Has 2020 really had no impact on early non-continuation?

Headline statistics, out today, from the Student Loans Company suggest that student withdrawal rates are low (around 5,000 in total) and in line with previous years.

What we are told is that Covid-19 and the associated measures have had no impact on how keen students are to stay on their course, although given the “irregular” start to 2020-21 (later starts, extended statutory “cooling off” periods, general disruption) conclusions need to be drawn with caution. Specifically, I think the actual figure – for students who were expecting to be on a course this year and drawing down fee and/or maintenance loans and are not – may be a lot higher.

I’d be delighted if this shows us that students have continued to study, learn, and thrive under such difficult circumstances. But definitionally, we need to be careful. For students we can read full time or part time undergraduate students domiciled in England, Wales, or Northern Ireland. “Withdrawal”, on the other hand, is linked to notifications of student withdrawal within data returned by providers to SLC.

What we’re looking at

Though a student accepting a place and registering at a provider makes them eligible for SLC fee loan funding (and, for Approved (fee cap) providers, maintenance funding), for the fee loan funding to actually flow the student needs to be confirmed as in attendance. The date or dates on which checks need to be made by the provider need to be at least two weeks after the start of term because of “cooling off” requirements. The generally accepted deadline to submit confirmation in term one would be the Thursday before the third Wednesday in October.

In 2020 providers could also initially confirm attendance for term 2 by this date – unlocking the advance payment of the second 25 per cent of this year’s SLC fees. Providers have a duty to let the SLC know as soon as they become aware that a student has left their studies (via the “Withdrawal” notification in today’s release) and in such cases will cease further payments of fees (worked out on a termly basis, 25 per cent in autumn, 25 per cent in spring, 50 per cent in summer) and maintenance loans – if the withdrawal date is before the census date for a fee payment then it will be repaid.

If you want to learn more about the complex system of higher education attendance monitoring that drives public reporting and fee payments, I’ve an article for you.

What is missing?

For first year students, there is an alternative coding, for students that registered on the course but never actually arrived: “Never Attended”. You might think quite a few prospective students might have looked at what was happening on campuses and in halls in early October and thought the better of arriving – they’d have this code applied to them. We don’t get this information.

Other students may choose to “suspend” their studies – because of the ongoing Covid-19 chaos or for other personal reasons. This is another route that makes some sense – if we see HE as a rite of passage the temptation would be to delay rather than skip it entirely. Providers don’t get fees with respect to suspended students until they start studying again. We don’t get data on “Suspended” students in this release.

The note on “cooling off” in the text is confusing, but students who made use of their statutory right to walk away from a contract without detriment would not be included in this data, as no money would have changed hands. If a provider decided to extend this period for 2020-21, this could include quite a lot of students.

Finally, in a stressful term, and using a system which is – in the nicest way possible – clunky and old fashioned (SLC got £40m in the Spending Review to update it!) it should not be unexpected that some student records contain manual errors. Maybe the “Withdrawal” code is correct, but the date given for the start of term is wrong (and with term dates moving quite late this may be pretty common). These records are not included in this release.

Looking at what isn’t there

Last week’s Student Loans Company’s 2020 release of “Student Support for Higher Education in England” would include early in-year data for 2020-21.

I’ve spotted a something of a financial hole – providers in total are receiving substantially less per student than otherwise expected when you look at the whole year – income is up this term in raw terms because of the early Term 2 payment, but when you scale that out there looks to be a £900m detriment, which is likely to mostly represent fee income.

It doesn’t really show us this, but by torturing the numbers we do get it appears to be possible that up to 7.5 per cent of full time undergraduate English domiciled students did not see fee loan payments made with respect to their study in the early part of this term – which could mean around 78,000 of this group of students left university before 31 October. This is a load bigger than the 5,040 that SLC have reported.

My figure is riddled with assumptions and caveats (do read on!), but is the first indication that the much predicted attrition – whether you cite the impact of the pandemic or A level grade inflation as the cause – is happening.

What we got

We get three sets of figures for 2020-21, each split by mode, level of study, and domicile (England, and EU students studying in England (Table 7c). The first is a Frankenstein’s monster of a headcount, combining the number of students paid fees or maintenance or anything else by the SLC with the number of students eligible for such payments where the former figure is not available. Even though we get it all the way back to 2013-14 it only really works as a volume measure.

