• Adnan Khan writes about state capacity and why it matters for the Economics Observatory.

• Daniel Yu discusses export manufacturing and labour mobility on the Africa Urban Lab Podcast.

• Y-RISE is calling for papers on the complexities of scaling up policy interventions in global development. Deadline: September 15.

Political polarisation is rising around the world. Social media echo chambers, divisive political rhetoric, and increasingly disconnected political realities may all be contributing – with potentially serious consequences for democracy, economic decision-making and, in some countries, the risk of conflict.

But can polarisation be reversed?

In this episode of Economics Unpacked, Cesi Cruz and Horacio Larreguy explore what research tells us about the causes and consequences of political polarisation – and the interventions that could help societies manage political disagreement.

The research suggests that tackling polarisation does not require everyone to agree. Instead, relatively simple interventions may help people see political opponents as fellow citizens rather than threats – making disagreement more manageable.

Transcript

Polarisation is often defined as the process by which a population divides into opposing camps with increasingly divergent attitudes, values, or identities, reducing the middle ground between them. This divergence is rising around the world, but the cost of polarisation may be even higher in developing countries. Why is this happening, and what, if anything, can be done about it? Welcome to Economics Unpacked, a VoxDev series where we try to answer the big questions in development.

Cesi Cruz: Most people treat polarisation like it’s an informational problem, a legal problem, even a political one, but I think fundamentally it’s a relational problem. Polarisation comes in different flavours. The one that we’re most concerned about when we see it in the news is something called affective polarisation – we view the other group as being a threat to us rather than just people that we disagree with. Other types of polarisation, such as policy polarisation, for example, we tend to see as less problematic, in the sense that some level of healthy disagreement is normal and even good for democracy.

Horacio Larreguy: Polarisation is certainly a phenomenon that we observe pretty much everywhere. Not so much in Africa, where historical levels of political polarisation have been quite high. In the European context, you see the rise of extreme right-wing parties. You always had a lot of left-wing populism in Latin America; now we’re also seeing extreme candidates on the right.

Cesi Cruz: In developing countries, the concern with political polarisation is that it’s not the worst thing on the spectrum. Something that’s even worse is that this political disagreement can lead to violence and civil conflict. In these places, there are fewer of these types of institutional guardrails; there might be less of a tradition, or less established democratic institutions, to be able to pull everybody back from that brink.

What is causing the rise in populism?

Horacio Larreguy: On the one hand, it’s the increased formation of echo chambers. This is largely driven by the rise of the internet and social media. People sort into ideological camps; they choose to consume certain types of political information, and that creates people having very disconnected realities, which contributes to ideological polarisation. At the same time, something that’s quite important is the rise of divisive elites. The internet and social media have really lowered the cost for new individuals to come to the political arena. But also, the fact that these people do not go through the traditional parties – they can bypass them, and they can have a rhetoric that is very, very divisive, that usually more traditional elites were not kind of happy to experiment with.

Cesi Cruz: If you think about a 20-second TikTok, in 20 seconds I can steer you, I might be able to make you laugh if you like corny jokes, but I probably can’t articulate an economic development plan unless I’m some sort of really gifted communicator. And so we’re privileging this more emotional messaging rather than policy messages.

And why does it matter?

Horacio Larreguy: First, democracy suffers. At the core of democracy is civil dialogue. Anything can be countered on the basis that it comes from the other group, or from a particular media outlet. The more polarised people are, the more they dismiss things that do not fit their view of the world. We get to a world where the core of democracy – basically, people forming attitudes on the basis of information and voting accordingly – is sort of broken down by these dynamics.

Cesi Cruz: Polarisation also has a lot of insidious, smaller effects that can be even more pervasive. Even the decisions we make to invest, or to engage in economic transactions, get seen through this lens. When people perceive policies to be less stable, then it makes it more difficult for them to make these kinds of economic decisions.

Is social media to blame?

Horacio Larreguy: It’s quite clear there’s been a contribution of social media and the internet to political polarisation. For example, you have a study documenting the algorithmic amplification mechanism of social media. Essentially, if you are on social media and you tend to follow certain types of media outlets, the algorithm tries to reproduce content for you that is ideologically similar to the type of outlets that you follow.

More recent studies try to replicate this by experimentally disconnecting people. There’s a very nice study by Allcott and co-authors, among Stanford undergraduates, ahead of a US election. They asked some of the students to stop using Facebook and others not, and actually that study shows that students who disconnected from Facebook for a month showed much lower levels of polarisation at the end of the study.

Are there any interventions that can reverse this trend?

Cesi Cruz: If you’re asking people to be more critical, or to question the information they’re getting, often that means they have to question information that’s coming from people that they love or respect. Imagine you’re watching your favourite team playing sport, for me that’s soccer. And I like to think that I’m a reasonable person in real life, but when the ref makes a bad call against my team, I’m yelling, I’m screaming, I’m angry. At that point, even if you showed me a replay, I’ll outright deny what I’m seeing in the replay.

A lot of times what we end up doing in these interventions is essentially showing people the replay, when maybe what we need to do is focus on how we get people to think past that. This was a field experiment that we did, delivering workshops in 160 villages in the Philippines, about how to help communities navigate disagreement. What we were trying to get is get people to discuss how they wanted their community to engage in things like sharing political information – just to help them think about the other side, or people that disagree with them, as fellow citizens, and people that just happen to disagree, rather than people with ill intent.

If we shift the goal of the interventions to where it’s okay to have some disagreement, then we actually find that some relatively simple interventions work very well. One in particular that I like very much is a very simple one that my co-author Horacio Larreguy did.

Horacio Larreguy: An experiment we conducted over social media, where we showed information about incumbent performance. This is in the context of COVID, in some instances, before we do that, we tell people: be aware that you tend to become emotional and politicised about these issues. And just by telling people about this, they’re much more receptive.

What’s the lesson here?

Cesi Cruz: What that tells us is that we don’t have to have the grand, big change. We don’t even have to get people to all agree with each other. But as long as we can get people to take a step back, and take the heavy politics out of their thinking, that already we’re going to see big improvements in the way that people take in information, process information, and ultimately in the way that they relate to each other.

