How to make a 'strategic brain' for innovation policy
Increasing its analytic capability would help government make better innovation policy
How to make a 'strategic brain' for innovation policy
“[UK research and innovation] must contain a small but highly effective strategic brain at its centre, with the aim of making sure that we invest every pound wisely.”
– John Kingman, 2 June 2016
Once upon a time, the government aspired to know an awful lot about the fine details of the economy. The 1965 National Plan, the high-water-mark of British corporatism, contains page after page of data about everything from the British leather industry to the production of transistors, compiled by officials who were expected to know a lot about the detailed structure of the industries and technologies.
Times changed. The industrial policy of the 1960s and 1970s seemed to be less good at picking winners than at subsidising well connected losers. Government loosened its controls on both the micro- and the macro-economy. With that, it chose gradually to forget most of what it knew about technologies and sectors. The new technology policy was mostly high-level, and sector-blind. The clearest example is the R&D tax credit, still the UK’s biggest innovation policy by some way: granting tax credit requires no knowledge on government’s part at all. There were some scattered centres of expertise: the Foresight programme in the Cabinet Office, which produced futurology; the Productivity Team in the Treasury; some technology experts in the new Technology Strategy Board. But the general trend was in the opposite direction.
By the mid-2000s, the change in what government knew about the micro-economy was complete: the idea that the Department of Business, Innovation and Skills should have deep knowledge of a wide range of specific sectors was largely a thing of the past*.
It turned out to be unfortunate timing. The Global Financial Crisis hit the UK with particular force, and the BIS Secretary of State, Lord Mandelson, announced that the department would support “New Industries, New Jobs” to help the recovery (a strategy entertainingly known as “NINJa”). This was a new ideological departure. What would have been an uphill struggle at the best of times was made harder by a department that to a great extent lacked the sectoral and technological level of expertise to provide the analysis this sort of policy required.
You might have thought this knowledge gap would be less of a problem under the Conservative-led Coalition, since perhaps laissez-faire would have reasserted itself, but in fact the opposite was true. George Osborne proved to be rather keen on policies to support and promote innovation, perhaps more so than his Liberal Democrat partners. The Government set up Catapult Centres to commercialise new technologies, made one-off investments like the Graphene Innovation Engineering Centre, and – remarkably in a period of austerity - steadily increased the budget of Innovate UK, the government’s innovation agency.
Gradually, Government realised that if it wanted to do this kind of hands-on innovation policy, it needed to get smarter. An early artefact of Number 10’s campaign of support of Tech City was the Tech City map, a not-perfect but very thought-provoking attempt to map tech businesses. BIS and the Research Councils funded the Enterprise Research Centre to map and research high-growth businesses. Plans were announced for a Smart Specialisation Hub, that would gather knowledge on what different UK regions were good at. Innovate UK began to improve its data and evaluations. After the 2015 election, the new science minister Jo Johnson called on local areas to map their innovative capabilities in Science and Innovation Audits**.
Despite these positive developments, innovation policy is still, for the most part, not very data-driven. When David Willetts set out the “eight great technologies” that the government would focus its support on, it’s not clear that the way the government machine was set up provided a lot of help with this kind of prioritisation process. And certainly the government’s knowledge of what is going on in the world of innovation is not a patch on the information it has access to on higher education and academic research, where decades of scientometrics and years of managerialism has created a system that was measured to a T, meaning that ministers could commission and peruse 100+ page document anatomising the efficiency of British universities in pumping out research papers and citations.
Into this mix comes this week’s very interesting comment from John Kingman, once leader of HM Treasury’s productivity team and now the incoming chair of UK Research and Innovation: that the innovation system needs “a small but highly effective strategic brain at its centre”.
This comment actually contains two very important points. First of all, the call for a “strategic brain”: If BIS and UKRI have ambitions to boost R&D and innovation in the UK by making direct investments like the Catapults, they need to know what is going on around the system – to be open to new ideas from researchers and businesses and to be aware of new developments.
