Researchers often attribute diminishing gender inequality to economic development. When different aspects of gender inequality are examined, however, evidence points to both cross-national convergence as well as persistent (or even growing) heterogeneity in women's status. To make sense of this contradiction, we examine the extent to which culture moderates the relationship between economic development and gender inequality. We consider two dimensions of gender inequality, gender gaps in educational attainment and women's share of parliament, using data for 150 countries between 1980 and 2010. We find convergence toward greater equality in education, independently of economic development. But cross-cultural differences in female political representation persist or even grow as a function of economic development. Our results imply that economic development is not a direct pathway to greater gender equality. Rather, cultural legacies play an important role in shaping developmental trajectories.

Many researchers attribute diminishing gender inequality to economic development, regardless of cultural traditions or legacies. Dorius and Firebaugh (2010:1957), for example, conclude that “declining gender inequality is a pervasive phenomenon” across countries with diverse religious and cultural traditions. Lesthaeghe (2010) similarly finds at least some global convergence toward more egalitarian values, explained by the simultaneous forces of modern values and economic development. As a result, a powerful and widespread assumption is that increased economic prosperity shifts countries predictably toward greater gender equality. This assumption underlies the logic of development programs such as USAID and the World Bank.1 

Others, however, argue that cultural gender norms moderate the effect of development on women's outcomes. At first glance, evidence for this position is easy to find. Although delegates from 189 countries at the Fourth World Conference on Women in Beijing in 1995 voted unanimously to “remov[e] all the obstacles to women's active participation in all spheres of public and private life through a full and equal share in economic, social, cultural and political decision-making” (United Nations 1995:7), the conference also revealed profound cultural and religious fissures. One-third of the countries that voted to adopt the Beijing Platform for Action, heralded as “the most progressive blueprint ever for advancing women's rights,” did so with reservations. The Holy See's delegation complained that the document “mark[ed] another step in the colonization of the broad and rich discourse of universal rights by an impoverished, libertarian rights dialect” (159). The Libyan delegation refused to “accept the right of any nation or civilization to impose its culture or political, economic or social orientations on any other nation or people,” and underscored the right of “every State to draw up domestic policies in keeping with its religious beliefs” (165).

How can this apparent paradox be reconciled? Development is widely thought to foster greater egalitarianism, but there are questions as to whether this is the case in all contexts. Existing theories tend to treat women's status as though it were all of a piece, such that all aspects rise and fall together. We consider whether some domains of gender (in)equality show greater cross-cultural variation than others. Deep-seated cultural norms and traditions may affect women's outcomes in the educational and political spheres differently. We offer a theory, inspired by institutionalist perspectives in sociology, that explains why some outcomes improve as a function of development across cultures while others continue to show cross-cultural heterogeneity.

To address these issues, we use data on gender gaps in educational attainment and women's share of parliament for 150 countries between 1980 and 2010. After allocating countries into different cultural zones based primarily on their predominant religious traditions, we find that women's educational outcomes generally become more equitable over time and converge cross-culturally. In contrast, cross-cultural differences in women's political representation persist and even grow as a function of development, producing increased divergence across zones.

DEVELOPMENT, CULTURE, AND GENDER: EXISTING APPROACHES

Social scientists have long debated how economic development affects gender equality. Early modernization perspectives expected gender inequality to diminish as societies developed. Among other things, development is thought to modernize individuals’ worldviews and attitudes, leading post-materialist values such as personal autonomy and self-expression to displace traditional and religious values (Inglehart and Baker 2000; Inglehart and Welzel 2009; Welzel 2013). Changes in the status of women are thought to accompany these broader shifts. Inkeles and Smith (1974:26) contend that “most of the traditional societies and communities of the world are, if not strictly patriarchal, at least vigorously male dominated.” As societies develop, however, attitudes toward women liberalize and gender relations become more egalitarian (Inglehart and Norris 2003a; Seguino 2010). Fertility rates fall, educational opportunities expand, and women enter public life in greater numbers. Popular support for the participation of women in school, politics, and the economy grows as well. Welzel (2013:3) summarizes the expected changes: “Patriarchy, the most enduring form of human oppression, is declining, and the status of women is steadily improving in all but a handful of societies worldwide.”

Some evidence lends support to these ideas. Forsythe, Korzeniewicz, and Durrant (2000) identify a positive relationship between economic development and women's status as measured by the Gender Development Index, an index of gender gaps in life expectancy, education, and incomes. Gray, Kittilson, and Sandholtz (2006) find that economic development improves the quality of life and status of women around the world. Among other things, they conclude that affluent countries tend to have larger shares of parliamentary seats held by women. Higher per capita incomes also correlate with greater parity in girls’ and boys’ educational attainment (Cooray 2012; Klasen 2002; Klasen and Lamanna 2009). And, economic growth has the potential to reduce poverty. Because poverty disproportionately affects women and girls, growth can reduce gender inequality through this channel as well (Dollar and Gatti 1999; Duflo 2005).

Many alternative perspectives modify or challenge the modernization thesis, on the grounds that enduring cultural legacies shape development processes and their consequences. The revised modernization approach of Ronald Inglehart and his colleagues predicts that women's outcomes will generally improve as a function of development, albeit at highly variable rates across cultural zones. In this view, economic development “seems to move societies in a common direction” toward increased tolerance and gender egalitarianism “regardless of their cultural heritage” (Inglehart and Baker 2000:30), but path-dependent religious traditions and cultural legacies condition the pace of change. Thus, culture interacts with economic development to move countries along different trajectories, albeit typically in positive directions.

A contrasting perspective, expressed most forcefully in Samuel Huntington's controversial book, The Clash of Civilizations (1996), expects to find pockets of what might be called reactionary modernization, whereby socioeconomic development produces a resurgence in non-Western identities and values. Economic development strengthens non-Western societies economically, politically, and militarily, emboldening them to reject “Western” values and empowering them to reassert their own values (72–78). Inglehart and Baker (2000:21) concur that “modernization can actually strengthen traditional values.” Examples of this phenomenon are common enough. Although democratization is widely regarded as an important consequence of modernization (e.g., Lipset 1959), Møller and Skaaning (2013:96) point out that the establishment of democratic institutions in India served to politicize caste identities, which in turn strengthened that society's traditional values. Moreover, Tipps (1973:215) posited that technological developments accompanying modernization may sometimes reinforce traditional values. Along these lines, a New York Times report in 1980 on post-revolutionary Iran noted that, “in a somewhat incongruous but highly effective juxtaposition, Iran's new religious leaders have been using the technology of the electronic age literally to amplify their message of a return to centuries-old ways” (Kifner 1980:1). In these ways, modernization may at times increase rather than reduce cross-cultural differences.

There is reason to believe that trends indicative of reactionary modernization (or, at least, persistent traditionalism) will be strongest with respect to gender roles in societies. A series of studies by Inglehart and Norris (2003a, 2003b; Norris and Inglehart 2001, 2002) finds large and persistent cross-cultural differences in perceptions of women and women's roles, despite convergence in other domains. Even though Muslims and Westerners share strikingly similar political values—support for democracy, for example, is as strong in predominantly Islamic societies as in Western countries—they remain deeply divided on the issue of gender equality (Norris and Inglehart 2002). People in Asian, African, and Eastern European countries also expressed less support for gender equality than Westerners, even after accounting for levels of economic development (Inglehart and Norris 2003a). Dollar and Gatti (1999) demonstrate that religious traditions matter for women's outcomes, with predominantly Muslim and Hindu countries having greater gender inequality net of other factors. These patterns led Inglehart and Norris (2003b) to conclude that conflicting attitudes toward gender equality represent the “true clash of civilizations” in the modern world.2 

Eisenstadt's (1999, 2000) “multiple modernities” approach attempts to stake a middle ground between these revised and reactionary approaches by emphasizing that elites, intellectuals, and even the “broader strata” of non-Western societies selectively adopt some elements of modernity while rejecting others. Although the norms and assumptions associated with development are Western in origin and content, they posture as universally applicable—and very often are treated as such in non-Western societies. At the same time, the developmentalist project does not function as a tightly packaged or coherent ideology. Different elements are subject to “resistance, adoption, and hybridization” (Thornton, Dorius, and Swindle 2015:292). Modernization is not accepted or dismissed wholesale: some elements of modernity are incorporated into local practices, whereas others are adapted, ignored, or disavowed. As such, “the question as to whether modern societies converge or diverge is not an either-or question” (Schmidt 2010:512). Development may cause societies to converge in some respects and diverge in others.

A NEW APPROACH: RECONCILING CULTURAL AND INSTITUTIONAL LOGICS

The question then becomes: in which domains of gender equality do we expect economic development to produce convergence or divergence across cultures? Existing theoretical approaches do not clearly specify which outcomes will improve with modernization, and which cross-cultural differences—if any—will persist or even grow as societies develop. It is possible that the modernization thesis and its reactionary antithesis are both correct, but in different domains.

We draw from institutionalist perspectives in sociology to hypothesize seemingly inconsistent patterns of cross-cultural convergence and divergence. Our explanation begins with the assumption that influential global models—including models of gender equality—coalesce around the twin ideals of progress and justice: progress understood in terms of economic growth and development, and justice conceived as human rights and empowerment (Meyer, Boli, and Thomas 1987; Meyer, Boli, Thomas, and Ramirez 1997). Models of progress rest on instrumental and functional rationalizations, or what March and Olsen (1998:949) call the “logic of anticipated consequences.” In our case, improvements in some dimensions of women's status are routinely depicted as effective means for achieving collective and universally valued ends such as economic growth.

Models of justice, on the other hand, are driven by a “logic of appropriateness and senses of identity” and hence framed in principled rather than pragmatic terms. When something is interpreted as a matter of justice, it becomes an appropriate or desired end in itself, rather than simply a means toward some ulterior end. In such cases, March and Olsen contend that “the pursuit of purpose is associated with identities more than with interests,” and ultimate goals come to be defined as “ethical dimensions, targets, and aspirations” (951).

We argue that the ease with which particular domains of gender equality can be linked to broader models of progress or justice helps to determine cultural understandings of their applicability or appeal.

Equality in Education

Gender gaps in school enrollment have diminished around the world (Bradley 2000; Bradley and Ramirez 1996; Ramirez and Wotipka 2001), in part because gender equality in education is framed as conducive to economic progress and development. Much international discourse portrays education for girls and women as an investment in human capital and hence a driver of economic growth (Berkovitch 1999; Berkovitch and Bradley 1999). Consider a few examples:

  • The World Bank declares that “girls’ education is a strategic development investment.”3 

  • The United Nations Population Fund (2005:10) affirms that “All girls and boys have the right to education. … Along with nutrition, health and skills, education is a pillar of human capital: These essential elements together enable people to lead productive lives and to contribute to their countries’ economic growth and development.”

  • The World Economic Forum makes a “business case” for gender equality in education: “The key for the future of any country and any institution is the capability to develop, retain and attract the best talent. Women make up one half of the world's human capital. Empowering and educating girls and women and leveraging their talent … are thus fundamental elements of succeeding and prospering in an ever more competitive world. … Maximizing access to female talent is a strategic imperative for business” (Hausmann, Tyson, and Zahidi 2012:v).

  • The Organization for Economic Co-operation and Development (OECD 2012:3) points out that “increased education accounts for about half of economic growth in OCED countries in the past 50 years, and that has a lot to do with bringing more girls to higher levels of education and achieving greater equality in the number of years spent in education between men and women.”

  • The Islamic Development Bank (IDB) laments that “enrollment rates in IDB member countries show disparities in education due to gender. … This situation cannot go on like this for long because there exists a powerful developmental case for achieving gender parity. … It is in the private and social interest to eliminate gender inequalities in education wherever they exist. If the IDB member countries have to create productive economies, build more cohesive societies, and engage actively in the global economy, they must make effort to close the gender gap in their educational systems” (Bashir 2004:12).

The conviction that women's education promotes development is widespread (Thornton, Dorius, and Swindle 2015), and many studies lend support to the notion that women's access to schooling promotes economic growth via multiple pathways (Barro 2002; Benavot 1989; Dollar and Gatti 1999; Hill and King 1995; Klasen 2002; Klasen and Lamanna 2009; Seguino 2000). Countries around the world have taken notice. A recent report by the Asian Development Bank (2015:7–8) “summarizes the empirical evidence on how gender equality can impact economic growth. In terms of the impact of education of women on economic growth, the evidence is very clear and positive.” Consequently, “efforts to increase educational access for girls and women [are] universally supported, ideologically and materially” (Berkovitch and Bradley 1999:481–82).4 In short, a logic of anticipated consequences propels the expansion of women's educational access. The presumed causal linkage between gender equality in education and national economic growth is highly theorized, providing a strong argument against cultural relativism. Countries everywhere, regardless of culture, aspire to grow and develop economically, and women's education represents a widely recognized means for achieving these universally valued ends. Efforts to increase gender equity in educational outcomes can therefore be rationalized in ways that circumvent or overcome potential sources of cultural resistance.5 

Political Representation

Paxton, Hughes, and Green (2006) demonstrate a trend toward greater political representation for women over the twentieth century, propelled in large measure by an international social movement that pressured countries to incorporate women into politics. Yet, compared with the educational domain, gender equity in politics is more often interpreted as a principled imperative, and thus shaped by local cultures. In much of the West, the desirability of gender equality in politics has (recently) become self-evident. Upon his election as prime minister of Canada, for example, Justin Trudeau pithily justified gender parity in his cabinet with the remark, “Because it's 2015.” In a similar vein, Joni Lovenduski (2012), professor of politics at the University of London, recently argued that “women should not have to justify their political presence on any other basis than justice.”

We suspect that such “self-evidentiary” principles remain culturally contested in much of the world. Women's equality in politics tends to be justified on ideological rather than instrumental grounds, by recognizing the formal and intrinsic status equality of women to men. To the extent that ideologies of gender equality vary across cultural divides, however, we expect to find persistent cross-cultural differences in gendered political outcomes.

Few studies have considered whether women's parliamentary representation varies across cultures. When culture is considered, however, it generally has tremendous predictive power. Kenworthy and Malami (1999:252) conclude that “culture is more important than level of development” in explaining the share of female parliamentarians, with predominantly Catholic and Muslim countries trailing Protestant nations. Paxton, Hughes, and Green (2006) confirm that predominantly Muslim, Orthodox, and Catholic countries are significantly less likely than predominantly Protestant nations to incorporate women into national parliaments. Others report that the primary distinction is between Christian countries—whether Catholic or Protestant—and countries with non-Christian religious traditions (Matland 1998; Reynolds 1999).

Even fewer studies examine the “instrumental” effect of women's representation in national parliaments on economic growth.6 Whereas the positive association between women's education and economic growth is both highly theorized and well documented, the corresponding relationship between gender equality in politics and economic growth is less direct and not as well established. Compared with gender equality in education, the supposed mechanisms linking women's political empowerment to economic growth are circuitous and tend to be grounded in essentialist understandings of gender. Female politicians are thought to be less corrupt, more ethical, and more risk-averse than their male counterparts; in turn, these characteristics are assumed to promote economic growth (see e.g. Dollar, Fisman, and Gatti 2001; Eckel and Grossman 2008; Rivas 2012). But even these relationships are ambiguous or contested. Until recently, for example, corruption was viewed as a routine and even normal way to conduct business in large parts of the world, including the West (Kaufmann 1997; Milliet-Einbinder 2000). Although some research suggests that women's share of parliament increases welfare state spending (Bolzendahl 2009; Bolzendahl and Brooks 2007), this relationship is not used as an argument to increase political representation for women.

