On Thursday morning, following the systematic unwind that precipitated the worst day for U.S. stocks since February and the worst day for the Nasdaq since Brexit, Nomura’s Charlie McElligott penned a post-mortem of sorts.
“Take this sloppy Equities deleveraging move for exactly what it is”, McElligott wrote, before pointing the finger at “the violent move in U.S. real rates and the re-pricing of UST term premium.” Below is a visualization of that point, which shows that the rates selloff has been driven by reals and the term premium trade, a scenario that’s harder for risk assets to digest than an inflation-led bond selloff.
He also attempted to put some numbers on CTA deleveraging and documented the role option gamma hedging played during Wednesday’s rout.
While dissecting Wednesday’s drawdown is important to the extent we can derive clues about whether more systematic selling/forced deleveraging is in the cards in the days and weeks ahead, the more crucial point is this: What happened this week is still further evidence to support the notion that the low vol. regime was but a veneer of stability masking a potentially dangerous buildup of risk beneath the surface.
A simple way to think about the low vol. regime is to view it as a byproduct of the “Goldilocks” macro narrative. Synchronous global growth and well-anchored inflation allowed market participants to harbor an upbeat view about the global economy while still insisting that because inflation remained stubborn, developed market central banks had the cover they needed to keep policy accommodative.
But beyond that, the two-way communication loop between central banks and markets was the key determinant in encouraging the proliferation of the short vol. trade. That same communication loop helped transform “buy the dip” from a derisive meme about retail investors into a virtually infallible trading “strategy.”
Central banks only acted with the market’s consent and markets became aware of their role in the decision calculus. I will never tire of reminding you that this dynamic was best described by Deutsche Bank’s Aleksandar Kocic in a truly brilliant note from September 2015.
“Fed’s communication strategy, it is becoming clear, is an equivalent of what in theater context is referred to as Removing the fourth wall whereby the actors address the audience to disrupt the stage illusion — they can no longer have the illusion of being unseen”, Kocic wrote, before elaborating as follows:
An unalterable spectator becomes an alterable observer who is able to alter. The eyes are no longer on the finish, but on the course — what audience is watching is not necessarily an inevitable self-contained narrative. The market is now observing itself from another angle as an observer of the observer of the observers.
Kocic revisited that in August of 2016 as follows:
The fed is observing the market and the market is observing the fed; the Fed has no maneuvering space and the market knows that, and the Fed knows that the market knows, but, in order to preserve the underlying symbolic order, both sides pretend that the other does not know it — their gaze can never meet, there has to be an implicit agreement that this can never happen; it is in no one’s interest to call the other side on it.
Because of this, nothing is likely to happen in the long run, the most we can see is some short-term disturbance that is self-correcting. If the Fed attempts to hike and this turns out to be disruptive for the markets, its action will be corrected quickly, probably before the next FOMC meeting. Similarly, if the markets appear too calm and complacent, the Fed might attempt to squeeze in a hike in order to free some maneuvering space without significantly raising volatility or disturbing the markets. One way or another, the excitement can be only short-term.
That relationship is what allowed the low vol. regime to optimize around itself; it is also what killed off “choice” and replaced it with “selection”.
Market participants could not rebel against the prevailing order unless they intended to burn premium waiting for something (i.e., a sustained move higher in volatility) that was not only unlikely to happen, but in fact could not happen in the absence of other market participants joining in and agreeing to stage a coup.
Theoretically, one could “choose” to rebel, but that would entail underperformance as everyone else rode the carry gravy train. Are you really “free to choose” if rebellion assures your destruction? Not really, because at that point, exercising your freedom is tantamount to suicide.
And so, “choice” became “selection”. Something like this: “Here are some ostensibly different options for expressing the carry trade, and you can select from this list.”
At that point, there was no conceptual difference between trades. It’s all one trade, and each option is distinguished only by the amount of carry on offer. Here’s Kocic from a January note:
Through their communication with the markets Central banks, and the Fed in particular, have become “good listeners” with their decisions and actions made with markets’ consent. After years of this dialogue, the markets have gradually surrendered to the ever shrinking menu of selections that converged to a binary option of either harvesting the carry or running a risk of gradually going out of business by resisting. Not much of a choice, really. In this process, Central banks have reached a point of enormous power and control where market dissent is practically impossible. We believe that such levels of market control remain uncontested with anything we have seen in recent history and that the markets’ dynamics have never been further from that of the free-markets. Low volatility is a perfect testimony of that.
That state of affairs creates an environment that seems stable, but in fact is not. Here again, Deutsche’s Kocic offers the best exposition. This is from a June 2017 note (more here):
In general, there are three types of equilibria to distinguish: stable, unstable and metastable. The bottom of the valley is stable; top of the hill is unstable; a dimple at the top of the hill is metastable (Fig). Metastability is what seems stable, but is not – a stable waiting for something to happen. [An] avalanche is a good example of metastability to keep in mind – a totally innocuous event can trigger a cataclysmic event (e.g. a skier’s scream, or simply continued snowfall until the snow cover is so massive that its own weight triggers an avalanche).
That is how this situation has evolved.
Why trace it in such painstaking detail? Well, because you need that frame of reference to understand what the above-mentioned Charlie McElligott meant on Thursday when he said this in the same note cited here at the outset:
As higher real interest rates reset term-premia, the cost of leverage, cross-asset correlations and asset price valuation across this post-GFC era market structure, Minsky Moments are a simple fact of nature–STABILITY BREEDING INSTABILITY.
In the hours after Charlie’s Thursday missive made the rounds, U.S. equities of course suffered another harrowing bout of selling and later that day, he made a cameo on Erik Townsend’s MacroVoices podcast. Unsurprisingly, all of the above came up.
