The widely varying economic effects of Higher Education

Posted on

The back of the head of a female student wearing a mortar board at graduation, among a sea of graduates.

When students are asked why they want to attend university, the most common answer is “to get a good job”. Luckily, there is a long and large evidence trail that suggests there is an economically important difference between the earnings of graduates and those of non-graduates – on average. But, averages can deceive. Not all degrees are equal when it comes to earnings in the labour market – and it turns out that the variance is huge.

Along with Professor Yu Zhu, at the University of Dundee, and other collaborators, I have been researching the variation in graduate earnings – across subjects and across institutions.

If you are thinking of spending three years, or more, at University A doing subject Y, as opposed to Institution B doing subject Z, there is no substitute for looking at the data – so our research looks directly at the earnings of all graduates, not just the average.

Subjects might matter because different curricula deliver different skills – some skills are very specific, like Medicine; some are quite general, like Management. Institutions might matter because you will mix with different quality peers and will be taught by different quality teachers.

But the data itself has to be viewed through a lens that makes for realistic comparisons across subjects and institutions. We cannot, for example, meaningfully compare Oxford graduates with those from Oxford Brookes, or Creative Arts graduates with those from Economics, because they are just too different. Here, our optics rely on weighting the data to focus on comparing like with like – both across subjects, and across institutions.

One can compare Oxford with UCL since there are a lot of similarities across their staff and across their students. So too between UCL and Manchester, Manchester and Lancaster, Lancaster and Sheffield Hallam, and Sheffield Hallam and Oxford Brookes. We can compare everywhere with everywhere else through a string of sequential close comparisons.

Similarly, we can, for example, compare between Creative Arts with English, English with History, History and Geography, and Geography with Economics – to build a meaningful comparison between Creative Arts and Economics. There are a huge number of these sequential comparisons to be made but, provided there are sufficient numbers of students who are either side of the links in the chains between them, we can reliably predict the counterfactual earnings of a student who did Y at A had they, instead, done Z at B.

In some labour market roles, it is obvious that understanding the subject matter is king. For example, hospitals insist that their surgeons have studied medicine, and legal services companies like that their lawyers to have been to law school. But what about the skills that English graduates have, for example – how well do they contribute to productivity in the labour market? As a general rule, employers pay more for higher skills relevant to what they do. It is skills that matter for earnings, not qualifications. Employers want to know what you can do (for them), not what you happen to know (about your subject) – unless what you know makes you better at doing what your employer wants you to do.

So – spoiler alert! – if you follow the chain that links Economics graduates to comparable Creative Arts graduates, the earnings differential is huge. Similarly, if you use our methodology to make meaningful comparisons between the earnings of graduates from the elite highly-selective institutions with graduates from the least selective ones, the differences are huge.

Indeed, our research was influential in showing the Augar Review of Higher Education Finance in 2019 that there is a large minority of graduates who do very badly in the labour market – begging the question of whether students could make better choices that deliver what they want to, and could, achieve.

Our analysis allows us to rank subjects and institutions in terms of the value they add to their students. When we do this, we find that our ranking is uncorrelated with the high-visibility rankings that regularly appear in the media. The latter do not provide any help to students interested in their future earnings.

In a few subjects, the government provides additional funding to institutions – overwhelmingly in Medicine, less so in STE (but not much for M for Maths), and not at all for LEM (Law, Economics, Management) or Arts. More importantly, the taxpayer has an interest in all of this because the government also provides large subsidies to low-earning students through the student loan system. In particular, to Creative Arts, Education, and Arts & Humanities; less so to Law and Management, and much less so for Economics. All of this begs the question of whether the loan system can be tweaked to generate choices that cost the taxpayer less. This is a thorn yet to be firmly grasped by policymakers – who are hindered by the independence of institutions.

Our research programme is going in two directions, so watch this space. One new strand of work looks at how degree class matters for earnings. Degree class measures a graduate’s understanding of the curriculum studied. In the data, we find that a first vs a 2.1 and a 2.1 vs a 2.2 have large effects on earnings – but only for those subjects that matter for earnings on average. Third class graduates do particularly badly – for all subjects.

It seems that, if a degree is worth doing then it’s worth doing well, but not otherwise. The fact that graduate understanding of their curricula is important reinforces our view that the curriculum matters for earnings. But we cannot tell to what extent student effort responds to degree class earnings differentials. If it does, then the income contingent nature of loans would undermine student effort in high-return subjects – which seems like an unintended consequence of the desire for the loan scheme to encourage higher education participation among low socio-economic background students.

While money matters, life is not just about money. So, a second strand of work is looking at the effects of graduate status on measures of “well-being”. It could be that Creative Arts students have high well-being despite their low earnings. And that Economics graduates are miserable despite their financial success. Our provisional results suggest these are not the case – rather, we find degrees that generate greater earnings also deliver higher well-being, even controlling for the earnings effect. So, our financial analysis of graduate status probably underestimates their broader effects.

For the future, we would like to understand the role of peer effects in higher education. How much of the financial return to attending an elite institution is due to the ability of fellow students is an important question for the effects of contextual admission policies being pursued by institutions at the behest of the Office for Students.


The opinions expressed by our bloggers and those providing comments are personal, and may not necessarily reflect the opinions of Lancaster University. Responsibility for the accuracy of any of the information contained within blog posts belongs to the blogger.

Back to blog listing