British politics has become increasingly volatile and fragmented, with no party able to command the substantial backing of the public that was once routine. The results from England’s local elections in May illustrate just how far politics has moved from the two-party dominance that characterised much of the 20th century. And at the centre of this change lies an important element: class.
In recent decades the electoral foundations of democracy have shifted, eroding the class basis that earlier voting habits were built upon. In the 2019 general election, Labour lost many working-class strongholds to the Conservatives, prompting questions about whether a class realignment was taking place. And in 2024, the rise of Reform UK in many working-class areas left some wondering if it is now the new party of the working class.
Our new research examines how the party system has developed across 37 general elections in Britain, from 1885 to 2024. By taking an unprecedented long-term perspective, we situated contemporary debates about the party system within more than a century of social and political change. This provided a historical baseline against which current patterns of competition can be better understood.
Based on a unique new dataset, we combined election results at constituency level with census data, using geographic boundary-matching techniques to track the rise, stability and eventual decline of class-based voting in Britain. Our findings offer important evidence about when – and why – this decline occurred.
The eroding class divide
There were three main findings. First, while there is some evidence of class voting between Liberals and Conservatives in the late Victorian era (1885-1901), this collapsed in the early 20th century. It was replaced by a rapidly emerging class divide between Labour and Conservative strongholds – a gap that crystallised within two general elections of mass enfranchisement in 1918.
Second, this class cleavage remained remarkably stable, showing little change from the 1920s to the 1990s. During this time people living in more working-class areas tended to vote Labour and middle-class areas Conservative.
And third, we show that the dissolution of the class divide is a relatively recent – and abrupt – phenomenon, unfolding almost entirely during and after the New Labour period from 1997.
The evolution of Labour’s support in seats with the most working-class populations can be plotted against seats with the least working-class populations. The first phase of class voting for Labour occurs rapidly over a short period of time.

Author provided (no reuse)
In the space of five elections, spanning 16 years, the difference in Labour support between the most working and middle-class constituencies grew from seven percentage points in 1906 to 34 points in 1922. This marks the start of the second phase.
From 1923 onwards, there were fluctuations in the overall level of support for Labour. Yet despite this, the difference between the most working-class constituencies and most middle-class constituencies remained remarkably stable until the early 1990s. We found a mean difference of 29 points.
After that, however, the third phase of Labour’s relationship with class began. There was a gradual erosion to the class basis of Labour’s support during the New Labour era from 1997 to 2010. The class difference dropped from 31 points in 1992 to 24 points in 2010, then dropped further to just 12 points in 2024. This was the smallest class difference in Labour’s support since 1910, and the party’s lowest share of the vote in working-class constituencies since 1918.
These findings have important implications. Many debates about class and voting patterns centre on whether the narrowing of the gap reflects a gradual process underpinned by social change, or whether it is a more abrupt shift brought about by the actions of political parties.
Our findings suggest that it was not a result of gradual social change. Our modelling identifies statistically those moments when a trend genuinely shifts direction – and we found that the Labour-class relationship remained stable until the 1990s. This is difficult to reconcile with the view that class dealignment originated in the 1970s.
Instead, our results strongly support the argument that the working class-Labour link was fluctuating and following no clear trend during the 1970s and 1980s. Its decline can be traced specifically to the New Labour era.
The pattern for the Conservatives is different. The Conservative-class relationship remained stable until around 2015, before unravelling rapidly in the wake of the Brexit referendum. This was when the Conservatives made dramatic inroads in working-class constituencies. By 2019, the class gap in Conservative vote share had fallen to just eight percentage points – the smallest in the entire 139-year span of our data.

EPA/FACUNDO ARRIZABALAGA
The timing of the decline – beginning in earnest under New Labour and accelerating with Brexit – points to the importance of the choices made by parties, rather than the slow erosion of class identities in the electorate.
Under Tony Blair, Labour moved to the right on issues that traditionally divided the classes, such as embracing free markets and privatisation. He also broke links with trade unions and selected more middle-class candidates. And he actively moved away from class rhetoric, declaring that the class war was over.
Once Labour turned away from explicit class politics, the alignment that had structured British electoral behaviour for most of the 20th century began to unravel. Despite Labour’s return to government in 2024, there is little evidence that these long-run trends have reversed. Support for the two main parties is now less divided by class than at any point since the era of mass enfranchisement began.
