Daily Commentary - Posted on Saturday, May 23, 2009, 11:19 PM GMT +1

10 Comments


May Saturday 23

Memorial Day, the VIX and Other (Un-)Favorable Seasonalities

WE031672-klein

Monday next week will be Memorial Day, and additionally we’re entering into the favorable (with a bullish bias) end-of-month period for stocks -at least as an adage says-.  As always I decided to take the quantitative approach concerning this (potential) end-of-month phenomena in oder to check if there is any historical and statistical evidence which supports the thesis, and to check it the sessions following Memorial Day and (almost, except Good Friday) all other exchange holidays as well show any potential peculiarities possibly providing a tradable edge.

I checked for those occurrences since 01/02/1990 where the following setups were triggered, assumed one would have bought the ES (E-mini S&P 500 Futures) on close of the session immediately preceding the respective exchange holiday for a holding period of 5 days (1 week). Therefore concerning Memorial Day one would have bought the ES on close of Friday immediately preceding Memorial Day (in this event Friday, May 22, 2009), and closed the trade on close of Monday one week after Memorial Day .

The US exchange holidays are celebrated …

  • New Year’s Day – January 1
  • Martin Luther King, Jr. Day – observed on the third Monday of January
  • Presidents Day – observed on the third Monday of February
  • Memorial Day – observed on the last Monday of May
  • Independence Day – July 4
  • Labor Day – observed on the first Monday in September
  • Thanksgiving Day – observed on the fourth Thursday of November
  • Christmas Day – December 25

____Left Out____

Good Friday

____No Exchange Holidays____

Columbus Day – observed on the second Monday in October

Veterans Day – November 11

_______________________

At first Table I below shows the E-MINI S&P 500 (ES)‘ performance (since 01/02/1990) concerning setups S1 to S5 -side by side- over the course of the then following 5 sessions assumed one would have bought the ES (S&P 500 futures) on close of the session immediately preceding the respective exchange holiday (E-mini S&P 500 Future’s day session). Setups S1 to S5 are defined as

Table I

  • Setup S1: New Year’s Day – January 1
  • Setup S2: Martin Luther King, Jr. Day – observed on the third Monday of January
  • Setup S3: Presidents Day – observed on the third Monday of February
  • Setup S4: Memorial Day – observed on the last Monday of May
  • Setup S5: Independence Day – July 4

20090523-ES-1

Table II

  • Setup S1: Labor Day – observed on the first Monday in September
  • Setup S2: Thanksgiving Day – observed on the fourth Thursday of November
  • Setup S3: Christmas Day – December 25
  • Setup S4: buy on close 3 sessions before the last session of the month, sell on close of the 2nd session of the new month

20090523-ES-2

With a win/loss ratio of 11:8 and a profit factor of 2.71 compared to an at-any-time profit factor of 1.15 for the ES, the week after Memorial Day not only shows the highest probability for a higher close 5 sessions later, but the highest profitability (pay-off) off all 8 exchange holidays listed above as well. Additionally the end-of-month period shows as well a win/loss ratio and profit factor (1.63) well above the respective at-any-time probability for a higher close 5 session later and profit factor (1.13).

But due to the VIX’ (S&P 500 Volatility Index) peculiar behavior on Friday’s session – the VIX posted an intraday low of -2.49% below Thursday’s close, but completely reversed course and close up +4.10% on the day although the SPX closed almost unchanged -, I had a hunch that it would be worth some time and effort in order to check if the VIX would -as it regularly does- mirror (but negatively correlated) the ES mini’s bullish bias during Memorial Day week and would show a bearish bias respectively. But now comes the surprise …

At first Table III below shows the VIX‘ performance (since 01/02/1990) over the course of the then following 5 sessions assumed one would have ‘bought’ (hypothetically and for statistical purposes only, the VIX itself is no tradable asset) the VIX on close of the session immediately preceding Memorial Day (regularly the Friday before). The ‘Trigger’ column shows the VIX’ value on close of the session immediately preceding Memorial Day (like last Friday). ‘next session’ is the session immediately following Memorial Day and so on.

20090523-VIX-2

It is especially remarkable -to say the least- that the VIX closed up the sesion immediately following Memorial Day on 18 (!) out of the last 19 years, for an extraordinary hypothetical profit factor of 37.93 compared to an at-any-time profit factor for the VIX of 1.09 only, especially due to the extraordinary winn/loss ration of 18:1 and the fact that the VIX closed higher the next session +4.00% or (significantly) more on 13 out of the last 19 years (and +8% or more on 9 occurrences). Additionally the VIX was trading above the close of the session immediately preceding Memorial Day one week after Memorial Day with a probability and hypothetical ‘profit factor’ (magnitude of change on the upside compared to the downside) significantly above the respective at-any-time probability for a higher VIX close 5 sessions later and ‘profit factor’ as well.

