Why The EU Pay Transparency Directive Is More Than A Compliance Exercise

David Rice: The European Union Transparency Directive isn't a reporting exercise. It's an operating model overhaul. But most organizations are only looking at the visible part of the iceberg, so to speak. On today's show, I'm talking with Tom Heys, he's a pay reporting expert at Lewis Silkin, about why he describes the PTD as the UK gender pay gap regime on steroids.

Because unlike previous reporting requirements, this one combines company-wide gap reporting with individual employee rights, new restrictions on what you can ask during recruitment, and the requirement to categorize jobs of equal value and defend those decisions to employee representatives. The last part is where a lot of organizations are going to struggle.

Accuracy and explainability are not the same thing. An answer can be statistically correct and still be impossible to defend. And this legislation doesn't just ask for answers, it asks for you to show your work. If an employee representative turns around and says, "Show us your model," and your black box generated the number, you're exposed.

Tom's broader warning is one that extends well beyond compensation. We're entering an era where leaders are making decisions partially influenced by systems they didn't build and don't fully understand. That shifts explainability from a technical issue to a leadership capability. So today, we're gonna cover what makes the PTD fundamentally different from prior pay reporting requirements, why job architecture is the foundation everything else sits on, the gap between documented work and what people actually do, why AI can accelerate this process but can't own the decisions, and why explainability is becoming as important as efficiency.

I'm David Rice. This is People Managing People, and if your organization is treating pay transparency as a compliance checkbox, this conversation shows you how much work is actually underneath that. So let's get into it.

If you're running a business, you probably didn't start it because you love bookkeeping. That's where QuickBooks Online can help. It automates everyday tasks like invoicing, expense tracking, and bookkeeping, so you spend less time on admin and more time growing your business. It's a simple way to stay organized and keep your finances working for you. If you're looking to save time and simplify your business finances, check out QuickBooks Online.

Tom, welcome to the show. It's good to have you.

Tom Heys: Thanks for having me on the show, David.

David Rice: When we were talking before this, you described the EU Pay Transparency Directive as sort of the UK regime on steroids.

What makes this different from previous pay reporting requirements, and, and why do you think so many organizations are underestimating the amount of work that's involved with it?

Tom Heys: I came at the, the PTD from a background of having done a lot of work over the past few years for UK companies on UK gender pay gap reporting.

That is at, at a high level. It's not comparing like with it's really more of a, a representation gap that you're looking at there. The PTD starts from a similar place in that you also have to report those organization-wide gaps. But then it's really much more focused on a bigger issue, equal pay, and it does that through reporting obligations, but also additional reporting obligations compared to what you, you see in, say, the UK.

So really looking at jobs that are comparable, comparing like with or at least like-ish with and then on top of that, it combines it with new individual rights as well as new bans on things, what you can and can't ask about during recruitment. So the approach that seems to have been taken is equal pay is a fundamental EU right.

It's part of the treaties, and how can we hit it from every single direction to try and ensure equal pay as much as possible? And that seems to be the sort of guiding philosophy behind how the PTD has ended up the way that it has. So it's a really big new piece of legislation, and I think it will be very disruptive over the next one to three years.

Kick up a lot of dust, and it will take time for that dust to settle.

David Rice: It seems it's not really as much a reporting exercise as maybe it is an operating model exercise.

Tom Heys: I think that's right. I think it's bringing in a new way of doing things, a new normal. And so it really forces a lot more discipline around things like pay setting and how you do and how you don't do it, what you can take into account, what you shouldn't take into account.

You know, ways of doing it to ensure that, that bias doesn't creep in. So, for example, I mentioned some of the, the new obligations around recruitment. So you won't be able to ask people their pay history, so you can't say to people, "What are you on now?" You can't ask that as part of recruitment. And I think for so many companies, I mean, that's just the standard way in which things have been done for a very long time.

So even on just that one very particular thing, it's a big change.

David Rice: Yeah, it is. The reporting part feels like the visible part of the iceberg, so to speak. You know, it's underneath there you've got like job architecture, leveling compensation, philosophy, governance, all these different things In some ways it feels similar to what's happened with AI readiness, you know, where everybody wanted outputs, but they had to fix the foundations first.

