The gendered dimensions of informal institutions in the Australian construction industry

The construction industry is the most male-dominated in Australia, despite companies implementing formal policies and initiatives to address this. While previous research has examined the role of workplace culture as a barrier to women in the industry, our research investigates the role informal institutions play in obstructing gender equity in construction. We examine the gendered dimension of informal institutions (practices, narratives and norms) in two multinational Australian construction companies using feminist institutional (FI) theory and rapid ethnography. The findings show that informal gendered institutions obstruct women ’ s recruitment, retention and progression in construction. Recruitment into and within the industry is framed by male sponsorship, cultural fit and traditional edu-cation pipelines. Retention is compromised by poor parental leave practices and norms, which tolerate sexism and gender stereotypes, as well as unsustainable work practices. Progression is impaired by the need for strategic alliances with powerful men and a lack of transparency around promotion opportunities.


| INTRODUCTION
The construction industry is the most male-dominated sector in Australia: in 2016, women represented only 12 per cent of the workforce, a decrease from 17 per cent in 2006 (ABS, 2006(ABS, , 2016. Among professional and managerial roles, women represent 14 per cent of staff (ABS, 2012). Men dominate senior and project delivery roles, while women congregate in junior, part-time and support roles such as human resources and marketing. Early enthusiasm by women entering construction professions and their future careers in the sector diminishes with increased exposure to the workplace as they experience relative disadvantage and inequality in pay and promotional opportunities compared to their male counterparts (Dainty, Bagilhole, & Neale, 2000).
These experiences take their toll with women leaving the construction professions almost 39 per cent faster than their male colleagues (APESMA, 2007).
In the construction industry, despite years of formal policies and research to address gender inequities (Powell, Bagilhole, & Dainty, 2009;Watts, 2007aWatts, , 2007b, male dominance and a masculine workplace culture remains. Past research has been extensive and details how the construction sector has attempted to address gender equity by raising awareness of women's career opportunities in the construction industry, establishing support networks for women, promoting the business case for gender equity, using female role models and mentoring, flexible work hours, childcare and reporting on the impact of policies and practices (Greed, 2006). Despite these initiatives, barriers to gender equity in construction remain, including a lack of networking opportunities, challenges in achieving a worklife balance and gendered discourses (Barnard, Powell, Bagilhole, & Dainty, 2010;Sang & Powell, 2012). Further problems include bias and discrimination in construction company policies and procedures and the heavy and dirty nature of work which takes place during unsociable hours and in remote locations (Wilkinson, 2006). The continued dominance of a culture which emphasizes long hours and 'presenteeism'understood here as the expectation that one is present on the job, regardless of whether or not the employee has work to be completedis also an ongoing issue (Powell et al., 2009;Watts, 2007aWatts, , 2007b. Recent research by Galea, Powell, Loosemore, and Chappell (2015) has been instrumental in demonstrating the value of new institutional (NI) theory in critiquing the 'robustness' and 'revisability' of formal policies designed to address gender equity. Through a documentary analysis and drawing directly on the work of Vivien Lowndes (2003;Lowndes & Roberts, 2013), Galea et al. (2015) showed a mismatch between core company values, company policies to address gender inequalities and the implementation of these policies. Often, these policies position women as both the problem and solution, adopting a 'fix' women response, with no attention to how men and masculinity operate on-site. Galea et al. (2015) also found that such policies frequently lack consistency, coherence and coordination across different policy areas, resulting in confused messaging which enables the dominant masculine culture to remain intact. A major identified problem was that there are few, if any, rewards or sanctions for achieving (or failing to achieve) formal gender equity objectives. Furthermore, the adoption of a 'one size fits all' approach, to a problem that operates in different ways across an organization, is problematic. Galea et al. (2015) concluded that formal policies often fail to address the problems they were created to resolve, in part because competing formal rules and informal practices and narratives interfere with the effective implementation of formal policies, making it difficult for these policies to 'stick'.
