New research from S&P Dow Jones Indices (S&P DJI) considers the benefits of multi-factor investing and proposes that, due to the cyclical nature of individual factor returns, multi-factor investing is an efficient means of reducing individual factor risk and thereby smoothing portfolio returns.

The report, titled “The Merits of Multi-Factor Investing,” argues that using multifactor strategies can help diversify individual factor volatility.

While smart beta factors are widely acknowledged to deliver excess returns over long time horizons in many different markets, significant deviations in factor performance over shorter time periods have been observed, making the timing of factor investments important.

The variability of factor returns versus a benchmark can be seen in the graph below, which shows the performance of single factors compared to the S&P 500 since 1995.

Source: S&P Dow Jones Indices.

Whether it is possible to effectively time factors is the subject of much debate, with smart beta grandees Cliff Asness of AQR Capital and Rob Arnott of Research Affiliates having a public disagreement on the matter over the last 18 months.

Arnott argues that factor timing is possible to a degree by looking at the current valuation of factors relative to historical averages and that a contrarian strategy of overweighting “cheap” factors and underweighting “expensive” factors can improve the performance of a factor portfolio. Supporting data from Research Affiliates shows that the performance of a factor over the previous five years is negatively correlated with performance over the next five years.

Asness, however, states that factor timing is very difficult, if not impossible, and that diversifying among uncorrelated factors is a much more efficient way to smooth the cyclicality seen in individual factors.

A new research paper from S&P Dow Jones Indices, “The Merits and Methodology of Multi-Factor Investing”, supports the Asness view that factor diversification is the best way to invest for market participants wishing to avoid the risk of choosing between single factor strategies.

The research shows that single factor strategies (value, momentum, low volatility and quality) based on the S&P 500 often outperform the underlying index over most time horizons during the period studied; however, the frequency of risk-adjusted outperformance was notably lower for shorter holding periods.

The research then showed that an equally weighted portfolio of factors performed as well or better than the best performing single factors over all time horizons. The diversification benefit of holding equal exposure between the four single factor indices contributed to outperformance of the portfolio 80% of the time compared to the S&P 500 over a one-year period and 97% of the time over a three-year period.

The paper goes on to examine whether a top-down or bottom-up approach is optimal for multi-factor investing. A top-down approach involves investing in a portfolio formed of single factor strategies like that described above. A bottom-up approach involves combing individual factor scores for each stock to create a multi-factor score, which is then used to select a more concentrated portfolio of “all-rounders” characterised by exposures that are fairly evenly distributed across all the desired return drivers.

For market participants without a factor viewpoint, both multi-factor approaches may offer a compelling solution, but each approach has its own pros and cons.

Due to the negative correlation of factor scores among stocks, a fund of funds top-down approach can suffer from factor exposure dilution, while a bottom-up approach may capture higher factor exposures.

The effect of higher factor exposures in bottom-up multi-factor strategies can be seen in backtests of the S&P Quality, Value & Momentum Multi-Factor Index, which uses a bottom-up approach and would have outperformed a top-down fund of funds approach as well as each of the individual factors over a period of more than 20 years. The index also has the highest average return during each five and 10-year period in the sample horizon, outperforming the top-down approach by an average of 1.6% and 2.4% per annum over each five and 10-year period respectively.

Source: S&P Dow Jones Indices.

A top-down approach is likely to offer a more sector-neutral breakdown compared to the benchmark than a bottom-up strategy. This finding aligns with the low tracking error of top-down portfolios owing to the relatively high number of constituents.

Ultimately, the paper argues that the choice comes down to whether an investor wishes to maximise risk-adjusted returns on an absolute basis or relative to the benchmark.