Concepts incorporating dynamics in systems thinking

All systems, both natural and social, experience change over time, involving both external and internal processes, acting at a range of scales. A system can be characterised by a reference state, described by the average condition, and expected variability around that average over a specified interval (Rykiel 1985; Ratajczak et al. 2018). A general definition of disruption is the violent dissolution of continuity, or forced change, in the reference state (Simpson et al., 1989). Abrupt change is characterised by a large amplitude or short period, relative to the expected magnitude and frequency of variability for the system (Ratajczak et al. 2018). A system’s response to disruption depends on both the characteristics of the disruption and the characteristics of the system. For this reason, theorists distinguish between disturbance (cause) and perturbation (effect) (Rykiel 1985). For the purpose of the present work, disruption events are considered to be discrete events in time that alter the components of the system relative to a prior reference state. They may include changes to government policy (Hossain et al. 2020), introduction of new technology, and social-economic forces such as employee group behaviour (Herbst 1957). Further, while the system contains many individuals, disturbances that have a large impact on individuals may not affect the function of the system as a whole (Rykiel 1985; Ratajczak et al. 2018).

Large organisations such as DOC can be considered to be complex adaptive systems comprised of multiple interacting components (Berbés-Blázquez et al. 2014). These systems are characterised by nonlinearity, feedbacks, emergent behaviour arising from the interaction of multiple processes, and other attributes that may be difficult to identify (Hossain et al. 2020). Such systems have the capacity to learn and adapt in response to both gradual and abrupt forcings (Bennett 2006; Chapin et al. 2009; Berbés-Blázquez et al. 2014; Manfredo et al. 2017) and those responses may also be either gradual or abrupt. For example, an employee group’s values may suddenly change in response to a significant context change, such as a change in business model (Manfredo et al. 2017). In this sense, our research is an ecological study of the limiting and supporting factors to female and male participation, achievement, and progression in the DOC system. Disruption events can also have more subtle effects, such as changes to power dynamics or opening of new pathways for employee advancement (Bennett 2006).

Systems theory offers theoretical frameworks for analysing and understanding the behaviour and structure of such systems (Stokols et al. 2013; Berbés-Blázquez et al. 2014; Hossain et al. 2020). This theory was developed when linear cause-and-effect approaches failed to generate robust explanations of systems characterised by nonlinearities, feedbacks and so on (Hossain et al. 2020). Holistic analysis of the entire system, including its environmental context, underpins the systems thinking approach (Berbés-Blázquez et al. 2014; Hossain et al. 2020).

Using the social-ecological systems model to examine influences on DOC operations, structure, and gender participation patterns

Combining social and ecological systems approaches, Bronfenbrenner’s (1992) social-ecological system (SES) model conceptualises a complex adaptive system using interconnected hierarchical layers (Bronfenbrenner 1992; Kilanowski 2017). The nested-circles SES schematic () places an individual as the focal organism within a layered hierarchy. The microsystem comprises interactions and relationships of the individual’s immediate surroundings and provides the strongest influence. The mesosystem comprises interactions between microsystems, such as between different groups of employees within an organisation. The macrosystem includes societal, religious, and cultural influences. Surrounding all other layers, the chronosystem encompasses time as it relates to both external and internal processes (Kilanowski 2017). Thus, the SES model provides a framework in which the interconnections among components of the system can be examined.

Processes operating at all levels of the social-ecological system, from individual responses to the environment at the focal organism level to cultural expectations at the macrosystem level, affect staff access to opportunities and equity. Equity involves identifying and responding to differences that are unfair, avoidable, and changeable, to give people the means to participate (Clow et al. 2009). Equity measures aim to support individuals so that the community can achieve equality of opportunity and participation (Clow et al. 2009). In western culture, gender has historically been defined dichotomously as either masculine or feminine. In recent years, the western definition of gender has evolved to recognise a phenomenological view which includes identities outside the gender binary (Castellini 2017). The literature and data providing the basis for this research relies on the historical dichotomous definition of gender, which has admitted only (mutually exclusive) male and female categories. This classification limits our insight into the diversity of gender identity and gendered experiences. New data collected as part of the annual StatsNZ Household Economic Survey indicates that 0.8% of the adult population identifies as non-binary or transexual (Stats NZ 2021). Other dimensions that would add depth to the understanding developed in this research include social-signifiers such as race, economy, religion, sexuality, age, and ability (Gay-Antaki and Liverman 2018), however, such information is not available as part of the data sets used here.