Having demonstrated the value of NI theory as a conceptual lens to explore the interaction of formal and informal rules, this article elaborates on the specific function informal rules play in maintaining gender inequity in construction and examines how these informal rules interact with the formal. Specifically, it investigates the gendered nature of informal rules in construction and asks what role informal rules play in obstructing gender equity across three career stages: recruitment, retention and progression. Our contribution to knowledge is threefold. Methodologically, we go beyond many existing studies on gender in the construction industry to collect data about gender from both women and men, rather than just women. We also adopt a rapid ethnographic approach, recognizing that access to the often 'taken for granted' is hard to collect from interviews and survey data alone. Empirically, our study operationalizes the feminist institutional (FI) concepts of gender rules, rules about gender and the role of gendered actors (discussed further below) and, in doing so, tests how far these concepts 'travel' outside the political sphere where they have predominantly been used to date (e.g., Krook and Mackay, 2011;Lowndes, 2019), guiding the way for others to apply this theory to a much broader range of cases, including other industries. Theoretically, the FI lens allows us to go beyond most existing research on women in male-dominated fields, much of which pays attention to either formal rules, such as flexible work practices, parental leave and so forth, and/or the impact of organizational culture on women (e.g., Dainty & Lingard, 2006;French & Strachan, 2013;Lingard & Francis, 2004, 2007Watts, 2009;Wright, 2016). It rarely looks at the interaction between the formal and informal, which FI enables us to do here, thus unveiling the constellation of rules that shape behaviours within an organization. Furthermore, FI enables us to tease out the processes and practices that perpetuate and maintain the masculine culture, which has been well documented by others.

| Informal institutions and gender equity
Organizations are not gender-neutral (Connell, 2005). Feminist institutionalism, a paradigm of 'new institutionalism', argues that in every organization, both formal and informal institutions in the form of rules, practices and narratives are gendered (Chappell, 2006;Mackay, Kenny, & Chappell, 2011). Institutions are rules that embed stable reoccurring patterns of behaviour that operate with a twofold effect in our workplace and society: they prescribe and proscribe which behaviours are acceptable and which are not (March & Olsen, 1984, p. 747). As such, institutions are different to organizations, the latter being the 'arena of action' in which institutions operate (Lowndes, 2019). Institutions can be both 'formal' (e.g., codes of conduct, contracts, procedures, policies, laws) and 'informal' (e.g., norms, practices and narratives) (Lowndes & Roberts, 2013;Powell & DiMaggio, 1991).
Formal rules are written down and made public in published organizational policies and guidelines and usually include explicit mechanisms for third-party enforcement (Helmke & Levitsky, 2004;Lowndes and Wilson, 2003).
In contrast, informal rules are often undocumented, tacit, taken for granted and not written down, although they can sometimes be better known and understood than formal rules (Helmke & Levitsky, 2004;Lauth, 2000;Lowndes & Roberts, 2013). For instance, understanding the appropriate attire to wear to a job interview; providing superiors with offices in open-plan workspaces; and for instance, in the Australian construction industry, working Saturdays regardless of level, duties or agreed work hours. Informal institutions perform a variety of positive and negative functions, substituting for formal rules that do not work or being used by vested interest groups to undermine formal rules that do (Helmke & Levitsky, 2004). Both formal and informal rules are an inevitable part of organizational life and are constantly evolving, being designed and constructed by actors over time through a process of negotiation, conflict and contestation. Together formal and informal rules produce 'stable, valued, reoccurring patterns of behaviour' (Huntington, 1968, p. 12) that determine 'how things are done around here' in organizations (Lowndes, 2014, p. 688). As Lowndes (2019) states, the real prize of feminist institutionalism is to better understand how distinctive modes of constraint work together in practice and to establish what this means for processes of institutional design and change.
