Sunday 11 April 2021

Effective Altruism Is Not Effective

Effective altruism is based on a very simple idea: we should do the most good we can. Obeying the usual rules about not stealing, cheating, hurting, and killing is not enough, or at least not enough for those of us who have the good fortune to live in material comfort, who can feed, house, and clothe ourselves and our families and still have money or time to spare. Living a minimally acceptable ethical life involves using a substantial part of our spare resources to make the world a better place. Living a fully ethical life involves doing the most good we can. (Peter Singer)

It is almost universally agreed that the persistence of extreme poverty in many parts of the world is a bad thing. It is less well-agreed, even among philosophers, what should be done about it and by who. An influential movement founded by the philosopher Peter Singer argues that we should each try to do the best we can by donating our surplus income to charities that help those in greatest need. This ‘effective altruism’ movement has two components: i) encouraging individuals in the rich world to donate more; and ii) encouraging us to donate more rationally, to the organisations most efficient at translating those donations into gains in human well-being.

Unfortunately both components of effective altruism focus on what makes giving good rather than on achieving valuable goals. Effective altruism therefore does not actually aim at the elimination of global poverty as is often supposed. Indeed, its distinctive commitment to the logic of individualist consumerism makes it constitutionally incapable of achieving such a large scale project. Effective altruism is designed to fail.

I. The No-Sacrifice Principle of Giving

In his best-selling defense of effective altruism The Life You Can Save: Acting Now to End World Poverty (2009, p.15) Singer provides this outline of his argument.

First premise: Suffering and death from lack of food, shelter, and medical care are bad.

Second premise: If it is in your power to prevent something bad from happening, without sacrificing anything nearly as important, it is wrong not to do so.

Third premise: By donating to aid agencies, you can prevent suffering and death from lack of food, shelter, and medical care, without sacrificing anything nearly as important.

Conclusion: Therefore, if you do not donate to aid agencies, you are doing something wrong.

Singer famously supports his second premise by reference to his ‘shallow pond’ thought experiment, in which nearly everyone agrees that we would have an obligation to rescue a drowning child even at some personal inconvenience. He argues that since we already seem to accept that principle, the moral challenge is to integrate it better into how we live by donating some of our ‘surplus’ income to charities. Effective altruism is thereby identified as a way of living better in accordance with reason and right, the correct answer to Socrates' challenge 'How ought we to live?'

What I want to bring out here is that Singer’s main concern is the question of how good to be in terms of how much we should be giving, i.e. the internal moral economy of the subject. The ‘bads’ of suffering and death identified in premise 1 are peripheral to this analysis. They may motivate our interest in altruism but their remediation is not the measure of our altruistic success.

On the face of it, premise 2 is a very demanding principle because it links our subjective moral economy to the prevention of significant objective harms. However, the way that Singer uses the principle severs that relation. Singer is concerned to help us calculate our personal budget for good works: how much we each can spare from our other interests and commitments. As Singer makes clear, altruism on this conception should not feel like a sacrifice because it is merely the harmonious integration of our moral with our other preferences. This generates a rather generic analysis of how much it is reasonable to expect people of different levels of affluence to contribute to good causes without having to make any real sacrifices, i.e. calculations of how much money we could easily do without. (Singer suggests a progressive rate of voluntary self-taxation starting at 5% of income for those earning more than $100,000.)

Such calculations are generic because they are fundamentally concerned with how to be an altruist, not with how to fix the world’s problems, and so they are unrelated to the significance of the specific problems our donations are supposed to address, nor with what would be needed to successfully solve them. For consider, even if global poverty were eliminated entirely, there will still always be causes you could contribute to that would be more valuable than pursuing your own interests (such as generating benefits to future generations). This is the paradoxical overdemandingness of utilitarianism identified by various philosophers (Bernard Williams; Susan Wolf; etc): that a world of utilitarians would be a world incapable of happiness. What I think Singer’s ‘no sacrifice’ principle actually offers is a (not especially convincing) way to reconcile our moral duty to doing good with our right to live a life of our own.  We are effectively asked to calculate our own voluntary moral tax-rate that delineates when we have done enough for others and can turn away, morally free to pursue our private projects and commitments. How much good this amount of giving will achieve in the world is irrelevant to what that tax rate should be.

