Minority Ethnic Staff in Universities: Organisational Commitments, Reputation and the (Re)structuring of the Staff Body

This article problematises whether organisational commitments impact the representation of ethnic minorities in the university workforce. In doing so, it considers the institutional context and the broader restructuring of universities’ personnel. The analysis is based on a longitudinal dataset of 120 universities, including university-level indicators of organisational commitments, institutional characteristics and ethnic minority staff numbers. The findings reveal that while on average, universities that are members of the Race Equality Charter exhibit higher shares of minority ethnic staff in higher-level contracts compared with those universities that are not members, joining the charter does not make a university more inclusive. Importantly, the share of minority ethnic staff is substantially lower in elite universities compared with all other universities, which indicates tensions between inclusion and university reputation. The results are discussed in terms of their relevance to sociological institutionalist and organisational theories, and to higher education policy.


Introduction
The representation of ethnic minorities in the UK higher education (HE) sector has been pursued systematically via the Equality Challenge Unit (ECU) and the Race Equality Charter (REC), introduced by the ECU in 2014. 1 The REC aligns with the statutory equality duties of the early 2000s, such as the Race Equality Duty introduced under the Race Relations (Amendment) Act of 2000.While the equality legislation made inclusion a social objective of the public sector, the REC played a role in the institutionalisation of inclusion as a university mission (Baltaru, 2018a).
Since the Race Equality Duty was promulgated and throughout the REC, the share of minority ethnic staff in universities increased about twice as much (Advance HE, 2021). 2  Nevertheless, ethnic minorities tend to be concentrated in lower-level contracts and underrepresented in higher-level contracts.This article considers the purported role of organisational commitments, such as the REC, in shaping staff composition.It then moves beyond it, taking into account how university reputation may foster (or inhibit) the representation of ethnic minorities working in HE, and how the share of ethnic minority staff has fared through changes in the university staff body (including, but not limited to the rise of fixed-term contracts).The article sheds light on how historical reputational concerns and new societal expectations to accomplish inclusive universities create powerful and often conflicting dynamics at the institutional level, which reverberate into the staff composition.
The institutional-level determinants of ethnic minority staff in universities concern policy making.In the UK, social science research (that includes descriptive and correlational evidence drawing on quantitative educational data) has been playing a key role in the pursuit of inclusion in higher education, in line with equalities legislation (Advani et al., 2020;Boliver, 2013;Desai, 2017).To this end, the current study can be used by HE practitioners and by universities in the UK to promote, maintain and monitor equality of opportunity among HE staff.In addition, the findings are of interest to researchers in the sociological institutionalist and organisational scholarship, who seek to understand how the institutional context shapes the organisational structure, and the possibilities and limitations of purposive organisational action; that is, the systematic pursuit of social objectives by means of organisational commitments and strategies (Baltaru, 2020; see also Keith, 2001).Such an investigation is lacking in the UK HE, as official reports do not go beyond descriptive accounts of staff composition (Advance HE, 2021;Equality Challenge Unit, 2015), while quantitative empirical research on ethnic inequalities has mainly focused on students (Boliver, 2013;Wakeling and Savage, 2015;Warikoo, 2018).The present study aims to fill this gap, looking into a specific dimension of minority ethnic staff inclusion: the participation of people from ethnic minority backgrounds in the university workforce as a share of the total staff, as opposed to other dimensions such as the experiences of minority ethnic staff in universities, studied elsewhere (Ahmet, 2021;Arday, 2018;Bhopal, 2022;Johnson and Joseph-Salisbury, 2018;Mirza, 2006).
