The writer is a science commentator

InnerCity Weightlifting, a non-profit gym in Boston, Massachusetts, recruits personal trainers from deprived backgrounds to get rich clients into shape. The unlikely pairings, aimed at steering youths away from a troubled life on the streets, produce more than just honed biceps: affluent gym-goers have spontaneously offered trainers extra job opportunities; shown up in court to support them; and even paid for their children to attend summer camps with their own more privileged offspring.

Those kinds of relationships across the class divide are critical for economic mobility, according to Raj Chetty, professor of public economics at Harvard University. Last week, Chetty published two papers in Nature demonstrating that children from low-income families who have lots of friends from high-income families are themselves more likely to earn higher incomes later in life. The analysis points to the importance of “economic connectedness” — in short, having rich acquaintances — as a valuable early rung on the economic ladder.

Chetty, a pioneering economist who uses big data to illuminate complex social questions, set about unpicking the idea of “social capital”, the loosely defined quantity that captures a person’s social network and environment which is often deemed the key to success. Along with colleagues from Stanford University and Stern Business School in New York, he turned to the world’s biggest hub of social connectedness: Meta, formerly Facebook.

The team combed through the anonymised data of all users aged 25 to 44 across the US, covering 72mn people and 21bn friendships. In the first paper, they sliced the data using three different definitions of social capital: cross-connections within friendship networks, as a proxy for social cohesion; levels of volunteering in neighbourhoods, as a proxy for civic engagement; and economic connectedness, a share of a person’s high socioeconomic status (SES) friends.

It was the last factor, particularly the number of wealthy connections made when young, that characterised upwardly mobile high earners. Rich children enjoy this advantage almost from birth. The average child growing up in a high-income family has seven out of 10 friends from wealthy families; for kids in low-income families, the figure drops to four in 10. “If children with low-SES parents were to grow up in counties with economic connectedness comparable to that of the average child with high-SES parents, their incomes in adulthood would increase by 20 per cent on average,” the authors conclude.

Exactly why growing up with affluent pals can boost later earnings is unexplained, though such connections plausibly open a child’s eyes to further education or more lucrative careers. Economic connectedness also appears to explain racial disparities in economic mobility. The team has launched a Social Capital Atlas, revealing levels of social cohesion, civic engagement and economic connectedness for every school, college and neighbourhood in the US. 

The second paper looks more deeply at how individuals connect across the income divide. Two factors contribute: the first is exposure, or whether rich and poor share the same space. But exposure alone is not enough; children in the same school can self-segregate along racial or class lines. Cue the additional factor: “friending bias”, or the tendency for kids from similar backgrounds to cluster socially. This is generally lower in religious institutions and recreational groups, such as the Boston gym, than in local neighbourhoods and schools.

Sir Richard Blundell, professor of political economy at University College London and also at the Institute for Fiscal Studies, said Chetty’s findings, though US-focused, had significant implications for levelling up in the UK: “The main takeaway is that being able to connect with people from higher SES groups has a strong and important association with upward mobility,” Blundell said. “If we’ve got communities that don’t have that connectedness then it obviously can’t be helping their social mobility.”

One obvious implication is that increasing the number of grammar schools — as Conservative leadership candidates Liz Truss and Rishi Sunak have pledged to do — may be incompatible with the levelling-up agenda. Grammar schools, as well as fee-paying schools, tend to separate more affluent children from their less well-off peers. If rich and poor need to mingle more, the truly radical path to wider prosperity may mean tearing walls down, not building new ones.