Horacio Larreguy: Other studies have actually tried to depolarise people. Look actually, people in your in-group and people in the out-group actually are quite similar; they have very similar prior policy priorities and views of the world. So whatever you think separates you from the others, actually it’s really not there. The other thing that people have tried is trying to find like a cross-cutting common thing; for example, national identity. There are studies that show this in the US and in Europe. We have also replicated this in Mexico, priming people to think about common identities.

And the last set of studies that try to counter polarisation actually use social media to do that. We have a very nice study in Turkey, where we enrolled people to consume cross-partisan media for a period of seven months, and after you see that people are much more likely to consume cross-partisan media that, at the beginning of the study, they were not consuming. They have much better views towards the other side.

And what’s the takeaway?

Cesi Cruz: We keep thinking about polarisation as a problem of individual politicians or voters, but we can’t correct our way out of it. If polarisation is fundamentally relational, then we need to build resilience at the level of communities. And it’s a way to think about how to make political differences and disagreements feel manageable, rather than threatening.

Is it all doom and gloom?

Horacio Larreguy: Fortunately, research indicates that we can counter political polarisation in three main ways. One is breaking the bubbles, showing cross-cutting pieces of information to individuals. Second, by actually priming common identities with the out-group. And third, by harnessing social media algorithms to expose individuals to cross-partisan information that usually they would not consume.

Cesi Cruz: If we just watch the news, it’s hard not to feel a little hopeless, but there’s also a lot of room for hope. The same types of social relationships, the same types of communities, that can make it easy to spread polarisation, can also be used to build resilience at a wider level. Even though polarisation travels through these relationships and communities, resilience does too.

You can find out more about this research at the link in the description. You’ve been watching Economics Unpacked. If you’ve enjoyed this episode, please like, share, follow, and subscribe.

Globally, women do less paid work than men, and childcare responsibilities are one important reason why. Could better access to childcare help more women enter the workforce, increase household incomes, and boost economic growth?

In this episode of Economics Unpacked, Lore Vandewalle and Rachel Heath explore what the evidence from Uganda, Kenya, Nicaragua, and elsewhere tells us about childcare, women’s employment, and household incomes.

Research reveals an important complication: while childcare increased work and earnings among single mothers, in two-parent households it was often fathers who worked and earned more. Looking only at women’s employment, therefore, can miss some of the wider benefits of childcare.

We also explore evidence showing that access to childcare can substantially increase women’s employment, why combining paid work with looking after children can reduce women’s earnings, and whether community-provided childcare could offer an affordable and potentially self-sustaining alternative to government-run centres.

Transcript

Lore Vandewalle: Globally, women do less paid work than men, and one likely contributor to this gap is that women are more likely to stay at home to look after their children. Providing access to childcare could enhance women’s economic agency, such as their empowerment, their autonomy, and their decision-making power. Furthermore, bringing more women into the workplace could also lead to a significant gain in economic output. So providing access to childcare could also give a significant boost to the global economy.

Welcome to Economics Unpacked, a VoxDev series where we try to answer the big questions in development.

Lore Vandewalle: My name is Lore Vandewalle. I’m a Professor of Economics at KU Leuven in Belgium. Access to childcare has been critical for increasing women’s workforce participation in high-income countries. But it is unclear whether this carries over to low- and middle-income countries, because there may be other barriers, such as restrictive norms around childcare or women’s work, or simply not enough jobs. So we randomly offered a childcare subsidy or an equivalent cash grant to mothers of a three- to five-year-old child, and then estimated the impact on mothers and on other household members.

And what did you find?

Lore Vandewalle: We find that childcare is an important barrier, but who benefits from the additional time to work depends on the household composition. For single mothers, who comprise about one-third of our sample, we observe that childcare subsidies indeed increase their labour supply and earnings. Among couples, however, it’s mainly the fathers who increase their labour hours and earnings, and this is likely driven by the large gender pay gap that favours men in our particular context.

This may seem like a counterintuitive finding. If the study had just looked at whether more women worked, it may have concluded that subsidised childcare wasn’t successful. Instead, looking at the household, what it showed is that overall the household worked more. How do you explain this?

Lore Vandewalle: If you then look at what childcare provision implies at the level of the household, we conclude it is better than giving them an equivalent amount in cash. There’s a larger effect on household income, consumption, and on child development. So this result shows that it is important to consider the impact on households and not just on mothers. From a household perspective, it may be the right choice for fathers to work more, as they can earn a higher income. Therefore, it is important to consider broader family dynamics and measure the impact on all household members, even if the barrier – in this case, childcare – is most important for mothers.

So providing childcare was shown to be successful in Uganda. What about elsewhere?

Rachel Heath: We know that childcare availability allows women to work – that’s a very common finding throughout the world, and the effects can be quite large in importance. To just take a couple of examples: in Kenya, when women got a voucher to use childcare, they were 10 percentage points more likely to work. In another setting, in Nicaragua, there was a public childcare centre that women gained access to, and they were 12 percentage points more likely to work. So, just really big increases in women’s labour supply.

So policymakers should aim to provide childcare, right?

Rachel Heath: Childcare – it’s an expensive industry, you need a lot of labour, but we found some recent promising evidence that community-provided childcare, a kind of bottom-up approach that originates in communities, had these beneficial effects at enabling women to work more. They obviously need some resources to get going, but when they do have those resources, it looks like families took advantage of them and women were able to work. The fact that they are enabling women’s labour supply and enabling women to get into better jobs means that families are willing to pay some for them, and so they can eventually be close to self-sustaining.

In many low-income countries, the type of work that women do is often self-employment, and women often do the types of employment where you could bring children along. So on the surface it might not seem like childcare is a big constraint to women working, but we do have evidence that when women combine work with childcare, they earn less than they would otherwise – which is probably not a surprise to anybody who’s ever taken care of a small child.

The research shows childcare works, but how do you provide it? What’s the takeaway?

Rachel Heath: Fundamentally, what we think families care about is having a safe, educational opportunity for their children. But what we learned in doing this literature review is that there are different ways that can be done. It can be done through the government hiring childcare providers and setting up centres. But if central governments don’t have the availability of resources or the capacity to do that, community-provided centres can be flexible in where they get funding from – through donors and other sources – and don’t need much, if any, government involvement to get started.