At the chronosystem level, organisational history has been demonstrated to influence gender participation through, for example, response to changes in leadership (Bruckmuller and Branscombe 2010). Power differentials are also evident through time, in which females are channelled into subordinate roles, particularly within economic and political spheres (Brownmiller 1975; Byrne 1993; Bennett 2006). Phenomenological barriers that exclude females from the pinnacle benefits of a given field are so well-documented that they have been synthesised into popular metaphors (Blickenstaff 2005; Ryan and Haslam 2005; Mandel and Shalev 2009; Cook and Glass 2014). Two such examples are the glass ceiling and glass cliff. The glass ceiling represents an invisible barrier to female advancement (Ryan and Haslam 2005; Mandel and Shalev 2009; Cook and Glass 2014) while the glass cliff describes a phenomenon in which females are more likely than males to rise to senior leadership in institutions that are in a state of disruption (Ryan and Haslam 2005; Bruckmuller and Branscombe 2010; Ryan et al. 2016). Although a combination of male and female decision-makers significantly increases the odds of a female being promoted to senior leadership (Cook and Glass 2014), a threshold of 30% female representation is required in an occupation or discipline for female representation to be perceived as normative (Byrne 1993). Below the 30% threshold representation is seen as non-traditional (Byrne 1993) and the organisational culture is unlikely to support females’ advancement (McPherson 2010).

The Department of Conservation is a hierarchical system (). Applying the SES model, or systems lens, to DOC engages historical (temporal), physical (spatial), environmental and social-political contexts so that the hierarchical network of elements that influence the organisation can be identified (Chapin et al. 2009; Stokols et al. 2013). In this paper, the systems lens is used to examine how disruption events have influenced DOC’s operations and structure since its establishment. Gender participation patterns within the organisation are then examined within that framework.

Figure 3. DOC organisational chart showing the five Tiers as at June 2019 (current to February 2020). Adapted from DOC (n.d.[a]).

Thematic analysis

Six themes, described in this section, were revealed through the corporate document analysis: leadership, finance, structure, strategy, legislation, and organisational management (Supplementals A–C, and 1, 2).

Gender participation patterns across the DOC organisational hierarchy

Tiers 2–4: gender participation patterns

DOC’s senior leadership in Tiers 2–4 has been dominantly male from the year 2000 (DOC 2018d, and ). A 1995 report indicates that males dominated both leadership and general staff roles prior to that time (O’Brien 1995). Of the 14 government Ministers responsible for DOC oversight between 1987 and 2019 (Parliamentary Information Service, pers. comm., January 21, 2021), 59% were male and every DG (Tier 1) since 1987 has been male (B. McDonald, pers. comm., November 27, 2018, Supplemental 2).

Figure 4. DOC populations in Tiers 2–5 (a–d), from 2000 to 2019. Solid colours represent female (left of 0) and male (right of 0) cohorts in each Tier. Hashed grey bars show F/M parity for each population size so that the un-hashed ends of bars indicate over-representation. ‘gc’ indicates a glass cliff pattern (see text). Arrows indicate where male representation increases (Data source: DOC 2018d).

Figure 5. Population growth in Tiers 2–5 (a), and female representation in Tiers 2–5 (b), 2000–2019. Each Tier population is normalised by the year 2000 population to facilitate comparison across Tiers, given their different sizes. Tiers 2–4 are DOC leadership Tiers. Stars highlight disruptive and change events from Supplementals A, B and C (Data source: DOC 2018d).

As of June 2019, there were almost 300 directors and managers across DOC Tiers 2–4 (). The population size of each of these leadership Tiers has varied over time, as has the proportional representation of females (). The population base state for Tiers 2–4 changed around the year 2012 (A). Prior to that year, all three Tiers were characterised by large-amplitude variation. After that time, Tiers 3 and 4 experienced growth and reduced variability, nearly doubling in size relative to the year 2000. The boundary between 2011 and 2012 is thus used as a chronosystem time marker in the analysis presented here.

Females have been, with rare exceptions, under-represented in every leadership Tier throughout the chronosystem record ( and ). Only in Tier 2 were females ever over-represented; in the year 2007, and across 2013 and 2014 (). The period between these boundary years is unprecedented with respect to female participation in top leadership roles and is thus also used as a chronosystem time marker in the analysis presented here.

The Tier 2 population size ranges between 6 and 8 individuals (A). Female and male cohort size is strongly negatively correlated over the whole time series at any population size ( and A). An episode of female over-representation during the time period was followed by decline in the female Tier 2 cohort size concurrent with an increase in the male Tier 2 cohort size.

Figure 6. Female and male cohort size within Tiers 2–5 (a–d). Tiers 2–4 are DOC leadership Tiers. The symbol colour indicates year in the data set as in the legend. Circle-shaped and diamond-shaped symbols indicate years before and after 2011, respectively. This is a chronosystem time marker discussed in the text. Years from 2007 to 2014, another chronosystem time marker, are highlighted using a heavy grey outline on the symbols (Data source: DOC 2018d).

