[Epistemic status/edit: After reading comments, no longer sure I agree with Part I. Part II still seems right but possibly a cost that can be outweighed by other factors. I continued to be worried about this without necessarily thinking it is a knock-down argument. I’m still not sure how to balance my support for some systemic change causes against my concern about others. Buck’s comment on morality seems important.]
One of the most common critiques of effective altruism is that it focuses too much on specific monetary interventions rather than fighting for “systemic change”, usually billed as fighting inequitable laws or capitalism in general. For example, Amia Srinivasan, in her review of Doing Good Better,
What’s the expected marginal value of becoming an anti-capitalist revolutionary? To answer that you’d need to put a value and probability measure on achieving an unrecognisably different world – even, perhaps, on our becoming unrecognisably different sorts of people. It’s hard enough to quantify the value of a philanthropic intervention: how would we go about quantifying the consequences of radically reorganising society? […]
Effective altruism, so far at least, has been a conservative movement, calling us back to where we already are: the world as it is, our institutions as they are. MacAskill does not address the deep sources of global misery – international trade and finance, debt, nationalism, imperialism, racial and gender-based subordination, war, environmental degradation, corruption, exploitation of labour – or the forces that ensure its reproduction. Effective altruism doesn’t try to understand how power works, except to better align itself with it. In this sense it leaves everything just as it is.
This same point has been made again and again and again. In response, many effective altruist leaders have gone on to say that they love systemic change and that the movement is entirely in favor of it.
I am not affiliated with the organized effective altruist movement and my opinion has no relation to theirs. They have spent a lot of work trying to convince everyone that they are entirely in favor of pursuing systemic change, I believe them, and nothing I say here reflects on that.
But I, personally, worry a lot about pursuing systemic change.
“Worry about” is not the same as “totally oppose”. This post’s Hansonian title is “Beware Systemic Change” rather than “Against Systemic Change.” But I’m pretty serious about bewaring of it.
First, what do I mean by “systemic change”? Traditional charity, like healing the sick, is almost universally viewed as good or at least neutral. Everyone agrees the sick should be healed; if there are unhealed sick people, it’s because we don’t have the resources to pursue our universally held goal. The same is true of feeding hungry children. It’s true of weird causes like AI risk – some people think it’s silly, but they’re happy to let other people work on them if those people want. It’s even true of things like cutting carbon emissions, sort of. When the Koch brothers say they oppose cutting carbon emissions, they mean they oppose laws mandating such cuts, or budgets that spend communal resources to enforce them. If a private donor offered to pay for scrubbers on every smokestack at zero cost to the rest of the economy, the Koch brothers would have no objection.
Some political issues are kind of like this. People from all over the political spectrum agree that corporate welfare is a bad idea; if we still have corporate welfare, it’s because there’s not enough attention and organization to force politicians to abandon it. In other cases, we all agree something is good but disagree on whether it is an optimal use of resources: for example, most people agree that aid agencies like UNICEF that help children abroad are doing good work, but not everyone agrees with funding them from the federal budget.
Other political issues are not like this. Some people believe that increasing the minimum wage is a laudable goal; other people believe it will hurt the economy or that it violates important moral rights. The reason we don’t have a higher minimum wage isn’t because passing laws costs a lot of money that no one has raised yet, or because no one is paying attention to the issue. It’s because a lot of people oppose it and so far those people are winning, or at least holding their own.
In terms of Freshman English Plot Devices, traditional charity like healing the sick is “man versus nature.” Political issues like the minimum wage are “man versus man”.
When I think of systemic change, I think of man versus man. Even if effective altruists helped governments increase their foreign aid budget, I don’t think Amia Srinivasan and Jacobin Magazine and the rest would think we were participating in “systemic change”. I think at the very least they’d want altruists out in the street demonstrating for higher minimum wages, and at most trying to eliminate global capitalism.
Which is a problem, because a lot of people like global capitalism. A dialogue:
Bob: Man vs. man conflicts raise some thorny issues which man vs. nature conflicts manage to avoid. I think we should be very wary about opening the door to political discussions.
Alice: What? Come on, Bob, you talk about politics all the time!
