Are Options on Index Futures Profitable for Risk Averse Investors? Empirical Evidence

George A. Constantinides, Michal Czerwonko, Jens Carsten Jackwerth, Stylianos Perrakis2010 · DOI

Authors: George M. Constantinides (University of Chicago), Jens Carsten Jackwerth (University of Konstanz), Michal Czerwonko (McGill University), and Stylianos Perrakis (Concordia University)

Abstract: American options on the S&P 500 index futures that violate the stochastic dominance bounds of Call bid prices more frequently violate their upper bound than put bid prices do, while violations of the lower bounds by ask prices are infrequent. In out of sample tests of stochastic dominance, the writing of options that violate the upper bound increases the expected utility of any risk averse investor holding the market and robust.

Introduction

In a perfectly efficient market, consistent outperformance without increased risk is theoretically impossible. However, the persistent presence of "volatility smiles" and "skews" in the options market suggests potential systematic mispricing. While previous research has explored these anomalies, many studies rely on restrictive models like Black-Scholes or overlook real-world frictions such as transaction costs and bid-ask spreads.

This paper addresses whether a generic, risk-averse investor—one who simply prefers more wealth and dislikes volatility—can identify and exploit mispriced options to improve their expected utility, regardless of their specific utility function.

Methodology: A Nonparametric Approach

The researchers utilize a nonparametric stochastic dominance (SD) framework, avoiding the assumption that stock returns follow a specific distribution. Instead, they use historical data to establish "fair" pricing bounds.

Stochastic Dominance Bounds

Following the framework of Constantinides and Perrakis (2007), the study calculates the maximum price a risk-averse investor should pay for an option and the minimum price they should accept to sell one. These bounds are tested against four volatility estimation modes:

  • Unconditional: Long-term historical volatility.
  • 90-day Historical: Recent volatility trends.
  • Adjusted Implied Volatility (IV): Derived from current market prices.
  • EGARCH: A statistical model for time-varying volatility.

Trading Simulation

The study compares two types of investors:

  1. Index Trader (IT): Holds the S&P 500 and a risk-free bond, rebalancing optimally.
  2. Option Trader (OT): Follows the IT strategy but writes (sells) options whenever the market bid price exceeds the calculated upper stochastic dominance bound.

To ensure statistical rigor, the Davidson and Duclos (2000, 2006) tests are applied to determine if the OT’s portfolio stochastically dominates the IT’s portfolio, implying the strategy is superior for all risk-averse investors.

Data and Robustness

The analysis uses a clean database of CME S&P 500 futures options from 1983 to 2006. Initial distribution estimates were derived from S&P 500 index returns dating back to 1928.

To ensure the results are applicable to real-world trading, the researchers incorporated significant "frictions":

  • Transaction Costs: A 0.5% cost was applied to all trades.
  • Execution Prices: Actual bid prices were used instead of mid-prices to account for the spread.
  • Execution Delay: A one-quote delay was introduced before entering trades to simulate realistic execution timing.

Key Insights: The "Good Sell" Phenomenon

The study identifies a significant asymmetry in the options market. Call options frequently violate their upper bounds, indicating they are "too expensive." Conversely, violations of lower bounds (underpricing) are rare.

The authors suggest this persistence is due to market segmentation. Speculators drive up out-of-the-money (OTM) call prices as leveraged bets on market rallies, while mutual funds drive up OTM put prices as insurance against market crashes. Because the results are based on Second-Order Stochastic Dominance, they apply to any risk-averse investor, independent of their specific risk tolerance level.

Results

The out-of-sample tests confirm that writing "good sell" calls consistently leads to portfolio returns that stochastically dominate an index-only portfolio.

  • Economic Impact: The identified "good sell" call options provided average monthly returns ranging from 12.3% to 43.2%.
  • Specificity of Strategy: Naïve strategies, such as selling all OTM options, failed to achieve consistent stochastic dominance. The specific SD bounds were critical for filtering profitable trades.
  • Portfolio Diversification: The findings remained robust even when the investor's portfolio included other asset classes like small-cap stocks, value stocks, or REITs.

Limitations and Conclusion

While OTM calls are frequently overpriced, the frequency of violations for put options was too low for strong statistical inference within this methodology. Additionally, the authors note that while these opportunities exist, large-scale exploitation might be constrained by margin requirements and market depth, which helps explain why these anomalies are not immediately traded away.

In conclusion, the S&P 500 options market exhibits inefficiencies that risk-averse investors can exploit. By selling call options that exceed theoretical stochastic dominance bounds, investors can significantly improve their expected utility in a robust, out-of-sample manner.

Summary updated

Abstract

American options on the S&P 500 index futures that violate the stochastic dominance bounds of Call bid prices more frequently violate their upper bound than put bid prices do, while violations of the lower bounds by ask prices are infrequent. In out of sample tests of stochastic dominance, the writing of options that violate the upper bound increases the expected utility of any risk averse investor holding the market and robust.
BibTeX
@misc{constantinides2010are,
  author = {George A. Constantinides and Michal Czerwonko and Jens Carsten Jackwerth and Stylianos Perrakis},
  title = {{Are Options on Index Futures Profitable for Risk Averse Investors? Empirical Evidence}},
  year = {2010},
  doi = {10.3386/w16302},
}
Figure