Thursday, December 15, 2022

(Re-)Assessing Effective Altruism and Earning to Give After Sam Bankman-Fried

by Michael C. Dorf

Just a few months ago, Sam Bankman-Fried was the most prominent proponent and (self-described) practitioner of a certain brand of effective altruism. Does Bankman-Fried's arrest on federal fraud and conspiracy charges discredit effective altruism? Should it? Let's dig in.

The term effective altruism refers broadly to the unassailable notion that in acting selflessly to benefit others, one should aim to do so effectively. For example, if you are deciding whether to make a $100 donation to either of two organizations that provide food and shelter to hungry unhoused persons, and organization A uses 100 percent of its charitable donations to provide goods and services whereas organization B uses 40 percent of donations to cover administrative costs (including high salaries for its executives), then, other things being equal, you should give your money to A rather than B. That "other things being equal" is important, however, because even with its much higher overhead, it's possible that B will do more good with the net $60 you give it than A will do with the net $100. And of course, your real choice is not between A and B but among all of the possible uses to which your $100 can be put, not to mention the possibility of giving more (or less) money. Proponents of effective altruism offer guidelines for making these sorts of decisions.

Philosopher Peter Singer's 2016 book The Most Good You Can Do offers a useful primer in effective altruism. With respect to the question where to give, Singer makes a powerful case that donations to such non-profits as museums and universities (gulp!) do less good than donations to effective organizations that combat hunger and disease in the developing world. I leave to readers interested in the effective altruism movement as a general matter the question of how successful Singer's argument is. I'll say for myself that I had a similar reaction to The Most Good You Can Do as I have had to Singer's landmark book Animal Liberation. With respect to both humans (the objects of the good Singer assumes you want to do as an effective altruist) and non-human animals, I find Singer's writing elegant, and I end up in more or less the same place as Singer does, but in each instance I get there by a substantially different route because Singer is a utilitarian whereas I am not.

Aside: After a colloquium Professor Singer hosted at Princeton on my book with Sherry Colb, Beating Hearts: Abortion and Animal Rights, Singer gave us a gift of The Most Good You Can Do, which led me to wonder whether that was really the most good he could do with that copy of the book; we could have borrowed a copy from the library and Singer could have sold the copy he gave us to earn some money to donate to an effective charitable organization. Even so, I'm glad he gifted us the book, both because I think very highly of Singer and because one of the two copies owned by the Cornell library system is checked out and the other was lost. I consulted my personal copy to write today's post. 

That's a lengthy and self-indulgent introduction to my preliminary conclusion: nothing Bankman-Fried is accused of doing remotely calls into question the general notion that one ought to be thoughtful about charitable giving, with respect to both the causes and organizations one supports. But Bankman-Fried's saga does potentially call into question one strategy promoted by some versions of effective altruism: earning to give.

Earning to give (which Singer describes in Chapter 4 of The Most Good You Can Do) is straightforward. If you're a public-spirited and talented mathematician deciding whether to pursue a career as a math professor or a hedge-fund manager, you should choose the latter. If successful as a math professor, you might have enough income to give thousands of dollars in annual charitable donations, but if you succeed as a hedge-fund manager, you can give millions to worthy causes. The more you earn, the more you can give. Hence, earning to give (which I'll abbreviate as ETG).

Bankman-Fried professed and purported to practice ETG. While he was a billionaire, he was the poster child for ETG, which relies on the concept of comparative advantage, familiar to anyone who has studied economics. If you're unfamiliar with comparative advantage, here's a simple example adapted from a typical introductory Econ class or textbook: Let's say that reigning two-time NBA MVP Nikola Jokic is very good at cleaning windows, in part because, at nearly seven feet tall, he can easily reach high spots. Jokic can clean the windows in his Denver apartment better and faster than any housekeeper he could hire. Nonetheless, Jokic should pay someone else to clean his windows, because, good as he is at cleaning windows, he is even better at basketball, which is a skill that the market values (much) more highly than it values skill at window cleaning.

ETG is premised on comparative advantage. Take our good-at-math hedge fund manager. She wants to do the most good she can, so she considers volunteering a few hours each week at a local soup kitchen. ETG/comparative advantage would say that's a mistake. Even if she's really good at ladling out soup, the market values hedge-fund managing (much) more highly than it values soup ladling, so by using those hours to do some additional hedge-fund managing rather than soup ladling, she will be able to donate money to pay for a whole lot of soup and also to pay the salaries of workers who, though individually not as good at ladling as the hedge-fund manager, can collectively ladle more soup.

Now suppose that instead of being a math whiz, someone I'll call Robin Hood is really good at both making balloon animals and at cracking safes. Robin wants to do as much good as possible. He could become a balloon-making clown who volunteers at a children's hospital when he is not working at birthday parties, or he could become a bank robber. As a bank robber, he will be able to steal millions, most of which he will give away (discreetly) to organizations that collectively do a whole lot more good than Robin could accomplish as a balloon-animal-making volunteer. ETG says Robin should become a bank robber.

