Swensen Six Update And Backtesting Results

by: Lowell Herr

Building diversity with six ETFs.

Staying out of trouble with the momentum model.

Performance results since 12/29/2006.

Current recommendations.

David Swensen, chief investment officer of Yale University, lays out a basic portfolio in his book, Unconventional Success: A Fundamental Approach to Personal Investment. In a section titled, The Science of Portfolio Structure, Swensen makes a strong argument for populating a diversified portfolio using only six ETFs. The asset classes are:

  • Domestic equity
  • Foreign developed equity
  • Emerging market equity
  • Real estate
  • U.S. Treasury bonds
  • U.S. Treasury Inflation Protected Securities

It is not difficult to find low-cost commission free ETFs for each of the above asset classes. In the current environment, with low inflation, I substitute U.S. Bonds for U.S. TIPs. Otherwise, all asset classes remain the same in the following analysis.

If one is a Vanguard client, below is the current portfolio of six ETFs. Regardless which discount broker you use, it is possible to find low-cost commission free ETFs to build a "Swensen look-alike" portfolio.

This portfolio includes equity ETFs for periods of strong market action. REITs provide for inflation protection while TLT and AGG are low volatile securities for periods when the market is weak. While Swensen recommends specific percentages be held in each asset class, there are periods when one does not want to be associated with specific portions of the market. Developed international equities and emerging markets currently fit this condition. Now is a good time to avoid international equities. How can one avoid weak asset classes in general and does it make a difference? More on this later.

Current Swensen Six Recommendations: Based on 8/2/2019 prices, the current recommendations are to purchase shares of VTI, VNQ, and TLT. Note that I have the Maximum Number of Assets set to three (3). The backtesting data explains why the number 3. The specific number of shares are recommended based on a $100,000 portfolio. The following worksheet flows out of a larger spreadsheet known as the Kipling.

The reason for selecting a maximum of three ETFs is based on backtesting data. See the graph later in the article.

Backtesting Performance Data: The following data runs from 12/29/2006 through 7/31/2019. For more historical testing of this momentum model, we need to find mutual funds with long records as the ETFs we use have not been in existence to test the bursting of the tech bubble. This momentum model has been tested using longer-running mutual funds. The same risk protection that shows up during the 2008 and early 2009 period is evident in the correction period at the beginning of this century.

The top-performing graph shows the results when three (3) assets are selected. The very bottom curve is the VTTVX target fund or the benchmark. It is difficult to distinguish the difference in performance when 5 and 6 ETFs are selected as the curves frequently rest on top of each other.

Here are the specific end dollar amounts for the different number of assets. The portfolio begins with $100,000. Rebalancing or portfolio reviews occur every 33 days so as to avoid the wash rule and short-term commission fees. In addition, the 33-day rotation results in the updates occurring on different days of the week.

  • One asset: $245,132
  • Two assets: $348,365,
  • Three assets: $402,304
  • Four assets: $335,963
  • Five assets: $317,723
  • Six assets: $320,629
  • VTTVX Benchmark: $202,712

Here are a few of my personal observations based on using this momentum model for a number of years:

  1. In a strong bull market, it is difficult to keep pace with the S&P 500.
  2. Overall, this momentum model throttles volatility. It is not unusual to see recommended portfolios come in with beta values below 0.5.
  3. The model performs very well in poor or less-than-stellar markets such as we experienced in 2018. 2008 and early 2009 are made for this model.
  4. If one weighs portfolio risk as important as portfolio return, this model is a good fit.
  5. No stock analysis is required, and the resulting portfolio ends up with greater diversity than any portfolio built around individual stocks.

I continue to evaluate, test, and monitor numerous portfolios using this model and slight variations of this momentum approach.

Disclosure: I/we have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.

Additional disclosure: I hold similar ETFs at Schwab such as SPTL and SPAB.