The Perpetual Motion Machine
- Tiago Figueiredo
- Mar 1, 2020
- 7 min read
The market may have severely underpriced the threat of the virus.
What a difference a week makes.
The spread of the COVID-19 virus outside of China coupled with the stunning rise of Bernie Sanders in the polls purged equity markets last week, resulting in the largest weekly decline since 2008. Bonds rallied, posting the best return since 2015 with yields down anywhere from 10-30 bps across the curve. Market participant's infatuation with duration was alive and well, although yield curves did steepen as the market began to price in more aggressive rate cuts. Oil prices were also down as the market repriced oil demand amid increasing city quarantines and travel restrictions. As a result of the price action, implied cross-asset volatility spiked, torching any yield-enhancing/income-generating strategies from the sale of volatility. The now faded dealer long-gamma position over the last few months helped bring momentum and risk-parity investors (who are generally short-gamma) into equities. These factors, in conjunction with a black swan tail event, like the COVID-19 virus, allowed the market to go from bubble to bust in a matter of days.
Money market traders have spent the last five days cornering the Fed.
The US yield curve was steeper throughout the week after the front end of the curve began to price in a rate cut for the Fed's next meeting in mid-March. The curve flattened once the outbreak became a reality in mid-march as a sign that inflation and growth expectations were falling. This week's steepner is the market's way of saying that the Fed has no choice but to cut rates now. The market is now pricing in a 100 percent probability of a 50bps rate cut at the next meeting in March. Meanwhile, 30-year yields hit a new all-time low.
It is not clear how a Fed rate cut will help stabilize global growth.
What happened this week was an amalgamation of the uncertain impact of the COVID-19 virus on economic growth and company earnings, triggering the familiar liquidity/volatility feedback loops I've mentioned in the past. What makes this time different is that it is not clear how a Fed rate cut will help shore up consumption during a pandemic. Sure, a Fed rate cut will help restore confidence in the market and ensure that borrowing costs and financial conditions more broadly don't become restrictive. However, the Fed can't cook up a vaccine to stop the virus. Let me put it another way, anyone who was afraid to get a plane isn't going to get on a plane now because stocks stopped falling. As Mike O'rouke put it, people aren't worried about buying a house or a car anymore; they are concerned about whether the virus will emerge close to them. This crisis will likely pass over the next few weeks or months but, that doesn't mean that people will all of a sudden buy 40 Starbucks coffees because they've spent the last couple months locked in their basements. With that in mind, any interest rate cuts will stay long after the virus is under control and should be supportive to risk assets. There have been some reports that the US administration is contemplating a targeted tax cut in the hope of stabilizing the situation. That's consistent with the approaches taken by China and South Korea and may indeed be more effective than a rate cut from the Fed.
Panic over a collapse in consumer demand sent oil prices tumbling 16 percent.
Apple had one of the worst weeks since 2008 as the company's stock was hit with a combination of concerns surrounding supply chain disruptions and consumer demand. These same concerns have hammered commodities and have created a stunning disconnect between stocks and growth-sensitive commodities (copper and oil). Indeed, high-yield energy bond spreads blew out as energy company's remain burdened by high debt loads, low commodity prices, and disappointing earnings. On the equity side, S&P500 energy stocks finished the week 5 percent lower than the broader index. Although OPEC (Organization of Petroleum Exporting Countries) will meet next week, even a production cut will not be enough to put a floor to oil prices if pandemic-esque headlines continue throughout the week.
What's going on with gold?
After going on an absolute tear two weeks ago, gold remained rangebound throughout this week after getting liquidated on Friday. That has puzzled some investors who view the yellow metal as a 'shit is hitting the fan" hedge. When you consider that the S&P500 erased four months of gains in just six days, colloquially, shit has hit the fan. So, why was gold down over the week? The rapid decline in equities has likely forced investors to sell their winners (gold) to cover margin calls on their losers (stocks). These selling pressures have probably kept gold prices suppressed.
The perpetual motion machine of modern market structure remains under appreciated.
To think this all started back in December.
