New data show that the class divide in Britain may not be so wide
“BUILDING SOCIAL capital will not be easy, but it is the key to making democracy work,” wrote Robert Putnam, an American political scientist, in 1993. He found that places where people trust and collaborate with others beyond their immediate social circles fared unusually well on a wide range of political and economic measures. His argument has gained traction as satisfaction with democracy has sunk to record lows globally. But Mr Putnam might strike a more optimistic note in his lecture at the Royal Society of Arts on March 24th. He is introducing research by Meta, the social-media giant that operates Facebook, showing that Britons are more “economically connected” than Americans—a measure of social capital.
Meta’s research draws on evidence from 6bn Facebook “friendships” in Britain (a sample size big enough to make statisticians blush) and builds on methods established in America. In 2022 researchers led by Raj Chetty of Harvard University used Meta’s data to show that friendships that cross class lines are associated with greater social mobility and reduced poverty. Meta’s new data set is the first of its kind outside America and paints Britain in a favourable light. Among Britons with low socioeconomic status, 47% of Facebook friendships are with high-status users, compared with just 39% in America. Although Britain is still divided (60% of high-status users’ friendships are with other high-status users), the results might come as a surprise to class-conscious Brits.
To investigate the matter, Meta (together with Behavioural Insights Team, a research consultancy) uses machine learning and Facebook’s extensive data to label users as as having low or high socioeconomic status, then calculates the share of low-status users’ friends who come from the high-status group. Meta finds that this index of economic connectedness is associated with a striking range of outcomes, including greater trust in other people, higher self-reported happiness and higher incomes among adults from lower-status backgrounds.
Although Britain is more economically connected than America on the whole, Meta’s data also show big geographic variation. Some of this can be explained by demography. Most friendships are local, so economic connectedness is necessarily lower in areas with fewer high-status users. Many cross-class friendships are formed in schools and universities, so areas with more educated residents tend to be better connected. Rural neighbourhoods, areas with fewer immigrants and those with more old people tend to be more economically integrated.
But even after accounting for these factors, analysis by The Economist finds that geography is the single best predictor of economic mixing. In particular, the legacy of heavy industry appears to cast a shadow over the structure of friendships in Scotland, Wales and northern England. Overlaying the locations of historical coal mines on the map of connectedness, for example, reveals a conspicuous relationship (see map).
These patterns pose a challenge for researchers. If economic connectedness is bound up with the legacy of historical industry, it is difficult to accurately estimate its true effect on incomes or well-being. Having accounted for local patterns, The Economist was not able to find a significant effect of economic connectedness on the incomes or emotional well-being of respondents to the British Election Study (BES) between 2014 and 2023.
Our analysis found much stronger evidence that friendships between rich and poor shape political attitudes. In the BES, connectedness appears to dampen populist, anti-rich economic beliefs, and to lead voters to be more socially liberal and accepting of immigration. Adding to the evidence for Mr Putnam’s theory, we found that (even after accounting for other variables) respondents in areas with the lowest economic connectedness were six percentage points less likely to agree that “most people can be trusted”, three points less likely to be satisfied with democracy and two points more likely to agree with the idea of a “strong leader who does not have to bother with Parliament or elections”
Although they are statistically robust, these effects are quite small. That could in part reflect Meta’s methodology. While Britain is relatively well-connected between the top and bottom halves of society, the measure used to construct Meta’s index, the divides are bigger at the extremes. Meta finds that people in the top 10% of the income distribution have over three times the share of friends from that decile than those in the bottom 10%.
Like John Cleese towering over Ronnie Corbett in their comic portrayal of the class system, the top and bottom rungs of society might not often look each other in the eye. But for the sake of liberal democracy, it is good to know that Britain is not as economically disconnected as you might have thought. ■