Triaging Pay Structures and Worker Categories Under the EU Pay Transparency Directive

BLOG OVERVIEW: Under the EU Pay Transparency Directive, worker categories must group jobs of equal value, and that value must come from a gender-neutral job evaluation or model built on skills, effort, responsibility, and working conditions. Five triage questions help employers test whether their job evaluation is documented and sound, whether pay structures preserve differences in job value, and whether worker categories rest on assessed value instead of occupation, pay arrangements, market rates, or actual pay. With the first data reporting period closing at the end of 2026, gaps found now can still be addressed before reporting and right-to-information requests test them.

In a previous blog, we discussed defensible worker categories under the European Union’s Pay Transparency Directive (the Directive). To be defensible, worker categories must at least derive from pay structures developed through gender-neutral job evaluation, or from gender-neutral models, either of which assign value to work based on the Directive’s criteria. In this piece, we’ll share five questions employers can use to assess the foundation for their worker categories and identify where further review may be needed.

Timing is of the essence for employers. Although the Directive’s first pay reporting deadline (for employers with 150 or more workers) isn’t until June 2027, the associated pay reference period closes at the end of 2026. Employers can retroactively construct their worker categories after that, but they cannot amend pay decisions made in 2026. Additionally, all employers must respond to worker right-to-information requests in nations with transposed legislation in effect. That requires having worker categories in place.

Employers are not all in the same position when it comes to the adequacy or defensibility of their worker categories. Some have a sound, documented job evaluation reflected in their pay structures and worker categories and may need to do little more than keep that foundation current. Others have a sound job evaluation, but worker categories are based on something other than the value assigned to jobs. Still others who have not gone through the exercise of assigning value to jobs know they have substantial work ahead. And a fourth group believes its existing job architecture and pay structures already satisfy the Directive, and in our experience that belief is not always a reliable indicator of where an employer stands. What each should do next is not the same.

In Part II, our goal is to help employers assess where they stand by examining how job value was determined, what evidence supports those assessments, and whether pay structures and worker categories reflect the results. These five questions can identify gaps and uncertainties that warrant closer review; the answers do not, by themselves, determine the remedy or the amount of work required. In a future installment, we will discuss how those findings can guide next steps.

Five Questions to Help Employers Triage Their Pay Structures and Worker Categories

The first three questions examine the foundation for assigning value to jobs; the last two examine how that value is reflected in pay structures and worker categories. Where the evidence is incomplete, an answer may remain unconfirmed. Employers using different approaches across functions, locations, or acquired businesses may need to consider the questions separately for each population.

Question 1: How is the value of different jobs determined in the organization?

This is the most important question employers need to ask themselves. Every question that follows presupposes an answer to this first question, and an employer that has never assigned value to work on the Directive’s criteria cannot have defensible worker categories no matter how well documented everything else is.

There are many possible answers to this question. For example, pay for jobs may be determined informally as new positions are created. Some organizations may have constructed pay structures that assign jobs to levels based on title and current compensation. Others may have priced jobs against market survey data, letting survey benchmarks determine where jobs sit relative to one another. None of these approaches, however, directly assign value to work based on skills, effort, responsibility, and working conditions, which is what the Directive’s criteria require (Article 4(4)). Market data, in particular, tells an employer what a job costs; it does not establish what that job is worth relative to other jobs in the organization.

Recall that Part I set out the two pathways the Directive points to for assigning value to jobs (Recital 26): gender-neutral job evaluation and gender-neutral models, the latter of which we argue includes vendor job architectures and national or sectoral frameworks (such as Italy’s CCNLs). An employer may rely on either, or on both for different populations. The Directive also appears to grant employers leeway in how they arrange jobs into pay structures, as long as some element(s) of those pay structures delineate jobs having different value. This is not to say employers cannot respond to the market (we’ll address that matter in a future installment).

Question 2: Can the job information, criteria, methods, and results used to assign value to jobs be furnished?

