¡°I want to see our universities competing on the quality of what they offer, value for money and strong positive outcomes for their students so that every degree is worth the investment.¡±

Those were the words of then-universities minister Sam Gyimah in November after the launch of the ¡®Absolute Returns¡¯ report from the Institute for Fiscal Studies, looking at earnings data of graduates to the age of 29.

Much ink was spilled by commentators across the media looking at the courses and institutions whose graduates earned the highest (and lowest) salaries.

Setting aside the lively and very valid debate about whether earnings are a complete reflection of the value of a qualification to the individual or to society, there seemed to be little dissention from the line that if earnings are equivalent to value then that means the higher the earnings the greater the value to the individual.

ÖйúAƬ

Every January, the Centre for Cities produces its Cities Outlook report examining the state of Britain¡¯s urban areas. The 2019 edition focused on the effects of austerity with a set of typically insightful analyses, but every year the Centre produces table after table of data on the economy of cities.

Table 15 is a personal favourite ¨C the Housing Affordability ratio, which measures the ratio of average house price to average local income. This year, Oxford came bottom again with an average local salary of ?29,600 and an average house price of ?511,110. In other words, the average house in Oxford costs 17.3 times the average salary in Oxford. The most affordable cities were Liverpool, Hull and Burnley, with ratios of 5.4, 5.4 and 4.3 respectively.

ÖйúAƬ

Why is this relevant? With the best will in the world, the graduate labour market in Burnley is not strong. But Liverpool is a different matter. In 2017, 2,445 first degree graduates got their first job in Liverpool.

The city is, by any standard, a significant, substantial graduate labour market, of both regional and national significance. It is also substantially cheaper to live and work in than Oxford.

This then begs the question: if you¡¯re a graduate from Liverpool who can get a good job in Liverpool, does it make sense for you to move to Oxford instead for a little more money? If the answer is ¡®no¡¯, then this has implications for the whole way we use salaries as a measure of the value of qualifications.

The problem is that this question turns out to be harder to answer than you might think.

We have decent data on early graduate salaries thanks to the Destination of Leavers?from ÖйúAƬ (DLHE) survey which, among a plethora of other data, asks for salaries six months after graduation. The sample size is large enough that it allows a view on these salaries for most of the major cities in the UK. Cost of living data is another matter.

Data on housing and rents exists and could be used to gauge which cities are most affordable for a new graduate but it doesn¡¯t give a rounder picture of other costs. Enter public data sources, in this case, Numbeo.

Numbeo is an international comparison website giving data on various aspects of cities across the globe. Data is harvested from public sources and then the selected measures are indexed to the figure obtained for New York. In addition, as indicators are updated, so are Numbeo's published measures. There are obvious issues with this data, its reliably and comparability, but it's possible to work with it as it allows us to make some illustrative calculations.

ÖйúAƬ

are very interesting. One of the measures Numbeo offers is ¡®Local Purchasing Power¡¯ (LPP), which examines how far the average local salary goes in each city. If we take average graduate starting salaries and compare to average local overall salaries, we can then work with the LPP measure to derive a figure that gives us a rough guide to how far a new graduate¡¯s pay packet might stretch in various parts of the UK.

ÖйúAƬ

Derby sits at the top; an unusual labour market dominated at graduate level by well-paid jobs in engineering and so hitting the sweet spot by having both high salaries and relatively low costs of living. Southampton has a relatively low cost of living for a skilled south-eastern labour market and so offers good value. Coventry also offers good local salaries in tech roles allied to relatively affordable costs of living, and so on. And the bottom four cities are the four UK cities with the highest Housing Affordability ratio in the new Centre for Cities report.

This data is far from definitive and could be improved if we had a good quality UK source of reliable and regular cost of living information. It strongly suggests that a graduate able to get a good job in Southampton, Coventry, Liverpool, Sheffield or Leicester would need to be earning a considerable amount more in Oxford, Cambridge or London for it to be worth their while moving.

It strongly implies that for many graduates, it may not actually be rational for them to try to maximise their salary if it means moving to a location so expensive that it more than wipes out any extra pay earned.

And that has obvious implications for salary-based metrics, for the idea of a graduate earnings premium and what that actually means in practice, for local economies, and for the information, advice and guidance offered to students and graduates.

With better cost of living data we could answer interesting questions like ¡®how much more do you need to earn in London for it to be worth moving and turning down the same job in Newcastle?¡¯

If the government is really serious about trying to ensure that students and graduates make the best possible choices for themselves to get the best returns on their degrees, they need to work on getting this data out there so that graduates can make rational decisions about what is best for them.

Charlie Ball is head of ÖйúAƬ intelligence at .

ÖйúAƬ