The Tier 3 population size varies between 25 and 48 individuals (B). Female and male cohort sizes are positively correlated in the period from 2000 through 2011 (). The positive slope less than 1 indicates that males are favoured by population change during this period, in this Tier. That is, when the population grows, proportionally more males are added, and when the population reduces, proportionally fewer males are lost. No statistically significant pattern is evident in the second part of the record (from 2012 to 2019). Female participation in the Tier 3 population first exceeded 30% in 2018 and was at 33% in 2019.

The Tier 4 population size varies between 116 and 240 individuals (C). Different statistical relationships are apparent in the intervals before and after the 2012 chronosystem time marker. No clear relationship is found between 2000 and 2011 (). In contrast, the regression slope for the 2012–2019 interval is 1.4, indicating that females benefited more than males from population change during this period. Female participation in the Tier 4 population first exceeds 30% in 2013, increasing to 44% in 2019.

Tiers 2–4: pay patterns

In 2019, DOC undertook a pay gap analysis as it developed a Gender Pay Gap Action Plan (DOC 2019b). Data from that analysis relevant to this study are presented and evaluated here ( and ). In contrast to some other Aotearoa New Zealand organisations (Pacheco et al. 2017), the DOC gender pay gap does not consistently increase up the wage continuum (). Males in Tiers 2, 3 and 4 have in general been paid more than their female counterparts (positive % difference in the tables and ), however, the pay gap has in general reduced over time. This change may be a direct result of macrosystem events identified in the Timeline(Supplemental 2). DOC’s People Plan 2012, developed in 2008 and refreshed in 2009, addressed pay equity in response to policy direction from the Department of Labour (Supplemental B, C). Grouping Tiers 2–4 together and removing an outlier (see B), the mean pay gap across all leadership Tiers was 6.8% (SD = 6.1%, n = 29) between 2000 and 2009, before the People Plan refresh, and 0.3% (SD = 4.1%, n = 29) between 2010 and 2019, after the refresh. Pay equity action prescribed in the People Plan 2012 may thus have reduced the pay gap (Supplemental B, C; DOC 2018d, 2019b). Processes associated with the chronosystem time marker 2011/2012, after which Tier 3 and 4 populations rebounded, may also be important. Policies implemented at the start of the 2012–2019 period of sustained growth may have resulted in more equitable pay.

Figure 7. (a) Annual average percent gender pay gap in DOC leadership Tiers 2–4 as a function of the proportion of females in each tier, 2000–2019. Positive percentage indicates a gap that favours males. (b) Box plots for the two time periods with 75th percentiles, medians, 25th percentiles. The whiskers are at ±2.7σ and an outlier is identified using Tukey's rule (Data source: DOC 2019b).

 

Table 3. Gender pay gap as a function of female representation.

 

Table 4. Tier 2–5 staff by Business Group (data from DOC 2019b).

 

Table 5. Tier 5 by occupation type and gender pay gap between male and female average and median salaries by occupation type (data from DOC 2019b).

The DOC gender pay gap appears to have been changing prior to the People Plan 2012 (). Qualitatively, the measured pay gaps are larger when female representation in leadership roles is below 30%, the threshold at which a demographic group’s presence becomes normative (Byrne 1993). With this threshold in mind, the relationship between female representation and the leadership pay gap is examined. Data from all leadership Tiers 2–4 are grouped and linear regression is used to find the slope and its statistical significance for the whole data set and for subsets below and above 30% female representation (). The relationship between female participation rate and pay gap size is highly significant below the 30% threshold and not significant above it. The implication is that increasing the proportion of women in leadership roles had a positive influence on pay equity for female leaders in DOC prior to specific equity action intended to address the pay gap.

The chronosystem

Both external and internal processes shape institutional behaviour and structure. DOC’s chronosystem, represented by the summary themed Supplementals A–C and 2, the Timeline, identifies specific disruption events including strategies, policies and actions that have affected the DOC system at all scales (from macrosystem to focal organism) (sensu Bronfenbrenner 1992) and at different rates and magnitudes (Chapin et al. 2009; Berbés-Blázquez et al. 2014). National scale economic conditions, election cycles and policies that play out over years to decades interact to condition the annual budgets, strategies and policies of organisations operating within the national context. This, in turn, conditions opportunities and barriers for individuals.

The macrosystem: social-political disruptive influences affecting gender participation patterns within New Zealand

A clear pattern emerges in the chronology of empirical milestones influencing gender participation and pay parity policy in Aotearoa New Zealand, in which various social policy innovations appear in other Western societies before their adoption in this country (Supplemental 3). The sequence in Aotearoa New Zealand appears to be a direct linear response to the international context. While Intergovernmental Organisation Treaties are not themselves part of domestic law, by ratifying the treaties, Aotearoa New Zealand is obligated to recognise various rights and make them enforceable by law (Rishworth 2016). For example, Aotearoa New Zealand is a party to seven of the nine core international human-rights treaties. Two of particular relevance to gender participation are the Convention on the Elimination of All Forms of Discrimination against Women (CEDAW) which obligates New Zealand as a signatory to end all forms of discrimination against females. The other is the International Covenant on Economic, Social and Cultural Rights (ICESCR) which protects the fundamental equal right of males and females to the enjoyment of all economic, social, and cultural rights that it articulates.