The second point is the idea that this capability should be located “at the centre”, rather than dispersed around Whitehall and various arm’s-length bodies. It seems to me that one thing that holds back the government’s ability to get good data on the innovation system is the fact that this analytic capability is widely dispersed: excellent analytic teams at HEFCE and the IPO; capability in various parts of BIS; the Smart Specialisation Hub; expertise in the Government Office for Science.
This dispersion brings two problems. First of all, it makes it harder to integrate, share and compare data sets. Each group tends to prepare “final” outputs (which are useful but time-consuming) and merging data sets requires hand-offs and takes time. Secondly, the groups are typically quite distant from the kind of policy decisions we would ideally like them to inform. This leads to a vicious circle of decreasing relevance: policy makers find it hard to get data, which means they are less likely to ask for it, which makes analysts less used to supplying it in useful formats and timeframes.
If we were to take John Kingman’s request seriously, what might we do? One option is to create an analytic capability from scratch. But this seems very wasteful at a time when BIS needs to make significant cost savings – especially since taxpayer is paying for analytic capability already in several places around the public sector.
A better alternative is to bring together some of the existing capabilities to make up this new smart centre. This would simultaneously improve the evidence for innovation investments, and increase the policy impact of the analysts. What capabilities might we want to bring together?
We might start with the Smart Specialisation Hub, the experienced BIS teams working on the science and innovation audits and enterprise analysis (such as the Innovation Directorate Knowledge and Analysis team), the strong HEFCE and RCUK analytics teams, and data and evaluation capabilities from Innovate UK.
The group would need qualitative as well as quantitative insight, so in an ideal world we would add in the tech foresight capability currently based in the Government Office of Science in the Cabinet Office.
Bringing together these capabilities would create a real centre of technological knowledge within government. But knowledge alone would not be enough. In government as in business, no matter how smart you are, the best people always work for someone else. So for this new group to work, it would need to be deeply connected to other centres of expertise. It would, I hope, be porous enough to have good relations with tech startups like Growth Intelligence (who know more than most about using machine learning to identify new sectors and firms) and Burning Glass (who scrape and analyse data on emerging jobs); with agencies like Bloom and Trampoline systems who have been mapping the tech sectors in specific places from Leeds to Cambridge; with initiatives like Open Corporates who are revolutionising business microdata; and with institutional experts like the Enterprise Research Centre and the IPO’s patent analytics team.
The new group would also need to be responsive, providing timely and actionable analysis to those making investment decisions to do with innovation.
The group would do a few important things, including:
- Systematically assessing new areas of technology development
- Creating new (and use existing) networks of external experts (e.g. those within GO Science)
- Identifying areas for priority innovation investments across Government. This might include new Innovate UK platforms, and new Catalyst programmes.
The question of where such a group would sit is an open one: it could be part of UKRI; it could be part of BIS if BIS is taking the lead on this sort of policy; in a very different political world it could be at the heart of the Treasury or an economic growth ministry superseding the Treasury.
But regardless of where it sits, there is an opportunity here: the government has subtly but decisively changed the way it makes innovation policy; it can no longer get away with technological ignorance or data-free decisions on technology investments (if it ever could); it is already spending quite a bit on the analysis of innovation, but disparately and disconnectedly. The Innovation Plan and the establishment of UKRI give an opportunity to bring this all together, and make better use of existing resources so as to make better policy. John Kingman is right: now is the time for a strategic brain.
* Some of this capability was in practice moved from BIS to Regional Development Authorities, which were of course abolished in the early days of the Coalition government, a story which deserves a blog post of its own.
** As we’ve discussed elsewhere, it’s not just policymakers’ demand for innovation data that is changing. Things are also changing on the analytical supply side. Talking about the importance of big and open data can sometimes be a cliché, but in this instance it really matters. We couldn’t have done the work we did on Tech Nation 2016 or the work we’re doing on Arloesiadur, the Welsh innovation dashboard, without the ability to pull in business data, to scrape websites for information on jobs and technologies, and to analyse this cheaply. It’s increasingly possible to understand the dynamics of innovation in a way that would have been impossible ten years ago.