Employees at Sodexo were horrified by the murder of George Floyd in Minneapolis in May 2020 and the food services and facilities management company felt it needed to do more to bring about meaningful change. One of its first actions was to sign an open letter to the Sunday Times newspaper from business leaders, pledging to set stretch targets for ethnic representation and report annually on its progress. The company then committed to reporting its ethnicity pay gap data, and was the first employer in the hospitality sector to publish this, in 2021.
“We believe data is one of the best places to start,” says Raj Jones, Sodexo’s head of diversity, equity and inclusion. “It shows you where people sit in the business, what pay equity is like. It enables conversations internally and holds us to account externally.” The company also set up a strategic taskforce made up of leaders from Black and other ethnic minority backgrounds to better understand the lived experiences of employees and start meaningful conversations about race. Jones says there is an “ongoing drumbeat” of activity around inclusion that brings in senior sponsors, employee networks and external speakers.
Sodexo is one of a growing number of employers choosing to voluntarily publish its ethnicity pay gap data as employee demands get louder around action on environmental, social and governance (ESG) goals.
According to Business in the Community, which asks organisations to sign up to pay gap reporting as part of its Race at Work Charter, the number of employers doing so has grown from 11 per cent in 2018 to 19 per cent in 2021. After a government consultation into making it a legal requirement concluded more than three years ago, there had been mounting pressure from MPs and business bodies to introduce legislation in April 2023. But this will no longer happen, as the government’s response to the 2021 Sewell report into racism in Britain has confirmed. Instead, the Inclusive Britain report said the government would support employers that want to publish their ethnicity pay gaps, and would publish guidance to help them do so this summer.
The report acknowledges that there are a host of challenges associated with ethnicity pay gap reporting that did not impact gender pay obligations – something that was highlighted by the then small business minister Paul Scully during a parliamentary debate on the issue last year.
First, we have a wide range of ethnic minority groups in the UK – ethnicity pay reporting is not ‘binary’ in the same way as gender pay reporting, so it is disingenuous to make sweeping comparisons between ‘white’ employees and all other ethnicities. This in turn affects statistical robustness, because minimum sample sizes would be needed to ensure data shows an accurate picture of pay comparisons in an organisation, and some ethnicities may have a far smaller representation than others.
Furthermore, any requirement would have to support reporting across demographically different areas – employers in parts of the country with smaller ethnic minority populations could struggle to produce a meaningful report because the numbers are too low.
Nigel Marriott, an independent statistician who has been involved in shaping the new guidance, agrees that it would be impossible to apply the same framework used for gender pay gap reporting to ethnicity. “With gender, the census shows it to be around 50/50 in the population while around 86 per cent of the UK is white. That means you’re comparing a majority category with multiple minority categories,” he says. “Also, ethnic distribution varies considerably across the country, so the end result could be that an employer might be all-white by virtue of where it is based. Similarly, if you’re doing a straight comparison between white and all other ethnicities, you could eliminate the pay gap by just recruiting a single minority – your ethnicity pay gap would be zero, but you’d have got there by excluding all other ethnic minorities.” Furthermore, because ethnic identity is something that has to be ‘shared’ willingly by an employee, anyone who felt uncomfortable with this could just choose to tick the ‘not applicable’ or ‘prefer not to say’ box when they are recruited, decreasing disclosure rates and the reliability of any analysis.
In its ethnicity pay reporting guide for UK employers, the CIPD advocates – as long as numbers allow – using the five broad ethnicity categories used in census data: white; mixed or multiple ethnic group; Asian or Asian British; Black/African/Caribbean/Black British; or other ethnic group. A recent survey by the HR body found that 47 per cent of those collecting ethnicity data used these classifications, and 30 per cent of this group used the 18 sub-categories from the 2011 census. But when it comes to statistical robustness and sample sizes, it’s likely that only larger organisations would be able to report with any granularity, Marriott suggests.
The Royal Statistical Society stated in its recommendations for improving pay gap reporting in 2019 that any category with fewer than 100 employees would become less and less reliable the smaller the sample. On this basis, only a small percentage of employers with fewer than 1,000 employees would be able to compare data based on the census ethnicity categories. Go down to the 250-employee threshold for gender pay gap reporting, and “around 95 per cent” would not be able to report, he says.
