Number 10 North is the core of Andy Burnham’s plans to remake UK government. As he says, it’s ‘the first major act of this government and what I hope will be seen as its most defining.’

This is because it’s part of a much bigger process of rewiring the state. Part 1 is expanding devolution to cities and towns, including, crucially, localising parts of the tax base. Part 2 is beefing up Downing St with a new Office of the Prime Minister and Cabinet. Part 3 is moving a big chunk of this out of London to Manchester. [1] Up to 150 civil servants will work in Number 10 North in the future; that’s about half the current Number 10 workforce. [2]

Here’s the PM again:

That’s the intention. In this post I want to look at how relocating half the heart of government might work out, and how we could follow its effects.

TLDR; there are four things to track:

• First, the microgeography — what happens when you split an organisation across two places?
• Second, will relocation really change the overall tone and direction of government policy?
• Third, how much of the wider media-policy complex will move too — journalists, think tanks, lobbyists, unions, big firms?
• Fourth, how might all this relocated activity change political dynamics, or reshape Greater Manchester’s economy?

I don’t have strong priors about how 1–3 will play out — existing work gives us clues, but not much more. On 4, we can learn something from the BBC’s move to Salford. My guess is that the economic impacts are likely to be small overall — but could also be highly localised, and could really shift things in a few neighbourhoods. We need to do the research!

NB I don’t want to talk devolution in detail here — others have said it better than me. [3] I do want to emphasise how much devolving locally depends on changing how central government is organised, specifically the No 10 ~ Treasury relationship. Setting up OPMC is a power move. Burnham’s team clearly know this: I’m glad they pulled the Devolution lever and the one marked Whitehall Change.

The power of proximity

Splitting the PM’s department across two cities brings up some tricky issues of microgeography. Historically, knowledge work needs people in the same room. Face to face interaction is richer; people can work more flexibly and intensively, responding to rapidly changing situations. One of the reasons cities still exist is because they help this kind of work get done better and more productively.

The same logic applies to large organisations where this kind of work is done. Recent evidence from lab moves and remote working / RTO suggests even quite small shifts in physical proximity have major effects on collaboration and learning. [4] Co-locating research labs increases collaboration by 3.5 times; separated colleagues collaborate less. Even sitting next to coworkers improves learning and productivity compared to teams split across buildings, especially for more junior staff. Studies showing remote workers are just as productive as before typically focus on routine tasks, like call centre work or patent examining.

All of which suggests things might go very badly. So what’s the more positive case? First, thanks to the massive forced experiment of COVID, civil servants understand how to do remote working. Second, remote teams work better if team members have already built working relationships through face to face. That suggests Number 10 North is going to build up rich working relationships very rapidly in the early stages. Even Andy Burnham may need to spend more than the planned day a week there. That’s good news for Avanti West Coast.

How will we know if it’s working? Systematically tracking No 10 workflows is basically impossible for researchers, since so much will be confidential. Instead, I suspect we’ll be relying on what political scientists call ‘process tracing’ and the rest of us call political journalism — insider accounts of how work gets done, who’s up and who’s down, and how far the physical split is help or hindrance.

Note that No 10 North involves shifting functions, as well as locations. If everyone working on devolution and local economic growth is in the same building in Manchester, in a relatively self-contained way, does the site split matter so much? Which brings me on to …

Different and better decisions?

If devolution is the answer to over-centralisation, why relocate parts of central government to change the direction of policy?

The reasoning goes like this. If we change where decision-making happens, different people make those decisions. A big argument for devolution is that it improves decision-making because local conditions vary, local knowledge is brought to bear, and elected local leaders have electorates to answer to. Similarly, central government decisions may improve too if different perspectives and knowledge bases are brought in.

These aren’t new arguments, and neither are they Labour’s: I’ve been paraphrasing Michael Gove’s 2020 Ditchley Lecture, which argued that relocation would change civil servants’ metropolitan mindset — and incentivise officials to make more effective policy:

In the case of No 10 North, the staff leading devolution and growth policy are GMCA veterans who’ve done it for their home city already, understand what it’s like to be on the receiving end of Whitehall, and know the challenges of implementation. There’s something to this!

What are the challenges in making it work?

• I’ve seen some discussion about whether the Manchester site can staff up — will civil servants want to move out of London? The ONS move to Newport is a cautionary tale — there was a lot of attrition — but in this case I’d be more relaxed. For one thing, this is a gradual buildup in a city with over 50,000 students — that’s a big potential workforce for the planned number of No 10 jobs. For another, the key staff are already on site — one of the logics is that they don’t want to move to London.
• The more serious concerns are split sites and symbolism. I’ve talked about the first already. Alex Thomas has a good piece on the second. He cautions against two risks. First, real power stays in No 10 South. Second, the Treasury continues to dominate economic decision-making.

