Defining Women-Friendly Public Policy

The most problematic component of theorizing the relationship between women's descriptive and substantive representation concerns establishing what constitutes, in comparative political research, women's issues.¹¹

The idea of women's substantive representation hinges on the notion of some kind of shared experience among women that fosters a sense of common social or political interests.¹² Because women's experiences are socially constructed and because they vary widely based on the specific processes of construction, women's issues differ from context to context, across and within states and across time.¹³ Women's issues in the U.S. in the nineteenth century included, for example, temperance, suffrage, child-labor, public libraries, etiquette, and children's discipline. At the same time in Britain, women's issues focused on women's rights as autonomous citizens in marriage, married women's property ownership, women's rights in divorce, and child custody. In Southeast Asia, women's issues include micro enterprise aimed at women; in parts of Africa and the Middle East, female genital mutilation, veiling, citizenship rights, and early marriage are the substance of political and social movements on behalf of women. In contemporary India, women's issues include dowry requirements, protection against domestic abuse, and employment discrimination, among others.¹⁴ Because women's substantive representation differs across time and space, what it means to be successful on behalf of women's interests is governed by dynamics of culture and history.

It is also problematic that the effects of gender may not be distinct from those of race and class. In other words, the intersectionality of gender with other identities may limit the applicability of a women's issues paradigm.¹⁵

Much of the criticism of first-world feminism by third-world activists, for example, focuses on the perceived commonality of gendered experiences that is expected to transcend histories of colonialism, and oppression, as well as different cultures and economic experiences.¹⁶ In addition, women's experiences in the same society at the same point in history may be very different, depending on other social forces, and women may not see themselves as similar or their interests as shared. They may even organize against each other.¹⁷

In light of these challenges, definitions of women's issues tend toward the general. Carroll defines women's issues as those that “disproportionately become the responsibility of women as a result of the sexual division of labor” and as issues “where policy consequences are likely to have a more immediate and direct impact on significantly larger numbers of women than of men”.¹⁸

Lovenduski defines women's issues as “those that mainly affect women, either for biological reasons (such as breast cancer screening or reproductive rights) or for social reasons (sex equality or child-care policy)”.¹⁹ In the European context, Hoskyns studies women's policy as “the area of political action and activity which particularly concerns (or targets) women or groups of women, or where issues are forced onto the agenda by women”.²⁰

In research on legislatures, the definition of women's issue legislation also tends to be broad. Thomas, for example, uses a definition of “policies favoring women, children and families” as women's issues.²¹

Bratton defines women's interest legislation as legislation that would “decrease discrimination or counter the effects of discrimination, or would improve the socio-economic status of women”.²² Bystydzienski refers to care policies and gender equality policies in her study of women's policy in Norway, writing that care policies

Heath, Schwindt-Bayer, and Taylor-Robinson consider the way that Latin America legislatures deal with women's issues in terms of committees that “deal explicitly with women and one of the primary locations of gender inequality—the family and home (e.g., equality in the work place, protection against violence in the home).”²⁴

Swers, in her analysis of women in the U.S. Congress, defines women's legislative issues as those “that are particularly salient to women because they seek to achieve equality for women, they address women's special needs, such as women's health concerns or child-care issues; or they confront issues with which women have traditionally been concerned in their role as caregivers, such as education, or the protection of children.”²⁵ As Swers points out, however, these definitions of women's interest policies are so broad that they could be read to include all policies or none at all.²⁶

There is also the problem of distinguishing subjective from objective interests and this is why much research on women's policy issues conflates women's policy with feminist policy.²⁷

For example, Hernes uses the term “woman-friendly” to refer to gender equality and social policies produced by state feminism.²⁸ Gardiner and Leijenaar also define woman-friendly as policy designed to “improve the status of women, the adoption of and implementation of which reflect the existence of a culture of equality.”²⁹ Borchorst considers woman-friendliness to describe an atmosphere in which “patriarchal patterns of power” are debated.³⁰ Lovenduski does not distinguish between feminism and women's movement activism yet means only feminist activism when she discusses women's interests in public policy.³¹ Gelb and Palley also conflate feminism and women's activism. They consider the different success rates of feminist public policy initiatives, calling them role equity and role change issues:

