Doing Good Better: Effective Altruism and a Radical New Way to Make a Difference

by William MacAskill

It’s the size of the drop that matters, not the size of the bucket, and, if we choose, we can create an enormous splash. We’ve already seen that we have the opportunity to provide a benefit for others that is one hundred times greater than the benefit we could provide for ourselves.

That we can’t solve all the problems in the world doesn’t alter in any way the fact that, if we choose, we can transform...

Key question 1: How many people benefit, and by how much?

James Orbinski

Under this regime, the Tutsi assisted the colonial rulers while Hutu were used as forced labour.

By the time Orbinski found himself at that Red Cross hospital, hundreds of thousands of Tutsi had been killed. The UN was stalling, not wanting to admit that a genocide was in progress, and had provided almost no support. Only a handful of non-profit workers remained in the country. Later in his life, Orbinski would become the president of Médecins Sans Frontières and accept the Nobel Peace Prize on its behalf, but at this time his role was simply to provide care for those who needed it, and with so many casualties, what could he do? He later recalled:

The reality of our world is such that, if we want to make the world a better place, we must make choices similar to those Orbinski had to make.

Effective altruism, at its core, is about confronting Orbinski’s dilemma and trying our best to make hard trade-offs. Of all the ways in which we could make the world a better place, which will do the most good? Which problems should we tackle immediately, and which should we leave for another time? Valuing one action over another is difficult both psychologically and practically, but it is not impossible. In order to make comparisons between actions, we need to ask: how many people benefit, and by how much? This is the first key question of effective altruism.

To begin to answer this question, we need to know the consequences of our actions.

This is such a typical state of affairs that you might not have thought about how astonishing it is. But imagine going into a shop where none of the products had prices on them. Instead, the proprietor asks: ‘How much would you like to spend today?’ You hand over some money and receive in return a selection of goods chosen by an assistant.

This, of course, is absurd. If this was how things worked, how could we ever do our shopping? One store could charge ten times the amount for the same product and we wouldn’t be able to tell until after we’d paid.

We can overcome this problem by thinking in terms of improving lives, rather than in terms of intermediate metrics like number of schoolbooks provided.

In order to truly make comparisons between different actions, we need to measure impact in terms of the size of the benefits we confer through those actions.

However, there are many harder cases: if you can prevent the death of a five-year-old or a twenty-year-old, which should you do? What if you can prevent ten people from suffering from AIDS or one hundred people from suffering from severe arthritis? What about preventing one woman from being domestically abused versus enabling one child to go to school?

For health benefits, economists have spent decades conducting research in order to answer questions like these. They have developed a metric called the quality-adjusted life-year, or QALY (pronounced ‘kwalee’) in order to help make decisions about how to prioritise among different health programmes.

‘save someone’s life’.

improve the quality of their life

The QALY combines these two benefits into one metric, using survey data about the trade-offs people are willing to make in order to assess how bad different sorts of illnesses or disabilities are. For example, on average people rate a life with untreated AIDS as 50% as good as life at full health; people on average rate life after a stroke as 75% as good as life at full health; and people on average rate life with moderate depression as only 30% as good as life in full health.

In the endnotes, I link to some official lists of quality-of-life estimates, to help you assess the severity of different conditions yourself.

QALYs are imperfect, as any measure of health benefit will be. For example, people who have never been on dialysis typically estimate that, if they were on dialysis, their health-related quality of life would be 39%, whereas people who actually are on dialysis on average rate their health-related quality of life at 56%. The same is true for other medical conditions: patients tend to regard their conditions as less bad than non-sufferers do. Is this because the general public doesn’t really understand what life with the medical condition is like and overestimates how bad it is?

In this book I’ll talk about QALYs quite a lot. That’s not because I think the only way to make a difference is to improve someone’s health. Rather, it’s because, as I’ll explain in the coming chapters, many of the best, most concrete and easy-to-measure ways of doing good involve improving global health. We’ve also got much better data for health programmes than for many other sorts of activity. Since the goal of effective altruism is to do the most good we can, health is a good place to start.

