The greater complexity of ethnicity pay reporting compared with gender pay reporting may be one reason for the UK government’s delayed response to its consultation. In 2018, the EHRC found that half (51%) of employers said they face barriers to collecting employee ethnicity data, including data collection being intrusive and employees not wanting to share the information. 

Challenges include: 

• Deciding on the most appropriate categories of ethnicity to report: reporting all those with ethnic minority backgrounds as a single category like ‘BAME’ (as well as often being an inappropriate term) masks the variations in labour and pay market outcomes between ethnicity groups. Using the ONS Census’s five ethnicity categories, or more detailed 18-category breakdown (see ‘The statistics: What ethnicity pay data to report’ below) on the other hand, risks skewed results because of low numbers. See below for more on terminology.  

• Intersectionality: intersectionality of ethnicity with other aspects of diversity – such as gender – makes unpicking any pay gap more complex.  

• Regional differences: the challenge of reflecting the local community’s ethnic mix in a workforce will be very different for, say, an employer based in Lincoln compared with one in London. This may also vary between the different nations within the UK.  

• Uncertainty and discomfort around language: this can stop people from having conversations about race.  

The obstacles to collecting, analysing and reporting ethnicity pay data are undoubtedly greater than for gender pay reporting. But there is evidence of progress, particularly post-COVID-19. In 2020, a survey of more than 100 employers carried out by PwC found that 67% collect ethnicity data (up from 53% in 2018), 23% calculate their ethnicity pay gap, and 40% of these have published it voluntarily.  

A 2019 CIPD survey of 243 members found the overwhelming majority supported the UK government’s compulsory ethnicity pay reporting proposal.  

The ‘top five’ benefits identified were:  

• To develop a reputation as a fair and progressive employer (66%).  

• To address workplace inequalities (61%).  

• To develop greater transparency and accountability (58%).  

• To offer ethnic minority employees equal access to development and progression opportunities (55%).  

• To create more inclusive workplaces (55%).  

This guide is intended to support employers who wish to take advantage of these opportunities and improve inclusion in their organisation through voluntary ethnicity pay reporting. 

We advise employers to be sensitive in the language and terminology used, and to engage and invite input from staff and external experts.  

The ONS defines ethnicity pay gaps as ‘the difference between the median hourly earnings of the reference group (white or white British) and other ethnic groups as a proportion of average hourly earnings of the reference group’. The CIPD recommends additional, more detailed analysis of pay by the more granular ONS ethnicity categories wherever possible.  

The CIPD follows the Race Disparity Audit’s recommendations in this guidance, referring to ‘ethnic minorities’ rather than using BME/BAME (black and minority ethnic/black, Asian and minority ethnic), which highlights some groups and omits others.  

Race Disparity Audit research suggests that very few people recognise these acronyms, while few people from ethnic minorities identify with them and many dislike them. Aggregate terms like BME and BAME encompass people from a diverse range of backgrounds, cultures and traditions. However, government departments, public bodies, the media and other groups use these terms. We therefore reference them in relation to research that has been conducted using these terms.   

While ethnicity pay reporting remains voluntary, it is likely that one of the main barriers will be failure to secure the necessary support from senior leadership. The rationale for voluntary reporting therefore needs to be strong. The CIPD puts forward the following three arguments in support of ethnicity pay reporting:  

The moral case  

There is an obvious moral argument for closing ethnicity pay gaps – no one should be disadvantaged by their ethnic origin in the workplace.  

The EHRC observes: ‘We have long known that people from a range of ethnic minorities are treated less fairly across all areas of life,’ with the Is Britain Fairer? report showing substantial differences in pay gaps across different ethnic groups. For example, white British employees earned a median hourly rate of £11.67, compared with £9.93 for Pakistani workers and £9.24 for Bangladeshi workers. Clearly, this is down to a whole range of social, economic and employment factors, but responsible employers cannot simply blame society. They need to do something to effect change.  

Ethnic differences also interact with other diversity strands. This means that we cannot, for example, fully address gender pay gaps or disability discrimination without also considering ethnicity. Any organisation that wants to improve inclusion and diversity must therefore consider how voluntary reporting and action plans will contribute to these aims.  

The business case  

There is growing evidence that reporting on and addressing ethnicity pay gaps has major economic and financial benefits for the UK economy as well as for individual employers and their employees. 

The Department for Business, Energy and Industrial Strategy (BEIS) analysis (2016, cited in EHRC 2019) calculated that improving employment rates and workplace progression for people from ethnic minorities could contribute £24 billion per year to the UK economy.  