II. Efficiency is Not the Same Thing as Effectiveness

Effective altruists ..... know that saving a life is better than making a wish come true and that saving three lives is better than saving one. So they don’t give to whatever cause tugs strongest at their heartstrings. They give to the cause that will do the most good, given the abilities, time, and money they have. (Peter Singer)

The problem with the first component of effective altruism was that it focuses on the internal moral economy of the giver rather than on the real world problems our giving is supposed to address. The second component of effective altruism might not seem to have that problem because it is explicitly concerned with maximising the amount of good that each unit of resources achieves. (This is also the component that has received more emphasis in the last 10 years as the movement gained traction among a younger generation of philosophers such as Toby Ord and William MacAskill.) However, this concern is better understood as efficiency than as effectiveness (the general idea of getting things done). This might seem an innocuous distinction since efficiency is about how we ought to get things done, i.e. a way of being effective. However, there are significant consequences for practical reasoning in the kind of cases effective altruism is concerned with.

If one takes the efficiency view promoted by the effective altruism movement then one assumes a fixed set of resources and the choice of which goal to aim for follows from a calculation of how to maximise the expected value those resources can generate; i.e. the means justifies the end. For example, in the context of global poverty, you would use evidence and careful reasoning to decide in which cause or organisation to invest your chosen amount on the basis of which generates  the most QALYS per dollar. This should ensure that your donation will achieve the most good, which is to say that you have done the best possible job of giving. However, despite doing so well at the task effective altruism has set you, if you step back you will notice that very little has actually been achieved. The total amount of good we can achieve with our donations is limited to the partial alleviation of some of the symptoms of extreme poverty, symptoms that will recur so long as poverty persists. But effective altruism supplies no plan for the elimination of poverty itself, and there is no way for a feasible plan for that goal to be developed and implemented by this method of reasoning at the margin.

The underlying problem is that effective altruism's distinctive combination of political pessimism and consumer-hero hubris forecloses the consideration of promising possibilities for achieving far more good.  First, effective altruists advance an ungrounded pessimism about political action that cuts off that obvious avenue for solving large scale problems. For example, in The Life You Can Save, Singer briefly discusses “whether it would be better to spend our time and money campaigning to eliminate trade barriers, rather than donating to agencies that give aid to the poor” before concluding sadly that “The powerful political interests allied against the elimination of trade barriers make political change unlikely.….Defeats like this suggest that our efforts are better spent elsewhere, where we can be confident of making a difference.” (p. 114)

The most charitable explanation of Singer’s dismissal of political action is that he is trying to sell being an altruist and he thinks a consumer -hero version is the one people are most likely to buy.  Singer and other effective altruist philosophers believe that their most likely customers find institutional reform too complicated and political action too impersonal and hit and miss to be attractive. So instead they flatter us by promising that we can literally be life-saving heroes from the comfort of our chairs and using only the super-power of our rich-world wallets. A big part of making this work is to make us feel in control of what happens to people around the world, and this in turn requires simplifying and personalising the logistical chain between action and outcome. Even if we end up achieving less good in total, we can be confident that the good that was achieved was due specifically to us - that we made a difference - and this may be a powerful psychological motivation for many people in our individualistic consumer societies and therefore result in more donations and more good achieved. Yet, however rhetorically powerful such a tactic seems, I think it represents a failure of consequentialist strategy. Firstly, it just doesn't work. Singer and others have been making this argument for nearly 50 years, yet the level of private donations remain orders of magnitude below what would be required to eliminate global poverty, however efficiently allocated. Secondly, it needlessly squanders the most obvious and powerful tool we have: the political sphere and institutions of government that we invented to solve complicated and large collective action problems.