The State Services Commissioner (note in 2020 the role title changed to the Public Service Commissioner [Hipkins 2020]) provides leadership and oversight of New Zealand’s state services, which includes promoting, developing, and monitoring Equal Employment Opportunities (EEO) within the public service (State Services Commission 2019). SSC initiatives can directly affect the leadership and capability opportunities offered to public servants, including DOC staff. The EEO objectives include, inter alia, identifying and developing leadership capability and promulgating a culture of stewardship in the public service (State Services Commission 2019). A current policy goal of the SSC is to facilitate a mobile public service leader workforce to move between agencies. Relevant fora include the Public Service Leadership Group composed of a cohort of leaders who network across the sector, and SSC career boards, established to support talent moves for senior leaders who have the aspiration and potential to move across the sector (S. Thomas, pers. comm., March 23, 2020).

The Human Rights Commission (HRC), working under the Human Rights Act 1993, seeks to promote and protect the human rights of all people in New Zealand. The HRC has up to five Human Rights Commissioners, including the Equal Employment Opportunities (EEO) Commissioner. The EEO Commissioner monitors access to employment to promote decent work for all New Zealanders, free from discrimination, harassment, bullying and victimisation (Human Rights Commission 2020).

The Ministry for Women supports the Minister for Women by providing advice on issues affecting females – such as financial security and full participation in New Zealand society – as well as managing international reporting obligations (Ministry for Women 2020).

Non-Governmental Organisations (NGOs) are influential in progressing gender equality within the country. For instance, the National Council of Women advocates for gender equality via the CEDAW process and on other international bodies (see National Council of Women of New Zealand 2015). Individuals also provide important political and social influence on the issue of gender equality, particularly via internet and social media connectivity (Dalziel 2018). Movements such as #MeToo have catalysed social change globally by challenging long-standing power structures. It is too early to say whether such challenges signify a tipping point in females’ political action (Dalziel 2018).

As a central government agency DOC’s priorities, policies and objectives are collectively decided between the Minister of Conservation and DOC’s Director General and must be consistent with government policy and Cabinet decisions (DOC 2006, 2008b, 2011b). Macrosystem policy can create both incentives and disincentives that either narrow or widen gender disparities (Hulbe et al. 2010), either directly or via feedbacks and emergent properties of the complex social-ecological system. The central government gender equity policies enacted in 2007 and 2017 are disruption events that should be expected to improve female representation within DOC. However, our analysis of the complete time series does not suggest a simple, uniform relationship between cause and effect in the organisation (). Instead, female representation is characterised by different patterns in different Tiers, suggesting that additional drivers may be important.

Thirty percent participation is recognised as a critical threshold for an under-represented group to be perceived as normative within a population (Byrne 1993; McPherson 2010). The proportion of leadership roles held by women rose to 30% in all three leadership Tiers for the first time in DOC’s history in 2019 and this may be a tipping point for the organisation. However, women have occupied more than 30% of roles in Tier 5 since at least 2000, but gender parity has not yet been reached (B). Altogether, the different patterns across Tiers and across time require both macrosystem and microsystem process explanations.

The microsystem: disruptive influences affecting gender participation patterns

A system’s history influences its response to disruption events and the nature of individuals’ interactions with the system (Harley et al. 2017). For example, institutions with a history of male leadership are more likely than institutions with more balanced gender history to exhibit glass cliff patterns, in which females are more likely to be hired to senior leadership roles during times of disruption while males are more likely to be hired during times of stability (Bruckmuller and Branscombe 2010). Such drivers on the operating environment of DOC, along with past structure, strategy, culture, and organisation management themes are explored here.

Historic drivers: Conservation was formalised in Aotearoa New Zealand within a colonial context and an essentialist ideological frame (Hodge 1998; Young 2004), in which female scientists had access to amateur spheres while male scientists had access to professional spheres (McEwan 1998). Gender essentialist beliefs hold that females and males have innately different worldviews, decision-making and leadership styles aligned with their respective biology and physical characteristics (Hoobler et al. 2018). In line with these beliefs accepted normative female behaviours then involve communalassociations, including the conveyance of concern, kindness, compassion, and modesty. Accepted normative male behaviours then include agentic traits such as strength, assertion, ambition, domination, and control. The agentic qualities are often viewed as synonymous with leadership qualities (Hoobler et al. 2018). Many contemporary scientists discount essentialist notions of gendered behaviour-traits (Fine 2014; Hoobler et al. 2018) for the reason that essentialism ignores environmental influences on brain development, human behaviour, and characteristics (Fine 2014; Allen 2018; Hoobler et al. 2018). Nevertheless, the explanations provided by essentialist normative views have been used to legitimise imbalances in social power between groups (Heyman and Giles 2006) and discount female contributions to science and the economy (Merchant 1989, 2006; Leach and Green 1997).