Bob: I watch porn all the time too; that doesn’t mean I’m proud of it, or think it’s the most good I could do with my resources. Suppose effective altruists get involved in the 2016 US presidential election – which isn’t prima facie a bad idea; think about how easy it would have been to make Gore win in 2000 and how much would have changed if he had. Lots of people work very hard and raise $10 million for the Democrats. Lots of other people work very hard and raise $10 million for the Republicans. Now the Democrats and Republicans are at exactly the same position vis-a-vis each other as they were before the effective altruists got involved, but we have wasted $20 million that could have gone to healing the sick or feeding the hungry. And I’m using money to make things obvious, but the same goes for donating time or advocacy or other resources. If this sort of thing started happening, we would want to promote a general cultural norm of “never spend resources on man vs. man conflicts”. If both sides were equally likely to follow the norm, then the conflicts would remain unaffected but everyone would have more resources to spend on the sick and hungry.
Alice: But effective altruists are very unlikely to donate to both sides of a political issues equally and cancel out. Political views are heavily shaped by demographics, and EAs are likely to skew left just like most other highly-educated groups. Even aside from this, their similar moral assumptions and thinking styles will lead them to converge onto the same side of an issue. Half of Americans are creationist, and almost as many oppose gay marriage, but I would expect fewer than 5% of EAs to hold either position. The idea of donations cancelling out is totally unrealistic. Instead, we should predict that on most issues, most EA donations will go to the same side. We end up not with a wasteful neutrality, but with a large sum of money going to one side, detracted from only slightly by a much smaller sum going to the other. If the cause is important enough, it might still be the most good we could do – the net benefit of (good from giving large amount of money to one side) minus (harm from giving small amount of money to the other side) would still be higher than the benefit of giving all the money to a traditional charity.
Bob: That just kicks the problem one meta-level up. Suppose that on each given issue, effective altruists converge dramatically around one or the other side. If half the time they converge around the right side, and the other half around the wrong side, then over a large number of issues their contributions will gradually even out and sum to zero.
Alice: That’s a ridiculous way of looking at it. We don’t just flip a coin to determine which side to back. We exhaustively study the argument for both sides, the evidence base, et cetera. Then we focus only on those issues where we can be most certain we’re in the right. The odds there are a heck of a lot better than fifty percent!
Bob: But you could make the same argument about picking stocks, couldn’t you? Do lots of research, focus on the ones where you’re most certain that they’re overvalued or undervalued, and then you have great odds of getting rich! But of course, we know that doesn’t work. Everyone else is trying the same thing, and the current position of the stock market reflects the consensus results of that process. You run afoul of the efficient market hypothesis.
Alice: Now you’re just being silly. There’s no efficient market hypothesis for politics!
Bob: But why shouldn’t there be? A lot of people mock rationalists for thinking they can waltz into some complicated field, say “Okay, but we’re going to do it rationally“, and by that fact alone do better than everybody else who’s been working on it for decades. It takes an impressive level of arrogance to answer “Why are your politics better than other people’s politics?” with “Because we want to do good” or even with “Because we use evidence and try to get the right answer”. You’d have to believe that other people aren’t even trying.
Alice: What you’re saying makes a certain kind of sense in Weird Platonic Spherical Cow Perfect Rationality Outside View World. But think about this from the Inside View perspective. Once again, half the country is creationist. Almost half oppose gay marriage. It’s like a stock market where half of the investors are throwing everything they have into the perpetual motion industry. Surely you can admit that even a little bit of intelligence, education, and rationality can actually take you a long way in politics?
Bob: Half the country is creationist, but there’s almost no easy gains from fighting them; any curriculum that federal politics can conceivably affect is evolutionary by this point, and it’s unclear we get any real benefits by going after the last few Alabama middle school students. As for opposing gay marriage, I think you’re going beyond your supposed reliance on evidence here. The strongest conservative case against gay marriage is that it reinforces a centuries-long redefinition of marriage from a strategic partnership focused on child-rearing to a ceremonial acknowledgment of romantic infatuation, potentially leading to a deep shift in the way people think about issues like who to marry, when to have kids, when to get divorced, and how to treat their family. That argument hasn’t been rigorously evaluated by statisticians and found wanting. It’s been found annoying and left untouched. Your differences are foundational assumptions and methodological disagreements about what sorts of issues to focus on, not simple “he made an arithmetic error when calculating the effects” style obvious superiority.
Alice: Really, Bob? You really want to go there?
