Using Asset Clustering and Relative Valuation to Time Crowded Trades

Summary

The article describes a framework for identifying crowded trading and distinguishing a bubble’s buildup from its unwinding. It estimates asset clustering from covariance and principal components, using the concentration of sector returns in important components as a proxy for crowding. It pairs this measure with relative valuation based on price-to-book ratios: high clustering with lower relative valuation is treated as a possible buildup phase, while high clustering and elevated relative valuation may signal crowded selling. The method is applied to sector rotation and timing equity factors.

The article reports historical examples, portfolio sorts, and backtests across US sectors, factor portfolios, and other equity markets. It says the combined measures performed better than either measure alone in the reported tests. These findings are historical and do not establish that bubbles can be reliably forecast. Clustering is an indirect proxy for trading flows, fundamentals can also drive large price moves, and the reported results depend on the chosen definitions, data, and backtest assumptions.

Key ideas

• Asset clustering is estimated from return covariance and principal-component concentration as a proxy for crowded trading.
• Relative valuation is combined with clustering to distinguish possible bubble buildup from unwinding.
• High clustering with lower relative valuation is associated with stronger subsequent portfolio performance in the reported tests.
• The article applies the framework to sector rotation and equity-factor timing.
• The evidence is historical and does not show that the indicators reliably predict every bubble or its turning point.

Tags

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