This week I spent three days in Cambridge as part of a policy fellowship at the Centre for Science and Policy (CSaP). I also spent three days there at the end of October, promised to write a blog, came home with a head full of portentous thoughts, went back to work and immediately got distracted by the day job. I pride myself on not working at weekends, but if I don’t write this now I never will.
In my six days in Cambridge I’ve had thirty meetings with academics or business folks. Most of them have roamed far wider than my five initial questions, and many didn’t even stay on the issues around science and evidence in policy-making but went off in all sorts of interesting directions. I shan’t attempt to summarise everything discussed – and I shan’t name every interlocutor though I am grateful to all of them – but will instead strive to separate out themes which seem significant to me. At some point I have to choose to focus down on one of them (or at least fewer), so any thoughts on importance and potential appreciated.
My overriding learning is that while obtaining evidence can be hard, it has nothing on working out what evidence you need in the first place. I realise that I already knew this by another name in policy: it’s not the answers that are difficult, it’s getting the questions right.
So perhaps the first theme to start on is data. It’s everywhere – we live in a world where more and more is being generated (and of course monetised), but it doesn’t yet appear to be making policy-making any simpler or easier. At least in my areas – the data that I think would help me with policy on manufacturing doesn’t seem to exist yet, though I think the combination of passive RFID and GPS is going to give one of my successors an easier time understanding supply chains in a few years.
Part of the problem could be that most of the data I want is about company activities, and that – just like data on individual activities – is hard to gather systematically without irritating people. Two years at BIS has made me acutely aware of any potential “burdens on business”, and six days in Cambridge is more than enough to appreciate that the reporting burden we bureaucrats put on funding is just as unpopular in academia as it is in industry. Several times I was reminded of an ongoing discussion in BIS about ensuring a “minimum irreducible bureaucracy” associated with a new funding model, and how this always takes us straight back to the rules on Managing Public Money and the ever-present threat of Daily Mail headlines and PAC grillings.
I was privileged to have the opportunity to discuss this issue with Lord Wilson – several of my prior interlocutors had touched on the dampening effect that these threats have on the entrepreneurial and creative spirits of civil servants, and I was musing about whether increasing demand for innovations like the use of RCTs in policy would necessitate gradual changes to the way risk is managed in the civil service. His view was that redesigning the public accountability framework was perhaps a little ambitious even for a policy fellow. I’m not so sure – I have over twenty years left before retirement (I’m so grateful for the extra seven) and if the recent changes to the Osmotherly Rules are anything to go by, quite a lot could happen in that time.
I’ve also realised how significant the 1968 Fulton report was, in a way that I didn’t before. Partly because of my experience at CMPS/NSG I had considered its impact primarily about management and professional development, but it also created a generation of civil servants who had total ownership of the policy process – and the resources to deliver it – and a second, distressed generation who feel that ownership but have been gradually losing it (probably since Margaret Thatcher asked Derek Rayner to create the Next Steps agencies but much more clearly since Tony Blair asked Peter Gershon to start a decade of belt-tightening). At least today this loss of ownership is not an inevitable result of an insatiable appetite for “efficiencies” but a properly ideological position that the Civil Service should not, and does not, have a monopoly on making policy.
Jeff Patmore gave me the superb insight that large organisations which develop good ways of nurturing the fairly and very talented necessarily struggle to deal with the truly exceptional, who need the rules bent to get the best out of them. So they often end up building their own organisations, which links to the conversations about the UK being excellent at nuturing start-ups and poor at translating them into behemoths (the vogue term for this being “scale-up“). There are many theories and discussions about this, but one that hadn’t occurred to me before was that successful entrepreneurs sometimes sell up not because they tire of running their company but because they – sometimes reluctantly – make a choice to extract their wealth from it. I had a couple of conversations about the excitement around secondary equity markets which allow people to realise cash from their holdings without needing to cede control of their company, but I couldn’t identify with the reasoning about needing to send their children to private schools and buy second homes. If that’s the reason so many UK start-ups get sold I’m not sure I have much of a problem with it.
Lord Hennessy will be pleased to know that history played an important role in many of these discussions. At one point I was discussing with Martin Daunton the aspiration to double UK exports (my first use of the word “trillion” without hyperbole) in the context of the Poor Law of 1834 – whether policy decisions are based on any real consideration of what kind of society we want to live in. (I struggle to explain why the aspiration is to double gross exports rather than net exports, but the point holds.) I was often reminded of Lord Heseltine’s views on the importance of industrialists to Victorian policy-making and how a sort of reciprocal or collective conscience tempered their self-interest. It will be interesting to see whether Devo Manc and any other mayoral deals tempt any business magnates to don the Chamberlain mantle and interrupt the era of the professional politician.