(pp. 9-10) Numerous studies have observed the gendered impact of both formal and informal institutions in determining the career experiences and fortunes of women in many different types of organizational settings. For example, O'Neil and Hopkins (2015) critique the gendered impact of organizational systems on women's career progression arguing that 'fixing women' argumentssuch as those that propose the need to boost women's confidence in seeking promotiondo not work. They argue such an approach ignores the many social, cultural and organizational norms that elevate men as 'natural' leaders over women, which cumulatively act to ensure that women remain underrepresented in senior positions in many organizations. As Chappell (2015) states gender operates as a set of norms and practices on and under the surface of institutions, in ways that profoundly shape their design, paths and outcomes, including future policy direction and opportunities for reform. (p. 19) Gender as an analytical term, has conventionally been used as a 'category' to describe the perceived differences between men and women and by extension, the cultural symbolism assigned to masculinity and femininity (Beckwith, 2005). However, feminist institutionalists, in the tradition of West and Zimmerman (1987), also show that gender operates as a 'process' that is embodied in everyday social interactions to organize social activity and power (Beckwith, 2005). Gains and Lowndes (2014) identify three main ways in which formal and informal rules intersect with gender in organizations and therefore impact gender equity: • Rules about gender allocate particular roles, actions, traits and benefits to men and women. For instance, maternity leave and quotas on women are formal rules about gender while separating out different roles as 'men's work' or 'women's work' • Rules with gendered effects are workplace rules that on the surface may not be gendered but when they intersect with rules from other social domains have unintended gendered consequences. For example, meetings held before work may seem to impact everyone equally yet for those with caregiving responsibilities (predominantly women) it may result in exclusion • Gendered actors working with the rules embody and perform gender and act as 'rule-makers, breakers or shapers' (Chappell & Waylen, 2013, p. 606). This would include having only heterosexual, white men holding all the senior positions, and therefore all the overt power, in the company Given this, feminist institutionalism provides a valuable conceptual lens to understand the career experiences of women in the construction industry, as well as how the over-representation of men is maintained and perpetuated. This is particularly the case given that research about the underrepresentation of women within and outside construction remains conceptually narrow. More recently, Bresnen (2017, p. 24) noted that despite the proliferation of work within construction management that draws upon management and organizational theory, one omission that stands out from the body of published work is 'the under-use of dominant perspectives in business and management research (such as institutional theory) to frame construction management and organizational issues' .

| METHODOLOGY
Building on work in feminist institutionalism and addressing the lack of application of this theory in the field of construction management, this research investigates the gendered nature of informal rules in construction, including how they obstruct gender equity in construction. In doing so, this article makes an important contribution to the feminist institutionalist literature, which until now has been limited in its focus on the political arena (Krook and Mackay, 2011;Lowndes, 2019) and in studies of the 'global north'. While this article does not broaden the geographic scope of this conceptual approach, it does test the applicability of key concepts within an entirely different industry sector, and demonstrates the broader value of feminist institutionalism for interrogating and exposing gendered practices and outcomes.
Studying informal rules and their gendered nature poses specific methodological challenges. In contrast to formal policies, informal rules are generally undocumented and enforced through 'subtle, hidden, and even illegal channels' (Chappell, 2006;Helmke & Levitsky, 2004). People may not recognize the existence of informal gendered rules because their taken-for-granted nature can also render them invisible for those who live and perpetuate them (Acker, 2006). To address this, we adopted a rapid ethnographic approach .
While rapid ethnography has been criticized by classic ethnographers as a second-rate approach to ethnography, it is a practical solution to the real challenges of fieldwork (see, e.g., Galea et al., 2017;Isaacs, 2013;Pink, Tutt, Dainty, & Gibb, 2010).
Our approach involved working in teams to undertake short but intensive case studies of two multinational construction companies using multiple and iterative methods. Company A was a multinational contractor and Company B was a publicly listed multinational developer and contractor. Both operated in the global commercial, residential, engineering and infrastructure markets and the structure of both companies was typical and representative of large global construction companies (managed by a board, CEO, regional managing directors, directors and project managers). The classic ethnographic tools of open-ended interviews and lengthy observations were replaced with condensed equivalents, which focused on specific issues of interest (Baines & Cunningham, 2013). Observations allowed us to witness naturally occurring practices rather than relying only on participant accounts via interviews. In total, our approach included participant observation of 14 company events (including new employee inductions, graduate recruitment assessment, leadership training, construction skills training, management 'roadshows' and gender equity-specific events); on-site observations and shadowing of 44 male and female construction professionals and 61 interviews across six major construction sites (Table 1). A wide range of roles and occupations were shadowed and interviewed including senior managers (such as operations managers and general managers), project directors (managers of individual construction projects), commercial managers, programmers, site engineers, foremen, project engineers, design managers and safety managers. The observations of company events focused on several factors including room layout, presenters, timing of events, who attended, who had a 'voice', tone of conversation . The observations on-site focused on work practices (e.g., what time people arrived and left the site), roles on-site (e.g., who does what roles, and whether roles were associated with particular work practices such as total availability or leadership).
