Blog/Analysis

Some voters matter more than others

Certain voter types in the UK are particularly common in swing seats (people who own their home outright, for example), whereas others tend to be concentrated in uncompetitive areas (social renters are disproportionately found in safe Labour seats). As a result, the major political parties tend to target voters in the former groups at the expense of the latter.

However, while some voters are efficiently-distributed geographically, they may not change their vote very often. Voters in the ‘A’ and ‘B’ categories of the NRS social grade system (defined as those with ‘intermediate [or] higher managerial roles, administrative or professional’) are some of the most efficiently-distributed in the country, but are also some of the least volatile.

Parties who want to optimise their election strategy under first-past-the-post would be wise to spend their time targeting highly-volatile, highly-efficient voters. These groups are the key swing voters in Britain – we call their characteristics ‘super-demographics’.

Defining 'super-demographics'

Super-demographics are determined as a function of two variables: volatility and efficiency. Volatility is a measure of how likely each group is to change their vote compared to the average voter. We calculate a group’s volatility by assessing the total net change in vote share for each party among each group over the last three general elections.

If we look at our How Britain Voted estimates of 2024 general election vote by 2016 EU referendum choice, we would calculate Remain voter volatility by adding the absolute value of each party’s change in vote share (6 + 0 + 2 + 4 + 4 + SNP/Plaid/other figures). We then do the same for the 2019 and 2017 general elections and add up the total figure for the three elections, which comes to ~68 percentage points. Finally, we calculate the same figure for the national result. Labour's vote share climbed by 2 points in Great Britain in the recent general election, the Conservatives fell by 20 points, Reform grew by 13 points, and so on. Adding all the party changes over the last three elections takes us to ~91 percentage points. Relative volatility is then calculated by dividing the group total by the national total, so for Remain voters, it would be 68/91 = ~75%, which tells us that Remain voters are less volatile than average.

The efficiency metric is slightly more complicated. We calculate efficiency by determining how many seats would change hands from the 2024 election result if there was a 10-point uniform swing between the Conservatives and Labour (in either direction) among all members of a particular demographic group. For example, in the Isle of Wight East constituency, 46% of voters own their own home, so we would perform a 4.6-point swing (46% * 10 points) from the Conservatives to Labour and vice versa, and see whether the winner would have changed. In this case, the swing would not be enough to change the outcome, but by applying this method to every seat, we add up the number of constituencies which could have a different winner. Finally, we assess how many seats would change hands under the same method if the voter type in question was uniformly-distributed across each seat (so homeowners would make up 32% of the electorate in Isle of Wight East, along with every other constituency). Our efficiency score is calculated as the total number of seat changes among a demographic group's real distribution divided by the number of seat changes under a uniform distribution. An efficiency score larger than 100% means that gains among that particular group will bring greater electoral rewards under first-past-the-post.

Under their current geographical distribution, a 10-point Conservative-Labour swing among homeowners would see 127 seats potentially change hands. If homeowners were instead uniformly-distributed across the country, the figure would only be 115. Our efficiency score for homeowners is therefore 127/115 = ~110%.

The efficiency and volatility of each demographic group tells us how important each is under first-past-the-post. Groups with above-average levels of both volatility and efficiency are super-demographics.

Six groups emerge from our analysis as super-demographics: 2016 Leave voters, those aged 55+, white voters, those without a university degree, voters in the C2 social grade, and those who own their home outright.

As you can see from the above graph, there is a small positive correlation between group volatility and efficiency. This dynamic is a recipe for electoral chaos, and partially explains the story of the last five years, when the country went from one large majority to another in such a short space of time (as outlined in our Sandcastle Politics piece). If the most volatile groups of voters are also concentrated in the most marginal seats, we should expect an electoral system to deliver larger changes in seat numbers than a uniform national swing would imply.

How super-demographics voted

The ultimate super-demographic voter is white, aged 55+, voted to leave the European Union, does not have a university degree, owns their home outright and is a ‘skilled manual worker’. Roughly 2–3% of voters match all of these criteria. From this description, you would correctly expect this voter to be much more likely to vote for a right-of-centre party, but a person does not have to be equally likely to opt for each of the major parties to be an important swing voter.

Super-demographics are key to determining general election outcomes. It is no surprise that – though the Conservative collapse played a major part in propelling Labour to a large majority – Keir Starmer’s party’s largest advances came among voters with higher numbers of super-demographics, despite their overall vote share not changing much from 2019.

The nature of Labour’s electoral coalition means that the party fares better with voters who have fewer super-demographics, but the difference in Labour's vote share between voters with no more than one super-demographics and those with five or more dropped from 49 points in 2019 to 32 points in this month’s election.

Between 2015 and 2019, Labour’s vote share grew slightly (significantly in the case of 2017), but its distribution was better suited to proportional representation – where every vote counts equally – than the current electoral system in which some count more than others.

The Conservatives still won the most recent election among voters with four or more super-demographics, but the swing against them was catastrophic. The governing party went from more than 4-in-5 voters (82%) with 5+ super-demographics backing them in 2019 to just 40% in 2024.

The story of the last 10 years is a story of the Conservatives building up their vote share among the most efficient voters – from 52% in 2015 to 82% in 2019 – before those gains were wiped out in the blink of an eye. Reform UK were the biggest beneficiaries of this collapse, and the party's relative strength among voters with a higher number of super-demographics indicates that they could be a dangerous force for the major parties if their vote share were to increase even further.

The collapse in the Conservative vote share among 5+ super-demographic voters is reminiscent of Labour’s 2024 collapse among Muslim voters compared to its peak in 2017. However, Muslim voters are less efficiently distributed across the country and thus made a significant impact in 2024, but only in a select number of seats.

A stream of losses became a waterfall

If we look at the voters the Conservatives lost between 2019 and 2024, we estimate that a majority have four or more super-demographics.

Among 2019 Conservative voters, Labour and the Liberal Democrats disproportionately gained among low and mid-level super-demographic voters, whereas Reform took large chunks out of the party’s vote share among the higher groups.

Though these super-demographics are politically volatile, their ‘super’ status has been a stable fact of UK politics for at least the last few general elections. Bar major geographical realignments in the next 10 years, we would expect the same constellation of voters to shape the elections in 2029 and 2034. Given the political upheaval of the last decade, though, it would be unwise to assume that anything will remain stable. If overall volatility continues to grow and other groups grow restless, we could see new super-demographics come to the fore.