Data analytics can make for better informed innovation policy. The UK should take note.
There are plenty of intelligent people making innovation policy, both in the UK and the rest of the world. But there is a shortage of intelligence. Too often, policymakers lack basic data on the innovation system. Sometimes they don’t even have information about the policy that the government itself makes to influence it. Whichever party forms the next government should change this.
If you want to make good innovation policy, information matters. A lot of innovation policy involves the government trying to support new technologies, technology businesses or clusters. In practice, this requires people within government taking a view on which different things are worth supporting, and which policies to use to support them. (If this strikes you as old-fashioned industrial policy, I can assure you it’s not: consider the current government’s support for Tech City, for graphene in Manchester, or for the Life Sciences sector through the Crick Centre or the Patent Box. Though many people don’t like to admit governments “pick winners”, they try to do just that.)
The problem is, the British government, and those of many other countries, lacks most of the intelligence it needs both to decide where to put its money, and even to know where its money has gone.
As far as I know, no one in the UK government can say how government funding for innovation is allocated among different sectors - how much goes to life sciences vs aerospace vs software, for example. The R&D tax credit, one of the most costly innovation policies, is almost completely opaque, with very little information available even within government about which companies use it.
Rich sources of data that government bodies maintain, such as Innovate UK’s database of grants, the IPO’s information on UK patents, or the Inter-Departmental Business Records data set are underexploited as tools for policymaking. The end result is that when politicians seek to get specific about where they are going to focus their innovation policy – as David Willetts did in his “eight great technologies” speech - they typically have to rely on subjective information and expert opinion*.
Expert judgment is not a bad thing. But it’s much more robust if it is supported and informed by a shared fact base. This is precisely what some other countries are beginning to do in their technology policy.
One large US government department, a long-standing backer of world-changing technologies, has for several years used quantitative tools to analyse patent data, business data and media stories to identify areas of tech they should be supporting. (They’ve used software designed by Quid, a San Francisco company with a British founder.) TEKES, Finland’s technology funding agency, has been using social network analysis and business records data to understand its technology clusters in order to improve its funding strategy.
There have been some promising first steps in the UK. Innovate UK, has done interesting experimental work with Growth Intelligence and the IPO’s patent analytics team to try to understand emerging sectors. Projects like the Cambridge Cluster Map (which Nesta co-funded), the DueDil/Tech City UK TechNation map, and Nesta’s Creative Cluster maps have attempted to measure and highlight tech clusters. But my impression is that so far these projects have been quite small scale, and not really designed to inform the policymaking process.
A new government should go further.
First of all, it should experiment with existing analytic packages to map emerging technology sectors in the UK. (This could be a good area for an SBRI contest – companies like Quid, Growth Intelligence and DueDil might well take part.)
Secondly, government should make more use of its own innovation data. BIS, either on its own account or through Innovate UK, should consider making regular reports on emerging sectors and technology fields. The team doing this could bring together existing data sources, making them open where possible and getting appropriate permissions to use them where not. For example, an open version of the IDBR, with confidential business information removed, would be a very valuable resource; it would be even more valuable if data on taxpayer-funded support and patents could be added to it. It would also make sense for HMRC to make it easier for BIS to analyse companies receiving R&D tax credits.
Analytics cannot replace sound judgment, deep relations with innovators, and expert opinion. But it can make the government’s decision-making process smarter, more confident, and hopefully more ambitious too.
* The IPO’s patent analytics work done to support the Eight Great Technologies speech is a notable exception – I’m clear whether it was done before or after the technologies were selected though.