The first part of the article offers an overview of the pursuit of inclusion in the UK HE, and the purported role of inclusion-oriented organisational commitments.The second part problematises the appeal of organisational commitments and considers how the institutional context, from a university's reputation to its staff composition, may account for variation in the share of minority ethnic staff.Minority ethnic staff are defined based on the 2011 census classification system, as an aggregate of the following groups: Asian (Asian/Asian British: Indian, Pakistani, Bangladeshi and any other Asian background), Black (Black/African/Caribbean/Black British: African, Caribbean and any other Black/ African/Caribbean background), Chinese, mixed and other ethnic backgrounds.The third part provides an empirical analysis of the determinants of minority ethnic staff in total staff in the UK universities.The results are presented and discussed in terms of their contribution to scholarship and policy.

Minority Ethnic Staff and the Institutionalisation of Inclusion as a University Mission
The pursuit of inclusion in the UK HE has been catalysed by global trends such as the post-Second World War emphasis on socio-economic development (Schofer and Meyer, 2005), and on individual empowerment, codified as human rights (Baltaru, 2018a;Soysal, 2012).Inclusion became a salient issue in HE, especially as the sector started being reorganised for mass provision.In the 1980s, New Public Management (NPM) policies strengthened university's managerial backbone as a way of enabling mass HE in more cost-effective ways than direct state intervention (Morley, 1997), and of increasing universities' competitiveness in the global HE market (Christensen, 2011).This has had an impact on inclusion, which became subject to 'diversity management' (Mirza, 2006).
In the UK, an important role in designating responsibility for inclusion at the institutional level was played by the equality legislation of the early 2000s, namely: the Race Equality Duty of 2001, the Disability Equality Duty of 2006 and the Gender Equality Duty of 2007, later combined into a generalised Equality Duty (Equality Act, 2010; see also Baltaru, 2018aBaltaru, , 2020)).Universities started designating in-house human resource teams to monitor the student and staff demographic, and developed their own inclusion policies, in line with their equality duty.Terms such as 'inclusion', 'equality' and 'diversity', started being used by universities on their websites, and in their prospectuses and brochures (Kimura, 2014;Baltaru, 2020).Equality charters were established to guide equality work in HE, while channelling inclusion-oriented organisational actions developed by universities into sector-wide databases of good practices, notably to support females (the Athena SWAN Charter as of 2005) and ethnic minorities (the Race Equality Charter as of 2014).
How did minority ethnic staff fare as inclusion became institutionalised as a university mission?Over the last couple of decades, the absolute number of minority ethnic staff in UK HE, irrespective of country of origin, has more than doubled from about 24,045 in 2003/2004, to almost 59,290 in 2018/2019, a remarkable growth compared with just under 40% increase in the total number of staff working in HE, irrespective of ethnicity (Advance HE, 2021).At the same time, ethnic minorities continue to be both underrepresented in universities, and concentrated in precarious contracts.In England, the region with the greatest number of universities in the UK, UK-based minority ethnic staff make up for 12.2% of total staff (Advance HE, 2021).This indicates an underrepresentation of ethnic minorities of working age (which make up for about 15% of the English population of working age, according to the 2011 Census), and a steeper underrepresentation of ethnic minorities of working age with higher-level qualifications (which make up for over 17% of the English population of working age with higher-level qualifications). 3 The figures should be understood in the context of a disproportionate growth of minority ethnic staff in lower-level compared with higher-level contracts in universities over the last couple of decades (Advance HE, 2021), also noted throughout the steady increase in the share of minority ethnic staff from 1995-1996to 2003-2004(HEFCE, 2008)).Now as then, minority ethnic staff are concentrated in lower-grade contracts, while being less likely to have permanent/open-ended contracts (Advance HE, 2021;Connor, 2008).The concentration of minority ethnic staff in lower-grade, less stable and less financially rewarding contractual levels, despite the significant presence of individuals from minority ethnic groups in higher education, highlights the importance of understanding the institutional-level factors at work.

