Does the student number cap incentivise over-recruitment?

I’m here to drink turmeric tea and assess the perverse impacts of draft statutory instruments - and I’ve just finished my tea.

Remember the Department for Education announcement on the 2021-22 recruitment cap, and how one half of the additional student number offer appeared to be aimed almost entirely (and entirely arbitrarily) at more prestigious providers?

Well, they’ve only gone and done it again.

Raising of the cap

The publication of The Higher Education (Fee Limits and Student Support) (England) (Coronavirus) Draft Regulations 2020 announces government intentions to modify regulations made in 2018 (fee levels) and 2019 (accelerated degree fee levels) under the Higher Education and Research Act 2019, and regulations made in 2011 (fee loans) under the Teaching and Higher Education Act 1998.

In essence, it enacts what is on the page of the publication earlier this month:

• If you over recruit English domiciled students by between 0-6 per cent, there’s a 3 per cent reduction in the maximum fee level (England) or loan amount (Scotland, Wales, Northern Ireland for 2021-22.
• If you over recruit by 6-12 per cent, there’s a 9 per cent reduction
• If you over recruit by more than 12 per cent, there’s a 15 per cent reduction.

To illustrate that, I’ve built a little model. The fictional Cow Eye University College has a recruitment cap of 1,000 English domiciled students, and usually charges undergraduates £9,250 a year (for clarity, it only teaches arts and humanities subjects so there is no OfS top up to complicate things). If it recruits 1,121 students it could charge a maximum of £7,862.50 in 2021-22.

Where this gets odd is when you realise these limits apply to all fee paying undergraduates in 2021-22, not just the over-recruited cohort of 2020-21. A cap of 1,000 students for last year means that the English domiciled recruitment for 2019-20 would have been about 935 (this year’s cap is last year, plus forecast - I’ve used the 1.5 per cent default - plus 5 per cent). We clearly know nothing about the 2021 cycle (or indeed the 2020 cycle!) so I’ve put a range of options on each axis. Black cell colouration means that the scenario is a net improvement on steady state (recruit to cap number in 2020 and 2021), red means a detriment.

[Full screen]

The range of fee income scenarios here extends from just over £23m (very naughty in 2020, recruit terribly in 2021) and £28.5m (recruit to cap in 2020, clean up in 2021).

An advertisement for breaking the law

But where we get super interesting is when we add in extra income from 2020-21. If Cow Eye University College overrecruits, it gets to keep all of the fee income for the year. There’s no indication of fines, despite how keen OfS has historically been to talk about fining people for transgressions around VC pay, grade inflation, conditional unconditional, and whatever else is on the front page of The Times that week.

It’s a bit like robbing a bank, and being punished via a cut in your interest rate for the next financial year. You still get to keep your ill-gotten gains.

So here we go. Figures are for the cash position from fee income for 2021-22 (2021-22 fee income plus additional income from 2020-21, again I’ve assumed a recruitment of 935 for 2018 and 2019 cycles).

[Full screen]

So, the absolute best thing a provider can do, financially, would be to over recruit by six per cent in 2020. Unless you are desperately unlucky in 2021, you are better off - by around £250,000 - than if you’d followed the rules in 2020 and recruited up to your cap.

If you are super-confident about 2021 (all those deferrals, demographic growth, being in a prestigious group of providers...) you might want to consider over-recruiting by 12 per cent in 2020. As long as you are at around the same level or more in 2021, you are only experiencing a small detriment - which will probably be cancelled out by the consequential rise in income in 2022.

Unit of resource

Aha - I hear you say - you may be financially better off, but as your income per student has fallen then surely you have less to spend per student? On the face of it, this is correct, but the marginal cost of teaching a small additional number of students is probably less for most institutions than the rise in fee income. Seminar groups will grow slightly, more people will watch lectures online. If there was a direct proportional link (and a constant unit of resource) there would be no point in growing recruitment, as the monetary gain would be directly offset by the growth in costs.

This doesn’t happen at the best of times - in a world where international recruitment is down it is likely to matter far less. Providers will have the capacity to teach more English students than last year, they’ll actually have an incentive to hold on to that capacity so they can make use of it in 2021, so this argument doesn’t apply.

Because TRAC isn’t an open dataset, we can’t model this directly, but put it like this. If you are a larger provider, with a sizeable international recruitment in a normal year, and have experienced year-on-year growth in English domiciled recruitment (or, indeed applications) over the past few years - you should be filling your boots, cap be damned. Now what kind of a provider does that sound like?

Such behaviour would, of course, potentially fall foul of new OfS condition of registration E6. It would work against the sector as a whole - so theoretically the regulator could levy sanctions. But only in England.

Unintended consequences?

DfE won’t be providing an impact assessment for these regulations. In the draft explanatory memorandum it is suggested that:

An Impact Assessment has not been prepared for this instrument because, while the fee reductions are judged to have a financial impact on those higher education providers that exceed their SNC, the direct effects of the instrument will last for less than 12 months”

This is not correct, as I have shown.

But the really important line comes a paragraph earlier:

It is expected that the financial impact of exceeding SNCs will be proportionately greater for more teaching-intensive providers (which have a greater reliance on domestic tuition fee income) and smaller providers (which have smaller and less diversified income streams than their larger counterparts). A similar analysis applies to institutions in the devolved administrations."

This refers to the impact on provider income - if fees from English domiciled recruitment is a bigger part of your income you are clearly going to get a larger financial detriment from a fee level cut. But we also need to think about the likelihood of student number growth in 2021 (which a headline fee cut is only going to help) and the financial cushion required to have a little bit of a gamble.

I recently criticised DfE for the impact of the additional student number plans, which overwhelmingly benefited larger, more prestigious, providers. We see exactly the same effect in the design of the student number cap. Once could be a coincidence, twice starts looking like policy.