Introduction
This article provides a realist perspective on the relevance of the ‘Business Case for Diversity’ (BCD) (Cox and Blake, Reference Cox and Blake1991; Schneider and Ross, Reference Schneider and Ross1992) to the British National Health Service (NHS). We broadly discuss the BCD in relation to race, not least because the BCD was originally developed here. The BCD has been broadly promoted in the UK for some time by a number of organisations inside and outside government (for example, Rutherford and Ollerearnshaw, Reference Rutherford and Ollerearnshaw2002; DTI, 2003; NAO, 2004; Jones, Reference Jones2006; Cabinet Office, 2008; Chartered Institute for Personnel Development (CIPD), 2005; Equality and Human Rights Commission (EHRC)/Trades Union Congress (TUC)/Confederation of British Industry (CBI), 2008), and also in the NHS (Esmail et al., Reference Esmail, Kaira and Abel2005; NHS Employers, 2009).
While the BCD underpins many recent proposals and initiatives (for example, Shah, Reference Shah2009; Davies, 2011), it has been little discussed in social policy. Bagilhole (Reference Bagilhole2009) contrasts the moral, economic or business and social cohesion cases, and presents a chronology of five eras of equal opportunity and diversity (EOD) (but see Powell et al., Reference Powell, Johns and Green2013). She writes that the business or economic case argues that EOD brings greater efficiency and economic growth to organisations and society in general. She views the 2000s as the ‘fairness tempered with economic efficiency era’.
However, we argue that there are major problems with the BCD. First, it may not be as strong as some of its supporters suggest (for example, Schneider and Ross, Reference Schneider and Ross1992; Robinson and Dechant, Reference Robinson and Dechant1997; Kandola and Fullerton, Reference Kandola and Fullerton1998; Centre for Strategy and Evaluation Services (CSES), 2003). This article does not, however, examine in any detail the empirical evidence for the BCD (see Milliken and Martins, Reference Milliken and Martins1996; Ryan et al., Reference Ryan, Hawdon and Branick2002; Anderson and Metcalf, Reference Anderson and Metcalf2003; Kochan et al., Reference Kochan, Bezrukova, Ely, Jackson, Joshi, Jehn, Leonard, Levine and Thomas2003; Dreachslin et al., 2004; Awbrey, Reference Awbrey2007; Curtis and Dreachslin, Reference Curtis and Dreachslin2008; Shore et al., Reference Shore, Chung-Herrera, Dean, Holcombe Ehrhart, Jung, Randel and Singh2009; Pitts, Reference Pitts2009; Bleijenbergh et al., Reference Bleijenbergh, Peters and Poutsma2010; Heres and Benschop, Reference Heres and Benschop2010; Haas, Reference Haas2010). Instead, we focus on the contingent nature of the BCD (Kochan et al., Reference Kochan, Bezrukova, Ely, Jackson, Joshi, Jehn, Leonard, Levine and Thomas2003). Even if the empirical evidence for the BCD stands in one context, it does not necessarily follow that these findings are transferable to other contexts (cf. Ram et al., Reference Ram, Theodorakopoulos and Worthington2007). We argue that the BCD has been stretched far beyond its original setting to cover a range of very diverse contexts. For example, the original claim was that increasing the ethnic workforce in for-profit companies in the USA would increase profits (for example, Cox and Blake, Reference Cox and Blake1991). However, translating empirical evidence on the BCD from for-profit business in the USA to a public service such as the NHS in Britain is problematic. Moreover, it has been stretched to cover issues as diverse as parliamentary representation, board membership, workforce diversity and leadership diversity. Mechanisms may be rather different for workforce diversity compared to leadership or ‘top team’ or boardroom diversity. Similarly, outcomes may be rather different for parliamentary representation (for example, more inclusive and better legislation) as compared with workforce diversity (such as, higher sales or share value). Finally, the BCD as applied to (say) ‘race’ and ethnicity may operate rather differently when applied to gender or disability.
