You are using an outdated browser.
Please upgrade your browser
and improve your visit to our site.
Skip Navigation

Facing the Facts Doesn't Always Change Minds

A new study on inequality shows that knowledge doesn't move the needle—with one exception

If Americans weren’t such ignoramuses, what would they think about income redistribution? I’m putting it a bit more rudely, but that’s basically the question addressed in a new study (“How Elastic Are Preferences For Redistribution? Evidence From Randomized Survey Experiments”) by Ilyana Kuziemko of Columbia's graduate school of business; Michael I. Norton of Harvard Business School; Emmanuel Saez, the Berkeley economist who (with Thomas Piketty) is our preeminent academic expert on income inequality; and Stefanie Stantcheva, an MIT economist. A nicer way to state the question is: To what extent does the American polity’s resistance to government programs that redistribute income reflect conviction, as opposed to mere lack of information?

Our starting point, of course, is that income distribution has become steadily more unequal in the U.S. during the past three decades (to learn why, read my book The Great Divergence) without creating the sort of political unrest you’d expect. About 100 years ago, when income share for the top one percent was also quite high—though lower than it is today—you had Wobblies, anarchist bombings, etc. “Why aren’t the bottom 99 percent marching in the streets?” I wrote in September 2010. A year later, they were marching in the streets as the Occupy movement took hold. But now they’ve stopped. We still hear a lot about income inequality, including from the president, but we don’t see much political mobilization to do anything about it. Indeed, “there has been a slight decrease in stated support for redistribution in the U.S. since the 1970s, even among those who self-identify as having below-average income [italics mine]." Liberals call this false consciousness and conservatives call it the deeper wisdom of the common man. Neither characterization tells you much about whether this belief is rooted in ignorance or basic values.

To find that out, the authors used an online tool called Amazon Mechanical Turk (mTurk) to quiz a randomized group of about 5,000 respondents, some of whom were, and others weren’t, provided “information on U.S. income inequality, historical data on the link between top income tax rates and economic growth, and the estate tax.” mTurk is a crowdsourcing site used to connect computer programmers with homo sapiens who can perform tasks that computers can’t do (not this week, anyway). Apparently it’s been found also to be a pretty good tool to perform social science research on the cheap while furnishing a sample that (according to a separate 2012 study published in Political Analysis) is “no worse than convenience samples used by other researchers in political science.” (Convenience samples are sample pools drawing on populations close to hand—for example, students at the college where the researcher is located.)

In general, what the researchers found was that it was not too difficult to change people’s opinions about inequality—that is, to persuade them that inequality is a much bigger problem than they realized. Respondents were asked what their income was and then shown where they fell in the income distribution. Then they were shown what their incomes would be today if economic growth since 1980 had been distributed evenly across all income groups. (Short answer: A lot higher for most.) They were shown that economic growth in the U.S. has tended to be higher in times when top marginal tax rates were high, and lower in times when top marginal rates were low. Simply catching respondents up on such facts closed 40 percent of the gap between conservatives and liberals on the question of whether income inequality was “a very serious problem.”

But it was very hard to change people’s policy preferences regarding what should or shouldn’t be done to create more equality—with one important exception that I’ll describe at the end of this column. 

Learning that inequality was a bigger problem than they knew made respondents only slightly more likely to support higher taxes on millionaires, or a higher top marginal income-tax rate, or a higher minimum wage. For example, the study’s mini-tutorial closed the gap between conservatives and liberals only by about 10 percent on the question of whether the top marginal rate should be higher. Opinions about the Earned Income Tax Credit and the Food Stamp program did not change to any degree that was statistically significant.

One explanation for the disconnect between changed opinions and largely unchanged policy preferences might be summed up as “knowledge breeds cynicism.” The researchers found that wising up respondents about the extent of income inequality decreased by nearly 20 percent the proportion who said they “trust government” most of the time. “Emphasizing the severity of a social or economic problem,” the authors write, “appears to undercut respondents' willingness to trust the government to fix it.” The idea, I guess, is that if the government was incompetent enough to let the problem fester, then it won’t likely be competent to fix it. Not exactly the can-do attitude needed to tackle this country’s problems.

Even more depressing, if you’re a liberal like me, was a divide between low-income and high-income respondents on policy preferences that’s exactly the opposite of what you’d expect. The toffs were more likely to support income transfers to the poor after they’d been briefed, while the proles were less likely to. “This result,” the authors write, “is consistent with recent findings by Kuziemko et al. (2011) suggesting that status anxiety can make low-status individuals wary of redistributive policies that may help groups below them.” Kuziemko calls it “last-place aversion.” It’s a familiar narrative to any student of George Wallace’s rise to national prominence in the late 1960s and early 1970s, though the authors don’t do much to clarify the extent to which this remains a divide defined largely by race. (A lot, I would guess, with lower-income Latino immigrants now joining lower-income blacks in the group consigned to Untouchable status.) 

Is there any good news in this study for liberals? Yes, in that exception I mentioned earlier. The one instance where providing information actually changed policy preferences quite a lot was … are you ready? The inheritance (“death”) tax! 

This is a real surprise because various academics have pointed out that Americans have loathed the inheritance tax for a very long time—so long that, in retrospect, it’s a little puzzling Republicans didn’t think to demagogue it until the late 1990s, nearly 20 years after the advent of the “Reagan Revolution.” The main reason Americans hate the inheritance tax is that they believe (incorrectly) that it applies to the non-wealthy. In 2010, a Gallup poll found that averting an increase in the estate tax was Americans’ top priority for Congress (56 percent said it was “very important”), well ahead of extending the Bush tax cuts (50 percent) or unemployment benefits (48 percent). 

“Would correcting this misperception produce widespread public support for the estate tax?” political scientist Larry Bartels of Vanderbilt wrote in 2004. “Probably not.” That turns out to be wrong. Kuziemko et. al. found—President Obama, please take note—that once people had been told the facts on this issue they were much more likely to change their mind than they were about other policies related to income redistribution. People really don’t know how the estate tax works, and once they find out, they stop hating it (unless they’re millionaires).

Why the big shift here and only here on policy preferences? The authors suggest the estate tax may be a rare instance where voters are “highly misinformed” about a matter that is not “linked to racial or other stereotypes.” That would distinguish the estate tax from welfare, a topic on which better information fails to change minds, presumably because it is linked to racial stereotypes (“welfare queens,” etc.). The question of how high the top marginal tax rate should be is similarly unlinked to “racial or other stereotypes,” so I guess the difference there must be that people aren’t as ignorant about it.

To summarize: Americans’ core convictions (not a lack of access to the facts) undergird most of their resistance to policies that redistribute income. Few of these convictions do their holders much credit. But many people can be moved at least a little bit in the liberal direction when provided access to good information (even as doing so deepens their cynicism about government’s ability to address big problems—perhaps the most harmful legacy of Ronald Reagan’s presidency). The demoralization and disaffection of lower-income Americans is a particular worry, one that didn’t happen overnight and won’t be solved overnight. But liberals are well advised to make a concentrated effort to dispel the distortions and outright lies peddled by conservatives about the evils of the “death tax.” Here’s one way to lower the deficit and reduce inequality (at least a little bit) that may turn out to be a much easier sell than anyone would have guessed.

Correction: A previous version of this article said Michael I. Norton works at Columbia’s graduate school of business. He works at Harvard Business School.