Doing Good Better: On effective and ineffective altruism

Will MacAskill’s Doing Good Better was one of the first books to explain effective altruism to a broad audience.

This excerpt (used with Will’s permission) opens the book, and tells a story about why it can be so important to evaluate the effectiveness of charitable interventions — even if they seem very promising on the surface.

The version of this excerpt featured in the EA Handbook has been lightly edited. You can find the original in the book.


Until 1989, Trevor Field was a typical middle-aged South African man who had lived a fairly normal life. He enjoyed fresh steaks, cold beer, and fishing with his friends. Working in advertising for magazines like Top Car and Penthouse, he had never thought seriously about using his skills for the greater good. When he discovered the PlayPump, however, everything changed.

That year, Field and his father-in-law, a farmer, visited an agricultural fair in Pretoria. There, he met a water engineer named Ronnie Stuiver, who was demonstrating a model for a new type of water pump. The demonstration reminded Field of a fishing trip he’d taken years before, during which he had watched the women of a rural village wait for hours next to a windmill-powered water pump. There had been no wind that day, but the women, who had trekked for miles, still needed to bring water back to their homes. So they simply sat and waited for the water to flow.

Field had been struck by how unfair this was. There simply must be a better way to do this, he thought. Now he was witnessing a potential solution. Stuiver’s invention seemed brilliant. Instead of the typical hand pump or windmill pump found in many villages in poor countries, Stuiver’s pump doubled as a playground merry-go-round. Children would play on the merry-go-round, which, as it spun, would pump clean water from deep underground up to a storage tank. No longer would the women of the village need to walk miles to draw water using a hand pump or wait in line at a windmill-powered pump on a still day. The PlayPump, as it was called, utilized the power of playing children to provide a sustainable water supply for the community.

“African [kids] have almost nothing — not even books in school, let alone playground equipment — and access to water is a huge problem,” Field later told me. “I thought it was just the best idea I’ve ever seen.”

Field bought the patent from Stuiver and worked in his spare time over the next five years to improve the design. Using his experience in advertising, Field came up with the idea of placing billboard advertisements on the sides of the water tank as a way to generate revenue to pay for pump maintenance. In 1995, he secured his first sponsor, Colgate Palmolive; installed the first PlayPump; and quit his job in order to focus full-time on the project, now a registered charity called PlayPumps International.

Progress was slow at first, but he persevered, paying for several pumps with his own money. At the same time, he developed connections with corporations and government bodies across South Africa to pay for more pumps. By the turn of the millennium, he had installed fifty pumps across the country. His first major breakthrough came in 2000 when, out of three thousand applicants, he won a World Bank Development Marketplace Award, given to “innovative, early-stage development projects that are scalable and/​or replicable, while also having high potential for development impact.” That award attracted funding and attention, which culminated in a site visit from Steve Case, CEO of AOL, and his wife, Jean.

“They thought the PlayPump was incredible,” Field said. “As soon as they saw it in action, they were sold.”

In 2005, the Cases agreed to fund the project and worked with Field to set up an American arm of PlayPumps International. Their aim was to roll out thousands of new PlayPumps across Africa. The PlayPump became the center of a massive marketing campaign. Steve Case used his expertise from running AOL to pioneer new forms of online fund-raising. The One Foundation, a British fund-raising charity, launched a bottled water brand called One Water and donated the profits to PlayPumps International. It was a huge success and became the official bottled water of the Live 8 concerts and the Make Poverty History campaign. The PlayPump became the darling of international media, who leaped at the opportunity to pun, with headlines like “PUMPING WATER IS CHILD’S PLAY and THE MAGIC ROUNDABOUT. In an article for Time in 2006, Bill Clinton called the PlayPump a “wonderful innovation.”

Celebrities, too, jumped on the bandwagon. Jay-Z raised tens of thousands of dollars through his “The Diary of Jay- Z: Water for Life” concert tour. Soon after, PlayPumps International secured its biggest win: a $16.4 million grant awarded by then First Lady Laura Bush, launching a campaign designed to raise $60 million to fund four thousand PlayPumps across Africa by 2010.

By 2007, the PlayPump was the hottest thing in international development, and Trevor Field was at the center of it all — a rock star of the charity world. “It has just gone berserk! … When I first looked at this water pump … I could never imagine that this is something that could possibly change the world,” Field said in 2008, reflecting on PlayPump International’s startling success. “It really rocks me to know we’re making a difference to a lot of people who are nowhere near as privileged as I am or my family is.” By 2009, his charity had installed eighteen hundred PlayPumps across South Africa, Mozambique, Swaziland, and Zambia.

