Volatility Compression

[dc]I[/dc] will write more about this in the future, but I want to introduce you to an important concept called volatility compression. From a theoretical perspective, markets tend to operate in one of two modes: in mean reversion mode extremes are quickly reversed. Traders buy dips and sell spikes, markets tend to be more or less rangebound, and price action is ruled by a very large random component. In range expansion mode, markets make directional movements. Feedback loops develop that fuel further extremes. (For instance, in a declining market, stops on long positions may be hit, which creates more selling, causing shorts to become even more aggressive.) Price action in range expansion tends to be more predictable, as it driven by a clear buying and selling imbalance.

There are many ways to measure volatility. Each of these ways defines something slightly different, and each method has its own idiosyncrasies and limitations. For now, let’s leave all theoretical concerns behind and just define volatility as the range covered in an average daily session. (Technically, I am using average true range in this example, which also includes any gap from the previous day’s close to the current day’s range. This is a minor technical refinement that is not overly significant.) Consider the following chart, which shows a daily chart of the E-mini S&P 500 futures, covering only regular session hours (9:30 AM – 4:00 PM New York time):

The dark red line underneath the price bars shows the ratio of the 5 day to 60 day (approximately one-quarter) ATR, with a reference line at 1.0. When the red line is under the reference line, the short-term volatility has contracted compared to the longer-term volatility. It is important to note that this is a relative measure that automatically adapts for the “baseline” level in any market. (Most good technical tools do this.)

I have marked the areas where the market was in short-term volatility compression with numbered blue boxes on the price bars. We normally expect that a directional move might emerge from these areas. This can be a multi-day thrust, or a strong trend day that opens and closes at opposite ends of the range. Note that there is not necessarily any clear directional bias from this tool; it serves a more important purpose, which is to quantify the most likely emerging volatility conditions / regime shift. Consider each of the boxes on this chart:

  1. This example is probably best categorized as a losing trade. No strong move developed after the condition triggered. No technical tool works all the time; the best we can do is to quantify tendencies.
  2. This is a good example of the way the best trades work. A strong trend day (shaved candle) followed the condition, which led to a two-week meltdown.
  3. Another strong trend day followed this area of compression, but there was no multi-day extension.
  4. We are once again in an area of volatility compression.

Please do remember that this extremely crude example serves only to illustrate the underlying tendency. In actual trading, other qualifying filters can increase the edge, and there are better ways to actually define the compression. Regardless, there is a clear lesson here for equity traders over the next few days: treat these markets as being in “breakout mode”, meaning that you do not want to be on the wrong side of a breakout. Up or down, we don’t know, but do not be buying weakness or shorting into strength–far better to “go with” these moves than to try to fade them. I am holding significant long exposure, and will be very quick to trim if the market is weak off the open. (Note that opening down is not the same thing.) On the other hand, I will be more reluctant to take profits on longs if the market breaks out to the upside. This is an example of how using an underlying tendency in the broad market can inform trade decisions in individual names, whether or not you intend to trade the actual setup itself.

(For reference, Toby Crabel uses several variations of this idea in Day Trading With Short Term Price Patterns and Opening Range Breakout. Linda Raschke gives a few more variations, centered around ratios of Historical Volatility instead of ATR in Street Smarts: High Probability Short-Term Trading Strategies. My forthcoming book will also feature a number of studies and trading ideas driven by range expansion following volatility compression.)

(Disclosure: I hold long positions in HLF and MAKO.)

AdamHGrimes

Adam Grimes has over two decades of experience in the industry as a trader, analyst and system developer. The author of a best-selling trading book, he has traded for his own account, for a top prop firm, and spent several years at the New York Mercantile Exchange. He focuses on the intersection of quantitative analysis and discretionary trading, and has a talent for teaching and helping traders find their own way in the market.

This Post Has 8 Comments

  1. Felix Ilium

    This is a very interesting article – Thanks. I like the comparison of the different volatility times. Just out of curiosity, any reason behind the 5 & 60 numbers?

    1. Adam Grimes

      Thanks glad you enjoyed it. The idea is to just use short and long term windows… i picked weekly to quarterly but you could use something else. Also, i created this indicator just for the article. I don’t usually use this particular way to look at vol compression… I recognized it another way and then created this as an illustration so focus more on the concept than the specifics. lemme know if you have any more qsts.

      1. Felix Ilium

        Thanks for the response. I was doing something similar using Bollinger Band Width on different time frames. I like your concept of using the ATR for this, seems better for this type of application. I’m going to have to do some research here – thanks for adding to my work load – LOL Now your answer has me wondering this mysterious “other way” is.

        1. Adam Grimes

          I’m not crazy about Bollingers for a lot of reasons, but I wonder about a Bollinger type band that uses mean absolute deviation rather than standard deviation. Might try that out just as a random thought.

  2. Anonymous

    Great post Adam! Very interesting stuff.

    I use a 8 period daily ATR on a regular basis, so am familiar with it as a volatility measure, but when you say 5 & 60 ratio, is that simply constructed by ATR(5) / ATR(60)? You’d mentioned other measures that might be easier to set up in my platform (RealTick).

    That’s a pretty compelling chart you posted, but using the above construction, I looked at Treasuries, Gold and Silver, the VIX (check that chart out!) and a few levered ETFs like TNA, EDC and FAS and in these cases, while such volatility compression is indicative of potentially a sharp new range extension, it often times stays compressed for extended periods of time before expanding again. In your example the sub 1.0 readings only last for a few days.

    Thanks again, each of your new posts is at the top of my daily must read list!

    1. Adam Grimes

      Hi. First, thanks for your kind words on the posts. Glad you find them useful.

      Correct, what is shown here is just ATR(5) / ATR(60). You can use ratios of historical volatility, and ideas like that too.

      Personally, I’d be careful of technical tools on levered products and VIX, but your point is absolutely, 100% correct: there are many times when volatility becomes compressed and stays that way. I think you can use this concept in a lot of ways…. perhaps as a filter to set up certain kinds of system entries, or just as a warning not to fade the first breakout.

      If you want to look at it systematically, use this as a qualifying filter and then find another entry trigger.

Comments are closed.