Bumpier Times May Be Ahead

The S&P 500 is at all-time highs, and pretty much all of the other major indexes are also at all-time highs or just now breaking even with respect to pre-COVID levels. Retail investors are apart of the game like never before, with roughly 20-25% of all NYSE volume coming from retail order flow. This is great, as investing services are giving people access to the markets like never before. However, this can also pose a challenge from a volatility perspective. In this blog post I will state why I believe we, with respect to the S&P 500, are likely to experience a decent amount of volatility within the next few weeks, if not more soon than that. 

The VVIX

For those who do not know, the VVIX is an implied volatility-of-implied volatility index. What this means is that it measures the implied volatility of the VIX index, which itself measures forward-looking implied volatility of the options listed for the S&P 500 index, hence the name "vol-of-vol index". This index is important for a few reasons, one is because it often front-runs major VIX moves, and two is that it takes into account different types of implied volatility predictions. This is natural to assume, since it measures the fundamental uncertainty OF the future uncertainty, of future front-month returns on the S&P 500 returns. Further, the VIX is expected to answer the question of "What is the expected best and worst case scenario for the SPX index level?", and the VVIX is supposed to answer "How likely are we expected to experience moves akin to the best and worst case scenario?". 

Construction of the VVIX Index follows the same methodology used by the VIX, but rather than referencing the price of SPX options, VVIX references the price of futures contracts. Thus, the VVIX reflects a constant 30-day maturity level by creating a weighted average value between a contract with more than 23 days remaining to expiry and a VIX futures contract with less than 37 days to expiry. Using a Black-Scholes framework, the VVIX reflects the weighted average implied volatility of all actively bid VIX futures contracts. Another key piece of information is that, as a result of the overweight towards at-the-money strikes and a bias towards puts, the VVIX will generally approximate the 95-97% implied volatility for options on the front-month VIX futures contracts (proving that it is an accurate enough indicator to utilize for additional information). 

In addition, VIX levels are extracted from the underlying derivative implied volatilities in SPX products that utilize the Black-Scholes option pricing framework. This is important because this pricing theorem does not take into account option skewness (options can have different implied volatilities along an option chain within the same maturity date), as well as what is called "jump risk" or "kurtosis" (sudden large changes in the underlying level, when the asset "jumps" in price). The only way traders can price in these two critical aspects of implied volatility is by inflating the actual number of said implied volatility AFTER they get their predicted value. This is shown in VVIX levels, as the VVIX, by construction of how it represents the implied volatility of the future uncertainty on a particular asset contains information on kurtosis risk (always has a higher spot value than the VIX, shows that it contains this jump risk), in addition to how it always has a higher spot value than the VIX. When we experience vol-of-vol in the markets (rising VVIX level), it is representative of kurtosis being factored into volatility products in the underlying return distribution, which in turn causes an increase in VIX levels. One important distinction of "kurtosis" (jump risk) is that it is not only representative of a sharp and sudden drawdown, but can also be a sharp and sudden rise in the underlying security (it does not represent directional bias). 

Furthermore, high levels of the VVIX can be representative of high levels of S&P 500 returns in either direction (as an absolute value basis). 

So, after that lengthy description of why the VVIX can be a better predictor of forward-looking volatility, I will show you a chart that displays a recent divergence between the VIX and VVIX, which are supposed to be extremely closely correlated. 

The top chart of of the VIX, and the bottom the VVIX. Now, you might notice something interesting. The VVIX bottomed out as of October 1, 2020, and has been making frequent higher lows ever since, progressing from 101.00 to 107.64 (+6.57%) (10/01/2020 - 2/06/2021), while the VIX has declined from 26.70 to 20.87 (-21.84%) over the same time period. Clearly, this an extremely important market event that has taken place, as the forward-looking instrument that contains more information in regards to forward-looking volatility has risen, while another volatility index has fallen. I believe that this represents a key idea: that the consensus implied volatility positions on the SPX index are underpricing a sharp move (most likely to the downside). This might represent an opportunity. Allow me to also represent the implied volatility term structure of the options embedded upon the micro-SPX index, ES futures contracts. 