Bob: Yes. In fact, I worry that this plays into exactly the potential flaws of the effective altruist movement. I can count up all the harms of banning gay marriage: exactly 1.13 million gay people regret not being able to marry, they rate their distress at 3.2/5 on the Likert Scale, that comes out to X QALYs lost per annum, but you have no way of easily quantifying the potential harms of gay marriage, therefore your argument is invalid. A lot of these issues involve trading off easily quantifiable harms on one side versus less quantifiable harms on the other: social trust, cultural cohesion, moral credibility, “freedom” broadly defined, ability to innovate. Highly educated people used to studying science might just be more likely to fall for the streetlight effect and go with the side that promises more quantifiability, rather than the side more likely to be right.
Alice: I…think you’re being deliberately annoying? It seems like exactly the same kind of sophisticated devil’s-advocate style argument we could use for anything. Sure, nothing is real and everything is permissible, now stop playing the Steel Man Philosophy Game and tell me what you really think! It really should be beyond debate that some policies – and some voters- are just stupid. Global warming denialism? Mass incarceration? Banning GMOs? Opposing nuclear power? Not everything is a hard problem!
Bob: I really do sympathize with you here, of course. It’s hard not to. But I also look back at history and am deeply troubled by what I see. In the 1920s, nearly all the educated, intelligent, evidence-based, pro-science, future-oriented people agreed: the USSR was amazing. Shaw, Wells, Webb. They all thought Stalin was great and we needed a global communist revolution so we could be more like him. If you and I had been alive back then, we’d be having this same conversation, but it would end with both of us agreeing to donate everything we had to the Bolsheviks.
Alice: Okay, so the smart people were wrong once. That doesn’t mean…
Bob: And eugenics.
Alice: Fine then. For the sake of argument, the smart people were wrong twice. That still doesn’t…
Bob: It does! A quick run through the history books shows that smart people trying to effect systemic change have an imperfect track record. I won’t say that they’re unusually bad compared to other demographics, but certainly nothing as stellar as the “let’s just not be morons” theory might lead one to expect. You like quantifiable things and specific examples, so let me give you one. I’ve sometimes thought that Friedrich Engels can be considered one of the fathers of effective altruism – at least of the earning-to-give variety. Wikipedia says:
Once Engels made it to Britain, he decided to re-enter the Manchester company in which his father held shares, in order to be able to support Marx financially, so that Marx could work on his masterpiece Das Kapital. Engels didn’t like the work but did it for the good of the cause.
And in one sense, Engels-as-altruist was utterly brilliant. He effectively zeroed in on the most influential thinker of his era, funded an otherwise-impossible level of output from him, and his work directly led to revolutions in a dozen countries with radical change in the lives of billions of people. But in the more important sense, the net effect of his contribution was global mass murder without any lasting positive change. If we count him as an effective altruist – and under the circumstances I’m not sure we can do otherwise – then the net contribution of the movement throughout history has been spectacularly negative. That should make us really concerned. Not “nod sagely and promise to think about it” level of concern, but more “run away screaming” level of concern. That’s why I’m so reluctant to accept your otherwise-reasonable points about the seemingly obvious issues.
Alice: On an emotional level, I get your point. But on a rational level, wouldn’t it be astounding if smart people trying to figure out the safest ways to do the most good consistently made things worse?
Bob: There are many more ways to break systems than to improve them. One Engels more than erases all of the good karma created by hundreds of people modestly plodding along and making incremental improvements to things. Given an awareness of long-tail risks and the difficulty of navigating these waters, I’m not sure our expected value for systemic change activism should even be positive, let alone the most good we can do.
Alice: So don’t go poking around super-complex systems with lots of variables as complicated as “capitalism” versus “communism”. Stick to well-understood things with fairly predictable effects. If we have a little bit more humility than Engels, maybe we won’t fall into the same trap he did.
Bob: All nice and well, except that I do not see even the tiniest sign of supra-Engels levels of humility in the effective altruist movement as it exists today. Recently I have had to deal with lots of our Facebook friends joining and sharing images from a group called “Muh Borders!” which exists to post memes making fun of anyone who opposes Open Borders as a stupid bigot who is not worth talking to:
In terms of “political causes that we can be totally sure won’t backfire and devastate entire countries for generations”, I would place open borders…well, let’s say somewhere in the bottom quartile. A thorough analysis by one of its strongest and most intelligent advocates concludes with “doubt that the American polity could survive and flourish under open borders” but has been mostly ignored in favor of constantly retreading the same old streetlight-illuminated ground of whether immigrants do or don’t affect native wages. And this is the community that is supposed to have solved the hard problem of getting mind-killed by politics, and can now be sure it’s genuinely pursuing the side of Good rather than the side that looks like Good but actually kills tens of millions of people?
