We observed how financial services firms are designing and embedding diversity and inclusion (D&I) strategies. We present our findings on these strategies and an overview of initiatives to improve diversity and inclusion, which industry leaders can consider in reviewing their own diversity and inclusion strategies.
Who this will interest
This review will be of interest to all regulated firms.
What you need to do
We encourage all regulated firms to consider these findings in the development of their diversity and inclusion strategies and practices.
Diversity and inclusion are essential for healthy firm cultures, enabling firms to deliver better outcomes for consumers and markets. We want to see an inclusive industry where the most capable people are able to progress, no matter what their background, and where diversity of thought is valued. Diversity and inclusion, founded on a culture where it is safe to speak up, is essential for firms to have healthy cultures that help to deliver consumer protection and market integrity. Although there has been progress over the last few years and most firms are publicly committed to change, there is still much to be done.
In July 2021, we published a joint Discussion Paper (DP) with the PRA and Bank of England. In this DP we discussed the current state of diversity and inclusion in the industry, set out the case that more progress advances our objectives, and proposed some areas for potential policy intervention. We will consult on these proposals in 2023.
We decided to get a better understanding of the current state of diversity and inclusion approaches in regulated firms. This work had 3 goals:
- To give firms and others a picture of the current position, allowing leaders to consider where initiatives might be relevant in their own firms.
- To encourage further industry action.
- To help us to develop a supervisory approach that we can use as the basis for future engagement with firms.
This review presents the findings from our qualitative research and Appendix 1 sets out some evidence for the effectiveness of actions, but it is not intended as guidance. Where our observations refer to ‘firms’, we are referring to firms that were part of this review.
1. What we did
We chose a sample of 12 ‘fixed’ (generally larger) firms across multiple sectors. We asked each firm for some basic information, including their diversity and inclusion policy and strategy, if they had them, their targets or goals and any data that they used. We also requested a 90-minute structured interview with each firm. We asked all firms to make a senior leader available for this interview, in addition to any specialists that they wanted to include.
We selected firms based on their gender pay gaps. We chose 8 firms with large pay gaps and 4 with relatively small pay gaps. However, as the largest pay gaps are disproportionately found in the investment banking and asset management sectors, we adjusted our sample to cover a wider range of sectors that would reflect the range of firms that we regulate.
In 2021, we issued a pilot data survey to understand better the diversity and inclusion data that firms were collecting. In this review, we are also providing some of the results of our pilot data survey (see Appendix 2).
2.1. General observations
We found a surprising degree of consistency among the firms we spoke to. All were early in the development of their approach on diversity and inclusion, typically having started serious efforts in 2019 or 2020. Some firms had made more progress than others but there was generally little correlation between how developed the firms’ approaches were and the scale of pay gap.
Almost all the people we spoke to were committed and passionate about making progress. There are a number of thoughtful initiatives underway. But many firms’ strategies were generic and did not take a holistic view. They lacked both a clear articulation of purpose and actions oriented to achieving their goals. Firms were not fully capitalising on the data they collect to identify the best remedies, nor tracking which remedies are most effective.
Very few firms seemed to have understood diversity and inclusion as a fundamental culture issue. Generally, we found much less understanding of and focus on building inclusive cultures than on actions to measure diversity and address specific issues.
None of the retail firms that we spoke to had undertaken substantial work on the diverse needs of their consumer base, though a few had recognised the need for this.
2.2. Key points
The firms that we spoke to are most focused on addressing gender representation, with ethnicity starting to receive more attention. Other demographic characteristics receive much less attention.
On both gender and ethnicity, firms tend to focus most on improving representation at senior leadership level. This is despite data showing that the biggest drop-off in representation is from junior to middle management grades. Such focus, in isolation, risks creating a culture where firms attempt to ‘poach’ diverse senior talent rather than develop their own pipelines. This is not a sustainable approach and is unlikely to bring meaningful, long-lasting change.
Firms’ diversity and inclusion strategies are not consistently based on a clear diagnosis of their specific circumstances and challenges. This means actions and initiatives may not be appropriately focused. Firms are also not systematically tracking the effectiveness of these measures and initiatives. This leads to a lack of understanding about what really works. Without a strategy informed by a diagnostic process and better tracking of initiatives, some firms risk expending considerable resource without seeing meaningful results.
There is wide variation in data quality. Firms with better diversity data had a better understanding of their position and were better placed to decide which actions to take. This variation was largely the result of differing levels of success with staff declaration rates. Firms with the best declaration rates have worked hard to achieve this, with focused initiatives to build trust and understanding, and optimising touch points with staff.
