Embedded Files
Helpful prior knowledge and learning objectives
Helpful prior knowledge and learning objectives
Helpful prior learning:
Section 1.1.1 The economy and you, which explains what an economy is and how it is relevant to students’ lives
Section 1.1.2 The embedded economy, which explains the relationship between the economy and society and Earth’s systems
Section 5.4.2 Poverty, which outlines the high-leverage recommendations from the Earth4All system model for reducing poverty
Section 7.1.1 Global exchange as a system, which describes global exchange as a system with parts, relationships, functions and emergence
Section 7.2.5 Who writes the rules of global exchange? which explain who shapes the rules of global exchange and how various sources of power influence which rules are written and whose interests they reflect
Section 7.3.1 Unequal global value exchange, which explains how unequal prices and unequal wages systematically transfer value from periphery countries to core countries
Section S.1 What are systems?, which explains what a system is, the importance of systems boundaries, the difference between open and closed systems, and the importance of systems thinking
Section S.2 Systems thinking patterns, which outlines the core components of systems thinking: distinctions (thing/other), systems (part/whole), relationships (action/reaction), and perspectives (point/view)
Learning objectives:
explain the role of randomised controlled trials in development economics.
evaluate the evidence on micro-interventions such as cash transfers, distinguishing between their measurable local effects and their ability to address the structural conditions that produce poverty
In the early 2000s, researchers working near the city of Udaipur in Rajasthan, India, observed a stark gap between policy and reality. The Indian state offered free vaccinations for young children, yet in some areas fewer than two per cent of children were fully immunised. State clinics existed, but nurses were often absent and buildings were locked more than half the time. Families had little reason to make the journey on any given day.
The researchers designed an experiment across 134 villages to test actions that could change this. Some villages continued with the existing state service. A second group received monthly camps with a nurse who was paid only when she showed up. A third group received the same camps, and also a bag of lentils for each visit and a set of metal plates when a child completed the full course. Full immunisation rates were 6% in villages with no change, 18% in villages with reliable camps alone, and 39% in villages with camps and incentives. Children who were vaccinated stayed healthier, attended school more regularly, and were less likely to fall into the cycle of illness and lost income that keeps households trapped in poverty. A parent freed from nursing a sick child also gained time for paid work and care. A small, local, carefully tested change had effects that rippled outward into families and communities.
Programmes like this one are often paid for by money from outside the community: states, international institutions, or private donors. This section looks at what these kinds of small-scale interventions, which can also come through global exchanges, can and cannot achieve.
Figure 1. A family receives lentils at a clinic where vaccines are given.
(Credit: J-PAL/IPA)
Why did economics turn to experiments?
Why did economics turn to experiments?
The lentils study was one example of a larger shift in how some economists approached poverty and development. To understand why this shift happened, it helps to look at what came before.
For most of the twentieth century, many economists who studied development believed that countries in the Global South were at an earlier stage of a universal path toward industrial growth (Section 7.3.10). Given the right conditions and policies, their incomes would eventually catch up with countries in the Global North. This view shaped the advice that international institutions gave to states across the Global South. It included the Washington Consensus policies of the 1980s, which pushed low-income countries to cut public spending (austerity), open their markets to global trade (trade liberalisation), and reduce the role of the state (deregulation and privatisation). In much of Latin America, sub-Saharan Africa, and parts of Asia, these policies worsened poverty instead of improving it.
This failure left some development economists looking for a different approach. Instead of relying on broad theories about how economies should be organised, they began testing specific programmes directly with the people they were meant to help.
Randomised control trials (RCTs)
Randomised control trials (RCTs)
One method for testing specific programmes is the randomised controlled trial (RCT). The method was borrowed from medical research, where it had been used for decades to test whether a treatment worked. Researchers take a large group of people and divide them randomly into two or more groups. One group receives the intervention being tested, such as a vaccine incentive, a cash transfer, or a new teaching method. Another group, the control group, continues with whatever was happening before. After a set period, the researchers measure the difference in outcomes between the groups. Because the division was random, any difference in outcomes may be caused by the intervention. This is what the lentils study at the start of this section showed.