You’ve been watching Economics Unpacked. If you’ve enjoyed this episode, please like, share, follow, and subscribe.

• The Yale Economic Growth Center (EGC) is hiring a Communications Associate.

• Morgan Fairless writes about a smart bets panel for AI in development.

• Taiwo Hassan Odugbemi and Meron Tesfaye on what five countries tell us about firm grid-defection, power sector health, and economic prosperity.

• BREAD-IGC is offering a virtual PhD course on political economy beginning on September 8.

• Justin Sandefur writes about the electoral dividends of growth in Southern Africa.

• PEDL currently has two funding calls open: 6th Special ERG Call for Proposals for LIC-based Researchers and 8th Special ERG Call for Proposals for PhD Students. Deadline: August 16, 2026.

• Saloni Dattani writes about why every disease is a policy failure for Works in Progress.

• Joel Khalili discusses the emergence of a new generation of philanthropists made rich by AI, and what we should make of their promises.

Can microfinance really help people escape poverty?

Twenty years ago, Muhammad Yunus and the Grameen Bank were awarded the Nobel Peace Prize for their pioneering work on microfinance. The idea held enormous promise: give people excluded from traditional banking access to small loans, which they could use to invest, build businesses, and improve their lives. But two decades of economic research have complicated that story.

In this episode of Economics Unpacked, Simon Quinn and Muhammad Meki explore the origins of microfinance, how the Grameen Bank model works, and what randomised controlled trials around the world have taught economists about its impact on poverty.

The evidence suggests that traditional microfinance is no miracle cure. While some borrowers benefit, a one-size-fits-all approach does not work for everyone. We look at how new technology and better data could enable more flexible financial products tailored to the needs of poorer households and small businesses – and why the next generation of microfinance may look very different from the last.

Transcript

Twenty years ago, Muhammad Yunus and the Grameen Bank were awarded the Nobel Peace Prize for pioneering microfinance as a route out of poverty. But has it lived up to its promise? To understand what happened, we need to go back to the beginning.

Welcome to Economics Unpacked, a VoxDev series where we try to answer the big questions in development.

Simon Quinn: So, one of the problems that every household and every business faces in different ways is that they need access to money. One way that we can do it is that somebody might give it to you. And in plenty of different settings, we have friends and family who might contribute to help us to make a purchase in life. Another way is that we might be able to save. So we might be patient enough to put aside, for example, some of our income each month in order to eventually accumulate a larger sum. Another way that we can do it is that we can take a loan. So we might go to a bank. We might ask for a large sum that we will then repay over time. Now, part of the problem in dealing with banks is that in many settings banks are unwilling or unable to deal with poorer households or smaller firms. And that’s where microfinance enters the picture.

But the fact that banks struggled to have enough information on poor households isn’t a new problem. Nor is microfinance a new idea.

Simon Quinn: Many people don’t know this. As early as the 1790s, Benjamin Franklin was pioneering provision of credit to try to help business growth in the United States. Some of the themes here are quite universal. So one of the most prominent early examples of microfinance was the Grameen Bank in Bangladesh. And the Grameen Bank, to their great credit, pioneered a model that had a few key features. Essentially, it was primarily targeting women and it was joint liability. In other words, women would receive individual loans, but they would each be liable if a member of their lending-borrowing group did not repay. So you’ve got some interesting features there. You’ve got the fact that there’s group lending or group liability, the fact that there are frequent repayments, which can help us with financial discipline – we get into a routine or a habit of repaying a loan – and then we’ve got the opportunity through that to invest, or to build up a useful lump sum. And essentially it’s designed to tackle a couple of the key problems. First is how do we help poor households to accumulate what we sometimes call ‘usefully large sums’? In other words, a lump sum that might be useful for making an investment in a business, maybe paying a household bill like school fees, and so on.

Microfinance was on the rise, and for a while seemed to be the most promising idea in poverty eradication.

Muhammad Meki: Many people had very high expectations for microfinance. In fact, the term ‘miracle cure’ was sometimes used, and the evidence suggests that it’s not a miracle cure.

Economists set out to capture the impact of microfinance in communities around the world. But the results painted a sobering picture.

Simon Quinn: About twenty years ago, a number of different economists in a number of different settings launched a series of randomised control trials. These were trials that sought to measure the impact of being provided access to microfinance, and many of those trials were thinking particularly about this kind of traditional model that the Grameen Bank had pioneered. Now, what did those studies show? Well, a number of different things. First, the headline result is that microfinance does not seem to have transformative effects, at least according to that traditional model.

So what went wrong?

Simon Quinn: Part of the original idea of microfinance is that we have this one kind of contract, and it’s going to be quite transformational. It’s almost the idea that one size fits all. But part of what we learned from those initial studies is that actually some borrowers seem to benefit from this contract, but many borrowers do not. And this immediately asks a really interesting question: what if one size does not fit all in the microfinance space?

So is microfinance not useful at all?

Muhammad Meki: The fact that the average effect isn’t transformative does not mean that microfinance is useless. The fact that it isn’t a miracle cure to poverty, again, doesn’t mean that it’s useless. It means that we need to think more carefully about the design of microfinance and the exact population that we’re focusing on.

Simon Quinn: Part of what these initial studies have done is to bring a sobering dose of reality to thinking about microfinance on the whole, while also encouraging us to think in more nuanced ways about how microfinance works, for whom microfinance can work, and in which ways microfinance can work better.

So how can the next generation of microfinance overcome previous issues of data availability on poor households and small businesses?

Muhammad Meki: The modern wave of microfinance research, and the frontier, is thinking about how technology and data can improve financial access, reduce the number of excluded borrowers, excluded groups, and also improve the products that we provide to borrowers – improve the diversity of products available to them. We have lots of data for small businesses from their transactions, and so we know much more about the income and expenditures of these small businesses.

Simon Quinn: That kind of data can be really useful for structuring the next generation of flexible microfinance contracts. And, one way or another, those contracts are going to look more like the equity contracts that large firms often seek when they need to make a much larger investment. That’s to say, they will be contracts in which the repayment that is owed to the capital provider will differ based on whether the business has had a good month or a bad month.

