“It’s not the number of assets that determines diversification, it’s the correlation among them.” – Richard Bernstein
Diversification through correlation analysis is another step in our efforts to lower volatility while maintaining a respectable return for the ITA portfolios. In the following example I’ve selected twenty-two (22) ETFs that are frequently used to populate the eleven portfolios tracked here at ITA Wealth Management. Platinum members have access to all the transactions of eight portfolios.
Two correlation tables are presented below. If you compare ETF to ETF you will not come up with exactly the same percentage or value since the Quantext Portfolio Planner (QPP) portfolio is set up with nearly equal percentages while the Hoadley correlation matrix uses equal shares. The 5% you see in the following table is rounded up from 4.6%. In the Hoadley correlation table I assigned a specific number of shares to each ETF. I did use five years of data for each table of data.
If you are unfamiliar with correlation tables, here is an example. In the QPP table, VTV and GLD have a correlation of 14%. This is classified as a low correlation between these two ETFs. The lower the number, the lower the correlation, and that is what we seek as we build well-diversified portfolios. If you scroll down to the Hoadley correlation table you will see the correlation is 0.27 or 27%. The difference is not something to worry over. Don’t attach too much value to the exact percentages, but look at the larger picture. Equity ETFs tend to be highly correlated with each other and equity ETFs tend to have much lower correlations with bond, commodity, treasuries, and precious metals. Both data tables underscore this correlation concept.
Quantext Portfolio Planner Correlations: As mentioned above, the following table is built using 4.6% assigned to each ETF with exception of SDS, a rarely used ETF. SDS is a short-ETF and I include it in the mix of securities as our “investment canary” since it provides a clue to market direction.
One thing I like about the QPP correlation matrix over the Hoadley data is that the QPP shows how individual ETFs are correlated with each other within the context of the overall portfolio. That is the column labeled, Portfolio. As we change the percentage assigned to each portfolio, the percentages in the Portfolio column will vary.
Hoadley Optimizer Correlations: The following table is based on five years of data and equal shares were assigned to each ETF. Emphasize the big picture. Equity ETFs are highly correlated with each other and equity ETFs have much lower correlations with treasury, bond, commodity, and precious metals ETFs. We need to include the possibility of investing in BIV, BND, PCY, DBC, TLT, etc. when the equity market begins to sag and decline. This is how we hold down portfolio volatility. Our momentum procedures also reduce portfolio volatility.
In order to construct a well-diversified portfolio, we need to include low correlated ETFs in our investment quiver.