Factor-based portfolio construction: combining factor exposures in a single systematic portfolio
Selecting individual securities is a noisy process. Thousands of variables influence stock prices, most of them unpredictable. Factor-based portfolio construction takes a different approach: instead of selecting securities directly, it builds portfolios with deliberate, systematic exposure to documented sources of return—factors that have been shown, across geographies and asset classes, to earn a premium over time.
What factor-based portfolio construction is
The empirical foundations of factor-based investing were laid by Fama and French (1992) in The Cross-Section of Expected Stock Returns, Journal of Finance, which showed that size and value factors explained returns that market beta alone could not account for. Carhart (1997) in On Persistence in Mutual Fund Performance, Journal of Finance, added momentum as a fourth systematic factor, establishing the basis for the multi-factor frameworks that underpin modern factor-based portfolio construction.
Factor-based portfolio construction is the practice of building portfolios by targeting specific risk factor exposures—value, momentum, quality, low volatility, size—rather than choosing securities on the basis of individual analysis. The distinction matters: security selection introduces idiosyncratic noise; factor exposure is systematic, rules-based, and verifiable. Asness, Moskowitz & Pedersen (2013) demonstrated that value and momentum factors explain a significant share of cross-sectional return variation across eight asset classes and four geographies, providing a robust empirical foundation for factor-based approaches.
How it works
There are two principal implementation approaches, each with distinct trade-offs.
The factor sleeve approach allocates a defined capital weight to each factor independently. A portfolio might hold 25% in a value sleeve, 25% in a momentum sleeve, 25% in a quality sleeve, and 25% in a low-volatility sleeve. Each sleeve is managed according to its own signal, and the combination is achieved at the allocation level. This approach preserves the purity of each factor signal and makes it straightforward to monitor individual factor performance—but it may result in offsetting positions across sleeves if one factor is long an asset that another is short.
The composite scoring approach ranks each asset by a combined factor score—a weighted average of its scores on each individual factor. A single portfolio is constructed from assets ranked highest on the composite. This approach avoids the cross-sleeve offsetting problem and typically results in lower turnover, but it dilutes each individual factor signal. Assets with moderate scores on every factor can crowd out assets with exceptional scores on one.
What the evidence shows
Asness, Moskowitz & Pedersen (2013) made two particularly important empirical contributions. First, they demonstrated that value and momentum premia are robust across asset classes—equities, fixed income, currencies, and commodities—and across geographies, including the United States, the United Kingdom, Europe, and Japan. Second, they showed that value and momentum returns are negatively correlated with each other: when value underperforms, momentum tends to outperform, and vice versa. This negative correlation means that combining the two factors into a single portfolio improves the Sharpe ratio relative to either factor held alone—a diversification benefit that does not require assumptions about which factor will perform better in a given period.
Limitations
Factor timing—adjusting allocations between factors based on expected relative performance—is difficult and rarely adds value. The same uncertainty that makes security selection unreliable applies to factor selection over short horizons. Composite scoring can dilute individual signals to the point where the resulting portfolio has limited meaningful factor exposure. Factor definitions vary across providers: two products claiming exposure to the same factor may construct it differently, producing substantially different portfolios. Finally, widely-followed factor strategies are susceptible to crowding—as assets concentrate in popular factor products, the premia they target can compress.
Factor-based portfolio construction in pfolio
pfolio uses momentum as the primary signal for asset selection across a multi-asset universe. The approach is rules-based and systematic, consistent with factor-based principles. You can read about the methodology in detail at how we build portfolios and explore the available strategies at pfolio portfolios.
Related articles
Disclaimer
Get started now