Hold on. You're probably thinking that in the foregoing example I forgot to subtract from the good Robin's stolen money does the harm it causes to the bank and its customers. But I haven't. For one thing, the customers' deposits are FDIC-insured, so the loss is spread broadly through a very slight marginal increase in the premium that banks must pay (which in turn will show up as a tiny decrease in the rate of interest that banks can pay depositors). But for the sake of argument, let's say that that cost is not trivial. Even so, Robin does better to become a bank robber. Why? Because the diminishing marginal utility of money means that the slight loss occasioned to each depositor in the developed world is more than overcome by the large gain that the donations do in the developing world--even accounting for some loss through administrative costs of FDIC insurance, marginal loss of confidence in banking, etc. So ETG really does support Robin's career choice to become an altruistic bank robber.

What about Bankman-Fried? In the view that Bankman-Fried has promoted since FTX's collapse, he set out to do good via ETG but made a number of very poor judgments when there was a bank run on FTX. In the view alleged in the indictment the Justice Department filed in the SDNY, Bankman-Fried was a nogoodnik from the start--someone who used ETG as public relations to obscure his fraudulent scheme. His eventual trial or guilty plea will shed more light on where the truth lies. For present purposes, however, it doesn't really matter. Either way, Bankman-Fried's fall would not undercut ETG.

Suppose that Bankman-Fried's self-description is accurate--that he started out with a pure heart and good intentions. If so, his comeuppance would not be a refutation of ETG. Rather, it would show only that he was not as good at making money as he anticipated: unable to earn, he was ultimately unable to give. Even so, people who are better at making money--my hypothetical hedge-fund manager, say--would be better able to practice ETG successfully.

Or suppose that, as the government alleges, Bankman-Fried was a fraudster who got caught. That too would not undercut ETG. Bad people can do harm by pretending at anything. If someone solicits funds for a non-profit children's hospital but uses the donations to purchase luxury beachfront property for his own use, he discredits his own character; he does not, however, discredit children's hospitals. Likewise with ETG. The fact that some people are bad at or pretend to ETG does not undermine ETG as practiced by people who are good at it and practice it sincerely.

So is ETG in the clear? Not necessarily. Suppose Bankman-Fried had been a successful practitioner of ETG. What would that look like? FTX was (and for now remains) a platform for trading cryptocurrency. Currently, more than a third of the market capitalization of cryptocurrencies is in Bitcoin, which is a driver of completely unnecessary climate change, because Bitcoin mining requires enormous amounts of energy to power supercomputers. The second most popular cryptocurrency, Ethereum, bills itself as a green alternative, but while its footprint is smaller than that of Bitcoin, it too uses energy for what are ultimately pointless tasks. By promoting and increasing demand for cryptocurrency, a successful trading platform causes serious environmental harm.

But what if the benefits that hypothetical non-fraudster successful Bankman-Fried produced through his philanthropy outweighed the environmental costs resulting from the promotion of cryptocurrencies? Would that justify ETG? Maybe, although now it's apparent that successful ETG looks a lot like my Robin Hood example. And I suspect--or at least hope--that most readers share my moral intuition that even if the good Robin does through redistribution outweighs the harm he does by robbing banks, his robbing of banks cannot be justified.

Part of the force of that intuition comes from the fact that bank robbery is illegal. But not all of it. Operating a coal-fired power plant to run the supercomputers for a Bitcoin-mining operation is legal but in my view still wrong--regardless of what one does with the money it yields. If you don't share that intuition, consider complaints about working conditions in Amazon warehouses. Let's take Jeff Bezos at his word that he plans to give away his billions at some point. Let's also assume for the sake of argument that improving the conditions for Amazon workers would yield a benefit for those workers of X but cut into the company's profits enough to reduce the size of the fortune Bezos can eventually give away by some amount greater than X. On the ETG logic, Amazon would be justified in rejecting workers' demands for better working conditions.

What these examples show is that the ETG logic is profoundly utilitarian and thus subject to the usual critiques of utilitarianism. Robin Hood, hypothetically successful Bankman-Fried, and Bezos are simply less extreme versions of the canonical example in which a doctor carves up a perfectly healthy patient to harvest his organs to transplant in five patients who would otherwise die.

As I said above, I'm not a utilitarian, but I acknowledge that utilitarians don't simply fold in the face of the organ transplant scenario. Rather, they typically proffer rule-utilitarianism in place of the act-utilitarianism that is the target of the transplant critique. Whether that move or others succeed is beyond the scope of this essay, but let's assume that it can succeed. What would that imply for ETG?

The short answer is that ETG might still be justifiable where one earns the money to give through an ethical business. Of course, if you're a classical Marxist, you will think that impossible. To amass a fortune, you need to exploit (at least in the technical sense) your workers, which is itself wrong. So if you think that ethical capitalism is impossible, then you will think that ETG is also impossible.

But I'm not a Marxist (of any sort), nor, I suspect, are most readers. I believe that profit-seeking firms can succeed without engaging in unethical practices. Still, there is a substantial gap between can and do. The imperative of maximizing shareholder value pushes publicly traded companies and their investors towards what my late colleague Lynn Stout aptly called psychopathic behavior. Publicly traded companies that do not maximize profits are vulnerable to takeover by vulture capitalists or wealthy megalomaniacs.

Bottom line: ETG is defensible in principle notwithstanding SBF. However, ETG is difficult to accomplish in practice. Moreover, the underlying legal structures and political economy that permit the amassing of giant fortunes probably do more harm than good, because ethical and successful practitioners of ETG are surely a minority among people who are motivated to become billionaires.

No comments: