A Quantitative Approach to Tactical Asset Allocation | #86

Episode Summary

Executive Summary: Meb Faber revisits his original 2007 tactical asset allocation paper and assesses a simple 10-month moving-average trend-following system over 10+ years. He argues the model’s main value is risk reduction, not market-beating returns, and shows it generally improved returns, volatility, and drawdowns across U.S. stocks and diversified global portfolios. He also discusses practical variants and stresses the need for a disciplined investing process.

Main Topics: Origin and evolution of the original tactical asset allocation paper (Priority: 5/5): Faber recounts how the first white paper emerged from his CMT requirements, was initially overlooked, then became highly influential after the financial crisis. The simple 10-month moving-average trend-following rule (Priority: 5/5): He explains the model: buy when monthly price is above the 10-month SMA, sell when below, using monthly rebalancing and cash/T-bill returns. In-sample and out-of-sample performance (Priority: 5/5): The transcript compares results before and after publication, showing improved returns and materially lower volatility and drawdowns in both periods. Diversified global asset allocation and timing (Priority: 5/5): Faber extends the analysis from U.S. stocks to a global portfolio of stocks, bonds, REITs, and commodities, showing timing reduced drawdowns sharply. Extensions and portfolio customizations (Priority: 4/5): He outlines three ways to modify the model: alternative cash management, alternative weightings (conservative/moderate/aggressive), and adding more asset classes or factor tilts. Practical implementation considerations (Priority: 4/5): Fees, commissions, slippage, and taxes are discussed as real-world constraints, with the recommendation that tax-deferred accounts are preferable. Long-term investing and the value of process (Priority: 4/5): The closing message emphasizes that investors need a plan for multiple market environments, especially when valuations and future returns may be less favorable.

Key Arguments: The strategy was designed as a risk-reduction tool, not primarily as a return-maximizing model, though it often improved returns as well. A simple monthly 10-month moving-average trend rule can remove much of the emotion from investing and help investors avoid catastrophic bear-market losses. Out-of-sample results after publication were broadly consistent with expectations, suggesting the model was robust rather than a data-mined anomaly. Timing can underperform in strong bull markets but adds value over full market cycles by avoiding deep and prolonged bear markets. Diversification remains powerful, but its benefits can weaken in crisis periods when many asset classes fall together; timing can help reduce that cross-asset drawdown risk. The model’s performance appears stable across nearby parameter choices, which supports confidence in the general trend-following approach. Investors can tailor the system through cash substitutes, asset-class weights, and factor tilts to better match their risk tolerance and objectives. Real-world frictions matter, especially taxes; the strategy is more attractive in tax-advantaged accounts and with low-cost implementation. Given elevated U.S. valuations and low bond yields, investors should have a flexible, rules-based process prepared for a wide range of future outcomes.

Data Points: Original paper publication year: 2007 - Faber says the paper was circulated in draft around 2005–2006 and published in 2007. Most downloaded SSRN paper: ~200,000 downloads - He notes the paper became the most downloaded of all time on SSRN. Trend filter: 10-month simple moving average - Monthly version of the classic 200-day trend-following rule. U.S. stocks in-sample return: 9.65% buy-and-hold vs 10.36% timing - 1901–2005 historical U.S. stock sample. U.S. stocks in-sample volatility: 18% buy-and-hold vs 12% timing - 1901–2005 historical U.S. stock sample. U.S. stocks in-sample Sharpe ratio: 0.33 buy-and-hold vs 0.55 timing - 1901–2005 historical U.S. stock sample. U.S. stocks in-sample max drawdown: 83% buy-and-hold vs 50% timing - 1901–2005 historical U.S. stock sample. U.S. stocks out-of-sample return: 7.6% buy-and-hold vs 8.5% timing - 2006–2016 period after publication. U.S. stocks out-of-sample volatility: 14.6% buy-and-hold vs 9.4% timing - 2006–2016 period after publication. U.S. stocks out-of-sample Sharpe ratio: 0.45 buy-and-hold vs 0.8 timing - 2006–2016 period after publication. U.S. stocks out-of-sample max drawdown: 51% buy-and-hold vs 16% timing - 2006–2016 period after publication. Global buy-and-hold portfolio return (72–05): 11.5% - Equal-weighted buy-and-hold of U.S. stocks, foreign stocks, bonds, REITs, commodities. Global buy-and-hold portfolio volatility (72–05): ~9% - Equal-weighted buy-and-hold global asset allocation. Global buy-and-hold Sharpe ratio (72–05): 0.6 - Equal-weighted buy-and-hold global asset allocation. Global buy-and-hold max drawdown (72–05): <20% - Equal-weighted buy-and-hold global asset allocation. Global buy-and-hold return (2005 onward): 3.5% - Out-of-sample performance for the diversified buy-and-hold portfolio. Global timing return (2005 onward): 4.8% - Out-of-sample performance of the timing model on the diversified portfolio. Global buy-and-hold volatility (2005 onward): 12.8% - Out-of-sample diversified portfolio. Global timing volatility (2005 onward): 6.5% - Out-of-sample diversified portfolio. Global buy-and-hold Sharpe ratio (2005 onward): 0.19 - Out-of-sample diversified portfolio. Global timing Sharpe ratio (2005 onward): 0.59 - Out-of-sample diversified portfolio. Global buy-and-hold max drawdown (2005 onward): 46% - Out-of-sample diversified portfolio. Global timing max drawdown (2005 onward): 9.45% - Out-of-sample diversified portfolio. Average cash exposure: ~30% - The timing portfolio is out of the market roughly one-third of the time on average. Alternative cash management uplift: ~1% per year - Using bonds instead of T-bills as the cash substitute increased return by about a percentage point. Alternative cash management uplift (more precise): ~200 basis points - He also describes the return improvement as roughly two percentage points in some comparisons. Conservative weighting: 40% bonds - One suggested lower-risk portfolio variant. Aggressive portfolio tilt: Top 3 of 5 asset classes by momentum plus trend filter - A more return-seeking version combining momentum and trend. Added return from smart beta tilts: ~150 basis points - He estimates extra return from adding value/momentum style tilts to the core allocation. Added return from combining tilts with timing: ~100 basis points - Incremental uplift from layering timing on top of the tilted allocation. Cash/implementation cost estimate: 3 to 70 basis points - He cites a range of ETF expense ratios, emphasizing the importance of low fees. Global stock declines in G7 history: At least one 75% loss in each G7 country - Used to illustrate the severity of historical equity bear markets. Rebound needed after 75% decline: 300% gain - Explanation of how hard it is to recover from severe drawdowns. Shiller CAPE: Above 30 - He cites elevated U.S. equity valuations as a warning for future returns. Long-term Shiller CAPE average: ~17 - Historical average cited for comparison. Long-term bond yield example: ~2.3% - He notes low U.S. government bond yields and weak forward return expectations. Illustrative future nominal return assumption: 4% stocks, 2% bonds, 2% inflation - Used to frame a potentially modest future return environment.

Pivotal Quotes: "The results suggest that a market timing solution is a risk reduction technique rather than a return-enhancing one." — Meb Faber: Core framing of the original paper and the episode’s thesis. "Price is unique as an indicator is that it can't diverge from itself." — Meb Faber: Explaining why the strategy is price-based and purely mechanical. "Do you have a plan or process that has prepared you for tomorrow's market in whatever condition you may find it?" — Meb Faber: Closing call for disciplined, rules-based investing.

Implications: For investors, the episode reinforces that simple trend-following can meaningfully reduce drawdowns and emotional mistakes, especially across long cycles. For the industry, it supports low-cost, rules-based implementation and flexible portfolio design over prediction.