Jun

17

In his studies of Life Histories, Eric Charnov likes to look at the sex ratio as it is feeds back with clutch size, weight, and resources.

He makes the point that often ratios are invariant and that there are many causes that contribute to the invariance, which must be teased out of the variability of what we would call demographic variables.

Inspired by Charnov, I looked at the sex ratio that follows periods without sunlight in the form of a daily rise or decline in the S&P, a factor almost as crucial as sex and its ratios to many market participants.

The period covered was the last 10 years of data, during which the unconditional chance of a daily decline was 0.52.

The sex ratio appears to be much more constant after days of sunhine for bulls.


Comments

Name

Email

Website

Speak your mind

6 Comments so far

  1. Russell Sears on June 17, 2009 10:25 am

    Looking at only the last two years, the ratio for days after sunshine are much different that the 10 year period. While the conclusion for cloudy days are similar. In hindsight, following yesterdays trend, would put you consistently on the wrong foot. No matter if it was sunny of cloudy. Perhaps this is consistent with a bear market. I will leave the stats and design of a test of the last statement as an excercise for the reader.

  2. Adam Grimes on June 17, 2009 1:52 pm

    I have some interesting work on individual stocks and the same concept. One of the keys to doing this kind of analysis is to say exactly X days up or down and not at least X days. Happy to share some of the work I've done. Drop me an email: grimes.111//at//gmail//dot//com

  3. vniederhoffer on June 17, 2009 10:15 pm

    harold thayer davis, a great comparable to bachelier, cowles, and the monthly weather review authors, has an excellent analysis, with formulas and empirical data for the dji in his great book 'the analysis of economic time series' – probably the best discussion of runs of exactly length n that i've seen. vic

  4. Adam Grimes on June 18, 2009 9:35 am

    Thanks for the book recommendation. I don't know that textbook at all but will look it up and get it. I've learned a lot from the books you've recommended over the years.

  5. Laslo Minks on June 19, 2009 9:05 pm

    I don’t know where else to post this so I put it here for now:

    I am interested in the study of what tends to happen when the correlation between 2 markets changes. Here I’m looking at the love/hate relationship between S&P/Oil. I have taken 20 day moving correlation coefficients from begin 2007 through present. As proxy for oil price I have used the USO ETF, for my own convenience.

    EOM S&P USO S&P/USO Corr
    07
    2 1407 51 .28
    3 1421 53 .25
    4 1482 51 .22
    5 1531 49 .04
    6 1503 53 .11
    7 1455 59 .19
    8 1458 56 -.02
    9 1527 63 .25
    10 1549 73 .27
    11 1481 70 -.14
    12 1485 76 .25
    08 1 1378 72 .18
    2 1331 80 .17
    3 1323 81 .22
    4 1386 93 .13
    5 1400 103 -.53
    6 1280 114 -.57
    7 1267 100 -.27
    8 1283 93 -.35
    9 1165 82 .61
    10 969 56 .65
    11 869 42 .81
    12 903 33 .20
    09 1 826 29 .58
    2 735 27 .45
    3 798 29 .47
    4 873 29 .86
    5 919 36 .71

    Note that since the 20 day corr are monthly they are (with minor exceptions) non-overlapping. The appearance, not yet quantified, is that there is corr persistence akin to volatility persistence.

    Note the cycle change at Sept 08. The hypothesis arising from this observational data is: does a dramatic change in love/hate relationship anticipate a reinforcing feedback loop? More data needed. Admittedly, the concept of feedback loops reeks of trend-following. Feedback welcome.

  6. vniederhoffer on June 21, 2009 6:26 pm

    Trustfully, Mr. Minks is using returns and changes as his data. And trustfully he will ask the more important question that Kendall first wrote about in his classic paper as to what the predictive correlations are. vic

Archives

Resources & Links

Search