# The Workforce Disclosure Reference: What Each Filed Number Proves

*Look up any filed workforce number and state its population, as-of date, trigger, what it proves, how it misleads, and which filing to cross-check it against.*

- Canonical URL: https://www.refolk.ai/guides/workforce-disclosure-reference
- Pillar: Market and talent intelligence
- Format: Reference
- Published: 2026-11-06
- Last reviewed: 2026-11-06
- Reading time: 22 min
- Keywords: competitor headcount from annual report, ESRS S1 employee data competitor analysis, gender pay gap report what it shows about a company, H-1B LCA data competitor salaries, 10-K human capital disclosure turnover, public workforce data sources for talent intelligence

## Key takeaways

- Only 23% of S&P 100 companies gave a specific employee turnover rate in their 10-K human capital disclosure, up from 18% a year earlier, so most turnover claims about a US rival rest on modelled data, not filings.
- A UK gender pay gap report covers one legal entity with 250 or more relevant employees on a single snapshot day, so it says nothing reliable about the size of the group.
- IQVIA's CEO pay ratio was 343:1 with non-US staff in the population and 230:1 using US employees only, which shows the ratio is driven by who was counted, not by how well the median employee is paid.
- The median employee behind a CEO pay ratio can be reused for three years, so Linde's 391:1 ratio for 2024 rests on a median employee first identified from 2022 population data.
- ESRS S1-5 employee counts cover only a rival's ten largest countries of operation, with a de minimis of 50 employees per country, so the small, new hubs a talent analyst cares about drop out of the filing.
- An LCA is a filing, not an approval or a hire, and certified-withdrawn cases sit in the same file, so filter case status before reading offered wages as anything.

This reference is for strategy and research teams, talent-intelligence analysts, and operators who have a rival's annual report, sustainability statement, pay gap filing, and visa wage records open and need to know which workforce numbers can go into a brief. Each row covers a number a company files because the law makes it, and states the population it counts, its as-of date and lag, the threshold that triggers it, what it proves, how it misleads, and which second filing to check it against. Jump to the row you need and leave.

## What every filed workforce number must carry

A filed workforce number belongs in a brief only once it carries five tags: scope, population, as-of date, trigger, and a cross-check. Without them, a group headcount from a 10-K and one entity's headcount from a UK pay gap report look like the same kind of fact. They are not.

Terms used throughout:

- **Scope**: the legal unit the number describes, which is the whole group, one legal entity, one country, or a single filing.
- **Population**: who is counted, including exclusions such as non-employees, casual staff, or non-US workers.
- **As-of date**: the day or period the number describes. A **snapshot** is a single day. A **period figure** averages or totals across a year or quarter.
- **Lag**: the gap between the as-of date and the day the number becomes public.
- **Trigger**: the threshold that forces the filing to exist, such as 250 relevant employees per entity.
- **De minimis**: a size floor below which a rule lets the filer leave something out.

> **Rule:** Population and date travel with the number
>
> A filed workforce number goes into a brief only with its scope, population, and as-of date in the same sentence. If you cannot state all three, the number stays in your working sheet.

#### The four scopes a filed workforce number can describe

1. **Group** - 10-K human capital section, ESRS S1 totals, CEO pay ratio population
2. **Legal entity** - UK gender pay gap report, registry-filed average employee count
3. **Country** - ESRS S1-5 counts for the ten largest countries, EU pay transparency reports per member state
4. **Single filing or person** - One LCA row, one identified median employee

*Most errors in a competitor workforce brief come from presenting a number from a lower layer as if it described the layer above.*

## Regime lookup: thresholds, as-of dates, and publication lag

Each regime has its own trigger, its own as-of basis, and its own lag, and those three facts decide what a number can support. Use this table to place a figure before you read its row.