Pawson and Tilley (Reference Pawson and Tilley1997) provide basic guidance about Programme Theory and CMO configurations. Theories must be framed in terms of propositions about how mechanisms (M) are fired in contexts (C) to produce outcomes (O): what might work for whom in what circumstances. In other words, the relationship between causal mechanisms and their effects is contingent rather than fixed. This article reports on a conceptually driven evidential review and it employs the CMO framework as a theoretical device to explore the relevance of the BCD for the British NHS. First it examines the BCD in general terms, and then turns to diversity in the NHS. It then moves to consider issues of context, mechanism and outcomes related to the BCD by interlacing the general case with specific illustrations drawn from the NHS. While the conclusions support those produced in much of the available literature, the use of realistic evaluation as conceptual frame allows the separate elements of the limitations of the BCD in this setting to be clearly identified for the first time.
The business case for diversity
During the late 1980s, diversity rhetoric in the USA shifted to a business case for supporting workforce diversity (Dickens, Reference Dickens1999; Riley et al., Reference Riley, Metcalf and Forth2008; Herring, Reference Herring2009; Pitts, Reference Pitts2009). The BCD has developed more slowly in Europe, but is gaining ground (Jones, Reference Jones2006). Cox and Blake (Reference Cox and Blake1991) argue that diversity can result in competitive advantage via six routes:
• resource acquisition: companies develop reputations as diversity employers;
• marketing: a diverse workforce understands diverse market segments;
• system-flexibility: tolerance for different cultural viewpoints leads to greater openness to new ideas;
• creativity: organisations with diverse perspectives are more creative;
• problem-solving: diversity promotes better decision-making;
• cost reduction: such reductions include lower staff turnover, reduced absenteeism and increased productivity.
Subsequent commentators give broadly similar sets of explanations, but sometimes in differing numbers, contexts and emphases (see Robinson and Dechant, 2007; Catalyst, 2004). The important point is that diversity is not about doing what is right, but rather achieving organisational results. It accepts that ‘social justice’ is insufficient, and that equal opportunities and diversity (EOD) initiatives have to appeal to self-interest. As Dickens (Reference Dickens1999) maintains, BCD arguments are inevitably contingent, variable, selective and partial. Perriton (Reference Perriton2009) argues that it works to frame, restrict and depoliticise the discussion of gender in the workplace, and downplays issues of discrimination (see also Wrench, Reference Wrench2005; Bagilhole, Reference Bagilhole2009).
Diversity in the NHS
Diversity in employment has been on governmental agendas since the passage of the Disabled Persons (Employment) Act in 1944, but the New Labour government took the biggest active step towards employment diversity, with impetus from the Macpherson Report (Macpherson, 1999). The Race Relations Amendment Act 2000 was supplemented by positive action for key government departments. While some organisations were seen to be inadequately diverse, the NHS was regarded as a public sector exemplar (Straw, Reference Straw1999).
There has been a long history of diversity initiatives in the NHS (Law, Reference Law1996; Oikelome, Reference Oikelome2007; Crisp, Reference Crisp2010; Bach and Kessler, Reference Bach and Kessler2012). According to the NHS Institute for Innovation and Improvement (NHS III) (2009), since the Race Relations (Amendment) Act 2000 there has been a plethora of NHS initiatives designed to promote workforce and managerial diversity. In 2000, The Vital Connection was published, outlining governmental expectations of the NHS in terms of equal opportunities, and was operational until 2010 (NHSE, 2000). This was supported by a framework strategy, and backed by national development programmes such as Positively Diverse (DoH, 2000a) and the Improving Working Lives Standard (DoH, 2000b). In 2004, a ten-point Race Equality Plan was implemented and an Equality and Human Rights Director was appointed. Diversity initiatives have also been established under the ‘Breaking Through’ programme, the National Leadership Council, the Equality and Diversity Council, NHS Operating Statements, the NHS Constitution and the Equality Delivery System. The NHS also features in wider initiatives aimed at diversifying the level of public boards across government (GEO, 2009) and is subject to the Equality Act 2010, which has very significant implications. However, while there has been a long history of gender and black and minority ethnic (BME) diversity in the NHS (Kalra et al., Reference Kalra, Abel and Esmail2009), this ‘proportional’ version of diversity, where workforces are broadly representative, was never matched by ‘structural’ diversity at senior levels (Esmail et al., Reference Esmail, Kaira and Abel2005; Johns, Reference Johns2006; NHS III, 2009; Crisp, Reference Crisp2010). As Trevor Phillips stated in The Guardian in 2004:
People from [BME] communities make up 35 per cent of its doctors and dentists, 16.4 per cent of the nurses and 11.2 per cent of non-medical staff. However at the top of each NHS organisation, the boss is almost always White. There are more than 600 NHS trusts, health boards, local health boards and health and social services boards in England, Scotland, Wales and Northern Ireland, and fewer than 1 per cent of them have a Black or minority ethnic chief executive. The contrast between snow-capped summit and the mountain base could hardly be more stark. (Carvel, Reference Carvel2003)
CMO perspectives in the BCD
Milliken and Martins (Reference Milliken and Martins1996) write that research on diversity is challenging because it spans disciplinary boundaries, assesses the effects of various types of diversity, focuses on many different dependent variables and employs a wide range of groups and settings. In short, the diversity literature is diverse. Realist perspectives stress that even if strong evidence exists to support BCD in different contexts, it does not follow that it ‘works’ for the NHS. In this section, we explore a number of issues concerned with context, mechanisms and outcomes.