But back to data. I came home convinced that government should not in most cases be deciding what data it needs to collect from any given project or scheme. The bidders / delivery partners / customers will make those decisions more easily and with better understanding of the implications. We should instead focus on stipulating how, where and when the data should be made available, if necessary providing the platforms and tools to ensure that it remains interoperable and accessible to everybody for as long as necessary – and then let crowdsourcing do most of the work to provide it. Ian Leslie reckoned we should look back (history again) at the connections between current innovations / drivers of growth and previous funding; Andy Neely reminded me that we could learn more from failure, not just in the traditional sense of understanding why projects fail but in seeing the patterns in unsuccessful bids – what are the Research Councils not funding, and is the UK missing the next big breakthrough?
This leads to both standards and paradigm shifts. I have yet to follow up with my BSI-facing colleagues but I know (again from working with Lord Heseltine) that the UK plays the standards game half-heartedly. International standards bodies are packed full of Americans, Germans and others, with the UK often under-represented. The explanation seems simple to me – the public sector lacks the resources and the private sector lacks interest – in most cases we Brits simply don’t see standards as important or relevant in the same way that the Germans do, and this comes right back to the different structural relationship between public and private sectors there (conscription to Chambers, the Mittelstand etc.) and the private sector’s role and influence over policy. Perhaps it is unfair to say that the UK private sector lacks interest in standards, but simply assumes that they are the job of government. With Industry 4.0 and The Internet Of Things just around the corner this would be a good time to up our game a bit.
I had several conversations about big companies being unable to cope with paradigm shifts (reading Thomas Kuhn was one of my few useful diversions as an undergraduate) – perhaps the best example being Kodak, who saw digital cameras coming yet still could not adapt in time. Jaideep Prabhu gave me the insight that this is more difficult for manufacturers precisely because investment in physical assets ties them in to their obsolescent business model. Steve Evans and Nathan Crilly helped me refine this – it’s not just investment in plant, but all the ancillary policies and procedures that a leading company evolves to keep just ahead of its competition – your HR strategy, your workplace strategy, your T&S policy and so on. Studies have shown (e.g. on chess problems) that if you have a solution to hand you are much less likely to find a better one than if you have none to start with – adapting to a completely new paradigm means having to reconsider everything, not just the production aspects, and it’s just too difficult for many whose systems have lost flexibility (Graham McShane and I wondered if this ossification was an inevitable successor to development and assurance, particularly for risk management systems – Laura James noted that it also happens to funky disruptive policy units so look out GDS). We decided that IBM would make an excellent case study, having reinvented themselves four or five times after these shifts without going under: typewriters to mainframes to minicomputers to PCs to (proprietary) software to services (for Linux). I also decided it will be very interesting to see whether ARM, who depend on few or no physical assets, find the next paradigm shift easier (I wonder whether smartphone to IoT counts as a new paradigm, or whether it will be something altogether different).
And so to IP. Ross Anderson – having taught me that cyber security is not in fact snake-oil but is about making online markets work properly – kindly invited me to his security group meeting, where there was lots of discussion about ARM and its TrustZone technology and its approach to openness and IP. I’m a big believer in openness – Apple’s approach is bad, Android’s is good (and morality aside, look at the growth rates of their app/developer bases – openness wins hands down). As HMG’s senior relationship manager for ARM it was very interesting to hear the views of academics at the forefront of the research curve on which ARM relies. At the moment I think ARM still manages to be seen as “not evil”, and I wonder how much of this is technocratic benevolence and how much is because ARM doesn’t do tax avoidance. The case studies of Google and Microsoft show that neither is sufficient on its own.
In 1997 at the height of the Napster furore I wrote advice to my then Secretary of State that the reason consumers were downloading music illegally was primarily because there was no legal way to do it, and predicted that the music industry would eventually stop lobbying against it and agree to a mechanism for digital distribution. Four years later iTunes was launched (by a tech company not by a music company, as the latter clung to their obsolete thinking that they made CDs rather than distributed music), and I wonder how much innovation was suppressed or lost during that half decade of demonising consumers. We’re now entering the dawn of the era of 3D printing and localised manufacturing, and I can foresee exactly the same arguments from big incumbent manufacturers and retailers that people shouldn’t be allowed to download schemata for vacuum cleaner nozzles, fridge hinges or whatever else they end up wanting to print down at the local corner shop. They’ll be unsafe, you see, and who will be liable then, eh?
So I was pleasantly surprised to discover a strong consensus among my interlocutors that demonising consumers and suppressing innovation are inherently bad things, and that we should seize the opportunity both to remind people why we have a system of IP protection and to set out clearly how our laws encourage invention and creation while enabling buyers and sellers to trade in a fair marketplace. Given that we’re starting from a status quo of “tolerated infringement” (it is technically illegal for me to print a Minnie Mouse picture from the internet for my daughter without paying something to Disney) – and of startups obtaining patents not to protect their IP (they would likely lose in litigation against big companies with better lawyers) but as a saleable asset and a badge of seriousness, it seems clear to me that the system is broken and needs extensive redesign before it will make sense to policy customers let alone to the man in the street. Strangely, it seems that various international treaties and agreements mitigate against doing precisely this.
