S&P500 stocks correlation with SPY

Discussion in 'Trading' started by caementarius, Feb 6, 2009.

  1. I've done some more analysis with a dataset to understand better how stocks relate to indices. The numbers I came up with are that the average correlation of a component with SPY is .62 with a stddev of .1265

    Here is a histogram of where the stocks fall on the distribution. 7 stocks I didn't have contiguous data for so they are left out of the analysis.

    Like before, I'm just posting what I'm looking into in the hopes of getting feedback and bouncing around trading ideas.

  2. Nice chart!

    You might get some interesting insights if you do the following. 1. Find the median return on the SPY. 2. Pull the data for all the days when SPY produced above median returns and create a version of this chart. 3. Pull the data for all the days when SPY produced below median returns and create a version of chart.

    From what I've read, you should see might higher correlations on the down days than on the up days.
  3. A more useful test IMO, rather than to use the median, is to apply it to data on days/weeks where S&P500 exceeded historical 5 sigma deviations. That's where you should expect average correlations to really shift towards the right.

    Some ideas along these lines are developed here:


    A very good read and a notch above most, however, I am a bit skeptical of some of the results.