Rethinking Asset Allocation

by: Lowell Herr

One of the primary reasons for developing a well-diversified portfolio is to reduce risk. This blog will focus primarily on the U.S. Equities portion of the portfolio, as we still recognize the importance of including emerging markets, bonds, international real estate, etc. The question to be analyzed is whether or not to include all “Big Nine” asset classes, or can we do as well if the number of U.S. Equity asset classes are reduced to six? To answer this question requires a little analysis, as we are changing multiple variables, and the changes are linked in complex ways.

Since the Bohr Portfolio is coming up for review tomorrow, and it currently holds a percentage in each of the “Big Nine” asset classes, I’ll use it as an example. Below is the familiar Bohr Dashboard. Thirty-four percent (34%) of the portfolio is allocated to U.S. Equities, and I will maintain that percentage through this analysis. Further, I will keep the same percentages for each cap size. For example, 12% will remain with Large-Cap, 11% with Mid-Cap and 11% with Small-Cap.

Bohr “Big Nine” Analysis: Here we have the QPP analysis using the Strategic Asset Allocation percentages. The five metrics to write down are the projected return (8.4%), projected standard deviation (17.6%), Return/Risk ratio (0.48), Diversification Metric (33%), and Portfolio Autocorrelation (21,5%). Keep these values in mind, as they will be compared with a Bohr “Big Six” analysis.

Original Bohr Correlation Matrix: To understand the diversification piece of the analysis, one needs to view the Correlation Matrix, shown below. Note the high correlation among many of the ETFs. Even VEU, developed international markets, is highly correlated with VTI or the U.S. Market. This is exactly why Considine is not including VEU in some of his portfolios. There seems to be little advantage to adding this holding.

Note that VUG, VOT, and VBR are also highly correlated with VTI. In the second half of this analysis, we are going to reduce the allocation to these three growth ETFs to zero and shift the excess percentage over to the value and blend asset classes. Will this shift make a difference? I don’t have a clue until I run the second part of the analysis.

Bohr “Big Six” QPP Analysis: Once more, examine the five critical measurements; projected return, projected SD, Return/Risk ratio (0.48), Diversification Metric (28%) or down five percentage points, and Portfolio Autocorrelation (21.0%). Other than DM, there is little difference between the two portfolios. The added complexity of adding the three growth asset classes does not make a lot of sense with this portfolio.

Bohr “Big Six” Correlation Matrix: If you scroll up and down between the two correlation matrix tables, you see slight changes, but nothing of statistical merit. Remember, we are only shifting nine percent of the total portfolio and the asset classes were initially highly correlated. There seems to be little advantage to adding the three specialized growth asset classes to the Strategic Asset Allocation plan.

Conclusion: Keep the portfolio simple. If we go back to looking at the Swensen Six portfolio, we will likely see better results and projections emerging from that simple portfolio.