Poor data quality also affected firms’ abilities to carry out intersectional analysis to understand the experiences of different groups. So they were not able to design or implement targeted interventions to address these issues. There is a risk that this leads to patterns or trends being missed.
The specific initiatives firms told us about included a number that we felt were likely to have a positive effect. But we also saw an overreliance, in some firms, on measures such as training, network groups and allyship, which although important, will not alone bring about the kind of systemic change needed.
Most firms told us that senior managers were accountable and that diversity and inclusion goals could affect pay and bonuses. But it was much less clear in many cases how this worked in practice.
Firms that are part of international groups had generally adopted a group-wide international strategy, without tailoring it to the circumstances of the UK organisation or the characteristics of the UK. These firms typically had less ambitious and well-defined strategies and were often reliant on global, rather than UK-specific, data.
2.3. Commitment to diversity and inclusion
Most firms approached the work positively. However, there were differing levels of commitment from firms, both to diversity and inclusion in general and specifically in their participation in our research. Some firms were reluctant to make senior business leaders available as we requested. Where they did, we spoke to people with a strong personal commitment to diversity and inclusion. But it was not clear to what extent the enthusiasm of the people we spoke to is more widely reflected across their organisations. In a few cases, they acknowledged that some parts of their organisations were harder to reach on this issue.
By and large, firms were open and candid about the challenges they are facing and focused on achieving meaningful progress. However, we saw several instances where firms focused almost exclusively on gender representation at senior levels because there are external targets and expectations for it. This suggests that a compliance approach, rather than a genuine commitment to diversity and inclusion, is driving some strategies.
2.4. Current performance
Most firms in our sample collected diversity data on their employees across all grades. In all cases where they had analysed this data in detail, it was notable that the step from junior to mid-level roles is where representation falls away most steeply, both for women and ethnic minorities. In the cases we saw, representation at senior levels is only marginally lower than at middle levels. This means that internal talent pipelines for senior representation will be limited, leading firms to look externally for top talent. This results in firms ‘cannibalising each other’, as one interviewee put it.
Large gender pay gaps persist across the industry, and this is particularly marked in some sectors. There is little sign that action to close these has yet been effective (see Appendix 1 on effectiveness of actions). However, even in sectors where pay gaps are most pronounced, some firms display relatively smaller pay gaps. We found bonus gaps to be even wider than hourly pay gaps. We conclude that this is indicative of the fact that the highest bonuses are paid at senior levels, where women and ethnic minorities are still under-represented.
Some firms had broken down ethnicity representation beyond a simple White/ethnic minority split. Where they had, the data showed there were clearly divergent outcomes for different ethnic minority groups.
Few firms have taken steps to address social mobility. Where they have, this has focused on the entry points, with less attention to the cultural experience of employees from less socially privileged backgrounds. This may also be a contributing factor to the lack of progress for some ethnic minorities.
Work around sexual orientation is often limited to supporting employee network groups and performative actions (eg, support for Pride). Similarly, few firms had given serious consideration to disability. Very few firms have paid attention to neurodiversity.
Firms we spoke to generally weren't considering whether there were compounded issues for people belonging to more than one minority group that could lead to disadvantages (‘intersectionality’). The Financial Services Culture Board found wide divergences in the experience of White women and women from ethnic minorities, for example.
2.5. Use of data
We found considerable variation in the range of data that firms are collecting and the level of analysis conducted on that data. Some gather detailed breakdowns of gender and ethnicity data by grade. Others don’t, in part due to poor employee declaration rates, and without clear strategies to improve these. Our findings were consistent with the results of our pilot data survey (see Appendix 2).
Few firms have actionable data beyond gender and ethnicity. Where firms had attempted to gather data on characteristics like disability and sexual orientation, they had all seen lower declaration rates than for ethnicity. We saw better declaration rates for data collected at recruitment/ onboarding stage compared to that from existing employees. Where firms achieved higher declaration rates, this was usually because they had made efforts to increase trust and make effective use of employee engagement. In one example, an action as simple as showing how to update diversity data at team meetings improved declaration rates.
We also found differences in the level of employee declaration which firms believe constitutes a reliable data set on which a strategy and targets can be set. This means that, in some cases, firms with worse declaration rates are doing more than those with better rates. It may be possible to draw tentative conclusions about representation even with lower declaration rates.
Crucially, it was not clear that firms - even those with the best data - are making full use of their data insights to inform their strategies. This is likely to mean that their interventions are not targeted on the most important issues. Firms were not generally making significant efforts to get to the heart of the reasons behind their issues in representation. For example, few firms were using detailed data about promotions processes or making use of qualitative feedback such as exit interviews to provide insight into the numbers.
