AUTOMATION and globalisation have brought drastic changes to Western labour markets. Middle-skilled jobs are disappearing fast. In America, wages for blue-collar workers have been largely stagnant since the 1970s, whereas those for university graduates have soared. Silicon Valley types frequently warn that advances in technology, especially in artificial intelligence, will be devastating for low-skilled workers. One prominent study, by Carl Benedikt Frey and Michael Osborne of Oxford University, estimated that 47% of jobs in America could be automated over the next two decades. The spectre of mass unemployment, along with increasing levels of income equality, has led many policymakers to see investment in university as crucial for economic prosperity.

Governments have plenty of reason to be bullish about higher education. Perhaps the best piece of evidence they have of the wisdom of investing more in universities is the graduate-wage premium—the difference in wages between those with university degrees and those without. In their book “The Race between Education and Technology”, Claudia Goldin and Lawrence Katz of Harvard University point out that this premium fell during the first half of the 20th century in America as universities expanded enrolment, but started rising sharply around 1980. Although the premium has started to level off in recent years, the fact that university graduates still make around 70% more than non-graduates suggests that demand for skilled workers still far exceeds supply.

Yet governments may actually be overestimating the economic benefits of higher education. While universities are places of learning, they are also social sorting mechanisms. Part of the reason why university graduates earn more is because they are brighter and harder-working to begin with. Some professionals, such as doctors or engineers, do require an extensive amount of technical training, but many do not. The fact that humanities graduates, whose courses often have little to do with their subsequent work, tend to earn more if they come from more prestigious institutions suggests that one reason for going to university is to get ahead of peers in the job market. Bryan Caplan, an economist at George Mason University, argues that earning a degree is like “standing up at a concert”. Doing so improves your view of the show, but it comes at a cost to those sitting behind you.

Across the OECD, a club of mostly rich countries, 43% of adults aged 25-34 now have tertiary degrees, up from 23% in 1995. Yet it is not clear to what extent these degrees have translated into economic gains. An analysis by The Economist of American labour-market data finds that since 1970 the share of workers with degrees has increased in virtually every occupation. But in around half of occupations with better-educated workers average wages have still fallen in real terms. The ubiquity of the degree means that for many workers going to university is more of an obligation than a choice. Moreover, university does not suit all learners. Estimates of the economic returns to higher education tend to assume that all students will graduate. In practice, around 30% of students in Europe and 40% of students in America will drop out before earning their degrees. This means that the expected economic returns of a university education for average students are far lower than commonly understood. Governments are right to fret about training future workers, but they should look beyond just universities.