How to configure growth policy across government is already a point of contention. No 10 North formally takes on devolution and local growth, but will also steer national growth policy:

But how will that happen in practice, since the Treasury is — to put it gently— used to doing that job? If devolution helps cities grow, that helps UK Plc. And industrial and science policy runs through places where priority sectors and research are clustered. From that point of view, all growth policy is place-based and No 10 should drive it. But if HMT continues to run macro policy, and keeps teams shadowing all spending departments, then it’s hard to see how No 10 can steer effectively.

From a machinery of government point of view, either this gets solved structurally — formally passing functions to No 10 — or it gets worked around politically, with Burnham, John Healey and key officials agreeing and driving through a shared agenda. Observers say this is still being worked out inside Government.

How will we know if it’s working? Again, this will be hard to parse; again, conversations with insiders will be valuable. But here I think we could push things further. In the jargon, we can think of No 10 North as an exogenous policy shock — not expected or foreseen, and not trailed before Burnham became leader. This gives us a relatively clean setting to try and back out impacts since the announcement.

• First, government language could look and sound different. We could track how official discourse is evolving in No 10 (physically decentralised) versus say HM Treasury (London-dominated) on the policy areas both departments keep working on. [5] Since Burnham became PM, have the two organisations diverged in — say — tone/sentiment, vocabulary, and the spatial distribution of examples/evidence? And how far? Natural language processing of policy documents would help quantify these differences.
• Second, patterns of public spending could become less London / South East-centric. An event study could look at the spatial distribution of departmental spend before and after No 10 North was set up, for example across business support, innovation and industrial policy. No 10 leads on devolution and local economic growth; in practice, it will need to steer decisions in other areas of policy too, including those where HMT or other spending departments may maintain influence. The trick would be to separate out budgets and programmes that stay national versus the increasingly large slices devolved to mayors.

Note that both of these designs track policy divergence — neither directly looks at whether outcomes get better, although both could be developed to give insight on that too. What will be challenging is that everything’s confounded with other machinery of government changes, especially the creation of OPMC, as well as the effect of a small number of key personalities across both sites.

Uprooting the wonks and the lobbyists

How No 10 North might affect Greater Manchester [6] itself? Specifically, how much of the media-political complex — journalists, wonks and think tankers, lobbyists, unions and industry groups, or even foreign governments — will now set up shop in the city-region?

Joshi Herrmann has a fun Mill piece speculating on how this might go. He’s joking-not-joking, but actually it’s an extremely helpful framework. It’s plausible at least some news orgs will set up more local offices; some London-based think tanks, unions and lobbyists and groups like the CBI might also open new branches, especially if their business touches on devolution, industrial strategy or local economic growth.

As Joshi suggests, it’ll be fascinating to see how firms — and workers — in the area respond, given their deeper local knowledge and connections. Some of that might come through hiring — key No 10 North staff have already been poached from the Combined Authority, and London firms might copy that playbook. Manchester has already birthed at least two high-powered consultancies, Metro Dynamics and GC Insight; will new firms spawn to compete with them? What will think tanks like CLES do? Will Manchester Uni spin out a group like City-REDI? (Full disclosure — I’ve worked in both these places.)

It’s also worth looking for changes in the wider political economy. It’s possible that patterns of lobbying, MPs’ outside interests and political donations will change, and become more GM-centric. As Jennifer Williams and others have said, devolution raises the need for proper local scrutiny. This also applies when central government moves. How will media in Manchester and the wider North West step up, building on the dynamism of outlets like The Mill? How will national outlets develop new ways of working to cover twin centres of power? Finally, as has already happened with Metro mayors, will voter awareness shift, and what will this to do turnout in elections?

How might we answer these questions? The first step is monitoring — checking counts of new and relocated companies, jobs and ownership structures, tracking job postings, following lobbying and MP financial interest data, and so on. I’ll say more about tracing the economic impacts below. The wider political economy impacts are outside my domain, so for now I’m sketching the landscape rather than plotting a detailed route. [7]

The economic effects of relocation

Governments around the world use public sector relocation to drive local economic growth. [8] How might this affect local economies? First, there are more jobs in the relocation site. Second, relocated workers spend money locally, and departments buy services from local firms. Third, if there’s a larger cluster, we might get spillover effects on innovation, or productivity, or via startups and other firms moving in. That’s the upside. Against this, relocated activity competes with existing businesses for workers, space and market share. Jobs and firms might just be shuffled across space, rather than new activity created. Incumbents could be displaced by newcomers. And there might also be wider effects on local housing markets — good for homeowners and landlords, not so good for first-time buyers or renters.