Alice: That’s not fair. Yes, there are some people who reflect poorly on the open borders movement, but they’re not all effective altruists, and even humble people who try their best to think about things rationally are allowed to let off steam on Facebook once in a while. The open borders movement has also done a lot of really impressive analysis, and even though there are risks, given the potential benefits it really can be thought of as a no-brainer.
Bob: I am a coward and will stick to buying bed nets, thank you.
There’s another problem with man vs. man: the people we want to recruit are men, and the people we want to make our movement out of are men.
(That came out sounding more sexist than I intended. You know what I mean.)
Several people have recently argued that the effective altruist movement should distance itself from AI risk and other far-future causes lest it make them seem weird and turn off potential recruits. Even proponents of AI risk charities like myself agree that we should be kind of quiet about it in public-facing spaces.
As someone whose own views on open borders are mixed (I should probably write a post), I am really turned off by memes like the one above. And since only seven percent of Americans fully support open borders, that’s a lot of potentially turned-off people. They’re going to go on effective altruist sites, see that a big part of the movement is arguing for a policy that they abhor, and notice their potential colleagues talking about how people like them who oppose that policy are stupid and parochial and hate foreigners. “We think you’re wrong and stupid, come join our movement” makes a really crappy recruiting pitch. But it is the pitch we are sending to anyone who isn’t a Silicon Valley libertarian, George Mason University economics professor, or Vox.com journalist – the only three groups from which I have seen a level of open borders support much beyond the lizardman level.
If effective altruists are split on an issue, then they’re working at cross-purposes, like the people above who donated $20 million to both the Democratic and Republican campaigns. But if effective altruists are not split on an issue, then they’re projecting a unified Effective Altruist Consensus on it which is going to look pretty intimidating to anybody who disagrees. And if there are enough of these issues, then a randomly selected person is almost certain to disagree with at least one of them. The more different from the EA stereotypical demographic they are, the more likely such a disagreement will be. Politics is the mind-killer and quickly takes over from everything else; I do not think political disagreements can stay quiet and harmless for long. If two people are both committed to healing the sick and feeding the hungry but one believes in open borders and the other in a more Bernie Sanders style approach to immigration – not even conservative, just a Bernie Sanders style approach! – they can peacefully coexist in an effective altruist movement focusing on traditional charity, but one focusing on systemic change is likely to get pretty heated.
If you think this is overly pessimistic, think back to the issues with the most recent EA Summit, which advertised fully vegetarian meals but added non-vegetarian options at the last second. This became a big enough scandal that I, who was two thousand miles away from the conference, got inundated with arguments about it on Facebook, Tumblr, and this blog. Several people threatened to quit effective altruism entirely, though I don’t know if any of them followed through.
This is a community that can literally almost tear itself apart over the question of what to have for lunch. I think there might be too much dynamite around to risk shooting off sparks.
And I also think effective altruism has an important moral message. I think that moral message cuts through a lot of issues with signaling and tribal affiliation, that all of these human foibles rise up and ask “But can’t I just spend my money on – ” and effective altruism shouts “NO! BED NETS!” and thus a lot of terrible failure modes get avoided. I think this moral lesson is really important – if everyone gave 10% of their income to effective charity, it would be more than enough to end world poverty, cure several major diseases, and start a cultural and scientific renaissance. If everyone became very interested in systemic change, we would probably have a civil war. Systemic change is sexy and risks taking over effective altruism, but this would eliminate a unique and precious movement, in favor of doing the same thing as everybody else.
If effective altruism became more political, it would likely fade seamlessly into something like the Brookings Institution (a top-tier think tank whose $100 million yearly budget is by my calculations well above what the entire world combined spends on deworming) or the Cato Institute (another top think tank whose $30 million budget is likely more than all AI risk charities and all effective animal rights charities combined). Probably the staff of the Brookings Institution go into work each day thinking “How can I best improve the world by giving it better policies?”, and I admire that, but they don’t have the same sort of moral mission as effective altruism and it would be disappointing to see the latter collapse into an annex of the former.