Malt Liquidity 63

Don't be Delta Neutral on Morals

10% Funded, 100% Reason to Remember the Name

Up until recently, I think it’s safe to say that the general assumption about Alameda Research was that they were good at trading (or at the very least, competent.) There have been rumblings about front-running and their relationship with FTX for a long time, of course, but there was nothing really concrete, and, if they were front-running, it would at least indicate some type of edge (as that isn’t technically illegal in the Wild Decentralized West.) In the last week, we’ve talked about the actual liquidity crunch and some potential fallout of what happened. Over the weekend, though, some details came out about FTX’s balance sheet itself - and personally, I think the story shifted heavily from “misusing leverage and getting caught out while trading” to “outright fraud.”

Some choice quotes:

Sam Bankman-Fried’s main international FTX exchange held just $900mn in easily sellable assets against $9bn of liabilities the day before it collapsed into bankruptcy, according to investment materials seen by the Financial Times.

This sounds bad enough on the surface, but given that FTX was responsible billions of dollars worth of customer deposits, this is begging to get blown up almost immediately. For all the vaunted bullshit spewed by the employees of the FTX-Alameda about probability and risk, it seems like even a simple understanding of tail risk would indicate that you, well, can’t do this.

The company’s biggest asset as of Thursday was $2.2bn worth of a cryptocurrency called Serum. Serum’s total market value was $88mn on Saturday, according to data provider CryptoCompare..

??? Unless there’s some definition of “total market value” that I’m not comprehending here, how exactly are you supposed to mark holdings of something that is actively traded at 22x the actual market value you could get assuming the liquidity of the entire market value? The more of a market you own, the less your holdings are worth relative to their “market value”. Wealth, net worth, and market cap are all calculations that ignore liquidity. Jeff Bezos might be sitting on $100b of amazon stock, but the cost of the liquidity to sell such a stake would render it worth magnitudes less. (Incidentally, this is a main criticism of private market valuations - a funding round that “values” a company at $50 bill based off of $500mm of transactions is unlikely to support the liquidation of anywhere near a sizable stake, which allows private investment funds to “mark-to-market” their returns at basically whatever they want.) To claim to be sitting on billions of this token (which happens to be another SBF project - even more self-dealing) is outright lying. The size of their holdings relative to Serum’s market value should be written down to close to zero - there was no shot they could liquidate even a single percentage of their holdings reliably.

A spreadsheet listing FTX international’s assets and liabilities, seen by the Financial Times, point at the issues that brought Bankman-Fried crashing back down to earth. It references $5bn of withdrawals last Sunday, and a negative $8bn entry described as “hidden, poorly internally labled ‘fiat@’ account”.

I’m no accounting expert, but from a cursory glance of public company financial statements, I can’t ever remember seeing an entry called “poorly internally labled ‘fiat@’ account,” let alone the number attributed to that account being in the billions. No wonder Binance’s due diligence was done within a day - this entry alone should prompt an immediate walk-out. I don’t even know where this would be placed on a typical balance sheet! If I had to take a stab at guessing, given that the entry is right above “Withdrawals on Sunday”, it’s an estimated hole of whatever future withdrawals FTX was expecting relative to their “assets”?

At least this heading is comforting.

There is also an obscure $7mn holding called “TRUMPLOSE”. There are no bitcoin assets listed, despite bitcoin liabilities of $1.4bn.

This entire balance sheet reads as if a particularly precocious 8th-grader skimmed the Wikipedia entry for Enron, shot himself up with meth, and proceeded to create “assets” out of thin air by creating a shitload of a token, selling a tiny percentage of it, and then marking their tokens at the value of the market they’ve just created just to borrow against it further. It’s borderline lunacy to even think about it this way, but then you realize they fueled this all with customer assets that are supposed to be safe. Honestly, this is grounds for an insanity defense.

Motivation

(Note: this section is based off of this thread that I highly recommend reading.)