Though it’s not a perfect parallel, the BBC’s relocation to Salford gives us some basic priors on the potential economic effects of No 10 North. Over 1500 staff relocated to MediaCity in 2011–12, with a further 600 new roles; around 4,000 staff currently work on site. In a study looking at economic impacts five years after the move, CEP colleagues and I find:

• Every 3 BBC jobs generated another job in Salford’s creative industries, and the number of creative firms (and firm size) also went up.
• Some job creation in adjacent tech activities. Incumbent firms were more likely to stay active.
• No evidence of any local spending effects. Non-creative jobs grew, but more slowly than without the relocation.
• No evidence of spillovers or displacement across the conurbation; effects were almost all highly localised to Salford.
• Effects on local housing markets were unclear, and hard to separate from larger shifts in flat-building around Salford Quays.

So what does that imply? The move adds up to 300 civil service jobs: this is far too few to move the dial, as this early evaluation of the Darlington Economic Campus makes clear. So any economic boost is going to come from the combined effect of government activity plus growth in media-policy world. Let’s imagine that in total, at least 1k jobs are created or move. That might be enough to trigger some halo effects through wonkish spillovers, supply chain shifts, local spending, or all three. As above, there will also be some creative destruction as arrivals from London compete with local companies. And we might expect some real estate shifts, both in the office market and in some layers of the GM housing market.

How big? I’d guess all of these effects will be fairly small — but a lot depends on how much activity, public or otherwise, moves or springs up around the Number 10 site. I’m more confident impacts will be highly localised to milieux around Heron House, New Islington and desirable neighbourhoods across the city-region. Journalists and policy types are highly clustered in London, both where they work and where they live, and you’d expect to see that in GM too. I’d also guess some of these effects, especially in office markets, will be visible before the new permanent site in Ancoats comes on stream.

To test the causal effects of all this, we need to compare changes in Manchester against what would have happened without No 10 North. The challenge is that — in the jargon — this is a one-off shock where there’s no obvious counterfactual. Just like the BBC’s move, it doesn’t look like any other city was considered as a site, even though there is Government real estate many cities outside London; we know Burnham’s desire to bring in trusted GMCA staff drove the decision. But this means we can borrow our BBC design — or something like it. This means a synthetic control, synthetic diff in diff or trajectory balancing type design, where you compare outcome changes in GM (or Manchester city) against a ‘synthetic’ control area built from a weighted average of other cities. Note that London can’t be used here, since it’s on the other side of the relocation. Given how London-centric media and politics are in UK, this may make it harder to build the synthetic GM we’d need.

Taking it seriously

All of this assumes that Number 10 North is a real change, not a gimmick. From a PM who cut his teeth in the New Labour era, you might well be cynical. However, from everything we’ve seen so far, Andy Burnham is deadly serious. What’s proposed is very close to Devo Max for England: it’s significantly beyond anything Mayors or devolution advocates have asked for. On top of that, tying decentralisation to a written constitution and universal basic services only underlines the scope of what’s in play. I hope I’ve clarified how to think about one part of it; how it will actually go is still wide open.

Edited 14/8 and 24/8 – many thanks to Kieron Flanagan, Simon Hooton and Jennifer Williams for helpful comments.

Footnotes

[1] The key machinery of government changes are as follows. Number 10 North will focus on local economic growth and devolution strategy, taking these functions out of MHCLG and HM Treasury. It will be led by Caroline Simpson, formerly head of GMCA; other key GMCA staff will also move over. In the future up to 300 civil servants will work in Greater Manchester. No 10 North will be part of the new OPMC, led by Louise Haigh, the Deputy PM, alongside the old No 10 functions, Cabinet Office functions serving the PM + Cabinet, and the new AI Taskforce. Other Cabinet Office functions will move to relevant spending departments. A OneGov Delivery Agency will be set up within the Cabinet Office. Burnham is also reviving the National Economic Council, bringing together ministers and city-regional mayors for regular meetings.

[2] Source for the 150 number. No 10 South numbers from IfG 2023.

[3] See e.g. Jennifer Williams on Rewiring the State, Oscar Selby and Ant Breach on fiscal devolution; Sam Freedman on devolution and public service reform, and Jack T Shaw on boundaries.

[4] For a recent review of other evidence, see Madaleno et al 2022.

[5] I’m assuming that the Darlington Economic Campus is essentially a satellite office, not a decision-making centre in its own right. Over 750 HMT staff now work in Darlington, around 25% of the total headcount. No 10 North will have around 300 staff, a much smaller number but likely a much larger share of the overall No 10 workforce. Sources here, here and here.

[6] I’m using the term for the conurbation deliberately — in practice I suspect any impacts will be very localised.

[7] I am working with colleagues at UCL and CERGI tracking some of these channels in Georgia, where Parliament was relocated twice during the 2010s.

[8] Relocation includes both moving capital cities (e.g. Bonn to Berlin) and more commonly, moving bits of government out of capital cities. Freitas (2025) reviews both types of move across OECD countries. Often there’s a cost-cutting goal too — governments save money by moving back-office jobs to cheaper locations.