Manipulations of social norms can legitimise imbalances between males and females in various professions over time (Bennett 1996, 2006; McEwan 1998). For example, during the nineteenth century, two spheres of scientific activity developed in Britain and its colonies; a professional, public, institutionalised sphere for males, and an amateur, domestic, private sphere for females (McEwan 1998). The social context, such as familial trade-offs (Abri-Am and Outram 1989), was critical to allowing (or not allowing) the participation of females in science. Around the turn of the twentieth century, there was a drive to redirect science education for girls and women toward domestic science which had the dual effect of protecting male autonomy over access to the scientific professions and exerting male authority over the home (Sheffield 2006). The patterns evident in female access to science have been argued to apply also to the environment and to indigenous people; which it is argued were attributed value according to their usefulness to males (Merchant 1989, 2006; Leach and Green 1997). In Aotearoa New Zealand, conservation efforts have become incrementally strengthened following the legal recognition of the interdependence of environmental, social, and economic values (Young 2004) and, over time, female representation in the sciences has overall increased (Bray and Timewell 2011; Pettorelli et al. 2013).

Disruptions: structure change within DOC

In general, organisational restructuring events can perturb staff wellbeing, commitment to the organisation and retention (e.g. Theissen 2004). At least six formal restructures are identified in the Timeline, in 1989, 1996, 2008, 2011, 2012 and 2013. Varyingly described in the corporate literature as restructures, reviews or refinements, the events were variable in scope, scale and impact (DOC 2010b, 2012a, 2012b, 2013b, 2015a, 2015b, 2018c, 2018e; State Services Commission, Treasury & DPMC 2014; Supplemental 2).

The first restructuring occurred in 1989, shortly after the establishment of DOC, and resulted in staff and budget cuts (Napp 2007). By 1993, DOC was presenting itself as a single coherent agency (Napp 2007). DOC’s second restructure, in 1996, was catalysed by a shock in 1995. A disaster occurred when a DOC-built platform collapsed at Cave Creek on the South Island’s West Coast resulting in multiple fatalities (Young 2004). DOC adapted by establishing line management functions, along with clear role responsibilities and accountabilities within its decentralised Field Centres, Area Offices, Conservancies and Head Office (Napp 2007). DOC underwent its third restructure in 2008, which again resulted in staff and budget cuts (DOC 2010a).

A number of strategic and budget drivers catalysed DOC’s fourth restructure in 2011 (Supplementals B and 2) (DOC 2011a, 2011b, 2014a, 2014b, 2014c): a cost-saving Business Improvement Programme was implemented in response to the government’s goal to lift the rate of economic growth (DOC 2011a), a recognition that conservation outcomes benefitted all New Zealanders (DOC 2005b, 2006), and the acknowledgement that despite DOC’s efforts, biodiversity was declining and the organisation needed external input to arrest decline (DOC 2005a, 2007c, 2009). Important external drivers of change came from the State Services Commission, Treasury, and the Department of the Prime Minister and Cabinet, who collectively led three Performance Improvement Framework Reviews of DOC, in 2010 (DOC 2015a), 2014 (State Services Commission, Treasury & Department of the Prime Minister and Cabinet [DPMC] 2014) and 2016 (DOC 2017a). The reviews initiated significant organisation and leadership change from the top down (DOC 2015a, 2017a; State Services Commission, Treasury & DPMC 2014). DOC was instructed to improve strategic leadership, clarity of vision, values, and strategy, and to align its priorities to the government’s (DOC 2015a; State Services Commission, Treasury & DPMC 2014). An Excellence Horizon was set out, prescribed by the second review (State Services Commission, Treasury & DPMC 2014), describing the changes to environment, business strategy and operating model required for DOC to be successful (DOC 2015a).

In response to the Performance Improvement Framework reviews, DOC explicitly reoriented its values outward, to support external partnerships and commercial activity (DOC 2011a, 2011b). The fifth restructure, in 2012, implemented a shared-services business model and staff numbers in the national and conservancy offices were reduced (DOC 2012a, 2012b, 2013a). The sixth restructure, in 2013, resulted in a new operational delivery structure. Existing offices were dismantled while new groups were simultaneously created and total staff were reduced by c. 9% from 2011 levels (State Services Commission, Treasury & DPMC 2014).