Thus, absent a clear instrumentalist account linking gender equality in politics to economic growth, efforts to increase women's share of parliament generally hinge on a logic of appropriateness rather than a logic of expected consequences. Again, we suspect that assessments of what is appropriate will display greater cross-cultural variability than understandings of what is instrumental. These arguments culminate in a simple and testable hypothesis: namely, that gender equality in education will improve and converge as a function of economic development across cultural divides, whereas cross-cultural differences in women's political outcomes will persist or even increase with development.

Preliminary evidence suggests the validity of this hypothesis. Dorius and Firebaugh (2010:1952) showed descriptively that while gender inequality in educational enrollment has narrowed over the past three decades, women remain severely underrepresented in national legislatures. However, they did not adjust these trends for political conditions, economic development, or other important factors that might render their trends spurious. Our analytic strategy, discussed below, puts these issues to rigorous testing.

DATA AND METHOD

Dependent Variables: Women's Educational and Political Outcomes

We analyze two dependent variables to measure gender equality. Educational equality refers to the gender gap in educational attainment, defined as the ratio of women's to men's average years of schooling. Data on educational attainment come from the Institute for Health Metrics and Evaluation, made available through the Quality of Government database (Teorell et al. 2013). For this variable, a score of 100 indicates that educational attainment is equivalent for women and men; scores below 100 mean that women attain fewer years of education, on average, than men. In the analyses that follow, positive effects therefore indicate a narrowing educational attainment gap between men and women.

Women's political equality is measured as women's share of parliament, or more specifically as the percentage of seats in a country's single or lower chamber of parliament held by women. Data before 1996 come from the Inter-Parliamentary Union by way of Melander (2005), updated using the World Bank's (2013) measure from the same source. A score of 50 indicates that women have reached parity with men.

Key Independent Variables: Economic Development and Cultural Zones

Our analyses follow standard practice by using gross domestic product (GDP) per capita, measured in constant 2005 U.S. dollars and logged to reduce skew, as a proxy for economic development (World Bank 2013).

To examine cross-cultural variation in women's outcomes we categorize countries into mutually exclusive and exhaustive cultural zones corresponding to their predominant religious traditions, as reported in the World Religion Dataset (Maoz and Henderson 2013). This data source estimates at half-decadal intervals the percentage of a country's population adhering to each of the following faiths: Christianity (Protestant, Catholic, and Orthodox), Islam, Buddhism, Hinduism, Confucianism, Animism, and none (i.e., no religious affiliation or faith). To assign countries to different zones, and in accordance with previous research (Cole 2016), we averaged these percentages over the entire observation period (1980 to 2010) and then coded the plurality faith as predominant.7 

On the basis of studies showing that “region trumps religion” when explaining liberalizing social reforms (Frank, Camp, and Boutcher 2010:885), democratization (Stepan and Robertson 2003; Teorell 2010), and a host of other outcomes (e.g., Eisenstadt 2001; Inglehart 2007; Inglehart and Baker 2000; Jenkins 2006; Spohn 2003; Welzel 2013; Welzel, Inglehart, and Klingemann 2003), we subdivided three of these religious groupings along regional lines, distinguishing Western and non-Western Protestant nations; Western, Latin American, and other Catholic countries; and Middle Eastern and other Islamic nations. Our framework results in 13 distinct cultural zones.  Appendix A reports the cultural affiliation of each country in the analysis.

Our method of allocating countries into cultural zones differs from Huntington's (1996) influential but controversial approach, which carves the world into several “civilizations” based on his assessment of fundamental differences in “religion, social structures, institutions, and prevailing values” (33). Although Huntington sees “religion as a central defining characteristic of civilizations” (47), he does not ground his framework in any objective measures of religious affiliation.

Inglehart and his colleagues (Inglehart and Baker 2000; Inglehart and Welzel 2005; Welzel 2013) base their macro-cultural classification schemes on Huntington's subjective designations. They find that his framework, with some modifications, maps onto patterns and changes in social values as measured by the World and European Values Surveys. Welzel (2013), for example, divides the world into “Western” and “Eastern” zones and then creates subcategories informed by various religious and historical factors.

Our framework departs from these subjective and inductive approaches in its use of an objective and a priori measure of religious composition to identify macro-cultural zones (Cole 2016). Nevertheless, the broad cross-cultural patterns in women's outcomes we report below are robust to the use of these alternative frameworks, as well as to regional classifications and indicators of colonial legacies (see  Appendix B for more details).

Control Variables

Despite the theoretical links between economic development and gender equality, one strand of research demonstrates that women's outcomes in a handful of affluent countries—those whose wealth derives from oil production—lag far behind. Women in these countries are less likely to work outside the home and consequently enjoy less political influence, less government representation, and fewer rights than their counterparts in other countries. Leite and Weidman (1999) argue that the dominance of natural resources fundamentally impacts gender dynamics in the economy and society, and Ross (2008) concludes that dependence on oil trumps Islamic cultural heritage in explaining women's negative outcomes. We therefore include an indicator for membership in the Organization of the Petroleum Exporting Countries (OPEC) as a control variable in our analyses.

In contrast with dependence on oil exports, overall trade openness, measured in terms of total trade as a proportion of GDP, is often associated with improved women's outcomes (Gray, Kittilson, and Sandholtz 2006; Richards and Gelleny 2007). This variable comes from the World Bank (2013).

We measure the extent to which a regime is democratic with the Polity IV institutionalized polity score (Marshall, Gurr, and Jaggers 2013), which ranges from –10 (most autocratic) to +10 (most democratic). A dummy variable for communist regimes accounts for their tendency toward better women's outcomes, at least nominally (Gray, Kittilson, and Sandholz 2006; Paxton, Hughes, and Green 2006).

Armed conflict can inhibit gender equality directly through the victimization of women and indirectly by “limit[ing] a society's capacity to cultivate and develop social norms and invest in social programs that promote gender equality” (Eastin and Prakash 2013:169). We include two binary measures, based on the Major Episodes of Political Violence dataset (Marshall 2006), that indicate the presence of international and/or civil warfare involving a country each year.

To control for demographic factors we include women's share of the total population and fertility rates from the World Bank (2013), as well as a measure of religious fractionalization from Alesina et al. (2003). The fractionalization score, which gives the probability that two randomly selected individuals within a country practice different religions, controls for within-country diversity to isolate the effects of between-country cultural and religious differences. It also accounts for “cleft” countries with multiple cultural or religious identities (Huntington 1996).

World society scholars contend that cultural and associational processes operating globally also shape country-level outcomes. Women's rights and gender equality have become highly institutionalized features of world-cultural discourse (Berkovitch 1999; Ramirez, Soysal, and Shanahan 1997), and women's outcomes might show greater improvement in countries with tighter linkages to world society. We include two measures of world-society influence. The first variable indicates whether a country is party to the International Convention on the Elimination of Discrimination against Women, a core women's rights treaty adopted by the United Nations in 1979.8 The second measure records the number of linkages countries have to international nongovernmental organizations (INGOs) focusing on women's issues (Berkovitch 1999; Wotipka and Ramirez 2008). A country is linked to an INGO if at least one of its citizens or civil society organizations belongs to it.

All of the preceding variables are common to analyses of both outcomes. When analyzing each dependent variable, we also include the other outcome variable as an additional predictor, to explore possible cross-domain dynamics (Jacobs 1996).

Finally, models analyzing women's share of parliament include an indicator for proportional-representation electoral systems (Beck et al. 2001; Keefer 2012) and a variable tallying the cumulative number of years since women were granted suffrage rights in national elections (Ramirez, Soysal, and Shanahan 1997), based on prior research (Gray, Kittilson, and Sandholtz 2006; Paxton, Hughes, and Green 2006; Reynolds 1999). Table 1 presents descriptive statistics for all dependent, independent, and control variables in our analysis.

TABLE 1.
Descriptive Statistics
VariableMeanSDMin.Max.
Gendered educational attainment ratio 74.56 25.26 11.76 143.64 
Women's share of parliament 12.23 9.68 56.30 
Protestant (Western) .08  
Protestant (non-Western) .06  
Catholic (Western) .12  
Catholic (Latin American) .15  
Catholic (Other) .08  
Islamic (Middle Eastern) .07  
Islamic (Other) .17  
Orthodox .06  
Buddhist .05  
Hindu .04  
Animist .09  
Non-religious .03  
Confucian .02  
GDP per capita (2005 US$, ln7.86 1.63 4.62 11.38 
Democracy/autocracy 2.76 6.97 −10.00 10.00 
OPEC member (1 = yes) .06  
Trade (% of GDP) 78.19 48.78 6.32 562.06 
International war (1 = yes) .01  
Civil war (1 = yes) .05  
Fertility rate 3.54 1.87 1.08 8.42 
Female population (%) 50.16 2.12 24.26 54.31 
Religious fractionalization .43 .24 .00 .86 
Communist (1 = yes) .05  
Women's INGO links 35.30 27.06 .00 200.00 
CEDAW member (1 = yes) .79  
Proportional representation (1 = yes) .51  
Years of female suffrage 49.25 21.96 118 
VariableMeanSDMin.Max.
Gendered educational attainment ratio 74.56 25.26 11.76 143.64 
Women's share of parliament 12.23 9.68 56.30 
Protestant (Western) .08  
Protestant (non-Western) .06  
Catholic (Western) .12  
Catholic (Latin American) .15  
Catholic (Other) .08  
Islamic (Middle Eastern) .07  
Islamic (Other) .17  
Orthodox .06  
Buddhist .05  
Hindu .04  
Animist .09  
Non-religious .03  
Confucian .02  
GDP per capita (2005 US$, ln7.86 1.63 4.62 11.38 
Democracy/autocracy 2.76 6.97 −10.00 10.00 
OPEC member (1 = yes) .06  
Trade (% of GDP) 78.19 48.78 6.32 562.06 
International war (1 = yes) .01  
Civil war (1 = yes) .05  
Fertility rate 3.54 1.87 1.08 8.42 
Female population (%) 50.16 2.12 24.26 54.31 
Religious fractionalization .43 .24 .00 .86 
Communist (1 = yes) .05  
Women's INGO links 35.30 27.06 .00 200.00 
CEDAW member (1 = yes) .79  
Proportional representation (1 = yes) .51  
Years of female suffrage 49.25 21.96 118 

N = 3,812.

SD = standard deviation; Min. = minimum; Max. = maximum; GDP = gross domestic product; OPEC = Organization of the Petroleum Exporting Countries; INGO = international nongovernmental organization; CEDAW = Convention on the Elimination of Discrimination against Women.

Analytic Strategy

To conduct our analyses, we use the heteroskedastic-robust OLS estimator with standard errors that are adjusted for first-order serial correlation.9 All analyses also include (but do not report) fixed effects for years that control for global changes in women's outcomes over time.10 We proceed in three analytic steps. First, we examine the independent roles of economic development and cultural context for each outcome variable: gendered educational attainment ratios and women's share of parliament. A second step considers cross-cultural variation in the relationship between economic development and women's educational and political outcomes by interacting cultural-zone memberships with per capita GDP.

A third and final way to explore patterns of cross-cultural convergence in women's political and educational outcomes draws upon a strategy used by economists to highlight the tendency of per capita incomes to grow at a faster rate in poorer countries than in affluent ones (Barro 1996; Sala-i-Martin 1996). This strategy documents income convergence by regressing each country's (or region's) rate of economic growth on its initial levels of economic development. In a similar vein, we regress changes in women's outcomes on baseline scores, to determine whether these outcomes improve at a faster rate in cultural zones with higher initial levels of gender inequality:

 
Y¯j,t0+TY¯j,t0=β0+β1Y¯j,t0+εj,t0,

where Y¯ gives the average outcome score (gendered educational attainment ratios or women's share of parliament) for cultural zone j; t0 indexes each zone's initial observation; and T specifies the length of the interval under consideration. Cross-cultural convergence occurs when β1 < 0, that is, when cultural zones with higher scores (i.e., more equitable gender distributions) at the start of a given interval improve at a slower rate than zones with lower initial scores. In the analyses that follow, we set T equal to 1, 5, 10, and 30, to analyze convergence at annual, half-decadal, decadal, and 30-year intervals, respectively.

RESULTS

Gender Gap in Educational Attainment

Table 2 presents the results of our analysis of gendered educational attainment ratios. We report three models. Model 1a, the baseline model, explores unadjusted differences across cultural zones. The second model adds per capita GDP to determine whether and to what extent cross-cultural variation in women's educational outcomes relative to men traces to differences in economic development. A third model is fully specified: it includes control variables alongside per capita GDP and the cultural-zone indicators. For ease of interpretation, Figure 1 plots coefficient estimates for the cultural-zone indicators from these models.