Muhammad Meki: It’s definitely a source of optimism that the data that we now have can be used to design better financial products that are better suited to borrowers and allow them to significantly improve their lives.

After twenty years of research, what are the key lessons?

Muhammad Meki: Economics suggests that often we need a big push to improve people’s lives significantly, and the reality is that microfinance can be too micro to really help people move out of the situation that they’re in. And so data can allow us to understand more about the exact financing needs that people have, and what kind of financing amount and financial structure of product may best enable them to improve their lives significantly.

You can find out more about this research at the link in the description. You’ve been watching Economics Unpacked. If you enjoyed this episode, please like, share, follow, and subscribe.

Education technology (EdTech) has expanded rapidly across low- and middle-income countries (LMICs), yet the evidence from the past two decades shows that technology itself is not typically an education input. Read here.

International migration is one of the most powerful tools available for poverty reduction and economic development. Read here.

Industrialisation has historically been central to economic development, driving productivity growth, structural transformation, and the creation of stable wage employment for workers with limited formal education. Read here.

Developing countries face the joint challenges of reducing poverty and adapting to a changing climate, while in some cases also needing to reduce their greenhouse gas emissions. Read here.

Technology improvement is central to development. In developing countries, this often means adopting and adapting technologies from the frontier rather than inventing new ones. Read here.

Air pollution is the world’s leading environmental health risk, and its burden falls overwhelmingly on low- and middle-income countries where populations are most exposed and least able to protect themselves. Read here.

See you next week, when we return to our regular schedule.

Oliver Hanney & Emaan Siddique - The VoxDev Team

Did one of history’s deadliest natural disasters help shape the birth of Bangladesh?

In this episode of Economics Unpacked, Mushfiq Mobarak explores the role of the 1970 Bhola cyclone in shaping Bangladesh’s independence.

Drawing on historical satellite imagery, archival records, and modern economic analysis, this research examines how the cyclone affected voting behaviour, support for the Awami League, and participation in the 1971 Liberation War. It also highlights how government responses to natural disasters can influence political stability, conflict, and public trust.

The findings suggest that while the cyclone alone did not determine Bangladesh’s independence, it strengthened support for the independence movement and increased participation in the war effort – offering important lessons for governments responding to today’s increasingly frequent and severe climate shocks.

Transcript

Do climate shocks affect politics? We are going to use historical data from my country, Bangladesh, to shed light on this question by studying the 1970 Bhola cyclone, one of the deadliest natural disasters in history. In 1971, Bangladesh ultimately won its independence. The creation of the eighth most populous nation in the world, Bangladesh, was one of the most important events of the 20th century. The question we ask is: was the cyclone pivotal to the birth of Bangladesh?

Welcome to Economics Unpacked, a VoxDev series. My name is Mushfiq Mobarak, and I’m a professor of economics at Yale University. A paper published in the journal Nature logs about 145 separate cases of separatist movements during the 20th century, and only 27 of them succeeded in achieving their primary objective of leading to some kind of separatist state.

When we look at this history of separatism – whether it’s Croatia’s independence from Serbia and Yugoslavia, or Irish independence from Great Britain – what you see is that it requires two critical steps. The first is that there has to be some kind of event that allows people to unify their voices and express their desire for separatism. The second step is people being willing to risk their own lives and take up arms against an incumbent government. At this step, a lot of these separatist movements fail. Some examples of this are the Tamil Eelam movement in Sri Lanka, which led to a 27-year civil war in Sri Lanka, and ultimately Sri Lanka remained intact, and the Tamils in northern Sri Lanka never got their own state. Another example is Biafra in the 1960s in Nigeria, which led to a two-and-a-half-year civil war in Nigeria, but ultimately Nigeria remained intact.

The fundamental question we’re asking in our research is whether a climate shock can change either or both of those conditions that lead to a successful independence movement.

I was born and raised in Bangladesh, so I grew up with these stories of our independence, and this particular history was taught in school textbooks. This is an event that has been studied extensively by the other social sciences, like history and political science, using qualitative methods. It’s an important event that just hasn’t received much attention from economists using their quantitative toolkit. So I wanted to understand the history better, and understand what were the critical factors that allowed for this separatist movement to turn into an independent nation. Together with Sultan Mahmud, we take the toolkit of economics to look at this moment in history.

Tumultuous crowds fill the streets, celebrating, singing and laughing. It’s Karachi’s first day of independence. After the British left the Indian subcontinent, Muslim-majority areas were put together as a new nation state of Pakistan. It had two provinces: West Pakistan, which is now known as Pakistan, and East Pakistan, which is now known as Bangladesh. These two provinces were separated by a thousand miles of India, a non-friendly state in between. There was really not much in common between the eastern and western provinces of Pakistan except for the same religion. The culture, the ethnicity, the language, the economies were quite different from each other. The centre of power always remained in West Pakistan, in Islamabad, the capital.

In 1952, Muhammad Ali Jinnah, known as the father of Pakistan, announced that Urdu, the language of West Pakistan, and Urdu alone, would be the unified language of all of Pakistan. There was a revolt centred around Dhaka University campus, where students came out into the street to protest this decision. So basically, our war for independence started almost 20 years before we actually got independence, and it started as a language movement. After that, the other grievances that accumulated were more economic in nature. East Pakistan accounted for about two-thirds of all the population of Pakistan, but it received only 29% of government revenue.

Then, in 1970, the first elections in Pakistan in about 16 years were scheduled for December 1970, and there was a party in East Pakistan called the Awami League, led by Sheikh Mujibur Rahman, who later becomes the founder of Bangladesh, or the father of Bangladesh. It was really a movement for separatism, not full independence. Now, what changed was this cyclone that happened just a month before the elections, and that hit the coast of Bengal – which remains the cyclone that has caused the largest number of human deaths in recorded history. There was no advance warning, and so people had not evacuated that area, and it killed an estimated 350,000 to 500,000 people.

