The research literature suggests that both short selling and options-trading activity serve to predict future market returns. Short sellers may be industry insiders or other informed market participants who are positioning themselves based on information that has not yet been discounted by the market. The addition of listed options for an asset has generally been found to provide additional, non-redundant information about future returns over and above the information contained in the equity order flow. A recent study evaluates which type of trading activity -- short selling or options trading -- is more predictive than whether the two are complementary.
In the study, "Who are more informed, short sellers or option traders?" R. Jared DeLisle, Bong Soo Lee, and Nathan Mauck attempt to confirm whether information from options markets actually is non-redundant and to determine which group of traders is more informed. One of the challenges for any comparison of short sellers and options traders is that there is likely significant overlap between the participants in those two groups.
Additionally, the introduction of liquid listed options affects the amount and the efficiency of short selling: Some short sellers buy puts or write calls instead, while some researchers have found that options reduce short-selling costs. Disentangling the two groups in order to analyze their respective effects is not easy, and the conclusions in the prior literature reviewed by the authors vary widely.
To test for a relationship, the authors hypothesize the following.
"If option traders act independently of or before short sellers, this would indicate informed traders engaging in the options market and supports the theory of non-redundancy. On the other hand, if option traders simply follow the lead of short sellers, then the short sellers are the more informed traders acting in the equities market, and this suggests the options market is not adding any completeness to the equities market and rejects non-redundancy."
The methodology of the test is as follows: a regression of option open interest against stock short interest in order to check for complementarity, and a time-series regression model against equity returns in order to examine which type of activity is more informative. The authors use monthly short interest ratios on U.S. stocks from 1996 to 2009 and monthly aggregate option open interest on the same underlyings.
Their conclusions run contrary to much of the existing literature. First, they find that monthly option open interest does not contain information about future equity returns, and that options markets generally "follow" short traders. So short sellers are relatively more informed. But they also find that, since short sellers react more intensely to past negative performance, they are also not especially informed.
We can think of two ways to refine the analysis of this study. In aggregating all option open interest and by doing so at a monthly horizon, the authors include a lot of necessarily "uninformed" activity that may obscure the relevant horizon of informed trading. So, first, they could consider working with data at a weekly or daily timeframe -- including short ratios, open interest, and asset returns -- in determining whether the presence of informed trading is detectable in the short term. Second, they could try to filter option open interest to focus on abnormal levels as a potential indicator of informed trading. For example, they could regress equity returns against open interest when it is at various percentage thresholds above or below normal.
Other recent studies of option open interest (Kehrle and Puhan 2012 and Fodor, Krieger and Doran 2011, neither cited here) have pursued both of these routes, and they've found economically significant returns for strategies trading on the information from option open interest. We will follow up on some of these more detailed studies in a future piece.