Singer’s great contribution to the academic and non-academic conversation about global justice was his forceful case for helping people suffering far away just as we would help those in front of us. With the help of the shallow pond thought experiment Singer brought many people to recognise that distance is irrelevant to whether you should help someone. We should not make a distinction between duty over here and mere optional charity over there. However, it seems to me that Singer fails to integrate this elementary point fully into effective altruism. When we identify a problem within our society that we think we should be solved - such as children not getting a proper education or unsafe roads - we do not merely pass a hat around for people to put as much money in as they individually feel they can without sacrificing anything important. Instead we organise ourselves as citizens to place this on the political agenda so as to get the improvements we want funded from compulsory (progressive) taxation and to have specific institutions and individuals assigned responsibility, authority, and accountability for implementing them. If this is the way we generally and mostly quite successfully approach the large scale problems we face in our society, then I do not see why – and effective altruists like Singer provide no good reason why - we should not do the same in the case of the problem of global poverty. We should not begin by asking ‘What should I do about global poverty?’ but ‘What should we do about it?’.

There are practical limitations to rejecting the political sphere and apparatus of government in favour of appealing to the conscience of individuals as consumers which make it unlikely to do much about global poverty. Singer advises affluent individuals who decide they can spare a few percent from their incomes to invest it where it will do the most good, using a charity comparison shopping service like GiveWell. Singer is appealing here to the idea that the rational consumer should allocate each dollar they spend to where it is expected to generate the most value, and therefore should never allocate a dollar to an organisation that would generate less value than another available option. For example, you should donate to a charity that can save a life for every $2,000 dollars before you should donate to a charity that needs $5,000 dollars to save a single life.

If only quite small amounts of money are donated, Singer’s advice makes sense since it directs that meagre flow to where it can do the most good (the ‘lowest hanging fruit’). The problem is that it is not good advice for disbursing very large amounts of money (in the hundreds of billions) which is the scale that would be needed to actually solve the problem of poverty. Micro-interventions don’t scale well since the whole point of them is that they are unusually easy to reach ‘low hanging fruit’. One could spend at most a few tens of millions of dollars on anti-mosquito bed nets before returns start dramatically diminishing because everyone who can be helped by them already has one. This points to the limits of the individualistic consumerist approach to ending poverty. The best – most beneficial – choice you can make as an individual spending $50 or even $5,000 is different from the best choice you should make if you have several hundred billion dollars to spend.

For example, a prominent philosopher in the effective altruism movement, William MacAskill criticises GiveWell's decision to recommend Give Directly which pays out its donations directly to people in very poor countries. MacAskill worries that simply giving very poor people money is an inefficient use of resources compared to investing in health micro-interventions like deworming pills, mainly because poor people won’t spend money as cleverly as people with PhDs in public health. I allow that such paternalism might be justified in consequentialist terms when thinking at the scale of an individual donor. However, if you have a larger, government sized budget to spend then a global basic income programme would appear as a cost-effective means to achieve a far more valuable goal: the elimination of the extreme poverty that drives the particular recurring health vulnerabilities targeted by micro-interventions.


Effective altruism is set up to focus on giving in a way that unfortunately neglects achievement. In order to be more appealing to more people it adopts an individualistic consumerist framing. But doing so forecloses political options and traps us in an inferior equilibrium. The advice effective altruists then offer us on how to help is true as far as it goes - these are the best ways to help, thinking as a individual consumer. But at the same time the advice is self-defeating. It is only useful under the pessimistic assumption that total contributions will be so small that only piecemeal contributions are possible that will leave the underlying causes of suffering unaddressed. It helps to select between programmes that cannot ever succeed in doing much good. Effective altruism is constitutively incapable of ending global poverty.

Note: This essay was previously published on 3 Quarks Daily

Creative Commons Licence
This work is licensed under a Creative Commons Attribution-NoDerivatives 4.0 International License. Attribution: Thomas R. Wells is a philosopher at Leiden University. He blogs on philosophy, politics, and economics at The Philosopher's Beard.