New Research Identifies Factors Behind Wild Market Swings of the Covid-19 Recession
By Guido Baltussen, Robeco Asset Management
In the last week of February 2020, as coronavirus surged, the U.S. stock market crashed by more than 10%, signaling the beginning of the COVID-19 recession. Market volatility rose, with the CBOE Volatility Index surpassing 82, its highest end of day closing, on March 16, 2020.
One major factor contributing to this increase in volatility was hedging by traders with short gamma positions. Gamma measures how much the price of a derivative accelerates when the underlying security price moves. Market makers in products with gamma exposure, such as options and leveraged ETFs, are commonly net short these products. As a result, they have to buy additional securities when prices are rising, and sell when prices are falling to help to ensure their positions are neutral even as the value of the underlying changes. Similar hedging activities have been carried out by other market participants, such as with dynamic hedging with portfolio insurance, for a long time. This trading/hedging in the direction of the market price movement can further exacerbate market swings and serve as an example of extreme market intraday momentum.
Market intraday momentum is a phenomenon whereby the markets tend to burst/continue to move in the same direction in the last half hour of trading, as they had been moving prior in the day. If, for example, the market is up from the previous day’s close, the theory of market intraday momentum suggests that it most likely will go up further in the last half hour of trading. If the market is moving down throughout the day, it will likely continue to go down further in the last half hour.
As both a researcher and an investor at an asset management firm, I and several of my colleagues had observed, for years, in the markets the following:
1. Increased volume and liquidity at close, with as much as 30% of daily trading toward the close, depending on the market
2. The market(s) appearing to structurally go up or down in the same direction at the close depending on the direction it was going during the day
We observed passive players (including ETFs and indexes) being active, and wondered: were the markets being efficient or systematic? Why did they tend to drift in the same direction in the last minutes? And was this phenomenon more pronounced at certain times over others?
In “Hedging Demand and Market Intraday Momentum,” Zhi Da, University of Notre Dame; Sten Lammers, Erasmus University Rotterdam; Martin Martens, Erasmus University Rotterdam, and I set out to explore these and other questions. We hypothesized that if the market contains a lot of short gamma exposure, intraday momentum with market makers hedging in the close would be present as a result. And, if the market is not short gamma, then there would not be a need for market makers to hedge in the close, and hence no market day momentum would be apparent.
To prove this theory, we examined the S&P 500, as the biggest, most widely tracked futures market, and 60 futures on equities, bonds, commodities, and currencies.
We looked at respective tick data on every major contract for every trade for over 45 years from 1974-2020. Additionally, we leveraged the TickData dataset and OptionMetrics data on options, going back to 1996. We have used OptionMetrics for other research as well and find it to be a great resource for which no comparable data set is available.
We took all data for the S&P 500 and other American equity uses, like the NASDAQ and Russell 2000. We performed a simple regression analysis to assess the markets at one half hour before the close, as well as at market close (3:30 and 4 p.m. for the S&P 500).
We determine that returns during the last 30 minutes before the market close are positively predicted by returns during the rest of the day. We document strong market intraday momentum in every asset type, market, and time period.
Why does the market tend to drift in the same direction in the last half hour?
As alluded to earlier, we conjecture that this has to do with the size of passive players of leveraged, or short ETFs being incentivized to move in the same direction as the rest of the market most of the time, because they are short gamma.
We know that on average market investors buy asset put options for market protection, especially as the market crashes. That means the other side of the trade, or the market maker (banks, hedge funds, etc.) hedge the risk or sell/short the put for the investors, and are, in essence, short gamma. Option market makers need to trade in the same direction as the underlying movement of the S&P 500 index if they have negative gamma exposure. The more negative their gamma exposure, the more aggressively they have to trade.
We seek to link this net gamma exposure (NGE) as a major force driving market intraday momentum. We use OptionMetrics to assess all options contracts outstanding at the end of every day from 1996-2020, measure options on them, and aggregate the short gamma exposure of market makers. Using a direct proxy of their negative gamma exposure, we confirm that market intraday momentum is present for the index when NGE is negative and becomes stronger when NGE becomes more negative.
What we find most interesting about these findings is the strength of market intraday momentum. We document a strong stylized fact for market intraday momentum that is larger than with other factor premiums. While we clearly show with OptionMetrics that market intraday momentum is absent when there is no hedging demand/the market doesn’t seem to be short gamma, the contrast is big, and there does, in fact, seem to be a very strong effect when the market is short gamma.
We observe this pattern to be especially strong in times of market crash, such as in March 2020 and in October 1987—interestingly, those times that investors lose on the equity portion of their portfolios and like to see returns coming from other parts of their investment portfolios. While passive investment options were not yet available in 1987, we show how portfolio insurance was similarly leveraged (in replicating options trading) contributing to the crash during that time period.
We document strong market intraday momentum in every asset class, in bonds, commodities futures, for every market.
How might investors leverage this information? What we now see happening, as a result of increased awareness of this phenomenon, are three effects:
- Investors might use intraday momentum as a smart way to trade. If one wishes to buy the S&P 500 index, for instance, he/she can buy now, or if he/she knows that if the market is up, it probably will continue to go up (or if it’s down, it will likely continue to go further down in the close), he/she can use this information to expose market intraday momentum to potentially trade to his/her advantage.
- Investment banks are increasingly offering exposure to options markets, with short and long volatility, and considering performing delta hedging, normally done at close, at earlier times in the day.
- There has been an increase, in the last six months, in investment banks selling intraday momentum as a standalone strategy for clients in which to directly invest. Goldman Sachs and Morgan Stanley, for instance, offer intraday momentum indexes.
In conclusion, what happens in the last half hour of the day is predictable based on what happened earlier in the day. While this pattern of market intraday momentum does not seem to align with the notion of an efficient market, it does offer opportunities for investors.
Guido Baltussen is Head of Factor Investing /Co-Head Quant Fixed Income at Robeco Asset Management and Professor of Finance at Erasmus University Rotterdam. He is responsible for a wide range of Robeco’s quantitative strategies, and as academic specialized in Behavioral Finance and Financial Markets. His work has been published in highly ranked academic journals, like the American Economic Review and Journal of Financial Economics.