Gender targets and trickle-down effects: Avoiding the ‘decoupling dynamics’ that limit female representation in senior roles

Women constitute the majority of the Australian public sector workforce, but their representation in senior roles is not proportional. Australian public services have gender targets to improve the representation of women in senior roles. Based on previous research, targets are expected to first increase female representation at the target’s focal level, such as executive level. Then they should initiate a trickle-down effect (TDE), increasing female representation at the level immediately below the target’s focal level, such as the executive feeder level. However, the TDE observed in a state public service decelerated after a gender target was imposed. We identified whether individual departments had a consistent or inconsistent TDE and conducted 13 semi-structured interviews with key stakeholders. Too many service-wide targets with low prioritisation of a gender target, as well as missing and ineffective practices, generated decoupling dynamics. Only departments with gender champions who had visible backing from the Chief Executive were able to keep the gender target coupled with practice to achieve its intended outcomes.


K E Y W O R D S
decoupling, female representation, gender champions, gender equality, gender targets, trickle down

Points for practitioners
• Gender targets in Australian public services may not be achieving intended outcomes due to decouplinga response to policies in which the policies are ignored and/or ineffective practices are implemented.
• Too many competing targets and limited accountability for achieving a gender target create a potential for decoupling by allowing individuals and groups to ignore or weakly adopt the policy.
• Integrated bundles of top-down (e.g.requiring at least two women on shortlists) and bottom-up practices (e.g.mentoring) can help avoid decoupling by ensuring women are appointed to senior roles and supported to progress through an organisation.
• Chief Executives are key to ensuring a gender target remains coupled with its implementation; Chief Executives must provide visible support to internal champions to make gender targets effective.

INTRODUCTION
Across sectors and around the world, female representation at senior organisational levels lags behind male representation.The Australian public sector is no exception.Australia has three tiers of government: federal, state or territory, and local.Approximately 75% of public sector employees work at the state level, with the federal and local public services accounting for 15% and 10%, respectively (Podger, 2017).Women represent 60% of the federal public service workforce but hold less than half of the most senior roles (Australian Public Service Commission, 2020).At state level, women make up approximately two thirds of the New South Wales (NSW), South Australian, and Queensland public service workforces but hold only 41%, 49%, and 50% of senior leadership roles, respectively (NSW Public Service Commission, 2020;Institute of Public Administration Australia SA, 2019;Queensland Public Service Commission, 2020).
To address the underrepresentation of women in senior levels, organisations have introduced a wide variety of organisational practices (Bilimoria, Joy & Liang, 2008;O'Brien et al., 2015).Some practices (e.g.boosting women's representation by using targeted recruitment practices and increasing the number of women on shortlists) focus on increasing gender diversity without addressing the systemic barriers that women face after hiring (Leslie, 2019).Other practices (e.g.targeting women for leadership training and mentoring or offering flexible work to accommodate childcare responsibilities) focus on supporting women to move up through organisations, thereby narrowing gender gaps in career outcomes.The framing of these wellintentioned programs can suggest that women need help, inadvertently positioning the cause of gender inequalities in women (a 'fix-the-women' approach; Ely, Ibarra & Kolb, 2011) and perpetuating negative gender stereotypes (Leslie, 2019).Neither group of practices, either separately or in combination, has resulted in anticipated gains at the top of Australian organisations.The glass ceiling is as strong as ever and continues to block the progress of women to senior levels.
Gender targets represent a strategy designed to drive gender improvements at a certain level (e.g.executive level) (Sojo et al., 2016), henceforth called focal level.Australian federal, state/territory, and local governments have imposed gender targets of 50% female representation in leadership positions (Australian Government, 2021;NSW Government, 2018;South Australian Department of Treasury and Finance, 2021;Queensland Government, 2021;Victoria Local Government, 2022).Unfortunately, targets may not be having the impact expected by governments.
Understanding the reasons behind ineffective targets is important if public services are to derive maximum value from gender targets.We investigate whether externally imposed gender targets become decoupled from internal organisational practices and their outcomes.In particular, we consider the organisational practices that might initiate and support a trickle-down effect (TDE) and the roles that gender champions and blockers play in (de)coupling dynamics.We begin by presenting the decoupling literature, followed by a discussion of gender targets and the TDE.We then explain the research design and present our findings.We use an Australian public service as a case study, and analyse interviews with senior departmental staff to identify situations where decoupling occurred and explain why a gender target did not consistently support a TDE.Our study provides insights into the role of the Chief Executive (CE) in keeping a target coupled with its implementation.

