So what happened on 2/5?

As time is passing by with more data coming out, it’s becoming increasingly more clear that the violent sell off had something to do with Bitcoin ETFs amidst one of the most violent days in the capital markets. We know this because IBIT had record trading volume (2x prior high at 10Bn+, wow what a record), and historic options volume activities (see below - highest contract count to date since launch of the ETF). Perhaps a little unusual relative to past activities, we can also observe the options activities were led by puts (vs calls) given the trading volume imbalance. (More on that later)

At the same time, we saw incredibly tight correlation of IBIT price action to software equities and other risk assets over the past few week. Goldman’s PB desk also published that Feb 4th was one of the worst daily performance events for multi-strategy funds, a z-score of 3.5. This is a 0.05% event, and 10 times rarer than a 3-sigma event. It was catastrophic. It is always after these types of events that risk managers at pod shops step in to ask everyone to de-gross indiscriminately and urgently. That would explain then why Feb 5th was such a blood bath as well.

For all the record setting activities and direction of the price action (down 13.2%), we would most likely have expected to see a net redemption number come through. Using historical data, like Jan 30’s record -$530mm redemption on IBIT when IBT was down 5.8% prior day, or February 4th’s -$370mm on a losing streak, it would have felt plausible to expect at least $500-1Bn in outflows. But in fact, the opposite happened - we saw broad based net creations: IBIT saw approximately 6 million new shares in creation, driving 230mm+ in AUM. The rest of the ETF complex added flows as well, counting at $300M+ inflows and going.

This is a little perplexing. It’s marginally imaginable that the outflows could have been reduced by the strong price action on Feb 6th, but to turn it into a net creation positive output is an altogether different story. It means that there are likely a confluence of various factors playing together here, but that they do not paint a singular narrative. Based on what we now know, there are several assumptions you can make, on the back of which I will propose my hypothesis:

• the Bitcoin sell off likely touched a multi-asset portfolio/strategy that is not exclusively crypto native (it could be a multi-strat hedge fund as noted above, or it could also be something like Blackrock’s Model Portfolio business that allocates between IBIT and IGV and had to auto-rebalance given sharp moves)

• the acceleration in Bitcoin sell off likely had to do something with the options market, particularly to the downside

• The sell-off did not result in end outflows of Bitcoin assets, which means that this was mostly driven by “paper money complex” activities led by dealers and market makers running generally hedged positions

Based on these facts, my current hypothesis is as follows:

1. The catalyst to the sell off was that there was a broad based deleveraging across multi-asset funds/portfolios due to the high downside correlation of risk assets reaching statistically anomalous levels.

2. This in turn led to a violent deleveraging that included Bitcoin risk, but a lot of this risk was in fact “delta neutral” hedged positions like the basis trade or RV vs crypto equities or other types that generally tend to box the residual delta across the dealer community.

3. This deleveraging then caused some short gamma to come into effect that compounded to the downside, which required dealers to sell IBIT, but because the sell off was so violent, the market makers had to net short Bitcoin without accounting for inventory. This creates new inventory, which reduces the large outflow expectations.

4. Then we saw positive IBIT flows come in 2/6 by IBIT buyers (the question is what kind?) taking advantage of the dip, and that created an additional offset to what would otherwise have been a small net outflow.

First, I tend to believe that the catalyst was driven by the software sell off given the tight correlation that it is exhibiting vs even gold. Take a look at the two graphs below:

This makes sense to me because gold is broadly not an asset that is held by multi-strat funds that are part of funding trades, though it could be part of RIA Model Portfolios, so this checks out to me that the center of drama are more likely multi-strat funds.

And then this starts to make even more sense on the second point, which is that the violent deleveraging includes hedged Bitcoin risk. Let’s consider the CME basis trade for example, an all time favorite of the pod shops:

Take a look at the full data set with the CME BTC basis on 30/60/90/120 days since January 26th through yesterday (h/t to the best research guru in the industry @dlawant). It can be observed that the near-dated basis trade jumped from 3.3% on 2/5 to a whopping 9% on 2/6. This is one of the largest jumps that we have personally observed in the markets since the ETF launch and speaks to what most likely occurred: a de-grossing of the basis trade by directive. Think about the Millenniums, Citadels of the world who had to forcibly unwind the basis trade (sell spot, buy futures), and given how large they are in the Bitcoin ETF complex, you can see how they would blow this out. I’ve written my hypothesis on this here.