| Regime | Threshold | As-of | Publication lag |
|---|---|---|---|
| UK gender pay gap | 250+ per entity | 5 Apr snapshot | up to 364 days |
| EU Pay Transparency Directive | 150+ first, 100+ later | calendar year | about 5 months after year end |
| Australia WGEA | 100+ | annual (snapshot unverified) | about a year after the period |
| ESRS S1-5 / S1-7 | 10 largest countries, 50 per country | financial year end | with annual report |
| LCA (US visa wage filings) | every filing | quarter | about 11 to 12 weeks (derived) |
| Companies House accounts | all companies | financial year average | 9 months private, 6 months plc |

Notes on reading the table:

- **UK**: the snapshot is 5 April for most employers and 31 March for public authorities, and the deadline is a year later, so a report can describe a day nearly a year old on publication.
- **EU directive**: the first reporting period is a full calendar year. Employers with 250+ workers report annually and those with 150 to 249 every three years, with 100 to 149 phased in later. National transposition sets the actual deadline.
- **WGEA**: the employer-level release covering the 2024-25 period came out in March 2026. I have not verified the snapshot date; read WGEA's methodology note before stating one.
- **LCA**: the lag is derived from one quarterly release (April to June data added on 20 September). Check the release note on the Department of Labor's OFLC performance data page for each quarter.

Lag and coverage together decide how a number can be used. Plot any figure on these two axes before deciding where it sits in the brief.

#### Where a filed number belongs in a brief

Horizontal axis runs from Narrow slice of the workforce to Whole workforce. Vertical axis runs from Stale, long lag to Fresh, short lag.

| Quadrant | What it means |
| --- | --- |
| Fresh but narrow | LCA offered wages; use for role and wage-level signal, never for headcount |
| Fresh and whole | Rare in mandatory filings; lead with it if you have it, and still date it |
| Stale and narrow | UK entity pay gaps, small-country ESRS gaps; context lines only |
| Stale but whole | 10-K headcount, registry averages, pay ratio; baseline figures, always dated |

*Fresh numbers tend to cover a narrow slice and whole-workforce numbers tend to be stale, so almost every figure needs a caveat.*

## US rows: 10-K human capital, median-employee pay, and LCA wages

US filings give a group-level narrative (the 10-K), one dated person (the median employee), and a stream of role-level wage offers (LCAs). None of them is a headcount you can trend without a second source.

### 10-K human capital: headcount, turnover, and workforce mix

- **Counts**: whatever the company chooses to state. The SEC rule did not define "human capital" or say what the disclosure should contain beyond a few examples in the rule text. Quote Item 101(c)(2)(ii) from eCFR rather than paraphrasing it.
- **As-of and lag**: the date the filing states, published with the annual report. If no date is stated, do not assume one.
- **Trigger**: any company filing a 10-K.
- **Proves**: what the company is willing to say about its workforce, under its own name, in a filing it signs.
- **Misleads**: turnover, above all. In a law-firm survey of S&P 100 filings, a majority discussed turnover only qualitatively. Only 23% gave a specific turnover rate, up from 18% the year before. One industry subset reached 41% in a later year. Full-time and part-time metrics were disclosed by 17% of the S&P 100. I have not confirmed the exact wording on whether a turnover formula is prescribed, but filings do not share one, and voluntary versus total turnover is often unlabelled.
- **Cross-check**: ESRS S1-5 country counts where the group reports under CSRD, and registry-filed average employee counts for its UK entities.

**23% - S&P 100 companies that gave a specific employee turnover rate in the 10-K**

Up from 18% the previous year, so about three-quarters published no turnover rate at all.

