The increasing participation of women in public life is a phenomenon that has been occurring worldwide for the last 30 years. As women enter into professions, they transform the institutions in which they work, including the institutions of government. Despite the increasing presence of women in professional and public life, there is a relatively small body of work examining the variation in women's participation cross-nationally and how institutional and structural factors may inhibit or favor women's participation. The paucity of work on women's participation in the institutions of government is most pronounced in the study of cross-national courts. The absence of a strong body of scholarly literature is a noteworthy omission as scholars increasingly consider the variation in the structure of high courts, focusing as well on the democratic nature of political institutions.

This article fills the void left by the literature on cross-national courts. We examine the variation in representation on the high courts of the countries in the Organisation for Economic Co-operation and Development (OECD) in 2007 by means of an original data set collected from 2006 to 2007. In considering this variation, ranging from no women on some courts to 60% of judicial seats on others, we explore the effects of the structure of the high court, the method of selecting judges, and the culture of the country on women's representation. We find that there is a significant amount of variation in women's representation across courts and across countries. Additionally, other factors, including selection and opportunity, affect women's representation on these courts. Finally, countries with a tradition of women's participation in public life, including those with quota laws for the national legislature, see greater participation by women on the high courts.

VARIATION IN WOMEN'S REPRESENTATION

There are a number of studies of women's participation in public life that provide insight into the factors that may affect women's representation on high courts. The framework for understanding representation tends to categorize these factors into those related to the demand side (the characteristics of the institution or country) and the supply side (the characteristics of the candidates).

On the demand side, quota systems increase women's participation within political parties, and that participation translates into representation in governmental institutions such as legislatures (Caul Reference Caul2001; Htun and Jones Reference Htun, Jones, Craske and Molyneux2002). Lower-prestige offices see greater representation of women, because the lower level of prestige makes the office easier to attain and less appealing to men. Additionally, the lack of a need to relocate to serve in these offices is particularly attractive for professional women, who often have young children. The method of selection for such office and the system of representation within the country also influence the number of women, with multimember districts and party lists favoring women's representation (Kenworthy and Malami Reference Kenworthy and Malami1999; Matland and Studlar Reference Matland and Studlar1998; Norris Reference Norris2004; Vengroff, Nyiri, and Fugiero Reference Vengroff, Nyiri and Fugiero2003). Voter selection of government officials can increase women's representation if the women are viewed as political outsiders, free of corruption (Valdini Reference Valdini2005). Other demand-side factors affecting representation include the structural factors within the country, such as the economic growth of the country, women's access to education, political ideology, and perceptions of gender roles (Kenworthy and Malami Reference Kenworthy and Malami1999; Matland and Studlar Reference Matland and Studlar1998; Norris Reference Norris2004; Paxton and Kunovich Reference Paxton and Kunovich2003; Vengroff, Nyiri, and Fugiero Reference Vengroff, Nyiri and Fugiero2003).

On the supply side, not only can the professional characteristics of individuals influence the number of women in public office but so can ambition, the resources of the candidates, and other social background characteristics apart from sex (Matland and Studlar Reference Matland and Studlar1998). Clearly, supply-side and demand-side explanations are interrelated in a way that affects representation. As Maria Escobar-Lemmon and Michelle Taylor-Robinson (Reference Escobar-Lemmon and Taylor-Robinson2005) note in their study of female cabinet officials, the increasing participation of women in one area of government, such as the legislature, can generate ambition and participation of women in other areas of public life. Not only can the supply side and the demand side affect the number of women participating, but they can also affect the types of positions within government that women hold.

While the vast majority of research on women's representation focuses on the executive and legislative branches of government, there is reason to believe that some aspects of such models can be applied to the judicial context. The importance of cultural attitudes, the openness of gatekeepers to recruiting women, and women's access to education are but a few of the structural factors within a country likely to affect participation both in the legislature and on the courts. The party ties that open access to the legislature are likely to be even more important for the judiciary due to the appointive method of selection that most countries employ to fill their courts.