One of the unique aspects of our use of rapid ethnography was that in all observations, we adopted a twinning strategy: two researchers were involved to represent the views of both 'insider' (extensive experience in the construction sector) and 'outsider' (academics from a sociology or political science backgrounds) and, where possible, a male and female. This approach ensured we had different perspectives on the data, but also provided the different researchers with 'access' to different types of data (Baines & Cunningham, 2013;Bjarnegård, 2013;Hayfield & Huxley, 2015).
Researchers made detailed field notes and collected artefacts (photographs of room layouts, corporate messaging posted on walls or issued to employees) to help identify informal gender rules. After each observation the researchers debriefed and reflected, recorded and transcribed their conversation, which we also included as data. Finally, in keeping with the traditions of reflexivity in ethnographic research (Pink & Morgan, 2013;Pink et al., 2010) researchers also kept analytical memos on how the participant observation impacted them as a researcher. Each of these steps helped the researchers identify patterns with and across sites to determine the nature of informal institutions, and the interaction between formal and informal rules in operation.
To complement the observations, participants were invited to take part in a semi-structured interview that explored how formal and informal rules shaped career pathways. Interviews were undertaken in person or by telephone and typically lasted 60-90 minutes. They were recorded with participants' permission and then transcribed. To address concerns around confidentiality, all participant interview and observation notes were anonymized.

T A B L E 1 Summary of observations and interviews
The data were analysed through a feminist lens, attentive to the gender power relations amongst both participants and researchers (Connell & Messerschmidt, 2005;Gains & Lowndes, 2014). Analysis began in the field through the construction of researchers' field notes and research pair debriefings (Ybema, Yanow, Wels, & Kamsteeg, 2009). Debriefing served as an opportunity to compare experiences and early interpretations of the data, as well as identifying points of commonality and difference in the practices, narratives, themes and views expressed by participants. Next, the fieldwork notes, transcribed interviews and debriefs were coded thematically using NVivo (a qualitative software enabling the organization of content-rich data text) (Richards, 2000).
The data were coded into themes including career pathways (including recruitment practices, retention, progression, etc.), work practices (including total availability, work hours, workload, geographical mobility, etc.), gendered behaviours (including aggression, authoritative, competitive, strength, etc.), institutions (including informal rules and practices such as the norm of working long hours and formal rules such as the employment contract, as well as narratives that describe the rules such as behaviours aligned to valued masculine norms and how sanctioning of the rules occurs) and gendered institutions (including gendered actors working with the rules etc.).
Thematic analysis involved: 'immersion' in the data (repeatedly reading the field notes, interview and debriefing transcripts to obtain a high level of familiarity with the data); categorization/coding (organizing and generating an initial list of codes from the theory); searching for themes (examining how codes combine to form overreaching themes in relation to the research questions); refining themes (continuing to search for data that supported or refuted proposed themes and connections between overlapping themes). The coding process was carried out by a team of researchers who reconciled the initial and corresponding thematic codes across each stage of analysis (for a detailed discussion of the methodology, see Galea et al., 2017).

| FINDINGS
Our findings are focused on the gendered informal institutional barriers that affect policy relating to women's recruitment, retention and progression, which the following discussion addresses in turn. We use examples from interview transcripts and fieldwork notes (including researcher debriefs) to illustrate our findings.

