Finding a sustainable way to fund postgraduate research isn’t easy
James Coe is Associate Editor for research and innovation at Wonkhe, and a senior partner at Counterculture
Tags
Funding postgraduate researchers is expensive. As UKRI’s new report by Pye Tait Consulting on the full economic costs of doctoral training shows, providers are recovering around 46 per cent of the full economic cost of PGR training and supervision.
For those unfamiliar with full cost recovery a university is required to estimate the full economic costing (FEC) of a research project. This is calculated using a method called Transparent Approach to Costing (TRAC). UKRI has a handy quick guide on TRAC but essentially the cost of research should be calculated against three areas
- Directly incurred costs – Things incurred specifically in the delivery of a project
- Directly allocated costs – Costs that are shared across multiple activities
- Indirect costs – Costs that are not specific to a project but make its delivery possible
The point of TRAC is that all providers have a single basis from which to work out the FEC of a project when it comes to bids, assessing the sustainability of their work, and understanding the impact of various resource allocations over time.
Cost recovery is then the proportion of funding needed to do a piece of research that is covered from a source that isn’t the provider’s own resources. Currently, research projects funded through research councils require universities on average to find £3 in every £10 for every project.
As the latest UKRI report demonstrates, the cost recovery for funded postgraduate research students (PGRs) is significantly lower.
It is worth noting that the extent to which universities are using FEC to calculate their own costs for supervising PGR is unclear. This may seem like quite a significant oversight to not measure the actual costs of student recruitment but this isn’t that unusual in the wider student recruitment landscape.
In some ways measuring the aggregate cost of PGR instruction is even harder. Around 20 per cent of PGR is funded by research councils while the remainder comes from a range of self-funding, sponsorship, and collaboration. No one source covers the full cost of teaching so the difference is made up from QR funding, partnerships, and like much else in higher education cross-subsidy by international student fees.
Among those working in research policy across the sector it is widely believed that recruiting PGRs is done at a loss. The obvious and immediate risk is that the future of research policy relies on a pipeline of researchers, while the future of universities requires them to get on a more financially sustainable footing. If those two imperatives are in tension universities will likely do the things that stop them shutting the doors, like recruiting fewer PGR students, irrespective of the wider consequences of the county.
In managing this dynamic providers are pooling their different QR pots to fund doctoral study. The report notes that QR has been variously used for “studentships, supervisor time, training activities, administration, and equipment, and is also on occasion used to support wider activities and investments related to sector reforms.” There is a wider debate to be had on the future of QR but one of its core purposes is to allow universities to invest flexibility into the things they are important to explore areas of research which may otherwise not be funded. This is how the blue-sky and experimental emerges. This is a very different use of QR than simply finding a new pool of funding to cover the basics of what it means to be a higher education institution.
And this is one of the key challenges the report raises. There is no policy appetite to say that all costs of supervision should be covered without universities using their own resources. However, the balance toward universities using their own resources seems to have shifted significantly because of the implied costs of PGR. Universities are investing in the research environment, staff time, and match funding to make their internal PGR ecosystem work. However, the things which are needed to make for a successful PGR experience only become more expensive over time. Time for staff to supervise PGRs properly. Training which is tailored and useful. Specific equipment for specific disciplines with instruction on how to use it. And the wider ephemera that comes with being a good employer and teachers like careers advice, pastoral support, and networking opportunities.
The research demonstrates there are limitations in understanding the true costs of PGRs. Providers reported that TRAC data did not cover the full range of ancillary costs associated with PGR. Providers wish to increase their PGR numbers but their ideal cost recovery rate would be between 80 to 100 per cent. There is no viable financial route to increasing cost recovery to these levels without reducing the overall research cost base and therefore these two ambitions are in conflict. The risk, as implied in the report, is that
If circumstances dictated, some HEPs [Higher Education Providers] – primarily TRAC peer groups A and B – said they would prioritise investment of resources (and therefore PGR student numbers) into subjects with a track record of research excellence. STEM subjects would likely be prioritised over arts, humanities, and social sciences, to an extent, with the strategic driver being the greater impact on the REF (Research Excellence Framework) score and subsequent QR funding, in addition to STEM subjects requiring greater investment in running costs, e.g. lab space and consumables.
This is the core issue. It’s not that the entire PGR ecosystem may fall apart but that the funding arrangements lead providers to make what are perceived as less risky choices. Maintaining similar numbers of PGRs to do similar things in ways to meet the incentives in front of them. The route to breaking this dynamic raised within the report is to recruit more international students which is also recognised as risky. A seemingly insurmountable gap then emerges between what the sector wishes to do and what it can actually afford to do
Ideally, most HEPs would like to increase the total number of PGR students, so long as they can continue to offer the expected levels of service and experience to students, as HEPs feel this may be a way of accessing more funding via QR RDP (although were all HEPs to increase PGR student numbers, QR funding would simply be spread more thinly). Increased student numbers could also be supported through increased funding levels and greater investment by HEPs themselves.
