Insights

EU Pay Transparency Directive Meets ESRS S1: Why 2026 Is the Year Your Pay Data Gets Checked Twice

Last updated · October 9, 2026

The transposition deadline for Directive (EU) 2023/970 on pay transparency passed on 7 June 2026. Only four member states - Slovakia, Italy, Lithuania, and Malta - met it. The remaining 23, including Germany, France, Spain, the Netherlands, Sweden, and Denmark, missed it by varying degrees. The European Commission has not opened infringement proceedings yet, but it has refused every request to delay or simplify the directive.

That fragmented picture does not let employers off the hook. The obligations are live in the countries that transposed, and the first gender pay gap reports for employers with 250 or more employees are due by 7 June 2027, covering 2026 pay data. Employers with 150-249 employees face the same June 2027 deadline but report every three years. Employers with 100-149 employees have until 7 June 2031 for their first report.

Here is the problem that most compliance teams have not yet named clearly: 2026 is simultaneously the reference year for ESRS S1 compensation metrics inside your sustainability statement. The pay data your HR team is capturing right now will surface in two places - a statutory pay transparency report to a national authority, and a sustainability disclosure reviewed by an assurance provider. If the two numbers differ and you cannot explain why, you have a credibility problem in front of a regulator, a works council, and an auditor at the same time.

Why this lands on the CSRD desk, not just the legal desk

The instinct is to route pay transparency to employment lawyers and leave ESRS S1 to the sustainability team. That split is the root cause of the problem.

Both regimes draw on the same underlying dataset: employee-level pay records, broken down by gender, employment type, and job classification. The legal team cannot produce a defensible pay gap report without clean, structured HR data. The sustainability team cannot produce an auditable S1-16 disclosure without the same. When those two teams work from different extracts of the same payroll system - using different population cuts, different pay definitions, and different reference dates - the numbers diverge. That divergence is not a legal problem or a reporting problem. It is a data governance problem, and it needs to be solved before 31 December 2026, when the reference year closes.

What the Pay Transparency Directive actually requires

The directive's obligations fall into two tiers: those that apply to all employers from the date of national transposition, and those that scale with headcount.

From the date of national transposition, every employer must:

Provide job applicants with the initial pay level or pay range for a role, based on objective and gender-neutral criteria, either in the vacancy notice or before the first interview. Phrases like "negotiable" or "market competitive" do not discharge this obligation - employers must name a figure or a range.
Stop asking candidates about their current or previous salary. The ban applies to verbal questions, application forms, background checks, and automated screening tools.
Remove pay-secrecy clauses from employment contracts. Workers must be free to disclose their own pay to colleagues.
Provide employees, on request, with information on their individual pay level and the average pay levels, broken down by sex, for categories of workers doing the same work or work of equal value.

For employers with 100 or more employees, the reporting obligations are:

Employers with 250+ employees report annually; employers with 150-249 report every three years; both groups file their first report by 7 June 2027 on 2026 data.
Reports must cover the gender pay gap across the whole workforce and within categories of workers.
Where a gap of 5% or more in any category of workers cannot be justified by objective, gender-neutral criteria and is not remedied within six months, a joint pay assessment with workers' representatives becomes mandatory. Note that the 5% threshold applies category by category, not to the overall average - a company with a 2% headline gap can still trigger the mechanism if one job family shows a 5.1% gap.
The burden of proof shifts to the employer: in equal-pay claims, it is now for the employer to prove that no pay discrimination occurred, not for the worker to prove that it did.
Workers who suffer pay discrimination are entitled to full compensation - back pay, bonuses, non-material damages, and interest - with no statutory cap.

National transposition determines the exact calculation method, reporting format, and enforcement body. A group operating across 15 member states may face 15 slightly different national methodologies — all of which must eventually feed one consolidated ESRS S1 figure. Build your data architecture to accommodate that variation from the start.

What ESRS S1 requires on compensation

Under S1-16 (Remuneration metrics), ESRS S1 requires two disclosures when own workforce is material - and for the overwhelming majority of reporters, it is:

The gender pay gap - defined as the difference in average gross hourly pay between male and female employees, expressed as a percentage of the average pay level of male employees. The methodology is prescribed: gross hourly pay for all employees, using the formula [(Average gross hourly pay - male) - (Average gross hourly pay - female)] / (Average gross hourly pay - male) × 100.
The annual total remuneration ratio - the ratio between the total annual remuneration of the highest-paid individual and the median annual total remuneration of all other employees.

