Malt Liquidity 62
Note: I am not a trained mind-reader
It just keeps getting bigger
Crypto exchange FTX lent billions of dollars worth of customer assets to fund risky bets by its affiliated trading firm, Alameda Research, setting the stage for the exchange’s implosion, a person familiar with the matter said.
FTX Chief Executive Sam Bankman-Fried said in investor meetings this week that Alameda owes FTX about $10 billion, people familiar with the matter said. FTX extended loans to Alameda using money that customers had deposited on the exchange for trading purposes, a decision that Mr. Bankman-Fried described as a poor judgment call, one of the people said.
All in all, FTX had $16 billion in customer assets, the people said, so FTX lent more than half of its customer funds to its sister company Alameda.
Yesterday, the number was about $8 billion. Today, it’s $10 billion. How levered was the FTX-Alameda complex?
As of Monday, Alameda owed $1.5 billion in loans to counterparties outside of FTX, the people said.
You know, I think what’s surprising about this is that everyone just sort of assumed that Alameda was good at trading. But it’s really looking like they just nuked $10 billion doing god knows what. So much for “risk neutral” philosophy, right? (It looks like Sequoia pulled the profile of SBF on the live site, but if you have ~15 minutes to waste, the archived article is linked in the sentence prior. I promise you it’s one of the most cringey things you’ll ever read.) But having been around quite a lot of people from similar market making firms to the backgrounds of Alameda traders, let’s maybe take a stab at figuring out what went wrong.
First, a quick reminder on what the core business of these firms is:
At its core, all market making does is facilitate transactions when there isn’t necessarily a buyer matched to a seller at the best bid/best ask price. Every transaction has to have an “other end” - if I wanted to buy Ford stock and had to wait for someone to come along who wanted to sell it precisely at the price I wanted to buy it at, transaction volume would be much lower, and prices would be much choppier, cause buyer and seller are in no way guaranteed to match. So all the market maker does is provide a bid and an ask on both sides, and when they are “hit” - one of those orders is filled - they try to seamlessly pair off an opposing trade, and the difference - the “spread” they make - is what makes this service worth providing.
The thing about capturing the spread in equities markets, though, is that it’s tiny - most stocks trade primarily a penny-wide, and although spreads on ETFs may be larger, the costs of trading the underlying relative to the ETF to successfully “arb” erodes away at the increased profit. Furthermore, every trade isn’t winning. Virtu CEO Doug Cifu said around the time of their IPO that “51-52% of their trades are profitable”. We can assume that the rest are scratches/small losses due to exchange fees, or cost the spread instead. So how does this turn into roughly 400 million of quarterly trading profit? Well, you lever up. It’s hard to know how much leverage is actually used, but I’ve heard some high frequency firms (not necessarily market makers) use anywhere from 8 to 17:1 leverage. Of course, there is a cap to how much capital you can deploy, based on number of signals, liquidity, volatility, and volume. It’s why the Medallion fund has a hard cap on the amount of assets it runs its strategy on. But equities markets are robust, and there isn’t really a lot of risk of products breaking apart with their underlying. SPY/ES arb has been around for ages, and with the multitudes of ETFs that have spawned over the last decade, arbing an ETF to its components is somewhat of a science at this point.
From what I’ve seen, Alameda’s most well-known traders didn’t seem to spend a lot of time at these market-making firms. While I’m not saying they’re inexperienced - clearly they know about and have done a lot more market making than I have - it does mean that they’ve only traded institutionally in one market regime - the ZIRP regime from 2009-2020 - and specifically some of the absolute least volatile years, 2014-2017. As we all know, during the pandemic, market conditions shifted rapidly, and then in the rising rates environment, they shifted rapidly again. The regime shifted significantly. And of course, as you spend more time in the industry, you learn how to adjust your thinking to various regimes. However, I’m not so sure Alameda did. Assuming that they were doing some sort of liquidity provision across exchanges and coins, which it certainly sounds like they did given the extent of the coins held on their leaked balance sheet, it almost feels like they took the same philosophy of leverage on ultra tight equity spreads and tried to apply it to a medium that, well, doesn’t support that. Crypto spreads are wider, the books are thinner, and exchanges even vary in prices significantly when high volume and volatility is being realized. My guess is that they tried to run similar strats to what they were doing prior - arbing coins against eachother is a good trade! they’re highly correlated! - at a leverage that just got turbofucked in one of the large selloffs of the past year. And then they went to FTX for a loan against all the illiquid coins they couldn’t sell without destroying their value. Of course, it seems like they lost money on other trades too - yield farming (probably with too much size), bailing out troubled crypto firms (and unable to recoup the value of the acquired assets), and acquiring a stake in Robinhood, for whatever reason - but blowouts don’t really happen like this unless you’re extremely levered. I guess I really expected more.
What’s the SEC been up to?
Given that the SEC doesn’t have jurisdiction over a non-US entity (FTX) with non-US clientele, it looks like the most the SEC can do is look into FTX.US and its ties to other SBF entities. Meanwhile, it seems to be slapping down random decentralized projects that raised funds by selling tokens. What does this say about their priorities? Well, let’s take a look at SEC v. LBRY.
Pretty much all cases involving token sales revolve around whether said tokens were unregistered securities. To determine whether an offering is a security, the courts developed a test called the Howey test, from a 1946 SCOTUS decision in SEC v. W.J. Howey, which states that “an investment contract … means a contract, transaction, or scheme whereby a person invests his money in a common enterprise and is led to expect profits solely from the efforts of the promoter or a third party.”
In LBRY, the Howey test was broken into three parts (“(1) the investment of money (2) in a common enterprise (3) with an expectation of profits to be derived solely from the efforts of the promoter or a third party”, with only (3) in dispute) to assess whether the LBC tokens LBRY sold to fund their ecosystem was a security. The token was obviously able to be mined, but was also supposed to be utilized to “publish content, create “channel[s]” that associate content with a single user, tip content creators, purchase paywall content, or “boost[]” channels or content in search results”, or do anything on the LBRY network other than view “free content”.
The SEC won summary judgment - a ruling made when a trial is unnecessary due to “no genuine issue as to any material fact” - by citing statements put out by LBRY that “led potential investors to reasonably expect that LBC would grow in value as the company continued to oversee the development of the LBRY Network.”
While the SEC has a bit of a history in going for lay-up cases instead of being able to land whales like Steve Cohen, it’s a little alarming to see summary judgment being granted - while ICOs were obviously securities, it seems that any token created for utility that has been sold as having value will be chased after and sued. It’s hard to see how any token with utility that has a fluctuating value can avoid being categorized as a security if the company makes any comment regarding its potential valuation. It really seems like we’re on the pathway to any tradable token being regarded as a security - the Judge obviously focused on the statements LBRY made promoting LBC, but he also talks about the “objective economic realities of the transaction.” It reminds me of an old adage about assessing the intent behind insider selling vs insider buying - when people sell, it can be for a multitude of reasons, but when people buy, it can only mean they think it will go up. One really wonders if $11 million of token sales should be the current priority of the SEC regarding the crypto space.