Organisational Commitments, Reputation and the (Re)structuring of the Staff Body
To uncover what drives the share of minority ethnic staff at the institutional level, the current study follows a two-pronged approach.First, it is important to factor in the potential impact of purposive organisational action, as underpinned by systematic, sector-level organisational commitments, such as the Race Equality Charter (REC).Second, we must question the appeal of purposive organisational action as a way of achieving inclusion, taking into account what is glossed over by the idea that a university can become inclusive by being organisationally committed to inclusion.The latter requires a wider consideration of the barriers faced by minority ethnic staff in HE institutions (HEIs), and of how universities transform following changing expectations about what they should look like.
It is possible that purposive organisational action helps develop inclusive environments where ethnic minorities feel encouraged to work and progress their careers.This approach is a key premise for the sector-level equality work in the UK HE, such as the establishment of the Equality Challenge Unit (ECU) to support equality and diversity in HE.Currently under the umbrella of Advance HE, the ECU introduced the REC, which is the largest, sector-level initiative to improve the representation and progression of minority ethnic staff and students.Its popularity can be seen in the number of HEIs joining the charter, which increased rapidly from 30 institutions as of 2015 to 71 institutions as of 2020 (Douglas Oloyede et al., 2021).Qualitative interviews with equality and diversity practitioners highlight the usefulness of the REC in guiding race equality work and in developing actions to tackle race inequalities (Douglas Oloyede et al., 2021;Bhopal and Henderson, 2019).Nevertheless, equality and diversity policies and strategies are often seen as tokenistic by the staff they are aimed at (Ahmet, 2021;Arday, 2018).Minority ethnic staff continue to report experiences of marginalisation, disproportional scrutiny and isolation (Bhopal, 2022;Rollock, 2019; see also Mirza, 2006), along with insufficient opportunities to enhance their promotion prospects and to participate in leadership (Arday, 2018;UCU, 2016).Thus, it is also possible that organisational commitments do not go beyond the surface, having little impact on the share of ethnic minorities working in universities.A sociological institutionalist account can help us understand why this might be the case.This perspective conceptualises formal organisation as ceremonial rather than functional (DiMaggio and Powell, 1983;Meyer and Rowan, 1977).Applied to the pursuit of inclusion in HE, it suggests that universities develop inclusion-oriented organisational commitments in order to meet epistemological expectations of agency and rationality projected on their behalf (Krücken, 2011;Ramirez and Christensen, 2013).As a result, formal policies can easily become 'decoupled' from the actual organisational practices (Meyer and Rowan, 1977).Baltaru (2020) notes that in the early 2000s, virtually all UK universities started formally organising for inclusion; for example by producing dedicated web pages and strategic action plans, despite persistent differences in the student and staff demographic across the sector.Organisational commitments to inclusion may serve the image of the university, rather than the staff they are aimed at.
Perhaps the institutional context of each individual university, such as its historically consolidated reputation and the composition of employees that populate it, plays a stronger role in the ponderance of minority ethnic staff than the organisational commitments we see across the board.The idea that how an institution evolves reflects its individual historical and structural conditions can be traced to some of the earliest sociological accounts of formal organisation (Parsons, 1958;Selznick, 1948;Weber, 1986).This postulate has been elaborated in organisation studies exploring how organisations adapt to external demands (Marquis and Tilcsik, 2013;Stinchombe, 1965;Thornton and Ocasio, 2008), and in HE research exploring the organisational expansion of universities (Baltaru and Soysal, 2017;Clark, 1983;Fumasoli and Stensaker, 2013).While a university may commit to inclusion as a way of adapting to this societal expectation (Ramirez and Christensen, 2013), the degree to which it becomes inclusive could vary depending on the compatibility between the newly acquired commitment and its institutional imprint; that is, what has purveyed the university with legitimacy since its foundation.For example, Gumport (2000) notes that in the USA, public universities adopting market-oriented principles risk losing the legitimacy coming from their reputation as public-oriented educational institutions.Oertel (2018) finds that in Germany, older, more established universities are less likely to adopt diversity management at the higher administrative levels in comparison with the newer universities.
