Say what you like about economists, but they do like to build models.
Not – I hasten to add – Airfix kits or renditions of the Houses of Parliament in matchsticks. To “do” economics in a way that generates useful predictions you have to model complex systems of interactions in a market.
Wiping the superglue from their hands Jack Britton, Monica Costa Dias and David Goll at the Institute for Fiscal Studies are now looking proudly at a model of the way applicants, providers, and courses find each other – and the long term salary implications for those applicants. They’ve used this to test a number of potential policies to improve intergenerational mobility.
Let’s dispense with the findings quickly:
- Policy 1 is scrapping tuition fees and converting loans to grants for disadvantaged students. The model says this would narrow participation gaps but not have much effect on mobility, and would cost about £1bn each year.
- Policy 2 is increasing cash support for disadvantaged students by £5,000 each. Again, the model says this would narrow participation gaps but not have much effect on mobility, and would cost about £1bn each year.
- Policy 3 is the same as Policy 2 but restricted to disadvantaged students at “high status” universities. The model says it would reduce access gaps at high status universities but would not be effective in improving mobility.
- Policy 4 is another variation on Policy 2, but restricting the grant to disadvantaged students studying a given list of subjects (law, economics, management/business, science, technology, engineering, maths). The model suggests this approach would substantially cut earning gaps between disadvantaged students and their peers.
- Policy 5 is the only supply side reform – enforcing contextual admissions to offer priority entrance to academically high performing disadvantaged students. The model suggests it would “dramatically” reduce access gaps to high status providers, reverse earnings gaps, and cost just £75m a year.
So the best thing for intergenerational mobility would be abandoning institutional autonomy, annoying the Telegraph, and giving admissions priority to disadvantaged students. There’s a lot of people that would argue for that approach anyway, and this IFS paper says they are right because – some numbers.
It’s a model (is it looking good?)
So, as so often in economics we need to critique the model in order to tease apart the findings. Economists do tend to chuck a load of symbols and equations around – in the main because they wish they were theoretical physicists (circa 1950) – but we’re just going to look at what is going in to the model and what has been left out.
This is a “two-sided matching model of sorting” into fields of study and university/provider within the UK higher education system, which takes into account the assumed preferences of students for particular universities, and the preference of universities for particular students.
What’s missing here – well, genuine individual student preferences for subjects is. The model uses some clever stuff around predicting individual skill levels across quantitative and communication axes (using GCSE attainment data!) to predict which of three big buckets of academic disciplines a student will go in to: law, economics, management (LEM), science, technology, engineering, medicine (STEM), or everything else (arts, humanities, social sciences and so on)
The model assumes that student preferences (on provider/subject combinations) are based on observable skills (plus an undefined unobservable skill used in determining which potential students are attractive to providers) plus background characteristics (here limited to parental socioeconomic status, private school attendance, and sex). The latter lot apply only to the choice of provider – and there’s distance and cohort variables (students like providers near them, and providers/subjects that their peers like). All this gets fed in to the latest version of the lifecycle model that we’ve played with before, via – for some reason – the Complete University Guide as a control for course quality.
In reports like this you look out for homo economicus – the mythical man (thank you Caroline Criado Perez) that is able and willing to make decisions solely to maximise future returns. As you slide down the multiple scales of disadvantage this idea holds less and less well – our IFS report does make some allowances for this effect (preference “shocks” to use the lingo), but arguably not enough.
As models go this is an interesting one – certainly one of the best I’ve seen of the complexities of the higher education market. It reminds me of the kind of things we used to talk about when the government periodically attempted to boil all this rich information down to a single sticker price. At least, in other words – people are trying to represent more of this complexity in policy thinking now. But we are along way away from something that can meaningfully test policies as the headlines suggest.