DECOUPLING DYNAMICS: THEORETICAL OVERVIEW
Across all sectors, including the public sector, external stakeholders are exerting an increased influence on organisational policy, especially regarding accountability and assessment (Bromley & Powell, 2012).External pressure does not always lead to outcomes that satisfy these stakeholders-a disconnect known as decoupling.Decoupling is traditionally understood as policy-practice decoupling whereby an organisation may adopt multiple, even conflicting, policies because of external pressures, with little change in actual practice (Hengst et al., 2020).Bromley and Powell (2012) also identify means-ends decoupling where organisational practices may change, but these practices do not actually achieve policy goals.Decoupling has been identified in the private sector (e.g.Crilly, Zollo & Hansen, 2012), not for profits (e.g.Bromley, Hwang & Powell, 2012), universities (e.g.Delucchi, 2000), and the public sector (e.g.Tilcsik, 2010).The public sector is subject to strong external pressure to improve gender diversity (OECD, 2019) and this includes the Australian public sector (Australian Government, 2021).In response to external stakeholder pressure, organisations have introduced gender targets to improve female representation, especially at senior levels (Sojo et al., 2016).However, despite the existence of gender targets across multiple Australian public services, female representation at senior levels continues to lag behind male representation, and this may be due to decoupling of gender targets from their implementation.Champions may have a role in keeping gender targets coupled with their implementation.Champions try to drive change (Stainback, Ratliff & Roscigno, 2011) and overcome blocker resistance to change (Tilcsik, 2010).It is important to understand what role champions play in keeping a gender target coupled with its implementation.

Policy-practice and means-ends decoupling
Policy-practice decoupling arises from symbolic adoption of policy; policies are either ignored or violated by internal actors (Bromley & Powell, 2012).Organisations can experience policy-practice decoupling when internal resistance to change results in failure to implement a policy or such weak implementation that the policy does not make a material difference to day-to-day activities (Battard, Donnelly & Mangematin, 2017).Policy-practice decoupling is likely when organisations see policy outcomes as more important for legitimacy than for operational purposes.Additionally, when limited resources are available to achieve a policy, managers must choose whether resources can be assigned to achieve the policy or whether the policy can be ignored.Finally, when internal actors do not see the value of a policy and believe they can resist external pressures, they may ignore or weakly adopt the policy (Bromley & Powell, 2012).
For example, flexible work should help women manage work and family roles (Chung, 2020).An organisation can signal its support for women to the labour market by adopting a flexible work policy that should achieve a specific outcome.However, the organisation will experience policy-practice decoupling if it adopts a flexible work policy without making substantive-and consistent-changes to internal practices (Dick, 2015).Policy-practice decoupling can occur when managers interpret the organisation's policy differently and implement practices inconsistently, creating a gap between the policy's intent and the practices that are implemented (Piening, Baluch & Ridder, 2014).Managers might be differentially enthusiastic about flexible work, so that some managers support flexible work but others do not-despite the organisation-wide policy that permits flexible work.Bromley and Powell (2012) identified a second form of decoupling, means-ends decoupling, where organisations direct resources to activities designed to achieve policy goals, with little evidence of effectiveness.Although policy-practice decoupling is symbolic adoption of policy, means-ends decoupling is symbolic implementation of policy.The organisation appears to be making changes to enact the policy, but it is unclear whether intended outcomes are achieved.
An organisation with a flexible work policy may avoid policy-practice decoupling by explicitly offering part-time work to employees (Bilimoria et al., 2008).But the organisation is still vulnerable to means-ends decoupling (Dick, 2015).Managers may view part-time workers as less committed to their work (Chung, 2020) and see part-time arrangements as difficult to manage (Dick, 2015).These views may be strongly held in relation to senior roles where jobholders are expected to prioritise their work role over their private lives.Part-time work is incompatible with the 'ideal worker' concept operating in many organisations (Mainiero & Sullivan, 2005).As a result, women may be aware of the part-time work option but choose not to take advantage of it; part-time work would be perceived negatively and put their careers at risk (Baird, Evesson & Oxenbridge, 2014).The organisation has avoided policy-practice decoupling because flexible work (policy) has been implemented through an organisation-wide part-time work option (practice).But the organisation has experienced means-ends decoupling because the policy's goal (helping women balance work and family roles) has not been achieved.