### Median-employee pay and the CEO pay ratio

- **Counts**: the annual total compensation of one identified median employee, set against the CEO's. The rule is 17 CFR 229.402(u).
- **As-of and lag**: the registrant picks a date within the last three months of its last completed fiscal year to identify the median employee, and must disclose that date. It needs to identify a new median employee only once every three years, provided nothing has changed that it reasonably believes would significantly change the ratio.
- **Trigger**: registrants covered by 229.402(u). Check the rule text for scope.
- **Proves**: the pay of one person near the middle of the counted population, on the date and measure the company chose.
- **Misleads**: the median can be up to three years old, identified on a narrow measure such as base pay alone, and drawn from a population that excludes up to 5% of total employees who are outside the US.
- **Cross-check**: LCA offered wages for comparable US roles, and the population size stated in the proxy against the 10-K headcount.

| Company | Median comp | Ratio | Median identified |
|---|---|---|---|
| Linde (FY2024) | $52,889 | 391:1 | 2022 data |
| Paychex (FY2023) | $69,165 | 101:1 | 2023 |
| ACI Worldwide (FY2022) | $78,168 | 139:1 | FY2022 |
| IQVIA (FY2024) | not used here | 343:1 vs 230:1 US-only | 2023 |

How to read the cases:

- **Linde** reported a CEO figure of $20,696,088 against a median of $52,889. The ratio is for 2024, but the median employee was first identified on 2022 population data.
- **IQVIA** reported 343:1 with the de minimis exclusion applied, against 230:1 using US employees only. Keeping low-cost non-US staff in the population makes the ratio 49% higher (343 divided by 230). A later proxy reused the 2023 median employee at 275:1.
- **Paychex** stated a global population of about 16,200 alongside its median, which gives you a headcount anchor to reconcile against the 10-K.

### LCA offered wage

- **Counts**: one row per filing, with case status, visa class, SOC code (the Standard Occupational Classification code for the role), worksite, wage level, and offered wage. The wage level field, PW_WAGE_LEVEL, takes the values I, II, III, IV, or N/A. Worksites ship as a separate file.
- **As-of and lag**: the fiscal quarter of the filing, published roughly 11 to 12 weeks after quarter end.
- **Trigger**: every filing.
- **Proves**: the employer intended to fill a role in a named occupation, at a stated wage level, in a stated place, and attested to an offered wage.
- **Misleads**: an LCA certification is not a petition approval and not a hire. Certified-withdrawn cases sit in the same file. No quantified public figure for how many LCAs become hires has been established, so do not assume a conversion rate. I have not verified the PERM field list; read its record layout before using it.
- **Cross-check**: USCIS approval and denial data, which usually updates once a year, and the proxy median for a population-level comparison.

> **Watch out:** An LCA row is not a salary or a hire
>
> Filter CASE_STATUS before you count anything, and never present offered wages as what a rival pays. Read the wage level next to the wage, because a Level I offer and a Level IV offer for the same SOC code describe different hires.

## UK, EU, and Australian pay gap rows

Pay gap filings prove the shape of pay inside a defined population on a defined basis. They do not prove company size, and they do not prove pay levels, except where the regulator publishes average remuneration by quartile.

### UK gender pay gap report

- **Counts**: relevant employees of one legal entity. The requirement applies separately to each legal entity with 250 or more relevant employees.
- **As-of and lag**: a 5 April snapshot, or 31 March for public authorities. The deadline is 4 April the following year, with public-sector deadlines at the end of March. Lag can reach 364 days.
- **Measures**: six of them.
  - mean pay gap
  - median pay gap
  - mean bonus gap
  - median bonus gap
  - bonus proportions by gender
  - quartile pay bands
- **Proves**: how pay and bonus are distributed by gender inside that one entity on that one day.
- **Misleads**: the report covers one legal entity, not the group. A low gap may only mean a small or unusual entity.
- **Cross-check**: the group's other entities on the government's reporting service, and the registry-filed average employee count for the same entity.

The UK government has said it intends to make ethnicity and disability pay gap reporting mandatory. Re-check its status before assuming those reports exist for a rival.