My discussions with Lionel Bently and others about IP led me to a deeper understanding of the hypocrisy at the heart of American foreign policy. The “fair use” exemption in US law is actually far wider and better for consumers than the exemptions in EU legislation, yet when arguing abroad the Americans will fight to the hilt to protect Apple, Google and any other US giants (hence the draconian WIPO agreements). This is precisely congruent with insisting at the WTO that developing countries “open up their markets” while remaining hugely protectionist about US steel and sugar. Several of my discussions raised concerns about the UK being alone in honestly operating a free market without protectionism – I don’t have an issue with this. I think the moral high ground will ultimately have significant economic value.
One thing the Americans do far better than us, said my interlocutors unanimously, is to empower their civil servants to make the right decisions to foster innovation and pull through new ideas and technologies. Leaving aside their vastly bigger budgets, the difference seems to be one of autonomy and risk appetite, bringing us right back to the Daily Mail and the PAC. American departments of state directly manage and contract for innovation projects, leading to a mutually beneficial dialogue between the researchers/early stage companies and the government. We created the Technology Strategy Board (now Innovate UK, which is awkward because IUK means Infrastructure UK) precisely because of the view that civil servants can’t really do that properly, and we need better-paid experts to have that dialogue for us. The opinions of academics on this arrangement seem to contrast utterly with those of industry, who see it as a huge success. I think this is partly because academics have more respect for civil servants than many business folks, and partly because the arrangement makes the research funding ecosystem more complicated without really changing the complexity experienced by industry. More than one person said we need much more use of systems like the US-inspired SBRI to bring the innovators closer to the customers in departments – particularly SMEs, who do not have anything like the voice that big companies have in these dialogues (though it is arguable that the gap is smaller in influence over TSB decisions than over policy decisions). The one thing everybody seems to agree on is that Swindon really wasn’t quite the right place to put them. Oh well.
The final theme I need to mention is network effects (at this point I ought to thank David Cleevely for setting up the programme in the first place – I’ve just read his essay in last year’s celebration of Sir Mark Walport’s appointment). Everyone knows about the Cambridge Network, and the bigger phenomenon of the small-n network which has resulted in so many innovative companies over the past few decades. Almost everyone wanted to discuss it and combining their inputs so that I can draw some insights into UK manufacturing supply networks is perhaps the biggest task ahead of me. Sir Mike Gregory reckons that different levels of contacts are important – at Cambridge the great and good have direct lines to each other but they also have an extremely efficient environment for setting up working level contacts and making progress quickly. Several people talked about the history of the university’s relationship with spinout companies (not helped by a period of unfortunate IP centralisation in HMG), and how it has settled into something notably different from Oxford, London and other UK universities. The Witty Review appears to have resonated not at all, but that could be unique to Cambridge.
I think also unique to Cambridge is the density of innovation. It’s really quite a small place, and with 9000 people pushing back the boundaries of knowledge it ought to be almost impossible not to find collaborators on whatever fantastic business idea you’ve had. The fact that each of the individual organisations (both within the university and without) is quite small helps with this issue of direct connections between people at the right level – good ideas don’t get lost in someone’s secretary’s inbox (much). From a manufacturing perspective one of the most interesting types of network is the “makerspace” experimenter/hacker community – I met Cambridge Makespace’s CEO Laura James twice (she’s also the CEO of the Open Knowledge Foundation so we talked about data a lot too), because I can see makerspaces having a significant impact on the people who will found tomorrow’s startups – inspiration from being able to experiment in a safe environment without scrutiny, access to information and people, and access to kit that you might not have the certainty or money to invest in at this point in your business’s cycle, solving a Catch-22. In many ways they’re like Catapults, for micro-businesses. Having discussed SME access with both MAS and the HVM Catapult I wonder if there is a gap in the market for something halfway between a makerspace and a full-fledged Catapult, with a bit more structure and kit than a hacker community meeting in a spare building but with lower barriers to entry than a Catapult.
The single most interesting insight came when I asked Laura about outcome data and case studies. Stewart McTavish (who runs ideaSpace, which is complementary but different) had mentioned how helpful it was that although ideaSpace’s funding agreement required them to report outcomes to BIS, the organisation wasn’t bound by hard targets. Laura said that the few case studies from Makespace successes had arisen from people proactively seeking her out to tell their story – “if I had to document systematically all the outcomes, they wouldn’t happen”. A bit like Heisenberg’s uncertainty principle applied to innovation: you can have the outcome or the data, but not both. Or perhaps Schrodinger’s cat – a system for capturing outcome data can materially affect that which is being observed. I think perhaps one of the greatest policy challenges is trying to gather evidence without doing this.