average American at the 50th percentile of income to what those Americans at the 90th percentile preferred, as well as the opinions of major lobbying or business groups, the researchers found out that the government followed the directives set forth by the latter two much more often. It’s beyond alarming. As Gilens and Page write, “the preferences of the average American appear to have only a minuscule, near-zero, statistically non-significant impact upon public policy.” In other words, their statistics say your opinion literally does not matter.

The Separate Policy Impact of Business-oriented and Mass-based Interest Groups

Interest group Influence coefficient
Average citizens’ preferences .05 (.08)
Economic elites’ preferences .78 (.08), p<.001
Mass-based interest groups .24 (.07), p<.001
Business interest groups .43 (.08), p<.001

N = 1,779
R2 = .07 [UPDATE: See the end of this post for a statement about low R-values from Prof. Gilens. –Des] All predictors are scaled to range from 0 to 1. The dependent variable is the policy outcome, coded 1 if the proposed policy change took place within four years of the survey date and 0 if it did not. Predictors are the logits of the imputed percent of respondents at the 50th (“average citizens”) or 90th (“economic elites”) income percentile that favor the proposed policy change, and the Net Interest Group Alignment Indices described in the text. Standard errors are asymptotically distribution-free, and all analyses reflect estimated measurement error in the predictors, as described in Appendix 2. That might explain why mandatory background checks on gun sales supported by 83% to 91% of Americans aren’t in place, or why Congress has taken no action on greenhouse gas emissions even when such legislation is supported by the vast majority of citizens. This problem has been steadily escalating for four decades. While there are some limitations to their data set, economists Thomas Piketty and Emmanuel Saez constructed income statistics based on IRS data that go back to 1913. They found that the gap between the ultra-wealthy and the rest of us is much bigger than you would think, as mapped by these graphs from the Center On Budget and Policy Priorities. Piketty and Saez also calculated that as of September 2013 the top 1% of earners had captured 95% of all income gains since the Great Recession ended. The other 99% saw a net 12% drop to their income. So not only is oligarchy making the rich richer, it’s driving policy that’s made everyone else poorer. [more]

UPDATE from Prof. Gilens: Jim, there are a number of factors that contribute to the low r-sq values for our analyses. First, there are a large number of proposed policy changes which do not elicit either strong support or strong opposition from the public or interest groups. This might be due to indifference, but often is the result of roughly equally strong forces in favor and in opposition. In these cases, we have little ability to predict the outcome (which is more likely to be determined by idiosyncratic features of the proposal and context than by systematic characteristics of the system that any “large n” analysis could hope to capture). Second, there are statistical limitations to the amount of variance that can be accounted for (which the r-sq reflects) because we are using continuous predictors to account for a dichotomous outcome. The underlying probability function is continuous (i.e., policies vary in the likelihood of being adopted) but we can’t observe that, only the outcome (adopted or not adopted). Finally, we have only proxy measures for our two “most important” predictors (economic elites and interest groups) which limits their predictive power. Of course, better measures of the factors captured in our analyses, or additional measures of factors that are not captured, might either lead to different conclusions or might reinforce the conclusions we drew from the data we have. All best, Marty Gilens

Princeton Concludes What Kind of Government America Really Has, and It’s Not a Democracy