| Recruitment
There were two gendered elements of recruitment. The first was recruitment into a company. Our data indicated that women and men were recruited through different processes and channels, with women more likely to be recruited through formal processes (i.e., formal advertisements, applications and interviews), while men were more likely to be recruited through informal networks including their family and family friends, schooling, sporting and industry connections. For example, one interviewee, Charlie (male, construction manager), said that he had never applied for a job other than the one he got the week after he left school; while Christopher (male, project engineer) remarked that, 'Dad was friends with a guy and he said, "I'll sort out an interview" …' The women we spoke to, on the other hand, had all been through formal recruitment practices. Bridget (female, safety manager) said that Company B was advertising, and she got the job; while Gail (female, commercial manager) said, 'I did the application on-line and I had one interview with the talent recruiter or the HR guy, and a follow-up conversation with him.' A strong narrative emerged in the data around the notion of 'cultural fit', based on how well people were perceived to align with established company and industry cultural values. These unquestioned values were largely shaped around men and masculine behaviours including reliability, authority, strength and technical competency, resulting in the recruitment of people who looked and acted like the incumbents. The second element of recruitment related to recruitment between construction projects. This was a highly informal practice, which appeared to determine the relative roles and opportunities afforded to men and women.
Throughout the research, we were unable to identify a formal procedure for internal project recruitment. Operations managers, most of whom were men, played a central role in appointing employees to projectsan example of gender actors working with informal rules. In practice, however, project teams were often selected by project directors (also mostly men) through a practice of 'picking your team' and taking 'your people with you' . Selection was based on previous working relationships, industry reputation/experience and an informal referral system. We found that men had greater awareness of this practice, which could be seen as a form of social capital. Men proactively negotiated their social networks and as a result were granted the best project opportunities. Few women were afforded the same access to these informal male networks and sponsorship relationships. For example, Monica (female, construction student) described how she found her site manager to be difficult. She went on to say, 'He'll always ask Tim for stuff but never me or anything like that and he was the one who was pushing Tim to get promoted.' Women in mid-to-senior roles recognized the importance of forming strategic alliances with others, yet few had access to these alliances or sponsors. One interviewee had recently experienced what she called 'true sponsorship' for the first time in her 25-year career and said that, '[it] would have made some other decades be a little bit easier to be honest' (Interview: Barbara, female, executive general manager). As a result, women were much less likely to be employed in high-profile roles on-site. This practice appeared to have a long-term impact on women's career progression, since having a high-profile role was critical to promotion.
There was also an expectation in both companies that employees would be able to move on short notice to nominated projects with little consideration of partner, family or care responsibilities. Men were typically able to relocate with their partners and families due to high salaries but for women, geographic mobility was often isolating, as their partners were less likely to relocate with them. Dorothy (female, site engineer) remarked that she had been given three weeks' notice to relocate for a new job and that she considered this 'good notice', since some people got moved within a week. This was not atypical, as noted in our fieldwork notes:

| Retention
We found normal work practices meant relentless high workloads, expectations of presenteeism, total availability and geographical mobility and very long work hours (escalating to 80+ hours a week towards the end of a project).
These norms, which persisted for both men and women, significantly impact the retention of women. On the other hand, while men rarely questioned whether they could pursue both a career and family, their work often resulted in long hours away from the home, which combined with job stress, often created, or added to, relationship difficulties, including marriage breakdown. A number of younger men reported that once they began a family they would consider a move out of contracting. For the few men who tried to fit greater care responsibilities around rigid work practicesby choice or due to a separation, divorce or the illness of their partnerthis was negotiated through informal short-term support from their managers. Finally, we found that the construction sites we studied were overwhelmingly a masculine space. Men dominated all senior management positions and traditionally masculine norms and practices, such as aggressive and combative exchanges, sexism, swearing and long work hours, were ever present. In our observations, women had to tolerate blatant sexism. We recorded, via photographs and observations, many examples of sexist graffiti and pornography as well as the use of sexist language. For instance, women were commonly called 'sweetheart' , 'babe' and 'girl' by managers, peers and subcontractors. Or their presence was ignored completely, as they were referred to as 'guys' , 'lads' and 'mate' in group emails and discussions. Women also reported to the research team cases of sexual harassment and sexual discrimination and sexist practices by managers and other workers.