Mazur, although explicit in acknowledging her focus on feminist public policy, limits her study to women's policy issues that involve at least three of five criteria:

Not all women's organizations or women's activism are feminist in the sense of challenging patriarchy; women often organize around nonfeminist and even anti-feminist goals.³⁴

Instead, women's organizations can be distinguished from other types of organizations, not by a litmus test of feminism but by “the primacy of women's gendered experiences, women's issues, and women's leadership and decision making.”³⁵ Much has been written on the negative connotations of the “feminist” label and this literature goes some distance to explain why individuals and movements may not identify with the label or with the framework of the broader agenda.³⁶ Some women's movements may deliberately (and strategically) avoid framing themselves as feminist;³⁷ in other contexts, rights-based approaches to pressuring the state would be inappropriate in the absence of a rights-based understanding of the state/society relationship. Indeed, some women's movements and organizations are pointedly anti-feminist (e.g., in the U.S., Daughters of the American Revolution, the Eagle Forum; in Chile, El Poder Femenino; in Nicaragua, the Association of Nicaraguan Women “Luisa Amanda Espinosa”). Furthermore, it would be a mistake to attempt to apply objective criteria of feminist content to determine if a movement is feminist or not or to parse out the goals of different women's movements for the purpose of discerning separate policy processes that pertain to each.

While we recognize these dynamics and acknowledge that women's movements and women's issues should not be conflated with feminism, we also acknowledge that nowhere in the comparative literature on women's substantive representation is a theoretical link established between women's movements that are non- or anti-feminist and the representation of issues that in an “immediate and direct way, are about women exclusively.”³⁸

Furthermore, no scholarship on elected women focuses on women in rightwing parties or theorizes about the policy produced by rightwing female legislators. Therefore, the relationship we aim to map between the descriptive representation of women and their substantive representation has most commonly (if not exclusively) been constructed around advancing women's status and equality. For the purposes of this project, we narrow the scope of women's issues and operationalize women-friendly public policy as a subset of policy advanced by women's organizations that both addresses issues that affect women exclusively and directly and that simultaneously advances their status in society. By this we mean specifically policies liberalizing divorce and reproductive rights; equalizing the civil rights of men and women in terms of education, employment, pay, training, property ownership and inheritance, marriage, mobility, and political representation; providing family and medical leave, subsidizing childcare, addressing domestic abuse, sexual assault, violence against women; and providing for women's health care; among others.³⁹

In short, our focus is on the interactions between sheer numbers of women and conducive policy-making contexts, based on the empirical political research to date. Our purpose is to theorize these interactions to identify combinations of conditions that might be expected to offer the greatest possibilities for converting women's descriptive representation into substantive representation, in the context of democratic political systems. Empirical evidence supports a general claim that the number of women elected to political office has a positive impact in terms of women-friendly public policy.⁴⁰

We identify three major sets of factors that provide the greatest analytical and explanatory leverage concerning women's substantive policy representation: sheer numbers of women, conducive parliamentary contexts, and conducive civil society contexts.

Sheer Numbers and Critical Mass

Research on the influence of sheer numbers of parliamentary women rely primarily on Rosabeth Moss Kanter's foundational study of women in a male-dominated corporate environment. Because of the structure of Kanter's argument, political scientists have discussed the impact of sheer numbers in terms of a “critical mass” of women and in terms of threshold percentages that are predicted to shift (that is, increase) women's policy representation. Kanter argued that women would have little influence as individuals or as a group in mostly uniform situations and that because they are still subject to performance pressures, role entrapment, and boundary heightening as tokens in “skewed groups,” they would not be able to muster influence within the larger corporation until they reached a representational threshold of 35 or 40 percent.⁴¹

Kanter hypothesized that in “uniform” majority/minority situations and “skewed” group contexts, where women constitute less than 15 percent of the overall group, their self-perception and the attitudes of their peers would interfere with their ability to interact effectively and would limit their influence. In “tilted” group contexts, where women constitute between 15 and 35 percent, Kanter hypothesized women would be seen as potential allies and that their influence would therefore increase. In “balanced” groups, where the groups are divided, at the most extreme, between 60 and 40 percent, Kanter argued, women would have their best opportunity to affect priorities and policies.⁴²