2.2
Gender targets and the TDE Gender targets are initiatives that aim to improve female representation at a focal level within an organisation.Targets allow a degree of flexibility when compared to quotas because they are designed to persuade rather than force cultural change (Klettner, Clarke & Boersma, 2016).Targets are most likely to increase representation when they create accountability (Whelan & Wood, 2012).
If the target is important to stakeholders, departments will comply with a gender target so they achieve legitimacy for their activities (Hillman, Shropshire & Cannella, 2007).However, the impact of these targets can go well beyond the focal level.Increases in female representation at a focal level (e.g.executive level) can also lead to increases in female representation at a lower organisational level (e.g.executive feeder level): a TDE.This effect has been found across sectors (public/private), countries (Australia, the United States), and levels (board/executive; senior management/middle management; executive/executive feeder) (Ali, Grabarski & Konrad, 2020;Bilimoria, 2006;Cohen, Broschak & Haveman, 1998;Gould, Kulik & Sardeshmukh, 2018a, 2018b;Kurtulus & Tomaskovic-Devey, 2012;Matsa & Miller, 2011).
A TDE requires women at the senior level.Therefore, top-down practices that increase the likelihood that women apply for, are considered for, and are hired into senior roles are essential.Screening job descriptions for gendered language (Gaucher, Friesen & Kay, 2011) and including women on interview panels who recruit through their professional networks (Kazmi et al., 2021) increase the likelihood that women will apply for senior roles.Lengthening shortlists (Lucas et al., 2021) and requiring at least two women on shortlists (Johnson, Hekman & Chan, 2016) make it more likely that women will be offered roles.Further, flexible work practices (e.g.parttime work) increase the likelihood that women will accept these senior roles (Firfiray & Mayo, 2017).
Once initiated, the TDE can be supported with bottom-up practices that help women progress from lower level roles into more senior ones.Leadership training and mentoring help women build the human and social capital required for advancement (Metz & Tharenou, 2001;Tharenou, 2005).However, there can be no TDE unless women are already at the top.Bottom-up practices can only support a TDE once women are already occupying senior roles.
Of relevance to this research, Gould et al. (2018a) found a TDE operating between executive and executive feeder levels in an Australian public service.The TDE in this public service was strongest when female representation at the executive level was between 15% and 35%.Gould et al. (2018a) predicted a gender target would accelerate the TDE because the external pressure of a gender target should increase accountability of individuals and legitimacy efforts of departments across the public sector.Surprisingly, even though the TDE continued after the target, it decelerated.The TDE deceleration may reflect decoupling dynamics.Individual departments might be failing to implement practices to achieve the service-wide target or adopting practices that are ineffective.If some departments are experiencing decoupling, they will be unable to initiate or support a TDE.

Gender champions and blockers
Gender champions, especially those who hold senior executive roles, use their power to drive change (Stainback et al., 2011).Champions challenge organisational norms and set priorities, including improvements in female representation (Cortis, Foley & Williamson, 2021).They facilitate the hiring of women from outside the organisation and the progress of women within the organisation.
Individuals can become gender champions because gender diversity lies within their job role or they may be intrinsically motivated to create change (Taylor et al., 2011).
However, individuals and groups within the organisation may act as 'blockers' who resist the change that champions are trying to progress, even if most employees view the change as positive (Malhotra et al., 2021).Blocker behaviour may be overt, talking negatively about women to reinforce gender stereotypes and maintain the status quo (Metz & Kulik, 2014).Blockers may also engage in more subtle behaviour ('oblique resistance'; Jones et al., 2021) that undermines champions' efforts by ignoring them or being slow to implement change.Blockers can be particularly difficult to manage when budgets are tight and managers must transfer resources away from existing operations to address new policy demands (Livanos, Yalkin & Nunez, 2009), or when Human Resource (HR) functions are devolved from a central authority to local management, creating a lack of accountability for outcomes (Williamson, Colley & Foley, 2020).
Because blockers operate 'under the radar', champions play an important role in recognising and addressing resistance (Dass & Parker, 1999).Over time, powerful champions can overcome resistance and establish new organisational norms (Tilcsik, 2010).For example, a powerful champion who hires part-time employees and assigns them high-profile clients contradicts the 'ideal worker' norm; the champion establishes an alternative ideology that part-time workers are highly valued.The support of champions may be needed over many consecutive years, as changes to well-established systems take time (Gonzalez, 2010).

RESEARCH DESIGN
Following the findings of Gould et al. (2018a), we investigated TDEs within an Australian public service and observed a service-wide TDE that decelerated after an executive-level gender target was imposed.To understand why TDEs might decelerate after a gender target, we conducted a department-level investigation.We wanted to investigate variations in the TDE across departments and identify factors that explained the variations.Understanding this variation is important if public services are to derive maximum value from gender targets.Gender representation data at executive and executive feeder levels were collected from departments within an Australian public service.The data spanned 11 years.In the sixth year, the state government imposed a gender target for executives across all departments.

Potential decoupling in the Australian public sector
Reform of the Australian public sector has devolved functions previously performed by large central agencies down to individual departments.Departments are now responsible for people management across the entire employment life cycle, from employee recruitment to employee exit (Williamson et al., 2020).HR activities are performed by line managers who may not have the necessary motivation or skills (Colley, McCourt & Waterhouse, 2012).Line managers may recognise the need for policies to improve female representation in senior levels but, despite good intent, not have the HR skills to implement such policies (Podger, 2017;Williamson et al., 2020).Additionally, where implementation is further delegated to line managers, variability in policy implementation can result (Crilly et al., 2012).Devolution has also reduced central agency monitoring of policy outcomes and lowered departmental accountability for achieving policy outcomes (Colley et al., 2012).
In light of external stakeholder pressure, it is important to investigate whether decoupling between gender targets and their outcomes is occurring in the Australian public sector.In response to gender targets, departments should be able to appoint women to senior roles with the expectation that this will initiate a TDE.But if decoupling occurs, then the same gender target will have variable outcomes at a departmental level.As the largest employer in Australia, it is important to understand what the Australian public sector can do to increase the likelihood that gender targets work with, rather than against, TDEs.The Australian public sector presents an opportunity for investigating decoupling dynamics; devolution of HR creates a risk of both policy-practice and means-ends decoupling in this sector.

