Foreign students are one of the most important types of immigrants. Across the world, there are nearly 5 million foreign students, and by 2025, OECD estimates predict that figure to rise to 8 million (ICEF, 2017). Many countries in North America and Western Europe welcome hundreds of thousands of foreign students each year (OECD, 2018). In the UK, for example, more immigrants moved to the country in 2018–19 to study than even to work (Sturge, Reference Sturge2019). Yet, amid these sizable numbers, many countries have witnessed strident anti-foreign-student calls. In the UK, a writer in one prominent newspaper refers to a “culture of hostility toward international students” (Paton, Reference Paton2013). In the USA, scholars identify “a backlash movement against further international student growth” (Miller-Idriss and Streitwieser, Reference Miller-Idriss and Streitwieser2015). The Washington Post editorial board (2019) says that America now sends a clear signal to foreign students: “get lost.” Even outside North America and Western Europe, foreign students have raised the ire of public officials. For instance, several Australian universities have restricted foreign student numbers (Ross, Reference Ross2019). Singapore has also capped its foreign students (Tan, Reference Tan2011).
Despite the controversy they provoke, the determinants of policy preferences over foreign students have received little attention.Footnote 1 Extensive research analyzes attitudes toward high-skilled immigrants (Hainmueller and Hiscox, Reference Hainmueller and Hiscox2010; Iyengar et al., Reference Iyengar, Jackman, Messing, Valentino, Aalberg, Duch, Hahn, Soroka, Harell and Kobayashi2013; Hainmueller and Hopkins, Reference Hainmueller and Hopkins2015; Bansak et al., Reference Bansak, Hainmueller and Hangartner2016; Wright et al., Reference Wright, Levy and Citrin2016; Valentino et al., Reference Valentino, Soroka, Iyengar, Aalberg, Duch, Fraile, Hahn, Hansen, Harrell, Helbling, Jackman and Kobayashi2019). Yet, foreign students are a distinct category. Foreign students are important not only because of their numbers, but also because they are a particular immigrant type that one might expect to be immune to backlash. Overwhelming evidence shows that foreign students enrich host countries economically (London Economics, 2018; IIE, 2019; Kennedy, Reference Kennedy2019). Unlike other immigrants (Scheve and Slaughter, Reference Scheve and Slaughter2001; Mayda, Reference Mayda2006), foreign students should not be perceived as posing an employment threat because they are only in the country to study. Additionally, unlike other immigrants (Hanson et al., Reference Hanson, Scheve and Slaughter2007; Facchini and Mayda, Reference Facchini and Mayda2009; Cavaille and Ferwerda, Reference Cavaille and Ferwerda2017), foreign students should not be accused of abusing public entitlements because they are not expected to rely on social insurance.
In this note, we conduct the first political economy attempt at isolating the types of information that activate anti-foreign-student attitudes. We theorize that even if foreign students neither vie for the same jobs as domestic workers nor cost taxpayers money by relying on public entitlements, there are analogous ways in which they may pose competition and impose fiscal burdens. First, concerns about competition may arise if foreign students are perceived to “crowd out” domestic students for scarce admissions slots at universities. Although empirical evidence on whether foreign students crowd out domestic students is mixed (Borjas, Reference Borjas2004; Zhang, Reference Zhang, Ehrenberg and Kuh2009; Machin and Murphy, Reference Machin and Murphy2017; Shih, Reference Shih2017), citizens may perceive that foreign students take away admissions slots from their children, relatives, and other domestic students. Second, concerns about fiscal burdens may arise if foreign students are perceived to cause “human capital flight.” When foreign students acquire state-subsidized skills but then depart without contributing to the national economy by working, this may be seen as subsidizing the labor force of other countries (Docquier and Rapoport, Reference Docquier and Rapoport2012; Haupt et al., Reference Haupt, Krieger, Lange, Gerard and Uebelmesser2015).
Anecdotal evidence suggests that concerns about both crowding out and human capital flight are salient in public debates. The media and other actors often report on the challenges of domestic students getting admitted to universities due to foreign students taking limited slots. Take, for example, headlines such as “Surge in Foreign Students May Be Crowding Americans Out of Elite Colleges” (Washington Post), “British Undergraduates at Oxbridge Fall amid Concern They Are Being ‘Squeezed Out’ by Foreign Students” (The Telegraph), or “Lucrative Foreign Students ‘Taking Oxbridge Places from State Pupils”’ (The Times) (Anderson, Reference Anderson2016; Turner, Reference Turner2019; Bennett, Reference Bennett2020). The Wall Street Journal (n.d.) reports that “many people worry the influx of international students is depriving qualified American youths of slots in top schools.” Writing in the Times Higher Education, Sir Keith Burnett (Reference Burnett2015) says that “[a]n obsession with cleansing the country of foreigners regardless of their contribution was once seen as a right-wing, crypto-racist issue.” Now, he laments, it is “a feeling in some families ... [that] their children are denied access to higher education because of ‘all these students from overseas’.”
