The impacts of immigration on anti-immigrant sentiment have been a theoretically and empirically controversial topic in sociology. This study aimed to contribute to the literature in a threefold manner: a comparison of the effects of immigration in an immigrant-salient (UK) and non-salient (Japan) context; the use of longitudinal data to more robustly test the causal assumptions of how immigration affects individuals; and an investigation into drivers of heterogeneity in immigration’s impacts (including financial status and left-right position on the role of government in society). Henrich et al. (2010) warned the community of social science that much work largely relies on subjects from Western societies. A decade later, it is difficult to conclude that this situation has improved, and studies conducted in non-Western societies are oftentimes required to justify the choice of the context and its generalisability to other regions of the world (mostly to Western societies). This situation is also applicable to migrant studies: for example, Pottie-Sherman and Wilkes (2017) reviewed studies on the impact of immigration; however, almost all the studies were conducted in Western contexts. As such, we still know little about whether natives in more Western and non-Western societies react differently/similarly to immigration into their societies, and thus how far theoretical perspectives developed in Western contexts are applicable to non-Western societies.
To remedy this deficiency, this study compared Japan and the UK, which while similar in several respects (e.g., economic development, geopolitics), have several key differences in experiences of immigration and immigrant policies (e.g., size and recency of immigration, integration debates and saliency of immigration-issues). On the whole, despite the differences in immigration-context between the countries, increases in the size of the immigrant-group appears to impact natives’ attitudes towards immigration in a similar manner.
The first key finding is that short-term changes in immigration appear to lead to a synchronous worsening of attitudes towards immigration in both contexts (at least at larger geographical units). This provides support for the group-threat theory in Japan and the UK. While the available measures of immigration and anti-immigrant sentiment differ between studies, the overall-effect of immigration on attitudes does appear stronger in Japan i.e., processes of threat may be more active in the Japanese context. This may stem, for example, from those outlined differences in immigration-context: for example, the relatively rapid change in immigration,Footnote 11 over a short period, where processes of ‘acculturation’ may be lagging behind the change, which could be driving this stronger negative impact. Alternatively, levels of segregation may differ between countries, which can cause more negative effects for the group size in Japan than the UK (e.g., Laurence et al. 2019).
The second key finding is that, in both Japan and the UK, drivers of heterogeneity also appear strikingly similar. Those who perceive their financial situation as being worse or worsening experience a stronger negative effect of immigration, while those who perceive themselves are being more financially secure experience a weaker (or absent) effect. Furthermore, and somewhat surprisingly, those who lean further left on their attitudes towards government intervention (employment and inequality) and support for the disadvantaged (e.g. welfare-support) experience a stronger negative reaction to a growth in immigration, while right-leaning individuals have stable attitudes against an increase in immigrant. This is again present in both Japan and the UK.
On one hand, the stronger negative impact of immigration among the more financially insecure is consistent with the group-threat theory, arguing that a larger immigrant-group size is perceived as a threat to in-groups’ resources, and those perceiving they have fewer resources are more likely to experience such threat (as previously suggested e.g., Quillian 1995). Thus, this study demonstrates this relationship longitudinally, robustly supporting the economic competition mechanisms, filling the theoretical and empirical gaps that most studies relying on cross-sectional dataset left. On the other hand, contrary to previous work (e.g., Graham et al. 2009), we observe that it is individuals with more left-leaning views towards a government’s role in society who experience a stronger negative effect of immigration on their attitudes, while right-leaning individuals hold relatively stable attitudes towards immigrants. Yet, this finding is present in both the UK and Japan and exists across several indicators of left-right position on government attitudes. Thus, this point requires further scrutiny.
While immigration does have a stronger impact on left-leaning individuals, a key point of note is that what this does is take the more pro-immigration attitudes of left-leaning individuals in low-immigration environments and reduce them to attitudes similar to right-leaning individuals in high-immigration environments. Right-leaning individuals, however, are more anti-immigration across all levels of immigration. Potentially, these results could be related to floor-effects i.e., conservatives are already more negative towards immigrants, and thus the rise of immigrant group size does not change their attitudes. In contrast, the threat-effects of immigration are only related to those left-leaning individuals who, tend to be positive towards immigrants, but when proximity increases, experience increased perceptions of threats, resulting in the stronger negative effect of immigration.