Research on women in national and regional legislatures has established that where women constitute less than 15 percent of a legislative body, women's influence will be constrained at best. There appears to be general agreement that a critical range of between 15 and 30 percent of women in a national or regional body, from parties across the political spectrum, is necessary for women to influence the agenda or style of business within that body or its policy outputs.⁴³

The impact of sheer numbers of women at the lower boundary of this critical range can be evaluated: women constitute 15 percent or more of the lower or single house members in 81 of the 185 nations about which the Inter-Parliamentary Union reported data.⁴⁴ For the purposes of assessing the critical mass argument, we employ the minimal 15 percent threshold, which we term the critical representation threshold. For empirical research purposes, setting any representational threshold higher than 15 percent becomes slightly problematic, given the problem of small numbers of cases. Although 53 nations meet a critical representation threshold of 20 or higher, only nine nations have a female delegation that constitutes more than 35 percent of their lower house.⁴⁵ Moving towards higher percentages by quintile, the numbers of nations within each quintile decreases (see table 1). Nonetheless, there is currently enough variation among nations in terms of their percentage parliamentary representation of elected women to assess the impact of sheer numbers.

It is likely that the higher the percentage of women elected to a national legislature, the more conducive the conditions for women's substantive representation. In addition, because most empirical research has found a strong relationship between women's representation within leftwing parliamentary groups or caucuses, we expect that high numbers of leftwing women elected to national legislatures will similarly effect women-friendly public policy. These claims in the literature can be hypothesized as follows:

H1: Where the percentage of women elected to a national legislature is 15 percent or higher, women's substantive representation in policy terms will be greater than in legislatures with a smaller percentage of elected women.

H2: Where the percentage of women representing leftwing parties in a national legislature is 15 percent or higher, women's substantive representation in policy terms will be greater than in legislatures with a smaller percentage of women elected by leftwing parties.

These, we argue, are the two most powerful variables concerning any critical mass of female legislators and the impact of sheer numbers of women on public policy outcomes.

Mapping the Numbers/Context Intersection

Mapping these seven factors provides a way to assess the impact of critical mass against sheer numbers, in specific political contexts. These seven factors are mapped interactively, albeit not exhaustively, in table 2. Although each factor should facilitate substantive policies for women, not all are necessary for substantive representation to occur. Instead, several combinations of intersecting factors are likely to effect women's substantive representation.

Table 2 maps five exemplary sets of intersecting factors with the potential to predict women's substantive representation. The first column of table 2 indicates the maximum facilitating conditions for women's substantive policies (Case 1). All seven facilitating factors are present: many women are in office, especially in left parties; there is a secure, leftwing government, with strong support from an active women's movement and from public opinion more broadly. In short, Case 1 relies on large numbers of elected women in the most auspicious context, hypothetically resulting in high levels of women's substantive representation. Although the intersection of these measures of numbers and context is likely to be rare, and data on some of the conditions are not available, Scandinavian parliaments since the 1960s come closest meet these criteria. By the early 1970s, Denmark, Finland, Norway and Sweden had all crossed the representational threshold of 15 percent women in the national parliament.⁵²

There were relatively high numbers of leftwing women in parliament, strong left parties in government, and popular support for women's equality. In each case, women's substantive representation was the result; each nation introduced laws regarding contraception and abortion, parental leave policies, equal pay for equal work, and equal opportunities at work, among other women-friendly policies, from the early 1960s to the mid 1980s.

Case 2 is one of few women but excellent conditions. In this case, of positive political context, a strong leftwing party governs with little effective opposition but with the support (or insistence) of a feminist movement encouraging women-friendly policy-making, and of women-friendly public opinion. Women's substantive representation is likely to occur under these conditions, even in the absence of sheer numbers of women. Examples of Case 2 can be found in the U.S. Congress, the German Bundestag, and the British Parliament in the late 1960s to the mid-1970s.⁵³

In each example, although there were few elected women, leftwing parties controlled the legislature, and feminist movements were active and visible. As a result, legislation concerning women's equality succeeded during this period. In the Federal Republic of Germany, abortion law and marriage and family law were reformed (1969–1976). In the U.K., the Labour government introduced social reforms that included the legalization of divorce, abortion, birth control and homosexuality, the establishment of the system of comprehensive education, and health care reform.⁵⁴ In the U.S., Congress approved federal funding for family planning (1967, 1970), proposed the Equal Rights Amendment (1972),⁵⁵ and approved national childcare legislation (1971).⁵⁶ In each of these instances, women's descriptive representation never exceeded six percent.⁵⁷