These included being filmed in the shower at work, forced to change phone numbers to avoid harassment, having their breasts commented on and being delegated administrative tasks not in line with their role. We

| Progression
Our data revealed a number of disparities in how career progression occurred for men and women. Firstly, we found women's capabilities were often singled out, questioned and discussed in relation to progression. As a result, and consistent with previous research, women routinely had to prove their ability, needing to better (not equal) men. The following quote describes a manager questioning the capabilities of women in general in relation to recruitment: The whole gender thing gives me the shits as well sometimes because … above me, they're like pushing me to, 'Oh yeah, more women, more women.' And I go, 'Well hang on. If I need to employ people then I will advertise for the role. I will interview for the role and I'll pick the best candidate.' Just because she's female … If she's crap, then I'm not gonna put her on because this guy might be better. But, if she's really good and better than him, then I'm gonna give her the job. (Interview: Raymond, male, senior manager) By comparison, men's capabilities were usually taken for granted and assumed to fit. As a result, we found that men were afforded more opportunities to demonstrate their capability that inevitably enhanced their career progression.
Similar to other research, we also found that women faced a double bind. Women who asserted themselves as managers, like men, to their line manager or peers, were routinely sanctioned or pulled into line 'for getting ahead of themselves' . Our findings also pointed to an informal rule that promotion into key senior positions required employees to have followed a specific career path. This career path involved project management and a proven ability to deliver construction projects successfully. As Angus (male, project director) explained, there's no substitute for 'project success' and a 'proven track record'. However, When you look at the trajectory of how people are promoted … That pond tends to be all of the same person with the same sort of history. (Interview: Paris, male, senior manager) While this may not seem gendered, our data indicated that men and women had different access to opportunities and resources to pursue such a career path and therefore progress professionally. Men were given greater opportunities to build and demonstrate their project delivery capabilities and to 'shine' in front of company leaders.
In contrast, women reported to us that they were often encouraged away from project delivery and into femaledominated career pathssuch as commercial and designthat reduced progression opportunities. This meant that many female participants felt undervalued, operating at diminished capacity and worn out from constantly having to prove their worth. In the following example a woman described how a male colleague was prioritized for a management course, despite having no supervisory responsibilities: Recently we had an odd situation where I had just been given [Name, Female]  We also found a lack of transparency around the process of promotion. Participants who were promoted said they: 'were in the right place at the right time' , were 'lucky' , were 'pulled through' by a manager or 'picked' for a team, The lack of transparency around promotion practices enhanced the need for employees to form strategic alliances and sponsorship relationships. Women who were often denied sponsorship relationships felt frustrated and bemused as to what was required to get ahead. They noted that having many different line managers over a relatively short space of time was a hindrance to career progression as they had to demonstrate and prove their capabilities again and again. For many women, communicating career goals and aspirations, having these taken seriously and supported or actioned by management, was problematic. Women with different levels of experience and seniority also highlighted a lack of mentoring, sponsorship and career support resulted in feelings of isolation and exclusion.

| DISCUSSION
Assessing the influence of informal institutions that operate in construction companies and their impact on women's recruitment, retention and progression, we found strong alignment with Gains and Lowndes' model of institutional gender rules, demonstrating that this model does indeed travel across different sectors. Our data revealed that informal rules, like formal rules, can be about genderthat is, aimed directly at specifying a logic of appropriateness for men and women or masculine or feminine forms of behaviour; informal rules can also have gendered effectsseemingly neutral on the surface, informal rules can create or reinforce gendered behaviours that advantage one sex over the other, or masculine or feminine norms; gendered actors work with rules in ways which can create or reinforce gender patternsfor instance, men dominate senior roles and women subordinate roles, leading to a lack of recognition of gendered disadvantage which produces or reinforce gendered outcomes (Gains & Lowndes, 2014).

