Thursday, January 5, 2023

Matt Levine's "The Crypto Story": Postscript

Source
Sam Bankman-Fried's implausible PR strategy since his companies collapsed has been to claim that he "f**ked up", that it was simply a mistake and no-one had evil intent. Matt Levine has a post-FTX postscript to The Crypto Story entitled How Not to Play the Game providing a somewhat less implausible explanation. Below the fold I explain why Levine is still too generous to SBF.

Levine analogizes cryptocurrency in general and Alameda/FTX in particular to a game that got out of hand:
One imperfect but useful way to think about crypto is that it allowed for the creation of a toy financial system. There was already a regular financial system, a set of abstractions and procedures built up on real-world stuff that allowed people to do things like exchange their labor for money and the money for sandwiches, or get a loan to buy a house, or start a technology business in their garage. That system grew up over time, in path-dependent ways; it was fragmented and complicated and embedded in society and history. Different bits of it had different cultures and practices and were regulated differently; the regulation had also accreted haphazardly over time, and it could feel arbitrary and constraining.

And then crypto came along with a new set of stuff to do finance to. This stuff is so clean and new and shiny. It lives entirely on computers; you never have to worry about how to foreclose on a house or take delivery of 5,000 bushels of soybeans. And it lives on really user-friendly computers: The assets are created and sent between users on permissionless blockchains, and anyone who has a clever idea for how to trade them can implement it. The culture of crypto skews young and tech-savvy and optimistic; people want to try stuff, and they want everyone’s stuff to work.
Here is one of the many points where Levine is too generous, because (my emphasis):
Also, for most of its history, the overall value of crypto has kept going up, which meant it was pretty easy to make your stuff work. If your strategy was “Buy a lot of crypto and then do some mumbo-jumbo,” the mumbo-jumbo might have been good or bad, but you probably made money just from buying a lot of crypto.
Note the lack of agency here. Why did the "overall value of crypto" keep going up? Because the whales were using a fraction of their HODL-ings to pay celebrities, journalists, politicians and regulators via the revolving door to massively pump the stuff. And that in particular includes SBF.

In Levine's analogy, these young people were playing a game and all was well while "number go up, but then "number went down" and problems started:
if the game stops going your way, you might be tempted to reprogram it, to cheat. In traditional finance, there are exchanges and clearinghouses and prime brokers and market makers, and they tend to be separate companies serving different purposes. This is part of what makes the traditional finance system feel slow-moving and annoying. To trade, you need relationships with all these different entities; there is so much bureaucracy, so many contracts, so many people who can object to what you are doing.

In crypto it is common for one exchange to do all of these things, to run the exchange that matches trades and also the website that takes customer orders and also the bank that lends customers money and also the market maker that buys what customers are selling and sells what they’re buying. This, in many ways, will feel like a better user experience; the customer can go to one website that does everything. It is also a better experience for the finance people building the game: You can just think of the best possible way to trade and offer it to customers, without dealing with any middlemen.

But then, if your market-making firm stops making money on the exchange that you run, you might tweak how the exchange is run so that you can make more money from your customers. Or, if your market-making firm loses a lot of money, you might tweak how the margin-lending function works so that, uh, you can take a lot of money from your customers and “lend” it to your market-making firm. This will not be a very good experience for your users in the end. But it’s all a game, anyway, to you.
This is the main reason I say Levine is too generous. He writes (my emphasis):
What happened at FTX? “They stole the money” seems to be a true but insufficient answer. I think that part of the answer is that they found, and helped to build, a toy financial system, and they played with it. They didn’t take the game too seriously; they didn’t spend a lot of energy hiring accountants and compliance people, because that is not the fun part of finance. They built clever systems for margin lending and risk management, because it is fun to build an idealized trading system from scratch. But they also exempted themselves—Alameda—from that system, because it was just a game. In the real world, if you run a hedge fund and your balance becomes negative, the game is over. At FTX, when Alameda’s balance became negative, it got to keep playing.
Writing for publication on 30th December, Levine had access to the charges to which Caroline Ellison pled guilty, which start:
The United States Attorney charges:
  1. From at least in or about 2019, up to and including in or about November 2022, in the Southern District of New York, and elsewhere, CAROLINE ELLISON, the defendant, and others known and unknown, willfully and knowingly did combine, conspire, confederate and agree together and with each other to commit wire fraud, in violation of Title 18, United States Code, Seciton 1343.
and Gary Wang, which starts:
The United States Attorney charges:
  1. From at least in or about 2019, up to and including in or about November 2022, in the Southern District of New York, and elsewhere, ZIXIAO (GARY) WANG, the defendant, and others known and unknown, willfully and knowingly did combine, conspire, confederate and agree together and with each other to commit wire fraud, in violation of Title 18, United States Code, Seciton 1343.
and the SEC's complaint, whose press release says:
The Securities and Exchange Commission today charged Caroline Ellison, the former CEO of Alameda Research, and Zixiao (Gary) Wang, the former Chief Technology Officer of FTX Trading Ltd. (FTX), for their roles in a multiyear scheme to defraud equity investors in FTX, the crypto trading platform co-founded by Samuel Bankman-Fried and Wang. Investigations into other securities law violations and into other entities and persons relating to the alleged misconduct are ongoing.

