Sage Journals: Discover world-class research

Abstract

A key feature of strategic action field (SAF) theory is the importance attached to collective identity, group membership, and social skills in understanding social orders. Gender is part of a SAF’s social order. This study investigates the relationship between credit union SAF choice and the percentage of women on boards. The setting is community credit unions in Northern Ireland where three distinct SAFs are explicitly identifiable, where board gender diversity is not subject to regulation, and where prior research identified financial management benefits from women on boards. The findings are that common understandings of the social order in community credit union SAFs are an important determinant of women on boards in credit unions. Credit unions belonging to the Orange Order SAF, a conservative religious organization, reveal a preference for a lower percentage of women on boards relative to credit unions not belonging to a SAF that emphasizes religious identity.

Keywords

cooperatives credit unions identity internal governance religion fields women on boards

A review of the non-profit literature identifies that irrespective of country, the percentage of women on non-profit boards is typically lower than national population proportions (Fredette & Sessler Bernstein, 2019; Y. J. Lee, 2014; Moore & Whitt, 2000), despite women being over-represented in chief executive officer positions (Faulk et al., 2021) and the workforce (Y. J. Lee, 2014). Few non-profit studies examine why under-representation exists and why it persists in some organizations given the societal shift toward gender equality in most countries (Abzug & Galaskiewicz, 2001; Themudo, 2009). Reasons noted include selection mechanisms (Moore & Whitt, 2000) and implicit discrimination resulting from social orders including engrained biases and prejudices (Y. K. Lee, 2023). Internal governance impacts economic performance (Forker et al., 2014) and is considered important for effecting change (Vehka & Vesa, 2023) yet no study to date has examined the influence of internal governance on the percentage of women on non-profit boards. In addition, the non-profit literature encourages the use of diverse theories from a variety of disciplines as they provide unique perspectives on specific dimensions of non-profit governance (Renz et al., 2023). In response, this study applies the sociological theory of strategic action fields (SAF) (Fligstein & McAdam, 2011) to investigate whether the percentage of women on credit union boards is influenced by SAF collective identity.

Fundamental to SAF theory is the view that affiliation with groups is an existential human motivation. SAFs rely on social skills to achieve collaborative common understandings about the purpose of the SAF and strategic actions such as power and resource allocation that contribute to the “emergence, maintenance and transformation of social order” within the SAF (Fligstein & McAdam, 2012, pp. 49–50). Gender is part of that social order. These common understandings become the SAF’s “rules of the game” and all actors/collective groups operate within these informal rules as this results in a stable SAF, which is beneficial for everyone (Fligstein & McAdam, 2011).

The research setting is community credit unions in Northern Ireland where religion and national identities dominate community life (McNicholl et al., 2019) and form a basis for SAF collective identity. In this jurisdiction, credit union directors serve on a voluntary basis (Mangan & Ward, 2024), choice of women on boards is not subject to regulation (GOV.UK, 2021), and prior research identified financial management benefits from having women on boards in terms of loan book quality and returns from assets (Ward & Forker, 2017). However, wide variation exists in the reported percentage of women on boards—some have no women, and some have no men (Ward et al., 2021). This is puzzling for two reasons, first given the importance of financial management quality for credit union sustainability and second, credit unions are non-profit cooperatives, an underpinning objective of which is equality.

SAFs are recognizable as they create internal governance units to oversee compliance with the SAF’s rules of the game (Moulton & Sandfort, 2017). They also perform an important advocacy role for their affiliated members (Vehka & Vesa, 2023) and have a stabilizing effect on SAF power structures and member behavior (Fligstein & McAdam, 2011). In Northern Ireland, three SAFs have the potential to influence the percentage of women on credit union boards. Two have created credit union focused internal governance units. These operate as trade associations: the Irish League of Credit Unions (ILCU) and the Ulster Federation of Credit Unions (UFCU). Both strategically promote credit union business models based on financial independence, competitive markets, and international cooperative principles, including “voluntary and open membership” which is defined as being “open to all persons able to use their services and willing to accept the responsibilities of membership, without gender, social, racial, political or religious discrimination” (International Cooperative Alliance, 2025). The third SAF, the Orange Order (OO), is a large, powerful, religious institution that focuses on promoting conservative Protestant religious beliefs. Patriarchal social orders are particularly prevalent in religious organizations (Bielefeld & Cleveland, 2013; Camarena et al., 2021; Dildar, 2015). Credit union connection to the OO is captured strategically by choice of common bond. A unique feature of credit unions is the legal requirement that membership is restricted to those within the chosen common bond. Common bonds ensure social links between members. This social connection reduces the information asymmetry prevalent in credit decisions, enabling loans to be awarded based on reputation (Forker et al., 2014).

Common bonds can be residential, associational or employment based.¹ In a credit union with a residential common bond, membership is open to any individual living within a pre-defined geographical area, regardless of religious or national identity. Credit unions that strategically select this common bond are better able to grow to achieve the economies of scale necessary for financial independence (McKillop et al., 2005). The ILCU and UFCU recommend that affiliated credit unions operate with a residential common bond, though the UFCU does not promote excessive large size due to the diluting effect it exerts on the common bond ethos (Forker et al., 2014). An associational common bond restricts membership to a particular association. Therefore, growth is limited and when linked to resource support, may impair financial independence. The OO provides sponsorship to credit unions that register with an associational common bond that restricts credit union membership to OO members and their families. Under SAF theory, where resource dependence within a field is high, a hierarchal relationship exists and that field will reflect the common understandings and purpose of the sponsoring SAF (Fligstein & McAdam, 2012 p.168). Therefore, when framed through SAF theory, choice of common bond is a strategic credit union action, that is either necessary for the credit union’s economic survival or is selected due to congruence with the purpose of the SAF, including identity and shared understandings of the purpose of and the social order within a SAF. In this paper it is assumed that the percentage of women on boards is a proxy for the gender social order within an organization, with low percentages reflecting a conservative social order and high percentages reflecting a moderate social order. Membership of a SAF is not exclusive, some credit unions are both affiliated to the UFCU and sponsored by the OO. In this paper, they are classified as boundary spanning and provide a unique opportunity to investigate the deeper significance on the percentage of women on credit union boards of the different purposes and common understandings of women’s position within the social order of these SAFs.

In the next section, SAF theory is applied in the context of community credit unions to develop hypotheses to investigate the “rules of the game” across credit unions fields in Northern Ireland. Whether SAFs with different cooperative and religious collective group identities explain variation in the percentage of women on credit union boards and if the percentages are stable over time. In the following section, econometric panel data analysis is applied to a unique hand-collected panel dataset covering 167 credit unions over the period 2003–2012. The results are then reported, findings discussed, and finally the conclusions are presented.

SAF Theory, Credit Union Fields, and Hypotheses

A widespread implication of an under-representation of women on non-profit boards that pursue social objectives, is damage to their legitimacy (Abzug & Galaskiewicz, 2001; Moore, 2000).² Women’s under-representation on boards is a complex issue as “. . . women do not behave as a homogeneous group towards a homogeneous non-profit sector” (Themudo, 2009, p. 679). For example, there is evidence of contextual variation across nations³ (Claus et al., 2013; Themudo, 2009), organization types (Stephenson et al., 2023; Themudo, 2009), community contexts⁴ (Abzug & Galaskiewicz, 2001; Stuart, 2007), and job roles⁵ (Council on Foundations, 2022).

Different disciplines have provided theories to explain under-representation. For example, in studies of organizational change, institutional theory explains how coercive isomorphism resulting from political influence and a need to maintain institutional legitimacy results in inertia and reduced organizational diversity (DiMaggio & Powell, 1983). Sociology and social movement theory subsequently focus on investigating the social dynamics between actors within social spaces, called fields, and identify that founding actors establish the social order, and design field rules to secure resources and retain power, which restricts diversity (Bourdieu, 1998; Bourdieu & Wacquant, 1992). Economic scholars, also note the importance of securing resources and identify the salient role of identity and explain that in some instances, identity, for example religious identity, may be more important than economic benefits (Akerlof & Kranton, 2000). Similarly, securing resources and legitimacy from communities served is also provided as an explanation for lack of board diversity in non-profits (Abzug & Galaskiewicz, 2001). Conservatism is particularly prevalent in religious organizations, which typically convey a preference for male hierarchical leadership (Bielefeld & Cleveland, 2013; Camarena et al., 2021). These theories provide explanations, from the viewpoint of an organization and its environment, for path-dependent board decision-making.