With this strategy, researchers could test specific programmes in specific places and measure whether they worked, rather than debating how whole economies should be organised. Donors and states funding poverty reduction could see clearer evidence about which programmes actually worked. Esther Duflo, Abhijit Banerjee, and Michael Kremer built an institution around this approach called the Abdul Latif Jameel Poverty Action Lab (J-PAL) at MIT, founded in 2003. In 2019, the three received the Sveriges Riksbank Prize in Economic Sciences in Memory of Alfred Nobel (Nobel Prize in Economics). By then, J-PAL-affiliated researchers had conducted over a thousand randomised trials across health, education, agriculture, and income support.
Figure 2. Esther Duflo and Abhijit Banarjee.
(Credit: Bryce Vickmark, for Nobel Prize)
Limitations of RCTs
Limitations of RCTs
Angus Deaton, also a Nobel Prize winner, and philosopher of science Nancy Cartwright, point out that an RCT can only answer a narrow question. Did this intervention change this outcome, for this group of people, in this place?
An RCT can show that an intervention caused a result in one place and time. It cannot show whether the same intervention would work somewhere else. That requires understanding why it worked, and whether the same conditions exist in the new setting. J-PAL researchers know this. They often combine RCTs with interviews, observation, and follow-up studies to understand why an intervention worked.
Deaton and Cartwright point to the importance of other kinds of research that focus on different questions. Historical analysis can show how poverty developed over time. Structural analysis can show why it persists today. Qualitative research can capture the experiences of people living in poverty in ways that numbers cannot. RCTs do not produce this kind of understanding by themselves.
RCTs also have practical limits. They are expensive and slow to run. They raise ethical questions too. Is it fair to give a (potentially helpful) programme to some people and not others, for the sake of research? Would the money be better spent running the programme itself, for everyone, rather than testing it on half a group?
RCTs also measure change only within wider economic systems. They ask whether one intervention helps, given how things already are. They do not ask whether those existing systems are the problem.
Do direct cash transfers work?
Do direct cash transfers work?
One of the most widely studied micro-interventions is the direct cash transfer. Some cash transfers are conditional. Households only receive the money if they meet certain requirements, such as sending children to school or attending health check-ups. Others are unconditional. Households receive the money with no conditions attached and decide for themselves how to use it. Organisations such as GiveDirectly, founded in 2011, have delivered unconditional cash transfers to hundreds of thousands of households across East Africa and elsewhere. They have evaluated the results using RCTs.
The evidence on unconditional cash transfers shows clear benefits at the local level. Studies in Kenya found that transfers increased the amount families spent on food, improved psychological wellbeing, and allowed households to buy tools, livestock, and other items that help earn income. A recent trial found that a single transfer of $1,000 reduced infant mortality by 48% among families who received it. Studies in multiple countries have also found benefits for surrounding households that did not receive the cash transfers. Money spent locally circulates through communities through spending. For every dollar transferred, around $2.50 of economic activity was generated in surrounding communities.
Figure 3. Three stories from GiveDirectly about how unconditional direct cash transfers help people.
(Credit: Give Directly)
One reason unconditional cash transfers work is trust. These interventions trust people living in poverty to understand their own needs. Many traditional aid programmes are designed by outside organisations that decide in advance what a community requires. An unconditional cash transfer lets each household direct resources toward what matters most in their own situation.
Unconditional cash transfers improve lives at the household level. But on their own, cash transfers and other kinds of micro-interventions are not likely to change the structural conditions that produce poverty, including the extractive global exchanges discussed in Subtopic 7.3. The money may also be unreliable. Cash transfers paid for by donors often depend on money from outside a country, and donors can withdraw it (Section 7.3.8).
When states are able to provide it, universal social protection is a longer-lasting alternative. This includes child benefits, pensions, and income support for people with disabilities, new mothers, and unemployed people. It also includes access to essential health care. A state pays for it, and it is a right for all citizens. But about 3.8 billion people, nearly half of the world's population, receive no social protection cash benefit of any kind. The 133 low- and middle-income countries would need to spend an extra 1.4 trillion US dollars per year to give everyone this basic level of protection.
