Does Gender Matter? A Quantitative Multi-Country Study on the Effect of Gender Representation in Senior Management Positions in Public Administration on National Government Performance

(1)

Does Gender Matter?

A Quantitative Multi-Country Study on the Effect of Gender Representation in

Senior Management Positions in Public Administration on National Government

Performance

Master Thesis

Public Management and Leadership

Amna Selimovic S1686194 Prof. dr. S.M. Groeneveld Leiden University Faculty of Governance and Global Affairs Netherlands, the Hague June 16th, 2020

(2)

Abstract

The purpose of this research is to investigate whether gender representation in senior management positions in public administration influences government performance at country level. Furthermore, this study focuses on situational factors having a moderating effect on the relationship between gender representation in public administration and government performance. The situational factors are gendered welfare regimes and gender roles. By conducting a multilevel regression analysis, using a fixed effect model, the results show that there is a significant effect of gender representation on government performance. However, contrary to the hypothesis, the results show that the relationship between gender representation in public administration and government performance is negative. The mixed model also shows that two situational factors, women’s employment and gender roles, have a significant moderating effect on the relationship between gender representation and government performance. But again, contrary to the hypotheses, the results show a negative moderating effect of women’s employment and a positive moderating effect of gender roles on the relationship between gender representation in senior management positions in public administration and government performance.

(3)

1. Introduction

1.1 Women and leadership

Females have been leaders throughout history. From the queens of England to the pharaohs of Egypt, female leaders are present in nearly every culture and time span (AAUW, 2016, p. 6). Yet, in almost all cases, male leaders outnumber female leaders to a considerable degree. Furthermore, customs and laws against female leadership can be found throughout history, particularly in all major religions (Christ, 2014). Females have acted as leaders in social movements, they have driven initiatives to enhance healthcare and sanitation, expand public schooling, build public libraries and construct social welfare programs. They also led societal change in varied environments such as the peace movement, trade movements, school reform (Keohane, 2012) and the civil rights movement (Barnett, 1993). While frequently unknown to the wider community, women have continued to develop significant institutions through their volunteerism, which in effect provided opportunities for women’s leadership in the paying sector. In short, female leadership is not a new phenomenon (AAUW, 2016, p. 6).

However, women are much less likely to be deemed leaders than men. The Global Gender Gap Report of 2018 showed that “when it comes to political and economic leadership, the world still has a long way to go” (WEF, 2018, p. 7). Across the 149 countries surveyed, only seven actually have women as heads of state, while on average 18 percent of ministers and 24 percent of parliamentarians worldwide are women. Likewise, women occupy just 34 percent of the managerial roles in countries where statistics are available, and less than 7 percent in the four worst-performing countries. Moreover, the reports’ key findings include that worldwide the average parity gap is 68 percent, which is a slight increase from the previous year. In other words, there is already an overall gender difference of 32 percent worldwide (WEF, 2018, p. 7).

Organizational, geographical, fiscal and political inequalities impact the decisions of and form incentives for both men and women. Underrepresentation of women in leadership has been described as a gap in which women are prevented from being leaders. The originally defined ‘glass ceiling’, the symbolic wall that women reach at mid-level management, can now be seen as a labyrinth (AAUW, 2016, p. 15). There is no lack of qualified women to occupy leadership, managerial or board roles. In the United States, women make up almost half of the labor force. They outnumber men in earning bachelor’s and master’s degrees and are also on the same level in gaining professional or medical certificates (AAUW, n.d.). Despite record breaking amounts of women elected to the public service in 2018 and the growing media

(5)

exposure of gender representation, women still remain underrepresented in industry-wide leadership positions (AAUW, n.d.). What a leader should look like, their qualities, as well the road to reach leadership roles are still overwhelmingly focused around an outdated male model that leaves out women (AAUW, n.d.). Since men have been leaders for too long, the features related to leadership are still viewed as masculine and not perceived desirable when presented by women (AAUW, n.d.). Men still outshine women in their networks to learn about chances and find advisors and sponsors to support their progress (AAUW, n.d.). Sexual assault, toxic workplace conditions and implicit biases still remain constraints, especially for women of color. They face more challenges to their development and are thus less able to step into leadership positions. Also, finding a balance between family and work can be a challenge which prevents women from pursuing leadership roles. A decades-old concept of male and female household positions is still reflected in the workplace (AAUW, n.d.). Just as the status quo excludes women from leadership positions, it also prevents men from accepting care-taking and supportive responsibilities. If all genders are able to make their own decisions, everyone benefits. With more women in top job positions, organizations profit from a broader variety of creative leadership talents, enabling them to hire and attract a more skilled workforce (AAUW, n.d.).

1.2 Gender representation and performance

Many studies emphasize the importance between gender representation and organizational effectiveness. Andrews et al. (2005) examined the effect of representative bureaucracies on organizational performance and argued that the public’s interest will be represented more effectively if the bureaucracy is similar to the population it serves. But there is still little and inconclusive evidence on whether and how representative bureaucracies, including women in senior management positions in public administration, affect government performance (Andrews et al., 2005).

Whether the involvement of minorities influences managerial decision making and the efficiency of the public sector generally depends on the institutional and national context. Thus far, only a couple of representative bureaucracy studies have specifically addressed whether there is a relationship between representativeness and performance of the public sector (Keiser et al., 2002). Many studies have pointed to certain factors, for example, the salience of a minority community affiliation to a policy problem (Dolan, 2000) or pre-entry versus institutional socialization (Wilkins & Williams, 2008). These two methods of socialization both influence when and to what degree representativeness drives work processes and their results.

(6)

These conditions have in common that they can be considered situational or contextual factors, that translates the individual’s values into behavior. This behavior in turn influences the relationship between minority representation and performance at the organizational level (Andrews et al., 2016). Nevertheless, scientific research about whether and how context influences representative bureaucracy and its outcomes remain sparse and diffuse. Current research on diversity management and performance has highlighted the need for studying situational factors that have an impact on the connection between diversity management and government performance, not only at the individual and organizational level but also at the country level (Andrews et al., 2016). This study focuses on gender representation in senior management positions in public administration, their effect on government performance and situational factors that impact the relationship between gender representation in senior management positions in public administration and government performance at the country level.

