The theory of a “critical mass” seems to live a life of its own, in spite of reservations expressed by researchers, my own reservations included. In debates about women's political representation, the importance of a certain number or percentage of women in political assemblies is often stressed. According to conventional wisdom, research has shown that it takes a certain minimum representation, for example, 30%, before the minority, here women, are able to make a substantial difference in politics.
In this essay, I examine the theoretical foundation and underlying assumptions of the critical mass theory. Already in the 1980s, the idea of a specific turning point, for example, 30% women, was rejected. Later the notion of irreversibility was questioned following an actual decrease in women's representation in, among others, the Eastern and Central European countries. I will, however, argue that while the theory has been contested in the scholarly literature, a story of the theory of a critical mass has played an important role for women politicians themselves, as well as for those advocating increased representation for women in politics, most recently for the many who advocate introducing electoral gender quotas. One may ask why 30% is the most widespread quota percentage.
Rejecting a specific turning point and the notion of incessant progress does not imply that numbers or proportions of minorities in political assemblies have no significance. More empirical research is needed on this theme. The theory of a critical mass is closely related to the discussion of whether “women politicians make a difference.” In this essay, the validity of this very question is scrutinized. It is finally argued that a distinction should be made between the policy outcome perspective and the politics as a workplace perspective when discussing the importance of the relative numbers of women (or other minorities).
The Critical Mass Theory of the 1980s
In the 1980s, the critical mass argument was primarily applied to situations where women had not reached 30% in parliaments or local councils. Consequently, the argument was that it was unrealistic to expect major changes until women's representation had reached a critical mass, because a small number of women in politics tend to be tokens.
Originally, the term critical mass was borrowed from nuclear physics, where it refers to the quantity needed to start a chain reaction, an irreversible turning point, a takeoff into a new situation or process. By analogy, it has been said that a qualitative shift will take place when women exceed a proportion of about 30% in an organization. In this way, the move from a small to a large minority is significant. Thus numbers, or rather percentages, count (Dahlerup 1988, 275–76).
Important points about minority behavior were taken from the organizational theory of Rosabeth Moss Kanter (1977). Even if she herself did not use the concept of a critical mass, her theory, as well as Helen Mayer Hacker's classic article on women's minority status in society at large (1951), gave inspiration to the discussion among feminist political scientists in the 1980s of applying a critical mass theory to our studies of women in politics.
In my 1988 article, I both constructed and critically reconstructed the critical mass theory. This article, which was based on the Nordic experience with over 30% women in parliament, concluded that no specific turning point can be identified, and that consequently the critical mass perspective should be replaced by a focus on critical acts that will empower women in general, for example, gender quotas. Dissatisfied with the contemporary, rather narrow, discussion of why women do not make more of a difference in politics, the article suggested broadening the research agenda by identifying six different aspects of changes that might follow from changes in the relative number of women and men (Dahlerup 1988, 283–84):
- Changes in the reaction to women politicians;
- Changes in the performance and efficiency of women politicians;
- Changes in the social climate of political life (the political culture);
- Changes in the political discourse;
- Changes of policy (the political decisions);
- Increase in the power of women (the empowerment of women).
The politics as a workplace perspective (points 2–4) was less developed before the 1980s because focus had been on the perspective of policy change (point 5). In a later section of this essay, I develop the point that the importance of the relative number of women seems to vary depending on the chosen perspective. However, systematic longitudinal research on changes of the more qualitative aspects of the gender structure of the Nordic or other parliaments has been and is still lacking to a large extent.
My 1988 article pointed to the difficulties in isolating the effect of the increasing proportion of women from what happens outside the parliament. Only when it comes to changes in the social climate of the political institutions (point 3), one can expect “a kind of ‘automatic’ change when the minority grows large.” Such changes occurred, however, even when the first women entered the meeting room, the argument went. Furthermore, I argued that with a large minority of women in politics, “the open resistance against women politicians is removed—now it seems hopeless to bring women back to the [home]” (Dahlerup 1988, 295–96). A rereading of my 1988 article makes evident that while the notion of a specific turning point, for example, at 30%, had been rejected by the 1980s, the idea of irreversibility, which matched the optimism of the 1980s, was not challenged until later, when we experienced actual examples of severe backlash in women's representation in the 1990s. With the rise of fundamentalism in many parts of the world, a backlash against feminism in general has become part of today's options, replacing the optimism of the 1970s and 80s.