2.6. Effectiveness of strategies and targets
Most firms did not have strategies that clearly linked diagnosis, action and measurement. The level of detail covered in strategies was variable. Many firms had high-level strategies that would benefit from more definition and struggled to give clear examples of how they were going to reach their goals. Many strategies were not specific to the firm and its particular issues. For example, although firms in different sectors have very different job roles and cultures, potentially raising specific issues, we did not see this generally reflected.
Many firms seemed unclear about their business rationale for better diversity and inclusion. Only 1 firm had made a clear connection with diversity of thought or recognised the potential benefits that this could bring to its business. Without a clear understanding of why firms are undertaking these efforts, there is a risk that diversity and inclusion is seen as an optional extra or that staff become fatigued and disengaged by ongoing initiatives.
Firms had the most developed strategies for gender and were most likely to set targets for it. We think this is, in part, due to the availability of data and the influence of high-profile initiatives, notably the Women in Finance Charter and 30% club. While we understand the need for firms to be realistic in what can be achieved, we felt some of these targets lacked ambition.
After gender, ethnicity received the most focus, although data availability stopped several firms from setting specific targets. As a result, ethnicity strategies often lacked the same level of focus, in terms of tangible actions, measures and accountability. Other characteristics received the least attention, with only a small number of firms in our sample articulating their importance and setting out measures to support them.
Firms did not consistently measure the effectiveness of individual initiatives. Although further effort would be needed to put evaluations in place, without them there is a risk of wasted effort and unintended negative consequences.
We found that some firms had launched numerous initiatives but had yet to see substantial improvements. We have 3 hypotheses as to why this might be:
- Diversity and inclusion initiatives take longer to deliver a visible impact than expected.
- Diversity initiatives alone, without meaningful cultural change driven from the top to embed them and drive inclusion, will not tackle diversity effectively.
- Some diversity and inclusion initiatives are not effective in delivering change.
These possibilities have very different implications. So it will be important for firms to understand the reasons where initiatives are not delivering change.
2.7. Specific initiatives
We found a lot of willingness to pursue initiatives for change. Some of the most positive initiatives we heard about were around supporting people returning to work. A specific example is ensuring that people who go on long term leave (eg, maternity) have their performance grades carried forward, to reduce the risk that their pay and future opportunities are negatively affected. Such initiatives require continued promotion to ensure awareness (both for the staff in question and managers) and to foster trust that their careers are not going to be adversely impacted.
The firms we spoke to had given more thought to recruitment processes than other phases of the employee life cycle (which include performance management, lateral opportunities to gain more experience, promotion, and the reasons for departures). They had considered how the process affected the diversity of their candidate pool and candidates’ success through the process. One firm had carried out a detailed recruitment audit which produced a list of detailed recommendations that they could implement.
But we also found firms’ strategies often heavily relied on measures such as training courses and information hubs. We believe that there is a role for these initiatives but that, if they are the centrepiece of a strategy, research suggests they are unlikely to tackle ingrained, systemic biases or bring about the kind of culture change needed (see Appendix 1).
Most firms had established employee network groups (also called employee resource groups). Some firms appeared to have devolved considerable responsibility to network groups and to the role of allyship. Network groups may be most effective where they are used as a source of culture feedback, giving essential insights about employees’ real lived experiences, and are empowered. They can have great value in providing challenge and can be used strategically to represent the interests of their members and ensure their views are considered. A few firms have allocated each group an executive sponsor, which is likely to be an effective way of giving them influence. Without this empowerment, they can still perform a useful role in providing support and opening up discussion about diversity and inclusion but they are unlikely to lead to significant change on their own.
Sponsorship and mentoring for under-represented groups were often used as tools to improve representation. We thought these were worthwhile actions, though there are limits to the number of candidates these programmes can feasibly serve and we did not speak to a firm that had tried to scale these up. There was little awareness that because selection for these programmes was discretionary, they could be subject to bias.
Some evidence and commentary on effectiveness of actions is in Appendix 1.
2.8. Governance and accountability
Most firms told us that their Boards were engaged and many were able to give examples of challenge from the Board. However, the scope of our review did not give us sufficient evidence to confirm that Boards were fully engaged across our sample.
Many firms said senior managers would be held accountable for progress and that it was part of their objectives. We saw a range of approaches among the firms in our sample. This ranged from firms that do not have any link to performance, to those who had diversity and inclusion as a performance objective or even as a key criterion in their performance grading. But even in the last case, it was often unclear how progress to goals would actually affect a performance grade or reward, and many firms could not give examples of situations that would call for a tangible adjustment to reward.
Some firms were candid with us that certain business areas are harder to influence than others. From what we were told, there were no obvious patterns in this. It is possible that it is due to sub-cultures or differing levels of commitment from senior leaders.