There is also a clear benefit to employers. McKinsey’s Diversity wins report found that the relationship between diversity on executive teams and the likelihood of financial outperformance has strengthened over time. The IES also believes that more diverse employment practices contribute to higher employee engagement and loyalty to the organisation. 

Current context  

The pandemic and social movements such as Black Lives Matter have increased the pressure on employers to address the employment and pay disadvantages and unequal treatment of non-white ethnic groups. At the same time, environmental, social and governance (ESG) priorities have moved up the corporate agenda. Greater attention is being paid to the corporate risks arising from the ‘E’ of climate change, but also on the ‘S’ of ESG, with investors and society increasingly focused on how companies treat all their stakeholders, especially employees. 

However, while the CIPD’s research found that the majority of employers (77%) believe that ensuring workforce diversity is a priority, only 36% of them collect and analyse data to identify differences in pay and progression for employees from different ethnic groups – which is even less than for disabled employees (44% data collection and analysis).  

While ethnicity pay reporting alone does not close or address ethnicity pay gaps directly, there is good evidence that it helps to prioritise and focus attention on actions which do. 

Although the continuing lack of legislative compulsion means UK employers currently have total flexibility on what to include in their voluntary ethnicity pay reports, this presents a significant risk. If employers decide to publish differently defined statistics, we will lose any sense of cross-employer and sector comparability.  

In the absence of legislation, the CIPD believes that all employers should publish annual ethnicity pay reports with three common components:  

• A uniform set of eight commonly designed statistics (see The six pay statistics, below) to profile pay by ethnicity in their organisation.  

• A supporting narrative to explain the nature and causation of any pay differentials and gaps by ethnic group.  

• An action plan of initiatives defined to reduce and remove these gaps over time. 

Principles and justification  

We have followed several core principles in selecting the most appropriate information for employers to include in their annual ethnicity pay reports: 

1 Align ethnicity pay reporting with gender pay reporting, but recognise the differences  

From a practical perspective, this alignment makes sense. Any compulsory reporting legislation on ethnicity will most likely be introduced as an extension of the existing gender pay regulations (reporting data drawn on the same snapshot date, using the same statistics and definitions – of employees, of pay, and so on). This will minimise the additional work and bureaucracy for HR and payroll (remembering that gender pay gap reporting is not yet mandatory in Northern Ireland). But there are differences in reporting on ethnicity which you will need to reflect, most notably the larger number of categories required beyond the male/female groupings on gender.  

2 Remember ethnicity representation is as important as, and linked to, ethnicity pay gaps  

The under-representation of ethnic minority compared with white employees in UK organisations’ senior management, and over-representation in low-paid jobs, is even more evident than the gender representation gap in these roles. We recommend that employers report on their ethnicity group representation breakdowns in each quartile, as well as their pay gaps by quartile (which many employers do voluntarily on gender).  

3 Recognise the value of simplicity and clarity  

While a single gender pay gap figure can be criticised for providing no indication as to how it can be reduced, it has been very successful in securing attention and action to address any gaps. We recommend that (alongside a narrative and action plan) the following two pay statistics are published by organisations:  

• Median ethnicity pay gap: The difference between the median hourly rate of pay of all white full-pay relevant employees and the median of full-pay relevant employees from other ethnic minority backgrounds.  

• Mean ethnicity pay gap: The difference between the mean hourly rate of pay of all white full-pay relevant employees and that of full-pay relevant employees from other ethnic minority backgrounds. This will highlight any difference between the median and mean average pay of their white and ethnic minority employees.  

4 Focus on action  

The CIPD believes that narrative reports to explain pay gaps and action plans designed to close them should be compulsory, alongside the mandatory publication of the six statistics in annual gender pay reports.  

5 Start and improve  

Employers vary in their ability and the resources they can devote to ethnicity pay reporting. Our recommendations focus on what we believe employers should be able to calculate and analyse, while suggesting more detailed categories, breakdowns and actions that could be done. It will take sustained action, over a number of years, to have a serious impact on any gaps revealed.  

6 Combine comparability in data with tailoring of analysis and actions  

There should be a minimum of core information communicated in every employer’s annual ethnicity pay report. This additional analysis – by detailed ethnicity category, by stage in the employment cycle, by ethnicity and gender, and so on – is crucial for employers to fully understand and address any ethnicity pay gaps evident in their stats. 

