2 ms·
Looks legit. On a quick skim, happy to see specific measures taken like - Population used as weights - Includes both nominal and real returns - Potential for
by tomrod 8y ago
Looks legit. On a quick skim, happy to see specific measures taken like
- Population used as weights
- Includes both nominal and real returns
- Potential for Simpson's paradox addressed
Critiques
- The Sharpe ratio is calculated for each group, but the weighted standard deviation is not reported. Given the higher volatility of the equities market versus the housing markets it would be nice to see the IQR.
- Population _growth_ may be a large factor in housing pricing growth -- you can't capture that with birth rates alone (need inflow from immigration) but level-population controls for some of that. But (and I've only skimmed) population may make for a poor time-series control due to serial correlation.
- While housing comes out on top, the p-value criterion shows only the slightest bit of statistical difference (t-stat 1.63) for geometric mean of housing v. equities -- it's on the border of statistical significance. [US Dollar returns, table A-14, geometric mean comparison]
Edit: Simpsons paradox, not ratio
- lordnacho 8y agoSimpson's paradox perhaps? Googling "Simpson's ratio" gives you some stuff about the aspect ratio of Simpson's episodes.
- ianai 8y ago“Simpson’s paradox – the effect that occurs when the marginal association between two categorical variables is qualitatively different from the partial association between the same two variables after controlling for one or more other variables…”. https://365datascience.com/simpsons-paradox/ https://365datascience.com/simpsons-paradox/ Addressed by per capita or inflation adjustments?
- tomrod 8y agoI typed faster than I intended. Yes, paradox. Another comment hit by autocorrect. Thanks for the correction.