Can High Options Volume Predict Future Stock Prices?

I have often seen and read claims that relatively high trading volumes in options are good predictors of future stock prices. Over the years, I have grown increasingly curious about these claims because I rarely see reports on where and how these analyses go wrong, only when these claims work. Moreover, most large stock moves do not seem to be preceded by a surge in options volume at all. So what do Speculators know from high options volume? (For this piece, we will not explore whether or not the SEC should pursue more cases or get more insider-trading convictions where surges in trading volumes for options and/or stocks precede major stock-specific news).

I recently ran a small test to determine whether I could find any predictive value in options trading volume. Consider this a sample analysis on the effectiveness of systematizing trades triggered by high options volume.

My analysis of options expiring February 20, 2009 produced two main observations:

1. The overall (downward) trend in the market dominates stocks' price performance regardless of the options trading. For example, stocks perform nearly the same (mostly negative) between the date of the high options volume and options expiration whether the high options trading volume is only in calls or only in puts. Also, the maximum price gain in a stock improves slightly if just puts are traded in high-volume instead of just calls.
2. When both puts and calls are traded in high-volume on the same underlying stock, the stock's price performance is at least equal to the cases where only calls or only puts are traded in high-volume.

In other words, in this analysis, the overall (downward) trend in the market served as a sufficient trading signal for bearish trades. Over a wide selection of stocks, options trading volume provided little additional information for triggering trades. There is little unique information in the options volume itself that informs the trader whether to place a bullish or a bearish bet beyond the market's prevailing trend. I was quite surprised by the consistency of these results. They imply that any advantage gained from watching options volume comes from augmenting the trading rules with other sources of information. Similarly, the success of the trade is not dependent on the market expressing a bullish or bearish bias in the options trading until expiration.

Before I present the quantitative results, here is a summary of the context and assumptions:

1. Trading data in options expiring February 20, 2009 started from December 30, 2008 - I am assuming that speculators who believe they have a unique informational edge look to profit from near-term moves and will use near-term options. Presumably, going into the last seven weeks of trading for a given options month, sufficient open interest has accumulated to make speculative trading volume more significant and noticeable.
2. This time period includes the January/February earnings season, but the first week of earnings season started immediately after the January options expiration. This means this analysis likely includes a lot of earnings-based speculation. Implied volatility in a stock's options tend to increase prior to earnings and then implodes afterwards. This analysis does not capture the significant impact such volatility shifts can have on option profitability.
3. I used the options filtering tools from OptionsXpress which forced a specific definition of "high" options trading volume. For example, I could define "high" as 50% of open interest. Instead, I had to work within ranges of open interest and trading volumes. The following three configurations defined "high" options volume on a daily basis: (A) Volume above 2,500 contracts with open interest between 100 and 1,000 contracts, (B) volume above 10,000 contracts with open interest between 1,000 and 5,000 contracts, (C) volume above 20,000 contracts with open interest between 5,000 and 20,000 contracts. These criteria create very high hurdles, especially for options whose open interest is already quite high. Yet, this filter still produced 856 individual options trades for analysis.

There are two other significant limitations of this analysis:

1. The data sample is very small. I am only reviewing one of many possible contexts. The seven weeks prior to February expiration were dominated by a downtrend in an overall bear market. As noted above, it also includes the entire winter earnings season. These are specific scenarios that limit the ability to extrapolate and generalize my findings to other contexts and seasons.
2. This analysis does not account for the quantity of money traded, just the volume of the contracts. The more money involved in the trading, the more significant that trading should be.

The most difficult aspect of building trading triggers from options volume is discerning the significance of any given options trading event. Options traders buy and sell options for many different reasons. Calls and puts can be used in either bullish or bearish bets. They can also be combined with stock holdings as hedges. For example, a call may be a hedge on a short stock position, and a put may be a hedge for a long stock position.

An investor or trader may also SELL an option for generating premium before committing capital to buy a long stock position. Shorted calls may hedge existing long positions (covered calls). Call and put spreads involve buying and selling options. Strangles and straddles involve taking long positions in both calls and puts. Pair trades can make things even more complex since they involve going long puts and/or calls in multiple stocks. This is enough to conclude that options volume is meaningless before even examining the numbers! However, some people claim they can differentiate amongst all these alternative motivations, so my curiosity got the best of me on this one. Again, it is very likely that traders must use additional knowledge sources to impart meaning to signals from the options volume.

