Abstract
Peer groups in cross-stock return tests are usually specified by the researcher rather than observed from investor behavior. We use active mutual fund holdings to define peer groups from buy-side portfolio choices. In U.S. equities from 1990 to 2024, holdings-based peer returns predict next-month focal-stock returns. Predictability is concentrated among nearest holdings-space peers, absent for the least similar stocks, and robust to industry, shared-analyst, common-ownership, and characteristics-based controls. Peer fundamentals and earnings news also forecast focal-firm fundamentals. A value-weighted long-short portfolio earns 0.67% per month with significant alphas. Post-formation returns continue rather than reverse, and the signal is not subsumed by common ownership. The predictive slope is larger among high-friction focal stocks. A simple model links these patterns to limited investor coverage and gradual price adjustment.
| Original language | English |
|---|---|
| Publisher | Social Science Research Network (SSRN) |
| Publication status | Published - 26 Jun 2026 |
Bibliographical note
Research Unit(s) information for this publication is provided by the author(s) concerned.Research Keywords
- Co-holding network
- Mutual fund holdings
- Active management
- Peer momentum
- Cross-firm predictability
- Information diffusion
Fingerprint
Dive into the research topics of 'Active Mutual Fund Co-Holdings and the Buy-Side Peer Network'. Together they form a unique fingerprint.Cite this
- APA
- Author
- BIBTEX
- Harvard
- Standard
- RIS
- Vancouver