The costs of factor strategies are not limited to the explicit implementation expenses of trading, such as commissions, taxes, and ticker charges. They also include the market impact of trading.

The authors seek to fully uncover the costs and capacities of factor investing by analyzing investible indexes to ascertain the real market impact costs and allow investors to make more-informed choices.

The authors analyze the trading costs of the most popular factor-investing strategies, including value, income, low beta (or low volatility), quality, momentum, and multifactor.

They examine the factor strategies of managers who track factor indexes, aiming to measure the immediate market impact of these managers’ trades. To do so, they source daily returns from Bloomberg for 2009–2016, creating a dataset of 49,867 trades worth US$56.6 billion.

The authors evaluate the costs of the strategies examined—all of which have varying turnover rates, trade sizes, levels of security liquidity, and numbers of rebalances. The resulting attribution framework maps costs onto the unique characteristics of each factor strategy. This approach enables a granular exploration of which characteristics of each investment factor have the most impact on costs.

Next, the authors define the natural capacity of each factor strategy that occurs, according to the study methodology, when costs hit 50 bps per year—a level seen as unsustainably high for an index fund seeking to attract new investors. 

According to the authors, the size of the price impact is directly related to a factor portfolio’s liquidity and the volume being traded. In addition, strategies that rely on frequent trading and those that buy or sell their entire holding of an illiquid security in a single trade generate higher costs.

Strategies with low portfolio volume, high turnover, high concentration of turnover, and low correlation with volume-weighted benchmarks have the highest trading costs.

Two momentum strategies stand out as the costliest. Momentum funds with an assumed $10 billion in assets under management that invest based on risk-adjusted Sharpe momentum and standard momentum strategies incur annual market impact costs of some 200 bps and 270 bps, respectively. The capacities of these two momentum strategies are limited to $2.4 billion and $1.8 billion, respectively.

Dividend strategies also incur noticeably high costs. High dividend and dividend growth strategies have annual costs of 61 bps and 76 bps, respectively. The main reason for these high costs is the strategies’ large tilt away from volume-weighted benchmarks because they allocate the portfolio to a small number of high-yielding stocks. The capacity of high-dividend strategies and dividend growth strategies is estimated at $8.2 billion and $6.6 billion, respectively.

Fundamental indexation is the least expensive strategy to implement because its turnover rate is low. Fundamental value and fundamental low-volatility indexes have the lowest market impact cost at just 2 bps and 5 bps, respectively, and the highest capacity at $291 billion and $108 billion, respectively.

Low-beta strategies have a notably wide range of costs, from some 190 bps a year for a low-volatility strategy to just 5 bps for a fundamental low-volatility strategy and 7 bps for a defensive strategy. The authors explain the huge disparity in costs among funds with the same primary objective as a matter of index construction.

Multifactor strategies—that is, combinations of single-factor portfolios—have only moderate costs despite the extra portfolio construction complexity. These combinations seem to lower costs because each factor portfolio can make use of a different pool of liquidity. Multifactor capacities range from $21 billion to $41 billion.

This cost attribution framework should help investors better assess how factor strategies are affected by trading costs and the importance of index design.

The proposed framework has the advantage of being relatively simple for investors to apply to various factor strategies, including those based on non-US assets. The authors acknowledge that in practice, costs can be reduced: Experienced managers will execute large trades over a number of days and avoid trading in thin markets. In addition, some managers use crossing networks or dark pools, trades for which obtaining information is difficult and that are not represented in the framework.

The authors’ attribution of costs to strategy characteristics can enable practitioners to more thoughtfully design factor products to reduce market impact costs.

As the authors explain, “Knowing what drives cost is perhaps more important than the cost estimations themselves.”