This interview was conducted by Saarthak Gupta of Exporters Without Borders (ESF), which incubates entrepreneurs to start new export industries in low and middle income countries. The video version of this podcast, an accompanying article, and their ongoing research are at exporterswithoutborders.com.

Thomas Kehler started his career teaching at a New England prep school, before he landed in Uganda, where he started guiding fishing tours of Lake Albert. Here he came to know a young Joseph Nye – later a key architect of US foreign policy – who got Thomas thinking about development economics, and how to have impact with his career.

In the decades to follow, Thomas would go on to start two export industries from close to nothing. Between them, Colombian cut flowers and Chilean farmed salmon are now worth ~$9 billion a year and employ over 300,000 people. On a combination of willpower and luck, he managed to solve one of the biggest open questions in development economics – how countries enter new industries, and capture the jobs that come with them – twice.

Something slightly different today, ahead of returning to normal scheduling soon!

I’m not sure why I committed to a weekly schedule. It has certainly made my life hectic, but overall I’m glad I did. However, given some recent trips / general busy-ness, I have run out of steam. So, I’m taking a month or so off to build the pipeline back up, and I’ll be back with more episodes in August.

Thanks to the team at Talk Normal Productions, and my co-hosts Kartik, Deena, and Kurtis, as well as Michelle, for all of their help.

I’m excited to finish off the current series on Evidence, and also cover new topics. Energy is the series I’d like to do next, and I’ll certainly return to AI and growth related topics in the future. But also very open to suggestions, on people, stories, and new topics. Hit me up at ohanney@cepr.org, or comment here, if you have thoughts.

I have really enjoyed the process of starting and running this podcast. But to be honest, I still can’t quite figure out if it’s worth continuing - I’m leaning yes, but not fully convinced at this moment in time. It’s relatively cheap financially, but expensive time/energy wise. Originally I had planned to do this for half a year, at which point I imagined I’d know if it was a keeper. But building an audience takes time, commitment, and a consistent schedule. Blowing up quickly is not really a realistic goal if you’re trying to build the right way, even if we’ve taken some liberties with our thumbnails. So I’ll decide at the end of the year.

Whilst it’s easy to get disillusioned when comparing your own ‘numbers’ against the content you see online, in reality those mask a much bigger ecosystem of success, even if it hasn’t ‘popped’ and reached most peoples’ feeds. Plus, numbers matter less for VoxDev than the specific people we reach, and how useful they find the episode. Scrolling through my subscribers on Substack, there are many cool, senior people who I respect and admire, which I’d like to think is a good sign.

The modal Ideas in Development podcast now does about the same as our VoxDevTalks series by Tim Philips, which is about/just over 1000 listeners across platforms. Though that doesn’t reflect the episodes which get a lot of traction, which depends on the mysterious YouTube algorithm, where some Ideas in Development episodes have organically gotten over 60,000 views (see this recent episode on Ethiopia for example). We also experimented with boosting certain content, which led some videos to do really well, like Deena and I’s episode on AI in Africa with Rose Mutiso, which has almost 300,000 views, almost exclusively from countries in Africa. But boosted views are less ‘real’ than organic ones, so the Ethiopia growth episode probably comes in as our most actually-listened to episode thus far.

In terms of my personal favourites, which is hard as I (along with my co-hosts) picked every guest, so they are already screened on being interesting and knowledgeable:

• Vietnam’s economy: The remarkable story of the last 50 years Madame Pham Chi Lan was so generous with her time, and it was a privilege to sit down and talk for so long with someone who has seen Vietnam’s transformation first hand over the last 50 years.

• The perfect city? Ed Glaeser just churns out insights, it felt like every other sentence contained a new thought, history, or topic I could read an entire long-form article about.

• Is the industrial revolution a good comparison for AI? This was a fun exercise with Bruno Caprettini, and I learned a lot about British history (it was probably in my head somewhere, but it’s been a while since I visited Ironbridge as a ten year old).

• Bill McRaith on building factories from 1990s China to present day Ethiopia, being pitched by governments, and the future of the apparel industry - a different perspective to the one we normally hear in development, full of interesting titbits.

Another surprise has been how well the Substack has done. It has over 1100 subscribers, and the write ups of episodes now routinely get over 1000 reads, about doubling the reach of the average podcast’s content and key takeaways. These are relatively easy for me to do, as I can feed the super long podcast transcript, with a good prompt and examples, into Claude, and get an average first draft, which then requires about an hour of editing. My workflow makes this relatively smooth, as I have to watch the episode back anyways (often twice!), so I can judge whether Claude did a good job at summarising from the transcript, or whether the draft needs sections adding or removing, as well as the general adjustments for tone and meaning that are always required (at least for now).

For the next month or so, as well as listening back to every episode and liking/commenting on them all, of course, I’ll leave you with the following.

• My interview in the FT last week (such a cool/lovely experience, thanks to Sarah O’Connor and John Burn-Murdoch for having me on). We talked about Nan Ransohoff’s recent piece on the third wave of American philanthropy, my own piece (there is no randomising a technological revolution), and much more! I ended by calling for the economic research teams currently being built out within Anthropic, OpenAI, and DeepMind to actually hire some development economists - though much more needs to be done beyond that short term fix if we are to better understand AI’s impacts on development, and help governments to navigate these uncertain times. (Sorry, subscription required, my guest link has run out).

• At VoxDev, Economics Unpacked is back today for our second series, seven videos on important topics in development. Our first episode is (imho) a banger, 150 years of growth in China in under 10 minutes. We are playing around with the format and scope of these explainer videos, so feedback welcome!

• And In Development is back this week, the next Issue (Issue 2) is full of really interesting articles, so make sure to subscribe.

So, lots to keep you busy, see you in August :)

What questions should researchers be working on? Has development economics lost its way by focusing too much funding and effort on the micro? When development economists have results, how should they package those in a way that encourages take up? And what types of institutions and organisations can help with that process of ensuring the best evidence informs policy decisions?

Some easy questions then!

Rachel Glennerster has spent her career working in both the places where evidence is made, and the places where it is used. She joined me to discuss all this and more.

Subscribe now

Radical simplification, a year on

Last year, Rachel joined myself and Stefan Dercon for a conversation on the future of foreign aid, in which she outlined her proposal for a radical simplification of aid. This started from her observation of how donors work: agencies spread themselves far too thinly, across too many countries and too many projects, reinventing the wheel with every new programme design. Rachel’s alternative is for a bilateral donor to pick perhaps six or seven areas of genuine comparative advantage – not broad sectors like ‘health’ or ‘education’, but specific, evidence-backed programmes such as structured pedagogy or the ultra-poor graduation approach – and to run them deeply, in fewer countries.

A year after she first made the case, she is more convinced, not less. The reason is political.

“Congress is much more supportive of areas where they really understand what USAID or the State Department was doing.”

PEPFAR retains support on Capitol Hill because it is legible: the public can see what the money bought and how many lives it saved.

“When you basically have an aid programme that says it’s all very complicated, leave it to the experts... it’s kind of unsurprising that you lose support for aid.”

Simplification, in this view, is not just an argument about efficiency. It is a survival strategy for aid budgets in rich countries.

What counts as growth?

Where does economic growth – the thing that actually lifts people out of poverty at scale – fit into a simplified portfolio of proven programmes?

Rachel pushes back on the premise of the question, arguing that the mental model in which there are ‘micro things’ on one side and ‘growth things’ on the other is a misreading of the evidence. Human capital, agricultural productivity and energy are not alternatives to growth policy; they are its ingredients.

Rachel points to recent work tracing the chain from the Green Revolution’s improved seeds through better health and education, demographic change, and on to structural transformation. Energy is another clear case, where evidence shows that supporting firms’ growth requires providing access to good, reliable, cheap energy.

She then adds a subtler point that she thinks is missed in almost every discussion of growth strategy. Not everyone can become Singapore, because the world economy only has so much room.

“There’s sort of a slot for doing cheap manufacturing goods at any one time... It’s not like every African country can move into cheap manufacturing at the same time, particularly when a huge country like China is occupying that space.”

Researchers can study the characteristics of whoever takes the slot next, but it does not follow that every country adopting those characteristics would get one. There is a hunger among policymakers for someone to name the sector that will deliver Singapore in a decade, and you can make a lot of money by going around saying that you can, but the honest answer is that no such single lever exists.

Subscribe now

The micro-macro divide – overstated?

For those cursed with a Twitter account, you would have noticed the recent social media row over whether development economics has lost its way by privileging micro over macro. I wrote some of my own thoughts here. Rachel, who also weighed in at the time, explained why she is not persuaded that there is a real dispute.

“In most of the academic literature, there isn’t a debate about should you do micro or macro. People just want good work.”

Rachel also points to feasibility. The discipline moved away from cross-country growth regressions because they were fragile – tweaking the specification would change the answer. What replaced them is not an absence of macro but a different kind of it – structural models disciplined by micro identification, cross-country trade papers – the type of work funded by STEG, for example.

And Rachel feels that some of the classic macro questions simply got solved. Hyperinflation in Latin America dominated Rachel’s undergraduate exams; today the profession broadly knows how to end hyperinflation and how to run monetary policy in developing countries, so the papers stopped.

“To some extent you have to go where you can answer the question. And we have answered a lot of macro questions, which is great.”

Ideas come from the ground

Asked what the macro-development agenda should learn from twenty years of micro ecosystem-building, Rachel’s first lesson is about where ideas come from. The great behavioural shift of the RCT era was that researchers stopped running regressions from their desks and started spending serious time in the field.

“That’s the story of a lot of people – their best ideas came when they were in the field and talking to people.”

She credits Doug Gollin, whose STEG programme (with Joe Kaboski) at CEPR funds exactly the bridging work she wants to see, by pairing big structural models with a deep understanding of what is happening on the ground.

The second lesson is about communities. The dense networks of researchers, shared data knowledge and skilled local staff that grew up around places like western Kenya meant a PhD student could arrive and start productive work immediately – it became a learning lab where everyone stood on the shoulders of everyone before.

Share

The consensus machine

We then turn to how evidence, once accumulated, gets packaged for governments – and specifically to the evidence panel behind the Smart Buys in education. The model convenes economists, psychologists, education specialists and practitioners from across the world, runs a rigorous systematic review, and then hashes out consensus recommendations that put cost-effectiveness and scalability at the centre.

“We don’t want to recommend things based on small pilots. We want to recommend things that governments can do at scale.”

Two features make these recommendations powerful. First, they are directional rather than prescriptive – reduce the ‘time to school’ can mean building schools, community schools or bicycles for girls, depending on context – and each comes with a short account of the contexts where it applies. School meals, for instance, look very different where nutrition is the barrier to attendance than where most children are already enrolled. Second, because the World Bank, UNICEF and FCDO sit in the convening group, a government hears the same advice from every advisor in the room. The consensus is hard to build, but once built it is solid, and the lists are re-evaluated every few years as evidence accumulates.

Rachel situates the panels within a three-box framework she uses to think through evidence-based policymaking:

1. Evidence on the need, which is local.

2. Generalisable knowledge about how people should respond, which is global and is the slot the panels fill.

3. Implementation, which is context-specific.

Could the panel model work for the economic impacts of AI in low- and middle-income countries, where policymaker demand is intense but the evidence base barely exists? Rachel splits the question. More panels, yes – a skills-for-jobs panel is underway with the World Bank, and she would love one on nutrition. But panels work where evidence is plentiful and not yet synthesised. AI is not there yet, and it moves too fast for the standard research machinery.

CGD’s growing AI team has instead built an evaluation framework for funders and governments – lighter-touch questions that can be answered quickly. Some are strikingly simple. Many AI learning tools fail because teachers never use them, and you do not need a three-year trial to measure usage.

Expert-based policymaking and the economics career ladder

In the last part of our conversation, we discuss who gets to advise and who gets to produce the research. Rachel acknowledges that the standard career arc in development economics looks something like – earn tenure first, do policy later. She recalls how (her now co-author) Kate Casey had an extraordinary ability to access data and work with government in Sierra Leone – but Rachel could not even mention that in Kate’s PhD recommendation letter, as it was seen as a negative!

Rachel resists the idea that senior academics are the only node of evidence-based policymaking, pointing to panellists like Ben Piper and Rukmini Banerji who came up through consulting and NGOs rather than the standard academic ladder. But the circle clearly remains too tight.

On the concentration of research production in the global north, she believes that we should separate the objectives before designing fixes. Research quality demands deep contextual knowledge:

“It’s not about where you’re based, it’s how much time you’ve spent on the ground and how well you know the institutions.”

Building institutions so that students in low-income countries can gain expertise is a different problem, with different solutions. Things like Leonard Wantchekon’s African School of Economics tackles the maths-training pipeline, while the Weiss Fund supplements the extremely low salaries that stop trained researchers returning to universities at home. The test for Rachel is not whether a representation box is ticked, but whether poor people in these countries end up with better policy, and better lives.

We’ve spent a lot of time on this podcast discussing the policy side of economic growth and attracting investment. But what’s it like to sit on the other side of the table, across from senior policymakers, as a representative of large private sector players? What are you actually looking for from government? How do you decide where to invest? And once you have decided, what goes into setting up operations from scratch in an entirely new setting?

In this episode of Ideas in Development we are joined by Bill McRaith, who has spent his career building supply chains across the world, working on both the manufacturing and retail side of the apparel industry. He started in the UK, in factories supplying Marks & Spencer, at a time when almost all of the supply chain was local, in the same time zone and market. But that model was breaking fast, and his career then took him around the world.

Setting up a factory from scratch in 1990s China

Bill McRaith arrived in Panyu in 1990, after flying to Hong Kong, taking a ferry, then a taxi along dirt tracks through paddy fields, he reached a small factory which he had been asked to turn into something capable of supplying Western markets.

He had grown up in an industry that was already a century old in Britain, surrounded by trained mechanics, supervisors and technicians whose only real need was orchestration. In China, he had to start almost from scratch.

Anyone in the industry under the age of 50 assumes that everything was always available in China. But when Bill arrived there was no textile base, no work-study engineers, no line managers able to run modern production. Materials were imported from abroad, and every design innovation came from elsewhere. It took roughly five years of hard work, and expatriate hires running training schools across the various disciplines, simply to reach a UK quality baseline before anything could be refined beyond it.

The binding constraints were often human, not technical. In Britain the consumer was also the producer, so an acceptable standard was easy to understand for the workforce. In China the prevailing standard was acceptable for China and unacceptable for the UK, and the hardest task was getting people who had never seen the destination market to understand why a higher bar mattered.

The social adjustments were just as demanding. A three-hour midday break was standard. A move from a flat weekly wage to an incentive scheme triggered a mass walkout – a failure he traces partly to interpreters, highlighting just how important their role is.

“The labour turnover in the first few years in China dwarfed anything that I ever dealt with from that point forward.”

Labour turnover was worse in those early China years than anything he later faced in other settings, a point worth holding onto for any government tempted to read too much into initial turnover.