Assuming an employer has conducted a job evaluation or used a vendor model to map jobs to a proprietary job architecture, this question asks whether that information is available. For a job evaluation, this can include: the criteria, factors, and subfactors used to rate or compare jobs; how factors were weighted or combined to produce a result; how the value of new jobs is determined; when the job evaluation was conducted; and who participated in the process. For a vendor model, the information may change slightly but would include: the criteria that formed the basis for the model; the process for mapping jobs to the model; when jobs were mapped; whether the same model is used to map new jobs; and who participated in the mapping process. In either case, the records should also include the job descriptions or other job information used, the resulting ratings or mappings, and the rationale supporting those decisions.

Having this information available is necessary to determine whether the approach for assigning value to jobs is methodologically sound and consistent with the Directive’s requirements. Without knowing what the evaluation factors are, an employer has no way to demonstrate that jobs were assigned value based on the Directive’s required criteria. The methods used to evaluate or map jobs, including factor weightings or who conducted mappings and how, for instance, can inform whether the approach was gender-neutral, which is also a requirement of the Directive. It’s possible that an employer may only have access to some of this information, which makes defending the foundation for worker categories difficult, if not impossible.

Before concluding that something is missing or unavailable, though, it’s worth taking an extra step to track down that information. For example, factor weights can sit in a scoring spreadsheet rather than a methodology document, degree definitions may live inside rater training materials, and key decisions may be recorded in meeting minutes or emails rather than in a signed agreement. Employers who worked with a vendor to map jobs to a framework should ask the vendor directly for the information needed to assess whether the framework meets the Directive’s criteria. Employers shouldn’t assume that material is unavailable without having asked.

Additionally, the age of a job evaluation or mapping to a vendor framework alone is not disqualifying. A 2018 evaluation that has been maintained (e.g., with new and changed jobs evaluated against the same instrument, and a record of those decisions) may be in better shape than a 2023 evaluation that has drifted since. What matters is whether the evaluation still describes the current jobs. Acquisitions, restructurings, and roles that have changed materially since the evaluation can all create gaps. Where jobs are assigned to a level by comparison to an already-evaluated job, employers should be able to produce a documented comparison of the jobs’ content that supports that placement.

Question 3: Is the assessment of job value sound and consistent with the Directive’s requirements?

Specifically, this question asks whether (1) the criteria used to evaluate jobs, or form the basis for a proprietary model, are objective, gender neutral, and include skills, effort, responsibility, and working conditions, as Article 4(4) requires; (2) the methods used to evaluate or map jobs themselves are gender neutral; and (3) the underlying job information, method design, and application provide a sound basis for comparing the value of jobs.

The European Institute for Gender Equality (EIGE) has published guidance for employers on conducting gender-neutral job evaluation. As mentioned in Part I, this guidance includes things like providing gender bias training for those participating in the job evaluation process, involving both men and women in the process, and ensuring that skills associated with female-dominated roles, which are often overlooked or underweighted, are appropriately represented and weighted in the process.

Including all four criteria does not, by itself, establish that an evaluation is sound. Job descriptions and other inputs should accurately capture the demands of the work, rather than the attributes of the incumbent. Factors, subfactors, and assessment levels should be clearly defined; any weighting should be appropriate to the work; and ratings or mappings should be applied consistently and supported by evidence. Weaknesses in these areas can distort the results, including by undervaluing work predominantly performed by one sex. These considerations apply to both internally developed methods and vendor models.

This is likely to be the most technically demanding of the five questions. Employers can identify missing information and apparent deficiencies themselves, but reaching a supported conclusion about the method and its application may require review by someone with job evaluation expertise.

Question 4: Does the pay structure preserve distinctions in the value of work assigned to jobs?

This question asks whether the value assigned to jobs carries through to an employer’s pay structure. With this question, we shift from asking about the job evaluation methods and inputs—the foundation for worker categories—to the outputs. A few things can disrupt the translation of a gender-neutral job evaluation based on objective criteria into a pay structure or make an otherwise gender-neutral model problematic within the context of the Directive.