I describe the results of this analysis using a series of histograms. Each histogram is structurally the same. The x-axis represents the range of price change in the underlying stock from the time of the high options volume to either (A) the point of the lowest stock price on or before options expiration on February 20, 2009, (B) the point of the highest stock price on or before options expiration, or (C) the day of expiration. The title of each chart (click on each to enlarge) explains which of these scenarios is relevant. The price range is inclusive of the top of the range and exclusive of the bottom of the range. The y-axis represents the number of times (frequency) that an underlying stock achieved the price change on the corresponding x-axis. I do not attempt to distinguish whether the maximum return precedes the minimum return or vice versa.

Each high-volume options event is recorded with an "action date", the underlying stock, and various price points for the stock: close on the day of the high options volume, expiration day, the maximum up to expiration, and the minimum up to expiration. I also record the high trading volume of a different option strike or same option on a different day on the same underlying stock as a separate
options trading event. This eliminates the need to determine why a particular option is traded. It also means that a stock that experiences broad-based options activity across multiple strikes and multiple days will count multiple times in this analysis. Fortunately, after eliminating options on the major indices, no single stock or sector-specific ETF dominated this analysis.

Finally, I separated the underlying stocks into three categories based on the kinds of options trading on them: (1) stocks on which calls represented all the high-volume options trading, (2) stocks on which puts represented all the high-volume activity, and (3) stocks on which high-volume option trading consisted of both calls and puts (note carefully that the action date was recorded from the time of the FIRST high-volume call or put activity). These distinctions allowed me to determine whether the type of options being traded had any significant predictive value on the resulting price change in the stock. Simplistically, we would want call buying to clearly signify a buying opportunity in the stock and put buying to clearly signify a shorting opportunity. It also means that each category contains a mutually exclusive set of stocks.

For all three price change scenarios - the maximum change, the minimum change, and the change to expiration - the distribution of price changes is nearly the same. (You can confirm this by converting the histograms' y-axis to percentage of observations). Note also that these data suggest it is typically not desirable to hold the options to expiration. When the underlying stock moves "significantly" in favor of the trade, it is best to take the profits.

At last, here are the results in graphical form...

Price change from the day of high-volume options activity to the day of maximum stock price before option expiration

31% of the time (49 out of 160 observations), buying stock after observing high-volume trading in calls generated a maximum return over 10% before options expiration (these are the first three bars on hte histogram below)...

However, 37% of the time (48 out of 130 observations), buying stock after observing high-volume trading in PUTS generated a maximum return over 10% before options expiration! The distribution of price changes is nearly identical to the call buying above.

Moreover, 41% of the time (103 out of 251 observations), buying stock after observing high-volume trading in puts AND calls generated a maximum return over 10% before options expiration. Again, the distribution is nearly identical the two histograms above.

Price change from day of high-volume options activity to minimum stock price before option expiration

79% of the time, buying stock after observing high-volume trading in CALLS generated a minimum return (maximum loss) of at least -10%.

80% of the time, shorting stock after observing high-volume trading in puts generated a maximum gain (listed below as minimum return) of at least 10%. Since this percentage is the same as the call-buying scenario, we conclude that the general downtrend in the market dominates these results.

Similarly, 76% of the time, buying stock after observing high-volume trading in calls and puts generated a minimum return (maximum loss) of at least 10%. Once again, all three histograms for this scenario are very similar.

Price change from day of high-volume options activity to option expiration

50% of the time, buying stock after observing high-volume trading in calls generated a minimum return (maximum loss) of at least -10% by options expiration.

46% of the time, shorting stock after observing high-volume trading in puts generated a maximum gain (listed below as minimum return) of at least 10% by options expiration. Again, we conclude that the general downtrend in the market dominates results.

53% of the time, buying stock after observing high-volume trading in calls generated a minimum return (maximum loss) of at least 10% by options expiration.

I am still collecting options trading data. I hope to provide updates on this analysis in the future and determine whether the results change under different contexts in the market. Note well that these analyses are a bit tedious, so I cannot give any guarantees!