Case 3 is distinctive by virtue of the presence of women in two arenas: sheer numbers of elected women within parliament, and extra-parliamentary activist women in feminist movement in civil society. In this case of numbers and civil society, women are active and positioned within and outside government, in the absence of a leftwing governing party. Strictly speaking, elected women are present in relatively high numbers only within leftwing parties, for two reasons. First, no scholarly literature theorizes about the strong presence of rightwing women vis-à-vis women's public policy outcomes. Second, leftwing parties, excluded from government, are unlikely to hold a large number of seats and hence are unlikely to have elected enough members, let alone female members, to enable women's representation within the entire legislature to surpass the 15 percent representational threshold. This dual positioning, although not ideal, may nonetheless be sufficient to wrest some women-friendly public policy from national legislatures. For example, in Italy in the late 1960s and into the early 1980s, activist women in leftwing parliamentary parties, supported by a growing and militant feminist movement and by shifts in public opinion, were able to enact or sustain legislation concerning civil divorce, abortion, and employment opportunities for women.⁵⁸

Case 4 is a situation of numbers and parliament in which the sheer numbers of women are present and the left party is in a strong position in government, with little or no support in civil society. Although we map this theoretical possibility, no empirical examples are available among parliamentary democracies. The sole empirical example appears to be that of the former Soviet bloc states under communism. In these cases, women's equal legislative representation was guaranteed as a principle of communism, in the context of a single governing left party; the percentages of women sitting in national legislatures in the Soviet Union and in the Warsaw Pact nations were consistently higher than in other nations. Strictly speaking, no civil society existed, independent of state control, and hence there was no social movement sector, and no feminist or opposing movements. In most of these states, in fact, women's descriptive representation declined sharply after the collapse of communism, in part because it was not supported by sustained women's activism.⁵⁹

These examples, however, fall outside the established democracies whose policies we seek to explain and address legislatures that had little to no policy impact; in this regard, the case is more a theoretical construct than an empirically operative one. A relatively large percentage of women are unlikely to serve in a democratically elected national legislature, leftwing government notwithstanding, in the absence of support from civil society. Similarly, women-friendly public policy is not likely to occur in parliamentary democracies in the absence of pressures from civil society.

Case 5, that of sheer numbers only, represents a theoretical ideal-type for which, again, there is no empirical example. In this case, strong descriptive representation of women is countered by unfavorable parliamentary and civil society contexts. We include this case for two reasons. First, the absence of such empirical examples can be attributed to the causal connection between sheer numbers and the context variables. Favorable civil society and parliamentary contexts are themselves implicated with favorable numbers conditions. In the absence of a feminist movement, supportive public opinion, or a strong left-party to adopt the agenda of women's descriptive representation, women are not likely to be elected in large numbers to the assembly. This relationship between numbers and political context variables is well-established in the literature on women's descriptive representation.⁶⁰

Second, Case 5 completes the mapping of cases across the seven factors and, to that extent, is an important theoretical construct. The theoretical contribution of its empirical invisibility is that, at some level, descriptive representation and substantive representation of women are deeply connected. It is empirically impossible, although theoretically conceivable, to have large numbers of women in a legislature where the other conditions for women's substantive representation are hostile. At some level, women actually embody women's policy issues, because when conditions are hostile to women's public policy issues, women are not elected to parliaments. Case 5—of sheer numbers of women in a politically hostile context—would be the least favorable, and least likely, case for women-friendly public policy.

Note that one final possibility has been excluded from this table: cases where none of the conditions obtain. That is, no women sit in parliament, no left parties govern, there is no feminist movement, and public opinion does not support women's issues. We have excluded this set of cases for several reasons which, while perhaps obvious, deserve mention. First, there is no explicit theorizing in the literature on women's policy issues or women's election to parliament concerning outcomes regarding women's public policy under these unfavorable conditions. It would be possible to do so, for example, insofar as these conditions might map to fascist states, which actually legislate in regard to women's issues but not in the direction we predict for favorable conditions.⁶¹

Hypothesizing about the absence of the seven factors we have identified in the literature and mapped in table 2 would require something other than a simple statement of null hypothesis; a different set of theorizing, which no one has yet undertaken, would be required.⁶² Second, the absence of such theorizing would nonetheless require the development of another set of hypotheses suggesting that the absence of any facilitating conditions would not predict an absence of women's public policy, but would predict instead a wave of anti-women/anti-feminist public policy around issues of abortion, reproduction, marriage, rape, wife-beating, purdah, and other similar issues, a project beyond the scope of this paper.

