There are plenty of things I’d like to see in the innovation landscape of 2030, from more public spending on innovation to greater democratic participation in the process. But in this post I’d like to focus on one particular thing: my fervent hope that by 2030 we’ll have a far more intelligent innovation policy system.
Let’s quickly recap the state of innovation policy in the UK. An important policy battle seems to have been won. The Government, the erstwhile Coalition and the Opposition all seem to agree (i) that the government needs to spend money to encourage innovation and (ii) that that money needs to be spent in specific fields and specific technologies, not just on generic interventions like R&D tax credits and university tech transfer schemes.
Osborne and McDonnell may disagree over how much to spend on innovation, but the idea that government should help fund and develop specific technologies, from spaceplanes to self-driving cars to precision medicine is now the consensus.
It’s great that the UK government is pursuing a more mindful innovation policy. After all, it’s what some of the most innovative countries in the world do, from the US and Germany to Israel and Taiwan.
But there are two practical problems:
- It’s not clear that the UK government has the analytical resources to make these policy decisions in an informed way. The selection process for Catapult Centres and other big capital assets seems at best to be based on qualitative analysis, at worst on guesswork and lobbying.
- There’s a surprising dearth of knowledge about which innovation policies actually work.
The analytical problem
My co-panelist David Willetts has done more than most to foster this new more mindful mode of innovation policy, not least in a seminal speech outlining the “Eight Great Technologies” the UK government would seek to support.
As far as I can tell, it seems that he had to do a lot of the analytical work on this project himself – the prioritization was certainly not something the government’s innovation policy set-up was designed to produce. (Entertainingly, a detailed quantitative report patenting activity in the eight great technology areas was published by the IPO’s excellent Patent Analytics team – in October 2014, nearly two years after Lord Willetts’s speech.)
This is not to say there aren’t lots of data available for an innovation minister. The university sector is notoriously quantified to within an inch of its life. The REF, the “metric tide” about which James Wilsdon has written so eloquently, and the abundance of patent metrics and R&D stats ensure that no innovation policymaker will ever be short of numbers.
But those numbers, focused as they are on academic research, rather than development, or on the generality of R&D spend rather than specific areas of expertise, tell us surprisingly little about the type of innovation policy we now want to do. They can dress up a report; they can impress a busy Treasury official demanding “evidence” to justify the next Budget appropriation. But in reality, they are like Potemkin villages: very impressive, but of little practical use.
The evidence problem
Our evidence problem is related. There are a whole host of government policies that can be used to support innovation, from tax credits to innovation districts to development grants. Working with the University of Manchester, Nesta produced a compendium of evidence of how useful they are.
The disappointing result was that the evidence for most innovation policies around the world was poor and incomplete. What’s more, many existing government innovation policy schemes have been set up in ways that make it hard to gather evidence on them. (The most egregious example is the R&D tax credit, the UK’s largest innovation policy by some way – because it is a tax break, it is extremely hard to find out which firms claim it in order to see how effective it is.)
The net result is we end up using policies whose effectiveness is uncertain to invest in areas we select based on intelligent guesswork.
A better world for 2030
In my ideal innovation policy world, we’d fix this.
First of all, we’d rigorously gather data on our innovation policy interventions and carefully evaluate them to learn what is effective. Randomised controlled trials have a role to play here – one that Nesta is helping pioneer through the international Innovation Growth Lab; sound qualitative research matters too.
Secondly, we’d gather data on how innovation is happening, in which sectors and technologies, and in which places.
Less than a decade ago, this sort of data was prohibitively expensive to gather. But better business records and better analytic techniques mean we should now hold ourselves to a higher standard. Nesta has trialed some of the sorts of analytic techniques that could be using in the Tech Nation 2016 report and in Arloesiadur, the Welsh innovation dashboard. These methods aren’t a silver bullet: they work best when combined with local knowledge (such as the recently announced Science and Innovation Audits) and qualitative approached like GO Science’s horizon scanning. But they would allow us to make better policy.
I’d like to see a system where innovation policymakers and funders, like the new UK Research and Innovation, gather these new sorts of data on innovation and use them to test and to inform the process of investment. And I’d like to see them take a more experimental approach to the innovation policies they deploy.
We’ve made innovation policy off the back of Potemkin data for too long: by 2030, I’d like to think we’ll be using the real thing.
Stian Westlake is Executive Director of Policy & Research at Nesta. His research interests include the measurement of innovation and its effects on productivity, the role of high-growth businesses in the economy, financial innovation, and how government policy should respond to technological change.
You can follow him on Twitter @stianwestlake.
Find out more about the Visions for 2030 PolicyLab event.