TABLE 2.
Heteroskedastic-Robust OLS Estimates for the Effect of Cultural Zones, Economic Development, and Control Variables on Gendered Educational Attainment Ratios, 1980 to 2010
Model 1aModel 2aModel 3a
Coeff.(S.E.)Coeff.(S.E.)Coeff.(S.E.)
Protestant (non-Western)a −31.035*** (6.416) −3.036 (6.053) 2.760 (5.562) 
Catholic (Western)a −5.440*** (1.163) −1.708 (2.022) −7.766** (2.729) 
Catholic (Latin American)a −7.027** (2.194) 12.475*** (3.717) 12.386*** (3.671) 
Catholic (Other)a −36.934*** (8.407) −6.771 (10.137) 8.183 (7.992) 
Islamic (Middle Eastern)a −36.081*** (3.902) −20.739*** (5.030) −8.582 (5.572) 
Islamic (non–Middle Eastern)a −44.810*** (4.115) −13.942* (6.529) −7.545 (5.725) 
Orthodoxa −11.938* (5.041) 8.388 (5.519) −6.241 (4.021) 
Buddhista −19.678** (6.674) 3.124 (5.928) −5.080 (5.548) 
Hindua −37.679** (12.051) −8.982 (9.588) −22.901** (7.211) 
Animista −44.700*** (7.565) −11.918 (9.138) −3.480 (8.284) 
Non-religiousa −6.919 (4.772) 12.447** (4.076) −5.533 (4.837) 
Confuciana −29.871*** (1.909) −11.894 (8.083) −34.925** (11.001) 
GDP per capita (2005 US$, ln  8.144*** (1.341) .162 (1.653) 
Democracy/autocracy     .273 (.231) 
OPEC member (1 = yes)     5.412 (4.226) 
Trade (% of GDP)     .068* (.030) 
International war (1 = yes)     5.753* (2.689) 
Civil war (1 = yes)     −1.510 (3.491) 
Fertility rate     −10.510*** (1.206) 
Female population (%)     .270 (.607) 
Religious fractionalization     11.106* (5.200) 
Communist (1 = yes)     −3.067 (4.307) 
Women's INGO links     −.017 (.045) 
CEDAW member (1 = yes)     −1.249 (2.092) 
Women's share of parliament     .133 (.078) 
Constant 89.493*** (1.220) 4.905 (13.806) 89.029* (38.751) 
F 21.417  24.658  46.352  
df 42  43  55  
Adjusted R-squared .494  .595  .766  
Mean VIF 1.85  2.47  2.25  
Maximum VIF 2.72  4.82  4.87  
Model 1aModel 2aModel 3a
Coeff.(S.E.)Coeff.(S.E.)Coeff.(S.E.)
Protestant (non-Western)a −31.035*** (6.416) −3.036 (6.053) 2.760 (5.562) 
Catholic (Western)a −5.440*** (1.163) −1.708 (2.022) −7.766** (2.729) 
Catholic (Latin American)a −7.027** (2.194) 12.475*** (3.717) 12.386*** (3.671) 
Catholic (Other)a −36.934*** (8.407) −6.771 (10.137) 8.183 (7.992) 
Islamic (Middle Eastern)a −36.081*** (3.902) −20.739*** (5.030) −8.582 (5.572) 
Islamic (non–Middle Eastern)a −44.810*** (4.115) −13.942* (6.529) −7.545 (5.725) 
Orthodoxa −11.938* (5.041) 8.388 (5.519) −6.241 (4.021) 
Buddhista −19.678** (6.674) 3.124 (5.928) −5.080 (5.548) 
Hindua −37.679** (12.051) −8.982 (9.588) −22.901** (7.211) 
Animista −44.700*** (7.565) −11.918 (9.138) −3.480 (8.284) 
Non-religiousa −6.919 (4.772) 12.447** (4.076) −5.533 (4.837) 
Confuciana −29.871*** (1.909) −11.894 (8.083) −34.925** (11.001) 
GDP per capita (2005 US$, ln  8.144*** (1.341) .162 (1.653) 
Democracy/autocracy     .273 (.231) 
OPEC member (1 = yes)     5.412 (4.226) 
Trade (% of GDP)     .068* (.030) 
International war (1 = yes)     5.753* (2.689) 
Civil war (1 = yes)     −1.510 (3.491) 
Fertility rate     −10.510*** (1.206) 
Female population (%)     .270 (.607) 
Religious fractionalization     11.106* (5.200) 
Communist (1 = yes)     −3.067 (4.307) 
Women's INGO links     −.017 (.045) 
CEDAW member (1 = yes)     −1.249 (2.092) 
Women's share of parliament     .133 (.078) 
Constant 89.493*** (1.220) 4.905 (13.806) 89.029* (38.751) 
F 21.417  24.658  46.352  
df 42  43  55  
Adjusted R-squared .494  .595  .766  
Mean VIF 1.85  2.47  2.25  
Maximum VIF 2.72  4.82  4.87  
a

Western Protestant countries are the omitted reference group.

Standard errors, adjusted for first-order autocorrelation, in parentheses.

Fixed effects for years included in the analyses but not reported.

CEDAW = Convention on the Elimination of Discrimination against Women; VIF = variance inflation factor score.

*p < .05, **p < .01, ***p < .001 (two-tailed).

N(country-year observations) = 3,812; N(countries) = 150.

FIGURE 1.

Cross-Cultural Variation in Gendered Educational Attainment Ratios

FIGURE 1.

Cross-Cultural Variation in Gendered Educational Attainment Ratios

Model 1a, the culture-only baseline model, indicates that female educational attainment in the Protestant West—the omitted reference category—was close to parity with men (constant term = 89.5) during the observation period. Gendered educational attainment ratios for countries in the non-religious zone did not differ significantly from the Protestant West, but the unadjusted educational attainment gap for all other cultural zones favored men, often by wide margins. However, Model 2a shows that much of these baseline differences trace to one factor: level of economic development. As GDP per capita increases, gendered educational attainment ratios also increase. With the addition of this measure of economic development, ratios in eight cultural zones—Animist, Buddhist, Confucian, Hindu, Orthodox, Western Catholic, other Catholic, and non-Western Protestant—no longer differ significantly from the Protestant West. Coefficient estimates on two additional cultural zones—non-religious and Latin American Catholic—become significantly positive, meaning that gendered educational attainment ratios in these zones are actually higher than in the Protestant West once economic development is considered. Only the estimates for Islamic countries, both in the Middle East and elsewhere, remain significantly negative, but even here the effect is much smaller.

Adding the full complement of control variables in Model 3a reveals a somewhat different pattern of cross-cultural differences in educational attainment for women, although estimated effects are significantly negative in only three zones: Confucian, Hindu, and (to a much lesser extent) Western Catholic.11 Three control variables—trade openness, religious fractionalization, and international war—correlate with higher attainment ratios, whereas educational attainment for women relative to men declines with higher fertility rates.12 GDP per capita is no longer a significant predictor of educational attainment for women in the fully specified model. Variance inflation factor (VIF) scores suggest that multicollinearity is not a cause for concern in our analyses.

Overall, we conclude that initially large cross-cultural differences in gendered educational attainment ratios shrink appreciably once countries’ overall levels of economic development and other political, social, and demographic factors are considered, with high levels of fertility serving as a particularly strong barrier for gender equality in education.

Women's Share of Parliament

Table 3 reports the results of analyses for women's share of parliament, and Figure 2 plots coefficient estimates for the cultural-zone indicators. The unadjusted estimates in Model 1b show that women's share of parliament is significantly and substantially larger in Western Protestant countries than in any other cultural zone. About 21 percent of parliamentary seats in the Protestant West were held by women during the entire observation period (as indicated by the constant term in Model 1b). On average, the percentage of parliamentary seats held by women in other cultural zones is roughly 10 to 20 percentage points lower, depending on the zone, than in the Protestant West; this range is bookended by Western Catholic and Middle Eastern Islamic countries, respectively. Only the estimate on the Confucian indicator fails to achieve statistical significance in the baseline model.

TABLE 3.
Heteroskedastic-Robust OLS Estimates for the Effect of Cultural Zones, Economic Development, and Control Variables on Women's Share of Parliament, 1980 to 2010
Model 1bModel 2bModel 3b
Coeff.(S.E.)Coeff.(S.E.)Coeff.(S.E.)
Protestant (non-Western)a −16.902*** (3.728) −18.243*** (4.116) −15.628*** (3.794) 
Catholic (Western)a −9.469** (3.424) −9.647** (3.452) −10.759*** (3.200) 
Catholic (Latin American)a −13.099*** (3.429) −14.033*** (3.570) −14.425*** (3.286) 
Catholic (Other)a −12.361*** (3.517) −13.805*** (3.565) −13.122*** (3.605) 
Islamic (Middle Eastern)a −22.865*** (3.323) −23.600*** (3.347) −19.221*** (3.461) 
Islamic (non−Middle Eastern)a −16.480*** (3.263) −17.958*** (3.567) −15.409*** (3.286) 
Orthodoxa −16.419*** (3.562) −17.392*** (3.668) −17.300*** (3.239) 
Buddhista −16.842*** (3.454) −17.934*** (3.527) −19.165*** (3.124) 
Hindua −12.789** (4.360) −14.163** (4.513) −11.742*** (3.547) 
Animista −15.708*** (3.500) −17.278*** (3.620) −15.368*** (3.531) 
Non-religiousa −10.319** (3.827) −11.246** (3.779) −14.909*** (3.212) 
Confuciana −10.412 (5.506) −11.273* (5.187) −17.425*** (3.743) 
GDP per capita (2005 US$, ln  −.390 (.365) −.342 (.453) 
Democracy/autocracy     −.269*** (.073) 
OPEC member (1 = yes)     −1.318 (1.295) 
Trade (% GDP)     .017* (.008) 
International war (1 = yes)     −2.150 (2.788) 
Civil war (1 = yes)     −1.499 (1.141) 
Fertility rate     .174 (.416) 
Female population (%)     .321 (.202) 
Religious fractionalization     .555 (2.316) 
Communist (1 = yes)     12.993*** (2.068) 
Women's INGO links     .048* (.022) 
CEDAW member (1 = yes)     1.464 (1.062) 
Proportional representation (1 = yes)     4.074*** (1.154) 
Years of female suffrage     −.005 (.026) 
Gendered educational attainment ratio     .041 (.026) 
Constant 20.678*** (3.219) 24.729*** (4.950) −.645 (11.619) 
F 11.292  11.428  13.835  
df 42  43  57  
Adjusted R-squared .403  .405  .521  
Mean VIF 1.85  2.47  2.44  
Maximum VIF 2.72  4.82  5.93  
Model 1bModel 2bModel 3b
Coeff.(S.E.)Coeff.(S.E.)Coeff.(S.E.)
Protestant (non-Western)a −16.902*** (3.728) −18.243*** (4.116) −15.628*** (3.794) 
Catholic (Western)a −9.469** (3.424) −9.647** (3.452) −10.759*** (3.200) 
Catholic (Latin American)a −13.099*** (3.429) −14.033*** (3.570) −14.425*** (3.286) 
Catholic (Other)a −12.361*** (3.517) −13.805*** (3.565) −13.122*** (3.605) 
Islamic (Middle Eastern)a −22.865*** (3.323) −23.600*** (3.347) −19.221*** (3.461) 
Islamic (non−Middle Eastern)a −16.480*** (3.263) −17.958*** (3.567) −15.409*** (3.286) 
Orthodoxa −16.419*** (3.562) −17.392*** (3.668) −17.300*** (3.239) 
Buddhista −16.842*** (3.454) −17.934*** (3.527) −19.165*** (3.124) 
Hindua −12.789** (4.360) −14.163** (4.513) −11.742*** (3.547) 
Animista −15.708*** (3.500) −17.278*** (3.620) −15.368*** (3.531) 
Non-religiousa −10.319** (3.827) −11.246** (3.779) −14.909*** (3.212) 
Confuciana −10.412 (5.506) −11.273* (5.187) −17.425*** (3.743) 
GDP per capita (2005 US$, ln  −.390 (.365) −.342 (.453) 
Democracy/autocracy     −.269*** (.073) 
OPEC member (1 = yes)     −1.318 (1.295) 
Trade (% GDP)     .017* (.008) 
International war (1 = yes)     −2.150 (2.788) 
Civil war (1 = yes)     −1.499 (1.141) 
Fertility rate     .174 (.416) 
Female population (%)     .321 (.202) 
Religious fractionalization     .555 (2.316) 
Communist (1 = yes)     12.993*** (2.068) 
Women's INGO links     .048* (.022) 
CEDAW member (1 = yes)     1.464 (1.062) 
Proportional representation (1 = yes)     4.074*** (1.154) 
Years of female suffrage     −.005 (.026) 
Gendered educational attainment ratio     .041 (.026) 
Constant 20.678*** (3.219) 24.729*** (4.950) −.645 (11.619) 
F 11.292  11.428  13.835  
df 42  43  57  
Adjusted R-squared .403  .405  .521  
Mean VIF 1.85  2.47  2.44  
Maximum VIF 2.72  4.82  5.93  
a

Western Protestant countries are the omitted reference group.

Standard errors, adjusted for first-order autocorrelation, in parentheses.

Fixed effects for years included in the analyses but not reported.

CEDAW = Convention on the Elimination of Discrimination against Women; VIF = variance inflation factor score.

*p .05, **p < .01, ***p < .001 (two-tailed).

N(country-year observations) = 3,812; N(countries) = 150.

FIGURE 2.

Cross-Cultural Variation in Women's Share of Parliament

FIGURE 2.

Cross-Cultural Variation in Women's Share of Parliament

These baseline cross-cultural differences remain robust to the inclusion of GDP per capita in Model 2b, whose effect is negative but statistically unreliable. Indeed, coefficient estimates for each cultural-zone indicator not only remain significantly negative, they also increase slightly in magnitude. The previously negative but insignificant coefficient on Confucian countries in the baseline model also becomes statistically significant in the presence of GDP per capita. These negative effects persist when evaluated alongside additional control variables in Model 3b, indicating that women's share of parliament in each zone continues to be lower vis-à-vis Western Protestant countries.

Several control variables attain significance in Model 3b. The share of female parliamentarians is greater in communist regimes and countries with proportional-representation electoral systems than in non-communist regimes and countries with other electoral systems. The share of parliamentarians who are women also increases with trade and linkages to women's INGOs. The democracy score, conversely, is inversely associated with women's representation in national parliaments, consistent with the “democracy paradox” highlighted by Fallon, Swiss, and Viterna (2012). Nevertheless, cultural differences appear to be the overriding explanatory factor.

These analyses demonstrate that cross-cultural inequalities in women's share of parliament are large and remain durable even when controlling for levels of economic development and other factors shown in past empirical research to explain gender inequalities in political representation.

Do the Effects of Economic Development Vary by Culture?

Table 4 presents analyses that include interaction terms between GDP per capita and each cultural-zone indicator, to examine whether the effect of economic development on the gendered educational attainment gap (Model 4a) and women's share of parliament (Model 4b) varies across cultural zones. Both models include the full set of control variables corresponding to each outcome. To render the main effects of cultural-zone indicators in these analyses substantively meaningful, we standardize GDP per capita to have a mean of 0 and standard deviation of 1.