Loess curves
Loess curves were produced for 13 departments to identify whether a TDE was consistently operating within each department during the 11-year timeframe.A Loess curve fits a smooth curve through points in a scatter plot to reveal the relationship between a predictor (female representation at executive level in year one) and an outcome variable (female representation in the executive feeder group the following year) (Cohen et al., 2003).The Loess curves showed five departments (38%) with a consistent TDE across all years and eight departments (62%) with an inconsistent effect (e.g. a negative effect then positive effect).We divided departments into two categories for analysis: those with a consistent TDE (five 'consistent TDE' departments) and those without (eight 'inconsistent TDE' departments).

Interviews
To understand why there was variation across departments, 13 semi-structured interviews were conducted with CEs (3), executives (9), and an HR manager (1) from 10 public service departments.As senior staff with oversight roles, these individuals were 'elite informants' with access to privileged and exclusive information about the public service and their particular departments (Solarino & Aguinis, 2021).Each interview was conducted during November and December 2014 by the first author, and lasted approximately 1 h.Two interviewees were men (15.3%) and 11 (84.7%) were women.Seven interviewees came from consistent TDE departments and six from inconsistent TDE departments.As recommended by Francis et al. (2010), we began looking for saturation after the first 10 interviews.The final three interviews confirmed previously identified themes and did not generate any novel codes (Aguinis & Solarino, 2019), suggesting that theoretical saturation had been reached (Bowen, 2008).The interview protocol was developed based on policies, practices, accelerators, and decelerators identified in the women's advancement literature (Bilimoria et al., 2008;Ely et al., 2011).For example, interviewees were asked about the use of flexible work arrangements, leadership training, and mentoring programs, as well as recruitment practices that specifically targeted women.They were also asked if they were aware of individuals or groups who championed gender diversity or who downplayed the importance of gender diversity.The first author was employed in four different departments in this public service across 8 years.This personal knowledge and experience working in the public sector also contributed to the development of the interview protocol.For example, interviewees were asked about their department's response to the introduction of an executive-level gender diversity target.Additionally, interviewees were asked about 14678500, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/1467-8500.12576by University Of South Australia, Wiley Online Library on [07/03/2023].See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions)on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License the general environment in which the target operated, including the public service's strategic plan (containing the gender target), relevant stakeholders, and the people responsible for achieving the target.The purpose of asking about practices in departments was to understand whether a TDE was likely to have been initiated with top-down practices to increase female representation in senior roles and then supported with bottom-up practices that help women progress into senior roles.Where interviewees did not spontaneously discuss flexible work practices such as part-time work, leadership training, or mentoring, they were specifically asked if those practices were present.
Interviewees were then shown their department's Loess curve and asked whether the curve surprised them or was expected.This gave interviewees an opportunity to provide personal insights into why their department showed either a consistent or inconsistent TDE during the 11-year time period.

Analysis
Each interview was digitally recorded and transcribed verbatim by a professional transcription firm.Interview transcripts were analysed using NVivo software.Interview data were analysed using two-stage directed content analysis (Hsieh & Shannon, 2005).This two-stage method begins with a deductive process to confirm the theoretical framework and then moves to an inductive process to extend this theory.In the first deductive stage, the literature on women's advancement was used to identify key concepts that became initial codes for analysis (Potter & Levine-Donnerstein, 1999).For example, a gender champion within a department should help initiate and support a TDE, since the existence of such a champion increases the likelihood that potential female candidates are identified and encouraged to apply for executive and executive feeder roles.We arranged codes hierarchically under the headings 'policies and practices' and 'accelerators and decelerators'.For example, 'recruitment' was placed under policies and practices and 'champion' was placed under accelerators and decelerators.In the first deductive stage, all transcripts were coded based on these initial codes and following the process described by Hsieh and Shannon (2005).
In the second inductive stage, we reviewed all transcripts, looking for sections where women's advancement was discussed.We noted content that was outside this initial set of codes: superfluous codes, changes to the scope of existing codes, movement of codes within the hierarchy, and potential sub-categories of codes (Hsieh & Shannon, 2005).Codes were added when discussions went outside our coding based on the women's advancement literature.For example, the initial single code 'champions' was divided into two subcodes: CE champions and other (non-CE) champions.