Conclusions

We have developed a theoretical model that is parsimonious, testable in all democracies, consistent with existing empirical research, and supported by the available theorizing. There are many other variables suggested in the literature such as incumbency and newness, levels and capacity of party discipline, ethnic and racial diversity, access to leadership and legislative committee positions, attitudinal diversity among the women and their willingness to represent other women, political and electoral stability, institutional newness, design, and culture, the presence or absence of women's policy machinery in the form of an executive agency dedicated to women's issues, that might provide additional explanation for women's public policy enactment.⁶³

The seven we have selected are, however, most likely to provide the strongest analytical leverage on the question of critical mass and on the relationship between numbers of women (descriptive representation) and women-friendly policy outcomes (substantive representation).

As suggested in the discussion of the cases, above, it is unlikely that a subset of cases can be mapped to every combination of sheer numbers and favorable contexts. Nonetheless, we advance table 2 as a site for plotting more specific hypotheses and applying case data. Among the five ideal-type cases we discuss, only two are likely to predict women's policy success: Cases 1 and 2, the maximum facilitating conditions and positive political contexts. An arrow diagram of these two cases reveals three possibilities of strong relationships between numbers and contexts.

ARROW DIAGRAM 1: DIRECT IMPACT OF CIVIL SOCIETY FACTORS ON WOMEN-FRIENDLY PUBLIC POLICY

ARROW DIAGRAM 2 (Case 1): CIVIL SOCIETY FACTORS AS ANTECEDENT TO SHEER NUMBERS + POLITICAL CONTEXT, WITH THAT COMBINATION DIRECTLY EFFECTING WFPP

ARROW DIAGRAM 3 (Case 2): CIVIL SOCIETY FACTORS AS ANTECEDENT TO POLITICAL CONTEXT ALONE, WHICH THEN DIRECTLY EFFECTS WFPP

In these models, we understand civil society context as antecedent both to sheer numbers and to political context; we also model civil society as directly affecting public policy. Two major paths, therefore, predict women's public policy outcomes: one directly through civil society, where the pressure from public opinion and feminist movements, in the absence of opposition, persuade (male) legislators to act for women (Case 2); and a second path indirectly from civil society through combinations of agency represented by favorable political conditions of leftwing governments with little opposition or sheer numbers of women (Case 1).

As data are mapped to these diagrams and to the table, competing impacts concerning sheer numbers and critical mass explanations should become apparent. A critical mass impact, relying on a percentage threshold, should emerge as a sudden or dramatic increase in women's public policy; as the percentage of elected women increases from one election to the next (in parliament as a whole or simply in a major leftwing party), a step change should not be evident until the percentage of elected women meets or surpasses the critical representational threshold. In contrast, the impact of sheer numbers can be evinced in two ways: within a nation across time, or across nations. First, as the percentage of women in a single nation's legislature increases across time, there should be a concomitant increase in women's public policy. Second, among nations with a high percentage of women in their legislatures, public policy in women's interests should be high; nations with a small percentage of elected women should have little success in legislating on women's policy issues. Similarly, women-friendly public policy is not likely to occur in parliamentary democracies in the absence of pressures from civil society, particularly in the absence of an organized feminist movement.⁶⁴

When and under what conditions does women's substantive representation occur? Drawing from the comparative literature on women-friendly public policy, we have proposed a means to predict women's representation in policy terms. In doing so, we take the claims of critical mass theory and contextualize them in other explanatory variables, showing the limits of the sheer numbers claim but demonstrating a way to approach the relationship between women's descriptive and substantive representation. Our purpose is to advance the theoretical discussion of the link between women's elected representation and women-friendly public policy beyond the idea of critical mass and open up avenues for further research into how women's interests are advanced in the policy process.