According to the SEC’s complaint, between 2019 and 2022, Ellison, at the direction of Bankman-Fried, furthered the scheme by manipulating the price of FTT, an FTX-issued exchange crypto security token, by purchasing large quantities on the open market to prop up its price. FTT served as collateral for undisclosed loans by FTX of its customers’ assets to Alameda, a crypto hedge fund owned by Wang and Bankman-Fried and run by Ellison. The complaint alleges that, by manipulating the price of FTT, Bankman-Fried and Ellison caused the valuation of Alameda’s FTT holdings to be inflated, which in turn caused the value of collateral on Alameda’s balance sheet to be overstated, and misled investors about FTX’s risk exposure.

In addition, the complaint alleges that, from at least May 2019 until November 2022, Bankman-Fried raised billions of dollars from investors by falsely touting FTX as a safe crypto asset trading platform with sophisticated risk mitigation measures to protect customer assets and by telling investors that Alameda was just another customer with no special privileges; meanwhile, Bankman-Fried and Wang improperly diverted FTX customer assets to Alameda. The complaint alleges that Ellison and Wang knew or should have known that such statements were false and misleading.
Source
All of these were published on 22nd December, and they all show that the fraud started in 2019. The cryptocurrency markets peaked in November 2021, so Levine's idea that SBF and others only started cheating after the market moved against them is wrong.

Given that unregulated exchanges inevitably exploit their home-field advantage to capture profits that should rightfully accrue to their customers, it appears that the reason SBF set up the FTX exchange was that Alameda was missing out on profits by trading on exchanges such as Binance. By establishing their own exchange and, as the guilty pleas of Ellison and Wang admit, exploiting FTX's home-field advantage by enabling Alameda to front-run trades, avoid liquidations, and borrow unlimited funds, Alameda could rip off others the way they believed they had previously been being ripped off. The fraud at FTX wasn't a response to Alameda no longer winning at the game it was playing, it was the whole point of FTX right from the start.

On the other hand, Levine is on the mark at the end of his article when he writes:
But, paradoxically, crypto is much more reliant on trust than the rest of finance and business. It only works if people believe in it. There is no external source of value.[6] Crypto prices go up when more people become more interested in crypto; they go down when people turn away from crypto.

One thing that this means is that, if you are running a scam, you will be drawn to crypto. You are running a confidence game, and crypto offers the most efficient market for turning confidence into money. “These people just made these tokens up and sold them for money,” you will say to yourself. “How do I get in on that?” There are more sophisticated versions. “You make a box and issue a token and get some trading action; everyone marks to market, and then you can borrow against it and never give the dollars back” would be one.

But another thing this means is that, if you are in crypto, and you are not running a scam, you rely on the trustworthiness of everyone else in crypto. If the highly trusted operator of a big crypto exchange—the public face of trustworthiness in crypto!—turns out to be running a big fraud, you can say, “Well, I wasn’t running a big fraud” or “That guy’s big fraud says nothing about the underlying blockchain technology,” and, fine. But, like it or not, you are in a collective social project, and that project is crypto as a whole, and you will be judged by the other people in that project with you. If everyone thinks “Ah, yes, crypto, that’s for scams and crime,” that is bad for the project.
But he is wrong again when he writes:
It is not, by any means, the end of the project. One possible future for crypto is to return to Satoshi Nakamoto’s vision of trustlessness and decentralization. Decentralized finance, DeFi, has come out of crypto winter looking relatively good; in fact, open-source smart contracts on the blockchain do seem less likely to steal customer money than centralized exchanges. The centralized exchanges do keep attracting money, though; people want somebody to trust.
As I write, Molly White's Web3 is Going Just Great has tallied $11.915B in losses. Much of this was via hacks and rug pulls from DeFi, illustrating that far from being trusless and decentralized, using DeFi requires trusting the developers and managers.

Finally, Levine fesses up to why, despite himself, he seems forced to give fraudsters and scam artists the benefit of the doubt (my emphasis):
Over the last few years crypto built a toy financial system. That was an accomplishment, both in a technical sense (crypto found smart ways to do financial trading) and a social one (crypto attracted a lot of smart finance people). It is an accomplishment that I personally appreciate, since I love a clever financial system. But it is in important ways a bad place to start. A cleverly designed exchange for trading magic beans will never get around the basic problem that the magic beans don’t work, and people might stop believing in them. If crypto is going to work in the long run, it will need to prove its real usefulness outside of finance. Finding new ways to trade the tokens is fun, but it is not enough; the tokens have to mean something, too.