FINDINGS
None of the interviewees were surprised by the shape of their department's Loess curve.All interviewees were aware of issues around the effectiveness of the gender target and had a clear understanding of how well (or poorly) their department was performing.Subsequent discussion revealed two overarching themes.First, policy-practice and means-ends decoupling arise from actions of external stakeholders, missing practices, a lack of accountability, and ineffective There was a potential for decoupling of the gender target from its implementation and intended outcomes.Policy-practice decoupling was likely for two reasons: external stakeholder actions and missing practices.First, the gender target was one of multiple targets imposed by the state government (external stakeholder), which meant departments might struggle to find sufficient resources to achieve the target.Second, interviewees revealed that important top-down practices to improve female representation were not even in play.Means-ends decoupling was also likely for two reasons.First, the state government prioritised some targets over others, which led to a lack of accountability ascribed to the executive-level gender target.Second, bottom-up practices were not implemented effectively, so intended outcomes of the target were not achieved.Discussions about practices were quite consistent across interviewees.Whether a department had a consistent or inconsistent TDE did not affect discussions of practices.

Policy-practice decoupling potential
External stakeholder action led to the potential for policy-practice decoupling.The executivelevel gender target was part of a service-wide strategic plan with approximately 100 targets in total.Departments had to make strategic resource allocation decisions and not all targets could be achieved.Where these decisions meant resources would not be allocated towards achieving the gender target, policy-practice decoupling occurred.Eight interviewees (61.5%) discussed the gender diversity target and how it was difficult to achieve given the plethora of strategic targets.
A typical example of interviewee discussion included: Yeah, I mean there's 90 [targets -note there were approximately 100] and it's one of the many.(Interviewee 8, Female Executive, Inconsistent TDE) Additionally, missing practices contributed to the potential for policy-practice decoupling.Every interviewee reported being involved in recruitment, but there were no practices designed to increase gender diversity by targeting women specifically.None of the interviewees used targeted recruitment practices, such as screening job descriptions for gendered language, including women on interview panels who recruit through their professional networks, lengthening shortlists, and requiring at least two women on shortlists, to appoint women to the executive level.Interviewees viewed job descriptions that included unnecessary technical skills as a barrier to appointing women to senior roles, with the following example showing a typical response: I sit on panels where they go 'we need 5 technical questions here, they need to know about blah, blah' -why?(Interviewee 2, Female Executive, Consistent TDE) Interviewees were asked whether there were any requirements to have a minimum number of women on a shortlist.None of the interviewees had observed this practice, so they were asked to think about a scenario where this minimum number was required.Three interviewees (23%) were quite supportive, with two examples given below: Well, just thinking in terms of the things where we could improve.I mean, you've sort of posed a couple of challenges almost, about should we be sort of planning in terms of our recruitment, should we have . . .diversity as a specific criteria?(Interviewee 13, Male Chief Executive, Inconsistent TDE) And the idea of a quota for the recruitment process, at least for the pool from which you draw, is a soft way to start the idea of quotas and I like it.(Interviewee 12, Female Executive, Inconsistent TDE) Interviewees were aware of the threat of policies being ignored and accepted this as being the result of too many targets.They also recognised that there were no practices designed to target women specifically.These findings were ubiquitous across the interviews; all departments had a potential for policy-practice decoupling.

