Daily Commentary - Posted on Sunday, March 7, 2010, 10:21 PM GMT +1

9 Comments


Mar Sunday 7

Seasonalities: SPY's ex-Dividend Days

First of all it seems that the bullish setup triggered on Friday, February 19, 2010 (five consecutive sessions with an RSI-High/Low (2-day) closing above 95, see my posting Modified RSI(2), Buying Power and Intermediate-Term Outlook) is currently playing out quite nicely.

The setup called for a higher close 5, 10, 15 and 20 sessions later based on an unbroken streak of sessions with a close at least in the upper half of the daily trading range, regularly close to the respective intraday high (buyers are  consistently overwhelming sellers during the second part of the respective sessions).

Table I below shows the SPY‘s historical performance (since 01/01/2005) over the course of the then following 1, 5 , 10, 15 and 20 sessions (1 month later) assumed one would’ve bought the SPY on the close of a session when the signal (5 consecutive sessions with an RSI-High/Low (2-day) above 95) had been triggered in the past:

Since 01/01/2005 (25 occurrences) the SPY was always trading at a higher level one month later (in this event on March 19, 2010), and never lost more than -0.20% three weeks later (in this event on March 12, 2010).

______________________________________

But now to another seasonality which I haven’t noticed (I didn’t find any postings about potential abnormalities around the SPY‘s ex-dividend days as well) until I made some adjustments (and minor corrections) to my Matlab coding in order to account for SPY dividend and cash deductions in the correct manner.

Table II below shows the SPY‘s historical performance on the then following session after the following setups had been triggered in the past (since inception of the SPY on 01/22/1993):

  • (Setup 1) the session immediatley preceding a SPY‘s ex-dividend day (the output is the SPY‘s performance on an ex-dividend day)
  • (Setup 2) SPY ex-dividend day (the output is the SPY‘s performance immediately following an ex-dividend day),
  • (Setup 3) SPY ex-dividend day, and the SPY closed up as well,
  • (Setup 4) SPY ex-dividend day, and the SPY and the S&P 500 posted contradictory signs concerning their respective end-of-day performance (SPY up and S&P 500 down, or vice versa),
  • (Setup 5) SPY ex-dividend day, and the SPY‘s and the S&P 500‘s end-of-day performance differed by at least 0.25%.

It is interesting to note that

  • probabilities and odds on a SPY‘s ex-dividend day are (heavily) lopsided on the upside, while probabilities and odds on a session immediately following a SPY‘s ex-dividend day are (equally) lopsided on the downside, especially in the event of a higher close on an ex-dividend day,
  • since inception of the SPY on 01/22/1993 there were 10 occurrences where the SPY closed up and the S&P 500 closed down (or vice versa) on a SPY‘s ex-dividend day, and on all 10 occurrences the SPY closed lower on the then following session, and last but not least
  • since inception of the SPY on 01/22/1993 there were 39 occurrences where the SPY‘s the S&P 500‘s end-of-day performance differed by at least 0.25% on a SPY‘s ex-dividend day, and on 31 out of those 39 occurrences the SPY closed lower on the then following session.

In order to account for any potential data and/or coding issues (especially with respect to the SPY‘s dividend and cash deductions and respective end-of-day performance), I thought it would be mandatory to verify these stats at least with one other index’ performance data (without any dividend and/or cash payments).

Table III below shows the Russel 2000‘s historical performance on the then following session after the setups listed above had been triggered in the past (means the SPY and the S&P 500 were only used as a trigger while performance stats are derived from the Russel 2000‘s historical performance):

The SPY and the Russel 2000 show only (if any) minor deviations (quite comparable to the at-any-time deviations in their respective end-of-day performance) in their respective performance stats, so the hypothesis of any data and/or coding issue with respect to the SPY‘s and/or S&P 500 end-of-day performance and data may be rejected.

At least based on the respective t-scores, chances that these abnormalities occurred by pure chance only are very small. A rationale for setup 1, 2 and 3 could be that commercial and institutional traders/investors seem to be interested to hold the SPY on an ex-dividend day (may be due to tax reasons), and to immediately get rid of their invest on the then following session (especially in the event of a higher close), but I don’t have a rationale for the statistical abnormalities behind setup 4 and 5.

Successful trading,
Frank

________________________________

If you might want to be instantly notified about what’s happening in the markets and at TRADING THE ODDS, I encourage you to subscribe to my RSS Feed or Email Feed, and (or) follow me on Twitter.

xx

Disclaimer: No position in the securities mentioned in this post at time of writing.

The information on this site is provided for statistical and informational purposes only. Nothing herein should be interpreted or regarded as personalized investment advice or to state or imply that past results are an indication of future performance. The author of this website is not a licensed financial advisor and will not accept liability for any loss or damage, including without limitation to, any loss of profit, which may arise directly or indirectly from use of or reliance on the content of this website(s). Under no circumstances does this information represent an advice or recommendation to buy, sell or hold any security.

Add to Technorati Favorites

Comments (9)

 

  1. kyle says:

    ok, hit the wrong comment area the first time (first version in previous day’s post). short version: when you say cash payment day do you mean ex dividend day? is the ex day the downside day?

    thanks

    Kyle

    • admin says:

      kyle,

      thnaks a lot for pointing this out.

      I meant the ex-dividend day (not the payment day, but the day when it goes ex-dividend). In order to avoid any mfurther misunderstandings, I made the respective corrections.

      Best,
      Frank

  2. kafka says:

    I couldn’t repeat ex-dividend test. I got bearish edge on ex-dividend day using SPY, Close to Close. I have these returns on ex-dividend day:
    2008-09-19 0.0337303240
    2008-12-19 -0.0123194087
    2009-03-20 -0.0282493033
    2009-06-19 -0.0019518543
    2009-09-18 -0.0041060097
    2009-12-18 0.0002722817

    Could you post yours?
    I didn’t find any edge on next day (ex-dividend+1). For example:
    2008-09-22 -0.022639381
    2008-12-22 -0.012813244
    2009-03-23 0.071828966
    2009-06-22 -0.029986962
    2009-09-21 -0.002529985
    2009-12-21 0.010162417

    • admin says:

      kafka,

      the first (most important) problem with your data is that on a SPY’s ex-dividend day you’ve to add the respective dividend and/or cash deduction to the SPY’s closing price in order to calculate the SPY’s performance on the respective day in a correct manner (instead of using the raw data as regularly published by any data provider.

      The 2nd problem is that with respect to the SPY’S closing data in 199x, 3 of (my) 5 data provider published deviating quotes (for an unknown reason).

      I’ll make an adjustment to the stats now utilizing slightly different quotes which are at least congruently reported by at least 3 data provider.

      Best,
      Frank

      • kafka says:

        I tried to use Close and Adjusted prices (Yahoo as data provider). I got only insignificant results…

        • admin says:

          kafka,

          don’t use Yahoo’s adjusted prices (Yahoo seems to adjust each and every closing). Simply add the dividend to the SPY’s closing price on the ex-dividend day in order to calculte the ex-dividend day’s performance, and use the reported close (without the dividend) for the next session’s performance. It’s only a one-time deduction, the next session’s performance is always based on the raw, reported data without taking into account the one-time deduction.

          Best,
          Frank

  3. Phileo says:

    Hi Frank,

    Just curious, is SPY ex-dividend date always 3rd Thurs. of every quarter end month (ie. Dec, Mar, Jun, Sep)?
    If so, this happens to line up with Quad Witching day (the next day)

  4. Nice post and this enter helped me alot in my college assignement. Gratefulness you for your information.

Leave a Reply

Your email address will not be published. Required fields are marked *