Impact · Essay
Universal vs. targeted: the wrong argument to have before anyone has money
The universal-basic-income-versus-targeted-cash fight is one of the loudest arguments in anti-poverty policy. It is also, today, mostly an argument about how to divide money that isn't reaching anyone as cash in the first place. Before funders litigate universality, they should move money to people who need it.
Key facts
- J-PAL's review of the cash transfer evidence finds cash raises consumption, food security and assets, with no increase in spending on alcohol or tobacco.
- The same review finds no systematic evidence that cash transfers reduce work.
- GiveDirectly's research in Kenya found large, lasting gains from unconditional transfers with no strings attached.
- Targeted programs like Mexico's Progresa/Oportunidades raised school enrollment — but targeting itself leaks: people in need are missed, and some aid goes to people who aren't.
- Direct cash remains a small fraction of global anti-poverty spending; most money still flows to in-kind programs, overhead and intermediaries rather than to people.
The debate assumes a faucet that's mostly off
Universal or targeted? The question only matters once cash is actually flowing to households at scale. Today, for most of the world's poorest people, it isn't. Anti-poverty spending overwhelmingly goes to things delivered on behalf of poor people: in-kind goods, services, contractor overhead, program administration. The direct-cash share of humanitarian and development aid remains small — GiveDirectly has long made this the core of its pitch, that only a tiny slice of aid dollars ever arrives as unconditional money a poor household can spend itself (GiveDirectly research).
So the universality argument is a fight over the plumbing of a pipe that barely exists. It's worth settling eventually. It should not be the gate that decides whether anyone gets water.
What the evidence says about universal-style cash
The strongest evidence base we have is for cash with few or no conditions attached. J-PAL, the Abdul Latif Jameel Poverty Action Lab at MIT, has reviewed dozens of randomized evaluations of cash transfer programs. Its bottom line: cash transfers "increase the consumption of food and other goods, and improve food security," reduce poverty, increase business activity, and — contrary to the most common objection — do not increase spending on alcohol or tobacco. Nor does J-PAL find evidence that cash transfers "discourage work" (J-PAL review).
GiveDirectly's flagship unconditional transfers in Kenya, studied by researchers including MIT's Johannes Haushofer, found large increases in consumption, assets and psychological wellbeing, with effects that persisted and even rippled through local economies. No conditions, no means test, no case manager deciding what the money was for (GiveDirectly).
In other words: the thing universal programs do — hand out money without screening — already works. That's the floor of the debate, and it's a solid floor.
What the evidence says about targeted cash
Targeted programs have real wins too. Mexico's Progresa (later Oportunidades), which attached conditions to transfers — school attendance, clinic visits — is among the most-studied anti-poverty programs ever run, and evaluations found it raised school enrollment and health outcomes. Conditional and targeted designs can direct more money per dollar to the poorest, at least in principle.
But targeting has costs the debate often skips: identifying who qualifies requires data, means tests and case-by-case administration; people get missed (undercoverage) or wrongly included (leakage); and conditions add compliance burdens on the very households with the least slack. The Niskanen Center's welfare policy work has argued at length that the design details of targeting — phase-out rates, cliffs, administrative burden — can eat much of the theoretical efficiency gain (Niskanen Center).
The honest summary: both models work. Cash helps whether or not you screen for it, and whether or not you attach conditions. The differences between them are second-order compared to the difference between cash and no cash.
Where the money actually goes today
That's what makes the universality fight so strange. It's an argument between two delivery mechanisms that both have strong evidence behind them, conducted by a funding world that mostly uses neither. A funder who refuses to send cash until the universal-vs-targeted question is settled isn't being rigorous. They're choosing the status quo — in-kind programs, intermediaries, and overhead — by default, and the evidence for that default is far weaker than the evidence for either side of the argument they're having.
What funders should do instead
- Move money now, pick the design later. Fund a direct cash program this quarter. Universal, targeted, conditional — the evidence says any of them beats not sending cash.
- Start targeted if you must, but measure the misses. If budget forces prioritization, target — and track undercoverage and leakage honestly, because those are the costs of targeting.
- Test universality where it's cheap. Pilots and RCTs are how this argument actually gets settled. Fund the study, not just the op-ed.
- Report what reached households. The metric that matters is dollars in hands, not dollars in programs.
The universal-vs-targeted debate is worth having — once the faucet is on. Until then, it's an argument about how to split money that hasn't moved. Move the money.
FAQ
Does universal basic income reduce work?
J-PAL's review of the cash transfer evidence finds no systematic evidence that cash transfers reduce work, and no increase in spending on alcohol or tobacco. Source
Are targeted cash transfers more efficient than universal ones?
In theory, targeting directs more money per dollar to the poorest. In practice, means tests cost money to run, miss eligible people, and wrongly include others. The Niskanen Center's welfare work argues these design costs can consume much of the theoretical gain. Source
What evidence is there for unconditional cash?
GiveDirectly's unconditional transfers in Kenya, studied by researchers including MIT's Johannes Haushofer, produced large, lasting gains in consumption, assets and wellbeing, with positive spillovers in local economies. Source
How much anti-poverty aid is delivered as direct cash?
A small fraction. Most aid and anti-poverty spending flows through in-kind programs and intermediaries rather than as money households can spend themselves — the core problem GiveDirectly was founded to point out. Source
Should funders wait for more evidence before funding cash?
No. The evidence base for cash — conditional or not, targeted or not — is already large and positive. The bigger risk today is funding neither model while the universal-vs-targeted argument continues. Source
Want a donate page your board and donors can actually read? Tell us about your organization by email at