Means-ends decoupling potential
A lack of accountability for achieving the gender target resulted in means-ends decoupling potential.Most of the strategic plan's targets were not related to gender but, in addition to the executive-level gender target, there was a government boards and committees gender target.Eight interviewees (61.5%) described strong stakeholder pressure for the boards and committees target, especially from a senior government council and government ministers.The pressure at the boards and committees level led to a lack of accountability for the executive-level gender target.Interviewee responses were consistent as shown below: I think the [senior government council] were riding [our minister] particularly hard on the boards and committees. . .I've never felt the pressure for the targets for executive; I've always only felt it for boards and committees.(Interviewee 4, Female Executive, Consistent TDE) We always dutifully put [target name] in our reports and things, but I've never seen an actual well thought out plan of how to achieve any of them.(Interviewee 5, Female Executive, Consistent TDE) Only one interviewee mentioned any form of gender metric as part of their individual performance evaluation.This metric did not have a strong impact, but at least it raised awareness of, and stimulated discussions around, gender representation in senior roles.Otherwise, there was little accountability for the target as shown in the following quote: . . .performance agreements with the Chief Executive, one of the mandatory measures was women in senior leadership . . .If you didn't meet some of those mandatory criteria, you simply weren't eligible for performance pay, not that there was ever much of that around.(Interviewee 6, Female Executive, Consistent TDE) Poorly designed practices also resulted in means-ends decoupling potential.Interviewees reported that bottom-up practices were ostensibly available to women in the public service.However, women were not consistently able to access these practices, which led to meansend decoupling; departments appeared to be helping women without actually achieving that result.For example, flexible work practices, such as part-time work, are designed to help women balance their work and family roles, making executive-level positions more attractive.Eight interviewees (61.5%) explained that it was difficult for executives to take advantage of these practices; executives were expected to put work before family so women selected themselves out of senior roles due to workload expectations.A typical response is provided below: . . . it was a quite deliberate decision [not to apply for an executive role] because of the work life, workload that they saw executives having. . . .(Interviewee 1, Female Chief Executive, Consistent TDE) Practices such as leadership training and mentoring help women progress through an organisation into executive roles.Six interviewees (46.2%) discussed leadership training.Training was generally not women specific, but one interviewee (7.7%) had developed a program specifically for women.Most training was ad hoc; individuals sought their own leadership development, whether within the public sector or outside.Only one interviewee highlighted cost as a barrier to accessing leadership training, suggesting budget is not a root cause of ad hoc training: . . .department's been quite proactive in promoting and funding those sort of things [leadership training] and providing opportunities and time, but I wouldn't say there's been a structured documented [approach].(Interviewee 3, Male Executive, Consistent TDE) Yes I mean there's [women's leadership training] stuff but I mean it always costs money and we're poor in our department.So you usually find lots of other departments go, but we don't tend to.(Interviewee 9, Female Executive, Inconsistent TDE) Four interviewees (30.8%) discussed mentoring.Women were encouraged to have mentors, but there was little departmental support for finding suitable mentors or taking time off for mentoring sessions.Three interviewees (23.1%) described formal mentoring programs, one of which focused specifically on women (two interviewees who discussed women-only mentoring were from the same department).Other mentoring was ad hoc, with departments supporting individual women who sought out mentoring opportunities, but not actively providing these opportunities.A typical example of the lack of mentoring support is provided: Yeah and you know at the end of the day the men who are mentored they weren't necessarily the best person for the job, but someone had faith in them and gave them a go and supported them and did the positive spin on how they were going.Women haven't had that and still don't have that.(Interviewee 11, Female Human Resources Manager, Inconsistent TDE) 14678500, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/1467-8500.12576by University Of South Australia, Wiley Online Library on [07/03/2023].See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions)on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Our analysis demonstrates that the lack of accountability, including a lack of metrics, meant that there were few consequences for not achieving the gender target.Departments were offering leadership training and mentoring, but not in a structured fashion.Departments appeared to be responding to the gender target, but their practices were ineffective.

Champions and blockers
Discussions with interviewees, regardless of whether they were from consistent or inconsistent TDE departments, identified a general risk of both policy-practice and means-ends decoupling.
All departments were at risk of decoupling, but departments diverged in the effects of internal champions and blockers.Departments that experienced a consistent TDE had internal champions who overcame blocker resistance.Specifically, CE support had a profound impact on non-CE champions' ability to improve gender representation at senior department levels.In departments with a consistent TDE, strong visible support from the CE gave non-CE champions the legitimacy to implement interventions even when they encountered resistance.In departments with an inconsistent TDE, the CE may have been supportive, but this support was not explicit and so blockers were able to disrupt champions' efforts.Blockers operated throughout the public service.If a department had gender champions, blockers resisted their change efforts and reduced the champions' impact.Blockers contributed to policy-practice and means-ends decoupling using both passive and active strategies.Blockers create decoupling potential and CE champions are necessary to maintain coupling.Eleven interviewees (84.6%) discussed the role of blockers, with three examples provided: . . .there were debates, particularly with some of the crusty old men about why you would do anything extra to encourage [women], because you know, they believed that it was purely merit based and couldn't see any argument in doing anything extra [for] women at all.(Interviewee 6, Female Executive, Consistent TDE) [Department] was pretty good at passive resistance, everyone would nod and say it's a great idea but then they'd just go on and do their own thing.(Interviewee 1, Female Chief Executive, Consistent TDE) We did get some resistance because, especially when we started the women's mentoring program, the blokes were saying 'what about us'? (Interviewee 11, Female Human Resources Manager, Inconsistent TDE) However, interviewees emphasised that the CE's behaviour was essential to keep the gender target coupled with its implementation.Some CEs (but not all) 'led loudly'.These CEs kept their departments focused on the target and ensured it was part of day-to-day operations.Champion CEs provided clear support for initiatives such as part-time work.Other CEs were less supportive.These CEs did not visibly embrace the gender target or support interventions that might have improved female executive representation.Six interviewees (46.2%) discussed the role of the CE as either a champion or non-supporter.The following quotes demonstrate these contrasting CE roles: 14678500, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/1467-8500.12576by University Of South Australia, Wiley Online Library on [07/03/2023].See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions)on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License He [the CE] said 'your permanent role isn't really a 2 day a week job it's a 5 day a week job so . . .I'll find a special project for you to do that's meaningful and real and has some teeth in it and let's do that as the transition' . . .I felt comfortable with that as a way to come back [from maternity leave].(Interviewee 5, Female Executive, Consistent TDE) [CE said] I don't know how you work as an executive part-time, I don't understand how you do that.(Interviewee 10, Female Executive, Inconsistent TDE) Without support from the CE champion, non-CE champions struggled to maintain momentum towards the gender target.These non-CE champions tried to raise gender issues but were unable to overcome the blockers.A challenge faced by non-CE champions is that improving gender representation at senior levels was often not an explicit part of their job descriptions; they made time for gender issues on top of their regular workload.One interviewee worked with other non-CE champions and conducted a survey to document gender issues in their department.When they presented the survey findings, the executive group reluctantly acknowledged the lack of gender representation at executive level.This acknowledgement reflected the high-quality unequivocal data provided in the champions' report.Yet the executive group took no proactive steps in response to the non-CE champions' efforts, and no improvements occurred.The quote from this interviewee is provided below: . . .what was the support like?Well my belief is that verbally [CE and executive group] would support it and if you went and requested, they were supportive but how proactive they were . . .so it's more of a culture issue that needs to be addressed.(Interviewee 11,Female Human Resources Manager,Inconsistent TDE) In departments with a strong CE champion, blockers did not necessarily embrace improvements in gender representation, but they did comply with change initiatives.Without CE support, as non-CE champions left the organisation, reduced their commitment, or lost resources, momentum was lost.The 11 interviewees (84.6%) who discussed non-CE champions emphasised that less powerful champions needed CE support to make progress towards the target, as shown below: [According to the CE] . . .Yes, so in order to make that change you need to just sometimes throw everything out and start over again which is what he's [CE's] done.(Interviewee 5, Female Executive, Consistent TDE) They were talking about [gender at executive level] and I know it was lip service, but they behaved and sometimes compliance and behaving is the best you can get . . .But the issue for me is the moment that you leave that gets left behind so I think [department] has gone backwards now.(Interviewee 2, Female Executive, Consistent TDE) Blockers existed in all departments.These individuals or groups made it more difficult for champions to make progress and were successful without strong CE support for the gender target.
Non-CE champions could not maintain momentum, with consistent CE championing to keep a department's focus on gender over extended periods.

