• Single factors have historically generated excess returns above a capitalization-weighted benchmark over the long term.
• Given the challenges of sticking with a single factor, investors might be tempted to rotate into favourable factors in search of the best possible returns. However, while seemingly intuitive, factor timing can be challenging and expensive.
• A multifactor blended approach can combine the benefits of the two without having to fully face the challenges inherent in either approach.

What follows is an excerpt of the full white paper. Download the full piece here.

Executive Summary

There are pros and cons to the various approaches to factor investing. We believe that investor interests are best served through a multifactor investing process. Factors such as Value, Size, Momentum, Quality and Minimum Volatility are employed to identify stocks with specific characteristics. Such stocks can then be used as part of an investment strategy to generate potential returns above a specific benchmark. For instance, stocks that exhibit lower volatility are thought to provide higher risk-adjusted returns to investors due to behavioural effects. Similarly, Value stocks are understood to be cheaper relative to their intrinsic value, and eventually provide higher returns when this intrinsic value is realised. As far as the factor investing process is concerned, single-factor exposure and factor-timing processes are among the most popular approaches.

We analyse the pros and cons of both of these approaches and suggest that a multifactor bottom-up blending process combines the benefits of both and offers the best risk-adjusted returns.

Single-Factor Approach — Pros and Cons

As mentioned before, factors tend to outperform capitalization-weighted benchmarks over long periods of time. For this reason, they are considered as key drivers of risk and return in equity portfolios.

We examined the performance of the five classic premium factors, namely, Minimum Volatility, Momentum, Quality, Size and Value as well as Growth, which is generally not considered as a premium factor, over several business cycles. Growth was added to the group in the context of the factor coming into sharp focus recently given its strong performance since the Global Financial Crisis.

The factors considered in our sample all individually outperformed the capitalization-weighted benchmark over a period of 25 years either in terms of return or risk or both (Figure 1).

To be sure, a two-decade long holding period is likely too long a period for many investors to wait for a single factor to outperform. Taking this into account, we calculated the historical success rate for each individual factor over a range of holding periods to estimate the necessary holding period required for a single factor to outperform the market (Figure 2).

What we see here is that historical success rates have varied a great deal by factor and that it may take up to 10 years or longer for factors to reliably realize their premia. The crucial element to consider here is the fact that this path to outperformance is not a straight line but is dotted with continuous quarters of underperformance that investors would have to brave.

Factor-Timing Approach — Pros and Cons

In theory, factor timing is the answer to the challenges that are inherent in single-factor investing. A perfect factor-timing strategy with quarterly rebalancing between the major MSCI factor indices would have compounded $100 invested in end-December 1998 to nearly $40,000 by end-June 2024. The annualized return of this strategy would be 26.4% versus just 6.4% for the capitalizationweighted benchmark. As mentioned before, this promise is not without its perils: the worst factor-timing strategy reduced $100 to $12, a painful -7.5% annualized return over the same period (Figure 3).

Where do investors need to place themselves on the spectrum of best-to-worst factor timers to make this timing strategy worthwhile? Our calculations show that a successful factor-timing strategy needs a prediction accuracy between 50% to 60% to outperform an equal-weighted factor allocation.¹ While this accuracy may appear only marginally better than the result of a random coin flip, such low prediction accuracy would require a high tracking error to generate meaningful alpha as suggested by the Fundamental Law of Active Management (FLAM) developed by Richard Grinold and Ronald Kahn.²

As stated in the equation below, FLAM asserts that the Information Ratio (IR), which is the ratio of benchmark-relative excess performance to tracking error, is approximately proportional to the product of the Information Coefficient (IC), which is a measure of the manager’s forecasting skill or prediction accuracy, and the square root of the Breadth (BR), which is used to measure the number of independent forecasts.

Equation: IR ≈IC x √BR

Given the high success rates of individual factors, using factors in a timing strategy has the advantage of increasing IC, but this comes at the expense of reducing Breadth. Remember that the initial motivation for factor analysis was Breadth reduction to address the dimensionality problem. Given the constraint on Breadth, a high IR, or risk-adjusted performance, is hard to achieve without near-perfect predictive power or high relative risk.

As we have shown, high prediction accuracy may not be easily possible given the time-varying relationship between factor premia and external indicators. Additionally, the distribution of factor performance is far from uniform, which means that factor outperformance can be squeezed into just a few periods. The implication is that missing any of these key periods could adversely affect the factor-timing strategy (Figure 4).

Missing the best 8 out of the 102 quarters of factor outperformance in our sample negated the alpha from the factor rotation process across all factors except for Quality, which had two more quarters left to end up below water. In principle, a factor-timing strategy aims to benefit from exceptional periods, but in this process, the strategy exposes investors to greater risk. Add turnover costs to this greater risk exposure, and the risk-return trade-off may appear unappetizing to investors.

United We Stand: Multifactor Approach

It is clear from the above illustrations that although factors in general offer meaningful alphageneration capabilities, single-factor as well as factor-timing strategies are fraught with risks that may dissuade even a seasoned investor from adopting such strategies. On the one hand, a simple and relatively inexpensive approach of sticking with a single factor may turn out to be ineffective owing to the unrealistic time periods involved in achieving outperformance. On the other, a seemingly rewarding factor-timing strategy could prove challenging due to the high accuracy requirements and increased turnover.

Is there a way out of this predicament where investors could reap the benefits of factor investing without having to choose between the options of single-factor or factor-timing strategies? To put it differently, as far as investors are concerned, the question should not be whether to employ factors as such, but how to employ them successfully by avoiding the risks inherent in both single-factor and factor-timing strategies.

One solution to this problem, in our view, is to diversify across a variety of factors taking into consideration their less-than-perfect correlations. Take the Momentum factor for instance, which has historically exhibited a moderately positive correlation with Size and Quality but a negative correlation with Minimum Volatility and Value. By blending these factors, which are less than perfectly correlated, it should be possible to improve the risk-return profile from a single-factor strategy, which could in turn cushion the impact of drawdowns.