5 comments:

David. said...

From the No Shit Sherlock department comes Allyson Versprille's FTX’s Venture Capital Backers Face ‘Serious Questions,’ CFTC Official Says:

"“What kind of due diligence did they conduct?” Commissioner Christy Goldsmith Romero said Friday in a Bloomberg Television interview. “Why did they turn a blind eye to what should have been really flashing red lights?”

If a fund entrusts millions of dollars and then a year later has to write it off completely, it raises such questions, the regulator said."

David. said...

Matt Levine has an interesting take on Mango Markets and the Eisenberg indictment:

"There is a fairly popular view in crypto that it is not illegal when it happens in crypto, that crypto markets are governed by their own rules, that “code is law,” that if you do a thing that is not explicitly forbidden by the terms of the crypto platform you are using then no government should have any authority to stop you. ... It is not, I think it is safe to say, the view of US federal prosecutors. Their view is that this stuff is all just as illegal in crypto markets as it is in the stock market, that it is fraud, and that if it catches their attention they can and will bring criminal cases.
...
Avi Eisenberg, ... was arrested in December; ... He was charged with commodities fraud: He traded futures on Mango tokens, crypto futures are plausibly “commodity futures” under relevant US law, and manipulating them is straightforwardly a crime. The US Commodity Futures Trading Commission also brought a civil case against Eisenberg.

But the futures and tokens that he traded — MNGO, the governance token of Mango Markets, and related futures on MNGO — are just obviously securities; MNGO is essentially a share of stock in the Mango protocol. ... And so last week the SEC brought its own market manipulation case against Eisenberg
...
The complaint covers the same basic facts as the Justice Department’s and CFTC’s cases, but it also has a long analysis of why the Mango tokens were securities. This does require the SEC to say, or at least hint, that Mango’s developers were also breaking the law:"

The SEC is really trying to establish that almost everything in the cryptosphere is a security.

David. said...

More evidence that Levine is wrong when he analogizes FTX to a "game that gout out of hand" comes in Sam Bankman-Fried Charged in US With Bribing Chinese Officials by Ava Benny-Morrison and Chris Dolmetsch:

"Bankman-Fried is accused of authorizing the payment of $40 million to one or more Chinese government officials in order to get them to unfreeze accounts at Alameda Research, a Hong Kong-based trading firm affiliated with FTX, holding more than $1 billion in cryptocurrency."

David. said...

Molly White tweets the highlights of John Ray's 43-page First Interim Report, which include:

"[FTX] kept virtually all crypto assets in hot wallets... [FTX] undoubtedly recognized how a prudent crypto exchange should operate, because when asked by third parties to describe the extent to which it used cold storage, it lied."

And:

"FTX generally didn't use multisigs. When they did, they stored all of the keys together in one place, thus defeating the purpose."

Matt Levine's commentary on the report, FTX Lost Track of Its Money is worth reading too:

"Look, if you run a financial institution, and you sometimes finding an extra $50 million of assets lying around, you might say things like “ooh money” and “that’s hilarious” and “such is life.” Finding money feels good! But it is indicative of a deeper problem. “If I sometimes find $50 million of assets lying around that I lost track of,” you might think in more reflective moments, “is it possible that I will sometimes find $50 million of liabilities lying around that I lost track of?” Because that would be bad. In the event, FTX was undone because, and I cannot emphasize this strongly enough, IT FOUND $8 BILLION OF LIABILITIES LYING AROUND THAT IT LOST TRACK OF, whoops."

David. said...

Another of Maat Levine's beautifully written pieces on FTX is SBF’s Defense Will Be Tough:

"Alameda owed FTX a bunch of money, but nobody else did — nobody else got wiped out by huge price moves leaving it in an uncollateralized negative position to FTX — because FTX did have reasonable risk management systems. If you ran some unaffiliated hedge fund and you wanted to put up $20 of your own money and borrow $1 billion from FTX to buy some illiquid volatile speculative crypto token, FTX’s computers would say “absolutely not.” This was a point of pride at FTX, a thing that they advertised, a thing that they touted to regulators and Congress and that Bankman-Fried tweeted about.

But Alameda was allowed to do that: There were settings in FTX’s code that allowed Alameda to have an infinite negative balance, to borrow freely with no collateral."

As I wrote above:

"By establishing their own exchange and, as the guilty pleas of Ellison and Wang admit, exploiting FTX's home-field advantage by enabling Alameda to front-run trades, avoid liquidations, and borrow unlimited funds, Alameda could rip off others the way they believed they had previously been being ripped off. The fraud at FTX wasn't a response to Alameda no longer winning at the game it was playing, it was the whole point of FTX right from the start."