The Core Behind Critical Mass Hypotheses

The belief that women politicians will have a substantive effect on political decision making is found within debates about the “politics of presence” (Phillips 1995), of which critical mass research can be seen as one part. In these debates, the messenger as well as the message is seen as important (Catt 2003), and women politicians are seen to be not only “standing as” women but also “acting for” women as a group once elected (Lovenduski and Norris 2003; Pitkin 1967). This assertion is based not on a belief in an essential link between sex and representation but on the way in which women experience the world and how this affects their actions if elected as political representatives. As Melissa Williams (1996, 106) puts it: “The representative who is capable of acting as an advocate for women's interest must have some understanding of the ways in which the lives of her constituents are shaped by the privilege of men, and the most effective starting-point for that knowledge is the fact of her own experience of exclusion and subordination.” Female politicians themselves have noted the importance of gender identity within the political realm. For example, during parliamentary debates on parental tax credits, New Zealand Member of Parliament (MP) Christine Fletcher noted (“Taxation (Parental Tax Credit) Bill” 1999, 16695) that: “there is a greater number of women in Parliament, and that allows us—as we approach the new millennium—to finally begin to debate some of the issues, which I see as the hard issues.”

The expectation that female politicians will represent women in political debates and decision making is not without problems. As Drude Dahlerup (1988, 279) notes, women as politicians are caught between two conflicting expectations. They have to prove they are just like male politicians and that they will make a difference when elected. Even without this double bind, there are problems if women politicians are seen to represent only “women's interests” (or perhaps “feminist interests”). The substantive representation of women is further complicated as women are by no means a coherent group and every female legislator will have cross-cutting identity characteristics that affects her worldviews. Despite these concerns, it is important to ascertain whether women have a “voice” or “voices” in the national legislature, whether this voice comes via female legislators, and what role numerical strength plays in enabling politicians to act as and for women.

Rethinking the Critical Numbers

The concept of critical mass came to the fore in political science after the publication of Dahlerup's 1988 article “From a Small to a Large Minority: Women in Scandinavian Politics.” On the basis of the idea that the form of a legislative assembly will shape the process and policies of that body, Dahlerup drew on a 1977 study by Rosabeth Moss Kanter of the interactions in groups composed of people of different cultural categories or statuses. Kanter (1977, 966–67) had presented a typology of four group types in order to investigate the effect of changing group dynamics on organizational culture. From her work, two group types have emerged as the most important in critical mass debates—the skewed group, where the minority constituted a maximum of 15% and are “tokens,” and the tilted group, in which the minority has between 15% and 40% membership and is “becoming strong enough to begin to influence the culture of the group.” While Kanter's article is central to the concept of critical mass, her proportions only rarely appear in critical mass inquiries, with Dahlerup's suggestion of 30% as the point of critical mass making it into both political science and into the political imagination of many nations (Childs 2004; Grey 2002; McAllister and Studlar 2002; United Nations Economic and Social Council 2004).

There is little evidence that 30% is a magical cure-all for ensuring the representation of women in national politics; authors (including Dahlerup) have suggested that the most important thing for the substantive representation of women in politics is “critical acts” (Dahlerup 1988; Lovenduski 2001) or “safe spaces” (Childs 2004). I suggest that given there is nothing in Kanter's work to indicate that a single figure is all-important, we need to move to the idea that different critical masses may be needed, depending on the outcome sought. Gaining 15% of the seats in a political body may allow female politicians to change the political agenda, but it may take proportions of 40% to have women-friendly policies introduced.

There is one further problem with the use of numbers in researching the substantive impact of women politicians that needs attention before moving on. It is difficult to explore fully whether critical masses are needed to ensure the substantive representation of women, due to the very lack of women in most democratic legislatures. This point can be illustrated by looking to a comparative study of 20 Organization for Economic Cooperation and Development (OECD) nations carried out by Ian McAllister and Donley Studlar (2002). While I agree with McAllister and Studlar that more caution is a needed regarding the concept of critical mass, I also note that few of the democracies in their study had anything near 30% women in their legislative ranks, let alone 40%.

Even if we recast the proportions at the heart of critical mass inquiries and allow for different critical masses to be considered as points of takeoff, however, it is likely that McAllister and Studlar (2002, 247) are right: The impact of critical mass has been inflated. What existing investigations of critical mass have signaled is that we need to look at developing a clearly articulated joint-effect model that more accurately reflects the factors that aid (or hinder) female politicians looking to act as and for women.

Rethinking the Critical Mass Expectations

Existing critical mass investigations have tested Dahlerup's hypotheses around changes to the parliamentary culture (Broughton and Zetlin 1996; Grey 2002; Kathlene 1994; Lovenduski and Norris 2003; Thomas 1991); feminization of political agendas and parliamentary debates (Broughton and Palmeri 1999; Freedman 2002; Sawer 2004; Thomas 1991); changes to legislation (Childs 2004; Thomas 1994); and reactions to women politicians (Kathlene 1994). Few have confirmed a straightforward causal link between female politicians reaching a critical mass and the substantive representation of women. From this body of work four factors seem most important to investigate alongside the role that critical masses of women have in bringing change to political process and outcomes: the position of the female politicians in question; their time in office; both their own and their political party's ideology; and finally, the reactions of, and to, the women politicians.