1.2.1 Gendered welfare regimes

Bureaucracies are transparent systems that are affected by their environments which in turn are shaped by their actions (Thompson, 1967). This means that country level government performance is shaped by the environment in question. Gender relations, reflected in the sexual division of labor, discourses and ideologies of citizenship, motherhood, femininity and masculinity form the character of the welfare state (Orloff, 1996). Additionally, the institutions of social provision, the collection of financial aid and social care services, universal citizenship entitlements and public services, which we refer to as the welfare state, impact gender relations in a number of ways (Orloff, 1996). There is plenty of literature about the effect of welfare systems on social outcomes (Kushi & MacManus, 2018; Esping-Andersen, 1990). The institutional and historical characteristics of the welfare state systems determine policies and social spending of the state, which influences who has access to benefits and shape government strategies (Kushi & MacManus, 2018). As many authors have argued, this means that the welfare state systems influence the position of women in society (Orloff, 1996). In some countries it might be that these systems encourage women to follow their careers, while in others it might be that a decades-old concept of male and female household positions is still reflected in the workplace. Because in some countries it is more common that there are more women in the labor market, the positive effect of representation will be bigger. When representation of women in these societies increases, there will be more diversity and creativity and therefore impacting the relationship between representation and performance – according

(7)

to the literature – in a positive way. In other societies there is an uneven distribution of men and women. When the representation of women in these societies increases, the system gets disrupted and this might stand in the way of government performance. Because various institutional contexts may indicate different meanings of representation and performance, it is possible that gendered welfare regimes have different impacts on the relationship between gender representation and government performance in different countries.

1.2.2 Gender roles

Because bureaucracies are affected by their environments, which are shaped by their actions, gender roles could affect governmental bureaucracies. For years, many arguments against women participating in politics and public administration have caused several barriers for women to participate. Political theorists like “Aristotle, Jean Jacques Rousseau, Thomas Hobbes, Avicenna, and John Locke justified the exclusion of women from politics, because of their assumed non-rational nature” (Paxton & Kunovich, 2003, p. 91). How people see the role of women and their position in society can either increase or limit the capacity to seek power (Paxton & Kunovich, 2003). It has been proven that the opinion of the public matters for political outcomes (Paxton & Kunovich, 2003). This means that, even when there is a favorable political system and/or a good supply of female candidates, public opinion and cultural norms can restrict chances for women to participate in politics and public administration. Because of this, it is likely that when they do participate, women will not be able to perform as well as they would in a society with less stereotypical gender roles. So, it possible that stereotypical gender roles have different impacts on the relationship between gender representation and government performance in different countries.

Based on the current state of the art of the literature on gender representation, welfare regimes and gender roles, the research question of this thesis is: How do gendered welfare

regimes and gender roles affect the relationship between gender representation in senior management positions in public administration and government performance?

1.3 Relevance

As mentioned before, plenty of research has been done on diversity management, gender representation and their effect on organizational performance. The academic relevance of this thesis is that it will test already existing theories on a higher aggregation level. There is little and inconclusive evidence on whether representative bureaucracies affect government

(8)

performance (Andrews et al., 2005). Current research on diversity management, gender representation and their effect on organizational performance have highlighted the need for studying the relationship between gender representation and government performance at the country level. Furthermore, this thesis will take into account two situational factors, which have not been tested in this relationship yet: gendered welfare regimes and gender roles. Therefore, this research improves our interpretation of situational factors that condition the relationship between gender representation and government performance. Lastly, this research will study the connection between gender representation and government performance in different countries over different years. It is important to examine the impact of public management on government performance over a range of time. Considering that studies on representative bureaucracies mostly focus on single-case studies, more comprehensive empirical work on countries with a bigger time span is required (Park & Liang, 2019). The social relevance of this research is that it is essential for the legitimacy of governments. Knowing that women make up half of the world’s population, it is fair to assume that men and women have the same opportunity to be involved in government decision making. Legitimacy consists of factors such as transparency, accountability and representation; gender balance requires representativeness (Clayton, O’Brien & Piscopo, 2019).

1.4 Structure

Following this introduction, the second chapter of this thesis will consist of a theoretical framework about gender representation, representative bureaucracy and government performance. Additionally, the contextual factors gendered welfare regimes and gender roles will be explained. It is of great importance to give clear definitions and explanations about what the concepts mean, in order to be able to provide a clear answer to the research question. This chapter will also present some hypotheses about the relation between gender representation in public administration and government performance and the two moderating variables. Subsequently, chapter three concerns methodology, explaining the methods, data and operationalization of the study. An extensive description of the quantitative research and operationalization will be given. This chapter will also address the validity and reliability of the current research. After that, the fourth chapter will provide an analysis of the results. Lastly, the fifth chapter will summarize the research and answer the research question. Furthermore, it will discuss the limitations of this research and provide some suggestions for future research.

(9)

2. Theoretical framework

This chapter will start with describing the current state of the art literature on government performance. After that, it will analyse the literature on gender representation in management roles, representative bureaucracy and the effect of gender representation on government performance. Subsequently, it will discuss two contextual factors: gendered welfare regimes and gender roles. This chapter will describe whether these factors are important for and have an influence on the relationship between gender representation and government performance.

2.1 Government performance

In the current state of the art literature about government performance there are many different definitions and explanations. There is not one single explanation that provides a clear definition of the concept. The most common studied dimension of government performance is effectiveness, which will be used to give a clear explanation of government performance (Lee & Whitford, 2009).

Government performance effectiveness has become a primary research line for scholars in the public and private sector (Lee & Whitford, 2009, p. 251). However, with the numerous and broadly accepted developments of the last quarter century, the quest for improving government performance has increased among academics and professionals of the public service, public management and public policy (Lee & Whitford, 2009, p. 251). In their article, Lee and Whitford (2009) talk about effectiveness as a dimension of government performance. They explain how this can be studied at the organizational level. Research has divided organizational effectiveness into two types of studies: those that create criteria to measure performance and those that forecast organizational effectiveness using a series of independent variables (Campbell, 1977). Relatively little research has been done on measuring government effectiveness (Lee & Whitford, 2009, p. 251). Although the definition of organizational effectiveness is fundamental to literature on organizations and management − because of its significance for the analysis of organizational systems, procedures and outcomes − it is still difficult to grasp what the concept means (Steers, 1975). This leads to several explanations. For example, Barnard (1938, p. 19) explains that an organization’s actions are effective when “a specific desired end is attained”. Furthermore, Osborne and Gaebler (1992, p. 351) describe effectiveness by answering the question “How well did it achieve the desired outcome?”. Another comprehensive idea is that effectiveness relates to how well the organization is performing; what it is meant achieve when individuals in the department are working well and

(10)

effectively; if the organization’s practices, procedures, leaders enable it to accomplish its task and whether it is really achieving its purpose (Rainey & Steinbauer, 1999). Throughout the whole public sector, nearly all organizations are concerned with success and productivity − at least indirectly − because efficiency in the execution of their goals impacts the wellbeing of citizens (Rainey, 2003).