Subscribe now

China grew without policy

“China had very little policy at that time… the reason China was so successful is there was a lack of policy. Policy didn’t get in the way of establishing industry.”

China, Bill argues, succeeded because there was very little policy getting in the way of building industry. It was neither restrictive nor lavishly enabling, with no great tax holidays on offer. You came in and you did business, properly, and the rules were built as scaffolding around industry as it grew, rather than ahead of it.

What mattered far more than regulations was responsiveness. When goods backed up at the Hong Kong border at month-end and year-end, or when his factory had no usable road, it was local government – at that time the heads of the local communist party – who cleared the way, and tackled bottlenecks.

“I have never worked with a group of people who were more willing to help us solve those problems. And I don’t mean that in a corrupt manner. They were focused on making business successful.”

Too many countries try to get the policy framework right first, pointing at modern China as the template, and forget that China grew the other way round, with industry first and policy fitted closely behind it.

Choosing where to build

By the time Bill moved to the retail and brand side, he had built many more factories across Asia where he encountered many of the same bottlenecks as China, but was no longer surprised when they arose. Now his job included scanning the world for the next place to build. The starting point is never the supply chain: it is the brand’s own strategy. If you do not know where the brand is going in the future, there is no point asking where the factories should be.

From there the screening widens. Companies look at free trade agreements and the trends in them, because a duty saving can dwarf any labour saving. He looks at regional blocs such as the East African Community, at a country’s position on existing supply routes, at its energy mix and its commitment to renewables, and at market size – Vietnam, for example, was rapidly overwhelmed by the electronics industry, leaving little room for apparel.

Above all he resists the instinct to chase the lowest wage. The labour differential between southern China and the UK has collapsed from roughly fifty-to-one to around three-to-one, and with automation rising, cheap labour is now a weak signal at best. His summary is that “the next China is a how, not a where”: the location matters less than the model you intend to build once you arrive.

Share

An industrial beachhead in Ethiopia

Bill’s first attempt in Africa failed. An office opened in South Africa in 2000 to scope opportunities on the continent and was shut within the year. Nepotism and corruption were too high in the countries they assessed, meanwhile China was opening up and the WTO accession made staying there far easier. The return came around 2014–15, prompted by Bill’s sister – then Deputy Director of Trade for South Africa – who argued that a new generation of ministers had arrived with a different mindset. He committed a vice-president to a year of scouting for a beachhead from which operations could later expand across the continent.

Ethiopia was not chosen because it was the most capable candidate: a landlocked country with almost no materials base and few real manufacturers. What set it apart was a government whose own vision for building an industry matched the investor’s, and a credible champion in Dr Arkebe – “Every country needs a Dr Arkebe”.

“It wasn’t because Ethiopia stood out as being the most capable… It was like China back in 1990. There was nothing there. But here was a government that had a vision to build an industry that matched our vision.”

Government officials were empowered to make things happen, and Ethiopia also held the most free trade agreements, sat within a workable regional zone, and lay on the supply lines between Asia, Europe and the eastern US.

The other decisive lesson from Ethiopia was sequencing. Rather than building first and being told five years later what had been done wrong, Bill brought everyone to the table before a shovel went in the ground: government, the manufacturers and mills he recruited, government agencies, and the NGOs whose work is usually remedial – repairing damage that earlier, more ignorant investors had caused. Defining the requirements for the factories, the industrial park, the neighbouring town and the surrounding environment up front is a key part of the process.

What governments get wrong when they pitch potential investors

Because the buyers control billions of dollars of allocation, no end of countries and companies want to pitch to them – and most, Bill says, have not done their homework. A weak pitch only lists assets: we have cotton, cheap labour, we can make a T-shirt for a dollar fifty. A compelling pitch starts from the buyer’s published strategy, articulates it, and explains precisely how this country beats the fifty others competing for the same work.

Price alone does not move the large strategic players, and a pitch built on replicating the thirty-year-old China model is selling something that no longer exists.

“The supply chains that we are using today were built for the world as it was 30 years ago. And we have not yet built the supply chains as the world will look in the next 10 years.”

Apparel as a first mover, and the automation question

“Every country that has not yet (started) on that industrialisation road should absolutely be looking to the apparel industry. It is a high consumer of labour… it helps people understand quality and what it means to do it right.”

For countries not yet on the industrialisation road, Bill believes that apparel is still the right entry point. This is partly because automation is harder there, so the labour-absorbing window stays open longer even as firms such as CreateMe push the frontier. Apparel also teaches the meaning of quality, and can act as a first rung of the ladder, enabling a move into higher-value sectors.

The caveat is that the apparel model itself must change. The old system tolerated vast overproduction – apparel’s second-largest source of waste – because huge labour cost gaps made markdowns affordable for producers. With the gap closing globally, overproduction nowadays simply destroys margin - the winners will be (are) those new firms that are built for a networked, sustainable, speed-to-market world rather than the legacy players optimising for the past.

“I fundamentally believe we are in what I call an extinction event right now… maybe the biggest players won’t disappear. But they’ll no longer be the big names… they’ll be replaced by new players that arrive that are building the future, not trying to keep going the past.”

This presents an opportunity to the governments who can articulate how they fit into these supply chains of the future.

A Bill of health for LMIC’s industrial strategies

1. Lead with industry, let policy follow. China’s lesson is not getting regulation right up front, but its sequencing: industry moved first and policy was fitted closely behind it.

2. Compete on responsiveness, not incentives. A decisive advantage for Bill’s work was local officials who cleared customs backlogs and quickly built missing infrastructure. A named, empowered problem-solver inside government is worth more than a tax holiday.

3. Bring your own vision and then match it to the right investor. Articulate where your country is going and pair it to firms whose strategy fits with your vision.

4. Stop pitching on price. Cheap labour alone no longer moves strategic buyers. Build the case around free trade agreements, regional blocs, position on supply routes, energy mix and speed to market.

5. Convene everyone before breaking ground. Getting government, manufacturers, mills and NGOs at the table from the outset prevents the costly remedial work that often follows when standards are set too late.

6. Treat the apparel industry as an escalator. Its resistance to automation keeps the labour-absorbing window open, and the potential path to electronics and automotive make it the classic first-mover sector – provided the model being built is the future networked one, not the old overproduction model.

A slightly different episode today. This is a repost of a recent episode on the Africa Urban Lab’s podcast, in which Kurtis Lockhart (who co-hosted our Cities series) interviews Stefan Dercon - very much an episode I wish I’d recorded!

Stefan Dercon, Professor of Economic Policy at Oxford and former Chief Economist at the UK’s Department for International Development (now FCDO), joins Kurtis Lockhart to discuss what it takes for cities and countries to kickstart economic growth, focusing Stefan’s recent work operationalizing elite bargains in the real world. It’s a wide-ranging conversation on why growth and development depend not only on planning and policy, but on the politics that make growth possible.

All AUL Podcast episodes available here.

We will resume our usual schedule next week!

In 2022, Dean Karlan became Chief Economist at USAID, tasked with steering the world’s largest bilateral aid agency towards evidence-backed approaches. He left in 2025, as the agency was being dismantled, with he and his team having moved roughly $1.7 billion of funding towards evidence-based approaches in the process.

It is a number he describes as simultaneously big and small – a source of pride given his team started with no office, no staff, and a congressional approval process to navigate, but a fraction of the agency’s $50 billion annual budget at the time, most of which has now collapsed.

Dean, Professor of Economics and Finance at Northwestern University and founder of Innovations for Poverty Action, joined me on this episode of Ideas in Development to discuss what the push for evidence-based policy actually looks like inside institutions, the rise of embedded evidence labs in developing country governments, and whether there are questions in development economics on which we now have enough evidence.

Subscribe now

How did they move $1.7 billion? And why not more?

The constraints on moving money towards evidence were partly institutional. Dean’s team had to build an office from scratch – a process that requires Congress not to veto it – and hiring within government was slow enough that the office raised philanthropic money to bring in staff quickly.

‘We spent a fair amount of time fundraising, we spent a lot of time working through the bureaucratic process.’

Without that outside support, it would have taken two years simply to hire a team. But the constraints were also about knowledge itself.

‘My first instinct in most situations is to start by fearing how little we actually know.’

There were areas where the evidence base was too thin to make strong claims, and his team was careful not to overstate what research could tell them. Equally, there were large areas – much of global health, for instance, including PEPFAR’s delivery of medicines – where the evidence already supported what USAID was doing. Those programmes were left alone, and deliberately excluded from the $1.7 billion figure, which only includes what they themselves moved.

Share

Moving funding inside a $50 billion agency

Dean’s office used a two-pronged approach: an in-house consulting team working directly with USAID teams, and a slower effort to change institutional processes, regulations, and culture. The consulting arm vetted opportunities against three criteria:

1. The size of the money at stake – with a small team, his office had to think about its own cost-effectiveness and leverage, which meant walking away from intellectually fascinating areas like carbon markets where USAID simply wasn’t spending that much money.

2. The delta: was there a meaningful gap between what teams were doing and what the best evidence suggested? Where teams were already close to the evidence, or where evidence had little to say, his office stayed out.

3. A willing and eager partner.

That last criterion was decisive, because in USAID’s structure the detailed decisions about how money got spent sat with foreign service officers and foreign nationals working in country offices around the world.

‘If they’re not bought in, then that was a recipe for disaster.’

An unenthusiastic partner had too many ways to quietly undo the work – taking advice at one stage of the process and discarding it at the next. This is why Dean still believes in a more collaborative approach over a more prescriptive one, even though it likely meant a smaller headline number. A directive approach was never actually on the table – but he argues it would also have been unhealthy, sacrificing the long game of building state capacity, both within USAID and in partner countries.

Taking goals as given – up to a point

Dean’s office took sector and country goals as given: those were decisions made by Congress, not by USAID, and his team’s job was to help achieve them more effectively. That stance is also why he was initially prepared to stay on into a Republican administration – the wonky policy people on both sides of the aisle supported hard-nosed measurement and walking away from things that don’t work.

But economics could still speak to whether a goal was doing what its advocates thought. He points to arguments for pressuring credit rating agencies to score African countries more favourably, to bring down their cost of borrowing. ‘As a matter of economic theory, that’s flawed,’ Dean argues. If political pressure distorts ratings, markets will simply discount them, worsening the adverse selection problem for investors and ultimately making borrowing harder for high-risk countries. The better path is the unglamorous one: helping ministries of finance manage debt, improve forecasting, and collect more revenue – work his team was directly engaged in.

Subscribe now

The logic of embedded evidence labs

‘Rather than saying, what’s the best way we can use our money in the provision of direct services, instead ask: what’s the best way we can use our money to help influence and improve the way local government uses their money?’

For all USAID’s scale, its budget in any given country was tiny relative to that country’s own budget. The leverage question follows naturally from this fact: rather than asking how to spend aid money best on direct services, ask how aid money can improve the way local governments spend their own.

‘That knock on effect creates leverage, creates much bigger impact if successful.’

Though Dean is candid that it is a riskier proposition, as influence one step removed is less likely to actually change decisions.

This is the logic behind embedded evidence labs. The labs are not cookie-cutter, but typically combine three components: helping governments review and adapt the global evidence base for their context; building the data infrastructure to manage programmes at scale; and deliberate, donor-funded innovation, where outside money pays for large-scale pilots so that ministries aren’t forced to cannibalise existing programmes to test new ideas.

He offers two examples, both from education. In Rwanda, a lab grew out of a randomised trial on teacher compensation, and its first task was building the national data infrastructure – attendance, test scores – that lets district heads manage schools and that underpins further testing, with the resulting policies now reaching all of Rwanda. In Peru, the focus was innovation: additional funding supported 14 randomised trials co-designed by researchers and the Ministry of Education, from which three successful ideas emerged and are now being scaled nationally.

The synthesis gap and the implementation gap

Dean’s forthcoming piece in the Journal of Economic Perspectives argues the academic-policy gap is smaller than often claimed but larger than it should be. Smaller, because in many cases he saw first-hand at USAID, research did speak fairly well to the goal at hand – it simply wasn’t being used. Larger, because nothing existed to bridge the last mile – in USAID’s case, there was no document a foreign service officer could pick up that synthesised the literature and translated it into what belongs in a request for proposals.

Dean sees the gap in two main parts. The first is synthesis. Economics, he argues, does not reward meta-analysis enough, and the discipline’s quantitative instincts desert it when it comes to aggregating evidence

‘We are a very quantitative oriented discipline, and yet when it comes to synthesis we shudder at the idea of being too quantitative — instead we want to take our 17 papers... put them up on the wall and then stare at it until we see a pattern.’

The result is that policymakers, donors, and academics alike put too much weight on individual studies – the one they funded, the one they wrote, the one in a top five journal.

The second is implementation. Papers are written to answer theoretical questions, not to equip an implementer to act on the findings. Dean thinks that papers could carry far more useful insight, even if it’s in the appendix. Things like the design decisions behind a programme, the pilots that failed and why, the details of the cash transfer amounts and savings components that get a few passing sentences in the main text. For a policymaker who has just read about impressive treatment effects, the immediate next question is how the thing was actually done.

When do we have enough evidence?

With 117 randomised trials comparing cash to no cash, Dean is comfortable saying we don’t need another study of whether cash transfers work for households. But ‘everything’s relative’ – he is himself running new cash trials, because how to deliver cash remains wide open, and the macro question also looms large: what happens to markets and multipliers when programmes hit a whole country rather than a district?

Even in medicine, where a drug’s efficacy is settled, the question of how to get pills into mouths and shots into arms is not. Asked where he would push the field if he were its general manager, Dean returns to his two gaps: more synthesis, supported by more systematic measurement across studies, and more research on the inner workings of organisations and programmes.

Ethiopia is arguably the most ambitious, and most closely watched, attempt to make export-led manufacturing work at scale in Africa. And that’s only one aspect of the country’s wide ranging set of reforms in recent years, which has seen the government embark on a range of macroeconomic policy changes, across exchange rates, monetary policy, taxation, and more.

For this wide ranging discussion, Kartik Akileswaran and I are joined by Mamo Mihretu. Having worked on trade competitiveness and finance at the World Bank, then as a senior economic advisor to the prime minister, and most recently as the tenth governor of the National Bank of Ethiopia, he has seen the same set of problems from different vantage points – both designing reforms, and actually implementing them. We start by discussing Ethiopia’s export-led manufacturing push, before getting into recent macroeconomic reforms that set the foundations for growth.

Subscribe now

An engine that had reached its limits

By the headline numbers, when Mamo joined the government in 2018, Ethiopia looked like a roaring success. Growth was near 10%, large public infrastructure projects were under way, and a bold industrial-parks programme was taking shape. But the engine underneath faced severe challenges. Growth had been driven overwhelmingly by public, debt-financed infrastructure investment, sustained by financing from domestic banks, monetary financing of the budget by the central bank, and heavy external borrowing, particularly from China.

