Why People Vote Against Redistributive Policies That Would Benefit Them

New explanations from economics research.
Source image: Element5Digital, via Unsplash
By: Stefanie Stantcheva | Nov 20

Understanding the connection between citizens’ information, beliefs, and political support for policies that can affect their lives in profound ways is both critical and difficult. Amid rising inequality and political polarization, uncovering citizens’ (mis)perceptions, views on fairness, and economic circumstances is an important first step in addressing problems that currently weaken U.S. democracy. A central puzzle is why so many voters seem to vote against redistributive policies that would benefit them, such as more progressive income taxes, taxes on capital income or estates, or more generous transfer programs, and why voters have tolerated policies that have contributed to a stark rise in inequality over the past few decades.

This article is excerpted from the book “Combating Inequality: Rethinking Government’s Role.”

The median voter model predicts that an increase in inequality, as captured by the gap between median and average income, should lead to an increase in support for redistribution and an increase in actual redistribution as policymakers cater to the median voter’s preferences. Yet, as shown by economist Ilyana Kuziemko and others (including myself), using the General Social Survey, there has been no increase at all in stated support for redistribution in the United States since the 1970s, even among those who say they have below average income.

A New, Innovative Research Tool

A promising way to answer these questions may be through large-scale online surveys and experiments with methodology I have developed and pushed forward over the years that reveal what’s obscured in other datasets. My guiding principle is that we need to listen more to people, not in an idealistic or wishful way but rather through the better use of surveys and online experiments as rigorous research methods. With surveys, I mean directly asking people for answers rather than trying to infer them indirectly from observational data. With experiments, I mean controlling part of the perception or information to see the effect it has on policy outcomes or attitudes.

Views on fairness, equity, and justice are much more complex and context-specific than we have grasped until now.

Surveys have been used for a long time in political science and sociology. In addition, “traditional” surveys were used for tangibles that are now much better recorded in high-quality administrative datasets, such as earnings, demographics, or program eligibility. I believe the online surveys and experiments I employ improve on earlier work in many important ways: They leverage new design and large-scale diffusion methods, they are quantitative and calibrated so they can be rigorously analyzed using econometric and machine-learning methods, and they present respondents with carefully designed, intuitive, and interactive questions. They allow me to reach a large number of people quickly and either target specific harder-to-reach subgroups (such as minorities, younger people, or residents of particular geographic areas) or collect substantial nationally representative samples in one or several countries. The most valuable contribution of such surveys is to reveal three types of intangibles that cannot be seen in other types of data, even high-quality administrative datasets or other “big data.”


The first of these intangibles is perceptions. What are the perceptions that people have about themselves, others, the economic system, and economic policies? Misperceptions may push people to wrongly vote in favor of or against certain policies, render them vulnerable to further misinformation, and damage democracy. Detecting misperceptions is the first step toward fixing them and creating better-informed citizens. As Bryan Caplan writes in his book “The Myth of the Rational Voter: Why Democracies Choose Bad Policies,” “The greatest obstacle to sound economic policy is not entrenched special interests or rampant lobbying, but the popular misconceptions, irrational beliefs, and personal biases held by ordinary voters.”

Views on Fairness

When people decide which policies to support, they weigh views on fairness, equity, and justice that are much more complex and context-specific than we have grasped until now. Views on fairness may be more important than gaps in knowledge; understanding these views and their interplay is key for preventing dangerous slides into populism caused by feelings of unfairness or injustice.

People’s Own Economic Circumstances

We can learn a lot about economic circumstances from administrative records, yet much still remains unknown, and the most direct way of uncovering it may simply be to ask people: Along which margins do people adjust to economic policies? How exactly do people benefit or suffer from policies? As Norbert Lloyd Enrick wrote in 1963, “Every man is ‘aware’ of taxes, especially in this year of 1963. The extent of this awareness has rarely been examined, despite the ever-increasing importance of the public sector. Given our ignorance about tax awareness or tax consciousness, it is surprising that some economists are so willing to predict the effects of changes in the tax structure on individual behavior. If we do not know people’s tax consciousness, how can we know the extent to which changes in their tax burden will affect their behavior?” Policymakers often must operate with limited data that obscure specific circumstances. This may render policies ineffective or, worse, hurt vulnerable groups.


These types of surveys and online experiments have already yielded interesting answers to the question of why people support or oppose redistribution.

A lack of information about the level or rise of inequality does not seem to be the culprit for the lack of support for redistribution. Kuziemko et al. use a series of randomized online survey experiments to show respondents’ personalized information on U.S. inequality, such as where people are in the U.S. income distribution and where they would have been if growth had been equally distributed. They find that respondents’ concern about inequality strongly increases in response to seeing this information, yet this has only a small effect on their support for more progressive policies.

It is worth noting that some type of targeted information about policies does work. For instance, respondents starkly overestimate the share of estates subject to the estate tax. Providing information about the extremely small true share strongly improves support for the estate tax.

Trust in government (or lack thereof) also seems to be a critical element in driving support for redistribution. When faced with negative information about inequality (i.e., that it is high and has increased), respondents tend to say that they trust the government less. This may stem from the belief that if it is politicians who let inequality become this bad, they should not be trusted to remedy it. In all surveys described here, trust in government in the United States in general is abysmally low. Over 89 percent of respondents agree that “politicians in Washington work to enrich themselves and their largest campaign contributors, instead of working for the benefit of the majority of citizens.” In addition, priming respondents to think about topics that they dislike about the government (such as asking them about their opinions on lobbyists or the Wall Street bailout) lowers their trust in the government experimentally. This significantly lowers support for most redistribution policies and increases support for “private charity” over government policies as a better way to reduce inequality.