In SAF theory, to reflect social decision-making in fields, the construct of collective identity (we first) replaces the focus on individual identity (me first) to explain how and why fields come together to create social orders that are acceptable to the collective. This shift of emphasis identifies the important role of skilled social actors in fashioning social orders in SAFs. SAFs are defined as:

meso-level social orders where actors (who can be individual or collective groups) interact with knowledge of one another under a set of common understandings about the purposes of the field, the relationships in the field (including who has power and why), and the field’s rules. (Fligstein & McAdam, 2011, p. 3)

Therefore, SAFs are settings where groups of social actors compete and negotiate an understanding on SAF rules of the game, including gender social hierarchies (Fagan & Teasdale, 2021; Fligstein & McAdam, 2011, p. 5). SAFs establish internal governance units to maintain SAF stability and hence social orders such as gender are inherently resistant to change. In the context of community credit unions in Northern Ireland, field boundaries impact social order and are characterized by differences in collective identity reflecting religious and political affiliation that are administratively maintained by choice of common bond.

In this paper, SAF theory is the lens applied to different credit union fields to develop the hypotheses tested in this study.

SAFs and Credit Union Fields in Northern Ireland

As portrayed in Table 1, three distinct SAFs influencing choice of female representation on credit union boards are evident in Northern Ireland.

Table 1.

Characteristics of SAF Internal Governance Units.

Strategic action fields
Internal governance unit	ILCU	UFCU	OO
Identity
Irish	√	-	-
British	-	√	√
Cooperative	√	√	-
Religious	-	-	√

Note. ILCU = Irish League of Credit Unions; UFCU = Ulster Federation of Credit Unions; OO = Orange Order.

The overarching purpose of the ILCU and UFCU internal governance units is similar, to promote the development and smooth functioning of the credit union sector for the benefit of their members. To this end, they provide business support services, guidance on corporate governance, training and insurance services to their affiliated credit unions and representation to external actors on their behalf (Forker et al., 2014). Credit unions can choose to affiliate to either trade association, or not. Common understandings include financial independence and cooperative principles. Both the ILCU and the UFCU strategically recommend that affiliated credit unions operate with a residential common bond. However, they differ in respect of size, sophistication, and national identity. The ILCU is older, larger, and better resourced than the UFCU. It operates as an all-Ireland body. In contrast, the UFCU identifies as a Northern Ireland body with a distinct British identity.⁶ At the time of this study, 45.1% of the Northern Ireland population were Catholic and 48.4% were Protestant (Northern Ireland Statistics and Research Agency [NISRA], 2011). In Northern Ireland, Catholics typically have an Irish identity, and Protestants a British identity. As religion and national identity are connected, UFCU credit unions attract few Catholic members. The situation is different in the Republic of Ireland, where both Protestants and Catholics identify as Irish. Therefore, in Northern Ireland the ILCU does not promote its Irish identity, is considered cross-community and has Protestant membership. Perhaps due to national identity preferences or for economic reasons (affiliation fees are payable to trade associations), no credit union in Northern Ireland is a member of both the ILCU and the UFCU.

The third SAF, the OO, functions to promote conservative Protestantism. Relevant common OO understandings include Protestantism, British identity and patriarchy. The OO strategically facilitated the emergence of the credit union sector in Northern Ireland within some Protestant communities (Quinn, 1999). Promoting and sponsoring credit unions enables the OO to meet its social objectives to its members (Roberts, 1971), strengthening its support within Protestant communities. To ensure that resources benefit the OO community, it strategically requires that sponsored credit unions register a common bond of association to the OO. These credit unions are typically small, and some are financially dependent on the OO.

SAF Collective Identity and Women on Boards (Hypothesis 1)

According to Stone and Sandfort (2009), the policy environment is important for the strategic action of non-profit organizations. In Northern Ireland, female representation on boards is within the scope of the Sex Discrimination (NI) Order 1976 (as amended). However, the provisions of the order do not apply to volunteer positions such as credit union directorships (GOV.UK, 2021). Therefore, a salient influence on the percentage of women on credit union boards in Northern Ireland is the common understandings of the position of women within a SAF’s social order.

SAF theory recognizes that individual credit union boards may have different views on women on boards to that promulgated by the internal governance unit of the SAF to which they belong but are willing to compromise and abide by the collectively negotiated common understandings in the interest of SAF stability (Fligstein & McAdam, 2011). Hence, the collectively accepted social order of women, including representation on boards, is reflected in the internal governance unit rules (Fligstein & McAdam, 2012). In the context of the study site, the ILCU, a member of the World Council of Credit Unions (WOCCU) strategically promotes female representation on boards (Cak Niederkohr & Ikeda, 2005). Therefore, as women are over-represented in the membership, staff, and volunteer stakeholder groupings of ILCU credit unions (McKillop et al., 2003), adherence with this rule should result in a dominance of women on boards. In contrast, the UFCU recommends that at least two directors on credit union boards are women and the OO does not have any formalized policy on women on boards for the credit unions it supports. Therefore, consistent with SAF theory, which predicts that field actors will conform to collectively agreed rules and the explicit gender inclusive representative policy promulgated by the ILCU, it is hypothesized that:

Hypothesis 1 (H1): ILCU affiliated credit unions have higher percentages of women on their boards relative to other credit unions in Northern Ireland (non-ILCU).

Religious Identity and Women on Boards (Hypothesis 2)

Gender discrimination in the workplace is attributed to the role of identity (Akerlof & Kranton, 2000), with patriarchal social orders particularly prevalent in religious organizations (Camarena et al., 2021; Dildar, 2015). This conservative view of gender identity may result from preference, where religious women internalize religious teachings with regard to the female role as being best preserved in the roles of wife, mother and carer (Davis & Gao, 2020; Moore, 2000) or may be the result of barriers to progression including obstacles to female professional development as reflected in the “think-management, think-male” world view and assumptions leading to formal assignments of “ministry” for men and “family care” for women (Stephenson et al., 2023).

Northern Ireland has a Christian sectarian history. Although division based on religious domination has diminished, religious organizations remain segregated and influential within society (McNicholl et al., 2019). The most prominent is the OO, which plays an important role in the religious, political, and social life of many in the Protestant community (McAuley & Tonge, 2007). Women play an increasing role within the OO, for example, permission for women to march in OO parades was officially sanctioned in Ireland in 1990 (Urquhart, 2018). Nevertheless, women are typically not represented at decision-making levels (MacPherson, 2016) and cannot vote at Grand Lodge Level where policy is determined (Urquhart, 2018).

Though not examined through a SAF lens, Bielefeld and Cleveland (2013) report that connections with faith-based organizations help to maintain religious identity with Wittberg (2013) identifying that umbrella bodies influence and are influenced by their members’ faith-based organizations and communities.

Under SAF theory, strategic actions by skilled social actors includes the careful design of internal governance unit policies to restrict entry to outsiders. This ensures that the existing SAF social order does not change (Fligstein & McAdam, 2011). This is observed within the OO SAF as all its sponsored credit unions register as having an associational common bond, which restricts membership to members or friends of a specific OO lodge. As credit union regulation mandates that directors are recruited from the membership, with the appointments agreed at the annual general meeting, this restrictive common bond strategically ensures that members and directors of each member credit union share the common understandings of the OO. All other community credit unions in Northern Ireland register with a residential common bond. Therefore, consistent with SAF theory which predicts that incumbent actors design SAF rules to maintain power and given the strength of the association common bond structure in thwarting change and the OO’s historical conservative view of women, it is predicted that:

Hypothesis 2 (H2): OO sponsored credit unions have lower percentages of women on their boards relative to other credit unions in Northern Ireland (Non-OO).

Collective Groups, Common Bond Strategic Choice, and Women on Boards (Hypothesis 3a–3c)

SAFs represent a collection of actors, including individuals or collective groups that collaborate for mutual benefit. Actors are classified as incumbents and challengers. Incumbents have power over the social order, common understandings, and purposes of the SAF. Challengers continually try to change these attributes of the SAF (Fligstein & McAdam, 2011; Moulton & Sandfort, 2017). Actors are not restricted to one SAF. In the context of the study site, a group of 30 “boundary spanning” credit unions belong to both the UFCU and the OO SAFs. To investigate the SAF power dynamics affecting choice of percentage of women on boards in credit unions affiliated to the UFCU and/or the OO SAFs, four distinct collective credit union groups are identified as portrayed in Table 2: ILCU-only; UFCU-only; UFCU&OO; and OO-only, respectively.

Table 2.

Strategic Action Fields and Collective Groups.

Strategic action fields (n)	ILCU (101)	UFCU (51)		OO (45)

Note. ILCU = Irish League of Credit Unions; UFCU = Ulster Federation of Credit Unions; OO = Orange Order.

The strategic influence of resource dependency and religion on the percentage of women on boards between these collective groups is investigated using three hypotheses, denoted H3a, H3b, and H3c, derived from SAF theory.

