Factor timing: can systematic rotation between factors improve portfolio outcomes?
Factor premia are not earned uniformly across time. Value dramatically underperformed between 2017 and 2020. Momentum crashed in 2009 as markets reversed sharply from their lows. Low volatility lagged in strong bull markets. If factor returns are at least partially predictable, systematically rotating toward factors expected to outperform—and away from those expected to underperform—could improve a portfolio's risk-adjusted return. This is the premise of factor timing. The practice is attractive in theory and difficult in practice.
What factor timing attempts to do
Static multi-factor investing holds all factors at roughly equal weight across all market conditions. Factor timing adjusts those weights dynamically, using signals believed to predict near-term factor performance. The signals typically fall into three categories: valuation spread (how cheap is the factor portfolio relative to its history?), momentum signal (has the factor been outperforming recently?), and macro regime (does the current economic environment historically favour or disfavour this factor?). Each approach has a theoretical foundation but mixed empirical support when applied out-of-sample.
Evidence on factor timing
The value spread—the dispersion between cheap and expensive stocks—has been proposed as a predictor of future value factor returns, on the intuition that a wider spread means more potential mean reversion. The empirical record is encouraging over long horizons but weak at shorter horizons where timing decisions are actually made. Momentum has been shown to exhibit time-series predictability in some markets, but the relationship is noisy. Macro regime approaches face a look-back bias problem: regimes are only clearly identifiable in hindsight, and real-time regime identification is substantially noisier. A review of the academic literature suggests that factor timing adds value in some studies and none in others, with the divergence often explained by transaction costs, look-ahead bias, or overfitting to historical data.
Rolling factor exposure
Even without deliberate timing, portfolios often exhibit factor rotation as a result of threshold-based rebalancing, changing factor loadings over time, or changes in the underlying universe. Monitoring rolling factor exposures—tracking how much momentum, value, quality, and low-volatility exposure a portfolio carries at any point—provides a cleaner picture of whether factor rotation is adding or subtracting value. This analysis complements deliberate timing and can reveal unintentional timing bets embedded in a supposedly static allocation.
Limitations
The most significant obstacle to factor timing is transaction costs. Rotating factor allocations requires buying and selling, and the factors that appear most attractive after transaction costs are often not the ones with the strongest raw signal. Factor timing also increases portfolio turnover and can create unintended factor interactions: rotating out of value into momentum in a rising market, for example, implicitly adds a growth tilt that was not intended. There is also a strong risk of overfitting: timing signals that worked in backtests may not persist out of sample. Most evidence-based investors prefer stable multi-factor exposure over active rotation, accepting short periods of factor underperformance as the price of a disciplined process.
Factor timing in pfolio
pfolio does not implement active factor timing in the core portfolio construction engine. The platform's factor exposure is stable and determined by the user's chosen strategy and asset class weights. Users interested in understanding their current factor exposures can view rolling factor attribution in Insights, which reveals factor tilts that emerge naturally from the portfolio's holdings rather than from deliberate rotation decisions.
Related articles
Disclaimer
Get started now
