Wednesday, April 20, 2016

Effective Altruism, Radical Politics and Radical Philanthropy

It can sometimes be difficult to discern precisely what's in dispute between Effective Altruists and their (leftist) critics. This is perhaps in part due to EA's being such a big tent that objecting to one proposal or proponent is not necessarily an objection to EA itself.  To clarify the latter, I see Effective Altruism as a matter of two core commitments:

(1) The "Altruism" bit: A commitment to making the world a better place -- including a willingness to expend some non-trivial proportion of one's own resources to this end.

(2) The "Effective" bit: A commitment to using these resources as effectively and efficiently as possible (based on the best available evidence, analysis, etc.).

And I guess there's a further 'movement-building' element that is perhaps common to all movements:

(*) The belief that others should do likewise.

Now, these core commitments seem pretty innocuous to me, so I'm always a bit baffled when I see people objecting to EA as such.  (Why would anyone be against making the world a better place?)

Of course, there's plenty of room for internal disagreement about what is actually most effective.  EAs vary between whether they prioritize (e.g.) traditional global poverty charities, (farm) animal welfare, meta-charities, global catastrophic risk mitigation, or public policy / lobbying.  The former cause area, with its robust supporting evidence, is emphasized in materials directed at popular audiences, for obvious pragmatic reasons (e.g. the prominence of aid skeptics, and the popular tendency to dismiss anything that sounds too "weird").  This has unfortunately led some academic critics to assume that EA is all about short-term thinking, which really couldn't be more wrong.

One of the internet's most, um, rhetorically strenuous critics of Effective Altruism is Brian Leiter, who in his most substantial post on the topic to date writes:
What if instead of picking worthy charities in accordance with Singer’s bourgeois moral philosophy, those with resources committed all of it to supporting radical political and economic reforms in powerful capitalist democracies like the U.S.; perhaps even committing their time and resources to helping other well-intentioned individuals with resources organize themselves collectively to do the same? Is it implausible that if all those in the thrall of Peter Singer gave all their money, and time, and effort, to challenging, through political activism or otherwise, the idea that human well-being should be hostage to acts of charity, then the well-being of human beings would be more likely to be maximized even from a utilitarian point of view? Do Singerites deny that systemic changes to the global capitalist system, including massive forced redistribution of resources from the idle rich to those in need, would not dwarf all the modest improvements in human well-being achieved by the kind of charitable acts Singer’s bourgeois moral philosophy commends? The question is not even seriously considered in the bourgeois moral philosophy of Singer. Although purporting to be concerned with consequences, like most utilitarians they set the evidential bar so high, and the temporal horizon so short, that the actual consequences of particular courses of action, including the valorization of charity over systemic change, are never really considered.

I don't see any objection to the core commitments of EA here, just a dispute about means. Leiter may insist that even if his proposals don't violate the letter of EA, they nonetheless remain contrary to the ethos of the movement, and would (in practice) be dismissed out of hand in a way that reveals the movement's fundamental shortcomings (such as the alleged focus on "modest improvements" over an unjustifiably short time horizon).  But I actually think this latter assumption is just false. In particular, Leiter -- like many critics of EA -- seems to have an unduly restrictive conception of "charity" compared to the wide range of EA-supported organizations I listed above. Whatever organizations you think could do the most good with additional funding (and I don't see anyone policing people's sincere and considered judgments here), you should be donating to them.

Should we prioritize lobbying for "massive forced redistribution"? Maybe!  It sounds to me like the kind of proposal that EAs would be interested in considering.  Of course, we'd want to hear more from reputable economists about the likely effects and possible risks (gotta be careful not to destroy the global economy, after all), and also more from skilled analysts on tractability: what our chances are of achieving such a goal.  My initial hunch is that this is unlikely to be the best bet out there, for broadly 'structural' reasons I'll explain below, but I'm certainly open to hearing more. (In particular, it'd be helpful to have some specific recommendations of putatively effective organizations pursuing this goal, along with their explanations of how they would use more funding, what kind of evidence they can offer that they're a well-run organization that would use the funding wisely, etc.)

Now, it's clearly not true that the EA movement is in general opposed to low-probability high-impact bets over long time horizons. After all, a popular point of ridicule is the prominence given to concerns about artificial intelligence and Bostrom-style reasoning about the overwhelming importance of existential risk ("X-risk") reduction. (Srinivasan's critical review, which Leiter praises, is even called "Stop the Robot Apocalypse"!)

So I'm a bit puzzled as to why Leiter and his ilk present themselves as external critics of EA, rather than hopping on board and making internal criticisms / suggestions about how best to achieve our shared goals.

