On Wednesday, May 15, the venerable Art Cashin on CNBC offered his observations on the strange consistency in the Tuesday trade since the S&P 500 (SPY) printed its low in 2012. He notes that since those lows, Tuesdays have generated 80% of the gains on the S&P 500 . For 2013, that figure is 67%. One of the theories is that geo-politics have turned weekends into a "wildcard" where Monday trades down and things turn-around on Tuesday. Cashin notes that what is now called "Terrific Tuesday" used to be known as "Turn-around Tuesday." During the 2008-2009 bear market weekends were particularly tumultuous as the Federal Reserve and other financial authorities made a habit of holding emergency weekend sessions.
In fact, Tuesdays have an extended history of special behavior. Last year, I read an article from 2011 that discussed this behavior and theorized, like Cashin, that Tuesday's strength comes somewhat at the expense of Mondays which suffer from weekend political and/or financial drama. A year ago this month, I wrote "S&P 500 Performance By Day Of Week And The Changing Nature Of Trading Tuesdays" to examine the daily trading patterns on the S&P 500 with a particular focus on Tuesdays. This anniversary is as good a time as any to update that analysis.
Although Tuesday's behavior changes from year-to-year, its relative consistency means that other daily trading scenarios co-exist with notable consistency. In this update, I found that as 2012 progressed the clear patterns from the first 5 months changed and in some cases weakened. I strongly suspect that 2013′s prominent patterns will also morph as the year progresses. Not surprisingly, the overall patterns since 1950 did not change as a result of the last year of trading.
Here is a summary of behaviors to note on the daily trading patterns for the S&P 500. All performance numbers are daily averages…
- Since 1950, Mondays tend to be the weakest day of the week.
- Since 1950, Tuesday's out-performance is most prominent following a down day, especially when the previous day closed with at least a 1% loss.
- 2013′s Fridays tend to deliver positive returns, second only to Tuesdays.
- In 2013, Fridays deliver the largest one-day return following any random down day. Wednesday is not far behind Tuesday.
- In 2013, Monday is the one day of the week that tends to follow selling with more selling.
- In 2013, Tuesday performs just as well after a down day as it does after an up day. Monday is similarly weak. The other days of the week experience minimal gains after an up day.
- In 2013, down days featuring losses of at least 0.5% have ONLY occurred before Tuesday and Friday trading. Both days perform quite well after such down days.
- In 2013, down days featuring losses of at least 1% have ONLY occurred before Tuesday and Thursday trading. This scenario generates the best performance out of Tuesdays while Thursday tends to follow-through on the selling.
These patterns are important because they can facilitate better timing for entering and exiting trades. Given the bullish bias of the market, it may also be reassuring to know that steep down days tend to be followed by reversals on Tuesday and/or Friday. The absence of such anticipated responses could signal an important (short-term) change in behavior in the market that will require a change in trading bias. After all, once regular patterns are well-known, accepted, and anticipated, traders tend to try to front-run those patterns or even fade them. Finally, these patterns are great reminders that the market's behavior is not completely random (perhaps the Federal Reserve's printing press only works best on certain days) even if the anomalies have a constrained shelf-life. These rhythms may be driven by the regularity of news releases, earnings reports, or some other periodic news. No matter the reasons, it should pay to at least monitor them for consistency and change.
Finally, note that as I updated the numbers, I noticed a few errors from last year, particularly in the graphs for 2011 data. Fearing a "Reinhart and Rogoff Excel Moment," I thoroughly rechecked all past numbers and the structure of my queries. I was relieved to find that almost everything else was fine. Also note that I did not re-confirm last year's finding that these daily trading patterns are statistically significant. I saw no reason to repeat those calculations.
Here are the charts (price data from Yahoo!Finance):
Average Daily Price Change of the S&P 500 Based On Day of Week and Performance of Previous Days (Jan 3, 1950 - May 15, 2013)
Average Daily Price Change of the S&P 500 Based On Day of Week and Year (Jan 3, 2007 - May 15, 2013)
Average Daily Price Change of the S&P 500 Based On Day of Week and Year When Previous Trading Day Closes Down (Jan 3, 2007 - May 15, 2013)
Average Daily Price Change of the S&P 500 Based On Day of Week and Year When Previous Trading Day Closes Up (Jan 3, 2007 - May 15, 2013)
Average Daily Price Change of the S&P 500 Based On Day of Week and Year When Previous Trading Day Closes -0.5% Or Lower (Jan 3, 2007 - May 15, 2013)
Average Daily Price Change of the S&P 500 Based On Day of Week and Year When Previous Trading Day Closes -1.0% Or Lower (Jan 3, 2007 - May 15, 2013)