Three Proprietary Strategies
A trio of proprietary strategies will feature regularly in The ETF Portfolio Strategist (ETF-PS), providing benchmarks for managing risk for a dynamic mix of the major asset classes. The goal: maximizing return for this opportunity set while attempting to keep the darkest of the drawdown demons at bay.
Details can be found below, but first let’s review the results (along with our standard portfolio benchmarks), based on numbers through Friday’s close, Aug. 7 (definitions outlined below):
For a quick visual comparison, here’s how the three proprietary strategies compare since Dec. 31, 2013 vs. our primary passive benchmark: Global Market Index (GMI), an unmanaged, market-value mix of all the major asset classes:
As the numbers above reveal, our three in-house strategies are weathering this year’s market volatility rather well. Why? Let’s dig into the details, starting with a quick housekeeping note: Global Managed Risk I and II are the same strategies with one difference: the “safe” asset for I is long Treasuries (iShares 20+ Year Treasury Bond (TLT)) vs. short Treasuries (iShares Short Treasury Bond ETF (SHV)) for II.
The dramatically higher performance for Global Managed Risk I vs. its counterpart is a function of the spectacular returns in long Treasuries so far in 2020 (and in recent years) compared with the cash-proxy results for SHV. The caveat, of course, is that using long-dated Treasuries as a “safe” asset for a dynamically managed portfolio is vulnerable if interest rates rise. So far, yields have been on a mostly downward slide, a trend that’s been magnified in TLT, which in turn has been great for returns (which move inversely to yields). Deciding if this payoff will continue (or not) is a critical question for deciding whether to favor Global Managed Risk I or II.
Meantime, here’s a quick review of the underlying risk-management methodology that drives Global Managed Risk I and II. The basic strategy: use drawdown for monitoring downside risk potential. When the risk is relatively high, shift to the “safe” asset (TLT or SHV); otherwise, hold the risk asset.
The portfolios for I and II are built with the following foundation for tapping into the major asset classes on a global basis:
These weights are used to rebalance I and II every Dec. 31. (Note that the allocation and funds outlined above are also used for one of our benchmark portfolios—Global Beta 15, albeit without a risk-management overlay other than resetting weights every Dec. 31).
The risk management for I and II is based on the following rule: if current drawdown falls below the 50th percentile (based on a rolling 50-day window), a risk-off signal is issued. A second filter applies before tactical trading takes place: If the risk-off signal persists for each of the last two trading days in a given week, the asset is sold at the open on the first trading day of the next week and the proceeds are moved to the “safe” asset.
The rule is applied to each risk asset independently. For example, the bucket for large-cap stocks transitions between between VTI and the “safe” asset (TLT or SHV), depending on the drawdown signal. Every Dec. 31, each bucket’s weight (regardless of whether it’s in the risk or safe asset) is rebalanced to the target allocation per the table above.
The goal at the portfolio is to nip steep drawdowns in the bud before they can unleash substantial damage. Not surprisingly, there is a fair amount of turnover with the strategy due to noise and so the number of false signals is non-trivial. But if a risk-off signal turns out to be wrong (i.e., the price decline stabilizes/reverses), the methodology will soon reverse and shift back to the risk asset, perhaps as quickly as the following week.
Overall, results are encouraging at the portfolio level. Note that I and II’s maximum drawdowns over the past five years have relatively mild vs. the benchmarks. Meantime, performance has been solid, resulting in relatively high Sharpe and Sortino ratios.
The other proprietary strategy is applying a minimum variance methodology to the portfolio based on the global assets in the table above (G.B15.MINV). Although this strategy is the weakest performer in terms of raw return, it’s also the smoothest portfolio, by design. Annualized volatility over the past five years is a mild 4.1% -- well below the other proprietary portfolios and the benchmarks. Meantime, the price tag for damping volatility hasn’t been overly harsh, earning an annualized 4.7% total return for the past five years. That’s modest compared with the benchmarks and the other two proprietary strategies, but for investors demanding a smoother ride, it may a price worth paying.
Finally, here’s a recap of the benchmark designs:
US.60.40:US stock/bond portfolio, rebalanced every Dec. 31
GMI:an unmanaged global portfolio that holds the major asset classes in market-value weights
G.B4:A twist on GMI that reduces holdings to four broadly defined ETFs that target global exposure to stocks, bonds, real estate and commodities. Weights: 60% stocks, 30% bonds, 5% real estate and 5% commodities. The portfolio is rebalanced every Dec. 31.
G.B15:An expanded version of G.B4 via 15 ETFs that slice and dice the world’s major asset classes into a more granular portfolio design. The overall allocation still matches G.B4: 60% stocks, 30% bonds, 5% real estate shares and 5% commodities.