Not Your Typical Vanilla Dispersion Trade

By Garrett DeSimone, PhD, OptionMetrics, and Jeff Corrado, Tom Corcoran, Dan Corcoran, Volos

Dispersion trading has experienced a significant surge in popularity in 2024. Once reserved for only the most sophisticated trading desks due to its complexity, this strategy has now gained broader attention. At its core, dispersion trading involves selling index options while buying options on individual constituent stocks, aiming to capitalize on the relative overpricing of index premiums. The profitability of this strategy hinges on a low-correlation environment—put simply, it benefits when stocks move in different directions.

Over the past several years, correlations have become structurally lower, leading to concentrated interest in dispersion trades and concerns about potential overcrowding. As with any popular trade with high asset growth, dispersion is vulnerable to left-tail or drawdown risk. This risk arises because correlations tend to move in sync during liquidity events. A stark example of this occurred on August 5th of last year, when the VIX spiked above 50 amid fears of a potential yen-carry trade unwind.

This raises an intriguing research question: can vanilla dispersion be modified to reduce drawdown risk during periods of shrinking profitability? To explore this, we turn to the economics underlying the strategy, which revolves around capitalizing on the difference between implied correlation (IC) and realized correlation (RC)—commonly referred to as the “correlation risk premium.” Our hypothesis is that when the gap between IC and RC narrows significantly, the compensation for taking on correlation risk diminishes. As a result, the strategy becomes riskier and its Sharpe ratio declines.

In Chart 1 below, we display the 91-day rolling difference between IC-RC. The most straightforward observation is that periods of market fragility are associated with negative correlation spreads. This is consistent with the notion that during market selloffs “all correlations go to 1.”

In order to further test this modification, we utilize implied correlation data from IvyDB Beta implemented within the Volos Strategy Engine, a systematic options strategy development platform. We then compare the vanilla dispersion versus active dispersion. Our analysis focuses on the dispersion strategy beginning in 2008: 

We structure the vanilla strategy as follows: 

• Each strategy sells ATM SPX straddles and purchases ATM straddles on 12 equally-weighted stocks.
• There are two stock reconstitutions to maintain security relevance in 1/2014 and 1/2021.
• The stock options positions are sized to run vega-neutral against the short index options positions. All the positions are daily delta hedged.
• We test strategies which target 1-month options and roll at expiration (1m1m) and strategies which target 3-month options and roll each month (3m1m).
• Mid-point prices are used for options transactions (i.e. no slippage included).
• Interest earned on the cash position is excluded to focus on pure options PnL. 

This is compared to the active strategy, where positions are deactivated when IC-RC falls below 5%, and reactivated when it is above 5%. 

In Chart 2 below we display the backtest results:

The implementation of an active dispersion incorporating a 5% (IC-RC) threshold demonstrates superior risk-adjusted performance compared to vanilla approaches. The active strategy, utilizing 3-month options with monthly rolls, achieved a cumulative return of 47.42% with an annual return of 2.33%, while maintaining lower volatility (2.49%) compared to both passive strategies. 

Most notably, the active approach exhibited a more favorable Sharpe ratio (Return/Risk) of 0.93, substantially outperforming both the passive 3m1m strategy (0.60) and the 1m1m strategy (0.14). The active strategy’s superior risk management is further evidenced by its reduced maximum drawdown of -5.33%, compared to -7.64% and -10.45% for the passive strategies, respectively. These results suggest that incorporating the IC-RC threshold as an active signal for position management can significantly enhance the risk-adjusted returns of vanilla dispersion trading strategies. Theoretically, an investor could utilize the IC-RC to evaluate other interesting modifications to a vanilla dispersion strategy such as adjusting levels of leverage based on IC-RC. 

One notable exception to utilizing IC-RC as an active dispersion trading signal is the August 5th, 2024 volatility event referenced above. Instead of a narrowing correlation spread leading into August 5th, IC-RC instead widened dramatically before falling and breaching the 5% threshold later that month on August 19th. While the backtests above may not indicate investors suffered significant losses (the active strategy having a Max Drawdown of ~3%), the anecdotal high leverage applied by a typically dispersion trader indicates that losses might have been substantial.  

Derivatives experts now point towards dispersion trading and its potential effects of dampening overall market volatility and creating environments conducive to extreme market volatility events. Efforts are now being made to demystify this traditionally opaque concept such as the launch of the Cboe S&P 500® Dispersion Index (DSPX) in September, 2023 which measures the expected dispersion in the S&P 500 over the next 30 calendar days. Despite this new indicator, dispersion strategies have no widely accepted benchmark as implementation remains largely discretionary in nature. Institutional investors can unlock unprecedented transparency and reframe dispersion as a science rather than an art form by utilizing robust historical implied correlation data and highly flexible options benchmarking software. 

For more insights on stocks, the options market, and volatility, visit the OptionMetrics blog.
 

There is an inherent risk involved with financial decisions. The information in this article is for informational purposes only and is not intended to provide financial advice. Views expressed are those of the author and are not necessarily those of the company.