Beware the smoke and mirrors of the university productivity dividend

Like most local growth policy advisors, I am passionate about the positive contributions universities can make to the places where they are located.

This orthodoxy underpins the Civic University Commission which proposed ‘Civic University Agreements’ to better define increased engagement with regions. It is also a major part of the rationale behind the Knowledge Exchange Framework (KEF).

However, the belief that higher education institutions are crucial to delivering local industrial growth, and the expectations they will play major roles in formulating and delivering Local Industrial Strategies (LIS) is simplistic at best and might even be counter-productive. What if the evidence just isn’t there to justify these articles of faith?

Overstating our reach

My colleague, Louise Kempton argues the importance of universities for regional innovation is probably overstated. This scepticism is increasingly shared by those who characterise universities as relatively wealthy anchor institutions whose local involvement is mainly in their own best interest.

This critical perspective is somewhat supported by two recent statistical publications on place-based productivity – the Centre for Cities (CfC) ‘Cities Outlook 2019’ in January, and the Office of National Statistics (ONS) regional and sub-regional productivity in the UK in February.

Both publications present evidence of places’ productivity. This is important because tackling the UKs productivity malaise is probably the principal purpose of the LIS. But what is striking about both is the poor headline correlation between the highest productivity places in the UK and university density.

High productivity, low university density

Leaving aside London and Edinburgh as major capital cities, the five of the six highest productivity cities by CfC’s primary measure (gross value added per worker) are places without a major indigenous anchor university – Slough, Milton Keynes, Worthing, Crawley and Swindon. Even the one university town in the top six – Reading – has relatively low student densities per capita on a sub-regional basis (i.e. under 2%) compared to a national average of around 3.5%.

One might counter that CfCs definitions of primary urban areas as cities and their use of GVA per worker are suspect. But ONS’ quite different per hour figures on NUTS3 geographies tend to support this observation. In terms of the latest nominal GVA per hour, again putting London and Edinburgh to one side, of the top five NUTS3 areas four are either low density or have no indigenous anchor higher education institutions. In a similar vein, the top ten areas for ‘real’ (as opposed to nominal) productivity growth over the ONS 2004-17 time series include Milton Keynes, West Sussex North East (around Gatwick), North Hampshire (so Basingstoke and Aldershot), East Derbyshire (so Chesterfield) and Walsall.

There is also a disconcerting representation of major university cities in the bottom ten of CfC’s productivity ranking – e.g. Sheffield, Nottingham, Stoke – whose Primary Urban Area includes both Keele and Staffordshire Universities – and Hull is 63rd/63. The ONS data is less dramatic, but amongst the 12 NUTS3 sub-regions in England whose real productivity has fallen by more than 10% since the global financial crisis (i.e. 2010-17) are five in London, plus Liverpool, Hull and Bristol!

Juxtaposing core productivity performance data with university location (or lack of one) might not be a devastating critique of the importance of universities even in the specific public policy area of the impending LIS. However, it does raise some questions for policy makers and academics.

Missing gap in current thinking

Arguably it has been a missed opportunity in the KEF as well. Attempts to cluster universities might be a useful way of improving comparisons, but the contextual definition of areas by local enterprise partnerships (LEP) as either higher or lower competitive is simplistic and misleading. For instance, five designated ‘higher competitive’ LEPs actually have productivity per hour below the UK average (with Gloucestershire and Coventry & Warwickshire both under 95%). And contextualising a huge LEP area like South East from Harwich to Hastings with multiple university hotspots and cold-spots as being in the same category as say Leeds City Region or Tees Valley seems opaque and highly unhelpful.

It is too easy to conclude with a call for greater investigation and analysis. However, if it is the case that universities at best are fairly marginal and increased engagement and collaboration might even be a red herring to Local Industrial Strategy attempts to drive productivity improvement, then surely local leadership teams will set about formulating their LIS and its priority interventions in a very different manner! If the KEF is intended to assist places to leverage university contributions and to incentivise universities to engage more effectively in priorities for place-based change, then it needs a far stronger and more nuanced place-based context in any clustering analysis.

Even more importantly, the impending civic university agreements need to be much more honest and explicit about both the impact and limitations of university influence on local growth and development. This is the sub-national opportunity and challenge – to blow away the smoke and re-shine the mirrors that suggest increasing university-anchor institution collaboration will provide a productivity dividend to places where they are located.

Leave a Reply