Replication Package for Star Firms and Industry Outcomes

Pseudonymized data and code replicate analyses from a finance paper on star firms and industry-level predictability. The package includes firm-level and industry-level datasets built from sources like CRSP, Compustat, and I/B/E/S. Vidhi Chhaochharia published this replication package via JFQA Dataverse in July 2026.

Use Cases

• Estimating earnings growth and surprise regressions based on firm-level data.
• Conducting macro-outcome regressions for GDP and employment using industry-level data.
• Performing dynamic panel GMM analyses to study information spillovers.
• Constructing portfolios for asset-pricing tests using Fama-French factors.
• Linking firm characteristics like patent data, markups, and vertical integration scores to industry outcomes.

Strengths

• Includes pseudonymized data capped at 1,000 observations for reproducibility.
• Provides SAS and Stata programs for constructing datasets and reproducing all empirical results.
• Integrates data from multiple established sources including CRSP, Compustat, I/B/E/S, BEA, BLS, and patent datasets.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Data is pseudonymized and capped, which may not reflect the full original dataset used in the paper.

Provenance

Source

JFQA Dataverse

Collection Method

Constructed from multiple financial and economic data sources (CRSP, Compustat, I/B/E/S, LinkUp, BEA, BLS, patent data, markup data).

Freshness

Last updated 2026-07-03 00:16:16; freshness should be verified.