Forskare pekar ut snabb väg att minska världens fattigdom

An age-old problem might have a surprisingly simple solution.

A world free of poverty would once have sounded fanciful. For most of human history it was barely even imaginable. Around 1800, when more than eight in ten people were destitute, Thomas Malthus, a particularly dismal dismal scientist, insisted it was an iron law that “Some human beings must suffer from want.” Herbert Spencer, similarly cheerless and an early contributor to The Economist, cast poverty as the “survival of the fittest”. The pessimism was not confined to gloomy Englishmen such as these two. The idea that poverty could be eliminated scarcely appears in print before the mid-20th century. Even Jesus, in three of the Gospels, tells his followers that the poor will always be with them.

At the turn of the last millennium, the world opted to try anyway. The 189 member states of the United Nations set a target to bring the share of people living on less than $1.25 a day to half its 1990 level by 2015. Astonishingly, given the size of the task, it was met and then some: 1.2bn people escaped penury in those 25 years, bringing the global poverty rate down from 43% to 13% (using today’s poverty line). Economic growth did nearly all the work. A booming China accounted for about two-thirds of the decline; red-hot India and Indonesia did much of the rest. It looked as though growth miracles might consign poverty to the past.

Miracles, however, are hard to repeat. Since 2015 the rate of poverty reduction has slowed sharply, to about a quarter of its previous pace. Roughly one in ten of the world’s people, or 830m, are still destitute today. The old playbook—providing aid in the hope it will boost growth—has disappointed, not least because the recipe for such growth remains elusive. Moreover, poverty is now concentrated in places where growth is harder to achieve, and population size is rising fast. Around seven in ten of the world’s poor are in sub-Saharan Africa; the Democratic Republic of Congo, Ethiopia and Nigeria alone account for a quarter of the total. If current poverty rates persist, rapid population growth means that these three could be home to more than two-fifths of the world’s poorest by 2050. Will Malthus be vindicated after all?

Not necessarily, because there is a simpler way to end destitution. Instead of turning Congo into China, it involves closing the gap between what people have and what they need. In principle, this means giving poor people enough cash to push them over the poverty line: if someone lives on $2 a day and the threshold is $3, they need $1 more. With perfect information, the cost of doing so would be strikingly low, and falling. Measured in today’s dollars, it dropped from $330bn a year in 1990 to $90bn in 2023.

One problem is that no one knows exactly who is below the poverty line and by how much. One solution would be a universal basic income—meaning payments to everyone, whatever they earn. Applied to the poor world at a level of $2.15 per day, this would cost 2-3% of global GDP each year. In reality, governments keep costs down by trying to direct support towards the neediest. Most of the over 130 countries with cash-transfer programmes rely on “proxy-means tests”, using crude indicators such as housing material or distance to markets to infer who is poor. This is cheap and scalable, but flawed. In one study of nine African countries, such approaches missed up to four-fifths of those below the poverty line, while wrongly targeting about half of those above it.

Ending extreme poverty therefore comes down to identifying who needs what. A recent paper by Roshni Sahoo of Stanford University and co-authors proposes a way to do so. Using data from 23 countries, they simulated a setting with limited information and asked how cash transfers should be allocated to drive poverty as low as possible. Rather than trying to predict who is poor, their method uses a machine-learning algorithm to assign transfer amounts—different for each person—so as to minimise the chance that anyone remains in poverty after receiving support. In other words, it targets the poverty gap, not the line itself, albeit less efficiently than would be possible with perfect information. The authors then used this framework to estimate the cost of (nearly) eliminating extreme poverty worldwide.

The result is a surprisingly small bill. The paper’s estimates suggest it would cost $318bn a year to reduce the global poverty rate to 1% at the $2.15-a-day line—roughly 0.3% of global GDP—with imperfect, real-world information. Even using the World Bank’s newer $3-a-day threshold, the bill rises to only $466bn, or about 0.5% of global GDP. That is less than a third of what the world spends on alcohol each year, and well below what rich countries devote to farm subsidies.

Naturally such estimates involve big simplifications. Many of the poorest countries could not deliver transfers reliably, especially in war-torn areas. Large cross-border flows of cash can distort incentives and strain exchange rates, and poorly targeted programmes risk fuelling resentment—as in Indonesia, where opaque eligibility rules have sparked unrest. And the bill is not fixed. Fuel and food shortages caused by conflicts, such as those in Ukraine and more recently Iran, can quickly swell the ranks of the poor.

None of this is insurmountable, though. As Alfred Marshall, a founding figure of modern economics, once observed, eradicating poverty is less a quandary for economics than for the “moral and political capabilities of human nature”. Yet he also noted that noble instincts can be stirred by facts. A survey by Abhijit Banerjee and Esther Duflo, two Nobel-prizewinners, and Michael Greenstone, of the University of Chicago, suggests as much: around 60% of rich-world respondents say they would be willing to give up 0.5% of their income if that were enough to end extreme poverty. Growth miracles are hard. Putting a price on the alleviation of poverty is easier—and may be enough to prompt action. 

© 2026 The Economist Newspaper Limited. All rights reserved.