Since rising to the top of JPMorgan’s research ranks, Marko Kolanovic’s role has felt more supervisory — notes “unmistakably Marko,” if you will, felt less frequent.
On Wednesday, though, he returned to form for a piece that might be described as “vintage Kolanovic.”
There were several sections to his latest, all of which were notable. But I wanted to highlight his stark warning on the 0DTE options phenomenon, which I’ve covered extensively in these pages including, as it happens, on Wednesday+.
Marko took investors on a trip down memory lane to February 5, 2018, a day that lives in infamy. It was Jerome Powell’s first day in the big seat at the Fed, and what a session it was for markets.
On the heels of a hot average hourly earnings print which accompanied that January’s NFP report (released the previous Friday), the VIX ETN complex imploded under its own weight in an event traders affectionately (or not) remember as “Volmageddon” or “Volpocalypse.”
After the close that day, the stay-at-home short-vol trade — immortalized six months earlier in a New York Times article about a Target logistics manager who quit a career in big box retailing to trade VIX products from his living room — was vaporized.
On Wednesday, Kolanovic asked if the proliferation of very short-dated option trading could lead to “Volmageddon 2.0.”
“While history doesn’t repeat, it often rhymes, and current selling of 0DTE (zero-day-to-expiry), daily and weekly options is having a similar impact on markets” as what unfolded in the lead up to Volmageddon, he said, before recapping the dynamics that presaged the VIX product fiasco. To wit:
In February 2018, a sharp increase of market volatility was triggered by a collapse in inverse VIX products and fueled by further systematic selling. The episode was eventually called Volmageddon. The rise of inverse volatility products prior to Volmageddon started as a virtuous feedback loop of volatility selling. Selling the VIX directly suppressed the level of implied volatility (boosting performance of short volatility products), as well as indirect suppression of realized volatility (via gamma hedging of underlying options). The decline of volatility and intraday hedging also manifested as buying the dip behavior. As the strong performance of volatility selling became self-fulfilling, leverage and tail risk in these products increased. On February 5, 2018, leverage was such that an increase of the VIX resulted in daily rebalance (closing of short VIX positions) that overwhelmed the market liquidity and led to an uncontrolled increase of volatility. This in turn triggered further selling from various other systematic investors such as volatility targeters, gamma hedgers and CTAs.
After that, Marko sketched a version of what he said could go wrong with the 0DTE option frenzy which, he pointed out, is manifesting in daily notionals approaching $1 trillion.
“These options are net sold by directional investors, and supply of gamma is likely causing a suppression of realized intraday volatility,” he said, cautioning that “if there is a big move when these options get in the money, and sellers cannot support these positions, forced covering would result in very large directional flows.”
Below, find his description of what could go wrong, along with the visual depicting the notional volumes, which I’ll present without further comment.
These are typically low delta options that rarely get in the money, and their impact is mostly through volatility suppression and an intraday buy-the-dip pattern that results from hedging. However, if there is a big move when these options get in the money, and sellers cannot support these positions, forced covering would result in very large directional flows. These flows could particularly impact markets given the current low liquidity environment. For instance, by estimating of how much of these short-term options are net sold by directional investors (and hedged by dealers), if there is a large market move, covering of short-term option delta could result in intraday selling on a large down move (or buying on a large up move) on the order of ~$30bn (Figure 3). One should also take into account that these flows would trigger further one-way flows from monthly option hedging as well as volatility control strategies and CTAs (particularly in a case of a large down move).
Not that it is unusual for me to not understand the options-vol-gamma stuff, but the comments of McElligot and Kolvanic seem a little at odds?
McElligot seems to say 0DTE has some role in the high intraday index volatility, Kolvanic seems to say 0DTE is suppressing intraday volatility (until it doesn’t)?
I’m tempted to shrug that intraday volatility has no bearing on my process – but I suppose any sort of ___geddon or ___pocalypse will affect everyone’s process, whether we (I) have any idea what is going on, or not.
You are correct on the first point. Nomura’s data suggests a net buy imbalance on most days, indicative of weaponized gamma, which feeds intraday reversals that are monetized same-day. The effect of that would be close-to-close vol suppression that “masks” intraday chop. I won’t try to reconcile the two, but I did want to acknowledge your comment because it demonstrates a keen eye. Good catch.
Why does any dealer sell those things? Hoping the majority expire worthless?
Not for the first time, I am thinking maybe capital gains tax rates should rise exponentially as holding period shortens to days, hours, minutes, seconds, etc.