Context
There are two main contextual issues. First, most of the literature is based on for-profit businesses, with few studies focusing on the public sector and health care. Little published research has considered the link between diversity management processes and public sector performance (Naff and Kellough, Reference Naff and Kellough2003). Pitts (Reference Pitts2009) claims that a handful of studies have addressed the diversity–performance link in public agencies. Most ‘case studies’ and ‘top tips’ are taken from the private sector (Cabinet Office, 2008; NHS III, 2009). According to Weech-Maldono et al. (2002), research on diversity management practices in healthcare organisations is scarce, numbering three studies.
In 1995, the Commission for Racial Equality (CRE) shaped the BCD to the requirements of the public sector via the ‘quality case’ (CRE, 1995). Its core was that where business is about profit in a competitive marketplace, this does not reflect the public service ethos. The motivating agenda for the quality case was the quality of provision to a diverse population, in effect serving diversity requires diversity. A related development for local government is the concept of ‘representative bureaucracy’. It is much wider, taking into account other identity dimensions than ‘race’ and ethnicity and using democratic rights as its principal justification (Andrews et al., Reference Andrews, Boyne, Meier, O'Toole and Walker2005). It is assumed that ‘passive representation’ (mirroring the service area demography) will lead to active representation, i.e. that people will actively represent ‘their’ communities. This may have underpinned New Labour's adoption of all-women short-lists in 1997 (the practice of fielding female candidates in safe parliamentary seats). However, this is a notion that has been questioned (Cashmore, Reference Cashmore2002; Johns, Reference Johns2006).
A further contextual factor is that most studies are from the USA. There are few studies in Europe (CSES, 2003) and the UK (Rutherford and Ollerearnshaw, Reference Rutherford and Ollerearnshaw2002; Riley et al., Reference Riley, Metcalf and Forth2008). It is difficult to transfer findings based on the different ethnic population structures of the USA and Britain, and their different historical constructions (Ram et al., Reference Ram, Theodorakopoulos and Worthington2007). Census data are obviously flawed, carrying combinations of ‘racial’, ethnic and national origin categories; nevertheless, they are the best sources available. According to the 2010 census conducted in the USA, the population included 79.6 per cent classed as ‘White’, with 12.9 per cent ‘Black’, 4.8 per cent ‘Asian’, ‘Native Hawaiian’ and ‘Pacific Islanders’, 1.0 per cent ‘American Indian’ and ‘Alaskan Native’ and finally 1.7 per cent reported as two or more ‘races’ (July 2009, US Census Bureau). By contrast, the UK in 2001 was constructed in the following way: 92.1 per cent ‘White’, 3.5 per cent ‘Asian’, 2.0 per cent ‘Black’, 1.2 per cent ‘Chinese and Other’ and 1.2 per cent ‘Mixed’ (see Craig et al., Reference Craig, Atkin, Chattoo and Flynn2012 for more details). A smaller BME population, differently constructed and with different historical relationships mean a quite distinct context, which suggests transferring policy initiatives is risky.
In addition, the legislative history and policy development is very different when we compare the UK with the USA. Affirmative Action (AA) which was initiated during the 1940s was arguably much more radical than the legislative responses to equal opportunities in the UK, although the AA apparatus has been eroded since the 1980s (Cohen and Sterba, Reference Cohen and Sterba2003).
