The Weekend Crypto Crash: A Lesson in Illusory Liquidity

If you’re a normal person, you likely saw a decline in crypto prices over the weekend that, while severe, looks in line with 10-15 other double-digit percent declines per year, and not too unexpected given the 3% drop in equities on Friday. It wasn’t great, but it wasn’t too abnormal either.

If you’re a fully on-chain crypto degenerate trader, however, you witnessed armageddon. 

The timeline of events indicates that a single tweet from President Trump about tariffs on Friday, after the equity market closed, triggered one of crypto’s largest liquidations on record, exposing structural flaws in leverage, collateral, and stablecoin design. We may never know how big the liquidations were, given some of the centralized exchanges underreport futures liquidations, but we know it was at least 5-fold higher than any other day in history.  

This was a painful day for many levered traders, and I don’t want to make light of that. Even for unlevered investors, many tokens are still down 20% or more from their Friday closing prices. But there are a few immediate takeaways from this event that are more important than the price action.

1. This was largely a technical, or mechanical crash, and not a fundamental crash.  Fundamental crashes (like the FTX bankruptcy, Terra Luna/UST unwind, the 2008 mortgage crisis, the dot-com bubble, etc) tend to happen because of a real decline in earnings or cash flows, leading to defaults and bankruptcies and a mass exodus of investors. Whereas technical or mechanical failures (often called flash crashes) tend to be short-lived, have little to no follow-through, and can almost always be blamed on some sort of market infrastructure or leverage. 

In this case, while some are quick to point out the failures of certain exchanges and market makers, or blame insider trading from Washington’s inner circles, the obvious culprit was simply too much leverage on highly volatile assets that have much worse liquidity than many acknowledge, and whose margin requirements are calculated based on untrustworthy price oracles. While the CEO of Tether is certainly biased, he’s not wrong: 

2. It is obvious that many investors and traders in crypto do not understand the market mechanics that broke this weekend.  This is not the exchanges’ fault.  This is not the market makers’ fault.  In fact, it’s nobody’s fault.  It’s simply a giant misunderstanding of how markets work. I’ve been saying for 7 years that crypto trades more like high yield bonds than it does equities/commodities.  Just because crypto has exchanges that look like equity markets, with very visible pricing and 24/7 markets, doesn’t mean it has the same liquidity as equities.  Equities have way more traders and investors, much more concentrated liquidity (on just a handful of exchanges), way more market makers with deeper pockets, and implicit backing from the government via both swap lines to banks and regulation.  This creates much greater liquidity and less volatility. 

Whereas the high yield bond market has similar fragmented liquidity across multiple dealers, just like crypto has fragmented liquidity across different exchanges.  High yield bonds often look very liquid, but in times of stress, that liquidity evaporates, similar to what we just saw in crypto.  Let me give you an example:

Suppose there are 10 major dealers (dealers are market makers, like JPM, GS, BofA, Barclays, etc) who are all making a “5x5 market” in XYZ high yield bond at 88-89. That means that a hedge fund or mutual fund could sell $5M worth of XYZ to them at 88, or they could buy $5M of XYZ bond at 89.  In theory, that’s a pretty liquid market. If a single hedge fund simultaneously contacted each of these 10 dealers to sell $5M of XYZ, in theory these dealers might take down a total of $50M of XYZ bonds ($5M each) at 88, and that would appear to be a pretty deep market. The market makers (dealers) would be unhappy, because they just got run over by the hedge fund and now they are all trying to re-sell these bonds to a new buyer, but it could in theory get done that way.  But let’s say instead that this hedge fund decided to sell $5M to Dealer A at $88, and Dealer A then dropped his market much lower to 87-88 and every other dealer followed suit. And then the hedge fund decided to sell another $5M at 87 to the same Dealer A or a different Dealer B, and all of the dealers started to get spooked by the aggressive selling. And then the hedge fund came back looking to sell another $5M.  Eventually, at least some of the Dealers would say “No more”, and would just stop making markets on XYZ bond, or he would take the market down to like 80-82, or 75-77, far enough down where the original seller would stop trying to sell to him so aggressively, and at a price low enough to find a new buyer. The market makers’ job is to provide liquidity in normal markets by setting prices where they believe they can find equal buyers and sellers of the bond, or where they can at least offset their risk using other assets. But they are under no obligation to just blindly bid at the same price while getting run over by sellers. If there are no buyers or a really aggressive seller, his job becomes to protect his own P&L, not to provide liquidity. As a result, the market that seemed large enough to take down $50M of bonds could actually support only 10-20% of that volume, and the more accurate clearing price for $50M bonds may be much lower than the starting price.

This is exactly what happened this weekend.  While BTC and a few other tokens like ETH and SOL have pretty strong liquidity to support an onslaught of selling pressure, most other tokens cannot handle it. There aren’t enough real buyers in the market at any given time, and the price is artificially supported by market makers. But when selling pressure intensifies, the market makers just disappear or lower their prices aggressively. Now, that normally would be no big deal – if a bond drops 20%, or a token drops 20%, and you find out there are no more sellers at those lower levels, the price of the bond or token would start to rise again until an equilibrium is found. There may not even be any trades at that lower level. It could have been a completely fake price, essentially used as a decoy to try to find the clearing price. While this discovery process can be painful, and some funds may have to mark their bonds down and take a paper (unrealized loss), it doesn’t wreak too much havoc because no one would be crazy enough to use a lot of leverage on a bond that could instantly fall 20-30 pts, nor would there be auto-liquidations on a price that is clearly pretty fake. The dealer did his job trying to figure out where the clearing price is, even if it meant a temporary pause in providing liquidity.