DISCUSSION
We investigated the TDE at department level in an Australian public service in which a gender target was imposed.We confirmed the existence of a TDE at a whole of public service level and distinguished between departments with a consistent TDE and those whose effect was inconsistent.Gender targets should help initiate and support a TDE through effective implementation of top-down practices that increase female representation at the senior level and bottom-up practices that help women progress.Our study shows that just setting targets is not enough; departments are susceptible to policy-practice and means-ends decoupling and only the CE's championing ensures a target achieves desired outcomes.

Decoupling dynamics
Our research contributes to the decoupling literature and the literature on creating a genderinclusive public sector.First, we draw on decoupling theory to understand why gender targets are not consistently effective.Our interview findings showed that there is a decoupling potential with a gender target; the gender target may be ignored (policy-practice decoupling) and/or not achieved due to the adoption of an ineffective combination of practices (means-ends decoupling).
Policy makers expect that gender targets should improve female representation at the focal level of the target.However, our findings demonstrate that targets alone will not lead to desired outcomes.Additionally, we demonstrated that a service-wide gender target did not generate consistent TDEs across departments.Decoupling risks are greatest when organisations are confronted with multiple stakeholder demands.When the state government (external stakeholder) set multiple targets, departments with limited resources had to make allocation decisions based on the importance of the goal and the power of stakeholders (Westphal & Zajac, 2001).The gender target at executive level was only one of almost 100 targets included in the government strategic plan, which led to policypractice decoupling.Multiple targets made it easier for departments to avoid targets they did not value (Colley et al., 2012).Further, the state government was focused on a boards and committees gender target, resulting in a lack of accountability for achieving the executive-level target and means-ends decoupling.
We contribute to understanding practices that help create a gender-inclusive public sector.Departments are more likely to cultivate supportive climates for women when they adopt integrated bundles of practices that send consistent signals about the value the department places on gender diversity (Ostroff & Bowen, 2016).We highlight the value of top-down and bottom-up practices in resisting decoupling and creating gender-inclusive work environments.Top-down practices increase the likelihood that women will apply for a senior role, will be chosen, and will accept the role (Johnson et al., 2016;Kazmi et al., 2021;Lucas et al., 2021).Without women in senior roles, the TDE cannot operate.Our findings indicated targeted recruitment practices were missing-evidence of policy-practice decoupling.Other top-down practices, such as parttime work, were ineffectively implemented-evidence of means-ends decoupling.Women had access to part-time work, but negative perceptions about working part time and executive roles (Mainiero & Sullivan, 2005) meant women did not take advantage of part-time work options.If a TDE is initiated, bottom-up practices, such as leadership training and mentoring, support the TDE by providing women with the human capital needed to progress through the organisation (Metz & Tharenou, 2001) and by helping women leaders counter systemic biases (Ely et al., 2011).In this public service, access to these practices was limited; women needed to seek out their own leadership training and mentors.
Finally, we highlight the role of key players in decoupling dynamics.Departments were all subject to the same gender (and other) targets and had the same reporting requirements.Further, department practices to improve gender equality were generally consistent across departments and all departments had individuals who resisted change (blockers).Essentially, all departments had the same potential for both policy-practice and means-ends decoupling.Nonetheless, some departments did experience a consistent TDE.The factor that distinguished departments with a consistent TDE was a powerful gender champion (the CE) who provided visible support for achieving the gender target.Without support from a CE gender champion, the TDE is inconsistent or absent, reducing the impact of a gender target on the level below the focus of the target; the gender target becomes decoupled from its implementation.

