Over the last 30 years innovation and entrepreneurship have become increasingly prominent concerns for successive UK governments. And yet our record is mixed, to say the least. The economist David Storey has calculated that we spend about £8bn a year supporting small firms in the UK. Having spent this money we should be asking: where are our Googles?
Innovation is often seen as originating from university research, which then migrates into start-ups incubated in science parks, before moving out into the wider world. There is also financial support to create geographically concentrated ‘clusters’ of networked, innovative small firms. But how many new global firms has the UK produced in recent years? A handful perhaps, ARM, Imagination Technologies, CSR, and the recently acquired DeepMind and Natural Motion are all excellent firms, but not yet at the level of Google, Apple or Cisco.
This is not just a British problem. In Europe and the US it is probably fair to say that there is not a single example of a successful cluster that has been created by government intervention.
This doesn’t mean that the UK economy isn’t innovative. On the contrary the UK has some fantastically innovative firms and entire sectors like automobiles, aerospace, pharmaceuticals, and oil and gas that are very innovative. However, these are sectors where innovation takes place in large, established firms and their supply chains, rather than in entrepreneurial start-ups.
Moreover, the main sources of innovation for these firms are customers and suppliers, with universities a long way down the list. When universities do engage with these firms it is not so much through science parks and spin-outs but in direct problem-solving collaborations and the training of future employees. In some instances – the University of Sheffield or the doctoral training centre in Financial Computing at UCL for example – that collaboration and training is arguably world leading. It is not properly appreciated how good UK universities are at this sort of activity. They get a higher percentage of their income from industry than US universities.
By contrast, looking at the bigger picture it is very hard to see any significant connection between the rate at which small firms are created, what one might call the degree of entrepreneurship, and either economic growth or technical innovation. The idea that the US is uniquely entrepreneurial compared to Europe is a myth. Levels of entrepreneurial activity are very similar, partly because there is nothing special about starting a firm – it is about as common as getting married.
Entrepreneurs are happier, have more job satisfaction and probably contribute towards job creation, but their public image as innovators doesn’t match reality. The reality is that most entrepreneurs fail, with half of firms failing within their first 3 years, and success is concentrated in a small, atypical minority.
Thinking the typical entrepreneur is someone like Bill Gates who will build a highly innovative global firm is very misleading.
The typical entrepreneur starts up a low-tech, low-growth, low-productivity firm that will typically be below the minimum efficient size needed to compete. They will often be in markets that are already saturated. So if they are around in five years, which is statistically unlikely, it is only because they have displaced a similarly poor performing firm. Rather than creating jobs and growth they are more likely to be creating unproductive economic churn, a phenomenon Demos Finance highlighted in its report Finance for Growth.
So why is there such a gap between the widely held view that entrepreneurial start-ups drive the economy and the economic reality? Three explanations spring to mind.
First, a political one. There are roughly four million small firm owners in the UK, which is roughly 10 per cent of the voting population, which in turn is enough to swing an election. No politician who plans to be re-elected is going to be critical of the politically active small-firm lobby and its £8bn yearly subsidy.
Secondly, because the failure rates for start-ups are so high and the returns are so skewed, the most visible entrepreneurs are extremely atypical. Lotteries also have skewed returns and high failure rates, but we don’t consider lottery winners representative of all lottery players in the way we consider the atypical successful entrepreneur representative of all people starting firms.
Thirdly, innovation and entrepreneurship policy have been heavily influenced by a misleading and outdated model developed by the famous Austrian economist Josef Schumpeter. Given his prominence – the Economist magazine’s column on innovation, for example, is called Schumpeter – it is worth looking at his ideas in some detail.
Schumpeter’s theory claims that most economic activity is mundane and stable. Occasionally a charismatic entrepreneur will find a new innovation and commercialise it in a new company. Because their innovative products are superior, the company will grow and displace existing firms through a process of ‘creative destruction’ to create new industries. As the firm grows it becomes more bureaucratic, creating the mundane conditions for a new cycle of entrepreneurial creative destruction.
In his later life Schumpeter changed his mind and recognised that innovation takes place in large firms, but his early theory still haunts innovation policy. Unfortunately, the theory is so poorly defined that it is easy to cherry-pick evidence to support it. Pick a well-known Silicon Valley entrepreneur, ignore all the other people and organisations involved in their success, and you have support for the theory.
Schumpeter’s theory of innovation is really a theory of invention with the innovation part left out. Invention is an event when someone first comes up with a new idea, while innovation is the process that generates the first commercial exploitation of an invention. If technologies were fully formed when they were invented, R&D statistics would mainly be research. In fact, two thirds of R&D spending is development. This is because technologies are typically in a very primitive state when they are first invented. They may work under purified laboratory conditions, but need to be modified, improved and adapted to work in commercial settings.
These incremental adaptations are uncertain, expensive and can continue for decades after the product is launched. Consider all the improvements in cars since Henry Ford. By confusing innovation with invention and over-emphasising radical change, Schumpeter’s theory misses all this activity, and therefore underestimates the time, money and skill required to develop and commercialise technology.
Schumpeter’s model also misses the difference between creating value and capturing value. As David Teece, a Professor of Strategy at the Haas School in Berkeley has highlighted, the firms that innovate are not always the firms that profit. Creating value and capturing value are two distinct economic activities and require different skills and assets. Firms can capture value with intellectual property protection, such as patents or copyrights, by having difficult to copy production processes, or strong brands, or global marketing reach or by selling complementary goods and services. All these activities take time and money and many are unavailable to small new firms.
