Poor Economics: A Radical Rethinking of the Way to Fight Global Poverty
Source: https://www.publicaffairsbooks.com/titles/abhijit-banerjee/poor-economics/9781610390934/ ↗
Banerjee and Duflo, later joint Nobel laureates, wrote the book that argued development economics had been asking the wrong question: not "what policies work" at the country level, but "what works for whom, under what conditions" at the level of specific interventions.
Their method — randomised controlled trials in the field — is the transposition of biomedical evidence standards to social policy.
The book is a patient tour through specific studies: bednets, school attendance, microcredit, vaccines.
For product direction the transfer is useful on two levels: the habit of treating large claims with specific scepticism, and the operational model of learning by experiment at small scale before scaling.
Read alongside Demirdjian for the medical history of the method.
Central argument
Banerjee and Duflo argue that development economics had failed not from lack of resources or political will but from operating at the wrong unit of analysis: sweeping country-level policy prescriptions substituted for evidence about specific interventions working for specific people under specific conditions. Their corrective is methodological — importing randomised controlled trials from biomedicine into field economics to test discrete interventions (bednet distribution, conditional cash transfers, microcredit schemes) and let results accumulate into a granular, revisable picture of what actually moves outcomes. The central finding is that the poor are not irrational or helpless but are making constrained decisions under informational and structural conditions that policy can alter, one tested lever at a time.
Critique
The RCT methodology that gives the book its rigour also sets a ceiling on what it can explain: randomised trials are well-suited to measuring the local average treatment effect of a bounded intervention, but they are structurally ill-equipped to evaluate systemic or institutional change — the kind that cannot be randomised into a control group. Critics including Deaton have argued that external validity is the book's unresolved problem: a bednet trial in Kenya tells you what worked in that context, not what will work when scaled nationally or transplanted, yet the policy appetite is always for generalisable prescriptions. There is a latent tension between the epistemic humility the authors preach and the confidence with which findings get absorbed into development orthodoxy.
Why it matters for product
The core reframe — from 'what is our strategy' to 'what works for whom under what conditions' — maps directly onto how a CPO should structure discovery: not seeking universal product truths but running tightly scoped experiments with defined populations, measurable conditions, and honest null results. The operational discipline Banerjee and Duflo model also challenges the scaling instinct common in product organisations; their framework implies that a result earned at small scale carries an obligation to re-test assumptions at the next order of magnitude rather than treating scale as automatic validation. For teams that conflate roadmap velocity with learning, the book is a corrective on what 'evidence-based' actually requires.