Earlier this week, we noted that the steep two-day slide in equities that unfolded on Tuesday and Wednesday (mirrored – and then some – across the pond) was likely exacerbated by systematic flows.
As Nomura’s Charlie McElligott wrote Wednesday, there was “a much deeper and more sizable Short $Gamma position for Dealers in both SPX and Nasdaq” this week. On top of that, spot SPX fell below deleveraging trigger levels on Nomura’s QIS CTA model, meaning some of the sell flow probably came from trend-following strats.
Here’s what Charlie wrote on Thursday morning, recapping Wednesday’s action, for instance:
US Equities were very much in the crosshairs of CTA deleveraging, with the ISM “macro catalyst” causing a fresh “growth scare” / “shock-down” impulse (later aided by the Bernie Sanders news and the heightened odds of a Liz Warren nomination–no comment), which then triggered mechanical “accelerant flows” that moved-us deeper into a more extreme dealer “Short Gamma” positioning that too was clustered with CTA deleveraging levels being broken to the downside (as the 3m window flipped outright “short” in a number of our models).
As a reminder, SocGen’s derivatives strategists wrote last month that most large daily moves in stocks (where “large” means magnitudes bigger than 1.5%) since May came “when the previous day’s aggregate gamma estimate was negative”. Wednesday’s action fit that description.
Read more: All Of This Summer’s Big Daily Stock Moves Had One Thing In Common…
Fast forward to Friday and JPMorgan’s Marko Kolanovic is out with a new piece that takes a look at this week’s action in the context of the same systematic flows.
“In less than 48 hours, many major markets lost ~5% (nearly an average annual gain)”, Marko writes, before recapping the sequence of events that began with Tuesday’s ISM manufacturing miss.
“Later in the night, significant selling came in futures as the S&P 500 broke through 50- and 100-day moving average signals as well as 1- and 12- month price momentum”, he notes, adding that “a nearly identical sequence occurred with the Eurostoxx 50, which broke 1M, 3M, 6M price momentum signals as well as 50- and 100-day moving averages”.
That catalyzed what Kolanovic describes as “one of the largest CTA reversals (from long to short)”, which helped drive markets lower during the Tuesday-Wednesday rout.
Here’s an annotated visual which shows the 4.6% drop in S&P futures from the 2,994 highs made just prior to Tuesday’s ISM manufacturing release and the knee-jerk selloff on Thursday morning following the non-manufacturing miss:
More consequential than the CTA deleveraging was option hedging, something that comes up again and again.
“Even more importantly, due to the significant recent increase of put options outstanding, dealers were caught significantly short gamma in both indices”, Kolanovic goes on to say, adding that “this resulted in selling related to hedging index put options that is likely more important than the selling by CTAs”. Here’s a snapshot of where things stood on Thursday morning via Nomura’s McElligott:
So, just how much selling pressure did dealer hedging and CTA deleveraging exert this week? Well, quite a bit, in all likelihood.
“Technical flows likely drove more than ~$100bn of equities selling in a 48-hour period”, Marko goes on to say. “Going into this week, the setup for indices (CTA levels and option gamma) was vulnerable, and it didn’t take much fundamental selling to trigger much larger technical selling”.
This is reminiscent of a cascade effect that’s played out on a number of occasions over the past two or so years. Take the August 14 rout for example. That day, the Dow plunged some 800 points on recession fears tied to the inversion of the 2s10s curve. The breakdown of equity flows suggested more than half of the selling was attributable to systematic flows, a large percentage of which came from hedging dynamics. At the time, Kolanovic broke things down as follows:
~$75bn of programmatic selling, with ~50% of it coming from index option delta and gamma hedging, ~20% from trend-following strategies, ~15% from volatility targeting strategies and the remaining ~15% from other products (e.g. levered/inverse ETFs, etc.). While these outflows would have represented ~25% of futures daily volume, in an environment of low liquidity they can be a dominant driver of price action.
To give you another example, recall that on the evening of Sunday, May 5, Donald Trump broke the Buenos Aires trade truce with a tweet. Minutes later, Nomura’s Charlie McElligott sent out a client blast warning that depending on how things panned out, SPX/SPY consolidated gamma could flip negative beyond which dealer hedging could exacerbate moves or, as he put it, things “could get sloppy”. And indeed they did.