I have a new working paper out — the most technical thing I’ve done. Written with a big team spanning UCL, LSE, Birmingham, Milan and Toronto, we build a large, unique data platform that tracks 800,000 workers across 10,500 UK firms and places over nearly two decades. It’s a major new piece of data infrastructure, with very wide applications, and that can be used alongside existing administrative microdata.

Here’s the link to the paper.

Below I talk more about why this matters, how we built the platform, and what we’re doing next.

Why this matters

Linked employer-employee datasets are a hugely powerful tool for researchers and policymakers. Being able to track workers through careers, across firms and across places enables us to answer a vast array of questions: for example, what drives firm productivity; how gender and qualifications influence what your earn and your progression, or lack of it; gains from working in big cities; how clusters form and evolve. Governments and national statistical agencies around the world build these data and use them for policymaking and research.

However, access to such data is typically highly restricted, with anonymised subjects and limited options for linkage; and where it exists, the quality and coverage varies substantially across countries. (The UK, in particular, has very limited administrative data tracking workers across firms, although this is slowly improving.)

Platforms built on open / online data can be a powerful complement to these high-quality but constrained administrative sources. We’re not the only people to have spotted this — there’s a growing literature linking online worker resumes with company registers to look at (for example) the effects of AI adoption, non-compete agreements and labour mobility, and the drivers of gender skill gaps.

What we did

To build GRAPH-EE, we combine information from a world-leading knowledge graph of the public internet — essentially, a database storing linked entities and their relationships — with open company data. The workflow is complex — see the full paper for details — but the core steps are as follows.

We generate the universe of larger UK companies from Orbis Historical and OpenCorporates between 2007 and 2023. The UK company registry, Companies House, is open data, and Diffbot includes these records in the knowledge graph. Using company IDs, we search Diffbot and pull organisation profiles and all-time workforce profiles. For companies where we can observe at least 25% of the all-time workforce, we build detailed worker profiles. Worker info includes age, gender, location, country of birth, full education and career histories.

We use machine learning to classify worker occupation codes from job titles, and to classify an array of specialist and non-specialist skills that workers have learned through their careers. (The figure below shows wordclouds for these skill clusters, and a radar map of how migrants and UK-born workers differ by major skill complexity).

At company level, we build measures of productivity and patenting, as well as company-level workforce characteristics. We do extensive validation on every critical step in the process, using a combination of human RAs and LLMs.

Next steps

There are many many things one could do with this data. For our UKRI project, we’ll shortly be putting out working papers on migrant workers, innovation and productivity in UK firms [early draft here]; the effects of the UK gender pay gap legislation [early draft here]; and the contribution of migrant entrepreneurs to high-growth company formation and performance [early draft here]. Beyond that, team members are already sketching out a lot of other ideas. What’s also neat is that many countries have open company registers, and our workflow could be adapted to build cross-country panels, opening up an array of other possibilities.

We’ll also be releasing a company-level version of this data, and the underlying code, so other teams can build on what we’ve made. Please keep a look out for that.

A quick PSA – I am super-excited to be joining the editorial board at Economic Geography. Many thanks to the editors for having me!

Economic Geography is the leading outlet in the field: it consistently puts out deeply-researched, frontier work on cities, urban economies, spatial inequalities, the forces that shape place trajectories, and how policymakers and citizens can grapple with these.

It’s also a journal close to my heart, as it’s where Neil and I published the first bit of work from my PhD: this paper on diversity, innovation and entrepreneurship in London firms, all the way back in 2013.

I’m really looking forward to shaping the journal’s future direction, and engaging with new ideas, data and debates.

Please keep sending us your best work!

Time capsules usually stay sealed, but I’m happy this one is open. Sleep Walk Sleep Talk is Suki Chan’s atmospheric film essay on post-crash London. Shot in 2009, it’s back online at FVU for the next month.

Suki’s film combines beautiful images of London after dark and the voices of Londoners talking about urban life, night work, freedom and routine.

She also interviewed me, in the kitchen of my old Hackney flat, and you can hear my voice weaving in and out of the audio mix, as I talk — a little hesitantly — about how cities work.

I was in the early stages of my PhD in 2009, reading a lot of urban economics and a lot of critical urbanism, both of which turn up in the commentary. Listening back today I sound both a lot younger and a lot less sure of my ideas — perhaps they weren’t really mine at the time — but it still holds up much better than I’d feared.

The film captures a particular moment in the capital, too, right down to the Burial-esque soundtrack. But it also manages to live outside of that time. As Steven Bode suggests in his superb companion piece, we see ‘the future non-stop city in embryo’. And Suki brilliantly captures that particular feeling of moving — immersed, but detached — through illuminated, emptied-out spaces, then back into the dark metropolis.