H3a investigates whether the social order of the OO or the UFCU dominate choice of the percentage of women on boards within the 30 boundary spanning credit unions that belong to both (UFCU&OO). The OO SAF’s purpose is religious, the UFCU’s is credit union focused. When credit union boards belong to more than one SAF the extent of adherence to the common understandings and purposes of the different SAFs depends on the negotiation skills of the collective group’s actors, the extent of resources received from each SAF, and collective identity alignment (Fligstein & McAdam, 2011). Both the OO and the UFCU are British and provide resources to member credit unions. The OO donates financial resources, including premises, volunteers, and marketing to sponsored credit unions and hence contributes to their financial stability (Ward & McKillop, 2006). For an annual fee, the UFCU also provide resources that contribute to financial stability, for example, corporate governance guidance, training, support, legitimacy, representation, and deposit protection insurance (Forker et al., 2014). UFCU&OO credit unions strategically identify SAF preference by choice of religious associational common bond over a residential common bond. This restricts their membership and potential for growth, and are typically financially dependent on the OO due to lack of economies of scale. Therefore, in line with SAF theory, which predicts that resource dependency and identity alignment influences hierarchical relationships (Fligstein & McAdam, 2012), choice of associational common bond identifies that the OO is the dominant SAF and hence its social order will be evident in UFCU&OO credit unions. To test this prediction, it is hypothesized that:

Hypothesis 3a (H3a): UFCU&OO credit unions have a similar percentage of women on their boards relative to OO-only credit unions.

Hypotheses 3b and 3c investigate whether a proximate relationship to the boundary spanning group of credit unions, that have chosen alignment with the OO, influences the percentage of women on boards in UFCU-only credit unions. First, hypothesis 3b investigates whether there are differences in choice of the percentage of women on boards between UFCU-only and UFCU&OO credit unions. The UFCU adopts the Raiffeisen Cooperative Principles (International Cooperative Alliance, 2020), which are underpinned by the International Cooperative Alliance’s principles (International Cooperative Alliance, 2025). In addition, the UFCU promotes financial independence and competition, and in support of this, recommend credit unions register with a residential common bond. The 21 UFCU-only credit unions chose to register with a residential common bond. This strategic choice suggests financial independence and alignment with common cooperative understandings, including gender equality (Fligstein & McAdam, 2012). Therefore, it is hypothesized that:

Hypothesis 3b (H3b): UFCU-only credit unions have a higher percentage of women on their boards relative to UFCU&OO credit unions.

Second, hypothesis 3c investigates whether the percentage of women on boards in UFCU-only credit unions has been moderated through collaboration with UFCU-OO conservative credit unions. Strategically, the cooperative democratic principles adopted by the UFCU gives each credit union member a vote (International Cooperative Alliance, 2020) and hence a say in the design of the SAF’s rules of the game (Fligstein & McAdam, 2011). This means the 30 UFCU&OO credit unions are incumbents and hold the balance of power over SAF decision-making. Under SAF theory incumbent and challenger groups within SAFs collaborate and compromise on the SAF’s informal social rules of the game to ensure the smooth functioning of the SAF.

Whether the social gender order of the 21 UFCU-only credit unions has been moderated through collaboration with the 30 conservative UFCU&OO credit unions is investigated by comparing the percentage of women on boards of UFCU-only credit unions to the control group comprising ILCU credit unions. The social order within the ILCU SAF should reflect the impact of broader environment, social and cultural processes (Bjerregaard & Jonasson, 2014; Lounsbury & Ventresca, 2003) and higher-level global cooperative SAF influence (Moulton & Sandfort, 2017) on the percentage of women on boards, as this group of credit unions has no links to a religious SAF. Credit unions gain legitimacy and hence a social license to operate within their communities by being affiliated to a reputable global cooperative body. The ILCU is a member of WOCCU and the UFCU adopts the Raiffeisen Cooperative Principles (International Cooperative Alliance, 2020). The principles promulgated by both bodies are underpinned by the International Cooperative Alliance’s principles (International Cooperative Alliance, 2025) and both recommend credit unions register with a residential common bond. Moreover, the ILCU and the UFCU both operate with similar common understandings as to the purpose of their SAFs—to support, promote, and represent affiliated credit unions for the benefit of their membership. As UFCU-only credit unions: do not have any direct connection to the hierarchal OO, promote Raiffeisen principles, and are registered with a residential common bond, a cooperative identity with similar moderate views to ILCU credit unions on gender is predicted. It is hypothesized that:

Hypothesis 3c (H3c): UFCU-only credit unions have a similar percentage of women on their boards relative to ILCU credit unions.

Field Stability (Hypothesis 4)

Incumbent actors within SAFs have power over SAF decision-making and hence design structures that through inculcation, indoctrinate domination over actors either with or without their complicity (McNay, 1999). Challengers remain alert to opportunities to challenge field rules. Nevertheless, change typically occurs only in response to destabilizing shocks that threaten the sustainability of the SAF and the position of the incumbents (Fligstein & McAdam, 2011), or when there is coercion from hierarchal social networks (McNay, 1999), for example, regulators (Stone & Sandfort, 2009) or other SAFs (Fligstein & McAdam, 2011).

In the context of the study site, Northern Ireland over the period 2003–2012, no legislation on gender quotas was introduced. Therefore, in the absence of destabilizing shocks in respect of the percentage of women on boards and consistent with SAF theory suggesting SAF stability over time, it is predicted that the percentage of women on credit union boards will remain a product of the SAF rules of the game which are a product of prior experience. Thus:

Hypothesis 4 (H4): There is no change, over the study period, in the percentage of women on credit union boards for the different SAF collective groups.

Data, Variables, and Methodology

Data

This study exploits data from 1,618 hand-collected observations on annual returns submitted to the Registrar covering the full population of 167 community credit unions in NI over a 10-year period from 2003 to 2012.⁷

Variables

A summary of the coding for dependent and independent variables is provided in Table 3.

Table 3.

Summary of Variable Descriptions.

Variable name	Measurement, coding, and transformation^a
Dependent variable %WOB	Percentage of women on boards is the ratio of the number of WOB divided by the total number of directors on the board multiplied by 100 (Gregorič et al., 2017; Ward et al., 2021).
Independent variables
SAF rules (H1)
ILCU SAF	Coded one if affiliated to the ILCU, zero otherwise.
Non-ILCU	Coded one if not affiliated to the ILCU, zero otherwise.
Religious identity (H2)
OO SAF	Coded one if affiliated to the OO (including boundary spanners), zero otherwise.
Non-OO	Coded one if not affiliated to the OO, zero otherwise.
Collective groups, strategic common bond choice and percentage of women on boards (H3 to H4)
ILCU-only	Coded one if affiliated to the ILCU, zero otherwise.
UFCU-only	Coded one if affiliated to the UFCU only (excluding boundary spanners), zero otherwise.
UFCU&OO	Boundary spanning: coded one where credit unions have UFCU and OO membership, zero otherwise.
OO-only	Coded one if affiliated to the OO only (excluding boundary spanners), zero otherwise.
Control variables
Multiple deprivation measure (MDM)	Comprises a low-income factor, a low employment factor, an education factor, a housing factor, an access to services factor, a social environment factor, and a health factor (NISRA, 2010). The higher the MDM score the more deprived the location. This variable varies between credit unions but is constant over time.
Size	Continuous measure of the natural log of total credit union assets where the log transformation is used to reduce the impact of outliers.
Board size	Continuous measure of the number of board directors.
Year effects	Ten dummy variables separately coded ‘one’ for each of the ten years, zero otherwise.

Note. ILCU = Irish League of Credit Unions; UFCU = Ulster Federation of Credit Unions; OO = Orange Order.

The source of each data variable is detailed in the Appendix.

Control Variables

The control variables are locality deprivation, credit union size, board size, and a set of year-specific effects. The expected impact on the percentage of women on credit union boards is as follows:

Multiple Deprivation Measure

Informational benefits when awarding credit from having women on boards in deprived locations are anticipated as women are better able to build trust relationships (Beck et al., 2013) and to form strong network links (Brown et al., 2012). Consistent with this view, Ward et al. (2021) report lower loan loss provisions in credit unions with higher percentages of women on their boards in deprived locations. Therefore, a positive association is predicted.

Credit Union Size

Existing non-profit studies have identified that males are more attracted to roles in larger organizations due to the better working conditions and higher status associated with such roles (Ostrower & Stone, 2006). Therefore, a negative association between credit union size and the percentage of women on boards is predicted.

Board Size

The turnover of directors is predicted to be higher in larger boards; hence, the potential for replacement by a woman is greater when the general societal trend is increasing representation on boards (Gregorič et al., 2017). Therefore, a positive association is predicted between board size and the percentage of women on boards.

Year Effects

Institutional arrangements for gender diversity in the labor market in Northern Ireland have evolved from overtly supporting a patriarchal society in the 1960s to supporting gender diversity as evidenced by the changes to gender roles in the public sector (Rouse, 2018). Therefore, a positive association is predicted between time and the percentage of women on credit union boards.