And I think it sounds to me like this might be something similar to that.

Tom Heys: I think you're spot on. The iceberg example wasn't something that I'd thought of before, but I think it's really appropriate. And you mentioned the job architecture. I mean, that underpins the two biggest things that companies seem to be most concerned about.

So in the, the PTD, it talks about categorization, so grouping people into jobs that are of equal value. So in terms of skill, effort, decision-making, responsibility rather and working conditions, grouping those jobs together. And then firstly, all employees have a right to the average pay of men and women within their category.

So straight away, that new right, which is underpinned by all of the really important job architecture stuff, that gives people the information that they need to be able to i-identify potential equal pay issues. 'Cause they can see, "Hang on, I'm only getting paid this much. All of the people doing similar enough jobs to me are getting paid that much, so I wanna know what's going on there."

And then secondly, the reporting bit that you mentioned. So there will also, as well as the, the broad company-wide gaps, the, the representation type of gaps, there will be a requirement to report i-internally the gaps within each category. And so you've got that combination of obligations on employers and also rights on individuals, and it's really putting all of the information out there for what people need to be able to bring equal pay claims.

And as you say, job architecture, the categorization bit sits right underneath both of those two really big things.

David Rice: One of the themes that kept coming up when we spoke before this was explainability, right? And this is even something I just wrote about when it comes to hiring screens. You know, how well can you explain what AI does, for example?

And the AI can produce answers very quickly, but this legislation doesn't just requires answers, it requires organizations to be able to defend them. And I'm curious, why would you say that distinction is so important?

Tom Heys: I think it's important because of the role the, the legislation envisages employee representatives have.

It's a really powerful role. They have rights and obligations, and they can, in some circumstances, effectively the arbiter of whether gaps are justified or not. They get access to methodologies and how things are done, and they're a human being, and they get to ask questions, and some of those questions might be, you know, perfectly logical and sensible, and some of them might come from positions of not so informed, because this is a new thing for lots of people.

It will take time for people to- upskill themselves on this. Because you've got to explain to a human, you've got to really know why you've done things in the way in which you have. What is the process that you've taken? Because someone will ask you about it, they can ask you about it, and it's almost certain that at some point they will ask you about it.

And unless you know why you've done all the things that you've done, it's not gonna stand up to scrutiny. The end result may be perfectly good and sensible, but unless you know how it's been done, then it leaves some underlying potential risk, a hostage to fortune for the future.

David Rice: One of the things I was thinking about as I was writing that article is accuracy and explainability are...

They're not the same thing. They're not even necessarily in the same category, and it's one of the two things can be true thing, right? An answer can be statistically correct but impossible to defend. And so, a lot of the AI conversations I have right now, it's like people talking about moving from output focus to outcome focus and The idea, I agree with it, but have warn against becoming outcome obsessed and because you still have to pay attention to reasoning, right?

Like, how did you get there? You're in-- We're in this time where trust is very low, and I'm like, trust isn't gonna be built because you have an answer. It's gonna be built because people understand how you do what you do or think or reason to make a decision. And so this is very much in that same lane.

Tom Heys: I think there's definitely a role for AI within all of this to help employees get ready because things like job architecture, it's a massive job.

Anything that can accelerate that process, particularly for lots of employees coming at this from a no architecture point, this is a whole new world that a lot of companies are entering into. So anything that can accelerate that process, then great. But I think there is a key difference between using AI to accelerate your decision-making or to improve your processes, but I think there's an important line between then deferring to it to make the decisions.

The decisions have to be yours.

David Rice: I think there's an obvious temptation here to, you know, you wanna... I think if you're in this position, you wanna upload the job descriptions, maybe run them through an AI model, and sort of accelerate job evaluation. But the thing about that is job descriptions often reflect what a role was supposed to be and not what it became or not what people actually do.

I'm curious, in your opinion, like, how big is the gap between documented work and real work?

Tom Heys: I think it can really vary, and it can vary by different types of organizations. I think particularly in smaller organizations where things are less rigid in terms of the job that you're doing, and also in things like startup, scale-ups, where roles can change very quickly, and the paperwork might not match up with what someone is actually doing.