The trouble, as Mamo describes it, was that those financing sources “reached their limit almost at the same time.” The result was a macroeconomic imbalance: a chronic shortage of foreign currency, inflation running close to 30% for three years, mounting debt-sustainability risk and a financial sector dangerously exposed to a single large state lender. COVID and domestic conflict compounded these problems, pushing the exchange-rate premium to almost 100%. This was the situation any reform programme would have to confront.

Getting industrial parks off the ground

The industrial-parks agenda had arrived earlier, in the early 2010sl. In a country with strong developmental-state traditions, the idea of a geographically demarcated area running liberal economic rules separate from the rest of the country was a hard ideological sell.

Three things, for Mamo, built consensus. First, policymakers spent close to a year studying how industrial parks had evolved in China, Vietnam, Singapore and South Korea – and, just as importantly, where they had failed, as in Nigeria. That study became a strategy document used as a consensus-building tool. Second, the parks were framed not as ideology but as a practical fix for binding constraints: the shortage of foreign currency, the low manufacturing base, and weak exports. And third – for Mamo the most important – there was “a genuine attempt to develop the institutional infrastructure” to implement the strategy, not merely write it.

The Ethiopian Investment Commission was modelled on Singapore’s Economic Development Board, a new industrial-parks corporation on Singapore’s JTC, and an investment board chaired by the prime minister gave strategic direction. These were staffed by technocrats who “put development at the centre” and had developed serious expertise – these institutions were what set Ethiopia apart from countries that wrote the same strategy and went nowhere.

Subscribe now

The most credible, not the cheapest

Nowhere is that clearer than in the Hawassa Industrial Park, built in under a year, with its first shipment reaching the US market less than two years after breaking ground. The strategy focused on landing an anchor investor – PVH, the apparel group behind Tommy Hilfiger and Calvin Klein – on the logic that the supply chain would follow.

Mamo recalls the investment commissioner going “to the airport at three o’clock in the morning” to receive investors in person, “because it matters to really project that desire.” What clinched it was not tax – incentives were “icing on the cake, but it was never the reason” – but credibility: a clear national vision, a focus on productivity and sustainability, and policymakers willing to “break into the logjam whenever a problem surfaced.” Ethiopia won investment, he says, “not because we were the cheapest, but because we were the most credible.”

At its peak around 2020–21, Hawassa alone supported roughly 100,000 jobs, but the suspension of Ethiopia’s AGOA access – its quota-free, tariff-free route into the US market – was “a serious blow” to textile competitiveness, which was compounded by conflict.

“Had it not been for the removal of the AGOA eligibility, I think Ethiopia’s textile exports would have increased significantly. It’s not so much that we lost in terms of actual export revenue because of the cancellation – but rather that the Ethiopian textile sector didn’t become what it could have been. That non-occurrence of what could have been is a significant development loss that I personally still regret.”

Getting the macro fundamentals right

For Mamo, none of this works without a sound macroeconomic foundation – and “the discussion about industrial policy and a competitive exchange rate is the same.” A misaligned exchange rate, with a premium of 80% or 100%, is simply a tax on every dollar an exporter earns; addressing it “is essentially creating a competitive manufacturing sector.”

To that he adds low and stable inflation, since “it’s very difficult for an exporter to enter into a five-year contract if inflation is hovering around 30%”; a financial sector able to supply working capital to manufacturers; and, above all, predictability. “If the business environment is not predictable, it’s very difficult for them to plan.” Deviate from these fundamentals, he warns, and “eventually, something will have to give.”

Share

How the exchange rate reform was done

When Mamo became governor in 2023, the situation was acute enough that an importer could wait two years for the foreign-exchange permit to buy $1,000 of inputs. So the central bank worked on and set its own strategy – price stability, exchange-rate reform, financial-sector stability, and strengthening the institution itself. It modernised monetary policy onto an interest-rate basis, opened banking to foreign competition, secured instrument independence for the central bank, and put guardrails around the monetary financing of the budget that had fuelled inflation.

And in July 2024, Ethiopia moved to a market-determined exchange rate, ending decades of a fixed regime. But Mamo stresses it was a wholesale reform: for example more than 80 accumulated foreign-exchange regulations were rationalised into a single Green Book, so that any bank, investor or importer needed to consult only one source.

What made it stick was the method. Ahead of this exchange rate reform, monetary tightening began in August 2023 – involving a credit cap, a new policy rate, and control of money growth. It was coordinated with the finance ministry through a committee where “even when there is disagreement, the disagreement stays within that room.” And it came with built in mitigation: higher public salaries, targeted subsidies on edible oil, fuel and medicine, a phased rather than overnight removal of fuel subsidies, and more spending on social protection.

The results have been striking: exports heading towards $10 billion within two years, gold exports up from around $200 million to $4 billion, inflation back to single digits, and reserves up sixfold. “It’s not just the reform,” he says. “How you do the reform matters as much.”

What reform requires

Asked for the single most important lesson, Mamo offers several:

• Start with honest diagnostics, because “the real tragedy in development policy” is spending reform energy on things that don’t matter.

• Address the actual constraint in your own context, not what worked elsewhere.

• “Personnel is always policy”: assemble a team that can learn, synthesise and act.

• Sequencing “is half policy reform.”

• Communication is integral — “treating the public as adults.”

• And research “informs, but research doesn’t decide”; what policymakers need most is judgment about the right policy “at that particular time, in that particular context.”

The deepest point is about ownership. “If the reform is owned, then it outlives the person who initiated it,” Mamo says; “if reform is externally imposed, then it will die.”

Having addressed what he identifies as the “self-inflicted” macroeconomic mistakes of the past, Mamo is optimistic for economic prospects of Ethiopia, so long as the reform stays the course and there is a foundation of peace and regional stability.

The conversation about evidence-based policy often focuses on the policy side. The questions we ask tend to be something like: why is this amazing body of evidence not already shaping decisions?

But what we should also be asking, on the research side, is whether the evidence base itself is actually worthy of shaping policy.

This week on Ideas in Development, I was joined by Rafe Meager, who is an Associate Professor at the University of New South Wales, and is one of the leading meta-analysts in economics. We discuss why the evidence base is messier and more uncertain than we would like, and how we can fix that by aggregating across studies.

Subscribe now

A paper is not a proof

It helps to start with what a single paper actually involves. Rafe describes research as a long, artistic process with a significant front-end thinking cost, including years of accumulating background knowledge, understanding, and insight before the idea for a project actually arrives. This is then followed by even more years of iteration, working papers and convincing expert referees that your result is new, plausibly true and relevant.

It can be tempting to treat the published paper that emerges out of this process as though it were a mathematical proof that something is true or false. But Rafe argues that this is the wrong mental model.

“It’s not right to think about papers as proving that something is happening or not happening, or true or not true… empirical social science is about the balance of probabilities”

Empirical social science deals in “the balance of probabilities”, and we accumulate evidence for or against fairly narrow propositions. And even a genuine mathematical proof only applies where its axioms hold, and whether those axioms describe your actual situation is itself a fuzzy, subjective judgement.

“If you do statistics enough, you come to respect the noise. You come to respect what’s not captureable as something… important alongside what we’re very happy that we can capture and explain.”

The noise we ignore

Even well-designed and well-executed papers can mislead.

“There’s just very little serious attempt to deal with noise. People do these power calculations that are mostly, to be honest, made up or kind of reverse engineered. And we wouldn’t accept that in other areas of our statistical practice.”

Rafe argues that while the discipline obsesses over internal validity and causal identification, we don’t hold ourselves to the same standards when it comes to dealing with noise. Precision, variance and statistical power are relegated to secondary concerns. The result is an evidence base far less precise than its confident abstracts suggest.

And every result is produced against a backdrop of existing knowledge, which shapes how a study is designed. So doing the best possible job given what is known, if what is known turns out to be wrong, means that a study can still mislead.

“You can do the absolute best job given the existing knowledge that you have. And yet if that existing knowledge is wrong, the study might still be misleading in a way that no one could know.”

Subscribe now

Social science is hard!

We then discuss external validity — whether a result learned in one place and time tells you anything about another. Classical statistical inference assumes you are drawing from a defined population by some known process. But in social science there is never really a fixed population.

“This idea of a population, especially in social science… there’s no population of adults in the UK. That’s changing every single day because people are moving in and out of the population.”

Everything flows. So even extrapolating from the past to the future for the ‘same’ population carries an error that classical statistics simply does not capture.

The temptation, faced with this, is is often one of two bad responses. One is to treat a single study as decisive: a zero effect on microcredit here means it works nowhere; a positive effect on cash transfers there means we should roll it out everywhere. The other is the fashionable pessimism that says it is absurd to imagine anything that works in Kenya could work in Ghana. Rafe rejects both. It is not absurd at all – it may or may not be true, and we have the capacity to study that and we should.

The whole premise of policymaking, as opposed to a tailored personal solution for every single individual, is that aggregation is possible. The question is how to do it honestly.

Separating real differences from sampling noise

If a trial in Ghana gives a different estimate from one in Kenya, that does not necessarily mean the underlying effects differ, because rerunning the Kenyan study would itself produce a different estimate. Sampling error alone leads to apparent differences, and on sensible assumptions it will overstate how different the effects really are. So what you need is a technique that separates true variation from sampling variation, and the regression tools of an introductory econometrics course will not cut it.

The job requires a deconvolution, a problem that astrophysicists also face, and Rafe’s tool of choice is Bayesian hierarchical modelling, which nests the uncertainty within each setting inside a second layer of uncertainty across settings.

And this does not require hundreds of studies! Often people who are trained on classical statistics incorrectly assume that meta-analysis demands a large number of estimates.

This is where Rafe’s microfinance work comes in, which aggregates seven randomised trials around the world that had a range of estimated effects. The microcredit literature is large, but most does not focus on the question which Rafe cared about: what happens if you expand access?

Only a few studies answer that, and with a small number of genuinely comparable data points, Bayesian hierarchical modelling, which does not lean on large-sample approximations, can help.

Share

Can your system actually deliver this?

Working with Noam Angrist and Youth Impact, which was scaling targeted instruction across Botswana and into Namibia and South Africa, Rafe used these methods to find which features of an intervention have to be preserved at scale and which can be compromised. Aggregation methods cannot answer this causally, but they can show which features correlate most strongly with larger treatment effects, telling you where to look.

What emerged was that implementation fidelity was very strongly predictive of effect size. Take-up is not a simple binary. Attending one remedial class is not the same as attending eight weeks of them, and the quality of teacher training mattered enormously for something as difficult as assessing children, splitting them by level and teaching each group accordingly. When Youth Impact then varied the intensity of targeting directly, more targeting produced more learning, corroborating the correlation.

Prior to the study, some economists had insisted that implementation could not matter much in studies, because trials are ‘gold-plated’ – everyone tries hard, so there is no meaningful variation to find. But basic contact with reality says otherwise: even smart people trying hard are usually barely getting these studies over the line, because they are attempting something ambitious.

Implementation, in other words, can be studied scientifically. For a policymaker, this work reframes the central question. Instead of asking “does this program work?”, the better question is “can my system actually deliver this program?”

What aggregation cannot fix

Rafe is equally clear about the limits of aggregation methods. Sites where rigorous evaluation is possible may differ systematically from those where it is not.

Publication bias persists, as the research community has preferences about what it wants to see that may or may not track the truth.

And there is no clean solution to any of this, which is precisely why Rafe calls for interdisciplinary collaboration and partnership between academics and policymakers. Neither group would really want to do without the other – more sources of knowledge are simply more.

Heuristics for the rest of us

Asked how they react to a single causal claim on social media or in an abstract, Rafe offered a set of rules of thumb.

• More than one study is always better than one – is this the first study of a question, or building on an established literature?

• The quality of the causal and statistical techniques matters.

• Was the study pre-registered and, if so, what was the pre-analysis plan?

• Ask people who know more than you do whether a finding passes their sniff test.

Seeking out the bits of knowledge you are missing is among the most useful things a reader can do.

The future of aggregation

Looking ahead, as AI drives up the volume of research, the need for disciplined aggregation will only increase. Careful attention is required on the quality of inputs and to whether studies are even conceptually coherent enough to combine.

Rafe worries about ‘cookie cutter’ approaches that apply meta-analytic recipes without care – more methodological progress is required for these aggregation methods.

Beyond that, we need to empower the public to understand evidence the way it understands football teams. Some evidence is simply better than others, the quality of studies varies, and saying so out loud need not be hard. The alternative we are left with – a public fed an endless stream of contradictory “studies say” headlines and given no tools to weigh them – is a recipe for losing trust in social science altogether.

Development economics has spent the past couple of decades building a large body of empirical evidence on a range of topics. Billions of dollars and thousands of careers have gone into attempts to causally estimate the impacts of cash transfers, deworming, microfinance, teacher incentives, nudges, conditional school grants, etc.

But we know very little about whether any of this evidence is used. There is an evidence gap on evidence use – the where, when, why, who, and how remain largely unclear.

That is the motivation behind our new Ideas in Development series on evidence, and for the first episode, I’m joined by Michelle Rao, a fellow at the Center for Global Development and one of a small but growing group of researchers trying to study evidence use empirically.

Her own research looks at 128 evaluations of conditional cash transfers in 17 Latin American and Caribbean countries between 2000 and 2015. Michelle finds no overall relationship between what evaluations find and governments’ subsequent spending decisions. We discuss that finding, the other research in this field, and more.

What development economists tend to mean by evidence

The word evidence means different things to different people. In the policy world, it has a broader meaning, encompassing descriptive data, case studies, and academic papers. Typically though, when a development economist is talking evidence, they are referring more narrowly to causal evidence – that is, the kind of evidence that tries to tell us what would have happened in a counterfactual world where a particular policy or programme had not existed.

A great deal of modern empirical economics is a set of techniques – RCTs, instrumental variables, differences-in-differences, regression discontinuity – designed to construct some form of believable counterfactual, so that we can credibly claim causality. This is the body of work that has grown dramatically in development economics over the past two decades, and it is the body of work which Michelle’s work focuses on.

Of course, other forms of evidence remain hugely important for both academia and policy. Raj Chetty’s descriptive work on mobility, Claudia Goldin’s on the labour market – none of it is causal in this technical sense, and all of it has reshaped how economists and policymakers see the world. But causal evidence has, in principle, a more direct route into policy. If you can credibly say that a given programme produced a certain effect that, in theory, maps more cleanly onto decisions – i.e. adopt it, redesign it, spend more, spend less.

From evaluation to spending

Conditional cash transfers (CCTs) are one of the most evaluated programmes in development. More than half of the 128 evaluations in Michelle’s sample were done in collaboration with the governments running the programmes, which means policymakers cannot claim not to have known about the results. If evidence-based policy is happening anywhere, it should be happening here.