Watch it here.

Do migrants, and urban diversity, help seed new ideas? Tom, Ceren, Abi and I have written an overview of the field.

It’s part of this excellent new book on migration and innovation, edited by Francesco Lissoni and Andrea Morrison, featuring many stellar contributors. You can read the introduction for free here.

Our chapter looks at how migration and innovation are linked, and how that plays out in city space. The urban scale is closely linked to the firm scale: economic activity clusters in cities, and cities create spillovers that help firms become more productive. At the same time, cities are hubs in larger systems of production and flow; so we talk about cities and globalisation too. We identify a number of gaps and spaces for future work. Finally we do some speculation how two rolling shocks to city economies — automation and hybrid working — might shift these urbanised diversity ~ innovation linkages.

Happy reading!*

[*] The piece is paywalled – so let me know if you’d like a copy.

30 years of hurt

Here is David Baddiel in the 1990s. Here is me.

I’ve been mistaken for David Baddiel since I was a teenager: well over 30 years now.

Friends at school thought the resemblance was hilarious. People in the street would often shout “David Baddiel!” to me, especially if they were drunk on nights out.

The level of Baddieling closely tracks David Baddiel’s public profile. It was especially intense in the 1990s, when David Baddiel was on TV a lot doing the Mary Whitehouse Experience, then singing ‘Three Lions’, doing football banter with Frank Skinner and so on. It peaks each time David Baddiel has a new book out or is doing another show.

Looking better than ever

Here is David Baddiel in the early 2010s. Here is me.

David Baddiel is 12 years older than me. When I still used Facebook, as an experiment I would occasionally swap my profile picture for one of David Baddiel. I would receive many compliments. “Looking better than ever!” people would say.

I have been Baddieled in many places: in the street, on public transport, in shops, online, when passing through security, and while receiving medical treatment.

In my early 30s I had a job near Covent Garden. One day after work I went to buy a pair of shoes. Hugh Fearnley-Whittingstall was also in the shop, trying on several pairs of Chelsea boots. When he saw me he smiled broadly, then looked offended when I didn’t respond. He continued to give me hostile stares while buying three pairs of the boots.

A while back I was travelling and had to get several painful vaccinations. The kindly nurse was convinced I looked a lot like someone on the TV, who also came to the clinic for jabs as he travelled regularly for work. As each needle went in, she would guess the wrong name. Finally she shouted “David Baddiel!” as she innoculated me against yellow fever.

It is a bit funny

Here is David Baddiel recently, and here is me.

A few years ago we lived in Crouch End, a London neighbourhood popular with people working in the media and TV. I would sometimes see very famous people walking down the street. Now and then one of them would catch my eye and stare back at me, looking puzzled. Sometimes they would wave.

I recently discovered that David Baddiel is married to Morwenna Banks. Morwenna Banks is smaller than David, blonde, wears glasses and is very pretty. In fact, she looks a lot like my wife. One of the many things I love about my wife is that she did not notice any Baddielishness when we met, and people had to point this out to her.

Are you thinking what I’m thinking?

But here is a puzzle. It strikes me, too, that I don’t really look much like David Baddiel in those early photos. And yet the regular Baddieling. Why might this be?

A scientific answer is that we tend to recognise and place people based on a few key features. Models of facial recognition often use this idea, and many studies now support it. I have similar glasses, dark hair, olive skin and now, a beard. Perhaps that’s enough.

A more sociological answer is that I look more like David Baddiel — to the Baddielers — than anyone else around me at the time. On the Wirral in the 1990s, in particular, there weren’t many dark-haired, olive skinned, vaguely Jew-ish looking kids around. Maybe people were trying to place me, and this was the quickest shortcut. Now I live in North London, maybe people expect that it really is David Baddiel they see on the street.

A more disturbing answer is that I am growing into the resemblance. Or growing myself into it: decades of Baddieling may have led me to try and look more and more like David Baddiel, without ever being fully aware of doing so. A therapist might say that I am trying to make myself intelligible — to myself, to society, or maybe both. A critical theorist would see this as a form of interpellation —society continually labels you, and eventually that’s you who are.

In fact, David Baddiel himself is also routinely mistaken for other people. So the most cosmic possibility is that both of us are part of huge chains of confusion that stretch around the world.

I sometimes wonder when all this Baddieling will end. Perhaps I will be Baddieled in the post office while drawing my pension. Or in hospital on my deathbed; a doctor will be checking my vital signs when suddenly a strange look of recognition and excitement will come over her face. Perhaps my last words will be: no, actually that’s not me.

PS: I don’t usually write about this kind of thing. I’m an academic: I post about urban policy and urban economies, not reflections on identity. Think of this post as a kind of cross-training: working through different ideas, and trying on other styles of writing.