VARIATION IN COURT CHARACTERISTICS

Other aspects of the models of representation developed from legislatures may not translate as easily to the judiciary. Electoral rules, known to influence women's representation in the legislatures, demonstrate how variation in selection can make institutions more or less representative of the people. Courts, on the other hand, are filled almost exclusively through appointment procedures, though the entity with appointment power can vary considerably at the national level. Because seats on high courts are unelected, women's participation should be more likely than for other institutions; the selectors, be they presidents or legislatures, seek the electoral benefit of appointing women (Slotnick Reference Slotnick1984).

Although the cross-national courts literature does not consider women explicitly, it does, however, provide insight into the variation in the design of these courts and how the structure affects the decision making of judges, especially within the context of the separation of powers (Ehrmann Reference Ehrmann1976; Epstein, Knight, and Shvetsova Reference Epstein, Knight and Shvetsova2001; Gibson, Caldeira, and Baird Reference Gibson, Caldeira and Baird1998; Ginsburg Reference Ginsburg2001; Herron and Randazzo Reference Herron and Randazzo2003; Schmidhauser Reference Schmidhauser1987; Schwartz Reference Schwartz1998). There is a significant amount of variation among courts both in terms of the power courts have within government and in the diversity of people who serve on the courts. Some countries divide the responsibilities of the judiciary among several high courts, instead of a unified high court, such as in the United States. Administrative courts render decisions on questions of judicial process and court procedure, while constitutional courts address matters of constitutional interpretation. Other countries include a separate category of high courts of appeals, which hear appeals arising from the lower courts in the country. The varying responsibilities of these courts influence the power and prestige of the institution. Administrative high courts, for example, are sometimes bureaucratic positions staffed by career civil servants trained to serve on such courts. Constitutional courts are often considered to be the most prestigious high courts because they address questions that have a significant impact on the country, and they are often staffed by well-known legal academics.Footnote ¹

In addition to the type of court influencing its power, country characteristics can also influence the prestige of the institution. Courts often wield more power in countries with a higher gross domestic product, and where the executive is weaker (Herron and Randazzo Reference Herron and Randazzo2003; Smithey and Ishiyama Reference Smithey and Ishiyama2000; Reference Smithey and Ishiyama2002). Additionally, courts have greater strength relative to the other branches of government when there is a stronger tradition for democracy and the rule of law (Schwartz Reference Schwartz1998) and where courts enjoy greater support from the public (Gibson, Caldeira, and Baird Reference Gibson, Caldeira and Baird1998).

As Lee Epstein, Jack Knight, and Olga Shvetsova (Reference Epstein, Knight and Shvetsova2001) note, the characteristics of the judges serving on cross-national courts are related to the independence and power of those courts. The structure of the court within a system of separated powers can make the court more or less prestigious, affecting the willingness of people to serve on such courts. Courts vary in their requirement of experience, the pool from which they draw members, and the way that members are selected, all of which affect aspirants for the high courts (Epstein, Knight, and Shvetsova Reference Epstein, Knight and Shvetsova2001; Ginsburg Reference Ginsburg2001). The prestige of the institution often affects the path people take to the judiciary, with transnational courts serving as a stepping-stone to national high courts (Kenney Reference Kenny2002).

The power and prestige of the courts within a country not only affect the professional backgrounds of those serving on high courts but can also influence the number of women serving within the institution. One of the few scholars to study the representation of women on these courts, Sally Kenney (Reference Kenney1998/1999; Reference Kenny2002), argues that understanding women's representation on judicial bodies is essential to comprehending the legitimacy of the decisions made by these bodies and the need for public accountability. While her focus is on transnational judicial bodies, the same argument can be made of high courts within nation-states. As Kenney (Reference Kenney1998/1999) finds, the length of term, the methods of selection, the influence of transnational bodies like the European Parliament, and the pool of female judges available for promotion all affect the number of women on the European Court of Justice.

Historically, the underrepresentation of women in governmental bodies has been blamed on lower levels of women's participation in public life generally. However, as women entered into professions, their representation in government did not proceed at the pace expected. While judicial office often requires additional education or specialized training, these requirements serve to limit the pool of potential candidates without limiting the number of women eligible to serve, especially as women increasingly participate in this type of training (see, for example, Anasagasti and Wuiame Reference Anasagasti and Wuiame1999). The lack of priority given to women's representation, persistent gender stereotypes, and systematic barriers to participation all work to limit women's representation in government, especially on courts (Linehan Reference Linehan2001).

