The UK has recently pledged in law to spend 0.7% of its gross national income on aid.  As a lawyer interested in effective altruism, I wanted to write about how the pledge could be vastly improved by having some sort of ‘effectiveness’ element to it. Ideally, the law shouldn’t just dictate the amount to be given in aid; it should also set aside a portion for only the most effective interventions.

First, however, I’d like to convey why cost-effectiveness is not just one more tick on a checklist of things donors want. If the effective altruism (EA) movement has persuaded me of anything, it is that cost-effectiveness may be the one of the most important considerations of all. For a fuller discussion of EA and the importance of cost-effectiveness, see here and here.


We are all familiar with cost effectiveness. We buy things all the time. Let’s say I want to buy some chocolate (a not unusual occurrence) and there are two sizes to choose from. The first is 100g and costs £1, and the second is 200g and also costs £1.

Hmm. If we make the optimistic assumption that there are no negative consequences to chocolate consumption, it is clear that I should buy the 200g. In fact, if I don’t buy the 200g bar, but choose the 100g instead, I have thrown away 50% of the chocolate I could have owned.

We can also think this way when buying stuff for other people. Perhaps my local school is asking for donations of books. I can buy 10 encyclopaedias for £100 from one bookstore, or I can buy 20 for £100 from a cheaper shop.

If I choose the less cost effective option, I’m not just wasting my own money any more; I am depriving school children of 10 books. I am throwing away 50% of the educational value I could have created.

I find type of counterfactual thinking the key to understanding cost effectiveness. The relevant question is always: What else could this investment have produced? Could I do more good investing in another place, for other recipients, or at another time?

Perhaps, for example, I shouldn’t have bought the books in the UK at all, but instead bought them for schools in Africa, where where £100 could buy over 350 books. And why stop with books? If my goal is to improve school experiences for children, there may be even better ways to help.

In his brilliant and highly-recommended new book Doing Good Better, philosopher William MacAskill compares the cost effectiveness of various charitable interventions.

He cites one study which showed that giving textbooks to schools in Africa actually had little impact on the attendance or performance of the children.  An initiative that deworms school age children, however, increased attendance by a huge 25%. The deworming programme proved one of the most cost-effective interventions in the world, with every £65 spent providing ten additional years attendance. Instead of buying 10 encyclopaedias for British schools, then, the same investment could have produced almost twenty additional years of school attendance in Africa. Differences in cost effectiveness can be enourmous.

The most staggering example of this I can remember was cited by philosopher Toby Ord in his paper The Moral Imperative toward Cost-Effectiveness in Global Health.

Suppose I have £30,000 to spend, and my aim is to fight blindness. I can use the amount to train a guide dog, which will help a blind person gain more independence and freedom from their disability. Alternatively,  I could also use the money to pay for operations to reverse the effects of trachoma in Africa. Trachoma is a treatable infectious disease that has blinded 1.2 million people worldwide. The operation costs less than £30 per patient.

Let’s simplify, and say these are the only two options I have with my £30,000 budget to fight blindness:

  1. Provide one blind person with a guide dog, or
  2. Cure 2,000 people of blindness

Setting out the options in this way makes it beyond obvious what is the right thing to do. If everyone counts equally, then it is 2,000 times better to fund the operations than the guide dog. To fund the guide dog would be a waste of about 99.95% of the value I could have produced.

Sometimes numbers like this are hard to picture . Imagine four Boeing 747 planes lined up on a runway, with every seat of every plane occupied. Choosing to fund the guide dog is choosing to help one passenger. Choosing to fund the surgery is choosing to help them all. And if we imagine that ourselves, or a friend or family member, could be in any position on that runway, it seems clear that cost-effective option is, by definition, the most compassionate.


So far we’ve only discussed cost-effectiveness without roughly the same area. But how can we compare interventions in radically different sectors? Is it possible to weigh up the benefits of micronutrient supplementation with improved access to the arts? It is better to provide an extra year of school to 10 children, or to prevent one person from contracting malaria?

It can feel wrong just to ask the question. The standard objection goes along the lines of “but who are you to say that preventing malaria is better than accessing art? They are totally different things, you can’t compare apples and oranges”.

But we compare apples and oranges all the time. If we make the assumption that when people decide what to do, they are trying to fulfill their preferences (here we can use the example of improving personal wellbeing) then we are in fact experts at assessing the cost-effectiveness of decisions across a huge variety of sectors.

There are, for example, lots of ways in which I could invest my money or time to fulfill my preferences: I could buy chocolate, vitamin pills, a gym membership, or tickets to the cinema or an art gallery. I could spend time hanging out with friends, working in the library, reading EA blogs online, or watching Netflix in my pyjamas.

Both my time and money are limited, so when I decide to do one of these things over another, my decision is based on an assessment of how much each activity will improve my wellbeing.

