Objective research to aid investing decisions
Menu
Value Allocations for September 2019 (Final)
Cash TLT LQD SPY
Momentum Allocations for September 2019 (Final)
1st ETF 2nd ETF 3rd ETF

Exploiting VIX Futures Predictability with VIX Options

Posted in Commodity Futures, Volatility Effects

Can traders use S&P 500 Implied Volatility Index (VIX) options to exploit predictability in behaviors of underlying VIX futures. In his June 2015 paper entitled “Trading the VIX Futures Roll and Volatility Premiums with VIX Options”, David Simon examines VIX option trading strategies that:

  1. Buy VIX calls when VIX futures are in backwardation (difference between the front VIX futures and VIX, divided by the number of business days until expiration of the VIX futures, is greater than +0.1 VIX futures point).
  2. Buy VIX puts when VIX futures are in contango (difference between the front VIX futures and VIX, divided by the number of business days until expiration of the VIX futures, is less than -0.1 VIX futures point).
  3. Buy VIX puts when the VIX options-futures volatility premium (spread between VIX option implied volatility and lagged 10-trading day VIX futures volatility adjusted for number of trading days to expiration) is greater than 10%.

He measures trade returns for a holding period of five trading days, with entry and exit at bid-ask midpoints. An ancillary analysis relevant to strategy profitability looks at hedged returns on VIX options to determine whether they are overpriced: (1) generally; and, (2) for the top 25% of VIX options-futures volatility premiums. Using daily data for VIX options data and for VIX futures (nearest contract with at least 10 trading days to expiration) during January 2007 through March 2014, he finds that:

  • Overall, VIX option buyers do not on average pay a material premium for the limitation of risk provided by long VIX options. However, VIX options may be substantially overpriced at times.
  • The long VIX option strategies are highly profitable with attractive risk-reward trade-offs for front contract VIX options across the five strikes nearest to at-the-money. Specifically:
    • Randomly entering 5-day long VIX call option trades generates average gross returns of -1.9% to -5.6%, with about three times as many losing as winning trades.
    • However, entering 5-day long VIX call option trades only when VIX futures are in backwardation (Strategy 1 above) generates average gross returns of 29% to 42%, with losing trades only modestly outnumbering winning trades.
    • Randomly entering 5-day long VIX put option trades generates average gross returns of 0.7% to 2%, with about equal numbers of winning and losing trades.
    • However, entering 5-day long VIX put option trades only when VIX futures are in contango (Strategy 2 above) generates average gross returns of 6.8% to 11%, with winning trades modestly outnumbering losing trades.
  • The specified threshold for the VIX options-futures volatility premium (Strategy 3 above) results in trading opportunities about one fourth of the time. Entering 5-day long VIX put option trades only when the volatility premium is high generates average gross returns of 10% to 21%, with winning trades modestly outnumbering losing trades.
  • Many of the above scenarios are not profitable after trading frictions including full VIX option bid-ask spreads. In other words, net profitability generally requires avoiding full bid-ask spreads, especially early in the sample period. However, many scenarios are robust to 0.1 point bid-ask spreads, close to the average quoted spread during the last two years of the sample period.

In summary, evidence indicates that predictable behaviors of VIX futures support strong gross profitability of the specified long VIX options strategies.

Cautions regarding findings include:

  • As noted in the paper, large VIX option bid-ask spreads are obstacles to exploiting VIX futures tendencies on a net basis. The study does not address strategy performance over time, failing to support belief that recently lower bid-ask spreads are helpful.
  • There may be data snooping bias in selection of parameter/threshold values for the specified trading strategies. Such bias would result in overstatement of expected returns.
  • The author does not examine cumulative, portfolio-level performance for the specified strategies over extended periods, as limited by frequencies of trading conditions (and the requirement to hold cash while waiting for opportunities).
  • Strategy 3 above is likely beyond the reach of most investors, who would incur fees when delegating data collection/processing to an investment advisor.
Why not subscribe to our premium content?
It costs less than a single trading commission. Learn more here.
Daily Email Updates
Login
Research Categories
Recent Research
Popular Posts