The most productive, fastest-growing and most heavily invested part of the UK economy is invisible in the data that policy-makers use for decisions. A new classification of frontier firms reveals the scale of the gap – and the consequences for interest rates, industrial strategy and growth policy.
As economists, the way we look at the economy is out of date. This blurs our understanding of what’s going on and it is a big constraint on effective policy-making.
Take these two examples. The UK’s latest unicorn company, Fractile, which is building what it describes as the world’s fastest artificial intelligence (AI) chip, sits in the same industrial category as IT support companies. And Oxford Ionics, a world-leading quantum computing business, is classified alongside manufacturers of computer printers. Cutting-edge companies are lumped in with run-of-the-mill ones in official data.
This is a structural feature of the UK’s Standard Industrial Classification (SIC) system, which was designed in the 1940s to categorise an economy of factories, farms and department stores. The SIC system has been updated periodically, but its architecture has not kept pace with an economy that is now driven by AI, synthetic biology, climate technology and advanced manufacturing. The result is that the most productive and dynamic part of the modern economy – what might be called the ‘frontier economy’ – is effectively invisible in official statistics.
New analysis using The Data City’s Real-Time Industrial Classifications (RTICs) – which identify frontier companies directly from the language that businesses use to describe themselves on their websites – shows why this is a problem for economists and policy-makers.
What is the frontier economy and how productive is it?
Frontier companies are businesses operating in emerging, cutting-edge sectors: AI, life sciences, ‘clean tech’ (the innovations to support reduced carbon emissions), advanced manufacturing and the dozens of adjacent technologies that are likely to define the economy of the next decade. What distinguishes them from other businesses is not simply their sector but their nature: they are typically highly innovative; they sell beyond their immediate local market; and they are capable of rapid scale.
Using company accounts data to compare output per worker across frontier and non-frontier companies shows that frontier businesses are around 60% more productive than non-frontier parts of the economy (see Figure 1). This holds across almost every cutting-edge sector individually, and ‘ad tech’ (the tools of digital advertising), engineering biology and biopharmaceutical companies are more than twice as productive as the non-frontier average.
The frontier average is higher than every broad sector. In fact, it is the most productive part of the economy.
Figure 1: Frontier sectors are considerably more productive than the national average
Source: The Data City.
Note: Estimates based on company accounts data. Frontier sectors with fewer than 5% of firms reporting full financials are excluded.
Frontier companies’ higher productivity is not surprising given their innovation intensity. Analysis of company website language – the samesses are at least four times as likely to exhibit signs of active innovation as other companies in the same SIC classification
This innovation premium is itself likely to be a driver of the productivity gap: innovating firms tend to grow faster, generate higher revenues per employee and attract more external capital than their non-innovating peers (Haskel and Westlake, 2018).
Why does the frontier economy matter for understanding the UK’s productivity malaise?
Static productivity estimates say nothing about productivity growth. And flatlining productivity is the economic challenge of our times. Frontier activities play an outsized role in this too.
To understand why, it first helps to divide the private sector into two broad groups:
- Export-base sectors, which sell beyond their immediate local market: car manufacturers, software developers, insurance companies and pharmaceutical firms. The revenues of these businesses flow in from elsewhere, bringing new money into an economy.
- Local service sectors, which sell primarily to people nearby: cafés, pubs, hairdressers and shops. These businesses recirculate money already in the local economy but cannot, on their own, grow it.
The distinction matters because of their varying contributions to productivity growth. As Figure 2 illustrates, local service sector productivity has barely moved since 1997 (when the data series begins). This is not surprising: a job serving tables in a café today does not look meaningfully different from three decades ago. The fundamental constraint – how many customers one member of staff can serve in a shift – is largely unchanged by technology.
Export sector productivity, by contrast, rose sharply between 1997 and 2008 – precisely when UK aggregate productivity grew – and it has flatlined since 2008, precisely when aggregate productivity growth stalled. It is exporting sectors, not local services, that determine the national productivity trend (this pattern holds internationally too).
Figure 2: Export sectors determine national productivity growth
Source: Office for National Statistics (ONS).
Note: Real output per hour, 1997=100. Definitions of export and local service sectors follow Swinney (2018).
Frontier companies are a subset of the export group, and the most dynamic subset at that. Identifying their specific contribution is difficult using official data alone, since they are not defined as a category in national statistics.
Figure 3 offers two proxies: the productivity performance of the SIC sector in which frontier businesses are most prevalent in absolute terms (computer programming – SIC 62); and a grouping of four frontier-intensive manufacturing and research sectors (pharmaceuticals, computer and electronic products, electrical equipment, and scientific R&D – SIC 21, 26, 27 and 72).
In both cases, these frontier-intensive sectors have comfortably outgrown the average for the export sector since 1997. But they too have stuttered since 2008.