2.9. Inclusion and culture
Most of the firms we spoke to value inclusion and want to take steps to develop inclusive cultures. But, in many cases, there was relatively little activity to build them. Few firms talked about the behavioural biases that affect inclusion or the role of systemic discrimination. Some firms did not seem to have recognised that fundamental issues, such as psychological safety and welcoming different perspectives, are critical to an inclusive culture. Interventions were usually limited in scope and likely effectiveness (eg, voluntary online training for managers).
We understand that measuring inclusion effectively is challenging. Most current approaches rely mainly on staff survey questions. Surveys have a role, but they have limits. They can give a high-level snapshot, but they cannot help with understanding culture in any depth. Firms used a wide range of questions as indicators of inclusion. A few firms were unable to explain what factors make up their inclusion metric. No firm we spoke to mentioned the Financial Services Skills Commission’s Inclusion Measurement Guide.
In general, firms with relatively high employee survey scores seemed to lack interest in asking themselves what they might be missing. Few firms have attempted to analyse their data by diversity characteristics, which could give them valuable insight on the lived experience of their employees. None of the firms we spoke to had used data to understand organisational dynamics in detail (for example, internal interviews or work allocation).
Generally, firms were making relatively little use of qualitative feedback, such as focus groups, using network/employee resource groups and exit interviews. Collecting this information could help organisations understand why there are divergent outcomes, not just where they occur.
In our view, firms are likely to struggle to make sustainable, meaningful change without greater attention to culture, including a long-term plan to deliver change, backed by senior level commitment.
3. Next steps
Firms that were part of this review have received written feedback letters. We will follow up with them through our usual supervisory activities to assess how they have considered our feedback.
We encourage all regulated firms to consider these findings and use them to assess their current diversity and inclusion strategies and practices.
4. Appendix 1: Effectiveness of actions
We note above that we found firms' strategies and actions are not consistently based on a clear diagnosis of their specific circumstances nor an evaluation of their effectiveness. We also note that there is a mixed evidence base for the effectiveness of some actions to address D&I. This section covers a selection of evidence on some of the most common actions firms told us about.
- Supporting career progression for staff returning to work
- Encouraging parental leave for men as well as women
- Mentoring and sponsorship
The Government Equalities Office lists these first three as promising actions for closing the gender pay gap. Mentoring and sponsorship is likely to have wider applicability.
- Setting specific business area targets
The Equalities Office also lists specific target setting as a promising action. They don’t identify targets for individual business areas but we think it is likely that this will help to focus action where it is most effective.
- Extending the reach of entry level recruitment beyond elite universities
It is highly plausible that this will increase the range of social and ethnic backgrounds of new employees but monitoring will be needed to ensure that it is working as expected, including retention of employees.
- Diverse shortlists
The Equalities Office notes that, in respect of gender, there is a positive effect as long as there is more than one woman on the shortlist.
- Diverse interview panels
Diverse panels may be expected to reduce collective unconscious bias in line with the diversity of their members but we are not aware of rigorous research on the topic. All panels are open to groupthink and may be unaware of other biases. Monitoring of effectiveness will be important.
- Anonymised CVs
There is evidence that anonymised CVs reduce gender bias throughout the interview process. The same study found a reverse effect on social mobility, but this has been explained by self-selection of firms participating.
- Mandatory training
The Equalities Office lists a number of types of training for which evidence of a positive effect is poor. Some diversity training can be counterproductive (see also the report of the Behavioural Insights Team; Iris Bohnet, “What Works” 2016). Some other types of training may be valuable but should be used with caution and results evaluated.
5. Appendix 2: Pilot Data Survey
Following the regulators’ Discussion Paper in 2021, we undertook a voluntary survey of firms to understand what diversity and inclusion data they collected and held. The results of the survey provide some support to the findings of our research. We emphasise that the survey was voluntary, so the sample is self-selecting, and that the sample size is small. The results should be treated as indicative and too much reliance should not be placed on them.
We received 158 responses from large firms and 514 from small firms. Large firms are defined throughout as having 251 or more employees; small firms have 250 or fewer.
5.1. Data collected on demographic characteristics
Does your firm collect employee data on any of the following demographic characteristics? (Selected characteristics shown.)
5.2. Setting targets
Does your firm set targets for underrepresented groups for the following?
If yes, which demographic characteristics does your firm set targets for?
5.3. Senior management accountability
Do one or more of your firm’s senior management have a diversity and inclusion objective?
If yes, how is it expressed? (All firms)
Which of the following actions have been taken in your firm to support progress on diversity and inclusion?
5.5. Intersectional analysis
Does your firm analyse data by more than one diversity characteristic at a time to help uncover and address intersectional barriers? (Large firms only shown)