The New York Times
By: Ilyana Kuziemko and Steganie Stantcheva
April 21, 2013
How much do Americans care about rising income inequality? Surveys send mixed signals.
In a poll released last year by the Pew Research Center, two-thirds of Americans agreed that there were "strong conflicts between the rich and poor" -- up substantially from when the question was asked in 2009 -- but in a Gallup poll taken at the height of Occupy Wall Street, in the fall of 2011, reducing the income and wealth gap was low on respondents' list of priorities for government action.
It seems like a paradox: Americans are increasingly worried about the gap between rich and poor, but are hesitant to have the government do anything about it.
An experiment we conducted, described in more detail below, may help explain the contradiction. Americans who were given more information about rising inequality expressed significant concern about the problem, but that concern did not always translate into greater support for redistributive policies. Our work identified a possible explanation for this seeming disconnect, and it is a sad one: the more people focused on inequality, the less they trusted the government.
How much Americans care about income inequality and the extent to which they want the government to address it are two distinct questions that are often conflated. Democrats and Republicans agree that America faces a long-run fiscal imbalance that in the coming decades will most likely require cutting social services, raising taxes or both -- policies that directly influence income distribution. Who will bear the brunt of this rebalancing will depend on whether the government uses tax and other policies to counteract rising income inequality with greater redistribution.
Since the 1970s, income inequality in the United States has increased at a historic rate. In 1970, the richest 1 percent of Americans enjoyed 9 percent of total national pre-tax income. In 2011, by contrast, that share had risen to 19.8 percent. And this large increase in inequality has not been softened by more progressive tax policy. Tax rates on the top 1 percent of taxpayers have fallen over the same period.
Such a reshaping of the income distribution was unlikely to go unnoticed, and indeed, surveys show that Americans are generally knowledgeable about the rise in income inequality. Using survey data from 2002, the political scientist Larry M. Bartels showed that three-quarters of Americans believed inequality has increased over the previous two decades. The majority of those respondents said this trend was a "bad thing."
And yet over the past 30 years, Americans have also become less supportive of government efforts to redistribute from high- to low-income households. Between 1991 and 2010, roughly 28 percent of Americans in the General Social Survey -- a continuing survey of opinions and attitudes in the United States, conducted by the University of Chicago -- agreed that the federal government should "improve the standard of living of all poor Americans." (Forty-five percent were neutral, and 27 percent agreed that "it is not the government's responsibility, and that each person should take care of himself.")
This was a sharp decline -- 5 percentage points on average -- in support for redistribution relative to those surveyed between 1975 and 1990.
The survey has also revealed that the share of respondents who believe that the rich should pay proportionally more of their income in taxes than the poor is substantially lower now than in 1987. Gallup polls reveal a similar decrease.
What explains Americans' unwillingness to translate their concerns over inequality into government action? Along with the economist Emmanuel Saez of the University of California, Berkeley, and the psychologist Michael I. Norton of Harvard Business School, we surveyed over 5,000 Americans online in 2011 and 2012. We randomly assigned half of our respondents to complete an online 10-minute tutorial about the increase in income inequality over the last decades. When respondents entered their current household income, the interface calculated a painful counterfactual: what their income would have been had economic growth since 1980 been equally shared among all American households.
In addition, the tutorial emphasized the role that public policy could play in reducing inequality, presenting charts showing that, historically, periods with higher top tax rates have not necessarily been periods of low economic growth. If lack of knowledge is responsible for lack of support for redistribution, our tutorial should have had a meaningful effect on respondents' policy preferences.
Our respondents did, in fact, react to our tutorial by increasing their already significant concern about inequality: the tutorial raised the share of respondents who indicated that inequality was a "very serious problem" by over 40 percent.
However, our results also revealed -- with one key exception -- that support for redistribution increased to a much smaller degree. Our tutorial had only a small effect on support for increasing taxes on millionaires and raising the minimum wage, and no effect on support for other policies that help low-income families, like the earned-income tax credit and food stamps.
What was the exception? Our tutorial informed participants that only about 1 of every 1,000 households was wealthy enough to pay the estate tax. (The tax is currently levied on individual estates worth more than $5.25 million.) Seeing this information sharply increased support for the estate tax. Other surveys have shown that Americans wildly overestimate the share of households subject to the estate tax. The effect of information on increased support for the estate tax has been demonstrated by researchers, in other contexts. It appears that in situations where the public is highly misinformed, providing simple facts can have a big impact on policy views.
But why didn't greater knowledge about inequality translate into support for policies to ameliorate it?
Our experiment suggested an answer: Those who saw our tutorial became 20 percent less likely to agree that government could be trusted at least "some of the time" -- a surprisingly large effect. By emphasizing to respondents the level and growth of income inequality over the last several decades, our tutorial appears to have simultaneously undercut their trust in government's ability to fix the problem. After all, if the government let things get this bad, respondents might logically conclude that it is also unable to do much to fix the situation.
Indeed, the General Social Survey has shown that confidence in government has dropped over the past 30 years. From 1976 to 1989, 23 percent of respondents, on average, reported having "hardly any" confidence in Congress. Since 1990, that number has risen by more than 11 percentage points. The survey has revealed an almost identical erosion of confidence in the executive branch of the federal government. Respondents reacted to our inequality tutorial by reporting lower trust in government, raising the possibility that Americans may have reacted to 30 years of rising income inequality by reducing their trust in government.
Whether or not the rise in inequality has itself lowered Americans' faith in government, the low opinion in which Americans hold their government may well limit their willingness to connect concern with inequality to government action.
Our results have differing implications for people on both ends of the political spectrum. On one hand, liberals can take heart in the news that Americans are deeply troubled about the current level of income inequality. On the other hand, conservatives may be glad to hear that despite this concern, Americans have a healthy skepticism that government can be trusted to do much about it.
Our research suggests that merely talking more about inequality is unlikely to change Americans' policy preferences. Americans are already aware of inequality and are troubled by it. Proponents of greater redistribution can probably save their breath pointing out that inequality is a problem. Instead, they face what seems to be a much more difficult task: convincing them that their government is up to the task of addressing it.