Gender champions overpowering blockers
Devolution of HR functions to departmental level in the Australian public sector creates a context where service-level policies can be ignored or ineffectively implemented (Williamson et al., 2020), resulting in the potential for policy-practice and/or means-ends decoupling.When there is a potential for decoupling, a powerful gender champion keeps focus on a gender target (de Vries, 2015).The champion can ensure that leadership training and mentoring take place, but these bottom-up HR practices may not be implemented in a meaningful way if blockers resist them.For example, leadership training may be offered predominantly to men, or women's mentors may not have the resources and networks needed for women to advance.Champions support policies and practices designed to overcome barriers to women's progression (Metz & Kulik, 2014), thereby supporting the TDE.Non-CE champions want to address gender issues, but they usually do not have this responsibility embedded in their job role.And if they leave the organisation, reduce their commitment, or lose resources, momentum is lost (Hekman et al., 2017).The role of CE champions was clear in our study.In departments with a consistent TDE, the CE was proactive and their support was visible.In the presence of a CE champion, blockers comply with change efforts even when they do not endorse this change.The CE's explicit support sends a message that change is inevitable, an important signal when resistance to change is active (Kellough & Naff, 2004).When CEs are supportive, blockers' concern for their personal positions leads to compliance (Herscovitch & Meyer, 2002).Without visible CE support, the blockers win.

RECOMMENDATIONS
To avoid policy-practice decoupling, the total number of targets should be limited so that departments can manage day-to-day departmental business while keeping focus on achieving targets.Too many targets can stretch the resources available to the department and dilute the focus on gender targets.Additionally, to avoid means-ends decoupling due to a lack of accountability, departments should be required to regularly report on gender representation at senior levels, ensuring consistent commitment across departments that is less dependent on CEs' personal values.While many managers support improvements in gender diversity, they may not have the HR skills required to effectively implement gender targets.Understanding decoupling theory and the risk of decoupling when a gender target is imposed will help managers recognise the importance of top-down practices to initiate a TDE and bottom-up practices to support a TDE.Departments should require evidence-based recruitment practices that specifically target women, such as removing gendered language from job descriptions, including women on interview panels who actively recruit via their professional network, lengthening shortlists, and requiring at least two women on shortlists.This will avoid policy-practice decoupling due to missing top-down practices.Additionally, bottom-up flexible work arrangements need to be effectively implemented so that means-ends decoupling does not occur.Overcoming negative perceptions associated with part-time work requires visible support from the top, including the CE.Departments should offer leadership training, ensuring that women are well represented in this training, and formal mentoring, helping women overcome a lack of access to suitable mentors via informal networks.
Limiting targets, ensuring accountability, and having the right practices are important for departments to reduce the potential for decoupling.However, departments also need gender champions who challenge the status quo and drive change.To overcome resistance to change from blockers, non-CE gender champions need visible support from the CE.Hence, the role of gender champion must be formally written into the CE job description and treated as a key indicator in performance reviews.

CONCLUSION
Organisations set gender targets to improve female representation at a focal level.The TDE provides an exciting opportunity for organisations to improve gender representation at the level immediately below the target's focal level.However, the TDE is vulnerable to decoupling.To enable a consistent TDE, public services need to ensure that gender targets are not lost in a multitude of targets and accountability is assigned to achieving these targets.Additionally, gender champions are needed to keep targets coupled with implementation.These champions must either be the CE or have explicit support from the CE.Without CE support, blockers will challenge gender champions and derail their efforts.

A C K N O W L E D G E M E N T S
14678500, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/1467-8500.12576by University Of South Australia, Wiley Online Library on [07/03/2023].See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions)on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 14678500, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/1467-8500.12576by University Of South Australia, Wiley Online Library on [07/03/2023].See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions)on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License practices.Second, internal champions, especially the CE, can address decoupling and overcome blocker resistance.
14678500, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/1467-8500.12576by University Of South Australia, Wiley Online Library on [07/03/2023].See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions)on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Open access publishing facilitated by University of South Australia, as part of the Wiley -University of South Australia agreement via the Council of Australian University Librarians.Aguinis, H., & Solarino, A. M. (2019).Transparency and replicability in qualitative research: The case of interviews with elite informants.Strategic Management Journal, 40(8), 1291-1315.https://doi.org/10.1002/smj.301514678500, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/1467-8500.12576by University Of South Australia, Wiley Online Library on [07/03/2023].See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions)on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License R E F E R E N C E S