To be able to understand organizational effectiveness Lee and Whitford (2009, p. 255) further discuss measurement criteria of organizational performance. They state that research found the best way to describe and evaluate operational effectiveness but did not settle on a common standard method for measuring performance (Daft, 2001). They discuss some methods for assessing effectiveness. The primary success criterium of an organizations’ performance is that an organization meets its goals (Pfeffer, 1982). Although this criterium can be applicable to both the private and public sector, it does not provide specific measurement criteria that fit the public sector very well. Therefore, they give some examples: Firstly, they state that the World Governance Indicators, “which measure the quality of public services, the quality of the civil service and the degree of its independence from political pressures, the quality of policy formulation and implementation, and the credibility of the government’s commitment to such policies, are useful for evaluating and improving the entire government’s effectiveness” (Lee & Whitford, 2009, p. 255). These indicators integrate viewpoints on government effectiveness of a big number of respondents from enterprises, citizens and expert surveys in developed and developing countries (The World Bank, n.d.). Additionally, they refer to a number of other authors who used different measurement criteria of government effectiveness. For example, Putnam (1993) in his study of Italian regional governments focused on three performance dimensions: policy processes, policy pronouncements and policy implementation. For policy processes he used indicators such as “cabinet stability, budget promptness, and statistical and information services” (Lee & Whitford, 2009, p. 256). As for policy promptness, he used some other indicators, such as reform legislation and legislative innovation. Finally, for policy implementation he used indicators such as the number of day care centres or family clinics, industrial policy instruments, agricultural spending capacity, local health unit expenditures, housing and urban development, and bureaucratic responsiveness (Lee & Whitford, 2009, p. 256). Another very popular example of government effectiveness indicators the authors mention is research by Boyne (2002), who did a study on English local governments. He used five dimensions of performance as measurement criteria of government effectiveness; “outputs (quantity and quality), efficiency (cost per unit of output), service outcomes (formal effectiveness, impact, equity and cost per unit of service outcome), responsiveness (consumer

(11)

satisfaction, citizen satisfaction, staff satisfaction, and cost per unit of responsiveness), and democratic outcomes (probity, participation, accountability, and cost per unit of democratic outcome)” (Lee & Whitford, 2009, p. 256).

Not only Lee and Whitford have studied government effectiveness. Governments in many countries provide important public services. A given country’s government performance can be measured in part by assessing the standard of public benefit services such as the public expenditure on childcare services, educational expenditures or benefit services on literacy, infrastructure or mortality (La Porta et al., 1999, p. 226). Good quality of such products is a result of valid government performance. A more controversial, yet still significant, indicator is government consumption and employment in the public sector. High government expenditure in these areas may reflect the willingness of the people to pay taxes, because they are happy with what the government does and therefore view the government positively (La Porta et al., 1999).

2.2 Gender representation and government performance

Donald Kingsley (1944) was the first to state that bureaucracies should be similar to the population it serves. It was argued that such a representative bureaucracy could “fill in the gaps left by the political institutions of government” (Andrews et al., 2005, p. 491). From the seventies onwards, the theory of representative bureaucracy has been understood as a theory that secures the interests of all groups in the bureaucratic decision-making processes. This means that a public organization should represent the people in terms of race, ethnicity and gender (Bradbury & Kellough, 2010). The core of representative bureaucracy theory states that representative bureaucracies are effective bureaucracies. The theory claims that some beneficial effects of representation will improve organizational performance. This is because the willingness of individuals to co-produce is affected by the perception of access to power in society and seeing group preferences represented in decision-making (Andrews et al., 2016). Two premises underlie that theory. The first premise is that individuals of the same social cultural background will share beliefs by socialization mechanisms (Long, 1952; Meier, 1975; Mosher, 1968). The second premise is that at the point of decision, bureaucrats will operate in compliance with their beliefs and principles and try to optimize the ideals that are important to them (Meier & Morton, 2015). When a bureaucracy reflects the public it serves, their decisions will match the public’s principles more accurately. By being more responsive to the preferences and wishes of the public, it is believed that representativeness has an impact on the performance and legitimacy of a bureaucracy (Long, 1952; Selden, Brudney & Kellough, 1998).

(12)

In his book, Mosher (1968), argued that a workforce can be representative in two ways: passively and actively. A passive bureaucracy occurs when a workforce exists of persons from specific groups − for example people from ethnic or racial minorities − and women. This means that a passive bureaucracy is representative when it employs ethnic or racial minorities and women proportionately according to the country’s population. The active type of representation occurs when the bureaucrats try to press the interests and needs of the people the bureaucrat is supposed to represent. Thus, bureaucrats will act and do their best to make sure that the interests of ‘their people’ − whom they share the same interests with − are not forgotten and put on the agenda. This type of active representation occurs when bureaucrats have the same core values, attitudes and beliefs as the social group they represent. Their interests are shaped by socialization experiences and the theory argues that their decisions are based on the values, attitudes and beliefs from their own backgrounds. This means that one expects passive representation to lead to active representation (Bradbury & Kellough, 2010).

Many studies have focused on how this passive representation turns into active representation (Meier & Morton, 2015; Schröter & Von Maravic, 2015). One example is that bureaucrats from minority groups affect the outcomes that benefit their groups directly. These bureaucrats also affect outcomes that benefit their groups indirectly, because they influence the behavior of group’s members. Furthermore, they influence the behavior of their colleagues or indirectly influence the policy of the organization. These mechanisms of active representation each impact overall public service performance (Meier & Morton, 2015; Schröter & Von Maravic, 2015).