With almost 23,000 employees, consulting giant PwC has the luxury of not only a large workforce sample but years of experience in collecting and analysing workforce data. The company has published details of its ethnicity pay gap since 2018, something that “creates a stronger sense of accountability to drive improvements internally”, according to Sarah Churchman, UK head of diversity, inclusion and wellbeing. One of the drivers for publishing this data before being under legal obligation to do so is the size and maturity of the company’s dataset. “We’ve focused on this for a long time so we have good data – around 97 per cent of people have shared their ethnicity,” she explains. “When you have robust data you can do good modelling.” PwC captures data during the recruitment processes to ensure they are fair, as well as other moments in the employment lifecycle such as performance management, promotions and participation on high-potential programmes. Employees are also invited to update their personal details – which includes a question about ethnicity – when they review their flexible benefits and complete regulatory training, so there are plenty of opportunities to disclose.
The benefit of such a high volume of good quality data is that the company has been able to break down its pay differentials further than comparing white employees with anyone from an ethnic minority background. So while its overall ethnicity median hourly pay gap is 2.9 per cent, this fluctuates according to different ethnicities within the workforce. “When we started breaking it down, we realised that 19 per cent of our Black and ethnic minority talent is Asian, and only 4 per cent Black. This obviously affects the pay gap down the line, so we thought about where we need to intervene,” adds Churchman. This year, PwC launched a scheme aimed at Black undergraduates where they undertake a three-day paid programme that could lead to an internship, work placement or graduate role depending on the stage of someone’s degree. Sharing the role of data in why such programmes exist is all part of building a culture of trust, she says. “We have a clear explanation as to how we use the data and that it’s held confidentially so we can look at the fairness of our processes. People like to believe we live in a meritocracy, but you need data to prove it.”
Sodexo broke down its own ethnicity pay data into four categories: white; Black; Asian; and mixed ethnic, as well as calculating an overall gap between white employees and those of other ethnicities. It was able to produce a figure for the overall gap for each of its legal entities (as is required in gender pay gap reporting) but the sub-categories had to be calculated across the business as a whole. “We’re now running a big initiative to get all of our diversity data up so we can overlay it and dig much deeper,” adds Jones. While the overall gap is quite small because around 20 per cent of frontline staff are from ethnic minorities, one of the key priorities has been to review processes across the employee lifecycle to ensure workers receive equitable promotion prospects within the company. “We want people to stay here and thrive,” she says.
But building this level of transparency also requires a degree of trust. “An initial hurdle is that some employees may not like to reveal to their employer how they categorise themselves, so be upfront about why you’re collecting the data and provide assurances that it will remain anonymous,” says Amanda Glover, a senior employment solicitor at Clarkslegal.
Sandra Kerr, race director at Business in the Community, says organisations need to “build confidence that the data will be used for something positive to benefit the organisation and the individual”. Many employers use internal campaigns to encourage people to self-declare, or share how past data insights have driven improvements, she advises. “Even if you don’t get everyone declaring, having people tick ‘prefer not to say’ means the message has been delivered rather than not getting through. If leaders share their stories on the ‘why’ and you encourage new joiners to disclose, you’re driving up data capture,” she adds.
Network Rail has disclosed its ethnicity pay gap for the past three years as part of a longer-term strategy to gain more transparency on the challenges for employees across a spectrum of protected characteristics. Two years ago, the diversity team launched a disclosure campaign offering employees the chance to update their personal details on ethnicity alongside other categories including their sexual orientation or if they have a disability. Office-based staff received an email, while frontline workers were written to and given a pre-paid envelope to return their responses. “The message was ‘you can share as much as you want’ – we’re clear that we use data to target resources and programmes and why it’s good to know the make-up of our workforce,” explains Jonathan Payne, who works in the D&I team. “But we’re also clear about what we don’t use it for – their line manager would not see it, nor would a recruiter.”
Gathering data across multiple diversity strands means the company can look at pay equity through an intersectional lens, he adds. “Many people might have one protected characteristic but we wanted to know the compound effect of being in more than one – for example, overlaying ethnicity pay with gender can show us whether males or females from certain backgrounds are treated differently.”