The undertaking may also disclose a voluntary breakdown of the gender pay gap by employee category, country, or segment, and by basic salary versus variable components. The revised ESRS draft (EFRAG, November 2025) retains both mandatory datapoints and tightens the rules on what counts as "pay" for the ratio calculation.

The double materiality gate applies, but own workforce is material for virtually every company with a significant employee base. If you are in CSRD scope, assume S1-16 applies.

For a full treatment of S1 disclosure requirements, see our ESRS S1 Own Workforce guide.

The reconciliation table: where the two regimes diverge

This is the practical heart of the problem. The two frameworks look similar on the surface - both measure the gender pay gap - but they differ on almost every technical dimension.

Pay Transparency Directive vs ESRS S1-16: Side-by-Side

Dimension	Pay Transparency Directive	ESRS S1-16
Legal basis	Directive (EU) 2023/970, transposed into national law	CSRD / ESRS delegated act (EU) 2023/2772
Who must report	Employers with 100+ employees (phased)	CSRD in-scope undertakings where S1 is material
Population covered	Legal-entity / national workforce in the transposing member state	Consolidated group workforce across all entities in the reporting perimeter
Pay definition	Gross hourly pay (base + variable components as defined nationally)	Gross hourly pay for gap; total annual remuneration for the ratio
Breakdown required	Categories of workers doing equal work or work of equal value	Overall gap mandatory; breakdown by category/country/segment voluntary
Mean vs median	Mean (average) gap per category	Mean gap (S1-16); median used for the remuneration ratio denominator
Reference period	Calendar year 2026 (first report)	Financial year 2026 (for FY2026 reporters)
Publication venue	National authority / public register (varies by member state)	Sustainability statement in the management report
First deadline	7 June 2027 (250+ employees)	Varies by wave; FY2026 reports due in 2027
Assurance / verification	None specified in the directive itself	Limited assurance by statutory auditor or accredited third party

The specific traps to watch:

Population mismatch. The directive covers the workforce in a given member state. ESRS S1 covers the consolidated group. A German subsidiary's statutory filing covers German employees only; the ESRS figure covers everyone. Rolling one into the other is arithmetically wrong.
Pay definition drift. National transposition may include or exclude specific variable pay components. ESRS AR 98 is explicit: gross hourly pay for all employees. If your national filing uses a narrower definition, the two numbers will differ even if the underlying data is identical.
15 methods, one ESRS number. A group operating across 15 member states may receive 15 national filings calculated under 15 slightly different methodologies. None of those figures can simply be averaged into the ESRS group figure. You need a single governed dataset at employee level from which both outputs are derived independently.

"Equal work and work of equal value": the real 12-month project

The directive's reporting obligation is straightforward in concept: report the pay gap within categories of workers doing equal work or work of equal value. In practice, defining those categories is the hardest part of the whole exercise - and it is the thing that takes 12 months, not the reporting.

"Equal work" is relatively tractable: same job, same level. "Work of equal value" requires a gender-neutral job evaluation methodology that assesses roles on skills, effort, responsibility, and working conditions. The directive names four primary criteria and 14 sub-factors for assessing work of equal value. Without a structured job architecture that maps every role to these criteria, you cannot produce a legally defensible category-level gap figure. You also cannot respond to an employee's right-to-information request with anything more than a headline number.

This is not a reporting project. It is a job architecture project that happens to produce reporting as a by-product. If your job families, levels, and pay bands are not documented and gender-neutrally evaluated, start there - before you worry about the calculation.

The consolidation problem

A group with subsidiaries in multiple EU member states will receive national pay gap filings from each entity. The temptation is to aggregate those figures into a group average for the ESRS statement. That approach is almost always wrong, for three reasons:

Different populations. Each national filing covers a different slice of the workforce. Averaging percentages across different-sized populations produces a meaningless number.
Different pay definitions. If Germany's transposition includes a variable pay component that France's excludes, the two figures are not comparable.
Different reference periods. Some national laws may use a different cut-off date or averaging window than the ESRS financial year.

