#872 1997 · Government of Mexico (Ministry of Finance; architect Santiago Levy) · Public policy / poverty programs
Mexico paid poor families cash — but only if the children went to school and to clinics
the problem
Decades of food subsidies for the poor were expensive and leaking, and poverty kept reproducing across generations
background
After the economic crisis that unraveled Mexico in 1994, poverty rose sharply and the government's established answer was showing its age: decades of in-kind food subsidies, expensive to run, leaking value between the state and the table, and doing nothing to change what poor households could become. The subsidies fed poverty without ending it, and every peso spent on them was a peso that could not be spent on anything else.
At the Ministry of Finance, deputy minister Santiago Levy produced a report on poverty alleviation that became the seed of something else: a pilot, then — after Levy pushed for proper monitoring and coordination before scaling — a national program launched in 1997 as Progresa. The money came from reallocating the ineffective food subsidies the program displaced. It started with 300,000 rural families and a budget of $58 million, four-thousandths of one percent of GDP, administered by the social-development ministry SEDESOL with a council that gave health, education and finance each a stake in its success.
what everyone would do
The standard toolkit was in-kind: subsidised food distribution, price supports, blanket transfers. Each consumed budget without changing what households did next — a fed child could still leave school at twelve — and each, once in place, was politically expensive to remove and almost impossible to evaluate.
what they saw
The scarce input wasn't food but human capital being formed in each household. Levy stopped subsidising consumption and started buying behaviour: cash to mothers, conditional on verified school and clinic attendance.
the move
Progresa paid cash directly to poor families — amounts equivalent to about a third of household income, delivered every two months to mothers — on two conditions: children attended school (at least 85% attendance, from third year of primary through third year of secondary, with larger grants for girls), and the family used a package of preventive health care. The rollout was built as an experiment from the start: 506 villages were randomised, 320 receiving the program and 186 serving as controls, with baseline data on some 24,000 households — the first time Mexico had evaluated a national anti-poverty program this way.
why it works
Cash respects the household's own knowledge of its needs better than any delivered good, while the conditions tie the payment to exactly the actions that break intergenerational poverty — schooling and preventive health — so the transfer funds the family's present and purchases its children's future at the same time. Paying mothers aimed the money at the household members most likely to spend it on children; paying for verified behaviour made the program auditable; and randomising the rollout produced the evidence that let the program survive elections, because each new administration inherited proof rather than promises.
the payoff
Poverty fell 1.8 points, child illness ~23%, secondary enrolment +3.4 points for girls — from $58m and 300,000 families in 1997.
where it breaks
It fails where the services on the far side of the conditions don't exist — a school to attend, a clinic to visit; Progresa's own targeting missed the poorest households that lacked access to either. It fails where the cash is smaller than the income a child's labour would bring, and wherever conditions can't be verified cheaply. And it remains hostage to politics: three renamings survived, but the program ended by decree, one election away from disappearance — its measured effects dying larger than they were born.
what came after
Renamed Oportunidades, then Prospera, the structure survived three changes of administration because each ministry owned a piece of it, grew to cover some five million families, and was replicated in more than 50 countries. Its end was political, not evidentiary: wound down abruptly in 2019, and a 2025 study found school enrolment fell immediately, losses concentrated among high-school-age boys — the program's effects at its death even larger than at its birth. The targeting was imperfect too: about 24% of eligible families were left out, 22% of recipients didn't qualify, and 90% of surveyed recipients thought the selection unfair.
references
- [1]From Progresa to Prospera: A Case Study on Mexico's Conditional Cash Transfer ProgramHarvard Ministerial Leadership Program, 2025ministerialleadership.harvard.edu
- [2]Rolling Back Progresa: School and Work After the End of a Landmark Anti-Poverty ProgramNational Bureau of Economic Research, 2025nber.org
- [3]PROGRESA Conditional Cash Transfer — registry entryThe Experiment Society, 2026experimentsociety.org