An alternative possibility, however, is that, on the whole, liberal-leaning individuals tend to value care, pro-sociality, structural reasons for disadvantage, and fairness, which can result in more support for groups that they perceive are in need, or as experiencing structural disadvantage (Graham et al. 2009). Accordingly, more left-leaning individuals may include immigrants within the ambit of whom they consider disadvantaged, and thus deserving of government support, welfare, and policies designed to improve their situations. This may result in more left-leaning individuals being more supportive of immigration. However, when left-leaning individuals actually experience an increasing share of immigrants in their environment, processes of in-group/out-group boundary-definition may increase and begin to align more along ethnic/cultural lines, as per ‘social identity perspectives’ (Hornsey 2008; Tajfel and Turner 1979). Here, when difference is encountered, ingroup identities can strengthen and natives tend to emphasise ethnic differences between their own cultures and immigrants’ culture. Prior to encountering immigrants in large numbers, more left-leaning individuals may be more likely to categorise immigrants based on their socio-political status, as an often disadvantaged and discriminated against group, falling within the parameters of other social-groups in need to support. However, when encounters with immigrants occur on a large enough scale, such social identity processes may enhance the salience of their ethnic/cultural identity over their socio-political status. As such, who left-leaning individuals consider ‘people like themselves’ (within their in-group), who are thus deserving of support, may shift from being inclusive of immigrants to being increasingly exclusive. In the UK, for example, it is only individuals who believe ‘the amount of money families on welfare receive is much too low’ who increasingly perceive that ‘immigrants are a burden on the welfare state’ under conditions of increasing immigration. Yet, if exposure to immigration triggers stronger in-group identities among left-leaning individuals, this may evoke welfare-threats for liberal-leaning individuals, such as when immigrants are perceived to benefit from the governmental redistribution more than they pay-in. This, in turn, may shift liberal individuals towards more negative immigration-attitudes in these contexts. In essence, immigration does not shift individuals further right on these government-issues (confirmed through further testingFootnote 12) but instead appears to decouple the link between left-leaning position and pro-immigration attitudes.
Such findings are not entirely unknown. Czymara (2020), for example, found an increase in foreigner-flow across Europe had stronger negative impacts on attitudes towards refugees on those with a greater willingness to help refugees. However, our posited explanation needs to be clarified in future work. Yet, given we observe these effects in both countries, across several indicators of left-right position, such heterogeneous effects by views towards the role of government could be observed more widely.
We did not find consistent evidence with the intergroup contact theory, possibly because of suppressing effects. Numerous previous studies have shown that the a larger immigrant-share provides intergroup contact between natives and immigrants (e.g., Pottie-Sherman and Wilkes 2017). In Japan and UK, it is possible that increases in immigrant-share increase contact opportunities. However, a larger immigrant-share also increases perceived threats among natives, and the threat effects may be stronger than the contact effects on attitudes towards immigrants. Consequently, we observe significant threatening effects. These speculations indicate that the negative effects of increased immigrant size could be larger if there was no intergroup contact.
This study attempts to explore the impacts of immigrant-share with novel datasets in two contexts, but future studies could benefit from incorporating a new longitudinal datasets. For example, future studies may be required to incorporate different geographical levels (not only prefectures but cities and district) and test threats and contact explicitly across different levels. Because the available longitudinal data do not contain threat/contact measurements, future projects could also include these variables. Questionnaires in such projects may contain, unlike this study, the same wording measurements across different countries, to test the generalizability of theories across further under studied contexts. In addition, ideally, the larger number of countries in different regions (e.g., Europe, Asia, South America) in a longitudinal style can further provide rigorous supports for the effects of immigrant-share on immigrant attitudes.
In sum, the similarity of results in both Japan and the UK is quite striking. These results indicate that the group-threat theory appears applicable to both of the contexts in similar ways, indicating a high-generalisability of the theory across different contexts (although effects may be stronger in the Japanese context). We thus see that, instead of exclusively focusing on contextual differences, which is often seen in social science studies, we may benefit from paying more attentions to similarity of findings across different contexts. To be sure, the negative findings observed may, in part, be tied to the prevailing media−/political-narratives occurring at the time (e.g. Hopkins 2011). In the UK in particular, the period of analysis covers the build-up to the 2016 EU Referendum, during which immigration issues became increasingly salient. Thus, further research, testing these findings across other periods, and potentially other countries, is required to examine how contextually-dependent they might be.
The paper’s innovative approach makes key contributions to the literature. In particular, the longitudinal, fixed-effects approach significantly strengthens our confidence that other processes (e.g., selection) do not account for our findings. In addition, demonstrating longitudinally that immigration exerts different effects for different individuals makes a strong claim that future work needs to account for such heterogeneity for understanding how immigration impacts attitudes towards immigration in society.