Many of the deaths were, of course, from the initial storm surge. But given the lack of state response and lack of relief, a lot of the deaths actually occurred after the fact, through disease such as cholera outbreaks, etc. The president at the time, Yahya Khan, did not visit the cyclone-affected areas immediately after. He was distracted with some diplomacy between President Nixon of the United States and Chairman Mao. So he stopped by Dhaka on the way back from a trip from Beijing, and then did a helicopter flyby over the cyclone-affected areas, where waterways were full of dead bodies, dead animals. The stench was overpowering, and against that backdrop, Yahya Khan is reported to have said, “Ah, it doesn’t look too bad,” and just flew back to Dhaka because somebody was organising a party.

So all of this meant that it became clear to Bengali, East Pakistan citizens, that this is a government that just didn’t seem to care. Those separatist sentiments that were pre-existing crystallised into an independence movement. The first step of that was to show up at the polls, and then what happens in those polls is that the Awami League actually wins the majority of all seats when you combine across East and West Pakistan, which meant that the military government was supposed to hand over power to the Awami League. But instead, under the guise of some negotiations, they fly Sheikh Mujibur Rahman, the leader of the Awami League, over to West Pakistan. Then they jail him, and then, on the night of 25 March 1971, they start something called Operation Searchlight, which was just basically starting to shoot people in East Pakistan. On 26 March, we declare independence, and the state of Bangladesh is formed.

There was a nine-month period of guerrilla warfare. Young men took up arms, and many of them were receiving the arms from neighbouring India, which was faced with a huge refugee crisis, with Bengalis pouring across the border. Faced with this overwhelming refugee outflow, the Indian military joined the war effort in early summer 1971, and once India joined, within a few days it became clear the war was over, and then the Pakistan military was going to surrender. The victory day we celebrate is 16 December 1971.

So Bangladesh appears to satisfy the two criteria for a successful separatist movement. But how do we prove that the cyclone was actually the inciting incident? To use quantitative methods and large-sample data to answer a question like this, the first thing you need is variation – variation in, say, the intensity of the cyclone as experienced by different people living in different parts of the country, and then variation in whether those different people reacted differently, given how exposed they were to this climate shock.

To understand this, we actually had to dig through and got very lucky with historical satellite data and other archival sources that we had to digitise. The first bit of good luck was that there was a satellite put into orbit by the US Department of Defense in January 1970. This was the world’s first satellite with a geosynchronous orbit, so it could take pictures constantly. This ITOS-1 satellite was put out of commission in mid-November 1970, due to a tape recorder malfunction. But we got extremely lucky, because the cyclone takes place a few days before the satellite goes out of commission. And then, recently, there have been some developments in physics and meteorology research that allow us to translate those images into a prediction about the cyclone’s intensity.

The general idea is that if there is very thick cloud cover, which is very correlated with how strong the winds were – the thickness of the cloud cover might prevent solar radiation from reaching the ground, and the images can actually measure that solar radiation as well. So, using an algorithm that translates the solar radiation and cloud cover into wind speeds, we are able to back out how intense the winds must have been in different parts of the country, or the province of East Pakistan. That meant that there were some districts, some political constituencies, that experienced the devastation in a much more intense way than others, and that is the variation that we end up using.

So we’re asking the question: did the people who experienced the cyclone more intensely vote differently a month later, in those December 1970 elections? And we also asked the question: were the people in the constituencies where they experienced the cyclone more intensely the ones more willing to take up arms and become guerrilla fighters, fighting against the Pakistan military about three months later?

We find that those constituencies that experienced the cyclone more intensely are more likely to vote for the separatist party, the Awami League. In those areas, the Awami League’s vote share increases from about 74 to 77%. We also see that these same constituencies that experienced the cyclone more intensely supplied more of the guerrilla fighters, or freedom fighters. So, when we look at the archival records of the birthplaces of the freedom fighters, we see that cyclone-affected areas are over-represented in those birthplaces. When you aggregate across the entire province, we estimate that there were an additional 46,000 fighters who joined the war effort as a result of experiencing a more intense cyclone.

Going back to the main question of whether this cyclone, or a climate shock, is pivotal to ensuring that a separatist movement successfully creates a new nation state. I think the first step, where we see the Awami League’s vote share increasing from 74 to 77%, makes it not very plausible that the cyclone itself was a pivotal event that made the difference between an independent Bangladesh versus not. The reason is that the Awami League was already quite popular, so they would have won the election anyway. They did win by a bigger margin due to the cyclone. However, the fact that the cyclone induced more people to join the war effort, and held off the Pakistan army, could have made a critical difference.

Can this research conclusively prove that the cyclone was the cause of Bangladesh’s independence? No, it cannot. What we show is that it did shift the balance, first politically, and then in terms of conflict participation, in the direction of people expressing their anger, both in the polls and on the battlefield. Some of those factors may have been critical. So what our research shows is that climate shocks not only cause these environmental and economic harms, but they may also have broader economic and conflict impacts as well.

Our research allows us to connect the dots and paint that picture a little more clearly. And what we learned, by painting the picture more clearly and looking at this precise mechanism, is that it’s not just the shock that matters – it’s also the government’s responsiveness to the shock, whether they choose to respond with relief for people who were affected, versus whether they ignore those shocks.

So, looking to the future, we are in a world where climate shocks are becoming more severe and more frequent. We are seeing more extreme weather events – more hurricanes, more typhoons, and cyclones – all over the world. What we need to understand is how governments should respond to these types of weather events in order to address citizen concerns. Maybe that will help give us guidance on what types of responsiveness to climate shocks we need, such that we can prevent more of these large-scale conflicts that cause a lot of unnecessary death and destruction.

If you like this video, please like and subscribe. And for more information about this research with Sultan Mahmud, check out the links below.

• Ethan Mollick writes an opinionated guide on which AI to use to do stuff.

• Ben Hyman discusses how to rigorously measure the impact of funding a for-profit company.

• Joe Hasell, Bertha Rohenkohl, Pablo Arriagada, Esteban Ortiz-Ospina, and Max Roser write about economic inequality for Our World in Data.

• The Center for Global Development is hiring a Research Analyst in London. Apply by August 6th.

• Dan Corry discusses economics and the art of influence on the Econ to Icon podcast.