Following the signing of the phase one trade deal and the UK election, many market participants were basking in good vibes going into the new year. Option delta on the S&P500 was reaching 90 + percentile going back to 2011, indicating that market participants were either reaching for upside in options or selling insurance on upside through call overwriting. The dealer gamma model, which assumes that all calls/puts are bought/sold by a market maker, was showing that dealers had a very long gamma position. A long gamma position means that market makers are selling as the market is rising and buying when the market falls. These flows act as a shock absorber in markets, leading to lower realized volatility. Any short gamma strategy, such as momentum or risk parity, uses volatility as a toggle to deploy leverage. As volatility falls, these systematic investors increase their exposure. The chart below replicates a simple risk parity strategy that rotates between equities and long-term bonds. We can see that risk parity funds increased their equity exposure going into the new year.
What about short vol?
Rolling down the term structure of the VIX Index has become a popular strategy post-financial crisis. When market makers are long gamma for prolonged periods, these strategies tend to perform very well as dealer hedging flows keep market volatility compressed. I've replicated the returns of the XIV short-volatility fund that went bankrupt in 2018 and show the relative performance of that fund to the S&P500 index below. The point of this is to illustrate how lucrative but dangerous these yield enhancement strategies can be. The more astute readers will realize that hedging these short volatility strategies is a harry endeavour. Hedging these types of strategies involve setting your original positions on fire due to positive feedback loops. It is these positive feedback loops that guarantee that whenever volatility increases, the VIX term structure will invert.
The COVID-19 tail event in South Korea/Italy collided with option expiry.
As fate would have it, plenty of call options expired the Friday before pockets of the COVID-19 virus emerged outside the US. That, coupled with some reactionary selling, pushed the dealer option gamma positioning into a net short. A short gamma position means that market makers are hedging with the market. Meaning that as the market falls, dealers are selling, and as the market rise, the dealers are buying. These flows exacerbated the initial sell-off, pushing spot and futures prices lower and triggering what was viewed as unthinkable a week ago; CTA deleveraging. Remember that CTA's (Commodity Trading Advisors - Momentum funds) use volatility as a toggle to deploy leverage. That volatility toggle was pushed higher by short-vol sellers hedging their exposure after being caught flat-footed basking in low-volatility for the better part of January.
High-frequency traders/market makers tend to stop trading when volatility spikes.
High-frequency traders play a significant role in liquidity in modern markets. Yes, their business model is predicated on front running your orders and using high-powered computers to buy the stock before you and sell it at a higher price but, their presence allows people to sell their positions almost instantly. Unfortunately, these carbon-based investors presence in markets is based on volatility. When volatility increases, it becomes too risky to front-run orders, and many pull the plug, quite literally. With that in mind, it's not surprising that top of book market depth in the S&P500 declined by nearly 50 percent on Monday and Tuesday, according to Goldman Sachs' Rocky Fishman. Although liquidity had not dropped as much in other episodes, thin trading leads to more violent moves. The chart below is from JPM and shows market depth in the S&P500 futures contracts. We can see that market liquidity never recovered from early 2018.
The perpetual motion machine.
So, to recap. Option expiry collided with pockets of COVID-19 outside of China to push dealers into short gamma. The hedging from dealers, which involved selling equities at lows and buying Treasurys/volatility at highs, drove the markets and volatility to even more extreme levels. At extreme levels, volatility control funds (CTA's and risk parity) started selling because volatility is their trigger. Unfortunately, volatility is also a trigger for market liquidity, which tends to dry up as high-frequency traders turn off their machines to avoid losses. That's the basic idea behind the vaunted liquidity/volatility feedback loop and the perpetual motion machine in markets. The chart below comes from JPM and shows the relationship between liquidity (y-axis) and volatility (x-axis).
How do we stop the feedback loop?
Similar to what happened in gold, as crash protection begins to increase in value, investors will likely start to monetize their positions to cover margin calls on their losses. The question is then whether investors choose to roll their crash protection to lower strikes. The chart below is an update to the dealer gamma model. The red line shows that the current spot is deep into short gamma. We can see a giant wall of puts scatted in 50 point increments between 3100 and 2800, with the majority of those puts expiring in the coming weeks. If these puts get rolled lower or in the same positions, it's likely dealers will remain short gamma. However, if these positions get rolled higher or flipped to calls, we could see dealer gamma flip positive and begin to act as a stabilizer to incoming news. More importantly, if news flow surrounding the virus gets better, markets could rebound swiftly as the same feedback loops that conspired to push markets lower can result in a rapid recovery. With that said, next week looks to bring more of the same, particularly following the beginning of what appears to be pockets of the virus in the US.
Tiago Figueiredo
Comments