Perhaps most critical is the need for the statistics to balance the requirement to be:  

• as simple, clear and easy-to-produce as possible, to minimise the administrative and resource burden on employers, especially SMEs  

• sufficiently detailed and meaningful, to highlight the exact location and nature of any gaps, and indicate how they have been caused and how they might be closed.  

In our view, the six compulsory gender pay reporting statistics (detailed below) successfully achieve this balance, with the single mean and median pay gap figures providing a focus for attention and measurement-through-time of overall progress. The additional bonus and pay quartile breakdowns highlight common causes of gender pay gaps, which are discretionary variable pay policies and a lack of female representation in senior and higher-paying roles/the top pay quartile in many employers.  

We recommend these same six statistics are also applied to ethnicity pay reporting, incorporating the same definitions of employees and hourly pay, the same snapshot reporting dates, and so on. But we also recommend two additional statistics to reflect the greater complexity and difficulty typical of ethnicity pay reporting and analysis compared with gender (as outlined in Table 2). 

The CIPD recommends all employers publish the following data in any voluntary ethnicity pay reports. 

Table 1: The six pay statistics

In addition, we propose two additional statistics that focus on encouraging the full representation of ethnic minority staff in the workforce (Table 2).

Table 2: Additional statistics that focus on encouraging the full representation of ethnic minority staff in the workforce  

The statistics themselves (mean, medians, and so on) and how to calculate them, as well as the relevant terms and definitions, such as employees, pay and bonuses, snapshot dates, and so on, are summarised below and are defined in detail in the CIPD’s Gender Pay Reporting Guide. More information on key terminology is also included in Appendix 2. 

Ethnicity pay gap  

This is calculated as the difference between the median (or mean) hourly earnings of the reference group (your white employees) and the median among your other ethnic minority employees, expressed as a proportion of average hourly earnings of the reference group. For example, a positive 5% ethnic pay gap between white British and employees from other minority ethnic groups would show that the median hourly earnings for employees in the latter category are 5% less than median hourly earnings of white employees. Conversely, a negative 5% pay gap would denote that they earn 5% more, on average, than white British employees. 

Employees  

The definition used for ethnicity pay reporting ideally should be aligned with your annual gender pay report, which is the employer’s population of ‘relevant full-pay employees’, that is, those employed on the ‘snapshot date’ and receiving their full basic pay or piecework pay in the pay period. This excludes anybody not receiving their full pay in the relevant period, for example, people on long-term sickness absence. Agency staff are included as employees of the agency and so should not be included in an employer’s calculations. Nor should a firm’s partners and directors if they are not employees. Where possible, though, the CIPD would encourage employers to include these people in a separate analysis in their accompanying narrative. 

Pay  

Pay is defined as hourly gross ordinary pay, including basic pay, piecework pay, shift premiums, paid leave pay and allowances, and so on. 

Bonus pay  

Bonus pay includes all variable payments, such as individual and team bonuses, profit-sharing, gainsharing payments, and so on. 

Hourly pay  

Hourly pay is calculated based on earnings and hours worked in the relevant pay period (pay period including the snapshot date, for example the month for monthly paid employees); bonus calculations are based on cash payments received in the 12 months leading up to the snapshot date (the [INT] CIPD’s Gender Pay Gap Reporting Guide contains some illustrative examples). 

Snapshot date  

The pay calculations are based on a ‘snapshot date’ of 31 March for public sector organisations, and 5 April for the private and voluntary sectors – again, we recommend the same date is used for ethnicity pay reporting. 

The median  

If you rank all of your people by their hourly pay, and then compare what the person in the middle of your ethnic minority population pay range received with what the person in the middle of your white population pay range received, and multiply by 100, you get the median ethnicity pay gap expressed as a percentage of the white employee pay. This is the preferred measure of the ethnicity pay gap, according to the ONS. 

The mean  

To calculate the mean ethnicity pay gap, add together all the hourly pay rates that people from an ethnic minority group earn, divided by the number of people from an ethnic minority group in the workforce, and then compare that with the same calculation for white people. Again, multiplying by 100 gives you the percentage difference. While the ONS statistically prefers the median as its primary measure, some argue that for employers the mean better captures significant differences at the top and bottom of the income distribution. 

Proportions and percentages  

A proportion is a number which is read in comparison with other numbers, in this case white and ethnic minority groups’ pay levels. It can be expressed as a numerical value (25 out of 100 ethnic minority employees received a bonus payment compared with 20 white employees), a fraction (a quarter of the ethnic minority people we employ received a bonus payment), or a percentage (25% received a bonus payment). The gender pay regulations require percentages, but it is worthwhile remembering that in some circumstances a numerical value may be a more effective way of communicating your findings. 

