A guide for HR and reward teams, works council members, and compliance officers at employers with staff in the EU. It explains what Directive (EU) 2023/970 requires at clause level: the gender pay gap indicators employers must calculate and report under Articles 9 and 10, who must act and by when, how each figure is computed, and where the UK calculation guidance differs. All arithmetic runs on a synthetic eight-person dataset so you can verify every number by hand.
Directive (EU) 2023/970 of 10 May 2023 strengthens the application of the principle of equal pay for equal work or work of equal value between men and women. It applies from 7 June 2026 (Art. 35), and member states had to transpose it into national law by 7 June 2026 (Art. 34(1)). It layers three duties on employers: pay transparency at hiring and on request (Arts. 5 to 7), gender pay gap reporting (Art. 9), and a joint pay assessment with worker representatives where a gap stays unexplained (Art. 10). This page concentrates on the two reporting duties.
Employers must publish the following seven items for each legal entity or group of entities (Art. 9(1)(a) to (g)):
| Item | What it measures | Clause |
|---|---|---|
| Overall gender pay gap | The difference in average pay levels between female and male workers, expressed as a percentage of the average pay level of male workers. | Art. 9(1)(a), Art. 3(1)(c) |
| Gap in complementary or variable components | The same calculation restricted to components paid in addition to ordinary basic wage or salary, such as bonuses or overtime pay. | Art. 9(1)(b) |
| Median gender pay gap | The same overall gap computed on median pay levels: the pay level at which half of the workers earn more and half earn less. | Art. 9(1)(c), Art. 3(1)(d) and (e) |
| Median gap in variable components | The median-based version of the variable-components gap. | Art. 9(1)(d) |
| Share receiving variable pay | The proportion of female and male workers receiving complementary or variable components. | Art. 9(1)(e) |
| Quartile distribution | The proportion of female and male workers in each quartile pay band: the four equal groups of workers divided according to their pay levels, from lowest to highest. | Art. 9(1)(f), Art. 3(1)(f) |
| Gap by category of workers | The gender pay gap broken down by categories of workers performing the same work or work of equal value, split by basic salary and by complementary or variable components. | Art. 9(1)(g) |
Three procedural clauses frame the numbers. Management must confirm the accuracy of the reported figures after consulting workers' representatives (Art. 9(6)). The report goes to the member state's monitoring body, which may publish items (a) to (f) (Art. 9(7)). Workers and their representatives may request and receive clarifications on any item, and an employer that cannot objectively justify a gap must remedy it within a reasonable time (Art. 9(10)).
The joint pay assessment is the escalation step. Where a pay report shows a difference in average pay levels of at least 5% in any category of workers (Art. 10(1)(a)), the employer must run a joint pay assessment with worker representatives unless the difference is justified on objective gender-neutral criteria (Art. 10(1)(b)) or remedied within six months of the report's submission date (Art. 10(1)(c)). The assessment itself is a structured document: the overall and median gaps, the gap by category, the distribution of male and female workers across quartiles within each category, and the gaps per quartile (Art. 10(2)(a) to (g)).
Deadlines scale with employer headcount. All dates below are set directly in the directive; currency note as of September 2026, with the primary source linked in the Sources section.
First report by 7 June 2027 covering the previous calendar year (2026 pay data), and subsequently every 12 months.
Art. 9(2)
First report by 7 June 2027 covering the previous calendar year, and subsequently every 36 months.
Art. 9(3)
First report by 7 June 2031 covering the previous calendar year, and subsequently every 36 months.
Art. 9(4)
Reporting is voluntary at this size under Article 9(5), and member states may still require it, so check the national transposition.
Art. 9(5)
Two scope notes. First, obligation and enforcement differ: the directive binds member states to transpose by 7 June 2026 (Art. 34(1)) and applies from 7 June 2026 (Art. 35), while penalties for infringements are defined nationally under Art. 23, so enforcement posture varies by jurisdiction. Second, the UK is outside this regime: its separate gender pay gap reporting duty (England, Scotland and Wales, 250 or more employees) continues under the Equality Act 2010 (Gender Pay Gap Information) Regulations 2017, SI 2017/172. The UK case matters here as the contrast case for calculation conventions, below.
Every percentage is a male-baseline gap: the male figure minus the female figure, divided by the male figure, times 100. The worked example below uses eight synthetic workers so each step is checkable by hand.
Compute average pay separately for female and male workers and take the male baseline. Using average hourly ordinary pay:
gap_mean = (mean_men − mean_women) / mean_men × 100
Article 3(1)(c) fixes the same male-baseline convention for the reported gender pay gap, so the sign and denominator are settled in the primary text.
Article 3(1)(d) defines the median pay level as the level at which half of the workers earn more and half earn less. With an odd count of workers that single middle value exists. With an even count the definition under-determines the figure, and implementations differ. The UK GOV.UK guidance resolves the tie by averaging the two middle values (Calculations 3 and 6 of the Making your calculations guidance, updated 21 May 2026). A second defensible convention takes the next-highest value instead, the upper of the two middle order statistics. Both satisfy a reading of the definition; they produce different published numbers, so an employer should state which convention its methodology uses. The worked example quantifies the divergence.