TABLE 4.
Heteroskedastic-Robust OLS Estimates for the Effect of Economic Development on Gendered Educational and Political Outcomes, as Conditioned by Cultural Zones, 1980 to 2010
Model 4a: Gendered educational attainment ratioModel 4b: Women's share of parliament
Coeff.(S.E.)Coeff.(S.E.)
GDP per capita (2005 US$, ln)b 11.350 (7.184) 19.372** (6.853) 
Protestant (non-Western)a 24.255* (11.295) 16.942 (11.710) 
 × GDP/capita (2005 US$, ln−5.949 (8.402) −16.401* (7.287) 
Catholic (Western)a 15.475 (11.095) 13.570 (11.909) 
 × GDP/capita (2005 US$, ln−16.197* (7.699) −14.389* (6.906) 
Catholic (Latin American)a 31.467** (11.988) 16.416 (12.203) 
 × GDP/capita (2005 US$, ln−21.043* (9.252) −16.608* (7.372) 
Catholic (Other)a 20.849 (15.281) 15.040 (12.122) 
 × GDP/capita (2005 US$, ln−18.678* (9.256) −23.499*** (6.969) 
Islamic (Middle Eastern)a 8.697 (13.289) 11.565 (12.123) 
 × GDP/capita (2005 US$, ln−8.025 (8.823) −23.773*** (7.213) 
Islamic (Other)a 5.764 (11.938) 14.072 (12.219) 
 × GDP/capita (2005 US$, ln−17.522 (10.877) −21.450** (6.904) 
Orthodoxa 11.534 (10.577) 13.385 (11.933) 
 × GDP/capita (2005 US$, ln−11.457 (7.274) −19.881** (6.672) 
Buddhista 12.299 (11.566) 11.833 (12.059) 
 × GDP/capita (2005 US$, ln−11.605 (7.935) −19.067** (7.037) 
Hindua 3.028 (12.249) 17.602 (11.644) 
 × GDP/capita (2005 US$, ln1.470 (10.639) −21.667** (7.320) 
Animista 33.787** (12.685) 15.075 (11.981) 
 × GDP/capita (2005 US$, ln9.504 (9.960) −20.144** (6.981) 
Non-religiousa 10.816 (10.556) 15.678 (11.934) 
 × GDP/capita (2005 US$, ln−9.320 (8.379) −18.582** (6.955) 
Confuciana −16.941 (10.831) 12.459 (11.990) 
 × GDP/capita (2005 US$, ln−23.566* (9.791) −22.234** (6.909) 
Democracy/autocracy .200 (.169) −.285*** (.068) 
OPEC member (1 = yes) 7.755 (4.579) −.059 (1.427) 
Trade (% of GDP) .086** (.030) .021* (.010) 
International war (1 = yes) 3.574 (3.152) −2.585 (2.723) 
Civil war (1 = yes) −.756 (3.483) −1.863 (1.197) 
Fertility rate −10.731*** (1.315) −.049 (.421) 
Female population (%) .671 (.792) .099 (.258) 
Religious fractionalization 10.559* (4.844) 1.911 (2.488) 
Communist (1 = yes) −1.855 (4.267) 12.534*** (2.148) 
Women's INGO links .017 (.046) .021 (.023) 
CEDAW member (1 = yes) −.916 (1.758) 1.850 (1.063) 
Proportional representation (1 = yes)   3.960*** (1.064) 
Years of female suffrage   −.001 (.024) 
Gendered educational attainment ratio   .031 (.028) 
Women's share of parliament .092 (.076)   
Constant 51.870 (44.756) −21.654 (16.833) 
F 103.744  40.467  
df 67  69  
Adjusted R-squared .802  .552  
Model 4a: Gendered educational attainment ratioModel 4b: Women's share of parliament
Coeff.(S.E.)Coeff.(S.E.)
GDP per capita (2005 US$, ln)b 11.350 (7.184) 19.372** (6.853) 
Protestant (non-Western)a 24.255* (11.295) 16.942 (11.710) 
 × GDP/capita (2005 US$, ln−5.949 (8.402) −16.401* (7.287) 
Catholic (Western)a 15.475 (11.095) 13.570 (11.909) 
 × GDP/capita (2005 US$, ln−16.197* (7.699) −14.389* (6.906) 
Catholic (Latin American)a 31.467** (11.988) 16.416 (12.203) 
 × GDP/capita (2005 US$, ln−21.043* (9.252) −16.608* (7.372) 
Catholic (Other)a 20.849 (15.281) 15.040 (12.122) 
 × GDP/capita (2005 US$, ln−18.678* (9.256) −23.499*** (6.969) 
Islamic (Middle Eastern)a 8.697 (13.289) 11.565 (12.123) 
 × GDP/capita (2005 US$, ln−8.025 (8.823) −23.773*** (7.213) 
Islamic (Other)a 5.764 (11.938) 14.072 (12.219) 
 × GDP/capita (2005 US$, ln−17.522 (10.877) −21.450** (6.904) 
Orthodoxa 11.534 (10.577) 13.385 (11.933) 
 × GDP/capita (2005 US$, ln−11.457 (7.274) −19.881** (6.672) 
Buddhista 12.299 (11.566) 11.833 (12.059) 
 × GDP/capita (2005 US$, ln−11.605 (7.935) −19.067** (7.037) 
Hindua 3.028 (12.249) 17.602 (11.644) 
 × GDP/capita (2005 US$, ln1.470 (10.639) −21.667** (7.320) 
Animista 33.787** (12.685) 15.075 (11.981) 
 × GDP/capita (2005 US$, ln9.504 (9.960) −20.144** (6.981) 
Non-religiousa 10.816 (10.556) 15.678 (11.934) 
 × GDP/capita (2005 US$, ln−9.320 (8.379) −18.582** (6.955) 
Confuciana −16.941 (10.831) 12.459 (11.990) 
 × GDP/capita (2005 US$, ln−23.566* (9.791) −22.234** (6.909) 
Democracy/autocracy .200 (.169) −.285*** (.068) 
OPEC member (1 = yes) 7.755 (4.579) −.059 (1.427) 
Trade (% of GDP) .086** (.030) .021* (.010) 
International war (1 = yes) 3.574 (3.152) −2.585 (2.723) 
Civil war (1 = yes) −.756 (3.483) −1.863 (1.197) 
Fertility rate −10.731*** (1.315) −.049 (.421) 
Female population (%) .671 (.792) .099 (.258) 
Religious fractionalization 10.559* (4.844) 1.911 (2.488) 
Communist (1 = yes) −1.855 (4.267) 12.534*** (2.148) 
Women's INGO links .017 (.046) .021 (.023) 
CEDAW member (1 = yes) −.916 (1.758) 1.850 (1.063) 
Proportional representation (1 = yes)   3.960*** (1.064) 
Years of female suffrage   −.001 (.024) 
Gendered educational attainment ratio   .031 (.028) 
Women's share of parliament .092 (.076)   
Constant 51.870 (44.756) −21.654 (16.833) 
F 103.744  40.467  
df 67  69  
Adjusted R-squared .802  .552  
a

Western Protestant countries are the omitted reference group.

b

Standardized to have mean = 0 and standard deviation = 1.

Standard errors, adjusted for first-order autocorrelation, in parentheses. Fixed effects for years included in the analyses but not reported.

*p < .05, **p < .01, ***p < .001 (two-tailed).

N(country-year observations) = 3,812; N(countries) = 150.

In these analyses, the coefficients for GDP per capita give the effect of economic development for the Western Protestant zone, the omitted reference category. On average and net of other factors in the model, a one-unit increase in (logged and standardized) GDP per capita is associated with increases of 11.4 points in the gendered educational attainment ratio and more than 19 percentage points in women's share of parliament among Western Protestant countries, although only the latter estimate is statistically significant. Because the effects of economic development for other cultural zones are difficult to discern from Table 4 alone, we plot the marginal effects of GDP per capita for each cultural zone in Figure 3, with all other control variables held constant at their mean values.

FIGURE 3.

Marginal Effects of Economic Development on Women's Educational and Political Outcomes, by Cultural Zone

FIGURE 3.

Marginal Effects of Economic Development on Women's Educational and Political Outcomes, by Cultural Zone

The left panel illustrates our findings for gender equity in education. Per capita GDP is associated with greater gender equity in educational attainment in the Animist and Hindu zones. Women's educational attainment relative to men decreases with GDP per capita in Confucian countries. For all other cultural zones, the marginal effect of economic development on gendered educational attainment ratios overlaps with zero, indicating statistically insignificant effects.

The patterns are somewhat different for women's share of parliament. We find a significant positive association between per capita GDP and women's share of parliament for only two zones: Western Protestant and Western Catholic. Even here, however, the relationship is much stronger for Protestant countries than for their Catholic counterparts. At the other end of the spectrum, increasing per capita GDP is associated with a decline in women's share of parliament in Middle Eastern Islamic and “other” Catholic countries. For all other cultural zones, the marginal effect of economic development is statistically indistinguishable from zero. These results suggest that the gap between Western and non-Western countries in women's share of parliament will grow as countries develop economically.13 

Our final set of analyses uses a different method to explore whether women's educational and political outcomes converge across cultural zones over time. Table 5 summarizes the results. Statistically significant negative coefficients indicate that outcomes are converging across zones over time. All coefficients in the analysis of gendered educational attainment ratios are significantly negative, indicating that educational outcomes converge across cultural zones at every interval (annual, 5-year, decadal, and the entire 30-year panel). Coefficients in the analysis of women's share of parliament do not achieve statistical significance.

TABLE 5.
Convergence in Women's Educational and Political Outcomes across Cultural Zones
Gendered educational attainment ratioWomen's share of parliament
Annual −.015*** −.002 
 (.002) (.012) 
5-year −.087*** .009 
 (.016) (.048) 
Decadal −.148*** −.017 
 (.024) (.106) 
Full panel −.390** −.224 
 (.096) (.431) 
Gendered educational attainment ratioWomen's share of parliament
Annual −.015*** −.002 
 (.002) (.012) 
5-year −.087*** .009 
 (.016) (.048) 
Decadal −.148*** −.017 
 (.024) (.106) 
Full panel −.390** −.224 
 (.096) (.431) 

**p < .01, ***p < .001 (two-tailed). Standard error in parentheses.

Coefficients give the estimated effect of initial scores for each outcome on changes in these scores across four intervals:

  • Annual—Modeling annual changes in average women's outcomes scores across cultural zones as a function of the previous year's score. N = 384.

  • 5-year—Modeling 5-year changes in average women's outcomes scores across cultural zones as a function of scores at the start of each interval: 1980–85, 1985–90, 1990–95, 1995–2000, 2000–05, 2005–10. N = 76.

  • Decadal—Modeling 10-year changes in average women's outcomes scores across cultural zones as a function of scores at the start of each interval: 1980–90, 1990–2000, 2000–10. N = 38.

  • Full panel—Modeling changes in average women's outcomes scores across cultural zones between 1980 and 2010 as a function of scores in 1980. N = 13.

Figure 4 illustrates one set of results, plotting changes in each zone's educational and political outcomes between 1980 and 2010 (y-axis) against their scores in 1980 (x-axis). The trend for gendered educational attainment ratios in the left panel is strongly negative: over the observation period, zones with ratios that favored men in 1980 closed the gender gap faster than zones with starting ratios that were more equitable. For example, compare the Protestant West with the Islamic Middle East. In 1980, average educational attainment for men and women in Western Protestant countries was near parity, at roughly 95, and the ratio increased by only 5 points between 1980 and 2010. In contrast, Middle Eastern Islamic countries’ lower ratio in 1980 (42) left much room for improvement. Over the period under observation, its ratio increased by nearly 39 points.

FIGURE 4.

Patterns of Beta-Convergence in Women's Outcomes across Cultural Zones

FIGURE 4.

Patterns of Beta-Convergence in Women's Outcomes across Cultural Zones

In contrast, the dashed trend line in the right panel shows that when the non-religious zone—a clear outlier—is removed from consideration, the relationship between a zone's initial share of parliamentary seats held by women and its change on this outcome over the entire period is positive. Accordingly, women's share of parliament increased at a faster rate in zones with greater levels of female representation in 1980, resulting in increased divergence across cultural zones over time. Compare again the results for Protestant Western countries and Middle Eastern Islamic nations. In the former, women held 14 percent of parliamentary seats, on average, in 1980, and the share of seats held by women increased by 20 percentage points over the next 30 years. In the latter, women held only 2.6 percent of seats in 1980, and their share grew by an anemic 4.7 percentage points over the period. In short, when we compare our findings across the two dimensions of women's societal position, we find reasonably clear evidence of cross-cultural convergence for gender equity in education, but cross-cultural divergence for gender equity in parliamentary representation.

DISCUSSION AND CONCLUSIONS

For decades, social scientists have developed theoretical frameworks linking economic modernization to a wide variety of structural transformations and psychological changes in society, including changes in women's status, attitudes about gender roles, and patterns of gender inequality. Various frameworks expect cultural differences to condition the effects of economic development on gendered outcomes, albeit in dissimilar ways. Modernization theorists predict that cross-cultural differences will diminish and ultimately disappear as countries develop. The revised modernization thesis suggests that some domains of gender equality will improve with development, but at highly variable rates that may not result in convergence across cultures. Reactionary modernization perspectives make the opposite prediction: in many non-Western societies, economic development revitalizes local cultures, leading countries to reject perceived Westernization and causing women's outcomes to diverge cross-culturally.

We have attempted to reconcile these competing perspectives into a coherent theoretical account that explains ostensibly contradictory patterns in the relationship between economic development and women's educational and political outcomes. Our account draws from institutionalist theories in sociology to argue that global models and prescriptions resonate most broadly across culturally disparate societies when they are theorized in functional or instrumental terms—for example, as rational means for achieving sustained economic growth. In such cases, “logics of expected consequences” transcend major cultural divisions. Models rooted in notions of justice, fairness, or empowerment find less support globally due to cross-cultural variability in “logics of appropriateness,” which are linked inextricably to culture and identity.

On the basis of this theory, we hypothesized that cross-cultural differences in women's educational attainment would diminish with economic development, inasmuch as education for women is widely regarded as an investment in human capital and, hence, an important driver of long-term economic growth. Gender equality in political representation, conversely, is more often defined in terms of justice and fairness than instrumentality. Because these standards vary considerably across cultures, we expected to find persistent or even growing cross-cultural disparities in women's share of parliament as a function of economic development.

The results of our analyses lend support to our core hypothesis. Gender inequality in educational attainment shows the most cross-cultural progress. As countries develop economically, gender gaps in educational attainment tend to close, albeit at variable rates across cultural zones. Indeed, our analyses suggest that these gaps have narrowed over time independently of economic development.

Cross-cultural differences in women's share of parliament show a markedly different pattern. GDP per capita is associated with higher rates of representation only in the predominantly Catholic and, especially, Protestant nations of the West. In all other zones, the gender gap in political representation remains durable or even grows with economic development (and also, in most cases, as a simple function of time). Thus, if current trends hold, cross-cultural differences in women's share of parliament will increase.

Our theory contends that women's educational outcomes improve more readily than women's political outcomes because the former could be framed as promoting economic growth and hence rationalized as a collective good. Alternatively, inequality in educational attainment might also decline because girls’ and women's education can be tailored to local cultures in ways that women's participation in parliament cannot. After World War II, “women's education was endorsed almost exclusively by emphasizing the benefits of educated mothers” (Berkovitch and Bradley 1999:486). Although this view has changed in the West, many non-Western countries continue to justify women's education in this manner. In Saudi Arabia, for example, girls’ education falls under the purview of the Department of Religious Guidance Education, which ensures that schooling prepared them to become wives and mothers or, at best, teachers and nurses (Hamdan 2005). Thus, education can be reconciled with women's traditional roles as mothers and caretakers, meaning that improvements in educational attainment ratios are not necessarily emancipatory for women.

With this in mind, a few caveats are in order. The measures analyzed here are imperfect proxies of women's social and political outcomes around the world. Women's educational attainment relative to men ignores important stratification dynamics with respect to fields of study (Bradley 2000; Charles and Bradley 2002, 2009). Much of the world has attained or even exceeded gender parity in overall enrollment, but women continue to be overrepresented in lower-status and less remunerative fields. The strong negative effect of fertility on women's education attainment relative to men also suggests that family structures and access to birth control remain crucial preconditions for gender equality. Similarly, parliamentary seats may not translate into meaningful political influence; this was true, for example, in communist countries, where women were well represented in legislatures but rarely occupied positions of real power such as party leadership roles or Politburo appointments (Paxton, Hughes, and Green 2006). Much work exploring cross-cultural differences in more fine-grained outcomes remains to be done.

Despite these limitations, this study sheds new light on the effects of culture and development on women's outcomes. Economic development does not uniformly promote gender equality in politics and education. Broad cultural differences continue to shape the direction and pace of developmental trajectories in these domains.