| Rules about gender
Our findings revealed two main informal rules where gender operated as a process in construction to maintain men's overrepresentation. Firstly, paid maternity leave may now be called parental leave but it continues to be modelled around informal rules that maintain traditional gender roles that reinforce women's role as carer and men's role as the breadwinner. As Banaszak and Weldon (2011, p. 268) observe, informal institutions in many organizations tend to relegate women to the homemaker role, enforce normative heterosexuality, and/or privilege men in the family and leadership positions. Watts (2009), in her interviews with women construction managers, similarly found that senior roles especially, were not seen by women as compatible with family. These roles are further reinforced through the lack of ownership of parental leave practices by senior leaders and flexible work options open to parents on their return to work. Paid parental leave was also found to carry a stigma and was seen as a cost to projects rather than a cost borne by the company as a whole. Senior managers and project directors had little regard for the cost of parental leave to women's pay equity and career progression. 'Lucky' women who found their way back to construction, often through informal male networks, were routinely denied flexible work arrangements on their return or have their new flexible arrangements met with resistance. Our findings showed that, despite formal institutional rules around parental leave, the informal rules and practices impact in at least three ways: reinforce women's traditional role of carer; individualize the issue, such that women need to negotiate their own pathway through; and impair women's retention and progression.
The second rule about gender was the tolerance of sexism that reinforced construction as 'men's work' and a masculine space. A lack of response towards sexism in the form of graffiti, language and task allocation by management on construction sites routinely reminds women (and men)subtly and overtlythat women are different or 'other'. It strengthens gender expectations, symbols and values, which over time become subliminal and normalized undermining women's legitimacy by casting them as the 'other' (Hawkesworth, 2005).

| Rules with gendered effects
Some rules appear to be gender-neutral but on closer inspection regulate actors' behaviours to produce gendered consequences. Across recruitment, retention and progression, we found rules that have different implications for men and women. In the first instance, the practice of recruiting through informal networks and using informal referrals of candidates, while sometimes supported by formal company policies, was found to have a gendered effect of advantaging male applicants over female applicants. Recruitment of employees based on their 'cultural fit' in construction, obligated actors in our study to behave in a masculine way. In other words, 'cultural fit' was framed by a gendered logic of appropriateness (Chappell, 2006) that is easier for male candidates to fulfil than female candidates.
Efforts to 'fit in' with this gendered logic by women (and other minorities) risks comment, criticism and exclusion. In the absence of formal promotion policies, not measuring up to 'cultural fit' impacts women's advancement into management roles. Similarly, Watts (2009) found that promotion appeared to be based on who you knew, rather than what you knew. In contrast, men's 'fit' is often assumed and unchallenged and gendered rules about women's capability as construction professionals and leadership skills remain a contested ground and see them guided into 'feminized' roles with less progression capital.
Furthermore, in the absence of effective internal recruitment policies or promotion policies, the informal practice of 'picking your team' and 'taking your team with you' acts to strengthen interpersonal links and male sponsorship.
Informal recruitment practices occurred despite the fact that both companies we researched had large human resource and talent acquisition teams and appears contrary to evidence in the UK that informal recruitment in construction is weakening (Clarke et al., 2015). Given the male dominance in construction, this practice produced a gendered effect that privileges men since they disproportionally hold positions of power. Women are effectively caught in a dilemma: women need men in their construction networks, but men do not need women.
The second notable rule with a gendered effect is the enforcement of inflexible and rigid work practices.
Presenteeism, total availability and the enforcement of long hours as a rule limits acceptable forms of flexible work available to women and men with care responsibilities. Wedged between rigid work practices and often inflexible care arrangements this rule is a major barrier to women's retention. Women's career advancement is further hampered when this practice is coupled with rigid views of the 'correct' career pathway into senior positions. This privileges those who have held key project delivery roles such as project managers which are less available to women compared to men. However, our findings also show that men are not left unscathed by the practice of inflexible rigid work practices. Their wellbeing and personal relationships suffer because of adherence to this informal rule. Lowndes (2014, p. 687) notes that rules distribute power by assigning particular roles including the opportunity to make, break and shape informal rules to particular individuals. As she has argued most recently: '[a]ctors reproduce rules as they use them, thus generating the regularities and stability associated with institutions' (Lowndes, 2019, p. 15). We observed this phenomenon across recruitment, retention and progression, where we saw evidence that senior managers and line managers (i.e., those who hold power and who are invariably men) can influence the practices that are enforced and the practices that are ignored. For example, in recruitment, the practice of male project managers 'picking their team', disadvantages women's recruitment onto different construction projects. This has some parallels with research by Wright (2016) who found that women in construction trades were restricted in their opportunities to consolidate workplace or on the job experience. This was reinforced in our research by the idea of cultural fit, which closely matched a masculine work style, and questioned women's competency.

