The CEO Pay Ratio Disclosure (Item 402(u)), Explained

Reading the Ratio Before You Walk Into the Room

You have a comp committee meeting on the calendar, and you have pulled up a peer company's proxy statement to understand how its board thinks about pay before your own renewal conversation. Buried past the Summary Compensation Table is a single line: a ratio, a number like 196:1, with a footnote pointing to a methodology you have never had to parse before. It looks like a compliance afterthought. It is not. That ratio is one of the few figures in a proxy statement built specifically to be compared — company to company, year to year, machine to machine. Read correctly, it tells you how a board thinks about internal pay equity, how it defines its own workforce, and how much weight it puts on optics versus substance. By the end of this piece, you will know exactly what the CEO pay ratio disclosure requires, where it lives, and how to read it the way a comp committee already does.

What Item 402(u) Actually Requires

The CEO pay ratio disclosure comes from Item 402(u) of Regulation S-K, adopted by the SEC on August 5, 2015 under Section 953(b) of the Dodd-Frank Act, per the Harvard Law School Forum on Corporate Governance. The rule requires most public companies to disclose the ratio of their CEO's total compensation to the total compensation of their median employee, calculated on the same basis used in the Summary Compensation Table.

That last clause matters. The pay ratio is not a separate compensation concept — it is a comparison built from the same total-compensation figure the proxy already discloses for the CEO, measured against a median-employee figure the company must identify and calculate itself. There is no single prescribed formula for finding the median employee; companies choose a methodology (full population, statistical sampling, or a consistently applied compensation measure) and disclose it. That flexibility is why two companies of similar size can report meaningfully different ratios even when their actual pay structures are not that different — the underlying method matters as much as the number itself.

How the Median Employee Gets Calculated — and Why It Moves

Before a board can compute a ratio, it has to define its own workforce, and that definition alone signals something. A company that includes only US employees will produce a different median, and therefore a different ratio, than one that includes its full global workforce. A company with a large hourly, part-time, or seasonal population will see its median shift depending on the measurement date it selects. None of this is disclosed as a single hard rule across all filers — it is disclosed per company, in the footnote to that company's own ratio, which is exactly why the ratio should be read alongside its methodology note rather than as a bare number.

This is also where scale matters for comparison. The Russell 3000 measures roughly 3,000 US companies and represents approximately 98% of investable US equity market cap, per LSEG / FTSE Russell (2025) — a useful reminder that a ratio calculated at a 300-person public company and a ratio calculated at a 300,000-person multinational are answering structurally different questions, even when the arithmetic looks the same.

Reading the Ratio Against a Real Benchmark

Once you understand the mechanics, the number becomes useful. The S&P 500 median CEO-to-median-worker pay ratio rose to 196:1 in 2023, with the median S&P 500 employee earning $81,467 that year, per the Equilar / Associated Press CEO Pay Study (2024, FY2023). The AFL-CIO's Executive Paywatch reported an average S&P 500 CEO-to-worker pay ratio of 285:1 for 2024, alongside average CEO pay of $18.9 million — a reminder that "average" and "median" ratios, calculated differently across studies, are not interchangeable and should not be read as the same figure.

If you are sitting across from a comp committee that points to a peer group's ratios as justification for a pay structure, the discipline is the same one that applies to every other figure in the proxy: ask what index, what year, and what methodology produced the number being cited. A 196:1 median from one dataset and a 285:1 average from another are both real, sourced figures — but they answer different questions, and conflating them in a negotiation weakens your position rather than strengthening it.

Where the Disclosure Lives — and How XBRL Changes What You Can Do With It

The pay ratio sits in the proxy statement, not the annual report, and it travels alongside a related but separate disclosure: Pay Versus Performance under Item 402(v), adopted August 25, 2022 and effective for 2023 proxies, per Mintz (2022). Unlike the pay ratio, PvP mandates Inline XBRL tagging — a structured data language that produces a single document readable by both people and machines, per SEC.gov (2024). That tagging requirement lives in the proxy or information statement itself, not in the Form 10-K, per Greenberg Traurig (2023), which means the machine-readable version of a company's pay data is filed in the same document you are already reading, not a separate technical exhibit.

Practically, this means the pay ratio and the PvP table can now be pulled, compared, and aggregated across companies at a scale that was not possible before XBRL tagging existed. SEC EDGAR provides free public access to every filing, per SEC.gov (2025), which is the same public record CEOSalary parses — cross-validating Summary Compensation Table figures against XBRL-tagged Pay Versus Performance and CEO Pay Ratio disclosures to build sourced, percentile-positioned peer comparisons. If you want the mechanics of that XBRL structure explained in more depth, see our guide on xbrl-tagged executive compensation data.

What the Ratio Doesn't Tell You

The pay ratio is a useful signal, but it is not a complete picture of how a board sets executive pay, and it should never be read as a proxy for whether a specific pay package is defensible. It says nothing about performance conditions attached to equity, nothing about how a board's compensation committee benchmarks against a peer group, and nothing about the "compensation actually paid" figure that Item 402(v) requires alongside stock and financial performance metrics — a distinct calculation covered in our companion piece on what compensation actually paid means. A company can report a moderate ratio while still structuring a CEO package poorly relative to performance, or a high ratio while paying strictly in line with a well-constructed, performance-tested peer group.

A pay ratio tells you how far apart two medians sit. It does not tell you whether either number is right for the role.

Boards also rarely face binding pushback on pay structure through the ballot box: only about 1.2% of Russell 3000 say-on-pay votes failed in 2024, down from 2.1% in 2023, per the Harvard Law School Forum on Corporate Governance (2025, FY2024). That low failure rate is itself informative — it means the pay ratio disclosure functions more as a research and negotiation input than as a mechanism that reliably forces board action on its own.

Turning the Disclosure Into Leverage

Reading a CEO pay ratio disclosure correctly means treating it as one input in a larger proxy-reading discipline: know where it sits inside the DEF 14A, know what methodology produced it, and know how it relates to the Pay Versus Performance table filed alongside it. If you have not yet worked through the full anatomy of a proxy statement, start with our guide on how to read a DEF 14A proxy statement, and if you want ratio benchmarks organized by sector before your next committee conversation, see ceo pay ratio by industry.

For a structured walkthrough of both Item 402(u) and Item 402(v) — built to help you decode a specific company's filing rather than rely on memory — the Proxy Disclosure Decoder: Pay Ratio & Pay-Versus-Performance Reader breaks down exactly where each figure lives and how the methodology footnotes change what a ratio means. It is built for the reader walking into the room with a filing in hand, not a headline number.