Concerns about women's participation, initially raised by women's groups within nation-states, caught the attention of the European Community, prompting a self-study of women's representation on courts in European Union countries (Anasagasti and Wuiame Reference Anasagasti and Wuiame1999). Using a survey of those in the legal community, the authors found that 50% of the judiciary in European Union countries was held by women, but the result was largely driven by their overrepresentation on less prestigious lower courts. Additionally, the researchers found that the lack of clear criteria for participation on the courts and the perception of unfair criteria were depressing women's participation. It is interesting to note that the perceptions differed significantly for women and men, and that there was a difference between the sexes in the proposed need for policies (such as quotas) to increase women's participation on courts.

WHY WOMEN'S PARTICIPATION ON COURTS MATTERS

While there are several similarities between legislative and judicial office, it is perhaps more important that we understand women's representation on courts because courts are the least democratic institution. The lower the number of women on the courts within a country, the less tied to the institutions of government women are likely to be (Pitkin Reference Pitkin1972). This is especially true of courts where there is no electoral mechanism for holding decision makers accountable (Phillips Reference Phillips1995). Litigants who are dissatisfied with the decisions of judges have little recourse beyond appealing the case. The costs of such appeals can be prohibitive, and in some countries individual litigants do not have standing to go before their high court seeking justice. Thus, the individual members of the judiciary, and the decisions they make, become all the more important.

The importance of the decisions of individual judges, then, makes the characteristics of those judges, and how they are chosen, all the more important. Scholars of the U.S. judiciary note the importance of the individual attitudes of judges, as well as their personal traits and characteristics (Segal and Spaeth Reference Segal and Spaeth2002; Tate Reference Tate1981; Tate and Handberg Reference Tate and Handberg1991), and there is support for such models of decision making outside the United States (Schubert Reference Schubert1977). There is some evidence that women decide cases differently from their male counterparts, or that case outcomes differ when women serve as judges, especially when issues of gender or women's rights come before the court (Allen and Wall Reference Allen and Wall1987; Boyd, Epstein, and Martin Reference Boyd, Epstein and Martin2007; Gruhl, Spohn, and Welch Reference Gruhl, Spohn and Welch1981; Gryski, Main, and Dixon Reference Gryski, Main and Dixon1986; Songer and Crews-Meyer Reference Songer and Crews-Meyer2000; Songer, Davis, and Haire Reference Songer, Davis and Haire1994; Walker and Barrow Reference Walker and Barrow1985), but other scholars find little evidence for a different voice for women (see, for example, Davis Reference Davis1992; Sisk, Heise, and Morriss Reference Sisk, Heise and Morriss1998; Westergren Reference Westergren2004). Scholars of courts outside the United States concur on the importance of individual characteristics on decisions by judges and argue that the exclusion of women suggests they lack the capacity for self-government (Kenney Reference Kenny2002). There is some evidence of an expectation that women will decide cases differently—an expectation based on the fact that the vast majority of women serve on courts dealing with human rights issues, including the rights of women (Linehan Reference Linehan2001). Conflicting explanations for women's representation and disagreement over the impact of these women suggest the need for further research examining this problem.

WOMEN'S PARTICIPATION ON HIGH COURTS

While the study of the effects of sex on decision making is important, such studies treat the women serving on these courts as predetermined without considering what factors influence the number of women judges. A key first step to understanding the effects of sex on decision making, and an area left unexamined by scholars of both the judiciary and gender, is to consider macro-level factors affecting the number of women who serve on these courts. Attempting to assess the impact of sex on decision making while ignoring macro-level factors that influence the number of women on the bench is tantamount to selection bias, and may produce biased estimates of sex on decision making. We seek to remedy this problem by considering how the characteristics of courts, and the country in which they sit, affect the representation of women.Footnote ² To understand this question, we examine the representation of women on the high courts in OECD countries. We define high court here as any constitutional, administrative, or appellate court of last resort in the country, meaning that for some of the countries in the OECD, we have more than one high court. We include all high courts where more than one exists in the country. To determine women's representation on these courts, we examined the Web pages of the court. If the information was not available (or was unclear), we contacted the administrative office for the court or the Ministry of Justice.Footnote ³

