This is a linkpost for

Cross-posted from the EA Forum

In this video, we outline a plan to eradicate global extreme poverty using cash transfers. The approach is based on GiveDirectly's internal research and documentation about the topic. Allen Liu, the first author of this post, has written the script. Writer, the second author, has given feedback. Give Directly provided significant help and feedback during the script-writing and storyboarding phases. 

On the top-right of the YouTube page, we included a fundraiser for GiveDirectly. Our goal is to raise 100,000 USD by the end of the year. 

Below, you can find the script of the video.

Over 600 million people worldwide can’t afford basic needs, like enough food, clean water, electricity, or basic life-saving healthcare[1].  Their day to day survival is under threat in ways unimaginable to most people watching this video. It's honestly kind of embarrassing  for humanity as a whole that hundreds of millions of us are suffering so much simply because they don't have enough money. Couldn’t we just give them some money? Could we actually end global extreme poverty with such an enticingly simple approach? Let's find out!

We would like to thank GiveDirectly for helping us write this video by providing their detailed research on the topic. They are a nonprofit that delivers donations as cash directly to people in extreme poverty across the world.

Giving money for people to spend without any restrictions is a type of aid called “unconditional cash transfers” or “UCTs”.  Like all forms of aid, there are both advantages and limitations to cash transfers as a whole, and UCTs in particular.

If we want to end global extreme poverty through cash transfers, we’ll need to raise enough funds, and send them to the people who need them.  We also need to be sure that we are actually helping the recipients, and not hurting people who are not receiving anything.  We also recognize in advance that there are a host of moral and ethical questions involved with such a large undertaking.  For example, who should be given money first?  And do we need to prevent people becoming dependent on aid?  Additionally, poverty does not exist in isolation from the world’s other major issues such as food security, human rights, and climate change.  Those questions won’t be the main focus of this video.  Instead, we’re going to focus on what can be achieved with cash transfers.

Let’s start with some basic facts about how we define global extreme poverty.  As defined by the World Bank in September 2022, the international poverty line is $2.15 per day in 2017 US dollars by purchasing power.[2]  This is roughly equivalent to $1 per day in US dollars in 1987.[3] At the end of 2022, up to 685 million people experienced poverty worldwide by this definition.[4] The international poverty line represents what we call “extreme poverty” as opposed to “poverty” in general. Countries can decide to use their own poverty line, but we are going to focus on the most extreme cases of poverty.   The international poverty line reflects the cost to cover basic human needs among the world’s poorest people, including food, shelter, and minimal healthcare and education costs. It's set based on the average price for this rough 'basket of goods' a person needs. To be clear, the international poverty line represents a very low standard of living, but one that lets a person avoid an immediate risk of death from starvation, exposure to the elements, or easily treatable diseases.  In terms of food, for example, the UN Food and Agriculture Organization estimates that in 2017 just 83 cents a day could buy enough food to avoid starvation, but a fully nutritious diet would require more than four times that, at $3.54 per day.[5] While many organizations consider more than just household income in defining poverty, including the World Bank’s own “Multidimensional Poverty Measure,”[6] the international poverty line remains a popular and useful tool for understanding the world’s poorest people as a population.  In particular, the international poverty line is the target used for the 2015 UN goal of ending extreme poverty by 2030 as part of the Sustainable Development Goals.[7] [8] [9]As GiveDirectly puts it, getting everyone above the international poverty line is not a high bar – it’s the lowest, and it’s one we should be able to clear.

And reaching this bar can change lives, and entire communities. Our team spoke by video call with one Rwandan woman, Valentin Oramahoro, in a village that had received cash from GiveDirectly several months prior.  She spoke about how she was able to invest in her store by moving to a new building with floors made of cement instead of uncovered ground. She also bought a tea kettle, and made new kinds of merchandise, like soft drinks, available to her community.  She mentioned that her proceeds from her store doubled from about 3 dollars to about 6 dollars per day, both thanks to her improvements to the store and from people in her community having money to spend after meeting their basic needs.  With the money they received, her neighbors could afford food every day, get cement for their floors, buy new livestock and farm tools, and more.

A 2016 review of the scientific literature by Francesca Bastagli and her colleagues at the Overseas Development Institute[10] helps clarify the picture in the aggregate.  As expected, household income and spending both increase for people receiving cash transfers.  Most studies also found that cash transfers improve a wide variety of other outcomes: including improving dietary diversity, increasing the number of medical checkups, and increasing the number of businesses and the productivity of farms.  Adults receiving cash in most of the studied programs spent about the same amount of time working, while children receiving cash spent fewer hours working for pay and more time in school.

