The manifesto, written in the discipline's own language. Banerjee and Duflo lay out why randomisation gives a credible answer to "does this work?" and address the standard objections head-on. Read it first so you understand the strongest version of the case the critics are responding to.
Deep Dive · Methods & MEL
Randomista Economics and Its Critics
The rise of the randomised trial, the 2019 Nobel, and the long argument about what experiments can and cannot tell us about development.
RCTs
Development Economics
14 readings
IM
ImpactMojo Editorial
Curated by the ImpactMojo team
This is the reading list our methods faculty use to teach the randomista debate — not as a verdict, but as an argument worth understanding from both sides. It pairs the foundational case for field experiments with the most serious critiques, so that an evaluator can hold the tension rather than pick a tribe. We welcome an invited curator (a development economist or applied methodologist) to extend it.
House Pick
Editor's Note
Over two decades, a small group of economists changed how development economics is done. Instead of debating grand theories of growth, the "randomistas" ran randomised controlled trials — borrowed from medicine — on specific, answerable questions: does charging for bednets reduce their use? Does deworming keep children in school? The movement's central claim was epistemic humility dressed as rigour: stop pretending we know what works, and find out, one clean experiment at a time. In 2019 the Sveriges Riksbank Prize in Economic Sciences went to Abhijit Banerjee, Esther Duflo and Michael Kremer "for their experimental approach to alleviating global poverty" — formal recognition that the credibility revolution had won the discipline's centre.
It did not win without a fight, and the fight is the interesting part. Angus Deaton — himself a Nobel laureate — argued that randomisation does not magically deliver unbiased, transportable knowledge, and that the method had displaced the harder questions of mechanism and theory. Lant Pritchett charged that RCTs are structurally biased toward small, targeted programmes and away from the growth and state-capability questions that actually move people out of poverty. Martin Ravallion questioned the field's unconditional preference for experiments on theoretical, ethical, and selection-bias grounds. From the political-economy and heterodox traditions, Sanjay Reddy and Ingrid Harvold Kvangraven argued that the randomista turn narrowed development economics and obscured its politics — that "what works" can never be separated from "for whom" and "according to whom".
This list moves through four stages: the rise of the method, the Nobel and its reception, the methodological critique, and the deeper political-economy critique. New to the debate? Start with Poor Economics and the Nobel popular-science summary, then read Deaton and Pritchett. Already running trials and feeling uneasy? Begin with Deaton & Cartwright and the Worm Wars.
Section 01
The Rise of the Randomistas
The method, its logic, and the texts that took field experiments from the seminar room to the centre of the discipline.
The book that brought the randomista programme to general readers, organised around case studies — bednets, deworming, microfinance, savings. The implicit argument is as important as the findings: that small, well-designed experiments can teach us more than sweeping theories of development.
The operational backbone of the movement, written by J-PAL practitioners. The chapters on threats to validity, sample size, and research ethics are what graduate students actually use when they design a trial — and what makes the method reproducible rather than charismatic.
The institutional engine of the credibility revolution. Beyond the trials themselves, J-PAL's "evidence to policy" case studies show the second half of the randomista claim — that experimental findings should drive how governments and NGOs spend money. Useful for seeing the movement's ambitions, not just its methods.
Section 02
The 2019 Nobel and Its Reception
The prize that crowned the movement — what the committee said it was rewarding, and how the discipline read the moment.
The committee's own four-page explanation of why Banerjee, Duflo and Kremer won "for their experimental approach to alleviating global poverty." The clearest short statement of the official case for the randomista programme — and a primary source for what the establishment thought it was honouring.
The technical companion to the popular summary, citing the laureates' key papers and situating the work in the wider literature. Dense but worth skimming to see exactly which contributions — bednets, deworming, education, credit — the committee treated as decisive.
A sympathetic but reflective account of what the prize signalled — not just a method but a shift in the profession's working culture toward field engagement and measurement. A good bridge between the celebratory coverage and the critiques that follow.
Section 03
The Methodological Critique
The careful, inside-the-discipline arguments that randomisation proves less than its advocates claim — about bias, transportability, and replication.
The most influential critical engagement with RCTs from within development economics. Deaton — a future Nobel laureate — argues that randomisation does not dissolve the problem of choosing what to study, nor does it deliver transportable knowledge without theory. The essay critics keep returning to.
The definitive statement of the methodological critique. Randomisation does not equalise everything but the treatment, does not by itself yield a precise average effect, and does not let us skip thinking about covariates or external validity. The single paper to read if you read only one critic.
A senior poverty economist's measured case against an unconditional preference for RCTs: the theoretical grounds are weaker than claimed, the ethical objections under-addressed, and insisting on randomisation introduces selection bias into the evidence base itself — we study what is randomisable, not what matters most.
The replication controversy that made the abstract debate concrete. When epidemiologists re-analysed the famous Miguel–Kremer deworming RCT and got weaker results, the field had a public argument about data, code, externalities, and what "the evidence shows." The best case study in why replication and transparency matter for the randomista programme.
Section 04
The Political-Economy Critique
The deeper objection: that the randomista turn narrowed the questions development economics asks, and stripped the politics out of poverty.
Pritchett's sharpest version of the argument: RCTs are structurally drawn to small, targeted, individuated programmes because that is what can be cleanly randomised — and away from growth, structural transformation, and state capability, which is where the largest gains lie. Read with charity even if you disagree.
A pointed review of Poor Economics from a development economist sceptical of the whole frame. Reddy argues that the randomista appeal rests on mono-causal reasoning and a faith that technology and trials can substitute for structural and political analysis. A bracing counter-read to Section 01.
A heterodox assessment timed to the Nobel. Kvangraven argues that, despite a rebellious self-image, the randomista enterprise produced a more exclusive, more technocratic development economics without improving our actual ability to fight poverty. The clearest statement of the political-economy objection.
A survey of the laureates' contributions by a leading practitioner, useful as the considered "defence" to read against the critics. It also documents how the field absorbed the critique — incorporating structural modelling, general-equilibrium effects, and external-validity concerns into experimental work.
Suggested citation
ImpactMojo Editorial (2026). "Randomista Economics and Its Critics." ImpactMojo Deep Dives. Retrieved from https://impactmojo.in/DeepDives/randomista-economics.html
Want to curate a Deep Dive?
If you teach, research, or practice in development and have a reading list worth sharing — pitch us.
Pitch a Deep Dive →