APPENDIX A: SAMPLED COUNTRIES AND THEIR CULTURAL ZONE AFFILIATIONS

Albania Islamic (Other) Gambia, The Islamic (Other) Nigeria Islamic (Other) 
Algeria Islamic (Other) Georgia Orthodox Norway Protestant (Western) 
Angola Catholic (Other) Germany Protestant (Western) Oman Islamic (Middle Eastern) 
Argentina Catholic (Latin American) Ghana Animist Pakistan Islamic (Other) 
Armenia Orthodox Greece Orthodox Panama Catholic (Latin American) 
Australia Protestant (Western) Guatemala Catholic (Latin American) Papua New Guinea Protestant (non-Western) 
Austria Catholic (Western) Guinea Islamic (Other) Paraguay Catholic (Latin American) 
Azerbaijan Islamic (Other) Guinea-Bissau Animist Peru Catholic (Latin American) 
Bahrain Islamic (Middle Eastern) Guyana Hindu Philippines Catholic (Other) 
Bangladesh Islamic (Other) Honduras Catholic (Latin American) Poland Catholic (Western) 
Belarus Non-religious Hungary Catholic (Western) Portugal Catholic (Western) 
Belgium Catholic (Western) Iceland Protestant (Western) Qatar Islamic (Middle Eastern) 
Benin Animist India Hindu Romania Orthodox 
Bolivia Catholic (Latin American) Indonesia Islamic (Other) Russian Federation Non-religious 
Botswana Animist Iran Islamic (Middle Eastern) Rwanda Catholic (Other) 
Brazil Catholic (Latin American) Ireland Catholic (Western) Saudi Arabia Islamic (Middle Eastern) 
Bulgaria Orthodox Italy Catholic (Western) Senegal Islamic (Other) 
Burkina Faso Islamic (Other) Jamaica Protestant (non-Western) Sierra Leone Islamic (Other) 
Burundi Catholic (Other) Japan Buddhist Singapore Confucian 
Cambodia Buddhist Jordan Islamic (Middle Eastern) Slovakia Catholic (Western) 
Cameroon Catholic (Other) Kazakhstan Islamic (Other) Slovenia Catholic (Western) 
Canada Catholic (Western) Kenya Protestant (non-Western) Solomon Islands Protestant (non-Western) 
Central African Rep. Animist Korea, Rep. Buddhist South Africa Protestant (non-Western) 
Chad Islamic (Other) Kuwait Islamic (Middle Eastern) Spain Catholic (Western) 
Chile Catholic (Latin American) Kyrgyzstan Islamic (Other) Sri Lanka Buddhist 
China Confucian Lao PDR Buddhist Sudan Islamic (Other) 
Colombia Catholic (Latin American) Latvia Non-religious Suriname Hindu 
Comoros Islamic (Other) Lebanon Islamic (Middle Eastern) Swaziland Animist 
Congo, Dem. Rep. Catholic (Other) Lesotho Catholic (Other) Sweden Protestant (Western) 
Congo, Rep. Catholic (Other) Liberia Animist Switzerland Catholic (Western) 
Costa Rica Catholic (Latin American) Lithuania Catholic (Western) Syria Islamic (Middle Eastern) 
Cote d’Ivoire Animist Luxembourg Catholic (Western) Tajikistan Islamic (Other) 
Croatia Catholic (Western) Macedonia, TFYR Orthodox Tanzania Islamic (Other) 
Cuba Catholic (Latin American) Madagascar Animist Thailand Buddhist 
Cyprus Orthodox Malawi Protestant (non-Western) Togo Animist 
Czech Republic Non-religious Malaysia Islamic (Other) Trinidad and Tobago Catholic (Latin American) 
Denmark Protestant (Western) Mali Islamic (Other) Tunisia Islamic (Other) 
Djibouti Islamic (Other) Mauritania Islamic (Other) Turkey Islamic (Middle Eastern) 
Dominican Republic Catholic (Latin American) Mauritius Hindu Turkmenistan Islamic (Other) 
Ecuador Catholic (Latin American) Mexico Catholic (Latin American) Uganda Catholic (Other) 
Egypt Islamic (Middle Eastern) Moldova Orthodox Ukraine Orthodox 
El Salvador Catholic (Latin American) Mongolia Buddhist United Arab Emirates Islamic (Middle Eastern) 
Equatorial Guinea Catholic (Other) Morocco Islamic (Other) United Kingdom Protestant (Western) 
Eritrea Islamic (Other) Mozambique Animist United States Protestant (Western) 
Estonia Non-religious Namibia Protestant (non-Western) Uruguay Catholic (Latin American) 
Ethiopia Orthodox Nepal Hindu Uzbekistan Islamic (Other) 
Fiji Protestant (non-Western) Netherlands Catholic (Western) Venezuela Catholic (Latin American) 
Finland Protestant (Western) New Zealand Protestant (Western) Vietnam Non-religious 
France Catholic (Western) Nicaragua Catholic (Latin American) Zambia Protestant (non-Western) 
Gabon Catholic (Other) Niger Islamic (Other) Zimbabwe Animist 
Albania Islamic (Other) Gambia, The Islamic (Other) Nigeria Islamic (Other) 
Algeria Islamic (Other) Georgia Orthodox Norway Protestant (Western) 
Angola Catholic (Other) Germany Protestant (Western) Oman Islamic (Middle Eastern) 
Argentina Catholic (Latin American) Ghana Animist Pakistan Islamic (Other) 
Armenia Orthodox Greece Orthodox Panama Catholic (Latin American) 
Australia Protestant (Western) Guatemala Catholic (Latin American) Papua New Guinea Protestant (non-Western) 
Austria Catholic (Western) Guinea Islamic (Other) Paraguay Catholic (Latin American) 
Azerbaijan Islamic (Other) Guinea-Bissau Animist Peru Catholic (Latin American) 
Bahrain Islamic (Middle Eastern) Guyana Hindu Philippines Catholic (Other) 
Bangladesh Islamic (Other) Honduras Catholic (Latin American) Poland Catholic (Western) 
Belarus Non-religious Hungary Catholic (Western) Portugal Catholic (Western) 
Belgium Catholic (Western) Iceland Protestant (Western) Qatar Islamic (Middle Eastern) 
Benin Animist India Hindu Romania Orthodox 
Bolivia Catholic (Latin American) Indonesia Islamic (Other) Russian Federation Non-religious 
Botswana Animist Iran Islamic (Middle Eastern) Rwanda Catholic (Other) 
Brazil Catholic (Latin American) Ireland Catholic (Western) Saudi Arabia Islamic (Middle Eastern) 
Bulgaria Orthodox Italy Catholic (Western) Senegal Islamic (Other) 
Burkina Faso Islamic (Other) Jamaica Protestant (non-Western) Sierra Leone Islamic (Other) 
Burundi Catholic (Other) Japan Buddhist Singapore Confucian 
Cambodia Buddhist Jordan Islamic (Middle Eastern) Slovakia Catholic (Western) 
Cameroon Catholic (Other) Kazakhstan Islamic (Other) Slovenia Catholic (Western) 
Canada Catholic (Western) Kenya Protestant (non-Western) Solomon Islands Protestant (non-Western) 
Central African Rep. Animist Korea, Rep. Buddhist South Africa Protestant (non-Western) 
Chad Islamic (Other) Kuwait Islamic (Middle Eastern) Spain Catholic (Western) 
Chile Catholic (Latin American) Kyrgyzstan Islamic (Other) Sri Lanka Buddhist 
China Confucian Lao PDR Buddhist Sudan Islamic (Other) 
Colombia Catholic (Latin American) Latvia Non-religious Suriname Hindu 
Comoros Islamic (Other) Lebanon Islamic (Middle Eastern) Swaziland Animist 
Congo, Dem. Rep. Catholic (Other) Lesotho Catholic (Other) Sweden Protestant (Western) 
Congo, Rep. Catholic (Other) Liberia Animist Switzerland Catholic (Western) 
Costa Rica Catholic (Latin American) Lithuania Catholic (Western) Syria Islamic (Middle Eastern) 
Cote d’Ivoire Animist Luxembourg Catholic (Western) Tajikistan Islamic (Other) 
Croatia Catholic (Western) Macedonia, TFYR Orthodox Tanzania Islamic (Other) 
Cuba Catholic (Latin American) Madagascar Animist Thailand Buddhist 
Cyprus Orthodox Malawi Protestant (non-Western) Togo Animist 
Czech Republic Non-religious Malaysia Islamic (Other) Trinidad and Tobago Catholic (Latin American) 
Denmark Protestant (Western) Mali Islamic (Other) Tunisia Islamic (Other) 
Djibouti Islamic (Other) Mauritania Islamic (Other) Turkey Islamic (Middle Eastern) 
Dominican Republic Catholic (Latin American) Mauritius Hindu Turkmenistan Islamic (Other) 
Ecuador Catholic (Latin American) Mexico Catholic (Latin American) Uganda Catholic (Other) 
Egypt Islamic (Middle Eastern) Moldova Orthodox Ukraine Orthodox 
El Salvador Catholic (Latin American) Mongolia Buddhist United Arab Emirates Islamic (Middle Eastern) 
Equatorial Guinea Catholic (Other) Morocco Islamic (Other) United Kingdom Protestant (Western) 
Eritrea Islamic (Other) Mozambique Animist United States Protestant (Western) 
Estonia Non-religious Namibia Protestant (non-Western) Uruguay Catholic (Latin American) 
Ethiopia Orthodox Nepal Hindu Uzbekistan Islamic (Other) 
Fiji Protestant (non-Western) Netherlands Catholic (Western) Venezuela Catholic (Latin American) 
Finland Protestant (Western) New Zealand Protestant (Western) Vietnam Non-religious 
France Catholic (Western) Nicaragua Catholic (Latin American) Zambia Protestant (non-Western) 
Gabon Catholic (Other) Niger Islamic (Other) Zimbabwe Animist 

APPENDIX B: SUPPLEMENTARY ANALYSES AND ROBUSTNESS CHECKS

This appendix presents supplementary analyses for our article, “Progress without Progressives? The Effects of Development on Women's Educational and Political Equality in Cultural Context, 1980 to 2010.” It is divided into three sections. The first section uses alternative cultural and regional indicators to test the robustness of our findings. We analyze five sets of indicators: Samuel Huntington's civilizations framework, Christian Welzel's cultural zone categories, the United Nations macro geographical region and sub-region classifications, and indicators denoting countries’ colonial legacies. Using each set of indicators, we re-analyze the models presented in Tables 2 (cross-cultural variation in gendered educational attainment ratios), 3 (women's share of parliamentarians), and 4 (the relationship between economic development on women's outcomes as conditioned by cultural zones). We also plot coefficient estimates and marginal effects for each set of indicators, akin to Figures 1, 2, and 3.

In the second section, we return to our framework for measuring culture and re-analyze our models on samples split by membership in the Organization for Economic Co-operation and Development (OECD). These analyses provide an alternative way to evaluate whether broad levels of economic development—conceptualized here as a binary variable distinguishing “highly developed” from “less developed” countries—conditions the relationship between cultural affiliations (and control variables) and women's outcomes.

The final section presents analyses that diagnose whether per capita gross domestic product (GDP) is endogenous to women's outcomes in our analyses. An important part of our analysis in the main manuscript considers whether GDP per capita conditions the relationship between cultural affiliation, on the one hand, and women's educational and political equality, on the other hand. There is reason to believe, however, that these analyses may suffer from endogeneity bias. Although we assume that economic development is causally prior to women's share of parliamentarians and gender equality in educational attainment, it is plausible that gender equity contributes to economic development. To mitigate this potential source of bias in our analyses, we estimate two-stage models with instrumental variables to diagnose and assess the presence of reverse causality.

SECTION 1: ASSESSING ALTERNATIVE CULTURAL AND REGIONAL INDICATORS

To operationalize cultural differences, we divided countries into mutually exclusive and exhaustive zones based primarily on their predominant religious traditions, as measured using the World Religion Dataset, and secondarily on geographical region. Yet culture is a notoriously difficult concept to measure, and scholars debate how best to operationalize it in comparative macro-sociological research. Here we conduct robustness checks using alternative frameworks for categorizing the countries according to their cultural affiliations, geographical locations, and colonial legacies. More specifically, we assess (1) Samuel Huntington's civilizations framework, (2) Christian Welzel's cultural zone categories, (3) the United Nations sub-regional classification scheme, (4) the consolidated U.N. continental region classification scheme, and (5) indicators denoting countries’ colonial legacies.

As noted in our article, Samuel Huntington (1996) allocates countries into several “civilizational” categories in his book, The Clash of Civilizations. He identifies nine such civilizations, along with a residual category for difficult-to-classify countries. These civilizations are African, Buddhist, Hindu, Islamic, Japanese, Latin American, Orthodox, Sinic, and Western. Although Huntington regards religion as the defining feature of civilizational identity, he does not ground his framework in objective measures of religious affiliation, as we do.

Drawing on the “revised modernization” theoretical tradition of Ronald Inglehart and his collaborators, Christian Welzel (2013) develops a somewhat different cultural classification scheme in his book, Freedom Rising. He begins by dividing the world into “Western” and “Eastern” zones, and then creates subcategories based on religious and historical factors. For example, Western countries are classified as “Reformed West” (northern European countries that served as centers of the Protestant Reformation), “Old West” (historically Catholic countries in Europe that industrialized and democratized later than the Reformed West), “New West” (predominantly British settler countries that extended the Reformed West overseas), and the “Returned West” (central and eastern European countries formerly under the sway of the Soviet Union). Eastern zones include Indic countries (centered historically on India), Islamic countries (centered on the Middle East), Sinic countries (centered on China), and Orthodox countries (centered on Russia). Two additional zones—Latin America and sub-Saharan—possess historical, colonial, cultural, and developmental qualities that distinguish them from both Western and Eastern countries.

In addition to these frameworks for allocating countries into distinct cultural or civilizational zones, we also examine patterns of regional variation in women's educational and political outcomes. Our own framework acknowledges that regional location, as well as cultural affiliation, shapes a host of social outcomes, including outcomes related to gender equality. In addition to the profound influence of religion, macro-cultural differences reflect colonial legacies, developmental trajectories, and diffusion processes that cluster historically by geographical region. For this reason, we distinguished between Western and non-Western Protestant nations; Western, Latin American, and other Catholic countries; and Islamic societies in the Middle East and elsewhere. Here we focus exclusively on regional distinctions, to the exclusion of cultural considerations. We use two regional classifications developed by the United Nations. The first classification divides countries into the following 20 fine-grained regions: Western, Eastern, Middle, Northern, and Southern Africa; Central, Northern, and South America; the Caribbean; Western, Eastern, Central, Southern, and South-Eastern Asia; Northern, Southern, Western, and Eastern Europe; Australia and New Zealand; and Melanesia. A coarser regional framework consolidates these classifications into the following six continental regions: Africa, Latin America, North America, Asia, Europe, and Oceania.

Finally, we recognize that cultural and institutional patterns outside the West often reflect colonial legacies. Using indicators from the Authoritarian Regimes Dataset (Wahman, Teorell, and Hadenius 2013), we identify countries based on the Western power that colonized them: the Netherlands, Spain, the United States, Britain, France, Portugal, Belgium, and Australia, with a separate code for countries that were never colonized. In cases where a country was colonized by several powers, only the last one is counted, but then only if that colonial regime lased for a decade or longer. British settler colonies in North America and Australasia—Canada, the United States, Australia, and New Zealand—are excluded, as are countries that were colonized exclusively by non-Western powers (e.g., Japan).

Using each of these coding schemes we estimate three sets of analyses: one that analyzes variation in gendered educational attainment ratios, another that analyzes variation in women's share of parliamentarians, and a third that examines how each block of indicators conditions the effect of economic development (operationalized as per capita GDP, standardized to have a mean of 0 and standard deviation of 1) on women's outcomes. These analyses correspond to those presented in Tables 2, 3, and 4 of our article. Tables A1 through A15 (three sets of analyses for each of five cultural, regional, and colonial frameworks) use the same heteroskedastic-robust OLS estimators as the main analyses to estimate these models. To facilitate interpretation, we also plot coefficient estimates and marginal effects from these analyses, as we did in our article. Figures A1 through A8 present these plots.

