Two years ago, in the November 2013 Autumn Statement, Chancellor George Osborne surprised us all by announcing the removal of the student numbers cap that had been in place since 2009. At the same time, he abandoned the short-lived high grades regime that encouraged institutions to maximise the numbers of those with AAB+ grades recruited at the expense of those with other A Level grades.
You may have read about it on Wonkhe here or in a report for the Higher Education Academy. The gist of that argument was that, given the relative failure of the AAB+ regime to engineer a price differential between the ‘best’ institutions and the rest, the removal of the numbers cap represented a Plan B: increasing supply so that it would match, and one day exceed, demand for places, thus dragging down the average fee level.
We now have the first tranche of UCAS acceptances data by institution since the removal of the numbers cap. Behind the headline figures (who is up and who is down) the data offers an early insight into how far a market effect in student places is emerging. Over the three years 2013 to 2015, 88 (73%) of the 120 providers on the UCAS list) have increased acceptances; 54 (45%) by 10% or more. At the same time, a further 23% (28 providers) have reduced acceptances over the three years. This offers some early evidence, perhaps, of the kind of volatility that we might expect given an open and dynamic market.
The big picture
While there is growth overall and some large movements by individual providers over the three years, acceptances have only grown overall by 3.9% in the first academic year since the removal of the numbers cap (i.e. between 2014 and 2015), only marginally more than the 3.4% growth between 2013 and 2014. The growth figure for the period covering the first three years since the new fee regime and other effects of the 2011 White Paper 2013-15 is 7% overall, which is unspectacular in comparison to the first three years after the previous fee-hike (2006-7 to 2008-9) when total acceptance numbers grew by 16%.
OK, so there are a few caveats we need to apply before we take this analysis any further. Firstly, the effects of demographic decline affecting the current number of 18-year-olds because UCAS data is effectively just a count of the numbers of school leavers entering higher education; most mature students, for example, are recruited outside the UCAS system and they end up representing around a third of all students in higher education at any one time.
As total numbers of young people fall we would expect percentage increases in the number entering higher education to be more modest. The other key caveat is that entry via the UCAS application route gives an incomplete picture of our institutions and who attends them. The data released in January cover only 120 higher education providers across the United Kingdom, ignoring the many hundreds of specialist HE institutions, colleges and alternative providers that mainly recruit older and part-time students directly. So any conclusions about the impact of the market reforms on the English system have to be tentative and contingent at this stage.
What we can learn from longer and shorter-cycle trajectories
However, we can make some sense of the data if we use a dual analytical lens, focusing both on trend data for the two dominant, distinct provider types – pre-1992 and post-1992 institutions – and on variations between growth patterns over ten years and the three years since 2013. What this reveals is that over the ten year cycle (2006-2015) post-1992 institutions on the UCAS list have been more likely to grow their number of acceptances than pre-1992s, and have also grown to a greater extent; however, analysis of the three-year cycle (2013-2015) show that this pattern of growth trajectory has been reversed; pre-1992s have been more likely to grow their number of acceptances and also grown to a greater extent than post-1992s.
The question, then, is how much of this is attributable to institutions responding to various market signals, and how much may be down to other factors? In some ways, the presence of non-English providers in the UCAS data helps isolate key variables, simply because of the absence of market incentives elsewhere.
Over the last three years, 43 of the 49 pre-1992s (88%) increased the number of acceptances: 27 (55%) by 10% or more, 15 (31%) by 20% or more, and five (10%) by 30% or more. Aberdeen, in Scotland, had by far the largest growth, at 67%; the other four big growers were in England and the 30-34% growth range. At five pre-1992s acceptances fell and one saw no change. Only two had also seen a longer term decline since 2006 (one St Andrews in Scotland, the other Bradford). Both saw larger falls since 2013 than their fall over ten years. Bradford’s acceptances are down almost 800 since 2009 – a 23% decrease.
A much larger proportion of post-1992s on the UCAS list (22 out of 67 or 33%) saw the number of acceptances fall over this period and three others recorded zero change in acceptances. Nine saw falls of 10% or more, six of which were in England. Five of these had recorded positive growth over the ten-year cycle, four between 10% and 20% growth (the outlier was London Metropolitan University, which has fallen 4% overall since 2006). Of the English post-1992s that have experienced falling acceptances over the whole decade, in each case their ‘high water mark’ was in fact 2011, before the fee increase – but also coinciding with the demographic decline in the number of 18-year-olds in society.
Big post-1992 fallers over the last three years (2013-2015) outside the English regime (two in Scotland and one in Northern Ireland) had peaks in 2013 or 2014, suggesting demography may be the major factor. Three post-1992s recorded zero growth over the last three years, one of which, Anglia Ruskin University, had been one of the fastest growers since 2006 (87%). In this case, the data may, of course, be explained by a planned consolidation after rapid expansion.
Among the post-1992s that have grown their numbers, a lower proportion (42 of the 67, 63%) increased their number of acceptances than of pre-1992s. The proportions of post-1992s exhibiting growth of 10% or more, 20% or more, 30% or more are all smaller than the proportions of pre-1992s exhibiting these levels of growth.
