February 25th, 2020 | by The Gerstein Fisher Team
Using Factors to Rebalance a Client’s Portfolio
We normally think about investment risk factors—such as value or momentum—at the individual security level. We decided to test whether the same factor logic works at the asset-class level. Specifically, we studied whether applying the momentum factor is a useful tool to manage quantitative, tactical asset allocation and its byproduct—portfolio rebalancing—in an effective way.
Momentum is the well-researched tendency for security winners to keep winning (for a period of time) and losers to keep losing. We measured momentum at the asset class level over 12 months, and used absolute and relative scores to determine what asset classes in particular to tactically over- and underweight in a 60/40 portfolio. In this scenario, large-cap growth stocks compete against large-cap value, for example; small caps against large caps; international fixed income vs. U.S. fixed income; stocks vs. bonds vs. alternatives. We then compared the long-term investment results of this systematic asset allocation (from January 1976 to June 2019) with those of two other typical methods of rebalancing.
In Strategy 1, we used a simple technique of rebalancing asset classes back to target allocation weights (e.g. 14.5% for large-cap growth, 12.5% for large value, 4% for global REITs, 31% for US bonds) each month. This method, which uses time rather than value as the trigger to rebalance, generated an annualized return of 9.3% for a 60/40 portfolio, 13.5% portfolio turnover, and an annual tax liability of 60 basis points for those investing in the presence of taxes.
For Strategy 2, we set bands, or ranges, around asset classes, a very common rebalancing technique in the investment industry. Specifically, we set 10% bands, which implies that if an asset class with a target weight of 20% appreciates to a 22% weighting in the portfolio (or falls 10% to an 18% position), then rebalancing is triggered. Compared with Strategy 1, this rebalancing strategy generated a higher annualized return (9.5% vs. 9.3%), but a higher tax cost (90 instead of 60 basis points) and much higher (29%) portfolio turnover. When viewed through the lens of returns over various rolling-year periods, Strategy 2 boasted quite a high win-rate, beating Strategy 1 pretax in 74% of the 5-year periods, 88% of the 7-year time frames, and 93% of the 10-year rolling periods.
For Strategy 3, we employed the momentum factor, allowing asset-class winners to drift higher with momentum and losers to drift lower than permitted by the target bands described in Strategy 2, and we permitted this allocation skew until the momentum signal reversed. For instance, by riding winners, the large-cap growth allocation (which had a benchmark allocation of 14.5%) had an average weighting of 15.7%, while in the second strategy it averaged 15.2%.
In our study, Strategy 3 generated the highest return (10.1% annualized), lower turnover (20%) than for Strategy 2, an 80 basis point tax cost, and the best risk-adjusted return (as measured by the Sharpe ratio*). In terms of batting average, Strategy 3 is a “slugger”, outperforming Strategy 1 in 97% of the 5-year rolling periods, and 100% of the 7- and 10-year periods.
*Sharpe Ratio is a measure used to help understand the return of an investment compared to its risk. Generally, the greater the value of the Sharpe ratio, the more attractive the risk-adjusted return.