These effects are quite persistent: studies of a 2008 lump sum cash transfer of 380 US Dollars in Uganda showed positive economic effects on recipients four years later.  Though a follow up at 9 years did not find a significant effect on incomes, in another study at 12 years during the COVID-19 pandemic the authors found that people who received cash transfers were able to better withstand the economic effects of the pandemic due to their type of occupation.[11]

The benefits to people receiving cash transfers are clear.  So how much money would we need to bring everyone above the international poverty line?  As a rough guideline, suppose we found every person in extreme poverty and added up the amount they’d each need to get just above the poverty line.  The Brookings institution estimates this quantity, called the "poverty gap", to be around $100 billion each year.[12] The real amount we’d need will be different, and we’ll discuss the reasons why in a moment, but we shouldn’t expect it to be many times lower or higher.

$100 billion is certainly a large amount of money, but given the scale of the problem, it’s actually quite manageable.  According to the OECD, governments worldwide gave over $201 billion of official development assistance in 2022.[13]  The OECD defines this type of assistance as resources sent from governments to developing countries with promotion of economic development and welfare as the main objective.  GiveDirectly estimates less than 5% of this is being given directly as cash to individuals; the exact amount is not being tracked by the governments giving this aid.  Of humanitarian aid, which aims to provide short-term relief from specific disasters rather than long-term development,[14] around $5.6 billion of cash aid was given in 2019, out of a total of roughly $30 billion.[15] Governments also don’t need to be the only source of funds: existing cash aid is dwarfed by the $1.5 trillion held by charitable foundations in 2018,[16] and total charitable donations from wealthy individuals, companies, and the general public surpass the $100 billion total many times over each year.[17]  So while the increase in total aid would be comparatively small, the increase in cash aid would be very substantial.

If we as a civilization decide to go through with this plan, how would we get the money to the people who need it, and how much additional funding would that require?

For starters, communications technology will be a big help.  GiveDirectly in particular sends money through mobile payments, with their field workers giving SIM cards and phones to recipients.  This helps recipients keep their money secure, simplifies the payment process, and gives the recipients an additional tool to help improve their lives.  GiveDirectly has even helped set up cell phone towers to make these payments work.[18]

The organizations that are in charge of the cash transfer program will also have some maintenance and delivery costs.  GiveDirectly’s financial releases[19] show that around 90% of the money they raise actually reaches recipients.  A similar style of program designed to reach everyone in extreme poverty would have some new challenges, such as reaching refugees or very remote communities, but would also benefit from economies of scale.

We’ll also have to determine who should get the money and how much.  This is often called “targeting”.  Any system for targeting will necessarily introduce inefficiencies.  Money we spend on finding who is eligible is money that we don’t give out.  Since our goal is to end global extreme poverty, we should be more focused on making sure that no one below the poverty line is excluded rather than avoiding including people who are just above the poverty line.  Additionally, many people close to the poverty line don’t primarily earn income as currency, so it might not even be possible or meaningful to identify them as just below or just above the international poverty line.  One popular method for targeting aid is “proxy means testing”, that is using information about people’s living conditions to estimate their income.  A statistical analysis by researchers at the World Bank[20] found that in places with a high poverty rate, giving based on proxy means testing can actually cost more and be less effective than paying the same amount to everyone, or nearly everyone.  For these reasons and others, it would likely be easier to spend our effort raising more money to give to those on the margin of poverty in poor communities rather than on meticulous targeting and means testing.

What about corruption? Corrupt actors – both local and foreign – are already skimming off at least tens of billions of dollars from the economies of Africa’s poorest nations.[21]  If we want to transfer money to hundreds of millions of people in these nations, the region’s governments will have to be important partners.  Giving money via mobile payments directly to recipients reduces the chance for funds to be lost to corruption, but we will still inevitably lose some, and we’ll have to spend some resources stopping corrupt actors within the project, given its large scale.  Of course, this is a problem with any aid program: cash is no riskier than other forms of aid in this respect.[22]

Overall, other than the 100 billion dollars estimated to close the annual poverty gap, most of the additional costs come from inefficiencies in the targeting method, with the other factors such as corruption and delivery costs accounting for smaller portions.  The people we spoke with at GiveDirectly weren’t ready to share an ultimate number, but their rough estimate is $200 to $300 billion dollars to reach everyone in extreme poverty for the first year of our project. This figure is in line with estimates from other organizations based on population data.  Once again, as big as this figure is, it’s less than one 500th of the world’s economy at $160 trillion dollars, and only about 1% of the government budgets of the world’s wealthier nations at $27 trillion.

So we know the amount of money, while large, is not insurmountable for developed nations.  It wouldn’t even be that large an increase over existing aid.  We also know that cash transfers improve recipients’ lives immensely.  But what would be the effects on recipients’ neighbors, and the economy as a whole?

