- The Big Picture - http://www.ritholtz.com/blog -

Seasonality Revisited

Posted By Barry Ritholtz On December 1, 2003 @ 3:27 pm In Finance | Comments Disabled

Last week [1], we made reference to the historical Seasonality factors impacting Thanksgiving week trading. The feedback this generated made us want to revisit this issue in more detail. There seem to be some misconceptions about these Seasonal factors, both over the short term (Santa Claus Rally) and over longer periods of time (Presidential Cycles). There are several thoughts you should keep in mind when considering Seasonality:

1) Statistics Are Not Guarantees: Seasonality issues take past market history, and generate a statistical probability – not a guarantee – of future performance. Example: we noted last week that about 65% of the time, the days before and after Thanksgiving are net positive ones for the market. It goes without saying (although perhaps not) that about 1/3 of the time, the markets have closed negatively. Certainly, that’s a nice statistical edge to have, its by no means a sure thing.

2) “Is seasonality working at present?” Look to the immediately prior period for strong clues as to whether any expected seasonality is likely to prevail. For example, the common expectation is to “Sell in May, then go away.” But that presumes the November to April period – the strongest 6 months of the year – performed to expectations. Since we didn’t see that this year, and the market had a serious sell off, and a rally off the lows 2 months prior to May, the odds for “Sell in May” working in 2003 was very low in our opinion (we mentioned this back in May). Anticipating when seasonality factors will fail is key. Conversely, some cyclical factors can confirm future market action. We previously discussed that the mid-term year of a Presidential cycle [2] tends to be when market lows occur. That happened in 2002, and provides validation for another cyclical phenomena: the 4th year of Presidential Cycle tends to be when markets peak. It’s a moderate probability bet that the market may top out sometime mid-2004.

3) “Will expected Seasonality Continue Working?” There are some aspects of the calendar based trading that seem to work most years: Consider “The January effect,” where small caps outperform large caps the first month of the year. Based on calendar tax selling before the year’s end, there’s little reason to think that this will not happen this year.

4) “Has Seasonality become too widely anticipated?” There are some issues – like the Santa Clause Rally – that have become well known – perhaps too well known. Traders position themselves in front of the expected activity. Thus, the Santa Claus Rally may come early this year.

Understanding how and why these factors work will improve your investment performance.

Seasonality Revisited (full report):
PDF [3]
Word Doc [4]

Data courtesy of the Stock Traders Almanac [5]


Article printed from The Big Picture: http://www.ritholtz.com/blog

URL to article: http://www.ritholtz.com/blog/2003/12/seasonality-revisited/

URLs in this post:

[1] Last week: http://bigpicture.typepad.com/comments/2003/11/chart_of_the_we_2.html

[2] mid-term year of a Presidential cycle: http://bigpicture.typepad.com/comments/2002/09/midterm_preside.html

[3] PDF: http://bigpicture.typepad.com/comments/december_1_2003.pdf

[4] Word Doc: http://bigpicture.typepad.com/comments/december_1_2003.doc

[5] Stock Traders Almanac: http://www.stocktradersalmanac.com/

Copyright © 2008 The Big Picture. All rights reserved.