In 2021, members of the International Labour Organization (ILO) proposed a Global Fund for Social Protection to help pay for social protection. The Global Fund for Social Protection was included in the 2026 Roadmap for Eradicating Poverty, published by the UN Special Rapporteur on extreme poverty. Such a fund would give low-income countries money they can count on for several years. It would draw money from aid, from debt swaps, and from special taxes on financial transactions and fossil fuels. Unlike project-based aid, it would help states build their own long-term systems. The fund is still a proposal.
Activity 7.4.7
Activity 7.4.7
Concept: Systems
Skills: Thinking skills (transfer, critical thinking), Communication
Time: Varies, depending on the option
Type: Individual, pairs, or small group
Option 1: Can an RCT answer this question?
Time: 15-20 minutes
Below are four research questions. For each one, decide whether an RCT could answer it, and explain your reasoning in one or two sentences.
Did providing free school meals increase attendance at primary schools in this district over one year?
Why have smallholder farmers in this region remained poor for three generations despite various aid programmes?
Does giving households a cash transfer reduce the number of days children miss school?
How do trade rules set by wealthy countries affect the ability of lower-income countries to develop their own industries?
After deciding and explaining for each question, discuss with a partner: what do the questions that RCTs cannot answer have in common? What kinds of research might be better suited to them?
Click on the arrow for sample responses, but give it a go yourself first!
1. An RCT could answer this question. It has a specific, measurable outcome (school attendance), a defined group (primary school students in one district), a clear intervention (free school meals), and a set time period (one year). Researchers could divide schools randomly into those that receive the programme and those that do not, and measure the difference in attendance rates.
2. An RCT could not answer this question. It asks why something has happened over a long period of time, involving historical, structural, and social factors that cannot be tested by dividing a group randomly and measuring a short-term outcome. Historical analysis, ethnographic research, and structural economic analysis would be better suited to this question.
3. An RCT could answer this question. Like question 1, it has a specific intervention, a measurable outcome, and a defined group. Researchers could randomly assign cash transfers to some households and not others, then compare school attendance data across the two groups.
4. An RCT could not answer this question. It involves large-scale economic and political systems operating across countries over long periods of time. It is not possible to randomly assign different trade rules to different countries and measure the results in a controlled way. Structural economic analysis, historical research, and comparative studies across countries would be more appropriate methods.
Option 2: Exploring a real RCT
Time: 60+ minutes if you have multiple groups presenting; can be shortened if there are no presentations. You could also run this in a jigsaw format.
In this activity you will explore a real randomised controlled trial from the J-PAL research database and summarise it for another student or group.
Go to the J-PAL website at povertyactionlab.org and browse the evaluation summaries. Choose one study that interests you. It can be on any topic — health, education, agriculture, income support, or another area.
Read the summary and prepare a short presentation covering the following four points:
Where and when was the study carried out, and who were the participants?
What was tested? Describe the intervention and how the groups were divided.
What did the researchers find? Summarise the main results.
What questions does the study leave unanswered? Think about what the RCT method cannot tell you about this situation: why the problem exists, whether the results would apply elsewhere, or what wider conditions might affect the outcome.
Your presentation should take around three minutes. You do not need slides. Be ready to answer one or two questions from your audience.
After all presentations, discuss as a class: looking across the different studies, what kinds of problems do RCTs tend to study? What kinds of problems are missing?
Option 3: A systemic impact for cash transfers?
Time: 15 minutes
Cash transfers help the households that receive them. This section also mentions a multiplier effect, where money spent locally circulates in communities through spending and reaches people beyond the original household.
Consider the following questions, and using one of the two system thinking options below them:
Could this local circulation ever add up to something bigger, a real shift in the structural forces described in Subtopic 7.3, such as unequal exchange between the Global North and Global South?
Or do cash transfers stay limited to local, short-term gains, no matter how many households receive them?
Does it matter whether they lead to large-scale systemic shifts or not?
