Abhijit V. Banerjee and Esther Duflo · 2011

Poor Economics

A Radical Rethinking of the Way to Fight Global Poverty

Rough AI truth score

83/100

Banerjee and Duflo's central move—replace stereotypes with close observation and testable questions—has durable evidentiary support. Randomized evaluations sharpen causal learning for bounded interventions, though transport, equilibrium effects, ethics, and macro institutions limit how far local results travel.

Based on three central claims · high confidence

The three claims

01Supported

Poor people make purposeful decisions under severe constraints rather than behaving according to one irrational or helpless stereotype.

Household studies repeatedly show context-sensitive tradeoffs involving risk, liquidity, information, health, and social obligations. Purposeful does not mean perfectly informed or bias-free; it rejects the stronger claim that poverty can be understood through a single deficit of motivation.

IMF profile of Banerjee and Duflo's poverty research Nobel Prize scientific background on experimental poverty research

02Mostly supported

Randomized evaluations can reveal whether specific anti-poverty interventions work and overturn plausible but false assumptions.

Random assignment offers unusually clear causal estimates when implementation and measurement are sound, and studies have changed policies in health and education. Attrition, spillovers, multiple testing, ethics, and implementation can still compromise inference, while null findings may not explain mechanisms.

NBER review of the experimental approach to development Nobel Prize scientific background on experimental poverty research

03Mostly supported

Accumulating many small, context-specific results provides a better route to poverty policy than relying on grand theories alone.

Incremental evidence improves program design and guards against intuition-driven waste, especially for service delivery. Local treatment effects do not automatically aggregate across settings or reveal macroeconomic, political, and general-equilibrium consequences, so theory and institutional knowledge remain necessary.

IMF discussion of external validity and RCT limits NBER experimental-development methodology review

Other claims worth checking

Small pricing differences can sharply affect take-up.
Microcredit is useful for some borrowers but not a universal escape.

What this number means. It is an AI-generated first-pass judgment of three central factual or causal claims—not a rating, exhaustive fact-check, or human peer review. Claim credits are 100% for supported, 75% for mostly supported, 50% for mixed, and 25% for weak, then averaged and rounded. Lower confidence means the score should move more as better evidence arrives.

Method three-central-claims/0.1.0 · checked 2026-09-01 · 3/3 selected claims assessed · method and source audit