Commentary: Factor investing – is keeping it simple shortsighted?

​The rapid proliferation of factor investing strategies raises important questions for investors. Are different factor implementations interchangeable, implying that fees alone should be the primary consideration in selection? Or can a more sophisticated approach materially improve performance?

Simple factor investing approaches, known as smart beta or alternative risk premium strategies, claim to capture recognizable returns premiums such as value, momentum and quality, while at the same time being cheap, transparent and virtually unlimited in capacity. But these approaches come with a cost. These implementations, which treat factors as simple commodities, forgo the potential benefits of sophistication at every stage of the investment process.

Momentum serves as a salient case study. Basic momentum formulations make use of each stock's past total returns, usually the period from 12 months to 1 month prior to investment. Stocks with higher past returns, by this measure, are deemed attractive to own, while stocks that have performed comparatively poorly are excluded from long-only portfolios or, in long-short implementations, sold short.

But strategies based on this rudimentary formulation might be vulnerable to unintended risks. They tend to display significant and pro-cyclical variation in their market exposure over time, as portfolios load up on high-beta names when the market is rallying and on low-beta names during declines. As well, they may unwittingly take on material sector exposures and participate in crowded trades. Simple momentum formulations also may respond to different aspects of investor behavior as market conditions change. They may reflect market-level sentiment when correlations among stocks are high and company-specific views when correlations are low.

Empirical evidence suggests that more sophisticated formulations of momentum may produce superior risk-adjusted returns than simplistic measures. One example, residual momentum, is a regression-based formulation that seeks to reduce time-varying risk factor exposures in the signal and better isolate stock-specific sentiment. Another example, a factor that our firm calls peer momentum, attempts to extract more nuanced sentiment information about a stock from price trends affecting groups of related companies.

Our work has shown that these formulations generate more stable performance and higher Sharpe ratios over time, experiencing smaller drawdowns and less vulnerability to crowded trades. In 2016, for example, a rudimentary long-short momentum formulation likely would have experienced material losses, whipsawed by short positions in high-beta industrials and oil-related stocks as stock and crude oil prices recovered from prior losses. Our analysis suggests that residual and peer formulations would have fared better during this episode.

Beyond factor selection and construction, we believe that benefits of sophistication extend across the entire systematic investing process. For example, most simple factor strategies don't generate an actual forecast of stocks' expected returns based on their factor exposures. This simplification forces reductive portfolio construction, because it becomes impossible to maximize expected return relative to risk. This helps to explain the confusing array of heuristic portfolio construction approaches in the smart beta/alternative risk premiums sphere.

The absence of an explicit return forecast also makes it impossible to incorporate a sophisticated transaction cost model into portfolio construction, because there are no expected returns to weigh costs against. This makes it difficult to exploit less liquid stocks, which we think tend to offer relatively large mispricings protected by limits to arbitrage. This reduction in universe breadth also inhibits risk management, restricting the flexibility to control undesired exposures (e.g., valuation risk in low volatility strategies during 2016 and 2017).

So why don't smart beta approaches incorporate greater sophistication? It's not, in our view, because their rudimentary approaches can't be improved upon. Rather, we believe the explanation lies in the requirements of smart beta's underlying mass distribution, ultra-low-fee business model, which requires simple transparent signals, low development and maintenance costs, low turnover, enormous capacity and easy execution. Such attributes preclude factor specifications — such as residual and peer momentum — that bring to bear more sophisticated analytical methods, require ongoing investment in factor research, and demand greater execution skill due to higher turnover and limited capacity.

We believe that rudimentary formulations are unlikely to deliver full value, and that nuanced factor constructions has the potential to materially increase expected returns and reduce unintended risks. Approaches that treat factors as commodities project a misleading sense of simplicity. They are best understood as products of a host of restrictions placed on the systematic investment process — active interventions that often have unintended and undesirable consequences and whose motivations may not be aligned with the interests of the investor. When it comes to factor investing, simpler is not necessarily better.

Seth Weingram is senior vice president and director at Acadian Asset Management, Boston. This content represents the views of the author. It was submitted and edited under P&I guidelines but is not a product of P&I's editorial team.