Is Historically Low Volatility About to Expand?

Seasonal Insights
August 16, 2017 | Dimitri Speck

You have probably noticed it already: stock market volatility has recently all but disappeared.

This raises an important question for every investor: Has the market established a permanent plateau of low volatility, or is the current period of low volatility just the calm before the storm?

When such questions regarding future market trends arise, it is often worthwhile to examine market history.

For the purpose of analyzing volatility I have used one of the longest time series available, namely the Dow Jones Industrial Average (DJIA) from 1915 onward – which represents more than a century of price history. In order to improve comparability, I have removed Saturdays, which used to be trading days in the distant past. I have calculated historical volatility as rolling and overlapping time periods of 21 trading days (equal to approximately one calendar month).

By the standards of the past 100 years volatility is very low

On August 7 volatility measured in this way had declined to just 3.99 percent, which is a very low level indeed.

By way of comparison, over the entire period of 25,734 days, volatility amounted to 15.05 percent on average!

However, this isn't the first time it stood below the level of 4 percent. Occasionally this has already happened in the past as well, for instance in 1944, 1964/65, 2011/12 and 2014.

The chart below shows the rolling 21-day volatility of the DJIA since 1915:

Dow Jones Industrial Average, 21-day volatility, 1915 to 2017

Seasonal Insights Chart - Dow Jones: 21-day volatility

Volatility is currently extremely low. Source: Seasonax

Volatility can reach much higher levels!

Levels of volatility as low as they are currently nevertheless are very rare occurrences. Over the past century levels below 4 percent were recorded in just 0.17 percent of all cases.

Conversely, the upside potential is considerable. To merely revert to the long-term mean of 15.05 percent would require more than a tripling from current levels.

However, historical extremes in volatility were much higher than the average: thus, in 2.99 percent of all cases volatility was at levels above 50 percent, and in 0.1 percent of all cases it even exceeded 100 percent!

The next chart illustrates the distribution of 21-day volatility levels of four percent and higher in the DJIA from 1915 to 2017 in increments of one percentage point.

Dow Jones Industrial Average, volatility distribution, 1915 to 2017

Seasonal Insights Chart - Dow Jones: Volatility Distribution 1915 - 2017

There is substantial potential for volatility to expand. Source: Seasonax

Once again it can be seen quite clearly how extraordinarily extreme volatility levels below 4 percent are.

Volatility in the course of the year

When should a surge in volatility be expected though? In order to find this out, we will examine the seasonal pattern.

The next chart shows the seasonal trends in the 21-day volatility of the DJIA over the entire time period under investigation from 1915 onward.

Note: While in the first chart above I have marked out the level of volatility at the end of the respective 21-day rolling time periods, as is usual practice, I have marked them out in the middle of the respective time periods in the chart below, as this shows the seasonal pattern more clearly.

Volatility of the Dow Jones Industrial Average, seasonal pattern, 1915 to 2017

Seasonal Insights Chart - Dow Jones Volatility Seasonal 1915 - 2017

A peak in price volatility is typically reached in October. Source: Seasonax

Despite the very long-time period under examination with its multitude of data points, there evidently exists a very distinct seasonal pattern in stock market volatility.

In July, the proverbial summer doldrums can be observed, with volatility levels averaging less than 14 percent. Thereafter volatility typically increases to a peak of almost 19 percent in October.

Low volatility is not going to persist forever!

Many investors are currently betting on a further decline in volatility. In view of its already very low level and the negligible additional downside potential it offers relative to the substantial upside potential, this is probably not the best idea ever. Moreover, as shown above, October is the month in which volatility typically reaches a seasonal peak. Taking both of these facts into account, it seems far more sensible to expect an expansion in volatility.


Past results and past seasonal patterns are no indication of future performance, in particular, future market trends. seasonax GmbH neither recommends nor approves of any particular financial instrument, group of securities, segment of industry, analysis interval or any particular idea, approach, strategy or attitude. seasonax GmbH hereby excludes any explicit or implied trading recommendation, in particular, any promise, implication or guarantee that profits are earned and losses excluded, provided, however, that in case of doubt, these terms shall be interpreted in abroad sense. Any information provided by seasonax GmbH or in this issue of the Seasonal Insights newsletter shall not be construed as any kind of guarantee, warranty or representation in a prospectus. Any user is solely responsible for the results or the trading strategy that is created, developed or applied. Indicators, trading strategies and functions provided by seasonax GmbH or in this issue of the Seasonal Insights newsletter may contain logical or other errors leading to unexpected results, faulty trading signals and/or substantial losses. seasonax GmbH neither warrants nor guarantees the accuracy, completeness, quality, adequacy or content of the information provided by it or in this issue of the Seasonal Insights newsletter. Any user is obligated to comply with any applicable capital market rules of the applicable jurisdiction. All published content and images in this Seasonal Insights are protected by copyright. Any duplication, processing, distribution or any form of utilisation beyond the scope of copyright law shall requirethe prior written consent of the author or authors in question.