How conditional cash transfers actually work.

When Mexico's Progresa program launched in 1997, it was one of the first large-scale conditional cash transfer programs in the world. Families in extreme poverty received regular payments — on the condition that children attended school and adults received preventive healthcare. The original randomized evaluation, published by researchers at the International Food Policy Research Institute, found strong effects on school enrollment, health checkups, and household consumption.

The program was subsequently renamed Oportunidades, then Prospera. It was evaluated again. And again. More than 200 papers have been published on its effects, making it one of the most studied social programs in history.

What the evidence shows

The short answer is: the conditions are probably not the mechanism.

Several subsequent experiments have found that unconditional cash transfers — programs that give money with no behavioral requirements — produce comparable or better outcomes on most measures. GiveDirectly's work in Kenya is the largest of these experiments: direct transfers to extremely poor households, no strings attached, show gains in assets, earnings, food security, and psychological wellbeing that persist for years.

Families who receive cash without conditions invest it in their children's education and health at high rates anyway. The conditionality may be politically necessary to build public support for the program. It may not be economically necessary to produce the results.

Why this matters for design

If the conditions are not the mechanism, then the mechanism is probably the cash itself — specifically, the increase in household liquidity and the removal of budget constraints that force families to make trade-offs they would prefer not to make.

This reframing has design implications. Conditional programs require an administrative apparatus: verification of school attendance, health checkup records, eligibility documentation. That apparatus has costs — administrative costs for the program, compliance costs for recipients. If the conditions aren't driving the outcomes, those costs are pure overhead.

It also has targeting implications. Conditional programs systematically exclude the households that most need transfers but are least able to meet conditions: families with disabled children, households in areas without schools or clinics nearby, parents working multiple jobs who cannot accompany children to checkups.

The political economy problem

The countervailing argument is that conditionality is what makes cash transfers politically sustainable. A program that requires recipient behavior can be framed as investment rather than charity. In contexts where unconditional transfers face political opposition, the conditionality may be a necessary feature even if it's not an effective mechanism.

This is a legitimate argument. It is also a testable one. What the evidence does not support is the belief that conditions produce outcomes by changing recipient behavior in ways that pure cash transfers would not.

What to do with this

For practitioners designing new cash transfer pilots, the implication is straightforward: run a multi-arm trial comparing conditional, unconditional, and in-kind transfers in your specific context. Don't assume the mechanism from the international evidence. Test whether the conditions add value in your setting.

The international literature gives you a strong prior. Your local context may differ — in ways the literature cannot predict.

Next issue: The null result problem — why the most important experiments don't get published.