General Counsel Pay at the Largest US-Listed Companies

Unit. One row per company: the top legal officer (General Counsel, Chief Legal Officer or equivalent) where that person is a named executive officer in the latest fiscal-year row of the Summary Compensation Table, including a GC who held the role for part of the year and has a table row. Where two people held the role in the year, the row is the holder at fiscal year-end. No person is named in the dataset.

Full-year sitting GC (the base of every statistic). A row counts only if the person held the top legal role for at least 9 months of the fiscal year and still held it at fiscal year-end, or left fewer than 30 days before it. Dates come from the filing (footnotes, compensation discussion, new-hire grant dates). Of 262 rows, 236 qualify; 14 joined or were promoted too late in the year, 10 left too early (including advisers after a handover) and 2 could not be dated from the filing and are treated as not qualifying. 8 rows in the statistics cover part of a year of pay because the GC started 9 to 11 months before year-end; their salary and total are understated.

Universe. The 500 largest US-listed operating companies by public float (SEC XBRL tag dei:EntityPublicFloat, latest value on the Form 10-K cover; funds, SPACs, non-operating trusts, subsidiaries and foreign private issuers filing Form 20-F are excluded). Float runs from $4.0tn to $14.3bn and is measured at each company's second fiscal quarter end (348 of 500 at 2025-06-30). Float is a proxy for market capitalisation, not the same thing.

Proxy statements. For each company the latest DEF 14A filed 1 October 2025 to 29 September 2026 (DEFR14A, DEFC14A or a 10-K/A carrying Part III where no DEF 14A exists). 494 of 500 had a proxy statement on file in the window; 4 of those had a Summary Compensation Table that could not be read and 6 companies had none in the window. The data therefore cover 490 companies, not 500.

Screen. The table was parsed from the filing text and every named executive's title, in the table and anywhere else in the proxy, was searched for legal-chief wording (General Counsel, Chief Legal Officer and variants; weaker hits: Secretary, Legal, Counsel). 266 companies were flagged (258 strong, 8 weak) and 224 had no legal-chief title among their named executives. The screen is text-based: a GC listed only by name in a table whose titles appear nowhere else in the proxy would not be flagged.

Extraction. Each flagged company's table and footnotes were read field by field: 262 rows are a top legal officer in the latest fiscal-year row; 4 flagged companies were not (the hit was another executive, or the GC's row was not in the latest year). Amounts are in US dollars as printed; a dash is zero. Checked mechanically: the seven components add to the Total within 1%, every amount appears in the filing text, outlier and year-on-year checks. A partial year is reported as printed and flagged, never annualised.

Verification. A second reader re-extracted 68 companies blind from the raw filing HTML, without seeing our extraction: 30 drawn at random with a fixed seed, plus every medium- or low-confidence row and every flagged company we had rejected. 691 of 693 numeric cells matched (99.7%) after adjudication; the mismatch was a non-equity incentive amount placed in an option column the table does not have. The same defect was then sought by comparing table headers with the extraction (210 of the 261 rows then in the data could be mapped mechanically), and one more company had it. Both were corrected (2 corrections), and every row is now checked for an amount in a column its table does not show. A further random draw of 10 extracted rows contained 0 that were not the top legal officer.

Statistics. Full-year sitting GCs only (236 of 262). Medians and quartiles use inclusive linear interpolation; a cut with fewer than 10 rows shows no quartiles. Outliers stay in medians and quartiles and leave only the trimmed mean. Fiscal-year labels of the rows: 2025: 231, 2026: 31.