We chose to study the OECD for a number of reasons. First, the OECD offers a significant amount of variation on the key variables of interest to this research, perhaps best exhibited in the table of women's representation (Table 1). While the representation of women clearly varies, so do key independent variables, such as a tradition of women's participation and level of development or economic prosperity. Unlike countries outside the OECD, the range for such variables is not so large that it will bias the estimates in a relatively small sample. However, if these variables affect women's representation here, they will also likely affect representation outside the OECD where the variables range more widely. In addition to the variation offered on the dependent and independent variables, the OECD is a group of countries on which it is relatively easy to gather data, making it a useful group for a first look at women's representation. The accessibility of the data on this group of countries is undoubtedly what makes it a common sample for analysis within the discipline.

Table 1. Women's representation on courts in the OECD

Table 1 examines the current variation in women's representation on high courts in the OECD countries. There appears to be a significant amount of variation in the representation of women both within court type and across court type. Administrative courts appear to have the greatest percentage of women serving, which is not terribly surprising given that most of the people serving on these courts come from the civil service. However, the presence of women on these courts should be studied with caution, as only five countries in the OECD have these courts. Constitutional courts appear to have the fewest women serving. The amount of variation across courts is also worthy of mention. Turkey, for example, has the lowest representation of women serving on its High Court of Appeals, 0%, and the highest representation of women on its Council of State, 60%. The minimum and maximum values of the dependent variable are thus represented by Turkish courts. The mean percentage of women serving on any court is 23%, and the standard deviation is 14. The huge range for the data, as well as the high variation, suggests that a more in-depth explanation is necessary.

In order to conduct a multivariate analysis of the factors affecting women's representation, we gathered information regarding the characteristics of the court, as well as information regarding the characteristics of the country (all demand-side explanations for women's representation). Regarding court-specific characteristics, the first factor likely to affect women's representation on high courts are the opportunities they have to serve on such courts, or the number of seats that exist on those courts. Not only are seats a measure of the opportunity that women have to serve on a high court, but seats can also serve as a proxy for the prestige of the court (Epstein, Knight, and Shvetsova Reference Epstein, Knight and Shvetsova2001). The more seats on the court, the less prestigious the position is, and lower-prestige institutions tend to see greater representation for women.

Other measures of court prestige commonly used in the literature on comparative courts include whether the judges on the court serve in life terms and if there is a mandatory retirement age for judges (Epstein, Knight, and Shvetsova Reference Epstein, Knight and Shvetsova2001; Herron and Randazzo Reference Herron and Randazzo2003).Footnote ⁴ Judges who serve for life are more independent of the other institutions of government, and thus have more institutional power, making the court more prestigious than those with renewable or fixed terms. Mandatory retirement requirements are used to protect the quality and distinction of the bench (Epstein, Knight, and Shvetsova Reference Epstein, Knight and Shvetsova2001).Footnote ⁵ Thus, we include dichotomous variables for courts with life terms and for those with mandatory retirement ages for judges.

In addition to the number of seats and other measures of prestige, the method by which judges are selected can affect women's representation, with appointers, such as presidents, seeking to gain electoral advantage by diversifying the bench (Slotnick Reference Slotnick1984).Footnote ⁶ Outside the United States, there is likely to be the same benefit for the person with the power to appoint high court judges, especially if he or she is a unitary actor. High courts in the OECD select judges through a wide variety of methods, from selection by the minister of justice to election by the national legislature or appointment by the executive. While there is a significant amount of variation in selection, one category, presidential appointment, is the most common (43% of cases). Thus, we include a variable measuring whether or not the president has the formal power to make appointments to the high court. All other selection mechanisms are the excluded category.Footnote ⁷

The characteristics of the country, like the characteristics of the court, are also likely to influence the representation of women. Especially important is the country's acceptance of women's participation in public life. We control for this factor in three ways. First, as Escobar-Lemon and Taylor-Robinson (Reference Escobar-Lemmon and Taylor-Robinson2005) note, the participation of women in one institution of government can have a spillover effect for another institution. Thus, we include a measure controlling for the representation of women in the national legislature. Secondly, countries that have quota laws have taken affirmative steps to increase the participation of women in public life. Thus, we include a dichotomous variable for whether there is a quota law in the country.Footnote ⁸ Third, because women's participation in public life is a tradition likely to build over time, we include a measure of the years since women were given the right to vote in the country. We expect to find that the longer the amount of time since women were given the right to vote, the more women serving in public life.

