Long-run contact with immigrant groups, prejudice, and altruism

Leonardo Bursztyn, Thomas Chaney, Tarek Hassan, Aakaash Rao 21 March 2021



Hostility against groups branded ‘non-native’ – Middle Eastern and North African refugees in Europe; Mexican and Central American migrants in the US; Rohingya in Myanmar; Kurds in Turkey – has risen sharply over the past decade, coinciding with a wave of far-right populism predicated on opposition toward minorities. The ‘contact hypothesis’ (Allport 1954) posits that greater intergroup contact can, under certain conditions, mitigate these prejudices and increase altruism. If the hypothesis is correct, then fostering the conditions under which contact can engender positive attitudes, and promoting contact under these conditions, may be among the most effective tools to reduce intergroup hostility. But testing the contact hypothesis is often complicated by concerns of reverse causality. If we observe that members of the majority group who report greater contact with minorities are less prejudiced against them, is it because this contact causally reduces their prejudice, or because they were less prejudiced to begin with?

A wealth of research spanning the social sciences has made progress on this question, often employing either randomised experiments or quasi-experimental variation to identify the causal effect of contact in different settings. Rao (2019) exploited the staggered introduction of quotas mandating that Indian schools admit a certain fraction of poor students, and examined how contact with these students affected wealthier students’ attitudes and behaviour. Contact with poor students made wealthy students more prosocial and egalitarian, more willing to socialise with poor students, and less likely to discriminate against poor students. 

Other work has found evidence of heterogeneity. Lowe (2020) and Mousa (2020) randomise the composition of sports teams, and although both suggest that cooperative contact (playing on the same team as a member of the out-group) leads to more positive social preferences. Lowe finds that adversarial contact (playing on an opposing team) has the opposite effect, while Mousa finds that these more positive social preferences do not translate to contexts beyond the sports pitch. The evidence on the persistence of these effects is also mixed. Some studies (Schindler and Westcott 2020, Bagues and Roth 2020) find that the positive effects of contact persist over years or decades, while others (e.g. Dahl et al. 2020) find that they fade relatively quickly.

Given these disparate findings, a crucial class of remaining questions concerns the aggregate effect of long-run contact. Summing up overall types of naturally occurring interactions over the course of decades, how does intergroup exposure shape beliefs and prejudices; how does this exposure translate into real-world outcomes; and to what extent are these effects consistent across different out-groups? We make progress on these questions in our recently published working paper (Bursztyn et al. 2021), in which we identify the causal effect of long-term intergroup contact on a comprehensive range of outcomes in the most natural possible setting – day-to-day interaction over the course of decades – and generalise our results to examine the effects of exposure to populations from several dozen countries.

Setting and empirical approach

Estimating the causal effect of long-run exposure to immigrant groups requires plausibly exogenous variation in such exposure. We draw upon the empirical approach of Burchardi et al. (2018), who construct a set of instruments for the present-day ancestral composition of US counties based on the confluence of ‘push’ and ‘pull’ factors in immigration. The logic is simple: if many people are emigrating from a given foreign country (e.g. Mexico) at the same time as settling in a given domestic destination (e.g. Detroit) is appealing to immigrants from all origin countries, then that domestic destination will receive large inflows of immigrants from that country. Over the course of decades (we draw upon 130 years of detailed Census data), this generates variation in the ancestral composition of counties that is plausibly unrelated to other factors that might independently affect the attitudes of the destination’s native residents.

Case study: Arab-Muslims in the US

We begin our analysis by examining the attitudes of the ‘majority group’ (white Americans) toward one of the most targeted minorities in the recent surge of nationalist authoritarianism – Arab-Muslims. We find that exposure leads to more positive attitudes, as measured by both explicit questions and the implicit association test (IAT). White, non-Muslim respondents who reside in US counties with (exogenously) larger populations of Arab ancestry are less explicitly and implicitly prejudiced against Arab-Muslims. Measured prejudice against Arab-Muslims does not simply proxy for more or less racial prejudice in general or more or less conservative political views; the estimated effect remains significant when we control for these factors. Exposure also affects political preferences – non-Muslim white Americans in counties with greater exposure to people of Arab ancestry are less supportive of a proposed ‘Muslim ban’ and, in 2016, were less likely to vote for Donald Trump, even controlling for voting behaviour in 2012.

Figure 1 Effects of exposure to Arab-Muslims on attitudes and political preferences

We next turn toward a revealed-preference measure of generosity. Using large-scale individualised datasets from two charity organisations – and applying machine learning techniques to donors’ names to focus on those likely of European descent (the majority group in the US) – we document that individuals from counties with exogenously larger populations of Arab ancestry are more likely to donate, and donate larger sums, to charitable causes in Arab countries following natural or manmade disasters.

