Paper: arXiv 2602.06198

Authors: Hangyi Zhao

Abstract

This paper examines whether SEC Form 4 insider purchase filings predict abnormal returns in U.S. microcap stocks. The analysis covers 17,237 open-market purchases across 1,343 issuers from 2018 through 2024, restricted to market capitalizations between $30M and $500M. A gradient boosting classifier trained on insider identity, transaction history, and market conditions at disclosure achieves AUC of 0.70 on out-of-sample 2024 data. At an optimized threshold of 0.20, precision is 0.38 and recall is 0.69. The distance from the 52-week high dominates feature importance, accounting for 36% of predictive signal. A momentum pattern emerges in the data: transactions disclosed after price appreciation exceeding 10% yield the highest mean cumulative abnormal return (6.3%) and the highest probability of outperformance (36.7%). This contrasts with the simple mean-reversion intuition often applied to post-run-up entries. The result is robust to winsorization and holds across subsamples. These patterns are consistent with slower information incorporation in illiquid markets, where trend confirmation may filter for higher-conviction insider signals.

Complexity vs Empirical Score

  • Math Complexity: 2.5/10
  • Empirical Rigor: 7.5/10
  • Quadrant: Street Traders — practical and empirical, lighter on theory

Why this score: The paper relies on straightforward ML models (XGBoost) and basic statistical metrics rather than complex mathematical derivations, placing it in the Street Traders quadrant. It provides strong empirical analysis with rigorous out-of-sample backtesting, proper time-series splitting, and useful practical insights regarding momentum signals in microcap Form 4 filings.

Research Flowchart

  flowchart TD
    A[Research Goal: Do SEC Form 4 Insider Purchases<br/>Predict Microcap Abnormal Returns?] --> B[(Data Inputs: 17,237 SEC Form 4 Purchases<br/>1,343 Issuers, $30M-$500M Cap, 2018-2024)]
    
    B --> C[Methodology: Feature Engineering<br/>Insider Identity, History, Market Conditions]
    
    C --> D[Model Training & Evaluation:<br/>Gradient Boosting Classifier trained on 2018-2023,<br/>Tested Out-of-Sample on 2024 Data]
    
    D --> E[Key Findings & Model Outcomes:<br/>• AUC: 0.70 | Precision: 0.38 | Recall: 0.69<br/>• Top Feature: 52-Week High Distance 36% Importance]
    
    D --> F[Key Market Discovery:<br/>Momentum Pattern: Purchases after >10% Price Gains<br/>Yield Highest Mean CAR 6.3%]