Immigration has occupied the news in the past weeks. We have seen disturbing images of immigrants suffering on overcrowded and unsafe boats, not far from European ports, while European leaders argued in the media and in all-night meetings over which country should let them in. The US discusses plans for a border wall while we watch heartbreaking images of children taken from their parents at the Mexican border.
On the Statue of Liberty, one can still read: “Give me your tired, your poor, and your huddled masses yearning to breathe free".
But this message seems very far from today’s attitudes to immigration.
Immigration policy is complex and involves economic, political, and moral considerations. Unfortunately, discussions are often not based on facts and data, but on stereotypes and misperceptions.
In a recent study (Alesina et al. 2018) we used commercial market research companies to run alarge-scale survey and experiment on a representative sample of more than 22,000 natives in six countries: France, Germany, Italy, Sweden, the UK, and the US, mostly between January and March 2018. The sample countries were chosen because they have different economic and social systems, but all have recently faced policy challenges around immigration.
We asked respondents detailed questions about how they perceived immigrants, such as how many they thought there were, where they had come from, and their economic circumstances. Following OECD (2015), we defined an immigrant as a foreign-born legal resident of the country. We also elicited natives’ attitudes towards immigrants, and their preferred immigration and redistribution policies.
Some striking misperceptions about immigrants
In five of the six countries, the average native believed that there are between two and three times as many immigrants as there are in reality. For instance, in the US legal immigrants are about 10% of the population, but US respondents thought the figure was 30%. Similar gaps existed in Germany, France, Italy, and the UK. In Sweden, the country with the highest proportion of immigrants, the public perception of 27% was closest to the true share (18%).
Natives also got the origins of immigrants wrong. They particularly overestimated the shares of immigrants coming from regions that have recently been described as 'problematic' in the media, and the share of non-Christian immigrants – Christianity being the mainstream religion in their country. In all countries except France, respondents overestimated the share of Muslim immigrants. The US and Sweden had the biggest misperception. In the US, respondents thought the share of Muslim immigrants was 23% when in reality it is 10%, and in Sweden they believed the share was 45%, when it is 27%. In the UK, Italy, and Germany, this overestimation ranged from 10 to 14 percentage points.
In all countries, including France, respondents underestimated the share of Christian immigrants by at least 20 percentage points. For instance, US respondents thought that 40% of immigrants were Christian, when 61% are. UK respondents believed 30% of immigrants were Christian, when the true figure is 58%.
In all countries, immigrants were viewed as poorer, less educated, and more likely to be unemployed than is the case. For instance, US natives believed that 35% of immigrants lived below the poverty line, while the real number is less than 14%. Natives also believed that immigrants relied heavily on the welfare state, with roughly one-third of all US, Italian, and French respondents, and one-fifth of all UK and German respondents, believing that an immigrant would receive more benefits than a native, even if both had exactly same income, family structure, age, and occupation. A large share of respondents also thought that immigrants were poor mainly because of lack of effort, rather than adverse circumstances.
These misperceptions were widely spread across all countries and groups of respondents. They were larger for respondents who are not college educated, who said they supported right-wing parties, or who worked in low-skilled occupations in immigration-intensive sectors. Respondents who personally knew an immigrant had less biased perceptions. But of course, getting to know an immigrant may have been the result of their views on immigration, rather than the cause.
Levels of support for immigration and redistribution are positively correlated
Respondents in all countries also greatly exaggerated the share of immigrants among the poor or the low-educated. For example, US respondents thought that 37% of the poor were immigrants; the true number is 12%.
These skewed perceptions may lead natives to conclude that immigrants are a burden on the public finances of their country, and that they disproportionately benefit from redistribution.
In fact, there is a strong negative correlation between the perceived share of poor who are immigrants and support for redistribution. This was captured by a redistribution support index that summarised the answers to all redistribution-related questions. Respondents who perceived that a larger share of the poor were immigrants supported less redistribution, even controlling for a detailed set of personal characteristics.1 Similarly, respondents who supported more immigration overall, as captured by an immigration support index that aggregated the answers to all questions related to attitudes towards immigration, also supported more redistribution.
Making respondents think about immigrants reduced support for redistribution
What are the consequences of these misperceptions? We performed four experiments to establish a causal link between views of immigration and support for redistribution.
