Representation, Electoral Systems, and Women

Although Hanna F. Pitkin develops a complex interpretation of political representation in The Concept of Representation (1967), more often works on political representation of women are only two-dimensional, limited to descriptive and/or substantive representations. This first conjures up the mirror concept: A legislative assembly is said to be representative if its composition reflects the whole (that is, the population). Substantive representation stresses and gives importance to what an elected individual thinks and does, rather than emphasizing who the person is; it requires deliberate actions. If political representation refers to normative theory and concepts, the way it is translated in practice depends on complex cultural and institutional settings specific to each country. One of the tools to operate the change from political representation as normative theory to political representation as practice is the electoral system.

Basically, there are three main types of electoral systems: majority (and plurality), proportional, and mixed systems. Each system conceptualizes representation differently. In PR, representation rests on the composition of democratic assemblies in order for them to reflect the characteristics that make up society (this reading is closer to Pitkin's conception of descriptive representation). In majority systems, representation is ensured by the representative and by her decisions in relation to the defense and promotion of interests associated with her electoral district (this is Pitkin's conception of substantive representation). Mixed systems borrow the idea of the political actor personifying the role of representation put forward by the majority systems with single-seat constituencies, and the idea of diversity and proportionality proposed by the PR with multiseat districts.

Analyzing the proportion of female legislators according to these three types of electoral systems makes it possible to reveal a relatively convincing relationship between electoral systems and the feminization of parliaments: Counting only the 89 countries that Freedom House considers free, in October 2005 parliaments convened from majority systems had an average of 10.8% women within their membership, assemblies formed on the basis of mixed systems counted 17.7% women, and parliaments stemming from PR showed an average rate of 21.1%. In other words, legislative assemblies formed from majority systems count two times fewer women than do parliaments born of PR (particularly with party lists). This report confirms the conclusions of many studies showing that PR offers a greater potential for feminizing parliaments than do majority systems (see, e.g., Matland 1998; Norris 2004, 179–208; Rule 1987, 1994). PR seems, therefore, better equipped than majority systems to support the forming of a critical mass¹

of women in parliament.

The state of the research does not lend itself to equal understanding of the influence that the electoral systems have on women's substantive representation. Certainly, a number of studies have shown that many elected women believe they have a responsibility to represent women—that is to say, to represent them substantively by acting for women's concerns and women's perspectives. For instance, a research project undertaken by the Inter-Parliamentary Union with 200 female members of parliament (MPs) from 65 countries shows that 89% believed they had a particular responsibility to represent the needs and concerns of women (IPU 2000, 133–41). Moreover, a number of studies undertaken in countries using a diversified range of electoral systems led to the same conclusion (for example, in Australia and Canada [Tremblay 2003], England and Wales [Chaney 2006; Childs 2002], France [Sineau 2001], Israel [Golan and Hermann 2005], New Zealand [Grey 2002], the Scandinavian countries [Skjeie 1991; Wängnerud 2000] and in the United States [Carroll 2002]).

Is it necessary to deduce from this uniformity that the electoral system would have no effect on substantive representation of women? This conclusion would be inadequate. As Marian Sawer (1998, 52) explains, in theory, proportional systems should be better able than majority systems to ensure women's substantive representation:

Pippa Norris, as well as Shaun Bowler and David Farrell, support this reasoning. Norris (2004, 260–61) demonstrated that candidate ballots, which are used most frequently in majority systems,²

invite individual politicians to develop connections with their electorate through local representation and to adopt a certain distance from their political party. By contrast, party ballots (and particularly those that do not include a preferential vote) invite politicians to hide behind their political party and to develop a representation style that is more sensitive to global than to local issues. In research carried out on members of the European Parliament, Bowler and Farrell (1993) noted that the MEPs elected under candidate-ballots systems were more inclined to maintain regular contact with their electorate, whereas MEPs elected under party-ballots systems were more likely to develop connections with particular electoral clienteles. In addition, while in majority systems with single-seat constituencies the selection process for candidatures is a zero-sum game in which only one person comes out the winner, in PR (notably with party lists) the selection process of candidatures gives political parties the opportunity to respond to the various interests and factions that make them up and to solicit targeted clienteles in the electorate (including women). Put simply, PR with multiseat districts depersonalizes the mandate of representation: While in majority systems representation means a clearly identified link between a representative and a local electoral district, in PR it implies a more diffused one between many representatives and a given electoral area (i.e., a region, a state, the entire country) or a community of interests defined by identity, gender, ethnicity, and so on. Finally, as mentioned previously, PR tends to generate higher proportions of women in parliament than do majority systems,³ and consequently supports the attainment of a critical mass of women in politics.

While relatively convincing connections can be seen among electoral systems, descriptive representation, and critical mass, the same does not apply when the substantive representation of women is called into question. Though the proportion (or critical mass) of women stems considerably from the mechanics of electoral systems,⁴

this latter does not necessarily entail a critical mass of women (whether in PR or in majority systems) in the representation of women from a substantive point of view. Certainly, in theory, PR seems to offer more fertile ground for the substantive representation of women. But in practice, studies of female politicians elected through a broad range of electoral systems and constituting extremely variable proportions within their national parliament (13% in France versus 50% in Wales, for example) point toward a common finding, that of recognizing the responsibility to represent women. For me, this is encouragement for reexamining the concept of critical mass so that it will better serve the study of the substantive representation of women.