### EU Pay Transparency Directive report

- **Counts**: workers of an employer in a member state, under Directive (EU) 2023/970. Employers with 250+ workers report annually, those with 150 to 249 every three years, and those with 100 to 149 are expected to start later.
- **As-of and lag**: a full calendar year, with the first reports due 7 June 2027 and smaller employers by 7 June 2031. Member states diverge: Slovakia moved its deadline to 15 April.
- **Measures**:
  - median and mean gap, company-wide
  - median and mean gap by category of workers doing work of equal value
  - variable pay gap
  - gender share in each pay quartile
  - share receiving variable pay, by gender
- **Proves**: closer to like-for-like comparison than the UK report, because of the equal-value categories.
- **Misleads**: transposition is national and uneven. A rival with UK and EU entities can show two different gaps for the same year, one from a single-day UK headcount and one from an EU calendar year.
- **Cross-check**: the UK report for the same group, and ESRS S1 if the group reports under CSRD.

### Australia WGEA employer gaps

- **Counts**: employers with 100 or more employees. WGEA publishes each employer's median gap for total remuneration and for base salary, plus average gaps and quartile composition with average remuneration.
- **As-of and lag**: annual. The 2024-25 release was published in March 2026. The snapshot date is not verified here.
- **Proves**: a regulator-computed gap for a named employer, plus a relative pay-level signal from average remuneration by quartile. The release covered more than 10,500 employers, and the midpoint of employer median total remuneration gaps was 8.0%, which gives you a peer anchor. A secondary source puts coverage at about 5,000 employers; WGEA's own report wins.
- **Misleads**: the 2022-23 calculation excluded CEOs, heads of business, and casual staff, so it is not comparable with later years.
- **Cross-check**: the 8.0% midpoint, and the same employer's earlier releases on a matching basis.

When a UK report does not explain how the entity defined relevant employees or bonus pay, the quickest answer comes from someone who has run that reporting. Finding them by title is slow, because the work sits under reward and compensation titles. In [Refolk](/) you describe the work instead:

I ran this search: `Reward managers who have run UK gender pay gap reporting at companies with 250+ staff` - [see the full result list](https://www.refolk.ai/s/q0mypk4t3n).

*Returns named reward and compensation practitioners with public profiles showing UK pay gap reporting work, so you can ask how a snapshot population is usually defined.*

## ESRS S1 and registry rows

ESRS S1 gives employee counts for a rival's ten largest countries of operation only, and registry accounts give an annual average, not a point-in-time headcount. Both are solid baselines, and both hide change.

> **Note:** Confirm S1 numbering in the Official Journal
>
> The revised standards were published in the EU Official Journal as Delegated Regulation (EU) 2026/1563. Secondary sources disagree on structure: one lists S1-1 to S1-16, another says S1 sub-topics merged into a combined S1/S2 block. Cite the Official Journal text, not a summary.

### ESRS S1-5: employee characteristics

- **Counts**: employees of the group, broken down by country, for the ten largest countries of operation, with a de minimis of 50 employees per country.
- **As-of and lag**: financial year end, published with the annual report. Application is mandatory for financial years starting on or after 1 January 2027, with optional early application for 2026. Re-check those dates against the regulation.
- **Trigger**: S1 applies only when own workforce is material. Once it is, S1-5 is required.
- **Scale**: EFRAG's draft list counts 49 "shall" datapoints in S1, and mandatory ESRS datapoints overall were cut by 61%.
- **First-year omissions** for wave-one reporters:
  - non-employee data
  - disability inclusion
  - certain health and safety indicators
  - work-life balance
  - social dialogue outside the European Economic Area
  - social protection coverage
  - training
- **Proves**: a group-signed, country-level count for the countries where the rival is largest.
- **Misleads**: countries under 50 employees and countries outside the top ten disappear. Those are the new hubs a talent analyst is usually hunting.
- **Cross-check**: reconcile the S1-5 total against the 10-K or annual report total. The gap between them is your upper bound on the hidden countries.

To size what S1-5 leaves out, count the rival's people in a candidate country from public profiles. Refolk returns named profiles by employer and location, which turns "is there a team in Lisbon?" into a list you can date.

