Could global poverty be eliminated simply by giving cash to every household that needs it? In theory, the arithmetic looks manageable: Identify everyone living below the poverty line and give them exactly enough money to cross it. In practice, governments rarely know precisely who needs help, and how much they need. That problem turns a simple calculation into a much harder policy challenge.
A new paper co-authored by Columbia Business School Professor Roshni Sahoo — along with Stefan Wager of Stanford, Paul Niehaus of the University of California, San Diego, and Joshua Blumenstock and Leo Selker of the University of California, Berkeley — takes steps toward overcoming that challenge.
“We don’t have direct measures for every household’s living standards in many developing countries,” Sahoo says. “What we tried to do in our research was to develop a realistic and actionable cost estimate that could be broadly applied.”
According to the researcher’s calculations, reducing the global extreme poverty rate to about 1% through targeted cash transfers could cost roughly $318 billion per year, the equivalent of 0.3% of global GDP. This finding suggests that the financial resources required to nearly eliminate extreme poverty may be more attainable than they appear.
Bridging the information gap
The new paper builds on a long history of efforts to estimate how much money would be needed to eliminate poverty. One common benchmark is the global poverty gap — the amount that would theoretically be required if policymakers knew exactly how far every poor household fell below the poverty line. But household surveys usually cover only a tiny fraction of a country’s population, and collecting detailed consumption data from everyone would be prohibitively costly.
At the opposite extreme is universal basic income. Under this program, everyone would receive the same payments, no matter what they earned. While that would avoid the problem of accurately targeting those below the poverty line, it would come at enormous expense.
Sahoo and her co-authors aimed somewhere in the middle. Rather than assuming perfect information, they asked how governments could allocate transfers using observable, verifiable household characteristics — the kinds of information that real social programs could collect. Using nationally representative household surveys from 23 countries that together account for about half of the world’s extreme poor, the researchers combined detailed consumption data with observable household characteristics, such as housing materials, education, geography, assets, and distance from markets. They then used algorithms to determine how much to transfer to different households to achieve a poverty-reduction goal at the lowest possible expense.
Targeting can dramatically lower costs
Across the 23-country sample, reducing the extreme poverty rate to 1% would cost about $170 billion per year using the researchers’ preferred targeting approach. That is roughly 5.5 times the theoretical cost under perfect information, but only about 19% of the cost of a universal basic income set at the poverty line.
The researchers also examined an important question about fairness. If the objective is simply to minimize the poverty rate, a policymaker has an incentive to prioritize people just below the poverty line because they are cheapest to lift above it. That can leave the very poorest households behind.
Instead, the authors emphasized minimizing the poverty gap, which provides larger transfers to people who are worse off. They found that this more equitable approach performs at least as well in practice at reducing the poverty rate.
From model to policy
While the findings suggest that the cost of dramatically reducing global poverty may be smaller than many people assume, implementation of such a plan would not be easy. In many of the poorest countries, the required transfers would represent a large share of GDP or government revenue, meaning substantial funding would likely have to come from abroad. Transfers on that scale could affect exchange rates, inflation, and local economic activity. Governments would also need up-to-date censuses and consumption surveys, along with reliable digital identification and payment systems.
There are political and social questions, too. Citizens may object to neighbors receiving different amounts, and people might change their behavior if they know which characteristics determine eligibility. For that reason, simpler rules may be easier to administer in some cases, even if they cost more.
Sahoo also stresses that cash transfers are not a substitute for more comprehensive anti-poverty policy. In places with poor access to markets, healthcare, or infrastructure, direct payments alone are not enough. A combination of cash transfers and other cost-effective interventions, such as investments in infrastructure, preventive healthcare, vaccines, and low-cost medicines, is likely to be more effective.
The research does not suggest that extreme poverty could disappear overnight. But it does put the scale of the challenge into a different perspective. “Raising people to the poverty line seems like a very expensive thing to do,” Sahoo says. “But when you actually do the calculations, it may not require as much money as you’d think.”
