August Volatility in the US Stock Market

Driving home the other day, I heard a segment on NPR that was discussing the recent market downturn and its possible causes. Interestingly enough, one of the guests during the segment asserted the following:

“Firstly, the markets are very thin right now because it is August and you have a lot of the seasoned traders away on vacation. You’ve got less volume than anywhere in the markets. And so there’s a long tradition of people overreacting during this vacation season.”

When pressed on this point, the guest analyst didn’t waiver:

“But the reality is that, you know, many of the most seasoned and senior traders tend to go on vacation this time of year. And yes, we can all remain in contact with our iPhones and our iPads and whatever else while we’re on the beaches. But traditionally, you have had a lot of pretty wild market movements in the month of August historically.”

The question of volatility seems to be an easy point to test. First, some assumptions:

• A market index such as the S&P 500 is an indicator of the broader market performance.

• Volatility is high if the closing price on a given day is markedly different (up or down) than the closing price on the day prior.

• A given month is more volatile if it has more high-volatility days than another month.

Based on the assertion that August volatility is caused by junior traders, I infer that with senior traders in charge, there should be less volatility. Otherwise, I claim that the causal relationship between volatility and the seniority of the traders is not provable.

To test the theory, I downloaded the S&P 500 closing prices for the last 10 years and performed some calculations. First, for each day I calculated the change in closing price, and percentage change in closing price. Here is an example:

Date D(i)	S&P 500 Closing Price P(i)	Change from Previous Day C(i) = P(i)-P(i-1)	Percentage Change from Previous Day R(i) = C(i)/P(i-1)
2005-08-25	1212.39
2005-08-26	1205.10	-7.29	-0.601%
2005-08-29	1212.28	+7.18	+0.596%
2005-08-30	1208.41	-3.87	-0.319%
2005-08-31	1220.33	+11.92	+0.986%

The variations in the percentage change in closing price, calculated by finding the standard deviation, will provide the measure of market volatility in accordance with the assumptions above.

Using the S&P 500 closing prices for the last 10 years, performing the above calculations, segregating the data into 12 separate monthly bins, and calculating the standard deviations for each monthly bin yields the following:

Month	Standard Deviation	Rank
Jan	0.0114	7
Feb	0.0109	8
Mar	0.0128	5
Apr	0.0096	12
May	0.0105	10
Jun	0.0109	9
Jul	0.0100	11
Aug	0.0136	4
Sep	0.0141	3
Oct	0.0191	1
Nov	0.0168	2
Dec	0.0122	6

Based on the daily variations in the S&P 500 closing prices over the last 10 years, it is true that August is one of the more volatile months. However, the variations in August are exceeded by the Fall months – September (#3), November (#2), and October (#1). Therefore, if the “junior varsity” traders being in charge cause volatility in August, what is the explanation for the even higher volatility in the Fall months?

Perhaps the guest analyst had some other measure of volatility in mind – inter-day swings in prices, daily trading volume, or something else I haven’t considered. Based on the volatility in daily closing prices, however, August is in the top half of volatile months, but isn’t as volatile as the Fall months. Therefore, I question the association of August volatility with junior traders. Based on data over the last 10 years, it doesn’t seem to be a viable explanation based on the assumptions above.

---

References:

Is The Stock Market Volatility A Correction Or A Full-Blown Crisis? August 25, 2015
http://www.npr.org/2015/08/25/434668675/is-the-stock-market-volatility-a-correction-or-a-full-blown-crisis

S&P 500 Historical Data, last updated 2015-08-26 6:51 PM CDT
https://research.stlouisfed.org/fred2/series/SP500/downloaddata

Image via pixabay.com
https://pixabay.com/en/statistics-chart-graphic-bar-810024/