Learning from FTX

It is both shocking and disappointingly familiar what has happened with FTX. I met Sam Bankman-Fried for the first and only time at Huobi APAC’s Singapore headquarters in late 2018 (incidentally at the same party I met Trent Barnes, which would lead to everything I’ve done with Zerocap). The impression I got was of immense self-confidence and an obsessive need for profit. I was thoroughly impressed and a little envious. It’s people like this that combine an insatiable appetite with towering intelligence that are the “Kobe Bryants” of the trading industry, and that make the big bucks. That was also where I met up with Constance Kang and Darren Wong, who would later join Alameda and eventually become COO and (ex) CMO of FTX, and who by all reports was kept out of the loop.

After parsing through the scuttlebutt, this is what I can piece together on what happened:

  1. Alameda used to be primarily a market-maker that aggressively pursued advantageous fee structures. At some point they started an OTC desk, using their market making infrastructure to hedge out risk. 
  2. After FTX launched, Alameda provided most of the liquidity. In order to reduce toxic flow to FTX users, liquidations on FTX would be sent directly to Alameda as an OTC transaction. This helped FTX grow by reducing toxic flow to 3rd parties. Alameda may have been used as a loss leader to bootstrap FTX. After all, toxic flows are still toxic flows, and it looks like people were keeping a lot of profits on FTX. Given Alameda was essentially running a B book on FTX, they were either the other side of the trade, or had to hedge on Binance (after the fact, since price discovery will almost always happen faster on Binance), either of which is, statistically and in the long run, likely to result in a loss. But maybe SBF et al figured it didn’t matter, since they would either make it back on the growth of FTX’s valuation, or simply via transaction fees (aka Alameda maker rebates).
  3. FTX launches FTT and Alameda serves as market maker and probably does some price management. As a result Alameda builds long exposure to FTT. This is a marked departure from its strategy up to now, which is delta neutral, however as a prop trading firm it exists to express the views of its principals. Since these people, who also operate FTX, are all long FTT both literally by holding the token and indirectly since the performance of FTT is positively correlated to the growth of FTX, and also because the price of FTT goes up, it’s all OK and everyone is happy. 
  4. Liquidity pools start making money and Alameda starts running prop strategies where they get tokens at low prices from pre-sales or OTC transactions, shove them into a liquidity pool, and dump the tokens on exchanges. Possibly to help Alameda hedge risk, FTX launches DeFi perps, however Alameda is probably still long some or all of these tokens. 
  5. SushiSwap pulls its little stunt and SBF steps in to save the project. In retrospect this may be because SUSHI going down would lead to a massive loss at Alameda due to being long DeFi. SUSHI’s liquidity and price action on FTX is crucial to the recovery, so Alameda steps in as a market maker again, and probably supports SUSHI’s price and builds a long position. But it’s OK because the price goes up eventually. The newfound confidence in DeFi energizes the market and Alameda begins printing shipping containers of money. 
  6. The thing about farming is that, if you’re doing it the way Alameda was, it’s highly profitable, but also highly capital intensive. SBF faces consistent frustrations getting capital, while sitting on a treasure chest of customer deposits at FTX. Now it would clearly be illegal to just give money to Alameda. So instead FTX extends a loan, which Alameda collateralizes with FTT. On paper this is a perfectly above-board, mutually beneficial arms-length transaction. FTX can monetize deposits and can theoretically always sell the FTT to cover a default. FTX customer deposits are now powering Alameda farming strategies. 
  7. At some point Alameda starts playing little games with their tokens, where they use their brand, social media presence and the minimal float of some tokens to pump the price up before dumping. SBF literally describes his process on a podcast with Matt Levine. However, since it’s in their interest to have minimal liquidity for some time, risk builds as they’re holding more and more illiquid tokens while getting rich on unrealized gains. 
  8. Over the years Alameda and FTX also made over 250 venture deals in the crypto space. Some of these were no doubt just one leg of a farming strategy. However, at the end of the day it’s more directional risk, much of which can’t be hedged. The biggest one is probably their bet on Solana, which appears to be over $1b in total, the majority of which is locked, and which led to a slew of follow-on projects to keep their gains. Alameda’s CEO has admitted to taking out loans to make some of these investments, for which they had to later repay using FTX deposits
  9. To summarize: at this point Alameda has gone from a highly profitable, bloodless, technically minded delta neutral market maker/arbitrageur, to a stupidly profitable, degenerate and cynical prop trading firm with its balance sheet and activities dominated by farming, and backstopping significant trading activity on FTX (and therefore FTX’s valuation). It has long directional bets on a number of illiquid tokens, foremost among these FTT. 
  10. What happens after this is kind of murky. It’s possible that, being smart guys, the good people at Alameda knew from the get go DeFi would collapse eventually. Then they watched Margin Call and decided to copy Goldman and be first. So they think about the best strategy and decide to crash LUNA. They can build their position in secret and when the time is right, trigger a coordinated attack spearheaded by dumping on-exchange and assisted by FUD across social media. The rumor is that CZ knew what they were planning and warned against it, but they did it anyway. 
  11. However I’m not sure it was Alameda, unless they were unbelievably careless, because the LUNA crash could have blown a hole in Alameda’s balance sheet in the form of liquidations. Since FTX launched a bunch of altcoin perps, people with long positions could be wiped in a wide DeFi crash, and as Alameda takes the other side of these liquidations directly onto its balance sheet, it literally catches the biggest falling knives on FTX. I don’t believe they would be able to hedge this, because how could they? The main if not only futures lit pool is on FTX and who’s going to sell them that many LUNA put options? This alone would be enough to end a smaller trading firm, but what about the chaos in the other futures markets on FTX, all of which is toxic flow? This would make it a very bad idea for Alameda to crash LUNA. But who knows.
  12. After the LUNA crash, not only has Alameda taken a loss, it turns out a bunch of people they borrowed money from (eg. Voyager) are now insolvent. Allowing them to die would take down Alameda as well, or at least expose their financial situation. To cover up the loss, and to control the fallout from the LUNA crash, Alameda bails them out, with the help of a loan from FTX collateralized with FTT. It’s worth noting that of FTX’s $1.4b bid to acquire Voyager, only $51m was in cash: https://www.cnbc.com/2022/09/30/ftx-is-paying-51-million-in-cash-for-voyager-assets-court-records-.html
  13. Then it gets worse. A bunch of the illiquid tokens on Alameda’s balance sheet are dropping like a rock. Because the market is illiquid Alameda can’t convert them and can only watch as their value plummets, threatening their solvency. If Alameda goes down it would expose FTX as well. So FTX gives them another loan collateralized with FTT.
  14. Unfortunately this keeps happening. Alameda is starting to run out of FTT. But that’s OK because they own most of the circulating float. They push up the price, which increases the value of the collateral, which means FTX can lend more to Alameda. Free money has been made out of thin air! But it also means their FTT position becomes more and more concentrated, and that they’re unable to sell any of it for fear of crashing the price.
  15. CZ hears about this and freaks out. He wants to get out. He may also have been looking for revenge. So he sells his FTX equity, which SBF buys. Then he announces he’s going to dump his FTT. Alameda offers to buy OTC to avoid impacting the market but he declines, maybe to teach them a lesson. Unfortunately this sparks a general run for the exits. Alameda tries to defend the price (further concentrating their position) but eventually runs out of ammunition and FTT tanks. 
  16. Now FTX/Alameda is collectively holding a ton of FTT that is rapidly declining in value and worth less than the loans, no market to sell into, not enough inventory to meet customer withdrawals, and probably not enough to match their other liabilities.
  17. The rest is well known: they try to put together a bailout, fail, and declare bankruptcy.