Table IV below now shows the VIX‘ intraday performance (since 01/29/1993) concerning the open, high, low, close (compared to the previous’s session close) and close versus open on the next session (in this event Tuesday, May 26) immediately following Memorial Day.

20090523-VIX-3

If history is any indication concerning the VIX’ performance on Tuesday, May 26, the VIX may be looking forward to -from a historical and statistical point of view- one of it’s probably most bullish sessions (means up-trending VIX, not the major market indexes) of the year, if not THE session (will be suibject to a follow-up) with the highest probability for a higher open/high/low/close and magnitude of change (on the upside) as well. During the last 8 years, the least higher open on the session immediately following Memorial Day was +4.44%; since 01/02/1990 the VIX always posted a high above the previous session’s close, and in only 4 out of 19 years did the VIX post an intraday low below the previous session’s close at all while the VIX left an unfilled gap on the upside in the other 15 years .


P.s.: Due to the significance of the findings I thought of a data or programming issue and manually (double-)checked (and verified) at least the last 9 occurrences (since 2000).

________________________________

Bottom line:

  1. From a historical and statistical point of view, the week after Memorial Day regularly shows a positive bias concerning the probability for a higher close 5 days later and the respective profit factor as well.
  2. But -again from a historical and statistical point of view- the ‘asset’ with the most bullish bias (on the upside) during Memorial Day’s week may probably be the VIX. Regularly there is an inverse relationship (negative correlation) between the S&P 500 and the VIX, and a signifcant bullish bias concerning the VIX would spell more than trouble concerning the SPX’ performance over the course of the then following 5 sessions, but Memorial Day’ week seems to be a period in the market where this (negative) correlation is completely decoupled.

Successful trading,

Frank

Disclaimer: (Net) Long VIX futures at time of writing.

P.s.: WordPress recently implemented a Twitter widget, so I’ll regularly make some intraday updates as well using Twitter. If you’re interested in, please have a look at the blog during the trading session as well or subscribe directly to Twitter (recommended).

Comments (10)

 

  1. moneyfriend says:

    Excellent post! Is it possible that the Vix has a tendency for this behavior over most long weekends??

  2. Steffen says:

    Good analysis, as always!
    Regarding your findings related to the VIX, one possible explanation is that implied volatily has the tendency to drop before long weekends. Options usually become cheaper the day before a long weekend in order to compensate to a certain amount for the 3 days of time decay.

  3. ADD says:

    Dear Frank,

    Consider the data, is it your intention to remain long on the VIX the next 3-4 sessions while monitoring daily for a possible reversal? The scenario you’ve presented is very inticing and, yes, the bulls are facing many headwinds in order to proceed but so many people are expecting a bounce in the VIX here, I wonder if we would bounce on Tuesday, then grind lower the remaining 3 sessions on below average volume and sideways consolidation in the S&P.

    If you wouldn’t be making the VIX trade, then would you otherwise be long BGU on 05/26 weakness and possibly holding market close on 05/28? What would that data look like?

    I guess the concern is about not knowing all the intricacies of the VIX even as overwhelmingly positive as the data illustrates. And maybe that’s the whole point of your fantastic blog: Having trust in the data for creating favorable odds even while not having the expertise in the trading vehicle–yet remaining vigilant in the trade as always.

    Any insight is appreciated. Thanks in advance.

    Thank you,
    -A.

  4. Adam says:

    Hey Frank:

    A couple things.

    The VIX holiday bias can go on at least week, knocking it down a bit before a holiday and then giving the illusion of a lift after. To me, that’s a lot of what’s going on. Not all however , some of the dip is real.

    And like you say, you can’t *buy* the VIX, only the futures or ETN’s. They haven’t existed for the 19 years you look at, and I imagine if they did, the VIX future or ETN would not have dropped that frequently.

    Now I’m being nitpicky, I love your site, your stats work is fantastic, just wanted to note VIX is quirky around holiday’s.

  5. CarlosR says:

    Frank,

    Thanks again for your blog, it’s very valuable and enlightening.

    I have one *extremely* small nit to mention: the correct spelling of the word is “losing”, not loosing. The only reason I mention this is because you (naturally) wind up using this word in almost every post you make, so you might as well spell it right.

    I think you may have mentioned that English wasn’t your primary language, but honestly, your English is better than that of many native-born English speakers I know!

    Keep up the good work!

Leave a Reply

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