Earlier this year, S&P DJI launched three S&P Quality, Value, and Momentum Top 90% Multi-factor Indices (the “S&P QVM Top 90%” Indices) across our large-cap, mid-cap, and small-cap universes. Each of these indices is tracked by an ETF.
Compared to S&P DJI’s other flagship multi-factor indices, this new series represents a differentiated approach to multi-factor index construction because it selects a high percentage of the universe and weights proportionally to float market cap. Based on internal back-tested research, this approach historically demonstrated moderate outperformance while retaining many of the core benchmark characteristics.
Methodology Overview
The S&P QVM Top 90% Indices are designed to track companies in the top 90% of their respective underlying index universe, ranked by their multi-factor score, which is based on the average of three separate factors: quality, value, and momentum.
Back-tested data shows that removing the lowest-ranked decile led to performance improvement. Exhibit 1 shows the cap-weighted return of each decile in the S&P 500®. Here, stocks have been ranked by their multi-factor score, placed into deciles (D1 = the lowest ranked, D10 = the highest ranked), and rebalanced quarterly.
For the S&P 500, the lowest-ranked decile exhibited the lowest performance over the period tested. Similarly, for the mid- and small-cap universes, the bottom decile has been the lowest, or close to lowest, performing decile.
Moreover, removing only the lowest decile also resulted in improved returns over the benchmark at relatively low tracking error. Exhibit 2 plots the ratio between excess returns over the S&P 500 and its resulting tracking error for a series of indices, each differentiated by the number of deciles removed. For example, T90% removes only the lowest-ranked decile (ranked by multi-factor score), T80% removes the two lowest-ranked deciles (i.e., the 20% lowest-ranked stocks), and so on.
Generally, as further deciles were removed from the back-tested strategy, the increase in tracking error did not result in proportional gains in excess returns. This is represented in Exhibit 2 by the slope of the line.
Headline Performance Statistics
The index construction methodology is such that the potential for substantial outperformance over the benchmark is limited, but so is the risk of significantly underperforming. Exhibit 3 shows the back-tested risk/return statistics for each of the S&P QVM Top 90% Indices. Since the start of the back-test period, the large-, mid-, and small-cap S&P QVM Top 90% Indices outperformed their benchmarks by 74, 107, and 107 bps per year, respectively, each with low tracking error.
Benchmark Characteristics
Removing only the lowest-ranked decile and weighting stocks proportionally to float market capitalization results in the S&P QVM Top 90% Indices having retained many of the qualities of the underlying benchmark. The back-tested analysis in Exhibit 4 shows that the active share, tracking error, and turnover historically remained low across the cap range.
Conclusion
The design of multi-factor strategies affects the expected performance characteristics and impacts positioning within a portfolio. For the S&P QVM Top 90% Indices, the construction methodology demonstrates moderate outperformance while retaining “benchmark-like” characteristics.
The posts on this blog are opinions, not advice. Please read our Disclaimers.