What this chart shows is the implied volatility values priced at various at-the-money option maturities for the same strike price versus the previous week. The important thing to notice here is that the orange line (current option settlement values) is below the grey line (option settlement values from a week prior). This means that options are less expensive, from an implied volatility pricing standpoint, than a week prior, further iterating the standpoint that people are, again, underpricing a sharp move relative to past values. 

Dealer Positioning

The first piece of information I would like to bring into question will be with respect to net positioning of market-maker's with S&P 500 E-Mini Futures products (ticker symbol: ES). The E-Mini Futures products are essentially futures that were developed to make trading the SPX index less expensive, and therefore provide better liquidity to trading this particular market. Now, there are various categories that one can analyze to take note of positioning on this particular product (that of Asset Managers, Leveraged Funds (Hedge Funds or CTAs), or other traders), but I choose to focus on the positioning of market maker's the most critically. The reason being is because they have the best access to rapid order flow data and short term predictions for the underlying security, given that they have to be able to provide a market at any given time, and still try to make money. Another important reason is because dealer's, as a percentage of total open interest on the ES products (long only positions), hold within 10-13% (from 11/10/2020 - 2/2/2021) of all open interest of ES products at any given time. That is a significant percentage of this market. Now that I have justified why I am using the following data, allow me to show you my findings and what it means. 

This visual represents the E-Mini S&P 500 (ES) Commitment of Traders report. This shows the net long, short, and spread positions of major players for this particular product. Long means that the person owns the underlying asset, and hopes the product will appreciate in value. Short means that the person sold short the asset (effectively 'borrowing' shares from someone else), hoping that the value will go down. A Spread means that the person (usually) has neither a long nor short directional bias for the product, and is what is said to be risk-neutral. We do not have to worry about the spread positioning as they are unrelated to this post, do not present any additional important information, and are relatively miniscule in absolute value relative to the broader long and short positions. The key information derived from this graphic is as follows: market maker's have increased their net short position from 42.20%, to 67.26% (8/11/2020 - 2/2/2021). This effectively means that they went from being net long by 7.20% (relative to their baseline of being marker-neutral, so 50% long ES contracts and 50% short ES contracts) to net short by 17.26% of all outstanding positions. This is a critical piece of information because, if we believe market maker's have the best positioning standards out of all major players (since they have to be able to play the long or short side at any given moment), then this is a clear warning signal suggesting that market maker's, on this particular product, are anticipating a draw down on the S&P 500 index, hence their short position. Just as a fun fact, dealer positioning went net short in 9/15/2020. Another fun fact, major market positioning as of 2/2/2021 had the same number of long and short contracts of ES (both 2,297,641 ES contracts). My take away is that there will be some big losers, and some big winners. If someone is very net short, such as dealers, then someone must be very net long, such as Asset Managers. Since this game is "essentially" a zero-sum game, the money has to come from somewhere, right?

Breadth of the S&P 500

In my previous post that I made of APPL, I included how it was important to consider a security's underlying breadth, not just the surface area of price. What I meant by that it how it is important to note what the volume and volatility are doing in junction with knowing the spot price of the security. For this next piece, I will present information on the SPY, the most popular ETF tracking security of the S&P 500 index. SPY volatility has declined as of recent, which is itself a bullish indicator (as the price of your asset increases, you do not want the underlying volatility complex to be rising at the same time). However, there are some volume issues following up this massive recent move upward on the SPY. For example, volume of SPY (as of 2/6/2021) was down -33.66% versus the 5 day prior average, down -31.71% versus the prior 15 day average, and down -26.45% versus the prior 30 day average. This, ladies and gentleman, is not good. When the price of a security moves higher, you want volume to increase behind it, as it represents more buying pressure, and a health uptrend in the security's rising price. However, it is the complete opposite for the SPY. Its breadth is very thin, and any sudden downward spike might have great consequences, as the volume indicates that there are not increasing amounts of buyers at recent levels. 

Conclusion

In summary, I believe that the market may be underpricing a sharp drawdown due to a variety of factors. In addition to that, the "smart money" has been building their short position for months now, and I think that there may be higher levels of realized volatility in the future. I hope that this post might bring some attention to these factors and possibly educate those who want a different perspective on the market, rather than that just preached by CNBC and Jim Kramer.