Each year more than 350,000 students start Higher Education (HE) degrees in England at a total cost of around £17 billion paid by graduates in repayments on student loans and the taxpayer (Belfield et al., 2017). This represents a significant investment and has the potential to have considerable implications for the students’ later-life outcomes. Students typically make the decision about where and what to study at university at an early age. This decision is influenced by many things, reflecting the various expected benefits associated with a higher education, but a key one will be the potential impact on their future employment prospects. However, currently there is little evidence about the differential impact different degrees might have on their medium term earnings or employment prospects. Accurate and timely estimates of the relative value of different degree courses are vital to ensure degrees represent value for money for both students and the government.
• This report is the first in a series of reports seeking to improve the information available to stakeholders on the value of different degree courses. The reports are making use of the new Longitudinal Education Outcomes (LEO) administrative dataset developed by the UK Department for Education (DfE), which tracks English students through school, college, university and into the labour market. The current report will provide estimates of the labour market return – measured by earnings and employment 5 years after graduation – to different subjects, institutions and degree courses relative to the average degree.
• Graduates’ earnings and employment prospects are affected both by their pre-university characteristics, such as their ability level or social background, and by the impact of studying a particular degree. As a result, subjects may have very high average graduate earnings simply because they take high-ability students rather than because of the impact of the degree itself. Raw differences between courses can therefore be misleading as to the actual return from doing a given degree.
• This report tries to disentangle the impact of a degree on earnings and employment outcomes from that of student characteristics, providing estimates of the impact of different degrees on graduate earnings. The LEO dataset provides a unique opportunity to do so by allowing us to account for differences in background and prior attainment between graduates who take different degrees.
• The labour market returns to different degrees vary considerably even after accounting for the considerable differences in student composition. Both the subject of degree and institution attended make a considerable difference to graduates’ earnings. All these estimates refer to differences in earnings 5 years after graduation (or expected graduation for dropouts).
• We know there is wide variation in the earnings between graduates from different degrees. Medicine, maths and economics graduates all typically earn at least 30% more than the
average graduate, while creative arts graduates earn around 25% less on average. A large proportion of these differences in raw earnings can be explained by differences in the char- acteristics of students taking these degrees. However, after accounting for these, significant differences in the relative returns to different subjects remain. Once these differences have been controlled for, medicine and economics degrees have returns around 20% greater than the average degree, and business, computing and architecture degrees all offer relative earnings premia in excess of 10% above the average earnings for graduates. Creative arts – which enrols more than 10% of all students – still has very low returns: around 15% less than the average degree.
• These differences in returns are large. By comparison, after conditioning on all other characteristics, degree subject and institution, graduates from independent schools and the top quintile earn around 7% to 9% more than those graduates from the lowest SES backgrounds. Similarly, adding an extra A at A-level increases earnings by around 3%.
• These figures represent the average returns based on the students who take these subjects. There is no reason to expect the returns will be the same for all students; for example, lower ability students would be unlikely to be able to achieve the high returns that we observe for medical degrees (even if they were able to gain access to the course). Indeed we do find evidence that degrees have a different impact on different types of students. Medicine, pharmacology and English have relatively higher returns for females than males. Computer science by contrast is more beneficial for males. Medicine and education have higher returns for students from lower socio-economic backgrounds, while economics and history have higher returns for students from higher socio-economic backgrounds. Social care and creative arts have a relatively higher return for students with lower levels of ability, as measured by their prior achievement.
• There is also considerable variation in raw earnings across institutions. High-status universities, such as the Russell Group and universities established before 1992, typically have higher-earning graduates. These universities however also typically take the highest-ability students. Once differences in the student composition between universities have been accounted for, the variation in returns is considerably reduced, but significant differences remain. Even after controlling for these differences, the traditionally high-status universities such as the Russell Group still provide the highest returns. This analysis cannot distinguish whether these differences result from the differences in the economic value of the skills provided by the universities or the signalling value of having attended a prestigious university. However, recent evidence on the issue of signalling vs. human capital effects of university education has suggested that the latter is important (Arteaga, 2018). Further, from a student choice perspective, this distinction might be less important.
• One of the key contributions of the LEO data including the full population of students is that we are able to estimate the returns to specific courses (a specific subject at a given university). Some of the estimates are imprecise due to small sample sizes. Nonetheless, the variation across courses is striking. The top-earning courses attract a 100% premium over
average graduate earnings, whilst the lowest-earning courses attract earnings that are around 40% below average graduate earnings.
• These findings imply that studying the same subject at a different institution can yield a very different earnings premium. For example, the best business studies degrees have returns in excess of 50% more than the average degree while the worst business degrees have below average returns. These are considerable differences in graduate earnings.
• There is also considerable variation in the impact of different subject choices on the probability of being in employment. The differences do not highly correlate with the differences in earnings. Some subjects appear to increase the probability of being in work and others increase graduates’ earnings conditional on being in work. For women, studying pharma- cology, medicine, maths, architecture, nursing and subjects allied to medicine increases the probability of being employed by around 2-3 percentage points over the average graduate. For men, studying social care and medicine increases the probability of being employed by up to 6-7 percentage points over the average graduate.
• There are also significant differences in the impact on employment between different insti- tutions. However, once again, the earnings and employment estimates look quite different; unlike the earnings estimates, it is not the high-status universities which have the largest effects on employment. In fact, institutions in the ‘Other’ university group appear to improve employment prospects more than Russell Group institutions. However, it should be noted that employment information on graduates who move abroad is not recorded and so these individuals count as not being in employment. This may well be more common amongst graduates from Russell Group institutions, for example.
Chosen excerpts by Job Market Monitor. Read the whole story at The relative labour market returns to different degrees