Foundationsdevelopment
Poverty: Measurement, Causes, and the Fight Against It
Overview
Over 700 million people still live in extreme poverty (under $2.15/day), even after the greatest poverty reduction in history. Understanding poverty — how it's measured, why it traps people, and what lifts them out — is development economics' core mission. This lesson covers measurement (headcount, gap, multidimensional), the mechanics of poverty traps, the record of reduction (dominated by China and India), and the interventions with the strongest evidence: cash transfers, health, and education.
Learning objectives
- Define absolute vs. relative poverty and major poverty lines
- Explain the headcount ratio, poverty gap, and MPI
- Describe poverty-trap mechanisms
- Summarize global poverty trends and their drivers
- Evaluate anti-poverty interventions by evidence strength
How poverty is measured
The World Bank's extreme-poverty line is 3.65 for lower-middle-income, $6.85 for upper-middle-income) reflect rising norms. Three statistics matter: the headcount ratio (share below the line), the poverty gap (how far below, on average — depth), and the squared gap (severity, weighting the poorest most). But income misses deprivations: the Multidimensional Poverty Index (OPHI/UNDP) counts overlapping shortfalls in health, education, and living standards — revealing 'hidden' poverty among the income-non-poor and vice versa.
A country can cut its headcount while the remaining poor get poorer. Always check the poverty gap alongside the headcount — depth matters as much as breadth.
Poverty traps: why poverty persists
Poverty traps are self-reinforcing mechanisms: being poor makes it harder to escape poverty. Nutrition-based: undernourished workers are less productive, earning less, eating less (the efficiency-wage logic). Credit: the poor can't borrow to invest because they lack collateral — missing credit markets block the climb. Health: malaria and worms sap energy and school attendance. Behavioral: scarcity taxes mental bandwidth (Mullainathan & Shafir), impairing decisions. Geography: remote regions lack markets and services. Each trap implies a different escape: nutrition, microcredit, health interventions, or infrastructure — which is why one-size-fits-all anti-poverty policy fails.
Banerjee and Duflo's famous diagram: below a threshold, people's future income falls (the trap); above it, it rises. Escaping requires a push past the threshold — the logic behind 'big push' and graduation programs.
The great reduction — and what drove it
Extreme poverty fell from ~36% of humanity (1990) to under 10% (pre-pandemic) — history's fastest escape. China accounts for the bulk (market reforms + industrialization lifting 800 million). India contributed hugely after 2000. The drivers: sustained economic growth (the dominant factor — growth's poverty elasticity is well-established), agricultural productivity (Green Revolution), and later, targeted social programs (Brazil's Bolsa Família, Mexico's Progresa/Oportunidades). Sub-Saharan Africa lagged: faster population growth, conflict, and weaker structural transformation meant absolute numbers of poor stayed high even as rates fell.
COVID-19 pushed ~70 million back into extreme poverty (World Bank) — the first sustained increase in decades, concentrated in South Asia and Sub-Saharan Africa. Recovery has been uneven.
What works: the evidence ranking
Randomized trials reshaped the field. Strongest evidence: unconditional cash transfers (GiveDirectly trials show lasting gains in assets, earnings, and child health); deworming (Kremer/Miguel — huge education and earnings returns per dollar); teaching at the right level (Pratham's Read India); and graduation programs (BRAC's asset-plus-training packages lift the ultra-poor sustainably). Moderate: microcredit (helps some businesses, no transformative average effects — the great microfinance disappointment); conditional cash transfers (Bolsa Família's health/education gains are real). Weakest: many training and 'empowerment' programs with no rigorous evidence. The meta-lesson: test, don't assume.
The striking finding of the 2010s: just giving poor people money often outperforms elaborate programs. Simplicity plus rigorous testing beats paternalistic complexity.
The remaining agenda
The 'last mile' is the hardest: extreme poverty concentrates in fragile states (conflict zones where programs can't operate), among marginalized groups (remote, ethnic minorities), and in Sub-Saharan Africa's fast-growing populations. Climate change threatens to push 100 million more into poverty through crop failures and disasters. The frontier: scaling what works (cash transfers via mobile money now reach hundreds of millions), adapting to fragility, and recognizing that poverty reduction increasingly means tackling inequality and climate vulnerability, not just raising average incomes.
Key points
- Poverty is measured by headcount, depth (gap), and multidimensional deprivations — use all three.
- Poverty traps (nutrition, credit, health, bandwidth) explain persistence and imply targeted escapes.
- Extreme poverty fell from 36% to under 10% since 1990, led by China and India — growth did most of the work.
- Strongest evidence: cash transfers, deworming, teaching at the right level, graduation programs.
- The last mile concentrates in fragile states and Sub-Saharan Africa; climate is the new threat.
Common misconceptions
✓ Truth: The evidence shows structural traps — missing credit, disease, poor schools, remote geography — dominate. Behavioral effects of scarcity itself impair decision-making; blaming the poor misdiagnoses the problem.
✓ Truth: Growth — especially in China and India — did the overwhelming bulk. Aid's macro record is ambiguous; its micro successes (vaccines, cash transfers) are real but smaller in scale.
✓ Truth: Rigorous trials find modest effects: some businesses benefit, but average transformative impacts don't materialize. Microcredit is a useful financial service, not a poverty cure.
Practice questions
Q1. The poverty gap differs from the headcount ratio because it measures:
- The number of poor people
- How far below the poverty line the poor fall on average (depth)
- Inequality among the rich
- Government spending
Headcount counts the poor; the gap measures depth — essential because a falling headcount can mask deepening destitution among those who remain poor.
Q2. A 'poverty trap' is best defined as:
- A government program that fails
- A self-reinforcing mechanism where poverty itself blocks escape (e.g., undernutrition reducing earnings)
- A type of tax
- Urban slums
Traps are feedback loops: poor nutrition → low productivity → low income → poor nutrition. Breaking them requires pushes past thresholds — the S-curve logic.
Q3. The intervention with some of the strongest RCT evidence for poverty reduction is:
- Microcredit for all
- Unconditional cash transfers
- Vocational training
- Tourism promotion
GiveDirectly-style cash transfer trials show sustained gains in assets, earnings, nutrition, and child health — often outperforming far more complex programs.
Click a question to reveal the answer and explanation.
Frequently asked questions
What is the $2.15/day line based on?
The average of national poverty lines in the world's poorest countries, updated to 2017 purchasing-power prices. It's meant to capture absolute deprivation — survival-level needs — not relative standing.
Why did Africa's poverty fall more slowly?
Faster population growth (diluting per-capita gains), more conflict, weaker manufacturing transformation, commodity dependence, and later demographic transition — structural headwinds beyond policy alone.
Do cash transfers make people work less?
Trials consistently find little to no reduction in work — and often increases, as recipients invest in assets and businesses. The 'laziness' fear is empirically unsupported at typical transfer sizes.
What are 'graduation' programs?
BRAC's model for the ultra-poor: a productive asset (livestock), training, consumption support, and coaching for ~2 years. RCTs across six countries show lasting earnings gains — the best-tested ultra-poverty intervention.