Methodology

The dataset is an unbalanced panel comprising time-series and cross-sectional observations on credit unions. Therefore, panel data techniques are used to investigate the hypotheses of interest. Singleton observations are excluded thus ensuring each credit union unit has a minimum of two observations. Given the panel dataset contains several key time-invariant variables, such as trade association affiliation, OO sponsorship and multiple deprivation measure (MDM), and other variables that change slowly over time (e.g., board size), a random effects model rather than a fixed effects model is appropriate (Ward et al., 2021). The random effects approach accounts, in an optimal way, for both within and between variation which is key in SAF theory. The dependent variable, percentage of women on boards (%WOB), is expressed as a linear function of the covariates to investigate the influence of the different SAF rules, ILCU and non-ILCU, on the percentage of women on credit union boards (H1) as:

\begin{array}{l} % {WOB}_{i t} = γ_{0} + γ_{1} {ILCU}_{i} + α_{1} {Size}_{it} + α_{2} {MDM}_{i} + α_{3} {Boardsize}_{it} \\ + \sum_{(t = 2)}^{10} q_{t} + u_{i} + e_{it} \end{array}

(1)

This approach is further utilized to investigate the influence of religious identity, OO and non-OO, on the percentage of women on credit union boards (H2) as:

\begin{array}{l} % {WOB}_{i t} = γ_{0} + γ_{2} {OO}_{i} + α_{1} {Size}_{it} + α_{2} {MDM}_{i} + α_{3} {Boardsize}_{it} \\ + \sum_{(t = 2)}^{10} q_{t} + u_{i} + e_{it} \end{array}

(2)

where i = 1, . . . N, and t = 1, . . . T, q_t represents the time effects, $u_{i}$ represents the random effects for each of the i credit unions, e_it is an idiosyncratic term, and all other variables are as defined above. (Note: “non-ILCU” and “non-OO” comprise respectively the base categories in estimation.)

To investigate the behavior of groups of SAFs with potentially conflicting social orders in respect of the percentage of women on credit union boards (H3a–H3c) and field stability (H4), equation (1) is re-specified:

\begin{array}{l} % {WOB}_{i t} = β_{1} + β_{2} {UFCU - only}_{i} + β_{3} {UFCU & OO}_{i} + β_{4} {OO - only}_{i} \\ + α_{1} {Size}_{i t} + α_{2} {MDM}_{i} + α_{3} {Boardsize}_{it} + \sum_{(t = 2)}^{10} q_{t} + v_{i} + ξ_{it} \end{array}

(3)

where v_i now represents the random effects, and ξ_it is the idiosyncratic term. The dummy variable ILCU-only is now the base category in estimation.

The regression models (1)–(3) are estimated using the generalized least squares procedure with a robust correction used to compute the asymptotic standard errors. A key assumption for the use of a random effects model is the independence of the random effects from the included set of time-varying explanatory variables. This proposition is testable using the Mundlak test, which augments the regression models with the group means of the two time-varying covariates and tests for their joint statistical significance (Mundlak, 1978; Wooldridge, 2010). This test is implemented below for the above regression models.

The asymptotic t-tests for the differences in the estimated coefficients for the collective group variables in Model 3 indicate whether: religious conservatism prevails resulting in similarities in the percentage of women on boards within the OO SAF groups as predicted under H3a (UFCU&OO = OO-only); moderate cooperative identity in the absence of association common bond prevails over conservative identity from the 30 incumbent boundary spanning credit unions resulting in higher percentage of women on boards as predicted by H3b (UFCU-only > UFCU&OO); and being part of a SAF (UFCU) where credit unions with religious identity are incumbent, does not weaken women’s position within the social order in UFCU-only credit unions as the percentage of women on boards are similar to those of the control group, ILCU-only credit unions, as predicted by H3c (UFCU-only = ILCU-only). Finally, we determine whether women’s position within the SAFs’ social orders is stable over time, as predicted by H4. The relevant tests for this hypothesis are based on separate Wald tests for the joint significance of the interaction of the set of year-specific (or time) dummies with the four groups (ILCU-only, UFCU-only, UFCU&OO, OO-only) of credit unions.

Results

As a prelude to the empirical analysis, the univariate results are provided in Tables 4 and 5 with the multivariate regression results reported in Table 6.

Table 4.

Average Percentage of Women on Boards Across SAFs, Groups and Time.

Year	CUs	Obs.	Percentage of woman on boards
Year	CUs	Obs.	2003	2006	2009	2012	Whole period	Growth2003–2012
Collective groups	n	n	%	%	%	%	%	%
SAFs
ILCU	101	982	41.35	43.83	45.19	44.92	44.00	8.63
UFCU^a	51	486	40.01	40.69	40.83	42.21	41.06	5.50
OO^a	45	438	33.93	34.36	35.06	36.75	35.05	8.31
Analysis by hypotheses
SAF rules (H1)
Non-ILCU	66	636	38.09	38.25	39.30	40.39	39.09	6.04
ILCU SAF	101	982	41.35	43.83	45.19	44.92	44.00	8.63
Total	167	1,618	40.08	41.64	42.89	42.13	42.07	5.11
Religious identity (H2)
Non-OO	122	1,180	42.35	44.21	45.86	45.63	44.67	7.75
OO SAF	45	438	33.93	34.36	35.06	36.75	35.05	8.31
Total	167	1,618	40.08	41.64	42.89	42.13	42.07	5.11
Collective groups, strategic common bond choice and women on boards (H3–H4)
ILCU-only	101	982	41.35	43.83	45.19	44.92	44.00	8.63
UFCU-only	21	198	47.52	46.18	49.39	49.59	48.09	4.36
UFCU&OO	30	288	34.94	36.55	35.42	37.83	36.28	8.27
OO-only	15	150	32.05	30.42	34.32	34.59	32.68	7.93
Total	167	1,618	40.08	41.64	42.89	42.13	42.07	5.11

Note. ILCU = Irish League of Credit Unions; OO = Orange Order; SAF = strategic action fields; UFCU = Ulster Federation of Credit Unions.

The UFCU and OO SAFs include 30 credit unions that belong to both.

Table 5.

Descriptive Statistics and Spearman Correlations for Study Variables.

Continuous variables	N	M	SD	1	2	3	4
1. %WOB	1,618	42.09	21.31	-
2. MDM	1,618	26.95	16.80	.11**	-
3. Size	1,618	14.43	1.53	−.08**	.29**	-
4. Board size	1,618	11.20	2.47	−.25**	−.21**	−.26**	-

Note. %WOB is percentage of women on the board. MDM = multiple deprivation measure. Details of each variable are provided in Table 3.

p ≤ .01.

Table 6

Multivariate Analysis Results From Random Effects Generalized Least Square Regressions.

Dependent variable^†	Model 1	Model 2	Model 3
Dependent variable^†	%WOB	%WOB	%WOB
Observations (n)	1,618	1,618	1,618
Credit unions (n)	167	167	167
Constant	94.13*** (4.52)	90.63*** (5.42)	92.98*** (4.42)
Independent variables ^†
Credit union SAFs (H1 and H2)
ILCU versus non-ILCU^α	9.61(2.27)*	--	--
OO versus non-OO^ꞎ	--	−12.67*** (−3.70)	--
Collective groups (H3 and H4)
UFCU-only^∞	--	--	−1.48(−.25)
UFCU&OO^∞	--	--	−13.28** (−2.75)
OO-only^∞	--	--	−13.10** (−2.71)
Control variables ^†
MDM^‡	0.20* (2.10)	0.19* (2.03)	0.19* (2.03)
Size	−3.51* (−2.48)	−3.28** (−2.85)	−3.43* (−2.40)
Board size	−0.58(−1.78)	−0.58(−1.80)	−0.57(−1.75)
Year effects^ϟ	Yes (+)	Yes (+)	Yes (+)
R-squared/overall	12.53	14.43	14.64
Wald chi² (13)	37.77	46.08	46.43
	(.00)	(.00)	(.00)

Note. %WOB = percentage of women on the board; ILCU = Irish League of Credit Unions; OO = Orange Order; SAF = strategic action fields; UFCU = Ulster Federation of Credit Unions.

†

Variable details are provided in Table 3. The base case in ^αModel 1 is Non-ILCU, in ^ꞎModel 2 is Non-OO, and in ^∞Model 3 is ILCU-only. ^ϟYear effects are significant and positive in all three models. Standard errors are adjusted for clustering at the level of the credit union.

Robust z-tests are reported in parentheses: *p ≤ .05. **p ≤ .01. ***p ≤ .001.

Univariate Analysis: Dependent Variable

The distribution of the average percentage of women on boards analyzed across the different hypotheses categories is recorded in Table 4.