Quartiles  

A pay quartile represents a quarter, or 25%, of your total workforce, ranked by pay. The pay quartiles then need to be broken down by ethnicity. For example, the top pay quartile would be the highest 25% of earners of all full-pay-relevant employees of all ethnicities, including white employees. Then for those top 25% you would work out and report the proportions that were white employees and the proportion that were other ethnicities/BAME. So, if you had a business with 800 employees in total, 200 employees would be in the top quartile. If 150 of them were white employees and 50 from other ethnic minorities, you would report an ethnicity mix in that quartile of 75% white employees and 25% BAME employees. And you would then repeat the calculation for the other three quartiles. Examples of these calculations are shown in the CIPD’s Gender Pay Gap Reporting Guide. 

More information on these definitions is available in Appendix 2. 

An additional complexity with ethnicity pay is deciding on the most appropriate categorisations of ethnic group to use. This is probably as important a decision as the actual statistics themselves.  

The CIPD employer roundtables and survey at the time of the 2019 government consultation found that calculating ethnicity pay information by the same quartiles as used for gender pay gap reporting was the most popular single option (favoured by 34%), with:  

• one pay gap ‘single figure’ comparing average hourly earnings of ethnic minority employees as a percentage of white workers being the preference for more than a third (30%)  
• almost a quarter (23%) supporting calculating and publishing several pay gap figures for different ethnic groups, using ONS standard ethnicity classifications (which we recommend employers consider in addition to the recommended statistics)  
• a fifth also agreeing that a mixture of measures could be used.  

Respondents from larger employers were slightly more likely to prefer using the more detailed ONS ethnicity categories and using the same quartiles breakdown as those used for gender pay gap reporting (37%); they were a little less likely to support using the single white employees/ethnic minority group employee pay gap figure.  

By contrast, employers with fewer than 250 workers were more likely to prefer one pay gap figure comparing average hourly earnings of ethnic minority employees as a percentage of all white employees (41%), reflecting the fact that most have not reported their gender pay gaps before. Even using the five-category ONS breakdown would leave many with categories with insufficient numbers to be able to report data.  

The single overall median/mean comparator figure was the subject of the most debate. On the one hand, it should be easy to calculate and, by being less granular, it would be harder to identify individuals even in the smallest employers. However, large employers felt it might obscure differences within the ethnic minority workforce (although it was acknowledged that this issue could be tackled through additional analyses).  

The advantages of the single headline figure, as one survey respondent said, was that:  

“The UK government has a large number of ethnicity classifications. Comparing each individual group means that you might be looking at very small populations versus a very large population, and that makes the output much less meaningful, as one or two people could make a large change in that group’s gap. While it won’t provide a detailed insight into the gap, a single comparison of non-white British employees with others would allow comparison of the gap in earnings between the majority ethnicity group and everyone else.”  

We also see great benefit in applying internationally recognised and standardised definitions of ethnicity pay and pay gaps. This ensures the quality and comparability of the information, and maximises its value relative to the cost and time involved in the reporting process.  

For the ONS national statistics, the headline measure of the ethnicity pay gap uses Annual Population Survey data and is calculated as: ‘The difference between the median hourly earnings of the reference group (white or white British) and other ethnic groups as a proportion of average hourly earnings of the reference group. For example, a positive 5.0% ethnic pay gap between white British and Indian ethnic groups would denote that the median hourly earnings for employees of an Indian ethnicity are 5% less than median hourly earnings of white British employees. Conversely, a negative 5% pay gap would denote that employees of Indian ethnicity earn 5% more, on average, than white British employees. Using this terminology ensures consistency with existing analysis of different pay gaps.’ 

The EHRC reflects this in the definition of ethnicity pay gaps it applies to individual employers as well, defining this as: ‘The difference between the average hourly pay of ethnic minorities and white British people. It is usually expressed as a percentage difference, with white British people’s earnings representing 100%.’

We recommend using this ONS binary definition of ‘white combined’ compared with ‘ethnic minority combined’ as the ethnicity categorisation for the core statistics that all employers should publish. Many employers will also want to carry out additional analyses using the ONS categories as they investigate the exact location and causation of gaps, which we recommend and describe in the next section of this guide.  

Table 3 profiles the ethnicity categories that were used for classifying people in the 2011 census, with Table 4 showing the numbers in each category in that 2011 census returns to give some idea of the relative scale of each group. The 2021 census has used an almost identical breakdown, with the only difference being that Roma people have been separated as a distinct additional category. Their numbers are so small that this distinction will make no difference for almost every individual employer. 

