Look inside, look around, look beyond: From productivity gaps to adaptive capacity

By Jun Du, Managing Director at The Productivity Institute

In September we held our 2026 conference, Building Connections for Growth, over two days in Manchester. The first day focused on research: trade and services, place, workforce skills, innovation, investment, and AI and the economy. The second, in partnership with the Industrial Strategy Advisory Council, asked how evidence can better inform policy.

We chose the theme of connections deliberately, and the two days reinforced why: productivity challenges cannot be addressed in isolation, and progress depends on stronger connections between people, firms, institutions, sectors and regions.

The harder question, and the one I want to reflect on here, is what this asks of us as researchers.

In August I wrote that the frontier no longer stands still, and that the question for TPI’s Phase 2 has moved from why the UK fell behind to whether it can adapt fast enough. I set out five priority directions: investment and productive capacity; human capital and skills; technology adoption and absorptive capacity; growth transmission; and trade, supply chains and global position. I stopped short of a detailed research agenda, because the first task was to listen, and the conference was designed in part to do that properly.

The conference did not tell researchers what questions to answer, nor should it. Research is not an instruction-taking exercise. But engagement with policymakers reveals where the evidence base is least useful for the decisions they have to make, and that is what I found most valuable about the two days.

The five priorities still stand. What changed is the question I would now put to all five.

What the conference changed

Across the sessions, the same problem kept recurring. We know considerably more about inputs and outcomes than we know about the mechanisms connecting them.

We know how much firms invest, what skills workers hold, whether firms adopt technologies and which places and sectors grow. We are much less able to explain why apparently similar firms, places or institutions respond so differently to the same opportunities and shocks.

That matters because productivity depends not only on the resources an economy possesses but on its adaptive capacity: the capacity of firms, places and economic systems to reallocate resources, reorganise activity and build new capabilities as technologies, markets and institutions change.

Why do some firms, sectors and places convert the same technologies, skills and investment into much larger productivity gains than others? Why do some shocks trigger innovation and restructuring while others lead mainly to contraction? Why do new capabilities spread rapidly in some parts of the economy but remain isolated in others?

Three gaps came up repeatedly, and I think they are versions of the same problem: what happens inside firms, institutions and places; what happens between them; and how both are conditioned by the UK’s position in the international economy.

These are the three lenses through which TPI’s Phase 2 now needs to ask the adaptation question. They are not questions policy has asked us to answer. They are questions our own evidence base cannot yet answer, which is why they belong at the centre of a research programme, not at the edge of a policy one.

Look inside

How firms actually adjust. We can often say with reasonable confidence what happens to the average firm following a shock, a new technology or a change in trading conditions. We can say much less about what individual firms did in response. One buys new equipment. Another retrains its workforce. Another reorganises production, or changes its suppliers, products or markets. Some combine several. Others do very little. An average effect conceals the route through which adjustment happened, and the route is usually what a business or a policymaker most needs to understand.

This matters most for technology. Both days called for faster adoption, and I would not disagree. But adoption is where our measurement stops: we record whether a firm has bought a technology and observe much less about what happens to the organisation once it arrives. The AI debate shows the cost of that. Adoption alone tells us little about productivity impact, because returns depend on whether firms change workflows, management practices, skills, job design and decision-making alongside the technology. Organisational change is therefore not a secondary consequence of adoption; it is often part of the investment itself.

Work is already under way at TPI on organisational capital and firm-level measurement through the Productivity Lab. Both run into the same limit: we can measure stocks of capability much better than we can observe capabilities being mobilised. The same applies beyond firms. We can describe differences between places and institutions, but struggle to identify what makes one local authority, university or business network substantially better at turning resources into outcomes.

The demand side of productivity. This is the part of my own thinking that changed the most over the Conference. Most productivity policy is framed around the supply of inputs: whether finance is available, whether skills exist, whether technologies are accessible, whether innovation support is provided. These are essential questions, but they are half the problem. What generates productive demand for investment in the first place? Where do commercially credible opportunities arise? What makes employers want higher-level skills, or a firm want a new technology? How do firms identify problems worth solving and organise themselves to solve them? None of this replaces supply-side analysis. It asks what determines whether the available supply is used at all.

Ageing belongs here too. An ageing workforce and the changing demand that comes with an ageing population alter both what firms have to adjust to, and what they can do about it.

Look around

Linkages, and the density of an ecosystem. Economic activity is highly connected, while much of our evidence is organised around discrete units: the firm, the industry and the region. A supplier may sit three tiers down a production chain. A manufacturer may depend on business services counted in another sector. A city may rely on surrounding towns for workers, housing and specialist suppliers. New management practices may travel through networks long before they appear in any statistics.

