A Holdings-Based Commitment Proxy and Reported Portfolio Rebalancing: Evidence from Form 13F Managers

Abstract

Form 13F filings are often invoked in behavioural arguments, yet they document delayed long positions rather than beliefs or executed trades. We examine whether reported additions to already material holdings, captured by a commitment proxy, are related to portfolio rebalancing during the following non-overlapping quarter. The analysis draws on official SEC records for 2022Q1 to 2025Q4: 2,332 manager-quarter observations from 261 large Form 13F reporting managers and 14 outcome quarters. In the primary model with manager and quarter effects, the coefficient is -0.0100 (standard error 0.0111; p value 0.364; 95 percent confidence interval [-0.0317, 0.0116]). Moving the proxy across its interquartile range is associated with -0.00190 in reported rebalancing, equal to -1.32 percent of the mean. The manager-effects result is also small (-0.0122, standard error 0.0112; p value 0.275). By contrast, the controlled pooled estimate is -0.2046 (standard error 0.0394; p value below 0.001; ordinary R-squared 0.1311), and the quarter-effects estimate is -0.1465 (standard error 0.0322; p value below 0.001). The specifications therefore separate a pronounced cross-manager association from a weak relation within managers. The non-overlapping construction makes clear that, in Form 13F data, the commitment–rebalancing pattern is chiefly cross-sectional.