Moreover, Selden (1997) argues in her book ‘The Promise of Representative Bureaucracy’ that there are some benefits of having a representative bureaucracy. She explains that when a bureaucracy is representative of the diverse population, this shows that there is equal access to power for its members (Selden, 1997, p. 6). When members of ethnic or racial minorities or women become public officials, they will have the access and power to influence public policy. They might add new information, which better reflects the preferences of their group. Selden (1997, p. 6) argues that when a certain group is underrepresented, its preferences are therefore brought to attention to a lesser extent. A diverse bureaucracy would take a wider range of perspectives into account. She also looks into Kingdon’s (1984) agendasetting-theory related to representative bureaucracy. According to Kingdon (1984, p. 160), bureaucrats put subjects on the agenda, because “incumbents in positions of authority change their priorities and push new agenda items; or personnel in those position changes, bringing new priorities onto the agenda by virtue of the turnover.” When the number of ethnic and racial minorities or

(13)

women change in a bureaucracy, they become a part of the decision-making process and thus issues from their groups are more likely to be put on the agenda (Selden, 1997, p. 7). As a result of this, some scholars have studied whether this would lead to increased responsiveness. Kranz (1976) argued that when a society’s demographic composition is reflected in the bureaucracy, this leads to a bigger range of interests in the agenda-setting and decision-making process and thus leads to more responsiveness from those groups (Kranz, 1976). Furthermore, bureaucrats of underrepresented groups will feel more connected to the organization when their representation increases and will therefore be more willing to co-produce and cooperate with bureaucratic agencies (Kranz, 1976). Because of this, it is more likely that possible clients will be more willing to take part in government programs when they identify themselves with the bureaucrats. Conversely, they may feel uncomfortable or intimidated by the bureaucrats when underrepresented and might avoid participating in government programs (Selden, 1997, p. 7). One last benefit Selden (1997) takes into account is related to human resources. Some excluded groups in society, such as minority women, are readily available in the labor pool. A representative bureaucracy would therefore lead to more efficient use of human resources (Selden, 1997, p. 7).

Pitts and Towne (2015, p. 368) state that diversity management or workforce diversity relates to “the extent to which a group is composed of individuals who are similar on relevant dimensions, whether based on observable or unobservable attributes.” There are two key forces that are important when studying the effect of workforce diversity and organizational effectiveness. On the one hand, when a workforce is more representative the possible benefits are related to having more diverse talents, new perspectives and ideas, broader knowledge, creativeness and innovation. Besides that, a diverse workforce encourages political and social legitimacy. On the other hand, one could relate a representative workforce to bad communication and coordination, which results in less trust and confidence between peers (Pitts & Towne, 2015).

Andrews et al. (2005) examined the effect of representativeness of bureaucracies on organizational performance and they state that the public’s interest will be represented more effectively if the bureaucracy is similar to the population it serves. However, it is still not clear whether this is the same at country level. Gender representation and its relationship to performance have already been extensively researched at the individual level of bureaucrats and the organizational level, but research is still limited at country level (Andrews et al., 2005). Based on the theory of representative bureaucracy, one can expect a positive effect of gender

(14)

representation in senior management positions in public administration on government performance. Therefore, the following hypothesis can be formulated:

H1: Gender representation has a positive impact on government performance

Figure 1: Gender representation and government performance

2.3 The role of context

There is still little and inconclusive evidence on whether representative bureaucracies affect government performance. This may indicate that there are moderating factors influencing the association between representation and performance. These have already been extensively researched at the individual level of bureaucrats and organizational level, but research is still limited at country level (Andrews et al., 2005). If the involvement of women in public administration influences managerial decision-making and the efficiency and effectiveness of the public sector generally depends on the institutional context. Thus far, only a couple of representative bureaucracy studies have specifically addressed whether there is a relationship between representativeness and performance of the public sector (Keiser et al., 2002; Andrews et al 2005; Park, 2013). Current studies on diversity management and performance have highlighted the need for studying situational factors that impact the connection between diversity management and performance (Andrews et al., 2016). Recent research argues that more attention will be paid to the role of context when gaining a more complete understanding of representative bureaucracy. Literature indicates that representative bureaucracy has been used across time and across national contexts as a solution to a variety of social, political and institutional problems (Groeneveld & Van de Walle, 2010). Likewise, Meier and Morton (2015) provide an analysis of the study of representative bureaucracy from a cross-national viewpoint, stating that the dominant roles and mechanisms of institutional representativeness differ considerably across nations. Because various contexts may indicate different meanings of representation and different performance conceptualizations, the relationship between representation and performance can also be influenced (Andrews et al., 2016). So, it is to be expected that gender representation varies across countries.

(15)

Situational factors can have a moderating effect in the relationship between gender representation and performance. Bureaucracies are transparent systems that are affected by their environments which in turn shape their actions (Thompson, 1967). This means that national government performance is shaped by the environment they are in. Gender relations, reflected in the sexual division of labor, discourses and ideologies of citizenship, motherhood, femininity and masculinity form the character of the welfare state (Orloff, 1996). At the same time, the institutions of social provision, the collection of financial aid and social care services, universal citizenship entitlements, and public services, which we refer to as the welfare state, impact on gender relations in a number of ways (Orloff, 1996). As bureaucracies are affected by their environments and these environments are shaped by their actions, gender roles could affect governmental bureaucracies. For years, many arguments against women participating in politics have caused several barriers for women to participate. How people see the role of women and their position in society can either increase or limit the capacity to seek political power (Paxton & Kunovich, 2003). It has been proven that the opinion of the public matters for political outcomes.

In light of this, the following two sections will analyse gendered welfare regimes and gender roles as contextual factors - which are essential to representative bureaucracy - and how these factors influence the relationship between representation and performance.

