- Weekend effect exists in the VIX
- Does not reflect in other volatility products
- /ES returns on Sundays are statistically significant
The VIX index is something referenced by many, but understood by few. The CBOE’s Volatility Index has acted as a fear gauge for the market since its inception in 1993, and still remains as the premier volatility product in the market. By tracking the volatility of 30-day at-the-money SPX options, the VIX is a great proxy for the expected volatility in the market.
One of the common misconceptions of the VIX is that when the SPX goes up, the VIX goes down at the same rate. The correlation between the returns of VIX and SPX is -0.53 (Spearman R), which is a light negative correlation indicating some semblance of inverse movement between the two indexes. It has long been thought that there was a statistically significant return on Mondays from the VIX versus other days of the week. Some have debated whether this would appear in other volatility products such as the UVXY (which tracks short term volatility of the VIX at 1.5x the daily return) and the VXX (which tracks short term volatility of the VIX, without the multiplier).
If this phenomenon exists, how could retail traders better understand it and benefit from it?
An Introduction to the Weekend Effect
The weekend effect, which is defined by the change in Friday’s close to Monday’s open, has been researched extensively in academic circles and private funds for decades, but in more recent literature has been thought to have disappeared. Very little public facing research exists examining other volatility products, and fewer incorporate futures data, and even fewer have researched the movements associated with the COVID-19 driven market of the past 2 years.
The VIX index relies on a calendar year computation, rather than a trading year computation (365 days vs ~252 days) as volatility can theoretically happen at any time with news cycles, world events, and business information released after market hours on weekends.
Because of this, the VIX behaves somewhat weirdly (statistically) on Friday’s heading into Mondays, with Monday’s open being the observed outcome.
This report analyzes price data for VIX, UVXY, VXX, SPX, and /ES from January 1st, 2020 to April 15th, 2022.
A note regarding the VXX: Though the VXX is now decoupled from the VIX index, for a large portion of the dataset, it was synchronized with the VIX index, so roughly ~95% of the VXX data will reflect the VIX’s behavioral patterns.
Does the weekend effect still exist, if so, does it appear in other volatility products?
Let’s start by plotting the average daily returns.
The VIX returns higher on Monday than other days. These changes are not reflected in the UVXY or the VXX. Average daily returns for GSPC (SPX) were positive overall. However, averages don’t tell the entire story, so let’s run some statistical tests.
We can use the Kruskal-Wallis H-test that compares the ranked sum of each member in different groups against each other, and does not assume equal variance between means.
What does that mean in English?
The Test Statistic is saying the higher the value, the more difference there is between groups, and the p-value tells you the significance of that difference. If the p-value is <.05, there is a significant difference between groups.
So now that we have this table, where do we go? Since the test statistic for the VIX is higher, and the p-value is way less than .05, we know that there is something there.
Let’s check what the data looks like if we compare VIX Monday returns against the other days of the week with standard t-tests.
VIX on Monday compared to other days
This is confirming the weekend effect on VIX is still, in fact, real and statistically significant.
So, to answer the first question: Does the weekend effect still exist, and can we see it elsewhere?
Yes, the weekend effect is still a very real phenomena, but we cannot see its effect in other volatility products, which means it is most likely non-arbitrable and there is no conclusive edge to trading other volatility products to benefit from the VIX’s weekend effect.
Does the weekend effect exist with /ES futures.
With the weekend effect still being present within the VIX, does it translate to /ES futures being statistically different on Sunday evenings/Monday mornings?
To examine the relationship between VIX’s returns and /ES’s returns, I tested the significance of returns over the period of January 2020 - April 15 2022, which is the same timeline.
The futures market opens on Sunday afternoons at 6 EST, and closes during the week at 4:15 EST for approximately 2 hours while contracts from the day settle, and then they’re back up and running. For example, Sunday’s returns are from Friday’s futures close to Sunday’s open.
The variance of the returns on Sunday is much larger than other days, most likely due to news cycles over the weekend, or Friday’s SPX close. To note Friday is not in the dataset as futures close at Friday’s open, so there is no data for that timeframe. Returns for /ES are positive overall, as would be expected.
We can get a better visual perspective with the returns by plotting with /ES, VIX, and GSPC (SPX)
As would be expected, returns for /ES and GSPC are positive overall, though the actual surprise lies in the VIX, which has positive returns overall at 1.27%. However, this isn’t a surprise, since the VIX is mean reverting to a flat level, compared to UVXY and VXX which are continually declining.
Though the averages may not state it, the difference in returns for VIX on Mondays (Friday close - Monday open) and /ES on Sundays (Friday futures close - Sunday futures open) is statistically significant with a test statistic of -7.6 and a p-value of 8.5 x 1013. With that, the returns between /ES on Sundays and /ES on Mondays is statistically significant with a test statistic of -2.7 and a p-value of .0069.
Why is it negative?