While it is yet to report its ethnicity pay gap data publicly, this year the Crown Prosecution Service (CPS) decided to conduct an analysis using the same methodology as the gender pay gap calculations, so it could get a baseline figure to work with. “While we collect ethnicity data on specific ethnic minority groups, we have undertaken the pay gap analysis on a more simplified level, using White and BME categories, because we know that if we were to analyse all minority groups separately such small sample sizes would provide little or no statistical validity,” says Andy Read, deputy director of people services. The exercise also threw up other statistical challenges, such as variations in employees’ length of service or their terms and conditions. “Overall, however, it remains important that we don’t just look at pay gap data in isolation: it is part of a much wider analysis of diversity in employment and directly informs our diversity and inclusion strategy,” he adds. Like Network Rail, the CPS embarked on a disclosure campaign called ‘Count me in’, including messaging in staff inductions, and inclusion and diversity events, such as awareness days. The organisation has increased the ethnic diversity of its workforce over the last five years – 21 per cent of employees are from ethnic minority backgrounds compared to the Civil Service average of 13.2 per cent – but Read adds that “data is imperative in sustaining our position as an inclusive employer”. “Having a strong set of accurate diversity data enables us to implement positive action in a properly targeted way,” he says.
Of course, simply publishing a set of figures does not translate to a greater sense of inclusion and belonging – the work that follows these insights is what makes the difference. “Even if you don’t report the data, there are actions you can commit to as a company that will reduce any pay gap, such as not asking for salary history or addressing unconscious bias,” adds Glover. Crucially, collecting and analysing data on pay across ethnicities can start an important conversation about where your business needs to be. Payne says Network Rail has never looked back since it started expanding its reporting processes: “Doing this demonstrates you take equity seriously – both to people outside your organisation considering you as an employer, and your existing employees,” he says.
Where should businesses start?
Although the government has now confirmed it will not introduce a legal requirement for companies to report their ethnicity pay gap, there are solid reasons to move forward and do so. Getting ahead on the data is “an opportunity to show you’re credible” about your diversity and inclusion promises, says Alison Woods, partner and co-head of employment at CMS Law. “We’re seeing ESG starting to have a snowball impact, and much more is going on without legislation having to be in place,” she says. “It’s people power; companies are looking at this issue because of wider pressures rather than a legal obligation.” So how should organisations set about collecting and reporting the data?
- Build trust: “People might not be willing to share if they already have concerns about their employer’s approach to diversity,” says Woods. “We’ve seen clients go on a journey of trust, but if you’re clear on data privacy and how you will use the data, you can build up the information.”
- Collect the data: ‘Keep asking’ is a common message among organisations that have achieved high disclosure rates (anything north of 70 per cent, according to Sandra Kerr, race director at Business in the Community). Use data-capture moments such as employee surveys, annual appraisals and onboarding to increase levels of data sharing.
- Run the numbers: In its Ethnicity Pay Reporting guide, the CIPD recommends organisations publish a) the difference between the median hourly rate of pay of all “full-pay relevant” white employees and the same for those from other ethnic minority backgrounds and b) the difference between the mean hourly rate of pay for the same groups.
- Break down the data: If numbers allow, run an additional analysis based on more detailed ethnicity categories alongside the headline figure. As with gender pay reporting, organisations could also look at the gap in bonus payments and present pay gap analysis across quartile pay bands.
- Present the context: The CIPD advocates two additional statistics that do not appear in gender pay reports: demographic data on the proportion of an employer’s total workforce from ethnic minority backgrounds; and the proportion of employees that chose to disclose their ethnicity. This data can then be used as contextual information in your report.
- Produce a narrative: A key recommendation of the Commission on Race and Ethnic Disparities is that organisations publish a diagnosis and action plan, setting out the reasons for and steps to address the disparities the figures have highlighted. “Having a duty to report does not fix the problem, as we’ve seen with the gender pay gap,” adds Woods. “Start to think about what your narrative would be, engage with your workforce and see if there is a wider commitment to change.”