In defence of effective altruism
Don't judge an idea by its worst advocates
With corrupt crypto king Sam Bankman-Fried as its most prominent representative, the Effective Altruism movement is not particularly popularly these days. And some other people associated with the Effective Altruism movement have bizarre and unappealing ideas. But the worst form of ad hominem argument is “someone bad asserts p, therefore p is bad”[1].
Whether under this name or not, most economic research on both welfare policy and development aid takes the premise of effective altruism for granted. The central premise of effective altruism is simple: if you want to help poor people, give them what they most need. The practical force of this premise arises from a lot of evidence showing that, in general, what poor people need most is money.
(Image generated with DALL-E)
The starting point for the premise is some version of consequentialism. It is most directly opposed to the idea that altruism should be evaluated in terms of personal virtue. To take a typical example, effective altruism would say that someone with professional skills that are highly valued in the market should not volunteer in a soup kitchen, they should spend the time working for high pay and then donate their pay to the people who are hungry. [2] And there are plenty of highly popular interpretations of personal altruism that are even less effective, such as “thoughts and prayers”.
Coming to the question: why give people money, rather than addressing their needs directly, I will quote from my book Economics in Two Lessons which presents the issue in terms of opportunity cost.
What would a desperately poor family do with some extra money? They might use it to stave off immediate disaster, buying urgently needed food or medical attention for sick children. On they other hand, they could put the money towards school fees for the children, or save up for a piece of capital like a sewing machine or mobile phone that would increase the family’s earning power.
The poor family is faced with the reality of opportunity cost. Improved living standards in the future come at the cost of present suffering, perhaps even starvation and death. Whether or not their judgements are the same as we would make, they are in the best possible position to make them.
There are plenty of qualifications to make here. Maybe the most important is that family heads (commonly men) may make decisions that are more in their own interests than in those of the family as a whole. Giving money to mothers is often more effective.
And sometimes delivering aid in kind is the only way to stop corrupt officials stealing it along the way.
In addition, there are some kinds of aid (local public goods) that can’t be given on a household basis. If people in a village don’t have clean drinking water, then the solution may be a well that everyone can use.
Nevertheless, whenever anyone advocates a policy on the basis that it will help poor people it is always worth asking: wouldn’t it be better to just send money?
[1] Ad hominem arguments aren’t always bad, as I’ve discussed before. If someone is presenting evidence on a factual issue, rather than a logical syllogism, it’s necessary to ask whether you are getting all the facts, or just those that suit the person’s own position .
[2} Contrary to the SBF example, it doesn’t suggest stealing money to buy an island in the Bahamas, then covering up by giving some of the loot away
By coincidence, I just saw this, suggesting some big supporters are EA are rethinking their commitment. So, nicely timed!
https://www.semafor.com/article/11/21/2023/how-effective-altruism-led-to-a-crisis-at-openai
Thanks for this thoughtful comment, Prof John. However, the proposition that poor people are in the best position to decide what to do with a gift has a number of counter-arguments. First, we know that a proportion of individuals will spend a windfall on disserviceable purposes. The Grameen model of micro-finance (at least in its original incarnation) granted more than 95% of its loans to women because the men in impoverished communities were more likely to spend a windfall on gambling or worse. Second, the proposition sounds uncomfortably like the proposition that brings economic rationalism undone: that everybody is rationally self-interested and can be entrusted to make decisions in their best interest in a free market. On the contrary, we know that there are such things as public goods that individuals can't supply for themselves and that are vital for well-being. (Of course, the proposition has different individual and collective dimensions. There is a large gap between what a competent bureaucracy would provide to a starving individual – a food voucher – and what it would provide to a food insecure population – better farming). Third, it would seem to demean the role of expertise – "We know what to do with our money, we don't need a bureaucracy to advise us ". A significant proportion of the public service spends its time trying to devise better solutions to social problems than individuals would apply by themselves. And finally, it would seem to be congruent with the claim that the reason for the gap in Aboriginal health outcomes is that Aboriginal people don't control the money ( confusing a necessary condition of consultation and involvement with a sufficient one).
Your post accurately identifies many of these exceptions/qualifications; I would argue that the exceptions are sufficiently powerful to render the proposition generally wrong, though sometimes appropriate.
Yes, opportunity cost is a powerful frame. So also is mediation by a competent public authority or civil society group with expertise in addressing the particular needs. The role of a competent entity seems to be missing from both effective altruism and personal virtue. It seems to be missing from the Bill Gates model that because he has lots of money, he knows best how to solve world health. If the central premise of effective altruism is "if you want to help poor people, give them what they most need", it avoids the question of who decides what they most need – the recipient (see above why this isn't always optimal) or the giver (see Bill Gates example above why this isn't always optimal).