“If I’m really stepping back and talking almost more philosophically, it’s the bigger picture here is that a higher real interest rate environment is resetting term premiums and, with that, the cost of leverage, cross-asset correlations, asset price valuation — all of these constructs built into the post-crisis quantitative easing era are now ripe to tip over”, McElligott said, responding to Erik’s question about a prospective shift in the landscape.
“We’re seeing these rolling Minsky moments as the pseudo-stability of lower interest rates, flatter curves, and suppressed volatility breeds instability through the leverage that’s had to be deployed on strategies over the past few years as yield was chased”, he continued.
Later in the chat, MacroVoices Co-Host Patrick Ceresna asked McElligott to elaborate on the role systematic strats play in accelerating declines. Specifically, Ceresna asked Charlie about the Thursday note excerpted in the second linked post above. Here’s what McElligott said:
I know that various negative convexity strategies which have proliferated in the prior quantitative easing era — those lower yields, flatter curves, suppressed volatility — have led to the growth in systematic trend, risk control, risk parity, vol targeting, annuities, leveraged VIX ETNs. That’s all part of the market structure that we deal with.
And, to my prior point, as we transition from QE to QT (quantitative tightening), we are now going to be forced to see structural unwinds of some of these forces.
So, the systematic community and the role that they play in the market, the highest impact of unemotional selling in the most violent fashion typically occurs through the CTA universe, the trend universe.
One, because it’s so structurally leveraged. And, most importantly, it’s a shorter-term vehicle, shorter term sign.
On Thursday afternoon and also on Friday, several analysts were keen to note that while the CTA de-risking had likely run its course, vol.-targeting strategies would probably continue to sell in the days ahead. As you’re probably aware, the vol.-targeting crowd is slower moving. This group includes risk parity, a strategy that leans heavily on leveraged bond positions and depends, at least in part, on the preservation of a negative stock-bond return correlation (positive equity-rates correlation). Do recall that a flip in the sign of that correlation was one of the key worries headed into this week.
Stock-Bond Correlation Breaks, Presaging Locust Plague, Pestilence, Famine
The positive equity-rates correlation that’s been a fixture of markets since 2000 has disappeared and as you can see from the following visual, it’s a rare thing that the sign flips negative:
That sign flip prompted some folks to speculate that risk parity exacerbated the unwind, and while that could end up being the case in the event vol.-targeting strats start to de-risk next week, McElligott reminds you that due to the slower-moving character of those strategies, it’s not correct to blame them for last week. To wit:
But when the bond equities correlation breaks down, as it is currently right now, people will jump to the risk parity side of the equation, which, per our construct, is a much slower moving vehicle. Ours, particularly, uses a two-year window. So there is a little bit of false attribution in my mind currently within the institutional marketplace as far as trying to pin responsibility on the risk parity community, when, in my mind, the much more powerful short-term force in the market are CTAs.
On Friday afternoon, JPMorgan’s Marko Kolanovic noted that “the biggest selling pressure [on Wednesday] was from option gamma hedging.” He went on to postulate that as long as the S&P held gains on Friday, “it could result in a squeeze higher by end of the day from gamma hedging flows.”There was indeed a squeeze higher into the close on Friday.
In the MacroVoices interview, Ceresna asks McElligott to elaborate a bit on how those flows impacted the market on Wednesday and what he (Charlie) sees going forward. Here are some excerpts from that exchange:
We saw an enormous jump day over day with the S&P futures options and SPY ETF options cumulative, both delta and gamma, on the day. So SPX net delta moves down $460 billion. That’s a 0.1 percentile move since 2013.
The day prior, that net delta was negative $55 billion. So just impossible, almost nine times growth over the course of the day with regards to how much delta was kicked off for sale from the options community yesterday in just SPX and SPY.
And what that means from the delta side of things is that — and this is as of yesterday’s numbers — but S&P gamma is now at $24 billion per 1% move plus or minus. And those big strikes there are 2,800 and 2,750.
And I think, judging by today’s spasms where it looked like we were going to break out, and then it looked like we were going to break down, and it looked like we were going to break out, and then it looked like we were going to break down, those levels kept us pretty well pinned.
But the danger here is that, on a close below that pretty heavy open interest line of 2,750, the more we start slipping below, the further out of position the short gamma is. And the more it slips, the more you have to sell to stay hedged.
And that’s always the danger of the options market.
When you think about all of this, it’s worth noting that we’re in the blackout period for buybacks. That said, Goldman’s buyback desk had its biggest day since February on Wednesday thanks to ASRs. To wit:
Although Corporates remain in the earnings blackout period until 11/6, a lot of companies are on 10b5-1 plans which allow them to buy back stocks through the blackout windows. Many of these plans include scales which get more aggressive the lower stocks go.
McElligott has variously warned over the past several weeks that “peak buyback blackout” could leave the market “naked” (so to speak) as the largest source of demand for U.S. equities (the corporate bid) disappears. Here’s what he had to say to MacroVoices about flows from 10b5-1 plans:
I want to be as black-and-white on this as possible, and totally clear. If there was going to be a period of pullback with this tape, it was going to come in this two-week window where we are at peak buyback blackout.
And that is absolutely where we are right now. The vast majority of S&P sub-industry levels are at effectively 100% blackout as of this week.
Now, 10B5-1 plans allow corporates to buy outside of the blackouts, but with a number of limiting factors there. The bottom line: There still is a reduction in net corporate flow. That is a critical facilitator allowing this risk-off trade to really proliferate.
Just to reiterate, the key takeaway from this discussion (no matter who is having it), is that the interplay between modern market structure and the low vol. regime created a self-reinforcing dynamic that embedded an enormous amount of risk beneath a veneer of stability. Wednesday should be viewed through that lens.