2.3.1 The role of context: gendered welfare regimes

Walby’s (2009) definition of gender regimes states that: “A gender regime is a set of inter-related gendered social relations and gendered institutions that constitutes a system” (Walby, 2009, p. 303). In their article, Roth and Saunders (2019) investigate how different policy responses to the financial crisis in three European countries are shaped by gender differences. For this they used Walby’s (2009) definition of gender regimes. Roth and Sanders (2019) combined gender, social policy and social movement scholarship to investigate how different policies resulting from the financial crisis in Spain, Sweden and the United Kingdom formed gender differences in anti-austerity demonstrations. By using countries with different gender regimes, they found that gender regimes do have significant impact on women’s and men’s participation in anti-austerity protests (Roth & Saunders, 2019). Their comparison of women’s and men’s involvement in anti-austerity demonstrations shows that, at the country level, biographical, structural and psychological resources are more important than, for example, grievances. Biographical resources refer to education, employment and age. Structural resources refer to time and place. Psychological resources refer to interest in politics (Roth &

(16)

Saunders, 2019). They link protest participation to gender regimes and argue that resources are shaped by gender regimes and gender regimes provide access to social support and decision-making. In other words, they show that there is a connection between gender regimes and demonstration participation (Roth & Saunders, 2019).

Gender regimes are quite complex and contain multiple levels of the overall social system, the social practices and the degree of gender inequality (Roth & Saunders, 2019). The indicators include political participation, education, gendered inequality in the workplace, equality legislation, gendered civil liberties and childcare public expenditure (Walby, 2009, p. 303). Walby (2009) argues that there are two types of gender regimes: domestic gender regimes and public gender regimes. They represent ideal types of regimes, which means that in reality, gender regimes vary between those two. Distinctions between the two can be found in the institutional domains of “economy, polity, civil society, as well as the level of the gender regime as a whole” (Walby, 2009, p. 304). Within the economy, the public gender regime entails that women are free wage laborers rather than domestic laborers. This can be studied by looking at the extent that women work in non-agricultural areas. High numbers of childcare services would also indicate a public regime (Walby, 2009, p. 304). The public form within the polity is the extent to which women have seats in parliament. The public form within civil society is the extent to which women have total access to education and the extent to which they have a full range of civil rights (e.g. the right to make use of abortion and contraception).

Using Walby’s (2009) explanation of the two regimes Roth and Saunders (2019, p. 4) describe domestic regimes as regimes where women do most of the unpaid care work. In such regimes, women have limited access to abortion and are less represented in the labor force and politics. In public gender regimes, women are more involved in the labor market and politics and have better access to abortion than women in domestic regimes. Walby distinguishes between two types of public gender regimes; social democratic and neo-liberal regimes (Walby, 2009, p. 307). Social democratic gender regimes entail the provision of public services such as employment regulations and childcare. In neo-liberal gender regimes, the market plays a main role in the provision services (e.g. regulation of employment and childcare) through market mechanisms (Roth & Saunders, 2019, p. 4). “The most important differences between the neo-liberal and social democratic forms of the public gender regime lie in the gendered depth of democracy and the degree of gender inequality” (Walby, 2015, p. 148).

The basis of Walby’s definitions of gender regimes lies in a theory that has been established years before hers, which is Esping-Andersen’s (1990) theory of gendered welfare regimes. In the nineties Esping-Andersen (1990) constructed a typology of welfare state

(17)

regimes consisting of a social-democratic welfare state regime, a conservative-corporatist welfare state regime and a liberal welfare state regime. This typology recognized the ideational power and importance of the three main important political movements of the 20th century in

Western Europe and North America (Hoekstra, 2003). The social-democratic welfare state is based on strong state control. Welfare services are available for a large part of the population. Income differences are very low as a result of the state’s re-distributive effects. This type of welfare state is based on the principle of universalism. There is a high degree of autonomy and low dependence on family and market (Hoekstra, 2003). In a conservative-corporatist welfare, state the state is active in the provision of welfare services, but the state is not responsible for income distribution. Furthermore, it is aimed at maintaining the existing hierarchy in society and because of this, different groups have access to different welfare services. Moreover, the state is not the only one to provide welfare services, but private non-profit organisations and the family also play a role in this (Hoekstra, 2003). The liberal welfare state regime is characterised by very little interference from the state and has a strong market focus. For the bigger majority, welfare services are provided by private companies. The state only helps out when necessary. For example, for low-income families they provide a safety net. Furthermore, the state is characterised by dualism: on the one hand there is equality, but there is also income inequality (Hoekstra, 2003). Similar to Esping-Andersen’s (1990) theory, Walby’s (2009) concepts of gender regimes address public services, women’s employment and women’s political involvement (Roth & Saunders, 2019). There is plenty of literature about the effect of welfare systems on social outcomes (Kushi & MacManus, 2018; Esping-Andersen, 1990). The institutional and historical characteristics of the welfare state systems impact policies and social spending of the state, which influences who has access to benefits and shapes government strategies (Kushi & MacManus, 2018). On top of that, many argue that welfare state systems influence the positions of women in society (Orloff, 1996).

It is of great importance to keep in mind that situational factors can have a moderating effect in the relationship between gender representation and performance. Bureaucracies are affected by their environments and these environments are shaped by their actions (Thompson, 1967). Important to note is that on the one hand in conservative welfare regimes, subsidiarity is central and therefore women’s dependence on family will increase (Orloff, 1996). According to the conservative-corporative welfare state of Esping-Andersen’s (1990) typology, the state is aimed at maintaining the existing hierarchy, which is why it is more likely that women do not fill up positions in governments. This is in line with Walby’s (2009) domestic gender regime, where women do most of the unpaid care work and are less represented in the labor

(18)

force and politics. In this domestic regime it is normal for women to have no paid work, let alone occupy the higher positions in public administration. Other types of welfare regimes are the social-democratic and liberal welfare regimes. Social-democratic regimes promote an individual model of entitlement and offering services, letting those who are responsible for care work enter the paid labor force (Orloff, 1966). In the social-democratic welfare state of Esping-Andersen (1990) the state promotes universalism, with women being more involved in the labor market and politics and having better access to services. Liberal regimes are indifferent to gender relations and leave service provision to the market (Orloff, 1996). There is a strong market focus, with little interference of the state. Both the social-democratic and liberal regime are line with Walby’s (2009) public gender regime. In other words, in a state with a public welfare regime it is more common that women are active in the labor force, politics and public administration.