By contrast, the UK has followed a path guided by formal equality of opportunity underlined by equal treatment; positive action was available from the 1970s but remained virtually latent (Bagilhole, Reference Bagilhole2009; Craig et al., Reference Craig, Atkin, Chattoo and Flynn2012). While Saunders (Reference Saunders2011) has argued that positive discrimination is common in the UK, the reality is that until the Disability Discrimination Act 1995 it was illegal to treat people differently except via positive action (outreach recruitment, goals and targets, and targeted training programmes). In the wake of the Macpherson Report (Macpherson, 1999), there was an intensification of positive action efforts and the Equality Act 2010 promoting identity as a tie-break criterion is taking the UK along a more radical path. It is clear that there are important contextual differences in population structures and very different legal and policy traditions that may hamper policy transfer between the US and the UK.
Mechanisms
Much of the BCD literature is unclear on mechanisms, and so specifying clear CMO configurations is problematic. The first issue is the extent to which routes to competitive advantage (Cox, Reference Cox1991; Cox and Blake, Reference Cox and Blake1991) developed in the context of for-profit business, apply to public service. Robinson and Dechant (Reference Robinson and Dechant1997) report a survey in which human resource executives from fifteen Fortune 100 companies were asked to identify the primary business reasons for diversity management. The top five reasons were: better utilisation of talent (93 per cent); increased marketplace understanding (80 per cent); enhanced breadth of understanding in leadership positions (60 per cent); enhanced creativity (53 per cent); and enhanced team problem-solving (40 per cent).
The BCD has increasingly been embraced in the NHS (Esmail et al., Reference Esmail, Kaira and Abel2005; Crisp, Reference Crisp2010). It is strongly supported by NHS Employers (2009) who suggest that the evidence for the BCD and its NHS relevance continues to grow. According to NHS Employers (2009), diversity enhances:
• an organisation's reputation – this attracts talent from all communities, helping to meet service delivery needs;
• staff recruitment and retention – valuing diversity enables employers to recruit and retain the best people;
• productivity – staff perform better in diversity organisations that are committed to employees’ well-being; and
• mitigating organisational risks – effective diversity management limits the risk of legal challenges.
The benefits of a more diverse leadership have been identified as improving morale and productivity, exploiting latent creativity and also generating trust within service populations (Kalra et al., Reference Kalra, Abel and Esmail2009). According to the NHS III (2009), interviews with Primary Care Trust (PCT) staff and workshops identified a variety of potential benefits of applying the BCD to the NHS. These benefits focused on financial gains (including reduced litigation), improved performance (better quality of provision) and a potential reduction in health inequalities.
If you speak to the finance director, emphasise the pure business case. If it is the public health representative then emphasise the impact on health inequalities. If you are speaking to frontline staff delivering the service, then emphasise making a difference to the way care is delivered; there is a business case for workforce diversity, but it is sold to different people in different ways. (Interview reported in NHS III, 2009: 21)
NHS Employers (2009) refer to a case study by University College London that claimed an NHS employer with 3,000 staff and a turnover of £150 million, treating 500,000 patients a year would save £3.8 million annually using policies informed by the BCD.
There are no obvious reasons why ‘labour’ factors (supply, talent, recruitment and retention, morale) should be irrelevant for the BCD in public services. However, the marketing argument – that a diverse workforce understands diverse market segments – may be more problematic. Companies with more BME employees that increase sales to BME customers will maximise revenue. The same might be true for an NHS Trust with more BME employees that treated more BME patients. Yet a commissioner (PCT or CCG) with more BME employees that tapped unmet needs in BME communities would increase its costs (at least in the short term NHS horizon). This would meet the NHS's equal access mission, but it also indicates that the simplistic transfer of the BCD is problematic for a public service.