One thing I am certain of is that nobody involved with FTX/Alameda is dumb. The interview process alone for the types of firms that people like SBF and Caroline (the Alameda CEO) worked at is extremely grueling, filled with probability theory, statistics, brain teasers, and IQ tests before you even really get to the “tell me about yourself” questions. If you’re one of the few that does get an offer, after a grace training period, your job safety is entirely dependent on performance. It’s not that easy to last in a trading job - quite a few of the people I’ve known in the space burn out over the course of a decade and look for something less strenuous.

No matter how good you are at trading, the total profit pool is somewhat out of your control. Each market making firm is competing with each-other for a share of a limited amount of money that is dependent on trading volume and volatility. The more volume and volatility, the larger the total profit pool is, and the more competitors will spawn. Conversely, when volume and volatility shrink, the lower the profit pool, and the harder it is for trading firms to sustain themselves. Infrastructure is ludicrously expensive when competing in the high-frequency arms race.

While Jane Street and their ilk always compensated employees well relative to the general populace, trading is an industry built on bonuses. You don’t know what your best year will be - all you can hope for is that you’re in a position to own a larger chunk of your earnings when your best years do arrive. You hope to cash out on a few big years to accumulate wealth.

The years of 2014-2017 were, save a few days here and there, ludicrously boring. Traders are moths drawn to volatility instead of light bulbs. I remember my own trading strategies for one of those years involved either sitting on an unlevered AAPL long or coin-flipping phase 3 trials on small-cap biotech stocks just out of a need for something to do. Keeping this in mind, it seems totally natural that these traders flocked to bitcoin (due to its volatility) and tried to replicate what they had learned at their firms. I myself knew a couple traders that had been recruited in the first surge in 2017 and jumped ship, one trader who successfully pulled off his own “Kimchi premium” trade, and I even interviewed at a couple crypto trading firms myself. I remember being struck by how unsophisticated crypto trading was relative to institutional trading then. I recall one trading interview for a firm of 6 people whose interview consisted of them flipping open an abstract algebra book to random pages and asking me theorems.

With this in mind, it seems highly likely that Alameda was one of the first shops to implement some institutional-type strategies. Spreads were extremely wide and a solid knowledge of how to provide liquidity and a decent code base probably provided for a lot of easy opportunities to pick off flow and arbitrage across exchanges. The biggest risk was probably counterparty risk. Furthermore, the actual amount you can get off in a lot of these trades simply isn’t that high. Even now, when I talk to people who have run arb strats in the past couple years, I generally get the sense that most coins support at most a deployment of low 6 figures for arb, if that. After a couple years, though, the profit pool had expanded sufficiently and the infrastructure had become just stable enough to allow actual institutions to invest into deploying their assets and IP into the crypto space. Markets become tighter and the first adopters, while they made some cash, simply couldn’t compete with a larger, better-equipped army. Crypto speed-ran the high frequency arms race, to some extent.

I generally think there are two major motivations for committing financial fraud. One is greed/hubris - “I know better than other people what to do with money, therefore I deserve it more.” The second is borne out of befuddlement - “I’m a smart person, and can’t believe I’m losing money! If I just had more to work with, I can make it back and then some.” Publicly, given the weird effective altruism cult mentality espoused by SBF, it’s easy to say it’s type one, and people who personally know him could probably attest to that if true. But as a trader, I’ve seen type two happen time and time again when objectively smart people become unable to turn a profit.

I particularly like the theory in the thread that Alameda had to operate at a loss to provide the image of liquidity on FTX. It goes a long way in explaining the catch-22 where they can’t stop unprofitable trading because it helps them raise funds, but the funds have to be used to bail out unprofitable trading. I think it goes one step further, though. At some point, anyone invested in private business knows that if it grows large enough, its finances will come under scrutiny, especially when trying to offload an ever-increasing-in-value stake in the company. So what do you do when you have a leaky trading strategy and an illusion of volume that drops off if you ever stop providing liquidity for better traders than you to pick off? Well, in the most extended bull market in history, you giga lever long and hope to plug up the money sieve. This seems to be backed up in the type of twitter threads we see from some high-level traders at Alameda, including SBF himself (where he seems to misunderstand the Kelly Criterion.) Eventually, however, the market turned, as it always does, and the bottom fell out. Even fraud requires liquidity to exit.