Findings show experimentally that simply making respondents think about immigrants in a randomized manner before asking them questions about redistribution lowers support for redistribution.

Being able to connect the concern for inequality with concrete public policy measures is critical as well. This idea, raised by the political scientist Larry Bartels, can be explored experimentally, too. One experiment consisted in showing respondents the budget constraint and spending of a household at the poverty line. The experiment was interactive and customized so that the household had the same composition as the respondent’s household. Then, respondents were shown concretely how different government programs (e.g., the minimum wage and food stamps) would alleviate the budget constraints of that household. Such an experiment did improve support for the policies that were shown but not for other redistributive policies that were not specifically mentioned. Taken together with the previous findings, this suggests that when trust in government is so low, the only way to move people’s views on a given policy is to explicitly tell them about their impact on specific families.

Fourth, it appears as if John Steinbeck was at least partially correct in his conjecture made in the 1960s that Americans do not support that much redistribution because the working poor perceive themselves as “temporarily embarrassed.” It does appear to be the case that people are willing to tolerate high levels of inequality if they think that opportunities are relatively equally distributed and that everyone has a chance at climbing the social ladder. Another recent survey experiment I was part of, with Alberto Alesina and Edoardo Teso, shows that more optimistic beliefs about intergenerational mobility reduced support for redistribution in five countries, but beliefs about mobility are not in line with reality. American respondents are in general too optimistic about the “American Dream,” the likelihood of making it from the proverbial rags to riches. On the contrary, Europeans are too pessimistic, specifically about the likelihood of staying stuck in poverty. There is also stark political polarization: Even when shown pessimistic information about mobility, right-wing respondents do not want to support more redistribution policies, because they see the government as a “problem” and not as the “solution.”

Furthermore, inside the United States, there is widespread geographical variation in perceptions of national intergenerational mobility, and these perceptions correlate negatively with the actual state levels of mobility. The South, for instance, has the lowest actual rates of intergenerational mobility in the United States, yet respondents there paradoxically have the most optimistic perceptions.

It does seem that generosity travels less well across ethnic, national, and religious lines and that people dislike redistributing toward people who are different from themselves. One such group is immigrants. Findings from large-scale survey experiments in six countries — France, Germany, Italy, Sweden, the United Kingdom, and the United States — show experimentally that simply making respondents think about immigrants in a randomized manner (which is achieved by asking them questions about immigrants without providing any information) before asking them questions about redistribution lowers support for redistribution. But, importantly, people seem to have strikingly large biases in their perceptions of the number of immigrants, their social and economic characteristics, and their reliance on government transfers.

In all countries in the sample, respondents think there are more immigrants than there actually are and that those immigrants are less educated, more likely to be unemployed, more reliant on government transfers, and more likely to benefit from redistribution than is the case. Respondents also think there are many more Muslim immigrants and many fewer Christian immigrants than there really are. It also appears that, in this instance, providing factual information does not have much power to convince people. Showing information about the actual shares and origins of immigrants to a subgroup of people merely makes the immigration issue more salient to them and further reduces support for redistribution. Telling people an anecdote about the day in the life of a very hardworking immigrant does somewhat better. On balance, it appears that, when it comes to immigration, narratives are more powerful than hard facts in shaping people’s views.

What Can Be Done? The Role of Economists

The first critical thing to clarify is what our goal here is. My strong conviction is that the aim is to give citizens the best tools to understand policies and the economic environment so they can make better decisions for themselves when it comes to policy choices and voting. The goal is by no means to push people in one direction or the other. On the contrary, it is to give them the means to think for themselves. The way forward, in my view, is through better, broader, and earlier economics education. We need outreach and education on economic policy issues for a much broader public, including for young people. This is a long, uphill battle and by no means a quick fix.

A lack of information about the level or rise of inequality does not seem to be the culprit for the lack of support for redistribution.

There are two main challenges to overcome. First, it is clear from this new body of research evidence that hard facts and pure information do not always work to correct misperceptions or improve understanding. Narratives sometimes have a strong hold, too. Hence, education must go beyond facts and explain workings, mechanisms, causes, and consequences in balanced, understandable ways.

Second, “experts” appear to be mistrusted more and more. Economists are no exception. Many scientists face large obstacles in getting through to the public, on issues such as climate change, vaccines, or evolution. Economists are perhaps in an even more difficult position because we do not always have perfect empirical evidence and randomized experiments to answer pressing questions. We are also particularly at risk of being considered biased and partisan. The fault is not wholly ours; “TV economists,” who are often quite nonrepresentative of academic economists, attract more attention in the media than do rigorous researchers. Although there are examples of stellar economic journalism out there, more often than not, the media takes up simple messages, which may be misleading and quite far from the nuanced view an academic economist would hold. But much of it is our responsibility; by not reaching out to a broader public, we are leaving the stage to self-proclaimed, partisan experts. By letting ideologies or political views slant our results, we are imposing a negative externality on all economists, as we hurt their future credibility, too.

Stefanie Stantcheva is a professor of economics at Harvard University and founder of the Social Economics Lab. This article is excerpted from the volume “Combating Inequality: Rethinking Government’s Role.”

Posted on