Monday, July 31, 2006

Is There a Bias to End-of-Month Trading?

Are market returns superior at month's end, when institutions may put capital to work? I decided to take a look.

Specifically, I went back to late December, 2002 (N = 903 trading days) and examined returns from the last three days of the month and the first three days of the following month. These six day periods averaged a gain in the S&P 500 Index (SPY) of .15% (156 up, 102 down). That is stronger than the average one-day gain of .04% (494 up, 409 down) for the entire sample.

A look at the 2006 figures suggests that this outperformance has not fallen prey to changing cycles. Since late December, 2005, the average gain in SPY during the last three days of the month and first three days of the next month has averaged .11% (25 up, 17 down). That is stronger than the average one-day gain in SPY of .02% (75 up, 72 down) over that same period.

It does appear that, during the recent bull market, the ends and beginnings of months have been positively tinged. My calculations show that the first three days of the month have been particularly bullish, averaging a gain in SPY of .20% (77 up, 51 down) since 2003.


cpptrader said...

I have not taken the time to do it, but it would be interesting to see if there is a significant difference in amplitude based on a month's stereotypical reputation, i.e. comparing the 6 day period of Dec/Jan to May/Jun. Maybe time for me to open up excel...
As always, thanks for the posts.


Brett Steenbarger, Ph.D. said...

Thanks, Matthew; that's a worthwhile idea to look into. Another would be to investigate the starts of financial quarters.


zaitzeff said...

A friend and I did a study of the following days over the following time period:
being long for the last trading day of the month and the first four trading days of the next month,
over the years 2000 to 2005. For the study, for example, the last trading day of December is counted with January; the last trading day of Jan is counted with February, etc. We did the study using S&P futures prices. The S&P was relatively flat during this period, so the results were not biased as a result of a particular uptrend or downtrend in the market itself.

During this time period, you would have had a drawdown in 2002, but overall, you would have had a return of about $6000 per year per e-mini contract. Next, we separated out the results of the months. That is, we looked at only Jan during the six years in question; then, only Feb over the same six years, then, only March, etc.

What we found was that several months in particular was chiefly responsible for the gains:
Mar, June and Nov. These months tended to be strongly up. Other months tended to be mildly up, and August was in fact down in this time for years 2000 to 2005.

Anyone who wished more information about that study, contact me.

Brett Steenbarger, Ph.D. said...

That's an excellent study, zaizeff. Thanks for passing that along. Your finding that certain months outperform others is particularly interesting--


yinTrader said...

Hi Brett

I have been trying to trade emini S&P500 futures intra day by factoring in the S&P 500 Market Probability Calendar 2006 in the Stock Trader's Almanac 2006.

It is quite useful to add to other studies for intra day or swing trading the S&P.

I intend to collate the outcome after doing 30 trades in due course.

Brett Steenbarger, Ph.D. said...

Thanks for the note. The Almanac is a very good resource for seasonality patterns in the market and is especially helpful for longer time-frame traders IMO. I find that my best trades often come when I see a historical edge to the market and then notice action in the market tape that supports this edge. I take fewer trades that way, but it's improved my ratio of winners to losers.