A negative test statistic means that there is an inverse result. In simple words, when the VIX goes up, /ES goes down. Verifying this with /ES on Sundays and Mondays means that the returns on Sundays are statistically significant, and thus are very different from other days of the week.
With that, we can conclude that the weekend effect is present within Sunday’s futures open, at least for this analysis.
Why are the UVXY and VXX different from the VIX?
The reason for this continual decline of the VXX and UVXY, rather than finding an oscillation range like the VIX, is due to how the indexes are required to function. Since the VIX is functionally a ratio of implied volatility, it’s not tradable with shares like the UVXY and VXX.
As a result, futures contracts on the VIX (/VX), trade at a higher premium value relative to the value of the VIX itself. The VXX and UVXY are built off of these futures contracts to be “coupled” to the VIX index. Coupled meaning they mimic the movement of the VIX with fairly high precision.
The VXX and UVXY hold one month and two month futures as assets for this coupling to happen, and since the VIX trades at a higher premium than what its actually worth, these indexes are bound to lose money over time. The VXX and UVXY have to roll the one month futures each month (close the one month and open another two month, as the current two month becomes the one month), which causes the two assets to be continually declining.
To reference again, the VXX is now decoupled from the VIX, meaning that Barclays (the creator of the VXX) is no longer putting new money into the product. Due to this, the VXX may behave erratically and not align to returns of the VIX or UVXY.
So, what does this information tell us?
When big bets on volatility cross the flow, don’t be fooled by thinking Armageddon is happening.
Most of these large bets on volatility are there for hedging purposes. Bear movements are typically faster and more aggressive than bull movement, and the whales use these products to hedge against these rapid downward movements.
On the Unusual Whales flow page (https://unusualwhales.com/flow), you can see the P/C ratio for these products being extremely low, along with a lot of sold and bought calls.
Again, this is due to volatility being mean reverting. In a bull market, volatility should theoretically always be decreasing. These betters may be selling a lot of calls due to thoughts of a reversal in this bear market, rolling out of one hedge and into another, longer dated hedge, or are outright hedging against volatility.
If you are trying to use volatility indexes to play an upward move, avoid paying the heightened premium on the VIX options by purchasing VXX or UVXY shares or calls
What does this have to do with futures?
The effect seen with the VIX and /ES futures means that an abnormality exists between the returns on the VIX index itself and futures on /ES. These returns can point to potential opportunities, such as calendar spreads over the weekends. If you don’t know what a calendar spread is, you can read about it on our Reddit. Being said, here is an explanation on how to use the Unusual Whales Options Profit Calculator to create a calendar spread on the UVXY.
Calendar spreads are composed, typically, of a near term short call, and a further term long call on the same strike. The expectation of a calendar spread is purely a volatility play. For the near term short call, you are expecting volatility to fall, therefore decreasing the value of the short call. Following expiration of the short call, you are expecting volatility to rise, which would increase the value of the long position.
Regarding the VIX and like indexes, we’re seeing a statistically significant return on the VIX on Mondays, meaning volatility is higher than normal, which allows for a great strategy surrounding the weekend effect with a calendar spread.
The UVXY is currently trading at $16.32 a share, so we would enter the short 16 strike for this week’s expiration on a Thursday or Friday, and the long 16 strike for next expiration.
To which, the expected payoff would be range bound. Since volatility, or in this case, implied volatility decreases with a lack of movement, the volatility of the short position should be negated with neutral or upward movement in the near term, with volatility peaking that next day after expiration for the long call and max profit coming if the UVXY is trading at $15.98 on that day as well.
Volatility bets are typically cheaper than buying index puts, and have the potential to return higher on significant moves. However, the appeal to index option puts on SPY or SPX comes from an account hedging perspective for these big betters, since they will typically hold a large amount of SPY shares or SPX futures, and index puts will be used as collateral. Maintaining hedging strategies, regardless of market conditions, should be centripetal to our trading plans, since we never know when volatility will strike.
As we enter a bear market, trading as we have known over the past couple years will be much different. There will be a lot of chop, and a lot of surprise movements, both up and down. As questions loom on supply chain issues, the FED and expected interest rates hikes, inflation and preliminary recession talks come out from banks, it is important to protect our securities and ensure that our portfolio maintains consistent, steady growth, even after the start of the Jan 2022 bear market / 20% pullback over 5 months.
That being said, hedging strategies with the VXX and UVXY are not perfect, and are typically better suited when paired with indexes or stocks such as $AAPL, $AMZN, $GOOGL, and the like due to their heavy weight within the indexes.
Understanding what volatility is will be a huge benefactor to trading in current market conditions, and though this report is a little math heavy, I hope you learned a thing or two!
What are your biggest takeaways on the weekend effect? Let us know on the Unusual Whales Discord community or @ us on Twitter. Also, feel free to DM me @falcon_fintwit with any questions you may have.