But in the crypto world, trading futures is very popular, and using leverage is even more so.  You could even say the market demands it.

Source:  Twitter / X

That sure sounds like fun. Except you can’t trade futures without putting up collateral. And if you misjudge the volatility of the asset that you are trading OR if you misjudge the volatility of your collateral, suddenly you can face liquidations. Worse, even if the prices drop to levels that clearly don’t make sense just because market makers are running for their lives trying to find a clearing price and no new buyers stepped in, that fake price is still used to calculate your margin, and can cause cascading liquidations. The ATOM token, for example, fell to $0.01. Clearly that wasn’t a real price where a lot of volume would normally happen, but when there are no bids, and you are getting margin-called, the forced selling happens at prices way below normal.

Source:  Binance

And this is exactly what happened this weekend.  Many, many, many traders lost everything because of fake prices. There’s a reason no one trades levered futures on high yield bonds. The same should probably be true for most crypto assets. Fragmented liquidity means that if any one of the exchanges quote a price at a much lower level because they are seeing unnatural amounts of selling and their market makers are disappearing, even if other exchanges aren’t seeing the same selling pressure, it could result in the oracle price falling to a level that liquidates long, levered futures positions.  

 3. A lot of people are complaining that “low-risk trading strategies” shouldn't have gotten liquidated.  By low-risk, these people are referring to market-neutral funds or market makers who have both longs and shorts at the same time to remain “neutral”, or who use different venues to go long an asset on one venue and short on another. There’s a problem with this, though. These are not “low-risk” strategies. They may have lower directional price risk, but they do not have lower risk. Counterparty risk is always a high risk, which is what you're taking when you cross-collateralize across multiple venues or concentrate all of your assets at a single venue. Crypto trading has incredibly fragmented liquidity, and these risks are always prevalent. Leverage is always a big risk, which is why you're using it whenever you have both a long and a short position at the same time.   Those running these strategies were taking high risks, and hopefully they knew that. Again, it doesn’t mean this weekend’s flash crash was an unexpected or even a high likelihood outcome.  But it certainly wasn’t a non-zero outcome either.

4. It’s unlikely that “retail” got liquidated here.  As my colleague Sasha Fleyshman points out frequently, the crypto market isn’t really split into “retail” and “institutional”.  There are 3 types of crypto traders:

a) True retail
b) “Crypto-native degenerates trading 7-8 figure portfolios who do yield farming and trade huge sizes across DEXs and CEXs. 

c) Institutional

True retail did not get rinsed this weekend. True retail trades on Coinbase or Robinhood in small sizes, many only own BTC, ETH, SOL and a few memecoins, and they rarely use leverage. Most don't even know what a Perp DEX is. Many don't even know crypto really broke this weekend, and will log into their account on Monday and will start buying with prices on sale.

It was the group from (b) that really got hurt this weekend. This is the group that uses massive leverage, understands all the different DeFi venues, and willingly accepts the risks that got them rich in the first place via being early to protocols and farming for airdrops. This does not make this weekend’s losses better per se, but it's not exactly the same as a bunch of Uber drivers losing their life savings.

Said another way, no one material to markets' functioning likely got hurt, and therefore, there is no systemic risk or contagion. If a Degen falls in the levered forest, does it even make a sound?

5. As we’ve stated repeatedly over the years, crypto and blockchain rails are largely experimental.  Since a blockchain is designed to move assets, but 99% of the world’s assets are yet to be on-chain (i.e. stocks, bonds, real estate, gold), we can’t really look at what exists today and draw any major conclusions about the assets themselves.  Most will go away over time as traditional assets begin to be tokenized.  BUT, the experimental infrastructure that supports these experimental assets may be utilized. And this weekend was another prime example of how well DeFi protocols and Layer-1 protocols, worked. Large sell-offs serve as good reminders that DeFi has been stress-tested numerous times and continues to work extremely well.  Uniswap closed nearly $9B in trading volume, well above the norm, with no stress or downtime.  Aave operated flawlessly, automatically liquidating a record $180mm worth of collateral in just one hour, without any human intervention or downtime. Hyperliquid had 100% uptime with zero bad debt, had massive volumes, and generated record fees for tokenholders.   Wall Street and Washington don’t care if your Fartcoin got liquidated, they care that the infrastructure that allowed you to trade this dumb asset will work well with real assets one day.

6. Unfortunately the only way to fix the structural problems is to start adopting the same processes from TradFi that everyone in crypto hates — bailouts, circuit breakers, leverage restrictions, delayed settlement.  Sure, most traders have more collateral to post if they had time to react, and they could therefore unwind their positions more slowly and manually. And sure, most dislocations in price would be closed by arbitrageurs if they had more time. But the point of blockchain is real-time settlement and less middlemen, and these solutions move away from that. There’s no easy answer. Either don’t offer the leverage in the first place or start adopting tried and true TradFi protection mechanisms. 

So where does that leave the market going forward?  As I’m writing this, equity futures are up, because equity investors had 2.5 days off to process the Trump tariff news that drove stocks down 2-3% on Friday, and can see that Trump is already softening his stance from Friday.  

If Trump-Xi reach an agreement, there will likely be no impact aside from a leveraged flush out, which ultimately helps to build a stronger base with new hands.  This is the likely path given Trump’s track record with China and in other negotiations. 

If Trump-Xi do not reach an agreement, and 100% tariffs on China kick in by November 1st, perhaps the world is headed towards a recession.  Thus far the odds on Polymarket are showing this to be unlikely, and the de-escalation is being priced in. 

Assuming the U.S. / China tariff issues do not escalate further, this past weekend will ultimately just be another blip, transferring risk from the short-term, levered traders to the more patient investors.