The average percentage of women on boards reported by the 167 credit unions in Northern Ireland over the period 2003 to 2012 is 42.07. Wide variation in this overall average is reflected by the large standard deviation of 21.31 relative to the mean value (Table 5). The minimum value is 0 and the maximum 100 (untabulated). The boards of the largest SAF with 101 credit unions, the ILCU, reports that on average 44% of directors are women. This SAF experienced the highest growth rate over the period at 8.63% (Table 4). The UFCU SAF has 51 credit unions members and reports that on average 41.06% of directors are women with a growth rate of 5.50% (Table 4). Finally, on average 35.05% of directors in the 45 credit unions connected to the OO SAF are women. This average grew by 8.31% from 33.93% in 2003 to 36.75% in 2012 (Table 4).

A closer analysis of groups within the UFCU SAF reveals that the 21 UFCU-only credit unions report on average that 48.09% of their directors are women, whereas the average is 36.28% on the boards of the 30 UFCU&OO credit unions. Finally, 32.68% of directors on the boards of the 15 OO-only credit unions are women (Table 4).

Univariate Analysis: Control Variables

Descriptive statistics and correlation coefficients for the sample are contained in Table 5.

Spearman correlation coefficient tests identified a significant positive association between the percentage of women on boards and MDM and a significant negative association between the percentage of women on boards and credit union size and board size, respectively. The variables MDM, size and board size are significantly correlated. However, the estimated correlation coefficients are all below 0.3 in absolute magnitude.

Multivariate Analysis

Table 6 presents the results based on the random effects models described earlier.⁸ Model 1 investigates the influence of the different SAF rules of the game, ILCU and non-ILCU, on the percentage of women on credit union boards as predicted under H1. Model 2 explores the association between the percentage of women on credit union boards and conservative religious identity (OO SAF versus non-OO credit unions) as predicted under H2. Model 3 investigates: whether the conservatism of the OO dominates the gender social order in boundary spanning credit unions (UFCU&OO) as predicted by H3a; whether in UFCU-only credit unions, credit union identity prevails despite proximate relationships with the 30 dominant conservative boundary spanning credit unions (UFCU&OO) resulting in a higher percentage of women on boards as predicted by H3b; and whether being part of the UFCU SAF, where credit unions with religious identity are incumbent, has moderated the credit union gender identity of UFCU-only credit union boards (H3c).

Women on Boards and SAF (Trade Association) Rules

Consistent with the univariate findings in Tables 4 and 5, the percentage of women on credit union boards affiliated to the ILCU, which promotes gender representation on boards, are found to be statistically significantly higher using a one-tailed test (Table 6: Model 1: ${\hat{γ}}_{1}$ = 9.61, p ≤ .05) relative to all other credit unions (the base case) in Northern Ireland. Therefore, H1 is supported.

Women on Boards and Links to Religion (the OO)

Consistent with the pattern reported in Table 4, credit unions sponsored by the OO, have a significantly lower reported percentage of women on their boards relative to the other credit unions in Northern Ireland (the base case) using a one-tailed test (Table 6, Model 2: ${\hat{γ}}_{2}$ = −12.67, p ≤ .01). Therefore, H2 is also supported.

Boundary Spanning Credit Unions

In Model 3, the coefficients between UFCU-only, UFCU&OO, and OO-only credit unions are reported relative to the base case, ILCU-only credit unions. UFCU&OO credit unions report .18 percentage points lower women on their boards ceteris paribus relative to OO-only credit unions (Table 6: Model 3, ${\hat{β}}_{3}$ - ${\hat{β}}_{4}$ = 13.28 – 13.10). An asymptotic t-test of the difference between the two estimates indicates that the differential using a one-tailed test is not statistically significant (t = .00, p = .97). Therefore, H3a is supported as the percentage of women on boards in both groups are statistically similar. The influence of OO religious identity outweighs the influence of UFCU’s cooperative identity in boundary spanning credit unions.

Consistent with the univariate statistics reported in Table 4, the percentage of women on UFCU&OO credit union boards is 11.8 percentage points lower than UFCU-only credit unions (Table 6: Model 3, ${\hat{β}}_{2}$ − ${\hat{β}}_{3}$ = −13.28 − (−1.48)). A t-test of the difference between these latter two estimated effects indicates that the difference is statistically significant using a one-tailed asymptotic t-test (t = 2.36, p ≤ .05). Therefore, H3b is supported - UFCU-only credit unions with a cooperative identity have a higher percentage of women on their boards.

Finally, as there is no significant difference in the percentage of women on boards between UFCU-only and ILCU-only credit unions (Table 6, Model 3, −1.48), H3c is supported. This further reinforces support for H3b indicating that despite UFCU-only credit unions’ proximity to UFCU&OO credit unions, its cooperative identity and moderate stance on the social order of women within their credit unions is sufficiently strong to deliver a similar percentage of women on their boards compared to those reported by ILCU-only credit unions.

The results of the Wald tests of the joint significance of the interaction of the time dummies with SAF categorisations as required to test H4 are given in Table 7.

Table 7.

Tests of the Joint Significance of the Interaction of the Time Dummies With Collective Groups.

Collective groups	Wald tests (probability value)
Observations	1,618
ILCU-only	9.66 (p = .379)
UFCU-only	6.77 (p = .661)
UFCU&OO	14.17 (p = .117)
OO-only	8.14 (p = .520)

Note. ILCU = Irish League of Credit Unions; OO = Orange Order; SAF = strategic action fields; UFCU = Ulster Federation of Credit Unions. Standard errors are adjusted for clustering at the level of the credit union. Coding is described in Table 3.

As shown in Table 4, the average percentage of women on boards increased in all four groups, with ILCU credit unions reporting the highest growth level at 8.63% and UFCU credit unions reporting the lowest growth level at 4.36%. However, the statistical variation across the years was not found to be jointly significant for all four groups (Table 7), hence these effects are stable across time. Therefore, H4 is supported.

Control Variables, Robustness Testing and Endogeneity

The findings are as anticipated for the control variables MDM, size, board size and year effects. Given the only difference between Models 1, 2, and 3 relates to a re-parameterization of the three SAF dummy variables to create four group dummy variables, the estimates for the control variables are similar, though the standard errors differ slightly. Thus, the interpretation here focuses on those reported for Model 1. Credit unions in more deprived locations have a higher percentage of women on their boards (Table 6, ${\hat{α}}_{2}$ = 0.20, p ≤ .05). However, the magnitude of the estimated effect is small as indicated by the size of the coefficient estimate. Larger credit unions have a lower percentage of women on their boards as denoted by the significant negative coefficients (Table 6, ${\hat{α}}_{1}$ = −3.51, p ≤ .01).

Robustness Testing

Credit union age is included as a standard control variable in many studies (Forker & Ward, 2012; Ward et al., 2021). Therefore, all models were run with the log of credit union age as an additional control variable. However, its association with the percentage of women on boards was statistically insignificant across all models and so did not alter the sign or significance of any of the other reported coefficients. Moreover, the Spearman correlation coefficient between size and age (0.82) was high and hence age is omitted given the potential for the introduction of multicollinearity. Though crude, Gujarati (1999) suggests a cut-off coefficient of 0.80. All robustness testing results are available from the authors on request.

Endogeneity of SAF Categories

An econometric concern regarding the empirical analysis is the potential endogeneity of choice of SAF category (trade association and OO affiliation) with respect to the percentage of women on boards. For instance, endogeneity is theoretically possible if the “unobservables” that determine the percentage of women on boards are correlated with the unobservables determining the choice of SAF category. This may occur if the change in the percentage of women on boards affects the SAF category choice. The presence of such endogeneity would potentially bias the estimated effects of SAF choice on the percentage of women on boards. The absence of necessary variables to perform the statistical role of identifiers to address this endogeneity problem in the current setting prohibits formal testing of this statistical proposition. However, we argue that although it is a theoretical possibility, it is unlikely to be relevant in the current application for two reasons. First, trade association and OO affiliation observed in the data is determined primarily by the national and religious identity of the communities the individual credit unions serve and not by gender on boards. Prior studies have identified the importance of identity in determining behavior (Akerlof & Kranton, 2000). This argument is particularly salient in the context of Northern Ireland, where religion and nationality are important (McNicholl et al., 2019). Therefore, over the short period of time covered by the data, the choice of SAF affiliation is likely to be exogenously determined given the glacial change in national identity affiliation in Northern Ireland across this period (Tonge & Gomez, 2015). This is supported by the fact that over the period 2003 to 2012 no credit union changed SAF affiliation.

Second, another source of potential endogeneity bias would be if the credit union type was measured with error. This possibility can be excluded given the credit union status is precisely known given their affiliation. Overall, we believe in the current context that it is difficult to argue in this application that the choice of SAF affiliation is endogenous with respect to the percentage of women on boards.

