Crowding and Relative Value Timing

Article BigQuant

Summary

The article describes a framework for identifying crowded sectors and estimating whether crowding may be building or unwinding. It measures asset clustering from principal components of sector return covariances, using the concentration of explained variance and sector exposures as a proxy for crowding. A relative-valuation measure compares each sector’s price-to-book ratio with its own long-run average and with other sectors. High clustering paired with relatively modest valuation is treated as a possible accumulation phase; high clustering paired with rich valuation is treated as a possible unwind phase.

Historical examples and portfolio tests are reported for sector rotation across the U.S. and other markets, as well as timing size, value, quality, and low-volatility factors. The article says the combined signals performed better than using either measure alone, and gives backtest results for sector and factor portfolios. These are historical findings, not proof that bubbles can be reliably identified in real time. Clustering is only an indirect crowding proxy, relative valuation can reflect fundamentals, and the framework depends on specific portfolio assumptions and sample choices. The article explicitly notes that bubble prediction remains unsettled.

Key ideas

The framework uses return covariance and principal components to estimate sector clustering as a proxy for crowded positioning.
A relative price-to-book measure is combined with clustering to distinguish potential accumulation from unwind phases.
Neither clustering nor relative valuation alone is presented as a sufficient bubble signal.
The framework is applied to sector rotation and timing common equity factors.
Reported results are historical backtests and do not resolve the broader uncertainty around predicting bubbles.

Using Asset Crowding and Relative Valuation for Sector and Factor Timing

Summary

Key ideas

Tags