Likewise, the concern that taxpayers partially foot the bill for foreign students—even though those students typically leave after graduation—is often featured in the press and other outlets. Fox News's Tucker Carlson, for example, has demanded that the USA “ceas[e] … subsidizing the education of the children of Chinese elites”: “Our colleges and universities—almost every one of which is supported by taxpayers in the end—educate, at a net loss, the children of the people who are trying to displace us. Why are we doing that?” he asks (quoted in Chapman, Reference Chapman2020). Similarly, one former government official in Canada has complained that “[w]hen an international student comes to a Canadian university …, he or she arrives to a system that has been bought and paid for by Canadian taxpayers. There's no logical reason to subsidize international students” (Rothenburger, Reference Rothenburger2019). “[T]he argument that tax payers are bleeding funding to pay for international students’ free education in Denmark is powerful and speaks into an existing agenda against foreigners that is increasingly dominating the European world,” observes a Danish university administrator (quoted in Smith, Reference Smith2015).
We test whether anti-foreign-student attitudes are activated by concerns about competition (“crowding out”) and fiscal burdens (“human capital flight”), when compared with merely making the topic of foreign students salient and providing basic facts about their numbers. To do so, we conduct a nationally-representative survey experiment in the UK, one of the world's top destinations for foreign students (UNESCO, 2016)Footnote 2 and a country where foreign students have become especially controversial (Parr, Reference Parr2012; Paton, Reference Paton2013; Buchan, Reference Buchan2018). Our objective is to understand how information about the use of a core public service by immigrants affects attitudes toward that service, and how these attitudes can vary depending on how the costs of foreign students are presented. We field an experiment because it enables us to simulate and unpack the kind of information about foreign students that citizens might be exposed to in real-life—for example, from the media, politicians, and activists. It offers insight into how political messaging and communication (Allen, Reference Allen2016; Haynes et al., Reference Haynes, Merolla and Ramarkrishnan2016) shapes public opinion toward policies regulating entry of foreign students.
The UK is a useful test case not only due to its large number of foreign students and the controversies that foreign students have elicited there, but also because UK universities charge among the steepest tuition fees in the world and offer limited financial aid to foreign students (Murphy et al., Reference Murphy, Scott-Clayton and Wyness2017; OECD, 2019). This means that foreign students contribute more to their own educations than in many other countries.Footnote 3 To the extent that our treatments still increase calls for foreign student caps, we should expect these results to be at least as salient in other contexts that provide more generous educational benefits. The UK is also a clear case where attitudes toward immigration have demonstrably affected government policy. As exemplified by Brexit, attitudes toward immigration are a significant force in politics, in some cases even shifting policymaking against the preferences of elites and the governing class (Hobolt, Reference Hobolt2016; Goodwin and Milazzo, Reference Goodwin and Milazzo2017). Public opinion toward foreign students may factor prominently not only in public discussions, but also in the policy levers that elected officials pull in response to constituent demands.
Our results reveal differential impacts of our treatments in shaping public support for foreign student caps. On average, the likelihood of respondents supporting a cap is 53 percent when not provided any treatment, compared to 56 percent when receiving a generic treatment that provides information neutrally about the large number of foreign students studying in the UK. As expected, priming respondents about crowding out significantly increases support for a cap—an 8 percentage point increase compared to neutral information about the large number of foreign students in the country. Priming respondents about human capital flight, however, has a smaller effect in raising support for a cap that is statistically indistinguishable from zero—a 5 percentage point increase over the generic treatment about the large number of foreign students in the UK. This is consistent with an ambiguous connection between perceptions of fiscal burdens and support for a foreign student cap. Additionally, with some exceptions, we find directional evidence that respondents who otherwise are less inclined to support a cap on foreign students absent priming are most responsible for driving this main result.