Moreover, in principle the same methods that were used to create the QALY could be used to measure the costs and benefits of pretty much anything. We could use these methods to estimate the degree to which your wellbeing is affected by stubbing your toe, or by going through a divorce, or by losing your job. We could call them ‘well-being-adjusted life-years’ instead. (Unfortunately, the acronym for this would be ‘WALY’, which might explain why it hasn’t yet taken off.) The idea would be that being dead is at 0% wellbeing; being as well off as you realistically can be is at 100% wellbeing. You can compare the impact of different activities in terms of how much and for how long they increase people’s wellbeing.

Suppose, hypothetically, we found that providing one guide dog (at a cost of $50,000) would give a 10% increase in reported wellbeing for one person’s life over nine years (the working life of the dog). That would be 0.9 WALYs. And suppose that providing 5,000 books (at a cost of $50,000) provided a 0.001% increase in quality of life for 500 people for forty years. That would be 2 WALYs. If we knew this, then we’d know that spending $50,000 on schoolbooks provided a greater benefit than spending $50,000 on one guide dog.

The difficulty of comparing different sorts of altruistic activity is therefore ultimately due to a lack of knowledge about what will happen as a result of that activity, or a lack of knowledge about how different activities translate into improvements to people’s lives. It’s not that different sorts of benefits are in principle incomparable.

Ken Berger,

Robert M. Penna

However, their view surely cannot be correct. If Berger and Penna were right then we couldn’t say that giving someone an extra dessert is a smaller benefit than saving someone’s life. Nor could we say that you do more good by saving a million lives than by saving ten. We would have to conclude that nurses who engage in triage – ensuring that doctors don’t spend their time treating mild coughs when they could be treating heart attacks – have no basis for their decisions. But that would be absurd. It might be difficult, both emotionally and practically, to weigh different people’s interests against each other, but it’s not impossible either in principle or, I believe, in practice.

But there was a problem. When I’d been in Ethiopia several years before, I’d visited this hospital. I’d hugged the women who suffered from this condition, and they’d thanked me for visiting them. It had been an important experience for me: a vivid first-hand demonstration of the severity of the problems in the world. This was a cause I had a personal connection with. Should I have donated to the Fistula Foundation, even knowing I could do more to help people if I donated elsewhere? I do not think so. If I were to give to the Fistula Foundation rather than to charities I thought were more effective, I would be privileging the needs of some people over others for emotional rather than moral reasons.

When making decisions, whether that’s where to volunteer, what career to pursue, or whether to buy ‘ethical’ produce, we should therefore consider (1) the cost of the activity, in terms of time or money, (2) the number of people it benefits and, crucially, (3) how much it improves people’s lives.

Key question 2: Is this the most effective thing you can do?

Dambisa Moyo,

William Easterly,

I endorsed something like aid scepticism for quite a long time. After I graduated from college, I decided against applying for jobs at non-profits partly because of stories I’d heard about food aid being stolen and sold by corrupt governments, and assumed there was no way I could have an impact under these conditions. I donated to development charities, but I always felt uneasy about whether I was actually helping others or merely alleviating my own sense of guilt about being born privileged in a world with so much need.

But to get a true picture of how much benefit the developing world has received from aid, one needs to focus instead on the best aid programmes.

A good contender for the best aid programme ever is the eradication of smallpox. Smallpox was a horrific disease.

It’s difficult to comprehend just how great an achievement this was, so let’s make a comparison. Suppose we’d achieved world peace in 1973. How many deaths would have been prevented? That timescale includes the killings by Cambodia’s Khmer Rouge, the Rwandan genocide, the two Congo wars, the 9/11 attacks and the wars in Afghanistan and Iraq. If you add up all the wars, genocides and terrorist acts that have occurred since 1973, the death toll is a staggering 12 million. Prior to its eradication, smallpox killed 1.5 to 3 million people every year, so by preventing these deaths for over forty years, its eradication has effectively saved somewhere between 60 and 120 million lives. The eradication of smallpox is one success story from aid, saving five times as many lives as world peace would have done.

