Is there an explanation why rebalancing the Feynman using the Hoadley optimizer does little to improve portfolio returns? The answer may partially lie in the correlation matrix. What I did was take one of the last asset allocation plans for the Feynman and paste it over into the percentage section of the Quantext Portfolio Planner (QPP) software program. The following correlation matrix table shown below may provide some clues as to why greater returns were not experienced when rebalancing took place every quarter.
Correlation Matrix: Many of the ETFs that make up the Feynman are highly correlated. Not always, but many of the lower correlated ETFs make up smaller percentages of the portfolio. If the optimizer calls for shifts in allocation between ETFs that are highly correlated, the returns of the portfolio will not change significantly. Developed international (EFA) and emerging markets (VWO) are highly correlated with the “Big Six” U.S. Equity ETFs (first six on the list), so we would not expect to see much change in portfolio returns if shares are moved around among this group of ETFs.
The question remains – will returns increase if we add in the momentum factor? So far the IRR values for most of the portfolios is increasing when measured against the ITA Index benchmark. Bohr and Kenilworth each lost 60 Basis Points to the ITA Index. All others gained from 20 BP to a high of 230 BP for the Gauss Portfolio.
Comments and questions are welcome.