In the UK, a distinction can be made between 'old', traditional elite 'sandstone' and 'redbrick' research-intensive universities, and 'new' universities.The latter are often identified with the 'post-1992' universities, established after polytechnics gained university status (Further and Higher Education Act, 1992).More recently, membership of the Russell Group has been used in sociological studies of ethnic diversity in higher education to indicate university reputation, as these 24 research-intensive universities have successfully marketed themselves as the UK's equivalent to the Ivy League in the USA (Baltaru, 2020;BBC, 2006;Boliver, 2013). 4Russell Group universities are the most selective across the sector (Raffe and Croxford, 2015), most sought-after by students (National Audit Office, 2017), tend to be highest rated in the Research Excellence Framework (REF) and have the highest income from endowments and invested income across the sector (Boliver, 2015).Reflecting on the stratified UK university sector, Ahmed (2006) notes that research-led, 'sandstone' universities are elite precisely because they are not seen as diversity-led (see also Mirza, 2006).She further argues that in elite, predominantly White universities, diversity work 'becomes about changing perceptions of whiteness rather than changing the whiteness of the organisation' (Ahmed, 2006: 118).
While it is well known that students from ethnic minority backgrounds are less likely to enter elite universities (Boliver, 2013;Warikoo, 2018), which are seen as 'white, middle-class bubbles' (Reay, 2018: 57), quantitative research on how minority ethnic staff have fared in these universities is scant.An isolated quantitative case study shows that less than 20% of Black academic economists work in Russell Group universities, and less than 30% of Pakistani and Bangladeshi academic economists do so, compared with 50% of the White counterparts (Advani et al., 2020).The study also finds that the ethnic minority economists who work in Russell Group universities are 45% less likely to hold professorial or managerial-level functions, compared with the White employees.Qualitative interviews with minority ethnic staff in elite universities help widen the picture.Bhopal (2022) compares the experiences of academics of colour in a US elite university and a UK ancient, elite university.Participants spoke about hostile behaviours from colleagues, such as being told that they have been hired 'to tick a box'.Based on interviews with minority ethnic academics and research students from Russell Group universities, Ahmet (2021: 161) identifies feelings of isolation, and scepticism towards diversity policies, which are seen as enabling universities 'avoid discussions of racism within their walls'.Based on semi-structured interviews with Black female professors in the UK, Rollock (2019) outlines the way in which historical hierarchies between universities shape their paths to professorship: a Black female senior lecturer may leave a post-1992 university because of being denied professorship and start all over again as a lecturer at a Russell Group university.
Beyond university reputation, it is important to account for how the share of minority ethnic staff has fared throughout the broader restructuring of universities' staff body over the last couple of decades.First, just like other public sector organisations who have been reorganised according to market principles, universities witnessed a decline in staff working in open-ended, permanent contracts, and an increase in more precarious, fixedterm contracts (Loveday, 2018).As of 2021/2022, HESA shows that 33% of academic staff are employed on fixed-term contracts, minority ethnic staff being more likely to have a fixed-term contract compared with the White staff (Advance HE, 2021).Second, UK universities internationalised, not only in their student body, but also in their staff.As of 2019, international students made up for 19.2% of the student population, while 20% of the staff were international (Universities UK, 2021).This has fostered diversification in the staff body (Carvalho, 2017).Third, the strengthening of universities' managerial backbone, following the NPM and the broader organisational expansion of universities, boosted the number of non-academic professionals (Baltaru andSoysal, 2017, Baltaru, 2018b).This trend may have implications for ethnic minority staff, who are slightly less prevalent in professional and support jobs compared with academic jobs (Advance HE, 2021).
To summarise, this article sets itself to answer the following research questions:

RQ1:
To what extent is the share of ethnic minorities working in the UK universities fostered by universities' organisational commitments to inclusion, such as the REC?

RQ2:
To what extent is the share of ethnic minorities working in the UK universities shaped by reputational differences between universities?
RQ3: How does the share of ethnic minorities working in the UK universities fare throughout the restructuring of the staff body in higher education, such as the increase in fixed-term contracts?