On the substantive merits of his proposals, though, the political radical is here in an awkward position.  If one wants a "safe bet" for improving the world, the global poverty or animal welfare wings of EA are clearly the way to go.  If one wants the "best bet" in expected value terms, and is willing to make optimistic estimates in the absence of a robust evidence base, the meta and X-risk wings of EA are clearly the way to go.  It's difficult to find any principled evaluative standpoint from which radical politics looks distinctively promising. (Personally, I'm inclined towards making the "best bet" in expected value terms with moderately pessimistic priors in the absence of a robust evidence base, which leaves me at least somewhat sympathetic to the whole range of EA focus areas, including policy reform, but rather skeptical of radical politics. Note that I would be much less enthusiastic about global poverty charities were it not for the potential flow-through effects.)

At least, so it seems to me at first glance.  But there's room for discussion, and I'm sure that most EAs would welcome a serious discussion about how best to evaluate the various options on offer to us as individuals (which include political actions).  Like McMahan, I haven't yet seen any such serious analysis from the philosophical critics of EA. (They generally seem more interested in signalling how very Left they are by complaining about how "bourgeois" or "consumerist" EA seems to them, and how horrified they are that anyone could ever possibly justify working in the financial sector, etc.) But it certainly would be welcome.  Critics' denigration of individual philanthropic action per se is, by contrast, just bizarre and unhelpful -- providing cover and solace to the morally complacent.


  1. I'm not against making the world a better place.

    But I think that people who wish to invest a substantial portion of their resources in making the world a better place in ways which do not affect them personally, and believe that this is because they care intrinsically about the condition of the world in itself, apart from their presence in the world, are factually mistaken about their own motives.

    It is possible that I am mistaken about this, but it seems to me a reasonable way to object to EA as such.

  2. To be clear, I accept that people can and do care intrinsically about the condition of the world in itself. But I don't think they care enough to justify the investment of a substantial portion of their resources -- when it seems to them that they do, this is really (in my opinion) an expression of other concerns that they have, such as being objectively good people and so on.

  3. It's not clear to me why -- even were it true -- "EAs are factually mistaken about their own motives" constitutes an objection to EA as such. (EA makes no claims about the motives of its proponents.)

  4. The objection is that "I should donate 20% of my income to such and such a cause, because it would improve the world in such ways, and I really want to make these improvements," is a false statement. That it would improve the world is true; but that they really want the improvements is not, and if they understood what they really wanted, donating 20% of their income would not be a very good way to achieve that thing, so it is also probably false that they should make the donation.

    (I'm not going to try to prove my opinion about people's motives. I may well be wrong, if people turn out more different from one another than I suppose.)

    1. The normative explanation "because... I really want to make these improvements" is no part of EA. (I, for one, reject your implicit assumption that we should only do what will achieve what we want.) It would be a bizarre objection to white abolitionists to say "Oh, but abolishing slavery will not get you what you really want." Who cares what the abolitionists, deep down, "really" wanted? It is of no relevance to the assessment of abolitionism as such.

  5. Am I right that, even though it is not entailed by your sparse formulation of EA (viz. 1, 2, and *), it is common if not universal to accept that effects should be calculated based on marginal difference? This seems like something a leftist might want to reject, while remaining consistent with `sparse EA'. That is, whether I assume that if I don't go into finance someone else will, so there's no net gain to my principled refusal to work in finance, is optional. Someone like Leiter appears to care more about the negative effects that someone in fact has by working in finance, rather than comparing those effects to a hypothetical world in which soeone else does the same job, or perhaps to care more about the farther-away worlds in which everyone refused to work in finance.

    Similarly, when calculating expected utility for X-risks, a leftist might want to reject various assumptions about how to think about the utility of future persons, and/or how to weigh the negatives of death against the negatives of suffering. Doing so might lead from a willingness to make low-impact/high-impact bets to trying to start revolutions rather than trying to prevent the singularity, again consistent with sparse EA but counter to what I take to be the orthodoxy.

    More generally, I wonder if leftists don't really just reject the crude form of act-utilitarianism that is a hidden #3 for most EA-proponents (at least, the one's with the loudest voices). In other words, you seem to be saying that even EA's critics really appear to accept 1 and 2, but they have essentially technical disputes about instrumental strategies, but it seems to me that while they may accept 1 and 2, the dispute is still about normative assumptions that, though not explicitly part of your sparse definition, appear to be a part of how EA is understood and practiced by its advocates.