TABLE A1.
Heteroskedastic-robust OLS estimates for the effect of Huntington's civilization indicators, economic development, and control variables on gendered educational attainment ratios, 1980 to 2010
Culture onlyNet of GDP/capitaFully specified
Coeff.(S.E.)Coeff.(S.E.)Coeff.(S.E.)
Africana −36.845*** (4.522) −15.475* (6.184) −1.752 (6.331) 
Buddhista −19.610* (7.749) 1.543 (8.223) −1.085 (7.046) 
Hindua −60.962*** (7.812) −34.713*** (8.117) −34.961*** (7.687) 
Islamica −37.063*** (3.748) −20.213*** (4.829) −9.623 (5.489) 
Japanesea 4.975** (1.730) 1.014 (1.397) 3.803 (2.451) 
Latin Americana −2.260 (2.760) 9.852** (3.498) 14.825*** (3.681) 
Orthodoxa −1.262 (2.200) 9.305* (4.516) −.932 (3.072) 
Othera −12.747 (7.827) 3.077 (6.203) −3.041 (4.419) 
Sinica −20.225*** (3.397) −8.710 (5.613) −21.448** (7.224) 
GDP/capita (2005 US$, ln  6.794*** (1.224) .116 (1.618) 
Democracy/Autocracy     .163 (.208) 
OPEC member (1 = yes)     5.250 (4.046) 
Trade (% GDP)     .050 (.033) 
International war (1 = yes)     5.550 (2.991) 
Civil war (1 = yes)     .522 (3.570) 
Fertility rate     −9.392*** (1.178) 
Female population (%)     −.047 (.571) 
Religious fractionalization     11.631* (4.569) 
Communist (1 = yes)     −4.759 (4.073) 
Women's INGO linkages     −.013 (.047) 
CEDAW member (1 = yes)     −2.070 (2.148) 
Women's share of parliamentarians     .197** (.073) 
Constant 84.391*** (2.086) 18.380 (12.267) 99.873* (40.036) 
F 23.131  34.211  66.113  
df 39  40  52  
Adjusted R-squared .531  .617  .758  
Culture onlyNet of GDP/capitaFully specified
Coeff.(S.E.)Coeff.(S.E.)Coeff.(S.E.)
Africana −36.845*** (4.522) −15.475* (6.184) −1.752 (6.331) 
Buddhista −19.610* (7.749) 1.543 (8.223) −1.085 (7.046) 
Hindua −60.962*** (7.812) −34.713*** (8.117) −34.961*** (7.687) 
Islamica −37.063*** (3.748) −20.213*** (4.829) −9.623 (5.489) 
Japanesea 4.975** (1.730) 1.014 (1.397) 3.803 (2.451) 
Latin Americana −2.260 (2.760) 9.852** (3.498) 14.825*** (3.681) 
Orthodoxa −1.262 (2.200) 9.305* (4.516) −.932 (3.072) 
Othera −12.747 (7.827) 3.077 (6.203) −3.041 (4.419) 
Sinica −20.225*** (3.397) −8.710 (5.613) −21.448** (7.224) 
GDP/capita (2005 US$, ln  6.794*** (1.224) .116 (1.618) 
Democracy/Autocracy     .163 (.208) 
OPEC member (1 = yes)     5.250 (4.046) 
Trade (% GDP)     .050 (.033) 
International war (1 = yes)     5.550 (2.991) 
Civil war (1 = yes)     .522 (3.570) 
Fertility rate     −9.392*** (1.178) 
Female population (%)     −.047 (.571) 
Religious fractionalization     11.631* (4.569) 
Communist (1 = yes)     −4.759 (4.073) 
Women's INGO linkages     −.013 (.047) 
CEDAW member (1 = yes)     −2.070 (2.148) 
Women's share of parliamentarians     .197** (.073) 
Constant 84.391*** (2.086) 18.380 (12.267) 99.873* (40.036) 
F 23.131  34.211  66.113  
df 39  40  52  
Adjusted R-squared .531  .617  .758  
a

Western civilization is the omitted reference group.

Standard errors, adjusted for first-order autocorrelation, in parentheses.

Fixed effects for years included in the analyses but not reported.

*p < .05, **p < .01, ***p < .001 (two-tailed).

N(country-year observations) = 3,870; N(countries) = 153.

TABLE A2.
Heteroskedastic-robust OLS estimates for the effect of Huntington's civilization indicators, economic development, and control variables on women's share of parliamentarians, 1980 to 2010
Culture onlyNet of GDP/capitaFully specified
Coeff.(S.E.)Coeff.(S.E.)Coeff.(S.E.)
Africana −7.039*** (1.878) −6.230** (2.308) −1.184 (1.977) 
Buddhista −8.853*** (2.428) −8.053** (3.003) −8.631*** (2.124) 
Hindua −8.911*** (1.674) −7.917** (2.568) −1.999 (2.737) 
Islamica −11.127*** (1.846) −10.489*** (2.279) −4.458* (2.031) 
Japanesea −8.742*** (1.639) −8.891*** (1.670) −8.505*** (1.518) 
Latin Americana −5.913** (2.158) −5.454* (2.311) −5.241** (1.994) 
Orthodoxa −8.555*** (2.293) −8.155*** (2.381) −6.512*** (1.945) 
Othera −3.421 (3.189) −2.822 (3.534) 2.772 (2.913) 
Sinica −3.819 (4.039) −3.383 (4.253) −5.615* (2.492) 
GDP/capita (2005 US$, ln  .257 (.546) .384 (.541) 
Democracy/Autocracy     −.234** (.076) 
OPEC member (1 = yes)     −2.148 (1.394) 
Trade (% GDP)     .007 (.009) 
International war (1 = yes)     −1.501 (2.329) 
Civil war (1 = yes)     −1.311 (1.011) 
Fertility rate     .594 (.426) 
Female population (%)     .277 (.236) 
Religious fractionalization     −2.470 (2.922) 
Communist (1 = yes)     14.600*** (2.156) 
Women's INGO linkages     .071** (.023) 
CEDAW member (1 = yes)     1.808* (.922) 
Proportional representation (1 = yes)     4.262*** (1.017) 
Years of female suffrage     .065* (.032) 
Gendered educational attainment ratio     .060* (.026) 
Constant 13.265*** (1.624) 10.767* (5.407) −18.842 (14.093) 
F 10.559  10.356  12.680  
df 39  40  54  
Adjusted R-squared .283  .284  .457  
Culture onlyNet of GDP/capitaFully specified
Coeff.(S.E.)Coeff.(S.E.)Coeff.(S.E.)
Africana −7.039*** (1.878) −6.230** (2.308) −1.184 (1.977) 
Buddhista −8.853*** (2.428) −8.053** (3.003) −8.631*** (2.124) 
Hindua −8.911*** (1.674) −7.917** (2.568) −1.999 (2.737) 
Islamica −11.127*** (1.846) −10.489*** (2.279) −4.458* (2.031) 
Japanesea −8.742*** (1.639) −8.891*** (1.670) −8.505*** (1.518) 
Latin Americana −5.913** (2.158) −5.454* (2.311) −5.241** (1.994) 
Orthodoxa −8.555*** (2.293) −8.155*** (2.381) −6.512*** (1.945) 
Othera −3.421 (3.189) −2.822 (3.534) 2.772 (2.913) 
Sinica −3.819 (4.039) −3.383 (4.253) −5.615* (2.492) 
GDP/capita (2005 US$, ln  .257 (.546) .384 (.541) 
Democracy/Autocracy     −.234** (.076) 
OPEC member (1 = yes)     −2.148 (1.394) 
Trade (% GDP)     .007 (.009) 
International war (1 = yes)     −1.501 (2.329) 
Civil war (1 = yes)     −1.311 (1.011) 
Fertility rate     .594 (.426) 
Female population (%)     .277 (.236) 
Religious fractionalization     −2.470 (2.922) 
Communist (1 = yes)     14.600*** (2.156) 
Women's INGO linkages     .071** (.023) 
CEDAW member (1 = yes)     1.808* (.922) 
Proportional representation (1 = yes)     4.262*** (1.017) 
Years of female suffrage     .065* (.032) 
Gendered educational attainment ratio     .060* (.026) 
Constant 13.265*** (1.624) 10.767* (5.407) −18.842 (14.093) 
F 10.559  10.356  12.680  
df 39  40  54  
Adjusted R-squared .283  .284  .457  
a

Western civilization is the omitted reference group.

Standard errors, adjusted for first-order autocorrelation, in parentheses.

Fixed effects for years included in the analyses but not reported.

*p < .05, **p < .01, ***p < .001 (two-tailed).

N(country-year observations) = 3,870; N(countries) = 153.

TABLE A3.
Heteroskedastic-robust OLS estimates for the effect of economic development on gendered educational and political outcomes, as conditioned by Huntington's civilizational indicators, 1980 to 2010
Gendered educational attainment ratioWomen's share of parliamentarians
Coeff.(S.E.)Coeff.(S.E.)
GDP/capita (2005 US$, ln)b −2.660 (4.686) 5.706*** (1.414) 
Africana −2.341 (7.709) 2.354 (1.463) 
 x GDP/capita (2005 US$, ln7.730 (5.940) −7.007*** (1.786) 
Buddhista 5.351 (11.397) −4.361 (3.177) 
 x GDP/capita (2005 US$, ln14.691 (10.228) −6.883* (3.200) 
Hindua −23.424* (10.705) −12.194* (6.144) 
 x GDP/capita (2005 US$, ln15.354** (5.297) −17.963*** (4.854) 
Islamica −13.079 (6.826) −2.071 (1.826) 
 x GDP/capita (2005 US$, ln2.942 (5.782) −9.514*** (1.902) 
Japanesea 1.412 (12.144) 33.140*** (5.124) 
 x GDP/capita (2005 US$, ln2.535 (8.500) −28.310*** (3.493) 
Latin Americana 11.773 (6.179) .176 (1.711) 
 x GDP/capita (2005 US$, ln−.844 (6.321) −5.421 (2.868) 
Orthodoxa −2.592 (5.147) −1.118 (1.862) 
 x GDP/capita (2005 US$, ln.421 (4.051) −6.348*** (1.919) 
Othera −4.266 (5.538) 7.324** (2.375) 
 x GDP/capita (2005 US$, ln9.759 (5.222) −3.488 (2.276) 
Sinica −22.594** (7.471) −1.060 (1.916) 
 x GDP/capita (2005 US$, ln−4.111 (6.957) −7.872*** (2.005) 
Democracy/Autocracy .123 (.199) −.265*** (.074) 
OPEC member (1 = yes) 5.160 (4.228) −.841 (1.375) 
Trade (% GDP) .052 (.033) .013 (.010) 
International war (1 = yes) 3.879 (2.219) −2.111 (2.188) 
Civil war (1 = yes) .681 (2.991) −.732 (1.088) 
Fertility rate −8.867*** (1.297) .312 (.386) 
Female population (%) −.036 (.680) −.149 (.248) 
Religious fractionalization 11.771* (4.890) −1.209 (2.925) 
Communist (1 = yes) −6.554 (3.903) 13.218*** (2.295) 
Women's INGO linkages .014 (.052) .050* (.022) 
CEDAW member (1 = yes) −1.789 (2.119) 1.559 (.888) 
Proportional representation (1 = yes)   4.166*** (1.061) 
Years of female suffrage   .041 (.030) 
Gendered educational attainment ratio   .067* (.027) 
Women's share of parliamentarians .228** (.077)   
Constant 100.689** (37.276) 1.271 (12.621) 
F 92.698  40.029  
df 61  63  
Adjusted R-squared .768  .505  
Gendered educational attainment ratioWomen's share of parliamentarians
Coeff.(S.E.)Coeff.(S.E.)
GDP/capita (2005 US$, ln)b −2.660 (4.686) 5.706*** (1.414) 
Africana −2.341 (7.709) 2.354 (1.463) 
 x GDP/capita (2005 US$, ln7.730 (5.940) −7.007*** (1.786) 
Buddhista 5.351 (11.397) −4.361 (3.177) 
 x GDP/capita (2005 US$, ln14.691 (10.228) −6.883* (3.200) 
Hindua −23.424* (10.705) −12.194* (6.144) 
 x GDP/capita (2005 US$, ln15.354** (5.297) −17.963*** (4.854) 
Islamica −13.079 (6.826) −2.071 (1.826) 
 x GDP/capita (2005 US$, ln2.942 (5.782) −9.514*** (1.902) 
Japanesea 1.412 (12.144) 33.140*** (5.124) 
 x GDP/capita (2005 US$, ln2.535 (8.500) −28.310*** (3.493) 
Latin Americana 11.773 (6.179) .176 (1.711) 
 x GDP/capita (2005 US$, ln−.844 (6.321) −5.421 (2.868) 
Orthodoxa −2.592 (5.147) −1.118 (1.862) 
 x GDP/capita (2005 US$, ln.421 (4.051) −6.348*** (1.919) 
Othera −4.266 (5.538) 7.324** (2.375) 
 x GDP/capita (2005 US$, ln9.759 (5.222) −3.488 (2.276) 
Sinica −22.594** (7.471) −1.060 (1.916) 
 x GDP/capita (2005 US$, ln−4.111 (6.957) −7.872*** (2.005) 
Democracy/Autocracy .123 (.199) −.265*** (.074) 
OPEC member (1 = yes) 5.160 (4.228) −.841 (1.375) 
Trade (% GDP) .052 (.033) .013 (.010) 
International war (1 = yes) 3.879 (2.219) −2.111 (2.188) 
Civil war (1 = yes) .681 (2.991) −.732 (1.088) 
Fertility rate −8.867*** (1.297) .312 (.386) 
Female population (%) −.036 (.680) −.149 (.248) 
Religious fractionalization 11.771* (4.890) −1.209 (2.925) 
Communist (1 = yes) −6.554 (3.903) 13.218*** (2.295) 
Women's INGO linkages .014 (.052) .050* (.022) 
CEDAW member (1 = yes) −1.789 (2.119) 1.559 (.888) 
Proportional representation (1 = yes)   4.166*** (1.061) 
Years of female suffrage   .041 (.030) 
Gendered educational attainment ratio   .067* (.027) 
Women's share of parliamentarians .228** (.077)   
Constant 100.689** (37.276) 1.271 (12.621) 
F 92.698  40.029  
df 61  63  
Adjusted R-squared .768  .505  
a

Western civilization is the omitted reference group.

b

Standardized to have mean = 0 and standard deviation = 1.

Standard errors, adjusted for first-order autocorrelation, in parentheses.

Fixed effects for years included in the analyses but not reported.

*p < .05, **p < .01, ***p < .001 (two-tailed).

N(country-year observations) = 3,870; N(countries) = 153.

