Eric Posner The strange case of Samuel Bankman-Fried
Effective altruists believe they should devote themselves to benefiting others the best they can. Some take this position to extremes by pursuing the highest-paying jobs they can find and then donating their riches to vetted charities that save the most lives per dollar deployed.
One of the most famous effective altruists, Sam Bankman-Fried, is currently on trial for fraud.
In Michael Lewis’s insightful new book, Going Infinite, Bankman-Fried is portrayed sympathetically as a brilliant, humane, but emotionally crippled figure. Though he seems to lack the capacity to feel empathy for those around him, he reasons himself into the effective altruists’ utilitarian outlook. He stops eating meat to minimise animal suffering, and he commits to minimising human suffering as well – including by supporting efforts to address existential threats like out-of-control artificial intelligence.
After earning a physics degree from MIT, Bankman-Fried joined the trading firm Jane Street, where he learned to arbitrage small price differences between assets. He proved to be quite good at quickly identifying, calculating, and executing gambles. But, as Lewis notes, and as Bankman-Fried surely understood, Jane Street simply leached rents from the financial system, rather than doing anything to benefit humanity. Even ignoring that, the millions that Bankman-Fried could ultimately earn there would do little to alleviate human suffering globally.
Bankman-Fried left Jane Street to find a better use for his talents, and soon noticed that price discrepancies were much greater in cryptocurrency markets, where poorly designed exchanges and the absence of regulation had kept professionals away. So, he founded the hedge fund Alameda Research, where he made millions of dollars by replicating Jane Street’s arbitrage strategies in more fertile hunting grounds.
Recognising that the crypto market needed a better-designed exchange that professional traders would be willing to use, Bankman-Fried enlisted his friend Gary Wang to write the code for FTX.
An exchange can be extremely profitable, because it takes a cut of every trade without taking much risk: when asset prices moved in the wrong direction, customers forfeited their collateral. FTX innovated by closing out trades extremely rapidly to minimise the risk that a trader’s collateral was insufficient to cover his or her loss, which would require FTX itself to spread that loss to other traders.
An exchange functions by matching buyers and sellers, but it requires a critical mass of traders to satisfy demand on each side. Because dozens of crypto exchanges already existed, Bankman-Fried needed both to lure their customers to FTX and bring in as many crypto-trading novices as possible.
With customary logic, Bankman-Fried overcame his previously solitary inclinations and invested vast amounts of his time and money to spread the word. He spent hours pitching FTX on television, and he threw money at celebrities like the NFL quarterback Tom Brady, online influencers, and financial gurus. He also donated large sums to political candidates – Democrats and Republicans alike – hoping that the United States government could be persuaded to license FTX’s business. His exchange would not be trusted by the masses, he recognised, until it was doing business from gleaming office towers in New York City.
To support FTX, Alameda was enlisted as a market maker, stepping in to buy or sell on the opposite side of the exchange when not enough traders materialised. But, of course, Alameda needed funds to take the losing sides of trades, and to hold onto the assets until they could be resold. Bankman-Fried allowed it to borrow an unlimited amount of money from FTX, effectively using customers’ funds to backstop those same customers’ trades. The risk-minimisation technology was a mirage.
This may explain how FTX gained its competitive advantage over the incumbent exchanges. FTX customers got good terms, never realising that they were exposed to massive risk. When they finally learned the truth, they withdrew their money, and the house of cards collapsed.
The government alleges that Bankman-Fried, the majority owner of both firms, deliberately gambled with customer funds. Bankman-Fried claims that he did not know that Alameda owed US$8 billion to FTX. He also contends that he was not making fraudulent statements when he or FTX offered various assurances that customer funds were protected, that Alameda enjoyed no special privileges, and that FTX was “fine” up to its collapse.
Lewis seems to believe him. But even if Bankman-Fried did not intend to deceive anyone, the law defines fraud to encompass reckless disregard of the truth. Bankman-Fried must have understood that he was taking extravagant risks by refusing to hire qualified financial and legal experts to oversee his business – risks that are well documented by Lewis, who also knows financial markets.
While charming or at least tolerable in a teenager, Bankman-Fried’s aversion to “grown-ups” likely will turn out to be criminal in a manager of a US$32-billion firm that had become integral to a growing financial sector. Moreover, as Lewis shows, because Bankman-Fried was never very honest or straightforward with people, few employees were in a position to spot trouble, bring it to his attention, or expect him to be responsive.
More to the point, there is reason to think that Bankman-Fried knew what he was doing all along. Lewis recounts his subject’s habit of calculating an ‘EV’ – expected value – for all his actions, including things as trivial as keeping a promise to attend a meeting or speaking event. He would mentally assign a probability that a particular meeting would generate revenue for his firm or possibly value for society, multiply it by that value, then compare it to alternatives.
If the 19th-century British philosopher Henry Sidgwick were alive today, he might point to this episode as an object lesson of the risks that arise when fallible humans employ utilitarian thinking. But that is probably not how Bankman-Fried would see it, even now. Remember, the ‘E’ in EV refers to ‘expected’. The question, then, is whether the collapse of his financial empire, and the damage inflicted on his friends, family, employees, customers, and investors was justified by the expected social pay-off had FTX survived.
Serving a long prison sentence for fraud was itself a part of the downside risk that, discounted to present value, surely shrank to nothing relative to the potential pay-off of saving humanity from an AI or asteroid.
FTX was just a gamble. It was only fair that the social costs should be weighed in the balance alongside the social benefits. And, as Lewis shows, Bankman-Fried was exceedingly confident of his ability to calculate the expected value of big gambles.
Eric Posner, a professor at the University of Chicago Law School, is the author of How Antitrust Failed Workers.
© Project Syndicate 2023