In section 2.1, it was explained that, according to the literature, gender representation has a positive effect on government performance. Furthermore, it was explained that in a public gender regime, women are more represented in the labor force and politics compared to a domestic gender regime. These regimes are situational factors that may have a moderating effect on the relationship between gender representation and government performance. National government performance in countries is shaped by its environments. In a public gender regime it is more likely that the system encourages women to follow their careers, while in a domestic regime it is more likely that a decades-old concept of male and female household positions is still reflected in the workplace. Because it is more common that women are more represented in a public gender regime, the positive effect of representation will be bigger. There will be more diversity and creativity and therefore impacting the relationship between representation and performance in a positive way. In a domestic gender regime there is an uneven distribution of men and women, meaning that women have to fall back on the existing norms of not working and staying at home. They will simply have no means (e.g. time, money, opportunities) to pursue a very successful career and support a family at the same time. When the representation of women in a domestic regimes becomes higher, the system gets disrupted, and this might stand in the way of government performance. Because of this the relationship between gender representation and government performance is influenced in a negative way. So, the effect of a public gender regime has a different impact on the relationship between gender representation and government performance, compared to a domestic gender regime. Based on the literature above, it is reasonable to assume that a public gender regime has a

(19)

positive moderating effect on the relationship between gender representation and government performance. Following from this, the second hypothesis can be formulated:

H2: A public gender regime is more likely to have a positive impact on the relationship between gender representation and government performance, compared to a domestic gender regime Figure 2: Gender representation and government performance affected by gendered welfare regime

2.3.2 The role of context: gender roles

Paxton and Kunovich (2003) described three types of factors that explain low levels of female participation in legislature; structural, political and ideological factors. The authors believe that the low representation of women in legislature is the result of the supply and demand of female candidates. The structural factors are related to the availability of female candidates. This means that structural factors such as education or professional opportunities are important for women to be able to run for office. If they do not have access to education or professional opportunities, they will not have the financial or human capital to participate in legislature (Paxton & Kunovich, 2003). These factors predict that the educational achievements of women and participation of women will have a positive effect on the level of women in public positions. Another part of the explanation consists of political factors that stand for the institutional differences in a political system. These differences account for a varying demand for women to participate in office. Crucial factors here are the type of dominant political parties and electoral systems in a country, because they have the ability to either obstruct or support the different groups in government to promote their interests (Paxton & Kunovich, 2003).

In a study by the Inter-Parliamentary Union (IPU, 2000) about female politicians, people did not mention structural or political factors as the main reason for low female participation in politics, rather, they mentioned ideological factors as the main cause. The way people see the role of women and their position in society can either increase or limit the capacity to seek political power (Paxton & Kunovich, 2003). It has been proven that the opinion of the public

(20)

matters for political outcomes. This means that, even when there is a favorable political system and ample availability of female candidates, public opinion and cultural norms can restrict chances for women to participate in politics (Paxton & Kunovich, 2003). For years, many arguments against women participating in politics have generated several barriers for women to participate. Political theorists like “Aristotle, Jean Jacques Rousseau, Thomas Hobbes, Avicenna, and John Locke justified the exclusion of women from politics because of their assumed non-rational nature” (Paxton & Kunovich, 2003, p. 91). In the IPU-study people mention a negative ideology much more frequently than the other two reasons (structural and political) to explain low female participation in politics. One respondent stated: “In spite of a

long tradition of active participation in the workforce by a vast majority of women, both women and men see motherhood and marriage as the most important goals in a woman’s life. A common standpoint is that “politics is a man’s business,” and that women are too emotional to deal with affairs of the state. The reasons for this are not to be found in education, with women being as educated as their male counterparts. It is simply because of the stereotyped and traditional structure of society” (IPU, 2000, p. 20).

Similarly, Arceneaux (2001) argues that culture, ideology and gender role attitudes relate to female representation in states. He found that gender role attitudes influence the level of female representation independent from political culture and ideology. He explains that the set of gender role attitudes are important for female representation. In some countries citizens or voters have been socialized to see politics as a man’s occupation. Negative gender role attitudes have consequences for women’s career outcomes (Arceneaux, 2001). Firstly, based on the gender-role attitudes women will not run for office, because they do not believe that they have the expertise for it. Secondly, women will not run for office because there is no demand for female candidates from the political elite and because the political elites is biased against female candidates. Thirdly, women who do choose to run for office, despite the negative voters at polls, are more likely to lose than female candidates in states where there are more positive gender roles are the norm (Arceneaux, 2001). Arceneaux (2001) also refers to Norrander and Wilcox (1998) in his article, who state that ideologically conservative states will have less women in legislation. According to them, women tend to run for office more in countries that are increasingly accepting of female candidates and where gender roles are less traditional (Arceneaux, 2001). Because of this, there is a bigger chance that more liberal and democratic countries will have more women in legislature (Arceneaux, 2001).

Furthermore, the absence of female participation in public administration demonstrates vertical and horizontal occupational gender segregation (Johnston, 2019). “Vertical

(21)

occupational gender segregation is often referred to as ‘glass ceilings’, where women struggle to reach leadership and senior decision-making positions” (Johnston, 2019, p. 156). Women continue to be clustered in lower-level and lower-paid positions inside the hierarchy of the public sector in many federal bureaucracies with no employment progression to the upper echelons (McTavish & Miller, 2006). Horizontal occupational gender segregation occurs when women are clustered in particular industries or public service occupations, such as the educational and health industries. This is referred to as ‘glass doors’: with women employed mainly in stereotypical female occupations, such as nursing (Guy & Newman, 2004; Kerr, Miller & Reid, 2002). The outcome of occupational gender segregation is often that women are being undervalued and underemployed (Johnston, 2019, p. 157).

Johnston (2019) continues with explaining how stereotyping gender role stereotypes are a barrier to female representation in public administration. Social gender inequality in society is the main cause for the existence of those barriers (Walby, 1989). Gender roles in culture and at work are transformed into gender stereotypes. Gender includes masculinity or femininity related ideals and attributes (Johnston, 2019, p. 157). The femininity of women is linked to their reproductive, maternal, caring and domestic roles. Social construction of sex creates gender roles in most societies where a patriarchal power structure is maintained (Nicolson, 1996). The majority of communities uphold the patriarchy, with men in the authority positions and women in the inferior roles. According to King (1995), societies favor masculine attitudes of assertiveness, hostility and leadership over feminine qualities, which relate to nurturance, submissiveness and dependency. Public life is deemed to be the realm of men from which women are excluded (Duerst-Lahti & Kelly, 1995; Mazur & Pollock, 2009). Public agencies are gendered like every other organization, because men’s corporate superiority and resource possession is to women’s disadvantage (Ferguson, 1985; Kelly & Newman, 2001; Savage & Witz, 1992). King (1995) described four ways in which male influence appears in public administration; 1) organizations are the male realm, because men are more likely to be leaders; 2) organizations are the male realm, because gender norms are rooted in a society that contributes to male dominance over female dominance; 3) the state is a masculine realm, and therefore government politics and administrative structures represent the cultural dominance of men over women; and 4) leadership and management is a masculine area, since the cultural dominance of society can be seen as assertive and violent in this concept of leadership (Johnston, 2019, p. 157).

