Comments
By Caitlin Tulloch, Senior Director of Research, Learning, and Product at GiveDirectly
Caitlin is an expert in evidence-based policy and cost-effectiveness, and formerly served as Deputy Director in USAID’s Office of the Chief Economist.
Summary
GiveDirectly has written a lot about what giving people money is consistently very good at (e.g., recipients earn more, spend more, own more assets, and the local economy gets a boost), the myths that aren’t true (e.g, cash doesn’t make people work less or drink more), and the fact that it’s what most recipients prefer.
But the fact that cash improves many things doesn’t mean it improves everything. And frankly, that’s an unrealistic bar to set.
I’ve spent the last 15 years doing cost-effectiveness analyses across education, health, nutrition, and livelihoods—first at MIT’s J-PAL, then at the International Rescue Committee and USAID, and now at GiveDirectly.
The major lesson I’ve learned is that no one intervention is “the best” for all outcomes—we have to grapple with tradeoffs. So, allow me to state the obvious part out loud: Cash is not a silver bullet. There are no silver bullets.
While increasing people’s purchasing power can improve their lives, individual spending choices alone can’t solve collective action problems.
🚫🛣️ Cash doesn’t build infrastructure.
Roads, electricity, and water systems are public goods that require many people to invest at once. While there are stories of cash recipients pooling funds for things like wells, the average person who receives $500 isn’t starting up a road-building co-op—and they shouldn’t have to. Building infrastructure is a government job, typically made possible by raising and managing public funds.
In the United States, where I live, our biggest infrastructure achievements—like the interstate highway system—weren’t built because people spent their individual money more wisely. They were built because the government invested at scale through public systems. That’s an essential part of the longer-term economic transformation that we think of as “development.”
And of course, there is evidence on how to make smart infrastructure investments (see: performance-based block grants). If your goal is to improve infrastructure, consider programs like that!
🚫🏛️ Cash doesn’t build or reform public institutions.
For communities to sustain long-term economic growth, they depend on functioning courts, civil registries, and tax systems—institutions built on trust, coordination, and legitimacy, none of which you can buy in a market.
There’s some evidence that cash reduces vote-selling in elections, which is a good thing. But we haven’t seen consistent gains in civic trust or participation. And without broader reform, it’s a stretch to think that giving people money will make governments more responsive or accountable.
That’s okay. There are other tools designed for institutional reform (e.g., government audits, pay-for-performance schemes). Fund those, if that’s your goal.
🚫🏫 Cash can improve access to services, but it can’t make them work.
Where money is a key barrier, cash can help people improve a variety of outcomes. We’ve seen this in both education and health: Cash (both conditional and unconditional) can improve school enrollment and attendance and can also help people access nearby healthcare. But being in school doesn’t mean you’re learning. And visiting a clinic doesn’t mean there are drugs there to treat you.
I spent years evaluating education programs, and the evidence is clear: in many countries, teaching quality, curriculum relevance, and time spent learning are low. If your only goal is to improve learning, there are far more cost-effective ways to do so than giving cash.
This isn’t a failure of cash, it’s a reminder that some outcomes, like child learning or maternal survival, depend on functioning public sectors. One of my favorite studies tested this directly: a government cash transfer program was rolled out in Nicaragua, with and without improvements to local health and education systems. The biggest impact came from doing both. Because access and quality are complements, not substitutes.
As my colleague and research expert Dr. Miriam Laker Oketta summarized a recent study on cash and child mortality in Kenya, “The takeaway isn’t that cash replaces health care. It’s that cash makes health care reachable.”
If your goal is poverty alleviation, cash works well. Period. It’s cost-effective and consistent across many contexts. Unconditional cash transfers are the index fund of global development—reliable returns, low management overhead. (And frankly, the returns to many other interventions are pretty disappointing *cough— job trainings—*cough*).
That’s why we advocate using unconditional cash as a default in poverty alleviation programs. If other programs cost more and deliver less than cash, that shouldn’t be a hard decision.
But when the goal is something else (like reducing mortality, improving child nutrition, or responding to climate shocks), we don’t assume lump-sum cash is best. We look at the evidence, and when the evidence doesn’t exist, we test. Here are some of the cool things we’re testing right now:
For each, we’re asking the same question: What’s the most cost-effective way to improve a particular outcome? If the answer is cash, we scale. If it’s cash-plus-something-else, we adapt our core cash program. And if cash doesn’t perform comparatively well, then we’ve learned something useful for the aid sector!
Even though we are fired up about what cash does really well, improving people’s lives shouldn’t be a turf war. Many others are doing critically important work to build the systems and services that cash alone can’t provide. Both things can be true.
At GiveDirectly, we focus on cash because it solves poverty effectively, scalably, and respectfully. We follow the evidence on where it works best, and we are honest about where it isn’t the best tool.
We’re not in love with cash. We’re in love with what works.
This was well written and informative. I'm interested in improving government performance to allow coordination and combat the problems that come with corruption, so thank you for linking away to those solutions as well.