Factor Investing and the Quant Approach to Alpha

The academic literature on asset pricing has identified dozens of characteristics that predict cross-sectional returns. Factor investing is the systematic application of these characteristics to portfolio construction, with the goal of capturing risk premia that are persistent, economically motivated, and implementable at scale.

The Core Factors

The five factors with the most robust empirical support are market beta, value (price-to-book, price-to-earnings), size (market capitalisation), momentum (trailing 12-month return), and profitability (gross profit margin). Each represents a distinct source of return. Momentum captures the tendency of recent outperformers to continue outperforming over three to twelve months. Value captures the mean-reversion tendency of cheap assets relative to expensive ones over longer horizons. Profitability captures the premium earned by high-quality businesses over distressed ones. The size factor has weakened significantly since its original documentation by Fama and French, a pattern consistent with factor decay through crowding and implementation arbitrage.

Building a Multi-Factor Model

The standard approach is to rank the investment universe on each factor, normalise the scores, and combine them with fixed or dynamic weights. An equal-weighted combination of momentum, value, and quality has historically produced a Sharpe ratio of approximately 0.8 to 1.0 in US large-cap equities, compared to 0.4 to 0.6 for the index alone. The practical implementation challenge is turnover: momentum requires frequent rebalancing, which generates transaction costs that erode net returns significantly in smaller portfolio sizes. A monthly rebalance with a 12-1 momentum window (trailing 12 months excluding the most recent month, which exhibits short-term reversal) is the standard institutional implementation.

Factor Decay and Crowding

No factor works in all environments. Momentum typically underperforms during sharp reversals — it was the worst-performing factor during the March 2020 drawdown and the 2022 rate-driven rotation out of growth. Value had its worst decade on record from 2010 to 2020 before staging a significant recovery in 2021 to 2022. The more widely a factor is known and implemented, the more its excess return is competed away. Monitoring factor crowding via positioning data and dispersion metrics — the degree to which high-factor-score stocks are trading at premium valuations relative to low-factor-score stocks — is as important as the factor signals themselves. A crowded momentum trade in a rising-rate environment is a different risk profile than the same trade in a low-volatility, low-rate regime.