As far as where things stand now, Marko on Friday writes that “dealers are still short gamma, and both the S&P 500 and Eurostoxx 50 are right below all of the key technical levels breached earlier this week”.
That, in turn, means that “if the market can move ~50bps higher during the day, it could spark a significant rally” driven by CTAs re-leveraging and “the same put options that helped push the market lower earlier in the week”.
Notably, McElligott made a similar point on Thursday, albeit with a number of caveats tied to his recent “binary”, “crash-down”/”crash-up” discussion. To wit, from Charlie:
If we do again “settle-down” with regards to the “macro shock” catalysts (e.g. Oct China / US meetings go well, or no further escalation in Trump impeachment noise), the daisy-chain of forced Dealer hedges they have had to take-on (due to the stuff they’re short to client in VIX OTM upside, or from other Dealers hedging the same flows) will then again likely need to be “puked” as they decay into expiration…
This would likely come in the form of any of the following–unwinding VIX upside hedges, unwinding S&P downside hedges or simply covering of “short Spooz,” and setting-up the potential for another “crash-UP” phase, similar to the +120 handle rally in S&P futures witnessed into- and after- the August expiries…”
FROM HERE, this is where SEASONALITY again matters, and makes the case for “HIGHER” market after making “local lows” in coming weeks–ESPECIALLY with these new “shorts” building, dealer “crash” building and again VERY “low nets” all as fodder for a move higher.
US stocks rose sharply on Friday, although the S&P still logged a third consecutive weekly loss as political and trade tensions conspired with evidence of a slowdown in the domestic economy to undermine sentiment.
8 thoughts on “Marko Kolanovic, Charlie McElligott See Scope For Market Bounce After Quant Flow-Driven Rout”
On the light side ,picture a herd of Buffalo grazing in a quiet pastoral scene… They are stalked by your favorite plains tribe of Native Americans.. Three of the Buffalo are named Kevin , Marko and Charlie ( just picking random names in this case).. They are spooked closer to a huge drop off pretty vertical I would add.. They ran .. Which ones of the three do you suppose did what?
Anyway on the more realistic side this employment report is influenced by a discernible trend in corporate America to chop overtime pay literally the moment the computers recognize it (immediately) for example . Lots of variables and promises of green grass to those Buffalo who tended ( trended) to be influenced not by the realities of the past but dreams of the future… To quote a frequent commentator on this site you can get a 2% move by just saying Boo…
I am no quant analyst, and I don’t read Greek, but apparently when stocks go down because of bad news people and the computer programs made by other people sell more stock and this could make stock prices go down more – unless it doesn’t. But the risk is still there.
I am also no physicist, but when stocks go up this makes people and their robots want to buy more stock – and this could make prices go up more. Unless something changes. Then who knows? Systemic flows in motion tend to stay in motion.
Stocks go up and down, because reasons, and if you don’t know this, you are flying blind.
Never thought of it that way…..”systemic flows in motion tend to stay in motion” That is a key to some of this Thanks Harvey.!!!
…unless acted upon by an outside force…
Now I am really risking getting my head caved in by you know who…
But I need to say this anyway.
On Day One, the markets go down 0.3% and there are all of these super technical notes explaining why. Flows and gammas and deltas and all of it. It is all very systemic. If the S&P hits 2XXX, Skynet releases the nukes. The next day, the markets move up 0.3% and there are these other notes and they have all of this jargon bordering on nonsense explaining why. Maybe 2020 will be the third straight year where the S&P hits 3000. Me, the n00b, the idiot, I look at two days where nothing happened. Now we are at two years where nothing has happened. All of the ups and their complex reasons cancelled out all of the downs with all of their complex reasons, and I am here thinking maybe none of it mattered. More precisely, that none of it was determinative. Maybe it is just noise. Chaos.
I make fun of squiggly lines, but I have taken a math class. I have taken statistics classes. I know what a slope is. I know about correlation. I have taken science classes. I have taken natural philosophy. Where is the predictive power in this stuff? Where is the falsifiability? What does any of this stuff have to do with the price of tea in China? Why is Chipotle $823 a share with a p/e of over 135? Why do we have poor liquidity having exaggerated price movements when the Fed has been pumping out cash for a decade, billionaires have been getting tax cuts, hedge funds have been underpaying taxes and raising cash, and stocks are at an all-time high? If you are cash poor, sell some Chipotle stock. It is a $23 BILLION company on net revenue of $250,000,000.