If you assume capturing value is easy and innovations emerge fully formed and therefore superior to existing products, it makes sense to think that universities can easily innovate and small entrepreneurial firms will be good vehicles for commercialising technology. The reality is that universities often lack understanding of markets and small firms typically lack the resources needed to either undertake long-term development or capture value.
The poor performance of university spin-outs, commercial licensing operations and science parks isn’t therefore just a UK problem that requires policy interventions to correct. It reflects a misunderstanding of how innovation takes place: if we understood innovation properly we wouldn’t expect them to perform any better. Even MIT makes more money from selling T-shirts than it does from licensing. Roughly 9/10 US universities lose money on their technology transfer offices. The most successful universities in the US, like the University of California, see the transfer of technology as a service they provide back to the community that provides their funding, not a source of profit.
When we look at spinoffs, we don’t find any significant difference in economic performance between university spinoffs and company spinoffs (Lofsten and Lindelof, 2005) or in R&D inputs and outputs (Colombo and Delmastro 2002, Westhead 1997). Even when studies find firms in science parks have better survival rates, they do not find evidence this leads to improved growth rates (Ferguson and Olofsson 2004).
This evidence suggests that the requirements of growing firms are very diverse. Most problems aren’t related to location, let alone being close to a university in a science park. While there are probably benefits from collective facilities management, and there is evidence that firms that have very closely related technologies benefiting from being together, that level of closeness is much more specific than being a university spin-out. The benefits that come from subsidies tend to be countered by encouraging more poor quality firms to enter the market too early. Collectively this suggests that science parks aren’t about growing the economy and reflect universities trying to signal their research prestige (Monck et al 1988).
By contrast, the innovative small firms that are most likely to succeed are the ones that have close relationships with customers and large firms or are backed by professional investors, like early stage venture capital funds, who help build their ability to capture value. On their own, small firms are likely to have a tough time.
What can be done?
If the UK wants to be serious about encouraging innovation-driven economic growth then we are going to need to be more realistic about entrepreneurial start-ups and the role of universities. The first thing we need to do is recognise that invention and innovation are not the same. Invention is an event, while innovation is a long term, expensive, uncertain problem-solving process that takes place in firms, and normally in established and often large firms.
The key issue is to ensure technologies match customer’s requirements and universities are particularly poor places to develop this knowledge. Universities play important supporting roles, but are much less important than firms’ connections to their suppliers and customers.
A lot could still be done to improve the role of universities in the UK innovation system. They could focus on their main role, which is to develop well-trained graduates and post-graduates who then go into firms and innovate. The main output of the university system is talent rather than technology.
When universities do provide technological inputs to innovation processes in firms they need to recognise that the economic value of what they contribute is probably pretty low. Technology transfer offices in universities shouldn’t get in the way of technology transfer by trying to extract rents for themselves if that conflicts with the generation of benefits to the wider economy. Universities are funded by the taxpayer and should focus purely on maximising their support for the UK economy.
Having a technology transfer office in each university covering all technologies and sectors doesn’t make sense, as individual offices don’t have the resources to form proper relationships with firms in every sector. Similarly, all the research a firm may need is unlikely to be only in one university. Technology transfer from universities into firms would probably be better if it was specialised by sector and organised at a national scale. That way firms would have only one point of contact for all the technology transfer they need. Given the use of the internet by scientists to engage with firms, it is not entirely clear that technology transfer offices are needed to make introductions any more. However, local university offices would still be needed to make those relationships work as smoothly as possible.
When it comes to science parks, their inherent limitations need to be acknowledged. There is a role for supporting small firms during their early stages, and good arguments for concentrating some activities, but it is not clear this needs to be subsidised or universities are the best organisations to do it.
This is part of a more general point. The number of start-ups in the UK is very high and possibly excessive. The vast majority fail quickly and add little to the economy beyond generating unproductive economic churn. It might be better to focus on encouraging poor quality firms to leave the market faster so that their resources and markets can be recycled by higher productivity firms. Too much subsidy can distort the market and make it harder for high potential firms to grow. A poor quality web design firm with subsidised rent in a science park can offer its services cheaper than a commercially viable firm paying market rent.
Finally, when it comes to the atypical high-performing firms, a case can be made for support given the disproportionately high economic value they generate. One way to do this would be to ensure support is aligned with commercial reality. We shouldn’t be setting up regional venture funds investing EU funds in firms that professional investors think aren’t worth investing in. Instead we should be directing more funding into large, professional venture capital funds operating on a global scale who have the ability to grow global firms. Support should be aligned with commercial incentives by co-funding with profit-seeking commercial investors. If they aren’t prepared to put money at risk the taxpayer shouldn’t.
In short, we need to move away from thinking that if entrepreneurial firms struggle it is because there is a market failure that needs a subsidy and recognise it could be because they are not very good. We need to move away from thinking universities are the main source of innovation and recognise that firms and their customers and suppliers play much more important roles. Instead of focusing exclusively on radical innovations and start-ups, we should also be paying more attention to the overlooked incremental innovations and diffusion of existing technologies that contributes so much to productivity growth in established firms.
While powerful political lobbies will defend the subsidies given to small firms, and entrenched cultural ideas will make it hard to be realistic about entrepreneurship, eventually the evidence will simply force us to rethink.