Wednesday, April 16, 2008

Volume in the Stock Index Futures Market - Part One

My recent post offered a simple tool for active traders who were having difficulty getting a handle on the character of the evolving trading day. This two-part series will outline a different set of tools for the same purpose, based on market volume. The core idea is that how the market will trade is a function of who is participating in the marketplace. If large, institutional traders are active, we will tend to see greater price volatility and larger market moves per unit of time. If those institutional traders are not active, the market is more likely to be quiet, with low volatility and smaller moves per time unit.

Those large traders are trading directionally, many times keying off fundamental dynamics (news, earnings reports, economic reports) and intermarket dynamics (shifts in interest rates, currencies, commodities). It's when these related markets are moving actively that we're more likely to revaluations of stocks, and hence greater institutional participation (volume). When news is relatively absent and those related markets are not undergoing shifts in their value areas (to use a Market Profile term), it is less likely that stocks will be revalued. That keeps volume in shares low and price action quiet.

A large part of understanding the character of the market day, then, is seeing what is happening in those related markets and seeing how stocks are trading relative to expectations regarding economic news, earnings reports, and the like.

Another way of capturing the character of the market day is to directly measure volume and compare it to the median volume for that particular time period over a lookback period. My research has found that, when this relative volume is elevated (i.e., when we're trading higher volume than normal in a particular time frame), the added volume almost exclusively comes from transactions of 50 contracts or larger in the S&P e-mini futures (ES contract). Clearly, those trades are not coming from small retail traders. Rather, it is the professional trader who is more active in the market when volume is elevated. The increased volume is the footprint that tells you *who* is in the market at the time.

For purposes of illustration, I went back to March 13th (when the June ES contract became active) and broke down each trading day into nine 45-minute segments. The correlation between the volume of the 45-minute period and the high-low price range for that period was a considerable .83. To give but one example, when the volume of the 45-minute period was above 170,000 contracts (N = 105), the price range for the period averaged .78%. When the volume was below that level (N = 102), the price ranged averaged only .42%. On average, price movement was nearly twice as high during busy periods as during slow ones.

Interestingly, 14 of the 20 highest volume periods occurred during the first or last 45-minute trading segment of the day. Conversely, 17 of the 20 lowest volume periods occurred during the midday periods from 11:45 AM ET to 2 PM ET. That tells us that *who* is in the marketplace changes significantly over the course of the market day. An active trader needs to have different anticipations of price movement early and late in the day compared with midday.

Finally, on a daily basis over the March 13th-present period, daily price range correlates a whopping .86% with daily ES contract volume. When the ES volume has been over 1,800,000 contracts (N = 12), the daily price range has averaged 3.16%. When the volume has been below that level (N = 11), the daily price range has averaged only 1.43%.

Clearly this has important implications for how traders manage trades. In a busier, more volatile market, it makes sense to place stops wider and to let profits run further (i.e., to place profit targets further from good entry points). In slow markets, it makes sense to keep stops tight and take profits aggressively, as these are less likely to run.

Who is in the market dictates how you should trade that market. That varies from one day to the next, and it varies from one time of day to another. I cannot think of a more important lesson for developing active traders.


Intraday Volume Patterns and Volatility


Ziad said...

I just have a quick question Dr. Brett. I've definitely seen the volume/range relationship first hand day after day, but one curious thing I have noticed is that a lot of times pure trend days that gap and drive from the opening bell (especially up trend days) often have LESS than average volume, and yet they typically will have a large range. Today was such an example, with volume on the mini dow coming in 21% below its 2 months average.

For almost all other types of days, low volume means small ranges, but for these trudging/unrelenting trend days, it doesn't seem to hold. Do you have any theories for why that is? One theory I have is that while the day as a whole has a large range, if you look at 1 minute bars they are usually quite small and reminiscent of tight chop, implying that volatility is actually low but directional bias is so high that these small bars add up as the day goes on thus producing a large range. But that, in itself, doesn't explain why overall volume would be low when institutions should be at work with such extreme value re-pricing. Do you have any ideas?


MikeH said...

The first thing I look at is the overnight volume (I trade the euro and yen futures primarily). In addition to adjusting targets and stops with relative volume, I find I also must adjust my entry and exit strategies. I feel that lower volume days have more "orderly" price level breaks. That is, they trade through a pivotal price and proceed directly to a new value area. So I adjust my entries to be more aggressive, usually using a stop to initiate entry. In contrast, on high volume days a price level will be tested repeatedly, and usually will have a brief pullback when it finally does start to break. Here I try to capture the edge from the spread, and have to remember that I can get multiple entries/exits off a single trade idea.

My theory here is that volume on the macro level correlates to overall price volatility, while on a micro level affects the price action. On a low day, the liquidity isn't there when the price starts moving. Even if it doesn't go far, your chance to get it is lessened. While on a high day, there are a lot more opposing ideas in play (fades vs. breaks), and possibly the wrong side stays in that little bit longer thinking they still have hope, before they have to unwind it a worse levels.

And thanks Dr. Brett for reiterating this idea. You don't find this mentioned anywhere. At best, other experts will advise you to pick a liquid issue if you will be actively trading, or just look for expanding volume. In another poker example, you can play based on just the strength of your hand, but you won't win much until you learn to switch gears based on the opponents you are currently playing.

BirdMan said...


Thanks for your post.

So the large players do not trade the pit S+P as opposed to the mini? I was thinking the increased volume by large players would show up in the pit... And when you look at daily ES volume is that from the day session only?

Thank you.


Brett Steenbarger, Ph.D. said...

Hi Ziad,

I'm not sure the mini-Dow is representative, as it's not widely traded institutionally. The ES contract captures the institutional interest better, I find. When you look at the day's action in ES, you'll see clearly that volume was *way* above normal during the periods of up movement. It went back to normal during flat periods in the day, but the greatly expanded volume during price surges told you that large traders were participating--


Brett Steenbarger, Ph.D. said...

Thanks for the insights, MikeH. Your theory makes a lot of sense--


Brett Steenbarger, Ph.D. said...

Hi Dave,

It would be interesting to look at pit action as an indicator. Much institutional volume has migrated to the screen, which makes ES viable as an indicator of large trader activity. And, yes, I'm looking at day session data only--