Our study sheds light on the drivers of public opinion toward policy governing foreign students. A considerable literature in political economy suggests that concerns about immigrants taking away jobs and abusing public entitlements can heighten anti-immigrant attitudes.Footnote 4 Less effort, however, has been made to analyze whether analogous effects apply to different categories of immigrants, such as foreign students, for whom such concerns do not directly apply.Footnote 5 We provide a framework for thinking about the calculations that inform voter preferences that could be adjusted and extended to study attitudes toward other specific immigrant types. Our study also contributes to a growing literature on the political economy of higher education (Ansell, Reference Ansell2008, Reference Ansell2010; Garritzmann, Reference Garritzmann2016, Reference Garritzmann2017; Jungblut, Reference Jungblut2016). Although this scholarship generally analyzes the origins of funding for universities and the redistributive aspects of resourcing tertiary education, we complement existing analyses by examining how citizens react to participation in the sector by foreign students. We show that political controversies can arise over the perceived beneficiaries of the globalization of higher education.
1. The survey
To test the activators of anti-foreign-student attitudes, we conduct an original survey experiment in the UK that primes respondents to think about the large number of foreign students who enter the country, as well as the competition and fiscal burden effects that they might induce. We fielded our survey in the UK in February 2018. Survey Sampling International (now known as Dynata), a global survey company, collected the data online from a panel of respondents who agreed to participate in surveys on various topics. UK citizens 18 years of age and older were eligible to take the questionnaire. Our final figures were nationally representative according to age, sex, and statistical regions of the overall population in the UK.Footnote 6 Our survey included completes for 3000 respondents, from a base of 3505 eligible individuals who started the survey.
1.1 Treatments
Respondents were randomly assigned to one of three main treatments or a control (see Appendix Table A1 for full text of vignettes). The Control group received no information. Treatment 1 (Simpleforeign) informed respondents about the large number of foreign students who attend universities in the UK. This was designed to absorb a residual foreign student effect. Treatment 2 (Crowdout) provided the same information as Treatment 1, but also informed respondents that competition exists for entry into UK universities, in which domestic students vie against foreign students for admission. Treatment 3 (HC flight) provided the same information as Treatment 1, but also informed respondents that most foreign students leave the UK after completing their coursework and take the skills they obtained with them.Footnote 7
1.2 Dependent variable
We derived our dependent variable from the question: “Should there be a cap on the number of foreign students who can study at UK universities?” Respondents could answer either “Yes” (coded 1) or “No” (coded 0). We kept the DV as a clear binary choice between supporting or rejecting a cap on foreign students because it provides the simplest representation of the option to limit foreign student participation and signals that respondents are broadly dissatisfied with permitting large numbers of foreign students into the country.
2. Empirics
We estimate linear probability models to measure the effects of our treatments on support for a cap on foreign students.Footnote 8 For robustness, we also re-estimate our main results using probit regression (see Appendix Table A7).Footnote 9
2.1 Aggregate effects
We first test for an aggregate foreign-student effect by comparing respondents who received any of the treatments to those who received the control. Model 1 of Table 1 reports these results. The coefficient on Any treatment is positive and statistically significant (0.08), suggesting that people are more likely to support a cap when assigned to one of the treatments. As shown in Figure 1, when holding the control variables at their average values, 53 percent of citizens support a cap on foreign students when assigned to the control, compared to 61 percent when assigned to one of the treatments. The size of these treatment effects should be viewed against a relatively high baseline of citizens supporting a cap even absent priming, suggesting less room for movement than on a policy receiving less initial support.
Fig. 1. Predicted probability of supporting a cap on foreign students (with 95 percent CIs), comparing respondents receiving any of the treatments with the control.
Table 1. Marginal effects of treatments on support for a cap on foreign students
This table displays results from linear regression models, with individual covariates as described in the text. Model 2 shows the effect of the Simple foreign student treatment (line 1) and the marginal effects of the Crowdout and HC flight treatments (lines 3 and 5, respectively), over and above the Simple foreign student treatment. Robust standard errors in parentheses.
+p < 0.10, *p < 0.05, **p < 0.01, ***p < 0.001.
2.2 Crowding out and human capital flight effects
Next, we turn to our main hypotheses by estimating the extent to which concerns about crowding out or human capital flight may explain these findings. Model 2 of Table 1 reports these results. Although the Simple foreign student treatment informing respondents about the large number of foreign students in the UK makes people more likely to support a cap, this effect (0.03) is not significantly different from zero. As expected, Crowd out has a large and statistically significant marginal effect (0.09, with a total effect when added to the Simple foreign student treatment of 0.12). HC flight also has a positive marginal effect, but it is not statistically significant (0.04, with a total effect with the Simple foreign student treatment of 0.07; the latter is significantly different from zero). The effect of Crowd out is larger than that of HC flight, and the difference is statistically significant.Footnote 10 Figure 2 plots the predicted probabilities of supporting a cap on foreign students, holding the covariates at their average values. Fifty-three percent of respondents favor a cap when receiving the control. That number increases to 56 percent when receiving the simple foreign student treatment, to 61 percent when receiving the human capital flight treatment, and to 64 percent when receiving the crowding out treatment.