### ESRS S1-6: non-employees

- **Counts**: a single essential datapoint on non-employees, required only when non-employees are critical to the business model.
- **Misleads**: absence does not mean the rival has no contractors. It means the company judged them not critical, or used a first-year omission.
- **Cross-check**: the 10-K narrative on contractors, if any.

### ESRS S1-7: probably collective bargaining coverage

- **Counts**: not verified. It probably covers collective bargaining coverage and uses the same ten-country, 50-employee threshold as S1-5. Confirm the topic in the Official Journal text before citing it.
- **Misleads**: social dialogue outside the European Economic Area is a first-year omission, so a blank there is not a finding.

### Companies House average number of employees

- **Counts**: the average number of employees of one UK company over its financial year, which is also the basis for its size tests. I have not verified the monthly averaging method in the Companies Act; read section 411 before stating one.
- **As-of and lag**: the financial year, filed within 9 months for a private company or LLP and 6 months for a public company.
- **Proves**: a statutory, entity-level baseline that exists even for subsidiaries with no other disclosure.
- **Misleads**: an average lags both a ramp-up and a layoff.
- **Cross-check**: the year-end headcount in the strategic report, where one is given.

### Layoff counts

I have not built a row for statutory layoff notices, because I have not verified their thresholds and lags. Until you check the local notice law, treat layoffs as visible only indirectly: a gap between the registry average and the year-end headcount, a drop in ESRS country counts year on year, or a stated turnover rate in the 10-K.

## Who can explain a filing's methodology

The people who can explain how a rival built a number are the reward, reporting, and immigration practitioners who produce these filings, and few of them carry a "talent intelligence" title. Search by the work they do, not by title.

| Title search in Refolk's index | Profiles |
|---|---|
| Talent Intelligence, US | 3 |
| Talent Intelligence, UK | 1 |
| People/Workforce Analytics, UK | 5 |

**1 - UK profile in Refolk's index with a Talent Intelligence title**

Five UK profiles carry People or Workforce Analytics titles, so the function mostly sits under other names.

In Refolk's index of professional profiles, the US Talent Intelligence profiles sit at JPMorganChase, Netflix, and Google, the UK one at Booking.com, and the UK analytics profiles at Visa, Fiverr, Barclays, TikTok, and Concentra Analytics. These counts are far too small to size a market, and title matching is narrow, so do not present them as population figures. What they do show is that a title search will miss most of the people you need. Describe the work instead, for example "CSRD sustainability reporting managers in Germany working on ESRS S1 workforce metrics" or "Immigration program managers at US tech companies that file H-1B LCAs".

## Cross-checking a filed number: the procedure

The procedure takes one figure from a filing to a caveated line in a brief in six steps, mostly under an hour plus one to two hours of cross-checking. Steps 4 and 5 can swap order, depending on which filing you open first.

#### From filed number to brief line

1. **Map each number to an entity and a regime** - Tag every figure with group, legal entity, country, or single filing, plus the regime that produced it. About 30 minutes; UK pay gap figures are always per entity.
2. **Record the as-of date and lag** - Write the snapshot day or period and the publication date for each figure. About 15 minutes; done when you can state each lag.
3. **Check the threshold trigger** - Record why the filing exists: 250 per UK entity, 150 or 250 for the EU directive, 100 for WGEA, ten largest countries and 50 per country for ESRS. About 10 minutes.
4. **Read the methodology note** - List the measure used to pick any median employee, every exclusion, and headcount versus FTE. About 20 minutes; done when exclusions sit next to the figure.
5. **Cross-check against a second filing** - Compare with an overlapping filing, such as 10-K headcount against ESRS S1-5 or LCA wages against the proxy median. One to two hours; done when the variance is explained.
6. **Write the brief with caveats** - Put population, scope, as-of date, and exclusions in the same sentence as the number. Done when no number can be quoted without them.