Data and Analytical Strategy
The article draws on a comprehensive dataset of university-level, organisational data, on 120 UK HEIs with university status.These are all universities with data available across all variables of interest, measured yearly from 2012/2013, the earliest wave of comparable data, up until 2018/2019 (prior to the COVID-19 pandemic).The sample includes various universities, large and small, more and less ethnically diverse, varied in terms of reputation, REC membership and staff composition, as illustrated in online Table 1a.Data from HESA, the official agency collecting data on the UK HE sector, are used to operationalise the first dependent variable; that is, the proportion of Black and minority ethnic staff in total staff.'Staff' refers to being employed under a contract of employment (academic or non-academic) at a HEI.HESA data on contractual levels are used to operationalise the share of minority ethnic staff in higher-level contracts; that is, academic and non-academic staff who hold professorial or senior management functions (that includes, vice-chancellors, university registrars, heads of department, directors of a major function group such as finance, corporate services or HR, and professors).
Online Figure 1 shows the two dependent variables over time.Note that while the share of ethnic minorities has increased steadily both in total staff and in higher-level contracts, minority ethnic staff remain concentrated in lower-level contracts.Online Table 1b provides a more granular representation of the sample data, differentiated by Russell Group membership.The increase in the share of minority ethnic staff across the board reveals an overarching move towards more inclusive staff demographics.When looking at the crude proportion of minority ethnic staff in total staff, the trend is slightly more pronounced in Russell Group universities.However, a closer look reveals that the increase in the share of minority ethnic staff in higherlevel contracts has been substantially slower in Russell Group universities, compared with all other universities.
To address RQ1, universities' organisational commitments to inclusion are operationalised in terms of membership to the Race Equality Charter (REC), the largest, sector-level initiative that centralises institutional-level organisational commitments to tackle the barriers faced by ethnic minorities in HE (Douglas Oloyede et al., 2021).The list of universities that are REC members has been obtained from Advance HE.Two indicators are used: (a) a binary indicator that differentiates between universities that are members of the REC (coded '1'), as opposed to all other universities (coded '0'), irrespective of when they have joined the charter.This helps us see if universities that joined the charter are generally more demographically inclusive compared with those that did not; and (b) a longitudinal indicator that captures a universities' transition from no REC charter membership (coded '0') to membership (coded '1').This indicator helps us to see if joining the charter is associated with change in the shares of minority ethnic staff.
To address RQ2, university reputation is operationalised based on Russell Group membership.This is a binary indicator, where university members are coded '1', as opposed to all other universities, coded '0'.As detailed earlier in this article, Russell Group membership is widely used as an indicator of university reputation in sociological research on higher education, as these 24 research-intensive universities have managed to market themselves as the UK's elite universities (Boliver, 2013; see also Advani et al., 2020;Ahmet, 2021;BBC, 2006).
To address RQ3, longitudinal HESA data are used to operationalise three indicators of staff composition: the share of international staff in total staff, the share of staff with fixed-term contracts (as opposed to permanent contracts) in total staff and the share of non-academic staff in total staff.These are key indicators for the restructuring of universities' staff, in the face of internationalisation, marketisation and organisational expansion, as elaborated earlier in the article.
The models control for university size, foundation era and region.University size is captured by the total number of students (Daraio et al., 2011).While large universities may be more demographically diverse, smaller universities may attract more ethnic minorities because of small organisations being more conducive to diverse group interactions (Fischer, 2008).Foundation era is operationalised as a binary indicator: whether the university was founded before 1992 (coded '0'), or post-1992 (coded '1').Newer universities could exhibit higher shares of minority ethnic staff as they are founded in an epoch that places a higher value on inclusion (Oertel, 2018).Finally, the models account for regional differences in the ponderance of ethnic minorities in the UK population, with London as the reference category.