Fig. 2. Predicted probability of supporting a cap on foreign students (with 95 percent CIs), comparing respondents receiving each of the treatments with the control.
Finally, we parse which respondents most contribute to the significant result by stratifying responses to the Crowd out treatment by our standard demographic characteristics, as well as additional political and contextual-level variables (see Appendix Table A9). Holding the control variables at their average values, Figure 3 plots predicted levels of support for a cap across a selection of these key variables.Footnote 11 Directionally, it shows that respondents who have the lowest initial support for a foreign student cap (males, non-whites, non-parents, non-UK-born residents, Brexit remainers, non-Conservatives, and non-middle-aged residents) generally have the most elastic preferences in response to the crowding out treatment.Footnote 12
Fig. 3. Predicted probability of supporting a cap on foreign students (with 95 percent CIs), by respondent subgroup, for those receiving the Simple foreign student treatment (left estimates) and the Crowdout treatment (right).
3. Discussion and conclusion
Foreign students are one of the largest categories of immigrants. In many countries, however, calls for restricting the number of foreign students have grown louder. This is the case even though foreign students yield large benefits for host countries, and despite typical criticisms of immigrants—that they take away jobs and abuse public entitlements—not applying to foreign students. In this note, we test whether anti-foreign-student preferences can be attributed to analogous ways that foreign students are perceived to generate competition and to impose fiscal burdens on taxpayers. In a nationally-representative survey experiment in the UK, we find that priming respondents to consider how foreign students compete with domestic students for finite university admissions slots significantly activates support for capping their numbers. Priming respondents to consider how foreign students impose fiscal burdens by leaving the country after receiving state-subsidized schooling, however, does not significantly activate support for caps. In general, we find that, directionally, citizens who are least supportive of a cap absent priming have the most elastic preferences in response to our treatments.
Our results suggest that different types of priming about the costs of foreign students can have asymmetric effects in activating anti-foreign-student attitudes. Simply framing foreign students in an ostensibly negative way does not automatically lead to greater support for capping their numbers. Instead, public opinion appears to be conditional on the types of information provided to citizens. There may be several potential reasons for our mixed results. With crowding out, for example, citizens may see competition for scarce admissions slots at universities as especially straightforward. Or, citizens may perceive that its downsides for students and families are particularly high-stakes and concentrated. By contrast, with human capital flight, citizens may simply not detect a link between foreign students and subsidizing the labor force of other countries, or educational subsidization in the UK may be too modest compared to elsewhere (e.g., the Nordic countries) to make a difference. Alternatively, citizens may detect a link, but not think that it is a major problem if their primary concern is ensuring that immigrants do not, in their view, take away jobs or abuse public entitlements.
In addition to providing a framework for analyzing attitudes toward a specific category of immigrant, our study may also have broader implications for scholarship on the political economy of education (Busemeyer and Trampusch, Reference Busemeyer and Trampusch2011; Gift and Wibbels, Reference Gift and Wibbels2014). Most analyses, including those on higher education (Ansell, Reference Ansell2008, Reference Ansell2010; Garritzmann, Reference Garritzmann2016, Reference Garritzmann2017; Jungblut, Reference Jungblut2016), examine the determinants of citizen support for education spending and reform. Proposed caps on foreign students, however, have received little scholarly attention, despite being an important policy over which considerable public disagreement exists. Going forward, scholars could probe whether our treatments bring to the fore latent preferences toward foreign students or actually create—or shift—attitudes. Scholars could also test how citizens react to positive, not just negative, information about foreign students. Another question is how perceived cultural threats of foreign students compare to the political economy concerns presented here. Understanding attitudes toward specific immigrant types should be a priority area for research. With foreign students, we found that citizen attitudes can be susceptible to activation.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/psrm.2021.23.
Acknowledgments
This project was generously supported by a grant from the British Academy. The authors thank anonymous reviewers for their helpful feedback on earlier drafts. They are also grateful to Ben Ansell, as well as participants at the 2018 Political Economy of Education workshop at Nuffield College, University of Oxford for their helpful comments and suggestions. The project was approved by the UCL Research Ethics Committee Project ID/Title: 10523/003.