Use this record card for every figure, then lift the last line into the brief.

**Filed workforce figure record**

```
FIGURE:
VALUE:
SOURCE FILING AND SECTION:
REGIME:
SCOPE: group / legal entity / country / single filing
POPULATION COUNTED:
EXCLUSIONS (with share if stated):
HEADCOUNT OR FTE:
AS-OF DATE OR PERIOD:
PUBLICATION DATE:
LAG:
TRIGGER THAT MADE IT EXIST:
CROSS-CHECK FILING:
VARIANCE AND EXPLANATION:
BRIEF LINE: [Rival] reported [value] for [population] in [scope], as of [date], per [filing]; excludes [exclusions]; reconciles with [cross-check] within [variance].
```

*One card per figure; the brief line is the only part that leaves your working sheet.*

## How filed workforce numbers mislead

Filed numbers mislead in predictable ways: wrong scope, stale dates, silent exclusions, and filings mistaken for outcomes. Each failure below has a tell and a check.

> A low pay ratio is a statement about who was counted, not about how well anyone is paid.

**1. UK pay gap headcount read as group size.**
- *Tell*: a brief says "Rival has 300 staff" from a pay gap report.
- *Why it lies*: the report covers one legal entity with 250+ relevant employees, not the group. A low gap may just mean a small entity.
- *Check*: look up the group's other entities on the reporting service and sum them, knowing entities under 250 do not appear.

**2. Median-employee pay used as a wage benchmark.**
- *Tell*: "Rival pays a median of $52,889" quoted against a current salary band.
- *Why it lies*: the median may be up to three years old, identified on base pay alone, and drawn from a population that excludes up to 5% of non-US staff. Linde's 2024 figure rests on 2022 data.
- *Check*: read the identification year and the exclusion list in the proxy, and label the figure as a dated global marker.

**3. Pay ratio read as a pay-level signal.**
- *Tell*: "a low ratio means the rival pays well".
- *Why it lies*: low-paid non-US staff in the population pull the median down and the ratio up. IQVIA shows 343:1 against 230:1 US-only.
- *Check*: note whether the population is global, US-only, or de minimis-adjusted before comparing ratios across companies.

**4. LCA offered wage read as actual pay or hires.**
- *Tell*: "Rival hired 40 engineers at $X" built from an LCA count.
- *Why it lies*: an LCA is a filing, not an approval. Certified-withdrawn cases sit in the same file, and no public LCA-to-hire conversion rate has been established.
- *Check*: filter CASE_STATUS, read PW_WAGE_LEVEL, and cross-check USCIS approval data.

**5. ESRS country counts read as the full footprint.**
- *Tell*: "Rival has no presence outside these ten countries".
- *Why it lies*: countries under 50 employees and outside the top ten drop out.
- *Check*: reconcile the S1-5 total against the 10-K or annual report total, then search public profiles for the gap.

**6. 10-K turnover compared across companies.**
- *Tell*: two rivals' turnover rates side by side in one chart.
- *Why it lies*: filings follow no shared formula, and voluntary versus total turnover is often unlabelled. Only 23% of the S&P 100 gave a rate at all.
- *Check*: quote the filing's own definition next to each rate, or drop the comparison.

**7. WGEA trend across a basis change.**
- *Tell*: a gap "narrowing" from 2022-23 to later years.
- *Why it lies*: the 2022-23 calculation excluded CEOs, heads of business, and casual staff.
- *Check*: trend only across years with a matching basis.

**8. Registry average read as point-in-time headcount.**
- *Tell*: a Companies House average used to date a layoff or ramp-up.
- *Why it lies*: an average smooths and lags both.
- *Check*: compare with the year-end headcount in the strategic report.

**9. UK and EU gaps treated as contradictory.**
- *Tell*: "the rival's two filings disagree".
- *Why it lies*: the UK report uses one day's headcount per entity and the EU report a calendar year with equal-value categories. Different numbers are expected.
- *Check*: state both bases in the brief and do not average them.

