EU Citizens, Non-EU Citizens, and the Brexit Vote

A widely known stylized fact is that citizens in places with few immigrants are more likely to express anti-globalist views. They may be more likely to think that immigration is a threat to their well-being, or that international economic integration is a problem, or any number of related views. This observation can be found across the advanced industrial economies.

As it turns out, this stylized fact also describes the Brexit vote. Those regions with few non-UK citizens are those that voted most strongly for Brexit.

This is, of course, well-known. But Brexit is interesting because local views of Brexit might be affected by two types of migrants: EU migrants (the “Polish plumber”) and non-EU migrants (American professors, former Commonwealth citizens, and so forth). It turns out that this distinction can matter in complex ways.

One way to visualize this is the following. I estimated a regression that predicts the local level Leave vote share as a function of the size of the electorate (a rough proxy for population), EU citizen share at the local level, and English statistical region fixed effects* interacted with EU citizen share. I then did the same, but with non-EU citizen share at the local level. From this, we can predict the local Leave vote share for different levels of EU versus non-EU population share, and across different regions of England. I’ve separated those regions into four “macro” regions so that the results are easier to visualize.

For each plot, I restricted the predictions to only the observed range of the non-EU/EU citizenship percentage within each “macro” region.

The thing to look at here is cases like Yorkshire. It’s true that places with more non-EU citizens were less like to vote Leave, but also places with more EU citizens were more likely to vote Leave. The same is true in the East Midlands; and yet the exact opposite is true in the West Midlands. It’s really only in the South that the general stylized fact of “more immigrants -> more Remain” is true without reference to whether we’re talking about EU versus non-EU immigrants.

Of course, it’s always possible to show that things are more complicated than some stylized fact—that’s why they are stylized facts! But these regional differences across England might give us some leverage on the social or economic foundations of the politics of resentment, labor market competition, and related factors that drive anti-internationalism.


* I broke up London into Inner and Outer London because it makes an enormous difference.