Gender roles can be described as the way people see the role of women and men in society (Paxton & Kunovich, 2003). Because gender roles in culture and at work are

(22)

transformed into gender stereotypes, this creates gender roles in most societies. Gender includes masculinity or femininity related ideals and attributes (Johnston, 2019, p. 157). For men, this entails attitudes of assertiveness, hostility and leadership and for women this entails attitudes of nurturance, dependency and submissiveness (King, 1995). These attributes may therefore create societies with more stereotypical gender roles compared to other societies. In societies with more stereotypical gender roles it may be more common for men to have positions with authority and for women to occupy inferior positions, than in societies with less stereotypical gender roles. As described in the section about the role of context, situational factors can have a moderating effect on the relationship between gender representation in public administration and performance. Stereotypical gender roles are considered situational factors. When a society has less stereotypical gender roles and it is more common for women to have senior administrator positions, it is more likely that they will be taken more seriously, which leads to a higher chance of succeeding at their job and thus influences the relationship between gender representation and performance positively. This may also works the other way around, when a societies has more stereotypical gender roles. In these societies it might be that women in senior administrator positions are taken less seriously, because of the stereotypical gender roles, and are therefore less likely to succeed at their job. Therefore, in such a society, it is more likely that stereotypical gender roles affect the relationship between gender representation and government performance negatively. Following from this logic, the third hypothesis can be formulated:

H3: Stereotypical gender roles have a negative impact on the relationship between gender representation and government performance

(23)

3. Methods, data and operationalization

This chapter will first describe what quantitative research method will be used in this study. To test the hypotheses, I will use several databases to find out whether gender representation in senior management positions in public administration has influence on government performance. This will be discussed in the second section of this chapter. Thereafter, the concepts from the theoretical framework will be operationalized. Finally, the validity and reliability of the research will be discussed.

3.1 Methods

The purpose of this research is to find out whether gender representation in senior management positions in public administration has influence on government performance at macro level and whether there are contextual factors that have a moderating effect on gender representation and government performance. To investigate whether there is a relationship between representation and performance I will use a quantitative research method. Because this thesis focuses on government performance at macro country level, this is the best choice. The variables that will be used have been researched extensively resulting in large amounts of data over various years. This makes the results more accurate (Carr, 1994). The dependent variable is government performance and the independent variable is gender representation.

Using SPSS, I will run multilevel regression analyses. Data of the 27 Member States of the European Union plus the United Kingdom from 2003 until 2018 will be studied. Thus, the data relates to multiple moments in time and multiple countries, specifically 16 years and 28 countries. The reason why I am running a multi-level regression analysis is because the observations of the countries from year to year are not independent from each other, thus violating one of the assumptions of a normal regression analysis. Multilevel analyses cast aside the assumption of homogeneity of regression slopes (Field, 2009, p. 729). Besides, as Andy Field says “mixed models laugh in the face of missing data” (Field, 2009, p. 730), because the model expects missing data and is therefore appropriate to use. First, I ran a normal regression analysis for the relationship between gender representation and government performance, but there was a lot autocorrelation, meaning the data did not allow this type of analysis. I thought this problem could be solved by adding the variable year as a covariate variable, but it did not. Therefore, I will run multi-level regression analyses. Starting point in this analysis is that there are various observations - the different years - that are nested in a unit at a higher aggregation level: the countries. The observations in the different years are related within the countries. I

(24)

will be using a fixed effect model. The data that I have allows for this type of model. With this model I am controlling for differences across countries. The coefficients of the fixed effect absorb all the action across the countries. And because of this, the threat of omitted variables is considerably reduced (Borenstein et. al., 2010). Using this type of research method, I will be able to state whether gender representation in senior management positions in public administration has impact on government performance and whether the two moderating variables, gendered welfare regimes and gender roles, have a moderating effect on this relationship.

3.2 Data collection

Because there is not one database that includes all the variables, I will use several databases and combine them. The databases used in the current research are the Worldwide Governance Indicators (WGI), the European Institute for Gender Equality (EIGE) database on gender statistics, the Gender Data Portal of the Organization for Economic Co-operation and Development (OECD) and lastly the European Values Study (EVS). In order to increase the reliability of the answer to my research question, I will study all the variables from the 27 Member States of the European Union plus the United Kingdom over multiple years.

Government performance will be studied by using the WGI. The WGI analyses composite and individual governance indicators for more than 200 countries and territories. They studied six government dimensions over the period 1996-2018: voice and accountability, political stability and absence of violence, government effectiveness, regulatory quality, rule of law and control of corruption (The World Bank, n.d.). For this study I will use the indicator government effectiveness from 2003 until 2018. This indicator integrates viewpoints of a big number of respondents in developed and developing countries from enterprises, citizens and expert surveys. The WGI dataset is credible and reliable. All the data of the WGI are available to the public and can be downloaded online. The WGI website demonstrates how the data is collected, why the data is collected and who worked on it, which makes it a very transparent dataset. Furthermore, many researchers use the WGI extensively, which implies that it is a good and credible dataset. And, the dataset has a lot of respondents from all over the world, which also makes the WGI dataset reliable (WGI, n.d.).