Article 3(1)(f) defines each quartile pay band as one of four equal groups of workers divided according to their pay levels, from lowest to highest. Sort ascending, split into four equally sized groups, and report the sex mix of each. The UK guidance sorts from highest to lowest instead and distributes remainders when headcount is indivisible by 4 (one leftover goes to the lower quarter; two go to the lower and upper middle quarters; three spread across the lower three quarters), details worth knowing because transposed national methodologies may inherit either ordering.
Women (4): 17.00, 19.00, 21.00, 25.00 (mean 20.50). Men (4): 20.00, 23.00, 26.00, 32.00 (mean 25.25). Bonus pay in the reference year: 3 of 4 men received a bonus (1,500; 2,000; 3,000) and 2 of 4 women received a bonus (1,200; 1,800).
(25.25 − 20.50) / 25.25 × 100 = 18.81%
Sorted women: 17, 19, 21, 25. Sorted men: 20, 23, 26, 32. With four values each, the middle pair is the 2nd and 3rd. The UK convention averages the pair; the next-highest convention takes the 3rd value.
| Convention | Median men | Median women | Median gap |
|---|---|---|---|
| UK GOV.UK (mean of middle two) | (23 + 26) / 2 = 24.50 | (19 + 21) / 2 = 20.00 | (24.50 − 20.00) / 24.50 × 100 = 18.37% |
| Next-highest (upper middle value) | 26.00 | 21.00 | (26.00 − 21.00) / 26.00 × 100 = 19.23% |
Same inputs, same clause, two published numbers 0.86 points apart. This is the convention risk the directive leaves to implementers.
| Band | Members (EUR/h) | Sex mix |
|---|---|---|
| Q1 (lowest) | 17.00 W, 19.00 W | 100% women |
| Q2 | 20.00 M, 21.00 W | 50% men / 50% women |
| Q3 | 23.00 M, 25.00 W | 50% men / 50% women |
| Q4 (highest) | 26.00 M, 32.00 M | 100% men |
Share receiving a bonus: 75.0% of men (3 of 4) against 50.0% of women (2 of 4), reported under item (e). Mean bonus among recipients: men 6,500 / 3 = 2,166.67 against women 3,000 / 2 = 1,500.00, so the mean bonus gap is (2,166.67 − 1,500.00) / 2,166.67 × 100 = 30.77%. The median bonus gap diverges by convention more sharply: male median bonus is 2,000; the female median is 1,500 under the UK convention (gap 25.00%) and 1,800 under next-highest (gap 10.00%).
The category's mean gap of 18.81% clears the 5% threshold of Art. 10(1)(a). If the employer cannot show the difference is justified on objective gender-neutral criteria and has not remedied it within six months of submitting the pay report, a joint pay assessment with worker representatives is mandatory, covering the item breakdown of Art. 10(2).
Processing pay data for Articles 7, 9 and 10 is processing of personal data, and the directive keeps the GDPR framework intact: Regulation (EU) 2016/679 applies in full (Art. 12(1)). The employer is the controller for that processing. If payroll rows leave the employer's infrastructure for an external service to compute, that service acts as a processor and a data processing agreement under GDPR Article 28 is required. Article 12(2) restricts processing of the data collected for the purposes set out in those articles, Article 12(3) requires that reports let the monitoring body check compliance while preventing the identification of individual identifiable workers, and Article 12(4) bars collecting additional personal data just to build the report. This is why the tools and examples on this estate run on synthetic inputs in your own browser: the arithmetic never needs real rows to be demonstrated or rehearsed.
Browser-based calculators on this estate already cover adjacent payroll arithmetic, all client-side and zero-PII. A dedicated Directive 2023/970 indicator calculator is not yet part of the live estate; this page documents the required computations so the definitions above can be rehearsed and checked meanwhile.
Compute gross-to-net payroll lines with FICA deductions step by step. The ordinary-pay layer the Article 9 indicators sit on, in browser form.
Open ToolRegular-rate and overtime arithmetic under the US FLSA, including the variable-pay weighting that parallels complementary-component gaps.
Open ToolRecompute certified payroll and prevailing-wage lines from raw inputs. Useful rehearsal for per-category pay aggregation under item (g).
Open ToolEmployer-affordance percentage arithmetic against safe harbors. A worked example of threshold logic akin to the Article 10 5% trigger.
Open ToolPrimary text retrieved 2026-09-25. Compliance dates on this page are obligation dates from the directive; enforcement is national under Article 23.
This page is an explainer for orientation and methodology rehearsal. It is not legal advice; confirm obligations for a specific employer with qualified counsel in the relevant member state. Currency of all clause references: as of September 2026.
v1.0 · September 2026 · Directive (EU) 2023/970 · Guides