Longer term analysis of the trajectories of the nine post-92s in England that have grown by more than 20% since 2013 shows that their acceptances peaked around 2011 and 2012 (possibly due to the new fee effect) and then surged once again only after the cap came off. The single fastest-growing Scottish post-1992 (Strathclyde) saw a steadier upward trajectory over the last decade, uninterrupted by changes to the fee regime; again this suggests that the fee increase and other market factors in England were key, assuming other factors being equal in their effect.
Looking at the three English post-1992s experiencing the highest rates of growth – Coventry (93%), Middlesex (83%) and De Montfort (93% – reveals different patterns of trajectory over the decade, again reinforcing the need for more institution-level analysis to make full sense of what is happening. For Coventry and Middlesex, increased numbers of acceptances formed a steady growth trend, but for De Montfort, the surge in acceptances after the cap was lifted (2014-15) represented only a return to its 2009 peak. Neither fees nor the demographic decline would explain such a variable pattern; in the absence of more information, it is safer to assume that the pattern was the result of the institution making decisions about specific market segments.
What are we to make of all this? What does this data really tell us about the potential for a marketised future?
Overall, the pattern of growth trajectory has flipped: post-1992s grew more than pre-1992s over the ten years between 2006 and 2015; a quarter expanded acceptances by 70% or more and five more than doubled their numbers. But over the three years between 2013 and 2015, most expansions happened at pre-1992 institutions: they were both more likely to expand, and more likely to expand at a greater rate than post-1992s. While 88% of pre-1992s increased their numbers of acceptances, less than two-thirds of post-1992s did so; more than half of pre-1992s increased numbers by 10% or more compared with a third of post-1992s. Various factors may explain this growth at the institutional level, but it seems clear that one constant is the new ability to expand the overall number of acceptances since 2014-15.
Research that myself and colleagues carried out into the impact of student number controls (specifically the ‘high grades’ AAB+ regime) showed that many pre-1992s in England felt pressured to move their student numbers away from programmes requiring lower UCAS tariff points to those – usually STEM and medicine-related subjects – requiring higher grades. This was seen as a threat to humanities, modern foreign languages and social sciences in some of the most prestigious universities (where ‘high grades’ are not always required) and some of the rise in acceptances by pre-1992s may well be taking the form of a rebalancing of these numbers in what remain high-demand courses.
Changing student acceptance numbers among post-1992s are harder to explain without institution-level analysis but in some cases they will be the result of a strategic programme review followed by a decision to expand overall, shrink overall or (more likely) to rebalance their offer by shifting away from one subject area and into others that may garner more applicants.
This sort of market-reactive behaviour may explain the greater variance in the patterns we see among post-1992s’ acceptances. UCAS acceptances data alone cannot show the extent to which other routes into HE are affected, or whether falling numbers represent numbers lost to pre-1992s, new alternative (or existing FE) providers or merely represent strategic decisions to consolidate or even downsize to enable a shift towards an emphasis on research or higher average UCAS tariff entry requirements; both of which may make more strategic sense to an institution wishing to enhance its league table position.
Three key variables: the market in action?
We do not yet know enough to explain these patterns. However it does seem clear that there are three variable factors in play: the 2011 White Paper reforms (new fee regime, informed choice and ‘high grades’ effects); demographic decline; and the removal of the numbers cap from 2014-15. The ability to grow numbers due to the removal of the numbers cap seems to offer more explanatory power in relation to growth among pre-1992s; post-1992s’ more rapid growth over the 10-year cycle obscures a degree of volatility (beyond what we may expect from the demographic decline) that suggests other market or strategic policy factors have been at play; they have clearly not been able to (or wanted to) expand numbers to the same extent as pre-1992s since the cap came off.
Finally, it seems that there is some evidence that the larger levels of expansion among pre-1992s is linked to both the cap coming off and the exhortations of the White Paper acting on both applicants (aiming higher due to better information) and institutions (expanding where they can from among those with the required entry qualifications).
A very interesting analysis. Two points come to mind: for the Scottish institutions, the influence of the negotiated Outcome Agreements with SFC reflecting Holyrood policy to recruit more students from WP Postcodes and Colleges may be having an effect. The extent of this varies between institutions and is likely to be come greater through time. This might make numbers of entrants through UCAS in Scotland look smaller, when virtually all institutions are full to capped numbers. Secondly, Strathclyde in not a Post 92.
I’d like to see the figures on which this analysis is based. Strathclyde University in Scotland, for example, is not a post-1992 university, but received it’s charter in 1964, which does not fill me with confidence that the classifications used are accurate.
Apologies for the error on Strathclyde. My bad for not fact-checking. The data the analysis is based on is Ucas acceptances data by institution which isn’t broken down by type. I applied the pre- and post-1992 classifications myself mainly to explore variations by type in the part of the sector affected by the market reforms, i.e English institutions. Such differentiation and its impact on widening participation is the subject of much of my recent research. I think the major arguments made about what this data may or may not say about the impact of the market stand scrutiny and anyone can request or purchase the undifferentiated Ucas data to retest the analysis.
This is an interesting analysis, as much for the within group variation as that between groups. It’s possible that the pre/post distinction (assuming the Strathclyde error is the only mis-classification) is becoming less useful as a crude indicator of market position. Capacity and desire to expand may be affected by many factors, including subject mix, space and resources for new buildings, as well as market ranking and aspiration.
The most interesting individual figure is Coventry’s apparently substantial expansion while moving up the rankings in (some} league tables. This is counter-intuitive.