The classic example of garbage in, garbage out is very relevant here because the AI is only as good as what information is going into it. And if you're feeding it out-of-date stuff, I mean, it can accelerate your processes and, and come up with some output, but that output is not gonna be... I mean, it's obviously, it's not well-informed because it's using out-of-date stuff, and so potential issues caused there.

David Rice: I hear this quite a bit, you know, the scope creep problem around a lot of people's roles is like, well, we hired Janet to do this But in actuality, as we got her into the company, and of course, she's been here three years, her role has now expanded out into all these other areas, and our documentation doesn't necessarily reflect that.

And then that context then is missing about actually the impact of Janet's role and all the different things that she does and all the things that have fallen into her purview. We talk all the time about how HR's role, for example, is ever expanding. It just keeps getting bigger and bigger. It's been that way for years.

But it doesn't necessarily get reflected back in the same way. So I think there's gonna be some issues there.

Tom Heys: Yeah. I wonder whether you might see perhaps employees being a bit more proactive as to ensuring their job descriptions are matching up with what they're actually doing. I can foresee people emailing HR and saying, "Well, this is what my job description said I should be doing, but I'm actually now doing X, Y, and Z, so that needs to be reflected."

David Rice: I would agree, and I think a lot of people are becoming more and more aware of how this stuff works. If they're aware that the technology will be used in that sense, then I do think they'll want that.

Tom Heys: That's certainly gonna increase as well. I mean, certainly across the EU, you've got two hundred and fifty thousand companies that the directive is applying to.

It's tens of millions of people, and now suddenly they're, they're hearing things about categorization and job levels and what all of that means. And I think there will become much more awareness of it. And as I said, it brings in new rights for employees as well. So it's sort of pushing people into a, a stronger bargaining position.

Maybe that's not quite the right word, but certainly pushing them into being at least greater advocates for themselves by giving them the rights that, you know, that they need to be able to do that.

David Rice: Quick pause. How's your cash flow looking right now? If you're a freelancer or small business owner, having a clear view of your finances can make all the difference. QuickBooks Online gives you real-time financial insights, helps you create and send professional invoices, and makes it easier to get paid faster. Plus, it keeps your income and expenses organized in one place, so you're always on top of where your business stands. Learn more about how QuickBooks Online can help you manage your business with confidence.

Yeah. We've had guests on this show that talk about the importance of documentation, and I think we're gonna continue to talk about documentation practices for a while longer because context with AI, in some ways, it's everything. A lot of vendors are positioning AI as sort of this quick path to compliance, but if employee rep or even employee asks why a pay gap exists, and the answer is essentially, you know, "The system said so.

The system flagged it," that's probably not gonna be enough. That's not gonna fly. I'm curious, do you think organizations are at risk of creating new compliance problems while trying to solve old ones?

Tom Heys: I think there is the potential for that, and I think the problems will come to light because, as I said, the role of the employee representatives is envisaged as being a powerful one, and there'll be a lot of publicity around the new rights as they come in.

It's supposed to all come in on the seventh of June, but there's a lot of fragmentation around implementation, so there's a large number of countries that will be implementing late and perhaps most around January 2027. But that aside, there's lots of publicity around the Pay Transparency Directive. The more people know, the more they will be asking of their employee representatives to hold their employer to account.

And so if companies don't have decent explanations for how they're doing things, then absolutely it's gonna come out.

David Rice: It's interesting 'cause we've seen this pattern w- everywhere with AI, right? Organizations want speed, and then that speed can sometimes create a new category of risk. And the temptation is, well, we should automate complexity away, but this to me sounds like a situation where the explanation's actually more important than what the answer is.

Tom Heys: I think so because the directive is... it's all about transparency, obviously, but transparency about the way in which things are being done. It's not dictating a way of doing things. It's not saying pay everyone exactly the same thing. Differences are fine. Differences are expected. There will obviously always be differences in pay, but What it's trying to get to is a point at which those differences are objective.