The many references to my 1988 article in the literature reveal that it no doubt contributed not only to the scientific debate on the critical mass theory but also to the construction of the story of the critical mass theory, which, in spite of reservations in the scholarly literature, nevertheless has been influential, as the following section will show.
The Use of the “Story of the Critical Mass Theory”
The “story of the critical mass theory” refers to the actual use of the theory of a critical mass. When feminist movements and female politicians themselves make use of this “theory,” it becomes important in itself, in spite of all scholarly reservations. Again and again one hears, as part of conventional wisdom, that research has shown that there must be at least 30% women, a critical mass, for women to make a difference in politics.
The theory—or, as I would prefer to label it, the story of the theory of a critical mass—is frequently used to cover two different situations. First, it is being used by female politicians who are defending themselves against the critique, primarily brought forward by the women's movements, that women politicians do not make enough of a difference once elected. It is important to note that feminists, ever since the election of the first women local councillors and parliamentarians, have expressed this kind of critique.
Second, advocates of enhancing women's political representation have frequently used the critical mass argument. It is said that political parties should nominate not just the obligatory lone woman but at least a critical mass of women, because the voters demand it or in order to ensure that the elected women are not just a token few. During the last decade or two, we have witnessed the critical mass argument being used among advocates of the introduction of electoral gender quotas in order to effect a rapid increase in the number of female politicians (Dahlerup 2006). This section focuses on the use of the story of the critical mass theory in debates about introducing gender quotas for public elections.
In the present research project, “Electoral Gender Quotas—a Key to Equality?” (2003) which concerns the global trend to introduce gender quotas for public election, we often find the story of a “critical mass” (see Dahlerup 2006; Dahlerup and Freidenvall 2005; “Global Database of Quotas for Women” 2006). In the last 10 to 15 years, almost 50 countries have introduced gender quotas by law or constitutional amendment (legal quotas). In a larger number of countries, some of the political parties have voluntarily introduced gender quotas for the party's own list for public election (voluntary party quotas). Such quota provisions may target the pool of aspirants, the composition of the candidate lists, or the actual number of women to be elected, also called reserved seats (for types of quota systems, see Dahlerup 2006, 21, 294).
Electoral gender quotas come in many forms and use many different percentages or numbers, from Nepal's 5% to Sweden's and France's 50%. When it comes to candidate quotas, however, legal as well as voluntary party quotas, 30% is the most widespread percentage chosen (Dahlerup 2006; www.quotaproject.org). It is worth looking at why this is the case and whether there is a connection to the story of the critical mass. It should be noticed, however, that it is often difficult to trace the reason for the choice of percentage in the individual cases.
In 1990, the United Nations Economic and Social Council endorsed a target of 30% women in decision-making positions in the world by 1995. This target was far from met, since in 1995 only 10% of the world's parliamentarians were women, in 2005, only 16%, still far from one-third (www.ipu.org). Consequently, many countries have opted for fast track policies like quotas. During the 1990s, several international declarations recommended a target or a quota of 30%.
International recommendations are important because they can render legitimacy to national advocacy for more women in politics (Krook 2004). In many countries, advocates of gender quotas in politics refer to the UN Platform for Action, Beijing 1995, and to the Convention on the Elimination of All Forms of Discrimination Against Women (CEDAW). Although the Platform for Action does not employ the controversial word “quota,” the platform is nevertheless cited as supporting gender quotas in politics because it advocates “special targets and implementing measures … if necessary through positive action” (United Nations Fourth World Conference on Women [henceforth FWCW] 1995, Art. 190a). Similarly, the CEDAW convention of 1979 speaks of “special measures.”
A shift is taking place from demands for 30%, or minority representation, to a more radical demand for 50%, or gender balance in politics. But while, for example, the Platform for Action represents this new line in its demand for equal representation, it nevertheless continues to argue for securing a “critical mass” of women in parliaments (FWCW 1995, Art. 181–95).