To test this, Michelle matches every evaluation to subsequent spending on the same programme and explores whether resources seem to flow towards what works. The answer is no. There is no relationship between individual evaluation findings, whether you look at the framing, the size of the effect, or the direction, and what governments do with their budgets in the years that follow. Allowing for the possibility that policymakers are more sophisticated than that, and learn from the evidence base as a whole rather than individual studies, she fits a Bayesian hierarchical model to estimate the cumulative evidence base in each country, and still finds zero effect.

While in these settings, credibility and generalisability are unrelated to spending, Michelle does find an important variation in impacts based on the timeliness of evaluations. When evaluations are completed quickly and the same political party is in power for both the evaluation period and the publication of results, evaluations are predictive of spending.

Share

Why nothing happens

The natural temptation is to read the null result as a failure of policymaker rationality or attention – they are not reading the papers, they don’t understand the methods, or they are captured by ideology. Michelle’s reading is different. The policymakers in her sample are not poorly trained. More than half of the finance ministers in her data have PhDs, mostly in economics. So the plausible problem is not a lack of understanding, but that even a policymaker who has read and understood an evaluation faces real political costs to acting on it.

Cash transfers are a particularly politically challenging setting for clean evidence-based adjustment. It’s unpopular to withdraw spending, and hard to increase it – so evidence has to clear a high political bar to translate into either expansion or contraction.

However, evaluations that arrive while the responsible party is still in office means they have ownership of what to do next, and can take credit for what worked. Along similar lines, Alix Bonargent finds that International Growth Centre research that is co-created with governments is more likely to translate into implementation. So an early insight from this evidence base on evidence use is the importance of the political conditions surrounding the production and arrival of research.

What development economics is missing

Development economics has invested enormously in generating evidence and almost nothing in studying what happens to it next.

The empirical literature on evidence use is genuinely small - I summarised what I could find here.

What is particularly striking, the more I’ve looked at this area, is that we lack even the basic descriptives we need. We don’t have good data on what kinds of evidence policymakers actually consult, how often, at what points in the policy cycle, or through which channels. While here is a great deal of institutional knowledge and useful anecdotes, on the importance of relationships, the importance of being there at the right moment, the importance of trust, very little of it is in a form that lets us test claims or design interventions. As Michelle puts it, “we need to know what we don’t know.”

This is the gap this series is trying to map out. The episodes ahead will cover both sides of the supply-and-demand problem: how research is funded, conducted, framed and communicated on one side, and what policymakers actually want, how organisations learn, and what would be needed to build genuine capacity for evidence use.

With aid budgets contracting and the political space for development cooperation narrowing, the cost of producing evidence that doesn’t have an impact is increasingly unacceptable to funders. Every dollar that goes towards funding a paper is a dollar that could have gone into a programme itself, and the case for donors funding research has always rested implicitly on the claim that the resulting evidence improves how the remaining dollars are spent. If Michelle’s data is pointing in the right direction, that claim needs more scrutiny than it has received.

Before getting into my write up, Dani wanted to highlight these open access sources as background.

• “The Past, Present, and Future of Economic Growth,” in Franklin Allen and others, Towards a Better Global Economy: Policy Implications for Citizens Worldwide in the 21st Century, Oxford University Press, Oxford and New York, 2014

• Premature Deindustrialization. Journal of Economic Growth. 2015; 21:1–33.

• “Prospects for Global Economic Convergence Under New Technologies,” in David Autor et al., An Inclusive Future? Technology, New Dynamics, and Policy Challenges, Brookings Institution, Washington, D.C., May 2022.

• “A New Growth Strategy for Developing Nations,” (with Joseph E. Stiglitz), in Dani Rodrik and Lili Yan Ing, eds., The New Global Economic Order, Routledge, London, 2025.

Dani Rodrik had been one of the most influential voices arguing that manufacturing is special. Manufacturing acted as a kind of shortcut: countries that could get workers onto the manufacturing escalator could grow rapidly without having to solve every other problem of governance, infrastructure and human capital first. It was the East Asian story, and it was, for a long time, the story Dani would tell.

Dani recently published an essay titled How I Became a Manufacturing Skeptic, and he joined me for a live episode of Ideas in Development to talk through what changed his mind, where growth might come from instead, and what an industrial policy fit for the new era should look like. Despite changing his mind of manufacturing, Dani feels the same underlying framework remains, just that it now needs to focus on services.

A shortcut that no longer exists

Dani’s framework distinguishes between two channels of growth. The first is what he calls fundamentals: human capital, governance, infrastructure, the slow business of broad-based capability building that delivers steady but unspectacular growth of around two to three per cent. The second is structural change: the rapid reallocation of labour from low- to high-productivity activities, which is how East Asian economies achieved decades of growth in the high single digits. Historically, manufacturing was the vehicle for that structural change. It was high-productivity, enabled catch up and was good for jobs.

“it’s a kind of an escalator that keeps rising, but it’s also very wide… you can put a lot of people on that escalator, because the skill requirements aren’t that demanding.”

That last property is what has eroded. Even the most successful current industrialisers, countries like Bangladesh, Vietnam, and Cambodia, are absorbing far less labour into manufacturing than their East Asian predecessors did. Beyond them, manufacturing employment in most of the world is flat or falling. Rodrik’s research on premature deindustrialisation has documented this directly: with each successive generation, developing countries are reaching peak manufacturing employment at lower shares of employment and at lower levels of per capita income. This manufacturing curve has been shifting down for half a century.

One key mechanism, in Dani’s account, is the changing nature of global value chains. Where South Korea and Taiwan exported manufactures based on the resource they had most of, low-skilled labour. But joining a modern global value chain now requires skills, infrastructure, logistics, governance and technological capability that developing countries by definition lack.

Trade in manufacturing, which was once a kind of a shortcut around weak fundamentals, no longer offers that bypass. The international reports on what countries need to do to plug into supply chains are essentially a checklist of fundamentals. The whole point of the manufacturing route was that you did not have to wait for those things.

One case that has helped to crystallise this for Dani is Ethiopia, a country that explicitly tried to replicate the East Asian model, with Chinese investors, industrial parks, and an East Asian-style industrial policy. There was some success. But the closer Dani looked, the clearer it was that it was not delivering enough employment. More on Ethiopia in a future episode, as we are joined by Mamo Mihretu to discuss Ethiopia’s ambitious, growth-oriented reforms of the past decade.

Subscribe now

The services that are actually absorbing workers

The paradox Rodrik flags is that, even as the manufacturing route was closing, the two decades before the pandemic were a strong period for developing country growth. The driver, increasingly, was services. He points to a paper by Fan, Peters and Zilibotti on India which clarifies an often-misunderstood point: India’s service-led growth is not really about IT and business process outsourcing, which employ a tiny share of the labour force. It is about non-tradable consumer services, things like hospitality, retail, personal care, local transport, and food services. The same authors find similar patterns in Africa’s high-growth episodes.

This is the model Rodrik now thinks developing countries should be planning around. It is, he is candid, a less powerful model than manufacturing-led growth. Because these services are non-tradable.

“You can’t have a strategy where you’re sequentially… emphasizing one service after another, because they all have to be sold domestically… You have to do it all at once… that necessarily reduces… the growth ceiling. You cannot grow nearly as rapidly in a services-driven model than in a manufacturing model, because of the role of non-traded versus traded.”

But two developments make him more optimistic than he was even a couple of years ago. The first is that services productivity itself is rising. In the United States over the last two decades, productivity growth in retail, food services and wholesale has actually outpaced manufacturing, reversing a very long-term trend. One recent study finds that the marginal product of labour in services in India is higher than in manufacturing. Mobile money, digital platforms, gig-economy infrastructure and now AI are all contributing to a quiet productivity revolution in sectors that were long dismissed as technologically stagnant.

The second is that AI, in Dani’s view, looks structurally different from earlier automation. Robots in factories directly displace workers. AI, by contrast, “gives the skills and experience and training of much more trained, much more experienced, much more educated professionals” and makes those skills available to those with less training.

I guess I am more pessimistic on this second point than Dani, as we discussed in our AI series of Ideas in Development. But Dani, like myself, is particularly interested in the possibility of low-cost AI-delivered business consulting for micro-enterprises – a productivity boost aimed squarely at the bottom of the labour market rather than the top. As our Training Entrepreneur Lit identifies, this does not just need to target the micro-enterprises, as this type of consulting could help firms move from small to large as well.

Share

Industrial policy for services, learning the correct lessons from East Asia

Dani wants to first dispel a caricature of industrial policy as top-down winner-picking enforced by a strong state. Even in East Asia, successful industrial policy was, as the sociologist Peter Evans put it, embedded – agencies engaged directly with sectors and firms, learning their constraints rather than picking from on high. This is a theme repeatedly emphasised by policymakers in our growth series. Experimental rather than directive.

“Letting the losers go rather than picking the winners.”

And subsidies were only one tool among many. Public inputs like vocational training, coordination across ministries, regulatory assistance, and technology provision, were often more important.

Translated to services, this looks concrete and local. Dani’s favourite example is a programme in the Indian state of Haryana, where the state government partners with Uber and Ola to expand local driver employment. There are no subsidies and no hard conditionality. The firms want to grow their reach, the state wants to create jobs for young people, and the cooperation is about easing the specific constraints, commercial licensing, access to jobseekers, regulatory frictions, that emerge only through dialogue. Another example, from Bogotá, organised street vendors via an app so they could coordinate purchases of supplies, sparing each vendor an early-morning trip to the wholesale market. A third, in Nigeria, used the YouWin entrepreneurship tournament to identify and back dynamic micro-entrepreneurs with finance, training and consulting.

The common thread is that the unit of analysis is small, local and often subnational.

“These services are very small, they require, sort of, you know, addressing the problems of very small scale and local, producers.”

The ministries that traditionally run industrial policy are still oriented towards trade, global value chains and large firms. Their gaze, Dani argues, needs to shift.

What this means for the jobs problem

One of Dani’s key motivations is creating good jobs. According to his own back-of-the-envelope numbers on the future of work, excluding China, roughly three-quarters of workers in developing countries are likely to be in occupations that neither engage with the world economy nor require college education.

“If we don’t have an answer to the question, how do we get those workers more productive… we don’t have an answer to development.”

In response to a question from the audience, Dani drew the comparison between manufacturing today, with its capital- and skill-intensive frontier, and natural resource sectors like oil, or copper. Dani increasingly sees manufacturing as resembling enclaves that might generate exports and tax revenues but does not absorb labour.

So while it can certainly still be worth pursuing manufacturing to benefit from the spillovers, build supply chain resilience or as a national security priority, as a strategy for lifting the majority of a country’s workforce out of low-productivity work, it no longer plays the role it once did. Therefore, the bulk of policymakers’ attention should be on the small firms, micro-enterprises and care workers that already populate the non-tradable service sector in many low- and middle-income countries.

This is a more modest vision of development than the one Dani had previously championed, which Dani makes no bones about. The ceiling is lower, the path is slower, and the policy instruments are potentially messier. But Dani remains hopeful for a new type of bottom-up approach, particularly against the alternative of waiting for an escalator that is not coming back.

The standard debate about technology policy in developing countries tends to dichotomise. Either a country adopts the best available technology from abroad, or it tries to innovate at home.

Two of the twentieth century’s most striking development stories – Taiwan’s transformation from assembly to home of the world’s most advanced chipmaker, and Brazil’s pivot from food importer to global agricultural powerhouse – suggest that this is a false dichotomy.

Taiwan and Brazil did both. And they did so by founding publicly funded research institutes in 1973, ITRI (Taiwan) and Embrapa (Brazil), which went on to transform each country. Karthik Tadepalli, a PhD student at UC Berkeley who has written about both cases for Asterisk magazine and on his Substack, joined me on Ideas in Development to explain what they did, why it worked, and what might generalise.

• The $16 million bet that created Taiwan’s chip industry

• The science policy behind Brazil’s agricultural triumph

A licensing deal and an investment in people

Taiwan in 1973 was lower-middle income, with an electronics sector that consisted largely of assembling components shipped in from Western firms. “Made in Taiwan” was a mark of cheapness. That year the government set up the Industrial Technology Research Institute (ITRI) under the Ministry of Economic Affairs, with a mandate to move Taiwan towards the technological frontier. It began with a few hundred employees, only ten of whom held PhDs.

ITRI’s first decisive move was to license chip technology from the Radio Corporation of America (RCA). The deal was hard-won: RCA was persuaded through the intervention of Pan Wen-Yuan, a Taiwanese-American engineer-turned-director who pushed the project. It was also expensive relative to ITRI’s modest budget. What made the deal transformative was not the blueprints but the training that came with them. A team of ITRI engineers spent extended time in RCA’s American factories, picking up not just the formal production process but, as Karthik put it, the kind of tacit knowledge and informal understanding that you only get by being on the production line. That early cohort would go on to lead UMC and TSMC.

The lesson Karthik draws is that adoption only works if you have the skill to absorb what you’re adopting.

“Without those engineers and without that training, the RCA deal would have been helpful,” he said. “It would have been nice to produce those chips when they weren’t producing anything before, but it wouldn’t possibly have led to the modern dominance that we see.”

The spin-off model and the politics of staying funded

ITRI’s second decisive move was institutional. As part of the RCA arrangement it built a demonstration plant, and that plant turned out to produce chips at a higher yield than RCA’s own American facilities. Rather than keep the plant in-house and settle for commercialising an already obsolete chip generation, ITRI spun it out. The plant, the blueprints and some of the engineers became the United Microelectronics Corporation (UMC). By 1985 UMC was Taiwan’s most valuable company. The same logic later produced Taiwan Semiconductor Manufacturing Company: Morris Chang was president of ITRI when he developed the plan, and the company was spun out of the institute in a natural continuation of an established model.

The spin-off model did political work as well as technological work. ITRI did not have the capital to commercialise on its own, so spin-offs were funded by soliciting investment from incumbent firms. This meant the rest of the industry shared in the upside rather than watching from the sidelines as a state-backed competitor took their market. It was a way of binding industry to the institute rather than setting them against each other.

That binding mattered because R&D support is politically harder to sustain than the more familiar tools of industrial policy. Subsidies and cheap credit are essentially free money in firms pocket, which they can pocket without actually changing what they do. Blueprints for a high-technology plant are not free money.

“If you give them a blueprint for a high technology plant, they can’t actually make any money off of that unless they’re willing to actually upgrade their capabilities.”

That is the in-built discipline of R&D support: it forces productivity growth because the technology cannot otherwise be monetised. But it also makes the policy easy to caricature as wasteful, which is why ITRI worked hard to provide tangible services to industry, duplicating photomasks, for instance, an expensive and difficult process the institute eventually spun out as the Taiwan Mask Corporation. Firms came to know exactly what they would lose if ITRI lost its budget.

