A little over a month ago, I was on Twitter, (@HowShouldWeVote – please have a look), and this tweet came up on my timeline:

Now I think that it was a bit over the top – and the Bregretters, while they are undoubtedly there, are hard to substantiate (see my previous post on Bregret) and they are almost certainly not that many at the moment. However, I picked up on the one point raised in the text of the tweet and forwarded it on with this comment:

Oh, what a little can of worms that opened!

The Reaction

Every comment imaginable came out of the wordwork – most saying what anyone with any experience of debating Brexit on Twitter will recognise: “You lost!” “Get over it!” “Wishful Remoaner thinking” (I’m paraphrasing a little here). If you stick to the more reasoned and polite comments, there was the clear tendency of Remainers to believe that it was true and of Leavers to believe that it wasn’t.

I suppose I was a little guilty of that as well as I didn’t bother to check (or ask) about the sources or calculations behind the figures. And some of the polite Leavers I debate with, pointed this out. They then showed their own preference by saying “even if there was such a trend it wouldn’t impact on the result for more than 10 or 15 years”; and “young voters change to old voters and they’ll change their views.”

I personally noted that there is no evidence of the second point and that there really does seem to be a huge generational difference on this. However, I had to concede that I didn’t actually know the truth of this supposed trend. Well, I’m an analyst by trade (amongst other things); so I said I would find out.

The Objective

With any analysis, it is important to keep things fixed apart from the variables that you are analysing (at that stage). So, for this, I was only going to analyse the demographic trend. As I said in my original tweet, I would ignore every other item and just focus on the deaths of existing voters and new voters coming of age. Using that I would then investigate how this would impact on the country’s overall view of Brexit.

It is important to note that I am assuming that there is no Bregret here, even though there is growing evidence of this. Not a huge amount (I’d estimate it to be currently about 4-5% of the Leave vote – and a larger proportion who are still hoping it will turn out ok yet are very concerned) but some. However, that’s a debate for another day.

I’m also assuming that there is no change in the way the various age groups behaved and that the same proportions would vote and that they would vote the same way. Again, that is debatable and I personally think there would be a higher turnout in the younger age categories than there was; but we are keeping these assumptions static and just looking at demographics.


Before we move on, I should address the accusation that discussing death in this manner is distasteful. I don’t actually see that I’m afraid. This is a purely analytical exercise of something that is going to affect us all for decades to come. If there is a significant generational imbalance then it should at least give us pause for thought. But the exercise in itself is just numbers – an abstract analysis that is no more distasteful than the job of an actuary. Someone has to work out these things that affect us all; but I understand it’s not everyone’s cup of tea.


Any reputable analysis should be 100% open about the sources used – and, ideally, they should be from a third party who has no interest or connection with the analysis being done. My sources therefore are:

While I think the use of the ONS data is fairly indisputable, others may challenge the two surveys that I’ve used. I don’t think any other ones produce anything materially different – from what I have seen so far – but I’m always open to new information. So if anyone believes there is something that I should use instead, please get in touch.


I’m not going to go through the detailed calculation here (it would undoubtedly bore most of you) but for those who really do care, here’s the workbook.

The main challenge was getting the data into a comparable format as I was using two surveys covering different age groups and had ONS data in different formats covering different years; and it all needed to be brought together. I think it has been done sensibly; but, again, I’d welcome feedback.


Purely from demographics, the Brexit majority disappears in 4 years after the vote (so just over 3 years from now).

So it’s double what that initial table said (I have no idea where it took its figures from); but it’s far from being way into the future. In fact, given we seem to be heading for a long period of transition out of the EU, we will still be heavily involved with it when the majority of voters (assuming no other changes) now want to stay.

In addition, since there is no indication that this trend is going to do anything other than continue, this situation is only likely to become more pronounced. This would indicate that we are committing ourselves to a course of action that – once it is completed – is not what the majority of people want!


This movement won’t show in polls.

I’m sure people will say that’s convenient; but I believe it to be accurate. It’s because polls need to be balanced  and weighted to make sure that the sample used reflects the population at large. Therefore, any poll asking about a general election or a referendum will make sure that the final sample used to show the current opinion is weighted to ensure that the sample reflects the nation as whole. If the underlying strucutral base of voters has fundamentally changed though – i.e. producing a 50:50 split with no changes of opinon – it will be rebalanced, ignoring some Remain until you get the 48:52 split of how the nation voted at the referendum. Because of this, Leave voters will be over-estimated and so you will not see this demographic change.

This is my understanding of how polls weight their samples. They simply do not expect huge generational changes. However, I am not an expert and if anyone has more knowledge of this, or would like to provide some evidence to show that I am wrong, I will happily listen.


I did this exercise because it hadn’t been done (or at least I hadn’t seen it done, which I appreciate is not the same thing); but the result is clear. Demographics – on their own – are likely to remove the mandate to leave before we have finally left. Perhaps that is just unfortunate; but on a practical level, it is just storing up trouble for the future. Making a major change that the majority of the population is against is likely to lead to us just having to undo it in the future with more cost. Which is exactly why big changes like this normally have supermajorities.

I know that there are a large number of Leavers that want to make the change knowing exactly that. They hope that the extra cost of changing back will dissuade people. I would tell them not to hold their breath. That’s an argument up there with saying “Get over it” “You have to accept it now” “We voted!” while the very essence of democracy is that right to change your mind.

But this is just one drop amongst many. Some Leavers are changing their minds – not many – but they are (virtually none go the other way). The Brexit we are getting is not the Brexit many were promised or wanted – and these are the ones who are still desperately hoping things will turn out ok. In short, we are heading for a future which only a shrinking minority want; and we are constantly being denied the chance to challenge that.

As a democrat – believing we should all be represented and listened to; as a patriot – wanting what is best for our country; and as a businessman – worried about the rising impact on trade, I therefore believe that this analysis helps to show that the question of Britain’s relationship with the European Union is far from settled.

3 thoughts on “Demographics”

  1. Hi Richard,

    Thanks for this analysis! Lots of work went into this and it gives us ‘Reminders’ as I prefer to call Remain voters like myself, hope. I know there are a couple of legal challenges in preparation and I am sure that, since Art50 is reversible, all is not lost. I’ll be in touch via PM once I know more.

  2. The problem is we did a massive survey of people’s opinions last year in something called a GENERAL ELECTION,80%,of those who voted,voted for parties committed to Brexit in their manifestos and the parties that supported halting Brexit(lib dems and SNP) lost ground…….which rather indicates that your research is unreliable.

    RT: Well, the research is all available from the links on the article and all from public sources; which means everyone can see the soruces.

    That line which is trotted out about the General Election is quite a desperate one. It misses completly three key points.
    1) A vote in a General Eleciton is never a ringing endorsement of every single point or policy in a manifesto – it’s more often a weighted decision of which is the least worst. Even the MPs of a party don’t agree with them 100%!
    2) The assertion could only be true under a fair voting system where every vote counts. In our medieval system, that does not happen. The only competition is in each constituency and so many (I think BES estiamted at about 6 million) had to vote tactically to have any influence.
    3) Finally, if we are going to try and judge and take a mandate from the result of the General Eleciton, Theresa May went into asking for support for her vision of Brexit and she lost seats. Yet still she carries on.

Leave a Reply to Magdalena Williams Cancel reply

Your email address will not be published. Required fields are marked *