The idea of a graduate premium – the ability of a freshly-minted graduate to earn more than a non-graduate peer – crops up all over the place in policy.
On the face of it, the idea is simple enough to understand. The equivalent of three years of sustained study allows for the development of a range of (soft and otherwise) skills that make someone more economically valuable and thus improves their earnings. Or the act of being selected for and completing a university course signals that a young person is among the cognitive elite of their cohort and thus makes them more likely to land a well paid job – the so-called “signalling effect”.
The modern conversation here also elides choice of subject, choice of provider, and academic performance (did you land that first?). And that’s before we get to the structural differences based on background and personal characteristics, and stuff that derives from that like A level performance.
So maybe it isn’t that simple.
The continuing research partnership between economics at Warwick and HESA has been a joy to watch – the papers that have emerged have been superb and nuanced interventions that sit a world away from the headlines and lists. Even if you are new to this end of policy there is a very readable summary from HESA, and the paper (here from Gianna Boero, Tej Nathwani, Robin Naylor, and Jeremy Smith) underpinning allows you to dive in to the plumbing in ways that isn’t always possible for sector agency publications.
If you read graduate pay you might think “LEO” and the issues that this entails – this research instead connects participants in two cohort studies (Next Steps for 1990 births, BCS70 for 1970 births) with the ONS Labour Force Survey (run quarterly). The headline finding – based on a sophisticated model examining earnings at age 25 and 26 – is a ten percentage point difference in the graduate premium between those in the 1980-1988 birth cohort (23 per cent), and those in the 1988-1993 birth cohort (13 per cent).
There has been other research that has identified a tailing off of the graduate premium happening alongside this growth in participation – notably work by Jack Britton’s team at the IFS. This is the first time we’ve been able to look at this effect outside of LEO data for the 1988-93 group – and the choice of a non-LEO methodology means we can compare this with earlier cohorts with some confidence.
More grads, better grades?
More than two decades of sustained policy attention paid to widening access mean there are more graduates in the population – up from 17 per cent in 1992 to 38 per cent in 2013 – than ever before. And this trend is coupled with a less enormous but still significant rise in the numbers of “good” degrees (a first class or upper second class) awarded – in 1996-97 about half of all degrees received were classified as lower second or below, but 2012-13 this was 32 per cent, and by 2017-28 we get to 24 per cent.
The paper suggests that, theoretically:
An expansion in the number of graduates entering the workplace is likely to weaken the extent to which having a degree can be relied upon by graduate employers as an informative selection criterion in the recruitment process … a higher proportion of graduates qualifying with at least an upper second class award will reduce the premium for at least an upper second class award relative to a lower second class award or below
And that theory seems to hold up. In 1990 the earnings premium for a first or upper second was 10 per cent higher than other degree classifications. This is four percentage points higher than in 1970, suggesting that expansion had indeed led to employers using classifications to discriminate between graduate candidates during that period.
The corollary to this might be that the rise in “good” degrees would be expected to lower the salary premium in future. We learn that for the 1990 cohort the premium for a lower second or below compared to no degree was just three percent, suggesting that for the moment doing better on your degree is a good investment in your future. But will that hold?
The earnings premium for graduates has dropped as more graduates have entered the workforce. This is to be expected – a larger and more diverse supply of graduates means that not everyone will land the highly paid “graduate jobs” as defined in the 70s. However I would argue that it has dropped less than you might expect. The effect is only first seen for those born in the late eighties and early nineties – a long time after the expansion in participation that characterised that period.
There is a cohort effect here – this is that generation that started work after graduation in the years around 2010, just in time for the effects of the 2008 financial crisis to take effect. This led to a decade-long plateau in wage growth, it is only this year that growth in pay has returned to pre-2008 levels. It would be ridiculous to blame graduates or the sector for these changes to the economy and job market.
Fundamentally, employers do want graduates. A decade of stagnation has allowed wages to become less competitive, a larger pool of graduates (due perhaps in part to the well known trend for people to train during a downturn) has dampened the tendency to raise wages to compete. Degree classifications are one way to perform an initial sift of applicants (subject of study or even provider type are also used in some cases) but it looks as if the value of this as a discriminatory tool is beginning to decline.
We’re likely then, to see employers turn to other tools – assessments, interviews, selection tasks – to choose the most promising candidates. And if this happens, non-graduates who have developed the required skills may get more of a chance to compete. The message to providers has to be that relying on the signalling effect in developing employable graduates is a strategy with a limited shelf life. This frankly, is good news for everyone.