The second issue involves the contested conceptualisation of diversity (Butt, Reference Butt2006; Bagilhole, Reference Bagilhole2009; Herring, Reference Herring2009; Craig et al., Reference Craig, Atkin, Chattoo and Flynn2012). Dandeker and Mason (Reference Dandeker and Mason2001) point out that the concept of representativeness conceals four distinct ideas: the statistical, the delegative, the symbolic and the value. ‘Diversity’ is often used to mean different things (Dudau and McAllister, Reference Dudau and McAllister2010), for example cultural diversity, linguistic diversity, ethnic diversity and gender diversity. It is also often used interchangeably with ‘race equality’ (Butt, Reference Butt2006). According to the CSES (2003), a ‘diverse workplace’ is difficult to define. There is no accepted way of distinguishing between diverse and non-diverse workplaces; similarly, the CIPD (2005) states that the definition of diversity is almost as diverse as the subject itself, making the interpretation of findings highly judgemental.
Broadly, there is a difference between identity and cognitive diversity (Page, Reference Page2007). Diversity is usually defined as a ‘variation of social and cultural identities among people existing together in a defined employment or market setting’ (Cox, Reference Cox1991) and separated into observable and non-observable characteristics (Milliken and Martins, Reference Milliken and Martins1996). The former includes characteristics such as gender, ‘race’, ethnicity and age. Non-observable characteristics include cultural, personality traits, cognitive, functional and technical differences. Thus, the concept of diversity in the workplace is increasingly represented as ‘the varied perspectives and approaches to work that members of different identity groups bring’ (Thomas and Ely, Reference Thomas and Ely1996).
Working on a joint project, the EHRC/TUC/CBI (2008) carried out forty-five in-depth interviews with employers, and found that the majority defined diversity in terms of workforce headcounts. Conversely, a minority said it did not matter how many BMEs or women they employ, rather it is the way people are treated and that procedures are non-discriminatory that are important – one employer conceded: ‘Just because the numbers are right it doesn't mean the culture is sorted.’
This leads to problems in the relationships between identity and cognitive diversity, and in examining dimensions in isolation. For example, would an all-female white workforce be ‘diverse’? If workforce diversity is examined in terms of the wider population benchmark, can a ‘minority’ group be ‘over-represented’ or a workforce ‘too diverse’?
Third, many national and regional documents (for example, DoH, 2009) are unclear about whether the benchmark for diverse leadership is the local population or the workforce. To use the example of gender, benchmarks would be different if they are based on population (about 50 per cent female) or NHS workforce (about 70 per cent female). A further problem concerns the benchmark of national or local population. Government benchmarks for public boards reflect uniform national figures (GEO, 2009), but most NHS documents use local figures. Gauss and Jessamy (Reference Gauss and Jessamy2007) argue that the best means of justifying diversity is to link it to local demography; employers should aim to reflect the local service population (what Johns (Reference Johns2006) referred to as local proportional diversity). However, they go on to say that if a local area is not diverse, for instance is essentially ‘White’, then there remains a moral duty to achieve workforce diversity. Yet, if it is about morality, the BCD becomes unnecessary.
Fourth, it is not clear whether the key mechanisms relate to workforce or leadership diversity. It appears unlikely that a diverse workforce alone delivers organisational benefits; otherwise, the NHS would already be reaping them. If leadership diversity is the key, it is not clear if diversity is more important in the executive and non-executive directors on boards or senior managers in general. According to Roberson and Park (Reference Roberson and Park2007), both researchers and practitioners have also assumed that leadership diversity equals financial success. They suggest that diversity in senior management helps to align business strategies with demographic and market trends to achieve organisational growth. According to upper echelons theory, a firm's leaders significantly determine performance given their organisational power (Finkelstein and Hambrick, Reference Finkelstein and Hambrick1996). However, they state that little attention has been given to the performance effects of ‘racial’ diversity in top management teams. In short, it is not clear if different or similar mechanisms underscore the competitive advantage of a diverse workforce and a diverse leadership.
Even if the broad BCD is accepted, we do not know what works or why (cf. Pawson and Tilley, Reference Pawson and Tilley1997). We require clearer ‘programme theory’ about specific interventions that work to deliver specified outcomes in particular contexts. Kalev et al. (Reference Kalev, Dobbin and Kelly2006) write that lists of diversity management ‘best practices’ have proliferated recently, but that these are best guesses. We know a lot about the disease of workplace inequality, but not much about the cure. Similarly, Curtis and Dreachslin (Reference Curtis and Dreachslin2008) argue that the literature on diversity interventions is too limited to provide significant human resources guidance. Although advocates of diversity management contend that interventions work and cite best practices, empirical support is lacking.