The second set gives you the total amount awarded or paid, using similar rules to the first set. And the third set gives you the average amount awarded or paid – again using the same range of years and caveats.

What I did

So we know the total SLC income with respect to FT UG England students for each academic year, and we have a number for the population of eligible students, for each academic year – including 2020-21 – based on the effective date of 31 October of each academic year (Table 7C).

It’s been fairly reliable over the past few years that around 55 per cent of all SLC payments for this group have been with respect to tuition fees (Table 2). This year, because of the payment of term two fees in term one (doubling the payment) it would be more like 71 per cent of the total.

We can get a “normal” figure for 2020-21 by taking 65 per cent (half of that 71 per cent plus 29 per cent of the original figure for maintenance) of the stated total income with respect to this group. This gives a figure lower than expected (£3,951.7m, the lowest since 2018-19) but not entirely out of line with previous years.

If we assume that eligibility for and use of maintenance loans remains largely proportionally equivalent to last year, and we assume that all providers have taken advantage of the advance fee payment system, we can calculate that the average total amount awarded per student was (on a comparable basis) £3,767 – significantly lower than last year’s average £3,950, and the lowest since 2016-17.

What it might mean

So if SLC are paying out less per student than in previous years what is going on? There’s a number of possibilities we can consider:

• Are students either entitled to less maintenance loans or claiming less maintenance loans? This seems unlikely – in cash terms and real terms the maximum available maintenance loan package has risen this year (Table 7A), and the economic slowdown, furlough, and job losses could reasonably be expected to lower family income on average.
• Are a sizable number of large providers not taking up the offer of receiving the term 2 payments in advance? As far as we know (this stuff is not often spoken about) most larger providers are getting next term’s payment this term.
• Has there been a large growth in uptake of provision that is not eligible for maintenance loans? This is unlikely – certainly the most recent data available (from the UCAS clearing analysis) suggests most growth has taken place in “high tariff” providers, which are large and established universities in the Approved (fee cap) registration category.
• Have less fee and maintenance payments been made at this point than in previous years? This could be linked to the “irregular” start of term, but would also be the case if more students than usual had either “Not Arrived”, “Suspended”, “Withdrawn”, or taken advantage of the providers “cooling off” period.

For the payments per student to be equivalent to last year we would need to have 78,000 less full time undergraduate students in England than are reported – 7.4 per cent of the total population. There is no way to know if this is actually the case, but it is fair to assume that something between the 5,040 marked as withdrawn and the 78,000 (which is an absolute worst case scenario) we would find the number of full time English undergraduate students who were expecting to be on a course this year, and are not.

What are we seeing?

If there is more attrition than usual, there’s two arguments doing the rounds that have explanatory potential – but we can’t confirm that we have proved either.

The first is that, for many students, this first term has been a painful experience far from what they may have expected, and at some point before the end of October some may have given up and gone home. This would be a shame and a waste of potential, but we can at least be grateful that these students would not be liable for tuition fee loan repayments.

The other, slightly less generous, interpretation is that the many additional students recruited after the adoption of Centre Assessed Grades (CAGs) in place of calculated A level results may not have the intellectual potential to keep up with a course that usual would have required higher grades from actual exams. This plays into a lot of dubious academic mythology: about how one day at a desk in the school gym says more about someone than a teacher’s impression based on two or more year of knowledge; how grades say meaningful things about aptitude that interviews and references do not; how university curricular articulate seamlessly to A level attainment.

This application round saw a roughly 14,000 increase in England domiciled undergraduates studying in England between over last year. The early year non-continuation indicated by SLC is substantially below this – obviously in suggesting that the actual number could be higher it is possible that it could cancel out this gain. Again, we don’t know this now, and we won’t have all the public data we need to know this until next summer at the earliest.

A further thought to bear in mind is that, although we’ve never known much about non-continuation in this early part of term, we could be seeing either additional attrition (on top of the 6.8 per cent we’d expect in the KPIs on average), or those same students leaving earlier rather than later. This earlier decision would have a financial impact where fee income is affected, but it would not necessarily say anything about the cohort if numbers overall turn out to be comparable with other years.