|Figure 1: Bottom Decile Irrelevant To Policy|
|Figure 2: Top Decile Much More Influential on Policy Than Median|
These striking figures are from Martin Gilens (2005). Gilens looks at 1,781 questions from opinion surveys given between 1981 and 2002 and soliciting opinions on policy changes that could be implemented by some combination of the president and the Congress. He codes the question based on whether the policy change was implemented in the four years after the survey was given. As I understand it, for each question he performs a regression based on opinions and income. This allows him to analyze the consistency, for each income percentile, of opinions and policy outcome.
Gilens then looks at questions where people at different income percentiles differ in opinion by at least 8% for his scale. He finds 887 such questions for the 10th and 90th percentile, and 498 questions for the 50th and 90th percentile. As you can see, poor people at the 10th percentile have virtually no impact on policy outcomes, and middle income people at the 50th percentile have only slight impact. Empirically, the United States is a plutocracy.
- Martin Gilens (2005). "Inequality and Democratic Responsiveness", , V. 69, N. 5: pp. 778-796.