Surrogate Representation By and Critical Mass of Women in Politics

According to Jane Mansbridge (2003, 522),“Surrogate representation is representation by a representative with whom one has no electoral relationship.” It implies that a representative “acts for the interests of voters beyond the boundaries of the representative's district” (Carroll 2002, 51). This model calls up Edmund Burke's notion of virtual representation: “Virtual representation is that in which there is a communion of interests and a sympathy in feelings and desires between those who act in the name of any description of people and the people in whose name they act, though the trustees are not actually chosen by them” (Burke quoted in Williams 1998, 35). In other words, surrogate representation dissolves the conventional electoral ties between a representative and a given territory postulated by all electoral systems (be it a local district or the entire country). Surrogate representation can even be expressed on the international level when, for example, a Canadian female MP puts pressure on her government for it to increase foreign aid for birth control or when she denounces rape as a weapon of war.

Surrogate representatives do not have to be descriptive representatives. But as Mansbridge (1999, 642) explains: “[I]t is in this surrogate process that descriptive representation often plays its most useful role, allowing representatives who are themselves members of a subordinate group to circumvent the strong barriers to communication between dominant and subordinate groups.” Though similarity may indeed inspire confidence, it does not necessarily bring representation, which is fostered by common experiences and weak presence: “That sense of surrogate responsibility becomes stronger when the surrogate representative shares experiences with surrogate constituents in a way that a majority of the legislature does not…. Feelings of responsibility for constituents outside one's district grow even stronger when the legislature includes few, or disproportionately few, representatives of the group in question” (Mansbridge 2003, 523; my emphasis). Far from requiring critical mass, surrogate representation works within contexts of low numbers.

These two criteria—common experiences and weak presence—help theorize women's political representation in terms of surrogate representation. Several political theorists have argued that historically marginalized social identities should be represented by members of those groups—in other words, women should be represented by female parliamentarians because they share an identity and experiences which, for the most part, remain foreign to male politicians (among others, Dovi 2002; Mansbridge 1999, 2001; Phillips 1995; Williams 1998; Young 1989). In 2006, the worldwide average percentage of women in the single or lower houses of national parliaments is 16.6%.⁵

In only 20 out of 187 countries in the Inter-Parliamentary Union do women make up at least 30% of the members.

Women's surrogate representation cannot be passive. It requires consciousness on the part of political women; that is to say, they must consciously commit themselves to representing women. At the least, this commitment calls for a gender consciousness, which implies recognizing that sex/gender differences are not neutral in that they structure the social interactions between women and men. In other words, women's surrogate representation is not synonymous with feminist representation. Although it can certainly adopt this orientation, it can also consist of wanting to consolidate the equity of traditional gender roles (for example, so that housewives are not thrown out on the street after divorce). Conservative MPs may work for gender equality and equity in a way congruent with the feminist ethos and movement (Schreiber 2002, Tremblay and Pelletier 2003).

To sum up, not only does surrogate representation emancipate political representation from its conventional geographical grounding (which many have interpreted as an asset of PR in representing women from a substantive point of view), but it also encourages representation based on social identities and gender consciousness by strengthening the relationship between presence and ideas, and does so, apparently, without reference to the electoral system. Can we conclude that electoral systems have no impact on the substantive representation of women? Certainly not.

Although a critical mass of women in politics does not in any way guarantee substantive representation for women, electing female surrogate representatives of women improves the chances of it happening because these political women feel that they have a responsibility to represent women. (This does not, however, imply that they necessarily follow through on this responsibility—but that is another debate.) Yet, as mentioned previously, a rich literature shows that historically marginalized social identities (such as women) tend to be descriptively better represented in parliaments elected through PR. Consequently, it is reasonable to suggest that because it generally produces a higher proportion of women in parliament, PR will also generate by default a larger number of female surrogate representatives of women. This is, however, a hypothesis that remains to be confirmed by empirical tests.

Why focus on the number of female surrogate representatives of women since, as already mentioned, surrogate representation is far from requiring a critical mass and instead readily adapts itself to low numbers? In fact, the number of female surrogate representatives of women is not crucial for enabling surrogate representation, since just one woman may suffice (Childs and Withey, 2006). However, larger numbers can support the recognition and consideration of the diversity of ideas, perspectives, and concerns both among women as politicians and women in the general population. This opening to diversity will feed the gathering and discussion functions unique to deliberative representation. It will also contribute to developing a flexibility of movement for women in politics, who see their mandate of representation structured by various constraints and opportunities. As Carroll (2002, 67) notes, however, the diversity and flexibility unique to surrogate representation present their own surprises:

A critical mass of female surrogate representatives of women can support but in no way guarantee the substantive representation of women, attenuating at best the still-unresolved tension between representation and responsiveness.

It is now generally accepted that a critical mass of women in politics does not necessarily translate into substantive representation of women, since this latter depends on many factors, such as gender identity, party affiliation, and the legislative roles of women. What is surprising is that electoral systems, which in practice solidify political representation, have not yet been taken into account in this interconnection of descriptive representation, critical mass, and substantive representation of women. In fact, it would probably be fairer to say that electoral systems have been taken into account for their mechanical effects on the proportion of women (or critical mass) in parliament. The concept of surrogate representation calls into question, however, the psychological effects of electoral systems on the behavior of women elected in politics.

Thus, future research should explore how the psychological effects of the electoral systems do or do not generate incentives for women to act in certain ways once elected. For example, since preferential voting encourages candidates to develop a personal vote, it should prompt them, more than for their counterparts on a closed list, to solicit women's groups and, once elected, to feel accountable in representing them—whereas closed lists encourage loyalty to the party that decides who shall have a seat in parliament. This is a line of questioning that deserves to be explored, if only to nuance a largely widespread—and unfounded—belief that closed lists would be more favorable to representation of women than open lists. (In fact, though this may be the case for descriptive representation, it is not so for substantive representation.) More generally, in a context where PR (and mixed) systems have more and more adepts on the international level, and their adoption is often accompanied by quotas to encourage women's access to parliamentary representation, it becomes imperative to understand the complex and multi-faceted effects of electoral systems on how female parliamentarians conceive of their role in the representation of women.