The correct approach is to build one governed pay dataset at employee level - a single source of truth that holds, for every employee: gender, employment type, job category, gross hourly pay, total annual remuneration, country, and legal entity. From that dataset, you derive the national statutory figures by applying each country's methodology to the relevant population, and you derive the ESRS S1-16 figure by applying the ESRS methodology to the full consolidated population. Both outputs come from the same source; they differ because the rules differ, not because the data differs. That explainability is what an assurance provider needs.

Governance and disclosure risk

When a joint pay assessment is triggered - because a category-level gap exceeds 5% and is not remedied within six months - it must be conducted with workers' representatives. In many EU jurisdictions, that means the works council. The assessment must identify the causes of the gap, propose remedial measures, and set a timeline.

The governance risk is this: if your ESRS sustainability statement describes a remuneration policy that is "fair, transparent, and gender-neutral," and your national pay gap filing simultaneously shows an unjustified 7% gap in three job categories, those two documents are in direct contradiction. An assurance provider reviewing the sustainability statement will ask for the national filings. A works council reviewing the national filing will ask for the sustainability statement. The connectivity expectation in ESRS - that narrative and metrics tell the same story - is not optional.

2026-2027 action timeline

Before 30 September 2026 — Audit your data architecture

Map every payroll system, HR system, and data source that holds employee pay records. Identify gaps: can you produce gross hourly pay by gender and job category for every entity in your group? If not, this is the gap to close first.

Before 31 October 2026 — Complete or validate your job architecture

Confirm that every role is mapped to a job family, level, and pay band using gender-neutral evaluation criteria. This is the prerequisite for category-level gap reporting under the directive and for the voluntary breakdown under ESRS S1-16.

Before 30 November 2026 — Run a shadow pay gap calculation

Calculate your gap at category level using 2026 year-to-date data. Identify any categories above 5%. Begin documenting objective justifications now — do not wait for the formal report. Flag any categories where justification is weak; those are your joint-assessment risk.

31 December 2026 — Reference year closes

Lock the employee-level dataset for the full calendar year. Confirm the population, pay definitions, and methodology for both the national statutory filing and the ESRS S1-16 figure. Document any differences between the two and the reason for each.

Q1 2027 — Produce and reconcile both outputs

Derive national pay gap filings for each transposed jurisdiction from the governed dataset. Derive the ESRS S1-16 group figure from the same dataset. Prepare a reconciliation note explaining why the figures differ (population, methodology, scope) — this is the document your assurance provider will ask for.

By 7 June 2027 — File and publish

Submit national pay gap reports to the relevant authority in each transposed jurisdiction. Publish the ESRS S1-16 metrics in the sustainability statement. Ensure the narrative on remuneration policy is consistent with both sets of numbers.

Out of CSRD scope but still inside pay transparency

The Omnibus proposal narrowed the CSRD reporting population significantly. Many companies that expected to be in scope for FY2025 or FY2026 are now out. That is a separate question from pay transparency, and the two scope tests are unrelated.

A company with 300 employees and €40 million in turnover is almost certainly out of mandatory CSRD scope under the revised thresholds. It is squarely inside the pay transparency reporting obligation - 300 employees puts it in the 250+ annual reporting band, with its first report due by 7 June 2027.

For those companies, the practical implication is that the ESRS S1-16 reconciliation problem does not apply - but the job architecture, the category-level gap calculation, the right-to-information process, and the joint assessment risk all do. The directive's obligations are employment law, not sustainability law. They apply regardless of what the Omnibus does to CSRD thresholds.

For companies that remain in CSRD scope, the message is the opposite: do not treat pay transparency as a standalone HR compliance project. The data you build for the directive is the same data your assurance provider will scrutinise in the sustainability statement. Build it once, govern it centrally, and derive both outputs from the same source.

The bottom line

According to Eurostat, the gender pay gap across the EU stood at 12.7% in 2023. Most employers operating across multiple EU member states will have category-level gaps that exceed 5% somewhere in their workforce. The directive gives them six months to remedy an unjustified gap before a joint pay assessment is mandatory. The ESRS sustainability statement will disclose the same underlying number to an assurance provider.

The companies that will navigate this well are not the ones that hire the most lawyers. They are the ones that build a single governed pay dataset before 31 December 2026, run the shadow calculation before the reference year closes, and ensure that the story told in the sustainability statement is the same story told in the national filing. That is a data project. Start it now.