Pay

1.1 The most straightforward measure of the contribution graduates make to the economy lies in the high salaries which, on average, they receive (although, as discussed later not all of the differences in pay between graduates and non-graduates may be caused by differences in education). This pay premium over non-graduates reflects the value firms place on the skills and aptitudes which a graduate can bring to an employer. Some jobs will require specific skills gained from studying for particular degrees. In other jobs, that do not per se require a graduate, graduates’ superior analytical and communication skills are likely to make them more productive than a non-graduate would be. As well as their higher pay, graduates are more likely to be in employment than less qualified individuals.

1.2 Our approach is to compare the average earnings of graduates with an appropriate comparator group – who are similar to graduates aside from not having a degree. For first degree graduates our comparator group is those with two or more A levels, (the typical entry qualification for a degree course), but no higher education qualification.

1.3 Chart 1.1 plots average weekly pay for graduates and the two or more A level comparators. The split by gender and age is illuminating. For men, both graduates and the comparator group benefit from a reasonably steady pattern of earnings growth over their working lives. While graduates’ earnings grow more rapidly their earnings advantage is limited in their 20s.

1.4 The pattern of earnings for young women graduates is broadly similar to men, only diverging after the age of 30 when growth in weekly earnings ends, reflecting a greater likelihood of part-time work and the consequences of a greater time spent out of the labour force. However, the difference between men and women from the two or more A level comparator group is sharper than it is for graduates. This means that the earnings premium for women graduates in their 20s is greater than for men.

Differences in employment

1.5 Differences in earnings between employed graduates and employed people with two or more A levels are only part of the story. There are also in differences in the likelihood of the two groups actually being in employment.

1.6 Chart 1.2 below sets out employment rates for graduates and non-graduates, split by age group and gender. These are the numbers employed in each group as a proportion of all people in the group including those economically inactive, employed and unemployed (but excluding full-time students).

1.7 There are several points to be drawn from Chart 2. Firstly, women are generally less likely to be in employment than men, particularly after the age of 25. Secondly, employment rates for men do not differ substantially between graduates and non-graduates after their early 20s until they are aged over 40. By contrast, employment rates for women differ considerably between graduates and non-graduates in the 25-29 years old age band.

How much of differences in pay reflects differences in education?

1.8 Of course the picture is slightly more complicated than this. Education is only one of several factors influencing pay. Innate capabilities and family background also play a strong role. Graduates are likely to have been more able on average than those with two or more A levels who did not continue to higher education. So part of the graduate pay premium will reflect factors other than education.

1.9 There is a range of research evidence on this topic.4 The broad conclusion from the literature is that education remains by far the most important factor in explaining earnings differences, even when we control for ‘ability’ and other background factors. This evidence suggests that between 60 per cent and 80 per cent of the graduate earnings premium reflects the skills gained through higher education, with the remaining 20 per cent to 40 per cent of the graduate earnings premium reflecting underlying ability and background factors. A range is given because this research evidence on the link between education and earnings is uncertain (with some studies falling outside this range).

1.10 Blundell et al (1997) is a particularly comprehensive recent UK study. Using the National Child Development Study which has tracked a cohort – now in their 30s – since birth allows the effects of higher education to be disentangled from ability or background with more confidence. The results support the 60 per cent to 80 per cent assumption for men but suggest almost all of the pay premium for women and the differences in employment rates for women reflect higher education.

1.11 In the analysis which follows we call the proportion of the graduate pay premium that reflects education ‘alpha’ and assume that ‘alpha’ is 60 per cent to 80 per cent for men. For women we also look at results assuming ‘alpha’ is 100 per cent.

1.12 It must also be stressed that the pay premium between graduates and the two or more A level category is an average across all graduates and all those with two or more A levels. Particular graduates may have benefited more or less from studying for a degree. Similarly particular individuals from the two or more A level category may have forgone more or less from not studying for a first degree than this average.

Variation by subject

1.13 Graduates’ earnings also vary systematically by their subject of study. For example engineering graduates earn more on average than arts graduates. This suggests that earnings are likely to increase more from studying for a engineering degree than from studying for an arts degree.