Subscribe now

Brazil, the Cerrado, and a surprising scarce input

Embrapa, the Brazilian Agricultural Research Corporation, was also founded in 1973, in a country that had sometimes been reliant on US food aid despite being a vast agrarian economy. Most of Brazil’s farming was concentrated in the subtropical south around São Paulo and Rio de Janeiro, where temperate techniques worked. The huge tropical biomes of the interior, including the Cerrado, were essentially closed to commercial agriculture. Rural extension services modelled on American practice had existed since the 1950s but had not delivered results, internal evaluations showed farmers receiving extension did no better than those who did not. The knowledge being transferred had been developed for Iowa, not for tropical Brazil.

Embrapa’s signature achievement was making the Cerrado farmable, unlocking an area roughly three times the size of Texas. The biome combined three problems: highly acidic soil that prevented crops from absorbing nutrients even when fertiliser was applied, inherently nutrient-poor soils, and a hotter, drier climate than Brazil’s agricultural south.

Embrapa attacked all three: liming the soil to neutralise its acidity, breeding heat-tolerant soybean varieties, and developing new fertiliser formulations matched to Cerrado deficiencies. The Cerrado now hosts roughly 70% of Brazilian soybean production, and Embrapa’s work effectively tripled the country’s available agricultural land.

That breakthrough relied on a long bet. Embrapa spent around 20% of its budget in its first decade sending staff overseas for PhDs, before producing any major results.

Both ITRI and Embrapa, treated knowledge as “a long-term capital investment rather than just, okay, we need to get the immediate output, so let’s burn our resources right now.”

You cannot do frontier work with a small team of generalists.

Embrapa’s research was also tightly directed. The organising framework was induced innovation: aim your R&D at the scarce, expensive input. In nineteenth-century America that input was labour, which drove the development of tractors and combine harvesters. In Brazil, counter-intuitively given the country’s size, the scarce input was land – because productive land was concentrated near São Paulo and Rio, where urbanisation was driving prices up. Cheaper food and competitive exports required cheaper land, and that meant making new land usable. The market signal told Embrapa what to work on, and a clear focus gave a nascent research institute the kind of focused and high-impact win it needed to justify its existence.

Share

Surviving the politics

The politics matter as much as the science. Embrapa had a sister agency, Embrater, focused on rural extension. Through circumstances that remain murky, Embrater officials backed Lula in the 1989 presidential election; Lula lost, and the new president shut Embrater down virtually overnight.

Embrapa, doing less visibly applied work, would have been an easier target still. It survived because it was relentlessly non-partisan, cultivated allies across the political spectrum, kept itself outside the university system (which was politically hostile to the military government), and invested heavily in public communication. It held seminars for journalists, mandated that scientists give interviews (even if they hated it), and on its tenth anniversary even ran a TV campaign by a New York agency showing a herd of cattle marching into São Paulo saying the results of Embrapa’s research are coming to town. The message to the political class and public was clear: you need us.

ITRI’s politics looked different, its constituency was industry, not the urban consumer, but the underlying discipline was the same.

• Build human capital before you need it.

• Identify a real political constituency for your work and serve them tangibly.

• Pick a focused bet that fits your context rather than copying a template.

“It wasn’t that they applied the same direct policies, it’s that they had the same thinking”

The risk in writing about cases like these, Karthik acknowledged, is survivorship bias. We know how these stories ended, and we are not learning from the institutes that quietly failed. His response is pragmatic, as technology policy success stories are rare enough that even biased lessons are worth having.

I am really excited to be joined by Dani Rodrik for this live episode of the Ideas in Development podcast, taking place next week on Wednesday May 20th at 2pm UK time.

Register here

During the hour, we will discuss his evolving views, the evidence that changed his mind, and what this means for future growth strategies. And we’ll leave lots of time for questions from the audience.

Timings where you are on Wednesday, May 20th:

• US East Coast: 9:00 - 10:00

• Rio de Janeiro: 10:00 - 11:00

• UK, Nigeria: 14:00 - 15:00

• Central Europe Time: 15:00 - 16:00

• Eastern Africa Time: 16:00 - 17:00

• India: 18:30 - 19:30

You can find the Dani’s recent article prompting this episode here: How I Became a Manufacturing Skeptic.

Across our series so far we’ve been asking big questions about fast-growing cities: how to build enough housing, how to make land legible, how to pay for infrastructure, how to prevent crime from taking root, and how to ensure young people see cities as ladders, not dead ends.

This week on Ideas in Development, Ed Glaeser, professor of economics at Harvard and one of the world's leading urban economists, joins us to think about how these fit together.

Subscribe now

Ed was quick to pour cold water on utopian thinking about perfect cities.

“The road to urban dysfunction is paved by dreams of urban perfection, and… frequently more incremental improvements to city life are a wiser course to follow rather than thinking that you’re gonna get it exactly right.”

He does, though, has a clear theory of what cities are for. They exist to enable possibilities for ordinary people that we would never find on our own. Whether the year is 1500 or 2050, whether the city is in sub-Saharan Africa or sixteenth-century Europe, the work of urban government generally boils down to three things. Get those right, and a city can become an escalator out of poverty. Get them wrong, and the demons of density will come back to bite you.

Safety, mobility, education – and a fourth thing

Glaeser’s three foundations are safety, mobility and education.

Cities exist “to bring people close to one another, to enable us to work together in different ways, to buy, to sell, to borrow ideas, to communicate.”

Each of the three foundations is a precondition for that to actually happen: law because density without law is just easier predation; mobility because proximity is meaningless if you cannot reach across it; and education because learning is what people in dense places trade with each other.

The metrics he would track to understand how cities are doing follow naturally. For safety, homicides (which remains the only metric broadly comparable across time and space) supplemented by victimisation surveys, with one particularly diagnostic question for the developing world: have you been held up by the police? Fear of police abuse, Ed notes, is a particularly useful sign that something’s gone badly wrong with the provision of safety.

For mobility, travel speeds (the kind Gilles Duranton’s team can now extract from Google street maps), augmented by cell phone data on how poor and rich residents actually mix across neighbourhoods and workplaces. For education, years of schooling, plus test scores like PISA.

But there is a fourth metric that matters at least as much, because it captures whether the other three are doing their job. Cities are both fun places to be rich but often less intolerable places to be poor.

This inequality is “only tolerable when cities are providing escalators out of poverty into prosperity.”

The thing to measure, then, is upward mobility itself: the dynamics of individual wealth over time. A city that is failing on safety, education or mobility will eventually show up here, in the gap between what poor residents arrive with and what they leave with.

Share

Infrastructure, incentives, institutions

A recurring theme of our conversation is that urbanists who focus on infrastructure alone are only focusing on part of the problem. Ed tells the story of nineteenth-century New York, which he grew up being told was a tale of engineering triumphalism: the Croton Aqueduct dispelled cholera and yellow fever from a fetid early-1800s city. But the truth was less heroic. For 25 years after the aqueduct opened, New York continued to have cholera outbreaks, because building the pipes did not solve the problem. Even with free hydrants, the water was heavy to carry, and poor residents were unwilling to pay for private connections. Mortality only fell once the city imposed fines on tenement owners who refused to connect.

“It wasn’t until the city started imposing fines on tenement owners who didn’t connect to the water system… that you really saw the cholera rates decline.”

The same dynamic shows up in his more recent work on Lusaka, where there is a strong correlation between how customers pay for water and how long water breaks stay broken. When customers pay by the litre, the company has a sharp incentive to fix breaks fast. When customers pay by the month, as they often do in poorer areas, repairs drag.

“If they don’t get fixed, the water company won’t get paid, so the water company fixes those things.”

The pipes are necessary. The incentives that surround them decide whether the pipes actually deliver water.

Then institutions, after the infrastructure itself, and incentives, represent the last I. There are many options available (direct public provision, parastatals, non-profits, fully private), and Ed cautions against being ideological, and remaining practical based on local context. But he does have one strong piece of advice on parastatals.

Only choose that route, he says, “if you can actually get someone who will lead this entity who has an international reputation that can be lost.”

Without that kind of figurehead, a parastatal is mostly a layer of plausible deniability that lets the mayor blame someone else for failures.

Subscribe now

Bus good, train bad

“Forty years of transportation economics at Harvard can be boiled down to four words. Bus good, train bad.”

Trains lock cities into enormously expensive, fixed investments that struggle to adapt, while buses, BRT systems, or even better-organised jitneys (privately operated minibuses) are flexible, cheap and serve poorer riders.

Within that, two more specific recommendations stand out. First, procurement discipline. The United States pays roughly $1.05 million for an electric bus that can be bought from Hyundai for $350,000, because American transit agencies insist on bespoke specifications.

“Basically any developing world city that’s customising their buses in the fashion of US transit agencies is making a huge mistake.”

So buy at scale, from large global producers.

Second, road repaving. With Gabriel Kreindler and Lindsay Currier, Ed has used US Uber data to show that in eight of eleven cities studied, there is no correlation between how rough a road is and whether it gets repaved next.

“We are repaving the rough roads just as frequently as we are repaving the smooth roads.”

A simple targeting rule, fix the bumpiest first, is sitting in plain sight.

Capacity eats policy as a light afternoon snack

The biggest constraint on actually doing any of this, Ed argues, is not knowledge but capacity. Doing his best management consultant impression, he says:

“in local government, capacity eats policy as a light afternoon snack.”

A mayor has 100 things hitting their desk every day and perhaps five people who will reliably do what they ask. The size of the civil service is misleading. He recalls former Rome mayor Ignazio Marino, who had over 100 projects he wanted to pursue and only two people who were actually willing to do anything that he told them to do.

That has implications for how academics work with mayors. Trying to redirect a leader’s scarce capacity toward priorities they did not choose is “a fool’s errand on a grand scale.”

Influencing policy, in Ed’s experience leading the IGC’s Cities That Work initiative, almost always depends on personal relationships with city leadership. White papers help, particularly because the people who read them often sit in finance ministries and can pose harder questions back to local governments, but they are not a substitute for a mayor or a chief of staff who actually wants to engage.

He names some of the leaders he has been most impressed by: Geordin Hill-Lewis in Cape Town, Jennifer Musisi in Kampala, Enrique Peñalosa and Claudia López in Bogotá, and Yvonne Aki-Sawyerr in Freetown.

The skill they share is balancing technocratic substance with political legitimacy. Comparing Michael Bloomberg and Tom Menino in the United States, Ed notes that Menino could build an innovation district and convince ordinary Bostonians it was for them; Bloomberg could do exactly the same thing in New York and be accused of catering to his rich tech friends.

Subscribe now

Plan the public realm, free the private one

On the perennial question of how much to plan, Ed splits the city in two. Public spaces, the road network, the basic infrastructure, the strategy for how mobility will evolve, need real planning. Private spaces, by contrast, do better with minimal planning, beyond bans on genuinely noxious uses. Ed contrasts post-war Tokyo, which rebuilt with light public direction and a well-designed land-readjustment process that allowed small owners to pool plots into larger structures, with post-war Beirut, where heavy central planning produced a polished downtown shopping district whose benefits to ordinary Lebanese citizens are hard to identify.

“I’m certainly on the side of the spectrum that favours maximal planning for the public spaces that are needed to provide mobility and to deal with sanitation and all those things, but fairly minimal control over what private individuals can do with their property”

That same instinct shapes his three priorities for sub-Saharan Africa over the next decade. Public health infrastructure first, with water at the centre – and not just the pipes, but the incentives and last-mile fines that get people connected. Education second, including the vocational skills that wrap around schooling, with humility about which interventions actually work. And third, mobility done well: smarter buses, better procurement, and congestion pricing built in from the start of any new road, on the model of Singapore in 1975.

“There’s no way to actually make streets smooth without actually charging people for the use of those streets.”

Politically, putting a charge on a road that used to be free is brutal. Putting one on a road that did not exist yesterday is not.

What unites all of Ed’s thinking a refusal to mystify cities. The demons of density – disease, crime, congestion – are ubiquitous, faced by Julius Caesar’s Rome and medieval London just as much as by Lagos or Lusaka today. The tools that work against them are mostly known. The harder problem is building enough state capacity, and enough trust between leaders and citizens, to use those tools well.

When young people in Kenya flooded the streets of Nairobi in protest against proposed tax hikes, and when post-election protests in Tanzania were met with an unusually heavy hand, the global story that filtered out tended to settle on a familiar frame: angry young people, restless cities, and fragile states.

Leonard Wantchekon, James Madison Professor of Political Economy at Princeton and founder of the African School of Economics, wants to push back against that frame. He joined Kurtis Lockhart and I on Ideas in Development to discuss why cities are best understood not as a youth problem but as a youth opportunity. For Leonard, the problems we do see in cities are usually a story about governance, not about the young people themselves.

Subscribe now

Cities as engines of culture, science and liberalism

Growing up in rural Benin, Leonard’s move to the city for secondary school was transformative, not because of school alone, but because of what came with city life. The French Cultural Centre, with its newspapers and films. Classmates from across Benin and visiting students from Nigeria.

Even before his move to the city, a nearby plant biology lab where scientists from Thailand and France ran conferences, the type of innovation hub that is more common in cities, broadened Leonard’s horizons.

“Walking around that place just made me wanted to think super big”.

For Leonard, cities are where cultural institutions can flourish, where science and technology can more easily be clustered into hubs that absorb young people regardless of background, and where peaceful politics is more likely to take root than in the countryside.

“Almost by definition, protests in cities are far more likely to be peaceful than protests in rural areas... You don’t have many places to hide. So as a result, you have to use your words.”

Reframing Kenya and Tanzania

Leonard views the Kenyan finance bill protests as sitting in the long shadow of a country whose politics has not fully reckoned with a colonial-era inheritance of violence. The remedy is not to pathologise youth mobilisation but to invest in the institutions that channel it, including independent media, NGOs that promote peaceful protest, and above all habits of dialogue between government and the young.

“What happened is not because of the youth,” he insists. “What happened is because of the government.”

Tanzania’s post-election repression worries him for a different reason. On top of the casualties, the longer-run cost is to a political culture that has historically been remarkably peaceful. He returns to Julius Nyerere’s warning, decades ago, that what governments do to the youth today, the youth may one day do back. To repress is to teach the next generation that violence is a legitimate tool of power. He recalls a story a colleague once told him about Ghana in the 1960s: a moment of acute political crisis, and a father’s response to his son’s question about what to do. “He said, let’s set up a newspaper.”

Leonard identifies that instinct, to reach for words rather than weapons, as the political culture worth protecting.

Share

Beyond the ‘jobs or unrest’ trap

The standard policy frame on urban youth is that they need jobs or things will fall apart. Leonard thinks this is too narrow. Plenty of young people in African cities are working; the question is whether their work is stable, dignified and connected to a sense of upward movement. Hunger, he points out, does not automatically produce crime. Instead, what produces social fragility is the absence of institutions that build social connection, the lack of credible hope, and a shortage of political education that teaches young people why peaceful organisation works.

Leonard also feels that too much of what passes for entrepreneurship policy in African cities is “all about survival”, involving the handing out small sums to street-corner businesses. The continent’s biotech labs, AI labs and university research centres are, for the most part, not being plugged into urban entrepreneurship in any serious way. The opportunity, in his view, is to push innovation-driven entrepreneurship as a real engine of urban prosperity rather than treating science and the high street as separate worlds.