The data are based on a balanced panel of UK universities observed yearly from 2012/2013 to 2018/2019.Two separate models are run; that is, to predict the share of minority ethnic staff in total staff (Model 1; N = 120), and the share of minority ethnic staff in higher-level contracts (Model 2; N = 106). 5The relationships between the share of minority ethnic staff (in total staff and in higher-level contracts respectively) and the explanatory variables are estimated using linear regressions.Time series are chosen over pooled ordinary least squares regression (OLS) after confirming that the clustering of observations over time is statistically significant (Breusch and Pagan, 1979). 6 Furthermore, running the regressions with random effects (RE) is preferred when there is interest in both time variant (within university) effects and time invariant (between university) effects (Bell et al., 2019).RE modelling can disentangle variation between and within units, making it possible to estimate the effects of both time variant predictors, such as the share of international staff in total staff, and time invariant predictors, such as reputation.Modelling alternatives to RE, such as fixed effects and first differences, do not allow for estimating time invariant regressors.The random effects model has been applied after confirming that there is no statistically significant association between the unique errors and the predictors, an important assumption of this modelling technique (Hausman, 1978). 7 The model becomes: where Y it is the share of ethnic minority staff for university 'i' at time 't', X it are the explanatory variables indicating the characteristics of the university and 'α' is the share of minority ethnic staff in each university, adjusted yearly by the residual u it .The gradient modelling the effect of the explanatory variables is marked as β , and  it represents the university-level residual.Note that the longitudinal indicator capturing the joining of the REC is lagged as X i t −1 , to account for a potentially delayed impact.The results from the Variance Inflated Factor analysis (VIF < 5) indicate that multicollinearity is unlikely (Hair et al., 1995). 8The Open University has been dropped because it is an outlier compared with all other universities. 9The models were run with robust standard errors.

Results and Discussion
The results from the two models are illustrated in online Figure 2, which displays all marginal effects at the 95% confidence interval, from the models predicting the share of minority ethnic staff in total staff (Model 1) and in staff with higher-level contracts (Model 2).Online Table 2 provides the exact estimates for the independent variables.The models control for dummy variables for regions and for years, all estimates being displayed in online Figure 2.
In online Table 2 we can see that Model 1 explains 49% of variation within universities, and 73% of variation between universities, while Model 2 explains 16% of variation within universities, and 35% of variation between universities.Thus, the models explain moderate to large amounts of variation in the share of minority ethnic staff working in universities.The smaller amount of within variation explained compared with between variation suggests that the share of minority ethnic staff is explained to a greater extent by differences between universities, than by changes within universities over the years under investigation.Nevertheless, the within variation explained is itself substantial.
The evidence for whether organisational commitments foster the representation of ethnic minorities among university staff (RQ1) is mixed.The results show that on average, universities that are REC members display larger shares of ethnic minorities in higher-level contracts compared with all other universities (Model 2: β = .024,p < .05).However, joining the Race Equality Charter (the lagged regressor capturing the transition to charter membership) is not statistically significant in any of the models. 10The combined findings show that while ethnic minorities are better represented in higherlevel contracts in the universities that are REC members compared with all other universities, this might be because these universities were more inclusive to begin with.
The results further show that the share of minority ethnic staff in the UK universities is shaped by reputational differences between universities, irrespective of organisational commitments and net of other institutional-level differences (RQ2).Minority ethnic staff are less represented in Russell Group universities, compared with all other universities, both in the total number of staff (Model 1: β = −.018,p < .05)and in higher-level contracts (Model 2: β = −.033,p < .01).These findings should be interpreted considering that ethnic minority students are also underrepresented in elite universities (Boliver, 2013).The apparent tension between university reputation and ethnic diversification should prompt further research into what comes to define university reputation beyond the formal emphasis on excellence and meritocracy.Elite status appears to be underpinned by resistance to demographic diversification (Ahmed, 2006), despite elite universities indicating that they are committed to inclusion; for example, via developing inclusion-oriented offices (Baltaru, 2020).