This brings to the third leg. As we know understand the mechanics for IBIT to be sold in a broad deleveraging, what is accelerating the downside? One possible “fuel” to the fire is structured products. While I don’t think the structured products market is nearly big enough to be solely responsible for the sell-off, I think it’s entirely possible that when everything is freakishly and perfectly aligned outside of what any VaR models could ever tell you, it can be the acute event that creates a cascade of liquidation style behavior. This immediately brings me back to my Morgan Stanley days, where KI put barriers can be devastating in causing situations where option delta increases larger than 1, something that Black Scholes does not even contemplate for vanilla payoffs.

Take a look at one of the notes JPM priced in November last year. You can see the barrier is right at 43.6. If the notes continued pricing into December when Bitcoin went lower by 10%, you can see that there would be a lot of barriers in the 38-39 range, which brings us to the eye of the hurricane.

In these barrier breaches, if the dealer hedged the Knock-In risk using some combinations of being short puts, the gamma moves so fast especially with negative vanna dynamics that if you are dealer you have to sell the underlying asset aggressively into weakness. And that’s what we saw with IV imploding to record levels, almost touching 90% that would read to catastrophic level of a squeeze that it’s then possible that dealers had to short IBIT to a level that ended up creating net new units. This part requires a little more imaginative stretch and hard to ascertain without more spread data. Yet it is entirely possible given the record volume traded that the APs were involved in that effort.

Now combine this negative vanna dynamic with the fact that because volatility has been low, we’ve generally seen put buying behaviors by clients in the crypto-native land over past few weeks. This means that crypto dealers are also short gamma naturally as well, and basically sold options too cheaply relative to what the outsized moves would eventually be, exacerbating the downside. You can see that positioning imbalance below as well where dealers were mostly short gamma on puts from 64-71k range):

That brings us to 2/6, when Bitcoin made its heroic 10%+ recovery. Here what we can point to be interesting is that the OI interest in CME expanded much faster than Binance’s (h/t again to @dlawant who looked at the hourly snaps to get it to 4pm ET to line the below up). You can see the drop from 2/4 to 2/5 as the OI collapsed (this again confirms that the CME basis trade was being unwound on 2/5), but that perhaps it came back yesterday to take advantage of the higher levels, neutralizing the effect of the outflow.

This puts it all back nicely together now, where you can now imagine IBIT is generally flat for creates/redeems because the CME basis net recovered, but the price is lower because Binance OI has collapsed and that means a lot of the deleveraging likely came from crypto native short gammas and liquidations.

So that’s my best theory on what happened 2/5 and 2/6 thereafter. It makes some assumptions and it’s not entirely satisfying in that there is no ‘culprit’ to blame for what happened (like an FTX). But the punch line is this: the catalyst came from non-crypto tradfi derisking, which happened to push Bitcoin down to a level where short gamma accelerated the downside due to hedging, but not directional, activities which in turn led to the requirement for more inventory- this then reversed quickly on 2/6 for tradfi market neutral (but unfortunately not for crypto directional). As unsatisfying as this may be, it’s also perhaps comforting to at least determine the possibility that yesterday’s sell off had nothing to do with 10/10.

Yes, I do not believe what happened last week is a continuation of deleveraging from 10/10. I read a post that alluded to the possibility of the debacle being a non-US HK based fund that is involved in a JPY carry trade that went wrong. There are two substantial holes to this theory. The first is that I simply do not believe there is a non-crypto prime brokerage that would service the complexity of the multi-asset trading and also provide a 90 day cushion to meet margin deficiencies that would not already have been underwater by a tightening of the risk framework. The second is that if the funding carry was used to purchase IBIT call options to “get out of the hole” then a decrease in Bitcoin would not necessarily cause acceleration to the downside - the options would simply just go out of the money and its greeks dead. This means that the trade had to involve downside risk, and if you were shorting IBIT puts while being long dollar yen carry, well - then that prime brokerage deserves to go out of business.

The next several days will be really important, because we’ll see more data towards whether investors are taking advantage of this dip and creating new demand which would be very bullish. For now, I am quite heartened by the potential ETF flows as I continue to believe that authentic RIA style ETF buyers (not RV hedge funds) are diamond hands, and there is so much institutional progress on that front that we are seeing across all the work that is being done by the industry and my friends at Bitwise. To see that, I am monitoring net flows that come without the expansion of the basis trade.

Lastly, it also goes to show you that Bitcoin is now integrated into the financial capital markets in a very sophisticated way, which means that eventually when we are positioned for the squeeze in the other direction, it is now going to be more vertical than ever before.

The fragility of tradfi margin rules is Bitcoin’s antifragility. Whenever the melt-up comes the other way, which is inevitable in my opinion now that Nasdaq increased the OI limits on options, it’s going to be so fantastic.