Why cities are the antidote to clientelism

Wantchekon’s research, including in Benin, finds that clientelist politics is overwhelmingly a rural phenomenon. In cities, voters can see beyond their own group, so opposition parties, which tend to run on programmes rather than patronage, typically do better. Even in deeply conservative regions, places that look more like cities behave more like cities politically. He reaches for a striking example from Wyoming, where small towns near Yellowstone vote noticeably bluer than the surrounding countryside, simply because tourism and migration have made them more cosmopolitan.

Leonard argues that anyone who wants to break the grip of clientelist, divisive or far-right politics should look to urban politics as their best ally. Cities are harder to divide and harder to mobilise around fear-based rhetoric, because the people in them are exposed to too much variety to fall easily for any single story.

Subscribe now

What civic engagement should actually look like

Asked what genuinely effective civic engagement looks like, Wantchekon points to deliberative democracy. His own experimental work has tested what happens when corrupt local governments and clientelist elections are nudged towards deliberation: town hall meetings rather than rallies, structured local consultations rather than money-driven mobilisation. The results, he says, are consistent: candidates who deliberate win more votes for less spending, and local governments that deliberate become less corrupt and perform better.

“We have solutions for those issues... And those are, they can be simple. We need to introduce them. We need to expand them. We need to strengthen them.”

The institutional minimum for cities, in his view, is built around two ideas: deliberation, which generates transparency, and decentralisation, which generates agency. Here, decentralisation is not just a budgetary slogan, it is the practical claim that neighbourhood-level organisations should be empowered to take ownership of cleaning, schooling, safety and social cohesion. He worries especially about the rise of gated enclaves where the wealthy retreat from the wider community behind their own private security.

A third strand, which is often underdiscussed in economics, is culture. Music festivals, public art, places where rich and poor can mingle around something they both enjoy. These, Leonard argues, are among the lowest-hanging fruit for building real urban social cohesion in Africa, where the cultural potential is abundant and undermobilised.

“Africa in particular is a place where you have so much talent, so many people value art… music festivals in Africa, for instance, should be every month… I mean, maybe that’s too much, but every three months, because it’s such an easy, low-hanging fruit to get people to join”

The African School of Economics, and rethinking universities

“First of all, originally, it was about bringing Black faces in the profession, in economics, because the numbers are super bad.”

Leonard traces founding the African School of Economics back to a stark statistic: only around 3% of papers in top economics journals have an African co-author. Without African voices in the discipline, the kinds of arguments he is making, drawn from lived experience and detailed local knowledge, simply do not make it into the literature.

ASE began in Benin and is now expanding to Zanzibar, Nigeria and Côte d’Ivoire, with a research hub at Princeton designed to put rigorous, multidisciplinary work behind specific institutional reforms in African cities.

The broader argument he wants to make to African universities is that the standard model, teaching factories disconnected from local life, is a dead end. African scholars should be partnering with farmers, firms and governments to commercialise their research and solve real problems. Promotion criteria should reward the solving of problems, not just the counting of papers.

“You shouldn’t get promoted simply based on how many papers you publish. You should be promoted (by the) real innovation you have made, what real problem you have set.”

Where the smart money should go

Leonard wants to see funding bet on institutions like ASE: research-driven, entrepreneurial universities whose ripple effects spread through both the business ecosystem and government. He also wants to see a focus on the institutional plumbing of urban social cohesion, things like the physical and virtual spaces where people meet and trust is built.

For a continent on the verge of one of the largest urban transitions in human history, that is an argument worth taking seriously, we need to think more about culture and politics, and not focus solely on the economic benefits of cities.

Organised crime is a core development challenge for a growing number of countries around the world. Where criminal groups gain a foothold, they have a host of destructive impacts on societies and economies, with many of these manifesting in urban areas.

In this episode of Ideas in Development, Chris Blattman joined us to unpack how organised crime evolves in cities, why it is so difficult to dislodge once entrenched, and what fast-growing cities can do to get ahead of this problem.

Subscribe now

The ladder from street gang to criminal confederation

Chris is careful not to describe the rise of organised crime as an inevitable evolution. In many cities, crime remains atomised, split amongst individual burglars, car thieves, small clusters selling drugs.

But in a subset of cities, a pattern repeats. Fragmented youth gangs, rooted in particular neighbourhoods and ethnicities, begin to take on more hierarchical structures as local drug markets grow. What starts as a social grouping becomes something more like a firm. Those firms then tend to face a choice: keep fighting each other, or collude. Hundreds of small gangs competing violently for territory is a miserable business model.

“Nobody’s making money in the middle of a drug war.”

The common solution, in the cities that get this far, is confederation. Neighbouring gangs are organised, often by an outside actor, into larger alliances that fix prices, allocate territory, and police each other’s violence. Medellín has around 400 ‘combos’ operating under roughly 17 ‘razones’, each of which Chris describes as something like a franchise system sitting beneath a couple of major alliances. At the top sits an informal institution the criminals themselves sometimes call La Oficina, the office. Combos that want to go to war must ask permission.

The terrible trade-off

A city with a homicide rate in the teens is, by the standards of most large cities in Latin America, a success story. Medellín now sits at roughly 10 or 11 murders per 100,000 people, lower than Chicago, and a fraction of what it was 15 years ago, when the city was one of the most violent places on the planet. And yet, on any given street in any neighbourhood in Medellín, someone is selling drugs for a combo. Someone is being extorted for a local monopoly on gas canisters. Every square inch of the city has been claimed.

That is the puzzle at the centre of organised crime as a development problem. Peace and capture can co-exist. Violence can fall while criminal power rises. And the policies that reduce one sometimes deepen the other.

“There’s sort of three things you want as a mayor. You’d like the gangs to be poor and not very powerful. You’d like there to be peace. And you don’t want the whole polity to be completely corrupted by organised crime. And you never really get all three. You can basically choose two.”

Medellín’s story is unusual. The city’s transition from chaotic violence to criminal peace was partly a product of its particular history, the civil war, the paramilitaries, the legacy of Pablo Escobar, and partly a product of deliberate tolerance by the state. Don Berna, a former paramilitary leader, built the first version of La Oficina. When he was extradited to the United States in 2008, a four-year war of succession followed. Only in 2012, with the Pacto del Fusil (the Pact of the Machine Gun) did the current criminal peace begin to stabilise.

Subsequent Medellín governments, Chris argues, have been quietly disciplined about maintaining this arrangement. They do not extradite senior criminal leaders. They tolerate, or at least do not actively prevent, leaders governing the street from prison. When violence starts to escalate, rival bosses are found to have been coincidentally moved into the same holding area to figure things out.

Chicago has taken the opposite approach. Decapitate, decapitate, decapitate. The confederations that once stitched together the city’s gangs, organisations the Black P. Stones and others called ‘nations’, have been broken up. What remains is a landscape of hundreds of fragmented youth gangs, and the escalatory spirals of violence that have returned in recent years.

This, Chris argues, is the terrible trade-off facing any mayor confronting organised crime. Three things you might want, low violence, weak criminal groups, and an uncorrupted political system, are incredibly challenging to achieve simultaneously, especially once crime has already taken root. Medellín chose one corner of the triangle. Chicago chose another.

“I don’t envy the politicians who have to make these decisions.”

Subscribe now

Why low violence can still mean deep capture

A city at ten homicides per 100,000 can still be a city being hollowed out. Combos in Medellín generate most of their rents from retail drugs, but perhaps 20 to 30% comes from extortion. This includes local monopolies on gas canisters used for cook stoves. A cut of the eggs, the arepas, the yoghurt, the bar shots sold on the street. Tiny markups individually, which add up rapidly when applied across an entire city. And drug profits must then be laundered, which pulls whole swathes of the retail sector into the orbit of organised crime.

Political corruption follows, as at elections, gangs can deliver turnout for favoured candidates.

Diagnosis, not prescription

The most common mistake Chris sees among policymakers is the search for a transferable template, a silver bullet, or a case study to copy. A mayor looks at New York’s homicide decline and wants CompStat and hotspots policing. Another looks at Chicago and wants violence interruption. Another looks at El Salvador and wants to emulate Bukele.

You wouldn’t trust a doctor who skips the diagnosis, and goes straight to prescribing you the exact same treatment as their last patient. But somehow, when it comes to crime, policymakers are turning off that part of their brain.

Bukele’s mass-incarceration approach to reduce crime worked (to the extent it did, as it’s still early days and there are many related legal/human rights issues), because El Salvador’s gangs were structurally weak. They were dependent on extortion, had no significant retail drug market, and no base of citizens clamouring for their return. Apply the same strategy to a city with an established drug trade and the demand for drugs remains untouched.

“If you did that in Medellín or in Chicago, and you just could wave a magic wand and pop all of these people into a Megamax prison somewhere, you would have thousands, tens of thousands of people who were buying their drugs who were like, ‘Where do I get my fix?’ And they’re going to be clamouring for somebody to come back and someone will fill that gap.”

What’s needed is proper diagnosis. And what’s striking is how few cities seriously invest in it. Senior security officials, mayors and police chiefs often do not grasp what is actually happening in their own cities, particularly in places like Ecuador, Peru and Chile, where violence has erupted in the past two years after long periods of relative calm.

Chris and his colleagues are now building what are in effect criminal diagnosis labs, in Medellín, Rio and Mexico City. By pooling interviews with gang leaders, middle managers and thousands of schoolchildren into open-source, anonymised datasets, the goal is to systematise how cities understand their own crime problems before the next mayor rotates in.

Share

The warning signs for African cities

“My worry is that in 30 or 40 years, Lagos looks like Rio.”

Several African cities are becoming attractive transshipment points for cocaine and pills moving between South America and Europe. Nascent gang structures, from scamming operations in Lagos to groups like the Mungiki in Kenya, look eerily similar to the decentralised youth gangs that eventually consolidated into Latin America’s confederations. Meanwhile, the state apparatus in Africa is often highly centralised, with security attention focused on rural insurgencies, and mayors having little real power.

The lesson from Medellín, Bogotá and Santiago is that retail drug markets are built slowly over a couple of decades, but once entrenched are almost impossible to dislodge. That time horizon is also the window for action. This involves keeping ports unattractive to traffickers, disrupting the spillover from transshipment into local retail, and, above all, investing in the kind of granular diagnosis that lets a city properly understand its own problem.

“These retail drug (markets) – they’re a little bit like herpes. Once you’ve got them, you’ve got them.”

If that window closes, the choice African mayors will not get to decide whether to tolerate organised crime, but only which corner of the trilemma they must accept.

“Not for sale” signs are commonplace in many African cities, reflecting people’s lack of trust in the systems that record and protect land ownership. In fast-growing cities, this breakdown in land governance sits beneath almost every other development problem, housing shortages, unplanned sprawl, infrastructure gaps, and weak investment.

“I would actually be hesitant to invest in such kind of environment… because I know I would be running a risk of losing my investment.”

In this episode of Ideas in Development, we are joined by Thierry Hoza Ngoga, one of the leading implementers of Rwanda’s landmark land regularisation programme, to walk through how Rwanda registered over 10 million parcels, issued more than 7 million title deeds, and did so at a cost of roughly $6 per parcel.

Subscribe now

What land reform actually means

Land reform is often reflexively associated with redistribution, i.e. taking land from one group and giving it to another. Thierry broadens the definition considerably. A comprehensive land reform agenda covers four areas: improving access to land (including making transactions easier); promoting productive land use (urban, agricultural, industrial); building the legal and regulatory systems that protect land rights; and establishing the institutional governance structures to administer all of this.

Security of tenure, trusted processes, a functioning legal framework, and clear institutional mandates are the preconditions for functioning urban land markets, and in many African cities, several are missing at once.

Rwanda before reform: 70% unplanned, titles for the few

Before the reforms, Kigali’s land management fell under the city mayor, who doubled as registrar of land titles. Outside Kigali, titles had to be signed by the minister responsible for land. Only a tiny fraction of parcels were titled at all, and around 70% of Kigali was considered unplanned, a level common across many African capitals.

Three pressures made reform urgent. First, Rwanda’s ambitious Vision 2020 development strategy placed land reform as a core pillar. Second, the aftermath of the 1994 genocide against the Tutsi had created enormous displacement, and the potential for land-related conflict was acute. Third, Rwanda is one of Africa’s smallest and most densely populated countries, with a majority dependent on land for their livelihoods. Using that scarce resource productively was existential.

Critically, land reform was embedded in the broader economic transformation agenda rather than treated as an isolated technical programme.

“In many cases you find land reform as an isolated program on the side. It’s not part of the broader economic transformation.”

Share

How Rwanda actually did it

The sequence was deliberate. A national land policy in 2004 set out the government’s position and ambitions. An organic land law in 2005 established the legal framework, categorising land types, defining rights, and creating institutional mandates. But policy and law alone don’t define how things work on the ground.

So the government ran extensive consultations with different land-use groups to gauge whether there was genuine public demand for reform and to understand existing informal systems of land exchange. The demand was overwhelming, so this was followed by a pilot exercise across four districts (chosen to represent the country’s diversity: urban, peri-urban, rural, and areas with larger holdings), which registered 14,000 parcels and produced the technical tools and secondary legislation needed for national rollout.

What made the programme distinctive was its community-led design. Demarcation of land boundaries and recording of ownership were carried out by locally trained people, rather than surveyors from the capital. This built trust, because people could see that no one else was deciding their land rights on their behalf. By the time demarcation and adjudication were complete, the programme had employed 1% of Rwanda’s entire population.

The approach was also deliberately affordable. The government subsidised those who couldn’t pay, and fees for everyone else were minimal. Rwanda used satellite imagery rather than expensive fixed-boundary surveys with concrete beacons, which is a major cost driver elsewhere.

The contrast with other countries is stark. Neighbouring countries attempting similar programmes have seen costs rise above $3,000 per parcel, but Rwanda achieved it for $6.

Learning by doing: What nobody anticipated

No country had done what Rwanda was attempting, in terms of scale and timeline, so there was no model to follow, or existing blueprint. Implementation surfaced many unforeseen issues. In fact, the very first parcel the team tried to demarcate belonged to a polygamous man, and since Rwanda’s constitution recognises only one legally married wife, the team immediately had to adapt.

Other challenges included cloudy satellite imagery that obscured boundaries in certain areas, and informal settlements where houses were so densely packed that physical boundaries were almost invisible. Each time, the team responded in the same way, by bringing people together, acknowledging the problem, and working out a solution on the spot.

“It was a learning by doing exercise”

Subscribe now

More winners than losers

Any land reform creates losers, particularly those who profited from the opacity of the old system. But Thierry argues the winners in Rwanda vastly outnumbered those losers.

Amongst the winners, women have perhaps been the most significant beneficiaries. Traditionally excluded from land inheritance, they now have their names recorded on titles – and if legally married, they hold a 50% share. The national land register shows that women are now the majority of sole landowners, which is a remarkable shift.

