Factors with Style

Abstract

We document significant spreads in style factors—value, size, quality, momentum, and low volatility—in each of the style box categories. This is also true even for the value and small size factors, which are reflected in the original definition of the style box framework. Some single factors stay within a given style box, like quality in Blend, while other factors drift across style boxes, like momentum and even the size factor! We build multifactor portfolios within each style box, giving access to five style factors that can stay within a style box category which have exhibited information ratios of about 0.8 over June 2003 to December 2020.

TOPICS: Analysis of individual factors/risk premia, portfolio construction, performance measurement, risk management

Key Findings

• ▪ Although the traditional style box provides categories based on size and value, there is still substantial factor dispersion within each of the style boxes. This may not be so surprising for the quality, momentum, and low volatility factors that are not directly captured in the style box framework, but it is also true for the size and value factors themselves.

• ▪ Active style box managers within the large cap US equity space have demonstrated dispersion in the value and size factors in line with their benchmarks, but the median active manager tends not to exhibit significant quality, momentum, and low volatility factor exposures. The range for these three factors is also relatively narrow, compared with size and value.

• ▪ Optimal combinations of the five factors—momentum, value, quality, size, and low volatility—can be constructed within each style box, allowing investors to harvest all five, and not just two, factor premiums while remaining in the same style box category. In historical data, these five-factor combinations resulted in information ratios above 0.8, which fall in the top decile of above-benchmark performance relative to active managers in the respective peer group, while exhibiting lower relative active risk.