Trivariate’s Proprietary Crowding Score
In this note, we analyzed several metrics and concluded through efficacy and parsimony that six signals – both level and change – across liquidity, volatility, and conviction by way of 13-F filings are good signals for identifying crowded stocks. We combined these six signals in multiple ways – equal-weighted, weighting liquidity more, weighting conviction more – into a set of proprietary crowding scores – in which we generally looked at beta-adjusted spread portfolios of the signal’s top quintile (“least crowded”) and bottom quintile (“most crowded”).