• Lee Crawfurd, Theo Mitchell, and James Hu write on how the lead industry is downplaying the harms.

• David McKenzie asks whether it matters how entrepreneurs are selected and how much funding they are given.

• Zainab Usman discusses why it is becoming more difficult for developing countries to achieve economic prosperity through trade.

• Wilson King writes about the mystery of short Indian children for In Development.

• James Robinson in conversation with Abdoulaye Ndiaye at the Africa’s Values Debate in Senegal.

• Oliver Kim writes about Robert McNamara for In Development.

What is the informal economy, and why does it matter?

In many developing countries, between 30% and 90% of workers are employed in the informal sector – running legal businesses or working jobs that exist outside government regulation and taxation.

In this episode of Economics Unpacked, Gabriel Ulyssea explains why informality is so widespread, how it affects economic growth, inequality and public services, and why simply cracking down on informal businesses often makes matters worse. Drawing on his research, he explores why the informal economy is often a symptom of deeper institutional problems and what governments can do to encourage more workers and firms to join the formal sector.

Whether you’re studying economics, interested in development policy, or simply curious about how economies work, this episode explains one of the biggest – and least visible – parts of the global economy.

Transcript

What if a huge part of the economy was hiding in plain sight, invisible to the government? It’s called the informal economy.

Gabriel Ulyssea: The informal sector is this massive part of developing countries’ economies that take place at the margin of all laws, regulations, and legal frameworks. In developing countries, this accounts for a vast share of workers. In Latin America, that ranges from 30–80% of the labour force. In sub-Saharan Africa, that can go up to almost 90–95% of the labour force.

So the informal economy is anywhere from 30–90% of the economy. How can this be possible? What’s the consequence of such a large informal economy? And what can be done about it?

Welcome to Economics Unpacked.

Gabriel Ulyssea: So the informal sector is every part of the legal economy, meaning it’s not criminal activities. So these are legal activities such as mom-and-pop shops, small textile manufacturers, all legal economic activities that are invisible to the government because they are not registered to pay taxes, they are not complying with the labour regulations. When you have so many firms and so many workers in the informal sector, they effectively are invisible to the government, and the government therefore cannot, for example, tax them. Now you might think, well, this is actually a good thing, and it might be potentially a good thing individually for a given worker or given firm, but on the aggregate, collectively, that’s very bad news for the economy, because that means that the government doesn’t have the ability to raise the resources that it needs to finance, for example, health expenditures or providing high-quality education, and these are, for example, two things that are crucial for the welfare of the population but also for long-term growth.

The informal sector also creates the wrong incentives for economic growth.

Gabriel Ulyssea: It provides bad incentives for both workers and firms in terms of, say, firm growth. In order to remain informal, firms cannot grow too much, because otherwise they become too visible to the government, that saves on taxes, but it creates this problem that, dynamically, you’re not growing, you’re not investing, you’re not becoming more productive. And when you have too many firms in the economy in this sort of distorted regime, that means that the economy as a whole also will not experience productivity gains on the whole, and will grow less, and will have lower levels of living standards. So that’s a major concern in developing countries.

The informal sector also includes housing.

Gabriel Ulyssea: Today there are over 1 billion people living in slums or informal settlements around the world. This is obviously concentrated in developing countries, but it accounts for more than a quarter of all people living in urban areas. In some countries, like in sub-Saharan Africa, the share of people living in slums can be as high as 60%. The property rights are not well defined – people don’t own their houses. They rent their houses in a very informal setting. So that means that typically these are very low-quality, poor settlements that often entrap people in a cycle of poverty.

How does the invisible economy create a challenge for economic analysis and policy?

Gabriel Ulyssea: Today there is no economic issue or policy discussion in developing countries that one can approach completely ignoring the informal sector. Even if one does not have the data to analyse what’s going on in the informal sector with informal workers and informal firms, one still has to consider and acknowledge the presence of this massive sector. Otherwise, policies can have massive unintended consequences and lead to adverse effects. Your conclusions can go almost in the opposite direction if you were to include the informal sector. I started working on this because, to me, it was striking – being from Brazil, and the informal sector being such a massive presence in that economy – that we had so little work done in mainstream economics, in development economics, studying this. And then the data that existed around the world suggested that that was the same in all low- and middle-income countries.

Here’s an example: when looking at minimum wages.

Gabriel Ulyssea: Let’s say you’re trying to understand the effects of the minimum wage in a developing country’s labour market, and you’re focusing only on the formal sector. You might see that when you introduce the minimum wage, actually wage inequality in the formal sector declines, because you’re making the low-wage earners better off. But if you don’t have data on the informal sector, and if your model does not include the informal sector, you might miss completely the fact that many people will transit from formal to informal jobs. And in fact, you might make overall wage inequality in this country worse, because you’re increasing the distance between formal sector workers and informal sector workers.

What’s causing the large informal sector in developing countries?

Gabriel Ulyssea: One key big-picture message that comes out of the research on informality. Well, the first thing to consider is it’s a symptom of badly designed regulations and institutions. So just promoting more enforcement, which means increasing inspections, making life in the informal sector a lot harder, can actually make things worse. The informal sector is also causing a lot of problems because it creates a huge reallocation of resources that should be in more productive formal firms away from these firms and into less productive informal firms. So that’s also hurting the economy. So the ideal combination would be to reduce the costs of being formal; for example, by reducing all the bureaucracy that individuals have to go through in order to open a formal business, or to reduce the tax burden once you’re formal. But together with that, you also have to increase enforcement. You have to increase inspections. You have to make life in the informal sector more difficult.

Why are governments not doing that? They actually have the technology, they have the ability to do that. One key thing most likely is the political costs of doing that, because associated to higher enforcement, there are always high social costs in the short run. Many of these individuals, these informal entrepreneurs, would not be able to find jobs in the formal sector immediately. They might experience long spells of unemployment. So all of that can generate high social costs and high political costs, and that’s one of the main challenges today in tackling these high informality levels in developing countries.

So how to change this?