Table 3: Survey ethnicity categories and structure

Table 4: Number of people in each category in 2011 census

  
As you can see, there are six larger groupings (sometimes referred to as five plus the ‘other ethnic group’ category, or ‘five plus one’). These are:  

• broken down further into 18 sub-categories, and  

• amalgamated into two overall ‘white’ and ‘ethnic minority’ groups, which are the basis for the headline single ethnicity pay gap figure published by the ONS and which we recommend are used for the core, common statistics published in all ethnicity pay reports.  

We asked in our CIPD ethnicity pay consultation survey whether respondents used the ONS standard ethnicity classifications for reporting. Of the 100 respondents who answered, almost half (47%) said that they did. Around a fifth (19%) said that they do not use ONS standard classifications, whereas 34% don’t know.  

Of the 47% of respondents who use ONS classifications for collecting data, 30% use the more detailed 18 standardised classifications from the 2011 census, 15% use the five/six standardised classifications from the 2011 census, and 13% say they use the 16 standardised classifications from the 2001 census. 

There was a consensus among respondents that using up-to-date census categories and figures was preferable, to reflect the current demographics in society. At some of the employer roundtables, there was a view that employers should use the classifications that mirror those used for the 2021 census, otherwise the danger is that employers will not be able to compare if their own employee population is not as ethnically diverse as the local population official data.  

The CIPD is acutely aware of the problems of combining disparate ethnicity groups in order to report a single all-white employees compared with all other ethnic minority group employees pay gap figure. It is often inappropriate culturally to blend groups together and present them as having a single view. In its own ethnicity pay report the CIPD uses this binary classification, but refers to the comparison as being between the ‘pay of white employees and that of employees from other ethnic backgrounds’, avoiding the BAME term. The CIPD also publishes a more detailed breakdown of all-white employees compared with three BAME sub-categories of black, Asian and other mixed ethnicity employees. PwC uses exactly the same categories and breakdown in its ethnicity pay report.  

Mathematically, doing this binary categorisation may also mask some of the differences between the pay of the white British majority in a firm and some of the various ethnic minority groups, hence the additional analyses that the CIPD recommends.  

So, while this blending together may often not be appropriate in conversations with individuals, many employers will often be in the position of needing to do this amalgamation for reporting purposes. Using all the ethnicity categories where an organisation has few employees in some groups could risk individuals being identifiable, or that any tracking of the success of an initiative will reflect a small number of individuals’ experiences.  

However, as with gender, we recommend that these simple binary median and mean pay comparison figures are published by all employers. It should provide a focus on the issue and help to drive more detailed analysis to get at the causes of any gap. It is also particularly suitable for the majority of the UK’s smaller employers. 

Ethnicity pay reporting should not be data-gathering for its own sake. As well as publishing the eight common statistics recommended in the previous section, the CIPD believes that every employer’s annual ethnicity pay report should also contain:  

• a narrative based on deeper analysis of the nature and causation of any ethnicity pay differences found in the eight headline statistics  
  
  
• an effective, targeted action plan of measures designed to close any of these gaps.  

Figure 2: Stages in the EPR process

Recommended stages in the EPR process are as follows: 

What are you trying to achieve, given that it is not currently a legal requirement in the UK? If it is to become a more equal, diverse and inclusive organisation, how are you going to achieve this? What level of resources will you need to produce your report? Having a clear and agreed statement of aims will help to address any difficulties.  

Being simple and clear in your approach are relevant here. In your first year of reporting, you could aim to publish accurate and representative core statistics on a binary all-white/other ethnic minorities categorisation. You can then move on to be more ambitious in future years. This could include more detailed analysis and driving action to address pay disparities, possibly setting and monitoring specific targets.  

Scope is an important consideration. In one sense the answer is obvious: ‘we will include all our employees and report on their pay by the most appropriate categorisation of their ethnicity.’ But do you know the ethnicity of all your staff? If not, what level of disclosure and coverage will you be comfortable with to produce and publish as your core stats, and regard them as valid and representative of your whole workforce?  

Critically, in terms of scope and content, how important is it to link work on ethnicity to other diversity strands? Do you agree that the link with gender is key and therefore to adopt the statistics and definitions from compulsory gender pay reporting and use these? Or are you going to adapt/modify them in some way? It is important to confirm this, as there are numerous options on what and how to report on ethnicity pay (as highlighted above). 