Then things went sour. Two damning reports were released, one by UNICEF and one by the Swiss Resource Centre and Consultancies for Development (SKAT). It turned out that, despite the hype and the awards and the millions of dollars spent, no one had really considered the practicalities of the PlayPump. Most playground merry-go- rounds spin freely once they’ve gained sufficient momentum — that’s what makes them fun. But in order to pump water, PlayPumps need constant force, and children playing on them would quickly get exhausted. According to the UNICEF report, children sometimes fell off and broke limbs, and some vomited from the spinning. In one village, local children were paid to “play” on the pump. Much of the time, women of the village ended up pushing the merry-go round themselves — a task they found tiring, undignified, and demeaning.

What’s more, no one had asked the local communities if they wanted a PlayPump in the first place. When the investigators from SKAT asked the community what they thought about the new PlayPump, many said they preferred the hand pumps that were previously installed. With less effort, a Zimbabwe Bush hand pump of the same cylinder size as a PlayPump provided thirteen hundred liters of water per hour — five times the amount of the PlayPump. A woman in Mozambique said, “From five A.M., we are in the fields, working for six hours. Then we come to this pump and have to turn it. From this, your arms start to hurt. The old hand pump was much easier.” One reporter estimated that, in order to provide a typical village’s water needs, the merry-go-round would have to spin for twenty-seven hours per day.

Even when communities welcomed the pumps, they didn’t do so for long. The pumps often broke down within months, but unlike the Zimbabwe Bush Pump, the mechanics of the pump were encased in a metal shell and could not be repaired by the community. The locals were supposed to receive a phone number to call for maintenance, but most communities never received one, and those who did never got anyone on the phone. The billboards on the storage tanks lay bare; the rural communities were too poor for companies to be interested in paying for advertising. The PlayPump was inferior in almost every way to the unsexy but functional hand pumps it competed with. Yet, at $14,000 per unit, it cost four times as much.

Soon, the media turned on its golden child. PBS ran a documentary exposing the PlayPump’s many shortcomings. (One thing that didn’t change was the media’s love of puns: the documentary was called Southern Africa: Troubled Water; The Guardian repeatedly referred to the PlayPump as “money down the drain.”) In an admirable response to this criticism, the US arm of PlayPumps International shut down and its sponsor, the Case Foundation, publicly acknowledged that the program had been a failure. Yet, despite its fall from grace, the PlayPump lives on. Under the name of Roundabout Water Solutions, Field’s nonprofit continues to install the same model of PlayPumps across South Africa, funded by corporations like Ford Motor Company and Colgate Palmolive.


Most people want to make a difference in their lives; if you’re reading this book, you’re probably no exception. As Trevor Field’s story illustrates, however, good intentions can all too easily lead to bad outcomes. The challenge for us is this: How can we ensure that, when we try to help others, we do so as effectively as possible? How can we ensure that we avoid accidentally causing harm, and instead succeed in having the greatest positive impact we can?

In 2007, at the peak of the PlayPump’s popularity, Michael Kremer and Rachel Glennerster launched an organization of their own, the culmination of decades of research into how to improve the lives of the poorest people in the world.

Glennerster had studied economics at the University of Oxford, graduating in 1988. She was interested in learning about poverty relief, so she decided to live in a developing country and spent a summer in Kenya. She spoke to people working in development, many of whom were deeply disillusioned. When she asked why, they told her to look at some of the ways development projects had backfired.

“I got sent down to big projects that had failed,” Glennerster told me. “I went to Lake Turkana, up in the north of Kenya. The Turkana people are basically nomadic, and various development projects had hoped to improve their quality of life by settling them on the lake, so they built a big factory for fish. They managed to get them to settle and fish in the lake, but then the lake got overfished, and the fish stock collapsed … It was depressing.” Disenchanted about the potential to have an impact in global development, she moved into domestic policy, taking a job at the British Treasury.

Michael Kremer also spent some of his young adulthood in Kenya, living there for a year after finishing his undergraduate degree. Like Glennerster, he was concerned by extreme poverty and wanted to learn more, so he lived with a local family, teaching English at a secondary school. He also saw some dramatic ways in which attempts to improve conditions there were failing. When he returned to grad school, he decided to figure out how things could be done better.

Kremer and Glennerster met at Harvard University in 1990. Kremer was a PhD student; Glennerster was visiting on a Kennedy Scholarship, having taken a sabbatical from her work at the Treasury. By the time Kremer became a professor at MIT in 1993, he and Glennerster were married. As a vacation, they returned to Kenya to visit the family Kremer had lived with several years prior. While there, Kremer spoke to Paul Lipeyah, a friend who worked for the Dutch charity International Christian Support (now called Investing in Children and Their Societies, or ICS).

ICS’s main program was child sponsorship, in which a donor paid a regular amount to help an individual child or a small community. ICS had been trying to improve school attendance and test scores. They provided a package of different things: new textbooks and additional teachers to schools, and free school uniforms to the students. ICS had received new funding, and Paul Lipeyah was about to roll out the program to seven new schools. Kremer urged Lipeyah to test his program using what’s called a randomized controlled trial: he would monitor and collect data for fourteen local schools, implementing the program in seven of them, while letting the other seven go about business as usual. By collecting data from all fourteen schools to see which fared better, he could find out if his program actually worked.

