Fish, chips and Brexit: Restaurant diversity and UK constituency voting patterns in the 2016 EU referendum
Update: the data are available here: https://docs.google.com/spreadsheets/d/1o8OwTf_xy1NyFCBz9OLY5yMsd9fPlpNWW8ozjNIUohc/edit?usp=sharing
When surveying the sunlit uplands of post-Brexit Britain and Northern Irealnd, the UK’s former Brexit Secretary and current Foreign Secretary Dominic Raab assured his hungry nations that post-Brexit, “there will be adequate food.” But what type of food, and how does this relate to the Brexit vote? Can food explain why the UK voted to leave the EU?
Many factors have been considered to explain why the UK voted by a margin of 52:48 in the 2016 referendum to leave the European Union. These include opposition to supranationalism, fears of immigration, levels of education, disinformation, age… the list goes on. Yet to date, no research has looked at the diversity of restaurants in parliamentary constituencies to see whether it relates to the Brexit vote. This blog aims to correct that.
Why might fish and chip shops be related to Brexit?
The argument here is a simple one. The presence of a large proportion of fish and chip shops in a constituency is an indicator of a low level of ‘diversity’, and that lower diversity means greater support for Brexit. The argument is not that the greater the number of fish and chip shops, the greater the Brexit vote; rather, that the greater the proportion of fish and chip shops relative to all other types of restaurant in a constituency, the greater the Brexit vote.
Naturally, there will be other factors which are better indicators of how pro- or anti-EU membership a constituency is; my interest in this blog is to test whether fish and chip shops have any relationship to the Brexit vote at all.
How do we find out whether fish and chip shops are related to Brexit?
Perhaps not surprisingly, there is no current data set listing the proportion of fish and chip shops in each UK parliamentary constituency which we can compare with the level of Brexit support. Well, there wasn’t, but there is now.
How were these data gathered? Well, the method to capture the proportion of fish and chip shops in a constituency is two-fold. First, we need to find a list of all of the postcodes in all of the 650 UK parliamentary constituencies. This is relatively straightforward; there are several sources of these data. Then, we need to feed these postcodes to an online source of restaurant information. The website which gave the best representation on a national level is the online version of the Yellow Pages: yell.com.
Doing this across the UK, we find that there are 57 main different types of restaurant. Of these, we can determine which is the most dominant in each parliamentary constituency. As is shown in Figure 1, six types of restaurant are most dominant: fast food, fish and chip shops, Indian, pubs, Italian and traditional restaurants.
As can be seen, of the six most dominant restaurant types, there is clear regional variation and clustering. Fish and chip shops are dominant in coastal regions in the south, but also across large parts of the north of England. A belt of Indian restaurants is identifiable from the south-east of England through to the midlands. Fish and chips are predominant in Wales and Northern Ireland, but in Wales we also see a lot of pub restaurants, and in Northern Ireland, “traditional” restaurants are also notable. Scotland likes fast food more than the other nations.
So how does this relate to the Brexit vote?
Of these six type of restaurant, there does seem to be something going on with the Brexit vote. Table 1 breaks this down by the four UK nations, with UK aggregate data on the bottom row.
As we can see in Table 1, there is a positive relationship between the Brexit vote and the proportion of fish and chip restaurants across the UK; this is most pronounced in England. Figure 2 gives a scatterplot of this relationship.
Stepping away from fish and chip shops, the relationship is reversed when we look at Italian restaurants. This becomes most pronounced in Scotland and Wales, as is show in Figure 3: the greater the proportion of Italian restaurants, the lower the Brexit share.
Ok, so based on the above, it looks like:
- there is a positive relationship between the Brexit vote and the proportion of fish and chip shops in a constituency
- there is a negative relationship between the Brexit vote and the proportion of Italian restaurants
But ‘looks like’ isn’t good enough to keep political scientists happy. We need to run some models to see how strong the type of restaurant is as an explainer of the Brexit vote.
Modelling Brexit and restaurant type
To test whether the proportion of fish and chip shops in a constituency remains significant when we consider other factors, we need to add some variables into the mix. We have quite a lot of information on UK constituencies, so we can plug these in quite easily. I’m going to start with population density, levels of home ownership, and the vote share for the three main political parties. The results are presented in Table 2.
As can be seen, the proportion of fish and chip shops remains significant across the models. It would appear that the proportion of fish and chip shops in a constituency is a good indicator of that constituency’s Brexit share.
One more variable we should consider is the percentage of consituents who have a university degree, as this has been found to be one of the most significant indicators of Brexit vote. Table 3 repeats the models from Table 2, with degree added.
Table 3 shows us that university degree is negatively significant across all models, but the proportion of fish and chip shops loses significance in the Liberal Democrat model. University degree, then, demonstrates a stronger (negative) relationship with the Brexit vote than the positive relationship for fish and chip shops.
So there we have it. There is a positive relationship between the proportion of fish and chip shops in a constituency and that constituency’s Brexit vote: a greater proportion of fish and chip shops means a greater Brexit vote. The proportion of fish and chip shops isn’t quite as robust as some other variables, such as the percentage of constituents who have a university degree. Nevertheless, the relationship is there and should not be discounted.
As for why this is the case: that’s harder to determine. Clearly there’s nothing in battered fish that would make people more leave orientated. But perhaps the dominance of fish and chip shops in a constituency is in some ways an indicator of that constituency’s diversity or cosmopolitanism.