So when I buy chocolate on my way from a long day, I do so because I think it will improve my wellbeing more than if I buy more lentils. When I buy a gym membership instead of cinema tickets, I am deciding that the investment I make in encouraging myself to exercise will further my well-being more than going out to see a few movies. I am able to compare the benefits of radically different activities – apples and oranges – all by reference to a singular preference or set of preferences. In the end, it can all cash out in term of wellbeing.

Of course, many of the decisions we actually make are very far from optimal for our wellbeing: We are all plagued with cognitive biases, motivational problems, and the frustrating human inability to sacrifice short-terms pleasure in pursuit of longer terms goals (I’m thinking of you, gym membership).

However, the fact remains that cross-domain comparison is not only possible but is the norm for decision-making, as long as we have a preference or set of preferences to refer to. There is no fundamental reason, then, why we cannot compare the effectiveness of different charities across the broadest possible spectrum of interventions, with the goal of reducing as much suffering as possible while counting everyone equally.

There are already a range of tools to facilitate such comparisons, though their use is mainly restricted to assessing health outcomes. The most commonly used universal measurement is known as a Quality Adjusted Life Years (QALY’s). If an intervention provides perfect health for one additional year, it is said to provide one QALY.

In the UK, for example, encouraging doctors to urge patients to stop smoking costs around £270 per QALY, a kidney transplant costs around £4,710 per QALY, and neurosurgery for malignant intracranial tumours costs around £197,780 per QALY.

This universal wellbeing measuring-stick is not without difficulties, but it is an immensely useful tool for ensuring that resources are directed in the fairest way possible. The idea behind QALYs could ultimately be expanded to encompass all aspects of wellbeing, not simply health alone (Ben Todd helpfully proposes well-being adjusted life years, WALY’s).

So it seems not only can we compare apples and oranges, but in a world where everyone deserve their fair share of fruit, we may be obliged to.


Finally, how should an understanding of cost-effectiveness affect the way we practice altruism?

First – consider donating to a cost effective charity! Just the knowledge that some charities are over 100 times more effective than others can be motivating: You can do the same amount of good with a £20 donation than someone else might do with £2,000, so not having a huge amount of disposable income does not need to be too disheartening.

GiveWell, a nonprofit charity evaluator, uses rigorous scientific analysis to recommend charities that save or improve lives as much as possible. The top rated-charities for 2015 are listed here. Animal Charity Evaluators does the same for animals, and their top-rated charities are here.

Second – perhaps consider changing the way you think about altruism. A few years ago, I spent some weeks as a fundraising caller for my university. My job was to call up alumni and ask them for for monthly donations to fund projects such as new university buildings, outreach, and hardship funds for students.

Around 70% of people I called chose not to donate, and their reasoning ranged from “Drop dead and never contact me again!”, to “Dave is out and will never be back home”. There was one objection, however, that made me pause. Sometimes people would tell me, “I already donate what I can to the Against Malaria Foundation, should really I switch and donate to an Oxford college?!”. I couldn’t, in good faith, tell them that they should. And so I didn’t.

Universities such as Oxford, however, routinely receive massive donations. Last year, higher education institutions in the UK alone received £485 million in donations.  In 2012, venture capitalist Michael Moritz gave Oxford University £75m to help it attract more students from poorer backgrounds. The donation was described as “the biggest philanthropic gift for undergraduate financial support in European history”.

Moritz’s generous donation will undoubtedly do an enormous amount of good. Those from wealthy backgrounds continue to have privileged access to top universities, and there is much to be gained from a more representative section of the public having a voice among elite decision-makers.

The same amount of money, however, could have provided almost 18 million insecticide bednets or 90 million schistosomiasis treatments – the equivalent of saving over 34,000 lives. This is almost double the number of students currently studying at Oxford University. The money could have also saved the lives of over 5 billion animals – ten times as many as all the cats and dogs kept as pets worldwide.

This is not to suggest that Moritz should obviously have donated elsewhere; scholarships supporting systemic change among developed nations may well be very effective in the long term, especially if shorter-term interventions such as insecticide-treated bednets become less chronically under-funded. It may well be better for large donors to fund research rather than direct interventions.

It is also counter-productive to condemn any less-than-maximally effective donation decision; we are all motivated to contribute to societies we associate with, and doing any amount of good is infinitely better than doing none at all.

I intend only to point out that a cost-effectiveness mindset challenges unquestioning and unreflective praise of of any type of charitable giving. If every individual matters equally, and some interventions are 100 times more effective than others, we should perhaps reserve the highest of our praise for those philanthropists who attempt to help as many as possible, by as much as possible. And in a new age of “super-donors”, such as Mark Zuckerberg and his $45 billion philanthropy initiative, this need for critical analysis may be more important than ever.