For a job evaluation, jobs are often arranged into levels or pay grades based on their scores, with jobs having similar scores being grouped together. Article 4(4) requires pay structures that enable the assessment of whether workers are in a comparable situation in regard to the value of work, which in our view means a structure must distinguish among jobs of different assessed value. If other, additional information is used to group jobs, adjust the groupings for some jobs, or create materially different pay structures for certain jobs (e.g., by business unit, job family), pay structures may no longer reflect distinctions in the value of work established by the job evaluation. This often happens when the levels or grades for some, but not all, jobs are adjusted based on market data. Additionally, the creation of very large pay bands from wide ranges of scored jobs may blur or erase meaningful distinctions in the value assigned to jobs.

A vendor’s job architecture may present similar problems. Even if its framework is based on objective, gender-neutral criteria and jobs are mapped to it in a sound and gender-neutral way, a framework with built-in market adjustments by job family or function that affect the levels where some jobs are placed can be problematic. This is particularly the case if those market adjustments favor jobs predominantly performed by men, for example, despite the underlying value of those jobs being equal to those performed predominantly by women.

Where existing pay grades do not preserve those distinctions, a separate mapping of jobs by assessed value may provide the basis for worker categories. What matters is that the groupings reflect equal value and that the employer can show how they were derived, not that the pay structure alone carries them. Employers taking this route should recognize that pay continues to be set by the existing structure, so jobs of equal value may still be paid differently.

Question 5: Are worker categories based on the pay structure, or something else?

What constitutes a worker category under the Directive is a choice for employers—a choice that must be defensible in and of itself. An employer’s approach to evaluating jobs and developing a pay structure may be wholly compliant with the Directive, but it can still run afoul of the Directive if its worker categories don’t group together jobs performing the same work or work of equal value. (As discussed in Part I, we understand both jobs performing the same work and jobs performing work of equal value to fall into the same worker category. And equal-value worker categories comprise jobs the evaluation treats as equivalent in value, whether that equivalence is expressed as a range of scores, a band, or a level.) It’s worth distinguishing here among four bases for constructing worker categories, the first three of which may plague employers.

I. Work and Pay Arrangements

Some features of jobs that reflect certain work or pay arrangements may seem like they should be used to define worker categories because they often relate to pay. This could include things like contractual hours, shift schedule, work location, bonus plan eligibility or participation, commission eligibility, or job-specific market premiums. These features alone do not establish differences in job value and should not factor into the creation of worker categories. Where an arrangement changes the actual demands or working conditions of a job, those differences should be assessed through the evaluation method. However, such features, if they are objective and gender-neutral, might be considered justification factors for any pay gaps that do appear within a worker category.

Understandably, this may frustrate employers because worker right to information figures (Article 7(1)) and reported pay gaps by category of workers (Article 9(1)(g)) do not account for any possible justification factors that explain why workers who hold the same jobs or jobs of equal value might have different earnings. The optics are the same whether a gender difference in pay is explainable. A compliant employer with a documented reason for a gap reports the same number as one with no reason at all. The reported figures do not tell the whole story. The explanation enters later, when a worker or workers’ representative asks for clarification, or when a reported gap triggers an obligation to justify it.

II. Occupational Similarity

Occupational similarity involves creating worker categories from jobs with tasks that are similar in nature. In practice this process looks like grouping by job family, job function, discipline, or occupational group—often crossed with a job level or pay grade. It’s probably the biggest trap for employers, particularly those used to practices in the United States.

Jobs do not have to resemble one another to be of equal value. Job evaluation already considers the content of the work by assessing its demands against the Directive’s criteria. Grouping jobs by occupation afterwards puts boundaries back between jobs that the evaluation may have established are of equal value. Those boundaries can conceal the very pay differences the comparison is meant to reveal—and that the Directive specifically aims to eliminate.

The pull toward occupational groupings is strong for US-based employers, and for understandable reasons. Pay equity analyses under Title VII often group similarly situated employees by job or by job family plus level. Market surveys for compensation are commonly organized by occupation or job family. Grades built from survey data inherit that architecture. And at the national level, there is no comparable equal-value standard in the US.