There are only a few high-quality academic studies that predominantly focus on the USA. Kalev et al. (Reference Kalev, Dobbin and Kelly2006) examined the effects of seven common diversity programmes: AA plans, diversity committees and taskforces, diversity managers, diversity training, diversity evaluations for managers, networking programmes and mentoring programmes on the representation of ‘White’ men, ‘White’ women, ‘Black’ women and ‘Black’ men in private sector management. They claim that diversity committees and diversity staff have been quite effective. On the other hand, diversity training was not very effective and showed adverse effects among non-contractors. Kalev (Reference Kalev2009) provides strong support for the argument that restructuring work to weaken job segregation improves the access of women and minorities to management. Both self-directed work teams and cross-training programmes have significant positive effects on the odds that managers are ‘White’ women, ‘Black’ women and ‘Black’ men, and a negative effect on ‘White’ men's odds of being in management. In contrast, programmes that do not expand workers’ opportunities to transcend job boundaries, problem-solving teams and job training do not have these effects. The results also indicate that ‘racial’ barriers are more resistant to change than gender barriers. However, there are few comparable studies in the UK. Moreover, programmes tend to be used in combination rather than in isolation, and there are debates surrounding the differences and different level of effectiveness between EOD and ‘Managing Diversity’ approaches (for example, Liff, Reference Liff1996; Kirton and Greene, Reference Kirton and Greene2007; Bagilhole, Reference Bagilhole2009; Ashley, Reference Ashley2010).
Outcomes
There are two major problems relating to outcomes. First, it is not clear whether outcomes refer to (process, intermediate) outcomes relating to the workforce or final outcomes relating to the population. The BCD refers to the ‘bottom line’ of organisational performance, which suggests that relevant outcomes should be seen in terms of greater efficiency, or effectiveness in terms of increasing population health or reducing health inequalities. According to the NHS III (2009), diversity of senior management should be seen as a means to an end, rather than an end in itself: unless such links between diverse senior management and successful health organisations are made, it will be difficult for the barriers described to be overcome. However, some commentators point to the outcome of greater leadership diversity, which can be seen as a means to an end rather than an end in itself.
Second, unlike the research on the effects of diversity on individual and group-level performance, where there are a large number of studies, there are relatively few studies assessing the relationship to the performance of the organisation (Jayne and Dipboye, Reference Jayne and Dipboye2004). Similarly, according to Kochan et al. (Reference Kochan, Bezrukova, Ely, Jackson, Joshi, Jehn, Leonard, Levine and Thomas2003), there is little research conducted in actual organisations that addresses the impact of diversity or diversity management practices on financial success. While there are several laboratory experiments that test specific diversity–performance hypotheses, there are few real world studies and fewer still that use objective performance measures.
A number of studies have explored the relationship between leadership diversity and organisational performance (for example, Catalyst, 2004; Krishnan and Parsons, Reference Krishnan and Parsons2008). Richard et al. (Reference Richard, Barnett, Dwyer and Chadwick2004) suggest that, rather than a simple linear relationship, management of diversity and firm performance may have a curvilinear relationship moderated by elements of strategic stance. Roberson and Park (Reference Roberson and Park2007) claim that previous research that has examined the relationship between top management team diversity and organisational performance has found equivocal results; they used longitudinal data for 100 US firms to test hypotheses related to the effects of diversity reputation and leader ‘racial’ diversity on financial outcomes. The results showed a positive relationship between diversity reputation and book-to-market equity, and a curvilinear, U-shaped relationship among leader diversity and revenues, net income and book-to-market equity, with the low point at about 22 per cent. In short, firm performance declines with increases in the representation of ‘racial’ minorities in leadership up to a point beyond which further increases in diversity are associated with increases in performance.
In the UK, Esmail et al. (Reference Esmail, Kaira and Abel2005) claim that in order to improve the quality of services to BME patients the NHS needs to embrace diversity as a central facet of its business plans. However, this is a very narrow claim – that leadership diversity would only benefit BME patients. They claim that there are three dimensions for an NHS BCD: improved patient care, improved performance and demographic imperatives, and that one of the strongest reasons to adopt a diversity approach is to reduce ethnic health inequalities. Yet no evidence is presented for this claim. Moreover, it is too vague: much of the BCD focuses on the workforce, and arguably the currently diverse NHS workforce has not fully delivered this objective.