On top of this gender dimension, orphans of the genocide now have documented rights that protect them from expropriation, foreign investors have clarity on where to go and what processes to follow, and many government institutions now use the land register as a backbone for their own sector development.

One question we had for Thierry, was what about credit markets, as per the famous Hernando de Soto thesis?

“A land title brings that certainty, that security that the lender needs, but it’s not sufficient.”

Thierry highlights that banks also want to see the borrower’s ability to repay and an active land market in case of default. In Kigali and secondary cities, many urban residents are indeed using land certificates as collateral for mortgages, but they tend to be people with formal employment. In rural areas, microfinance borrowers use certificates for small business loans.

Land reform lessons for other African countries

Rwanda’s reform, a process started in the aftermath of the genocide, was shaped by a number of unique circumstances, including it’s small and relatively homogeneous territory. So Thierry is realistic about the limits of direct replication, but emphasises that the core lessons are transferable.

Start small. Countries don’t need a national programme or even a completed legal framework to begin their land reform journey. A single municipality can pilot land recording, learn from the experience, and build evidence for scaling.

“If you start small, you perfect your model” Thierry argues. “You design it in a way that is building on the reality on the ground.”

Make it a national priority, not a side project. When land reform is treated solely as the mandate of an underfunded, understaffed land ministry, it fails. But if it is embedded in a country’s broader development strategy, with coordination across central and local government, it gains momentum and resources.

Leverage technology. Drones, improved satellite imagery, and digital registration systems have made mapping and recording far cheaper and faster than when Rwanda began decades ago.

Build trust before expecting demand. One puzzle in this space is the seemingly low demand for titles in some African cities, where officials report spending money on cadastral surveys only to find that nobody comes to collect their title. Thierry explains why a title alone is not enough.

“People will never come if they don’t trust the system in place.”

People need to see that the system issuing titles is reliable, that disputes will be resolved, and, crucially, that titling comes with broader benefits: infrastructure, services, and a genuine improvement in their lives. Without that package, a piece of paper is just seen as an invitation to pay more tax in the future.

Every generation has some version of a ‘build or be priced out’ story. In cities like Los Angeles, San Francisco and New York, jobs and opportunity have piled in, and housing has not kept up. In response to this, the YIMBY movement – “Yes In My Back Yard” – has started to become a real political force in some parts of the world.

But this story is playing out across the world, and in some regions, the wave of residents arriving in cities dwarfs that of US cities. Where entire neighbourhoods that don’t exist yet will have to be built from scratch. And where being pro-housing is far more complex, touching on land, roads, water, electricity, tenure, finance, construction systems, and the basic institutional plumbing that makes building at scale possible.

Over the next 25 years, Africa is expected to add close to a billion new urban residents. That’s the largest wave of urban population growth the world has ever seen. A defining question of this century is therefore: will Africa’s cities be able to absorb this growth in a way that creates prosperity, or will high housing and infrastructure costs choke off the benefits of urbanisation?

Subscribe now

In this episode of Ideas in Development, we are joined by Kecia Rust, executive director and founder of the Centre for Affordable Housing Finance in Africa (CAHF), to discuss the full housing chain in Africa – from land and finance to construction and rental markets – and what it would take for African cities to absorb nearly a billion new residents over the next quarter century.

Africa’s housing chain and its weakest links

Housing delivery is not a single act but a chain of interdependent links. Kecia walks through two intersecting sequences. The horizontal chain runs from land and tenure, through infrastructure, construction, and on to management and maintenance. The vertical chain is finance: every link in the delivery process has to be paid for, and the financial products that make this possible – construction finance, mortgage finance, housing finance (mortgages and/or micro-mortgages) – depend in turn on capital markets and institutional investors higher up.

The chain is only as strong as its weakest link. And in African cities, Kecia argues that the most binding constraint cutting across every link is local government. Municipalities control tenure processes, approvals, bulk infrastructure connections, and regulatory compliance. When they lack the capacity to perform these functions, which is common, the entire system slows down, costs rise, and development spills into informality.

A housing development in Kenya is an illustrative example. After receiving municipal approval, the developer waited for the promised electricity connection, but it never came. They had to sacrifice one of their 20 planned plots to build their own substation and water treatment plant, meaning the cost of the entire project now had to be recovered from 19 houses instead of 20, pushing prices up for buyers.

Share

Informality as opportunity, not just failure

This is where Kecia parts ways with a common policy instinct. Informality in housing is often treated as purely negative, equated with slums and deprivation, and seen as something to be eradicated. Kecia sees it differently.

In Nairobi’s Pipeline area, informal developers have built eight-storey tenement walk-ups without government approval. The government is now pursuing regularisation, which is reasonable.

“That housing is serving a target market that no one else in Kenya is serving.”

It sits in a price bracket above informal settlements but below anything the formal market or government subsidy programmes deliver, and it is in a prime location.

The same logic applies to South Africa’s backyard rental market, where homeowners build small rental units on their properties. This sector delivers more housing than any other form of supply in the country, and it creates small-business income for low-income homeowners who would not otherwise be on the supply side of the market.

Kecia cautions against romanticising informality, but emphasises that the energy, capacity, and market responsiveness within the informal sector represent an enormous opportunity. The policy question should not be how to eliminate informality but how to work with it: regulate what matters for safety, reduce barriers to entry, and recognise what is already working.

Mortgages, micro-mortgages, and the missing middle

Mortgages are a powerful instrument, enabling people to buy a house before they can pay for it outright, who then repay over subsequent years while living in the property. But they depend on conditions that are often absent in African markets: clear title, functioning foreclosure systems, formal employment, and affordable interest rates.

In many African economies, base interest rates are so high that a long-term mortgage becomes the equivalent of buying a house on a credit card. In these circumstances, this instrument often doesn’t make sense, so shorter-term ‘microloans’ and incremental building is, in practice, how most housing on the continent is already built.

The real frontier is therefore the micro-mortgage: loans of under $10,000, with shorter tenors, that can serve the vast majority of the market that conventional mortgage products cannot reach. India is making progress here, but in Africa the space remains almost empty. A global fund set up to invest in housing micro-lenders could find only one eligible institution on the entire continent.

“People earn their money differently now than when that mortgage was first invented.”

There are bright spots. In South Africa, First National Bank has started targeting the resale market for government-subsidised housing – properties now entering a second generation of ownership at prices that suit a micro-mortgage product of $10,000 to $15,000. A company called Empowa is using blockchain technology to recognise the irregular income patterns of borrowers who earn from multiple sources, enabling instalment-sale agreements that a conventional bank would not touch.

Subscribe now

Rental markets: The unsung engine of African housing

While micro-mortgages are the frontier for ownership, rental markets are the under-appreciated backbone of African housing systems. In Kenya, ~90% of renters rent from individual landlords, not institutions. In South Africa, backyard rental is the single largest source of new housing supply.

Rental serves multiple functions. It is the entry point to city life for new migrants. It provides income to low-income homeowners who become small-scale landlords. And it creates the liquidity that makes the broader housing market function, as people need to be able to move in and out of properties at different life stages.

Kecia also makes a compelling environmental case. African inner cities are full of derelict or abandoned buildings that represent enormous untapped potential. In Johannesburg, a mortgage lender called Tuff has carved out a niche providing mortgages to landlord-entrepreneurs who buy, refurbish, and rent out units in neglected inner-city buildings.

“The greenest housing we’ve got is the housing that’s already built because it doesn’t add any more carbon.”

What YIMBY actually means in Africa

The YIMBY movement in high-income cities essentially boils down to legalising density and reducing veto points. Translated to Africa, the spirit is similar but the specifics differ enormously. The majority of the population is already locked out of formal housing systems, so the question is not overcoming nimbyism among wealthy incumbents, but making the formal system recognise and work with the enormous capacity that already exists outside it.

Kecia points to Cape Town as an example. A new bylaw recognises micro-builders as legitimate participants in the housing delivery system and classifies backyard rental as a formal house type.

“All it took was a bylaw. They’re still doing it the same way but now they can actually be regulated as they exist and then they can be counted.”

The city also set up a building control office in Khayelitsha, a low-income township, which became the highest-performing office by volume of plans processed.

Other bright spots include Development Workshop Namibia, which formalises site-and-service delivery so that households can build incrementally on serviced plots; South Africa’s growing multifamily residential rental sector, which is creating a new asset class that attracts institutional investors to moderate-income housing; and Nigeria’s Family Homes Fund, which has developed a suite of instruments addressing different links in the housing chain simultaneously.

What a pro-housing government should do

If a mayor came to Kecia and said they wanted to be a YIMBY government, her answer would be direct: make it easy. Finance is a constraint, but it is a constraint largely because risk and cost are so high. If local governments streamlined approvals, delivered the bulk infrastructure they promised, and reduced the regulatory friction that inflates development timelines, private capital would follow.

“The only reason that finance is a binding constraint right now is because it’s so expensive because the risks are so high.”

Beyond reducing friction, governments need to work across the full housing finance chain. That means getting macroeconomic policy right – high treasury bill rates crowd out mortgage lending, so there is a long-term growth case for keeping interest rates low enough to support a functioning housing market. Creating regulatory space for micro-mortgages is also key. As is supporting finance for those constructing housing, not just end-user finance. It also means finding ways to bring pension funds and institutional investors into the housing space by managing the risks that currently keep them out.

Above all, this requires recognising the breadth of supply that already exists.

“from a mother whose kids have gone to school and now she can offer rental accommodation… all the way through to somebody who can go and buy a block of flats.”

African housing markets are far more dynamic than policy frameworks give them credit for. The task is not to try and replace that dynamism with top-down delivery, but to create the conditions in which it can operate safely, at scale, and to the benefit of the majority.

Africa’s cities are growing at an unprecedented rate. So, understandably, African mayors currently have a long shopping list. They need to build roads, drainage, water, power connections, housing infrastructure, public transport, waste systems, and the other basic public services that make dense urban life actually liveable. But there is an important paradox here.

Despite the pressure on mayors to deliver on all of these priorities, in many places they don’t actually control the big fiscal levers. City revenues are low, transfers from the central government can be unpredictable, and city borrowing is either legally restricted or wildly expensive. But hard is not impossible. Other countries have been able to build the institutions that allowed subnational governments to access long-term finance, raise revenues, and invest at scale.

In this episode of Ideas in Development, we are joined by Astrid Haas to discuss why African cities are so fiscally constrained, what reforms have worked elsewhere, and what it would take to unlock the finance that fast-growing African cities urgently need.

Subscribe now

Being an African mayor is hard!

Astrid starts by outlining just how hard being the mayor of an African city is. Imagine managing a city of four million people on an annual budget of $25 million. Astrid recently attended a panel discussion where this was the reality described by the mayor of a large African city – their European counterpart on the panel could not comprehend such a low budget. It captures the central paradox of urban governance across much of Africa: mayors are locally accountable for delivering infrastructure but they don’t control the fiscal levers needed to pay for any of it.

The fiscal paradox of African cities

African cities occupy an awkward position in the architecture of the state. On paper, they are legally recognised entities with responsibility for infrastructure and services. In practice, they are decentralised entities trapped in fiscally centralised systems. They carry enormous delivery obligations but have limited powers to raise revenue, borrow, or plan with confidence.

“You are responsible and you are locally accountable, but you don’t have the money to do so”

The result is that most African mayors function less as CEOs of their cities and more as delivery managers for budgets they didn’t set and can’t control.

But Astrid argues that the best city leaders refuse to accept scarcity as the endpoint. The successful ones treat it as a starting point, working to build coalitions, expand fiscal space, and find ways to do more with what they have. That requires leadership skills more commonly associated with entrepreneurs than bureaucrats: coalition-building, communication, and the ability to generate trust.

The social contract in cities

The ability to raise revenue locally is fundamental to the social contract. If a city is responsible for delivering services but someone else raises the money, the accountability link between citizens and their local government breaks down. Conversely, when cities raise their own revenues, citizens can see what they’re paying for and hold leaders to account.

There is also a practical benefit. A steady stream of own-source revenue is the foundation for borrowing. Without it, cities cannot demonstrate the creditworthiness needed to access longer-term finance for infrastructure investment.

That said, Astrid is careful to note that no city can survive on own-source revenue alone. City tax bases are inherently smaller than national ones, so intergovernmental transfers will always be part of the picture. The question is getting the balance right between the different sources, and making sure transfers are predictable enough that cities can actually plan around them.

Share

Why cities aren’t borrowing – and when they should

Given the scale of Africa’s infrastructure needs, it might seem obvious that cities should be borrowing more. But Astrid pushes back on the assumption that borrowing is always desirable. Cities need the fiscal capacity to repay without compromising service delivery. And not all projects are suitable for debt finance. Public goods are public goods for a reason, and not everything can be packaged into a bankable project.

But for those cities that can and should borrow, the barriers are formidable. City finances are often opaque, with weak auditing and financial reporting that make them unattractive to investors. National legislation can be overly restrictive. Political dynamics can play a role too, as national governments sometimes resist cities becoming financially independent, particularly where cities are opposition strongholds. And critically, there is a shortage of local currency finance – cities should not be borrowing in hard currency because, unlike national governments, they cannot hedge against foreign exchange risk.

On the supply side, domestic financial markets in many African countries are simply too shallow. Pension funds and savings exist but are not being channelled towards local infrastructure. So Astrid argues for strengthening financial intermediaries institutions that can take on foreign currency loans and convert them into local currency lending for cities, while also bringing an understanding of the domestic market.

Lessons from Mexico, the Philippines, and France

The good news is that other countries have faced similar problems and found institutional solutions.

In Mexico, the key reform was stabilising intergovernmental transfers. Once subnational governments knew their transfers would arrive predictably, on time and in a known amount, they could borrow against them. Investors, in turn, had confidence they would be repaid, because the first portion of each transfer was ringfenced for debt service. The reform unlocked a new source of finance by simply making an existing revenue stream reliable.

The Philippines took a differentiated approach in its 1996 local government framework, recognising that not all cities are in the same position. Large, fiscally sound cities were allowed to go to market. Mid-sized cities with emerging creditworthiness could borrow from national financial intermediaries. Small cities that were not yet creditworthy were prioritised for national transfers and subsidies. This freed up the public budget to support the places that genuinely needed it, rather than spreading resources thinly across all cities regardless of capacity. Alongside this, the Philippines created a municipal development fund office and a local guarantee unit that provided government guarantees to crowd in private finance and, crucially, conducted local credit ratings, assessing cities against each other rather than against international standards.

France’s ‘caisse des dépôts et consignations’ model offers another template. This is an institution linked to the national treasury that accumulates domestic savings and lends them locally. Variants of this model exist across Francophone Africa, and similar principles underpin Morocco’s FEC and the Development Bank of Southern Africa, which has one of the largest municipal lending portfolios in the developing world.