They're explainable. They're not for biased, subjective, unfair, illegal reasons, but reasons that make sense, that are proportionate. So the fact that someone has more experience than another, yeah, okay, yeah, that's a good reason why someone might be getting more pay than another. If that experience is, you know, someone has 30 years experience in a particular role and someone has five, and you get as good as you're ever gonna get in this particular type of job when you've been doing it for three years, you're still paying the 30 years experienced guy a lot more, then you start getting into, to issues, because whilst you start with a potentially objective reason, you're not applying it in a proportionate way.

And, you know, these nuances, they will come out over time and, you know, as I say, it's shining a light. It's exposing the way in which things are being done.

David Rice: Yeah. It's interesting 'cause historically, I mean, well, at least in the US, so correct me if I'm wrong about in Europe, but, you know, compensation, it's been one of the least transparent parts of organizational life for a long time What do you think happens when people suddenly have the right to ask all these questions that a lot of organizations have never really had to answer before?

Tom Heys: Well, I think that the tendency to not talk about money is the, is something we see this side of the pond as well, for sure, and certainly across a lot of... Not exclusively. There are some countries where people are not afraid to say, "I get this much. How much do you get?" But in a lot of places, there is certainly a tendency to shy away from talking about the money.

But in terms of what happens, so what will the outcome be once people get all of these rights and once companies have all of these obligations and the sun is shining down on all these ways of doing things? Well, I think the UK offers a very cautionary tale because about 10, 15 years ago, we had a lot of claims against local authorities, councils across different boroughs and cities across the country.

And these were claims being brought, not by people doing the same job, but by people, say, mostly female-dominated jobs, like cleaners and carers, comparing themselves against mostly male jobs. So, street sweepers, gravediggers, these are all the classic examples, mostly male jobs. The mostly male jobs that had large bonuses, which became de facto part of their pay, they weren't motivating extra productivity or performance or rewarding anything.

It was a de facto part of their extra pay. And the amount of these bonuses was very big. And in some places, it was multiples of the pay of women. That was the extent of the difference. This all came out, though, 'cause of the nature of the sector. So public sector, freedom of information applies. People could ask, "What's the average pay for these types of jobs and those types of jobs?"

And also, it had all been negotiated through trade unions. The information was out there. People knew that there were these pay differences and there were these different bonuses. And then what happened? Well, obviously, the underpaid people wanted their pay to be corrected, and so Claims were brought, big multi-party claims hundreds, thousands of people in some cases, and the impact of that was massive.

So Birmingham City Council, one of the largest employers in the UK, possibly the l- largest local authority in Europe as well, something like a million people very big, their equal pay liability was huge. And we're talking I think it was over a billion. And so I think that just shows the consequence of what happens.

The pay differences in, in these cases were quite often big, but it doesn't even have to be a very big pay difference, particularly when you're talking about a sort of workforce where there are lots of people. And what we've seen in a sort of similar situation is retailers in the UK have faced large multi-party claims.

The differences there have been smaller, but because you're talking about tens of thousands of employees, forty hours a week, you can get six years back pay. The amount of the difference doesn't have to be very much, you know, two, three pounds here or there. But when you're multiplying it by so much, the extent of the liability becomes massive.

And what those sorts of cases are, are looking at is, say, shop floor people comparing themselves against warehouse people. Different jobs, jobs which when looked at through the equal value lens, so the amount of skill, effort, responsibility, working conditions, in exactly the same way the Pay Transparency Directive tells employers to look at their jobs.

What happens is that the, the differences become exposed, people bring the claims, and I think there will be a lot of those types of claims spreading across Europe In the UK, you can claim six years back pay. I know that in other parts of Europe, in parts of the EU, it's longer. In some places it's unlimited.

So someone who's been there thirty years underpaid could get their thirty years back pay. So I think that there will be a lot of those sorts of claims being brought over the next few years.

David Rice: It's interesting 'cause there's been-- it's sort of the the bill's come due for sort of a trust us culture around compensation is what it feels like, right? So-

Tom Heys: Yeah ...

David Rice: now we're moving into this like show us environment, and I imagine that's gonna be pretty uncomfortable for a lot of organizations that never really had to articulate a compensation philosophy in any sort of plain language externally. I think that's gonna be fascinating to see.