    1. That should have been "low-probability/high-impact bets" in \P2. I should also add that a leftist might want to use a value-pluralist version of consequentialism as well. If you care less about the positive utility of future persons existing (rather than never having been), use a version of rule-consequentialism (or maybe "sophisticated" act-consequentialism, like Railton's), and assign greater value to things like art, crticisism, and non-alienated work as intrinsic goods, you might wind up with 1 and 2 speaking heavily in favor of seizing the means of production rather than trying to insure that future AIs will be friendly to us. This would indeed be an internal disagreement with sparse EA, but not just about means, rather about normative principles that are part of the definition of non-sparse (plentiful?) EA.

    2. Thanks Jack, that seems right. Iason Gabriel has developed some of those normative disagreements in his helpful paper 'Effective Altruism and its Critics' (including the sort of group-egalitarian objection that I criticize here), albeit from a more "internal" standpoint.

      In general though, it takes a lot more than "just reject[ing] the crude form of act-utilitarianism" to find EA precepts broadly objectionable. I'd think that even, say, a Rossian pluralist could get on board with them. (It's not like we're advocating organ harvesting or pushing people off bridges here! Just saving more people rather than fewer, and stuff like that. Even the recommendation to go into finance is on the assumption that there are plenty of morally innocuous jobs in the financial sector. Actually doing harm is not, as far as I'm aware, sanctioned by 80k Hours or any other major EA organization.)

      There are certainly additional widespread assumptions that might be questioned, e.g. that extinction or low welfare for future generations would be a very bad thing. It seems to me pretty difficult to deny that, but some philosophers do accept radical person-affecting views with such implications (though presumably not anyone who also worries about the impact of climate change on future generations), and it could be interesting for such discussions to develop further.

  6. A libertarian might reasonably query how charity is consistent with a movement whose success would be to take property from one group to give to another and, if they are right, lead to lower economic growth/wealth thereFter. The kist of socialist countries where the poor are better off than thse in decadent capitalist countries is famously short.

    1. Hence my line about how EAs might be wary of endorsing the radical Leftist's politics until we've heard "more from reputable economists about the likely effects and possible risks" of their policy goals. Of course, there's a big difference between, say, Scandanavian-style market socialism vs., say, soviet communism! I don't think anyone (sane) is advocating for the latter these days...

      Just to clarify though, EA as a movement contains people from across the political spectrum, including libertarians.

  7. I appreciate your typically clear statement of how you understand the dispute, and I'll have more to say about it when I have more time. Regarding your sniping at the end about the use of the label "bourgeois" to describe yours and Singer's moral philosophizing, it has a quite precise meaning that I discuss in the Analyse & Kritik paper from which I quoted the excerpt in the blog post to which you link: it means philosophical inquiry that does not challenge the perquisites of the capitalist class or capitalist relations of production. I use it in a pejorative sense, because I think it is a defect of a philosophical approach that it is conservative in this regard.

  8. Effective altruism is an evidence-based movement. So any discussion of forced redistribution should include a look at the evidence: what is the track record of forced redistribution and what sort of effects has it had historically? The track record is very disappointing:

    The US government rarely uses evidence to select programs and seems virtually incapable of cutting programs that don't work. Millions who died under communism are an even more salient example. If someone is pushing forced redistribution, the onus is on them to explain their proposal in detail and show that (a) it will actually work (b) it will not fall prey to the issues forced redistribution has had historically:

    I doubt having EAs in charge of the forced redistribution would solve the problem, for instance. Sociopaths then have an incentive to infiltrate the EA movement, the same way they've infiltrated the administration of forced redistribution in the past.

  9. Wonderful, if long, piece by Deidre McCloskey, author of the Bourgeois Virtues worth a read in this regard :
    "The contribution the non-economist
    clerisy can make to an ethical change is to cease talking of voluntary exchange as
    exploitative, or as easily second-guessed by the better Swedish bureaucrats, as Nussbaum
    does. Prudence Only at the level of an ideal bureaucracy is just as partial and unethical as
    Prudence Only at the level of individual motivation. We need to inquire into how to make
    good people, including our governors, in the world as it is."

  10. A somewhat belated further reply. I have doubts about #1, but those are not at issue in the current discussion, You're quite right all the doubts are about #2. Your invocation of "reliable" arbiters of relevant effects (namely, economists) is, I think, a giveaway and consistent with my doubts. That being said, I do not disagree at all that those who feel like they want to make a quantifiable difference in the very short-term, should embrace EA. My worry is, in part, about the effects of this attitude. I don't see that anything you said is responsive to that.

    1. How about, "it's clearly not true that the EA movement is in general opposed to low-probability high-impact bets over long time horizons..."


Visitors: check my comments policy first.
Non-Blogger users: If the comment form isn't working for you, email me your comment and I can post it on your behalf. (If your comment is too long, first try breaking it into two parts.)

Note: only a member of this blog may post a comment.