The Socialist International and especially Socialist International women have been very active in advocating gender quotas in socialist parties all over the world. This organization played an important role in the quota debate on the Balkans, to mention one region (Antić and Lokar 2006). Even if its recommendations argue for “the assurance of gender-balanced representation in party structures” and “the preparation of more gender-balanced lists of candidates for elections of any kind” (Socialist International Women 1998), the demand for gender balance is, in most member parties, translated into a demand for 30% women. Very few socialist/social democratic parties employ 50% quotas, among them the Social Democratic Party in Sweden and PRI (Institutional Revolutionary Party) in Mexico, while some additional parties have 40% quotas for their electoral lists, among them the Social Democratic Parties in Croatia, Norway, Germany, and Spain. The most frequent quota among those member parties of Socialist International that have adopted voluntary party quotas for their electoral list and for the internal party structures is 30% (“Global Database of Quotas for Women” 2006).
In its 1997 “Declaration on Gender and Development,” the Southern African Development Community, SADC, made a commitment to achieve a 30% representation of women in politics and decision making by the year 2005. A 1999 conference, organized by SADC's Gender Unit and titled “Beyond 30 Percent in 2005: Women and Decision-making in SADC,” adopted a far-reaching program for gender parity in the region (Lowe-Morna 2004).
More research is needed at the level of individual countries in order to answer the question of how and why 30% became the most widespread quota regime. Latin America may serve as an example. Latin America is the leading region in the world when it comes to legal candidate gender quotas. Why have Argentina and, after Argentina, so many others of the quota countries in the region chosen 30% as their target? In her overview of electoral gender quotas in Latin America, Jacqueline Peschard writes: “Quotas result from [sic] the greater presence of women, and are useful for encouraging a larger mobilization along these lines, in order to constitute a ‘critical mass,’ which is believed is reached with about 30 per cent representation” (Peschard 2003).1
Here, J. Peschard refers not to Latin American experiences but to the Spanish translation of my 1988 article (Dahlerup 1993).
Even if the story of the critical mass no doubt has been important for the new trend to introduce electoral gender quotas and for the preference for 30% quotas, the occurrence of other motives cannot be excluded—such as increasing the number of women in order to look modern and democratic or simply in order to avoid the more radical demand for real gender balance as those expressed in the 50/50 campaigns of the Women's Environment and Development Organization (WEDO 2005).
The wish that women will “make a difference” in politics is the agenda of the women's movements. Generally speaking, party leaderships recruit women not to get a different politics but to gain votes. Once elected, most parties want “their” women members of parliament (MPs) to work as loyal party representatives. But for the women's movement this question has been crucial.
The Validity of the Question of “Making a Difference”
The concept of a critical mass is closely related to the debate about whether women in politics will “make a difference.” Even if we limit the discussion to changes in public policy (see point 5 in the first section), we are still left with the fundamental problem that making a difference is in fact a very diffuse concept. Within the broad range of women's organizations, we will not find any agreement about what making a difference implies. At play in this discussion are concepts like “women's issues,” “feminist issues,” “gendered issues,” and “women-friendly policies” (Beckwith and Cowell-Meyers 2003); “accountability to women,” “gender-sensitive policies,” “women's political effectiveness” (Goetz 2003); “strategic gender interests” (Molyneux 2001); and “feminizing politics” (Lovenduski 2005). When we move from the theoretical discussions of “politics of presence” and of “descriptive” versus “substantive” representation to empirical analysis of the relation between the proportion of women (or any other minority in politics) and “making a difference,” we are still in need of a more precise definition of the dependent variable, the notion of making a difference, as Beckwith and Cowell-Meyers (2003) and many others have also argued.
The Difference Fallacy
Methodological problems are also involved, in what I label “the difference fallacy.” It is often an underlying assumption in empirical studies that female politicians have proven to make a difference in politics, if the researcher is able to identify some difference between women and men (for instance, differences in attitudes, issue priorities, the number of petitions on certain subjects, the way of speaking in politics, and contacts with women's organizations). Often this research has, in fact, been able to show various differences between politicians on the basis of gender. For instance, there is enough evidence to make it possible to conclude that women politicians all over the world tend to be more active than are their male colleagues when it comes to placing equality policy on the political agenda (Goetz and Hassim 2003; Lovenduski 2005; Skjeie 1992). This type of research also reveals that not all women agree and that party affiliation is extremely important in most political systems, even if some gender differences remain after controlling for party affiliation.2
It should be noted that such differences in attitudes and actions in this type of research are not seen as an essentialist, sex-specific feature, but rather as historical constructs, with large differentiation according to time and location.