In the first, which we called an 'order experiment', we randomly inverted the order in which respondents were asked the questions on immigration and the questions on redistribution, without providing any information on immigrants. Respondents who saw the immigration questions first were merely made to think about immigration, and the demographic and economic characteristics of immigrants, before they answered questions on redistribution. These respondents would have naturally taken immigration into account when answering questions about redistributive policies.
It turns out that simply making respondents think about immigrants and their characteristics made respondents much more averse to redistribution. These respondents also decreased their actual out-of-pocket donations to charities that support low-income groups but do not target immigrants.
Our other three experiments involved showing respondents information about the true characteristics of immigrants – their share, their origins, and their work ethic:
- the first treatment showed the true overall share of immigrants in the country,
- the second showed the true shares of immigrants coming from different parts of the world,
- the third tells an anecdote about a day in the life of a hard-working immigrant.
All these informational treatments significantly increased support for immigration policies.
What about support for redistribution?
Showing the respondents a day in the life of a hard-working immigrant fostered support for redistribution – confirming the importance of views about effort and 'deservingness' of the poor, as highlighted in the case of poor natives in Alesina and Glaeser (2004) and Alesina et al. (2018). But the experiments that showed respondents the true share and origins of immigrants did not generate significantly more support for redistribution.2
We also found that none of these treatments and pieces of information, not even the 'hard work' treatment, was able to overturn the decline in support for redistribution that occurred when respondents were asked the immigration questions first. When respondents were made to think about the demographic and economic characteristics of immigrants, their very negative priors dominated in subsequent answers to redistribution questions, even when they also received favourable information about immigrants.
An important debate in a world of misinformation
One implication of our results is that anti-redistribution politicians, even if they were not averse to immigration per se, can focus on immigration to generate a backlash against redistribution. It also means that this focus on immigration in the political debate – without a preceding correction of the striking misperceptions respondents have about immigrants – could have the unintended consequence of reducing support for redistribution, as well as reducing support for more open immigration.
We have not addressed the important question of where the enormous misperceptions of natives have come from. Perhaps the media overwhelmingly emphasises the negative actions of some immigrants – for instance sensational articles on crimes committed or abuses of the welfare system by immigrants. The media rarely writes about the majority of lawful, honest, and hard-working immigrants who contribute to their host country, perhaps because of the nature of news. The information provided may itself be endogenous if the media or policymakers cater to the views of specific audiences, who then look for confirmation of their biases. This vicious circle perpetuates the diffusion of misinformation if it is in the interest of anti-immigration groups to aggravate misperceptions rather than correct them.
Whatever the source of these misperceptions, the debate about immigration today takes place in a world of misinformation.
The problem of immigration is important in many developed countries. It does not have simple solutions. There is no way of magically resolving the global inequalities that contribute to large immigration flows, or to seamlessly integrate millions of economic immigrants in Europe or the US, or to ensure that multi-ethnic communities live together peacefully.
Difficult choices have to be made in all these countries. People may honestly and reasonably disagree on the correct degree of openness to immigration. But these important and much-needed discussions have to be based on reality, not on misperceptions.
Alesina, A, and E Glaeser (2004), Fighting poverty in the US and Europe: A world of difference, Oxford University Press.
Alesina, A, A Miano, and S Stantcheva (2018), Immigration and Redistribution, NBER Working Paper 24733.
Alesina, A, S Stantcheva, and E Teso (2018), "Intergenerational Mobility and Preferences for Redistribution", American Economic Review 108(2): 521–554.
Kling, J R, J B Liebman, and L F Katz (2007), "Experimental Analysis of Neighborhood Effects", Econometrica 75(1): 83–119.
OECD (2015), Indicators of Immigrant Integration 2015: Settling In, OECD Publishing.
 The redistribution and immigration support indices consisted of equally weighted averages of all policy outcome variables related to redistribution and immigration, following Kling et al. (2007). The variables were all normalised by subtracting the sample mean and dividing by the sample standard deviation.
 Note that these treatments, by necessity, primed respondents to think about immigrants along the particular dimension on which they were focusing (their share or their origins). In the 'order' treatment, making people think about detailed characteristics of immigrants generated a negative reaction to redistribution. Therefore, this priming on its own could have had negative effects on views on redistribution. Thus, the positive information provided may have just counteracted a negative priming effect.