Discussion
Cox and Blake (Reference Cox and Blake1991) argued that the logic of ‘valuing diversity’ is rarely made explicit. Richard and Johnson (Reference Richard and Johnson2001) term it a ‘theoretical abyss’ (but see Curtis and Dreachslin, Reference Curtis and Dreachslin2008). Pitts (Reference Pitts2006) points to the atheoretical nature of diversity research. Very few scholars have attempted to establish theoretical frameworks for diversity in public sector organisations (Thomas and Ely, Reference Thomas and Ely1996). In addition, much of the work on diversity stems from a normative view that any diversity leads to positive consequences, and the ‘diversity-as-panacea’ view has crept into the scholarly literature. With few exceptions (Wise and Tschirhart, Reference Wise and Tschirhart2000), research has not attempted to assess the real value of diversity.
Esmail et al. (Reference Esmail, Kaira and Abel2005) argue that the diversity approach is organisationally universal, but most of the literature claims that diversity is contextual, which suggests that a simplistic policy transfer is untenable. Many commentators argue that organisational context is crucial (Richard, Reference Richard2000; Richard et al., Reference Richard, Barnett, Dwyer and Chadwick2004; Andrews et al., Reference Andrews, Boyne, Meier, O'Toole and Walker2005; Awbrey, Reference Awbrey2007). This has a number of implications. Shore et al. (Reference Shore, Chung-Herrera, Dean, Holcombe Ehrhart, Jung, Randel and Singh2009) state that research on different dimensions of diversity has mostly evolved independently. We do not know if conclusions for one dimension (for example, gender) can be applied to others (such as disability), or how the dimensions interact (for example, BME women) (cf. Pitts, Reference Pitts2009). Most research examines diversity at the workforce level, and less at the leadership level. It is not clear whether conceptual arguments or empirical conclusions can be applied to the different levels. Most work has been carried out in the USA, and the relevance for the UK is unclear. Most work has been on business, with few studies on public services and health care. This suggests that the work on representative bureaucracy might form a more solid theoretical foundation than the BCD, and that more stress should be placed on studies in public services and health care. In short, context truly matters (Haas, Reference Haas2010; Bleijenbergh et al., Reference Bleijenbergh, Peters and Poutsma2010). As Riley et al. (Reference Riley, Metcalf and Forth2008) put it, whether diversity results in net benefits or costs will vary by organisational context.
Conclusions
Our main conclusion is that, in addition to being complex and elusive (cf. Shapiro and Allison, Reference Shapiro and Allison2007), the relationship between diversity and the ‘bottom line’ is contextually contingent, which means that a simplistic policy transfer from US evidence on the for-profit sector to UK public services, which is often – at least implicitly by ignoring contextual issues – asserted, is unwise.
In short, we argue that more consideration is required on the ‘context–mechanism–outcome configurations’ or ‘programme theory’ (Pawson and Tilley, Reference Pawson and Tilley1997) associated with the BCD. Put another way, we argue that it is necessary to develop contingent ‘best fit’ practices (if and where the BCD works) rather than assume a universal ‘best practice’ approach to ‘what works’. This fits with the concerns of some commentators on the BCD (such as, Ryan et al., Reference Ryan, Hawdon and Branick2002; Bleijenbergh et al., Reference Bleijenbergh, Peters and Poutsma2010; Heres and Benshop, Reference Heres and Benschop2010; Johns et al., Reference Johns, Green and Powell2013). Noon (Reference Noon2007: 781) puts it nicely when he says: ‘the argument for the moral case based on the human rights of all employees and job seekers must not be abandoned for the current fashion of diversity and the business case’. This is certainly true for the NHS, where Esmail et al. (Reference Esmail, Kaira and Abel2005) underline that the BCD alone will not be enough to drive forward diversity as an outcome. We suggest a research agenda that draws on realistic perspectives in order to critically explore how policy makers develop (implicit) programme theory in policy and practice documents that seeks to examine how the mechanisms of the BCD are linked with particular outcomes on particular circumstances.