Gabriel Ulyssea: So one key dimension that is very much overlooked in the economics research about the informal sector is the fact that informality, associated to all these countries, are now in a type of equilibrium where the social costs of being informal are very low or zero. What I mean by that is that no one is ashamed of not complying with certain laws and regulations, because they are seen as useless or they are seen as inconsequential, and that contributes to reinforcing this perpetuation of high levels of informality. Whereas you have actually other countries, typically richer, more developed, where the social costs of non-compliance are quite high. You wouldn’t be seen with good eyes by your neighbours or friends if you evade your taxes, you don’t comply with your council tax, etc. So in that sense there are multiple possible equilibria in these economies, and even with the existing resources, if everybody agreed that not paying taxes is a bad thing, this in itself would create a reinforcement such that everybody would be complying and paying taxes, even though with the same environment around them. So this dimension is key, but it has not been studied in economics. One reason for that is it’s extremely hard to find data to support this analysis.

So what’s the takeaway?

Gabriel Ulyssea: I think the first basic takeaway is that informality is a symptom of other underlying issues – it’s an outcome, as is high unemployment, and it should be seen as such. So then what we need to do is basically tackle this head-on: increasing the costs of informality and reducing the cost of being formal, and increasing the benefits of formality, even if it may lead to short-run costs, because in the longer run the whole society will be better off.

You can find out more about this research in our VoxDev Lit link in the description. You’ve been watching Economics Unpacked. If you enjoyed this episode, please like, share, subscribe.

• Esteban Quiñones, Marie Gaarder, and Thomas Kelly write about evidence, method, and what the RCT debate is missing.

China’s rise from a largely closed economy to one of the world’s biggest trading nations is one of the most remarkable economic stories in history.

In this episode of Economics Unpacked, Hanwei Huang explores more than 150 years of China’s economic development – from the Opium Wars and the opening of China’s ports, through central planning under Mao, the market reforms of Deng Xiaoping, WTO accession, and the country’s emergence as a global manufacturing and technology leader.

We’ll also examine the role of trade, industrial policy, foreign investment, and economic reform in shaping modern China, before asking an important question: can other countries replicate China’s success?

Transcript

Over 150 years ago, China was a mostly self-sufficient economy that remained closed to the rest of the world. One and a half centuries later, China is now the world’s largest economy in international trade. How did that happen?

Welcome to Economics Unpacked, a VoxDev series where we try to answer the big questions in economic growth. In the next few minutes, I’m going to try to explain the economic history of over 150 years. That sounds crazy, but let me give it a try.

China’s growth story started more than 100 years ago when China lost the Opium War. The cause of the war was due to China’s rejection of the UK and France’s demand to trade opium in China. Before then, China was basically a closed economy. As a result of the Opium Wars, China was forced to open more than five trading ports, including Shanghai. At the same time, territory including Hong Kong was annexed by the UK. The emperor realised China must engage with the West and started introducing Western technologies and ideas to China. At the same time, Westerners started to come to China, introducing modern technologies and setting up enterprises in China.

The next stop is the interwar period. The Western countries were busy fighting wars. It provided a brief opportunity for China to make use of modern technologies, building up its own production and then exporting to the rest of the world.

The next key date is 1949, the establishment of the People’s Republic of China, or the PRC. The PRC adopted the Soviet economic system, which was dominated by state-owned enterprises in the urban area and collective farming in the rural area. The planned economy helped China to build its heavy industry base, which was facilitated by 156 projects transferring Soviet know-how and technologies to China. However, the planned economy also featured low efficiency and high distortions, which brought disasters.

For example, the Great Leap Forward movement in the 1950s. The reason why this system didn’t work was that the collective farming system didn’t provide people enough incentive to work on the farm. What you get out of this system depends very little on what you do. People weren’t rewarded for how much they worked, and there was a severe free-riding problem: the system rewarded collective outputs, so people could free-ride on other people’s efforts.

1978 is a watershed moment in China’s growth history. It marked the start of the era of reform and opening.

Two important things happened. First, China transitioned from the planned economy to the market economy. State-owned enterprises gave way to private enterprises and firms from foreign countries in the rural area. A household responsibility system was introduced, which basically replaced collective farming. The other important thing that happened is the opening of China to the rest of the world. China started to engage more in foreign trade and attract foreign investment, and this was helped by important reforms such as the establishment of the special economic zones, which allowed foreign firms to do processing trade in China.

Trade is important for three key reasons. First, it allows countries to specialise in sectors in which they have a comparative advantage. Second, it allows countries to exploit economies of scale. Firms can sell not just to buyers from their own countries, but also to buyers from the global market, which allows them to lower their costs. And third, trade spurs innovation and technological advancement. By trading with other countries, you get to learn from them, and it gives firms more incentive to do R&D. During the 80s and 90s, many of the previously state-owned enterprises were privatised during this reform period as the country transitioned to the market economy.

After decades of negotiation, China finally got into the WTO in 2001. This helped China to rise for three reasons. First, the WTO enabled China to get access to more than 100 member countries of the WTO with reduced trade costs. Second, WTO membership implied more certainty in the trade policies faced by Chinese exports, which gave firms more confidence to invest in China. And third, as a commitment to becoming a WTO member, China had to deepen its reforms. For example, in certain industries it was previously not allowed for foreign investment, but once China became a WTO member, it had to level the playing field for foreign and domestic firms. One example is the financial sector.

Since the WTO accession, China has basically become the world champion in trade and manufacturing. At the very beginning, China specialised in low-value-added sectors such as textiles, and now China has gradually moved up the value chain, making sophisticated products such as cell phones, electric vehicles and chips.

Another important moment came in 2015, when China came up with the Made in China 2025 growth initiative, targeting certain advanced industries such as the semiconductor industry and biotechnology.

A good example would be electric vehicles. Since the early 2000s, China has started to invest heavily in the electric vehicle sector, providing subsidies for production, consumption and R&D. It’s very difficult for China to break into the traditional auto sectors, but in this new and growing electric vehicle sector, China has been able to gain more brands and more patents, and to outcompete Japanese and German competitors.

So here are the big three lessons for today. First, trade openness brings prosperity. Second, industrial policy can work if the government does not simply pick winners but provides the right incentives to firms. And third, the government can play a role when the market is not functioning properly.