The DOC system response to restructuring is characterised by different gender participation patterns within each Tier. The period between 2007 and 2013/14, chronosystem time markers in the Tier 2 gender participation rate, is concurrent with the period of active restructuring from 2008 to 2015. In the context of microsystem processes, the rapid rise in female representation rate at the start of the period and rapid decline in female representation at the end of the period (A and 4B), are interpreted as a glass cliff pattern (Ryan and Haslam 2005; Bruckmuller and Branscombe 2010; Ryan et al. 2016). Initially, restructuring appears to create opportunities for females at Tier 2, but they do not persist.

An overall upward trend in the Tier 3 population is interrupted by the 2008–2015 period of restructuring, during which the Tier 3 population declines. Overall, population change in this Tier favours males ( and B), but shorter intervals display markedly different patterns (B). Women’s participation grows to 22% in 2004, after which a period of decline begins. The downward trend reverses during the period disrupted by restructuring, and participation grows from a low of 14% in 2009 to a high of 26% in 2015, after which it again declines, even as the Tier 4 population size grows. In the context of microsystem processes, this may be interpreted as a glass ceiling between Tiers 4 and 3 (Ryan and Haslam 2005; Mandel and Shalev 2009; Cook and Glass 2014). The participation rate rebounds in 2018 and rises to above 30% for the first time in the data record (B).

The Tier 4 population changes in an opposite sense to the Tier 3 population prior to and during the period of restructuring. The population declines prior to this period, is variable throughout restructuring, and begins a nearly steady rise near the end of the period (A). Population decline prior to restructuring is associated with declining female representation ( and B). When population growth resumes it benefits females more than males and the proportion of females in Tier 4 roles rises to above 30% by the end of the period of restructuring and remains above that threshold through the rest of the data record (B and ). The growing proportion of women in Tier 4 roles corresponds with the unprecedented international trend of females participating in middle-management (Ryan et al. 2016).

Overall, women gain access to Tier 5 throughout the data record but the process responsible for that access appears to change over time. Early in the record, both male and female cohort sizes both increase as the Tier population grows. Women are at a slight advantage during this phase and the proportion of women in the Tier grows slowly (, B and ). While women’s increasing representation initially stalls during the period of disruption due to restructuring, by the end of the period, males are being replaced with females in the cohort (D). However, stratification within occupation types acts against pay gap gains even as women’s representation improves (). Women’s increasing access early in the data record may be a response to growing numbers of individuals with appropriate training in the candidate pool (Pacheco et al. 2017) together with microsystem policy responses to government equity directives (Supplemental C). The change in pattern at the time of Tier 5 population decline requires a different process explanation. DOC’s strategic direction may provide an explanation.

Disruptions: strategy change within DOC

Starting in 2009, the premise of DOC’s strategy and policy frames begin to change as conservation is described as an economic investment rather than solely a social good cost (DOC 2009). New emphasis was placed on external partnerships and commercial activity (Supplementals B and 2) and both business plans and restructuring activities reflected this shift. By 2015, DOC’s strategic direction had explicitly changed, with external partnerships considered to be as important as conservation biodiversity work (DOC 2015a, 2015b; Supplemental 2). DOC anticipated its function would move away from delivery work towards supporting and advising external partners to deliver conservation work (DOC 2015a, 2016b). Recreation and tourism were emphasised over biodiversity investment (DOC 2015a, 2017a) and landscape-scale projects involving external partners were promoted as the new way of working on public conservation land, in urban areas, farmland, catchments and regions (DOC 2017a). An organisation-wide behavioural shift was considered necessary to support DOC’s transformation (DOC 2015a; State Services Commission, Treasury & DPMC 2014). Agility, adaptability, flexibility, and innovation became explicit core competencies for staff (DOC 2015a). Together, these changes in DOC strategy and business model that began in 2010 shifted its staffing requirements away from business groups and occupations dominated by males and toward groups and occupations with larger female representation (DOC 2018d, 2019b). Viewed through the SES lens, patterns of increasing female representation in Tiers 4 and 5 can be seen as at least in part a result of changing workforce requirements ().

DOC’s annual budget grew between 1999/2000 and 2001/2002, after which it experienced a long period of decline, in inflation-adjusted dollars, irrespective of government (Supplemental 4). The 2017 incoming government delivered substantial new funding (Budget 2017 $76 million over 4 years; Budget 2018 $181.62 million over 4 years; Budget 2019's International Visitor Levy contributing $40 million per annum), which enabled a significant capacity augmentation, called Grow-Conservation (DOC 2019a). With the incoming government in 2017 came a change in expectation and DOC returned to its central position for providing conservation services and leadership to protect and enhance natural and historic heritage while building partnerships and technological innovation remained important foci for increasing the scale of biodiversity management (DOC 2019a).