From the body of critical mass research, there is also an indication that researchers should not look merely for positive impacts occurring as the numbers of women in a legislature increase. Take, for example, Dahlerup's (1988, 283–87) hypothesis that there will be a change in the reactions to women as politicians, both inside and outside legislatures, once their numbers reach a critical mass. The reaction to rising numbers of women in powerful positions may be hostile (see Kathlene, 1994); this could, in turn, impact the behavior of elected women.

New Zealand provides a natural case study from which to begin development of a joint-effect model into the factors behind the substantive representation of women in politics, for a number of reasons. First, the number of women in the national legislature has been relatively high for nearly a decade, with women occupying around 30% of seats in the House of Representatives since 1996. There have also been relatively high proportions of women in three successive government executives (in 1999, women held seven of the 20 cabinet positions; in 2002, six out of 20; and in 2005, five out of 20—the number of women in the cabinet includes Prime Minister Helen Clark). The New Zealand case also allows researchers to look at whether the rules of the game influence group representation, due to the change in electoral systems in 1996 from first-past-the-post (FPP) to mixed-member-proportional (MMP) representation.

To begin development of a joint-effect model, I draw on research that evaluates the substantive impact of New Zealand women politicians since 1970 through textual analysis of parliamentary debates on three topics: child care, pay equity, and parental leave. These issues were chosen because empirical research has shown that gender differences in politics are more obvious on women's issues or issues to do with women's autonomy (Lovenduski and Norris 2003; Norris 1993; Reingold 1992). The textual analysis allowed evaluation of occasions during political debates when women politicians stated that they were acting as or for women.

A long time frame for research into the substantive representation of women is necessary, despite a common misconception within critical mass inquiries that there will be an immediate takeoff into a new situation once women reach a critical mass. What the critical mass hypothesis calls for is detection of an irreversible takeoff into a new situation. The New Zealand case highlights that where human interactions are concerned, time may be a crucial factor in gaining lasting change. Many women-friendly policies in New Zealand were passed almost 20 years after women first occupied 15% of the seats in the national legislature.

Textual analysis of parliamentary debates has limitations. Actors in Westminster-style parliaments are bound by standing orders (written rules) and by unwritten conventions. In addition, much of the activity of MPs is performed outside the debating chamber, in parliamentary committees, in their geographical constituencies, and in the community. Nonetheless, textual analysis provides a manageable way to undertake analysis across a long time frame, and provides a window into the actual behavior of politicians. Numerous studies have used survey results to show that female MPs hold substantially different views than their male colleagues, and infer that this will lead to different political behavior. While it is important to discuss politicians' motivations, surveys gauge the subjective attitudes of MPs rather than objective behavior within the bounds of legislative assemblies. As Joni Lovenduski and Pippa Norris (2003, 97) note in their study of the politics of presence in Great Britain: “without independent verification, self-reported claims expressed during interviews that women politicians will prioritise women's interest and concerns more than men cannot be accepted at face value, any more than we would accept without demonstrable evidence any claims that Labour MPs speak for and defend the interest of the poor, or that Conservative MPs represent the business community.” These concerns are validated by the work of Beth Reingold (2000), who found that legislative behavior of women politicians often did not match their expressed interest in women's issues. Textual analysis of parliamentary debates allows for independent verification of women politicians' behavior in parliamentary debates, even if these actions are constrained by the institutional rules of the legislature.

Methodologically, another issue arises from using textual analysis as the tool of inquiry, given that the very way issues are categorized in political debates affects how they are discussed and by whom. For example, in New Zealand, parliamentary debates on child care are seen as being part of the education portfolio, while parental leave is linked to the employment rights of women, a division which had an impact upon parliamentary debates. Between 2000 and 2005, men dominated parliamentary speaking time on the topic of child care, while female MPs dominated parental leave discussions. This limitation must be borne in mind when trying to unravel whether female politicians have acted as and for women.

Numbers, Roles, and Ideologies

Textual analysis of New Zealand parliamentary debates provides no evidence that there is a single “critical mass” that brought about wide-sweeping feminization of political processes and outcomes. Changing numbers of women in politics, however, do at times correlate to changes in the political agenda and in the reactions of legislators. Once female MPs move from being a token to a minority group in the New Zealand House of Representatives in 1984, there is an increase in the level of debate on issues such as child care and parental leave, and female politicians claim a greater stake in these debates. These findings on changes in the political agenda echo the results of studies from the United States, Europe, and Australia, though the threshold for change varies, ranging from as low as 15% to 30% (see Freedman 2002, 186; Sawer 2000; Thomas 1991). But are the agenda changes evident in New Zealand and in other democracies due to rising numbers of women in the national legislature alone?

Researchers have long focused their attention on the numerical strength of women in national legislatures as the linchpin in improving the substantive representation of women, and in doing so have ignored the complexity of power relations in politics and how this affects the ability of female politicians to act as and for women. Women's numbers and their roles in government executives and in political parties are likely to affect any attempts to represent women as a group. The need to look at women's numerical strength within political parties is particularly true of strong party systems like New Zealand.

