The new Labour government has bet everything on reviving economic growth. This makes a lot of sense after over a decade of stagnant productivity and GDP growth, since there is little chance of sorting out net zero, the care system or public services without growth.[i]
Much has been written about whether the plans make sense, whether taxes will have to rise, whether relaxing planning rules will have the promised impacts and more.
But surprisingly little of the strategy, or of the commentary, addresses what may turn out to be a crucial element of what’s needed: a steady improvement in the ability of Britain’s firms, public services and civil society to adopt more effective techniques and technologies.
This is the heart of any economy – the capability of firms to make aircraft, run shops, orchestrate software and thousands of other tasks. It depends on financial and human capital but is not reducible to them.
It’s also very variable in the UK. A minority of firms are as productive as the best in the world. A majority are not, and the differences are widening, with workers in firms at the 90th percentile of productivity produced around 3.7 times as much output as workers in firms at the median of the distribution. The widely shared diagram below, from a few years ago, captures the problem - the large proportion of firms that are simply much less productive than they should be:
Without improvements to their capabilities, we’re unlikely to see much improvement in growth. But as I’ll show the absence of ideas on how to fix this is an important gap in Labour’s otherwise sensible plans, and it’s a gap that reflects a broader weakness in the knowledge base of many decision-makers and commentators.
These issues are not glamorous. But they are important. No-one would win an election by focusing on them. But they could be crucial to helping the government achieve its goals, and so, indirectly, crucial to its chances of re-election.
Theories of growth and their limitations
To understand what has gone wrong in terms of both diagnosis and prescription we need to go back a step. For most of the 20th century mainstream economics argued that growth was essentially driven by inputs of capital and labour. These were the key factors, and economic models of all kinds focused on these.
Labour’s approach reflects this in that much of it focuses on raising investment. This is wise given that UK investment has been low by international standards, both in the private sector and the public sector, and in relation to infrastructure. More investment is indeed a necessary if not sufficient condition for higher growth (and has recently been endorsed by the OBR).
Labour also, rightly, addresses inputs of labour both in terms of quantity and quality. The UK has done relatively well in increasing participation in recent years, but there are still big problems of quantity (the several million inactive) and of quality (the results of longstanding weaknesses in training which got worse in the 2010s). Again, fixing this a necessary, albeit not sufficient, condition for higher growth. For many economists, more investment and more skills, combined with a favourable macroeconomic and regulatory environment, are the keys to growth and there’s not much more to say.
How economics shifted
Towards the end of the last century, however, some of the brightest economists became increasingly uncomfortable with the idea that capital and labour are the only real inputs to an economy. The dominant models turned out to have a large ‘residual’, which meant those aspects of growth which weren’t explained by capital and labour inputs. These were given the deliberately opaque label of ‘total factor productivity’, a scientific-sounding euphemism to label the factors that couldn’t be explained (in one definition TFP is ‘the portion of output not explained by the amount of inputs used in production… its level determined by how efficiently and intensely the inputs are utilized’).
To many economists it was obvious that the dominant theories simply didn’t fit either the leading-edge economies, like the US, which were increasingly founded on knowledge, or examples like the USSR which did well in mobilising capital and labour but much less well in achieving growth. Meanwhile, analytic work showed just how inadequate the dominant models were: twenty years ago, for example, William Easterly and Rob Levine estimated that TFP explained 60% of growth in output per worker. So the importance of understanding it, and how to influence it, became urgent.
Knowledge, innovation, institutions and adoption
The crucial advances came in three main areas of all which had been absent from the mid-20th century economics mainstream: knowledge and innovation, institutions, and adoption. I’ll cover each briefly in turn.
First, innovation. Perhaps the most important work here was done by Nobel Prize winner Robert Solow who showed that innovation – by which he meant new ideas and their adoption – probably explained the majority of economic growth and should therefore be a priority for policy. He became famous for his comment that computers could be seen everywhere but in the productivity statistics – but after he had made that comment, as firms learned how to make the most of computers, the impacts on productivity proved to be far-reaching.
Another Nobel Prize winner, Paul Romer, also played a decisive role in shifting thinking. He showed that economic growth is largely endogenous, again, driven by rising levels of knowledge and in particular helped by the impact of what economists called the non-rival nature of knowledge, and its spillovers, meaning that when one firm uses new knowledge it remains available for others to use as well.
Several decades of intensive work has built on these foundations through the work of figures like Philippe Aghion and Jonathan Haskel, Ricardo Haussman and Cesar Hidalgo who have mapped economies in terms of their complexity (which turns out to be quite a good predictor of growth). Organisational researchers such as Linda Argote have focused on learning curves and patterns, complementing the work of economists like Wesley Cohen and Daniel Levinthal studying the absorptive capacity of economies, the ability of firms to absorb new ideas.
