In another question I asked for an estimate of how British citizens voted in the Brexit referendum, considered separately from the approximately 2% of voters (around 1 million people) who were not British citizens. Information given in an answer suggested that a majority of British citizens who voted in the referendum voted for Britain to leave the EU.

I am now asking how the voting broke down by ethnicity, or more particularly I would like estimates for the following four groups of voters:

  • white British citizens
  • non-white British citizens
  • white people who were not British citizens
  • non-white people who were not British citizens

I am especially interested in the figures for the first group and for the other three groups combined, and whether they might shed light on the following statistics:

BRITAIN AS A WHOLE (i.e. England, Scotland, Wales and Northern Ireland)
Leave           52%   Remain  48% (percentages of voters)
white British   82%     other 18% (percentages of population)

Leave           40%   Remain  60% (percentages of voters)
white British   45%     other 55% (percentages of population)

Voters in London voted 60-40 for Remain. How did a majority of white British voters in that city vote - for Remain or Leave?

  • 3
    Bear in mind that the only official data on this is not broken down by nationality, or ethnic identity, so any answer would have to look at polling data. This tends to get less reliable as the sample size gets smaller, so the regional breakdown of national polling is likely to have error bars on the order of ±10%, plus any systematic polling error.
    – origimbo
    Jan 30, 2018 at 13:25
  • Also, ethnicity is self-reported and largely nonsense anyway, e.g. is "Jewish" an ethnicity or a religious affiliation or a synonym for "white"? What groups do you lump together, e.g. is "British Asian" a thing even though it covers people from a variety of backgrounds and a huge area of the world, and some people use the American definition of "Asian" which includes Han people etc. etc. Even if this data existed it would be meaningless.
    – user
    Jan 30, 2018 at 15:17
  • 2
    Incapability of precise definition does not in itself render notions "meaningless" or "nonsense", especially ones that are in constant use. Consider a notion such as "adult".
    – user9876
    Jan 30, 2018 at 18:37
  • This is not computer programming.
    – user9876
    Jan 30, 2018 at 18:44
  • 1
    @ruffle "adult" has a very clear, precise definition. You don't seem to understand the point you are trying to refute. In any case, the question asks about "non-white" people, without defining what "white" is. Are Jews white, because there are a lot of them in London... Honestly, the question itself sounds pretty racist, trying to arbitrarily divide people up like that.
    – user
    Feb 1, 2018 at 11:52

1 Answer 1


After some digging, the only large polling data set with a recognisable methodology I can find which comes anywhere near your question is this one by Lord Ashcroft, which did its fieldwork in the days after the event. This provides a breakdown of self identified ethnic group into categories of White, Mixed, Asian, Black, Chinese and Other, (as well as a "prefer not to answer" option).

The headline weighted figures are:

White : Remain 47% Leave 53%
Net BAME: Remain 68% Leave 32%

In the absence of other information, this makes it highly likely that your census identifier "White British" group voted Leave. This isn't hugely surprising (the "White" group makes up the large majority of those who voted, so it would be fairly unlikely to obtain the final result we did if it didn't favour Leave).

The tables don't provide a vote by double breakdown by region and ethnicity, but do include cross tables on ethnicity versus region. If we do some junk statistics, we can note that to obtain the observed poll breakdown for London (42%-58%, off by a couple of points from the real result) with an even split of white voters would require something like a 10%-90% split in BAME voters, which doesn't fit with the observed figures nationally. This provides (admittedly methodically weak) support for the hypothesis that white voters in London split for Remain along with their BAME neighbours, but it's not something I'd put up to peer review.

  • 1
    Thanks. I'd found those figures from Ashcroft. If we assume that the white British voters voted for Leave and Remain in the same proportions as the white non-British, and read that 53% as 53.0%, and then only consider the white British, which is very tenuous, we get that the white British 82% voted 53% Leave so the combination of the other three groups voted 45% Leave (to get the overall Leave vote down to 51.6%). Which is interesting, even if based on such shaky methodology. It suggests this: white British, pro-Leave; others, pro-Remain.
    – user9876
    Jan 30, 2018 at 19:13
  • 1
    @ruffle Note unless your source corrects for it, some of those "not White British" will have been ineligible to vote (e.g. EU citizens not from Ireland, Malta or Cyprus), so your 45% is probably an upper bound, which would better match with the polling estimate.
    – origimbo
    Jan 30, 2018 at 23:34
  • Thanks for this - you are quite right. They include around 800,000 citizens of Poland.
    – user9876
    Jan 31, 2018 at 1:35
  • @ruffle The survey did not check if the respondent was British or not, only their self-identified ethnicity. It is likely that the majority of them were British, since relatively few non-British people were allowed to vote in the referendum.
    – user
    Feb 1, 2018 at 15:40
  • 2
    @user It's even more complicated than that; It's perfectly possible to identity as white, be a British citizen, but not identity with the census category of "White British" so none of these figures match up with ruffle's requested breakdowns. See e.g. the ONS discussion on the 2011 "Gypsy or Irish Traveller tick box ons.gov.uk/peoplepopulationandcommunity/culturalidentity/…
    – origimbo
    Feb 1, 2018 at 19:07

You must log in to answer this question.