To study gender representation I will use the database on gender statistics of the EIGE. The data that I will be using is from 2003 until 2018. This database provides data on the number of women and men in important decision-making roles in various life contexts including reliable statistics that can be used to track the present situation and developments over time

(25)

(EIGE, n.d.a). The data on decision-making was collected from the 27 EU Member States, the United Kingdom, five EU candidates, two potential candidates and the other three EEA countries. The areas covered include: “politics, public administration, justice, business and finance, social partners and NGOs, environment and climate change, media, science and research, and sports” (EIGE, n.d.a). The EIGE database has been funded by the European Commission, which is a reliable source. The EIGE’s aim is to provide comparative information that can be used to reinforce and supplement the Strategy on Gender Equality of the European Commission and to assist Member States in tracking their development. The added value of this database is that it is easily accessible, up to date and contains all the gender statistics of the Member States. “The main purpose of the database is to act as a reliable resource in formulation and monitoring of policies that are beneficial for both women and men and will facilitate making appropriate decisions toward the advancement of gender equality” (EIGE, n.d.b). It uses several sources to collect their data: Eurostat, Eurofound, the European Commission, HBSC research network and the WHO. Just like the WGI dataset, this dataset is very transparent, as it is clear where all the data comes from.

One of the moderating variables is gendered welfare regimes. This concept is hard to measure directly and therefore I will use proxy indicators, which are related to gendered welfare regimes. All three proxy indicators will be studied from 2005 until 2018. For this I will use the Gender Data Portal of the OECD. The Gender Data Portal is part of the OECD Gender Initiative, which explores structural barriers to gender equality in education, employment and entrepreneurship. The OECD Gender Initiative tracks governments in promoting gender equality in OECD and non-OECD countries, and offers good practices focused on analytical tools and reliable data (OECD, n.d.a) The OECD’s goal is implementing policies that would enhance people’s economic and social well-being across the globe. The OECD is one of the biggest and most reliable sources of comparative statistical and social data in the world. It tracks developments, collects evidence, analyses, predicts economic growth and investigates emerging dynamics in a broad variety of public areas, such as gender (OECD, 2013). The Gender Data Portal demonstrates to what extent countries are reaching gender equality and where change is really required. The goal is to provide data for users and providers with knowledge regarding the OECD’s direct and indirect data collection. Data on planning, selection of methods and procedural criteria adopted by the OECD countries are available on the website (OECD, n.d.b).

The second moderating variable is gender roles. To study this, I will use a survey question from the EVS from two respective years: 2008 and 2017. I studied another database

(26)

before deciding on the EVS, namely the database on Family and Changing Gender Roles 2012 of the ISSP. The EVS is a large scale, cross-national and longitudinal research on human values, based on the European Member States (EVS, n.d.). The dataset includes more years with questions on human values and is therefore more accurate, compared to the ISSP (ISSP, 2016). In short, it was a very deliberate choice to use the EVS. The EVS started in 1981 and is repeated every nine years. The EVS gives insight into peoples’ thoughts, views, desires, perceptions principles and viewpoints throughout Europe. It is a reliable dataset on how people in Europe think about life, families, jobs, culture and gender roles (EVS, n.d.). All data are provided by GESIS (GESIS, 2010; GESIS, 2020). GESIS transparently provides information on how and where the data is collected. Furthermore, it is a very popular dataset that is used by a lot of researchers, which implies that it is credible (EVS, n.d.).

3.3 Operationalization

Government performance can be analysed in multiple ways with varying objectives. Due to the daunting complexity of measuring performance in the public sector, it is advised to use several indicators including archival (objective) and perceptual (subjective) measures, in order to improve the significance of the results (Andrews et al., 2005). In order to measure government performance I will use the Worldwide Governance Indicators (WGI). As explained in the section above, the WGI studies six government dimensions from which I will use the indicator government effectiveness. This indicator integrates several indicators into one. “Government effectiveness captures perceptions of the quality of public services, the quality of the civil service and the degree of its independence from political pressures, the quality of policy formulation and implementation, and the credibility of the government’s commitment to such policies” (The World Bank, 2019, p. 1). It combines several indicators, and ranges from -2.5 (weak) to 2.5 (strong) government performance.

Kingsley (1944) defines representative bureaucracy as one whose demographics resemble those of the socially ruling class. Most literature on representative bureaucracy classifies identity in specific demographic terms including race and gender. In this thesis I describe sex as binary variable, which only contains men and women, leaving non-binary individuals out. Using the database on gender statistics of the EIGE, I will study gender representation by data on the number of women in important decision-making roles. More concrete, the indicator I will use is the percentage of women in public administration as senior administrators at national level.

(27)

Using the OECD Gender Data Portal, I will study gendered welfare regimes. As mentioned before, this concept is hard to grasp directly, and I will therefore use proxy indicators. In the theoretical framework, I described the two different gender regimes of Walby (2009). She makes a distinction between a public and a domestic gender regime. She describes domestic regimes as regimes where women do most of the unpaid care work, women have limited access to abortion and are less represented in labor force and politics. In public gender regimes women are more involved in the labor market and politics and have better access to abortion than women in domestic regimes. The OECD dataset provides several variables that I will use to find out whether gendered welfare regimes have a moderating effect on gender representation and government performance: women in the workforce, childcare public expenditure and gendered education. Firstly, about women in the workforce: a high percentage of women in the workforce would indicate a public gender regime. A low percentage of women in the workforce would indicate a domestic gender regime. Secondly, childcare public expenditure as a percentage of GDP is an indicator of the gendering of state welfare. A high level would indicate a public gender regime and a low level would indicate a domestic gender regime. Thirdly, gendered education as a percentage of tertiary education of women. High levels would indicate a public gender regime, while low levels would indicate a domestic gender regime.

To study gender roles in the European countries I will use the dataset of EVS. In the theoretical framework, I explained how stereotyping gender roles form a barrier to female representation in public administration. Barriers to women in occupations mainly derive from social gender inequality in society (Walby, 1989). Gender roles in culture and at work are transformed into gender stereotypes. Gender includes masculinity or femininity related ideals and attributes (Johnston, 2019, p. 157). Knowing this, the EVS asks their respondents questions on how they feel about gender roles. There is one question that was included in the fourth wave (2008) and the fifth (2017) that I will use; “When jobs are scarce, men should have priority over women” (EVS, 2017, p. 9). I recoded the answering options into; -1: do not know, 1: agree, 2: disagree, 3: neither.

3.4 Validity and reliability

As I mentioned before, this research investigates whether the dependent variable, gender representation in public administration senior management positions, has a positive impact on the independent variable, government performance at the country level. Adcock and Collier (2001, p. 530) explain that “validity measurement is achieved when scores meaningfully

Does Gender Matter? A Quantitative Multi-Country Study on the Effect of Gender Representation in Senior Management Positions in Public Administration on National Government Performance