EOP Frequently Asked Questions

We ask: “What would it actually cost to end extreme poverty through direct transfers?” While this question has been asked for decades, prior answers have relied on back-of-the-envelope calculations that make unrealistic assumptions about the information available to program administrators, and/or on a number of untested causal assumptions. This study uses modern statistical learning and household survey data to produce more realistic and actionable estimates.

New estimates suggest that lifting everyone above the $2.15/day extreme poverty line would cost about $170 billion per year, or 0.16% of global GDP, in the 23 poorest countries studied, which collectively account for around half of the world’s extreme poor.

The analysis then uses the results from these 23 countries, all of which have recent and reliable household survey data, to estimate the cost of ending poverty in countries where such data do not exist or are not publicly available. This extrapolation implies that it would cost $318 billion per year to reduce the global poverty rate to 1%. This amount corresponds to 0.3% of global GDP.

In recent decades people living in extreme poverty have made dramatic progress, with the global headpoint poverty rate falling from 41% in 1981 to 8% in 2024. This potentially makes ending poverty outright far more affordable than it has ever been in the past. At the same time, however, the world remains hundreds of millions of people off track from meeting the Sustainable Development Goal of eradicating poverty by 2030. And projections based on current growth trends indicate that progress against poverty is likely to slow to a halt, because the countries home to the most poor people are expected to grow slowly if at all. This makes it all the more urgent to identify ways to accelerate poverty reduction.

Earlier figures have sometimes equated the “poverty gap” with the cost of ending poverty, assuming governments knew exactly what each household consumed. That’s unrealistic. This study reframes the problem as one of statistical learning, producing cost estimates that reflect what policymakers actually observe. The result is estimates that are not just theoretical numbers, but actionable benchmarks.

While the costs are significant, they compare favorably to other global expenditures. For example: 

• The cost of eliminating poverty (our estimate): 0.3% of global GDP
• Annual expenditures on alcoholic beverages: 2.2% of global GDP
• Annual expenditures on cosmetics: 0.6% of global GDP

• For someone earning the U.S. median income (~$45,000/year), the global cost corresponds to a contribution of just $135 annually. Modest sacrifices could add up to ending extreme poverty.
• At current levels of private wealth, over a hundred individuals could end poverty in an entire country by pledging the returns on their assets. For high-net-worth philanthropists, this represents a rare opportunity for a truly transformative legacy.

Some low-income countries could contribute significantly, but most would not be able to finance full poverty eradication on their own. International financing will remain essential, particularly if it supplements — rather than replaces — existing high-impact programs.

Yes! According to the 2025 Forbes list, there are over 100 people whose net worth exceeds $18B. In an average year, each of those people might earn over $1.4B in investment income. If they directed that investment income toward the fight against poverty, it would be sufficient to eradicate poverty in one of the countries in our sample, such as Senegal or Togo.

No. These estimates show the direct transfer costs of ending poverty assuming that other beneficial programming continues as-is. Reallocating funds away from high-impact programs like health, education, or malaria prevention would undermine broader development gains. As Prof. Joshua Blumenstock, a co-author of the study, notes, “Defunding a high-impact program like malaria prevention in Nigeria would be a false savings. Such an action would hurt, not help, the bottom line in terms of poverty reduction.”

The study models the annual transfers required. That said, its estimates can be used to calculate the reductions in poverty that could be obtained from one-time contributions (or the creation of endowment funds), with the size of the gains depending on investment returns. Evidence from similar programs shows that for every $100 transferred, recipients’ annual earnings increased by about $17.

Yes. Using the World Bank’s new $3/day poverty line (adopted in 2025) would raise costs by roughly 50%.

In our 23-country sample (accounting for 50% of the world’s extreme poor): we find that data-driven targeting can achieve nearly the same poverty reduction as UBI at a fraction of the cost.

• Perfect targeting (impossible to implement, but a common benchmark): ~$31B per year.

• Data-driven targeting (the innovation of this study): ~$170B per year.

Universal Basic Income (UBI): ~$895B (flat $2.15/day PPP)

No. In the very poorest countries, the cost of ending poverty via UBI are about the same as using targeted transfers. And UBI may have political appeal because it is simple and inclusive. Outside of the very poorest countries, however, UBI is far more expensive than targeted approaches. Hybrid designs that combine wide coverage with data-driven prioritization may offer a viable compromise.

Perfect targeting (or “oracle targeting”) assumes a government could give each household exactly the amount needed to bring them to the international poverty line ($2.15/day). For example, if a household consumed $2.10/day, the transfer would be precisely $0.05/day. In reality, no government observes household consumption at this level of detail. Instead, they see only proxies — such as neighborhood, housing quality, or household characteristics — which are imperfect indicators of need.

Perfect targeting is therefore a useful benchmark, but not a feasible policy: it represents the lower bound on what it might cost to eliminate poverty if governments had omniscient information. Our approach is more pragmatic. We assume policymakers face the same informational limits they do in practice and then ask: “Given these constraints, what would it actually cost to end extreme poverty?”

We selected 23 countries with recent, high-quality surveys that are significant contributors to global extreme poverty. Together, they represent about half of the world’s extreme poor. We then scale results using global poverty data (e.g., World Poverty Clock) to estimate global costs. These extrapolations should be read as order-of-magnitude estimates — the true global figure depends on country-specific conditions, but the takeaway is that ending extreme poverty would require less than one  percent of global GDP.

No. Our calculations focus on the direct cash transfers needed to lift households above the poverty line. They do not include administrative costs (such as targeting infrastructure, delivery systems, or leakage) or broader macroeconomic effects (such as changes in labor supply, prices, or growth). Those factors could raise (in the case of administrative costs) or lower (if large economic stimuli have a multiplier effect) the true cost in practice. Our contribution is to provide clean benchmarks for the transfer costs themselves, which policymakers can then adjust for their specific context.

No. They are order-of-magnitude estimates intended to guide policy discussions. The true cost will depend on context: local conditions, available data, political choices, and program design. Our contribution is to provide credible benchmarks that are more realistic than past back-of-the-envelope figures.

The study was carried out by Roshni Sahoo (Stanford University), Joshua Blumenstock (UC Berkeley), Paul Niehaus (UC San Diego), Leo Selker (UC Berkeley), and Stefan Wager (Stanford University).

Please contact Joshua Blumenstock at jblumenstock@berkeley.edu, or 510-816-4911.