The importance of institutions
If these economists emphasised intangibles and learning, and their central importance to economic growth and productivity, a parallel stream of economic revisionism emphasised institutions. There was little value having lots of capital and labour to feed into the economy if you didn’t have institutions to sustain the rule of law, block corruption and deliver public goods. Douglass North, yet another Nobel Prize winner, was the most prominent advocate of this argument which became almost an orthodoxy by the beginning of this century (the eminent economist Dani Rodrik wrote that “the quality of institutions “trumps” everything else”). China’s successes have partly tempered this confidence, but few would now dispute the importance of institutions in helping economies grow and adapt.
Adoption and diffusion at the heart of economic history
Finally, these shifts were echoed by a sea-change in thinking about economic history which is also very relevant to the UK in the 2020s. An earlier generation of historians had emphasised which countries dominated the cutting edge of technology, from a Britain that dominated in textiles, iron, steel and ships in the 18th and 19th centuries, to US dominance in cars and planes in the 20th and the potential for Chinese dominance in AI in the 21st.
But more recent historians have put much more emphasis on patterns of adoption and diffusion. The UK story is now seen as much through the depth of its engineering skills: France had more advanced technical institutes in the 18th century, but the UK did well through mechanic’s institutes, clubs and networks that spread the ‘tinkering’ expertise that fuelled industrial revolution. This was a strength in depth that more than made up for deficiencies at the elite level.
A century later the US prioritised diffusion and adoption, notably through agriculture and land grant colleges and manufacturing advisory services, and these gave the US a dominance in productivity long before it achieved dominance in science and technology. Then in the second half of the 20th century the story was repeated with Japan, South Korea and later China, all of which prioritised adoption. Singapore is a particularly successful example of policies, and tax credits, sharply focused on adoption and productivity improvements.
None of this is to discount the value of being at the cutting edge of science and technology. But the use of new ideas matters just as much as their generation, not just in terms of productivity but also in terms of regional balance.
So to summarise: the last few decades have brought a profound shift in economic thinking, which recognises that while growth depends on inputs of capital and labour, it also depends on the creation and sharing of knowledge, the spread and adoption of new technologies and techniques, and the quality of institutions. Any coherent strategy needs to pay attention to all of these.
What follows? What should be done?
The analysis above may seem obvious in an era when so much of the economy is shaped by high technology, and many of the world’s most valuable companies are founded on data and platforms.
What’s odd, however, is the extent to which these points are missed out in the economic debate. There is plenty of rhetoric about science and high technology, and some on adoption of AI, but much less about what these mean for a typical firm in a typical place. Studies of productivity do mention the importance of adoption of new techniques and technologies, but are rarely confident in proposing any solutions (partly because the neglect of researchers also means that there are few evidence-based policies to advocate).
One reason for the relative failure to engage with recent economic thinking may be that most of the key policy-makers and commentators in the UK received their economics training before these big shifts in the discipline and have had little engagement with recent advances (John van Reenen, who’ll chair Rachel Reeves advisory council, is a rare exception).
As a result most decision-makers and commentators are more comfortable with traditional macro and microeconomics, on which they are often impressive and eloquent, or making the case for raising investment as the key to higher growth. They’re much less eloquent discussing the efficiency of the everyday systems used by delivery firms and supermarkets, paper factories or rail companies. Yet these are the places where productivity either stagnates or rises.
These issues are largely invisible in the broadsheet newspapers, or on the BBC. The recent advances in economics simply don't feature much in what could be called 'media economics' - the ways in which news media typically cover the economy (focusing on finance, interest rates, spending, pay, employment rates etc but rarely touching on knowledge, business models, techniques or adoption, even though these arguably have more impact on daily life than whether the FTSE went up or down yesterday).
So, if Labour is serious about raising productivity it needs to complement its existing plans with new elements.
Take adoption and diffusion seriously
First, government needs to pay much more attention to the practicalities of adoption and diffusion, since without better methods being used in garages and shops, by plumbers and doctors, schools and bus companies, productivity is bound to remain stagnant.
How ideas spread is complex. It’s shaped by many factors. They include investment (which enables adoption of new software for example), competition, regulations and laws, training, incentives, supply chains, online support, coaching, peer influence and more. The spread of new digital technologies is not the same as new business models or new marketing and production techniques.
But there are some clear policy implications and there is a long history across the world of successful policies designed to boost adoption. In the UK a small but important programme called ‘Business Basics’ was run a few years ago by BEIS which used experimental methods to test out what works in boosting adoption (Nesta, which I ran at the time, was closely involved). It tested out promising approaches to coaching, scientific methods of decision-making, and microbusiness growth methods, and found positive results (all summarised in a final report published by the government earlier this year).
The key point was that many different methods needed to be used to drive up both the supply and demand of new techniques – and that these needed to be tailored to different contexts.
Something of this kind is badly needed on a much larger scale, so that it can tailor programmes to the very different needs of construction, schools, retail and so on, and also tailored to priorities like net zero. Better data are also needed so that we can track in more detail how quickly new methods spread in different industries (current use of AI by small firms varies greatly by sector, for example). Many existing tax incentives (including some R&D tax credits and the patent box) do little for this kind of productivity enhancement. So, there are plenty of options for cost-neutral adjustments.