Responses to disruptions: culture change within DOC

An organisation’s culture and climate of staff engagement are critical to performance and retention (Deloitte 2016). Culture emerges through staff values, beliefs, norms, behaviour, and actions (Schneider and Barbera 2014), while the climate of engagement reflects commitment to both the organisation and work (Deloitte 2016). Understanding DOC’s culture and climate is important to identify barrier patterns to internal career progression and aspirations that might be experienced by DOC staff and in turn explain variation in participation rates.

In 2015, DOC’s Senior Leadership Team formed a programme to align culture with the new strategic trajectory using the strong leadership style based on the State Services Commission’s Leadership Success Profile (DOC 2016a, 2016b). Rule-driven social imperatives were introduced to train staff to have a common understanding of good team member and leader behaviour and accountabilities (DOC 2016a, 2016b; Supplementals C and 5). The plan was motivated by the scale of DOC’s transformation, purported to be the most significant change in DOC’s history (DOC 2016a, 2016b; State Services Commission, Treasury & DPMC 2014). Other reasons given for the proactive culture shift were a misalignment between some staff views and DOC’s strategic change (State Services Commission, Treasury & DPMC 2014). Anecdotally, some staff resisted competition between DOC’s new commercial and partnership arms and its biodiversity conservation work (Stowell 2013; Williams 2018a). Many people who work for DOC prioritise biodiversity protection over other DOC objectives (Kennedy 2003). The values and belief systems that people hold bind groups, result in collective behaviours and are typically stable and persistent (Manfredo et al. 2017). To support the culture shift, new recruits skilled in working with and influencing others (State Services Commission, Treasury & DPMC 2014) and young emerging leaders were sought (DOC 2013a).

It is not possible to quantify how the structural, strategic, and cultural changes were received by staff during the period of DOC’s transformation due to a hiatus in collecting staff engagement information during the active restructuring years (2011–2015) (D. Petersen, pers. comm., May 14, 2019). However, a mini engagement review undertaken by Public Service Association (PSA) delegates, examined staff opinions at several sites in March 2014 and found variable staff morale (State Services Commission, Treasury & DPMC 2014). That a degree of dissonance was present was illustrated as staff raised their objections with the media (Stowell 2013; Mitchell 2018, 2019; Williams 2018a, 2018b). A workplace engagement survey undertaken in October 2018 showed improved staff perceptions about their job and context since the last survey in 2014 (DOC 2018c).

Responses to disruptions: gender participation and equity initiatives within DOC

Processes that perpetuate gender participation barriers within workplaces can vary and are not well understood (Gallard Martínez et al. 2019; Tajmel 2019). For instance, the roles of males and females can be segregated in the labour force via structural mechanisms (Hoobler et al. 2018) that influence or limit the choices and opportunities available for certain individuals or groups (Fotaki 2011; Yang and Liu 2019). Structural barriers can be controlled at the microsystem level. For example, an absence of flexible work arrangements, limited attention to equity during recruitment and performance review, and a climate that alienates or isolates female employees all create barriers (Johns 2013; US Department of Labor 1995). Structural barriers can also be presented by the macrosystem level, such as an absence of parental leave laws (Johns 2013; US Department of Labor, 1995). However, individuals’ agency, their capacity to act autonomously, using their own freewill independent from structural or social constraints, must also be taken into consideration (Fotaki 2011; Yang and Liu 2019). Gender imbalance can occur within institutional hierarchies via structural mechanisms, leading to power imbalances (Hoobler et al. 2018). Such mechanisms can limit the options and choices for female agency (Hoobler et al. 2018). Organisational leadership decisions directly affect the extent to which individuals perceive their employer’s policies, procedures and events unfairly favour some groups above others (Hoobler et al. 2018).

Two equity pulses can be identified; in 2007 (DOC 2007a, 2007b) and 2018 (DOC 2017a, 2017b, 2018b; Supplemental 2). In 2007, the New Zealand government initiated a public service plan to address gender pay and employment equity issues (DOC 2007a) and in response DOC undertook an Equity Review to investigate rewards, participation and people’s experience of respect and fairness (DOC 2007b, 2008a). The Equity Review found that although there had been a positive change since 2003, a gender pay gap existed, there was an under-representation of females in management across the organisation and a higher rate of turnover for females than for males (DOC 2007b, 2008a, 2009). In 2007, equity, rewards and fairness were addressed explicitly in DOC’s corporate documents (DOC 2007b). DOC’s People Plan 2012, developed in 2008 and refreshed in 2009 (DOC 2009; Supplementals C and 2), set out leadership, organisational development, and cultural goals to help address gender equity (DOC 2008a). DOC’s other workforce systems, such as its remuneration system, learning and development framework and national succession planning, were also expected to help address equity issues and progress the response plan actions (DOC 2008a). In 2010, a further assessment of females’ experiences was undertaken to understand retention and progression issues. From this work, the People Plan 2012 was updated to the People Strategy 2020. The new strategy focussed on improving culture, leadership, capability, processes, and systems (DOC 2010a). The analysis presented here shows the effects of these initiatives differed between the Tiers.

