Using Crowding and Relative Valuation to Time Sector and Factor Rotations

Summary

This review describes a method for identifying crowded markets and distinguishing potential bubble build-up from unwinding. It measures sector or factor clustering from return covariances and principal components, then pairs that measure with relative price-to-book valuation. High clustering with relatively low valuation is treated as possible accumulation; high clustering with elevated valuation is treated as possible distribution. The article applies the framework to historical bubbles, sector rotation, and factor timing.

The reported backtests show stronger relative results for highly clustered but less expensive sectors and factors than for highly clustered, expensive ones. The review also describes portfolio construction using assumed expected returns for the two states and a mean-variance optimizer, with related tests in other equity markets. These findings are historical and do not establish that bubbles can be reliably forecast. Clustering is only a price-based proxy for crowded positioning, valuation alone does not identify its cause, and the summary notes that the method does not precisely date bubble peaks or breaks.

Key ideas

• The method estimates asset clustering from covariance structure and principal-component exposures.
• Relative price-to-book valuation is combined with clustering to distinguish possible bubble accumulation from unwinding.
• High clustering with lower relative valuation is treated as a more favorable rotation signal than high clustering with elevated valuation.
• The article reports historical backtests across sectors, factors, and several equity markets, but these do not guarantee predictive performance.
• Clustering is a proxy for crowded trading and cannot by itself establish that price moves lack fundamental causes.

Tags

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.