Ensure science and technology policy is more balanced - supporting use of knowledge as well as its generation
Second, Labour’s science and technology strategy needs to be broadened out and balanced. Labour is right to encourage investment on the frontiers of science, one point of continuity with the last government. But it would be a big mistake to focus only on new knowledge rather than its use. If you care about growth and opportunities for the majority, these matter just as much as the successes of DeepMind or GSK.
This has never been an interest or priority for the south-east based science elite and it has long been one of the glaring gaps in the otherwise thoughtful reviews done every few years by eminent scientists. These questions of adoption are seen as somehow a bit grubby, lower status, a topic for engineers, not the most illustrious professors.
The dominance of the biomedical sciences and industries – whose members have filled the top jobs in science policy for decades and are even more entrenched now - has probably reinforced this myopia, since the biggest pharma and biotech companies are very good at keeping up with the world’s best and have little feel for the challenges faced by SMEs. But a science and technology policy that doesn’t address adoption, and in particular adoption of new methods using data and AI, risks entrenching imbalances rather than solving them.
Create new institutions to support adoption and diffusion
Third, this all needs institutions. The UK currently has powerful institutions for science funding (UKRI and now ARIA) and even more powerful institutions for finance (the Bank of England, Treasury and the City). But it has no institutions dedicated to the tasks I’ve mentioned above.
In the past, Britain often looked to Germany as a role model in terms of technology and technique diffusion in its Mittelstand, the layer of small and medium sized companies that sustain German industrial dominance in fields like machine tools and chemicals (Germany is also home to 12 of the 20 most R&D intensive firms in Europe).
Yet the German government is now creating a new agency, DATI (Deutsche Agentur für Transfer und Innovation, the German Agency for Transfer and Innovation), to focus solely on adoption (I’ve been one of its advisers). Its role is not just to help the private sector adopt new methods in data, AI and other fields, but also public services and civil society. The UK should be doing something similar too, and urgently. If driving up adoption isn’t a central part of someone’s job, it’s unlikely to be done well. In my view we will also need a clutch of other new institutions to shape the economy of the next few decades – whether in skills, net zero or the governance of AI (and I’ve written elsewhere on what these might look like). But this should be a relatively easy place to start.
Incorporate innovation and adoption into place-based policies
The final priority is to rethink place-based and regional strategies in terms of knowledge, organisational capacity and adoption. Low productivity in much of the UK reflects weaknesses in organisational capacity, linked to all the factors already mentioned. To achieve more balanced growth requires that there are multiple centres of innovation and diffusion. In a few weeks time (23 October) I and colleagues will publish a major review of urban innovation districts, and the lessons to be learned from hundreds around the world, in cities as diverse as Boston and Barcelona, Medellin and Melbourne (we’ll be doing this jointly with the UK’s Innovation Districts Group).
Getting this right isn’t easy. The world has seen dozens of failed attempts to copy Silicon Valley. There’s a tendency to herding (everyone now wants to be an AI or quantum hub), and to an exaggerated emphasis on buildings and PR. In some cases, too, these districts create pockets of innovation that have little effect on the wider economy and do little to create opportunities for local people.
But where these are carefully tied to existing capabilities and advantages, and where they are linked up with schools and apprenticeships, supply chains and inward investors, they can be powerful tools for accelerating growth. In these cases doing more innovation and R&D also has the side-effect of improving adoption.
A shift in mindset?
The dominance of London and of finance – the Bank of England, Treasury, and the City – has for years given the key London-based decision makers a distorted view of how the economy actually works. They see it in abstract terms, primarily through the available numbers, and through a monetary lens, and are relatively uninterested in the everyday reality of making and maintaining things, providing services, orchestrating thousands of people in complex systems.
Some also genuinely believe that none of this matters. If you are a true believer in market orthodoxy, then it’s obvious that market forces will pressure companies to adopt the best available technologies and techniques, and if they don’t, they’ll go out of business. In theory, rational firms will naturally adopt any method that will improve profits. Unfortunately, this is one of many cases where the facts and the theories simply don’t align.
To conclude: if the government’s ambitious growth plans are to have a good chance of success, Britain needs both a change of mindset and a change of practice. It needs a mindset that puts knowledge and learning at the core of how we see the economy, and a shift towards policies and programmes that boost the ability of people and organisations to adopt, adapt and learn.
We need the UK to thrive in rocket science and its equivalents. But a modern economy isn’t only about rocket science. Widespread prosperity depends on the strength and depth of human and organisational capacities in everything from shops and car manufacturers to airports, primary schools and hospitals to farming. None of this is as exciting as rocket science. But it is essential. And at the moment it’s largely missing.
[i] I‘ve written elsewhere on the separate question of what it’s desirable to grow, and the limits of the still-prominent, but anachronistic argument, between those who are pro and anti-growth: https://www.nesta.org.uk/blog/what-kind-growth-do-we-want-and-why-are-we-still-having-wrong-arguments/.
Comments