A little over four months ago, Mebane Faber published A Quantitative Approach to Tactical Asset Allocation, a paper that updates his earlier 2006 publication by adding current market results. If one is not familiar with Faber's risk reduction model, another source of information is The Ivy Portfolio: How to Invest Like the Top Endowments and Avoid Bear Markets.
The following article explains how I am using a slight modification of the Faber model with several portfolios in an effort to test the viability of reducing risk. The key ingredients of the Faber model are the following.
- Use a 10-month simple moving average as a buy-sell trigger.
- Review the portfolio at the end of the month.
- Take advantage of the momentum effect.
- Reduce portfolio risk.
- Moderate the whipsaw problem.
The driving motivation behind this model is to prevent huge losses of the type investors experienced from the late 1960s through the mid-1970s and again from 2000 to 2002. After recovering from the losses in 2000-2002, investors were slammed again in 2008 and early 2009. This risk reduction model is designed to moderate bear markets we are sure to see again.
Here are the modifications I made to the Faber model. Instead of using a 10-month SMA (approximately 200-day), I use a 195-Day Exponential Moving Average as it is a little faster acting. The idea is to move ahead of the herd who follow the 200-Day moving average. Portfolios are reviewed every 33 days instead of every month for several reasons.
- To avoid the wash sale rule.
- To avoid the short-term trading expense should the same ETF be bought and sold within a 30-day period.
- To minimize the whipsaw problem.
- The primary reason for using a 33-day review period is to shift the portfolio review throughout the month rather than always make decisions at the end of the month. Mutual fund managers are tweaking their portfolios at the end of the month and the 33-day review period avoids last of the month market moves.
In Faber's recent paper he speaks of Global Tactical Asset Allocation where he builds a portfolio using five asset classes with equal percentages assigned to each. Those asset classes are: U.S. Equities, Foreign Stocks, Bonds, Real Estate, and Commodities. To cover those asset classes, one can use Exchange Traded Funds (ETFs) such as VTI, VEU, BND, VNQ, and DBC.
The following sample portfolio illustrates how one can implement the Faber model. In the following screen-shots will see an efficient frontier, ranking of commission free ETFs, and buy-hold-sell recommendations.
Efficient Frontier: The efficient frontier graph paints a quick picture of how the current portfolio (a sample in this example) is constructed with respect to an optimized portfolio. Faber states that he makes no attempt to optimize the asset allocation. Portfolio optimization is one of the overlay modifications to the Faber model.
To move the current portfolio (diamond dot) up the efficient frontier toward the optimized portfolio (small red dot), one needs to sell off some of the bond holdings and replace them with equity oriented ETFs. Those recommendations are spelled out in the third screen-shot shown below.
ETF Rankings: Instead of confining the portfolio to five asset classes as Faber does in his Global Tactical Asset Allocation plan, a few more asset classes are added such as emerging markets, metals, international real estate, and international bonds. Faber discusses this option in the later pages of his recent paper so these additions are nothing new.
The following rankings are based on three factors. A 50% weight is assigned to the most recent three-month performance. The second factor is the six-month performance and 30% is allocated to this value. The third metric is ETF volatility and 20% is apportioned to this calculation as we seek lower volatile ETFs.
In addition to ranking the ETFs, users of this table can easily spot which ETFs are currently priced above or below their 195-, 13-, and 34-Day EMAs. The last column measures ETF momentum. This is a relative ranking showing which securities are growing fastest on a daily basis over the most recent three months compared to the most recent six months. SDS, an ultra-short ETF, is showing the greatest momentum and that is directly related to recent drops in the broad market.
All these ETFs are commission free ETFs from TDAmeritrade with exception of GLD (GOLD) and SDS. Most investors will not choose to use SDS, but I include it here should anyone want to see what a hedging ETF looks like in this ranking system.
Buy-Hold-Sell Recommendations: If one is following the Faber model and using only five global asset classes, current investments would include U.S. Equities (VTI), Foreign Stocks (VEU), and Real Estate (VNQ). Since commodities (DBC) and bonds (BND) are priced below their 195-Day EMA, one is in cash with that 40% of the portfolio.
The model (Momentum-Optimization Model or MOM) I use is slightly more complicated as it includes the optimization overlay. For example, if this half-million dollar portfolio were real, one would invest in more VTI shares and sell off all BKF shares. 500 shares would be added to VTV and 180 shares would come out of RWX. Positions in TLT and DBC would be eliminated. These are examples of recommended changes if one is following the MOM plan.
Over at ITA Wealth Management I am applying MOM principles to five portfolios and may expand it to more. The Schrodinger portfolio is used as a standard buy and hold portfolio as it is passively managed and has been for over twelve years.
To test the MOM model, Internal Rate of Return (IRR) values are available for each test portfolio as well as the IRR values for four benchmarks. These are discussed at ITA Wealth Management. Over time the difference between the portfolio performance and benchmark IRR values will be compared to see how well the MOM plan works. As mentioned above, a passively managed portfolio will also be used as a benchmark for the buy and hold model.
Readers interested in following this experiment need only track the Gauss, Maxwell, Euclid, Kenilworth, and Madison portfolios. In the near future I may add the Einstein, Bohr, and Kepler portfolios to the MOM plan.