Subscribe now

Freetown proves it can be done in Africa

The most striking example in the episode comes from Freetown, Sierra Leone, where Mayor Yvonne Aki-Sawyerr inherited a city with high urbanisation rates, crumbling infrastructure, and a broken fiscal contract. She needed money to enact reform, but citizens had no reason to trust that their taxes would be well spent.

Her approach started not with raising more revenue, but with spending existing money more effectively, through a plan called Transform Freetown, which was built with extensive input from residents. Only once citizens had bought into what their money would be used for did she push through a property tax reform – often called the tax people love to hate, given its visibility. She did so during COVID, when most cities worldwide were waiving or reducing taxes, overcame resistance from the national government, and ultimately succeeded.

“People are actually willing to pay tax if they know what they’re going to get from that tax”

The lesson is that revenue reform cannot be separated from expenditure reform as the two go hand in hand.

“You don’t need a perfect system… you need to be able to do reforms that build credible momentum over time and then build on that momentum.”

Freetown is now in the process of raising finance for the city’s first cable car, which reflects how early fiscal credibility can compound over time and open up larger possibilities.

What Africa can, and can’t, learn from urban finance in China

China is often held up as the gold standard of rapid urbanisation. And Astrid identifies some genuine lessons. For example, China allowed local governments to experiment within centrally set parameters, rather than imposing a single model; it leveraged land-based financing tools to fund infrastructure; and it built infrastructure ahead of settlement, avoiding the far higher costs of retrofitting.

But hindsight reveals serious some risks. Land was over-leveraged in many places, and once it was leased there were few alternative revenue sources. China still does not have a property tax, and introducing one now, when property values are already high, is politically far harder than it would have been at the outset. African cities should take note of this – delaying a property tax makes it more difficult, not less. And China’s growth path was built on carbon-intensive infrastructure that African cities cannot afford to replicate in a climate-constrained world.

What national and city governments should do now

Astrid’s concluding recommendations run in two directions.

For national governments, the priorities are threefold:

1. Make intergovernmental transfers predictable in both timing and amount, so cities can plan and potentially borrow against them.

2. Create differentiated frameworks that allow creditworthy cities to borrow while directing subsidies to those that cannot.

3. Develop or strengthen financial intermediaries that can channel domestic and international capital into local-currency lending for cities.

For city governments, the task is to get their own house in order, by improving financial transparency, investing in administrative capacity, and demonstratin that they can raise and spend money effectively. Creditworthiness needs to be built, incrementally, through credible fiscal management.

No country ever got rich without also urbanising. Cities make people more productive, more innovative, better paid, and, in the modern age at least, healthier and longer-lived. A rough rule of thumb from the research is that every 10% increase in a country’s urban population share is associated with an increase in output per capia of around 30%. Between 2010 and 2020, 80% of private sector jobs created globally were generated in big cities.

But urbanisation alone is no guarantee of development. And in the world’s fastest urbanising region, the link between cities and prosperity is breaking down. In this opening episode of our new Ideas in Development series on cities, we are joined by Kurtis Lockhart, founder and executive director of the Africa Urban Lab, to set the scene. We discuss why everyone who cares about development should be thinking about Africa’s urbanisation, what’s currently going wrong, and where policy changes should happen.

Subscribe now

The scale of what’s coming

By 2050, roughly 900 million people will be added to African cities, more than doubling the continent’s current urban population. That’s the equivalent of building the entirety of American and European city systems from scratch, in 25 years.

African cities are growing at roughly 3.5 to 4% per year, meaning many will double in size every 18 to 20 years. Some faster-growing cities will triple or quadruple in that period.

And this building boom will happen at income levels far below those at which other regions urbanised. When the United States hit 50% urbanised around 1920, per capita incomes were roughly $7,500. When South Korea crossed the same threshold in the late 1970s, incomes were around $4,500. Much of Africa is approaching 50% urbanised at per capita incomes of around $2,000 to $2,500. This means a far smaller tax base to finance the water, sanitation, roads, power, and housing that functional cities require.

About half to two-thirds of the physical urban space Africa will need by mid-century does not yet exist. It all needs to be built, and getting that right is crucial, as once it’s built, it’s sticky. Urban form has a lock-in effect that makes poor early choices extremely costly to reverse.

Why Africa’s urbanisation is different

In a paper titled Urbanisation With and Without Industrialisation, Doug Gollin, Remi Jedwab, and Dietrich Vollrath show that across every region in the world bar one, urbanisation is positively correlated with structural transformation, the shift of economic activity from agriculture into manufacturing and services. Every region, that is, except Africa, where the relationship is a flat line.

“if this link is not restored, it effectively means that the nearly one billion people being added to these cities between now and mid-century… they lead worse lives.”

Without structural transformation, rapid urban growth produces not prosperity but informality, slums, disconnected sprawl, low productivity, and limited upward mobility. So, getting African urbanisation right is, in Kurtis’s view, “one of the biggest policy imperatives of the 21st century.”

Cities as engines of growth

Kurtis’s interest in cities ultimately boils down to economic growth itself. Cities, he argues, are a primary driver of that growth. The evidence is extensive: cities boost productivity at both the individual and firm level, increase human capital, concentrate innovation, and generate higher wages. The mechanisms are the classic Marshallian forces of agglomeration, i.e. thick labour markets, supplier networks, and knowledge spillovers, amplified by the density of interaction.

But Kurtis is quick to add that this is not purely a laissez-faire story.

“Just as there are no atheists in the foxhole, there are no libertarians in the city” Kurtis Lockhart, borrowing a quip from Ed Glaeser.

The externalities that arise from density, both positive and negative, are so large that effective urban governance is indispensable. Markets alone will not produce functional cities.

Bridging planners and economists

One of the Africa Urban Lab’s core missions is to close the gap between urban planning and urban economics, two disciplines that study the same object but often talk past each other.

Kurtis sees planners as tending to place too much faith in top-down rules, without adequate tools to anticipate the second and third-order economic consequences of their decisions. Zoning rules, building codes, and land-use regulations all have enormous implications for housing costs, firm location, labour markets, and growth, which planning curricula often fail to teach.

Economists, on the other hand, can be too comfortable with the idea that markets will sort things out. In cities, where congestion, pollution, crime, and infrastructure all involve massive externalities, they won’t.

The Lab’s professional diploma programme, currently training 50 students from 15 countries, is designed to sit at this intersection, equipping urban planners and city leaders with the economic reasoning they need to make better decisions, and grounding economists in the institutional realities of how cities are actually governed.

Share

The Africa Urban Lab: Research, training, and implementation

Housed at the African School of Economics in Zanzibar, the Africa Urban Lab operates across three pillars.

The first is education. The diploma programme trains mayors, deputy mayors, senior planners, and ministry directors from across the continent. A key pedagogical commitment is integrating the insights of urban economics, such as the work of Alain Bertaud, whose book Order Without Design Kurtis recommends as essential reading for any planner wanting to understand the economic consequences of planning decisions.

The second is research. The Lab conducts rigorous academic work across four research themes (detailed on their website), and is launching a network of fellows this month for scholars working on African cities.

The third, and Kurtis emphasises this as critically important, is policy implementation.

“We realised that we can write these amazing research papers and these amazing policy recommendations, but nobody reads them”

Two active workstreams translate research into practice: an urban expansion programme working with around 30 fast-growing East African cities on proactive planning for peripheral growth, and an urban finance programme focused on helping cities raise revenue, improve budget execution, and eventually tap into domestic capital markets.

The stakes are high

Africa is home to 16% of the world’s population but roughly two-thirds of the world’s extreme poverty. By the end of the century, it’s estimated that four out of every five human beings on the planet will be African. Post-2050, Africa will be the only region driving global population growth.

These demographics make the urbanisation question inescapable. If the link between cities and prosperity can be restored in Africa, through better planning, smarter finance, functional land markets, and effective governance, the payoff will be transformative. If it cannot, the consequences will be felt not just on the continent but globally. As Kurtis puts it, this is not an abstract policy debate. It is about whether nearly a billion people arriving in cities over the next quarter century lead worse lives or are given the chance to build better ones.

Across our series on AI, Deena Mousa and I spoke with researchers and policymakers on topics relevant to the AI and development debate. In this episode we took stock of our main takeaways from these conversations, explored some of the many unanswered questions, and recapped what Deena learned at the AI Impact Summit in India.

Subscribe now

What happens when AI meets the world?

There are fifty million cases backlogged in India’s courts. An organisation called Adalat AI is building AI tools to help, such as their transcription tool. When the founder, Utkarsh Saxena, described the deployment experience, one detail stood out. Judges, confronted with a new piece of technology that could transform their work, were stuck on the most basic questions.

“How do I update Chrome?”

This is a good summary of what this series has been about – what happens when AI meets the real world? This tends not to be the question technologists are answering when they talk in broad strokes about the capabilities of their models, and how fast they are improving, against different benchmarks. But economists ask a different set of questions, what actually changes in an economy when a new technology arrives? Who captures the gains? How long does it take? What gets in the way?

Those second-order questions are the ones that determine whether a technology that works in a lab changes anything for people living in Nairobi or Karachi.

What is the binding constraint?

This is a useful framing to have in mind when looking to apply AI in any development context. Ask: what are the binding constraints here, and does AI relax them?

There is an instinct to layer AI onto whatever process already exists, but the more useful approach starts from the problems which AI is actually good at solving, and how to put AI in a position to help.

The series gave several clear answers about where the constraints actually are. Umar Saif put the first most sharply: large parts of the world are data deserts, where the foundational information AI needs to be useful simply does not exist. No patient records to train a diagnostic model on. No land registries. No systematic crop yield data. Rose Mutiso made the same point from an infrastructure angle: the absence of underlying data isn’t a problem of being behind on adopting the latest model. It’s a problem of missing inputs.

This puts the national AI strategy conversation in a different light. Announcing a national AI policy produces summits and documents, and a new technology park might play well in a TV ad. Digitising land records is not sexy, and bumps up against a lot of political economy constraints. But the latter is what actually creates the conditions for AI to be useful.

Share

What universities actually do

Josh Lerner’s research added another dimension to the foundational investments that matter. The finding that stuck was about place: faculty members who moved from Duke to Stanford became more commercially oriented, while the reverse move had the reverse effect. Roughly 20 - 25% of that shift was attributable to the environment itself, i.e. the norms, the networks, proximity to investors, and the institutional culture around risk-taking.

For LMICs, the implication is that funding more research or training more scientists is not sufficient on its own. The environment in which those scientists work determines whether their ideas become companies, and whether those companies stay. Building that environment is a slower and less legible investment than building a data centre, which is precisely why it tends to be underinvested.

The disruption to development escalators

Perhaps the most uncomfortable finding across the series was not about the obstacles to adopting AI, but about what AI may do to growth models that currently serve as a ladder of development.

The traditional path ran from agriculture to manufacturing to services. India rode services, including software and back-office work. Many other countries had been hoping to board a similar escalator, particularly given the difficulties of replicating East Asian export-led manufacturing growth in 2026. But as Umar pointed out, many basic Fiverr-style tradeable services – copyediting, logo design, SEO-optimised content, basic coding etc. – will be displaced by AI relatively rapidly. The door is slamming shut on the opportunities which India took advantage of around the Y2K scare, and it’s unclear what, if anything, will replace that bottom rung.

Raghuram himself was optimistic about India, as their labour cost differential persists even when both workers have AI. But India has progressed beyond the bottom rung of basic service exports. Whether that optimism extends to lower-income countries just trying to get onto the ladder is a genuinely open question.

So developing countries are in a double bind: AI is harder to adopt in LMICs because the prerequisites are missing, and simultaneously it is threatening the growth model they are already running.

What is the bottom rung in 2030?

Manufacturing worked for a very specific combination of reasons: it was tradeable, scalable, absorbed large numbers of workers with modest skills, and generated productivity growth through learning-by-doing. Some services were able to replicate these properties. What AI does to services, and could do to manufacturing (pending cheaper, better robots on the factory floor), is precisely target the tradeable, routine, scalable parts of both.

Bruno Caprettini’s research on the industrial revolution is useful here. The disruption from threshing machines was less damaging near industrial towns, because conditions existed for workers to move toward emerging alternatives. The policy implication is therefore not to identify the new escalator and build toward it. It is to make sure the conditions exist for workers to match with whatever opportunities do emerge. Strong education systems, social protection, connectivity, and urbanisation that puts people in proximity to economic activity.

Access versus capture

Josh Lerner’s research also pointed to a distinction that tends to get lost in AI and development discussions: the difference between access and value capture. Access means you can use a technology. Capture means you participate meaningfully in the value it creates, the jobs, the tax revenues, the firms. The historical pattern suggests those can diverge for decades. A world where every country gets access to excellent AI tools is entirely compatible with a world where the majority of financial gains accrue to a small number of hubs for decades.

The partial counterpoint Josh identified is South-South knowledge exchange. When China rose as a major innovation hub, the effect spread: entrepreneurs in other emerging markets began adapting business models forged under similar constraints in China, e.g. limited credit card penetration, unreliable infrastructure, sparse data. The constraints in one emerging economy are often much closer to those in another than to those in the United States, so Chinese innovations travelled in ways Silicon Valley innovations did not.

AI working through humans

Across the concrete cases in the series, a consistent pattern emerged: AI performing best when it works through, rather than instead of, people. Adalat AI is not replacing judges, it is removing the administrative burden that essentially turns them into scribes, enabling them to hear more cases.

And for the smallholder farmer who has never had access to an agronomist, the student whose teacher has forty children in a class, the counterfactual to AI assistance is not a professional being displaced. It is nothing. That is a fundamentally different economic calculation, and one that points to AI’s particular potential in many development settings.

What remains unresolved

The measurement problem Anton Korinek and Josh Lerner both flagged is a key challenge. Counting job postings, running benchmarks, and tracking AI usage data only captures who uses AI and how often, not what actually changes inside a job when someone does. A lawyer using an AI drafting tool might bill the same hours, work less hard, or take on more clients. And usage data is almost entirely invisible on displacement: the worker who was not hired because the firm used AI instead, the BPO contract that was not renewed. Those effects may be the most economically significant ones in the short term, especially in LMICs.

The cost trajectory changes many calculations, but perhaps not in the direction one might hope. Umar’s point that capable models may be running on phones within five years is plausible. If that happens, the access barrier largely dissolves. But that makes the data and infrastructure argument more urgent, not less. Cheap models running on bad data, in languages they were not trained on, without institutional scaffolding, do not represent a development opportunity.

And then there is the question of whether development economics as a discipline is positioned to grapple with any of this at its full scale. Running RCTs on individual AI applications is genuinely valuable, and there will be hundreds of them. But the big-picture questions – scenario planning, macro implications, growth model disruption – require the kind of speculative, system-level thinking that academic incentives currently work against. I’ve expanded on this in my new blog - there is no randomising a technological revolution.

If you’ve enjoyed this series, please do share the episodes with your friends and colleagues.

Next up, seven episodes on cities with Kurtis Lockhart!