We conclude our analysis of Arab-Muslims by presenting the results of a large-scale custom survey designed to examine the underlying mechanisms. We find that an exogenously larger Arab-Muslim population in a respondent's county substantially increases both the probability that the respondent has visited a Middle Eastern restaurant and the probability that the respondent knows an Arab-Muslim friend, neighbour, or workplace acquaintance. A larger Arab-Muslim population also substantially increases respondents' knowledge of Arab-Muslims and Islam in general, and decreases the probability that respondents believe that “holy war against non-believers” or the “subservience of women and children to men” are fundamental tenets of the faith.

Figure 2 Effects of exposure on contact with Arab-Muslims and knowledge about Islam

Generalising results

We then expand our analysis beyond Arab-Muslims and examine all of the several dozen countries in our donations data, seeking to understand whether our results generalise. We show that exposure to a local population of a given foreign ancestry increases generosity toward that country. Our estimated effects of exposure operate on both the extensive and the intensive margin of donations. For example, white residents of counties with an exogenously larger Haitian minority are more likely to donate to causes in Haiti (relative to their donations to other countries in general), while white residents of counties with an exogenously larger Dominican minority are more likely to donate to causes in the Dominican Republic (relative to their donations to other countries in general). The estimated effect is positive for all recipient countries across our two datasets – several dozen countries in total. Moreover, the positive effect of exposure is stronger for ancestries that are genetically more distant, which we interpret as suggestive evidence that the effects of contact may be largest for populations that look most visibly ‘foreign’.

Figure 3 Effects of exposure to local population on donations to ancestral country


Taken together, our results indicate that long-term contact – across all types of everyday interactions over the course of decades – makes the majority group less explicitly and implicitly prejudiced, less politically hostile, and more altruistic toward minority immigrant groups. Our work thus speaks to a literature on the historical and present-day effects of immigration and intergroup contact (Sequeira et al. 2017, Bazzi et al. 2018, Tabellini 2019) and to a literature on prejudice reduction more broadly (e.g. Paluck et al. 2021, Bursztyn et al. 2020). Yet our paper also suggests several directions for further research. For example, are the positive effects of exposure on altruism muted, or even reversed, when local economic conditions are poor and out-groups may be seen as competitors for scarce jobs? And how do horizontal and vertical transmission of beliefs about immigrant groups – for example, transmission from neighbour to neighbour or from parents to children – mediate the effects of exposure? We hope that as both academic and policy interest in these topics grow, and as new fine-grained data become available, researchers will examine these important questions.


Allport, G W (1954), The Nature of Prejudice, Oxford.

Bagues, M and C Roth (2020), “Interregional Contact and National Identity”, Working Paper 3753981, Social Science Research Network.

Bazzi, S, A Gaduh, A Rothenberg and M Wong (2018), “How intergroup contact can foster nation-building”, VoxEU.org, 7 January.

Burchardi, K B, T Chaney and T A Hassan (2019), “Migrants, Ancestors, and Foreign Investments”, The Review of Economic Studies 86(4): 1448–1486.

Bursztyn, L, I Haaland, C Roth and A Rao (2020), “Disguising Prejudice: Popular Rationales as Excuses for Intolerant Expression”, NBER Working Paper 27288.

Bursztyn, L, T Chaney, T A Hassan and A Rao (2021), “The Immigrant Next Door: Exposure, Prejudice, and Altruism”, NBER Working Paper 28448.

Dahl, G B, A Kotsadam and D-O Rooth (2020), “Does Integration Change Gender Attitudes? The Effect of Randomly Assigning Women to Traditionally Male Teams”, The Quarterly Journal of Economics.

Lowe, M (forthcoming), “Types of Contact: A Field Experiment on Collaborative and Adversarial Caste Integration”, American Economic Review.

Mousa, S (2020), “Building Social Cohesion between Christians and Muslims through Soccer in Post-ISIS Iraq”, Science 369(6505): 866–70. 

Paluck, E , R Porat, C S Clark and D P Green (2021), “Prejudice Reduction: Progress and Challenges”, Annual Review of Psychology.

Rao, G (2019), “Familiarity Does Not Breed Contempt: Generosity, Discrimination, and Diversity in Delhi Schools”, American Economic Review 109(3): 774–809.

Schindler, D and M Westcott (2020), “Shocking Racial Attitudes: Black G.I.s in Europe”, The Review of Economic Studies 88(1): 489–520

Sequiera, S, N Nunn and N Qian (2017), “The immigrants who made America”, VoxEU.org, 17 May.

Tabellini, M (2019), “Gifts of the immigrants, woes of the natives: Lessons from the Age of Mass Migration”, VoxEU.org, 25 May.



Topics:  Migration Politics and economics

Tags:  prejudice, hostility, immigrants, intergroup contact, Arab-Muslims

Professor, Department of Economics, University of Chicago

Professor of Economics, Sciences Po; CEPR Research Fellow

Associate Professor of Economics; Research Affiliate, CEPR; Faculty Research Fellow, NBER

PhD student in Economics, Harvard University


CEPR Policy Research