What precise role the frontier has played in the overall productivity slowdown of exporting activities is hard to say. But that’s precisely the point – because they aren’t defined in official statistics, we aren’t able to get this insight, with implications for how we understand the UK’s productivity slump.
For example, the absence of data has allowed the distracting narrative of the ‘long tail’ of low productivity firms to take hold as a reason for the UK’s productivity problems. But what is shown by the data that do exist is that it is the slowdown of high-productivity firms that has been the problem.
Figure 3: SIC sectors intensive in frontier companies have seen higher productivity growth than the export average
Source: ONS.
Note: Real output per hour, 1997=100. Frontier-intensive sectors: SIC 21, 26, 27 and 72. Telecommunications (SIC 61) also qualifies, but its figures are distorted by spectrum licence sales and it is excluded.
Why does the rest of the economy depend on frontier success?
Frontier companies also play an outsized role in employment, accounting for just 3% of UK businesses but 14% of their employment. This of course means that the vast majority of workers don’twork in these sectors. And so it’s not uncommon for people to claim that these sectors have no impact on other parts of an economy.
But their economic importance extends beyond their direct employment through what economists call the local multiplier effect. Like all exporting sectors, frontier businesses bring money into economies from sales made elsewhere – to customers in other cities, other countries and other industries. That income circulates through local labour markets: frontier workers spend wages in local restaurants, shops and services, generating indirect employment in sectors that do not themselves export.
Figure 4 makes this relationship concrete, echoing the work of Enrico Moretti (2010). Across the UK, places with higher concentrations of frontier businesses have substantially higher levels of local service employment.
Figure 4: Places with more frontier businesses have more jobs in local services
Source: The Data City; ONS.
Note: RTIC companies per 10,000 population, 2026, against employment in local services per 10,000 population, 2024. The four ITL-2 areas in Greater London are combined to have a better approximation of London’s economy.
What does the investment evidence tell us?
If the productivity and growth arguments above are not fully visible in official statistics, the signal from private capital markets is considerably clearer. Equity investors – with money at stake – have already formed a view about which part of the economy is most likely to generate future value.
Frontier companies have attracted 41% of all equity investment deals into UK companies since January 2025 (see Figure 5) and 56% of total deal volumes by value over the same period.
Figure 5: Frontier sectors account for disproportionate shares of jobs, output and investment
Source: The Data City; Specter.
What are the policy implications of frontier invisibility?
The employment and productivity evidence shows that far from each constituent part of the UK economy being the same, they play different roles. And there is a hierarchy to them, with the fortunes of the frontier in particular having a downstream impact on other parts of the economy. In this sense, the frontier is like the keystone of an economy, which makes understanding its performance very important. The success of the UK economy is driven largely by activities hidden in official data.
Given this, it is clear why the fortunes of the frontier should be factored into economic commentary and policy decisions. For example, the setting of interest rates by the Bank of England’s Monetary Policy Committee (MPC) could consider the prospects of frontier companies. Tax policy devised by the Treasury could be influenced by it too.
It could shape forecasts produced by the Office for Budget Responsibility (OBR). It could define the purpose of industrial policy for the Department for Business, Innovation, Science and Trade (DBIST). And it could shape sub-national growth policy by Number 10 North and local government. But for now, it isn’t because the frontier isn’t classified in official data.
The business department (under its previous name) has itself acknowledged this problem: the Industrial Strategy notes explicitly that ‘the SIC system does not adequately cover many of the sectors’ (Department for Business and Trade, 2025). Whether the right companies are receiving support; whether clusters are being correctly identified; and whether investment is flowing to the most productive firms – none of these questions can be answered from official data alone.
How can the frontier economy be made visible?
There are now techniques and technology that make it possible to measure and monitor the frontier of the economy. But for this approach to influence the policy-making process, there has to be demand to use it too.
So I am looking to build a coalition of economists and policy-makers to make the case for the requirement of these data. If you are persuaded by the argument above and would like to join this coalition, then please get in touch.
Where can I find out more?
- The wrong tail? Why London isn’t too big – and how the rest of the UK could compete: 2018 article by Paul Swinney for the Centre for Cities.
- Capitalism Without Capital: The Rise of the Intangible Economy: book by Jonathan Haskel and Stian Westlake, published in 2018 by Princeton University Press.
- The UK’s Modern Industrial Strategy: 2025 plan from the UK government’s Department for Business and Trade.
Who are experts on this question?
- Diane Coyle, Bennett Professor of Public Policy, University of Cambridge
- Anna Valero, Director of the Growth Programme, Centre for Economic Performance, LSE
- Stian Westlake, Executive Chair, Economic and Social Research Council
- John Van Reenen, Professor of Economics, LSE