Where was I? Anyway, at the macro-level, if it is not an accumulation of idiosyncratic micro-level things like company earnings, then I think the obvious headline news matter more. iPhone sales. Unemployment rate. Fed policy. Trade policy. Middle East wars, that sort of stuff. If that below the hood super-technical stuff mattered, I don’t think it would be publicly available. I believe the scientists would predict the future and beat the market with proprietary knowledge.
Why does my coffee tas…
Couldn’t agree more. It is more noise than insight.
Guys, the bottom line is that every single selloff of 1.5% or more on SPX over the past six months unfolded because of these dynamics.
That’s not an “opinion”. It is a fact.
I get that it frustrates some people to read these type of posts, but the thing is, three quarters of my readership are here only for this. They don’t care about Trump or satire or macro. They care about daily market commentary. And I’m going to give it to them not just because they want it, but because I enjoy writing it.
More to the point, though, some of this is just math. You’re literally arguing against math.
There’s not going to be any convincing you, which is fine, but trust me when I tell you that the dealer option hedging discussion is the opposite of “nonsense” — it is more “mathematical certainty”.
So no, it’s not “nonsense”. It’s the opposite of “nonsense”.
And Harvey, that comment about liquidity is incorrect. You’re talking about top-down liquidity.
Bottom-up liquidity is something completely different. It has nothing to do with “cash poor” or “selling Chipotle stock”. It has to do with the fact that market making algos aren’t required to make markets. Rather they are profit-making machines (literally). So, when volatility goes up, they adjust quotes, widen their bid/asks, reduce size or pull quotes altogether. That leads to exaggerated price action (i.e., volatility). As volatility spikes, it makes this situation worse.
That’s what “liquidity” means in this instance.
Again, readers should know that the comment above is not correct. It is a factually inaccurate comment disguised as an opinion.
The combination of fundamental data and market action brings to mind a frog boiling.
Ooops, guess I was not logged in.
– in the last year, global and US econ data has unquestionably gotten worse and worse.
– the typical Street strategist now concedes the US is in a “mini-recession” and ROW is entering “recession” (no mini qualifier) or getting close.
– the remaining and much-cited source of strength is the consumer, specifically the US consumer.
– US consumer data is “softening from a high level” – that is a generic characterization that describes everything from job data to confidence survey
– normally stocks don’t wait for the last area of positivity to go fully negative, before rolling over. In fact, by the time that happens, stocks are usually bottoming. Normality stocks respond to the second derivative.
– This econ deterioration is happening when the US govt is already pulling most of its stimulus levers.
– $1TR+ deficit means there isn’t too much more fiscal stimulus to give; Treasury is already starting to feel the limits of how much net UST issuance it can do, without risky impacts to yields and liquidity.
– By the end of Oct, the Fed will have delivered 75 bp cuts and will in effect have restarted QE; there isn’t that much harder the string can be pushed without setting off the fire alarm.
– Discretionary investors are increasingly moving to the sidelines, with cash allocations and recession expectations high. Defensives continue to outperform. That value-rotation got snuffed out pretty damn quick.
– Strategists (some) are saying the lagged effect of coordinated global easing will “soon” start lifting risk assets. I don’t think (most) investors are betting on that. Instead, I think most are simply afraid of waking up to a “trade deal” and being too defensively positioned.
– So right now, it probably makes sense to be quite defensive, and maybe even a little short risk assets if you play that game, but being way short is too dangerous when one Trump-Xi smooch can blow you up.
– That’s how it seems to me but maybe I’m a coward. Anyone disagree? You think the market reaction to a “trade deal” will be more like “sell the news”?
– The impeachment chaos, in and of itself, is probably more or less neutral for markets. Warren is a threat to billionaires, big tech and health care, but Trump is a threat to manufacturing and trade while having no prospect of delivering more business-friendly legislation, so it’s hard to say that stock indicies would clearly prefer one or the other. Warren + GOP Senate is probably the “best” outcome for the typical portfolio, and that’s not all that unlikely.
– But if the impeachment chaos has an effect on the trade war/deal odds, that would be huge, and if anyone is making clear calls on that, I’ve missed it.
– Suppose econ data keeps deteriorating, to the point that the US consumer data joins in the steep decline party. At some point, do bearish-minded investors stop “fearing” a trade deal because it’s too late for return of trade sanity to stave off the recession?