There have been several significant changes to the way higher education (HE) is funded in England over the past 20 years, moving from a heavily grant-based system to a heavily loan-based system. All students can borrow up to £9,250 per year to cover their fees and, on average, around £6,500 per year in maintenance loans to help with their living expenses. Because the loans are income contingent – meaning graduates only repay 9% of their income above £25,000, with any outstanding debt written off after 30 years – the government can expect to write off around half of loans issued. In fact, loan write-offs now account for more than 90% of government spending on undergraduate HE. As a result, the government is much less able to target the money it spends on HE and, instead, the subsidy mechanically accrues to those graduates with the lowest lifetime earnings. Although there are several very good reasons for the government to subsidise HE, this distribution of spending may not necessarily align with the students, or subjects, that the government wishes to prioritise.
This work estimates how government spending is distributed by subject studied and university attended, based on grants and unrepaid student loans (including both tuition and maintenance loans). This has not been previously possible due to data limitations, but we are able to circumvent those limitations using a specifically created linked administrative data set. We estimate the implied levels of spending for each subject area via unrepaid loans and direct teaching grants, noting that in practice this may not reflect the true distribution of spending because universities are likely to cross-subsidise courses that are expensive to teach with courses that are relatively cheap to teach. It is also important to note that this work is not estimating returns to different degrees, and is instead estimating the value of loan repayments, which is of course determined by many factors other than the degree itself, such as gender and prior attainment.
Estimating loan subsidies – formally, the total amount the government issues in loans, minus the discounted present value of all loan repayments made by all students across their lifetimes, all divided by the total value of all loans issued (in other words, the share of all loans the government issues that it expects to write off) – is an inherently speculative exercise. It requires the forecasting of earnings of graduates over a 30-year period and relies heavily on how graduate earnings have evolved in the past, as well as economic forecasts over the next 30 years. Our results are sensitive to these assumptions and therefore should be treated with caution.
Key findings
Our best estimates suggest considerable variation in loan subsidies by subject area. The government only expects to write off around a quarter of the value of the loans it issues to economists, while for many subjects the expected loan subsidy is in excess of 60%. For creative arts, it is around three-quarters. The subject area with the lowest loan subsidy is medicine & dentistry, with around a fifth of loans written off.
Differences in loan write-offs across subjects largely reflect differences in loan repayments, rather than differences in the size of the loans. The size of the loans students are eligible for does not depend specifically on the subject they study, but rather on the length of their course and their parental income. In fact, annual tuition fees are the same (£9,250 per year) for almost all students, regardless of course or institution. Tuition and maintenance debts are treated as indistinguishable by government (i.e. one is not repaid before the other), so our estimates therefore include write-offs from both. This is appropriate since we think of government contributions to living costs during study as a cost of funding HE.
The subjects with the highest loan write-offs typically receive the highest government spend per student. The cost to government is around £11,000 per economics student taking out full tuition fee and living cost loans. The equivalent figure for an engineering student is roughly £27,000, while for a creative arts student it is around £37,000. Despite having a low loan subsidy, medicine & dentistry is still one of the higher-cost degrees to government, at around £45,000 per degree, due to large teaching grants.
The government cost per student also varies a lot by institution type. Because students from Russell Group universities typically have relatively high earnings, the government cost per borrower at a Russell Group university is around £24,000, while for ‘post-1992’ and ‘other’ universities the cost is around £31,000. Again, this is the total cost to government, including its contributions towards living costs, and does not include the contributions of graduates (in fact, total funding received is extremely similar across different universities).
The distribution of spending by subject and institution has been hugely affected by reforms since 2011. For example, the cost to government of providing engineering degrees decreased by around £9,000 per student during this period, while the equivalent figure for creative arts degrees increased by more than £6,000. The government now spends over 30% more per creative arts degree than it does per engineering degree, whereas if the 2011 system were still in place today, it would spend nearly 20% less per degree on creative arts than engineering. Similarly, government spending per borrower at Russell Group universities is around £6,000 lower under the 2017 system than under the 2011 system, while it increased for the ‘post-1992’ and ‘other’ university groups by more than £2,000.
Consequently, the distribution of total government spending on HE is very different today from what it would have been with no reforms. Holding the set of students and their earnings fixed, if the 1999 system were still in place today, around 57% of government spending on undergraduate HE – teaching grants to universities, grants to students and unrepaid fee/maintenance loans – would be going towards students studying science, technology, engineering and maths (STEM) courses, with 30% going towards arts and humanities (AH) students. Under the current system, 48% goes towards STEM students, with 37% going to AH students.
The recent ONS review on the accounting treatment of student loans dramatically affects the impact of different subject areas on the deficit. Under the old accounting treatment, grant spending today counted towards the deficit today, while write-offs from loans issued today only affected the deficit 30 years down the line. Under the new system, expected write-offs from loans issued today count towards the deficit today. Consequently, many subject areas go from adding almost nothing to the deficit today under the old system to adding significant sums to it under the new one. For example, we estimate the deficit impact per cohort of creative arts students will increase from around £25 million to around £1.2 billion as a result of the change. These changes could dramatically increase scrutiny from policymakers concerned about deficit spending today.
Combined with the design of the finance system, the recent removal of controls on student numbers exposes the government to risk of spiralling costs. Until 2014, there were tight restrictions on student numbers to help control costs. Under the current system, total spending and the distribution of that spending both depend on student choices: large increases in the number of students doing degrees that currently result in lower earnings could dramatically increase government costs, especially given recent increases in the costs to government associated with lowerearning subjects.
Lowering the fee cap from £9,250 to £6,000 could give the government more flexibility to target spending and reduce exposure to risk. This change would save the government around £7,000 per borrower due to lower loan write-offs, with most of the savings coming from lower-earning subjects. The money saved from lower loan write-offs would free up funds for the government to target other priority areas more directly – for example, if all of the government savings were put into grants for STEM courses (keeping total long-run government spending constant), the share of government spending on STEM would increase from 48% to 62%. University funding overall would drop under this policy due to lower contributions from higher earning graduates that are not fully replaced by increased grants.
Variable fee caps could also allow the government to regain flexibility in where it targets spending – but there are significant caveats. Reducing the fee cap for non-STEM subjects to £6,000 per year would reverse part of the funding increase that AH subjects have seen in the last 20 years and would reduce the government exposure to big changes in student choices. But it might increase demand for non-STEM courses, or perversely lead to a reduction in funding for STEM due to subsequent reductions in within-university cross-subsidisation.