Occupational similarity comes a little closer to the mark where jobs are identical, but even then, the test is comparability across the nature of the work, training requirements, and working conditions, not task similarity alone.¹ So, it’s not entirely clear that worker categories based on job family and level, for example, would satisfy a “same work” definition. For work of equal value, occupational similarity is the wrong dimension entirely.

III. Other Factors Unrelated to the Work Performed

This grouping basis rests on nothing about the assessed value of the work but perhaps instead on something that may seem reasonable to an employer. Grouping workers by cost center, for example, or by headcounts to reach (or avoid) reporting thresholds would qualify as arbitrary, which the Directive’s definition of a category of workers excludes: Article 3(1)(h) requires that workers be grouped “in a non-arbitrary manner” on the criteria referred to in Article 4(4).

Perhaps the most egregious example is grouping jobs by actual worker pay and not on the assessed value of the work being performed. This approach is circular: the categories are defined by the exact variable they exist to test, so any difference between jobs of equal value is absorbed into the category boundary rather than appearing within it. It can produce near-zero pay gaps across groupings that an objective evaluation would find to be of materially different value. In our view, this circumvents the Directive rather than complying with it, though it may be a tempting approach for employers with no formal pay structure to fall back on.

Grouping jobs based on market rate is similarly arbitrary under the Directive: market rates reflect what employers pay for labor under prevailing supply, demand, and bargaining conditions, rather than an assessment of the work against the Directive’s criteria. Jobs that command similar pay in the market do not necessarily have equal value, and jobs of equal value may command very different pay. This is particularly problematic where market rates reflect systemic undervaluation of work predominantly performed by women (or men); using those rates to construct worker categories can embed that undervaluation in the very groupings intended to reveal it. This is not to say that market data has no place in an employer’s compensation practices, only that it cannot supply the basis for worker categories. We’ll take up how employers can respond to the market while complying with the Directive in a future installment.

IV. Value

Worker categories should comprise jobs based on their assessed value, according to the Directive’s criteria, and really nothing else. That is a single constraint, and it leaves more open to choice than employers often assume. The Directive does not tell employers what counts as skill, effort, responsibility, or working conditions; how finely to differentiate within each; or how much weight any of them should carry (Recital 26). Nor does it prescribe a pay structure design. Employers using a vendor model choose the vendor and often have some latitude over which factors are assessed and how they are weighted.

What the Directive constrains is the basis for grouping jobs: assessed value. Employers define what value means in their organization, subject to the requirement that the definition be objective, gender-neutral, and include all four criteria. Risk enters at this step when categories are built on some basis other than the value the evaluation produced—even when the evaluation itself was sound.

Table 1. Four Bases for Defining Worker Categories

Considerations for Employers

The five questions identify an employer’s first unresolved issue, not the remedy. What follows from the answers depends on what a closer look reveals.

The first three questions establish whether a usable foundation for worker categories exists; the last two ask whether what was built upon it holds. The answers can diverge, and an employer with a sound job evaluation but poorly drawn categories needs something quite different from an employer with neither.

Employers should consider having their job evaluation criteria, methods, and results reviewed for consistency with the Directive and applicable national requirements, as well as sound professional practice in job evaluation. That review may draw on internal or external job evaluation expertise, with legal input where needed. It should also examine whether job descriptions and other supporting information are accurate, current, and sufficiently complete to support the assessments. A well-designed method cannot compensate for inaccurate or incomplete information about the work.

In a future installment, we’ll map the findings from these questions to potential next steps for employers and discuss what could change those steps or their sequence. This includes whether an existing evaluation can be revised, whether revised results support current structures and categories, and how Member State requirements, national or collective frameworks, and the employer’s size and footprint may affect the options. We’ll also discuss timing and how employers can accommodate market forces while maintaining defensible equal-value groupings.

¹ In Angestelltenbetriebsrat der Wiener Gebietskrankenkasse v. Wiener Gebietskrankenkasse (C-309/97, 1999), para. 17, the Court of Justice of the European Union held that whether two groups perform the same work turns on whether they are in a comparable situation, taking account of the nature of the work, the training requirements, and the working conditions—not on task similarity alone. ↩