The strength of any report is dependent on the quality of the information and data that goes into it. Many employers have been able to significantly improve the coverage and quality of their data required for their gender pay reports since compulsion was introduced and the statistics and definitions finalised in the 2017 legislation and associated guidance. But as already noted, the challenges on ethnicity are often more significant, with many employers starting from a lower initial point of coverage.  

The process of data-gathering and cleaning and the data that you need to record to calculate is illustrated in Figure 3. 

Figure 3: The data required

Employees and full-pay relevant employees 

You will need to have a list of all employees and all full-pay relevant employees. You will already have these if you produce annual gender pay reports. Remember that you cannot use full-time equivalents but must base your calculations on the number of individual employees employed on the snapshot date. A full-pay relevant employee is one employed and receiving full pay during the specified pay period. An employee taking parental leave or sick leave on full pay or on unauthorised absence would be included, but employees on reduced pay, for a reason such as maternity leave or any other form of long-term special leave, should not be included. Although not required by the regulations, the CIPD also recommends that you record employees’ contract status. This will help to decide who to include in the calculation of your gender pay gap (salaried partners, for example). It will also help with year-on-year comparisons and seeing the impact of any changes in status of workers, if say, for example, your organisation decided to give permanent contracts to a team of self-employed or agency staff.  

Employees’ ethnicity, broken down into the most appropriate categories 

You need a record of the ethnic group of your relevant and full-pay relevant employees, which involves you determining which ethnicity categorisation you are going to use. Even if you decide to only publish the average pay for the binary all-white/other ethnic minority categories in your report, it is usually worth gathering data using the more detailed ‘five plus one’ and/or 18 ONS categorisations. Most employers already attempt to capture this information when they are recruiting their employees and it makes sense to align these categories across the employment journey of your staff. The main problem tends to be low levels of disclosure and coverage within many employers and a high proportion of ‘don’t knows/not stated’ in your data. Therefore, the CIPD recommends you publish this statistic too, as an indication of the quality of your data. 

Hours worked 

A record of the weekly working hours for full-pay relevant employees – this information underpins the calculation of the hourly rate, used in calculating both the mean and median pay gaps.  

Ordinary pay and hourly pay 

A list of all pay received in the relevant pay period for full-pay relevant employees only – this information underpins the calculations of the mean and median pay gaps. You will also need to record compulsory employee deductions, such as court orders requiring child support payments, or other forms of attachment of earnings. This is because employees’ ordinary and bonus pay is calculated before deductions are made at source. You also need to record any benefits which are paid for by the employee on a salary-sacrifice basis, as under the gender pay gap regulations ‘the employer should use the employee’s gross pay after any reduction for a salary-sacrifice scheme’. As with the data on hours, many employers will already have this data and have done the calculations for men and women and so the same information can be used for ethnicity – all you need to do is relate the pay information to ethnicity rather than gender categories.  

Bonuses received and level of payment 

A record of all bonuses received in the relevant bonus period (12 months prior to the snapshot date) for all relevant employees and full-pay relevant employees – this information underpins the calculation of the bonus pay gap as well as the mean and median pay gaps.  

Quartile pay bands 

This information underpins the reporting of the proportion of different ethnicities within each of the quartile pay bands and of the average pay levels and any gaps for all white and other ethnicity employee categories within each quartile. 

Consider adding additional employee characteristics that may have an impact on ethnicity pay levels, such as gender, age, education, and other personal information that you hold. The ethnicity pay gap increases with age and is wider for men than women, and the level of education appears to be one factor linked to it. This information will help with your additional analysis and narrative drafting. 

If you don’t know the ethnicity of your employees, you can’t start reporting on their pay. New HR information systems are improving analytical capability but they depend on the quality of the data held. So, if your current coverage is low and you don’t know the ethnicity of the majority of employees, how can you report meaningfully on your ethnicity pay?  

Fortunately, there is a lot of recent research to highlight how employers can improve in this regard, and there are now many examples of employers having made rapid improvements by paying attention to these factors. For example, the CIPD’s report, Encouraging Ethnicity Data Disclosure, cites many useful cases and tips.  
 