Relocating big firms and scientific infrastructure can lead to large spillover effects on local economies, especially over the long run. Can shifting public sector activity also generate halo effects?

Public relocation programmes are more widespread than you might think: Austria, Canada, Denmark, France, Germany, Korea, Mexico, Norway, Spain, Sweden, the UK and Zimbabwe have all tried — or are trying — moving bits of the state around to boost local economic development.

So it’s important to understand how effective these tools are. We have some evidence, but it’s still quite thin.

I have a new CEP working paper — with Henry Overman, Capucine Riom and Maria Sanchez-Vidal — which looks at the early impacts of one of the UK’s biggest recent relocations: the BBC’s move of key departments, and over 1,700 staff, from London to MediaCity in Salford.

Here’s the abstract:

This paper considers the impact of a major public sector relocation: the British Broadcasting Corporation’s partial move from London to Salford, Greater Manchester, starting in 2011. We identify effects of the move using synthetic control methods applied to plant-level data at Local Authority and Travel to Work Area level. Each BBC job creates on average 0.33 additional jobs in the creative industries, rising to 0.55 additional jobs by 2017, and the relocation had an impact on sectoral and firm composition. We find no significant effect on total employment but a small positive effect on Local Authority average wages.

The full paper is here.

If you don’t want to read the whole thing, you have a few options:

• Henry, Capucine, Maria and I wrote an explainer for the Economics Observatory. This was fun to do. We cover the paper, and look at multipliers and cluster-building efforts more broadly — drawing on historical case studies and more recent evidence from across the OECD.
• I gave a short talk on the paper for What Works Growth earlier this month, with Victoria talking through multipliers more broadly — link to the video is here.
• Or you can check out the media coverage — The Observer and the Financial Times have both done nice write-ups.

Urban economists vs. economic geographers, and economic geographers vs. themselves

Last month Neil Lee, Helen Simpson and I recorded a Centre for Cities podcast on the state of urban economic research, expertly helmed by CFC boss Andrew Carter.

Two big things we didn’t have time to talk about were 1) how to define ‘research about urban economies’; and 2), how economic geographers and urban economists relate to each other.

I think this stuff is important — for academics, certainly, but also for those who are users of academic work.

We did try out some things we might say before hitting record, and if you *are* interested in this stuff, here’s what I would have said …

In the podcast we kicked off by reflecting on why the urban economies research space has got so much bigger in the last 20 years.

We agreed that a big part of that story has been the spread of better, more powerful quantitative frameworks, tools and data (new and old) to explore urban places. It also reflects big shifts in cities’ material reality and public perception. The 1990s and 2000s saw vast urbanisation, especially in South and East Asia, which has helped take the world’s population majority-urban. In the global North, many large cities stopped losing people and began economic recovery alongside high-profile physical regeneration. Both of these shifts helped thrust cities into the public, media and policy conversation, especially in the US and UK.

We could even — reaching a bit — trace a Line of Urban Influence through the commentariat, from the foundation of Brookings Metro in 1998 and The Rise of the Creative Class in 2002, the launch of The Atlantic Cities in 2011, the publication of The Triumph of The City in 2011, and the launch of Guardian Cities a few years later, in 2014. Centre for Cities itself falls neatly into this schema, launching within IPPR in 2005 and spinning out in 2007.

Academics also notice and respond to these shifts! But different urban research tribes have reacted in different ways. So I think it’s also useful to think about *how* the urban economies topic space has got bigger in different parts of that space. In turn, that helps give us a clearer idea of what we talk about when we talk about ‘research about urban economies’.

The two biggest parts of the topic space are urban economics and economic geography. Cards on the table: I’m an economic geographer, but I use a lot of economics in my work, both frameworks and methods. I like data, and mainly use quantitative designs, but I also do mixed methods work with co-authors from qualitative backgrounds. So in this debate, I’m econ-adjacent — but not a proper economist.

Over the past two decades, urban economics has become hugely larger as a field in econ. One metric of that is conferences. I was at the first Urban Economics Association event in 2009 in San Francisco — it was a couple of rooms in a much larger regional science conference. There were maybe 50 people there. Now UEA runs two conferences a year, one in the States and another in Europe, with hundreds of people at each.