Tom Heys: Yeah. I mean, I, I think of it in you know, maths exams when you're at school.

You get marks for getting the right answer, but you get more marks for showing your working. And showing your working is where, i-is where the real substance is here.

David Rice: That's a great way to put it, actually. It sounds like a lot of organizations are sort of discovering they can't comply with pay transparency unless they first clean up the job architecture, the grading, the leveling systems.

Would you say this legislation is really a sort of a compliance challenge, or is it Really getting at exposing some deeper organizational issues that have been ignored for a long time.

Tom Heys: I think it's both. I think the deeper issues cause the equal pay risk, and the, the PTD is surfacing that risk, so it's putting everything out in the open.

And so as a result of that, the, the compliance problems become apparent. I don't think that the PTD should be thought of in terms of a compliance tick the box exercise. We've done it now, we don't need to worry about that anymore. I think it's really about throwing away the old way of doing things and replacing it with a new normal.

I think it's going to be incredibly disruptive. A lot of organizations, it will, as you mentioned, kick up a lot of dust. When the dust settles, there'll be a new normal, normal 2.0, which will be a, a more objective, clearer way of pay setting, more honest potentially. Like you said, it's not just, "Trust us, we're, we're doing it right," but, "Show us."

And you know, it's not just trust us, it is showing as well. So I think that's the way in which it should be thought of. It's ushering in a new era. It's not a tick the box job.

David Rice: Yeah. It feels almost like a, like a organizational MRI, so to speak. Re- sort of revealing structures and inconsistencies that may have been around for a long time and, you know, it kinda looks like compensation problems on the surface, but I really think it's more of a governance problem than anything.

Tom Heys: No, I think that's a sensible way of putting it. 'Cause, I mean, ultimately it is all about how much people are paying, but where the problems creep in is where you get things like too much manager discretion, I suppose, and not enough guide rails on how things are done. So if you have managers without a clear way in which they should be awarding bonuses, say, or making use of whatever budget they have to reward their team, then you know, you've got unexplained things going on which has a pay impact, which then has a, an equal pay potential risk.

David Rice: You've mentioned the reporting obligations under the PTD. I'm curious What are the issues in, you know, using AI to solve X? Like, whatever it is.

Tom Heys: Yeah. So I think this is relevant in terms of the, as I mentioned before, reporting of gaps by category. So employers have to explain any gaps which within a category are greater than five percent.

And we're gonna need to get into the maths a bit for me to, to answer this question. So you can explain your gaps by reference to objective reasons. So potentially experience, like I mentioned before, but maybe also things like performance, potentially, you know, market factors, lots of different potential reasons.

So the objective really is to say, "Okay, we know that we have an eight percent gap, say, within level ten. But we know, thanks to our analysis, we can see that actually five percent of that can be explained, which only leaves three percent left unexplained." There's a threshold of five percent for gaps within a category.

The PTD sets this out. And where the unexplained gaps are greater than five percent, it causes a requirement to undergo a joint pay assessment, essentially an equal pay audit, opening up all of your books to employee representatives. They get to see how you do everything. Big distraction away from actually being productive and trying to explain how you're doing things.

So if you can avoid that, then great. So the standard way of being able to derive the amount of explained, unexplained gap would be through regression analysis. So at its simplest, it's, it's drawing a line through a set of points, the line that fits the best But the thing is that this sort of analysis regression is subjective.

It's not objective. We're not talking really about two plus two equals four. My analogy is that it's a lot more like mixing paint. Think red plus green equals brown. So you know how much of each color are we talking forest green and crimson or lime and scarlet? How thick are we putting the paint on?

We're using oils or watercolors. In the same way you can think of regression analysis. How much of each variable? What's the model specification? Which variables are we considering here? Lots of different decision points, and it's not the case that there is a right way of doing it for any particular category that you're looking at.

There are degrees of right, I suppose, degrees of defensibility, because as we've talked about, you've got to explain it all to an employee representative and there might be different models that work, then there might be other ones that also reasonably work. And so your choice is which ones are you gonna use, which ones are you gonna not use?