Increasingly, firms compete as parts of supply chains and ecosystems, yet these connections remain among the least well understood parts of the productivity picture. One reason is data. Another is disciplinary fragmentation: supply chains are central to management research, operations and international business, and much less visible in mainstream productivity analysis. The same is true of the density of an economic ecosystem, the thickness and quality of relationships between firms, suppliers, institutions, workers and universities. We recognise dense ecosystems when we see successful places. We are much less able to explain how they form, how they evolve, and whether policy can help reproduce the conditions that sustain them.

Where growth lands, and who gets it. If national growth is increasingly concentrated in a small number of strategic sectors, the transmission mechanism matters enormously. Do those sectors pull the wider economy with them, or become highly productive enclaves, connected more strongly to global markets than to the places and supply chains around them? Research should follow the eight growth-driving sectors, because that is where the transmission question is being tested at scale, and it should follow foundation sectors with equal seriousness: many employ large numbers of people, many have been disrupted by the shocks of recent years, and the strategic sectors depend on them for inputs, services, infrastructure and labour.

The question is not which sectors grow but how growth travels. How does growth in a strategic sector reach suppliers further down the chain? How much value stays in the UK? What determines whether smaller firms gain from the expansion of frontier firms? Distribution belongs here too. If the gains are weakly transmitted across firms, places and households, their economic and political durability will be limited.

Why places need each other. The UK is not a large country, but its administrative geographies are often smaller still. Many of the economic functions that support productivity, such as deep labour markets, specialised supply networks, knowledge diffusion and markets large enough to justify investment, operate across administrative boundaries, so places can benefit from combining scale and complementary capabilities. But connection alone does not guarantee mutual benefit. Two places gain only when each brings something the other can use, and even then the gains may be distributed unevenly. So, how can two places combine capabilities in ways that increase scale and scope for both? Which forms of connectivity matter most? How can growth strategies be designed around economic relationships rather than administrative boundaries?

Devolution makes these questions more urgent. Greater local powers and budgets can strengthen accountability and experimentation, but they can also set places competing for investment, government funds and the same firms, even where the economics points towards collaboration. Manchester and Cambridge have begun exploring closer cooperation. It is too early to draw conclusions, but the conversation raises exactly the kind of question TPI should examine. We will be working on local governance capability and on how central government, local government and civil society coordinate to produce better outcomes, and our Productivity Forums give us a standing relationship with places across the UK to ask what they need from one another instead of inferring it after the event.

Look beyond

Positioning in global value chains. The UK is a mid-power economy, highly open, services-intensive and dependent on international trade, investment, technology and knowledge flows. Yet much productivity research still takes the domestic economy as the natural frame and treats the international dimension as one variable within it. I think that this framing is inadequate. I have written about this before: a more useful starting point is the UK’s position within global value chains and international production networks, and what forms of positioning are feasible for an economy of this size and type under the constraints it faces.

Recent debate has understandably focused on resilience, dependency, exposure and economic security. Those questions matter, but they are largely defensive. The productivity question also has an offensive dimension: where can competitiveness be created and sustained in a world in which technology, trade policy, industrial strategy and geopolitics are all changing at once? Where does the UK currently create and capture value within global production networks? Which capabilities allow firms and sectors to move into more productive and more defensible positions within them?

These questions change how we think about domestic policy. An investment may have very different returns depending on where a firm sits in an international production network. Skills may matter differently depending on whether they support activities where the UK can retain value. Technology adoption may raise productivity without raising domestic value capture, if the critical complementary assets lie elsewhere. International positioning is not a topic to be added once domestic productivity has been analysed; it helps determine the returns to the domestic capabilities we are trying to build.

I should be straightforward that this remains the least developed of the three areas in TPI’s portfolio, and the one where I would most like to work with researchers studying similar questions in other open, mid-power economies.

What this changes, and an invitation

Inside, around and beyond are not three new priorities to sit alongside the five. They are three lenses through which those five should be examined. What changes is the question we put to them: less about what the economy has, more about how it adjusts.

The reason to say this now, and not in a year, is that the decisions it bears on are already being taken: how growth in strategic sectors spreads through the economy, how devolution is implemented, how AI is absorbed into the organisation of work, and how the UK positions itself in a global trading system being reshaped around it. None of these questions will yield to a single discipline, dataset or institution, so we will build this agenda with others.

This is the direction TPI’s Phase 2 will take. If you are working on these questions, or would like to, please get in touch.

My thanks to everyone who presented, chaired, challenged and argued with us in at the Conference in Manchester. Much of what is written here began in those conversations.