An area of concern may be that Russell Group membership is a restrictive measure of reputation.Moreover, the models were re-run with an alternative operationalisation of reputation, proposed by Soysal et al. (2022).In the new, binary indicator, the highest category includes those universities consistently ranking in the TOP in the earliest years documented by The Times ranking tables (1992, 1994, 1995 and 1996), compared with all other universities.The negative and statistically significant coefficient for university reputation was replicated, specifically in the higher-level contracts. 11The share of ethnic minorities in higher-level contracts is consistently lower in elite universities across both operationalisations of reputation, while being higher in universities that are REC members.This finding reveals that the tension between sector-level organisational commitments to support historically underrepresented social groups and the institutional barriers that prevent their representation plays out especially in the higher-level contracts.
The findings also document the impact that the restructuring of the staff body in universities has had on the share of minority ethnic staff (RQ3).First, the internationalisation of the university's staff body is a very strong driver of the share of ethnic minorities in total staff (Model 1: β = .483,p < .001;Model 2: β = .397,p < .05).Second, the proliferation of fixed-term contracts inhibits the presence of minority ethnic staff in higherlevel contracts (Model 2: β = −.078,p < .001).This finding supports existing research suggesting that precarious working conditions in HE disproportionally affect ethnic minorities (Mirza, 2006).
Online Figure 2 additionally displays the estimates for dummy variables for years and for region.As expected, the share of minority ethnic staff in total staff tends to be lower in the various UK regions compared with London, which is the reference category and most ethnically diverse region in the population.We can also see that the share of ethnic minorities working in universities has increased over the years under investigation as compared with 2012/2013, the baseline year.

Further Analyses and Limitations
Additional analyses have been carried out to ensure the robustness of the results. 12 First, a quasi-experimental, difference-in-differences model was employed to compare the changes in the shares of minority ethnic staff between the universities that were REC members (the treatment group) and those that were not REC members (the comparison group).The results confirmed our original findings, as follows: (a) universities started hiring more minority ethnic staff over time; (b) universities who opted to join the charter had higher shares of minority ethnic staff regardless of the charter; and (c) the charter on its own did not have a statistically significant effect on the share of minority ethnic staff.
Second, the models were re-run with fixed effects, which provides an appropriate estimation for the lagged dependent variables.The lagged dependent variable was consequential for the share of minority ethnic staff in total staff, but not for the share of minority ethnic staff in higher-level contracts.This finding shows that the progressive growth in the number of ethnic minorities working in universities is an important driver for the overall share of minority ethnic staff.Unsurprisingly, these demographic pressures have been slower to penetrate the higher-level contracts.All other coefficients among the time variant indicators were replicated.
Third, because the dependent variable is a proportion, the results from the linear estimator have been compared with results obtained when using a non-linear model that considers the bounded nature of the outcome (Papke and Wooldridge, 2008).The models have been re-run in two ways: (a) with a two-limit tobit model, which treats the dependent variable as a censored continuous variable; and (b) with a generalised linear model, specified with a logit link, from the binomial family, and robust standard errors.Both main results -that is, the negative and statistically significant coefficient for reputation, and the positive and statistically significant association between being in the REC and the shares of minority ethnic staff in higher-level contracts (distinct from joining the REC, which is not statistically significant) -were replicated.
Note that while this longitudinal analysis of the association between organisational makeup and the shares of minority ethnic staff has undergone the above robustness checks, one should still be cautious of drawing causal conclusions.As time passes and more data become available, future research may draw on longer timespans that can afford the extensive analyses required for establishing causal links.

The Role of Organisational Commitments, Institutional Context and Societal Expectations
The findings show that the share of ethnic minorities working in universities does not evolve in isolation of the institutional settings.The representation of minority ethnic staff in universities is underpinned by organisational commitments and by the institutional context, which are themselves a reflection of the wider society in which universities operate.