Discussion

As predicted credit unions within the SAF promoting gender representation, the ILCU, have a significantly higher percentage of women on their boards when compared to the average reported by all other credit unions in Northern Ireland. This is consistent with SAF theory which asserts that SAF rules of the game are important in both reflecting and enforcing the accepted social order of the SAF within individual fields, such as credit union boards. At 44%, the average percentage of women on credit union boards is high relative to the average percentage of women on for-profit boards at the time (Potter, 2015).⁹ However, consistent with earlier findings for non-profit boards (Fredette & Sessler Bernstein, 2019; Lee, 2014; Moore & Whitt, 2000), the proportions are under-representative of the general population in Northern Ireland, the gender mix of the membership, the staff, and the operating volunteer workforce of credit unions in Northern Ireland (McKillop et al., 2003). Therefore, despite the rules promulgating gender representation, this is not reflected in individual ILCU credit union choice of percentage of women on boards.

This may be explained by the view that social orders, once established are difficult to change (Bielefeld & Cleveland, 2013) as confirmed by our results. SAF social orders are influenced by the broader environment, which encapsulates social and cultural processes (Fligstein & McAdam, 2011; Lounsbury & Ventresca, 2003). Therefore, consistent with Bjerregaard and Jonasson (2014), it may be that social orders already in existence within society in general, have moderated the practical application of the ILCU’s gender representation policy. As an all-Ireland body, the position of women within the social order of ILCU credit unions would have been influenced by the social order within Irish society. When the credit union sector emerged in Ireland in the 1950s and 1960s Irish society was arguably patriarchal. For example, a marriage bar that blocked married women from obtaining employment in the civil service (public sector) was removed in Northern Ireland in 1946 but remained in place in the Republic of Ireland until 1973 (Rouse, 2018). This provides a possible explanation for the lower representation of women on ILCU credit union boards, despite cooperative identity and ILCU recommendations on women on boards, to that observed in the membership, employee, and volunteer stakeholder groupings (McKillop et al., 2003).

A key finding is that the 45 credit unions sponsored by the OO, a conservative religious organization, have significantly lower percentages of women on their boards relative to other credit unions. This finding is consistent with prior studies that report religion as having a constraining effect on gender representation (Dildar, 2015; Racko & Burchell, 2013). However, despite the significant difference, the average percentage of women on credit union boards with links to the OO is high, at 35.05%, relative to other organizations in Northern Ireland, such as private sector companies during this timeframe (Potter, 2015). This would suggest a more permissive social order to that historically promoted by the OO.

A closer analysis of the 45 OO credit unions with a religious association common bond identifies two collective groups: 30 credit unions that are also affiliated with the UFCU (boundary spanning), the remaining 15 credit unions only have links with the OO. The results confirm that the OO SAF is the dominant influence over the boundary spanning group. This finding is explained by SAF theory which predicts that incumbent actors create SAF social orders that enable them to retain power over decision-making (Fligstein & McAdam, 2011). The OO requirement, that OO-sponsored credit unions officially register with an associational common bond, restricts membership to OO members, family and friends. This reduces the potential for challenge to the status quo by ensuring that the membership supports the conservative religious identity of the incumbent actor, the OO.

The significant difference in the percentage of women on credit union boards between the UFCU’s two collective groups reported in Table 5 (UFCU-only = 48.39%, and UFCU&OO = 36.28%) is explained by SAF theory, which predicts that incumbent actors (UFCU&OO) use social skills to design SAF rules to ensure they retain power and resources, while also ensuring the rules are acceptable to challenging actors within the SAF (UFCU-only). This may explain the UFCU SAF’s conservative recommendation, that boards should comprise at least two women. As the average board size during this period of UFCU credit unions is 12.7 directors (untabulated), a minimum threshold of two, represents 15.7% which is unlikely to represent a critical mass. Critical mass is the tipping point at which female voices become relevant to decision-making. Although Torchia et al. (2011) suggest that two or more women on a board is considered to create a critical mass, most studies and policy recommendations use between 30 and 40 percent of the board as a minimum threshold (European Commission for Justice, 2012). Therefore, the UFCU set a gender target that may not weaken male dominance on boards for credit unions with links to a religious organization (UFCU&OO), yet it provides sufficient flexibility for greater female representation on the boards of the challenging group of actors, the UFCU-only credit unions.

Interestingly, the average number of women on OO-only credit union boards is higher at 4.34 women (untabulated) than the minimum threshold of two. The higher number may reflect progress within the OO, responding to calls for improved gender equality from women within its ranks (Urquhart, 2018). Alternatively, it may be that representation of 36.28% does not reflect “female voice” in OO credit unions due to a preference of women within the OO of assisting men in their roles as opposed to taking their roles (Urquhart, 2018). Davis and Gao (2020) found that religious women preferred a traditional female role, yet when controlling for preference, conservative religious norms constrain the percentage of women on boards. Further research taking collective identity into account is thus warranted. Conservative religious identity is confirmed as having a constraining influence on the percentage of women on credit union boards in Northern Ireland. Furthermore, the results confirm that cooperative identity, irrespective of the type of global organization affiliated to, results in similar choices of percentage of women on boards.

An overarching finding for the credit union sector is the overlooked implications of the associational common bond. Theoretically the common bond is a social glue that reduces the information asymmetry in credit decisions, enabling credit unions to award credit based on character knowledge. The tight social bond also reduces moral hazard, hence increases the likelihood of repayment (Ward & McKillop, 2010). However, in this study we uncover that it may also be used to maintain a hierarchal relationship and hence stifle the potential for diversity which is contrary to the cooperative principle of “open and voluntary membership.”

Finally, prior studies identify that social structures are difficult to change (Bielefeld & Cleveland, 2013). A feature of SAF theory is that it explains how and why actors use social spaces, SAFs, to maintain the status quo (Fligstein & McAdam, 2011). The “how” is the carefully negotiated rules of the game designed to provide order for the SAF, as it is in all actors benefit to keep the SAF stable (McNay, 1999). The “why” are benefits from resources, whether economic or symbolic, for example identity or legitimacy. The results in the final hypothesis identify that the position of women within the social order is stable for all SAF groups. Though the percentage of women on boards increased in every SAF as noted in Table 4, the change over the 10-year period was not significant. This suggests path dependency and social order inertia concordant with the conservative percentage of women on boards reported in other organizations in Northern Ireland at this time (Potter, 2015). Also supported is the view that historic social structures moderate the percentage of women on boards (Bjerregaard & Jonasson, 2014) and that change to the social order of percentage of women on boards will not arise within boards unless they experience a dramatic shock (Bielefeld & Cleveland, 2013; Fligstein & McAdam, 2011). Moreover, our findings suggest that progress toward representative proportions of percentage of women on boards will be glacially slow and intervention may be required for more timely alignment, such as the changes in Australia, wherein the Sex Discrimination and Fair Work (Respect at Work) Amendment Bill 2021 extends legal protection against gender discrimination to unpaid workers including volunteers, interns and the self-employed (Parliament of Australia, 2021).

Finally, a key limitation of this study relates to the data. Richer detail on each director, would have enabled an evaluation of whether differences in the percentage of women on the board was attributable to differences in human capital (Camarena et al., 2021). In addition, annual returns have not been publicly available since the credit union registration function for Northern Ireland credit unions transferred to Financial Services Authority in 2012 (now the Financial Conduct Authority). Thus, our findings are best interpreted as historical.

Conclusions

This study contributes to the existing literature in several ways. The first contribution is the framing of Fligstein and McAdam’s SAF theory to conceptualize the social structuring of fields as it applies to the percentage of women on boards. The second contribution is context. For the first time the study investigates influences on the percentage of women on boards in community non-profit organizations including the role of collective groups and hierarchal SAFs in shaping the social order of women in credit unions as reflected by the percentage of women on boards. The third contribution is empirical. SAF theory is rarely investigated as field boundaries are difficult to define (Fligstein & McAdam, 2012). This unique study-site frames the percentage of women on boards as a social construct. Specifically, as directors are volunteers, the Northern Ireland gender equality regulation is not applicable (The Fawcett Society, 2018). This oversight means that choice of the percentage of women on boards in credit unions in Northern Ireland is influenced by common understandings of the gender social order within SAFs to which they belong, as opposed to compliance with a regulatory target or quota. Moreover, because of Northern Ireland’s unique history, credit union SAFs based on credit union and religious identities respectively are clearly identifiable, thus enabling formal testing of an array of predictions relating to SAFs.