I think there’s a lot we can take away from this. One thing that really stood out to me was the insidious allure of directional bets. Especially if they make money. I’ve seen this again and again, including in myself, where even the most meticulous trader can start to be convinced of their infallibility and start doubling down on a profitable directional strategy. It seems they can do no wrong and all they need is more capital and more leverage – until this is no longer true.

Already people are blaming the incident on “growing pains” and lack of regulation. But I don’t think this is fair at all. Crypto is not the same new asset class it was in 2014, or 2017. There are a lot of institutional investors and a lot of sophistication, and the “growing pains” excuse can’t be used forever. Morever, there are massive scandals like this even in highly regulated industries. LTCM famously lost over $4 billion in 1998 ($8.48 billion today), and Enron’s market cap of over $60 billion went to zero in a matter of months. The simple fact is that power corrupts, and so do profits.

However, the more I read into this, the more I felt Sam, Caroline and the rest of the tight-knit crew that lived together in the Bahamas were also victims. I get the feeling that, since they were very young, they have been told by everyone around them that they were smart. That they were gifted, and special. Nowhere fetishizes intelligence more than a quant trading firm and they were at one of the best, starting their careers at a place that is more selective than Harvard or MIT. 

Here’s the thing about seeing yourself as someone that’s smart, gifted and special: when it becomes part of your core identity and sense of self, you will try to be smart, gifted and special all the time. You’ll always be looking for the clever, unique solution that no one else thought of, because that’s your thing. It doesn’t help that, when trading, you’ll never capture alpha doing the same thing as everyone else. Sure, you can make some money off beta, but the whole point is to be smarter than the market and find patterns other people miss to get uncorrelated returns. And that’s what they did, again and again and again. They routed toxic flow to Alameda to help FTX grow. They thought they understood farming and poured massive amounts of capital into it. They thought of a loan structure that, on paper, was equitable to both parties. They probably thought that, if they understand trading and farming, then venture capital would be a piece of cake and jumped wholeheartedly into that too. And it worked! FTX’s valuation grew, the price of FTT went up, FTX monetized customer deposits, their big bet on Solana paid off and they made massive, massive profits from farming. If any of it was distasteful, they could distance themselves intellectually by saying the ends justified the means, and eventually this will benefit humanity. All of this became a self-reinforcing cycle, producing empirical proof that they really were special. But the thing about financial markets is that everything that people do, they do for a reason, and those reasons are usually written in red. No one outsmarts the market forever. 

I deeply empathize with their choices, because I’ve been the same way. Not to their degree, but when I was growing up, everyone told me I was smart too. I also work at a quantitative trading firm, and I’ve also tried to find a competitive edge in the crypto markets and for my fintech startup. And I’ve also found myself falling into the same trap, of looking for the hidden path when sometimes, the conventional answer works just fine. Lao Tzu knew what he was talking about when he said “If nothing is done, then all will be well.

Above all, I saw in their experiences two lessons that were incredibly painful for me to learn and accept, and that I hope they’re learning now. The first is that, in the long run, there is no strategy more profitable and no risk more important to manage than integrity. And second, sometimes it’s OK to make money the slow way.