In the New Zealand political debates analyzed, politicians from both left- and right-leaning political parties made speeches in which they claimed to speak as a woman or on behalf of women; however, women of the left-leaning Labour Party made more gendered claims than their conservative colleagues. And it is during Labour government terms that there is evidence of an increase in debate on women's issues and the adoption of policies advancing the autonomy of women (such as paid parental leave).

Overt gender recognition by New Zealand's left-leaning women politicians and their support for women-friendly legislation is likely to be related to a number of factors, including the links that these women had with feminist organizations. As McAllister and Studlar (2002, 248) note, the substantive representation by female politicians requires an increase in feminist attitudes in the legislature (not just a rise in the number of female politicians). The ideologies of the major political parties are also likely to have affected the way female politicians talked about women in society. The Labour Party has traditionally been aligned with social democratic principles and with labor unions (although this relationship was strained in the 1980s when the Fourth Labour Government pursued neoliberal reforms of the economy and state sector). The National Party is, on the other hand, historically associated with liberal ideals and with the business and farming sectors. And as National MP (and former Prime Minister) Jenny Shipley (“Human Rights Amendment Bill” 2001, 12996) noted: “we [the National Party] do not believe in collectivism.”

Acknowledging the role of party ideology and personal beliefs on the activities of women politicians, in the New Zealand case, the numbers of women in the two major parties' parliamentary caucuses also appear to have affected the attitudes of female legislators. Once Labour women MPs occupied minority status (over 15%) in their own party's parliamentary team, they frequently commented about the “team spirit” this engendered (Helen Clark, “General Debate” 1988a, 3643; Anne Fraser, “General Debate” 1988b, 5812). What we see is the creation of a safe space for Labour Party women, a safe space bolstered by numerical strength and the ideological leanings of both the female politicians and their party of choice. While National's female MPs set up formalized women's caucus meetings once their numbers in parliament reach eight (Shipley cited in Baysting et al. 1993, 99), their numbers have never been as high as those of their Labour counterparts, and there is no evidence from the debating chamber that National Party women saw themselves as a team.

The change in the reactions of the Labour women politicians is not the only such change seen in New Zealand as the number of women in the legislature rises. The increase in women as legislators in New Zealand has coincided with a rise in statements by politicians (particularly male politicians) expressing hostility toward feminism. However, as with other critical mass hypotheses, there is no indication of a single critical mass figure in the New Zealand House of Representatives that has led to this backlash toward feminist ideals. The backlash came at a time when women occupied 30% of both the House of Representatives and cabinet, and when a substantial number of women who frequently identified as feminists held prominent positions in the ruling Labour Party, including the post of prime minister.

This rise in hostility does appear to have consequences with regard to the substantive representation of women, with the textual analysis showing a drop in overt claims by New Zealand's female politicians that they are “acting for women” during the last five years of debates scrutinized (2000 to 2005). The backlash against feminism appears to have made female politicians (and particularly women in the Labour Party parliamentary caucus) reluctant to speak for women in parliamentary debates since 2000. This result is similar to the findings of Lyn Kathlene (1994, 573): In the U.S. legislatures she scrutinized, the more women who served on a committee, the more silenced women became, revealing the powerful impact of gender socialization.

The backlash to feminism, and the apparent reluctance to act as and for women in the following legislative debates, are not the only negative impacts that followed the rise in the numbers of women in New Zealand politics. Far from leading to increased claims about women's collective needs, the increased diversity in the debating chamber (ushered in by the shift to a proportional representation system in New Zealand) has coincided with claims by many female MPs that they represent groups other than women or that they speak for subgroups of women. This diversity diluted the strategic essentialism of earlier debates. This finding around the increased diversity in the groups that female politicians claim to represent echoes the work of Beth Reingold (1992) and of Sue Thomas and Susan Welch (1991). This change, however, may not be seen as a negative consequence by all, for as Thomas and Welch (1991, 453) suggest, what is occurring is that women are no longer tokens, assigned to fill the women's position on certain committees, but are free to serve on a wider variety of committees.

Conclusion

A joint-effect model into the substantive representation of women in politics would consider how the numbers of women in political parties and government executives, the positions of these women, and the reactions to their presence (both their own and others) impacts upon the ability and willingness of women politicians to act as and for women. The single case study of New Zealand indicates that women politicians have more readily acted as and for women when they have a team (of sufficient size), whose members have feminist leanings, and when they find themselves in a general environment supportive of feminist ideas. However, there is a point at which a new set of inhibitors, some linked to numerical strength, takes hold. The substantial rise in the numbers of women in the New Zealand parliament after 1996 acted to diffuse the strategic essentialism used by female politicians in earlier debates to advance women-friendly policies. This increased female presence in both the legislature and in government executives also correlates with a rise in overt hostility toward feminist agendas. Nonetheless, New Zealand is only one case, and there needs to be more work developing a joint-effect model through comparative work. The difficulty, ironically, is finding enough countries where women make up significant numbers in the legislature, political party caucuses, and in government executives—and all this over a significant time frame.