One reason those events have got so large is that we have better theory, tools and data about cities to work with, a point we picked up in the podcast. The 90s and 00s wave of urban economics partly came about through a highly productive splicing together of 90s New Economic Geography models (coming out of trade and endogenous growth theory) with an older literature on urban systems and city models, dating back to the 60s and 70s (see here, here, and here). The current wave of urban econ has also got bigger because it’s more and more woven into other parts of economics, in very productive ways. In urban conferences — and on the job market — you now find macro people (often working on housing markets); public economists (working on government or place-based policies), IO people (working on firm structure and location, but also on urban structure), as well as labour and innovation folks. In other words, urban economics is now far more than its core interests of transport, infrastructure, real estate and local government.[1]

Back in 2004 Henry Overman identified two main weaknesses in urban economics — a lack of precision about places, and an under-use of case studies (and related bodies of evidence). Are these still problems? The first concern is attenuating somewhat: as spatial datasets have got richer, research designs are now better able to precisely specify appropriate units of observation (and indeed, to use these to define urban areas). The second issue is still live. To me at least, urban economics seems less insular in terms of citing outside the field, and drawing on ideas and frameworks beyond economics proper, even if these insights end up being converted into econ frameworks or formalised into models.[2] Jane Jacobs and Anna-Lee Saxenian are the classic examples of non-economists who’ve become canonical for urban economists. But it’s still hard to see many *spaces* today where urban economists and (say) urban sociologists might encounter each other and their ideas. In turn, that may limit cross-fertilisation of new ideas, as opposed to rediscovery of older ones. I come back to this at the end.

Economic geography is also larger and richer, but has evolved in a much more complex way. One way of understanding this is that economic geography is not just ‘where economic things happen and why’, but also the spatial footprint of many other fields. Everything has a geography! Again, conferences are a useful metric: you can see the breadth and richness of economic geography in the diversity of disciplines and methods on display at the huge, bi-annual global economic geography conferences.

This has consequences. Economic geographers proper have often reacted to that disciplinary overwhelm by working almost a-spatially; geography, or space, is in the background but is implicit, and rarely discussed either as a passive or active feature. This is not a new tendency — Andrés Rodríguez-Pose discusses it extensively in this 2011 piece — but it’s become increasingly embedded.

For example: I co-convene the UCL Economic Geography Group, an array of researchers interested in topics around urban economies. It’s a very rich and productive group. It’s also striking how few of us actively identify *as geographers*.[3] We have people who make maps and work with spatial data, think about urban governance, or explore economic / social formations in particular places. But usually these are planners, not geographers. The majority work on topics like global value chains; care homes; labour organising, industry 4.0; financialisation, or platform economies. Space and place are here, but are not not front and centre.

For me, economic geography’s heterodox nature is one of its major strengths: after all, the questions we look at mean that researchers will often need to draw on an array of perspectives and frames to think about a problem and develop a research design. This should lead to richer and better-grounded insights. But the shift to a-spatiality, coupled with that lack of a single, foundational body of theory, also creates a lack of definition and coherence.

The protagonists have responded to this in a variety of ways. Some have tried to create their own syncretic language or unified field theories; others circle the wagons, often defensively, especially against the more econ-adjacent parts of the discipline.[4]

Again, this reaction isn’t new, but it is still going. Last year, Henry Yeung’s otherwise magisterial review of the state of economic geography was notable — to me at least — for not engaging at all with quantitative geography, spatial data science, network/complexity approaches, or urban economics. Neither, tellingly, did any of the responses by Trevor Barnes and Shaina Potts published alongside it. We can also see these dynamics in the ongoing flame war about whether it’s even possible to do comparative urban analysis (the blue corner; the red corner): a dispute best understood as a proxy war between quant-adjacent and qual/critical parts of the discipline.

Where does this leave urban economists and economic geographers? Formally, it’s a situationship, fraught, but also very one-sided. Geographers — including me — spend much more time thinking about this than economists seem to. The last time I’m aware actual economists got involved is a while back now — Henry’s provocative 2004 piece in JOEG, and this thoughtful 2005 reaction piece from Gilles Duranton and Andrés in EPA.

But as I suggested above, this isn’t just about economists vs. geographers. The positions and dynamics within economic geography reveal another fraught relationship: economic geography vs. itself, or more specifically, the largely quantitative, econ-aligned vs. largely qualitative, critically-aligned parts of the discipline.

That might all sound a bit bleak. But I think the current picture is more encouraging, for three reasons. First, if we think about non-economists who have become canonical in urban economics, I suspect this is lagrely because their granular, qualitative work gives the microfoundations that quantitative researchers need to run regressions and build models. These kind of qual-to-quant designs can be both very rich, and an approach where boundary-spanning economic geographers should have a clear comparative advantage. Of course quants can do this too — one of my go-to econ papers on clusters, Arzaghi and Henderson’s 2008 study of ad firms in Manhattan, builds its analysis off of dozens of semi-structured field interviews with managers and workers in firms on the ground. With Georgina Voss and Emma Vandore, I tried to follow the same qual-to-quant pipeline in exploring the Tech City cluster in London (though we needed three papers to do this).[5]

Second, there is a connective web of individuals whose work and methods bridges these worlds — Maryann Feldman, Michael Storper, Dani Rodrik, Karen Chapple, Simona Iammarino and Ian Gordon, for example. All of these have been inspirational to me as I’ve threaded my way through academia. And third, just as important, there are many groups, departments and events that bridge. To name a few: the EGSE cluster at LSE, the LSE-Oxford International Growth Centre, the GEOINNO conferences, Liverpool’s Geographic Data Science group, the Turing’s SAD podcast and seminars, and the (sadly-missed) SEG group of early career researchers, have all created spaces for presentations and productive exchange across urban economics, economic geography and related fields like complexity/networks and data science. Newer urban labs like UToronto’s School of Cities are also starting to play these roles. And there will be many others I’ve forgotten or don’t know yet.