In addition to the measure of women's participation in public life within the country, we also include three control variables for country-specific characteristics likely to affect women's participation, as well as the functioning of courts more generally. We include measures to control for the participation of women in the labor force, the fertility rate, and the gross national product per capita.Footnote ⁹ We expect that the greater the percentage of women participating in the labor force, and the lower the fertility rate, the higher the political participation of women. Women will already have experience working outside the home in other areas, making participation in the legal profession also more likely. Fewer children will also make it easier for women to participate in public life. Wealthier countries are also more likely to see women's participation; as wealth increases, women's educational and career opportunities also increase.Footnote ¹⁰

To estimate the effects of these independent variables on the number of women serving on high courts, we include two negative binomial analyses.Footnote ¹¹ The negative binomial models are appropriate for count data such as these. Included in each model is a measure of the number of seats on each court (discussion follows) that prevents a high number of seats (and thus a high number of women) on one court from inflating the results. Tests were conducted to measure the dispersion of the data and verify the appropriateness of the negative binomial model. For both models we reject the null that alpha = 0 (the null that the data are equi-dispersed). Instead, the data showed evidence of overdispersion, suggesting that a woman holding one seat on the court influences the next seat being held by a woman. Both the tests for equi-dispersion and the significance test for the model suggest that the process is measured properly.Footnote ¹²

Table 2 shows the results of two models examining women's representation on courts. Because of the variation in the bivariate results across the courts (see Appendix B), the first model includes dummy variables for the different types of courts, with a unified high court as the baseline. The second model shows the results without the court dummies. The results are not substantially different across the two models, and both models are statistically significant. The higher the number of seats on the court, the presence of quota laws, and the more years since women were given the right to vote, the higher women's participation on the high courts. Countries where the president appoints judges to the bench also increase the representation of women on the courts, but only in the first model. In order to determine the impact of these variables, predicted counts were estimated.Footnote ¹³

Table 2. Negative binomial models of women's representation on high courts

* * = p < .10, ** = p < .05, *** = p < .01

Model 1 Significance: Wald chi2(13) = 127.96, Prob > chi2 = 0.0000

Model 2 Significance: Wald chi2(10) = 102.10, Prob > chi2 = 0.0000

Model 1 Likelihood-ratio test of alpha = 0: chi2(01) = 5.65 Prob > = chi2 = 0.009

Model 2 Likelihood-ratio test of alpha = 0: chi2(01) = 9.61 Prob > = chi2 = 0.01

N = 38 countries (There are 47 courts in this data set, 38 of which have complete information, for a total of 1,013 seats on these courts.)

Table 3 shows the predicted counts for the two models. The baseline model predicts an average of three women serving on the court for the first model. Increasing the number of seats one standard deviation increases the number of women serving on the court to nine. A court with five seats, holding all other factors constant, has a predicted count of two women judges.Footnote ¹⁴ Countries without quota laws see an average of one woman fewer serving on the court, suggesting that the political climate for women does affect women's representation on high courts. In countries where the president appoints the members of the high court, five women serve on the bench, an increase of two. Finally, the years since women were given the right to vote also affects women's representation on the courts. A one standard deviation increase in the years since suffrage increases the number of women on the court to four, while a decrease of one standard deviation reduces the number to two.

Table 3. Predicted counts of women judges

While the predictions for the second model are not significantly different from those of the first model, some of the variables lose their statistical significance when the court dummy variables are eliminated. The number of seats, quota laws, and the years since women attained the right to vote all remain significant in the second model. Increasing the number of seats by one standard deviation increases the number of women serving on the court from a baseline of two to seven. Reducing the number of seats on the court to the minimum of five decreases the number of women serving on the court to one. Countries without quota laws see a modest decrease in the number of women serving; about one woman fewer serves on the courts in these countries. Finally, increasing the number of years since suffrage by one standard deviation increases the number of women serving on the court to three, while a decrease of one standard deviation does not substantially change the number of women serving.