The current vitriolic debate about immigration may appear light-years away from the poem on the Statue of Liberty: “Give me your tired, your poor, / Your huddled masses yearning to breathe free.” The Economist recently called immigration “perhaps the defining issue of the 2016 election” in the United States, and it has also been an incandescent campaign topic in many recent elections in Europe. At the same time, despite a sharp increase in inequality, sustaining a generous level of redistribution in light of stagnating growth and aging populations is becoming increasingly difficult. The conflicts over scarce resources become even more intense when they are intertwined with national, ethnic, and religious fragmentation.

We examine native citizens’ perceptions of and attitudes toward immigration, and how these relate to support for redistribution. In what ways do people (mis)perceive the number and the characteristics of immigrants? Does a surge in real or perceived immigration flows reduce support for the welfare state? Are people worried about the number of immigrants or rather about their composition—in terms of origin, religion, or economic circumstances? We uncover large misperceptions about the quantity, origin, and characteristics of immigrants, and these misperceptions are related to lower support for redistribution among natives.

We design and run large-scale international surveys on a representative sample of around 22,500 respondents from six countries: France, Germany, Italy, Sweden, the United Kingdom, and the United States. These countries are different economically and socially, but in many ways have had the immigration issue at the center of their political arenas. We elicit the respondents’ perceptions of immigrants, such as the number, origin, or economic circumstances of the latter; we then explore natives’ attitudes toward immigrants and their views on immigration and redistribution policies. To investigate the causal link between perceptions of immigration and redistribution, we also randomly give respondents three “information treatments,” which provide different sets of information about the true share, the origin, and the work ethic of immigrants.

In the survey we define an “immigrant” as somebody legally living in the country of the respondent but born abroad. The surveys—which are restricted to natives—ask detailed questions about the background information of respondents, about their perceptions of immigrants, and about their views on their country’s immigration policies. The respondents’ perception of immigration and their attitudes toward immigration are referred to as the “immigration block.”

The next set of questions asks respondents about their views on certain policies with a focus on redistributive policies, such as how to allocate the government’s budget or how much of the total tax burden people with different incomes should bear. To get at the question of private (non-government-based) redistribution, as well as to test for a real effect of the treatments, we also tell respondents that they are enrolled in a lottery to win $1,000—but before knowing whether they have won, they have to commit a share (zero or positive) of their gain with one or two charities that help low-income people.

Natives overall have striking misperceptions about the number and composition of immigrants. In all countries surveyed, the average and median respondents vastly overestimate the number of immigrants. For instance, in the United States, the actual number of legal immigrants as defined above is 10 percent of the total U.S. population, but the average perception is 36 percent; in Italy, the true share of immigrants is 10 percent, but the perceived share is 26 percent. Respondents also systematically misperceive the composition of immigrants. They believe immigrants are more likely to come from more culturally distant regions (which is often branded as problematic in the public debate) and that they are economically much weaker and less able to contribute to their host country than is the case. For instance, respondents starkly overestimate the share of Muslim immigrants, immigrants from the Middle East and North Africa, and strongly underestimate the share of Christian immigrants. They believe that immigrants are less educated, poorer, more likely to be unemployed, and more likely to receive government transfers than they are in reality.

What is perhaps most striking is that these stark misperceptions hold across all groups of respondents, whether we split them by income, age, gender, education, political affiliation, or sector of work. While substantial heterogeneity exists, and some respondent groups are more accurate than others, they are still substantially wrong. Respondents who have the largest misperceptions along most dimensions we ask about are the low-skilled who work in sectors more exposed to immigrants, the non-college-educated, women, and rightwing respondents. While left- and right-wing respondents misperceive the share of immigrants to the same extent, they have very different views about the composition of immigrants: right-wing respondents in all countries systematically consider immigrants to have less desirable characteristics.

Those who live in a commuting zone in the United States with a high share of immigrants have larger misperceptions. The perceived share of immigrants alone is not a key driver of support for either immigration or redistribution policies, but the perceived characteristics of immigrants are. Controlling for the full array of individual respondent characteristics, including political affiliation, we see that support for immigration and redistribution are strongly positively predicted by the perceived work ethic of immigrants and the share of immigrants that are highly educated, as well as by knowing an immigrant personally. They are significantly negatively predicted by the perceived share of immigrants who are freeriding, low-educated, unemployed, or Muslim.

We then turn to our experimental part and our informational treatments. We begin with our “order of the questions” treatment, whereby half of the respondents are randomly shown the immigration block before the redistribution block and vice versa. This allows us to study the effects of purely making respondents think about immigration and the characteristics of immigrants on their answers to questions related to redistribution policy. We find significantly negative effects of simply prompting respondents to think about immigrants and their composition: respondents who are asked first about their perceptions of immigration (without receiving any information on immigrants) and only then about redistributive policies show a significantly larger aversion to redistribution—including actual donations to charity—than those who are asked about redistribution first and immigration second. This is to be interpreted in light of our aforementioned findings regarding the negative views that respondents hold about immigrants, their difference to them, and their economic contributions to their host country. Respondents are also randomized into one of three informational treatment groups. The first informational treatment informs respondents about the true number of immigrants in their country; the second treatment informs them about which regions immigrants in their country come from; and the third one shows them an anecdotal day in the life of a low-income, hard-working immigrant. Our three informational treatments have strong first-stage effects: treated respondents’ perceptions on the number, origin, or hard work of immigrants are significantly different from those of the control group and, expectedly, the informational treatments reduce misperceptions. We also conducted a followup survey in the United States to show that the effects on perceptions of the informational treatments persist after one to three weeks.

The “hard working immigrant” treatment on its own has strong effects on support for redistribution: treated subjects become significantly more favorable to redistribution when reminded that at least some immigrants are hard-working. However, when respondents are shown the immigration block first, and are thus asked to go through detailed questions about the number and characteristics of immigrants, their negative perceptions dominate; none of the favorable informational treatments are able to overcome the negative perceptions of redistribution raised by prompting people to think at length about immigrants’ characteristics. Since all groups of respondents have negative and biased baseline views of immigrants, all of them react negatively to being made to think about immigrants. Groups which have more negative baseline views (e.g., the non-college-educated, the low-skilled in immigration-intensive sectors, and rightwing respondents) react more strongly to the order treatment and are less inclined to change their views after viewing the favorable hard work treatment.

Our results imply that people’s attitudes on immigration and redistribution are formed in an environment of misinformation—perhaps even disinformation. Rather than being corrected, as we attempt to do here, these misperceptions may be strategically manipulated or even fostered by parties or interest groups averse to immigration or redistribution.

NOTE
This research brief is based on Alberto Alesina, Armando Miano, and Stefanie Stantcheva, “Immigration and Redistribution,” NBER Working Paper no. 24733, June 2018, http://​www​.nber​.org/​p​a​p​e​r​s​/​w​24733.