The study highlights SAF-based research design as a fruitful and exciting avenue to inform social policy formation, particularly for community-level organizations. Moreover, the findings are generalisable to jurisdictions where requirements for the percentage of women on boards are low or non-existent and where organizations are inert and influenced by conservative SAF links. For example, according to the World Bank, only 10 countries in the world enshrine gender equality in legislation (World Bank, 2021). Although many developed countries have gender equality legislation, it typically refers to employees, and not unpaid or uncontracted workers. Examples include the US (Tsuruda, 2018) and Britain (GOV.UK, 2021).

Policymakers typically focus on introducing initiatives to encourage companies and institutions to increase the percentage of women on boards. Consistent with Vehka and Vesa (2023), this study highlights the importance of internal governance units as conduits for change. They provide the social space for salient actors to contribute to the design of SAF “rules of the game” and to effect change. Moreover, a holistic approach is required as boards can belong to several hierarchal SAFs and alignment with the collective identity of these SAFs is influenced by resource dependency.

Footnotes

Appendix

Sources of Variables Making Up the Database.

Variable name	Source
Dependent variable
Percentage of women on the board (%WOB)	The annual return provides a list of directors’ names. These were categorized into male or female depending on prefix (Mr was assumed to signify male and Ms/Mrs/Miss female). When there was no prefix name was used (In Northern Ireland names are typically sex determined). When there was doubt about the sex of the underlying director, the credit union was telephoned for clarification. In the period 2003 to 2012, a binary classification for gender is reflective of the norms at the time. This would no longer be the case as Northern Ireland society has a more fluid view of gender.
Independent variables
Trade association affiliation	Credit unions choose to affiliate to a trade association. Yearly, lists of affiliated credit unions were made available to the researcher by the two trade associations. These lists were then used to create the variables for ILCU and UFCU membership.
Religion	During the period 2003–2012, regulation required that credit unions report their common bond type to the Registrar for Industrial and Provident Societies (NI). The document registering common bond type or changes to common bond type was sourced from the Registrar’s website for each credit union¹. In Northern Ireland, all credit unions registered with an associational common bond state that the OO is the relevant association; hence, this is taken as an indicator of religious identity. All other credit unions registered with a residential common bond, which is non-religious.
Control variables
Multiple deprivation measure (MDM)	This data was provided by the NI Statistical and Research Agency (NISRA) (2010). Credit union postcodes were used to manually match the deprivation data at ward level to individual credit union locations.
Size	This data was obtainable from softcopy annual returns that were registered with the Registrar in Belfast.
Board size	This data was obtainable from softcopy annual returns that were registered with the Registrar in Belfast. The return required the credit union to list the names, and roles, of those on the board of directors. The variable is a sum of the names.
Year effects	The date on the annual return specifies the year that the underlying data disclosures relate to.

Acknowledgements

Special thanks to Neil Fligstein and Marc Jegers for helpful detailed comments on earlier drafts of this paper. We also thank participants for feedback at numerous seminar presentations and conferences and the anonymous reviewers and editors for advice and insights that considerably improved the paper.

Data Availability

The data are not publicly available due to ethical, legal, or other concerns.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iDs

Anne Marie Ward

Barry Reilly

Notes

Author Biographies

Anne Marie Ward is a Professor of Accounting at Ulster University. Her current research interests include the role of regulation, institutions and women within non-profit organizations. She has also published on the performance, efficiency, governance and role of volunteers and subsidization in non-profit organizations.

John Forker is an Emeritus Professor of Business and Management at the University of Sussex. His research interests focus on the contribution of financial reporting, auditing and governance to performance evaluation and accountability in private and public sector organizations and on the role of regulation, institutions and women in non-profit organizations.

Barry Reilly is a Professor of Econometrics at the University of Sussex. His research interests lie in the field of applied econometrics and labor economics. The research emphasis has been on labor market outcomes in developing economies with a focus on gender issues, ethnicity, and informal sector activity.

References

Abzug

Galaskiewicz

(2001). Nonprofit boards: Crucibles of expertise or symbols of local identities. Nonprofit and Voluntary Sector Quarterly, 30(1), 51–73.

Akerlof

Kranton

(2000). Economics and identity. The Quarterly Journal of Economics, 115(3), 715–753.

Beck

Behr

Guettler

(2013). Gender and banking: Are women better loan officers? Review of Finance, 17(4), 1279–1321.

Bielefeld

Cleveland

W. S.

(2013). Defining faith-based organizations and understanding them through research. Nonprofit and Voluntary Sector Quarterly, 42(3), 442–467.

Bjerregaard

Jonasson

(2014). Managing unstable institutional contradictions: The work of becoming. Organization Studies, 35(10), 1507–1536.

Bourdieu

(1998). Practical reason: On the theory of action. Polity Press.

Bourdieu

Wacquant

(1992). An invitation to reflexive sociology. Polity Press.

Brown

W. A.

Hillman

A. J.

Okun

M. A.

(2012). Factors that influence monitoring and resource provision among nonprofit board members. Nonprofit and Voluntary Sector Quarterly, 41(1), 145–156.

Cak Niederkohr

Ikeda

. (2005). Credit union governance: White paper. WOCCU. https://www.woccu.org/documents/CUGovernance

10.

Camarena

Feeney

M. K.

Lecy

(2021). Nonprofit entrepreneurship: Gender differences in strategy and practice. Nonprofit and Voluntary Sector Quarterly, 506(6), 1170–1192.

11.

Claus

V. A.

Callahan

Sandlin

J. R.

(2013). Culture and leadership: Women in non-profit and for-profit leadership positions within the European Union. Human Resource Development International, 16(3), 330–345.

12.

Council on Foundations. (2022). The state of change: An analysis of women and people of colour in the philanthropic sector. https://cof.org/sites/default/files/documents/files/2017-Gender-Diversity-Report.pdf

13.

Davis

Gao

(2020). Preferences or patriarchy: Why do religious women work less? Social Indicators Research, 147(1), 287–310.

14.

Dildar

(2015). Patriarchal norms, religion, and female labor supply: Evidence from Turkey. World Development, I, 7, 46–61.

15.

DiMaggio

P. J.

Powell

W. W.

(1983). The iron cage revisited: Institutional isomorphism and collective rationality in organisational fields. American Sociological Review, 48, 147–160.

16.

European Commission for Justice. (2012). Women on boards: Commission proposes 40% objective. https://ec.europa.eu/commission/presscorner/detail/en/IP_12_1205

17.

Fagan

Teasdale

(2021). Women professors across STEMM and non-STEMM disciplines: Navigating gendered spaces and playing the academic game. Work, Employment & Society, 35(4), 774–792.

18.

Faulk

Kim

Derrick-Mills

Boris

Tomasko

Hakizimana

Nath

(2021). Nonprofit trends and impacts 2021. https://www.urban.org/sites/default/files/publication/104889/nonprofit-trends-and-impacts-2021_2.pdf

19.

The Fawcett Society. (2018, January). Sex discrimination law review. https://www.fawcettsociety.org.uk/Handlers/Download.ashx?IDMF=e473a103-28c1-4a6c-aa43-5099d34c0116

20.

Fligstein

McAdam

(2011). Toward a general theory of strategic action fields. Sociological Theory, 29(1), 1–26.

21.

Fligstein

McAdam

(2012). A theory of fields. Oxford University Press.

22.

Forker

J. J.

Grosvold

Ward

A. M.

(2014). Management models and priorities in member associations: Is credit unions’ community involvement crowded-out? Nonprofit and Voluntary Sector Quarterly, 43(25), 105S–123S.

23.

Forker

J. J.

Ward

A. M.

(2012). Prudence and financial self-regulation in credit unions in Northern Ireland. The British Accounting Review, 44(4), 221–235.

24.

Fredette

Sessler Bernstein

(2019). Ethno-racial diversity on nonprofit boards: A critical mass perspective. Nonprofit and Voluntary Sector Quarterly, 48(5), 931–952.

25.

GOV.UK. (2021). Volunteer opportunities, rights and expenses: Volunteers’ rights. http://www.gov.uk/volunteering/volunteers-rights

26.

Gregorič

Oxelheim

L. C.

Randøy

Thomsen

(2017). Resistance to change in the corporate elite: Female directors’ appointments onto Nordic boards. Journal of Business Ethics, 141(2), 267–287.

27.

Gujarati

D. N.

(1999). Essentials of econometrics (2nd ed.). McGraw Hill.

28.

International Cooperative Alliance. (2020). Friedrich Wilhelm Raiffeisen. https://www.ica.coop/en/friedrich-wilhelm-raiffeisen

29.

International Cooperative Alliance. (2025). Cooperative identity, values & principles. https://www.ica.coop/en/cooperatives/cooperative-identity

30.

Lee

Y. J.

(2014). The feminine sector: Explaining the overrepresentation of women in the non-profit sector in the USA. International Journal of Social Economics, 41(7), 556–572.

31.