## Keeping this reference current

Most rows here change by mechanism, not by data, so re-check the mechanism on a schedule rather than trusting a remembered value. Set reminders against these sources:

- **ESRS**: the Official Journal text of Delegated Regulation (EU) 2026/1563, for final S1 numbering, the S1-7 topic, and application dates.
- **EU directive**: each member state's transposing law, for deadlines and any local divergence like Slovakia's.
- **UK**: the government's position on mandatory ethnicity and disability pay gap reporting.
- **WGEA**: each release's methodology note, for the snapshot date and any change in who is excluded.
- **LCA**: each quarter's record layout on the OFLC performance data page, since fields can change between releases. USCIS approval data usually updates once a year.
- **10-K**: Item 101(c) on eCFR, for any rule change on human capital content.

#### Before a filed workforce number leaves your sheet

- [ ] The figure is tagged with one scope: group, legal entity, country, or single filing.
- [ ] The population and every stated exclusion are written next to the figure.
- [ ] The as-of date and publication date are both recorded, and the lag is stated.
- [ ] You can name the threshold that made the filing exist and what a company below it would show.
- [ ] Headcount versus FTE is stated, or marked as not stated by the filer.
- [ ] The figure has been compared with one second filing and the variance is explained.
- [ ] Any median-employee figure shows its identification year.
- [ ] Any LCA-derived figure is filtered by case status and labelled as offered, not paid.
- [ ] No turnover rate is compared across companies without each filing's own definition.
- [ ] The brief line carries population, scope, and date in the same sentence as the number.

## Frequently asked questions

### Can I get a competitor's headcount from its annual report?

Yes, but only the headcount the company chose to state, on the date and basis it chose. The SEC rule behind the 10-K human capital section does not define human capital, so check whether the figure is global or US-only and headcount or full-time equivalent. Then reconcile it against ESRS S1-5 country counts if the group reports under CSRD, and against registry-filed average employee counts for its UK entities.

### What does a UK gender pay gap report show about a company?

It shows the pay gap shape inside one legal entity with 250 or more relevant employees, on a snapshot day of 5 April, or 31 March for public authorities. It gives six measures: mean and median pay gap, mean and median bonus gap, bonus proportions, and quartile pay bands. It does not show group size, group pay levels, or entities below the threshold.

### Does H-1B LCA data show what a competitor actually pays?

No. An LCA row shows the wage an employer offered on a filing for a role, SOC code, worksite, and wage level, not the wage paid or even whether anyone was hired. Certification is not petition approval, and certified-withdrawn cases sit in the same file. Filter by case status, read the wage level field, and cross-check against USCIS approval data before using it.

### Is the CEO pay ratio median employee a good salary benchmark?

Usually not. The company may reuse the same median employee for three years, may identify it on a narrow measure such as base pay, and may exclude up to 5% of total employees who are outside the US. Linde's 2024 median rests on 2022 population data. Use it as a dated, global-population marker, never as a role-level wage benchmark.

### What employee data does ESRS S1 require after the simplification?

If own workforce is material, the company must disclose employee characteristics under S1-5, using a significance threshold of its ten largest countries of operation with a de minimis of 50 employees per country. Non-employee data under S1-6 shrinks to one datapoint, required only when non-employees are critical to the business model. Confirm the final numbering in the Official Journal text.

### How often do US companies disclose turnover in the 10-K?

Rarely as a number. In a law-firm survey of S&P 100 filings, only 23% gave a specific employee turnover rate, up from 18% the year before, and most discussed turnover qualitatively. One industry subset reached 41% in a later year. Where a rate is given, check whether it is voluntary or total turnover, because filings often leave it unlabelled.

---

*From the Refolk guide library. I revise these guides rather than replacing them, so the current version is always at https://www.refolk.ai/guides/workforce-disclosure-reference*