The risk is that when you defer this sort of analysis to AI, and we're talking really here about analytical or predictive AI rather than the sort of generative AI. But when you say to a black box, I wanna know how big my adjusted gap is, and it spits out an answer, and you then tell your employee representatives, "Well, it's three percent, so it's fine."

And they turn around and say, "Well, you know, show us your model, show us your working." And that's the point at which if your black box has, has done it in a not very defendable way, so maybe it's an overspecified model, maybe it has too many variables that doesn't match the... You know, I think it's inevitable that all employers will have at least one small group that they have to look at.

And when you start getting into small groups, regression stops working well because you don't have enough data to rely on. There's not enough points to be able to draw the line through in any clear or reliable way. And unless you know that, unless you can see that, then you can't rely on, on any output as being something that would stand up to scrutiny.

So I'd exercise some caution around the use of, of automating the analytical process, not just looking at the answer at the end of the maths paper, but also un-- making sure you understand the working as well.

David Rice: The old garbage in, garbage out problem, right? If your data that you've got that you're, you're doing this analysis based on is, is obviously a risk.

And then there's a risk that like, you know, a lot of times we put things in and we feel more certain than we should based on the output. You know, you get a polished answer, whether it's a, a chart or a recommendation or whatever it is, it sort of feels authoritative But if the underlying assumptions and judgments and decisions that somebody has to own are based on something that isn't real...

I always talk about this. I'm like, "You can't hold a machine accountable, right?" So it can accelerate your analysis, but it's not gonna absorb your accountability at the end of the day.

Tom Heys: The machines are doing what you ask them to do, but we need to make sure you're asking them to do the right things.

David Rice: That's exactly it, and we've got to give it the right context, not just, "Oh, I fed it a bunch of data." It's like, "Cool. Was any of that meaningful?"

Tom Heys: What data have you fed it, and what's it done with it? It's like you throw a ball for the dog, and it brings it back. The machine brings the data back, but, you know, you, you don't know whether it's brought the ball back by going out of the park and down the street and round back the other way or whether it's brought it straight to you.

David Rice: Exactly. Well, and there's a broader lesson here, I think. It extends beyond compensation, and that's as AI becomes more involved in our decisions, do you think we're entering an era where explainability is just as important as efficiency?

Tom Heys: I think so. I think efficiency will become-- Using AI to become efficient, then great.

We've all seen it's super impressive and does lots of things that can really speed up decision-making. But I think that's the key. Use it to accelerate your decision-making. Don't outsource your thinking to it. Use it to enhance your thinking.

David Rice: Yeah. I think this is one of the biggest sort of challenges leadership challenges of the AI era, right?

It's like for years we've just been, like, rewarding faster decision, faster process, faster execution, go quick, go-- move fast and break things, right? But we're, we're entering this world now where leaders may have to explain decisions that were partially influenced by systems that they didn't build and, in a lot of cases, don't personally understand that well.

And so that shifts explainability quite a bit from being sort of a technical issue to being a leadership capability.

Tom Heys: As you say I mean, the importance of governance around all of this stuff, I think definitely will be increasing focus on it.

David Rice: Well, Tom, this has been a fascinating conversation. I feel like I learned a lot from you 'cause I didn't know that much about this, this regulation.

So this is good for our audience to kinda get a good grip on this. Thanks for coming on the show today.

Tom Heys: Oh, no problem. Thanks for having me.

David Rice: All right. Well, listeners, be sure to look out more for this type of content and for more on developing regulations around AI, pay transparency, and everything that's going on in the workplace right now by following People Managing People on all your social platforms as well as heading on over to peoplemanagingpeople.com/subscribe. Get signed up for the newsletter. You'll get all of this content straight to your inbox.

And until next time, think about how you're gonna explain things.

Before we wrap up, here's one last tip. Staying organized throughout the year makes tax time a whole lot easier. QuickBooks Online helps keep your income, expenses, and financial records organized while simplifying bookkeeping and reporting along the way. Whether you're a freelancer, a sole proprietor, or a growing small business, it gives you the tools to stay on top of your finances with less manual work. See how QuickBooks Online can help simplify your business finances today.