In comparison, government departments in the US will pay for infrastructure to improve safety if doing so costs less than about $7 million per life saved:

This calculation not only shows that aid has ‘worked’, but also that it’s been cost-effective on average. Moreover, this calculation significantly underestimates the positive impact of aid. Thanks to immunisation, annual deaths from preventable illnesses have declined from 5 million in 1960 to 1.4 million in 2001, despite world population doubling in that time. Annual malaria deaths have declined from 3.8 million to about 0.7 million. Annual diarrhoeal deaths have declined from 4.6 million to 1.6 million. Aid isn’t responsible for all these reductions, but it is responsible for a proportion of them, and that’s despite the fact that the amount of money spent on aid is tiny in global terms.

We’re used to thinking of what’s typical and what’s average as being the same. For example if you measured the height of all women in North America and plotted those heights on a graph, the result would be something like Figure 7. The height of a typical woman in North America (that is, a woman who is taller than 50% of people and shorter than 50% of people) is 5 feet 5 inches; the height of the average woman (which is equal to the total height of all women divided by the number of women) is also 5 feet 5 inches. In the case of height, what is typical and what is average is the same. This sort of distribution is what we’re most familiar with, so it’s aptly called a normal distribution.

Distributions that look like this are called fat-tailed distributions. (You might have heard of the ‘80/20’ rule: that 80% of the value of an entire set of activities can be achieved by performing the best 20% of those activities. That rule describes a fat-tailed distribution.) Fat-tailed distributions are interesting because they’re marked by extreme events. Whereas there are very few extremely small or extremely tall people, there are a relatively large number of extremely rich people. (If height were distributed like income is, we would regularly see people towering 270 feet tall, peering over skyscrapers.) That’s why the world’s average income, which is $10,000 per year, is so much higher than the typical income, which is only $1,400 per year: the richest people bring up the average.

because the best programmes are so good, they make aid very effective on average.

In the context of helping others, the difference between a good use of money and a great use of money is huge. We shouldn’t just ask: ‘Is this programme a good use of money?’ We need to ask: ‘Is this programme the best use of money?’

The same phenomenon occurs with respect to developing-world health. Figure 10 lists the estimated cost-effectiveness of different health programmes, measured in QALYs (where one QALY, remember, represents a benefit equivalent to giving one person one year of life in perfect health).

Imagine saving a single person’s life: you pass a burning building, kick the door down, rush through the smoke and flames and drag a young child to safety. If you did that, it would stay with you for the rest of your life. If you saved several people’s lives – running into a burning building one week, rescuing someone from drowning the next week, and diving in front of a bullet the week after – you’d think your life was really special. You’d be in the news. You’d be a hero. But we can do far more than that.

According to the most rigorous estimates, the cost to save a life in the developing world is about $3,400 (or $100 for one QALY). $3,400 is a small enough amount that most of us in affluent countries could donate that amount every year while maintaining about the same quality of life. Rather than just saving one life, we could save a life every working year of our lives. Donating to charity is not nearly as glamorous as kicking down the door of a burning building, but the benefits are just as great. Through the simple act of donating to the most effective charities, we have the power to save dozens of lives. That’s pretty amazing.

Indeed, medicine is the banner career for people who want to make a difference. Every year about 20,000 people in the US and 8,000 people in the UK go to medical school, and the number is growing year on year.

Greg Lewis.

As a result of that research, he developed a different view on how he could best make a difference in the world. To explain the reasoning behind that view, we need to look at the third of effective altruism’s key questions: Is this area neglected?

This ‘water and diamonds’ paradox shows the importance of what economists call thinking at the margin: assessing the value of an additional thing – what is known in economics as its marginal utility – rather than thinking about the average value of that thing.

The first slice of cake is delicious, but by the third, you’re feeling a little sick. Having one copy of this book might provide you with an interesting and entertaining experience, but having a second might just provide you with a makeshift doorstop. This is what economists call the law of diminishing returns.