Actions for encouraging higher self-reporting rates focus on reassuring employees that they can trust how the employer will be using the data and that it will be held confidentially. Employers should also make the data easy and ‘normal’ to provide and update. Specifically, the CIPD recommends:  

• Collect ethnicity data at continuous stages throughout employment: for example, in recruitment, at induction/onboarding, during annual appraisals, and so on.  
• Consistent communication campaigns: encourage employees to self-report and outline how this information will be used to embed equality, inclusion and diversity in organisational culture and change.  
• Provide different opportunities for staff to give this information: check and update the information on a regular basis.  
• Clearly explain to employees how the data will and will not be used: employees are likely to have many questions and potential fears around this, such as: ‘Where is this information being held and who can have access to it?’, ‘Might this be a basis for, rather than a means of preventing, discrimination against me?’ These need to be addressed.  
• Use staff networks to act as champions: the CIPD’s survey respondents talked about the value of using staff networks to answer any questions employees had, particularly those they may not feel comfortable asking of their manager or HR.  

It’s worth noting that the CIPD believes this provision of information should always be voluntary. So, employers should always include a ‘prefer not to say’ option.  

Individual confidentiality is also obviously key here. Employers must ensure that all such special category data is kept confidential, stored securely and complies with relevant legislation. In their actual report, there should be a clear minimum threshold for the number of respondents from any one ethnic group being reported, and responses should be suppressed if they fail to meet this threshold to protect their confidentiality. Ten respondents per category might be a suitable threshold. Individual responses should also be anonymised.  

The ONS has published guidance and recommended survey questions in order to gather data on ethnicity in the most effective and consistent manner. 

A narrative section in your annual ethnicity pay report would typically have the following aims and components:  

• Additional analyses and breakdowns of data gathered, to more fully understand the nature and location of any ethnicity pay gaps and differences between different ethnic groups.  
• Analyses of a wider range of diversity and other data to identify the inter-relationships and likely causation of any ethnicity pay differences evident in your published statistics.  
• A summary of some of the existing policies and practices that you have in place to combat the risks of different pay between people in different ethnic groups. 
• An evidential base, justification and support for the subsequent targeted action plan.  

These additional analyses, some or many of which may not actually be published for reasons of space and/or confidentiality, need to be tailored to each employer. It represents ‘a detective’s job’, to highlight, with as much accuracy as possible, where gaps are located and what associated and underlying factors might be driving them.  

Considering existing research studies on the factors that influence ethnicity pay variations, we would expect the following analyses to be useful.

Pay figures broken down by more detailed categorisations of ethnicity  

To really understand and address any differences in pay based on ethnicity will require going deeper than the headline binary comparison figure. We recommend smaller employers do so using the ‘five plus one’ ONS categories and larger employers segment the analysis using the full 18-category ONS classification wherever possible.  

This type of analysis invariably finds that some ethnicities have small or no differences with the average pay of white employees, but some may display significant shortfalls. This will enable much more effective targeting and communication of initiatives developed to remove them. 

Example: The Greater London Authority (GLA) 

The Greater London Authority publishes an overall all-white/BAME headline categorisation of pay comparator figures, but then breaks this down further to show actual hourly pay rate averages and percentage differences for each of the six ONS categories. The 2018 analysis shows that the gap for black/black British employees at 24% was more than double the overall BAME average of 11%. 

To balance the demands of analytics pushing for more detailed breakdowns and confidentiality considerations encouraging more aggregation, both the CIPD and PwC break down the same binary categorisation and headline pay data into three – those employees of Asian, of black and of mixed ethnic background – and compare these groups with their white employees. 

Statistics broken down by more detailed categories of pay  

This will often help to better locate and understand the nature of any ethnicity pay gap. We have already recommended that you calculate and show any ethnic differences by pay quartiles. A more detailed and commonly carried out analysis is to review the data by organisational level or grade. Certain types of pay, such as market pay supplements and management bonus plans and performance-related payments, are also, we know from research, prone to significant differences and advantages for white male employees. So, you may want to analyse these separately from your amalgamated ordinary total pay and amalgamated bonus pay figures. 

Relevant external comparator data  

This is often gathered and presented to put each organisation’s findings into better context. This might be compared with national figures, or with the relevant local area or sector. 

In terms of analysis aimed at identifying underlying factors of causation, two types of analysis seem to be most common and helpful.  

First are analyses of variations at different points in the employment cycle, for example in recruitment applications and acceptances, rates of promotion and associated pay levels, performance appraisal ratings, positioning within pay grades and ranges, and so on. 

Second are analyses of the relationships between pay levels and those demographic and other factors which have been found in research to help to explain ethnicity pay gaps at the national level (which we have summarised above).  