First, the article finds some limited evidence that organisational commitments, in the form of equality charter membership, are not only giving universities opportunities to align with societal expectations, as theorised by sociological institutionalists (Baltaru, 2018b;Krücken, 2011); they also bring together inclusive universities, enabling them to lead by example.More exactly, while joining the Race Equality Charter is not a statistically significant predictor for the shares of minority ethnic staff, being in the charter is a common feature that the more inclusive universities share.This is because on average, universities that are charter members employ more minority ethnic staff in higher-level contracts compared with those universities that are not members.This may transform the idea of what a university should look like, towards a greater emphasis on demographic representation, but there is no evidence that joining the charter on its own makes a university more inclusive.
The results apply to organisational commitments that stem from systematic, sectorlevel action, such as the REC.Organisational commitments expressed in other forms, for example, a university merely presenting itself as inclusive on its webpages, are already decoupled from variation in staff composition, as virtually all UK universities do it (Baltaru, 2020).
Furthermore, the institutionalisation of inclusion as a university mission is not the only force shaping the share of minority ethnic staff.The current findings show that growth in precarious contracts in higher education, as normalised by market-oriented concerns of cost-efficiency (Loveday, 2018), and the influx of international staff, impact minority ethnic staff in different ways.Note that the positive and statistically significant association between the shares of international staff and the shares of minority ethnic staff is underpinned by complex internationalisation processes (Carvalho, 2017).While international staff may boost the share of ethnic minorities in universities, this does not translate into enhanced representation for the UK-based staff.For example, the large share of minority ethnic staff in total staff observed in the Russell Group universities compared with other universities (see online Table 1b) is likely an artefact of the large share of international staff in these universities. 13Indeed, after controlling for international staff among all other variables (see online Table 2), we observe the reverse; that is, a negative association between Russell Group membership and the shares of minority ethnic staff in total staff.
Second, the institutional context matters for the share of minority ethnic staff, especially when it comes to university reputation.We have seen that minority ethnic staff are less prevalent in elite universities compared with all other universities.Of special relevance here may be that 'excellence and meritocracy' narratives have become separated from 'equality and diversity' to the extent that they have started to be perceived as antagonistic when it comes to university reputation (Ahmed, 2006; see also Deem, 2009).The low shares of minority ethnic staff in the UK's elite universities may indicate a conflict of institutional adaptation that reverberates into their personnel structure, as theorised in organisational research (Marquis and Tilcsik, 2013;Oertel, 2018;Stinchombe, 1965).Elite universities are expected to become inclusive, but this implies a shift away from the specific understandings of excellence and meritocracy that have historically fuelled their reputation.Baltaru (2020) notes that such a shift is underway, as equality charters slowly integrate inclusion as a dimension of university reputation; for example, ranking universities in terms of their inclusion-oriented outputs, conditioning research funding on inclusion progress (until recently, the National Institute for Health and Care Research required HEIs to hold at least a Silver Athena SWAN award to qualify for competitive research funding).
Overall, this study answers the need for an institutional-level investigation into what drives the representation of minority ethnic staff in universities, explaining a substantial amount of variation in the share of minority ethnic staff in universities and across contractual levels.It shows that a limited amount of variation in the share of minority ethnic staff in universities can be traced to the institutionalisation of inclusion as a university mission (specifically, to more inclusive universities joining sector-level actions, such as the REC), but also to the institutional context, in terms of university reputation, and to the broader societal expectations that transform universities and propel the restructuring of their personnel.
Further research may explore what encourages (or discourages) ethnic minorities from applying to work in universities in the first place, the selection and the interview stage, and the micro-level social processes that affect the progression and retention of minority ethnic employees in universities.This research may afford a more granular approach, zooming into the representation and experiences of individuals from distinct ethnic minority groups, as well as minority ethnic individuals who may face additional barriers because of their social background, such as their gender or social class.Finally, while the aggregate data utilised here cannot be used to extrapolate on the micro-level social processes that work to include (or exclude) ethnic minority staff in universities, further theoretical syntheses are needed to integrate insights from institutional and individual-level research on the topic.