Absent more funding there are only a range of third-rail issues to bridge the expansion-funding chasm. Either reform QR funding to be more directed toward development of PGR which will hurt the wider ecosystem. Fund PGR more generously but in fewer places or disciplines. Redirect organisational resources toward PGR from other areas. Or, and what ends up seemingly the most likely, continue to push for a mismatch of funding pots and programmes that in total make PGR just about sustainable if always seemingly below its collective potential.
There is a curious feature in UK academia in which there is no formal tuition in the form of classes/modules at the PhD level (unlike the USA and other countries), however a ‘tuition fee’ is charged. Ostensibly this is for ‘supervision costs’, however it is doubtful if PIs actually do great deal of formal supervision duties. Most likely they include the PhD student in their own research activities, thus providing a manner of informal supervision (learning through apprenticeship, as it were). The same overlap applies to resources. The university does not buy lab equipment specifically for the PhD student. The student uses equipment that exists in the PI’s lab, so charging a bench fee does not seem defensible. Many software resources (e.g. Matlab) are nowadays purchased by site license, so an additional 50-100 PhD students is not going to increase costs in a noticeable manner. This then brings into question the actual cost of running PhD studentships and whether the statement ‘…recruiting PGRs is done at a loss.’ holds water. @James Coe, I would be interested in your opinion on this.
Morning!
There’s probably three different versions of the truth that universities express to me on that one
1) There is a difference between the actual cost and the marginal cost. So it might be that in the example you have if you distribute the cost of a license between the first 50 students the programme makes a loss but for 51 onwards then it makes a loss. This is overly simplified but it’s really hard to think about at what point a programme might tip into profitability
2) The way that costs are accounted for works across multiple levels all of the time which might mean they make a loss on paper. Given the relatively small number of PGRs compared to other forms of study apportioning global costs will make them appear more loss inducing. Whether that is actually the case us much harder to tell.
3.) Then you are back to our old friend the programme mix. Your example makes sense there might also be some examples where the exact opposite is true where there is a one for one cost in new equipment which balances it out. This might even exist within the same family of programmes.
It’s a really good provocation on whether the perceived wisdom is correct. I think we can be confident it wouldn’t make sense to look at PGR recruitment as the cornerstone of a fees income strategy but beyond that the evidence depends on internal ability to follow cost and expenditure, which we then know is variable.
Say the PGR fee is £4400 per year and the supervisor salary plus on costs is £110 per hour, then the fee pays for 40 hours of the supervisor’s time. Supervisions are 1-1 meetings rather than classes – 1 hour a week for 43 weeks goes over budget, and doesn’t include the examination, annual progress reviews, training sessions and other support.
Bench fees are generally for consumables that are used with equipment, rather than the equipment itself? Or the equipment could be hired out if the PGR student wasn’t using it, so the bench fee offsets that?
‘Most likely they include the PhD student in their own research activities, thus providing a manner of informal supervision (learning through apprenticeship, as it were).’ – wow this is light years away from PGR supervision in the Arts and Humanities, goodness…
Even within STEM, there’s huge variation between disciplines in the norms for PhD supervisor time commitment. In pure maths it is not unheard-of for a supervisor to devote an afternoon every week to working 1:1 with a single PhD student at a blackboard (and it is consequently rare for a supervisor to have more than 1 or 2 students at a time). In some lab subjects, the supervisor does little beyond suggesting an experimental programme, leaving any day-to-day supervision to postdocs (their time not being free of course, but not necessarily falling on the university baseline).
It might also be worth to to consider the economic benefit of PGRs’ contribution to the institutional workforce in this equation: how much work do PGRs do which would be more expensive if done by other institutional staff? And how could considering this offset some of the ‘costs’. This does not only include graduate teaching assistantships but things like senior students/assistant wardening in student accommodation, short-term research assistantships which would be difficult to recruit to otherwise, or even voluntary labour in the form of organising the postgraduate research community etc.
Interesting! The complementary bit would be how much additional staffing does the recruitment of PGRs drive? If the answer was that PGRs exist in a sort of light touch community of peers with some occasional oversight then I suspect the contribution would be big. If the existence of PGRs if then a primary driver for lab techs, admin, research staff, and so on, then maybe little. BUT the other way to look at that would be that PGRs produce new economic benefits in the forms of wages and infrastructure by acting as a kind of internal pressure to reallocate resources. There is definitely an interesting HESA question on whether there is a link between increase PGR spend and increase research spend overall, and whether there is any causation between the two.
It’s also worth considering the impact of PGRs on research productivity more generally: 79% of doctoral supervisors say the quality of their own research is improved by being doctoral supervisors. And it would be interesting to see the proportion of co-publications with doctoral candidates.