The most significant of these factors include:  

• Gender: At the national level the IES’s research with the EHRC showed not only different pay gaps for different ethnic groupings compared with white employees, but also significant differences within these groups according to gender. Nationally, the ethnic minority pay gap is significantly higher for males than females. This linked analysis is particularly important to highlight employee groups where these factors interact and compound to depress pay levels. The UCEA study found that ‘being a woman from an ethnic minority background could compound this disadvantage significantly, depending on the ethnic group in question’. Disabled employee pay analysis is also recommended by the CIPD and this factor can have similar compounding effects on ethnic group differences.  
• Overseas-born status: An employee’s UK-born or overseas-born status also has a major impact on pay at the national level, with gaps on average reducing significantly for second-generation and locally born ethnic minority people.  
• Location: There are significant variations in levels of representation and pay for ethnic minority groups across the UK according to location. There was a strong consensus from CIPD members that it is often helpful for an organisation to be benchmarked against their local demographics, instead of a national average, and we believe that universal generic targets on this are neither fair nor effective. This often also supports more tailored and specific self-targeting of improvements in closing ethnicity and gender pay gaps by individual employers as part of their action plans.  
• Age: We know that nationally the ethnicity pay gap increases with age. Is that the case in your organisation and, if so, how does that impact on your action plan to reduce and eliminate ethnicity differences?  

Employers with more advanced HR data analytics capabilities may want to model and consider the combined effect of these factors in the development of their action plans, replicating research that has been carried out on the national data.  

Improved pay transparency, in order to address ethnicity pay differences, is perhaps the key objective behind support for ethnicity pay reporting. Most organisations will want to publicise their results externally, for example, by:  

• publishing their report on their website and including key findings in their annual reports 
• uploading their core statistics onto websites that enable organisational comparisons.  

The CIPD’s race inclusion strategy reports recommend the following on internal communications:  

• Develop a communications and involvement strategy as part of your ‘comprehensive and well-resourced race inclusion strategy’ to communicate aims and achievements, to share the organisation’s diversity data and to explain the race equality and inclusion and diversity (I&D) strategy. 
• Collect, analyse and publish a framework of relevant ethnicity data and statistics across all the HR processes in your organisation, including pay and reward.
• Communicate your action plan for tackling any ethnicity-based disadvantage that your data exploration reveals.  
• Support senior leaders to lead the way in initiating positive conversations about race, to be clearly visible as part of the I&D strategy and to communicate the importance of race and ethnicity in the organisation.  
• Support people managers to manage their teams in a non-discriminatory fashion and support the career development of those who report to them. 
• Establish employee network groups to facilitate conversations about race and other diversity strands in a safe space.  
• Consult employees, for example via employee network and ethnicity groups, on the most appropriate individual and collective terminology to use in relation to ethnicity. 

Much of the communications advice contained in the CIPD’s Gender Pay Reporting Guide is highly relevant to ethnicity pay reporting. It suggests a specific communications plan to inform and ensure consistency. It typically needs to contain:  

• Your aims. 
• Your key messages.  
• A breakdown and analysis of your key audiences.  
• Planning actions and messages through each of your key channels of communication, including their timing and sequencing.  

You will also want to consider the best way of presenting your ethnicity pay data. Stating figures in pounds and pence is often more helpful than percentages, while illustrations and visuals sometimes work better than long paragraphs of text.  

As with any form of publicity, there are risks with greater pay transparency. They include:  

• Not understanding the ethnicity pay gap: It’s important that your spokespeople understand what the ethnicity pay gap is, how it compares externally and how it plays out in your organisation. It’s also important that your employees understand what ethnicity pay reporting is, its limitations, and why you are using eight different measures. 
• Inadequate data: If you are missing key data, say so, but also say what you intend to do to ensure this data is captured for future reports. 
• Being unprepared for questions and questioning of your ethnicity pay report and the data it contains: Even if you don’t think you are going to be questioned, it’s a good idea to be prepared. Draw up a Q&A and ask internal stakeholders for their input.  

It is good practice to determine a process for reviewing progress and to allocate responsibilities, directly involving your leaders and managers and holding them accountable. Common progress measures obviously include any overall reduction in ethnicity pay gaps, but can also include:  

• Progress over a number of years towards a set target ratio of senior posts being filled by employees from ethnic minority groups.
• Progressing towards a target percentage of ethnic minority staff in particular employee populations, such as professors in universities or traders in financial institutions; improving the ethnicity mix in recruitment and promotions is another target to consider.
• Monitoring the impact of specific initiatives, such as anonymous recruitment, representative selection panels on the mix of internal promotions, the removal of performance-related pay on pay differences within grades, and so on.  

Principal associate at the Institute for Employment Studies.