Please rotate your device to landscape mode to view the charts.

Thousands of Alpha Tests

Journal: Review of Financial Studies

Date: 20210701

Author: Giglio, Stefano; Liao, Yuan; Xiu, Dacheng

Abstract:
Data snooping is a major concern in empirical asset pricing. We develop a new framework to rigorously perform multiple hypothesis testing in linear asset pricing models, while limiting the occurrence of false positive results typically associated with data snooping. By exploiting a variety of machine learning techniques, our multiple-testing procedure is robust to omitted factors and missing data. We also prove its asymptotic validity when the number of tests is large relative to the sample size, as in many finance applications. To improve the finite sample performance, we also provide a wild-bootstrap procedure for inference and prove its validity in this setting. Finally, we illustrate the empirical relevance in the context of hedge fund performance evaluation.

Link: Google Scholar

Key Findings

Enhanced Hedge Fund Selection

Novel FDR control methodology improves hedge fund selection by identifying skilled managers while controlling false discoveries. Out-of-sample performance shows 22bp monthly alpha compared to 15bp using standard approaches.

Robust to Missing Data

Matrix completion technique effectively handles missing data patterns in hedge fund returns (>70% missing), enabling more accurate factor estimation and alpha testing.

Superior Factor Model Integration

Combined observable and latent factor approach outperforms traditional methods, selecting larger fund portfolios while maintaining higher alphas.

Out-of-Sample Hedge Fund Performance

  • Mixed FDR approach achieves highest monthly alpha of 22.3bp
  • Traditional observable factor approach yields 15.1bp alpha
  • Simple p-value screening produces only 10.7bp alpha

Fund Selection Coverage

  • Mixed FDR approach selects 25% of available funds
  • Observable factor FDR selects only 17% of funds
  • Selected funds maintain higher AUM on average

Model Robustness Across Specifications

  • Performance remains stable across different rebalancing frequencies
  • Results robust to different benchmark models
  • Consistent outperformance with both annual and monthly rebalancing

Contribution and Implications

  • First rigorous framework for multiple testing in asset pricing that handles omitted factors and missing data
  • Practical methodology for identifying skilled fund managers while controlling false discoveries
  • Demonstrates importance of incorporating both observable and latent factors in fund selection

Data Sources

  • Performance comparison charts based on Table 3 out-of-sample results
  • Fund selection statistics derived from Table 4 Panel A
  • Robustness analysis uses data from multiple specifications in Table 3 rows