In 2018, DOC committed to gender balance across its leadership roles (i.e. Tiers 2–4) by 2022, as part of the government’s public service milestones (DOC 2018a, 2018b). By 2019, a project was initiated to understand why females were under-represented in leadership (DOC 2019b). Work on recruitment systems began to support diversity initiatives broadly, and all managers and supervisors were required to undertake Unconscious Bias Training (DOC 2019a).

Viewed in isolation, changes in female representation may be interpreted as system responses to macrosystem disturbances, such as state-sector equity initiatives. However, dialogue within DOC's corporate documents suggests that government leadership was not the primary driver of the organisation's equity action. For example, in the 2017 Annual Report, DOC stated it refreshed its Diversity Plan in response to feedback from staff through the annual engagement survey (DOC 2017c). DOC’s Organisational Management Human Resource Projects 2000–2019 (Supplemental 5) show equity endeavours are intermittent, with a tendency to be incorporated into general systems and diluted with other goals. This pattern of variable progress, non-linearities and emergent behaviours at the microsystem scale are consistent with Dalziel’s (2018) argument, that equity initiatives need constant attention, otherwise progress falters.

In 2017 DOC deliberately steered away from nation-wide diversity strategies or targets, instead supporting regional diversity strategies to guide staff hiring (DOC 2017a, 2017b). It was believed that the implementation of regional diversity strategies would ensure regional populations in relation to age, gender and ethnicity would be mirrored within DOC’s staff demographic (DOC 2017a, 2017b). However, when equal employment opportunity targets became a focus of the new government in 2018, DOC changed direction on department-wide diversity strategies and introduced top-down national equity imperatives in addition to the bottom-up regional initiatives (DOC 2017b, 2018b, 2018c).

Organisational setting: staff opportunities within DOC

Forty-two Organisational Development and Human Resource initiatives tallied from the DOC corporate grey literature from 2000–2019 are synthesised in Supplemental 5, for the purposes of examining mechanisms that might affect gender participation in DOC. The initiatives are a combination of people strategy, leadership development, diversity, and inclusion, learning and development, reward and recognition, talent management and organisational change initiatives. Development of DOC's integrated employee system – Suite of People Systems (Supplemental 5) – began in 2018.

DOC offers explicit talent opportunities for leaders and youth (Supplemental 2). DOC’s Talent Management Framework, rolled out July 2019, facilitated leadership opportunities for managers as part of a wider State Services Commission goal to facilitate a mobile workforce of public service leaders to move between agencies (State Services Commission 2016).

Internally, the mechanisms for offering development opportunities to the staff workforce are provided by discussions between a staff member and their manager at the time of Annual Development Plans and Monthly Operating Reviews (DOC, n.d.[b], n.d.[d], n.d.[e]). Development Plans and Monthly Operating Reviews record staff aspirations, learning and development needs along with discussions on work objectives, standards and employee wellbeing (DOC, n.d.[b], n.d.[d], n.d.[e]).

Workforce systems within DOC

Workforce systems can perpetuate gendered labour segregation and create a gendered power imbalance within an institution’s hierarchy (Hoobler et al. 2018). Workforce systems in DOC were examined for insights into the organisation’s patterns of gender participation; particularly those that might be conditioning decisions by female staff. Specifically, workforce agreements, recruitment processes, job roles, remuneration bands and staff review mechanisms were evaluated (Supplemental 6).

No explicit gender structural barriers were identified within in any of the workforce systems reviewed here and several generations of equity-oriented initiatives have been applied. Nevertheless, gender imbalance is an emergent property of the DOC system, and specific patterns such as the glass ceiling and glass cliff are observed. Various attributes of the DOC microsystem must then be interacting in ways that produce this outcome. For example, a flexible work policy and procedure driven by the public sector at the macrosystem (external to DOC) level, were formalised only recently in 2019 (DOC 2021). The strategic focus on external leader recruitment (DOC 2015a) may have diminished internal progression advancement opportunities, for example, by creating the apparent glass ceiling between Tiers 4 and 3. Performance and development systems have been inconsistently applied (DOC 2015b). DOC’s performance review process was evaluated as being restrictive in its ability to acknowledge strong performance (State Services Commission, Treasury & DPMC 2014). Talent opportunities for staff were found to be nebulous, despite the raft of tools made available (Supplemental 5; State Services Commission 2016). The systems and managerial influence show that both individual and microsystem drivers shape the experiences of staff within a regime.