Public-firm analysis dataset

This unified public-firm panel combines company-year MQ with accounting, market, and patent outcomes for table8_firm_performance and table9_innovation.

flowchart TB
    mq[(Company-year MQ)]
    mapping[(Revelio-Compustat-CRSP mapping)]
    controls[(Accounting and returns controls)]
    patents[(Patent firm-year data)]
    build[Assemble linked public<br/>firm-year panel]
    output[(Firm-year analysis data)]

    mq --> build
    mapping --> build
    controls --> build
    patents --> build
    build --> output

    classDef source fill:#e8eef7,stroke:#52739e
    classDef process fill:#fff0cc,stroke:#a66b00
    classDef data fill:#eee5f5,stroke:#76528c
    class mq,mapping,controls,patents source
    class build process
    class output data

Grain and key: public gvkey, year
Location: /data/full/analyses/full_training/table89_dataset.parquet
Config: analyses/config/config_tables89_full.yaml

The build uses several analysis-support inputs:

• patent_panel.parquet, containing patent counts, citations, and values;
• firm_mq_stats.parquet, containing company-year MQ distribution measures;
• returns_volatility.parquet, containing annual monthly-return volatility.

It also uses the processed Revelio-Compustat-PERMNO mapping. The final dataset joins those products to Compustat controls and processed Revelio company attributes.

Rscript pipelines/run_build_mapping_permno.R --config config/config_tables89_full.yaml
Rscript pipelines/run_build_patent_panel.R --config config/config_tables89_full.yaml
Rscript pipelines/run_build_table89_dataset.R --config config/config_tables89_full.yaml

Before estimation, verify unique firm-year keys and inspect the reported coverage of MQ, accounting controls, returns, and patents.

The current consumers are the firm-performance and innovation analyses.