Can other countries copy China’s success? I think the answer is both yes and no. No, because China is a large country with a unique culture, history and its own unique political system. But at the same time, yes, because there are two lessons that we can take from China. First, China has been quite successful in adopting an experimental approach to policymaking — as Deng Xiaoping famously put it, “crossing the river by feeling for the stones.” By experimenting with different policies in certain areas and sectors, this allows policymakers, firms and workers to adapt to the policy and find the policy that can work, and this approach has been adopted by countries like Vietnam. Second, the lesson that we can learn from China is the long-term vision in policymaking. China has been investing in the green sector, for example, for more than two decades, and this has given firms certainty in their investment, which I think is one key reason China is now leading in this sector.

There we have it: 150 years in less than 10 minutes. You have been watching Economics Unpacked. If you enjoyed this episode, please like, share, follow and subscribe.

• Africa Urban Lab was featured in The Economist.

• Han Sheng Chia discusses his underlying priors on AI.

• Oliver Hanney was interviewed by the FT on whether AI will trigger a new wave of philanthropy.

• Oliver Hanney provides an updated list of resources on AI and development economics.

• IGC has released an AI Playbook for government.

• Coefficient Giving is requesting proposals for projects related to “global health and wellbeing in an era of transformative AI“.

• Pranab Bardhan and Raghuram Rajan discuss the international economic order.

• The Economics of Environment and Energy Programme (EEE), the Global School of Sustainability (GSoS) and the International Growth Centre (IGC) within LSE are convening the fifth Environment Week at the LSE on 21-24 September. Submit your paper here.

• On Monday, we released our VoxDevLit on Technology and Development with Senior Editors Julieta Caunedo and Tommaso Porzio, and Co-Editors David Argente, Jaedo Choi, Yulu Tang, Danial Lashkari, Jacob Moscona, Karthik Sastry, Deivy Houeix, Federico Rossi, Luisa Cefala, Erin Kelley, and Conor Walsh.

• On Wednesday, we published our VoxDevLit on Air Pollution with Senior Editors Teevrat Garg and Anant Sudarshan.

Dean Karlan, former Chief Economist of USAID, joined us this week on 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.

What does the standard seat count really tell us about women’s political power? In this week’s episode of VoxDevTalks, Soledad Artiz Prillaman discusses her review of the research into non-elite women’s political participation. The conversation explores who counts as ‘elite’, why that distinction matters for policy, and what the evidence actually shows about the barriers women face when trying to engage with politics.

As part of the National Reimbursement Drug List reform, the Chinese government negotiated steep price cuts with pharmaceutical firms for innovative, high-priced therapies in exchange for guaranteed coverage and near-universal patient access. Panle Jia Barwick, Hongyuan Xia, and Tianli Xia find that this dramatically expanded the country’s pharmaceutical innovation, offering a rare policy model that reconciles drug affordability with R&D incentives.

A land formalisation programme in Benin significantly increased the likelihood that widows – especially those without a son – remained in their villages, offering formal institutional protection where customary norms left women most vulnerable. Ioana Botea, Markus Goldstein, Kenneth Houngbedji, Florence Kondylis, Michael O’Sullivan, and Harris Selod discuss.

Leila Pereita, Rafael Pucci, and Rodrigo R. Soares examine how an air interdiction policy in early 2000s Brazil shifted cocaine trafficking from air routes to Amazonian rivers, increasing violence in municipalities along these waterways. They find that the policy redirected trafficking to river networks by raising the risks of aerial transport, bringing criminal organisations into closer contact with local communities.

Instant payments can substitute for cash when adoption moves quickly beyond high-income early users. Evidence from Brazil, Costa Rica, and Mexico finds that the key is a rapid low-income gradient: systems must combine low adoption costs, dense networks, supply-side coordination, awareness, and trust. Without them, even sound platforms can remain marginal. David Argente, Paula González-Alvarez, Esteban Méndez, and Diana Van Patten explain.

Elsewhere in development:

• Ivan Manhique, Felix Mambo, and Jennifer Mascarenhas discuss how Mozambique became one of the world’s poorest countries.

• The South African Economics Pipeline invites students from South African universities interested in economics to apply for their bursary and training programme.

Read and download here

Check out the first issue of our VoxDevLit on Air Pollution by Senior Editors Teevrat Garg and Anant Sudarshan.

Later today, at 4pm (UK time), the Senior Editors will outline the key policy takeaways from this review in our one hour launch event.

Register for today's webinar

This VoxDevLit, which will be updated as new evidence is released, summarises key policy priorities:

• Damages are large and are almost certainly understated

• People respond to information, but private protection is unequal and often insufficient.

• Low demand for clean air among low-income households is not primarily explained by standard market failures.

• Institutional investments in indoor air quality offer unusually high returns, but sustained operation is the binding constraint.

• Emissions markets can work in developing countries.

You can now click between chapters interactively on our website! Make sure to check out this new home for our library of VoxDevLits.

Library of 24 VoxDevLits

• Susannah Hares, Jack Rossiter, and David Evans on rethinking how we report learning gains.

• Han Sheng Chia and Tim Ohlenburg have written a policy brief on generative AI evaluation.

• Deena Mousa on why AI won’t be the next leapfrog miracle.

• Ben Hyman makes a case for international labour mobility.

• Pranab Bardhan writes about the neglected aspects of inequality in India.

• In Development is calling for pitches for narrative-driven essays on development. Submit these by July 1, 2026.

• CGD has a request for proposals for Up to $5 Million for Research on Global Lead Poisoning.

• David McKenzie has written about the dangers of taking theory too literally, of log(x/a+1) transforms, and place versus cows in determining poverty dynamics.

• Jishnu Das has started releasing a new series of blogs titled Economists and Mental Health.

• And Ken Opalo has written a caution against faddist approaches to industrial policy (the old industrial policy rules still apply).

And last but not least! Aaditya Mattoo, Anna Tompsett, and Eddy Zou are working on a project about complementarity that asks whether and when development interventions work better together than separately. Their project combines a review of existing evidence with an expert survey and observational data about policymaker choices. They’d be super grateful if you’d take 10 minutes to share your views via this survey.