Be careful out there!

Full disclosure: No positions. For other disclaimers click here.

Comments

[AU]

My 2 cents worth with the intent it helps your study.

I took a simiular point of view in the late 70's :
It's not "money makes the world go round".
It's the leverage on the money that makes the world go round.

My results were much like yours. I found that open interest (contracts actually sold) skewers the data !

What your looking for (raw data) is the number of contracts being brought to market to be sold to the public not the number of contracts actually sold !

Example: I had Monsanto 85 calls and the stock price would not go over the strike price nearing expiration when my broker informed me that 10,000 Monsanto 85 calls had been brought to market. After expiration the stock price leaped over 85. I asked what was his source of information - responce "Bridge Data in St Louis" is what his brokerage used. I could not afford Bridge Data's monthly fees ($1000 / mo) so I dropped the project.

Wishing you the best of luck with your endeavor.

2009 Mar 13 11:47 AM

[Jack Walker]

Dr. Duru,

You have done a lot of work and your results are interesting.

Have you considered adding the relationship between the stock’s historical volatility and the options implied volatility to the analysis?

Jack

2009 Mar 13 11:48 AM

[bozak]

Take a look at the journal article "Option volume and stock price: evidence on where informed traders trade" from April 1998 in The Journal of Finance.

This was a large scale study, the primary finding was: "Our main empirical result is that negative and positive option volumes contain information about future stock prices."

2009 Mar 13 04:13 PM

[bozak]

Also look at "The information in option volume for future stock prices" from The review of financial studies, June 2006

2009 Mar 13 04:34 PM

[Urbane Gorilla!]

You went through a lot of work. However, I'm not sure how you applied your findings to your test model. All large interest really indicates is that there is an expectation that **something** is going to happen.

I suspect that 'bozak's' references are correct. Possibly a more successful approach would be to scan for high activity and then apply technicals for an appropriate entry and a tight stop. I suspect that just knowing that a stock is ready to pop and being able to take advantage of that movement in some cases, while avoiding a killer drop with an appropriate stop, would give you quite and advantage.

Thanks for the article. I;ll have to check up on it.

2009 Mar 13 08:53 PM

[Urbane Gorilla!]

I found this article interesting enough that I checked CXO Advisors.

This is their finding: "In summary, put-call option ratios as defined have significant predictive power for individual stocks, with high (low) ratios indicating short-term underperformance (outperformance). However, this effect relates predominately to data that is not publicly available. Also, the effect does not work for index options."

You can't post links on SeekingAlpha, so you'll have to go to cxoadvisory.com and use the search box. Type in 'option volume' and read the article titled: "
June 22, 2007 - The Predictive Power of the Put-Call Ratio for Individual Stocks"

BTW, If you ever come across any obscure concept. These guys have pretty much already back-tested or reviewed it. Just search their site.

2009 Mar 13 09:34 PM

[Dr. Duru]

Thanks everyone. Just to be clear - my first interest was in testing the claims I see in various media that options volume has predictive value on a stock's future price direction. I strongly suspect that adding volatility measures would improve the analysis (I hinted at that in this piece), but I simply did not have the "wherewithal" to try to add it in. For example, as I menetioned, the intersection with earnings season in this data really distorts the volatility picture. I will definitely look into the journal references and CXO. Thanks everyone for the constructive comments.

2009 Mar 15 12:51 AM

[Stock Yid]

I trade options all the time.

I look at the unusual option activity reports from Schaeffers Research everyday.

I have found, alot of times the unusual option activity or open interest is a hedged trade. Either a covered call, a vertical spread etc. Whenever I see a number over 10,000, I will pull up an intraday chart of both the stock and the option and then show the bid and the ask to see if the option was bought or sold and what there a corresponding stock trade to go with it.

The one thing i have noticed that works very often is the anticipated price move before an earnings announcement. That can be seen from the price of the ATM straddle.

For example XYZ stock is trading at $30 pre announcement. The $30 call = $1.50 and the $30 put = $1.50, the stock will move $4 after hours, by the time opening bell rings at 9:30, the stock has only moved $3 and it will not continue its pop or drop until enough options have traded for the straddle selling market makers to cover their position.

2009 Mar 15 09:52 PM