Now, perhaps it’s no surprise that a lot of the people and groups I’ve named are at LSE, which has a critical mass of spatial economists AND economic geographers. Clusters like this matter! But just as innovations slowly diffuse — through collaborations, interaction and career moves, and become mainstream over many years — so we can imagine a set of projects, groups and institutions gradually seeding more inclusive norms and ways of working across the wider system.

If you want to do work on urban economies that draws on geography *and* economics, in other words, there are many ways — and people, groups and places — to draw on. There is still a way to go. But this way is better for everyone.

Many thanks to Dani Arribas-Bel, Neil Lee and Liz Moor for thoughtful comments. The usual disclaimer applies.

NOTES

[1] The other crossover point for urban economics has been with the urban scaling literature developed by statistical physicists and empirical mathematicians, and the search for universal (or high-generality) laws governing systems of cities. I wrote about this productive but confusing topic space over here.

[2] That’s a pretty strong claim, and needs some caveats. In my experience this is a relatively recent change. It’s notably different from the picture Marion Fourcade paints a decade ago. We’re also going to see more cross-pollination outside the Top 5 journals (which Fourcade focuses on), and in specific subfields like urban. (It also strikes me as interesting that Fourcade’s highly critical take on top tier economics was published in a top tier Econ journal. You see the same with behavioural economics papers – many of Kahneman and Tversky’s key papers were published in Top 5s.) Of course econ as a whole has other challenges – with internal culture, especially toxic speech [see Alice Wu’s papers here and here, and Ederer et al here], and with discrimination in hiring and progression [punchily summarised by Nikita Sharma]. One bright spot is that urban economics is one of the most gender-balanced subfields, especially at entry level.

[3] My UCL colleague Lin-Fang Hsu has mapped this using keywords scraped from group members’ public profile pages.

[4] Anther strong claim. I’m going to speculate that this stance sometimes embodies: a) fear of number, and a worry that quantitative analysis (read: economics) is b) overly reductive, c) overly influential and d) typically used to assert right-wing political positions.

I feel sad about a). I talk about b) above; I’m sympathetic but fundamentally think this is a false binary. On c), I’m somewhat sympathetic, but I think it’s a lot more complex than this. First, I think it overstates the impact of academic economists on policy, as opposed to economists in government. Second, in some areas it’s obviously not true: the What Works Centre’s evidence reviews, for instance, show that the vast majority of economic development policy evaluations aren’t quantitative, and don’t use methods most economists would consider robust. Third, timing matters. In the short term, policymakers have set agendas, and are going to be more interested in evidence that helps them design and implement interventions. Over the longer term, there is usually greater interest in work that challenges or ‘unsettles’ that agenda. I find d) harder to grapple with, not least as studies of economists regularly show that the majority hold politically progressive views. My UCL colleague Mariana Mazzucato is a very visible example.

[5] Dani Rodrik’s book is also full of helpful ideas about how to do boundary-spanning work. See his 2023 industrial policy paper with Reka Juhasz and Nathan Lane for some real-world applications.

Academic news! I’m now — officially — Professor of Economic Geography at UCL.

Like most universities, UCL announces these in the summer; but in true academic style, new titles only go live today.

I’m delighted to get promoted. I also want to say that for me, academia is a team game. So rather than risk a Father Ted-style speech, I’d like to say thank you to my people: my family, co-authors, colleagues at UCL CASA, The Bartlett, UCL’s Faculty of the Built Environment and Centre for Economic Performance, and collaborators and community beyond. I’m hugely grateful.

I have a new City Talks podcast with Helen Simpson (University of Bristol), Neil Lee (LSE) and Centre for Cities CEO Andrew Carter.

We take on a big question: what is the state of research about urban economies? We talk about the huge growth of urban economics and quantitative economic geography as academic fields over the past two decades, and what’s behind this. Among other things, we highlight which new topics are emerging, the enduring relevance of some core urban questions, and how new sources of data — and innovative tools for processing it — are are helping researchers push forward the frontier. We also discuss the increasing demand for evidence from policymakers, and how effectively academics are taking their ideas out into the world.

The podcast is about 70 mins long — there’s a lot to talk about, and we had a lot of fun doing it. Happy listening!