TABLE A4.
Heteroskedastic-robust OLS estimates for the effect of Welzel's cultural zone indicators, economic development, and control variables on gendered educational attainment ratios, 1980 to 2010
Culture onlyNet of GDP/capitaFully specified
Coeff.(S.E.)Coeff.(S.E.)Coeff.(S.E.)
New Westa 1.725 (1.454) 3.478* (1.709) 6.243* (2.766) 
Old Westa −5.271** (1.679) −2.726 (2.048) −4.577 (2.852) 
Returned Westa −1.597 (2.023) 8.697* (3.772) 3.421 (3.670) 
Orthodox Easta −4.530* (2.079) 16.147** (6.100) 13.955** (5.164) 
Indic Easta −32.409*** (7.634) −10.609 (7.930) −4.915 (7.454) 
Islamic Easta −39.994*** (4.996) −22.878** (8.464) −9.867 (7.096) 
Sinic Easta −16.574** (6.318) −1.097 (5.087) −.708 (4.143) 
Latin Americaa −6.203* (2.990) 8.645 (4.680) 12.786** (4.050) 
Sub-Saharan Africaa −39.004*** (5.742) −10.472 (9.512) 6.274 (7.740) 
GDP/capita (2005 US$, ln  6.666*** (1.798) 4.696*** (1.391) 
Democracy/Autocracy     .375 (.210) 
OPEC member (1 = yes)     4.371 (5.086) 
Trade (% GDP)     .007 (.022) 
International war (1 = yes)     −2.972 (3.292) 
Civil war (1 = yes)     11.382 (7.223) 
Fertility rate     −4.489** (1.433) 
Female population (%)     2.578** (.892) 
Religious fractionalization     −3.481 (4.642) 
Communist (1 = yes)     7.159 (4.112) 
Women's INGO linkages     .020 (.040) 
CEDAW member (1 = yes)     2.446 (2.009) 
Women's share of parliamentarians     .056 (.067) 
Constant 88.365*** (1.613) 19.423 (18.470) −84.125 (52.836) 
F 17.377  32.163  62.187  
df 39  40  52  
Adjusted R-squared .667  .739  .825  
Culture onlyNet of GDP/capitaFully specified
Coeff.(S.E.)Coeff.(S.E.)Coeff.(S.E.)
New Westa 1.725 (1.454) 3.478* (1.709) 6.243* (2.766) 
Old Westa −5.271** (1.679) −2.726 (2.048) −4.577 (2.852) 
Returned Westa −1.597 (2.023) 8.697* (3.772) 3.421 (3.670) 
Orthodox Easta −4.530* (2.079) 16.147** (6.100) 13.955** (5.164) 
Indic Easta −32.409*** (7.634) −10.609 (7.930) −4.915 (7.454) 
Islamic Easta −39.994*** (4.996) −22.878** (8.464) −9.867 (7.096) 
Sinic Easta −16.574** (6.318) −1.097 (5.087) −.708 (4.143) 
Latin Americaa −6.203* (2.990) 8.645 (4.680) 12.786** (4.050) 
Sub-Saharan Africaa −39.004*** (5.742) −10.472 (9.512) 6.274 (7.740) 
GDP/capita (2005 US$, ln  6.666*** (1.798) 4.696*** (1.391) 
Democracy/Autocracy     .375 (.210) 
OPEC member (1 = yes)     4.371 (5.086) 
Trade (% GDP)     .007 (.022) 
International war (1 = yes)     −2.972 (3.292) 
Civil war (1 = yes)     11.382 (7.223) 
Fertility rate     −4.489** (1.433) 
Female population (%)     2.578** (.892) 
Religious fractionalization     −3.481 (4.642) 
Communist (1 = yes)     7.159 (4.112) 
Women's INGO linkages     .020 (.040) 
CEDAW member (1 = yes)     2.446 (2.009) 
Women's share of parliamentarians     .056 (.067) 
Constant 88.365*** (1.613) 19.423 (18.470) −84.125 (52.836) 
F 17.377  32.163  62.187  
df 39  40  52  
Adjusted R-squared .667  .739  .825  
a

Reformed Western countries are the omitted reference group.

Standard errors, adjusted for first-order autocorrelation, in parentheses.

Fixed effects for years included in the analyses but not reported.

*p < .05, **p < .01, ***p < .001 (two-tailed).

N(country-year observations) = 2,234; N(countries) = 85.

TABLE A5.
Heteroskedastic-robust OLS estimates for the effect of Welzel's cultural zone indicators, economic development, and control variables on women's share of parliamentarians, 1980 to 2010
Culture onlyNet of GDP/capitaFully specified
Coeff.(S.E.)Coeff.(S.E.)Coeff.(S.E.)
New Westa −10.559** (3.968) −10.740** (3.948) −7.744* (3.854) 
Old Westa −13.953*** (3.500) −14.216*** (3.538) −13.200*** (3.050) 
Returned Westa −14.184*** (3.016) −15.245*** (3.210) −15.162*** (2.586) 
Orthodox Easta −17.378*** (3.483) −19.508*** (4.173) −16.745*** (3.224) 
Indic Easta −17.937*** (3.072) −20.184*** (3.789) −13.308*** (3.224) 
Islamic Easta −24.415*** (3.027) −26.179*** (3.477) −21.184*** (3.351) 
Sinic Easta −13.367** (4.657) −14.962*** (4.440) −18.571*** (3.035) 
Latin Americaa −16.846*** (3.196) −18.375*** (3.619) −15.204*** (3.183) 
Sub-Saharan Africaa −14.303*** (3.596) −17.243*** (4.397) −9.604* (4.061) 
GDP/capita (2005 US$, ln  −.687 (.628) 1.038 (.615) 
Democracy/Autocracy     −.237 (.125) 
OPEC member (1 = yes)     −2.997 (1.776) 
Trade (% GDP)     .007 (.010) 
International war (1 = yes)     −1.334 (1.755) 
Civil war (1 = yes)     1.834 (1.189) 
Fertility rate     .990 (.572) 
Female population (%)     .192 (.506) 
Religious fractionalization     −4.449 (4.783) 
Communist (1 = yes)     19.487*** (2.106) 
Women's INGO linkages     .003 (.031) 
CEDAW member (1 = yes)     1.649 (1.425) 
Proportional representation (1 = yes)     3.306 (1.779) 
Years of female suffrage     .060 (.043) 
Gendered educational attainment ratio     .034 (.041) 
Constant 22.180*** (3.140) 29.283*** (7.270) −7.550 (26.315) 
F 17.722  21.370  26.653  
df 39  40  54  
Adjusted R-squared .492  .495  .617  
Culture onlyNet of GDP/capitaFully specified
Coeff.(S.E.)Coeff.(S.E.)Coeff.(S.E.)
New Westa −10.559** (3.968) −10.740** (3.948) −7.744* (3.854) 
Old Westa −13.953*** (3.500) −14.216*** (3.538) −13.200*** (3.050) 
Returned Westa −14.184*** (3.016) −15.245*** (3.210) −15.162*** (2.586) 
Orthodox Easta −17.378*** (3.483) −19.508*** (4.173) −16.745*** (3.224) 
Indic Easta −17.937*** (3.072) −20.184*** (3.789) −13.308*** (3.224) 
Islamic Easta −24.415*** (3.027) −26.179*** (3.477) −21.184*** (3.351) 
Sinic Easta −13.367** (4.657) −14.962*** (4.440) −18.571*** (3.035) 
Latin Americaa −16.846*** (3.196) −18.375*** (3.619) −15.204*** (3.183) 
Sub-Saharan Africaa −14.303*** (3.596) −17.243*** (4.397) −9.604* (4.061) 
GDP/capita (2005 US$, ln  −.687 (.628) 1.038 (.615) 
Democracy/Autocracy     −.237 (.125) 
OPEC member (1 = yes)     −2.997 (1.776) 
Trade (% GDP)     .007 (.010) 
International war (1 = yes)     −1.334 (1.755) 
Civil war (1 = yes)     1.834 (1.189) 
Fertility rate     .990 (.572) 
Female population (%)     .192 (.506) 
Religious fractionalization     −4.449 (4.783) 
Communist (1 = yes)     19.487*** (2.106) 
Women's INGO linkages     .003 (.031) 
CEDAW member (1 = yes)     1.649 (1.425) 
Proportional representation (1 = yes)     3.306 (1.779) 
Years of female suffrage     .060 (.043) 
Gendered educational attainment ratio     .034 (.041) 
Constant 22.180*** (3.140) 29.283*** (7.270) −7.550 (26.315) 
F 17.722  21.370  26.653  
df 39  40  54  
Adjusted R-squared .492  .495  .617  
a

Reformed Western countries are the omitted reference group.

Standard errors, adjusted for first-order autocorrelation, in parentheses.

Fixed effects for years included in the analyses but not reported.

*p < .05, **p < .01, ***p < .001 (two-tailed).

N(country-year observations) = 2,234; N(countries) = 85.

TABLE A6.
Heteroskedastic-robust OLS estimates for the effect of economic development on gendered educational and political outcomes, as conditioned by Welzel's cultural zone indicators, 1980 to 2010
Gendered educational attainment ratioWomen's share of parliamentarians
Coeff.(S.E.)Coeff.(S.E.)
GDP/capita (2005 US$, ln)b −4.563 (6.387) 3.895 (11.226) 
New Westa 8.393 (14.131) 15.943 (19.573) 
 x GDP/capita (2005 US$, ln−2.203 (9.845) −15.204 (11.811) 
Old Westa −23.027* (9.707) −12.077 (19.124) 
 x GDP/capita (2005 US$, ln10.803 (6.477) −.572 (11.192) 
Returned Westa −4.059 (9.664) −9.018 (19.043) 
 x GDP/capita (2005 US$, ln−3.150 (6.363) −5.957 (11.256) 
Orthodox Easta −8.969 (9.898) −12.931 (19.163) 
 x GDP/capita (2005 US$, ln−1.139 (6.822) −.310 (11.278) 
Indic Easta −25.175* (10.832) −11.139 (19.731) 
 x GDP/capita (2005 US$, ln19.253* (8.149) −5.049 (11.586) 
Islamic Easta −35.895** (11.593) −17.812 (20.024) 
 x GDP/capita (2005 US$, ln6.336 (7.985) −4.418 (11.499) 
Sinic Easta −21.743* (10.067) −15.552 (19.064) 
 x GDP/capita (2005 US$, ln13.242 (6.889) −1.459 (11.433) 
Latin Americaa −12.335 (11.768) −14.322 (19.505) 
 x GDP/capita (2005 US$, ln18.064 (10.830) 7.341 (11.430) 
Sub-Saharan Africaa −8.190 (11.408) −7.010 (19.332) 
 x GDP/capita (2005 US$, ln23.696** (8.094) −2.493 (11.978) 
Democracy/Autocracy .389* (.190) −.284* (.116) 
OPEC member (1 = yes) 3.910 (5.232) −3.513 (2.018) 
Trade (% GDP) −.012 (.020) .016 (.014) 
International war (1 = yes) −3.612 (3.363) −.071 (1.524) 
Civil war (1 = yes) 11.783 (7.208) 1.770 (1.275) 
Fertility rate −2.842 (1.650) 1.264* (.621) 
Female population (%) 2.043** (.791) .193 (.533) 
Religious fractionalization −8.322 (4.826) −4.130 (4.672) 
Communist (1 = yes) 8.338** (2.626) 20.252*** (2.660) 
Women's INGO linkages .025 (.042) .005 (.030) 
CEDAW member (1 = yes) 3.348 (1.778) 1.496 (1.455) 
Proportional representation (1 = yes)   4.138* (1.694) 
Years of female suffrage   .048 (.048) 
Gendered educational attainment ratio   .042 (.050) 
Women's share of parliamentarians .034 (.065)   
Constant −1.385 (42.536) −5.196 (31.708) 
F 91.442  28.579  
df 61  63  
Adjusted R-squared .852  .635  
Gendered educational attainment ratioWomen's share of parliamentarians
Coeff.(S.E.)Coeff.(S.E.)
GDP/capita (2005 US$, ln)b −4.563 (6.387) 3.895 (11.226) 
New Westa 8.393 (14.131) 15.943 (19.573) 
 x GDP/capita (2005 US$, ln−2.203 (9.845) −15.204 (11.811) 
Old Westa −23.027* (9.707) −12.077 (19.124) 
 x GDP/capita (2005 US$, ln10.803 (6.477) −.572 (11.192) 
Returned Westa −4.059 (9.664) −9.018 (19.043) 
 x GDP/capita (2005 US$, ln−3.150 (6.363) −5.957 (11.256) 
Orthodox Easta −8.969 (9.898) −12.931 (19.163) 
 x GDP/capita (2005 US$, ln−1.139 (6.822) −.310 (11.278) 
Indic Easta −25.175* (10.832) −11.139 (19.731) 
 x GDP/capita (2005 US$, ln19.253* (8.149) −5.049 (11.586) 
Islamic Easta −35.895** (11.593) −17.812 (20.024) 
 x GDP/capita (2005 US$, ln6.336 (7.985) −4.418 (11.499) 
Sinic Easta −21.743* (10.067) −15.552 (19.064) 
 x GDP/capita (2005 US$, ln13.242 (6.889) −1.459 (11.433) 
Latin Americaa −12.335 (11.768) −14.322 (19.505) 
 x GDP/capita (2005 US$, ln18.064 (10.830) 7.341 (11.430) 
Sub-Saharan Africaa −8.190 (11.408) −7.010 (19.332) 
 x GDP/capita (2005 US$, ln23.696** (8.094) −2.493 (11.978) 
Democracy/Autocracy .389* (.190) −.284* (.116) 
OPEC member (1 = yes) 3.910 (5.232) −3.513 (2.018) 
Trade (% GDP) −.012 (.020) .016 (.014) 
International war (1 = yes) −3.612 (3.363) −.071 (1.524) 
Civil war (1 = yes) 11.783 (7.208) 1.770 (1.275) 
Fertility rate −2.842 (1.650) 1.264* (.621) 
Female population (%) 2.043** (.791) .193 (.533) 
Religious fractionalization −8.322 (4.826) −4.130 (4.672) 
Communist (1 = yes) 8.338** (2.626) 20.252*** (2.660) 
Women's INGO linkages .025 (.042) .005 (.030) 
CEDAW member (1 = yes) 3.348 (1.778) 1.496 (1.455) 
Proportional representation (1 = yes)   4.138* (1.694) 
Years of female suffrage   .048 (.048) 
Gendered educational attainment ratio   .042 (.050) 
Women's share of parliamentarians .034 (.065)   
Constant −1.385 (42.536) −5.196 (31.708) 
F 91.442  28.579  
df 61  63  
Adjusted R-squared .852  .635  
a

Reformed Western countries are the omitted reference group.

b

Standardized to have mean = 0 and standard deviation = 1.

Standard errors, adjusted for first-order autocorrelation, in parentheses.

Fixed effects for years included in the analyses but not reported.

*p < .05, **p < .01, ***p < .001 (two-tailed).

N(country-year observations) = 2,234; N(countries) = 85.