In hindsight, Kremer’s idea seems obvious. Randomized controlled trials are the gold-standard method of testing ideas in other sciences, and for decades pharmaceutical companies have used them to test new drugs. In fact, because it’s so important not to sell people ineffective or harmful drugs, it’s illegal to market a drug that hasn’t gone through extensive randomized controlled trials. But before Kremer suggested it, the idea had never been applied to the development world.

With the help of collaborators, Kremer tested the different ICS programs one by one. First, he looked at the efficacy of providing schools with additional textbooks. Classrooms would often have only one textbook for a class of thirty, so it seemed obvious that providing more textbooks would help students learn. However, when Kremer tested this theory by comparing test scores between schools that received books and those that didn’t, he found no effect for all but the most high-achieving of students. (He suggests the textbooks were written at too high a level for the children, especially considering they were in English, the pupils’ third language after Swahili and their local languages.)

Next, Kremer looked at providing flip charts. The schoolchildren couldn’t understand the textbooks, but having flip charts would allow teachers to tailor lessons to the specific needs of the students. Perhaps these would work better. Again, however, no effect.

Undaunted, he took a different approach. If providing additional materials didn’t work, maybe increasing the number of teachers would. After all, most schools had only one teacher, catering to a large class. But, again, he found no discernible improvement from decreasing class sizes.

Over and over again, Kremer found that seemingly obvious programs to improve education just weren’t working. But he persisted. He refused to believe there was simply no way to improve the education of children in Kenya. At that point, a friend at the World Bank suggested he test deworming. Few people in developed countries know about intestinal worms: parasitic infections that affect more than one billion people worldwide. They aren’t as dramatic as AIDS or cancer or malaria, because they don’t kill nearly as many people as those other conditions. But they do make children sick, and can be cured for pennies: off-patent drugs, developed in the 1950s, can be distributed through schools and administered by teachers, and will cure children of intestinal worms for a year.

Kremer did an experiment to see whether treating children for these intestinal worms had an impact on education. The results were striking. “We didn’t expect deworming to be as effective as it was,” Kremer told me. “It turned out to be one of the most cost-effective ways of increasing school participation.” Absenteeism is a chronic problem in schools in Kenya, and deworming reduced it by 25 percent. In fact, every child treated spent an extra two weeks in school, and every one hundred dollars spent on the program provided a total of ten years of additional school attendance among all students. Enabling a child to spend an extra day in school therefore cost just five cents. It wasn’t merely that deworming children “worked” at getting children into school. It worked incredibly well.

What’s more, deworming didn’t merely have educational benefits. It had health and economic benefits, too. Intestinal worms can cause a variety of maladies, including anemia, intestinal obstruction, and a suppressed immune system that can increase the risk of other diseases like malaria. Deworming decreases all these risks. Moreover, when Kremer’s colleagues followed up with the children ten years later, those who had been dewormed were working an extra 3-4 hours per week and earning an extra 20 percent of income compared to those who had not been dewormed. In fact, deworming was such a powerful program that it paid for itself through increased tax revenue.

By the time his work on deworming was published, Kremer’s revolutionary new approach to development had spawned a following, with dozens of the brightest young economists running hundreds of trials of different development programs. Meanwhile, Glennerster had quit her job and become the executive director of the newly founded Poverty Action Lab at MIT, where she used her knowledge of policy to ensure the research Kremer and his colleagues were conducting would have real-world impact. In 2007, on the basis of this research, Kremer and Glennerster co-founded the nonprofit Deworm the World Initiative, which provides technical assistance to the governments of developing countries, enabling them to launch their own deworming programs. The charity has provided more than forty million deworming treatments, and the independent charity evaluator GiveWell regards them as one of the most cost-effective development charities.


When it comes to helping others, being unreflective often means being ineffective. The PlayPump is the perfect example. Trevor Field and everyone who supported him were driven by emotions — the appeal of seeing happy children provide their communities with clean water through the simple act of playing — rather than facts. The Case Foundation, Laura Bush, and Bill Clinton supported the PlayPump not because there was good evidence to believe it would help people but because it had the thrill of a revolutionary technology. Even critics of the campaign would stop short of accusing Field and his supporters of bad intentions — they no doubt genuinely wanted to help the people of rural Africa. But relying on good intentions alone to inform your decisions is potentially disastrous.

It would be nice if the PlayPump were an isolated example of unreflective altruism, but sadly it’s just an extreme example of a much more general trend. We very often fail to think as carefully about helping others as we could, mistakenly believing that applying data and rationality to a charitable endeavor robs the act of virtue. And that means we pass up opportunities to make a tremendous difference.

No comments.