We crunched voters’ data from 36 years of elections. Here’s what we found

At every election, politicians seem to obsess over a different group of voters. In 2019, it was northern voters in so-called ‘red wall’ seats. In 2017, it was the ‘youthquake’ voters. In 2015, it was ‘shy Tories’. But were any of these groups actually pivotal?

In a new research paper, we answer that question by looking at the data. This systematic approach shows us how many votes each demographic group actually contributed to a party in a given election. It also helps identify which demographic groups are likely to contribute more votes to a given party in the future.

To do this, we collected two types of data, for every general election from 1983 to 2019. From census data, we collected the size of each demographic group; and from survey data, we got the turnout percentage for each group, and the percentage that voted for a given party. We multiplied these three numbers to find the total number of votes that each party earned from a given demographic group.

For example, if there were a million trade union members in 1983, and 50% of them voted, and 75% of those people were loyal to Labour, then union members would have contributed 375,000 votes to Labour. If Labour went on to get 3.75 million votes in the election, then union members made up 10% of their ‘coalition’.

By repeating this calculation for every election, we can see whether union members are growing or shrinking as a percentage of Labour’s coalition. Of course, these percentages are affected not only by what union members do, but also by what every other demographic group is doing. If Labour lost everyone’s support except union members, then the union coalition contribution would shoot up to 100%, even if union members did nothing different.

So for this step of the analysis, we are only focused on answering questions about the relative importance of different demographic groups. The graph below shows these relative coalition contributions for some of the groups that turned out to be most important in UK elections.

Labour v Conservative: change v consistency

Right away, some obvious patterns emerged. In the 1980s, the most important demographic groups were those related to social class. For example, Labour was getting around half its votes from manual labourers, while around 70% of the Conservatives’ votes were coming from people who owned their own homes.

But these classic patterns began to change over time, especially for Labour. The Conservative coalition has stayed more consistent.

In the 1980s, 97% of Conservative votes came from white people; in the 2000s, this figure was still above 90%. Those in ‘professional/managerial’ jobs (which are the two highest groups in the eight-group scale of social class used by the Office of National Statistics) have become slightly more important to their success, as have the self-employed, while Church of England members have become slightly less important. Homeowners, and people living in England, are still vital elements of the Conservative coalition.

This consistency has helped the Conservatives. After 1997, turnout and loyalty declined among several core Labour groups, notably those living in the north of England. This decline, coupled with the more consistent support of their own groups, has helped the Conservatives win an increasing share of the electorate over time.

However, when we dig into the underlying data, we find that this is only part of the story. The long-term trajectory of the parties depends not just on which groups vote, and who they vote for, but also on the size of those groups.

For example, manual labourers continued to turn out to vote, and even became more loyal to Labour under the leadership of Jeremy Corbyn (as there have been no general elections under Keir Starmer’s leadership, it remains to be seen if this will extend to the new leader). But there was a steep decline in the actual number of manual labourers over the same period, resulting in a lower contribution.

Conversely, there has been a steep decline of Labour loyalty among Scottish voters. But the group is relatively small, and so the contribution of Scottish voters to Labour only slightly declines. Changes among religious groups have also affected Labour. Turnout among Muslim voters has been increasing, from 58% in 2001 to 77% in 2019. Their loyalty to Labour has also increased, from 61% in 2001 to 71% in 2019. Over the same time period, however, the loyalty of Catholic voters to Labour has declined almost 20 points, from 57% to 38%.

Overall, then, we find that the Labour coalition is a large and shifting set of small groups, whereas the Conservative coalition is a small and stable set of large groups. We also find that it is the size of the demographic groups that is driving overall contributions, much more than changes in turnout or vote choice.

Both Labour and the Conservatives should consider whether they are appealing to the demographic groups of the future or the demographic groups of the past

This suggests that demographic changes are crucial for the future electoral success of each party. A lot of what we think of as normal campaign tactics – trying to get your base to turn out, trying to persuade floating voters to come over to your side – are actually only part of the story. They dominate the headlines, but they don’t dominate the outcome. The raw size of different demographics groups is also important.

This means that, from the perspective of the parties, both Labour and the Conservatives would do well to consider whether they are appealing to the demographic groups of the future or the demographic groups of the past.

College graduates and the non-religious are some of the fastest-growing groups. This bodes well for Labour. But a lot of these groups overlap: the college-educated, the non-religious, living in London. Labour is increasingly reliant on those who fall into these categories. This could prove electorally dangerous if those demographic groups stop growing. It could prove politically dangerous if it fundamentally changes who feels represented by Labour.

For the Conservatives, the decline of the Church of England and the increase in Britain’s racial diversity are particularly important trends. Historically, the Conservatives have not done well among Black, Asian and minority ethnic voters, or among Muslim, Catholic or non-religious voters. To continue to assemble an electoral majority, they may need either to reach out to these groups, or to double-down on their existing support among groups such as homeowners and the highest social classes.

From the perspective of voters, our findings show the rigidity of the parties. The continued prominence of a few large groups, especially within the Conservative coalition, is reminiscent of the idea of a ‘tyranny of the majority’ whereby smaller and marginalised groups are consistently ignored. Our findings also highlight the importance of first-past-the-post. Under the UK’s current electoral system, demographics concentrated in ‘swing’ districts are likely to get more attention.

The most recent elections have seen a realignment, away from purely class-based voting to new cleavages based on education, religion and region. This is being reinforced by internal migration to the cities, as well as immigration, access to higher education, and reduced adherence to religion.

Either way, the coalitions are finely balanced. Demographic calculations are likely to remain absolutely crucial for assembling a winning coalition.