Making a difference is not the same as staying different. The difference fallacy implies underestimating the obvious possibility that women, as politicians, perhaps especially when there are many of them, have been able to influence their male colleagues and thus change either the overall political agenda or the agenda of their individual parties (or both), with the consequence that only few gender differences in attitudes and actions can be identified. Since cross-party cooperation among women MPs is usually considered suspicious, female politicians will see it as a success if they have managed to place new issues, like violence against women or child care, on the political agenda in consensus with their male colleagues in parliament.3
In one-party systems or in political systems with one dominating party, cooperation among women parliamentarians or the establishment of a women's caucus in parliament is not considered illegitimate to the same extent as in multiparty party systems. Women's joint actions in Rwanda and within the African National Congress are examples of this.
Even if researchers should choose to abandon the very question of whether women make a difference, Marian Sawer rightly states that we cannot expect campaigns for more women in politics to give up on the “making a difference” discourse (Sawer 2000, 377). The justice argument was never the only argument put forward by women's movements, which always have envisioned that women would change politics for the better for women and for humanity at large.
Reformulating the Hypothesis on Minority Representation— Do Numbers Matter?
Having rejected the idea of a specific turning point, a takeoff into a new situation when the minority increases, say, to over 30%, and having rejected the notion of irreversibility, what is left of the theory of a critical mass? The answer is that we need more specific, empirical-based research about the importance of the size of the minority under different conditions.
It seems crucial to make a distinction between the following two research questions: First, does the proportion of women in political institutions make a difference for the content of the political decisions—the policy outcome perspective; and, second, does the proportion of women in political institutions make a difference for the possibility that female MPs will perform as representatives the way they want—the politics as a workplace perspective? These questions correspond to points 2–4 in the first section list. The overall research question is that of the importance of the relative number of the minority, everything else being equal. Larger questions, such as the importance of women's representation for the legitimacy of democracy, are omitted from this more specific discussion.
The relative number of women as a minority in political institutions is more important, I argue, regarding the second question concerning politics as a workplace. With this question, we return to Kanter's organizational perspectives. My formulation of this question deliberately avoids linking the question of effectiveness of women politicians to the pursuit of a feminist agenda. It deals instead with the importance of women's relative number for their ability to become effective in their work, to perform their tasks as politicians the way they individually prefer, in spite of being a minority inside and outside parliament.4
While Goetz's concept of “women's political effectiveness” includes voicing political issues of concern to women (Goetz 2003, 29), my concept of effectiveness is open to any kind of agenda that the individual or group of female politicians wants to pursue, including nonfeminist goals.
Concerning the first question, I argue that number and percentages are of only minor importance for the policy outcome. Even a few women in politics, under the right circumstances, can make a big difference, while a large minority of women parliamentarians may not wish or be able to change the political agenda in a certain political system at a certain time. At any rate, numbers interrelate with other factors, and consequently, isolating the effect of sheer numbers is almost impossible.
It may be that the attempts to link the relative number of women to policy outcomes are largely misplaced in research. As Joni Lovenduski points out: “[F]eminizing politics is like many other political processes” (2005, 180). I argue that the adoption of feminist public policies depends largely on the same type of factors that we study concerning other political issues, like environmental policies or social policies: the political context, state feminist machineries, prevailing discourses, framing of the issue, coalition building, and movement strength, among others. The number of women in parliaments is probably not the most crucial factor, even if numbers, of course, are important for any parliamentary majority building.5
The Danish case illustrates that numbers are of less importance than the political context. Even if the relative number of women in the Danish parliament increased gradually over time, there were more feminist women parliamentarians in the 1980s than in the 1990s. Now, experienced feminists have begun to keep a low profile in parliament, and young women politicians have been eager to stress that they were not feminists. In a comparison with Sweden, a country that did not experience such a backlash in the 1990s, I explain this remarkable and recent differentiation between the two neighboring countries with the variations in the strength of their feminist movements and of state feminism (Dahlerup 2004).
This reasoning, however, does not imply that increasing women's political representation is unimportant for issues of citizenship, democracy, and development. The new trend of introducing gender quotas in postconflict societies often rests on discourses of the importance of the inclusion of women for the development of democracy (Dahlerup 2006). For example, “GEMSA is aware that the mere representation of women in politics is not equivalent to gender transformation. However, we also firmly believe that transformation cannot begin to take place when over half the population is effectively excluded from decision-making!” (Gender and Media 2006).