| Gendered actors working with the rules
In another example we found male project managers were often unwilling to implement and support flexibility initiatives and different flexible work practices such as part-time and shared roles. This directly impaired women's retention in construction particularly in and around parental leave. Also, in relation to retention, project managers' tolerance of sexist practices, language and behaviour on-site, some of which is illegal, keeps construction a masculine space and women's legitimacy and career satisfaction are eroded. In terms of progression, informal male sponsorship by senior managers acts to maintain men's position in power.
Applying Gains and Lowndes' (2014) feminist institutionalist framework has provided the conceptual tools to investigate the operation of informal rulesand their interaction with formal rulesin the Australian construction industry. Instead of drawing on a vague notion of 'culture' or on formal policies, it has enabled new insights into what's happening 'under the surface' in these workplaces focusing on gender logics of appropriateness and repeated patterns of behaviour that are enforced through rewards and sanctions . Identifying the unwritten codes, practices and norms that are being replicated across construction sites contributes greatly to our understanding of why 'positive' formal gender policies such as improved paid parental leave conditions or flexible work policies often do not 'stick'; it shows that despite their codification and compliance procedures, formal rules can easily be undermined or over-written by existing entrenched informal rules. By paying attention to the gendered aspect of these rules in terms of their inherently gendered nature, gender effects and gendered actors' engagement with these rules, we were able to identify the multiple layers through which gender operates to maintain male privilege and women's disadvantaged position in the Australian construction industry.

| CONCLUSION
This article extended Galea et al.'s (2015) research into formal gendered institutions using feminist institutionalist theory. The article introduced a new conceptual lens to the construction literature to explore the three ways informal institutions intersect with gender. Using a rapid ethnographic approach, we have revealed new conceptual and practical insights into the different ways informal institutions act to undermine the recruitment, retention and progression of women in the construction industry. Significantly, by operationalizing Gains and Lowndes' (2014) model of gendered institutions, we have unpacked the interaction between formal and informal rules which influence gender outcomes that are often masked in research that examines the gendered organization of workplaces through the vague concept of 'culture'. This approach goes beyond existing studies of women in the construction sector (and indeed other male-dominated industries) by unveiling the constellation of rules that shape behaviour within an organization. It reveals how the rules in use intersect with gender to maintain the gender status quo. As a result, we have demonstrated that both formal and informal workplace rules, norms and practices are gendered and are enforced by gendered actors.
In summary, we found women were more likely than men to be recruited into companies through formal mechanisms, while recruitment onto projects (within companies) routinely operated through informal practices of male sponsorship and 'picking your team'. We found that informal rules around work practices (i.e., long hours, presenteeism), while seemingly gender-neutral, have a disproportionate gendered effect. Further, such practices undermine employee wellbeing for both women and men, leaving us to question why these informal rules remain intact. We also found that career progression is highly dependent on past success, yet men and women have unequal access to opportunities to demonstrate this, with men given greater opportunity to shine in front of leaders, thus reducing progression opportunities for women.
Our effort to unearth the informal rules in this sector required the deployment of an innovative methodology. It is not easy to study invisible informal institutions which participants may not acknowledge exist. We also acknowledge the limitations of this research which are its Australian context and the case study approach which limits the generalizability of the findings. Nevertheless, our approach produced depth of insight and the rapid ethnographic methodology was effective in revealing the complex and hidden informal gendered institutions which undermine the intent of formal gender equity and diversity policies.
Looking more broadly, this article has contributed to the feminist institutionalist literature by demonstrating that key concepts developed and applied in political settings can also travel to other sectors, such as the Australian construction industry. Importantly, this opens the possibility of more widespread application of these concepts beyond the political sphere. However, as a study of gender rules in the context of a high-income developed country, the study has not tested the value of these concepts across different geographic settings. A comparative study of construction across cases such as Australia and Indonesia, for example, would be an