Centre county Pennsylvania, USA | "Stock market returns aren’t normally distributed. Tails are fatter. Even if they were not sure what such a small sample size would be good for. Your example is more likely to be a contrarian indicator than anything after the robust returns we’ve had. Maybe that is what you are intending to use it as?"
Agree with all of your thought provoking comments, thanks. Obviously more work is needed before portfolio CAGR can be used as a (reliable) technical indicator.
Although stock market returns are not normaly distributed, portfolio CAGR may be much more normally distributed (probability of portfolio CAGR is the statistic we are seeking). As a check of that hypothesis, trailing CAGR for 19 years of that portfolio was computed. Portfolio standard deviation for each of those 19 years was normalized to 19 year average standard deviation and scatter plotted as shown in attached chart. The chart shows that portfolio trailing CAGR standard deviations were all within +- 2 standard deviations of the porfolio's mean value (the expected value). For this check, sample size was increased from 6 independent samples (6 years) to 19 independent samples (19 years) to address the sample size comment. My available data is limited to 19 years of trailing returns.
This check also shows that the recent 6 year robust returns are not a true indicator of longer (19 year) returns where trailing years 10,16,17,and 18 produced strong negative CAGR standard deviations.
(portfolio_6 CAGR std dev scatterplot (full).jpeg)
Attachments ---------------- portfolio_6 CAGR std dev scatterplot (full).jpeg (38KB - 57 downloads)
|