Revelio–Compustat company mapping

This crosswalk connects Revelio firms (rcid) to Compustat firms (gvkey).

flowchart LR
    revelio[(Revelio company reference)]
    comp[(Processed annual Compustat)]
    match[Validate supplied gvkeys<br/>and match strong identifiers]
    mapping[(Revelio-Compustat mapping)]

    revelio --> match
    comp --> match
    match --> mapping

    classDef source fill:#e8eef7,stroke:#52739e
    classDef process fill:#fff0cc,stroke:#a66b00
    classDef data fill:#e2f2e7,stroke:#3b7a4b
    class revelio,comp source
    class match process
    class mapping data

Grain and key: one retained rcid
Location: /data/full/processed/mapping_revelio_compustat.parquet
Builder: shared/data/build_mapping_revelio_compustat.py

The mapping first retains Revelio-provided gvkeys found in the North American Compustat extract. For remaining firms, it accepts CUSIP or CIK matches when no identifier disagrees. Ticker alone is not accepted because tickers collide across exchanges.

Column	Meaning
rcid	Revelio firm identifier
gvkey	Compustat firm identifier, stored without required zero padding
match_method	revelio_provided or multi_identifier
match_score	Current v1 confidence field

The documented v1 build contains 24,897 mappings to 21,844 distinct gvkeys: 16,728 supplied by Revelio and 8,169 recovered from identifiers. These counts should be regenerated when the source files change.

Important limitations:

• several rcid values can map to one gvkey;
• Revelio has limited coverage of old delisted firms;
• Revelio-provided global gvkeys outside the North American extract are not rematched;
• identifiers are not time varying in the current mapping.

python shared/data/build_mapping_revelio_compustat.py

When aggregating to gvkey, downstream code must explicitly handle multiple Revelio firms per Compustat firm. The Revelio-Compustat-PERMNO mapping extends this crosswalk for CRSP-linked analyses.