Graduate returns, participation and ability composition

We investigate the hypothesis that higher education expansion in the UK has induced changes in graduate returns, operating through an omitted-ability composition effect. We look at both the average return and the difference in the return according to degree class achieved. Using a variety of complementary datasets for individuals born in Britain around 1970 and aged between 30 and 40, we estimate a premium for a ‘good’ (relative to a ‘lower’) class of degree of 7% to 9%. Our estimate of the premium for a ‘lower’ degree class (relative to A-levels) of 11% at age 30 indicates a wide spread around the average graduate premium according to academic achievement. We also estimate the premium for a good relative to a lower degree for different cohorts (ranging across those born between the late 1960s and 1980) and find support for the hypothesis that a rising premium reflects expansion-induced changes in omitted-ability bias.