A 2022 study led by Dennis Egger[23] found that every $1,000 of cash given actually had a total economic effect of $2,500 thanks to “spillover” effects growing the local economy, as recipients spent more money at their neighbors’ businesses, and those businesses spent money, and so forth.  Not only did recipients’ incomes increase, their neighbors’ incomes also increased by 18 months later.  Even neighboring villages without any recipients saw slightly increased incomes, which could have been from a “spillover” effect as well.  These effects mean our cash transfers will go farther, and we may find that we’ve reached our goal of ending extreme poverty sooner – and for less money – than we would otherwise expect.  The research suggests that the $200 to $300 billion figure we’d need to give for the first year will decrease every year thereafter as the economies of entire regions and countries grow and lift their poorest residents out of extreme poverty.

What about Inflation?  As it turns out, the effect is minimal.  Average price inflation was only 0.1%, and no category of goods saw a price increase of more than 1.2%.  GiveDirectly thinks that price increases are so low because places with the most extreme poverty have a lot of economic “slack”, so local businesses and farms can easily grow to supply more goods to their new customers.

We know less about unconditional cash transfers at the largest scales, with millions of recipients.  The best examples of this kind of transfer so far might be economic stimulus programs run by governments, such as during the COVID-19 pandemic.  Several wealthy countries spent over 15% of their GDP on stimulus[24] and some such as the USA and Japan provided direct cash payments to their residents.  This of course has had many far reaching effects on the economy of these countries and the world, some of which remain to be seen, and which are hard to disentangle from the effects of the pandemic itself.  A cash transfer of $100 billion to the world’s poorest people would not have nearly as large a price tag as these multi-trillion dollar programs, and would not have as big an impact on world finance.  It would, however, cause a much larger improvement on the lives of the recipients, because they are the people most in need of it.  So while we’ll need to monitor the indirect effects carefully, it does seem like giving unconditional cash transfers to the world’s poorest people is a viable path for ending extreme poverty.

All of this shows that we need to look at ending global extreme poverty not as some distant unreachable dream, but as a feasible goal that we could accomplish with global coordination.  Cash transfers are only one way to approach the problem, but they are incredibly effective, and could be the most useful tool for radically reducing global poverty we’ve seen yet.  We can act, and we should.  Hundreds of millions of people’s lives and livelihoods are at stake.

If you're inspired to take action, we've started a fundraiser to send money directly to families in need and do our part to help accelerate the end of extreme poverty. You can donate directly through YouTube, or at  Every donation - even a small one - is impactful, as it helps GiveDirectly reach more families and demonstrate to policymakers that there's popular support for trusting people in poverty.  If every one of our subscribers gave just $1 today, that could mean reaching nearly 1,000 more people with the funds to help meet their needs and transform their lives.  To stay informed, you can sign up for GiveDirectly’s newsletter[25] or follow them on social media.  If you're not able or ready to give, you can help spread the word by sharing this video and encouraging others to take action.

  1. ^

  2. ^

  3. ^

  4. ^

  5. ^

  6. ^

  7. ^

  8. ^

  9. ^

  10. ^

  11. ^

  12. ^

  13. ^

  14. ^

  15. ^

  16. ^

  17. ^

  18. ^

  19. ^

  20. ^

  21. ^

  22. ^

  23. ^

  24. ^

  25. ^

New to LessWrong?

New Comment
5 comments, sorted by Click to highlight new comments since: Today at 10:08 AM

How effective is this compared to just investing in developing markets?

Searched for this on EA forums, but found nothing. Quite surprised that no one else has done an analysis yet. 

No idea, but intuitively, the money seems to flow better in some directions compared to others:

If you give money to homeless people, I would expect the profits of local supermarkets to grow.

If you give money to supermarkets, I would expect the situation of local homeless to remain the same.

Investment is an asset, but donations are just gone. Hence, you can put a lot more investments into specifically develping market funds than you can donations. 50% of disposable income into funds may well be both less painful and more effective than, say, 10% into donations.

I would say that a $5 loan towards a local super market has much more impact on local homelessness than a $1 direct handout, since that money is going to be spent in the local economy anyways.

It seems to me that poor people spend their money, rich people save them.

Technically, saving is investment, but it can be an investment on the opposite side of the planet, while spending is typically near. And the point is to make the money bubble from one poor person to another, not to end up somewhere in Silicon Valley after two or three jumps.

Also, from the point of "what if a dictator takes the money", taxing the supermarkets is probably easier than taxing the poor.

That said, ¿Por qué no los dos?


You might be interested in the debate here.