Lee

Y. K.

(2023). Board gender diversity and nonprofit CEO compensation: Implications for gender pay gap. Nonprofit and Voluntary Sector Quarterly, 531(1), 257–273.

32.

Lounsbury

Ventresca

(2003). The new structuralism in organizational theory. Organization, 10(3), 457–480.

33.

MacPherson

D. A. J.

(2016). Women and the orange order: Female activism, Diaspora and empire in the British world, 1850-1940. Manchester University Press.

34.

Mangan

Ward

A. M.

(2024). Co-operative resilience: Towards a heuristic model of collective action in a crisis. Financial Accountability & Management, 40(3), 308–325.

35.

McAuley

J. W.

Tonge

(2007). “For God and for the Crown”: Contemporary political and social attitudes among Orange Order members in Northern Ireland. Political Psychology, 28(1), 33–52.

36.

McKillop

D. G.

Briscoe

McCarthy

Ward

Ferguson

(2003). Irish credit unions: Exploring the gender mix. Voluntas: International Journal of Voluntary and Non-Profit Organisations, 14(3), 339–358.

37.

McKillop

D. G.

Glass

C. J.

Ward

A. M.

(2005). Cost efficiency, environmental influences and UK credit unions, 1991 to 2001. Managerial Finance, 31(11), 72–86.

38.

McNay

(1999). Gender, habitus and the field: Pierre Bourdieu and the limits of reflexivity. Theory, Culture & Society, 16(1), 95–117.

39.

McNicholl

Stevenson

Garry

(2019). How the “Northern Irish” national identity is understood and used by young people and politicians. Political Psychology, 40(3), 487–505.

40.

Moore

(2000). Gender identity, nationalism, and social action among Jewish and Arab women in Israel: Redefining the social order? Gender Issues, 18(2), 3–28. Fhttps://www.proquest.com/scholarly-journals/gender-identity-nationalism-social-action-among/docview/217431938/se-2

41.

Moore

Whitt

J. A.

(2000). Gender and networks in a local volunteer-sector elite. Voluntas: International Journal of Voluntary and Nonprofit Organizations, 11(4), 309–328.

42.

Moulton

Sandfort

J. R.

(2017). The strategic action field framework for policy implementation research. Policy Studies Journal, 45(1), 144–169.

43.

Mundlak

(1978). On the pooling of time series and cross-sectional data. Econometrica, 46(1), 69–85.

44.

Northern Ireland Statistics and Research Agency. (2010). Northern Ireland multiple deprivation measure 2010. https://www.nisra.gov.uk/statistics/deprivation/northern-ireland-multiple-deprivation-measure-2010-nimdm2010

45.

Northern Ireland Statistics and Research Agency. (2011). Census 2011: Key statistics for Northern Ireland. https://www.nisra.gov.uk/sites/nisra.gov.uk/files/publications/2011-census-results-key-statistics-statistics-bulletin-11-december-2012.pdf

46.

Ostrower

Stone

M. M.

(2006). Governance: Research trends, gaps and future prospects. In Powell

W. W.

Steinberg

(Eds.), The nonprofit sector: A research handbook (2nd ed., pp. 612–628). Yale University Press.

47.

Parliament of Australia. (2021). Sex discrimination and fair work (respect at work) amendment bill 2021. https://www.aph.gov.au/Parliamentary_Business/Bills_LEGislation/Bills_Search_Results/Result?bId=s1306

48.

Potter

(2015, October 27). Women on boards of the top 100 companies in Northern Ireland (Northern Ireland Assembly: Research and Information Service Briefing Note, Paper 113/15. Women on Boards of the Top 100 Companies in Northern Ireland). https://www.niassembly.gov.uk/

49.

Quinn

(1999). Credit unions in Ireland. Oak Tree Press.

50.

Racko

Burchell

(2013). The role of technical progress, professionalization and Christian religion in occupational gender segregation: A cross-national analysis. Work, Employment & Society, 27(4), 581–599.

51.

Renz

D. O.

Brown

W. A.

Andersson

F. O.

(2023). The evolution of nonprofit governance research: Reflections, insights, and next steps. Nonprofit and Voluntary Sector Quarterly, 52(Suppl. 1), 241S–277S. https://doi.org/10.1177/08997640221111011

52.

Roberts

D. A.

(1971). The Orange Order in Ireland: A religious institution. British Journal of Sociology, 22(3), 267–282.

53.

Rouse

(2018). Gendering the institutional legacies of the Northern Ireland senior civil service. Administration, 66(3), 55–74.

54.

Sex Discrimination (Northern Ireland) Order (1976). https://www.legislation.gov.uk/nisi/1976/1042/contents

55.

Stephenson

A. L.

Dzubinski

L. M.

Diehl

A. B.

(2023). A cross-industry comparison of how women leaders experience gender bias. Personnel Review, 52(1), 145–165.

56.

Stone

M. M.

Sandfort

J. R.

(2009). Building a policy fields framework to inform research on nonprofit organizations. Nonprofit and Voluntary Sector Quarterly, 38(6), 1054–1075.

57.

Stuart

(2007). Institutional change and embeddedness: Caste and gender in financial cooperatives in rural India. International Public Management Journal, 10(4), 415–438.

58.

Themudo

N. S.

(2009). Gender and the nonprofit sector. Nonprofit and Voluntary Sector Quarterly, 38(4), 663–683.

59.

Tonge

Gomez

(2015). Shared identity and the end of conflict? How far has a common sense of “Northern Irishness” replaced British or Irish allegiances since the 1998 Good Friday Agreement? Irish Political Studies, 30(2), 276–298.

60.

Torchia

Calabró

Huse

(2011). Women directors on corporate boards: From tokenism to critical mass. Journal of Business Ethics, 102(2), 299–317.

61.

Tsuruda

(2018). Volunteer work, inclusivity, and social equality. In Collins

Lester

Mantouvalou

(Eds.), Philosophical foundations of labour law. Oxford University Press. https://philosophy.ucla.edu/wp-content/uploads/2016/08/Tsuruda_VolunteerWork.pdf

62.

Urquhart

(2018). Unionism, Orangeism and war. Women’s History Review, 27(3), 468–484.

63.

Vehka

Vesa

(2023). Internal governance of interest organizations: The role of boards of directors in advocacy performance. Nonprofit and Voluntary Sector Quarterly, 52(2), 474–492.

64.

Ward

A. M.

Forker

J. J.

(2017). Financial management effectiveness and board gender diversity in member-governed, community financial institutions. Journal of Business Ethics, 141(2), 351–366.

65.

Ward

A. M.

Forker

J. J.

Reilly

(2021). The effect of bundling trade association monitoring and female board representation on loan losses in community credit unions. Nonprofit and Voluntary Sector Quarterly, 50(6), 1193–1215.

66.

Ward

A. M.

McKillop

D. G.

(2006). The pattern of subsidisation and assistance received by credit unions in Northern Ireland. Journal of Public and Nonprofit Services, 34, 89–100.

67.

Ward

A. M.

McKillop

D. G.

(2010). Profiling: A strategy for successful volunteer recruitment in credit unions. Financial Accountability and Management, 26(4), 367–391.

68.

Wittberg

(2013). Faith-based umbrella organizations: Implications for religious identity. Nonprofit and Voluntary Sector Quarterly, 42, 540–562.

69.

Wooldridge

J. M.

(2010). Econometric analysis of cross section and panel data (2nd ed.). MIT Press.

70.

World Bank. (2021). Women, business and the law 2021. https://www.worldbank.org/content/dam/sites/wbl/documents/2021/02/WBL2021_ENG_v2.pdf

Strategic Action Field Theory: Explaining Variation in the Percentage of Women on Boards in Credit Unions

Abstract

Keywords

SAF Theory, Credit Union Fields, and Hypotheses

SAFs and Credit Union Fields in Northern Ireland

SAF Collective Identity and Women on Boards (Hypothesis 1)

Religious Identity and Women on Boards (Hypothesis 2)

Collective Groups, Common Bond Strategic Choice, and Women on Boards (Hypothesis 3a–3c)

Field Stability (Hypothesis 4)

Data, Variables, and Methodology

Data

Variables

Control Variables

Multiple Deprivation Measure

Credit Union Size

Board Size

Year Effects

Methodology

Results

Univariate Analysis: Dependent Variable

Univariate Analysis: Control Variables

Multivariate Analysis

Women on Boards and SAF (Trade Association) Rules

Women on Boards and Links to Religion (the OO)

Boundary Spanning Credit Unions

Control Variables, Robustness Testing and Endogeneity

Robustness Testing

Endogeneity of SAF Categories

Discussion

Conclusions

Footnotes

Appendix

Acknowledgements

Data Availability

Declaration of Conflicting Interests

Funding

ORCID iDs

Notes

Author Biographies

References