Large Caps Tend to Outperform in October

28Sep11

One last bit of October seasonality to share.

As shown in my previous post, October hasn’t been a particularly strong or weak month historically, despite turning in abysmal performances in 1987 and 2008.

But October has been very strong for large cap stocks relative to small caps.

Like all things market-related, this observation isn’t as straightforward as it might appear, so read on to understand just what that statement means.

[linearly-scaled, growth of $1]

The proper way to analyze this subject isn’t to simply compare the raw returns of large caps to small.

Small caps are inherently more volatile, and because of that difference in volatility, small caps will always tend to outperform when the market goes up, and underperform when the market goes down.

What we really want to understand is performance after accounting for that difference in volatility. So in the graph above I’ve assumed a pairs trade, long large caps and short small caps in October (red) versus the average month each year (grey), since 1930.

The pair isn’t split 50/50%. The S&P 500 leg is larger to compensate for its smaller expected volatility. I explain at the end of this post how I split the pair, but the important point is that if indeed there were nothing special about October, the red line would look flat’ish like the grey one.

The graph shows that since the late 1940’s, October has been consistently bullish for large caps.

The effect has waned a bit in the last decade and so I’m cautious as to whether it continues in the future, but there’s no doubt it has existed over the last 60 years.

Again, the point of this post is not that large caps will (a) go up or (b) outperform small caps in the absolute sense. The point is that controlling for differences in volatility large caps tend to outperform in October.

For more stats on the monthly performance of large vs small caps, see this post from last year.

. . . . .

How I split the pair (for the geeks)…

I’m KISS’ing here:

First, note that I used the Fama/French benchmarks to represent large and small caps. While these exhibit high correlation to investable indices like the S&P 500 or Russell 2000, they are not themselves investable. That means that an investor who actually traded any of these observations would receive slightly different results than what I’ve presented here.
Each month, calculate the monthly standard deviation (of log returns) for both large and small caps over the previous 3 years.
Determine what % you would need to allocate to each leg to make the expected volatility of each leg equal (Alloc1 = (SD2 / (SD1 + SD2))) and (Alloc2 = (SD1 / (SD1 + SD2))).
The monthly return of the pair would then be = (this month’s large cap return X previous month’s large cap allocation) – (this month’s small cap return X previous month’s small cap allocation).

Happy Trading,
ms

. . . . .

To stay up to date with what’s happening at the MarketSci Blog, we recommend subscribing to our RSS Feed or Email Feed.