If you are familiar with causal loop diagrams (Section S.5): Draw a causal loop diagram that traces what happens after a cash transfer enters a community. Include the local spending multiplier, and try to show where the loop might reinforce itself, where it might reach a limit, and where it connects (or fails to connect) to the larger structural forces discussed in Subtopic 7.3.
If you prefer to work without a diagram: Talk through the question with a partner. Consider: What would have to be true for cash transfers to add up to structural change? What stays the same no matter how much money circulates locally?
Write down your reasoning in whatever form makes it clearest, a diagram, a short paragraph, or a list of points. Be ready to explain your thinking to another pair or the class.
Ideas for longer activities and projects are listed in Subtopic 7.5
Checking for understanding
Checking for understanding
Coming soon!
Further exploration
Further exploration
What's the best way to lift people out of poverty? — A TED-Ed animated video examining the evidence on direct cash transfers as a poverty intervention. It covers what the research shows, why results are more complicated over longer time periods, and why cash transfers work. Difficulty level: easy.
The problems with Randomised Controlled Trials — A short video interview with Nobel Prize-winning economist Angus Deaton, covering the main limitations of randomised control trials. Difficulty level: medium.
Designing and Running Randomized Evaluations — A free online course from MIT and J-PAL covering how RCTs are designed and conducted in development economics. Useful for students who want to understand in more detail how the research behind interventions like those discussed in this section actually works. Difficulty level: hard.
From poverty to progress — 1 year of cash impact: A ca. 3 minute video from Give Directly about how unconditional direct cash transfers are transforming lives in Malawi. Difficulty level: easy
J-PAL - The website of the Abdul Latif Jameel Poverty Action Lab at MIT, the research centre founded by Duflo, Banerjee, and Kremer discussed in this section. The site summarises hundreds of RCT studies across health, education, agriculture, and income support, and explains the findings in accessible language. Difficulty level: easy to medium depending on the study.
Sources
Sources
Abdul Latif Jameel Poverty Action Lab. (2025). Giving directly to support poor households [Case study]. https://www.povertyactionlab.org/case-study/giving-directly-support-poor-households
Banerjee, A., Duflo, E., Glennerster, R., & Kothari, D. (2010). Improving immunisation coverage in rural India: Clustered randomised controlled evaluation of immunisation campaigns with and without incentives. BMJ, 340, c2220. https://doi.org/10.1136/bmj.c2220
Banerjee, A., & Duflo, E. (2011). Poor economics: A radical rethinking of the way to fight global poverty. PublicAffairs.
Cattaneo, U., Schwarzer, H., Razavi, S., & Visentin, A. (2024). Financing gap for universal social protection: Global, regional and national estimates and strategies for creating fiscal space (ILO Working Paper 113). International Labour Organization. https://www.ilo.org/publications/financing-gap-universal-social-protection-global-regional-and-national
Center for Effective Global Action. (2025, August 18). Cash transfers dramatically cut infant mortality in Kenya, new research finds. https://cega.berkeley.edu/article/cash-transfers-cut-infant-mortality/
Chang, H.-J. (2014). Economics: The user's guide. Pelican Books.
Deaton, A., & Cartwright, N. (2018). Understanding and misunderstanding randomized controlled trials. Social Science & Medicine, 210, 2–21. https://doi.org/10.1016/j.socscimed.2017.12.005
De Schutter, O. (2026). The roadmap for eradicating poverty beyond growth. Mandate of the Special Rapporteur on extreme poverty and human rights, United Nations Human Rights Council. https://www.neep-poverty.org/wp-content/uploads/2026/06/The-Roadmap-for-Eradicating-Poverty-Beyond-Growth.pdf
Egger, D., Haushofer, J., Miguel, E., Niehaus, P., & Walker, M. (2022). General equilibrium effects of cash transfers: Experimental evidence from Kenya. Econometrica, 90(6), 2603–2643. https://doi.org/10.3982/ECTA17945
International Labour Organization. (2024). World social protection report 2024–26: Universal social protection for climate action and a just transition. https://www.ilo.org/sites/default/files/2024-09/WSPR_2024_EN_WEB_1.pdf
Terminology
Terminology
Coming soon!
Page updated
Google Sites
Report abuse