Fourteen years after the Brinson, Hood, and Beebower paper "Determinants of Portfolio Performance" appeared in the Financial Analysts Journal, the opportunity arose to set up a passive portfolio using ETFs. In the late 2000s the number of ETFs were limited and iShares was the primary provider. Nevertheless, there were a number of broad-base index ETFs, sufficient to construct a well-diversified portfolio. As a guideline for the asset allocation plan, various portfolios found in William J. Bernstein's first book were studied. "The Intelligent Asset Allocator" was an valuable source of information and I continue to rank it in the Top 10 Investment books of the past decade.
From Bernstein's book and Robert A. Haugen's "The New Finance," I learned of the Fama and French celebrated study, "The Cross-Section of Expected Stock Returns," published in The Journal of Finance, June 1992. Bernstein emphasized diversification and Fama & French made clear the importance of value loading in a portfolio. Using these principles, a Strategic Asset Allocation (SAA) plan, very similar to the dashboard screen shot below, was established in late 2000. Value loading involves tilting the asset allocation percentages toward the value side of the investing spectrum. In addition to skewing toward value, the portfolio SAA also carries a heavier weighting in mid-cap and small-cap asset classes compared to what one finds in a total market index fund such as the VTSMX.
Initially, the portfolio did not contain commodities and international REITs as the ETFs were limited if non-existent. For the first few years of operation, this portfolio contained no bonds and this hurt performance during the "lost decade" from 2000 through 2009. Today the portfolio is made up of 26 ETFs. Breaking it down into major sections of equities, bond-income, international, commodities, and REITs we have the following.
Equities: VTV, VOE, VBR, VV, VO, VB, VUG, VOT, and VBK. These are the "Big Nine" U.S. equity asset classes.
Bond-Income: BND, TIP, IEF, HYG, and DVY (dividend oriented)
REITs: ICF, VNQ, and RWX
International: EFA, EPP, EEM, VEU, and VWO
Commodities: GLD, SLV, DJP, and GSG
Yes, there is overlap in certain asset classes and that is due to rebalancing, a rare event with this portfolio. When an asset class is above target and it holds overlapping iShares and Vanguard ETFs, we sell the ETF with the higher expense ratio and move to the ETF carrying the lower fee.
Below is the SAA plan and it closely resembles the original plan of 2000. Percentages with the white background are the target percentages and the color coded backgrounds are the actual percentages held in the 16 asset classes.
Over the 10 years of operation, this portfolio outperformed the VFINX and VTSMX by 3.2% and 2.0% points respectively. The percentages displayed below are all annuallized values.
The ITA Index is a customized index that is designed to track this particular portfolio. If each asset class is on target the ITA Index will closely match the internal rate of return of the portfolio. Note that VTSMX topped the VFINX by 1.1% points annually. Since the portfolio also held a higher percentage in value and smaller cap ETFs, the advantage went to the portfolio over the benchmarks. Value loading, as advocated by Fama & French, worked to the benefit of this portfolio.
Critics of the above portfolio will argue that it carries too high a risk or excessive volatility. The counter point is presented in the Sortino Ratio (SR) and Retirement Ratio (RR) data. Note the high SR and RR values. In these calculations, the portfolio manager is only penalized for downside volatility. Upside volatility is ignored as we want investment movement to the upside.
Will this portfolio and the skewed SAA plan continue to work in the future? No one can say for sure, but over the last decade several principles worked very well: Index ETFs with a tilt toward value was beneficial, global diversification served this portfolio very well, the portfolio was passively managed, as there were very few trades after the original investments, and the SAA was weighted toward smaller cap investments.
To continue to track the performance of this portfolio, known as the Schrodinger, go to this site. Portfolio performance data is published nearly every week on this portfolio and 20 other active portfolios.