TABLE A7.
Heteroskedastic-robust OLS estimates for the effect of U.N. region indicators, economic development, and control variables on gendered educational attainment ratios, 1980 to 2010
Culture onlyNet of GDP/capitaFully specified
Coeff.(S.E.)Coeff.(S.E.)Coeff.(S.E.)
Western Africaa −60.394*** (2.186) −42.480*** (5.564) −27.688*** (5.316) 
Eastern Africaa −47.035*** (3.731) −28.245*** (6.542) −16.939** (5.812) 
Northern Africaa −47.593*** (6.688) −34.505*** (7.696) −26.086*** (7.350) 
Middle Africaa −54.148*** (3.697) −40.136*** (4.765) −22.296*** (5.172) 
Southern Africaa 2.828 (7.702) 13.656 (8.530) 20.985** (7.921) 
Central Americaa −9.359* (3.853) 1.161 (4.935) 5.402 (3.915) 
South Americaa −8.293** (2.795) 1.689 (3.686) 1.018 (3.625) 
Northern Americaa .216 (.763) −.808 (1.465) −3.534 (2.889) 
Caribbeana −8.410* (3.963) .899 (3.690) .003 (3.503) 
Eastern Asiaa −15.505** (6.010) −6.893 (5.049) −10.836 (6.221) 
Southern Asiaa −51.827*** (9.173) −34.728*** (9.908) −35.660*** (8.925) 
South-Eastern Asiaa −28.094*** (5.444) −14.883* (6.198) −16.920* (6.600) 
Southern Europea −10.455*** (1.224) −6.442*** (1.935) −9.349*** (2.344) 
Australia and New Zealanda −1.559 (.964) −1.459 (1.597) −3.780 (2.832) 
Melanesiaa −31.934*** (9.063) −18.738* (8.617) −13.467 (8.006) 
Central Asiaa −9.854*** (1.985) 6.080 (4.974) 8.502 (4.504) 
Western Asiaa −29.674*** (5.614) −22.440*** (5.755) −14.168** (5.171) 
Eastern Europea −4.794*** (1.263) 3.572 (3.178) −4.065 (3.160) 
Western Europea −6.988*** (1.436) −8.526*** (2.071) −10.394*** (2.323) 
GDP/capita (2005 US$, ln  4.545*** (1.251) −.195 (1.388) 
Democracy/Autocracy     .392** (.121) 
OPEC member (1 = yes)     8.639 (4.809) 
Trade (% GDP)     .024 (.024) 
International war (1 = yes)     2.341 (3.908) 
Civil war (1 = yes)     2.642 (3.298) 
Fertility rate     −6.615*** (1.028) 
Female population (%)     −.164 (.497) 
Religious fractionalization     6.497 (3.915) 
Communist (1 = yes)     −.979 (3.375) 
Women's INGO linkages     .058 (.041) 
CEDAW member (1 = yes)     2.646 (1.533) 
Women's share of parliamentarians     −.007 (.072) 
Constant 90.866*** (1.054) 44.890*** (12.572) 109.193*** (32.956) 
F 56.634  63.548  86.762  
df 49  50  62  
Adjusted R-squared .752  .777  .841  
Culture onlyNet of GDP/capitaFully specified
Coeff.(S.E.)Coeff.(S.E.)Coeff.(S.E.)
Western Africaa −60.394*** (2.186) −42.480*** (5.564) −27.688*** (5.316) 
Eastern Africaa −47.035*** (3.731) −28.245*** (6.542) −16.939** (5.812) 
Northern Africaa −47.593*** (6.688) −34.505*** (7.696) −26.086*** (7.350) 
Middle Africaa −54.148*** (3.697) −40.136*** (4.765) −22.296*** (5.172) 
Southern Africaa 2.828 (7.702) 13.656 (8.530) 20.985** (7.921) 
Central Americaa −9.359* (3.853) 1.161 (4.935) 5.402 (3.915) 
South Americaa −8.293** (2.795) 1.689 (3.686) 1.018 (3.625) 
Northern Americaa .216 (.763) −.808 (1.465) −3.534 (2.889) 
Caribbeana −8.410* (3.963) .899 (3.690) .003 (3.503) 
Eastern Asiaa −15.505** (6.010) −6.893 (5.049) −10.836 (6.221) 
Southern Asiaa −51.827*** (9.173) −34.728*** (9.908) −35.660*** (8.925) 
South-Eastern Asiaa −28.094*** (5.444) −14.883* (6.198) −16.920* (6.600) 
Southern Europea −10.455*** (1.224) −6.442*** (1.935) −9.349*** (2.344) 
Australia and New Zealanda −1.559 (.964) −1.459 (1.597) −3.780 (2.832) 
Melanesiaa −31.934*** (9.063) −18.738* (8.617) −13.467 (8.006) 
Central Asiaa −9.854*** (1.985) 6.080 (4.974) 8.502 (4.504) 
Western Asiaa −29.674*** (5.614) −22.440*** (5.755) −14.168** (5.171) 
Eastern Europea −4.794*** (1.263) 3.572 (3.178) −4.065 (3.160) 
Western Europea −6.988*** (1.436) −8.526*** (2.071) −10.394*** (2.323) 
GDP/capita (2005 US$, ln  4.545*** (1.251) −.195 (1.388) 
Democracy/Autocracy     .392** (.121) 
OPEC member (1 = yes)     8.639 (4.809) 
Trade (% GDP)     .024 (.024) 
International war (1 = yes)     2.341 (3.908) 
Civil war (1 = yes)     2.642 (3.298) 
Fertility rate     −6.615*** (1.028) 
Female population (%)     −.164 (.497) 
Religious fractionalization     6.497 (3.915) 
Communist (1 = yes)     −.979 (3.375) 
Women's INGO linkages     .058 (.041) 
CEDAW member (1 = yes)     2.646 (1.533) 
Women's share of parliamentarians     −.007 (.072) 
Constant 90.866*** (1.054) 44.890*** (12.572) 109.193*** (32.956) 
F 56.634  63.548  86.762  
df 49  50  62  
Adjusted R-squared .752  .777  .841  
a

Northern European countries are the omitted reference group.

Standard errors, adjusted for first-order autocorrelation, in parentheses.

Fixed effects for years included in the analyses but not reported.

*p < .05, **p < .01, ***p < .001 (two-tailed).

N(country-year observations) = 3,878; N(countries) = 154.

TABLE A8.
Heteroskedastic-robust OLS estimates for the effect of U.N. region indicators, economic development, and control variables on women's share of parliamentarians, 1980 to 2010
Culture onlyNet of GDP/capitaFully specified
Coeff.(S.E.)Coeff.(S.E.)Coeff.(S.E.)
Western Africaa −17.096*** (3.436) −19.073*** (3.946) −16.106*** (4.096) 
Eastern Africaa −13.561*** (3.889) −15.635*** (4.261) −12.224** (4.322) 
Northern Africaa −18.998*** (3.590) −20.442*** (3.889) −17.409*** (3.985) 
Middle Africaa −16.132*** (3.530) −17.679*** (3.860) −16.703*** (4.067) 
Southern Africaa −12.373*** (3.691) −13.568*** (3.947) −11.257** (3.833) 
Central Americaa −12.328*** (3.693) −13.489*** (3.916) −12.041** (3.735) 
Northern Americaa −12.784** (4.397) −12.671** (4.421) −10.607* (4.813) 
South Americaa −14.364*** (3.717) −15.465*** (3.788) −13.607*** (3.600) 
Caribbeana −8.485 (5.847) −9.513 (5.898) −10.086* (4.428) 
Eastern Asiaa −13.248** (4.433) −14.198*** (4.283) −16.199*** (3.689) 
Southern Asiaa −18.028*** (3.491) −19.915*** (3.922) −15.096*** (4.459) 
South-Eastern Asiaa −13.159*** (3.716) −14.618*** (3.894) −14.008*** (3.979) 
Southern Europea −12.325*** (3.612) −12.768*** (3.680) −12.268*** (3.492) 
Australia and New Zealanda −3.951 (3.354) −3.962 (3.414) −4.497 (3.475) 
Melanesiaa −22.035*** (3.821) −23.492*** (4.177) −17.442*** (4.213) 
Central Asiaa −14.924*** (3.620) −16.683*** (4.058) −12.961** (3.949) 
Western Asiaa −22.330*** (3.514) −23.128*** (3.614) −19.888*** (3.488) 
Eastern Europea −12.796*** (3.682) −13.719*** (3.818) −13.454*** (3.564) 
Western Europea −5.967 (4.192) −5.798 (4.242) −6.508 (3.997) 
GDP/capita (2005 US$, ln  −.502 (.453) .270 (.543) 
Democracy/Autocracy     −.263** (.082) 
OPEC member (1 = yes)     −1.685 (1.307) 
Trade (% GDP)     .006 (.011) 
International war (1 = yes)     −.854 (1.364) 
Civil war (1 = yes)     −.670 (.864) 
Fertility rate     .369 (.493) 
Female population (%)     .087 (.203) 
Religious fractionalization     1.887 (2.170) 
Communist (1 = yes)     11.762*** (1.994) 
Women's INGO linkages     .044 (.025) 
CEDAW member (1 = yes)     1.551 (.876) 
Proportional representation (1 = yes)     3.946** (1.231) 
Years of female suffrage     .012 (.023) 
Gendered educational attainment ratio     −.001 (.033) 
Constant 20.784*** (3.375) 25.859*** (5.719) 7.359 (13.330) 
F 56.509  58.099  19.571  
df 49  50  64  
Adjusted R-squared .421  .423  .520  
Culture onlyNet of GDP/capitaFully specified
Coeff.(S.E.)Coeff.(S.E.)Coeff.(S.E.)
Western Africaa −17.096*** (3.436) −19.073*** (3.946) −16.106*** (4.096) 
Eastern Africaa −13.561*** (3.889) −15.635*** (4.261) −12.224** (4.322) 
Northern Africaa −18.998*** (3.590) −20.442*** (3.889) −17.409*** (3.985) 
Middle Africaa −16.132*** (3.530) −17.679*** (3.860) −16.703*** (4.067) 
Southern Africaa −12.373*** (3.691) −13.568*** (3.947) −11.257** (3.833) 
Central Americaa −12.328*** (3.693) −13.489*** (3.916) −12.041** (3.735) 
Northern Americaa −12.784** (4.397) −12.671** (4.421) −10.607* (4.813) 
South Americaa −14.364*** (3.717) −15.465*** (3.788) −13.607*** (3.600) 
Caribbeana −8.485 (5.847) −9.513 (5.898) −10.086* (4.428) 
Eastern Asiaa −13.248** (4.433) −14.198*** (4.283) −16.199*** (3.689) 
Southern Asiaa −18.028*** (3.491) −19.915*** (3.922) −15.096*** (4.459) 
South-Eastern Asiaa −13.159*** (3.716) −14.618*** (3.894) −14.008*** (3.979) 
Southern Europea −12.325*** (3.612) −12.768*** (3.680) −12.268*** (3.492) 
Australia and New Zealanda −3.951 (3.354) −3.962 (3.414) −4.497 (3.475) 
Melanesiaa −22.035*** (3.821) −23.492*** (4.177) −17.442*** (4.213) 
Central Asiaa −14.924*** (3.620) −16.683*** (4.058) −12.961** (3.949) 
Western Asiaa −22.330*** (3.514) −23.128*** (3.614) −19.888*** (3.488) 
Eastern Europea −12.796*** (3.682) −13.719*** (3.818) −13.454*** (3.564) 
Western Europea −5.967 (4.192) −5.798 (4.242) −6.508 (3.997) 
GDP/capita (2005 US$, ln  −.502 (.453) .270 (.543) 
Democracy/Autocracy     −.263** (.082) 
OPEC member (1 = yes)     −1.685 (1.307) 
Trade (% GDP)     .006 (.011) 
International war (1 = yes)     −.854 (1.364) 
Civil war (1 = yes)     −.670 (.864) 
Fertility rate     .369 (.493) 
Female population (%)     .087 (.203) 
Religious fractionalization     1.887 (2.170) 
Communist (1 = yes)     11.762*** (1.994) 
Women's INGO linkages     .044 (.025) 
CEDAW member (1 = yes)     1.551 (.876) 
Proportional representation (1 = yes)     3.946** (1.231) 
Years of female suffrage     .012 (.023) 
Gendered educational attainment ratio     −.001 (.033) 
Constant 20.784*** (3.375) 25.859*** (5.719) 7.359 (13.330) 
F 56.509  58.099  19.571  
df 49  50  64  
Adjusted R-squared .421  .423  .520  
a

Northern European countries are the omitted reference group.

Standard errors, adjusted for first-order autocorrelation, in parentheses.

Fixed effects for years included in the analyses but not reported.

*p < .05, **p < .01, ***p < .001 (two-tailed).

N(country-year observations) = 3,878; N(countries) = 154.

TABLE A9.
Heteroskedastic-robust OLS estimates for the effect of economic development on gendered educational and political outcomes, as conditioned by U.N. region indicators, 1980 to 2010
Gendered educational attainment ratioWomen's share of parliamentarians
Coeff.(S.E.)Coeff.(S.E.)
GDP/capita (2005 US$, ln)b −4.446 (3.576) 13.232*** (2.252) 
Western Africaa −29.368*** (7.254) 6.239 (3.644) 
 x GDP/capita (2005 US$, ln) 6.719 (6.088) −10.762*** (2.879) 
Eastern Africaa −24.660*** (6.480) 2.733 (4.237) 
 x GDP/capita (2005 US$, ln) 1.304 (5.888) −17.081*** (3.294) 
Northern Africaa −27.699* (13.093) −.055 (3.617) 
 x GDP/capita (2005 US$, ln) 7.925 (16.169) −16.968*** (4.674) 
Middle Africaa −21.697*** (6.240) 3.337 (3.315) 
 x GDP/capita (2005 US$, ln) 11.812* (4.861) −12.405*** (2.641) 
Southern Africaa 15.171** (5.526) 7.117* (2.866) 
 x GDP/capita (2005 US$, ln) −30.413*** (6.038) −10.973*** (2.548) 
South Americaa −4.281 (5.046) 5.186 (3.166) 
 x GDP/capita (2005 US$, ln) 12.307* (5.357) −18.562*** (4.839) 
Central Americaa −.108 (5.746) 6.991* (3.055) 
 x GDP/capita (2005 US$, ln) −2.290 (4.951) −11.642** (4.275) 
Northern Americaa 1.605 (18.176) 27.168* (13.308) 
 x GDP/capita (2005 US$, ln) −3.393 (12.148) −25.189** (9.294) 
Caribbeana −5.518 (5.098) 7.099* (3.193) 
 x GDP/capita (2005 US$, ln) 8.930 (5.833) −3.355 (2.609) 
Eastern Asiaa −15.857 (8.641) 3.310 (3.214) 
 x GDP/capita (2005 US$, ln) 4.861 (6.507) −15.646*** (2.662) 
Southern Asiaa −16.509 (14.255) −2.317 (4.656) 
 x GDP/capita (2005 US$, ln) 30.459** (9.757) −19.974*** (3.690) 
South-Eastern Asiaa −21.407*** (5.106) 4.463 (3.002) 
 x GDP/capita (2005 US$, ln) .787 (5.498) −13.550*** (3.313) 
Southern Europea −10.930* (5.080) 6.919* (2.907) 
 x GDP/capita (2005 US$, ln) −.030 (3.964) −13.840*** (3.325) 
Australia and New Zealanda 16.855* (8.136) 9.120 (15.767) 
 x GDP/capita (2005 US$, ln) −15.284** (5.309) −9.305 (11.689) 
Melanesiaa −8.700 (5.188) 4.012 (3.210) 
 x GDP/capita (2005 US$, ln) 29.026*** (8.714) −6.901* (3.327) 
Central Asiaa 7.555 (4.112) 6.098 (3.676) 
 x GDP/capita (2005 US$, ln) 5.657 (4.534) −11.047** (3.609) 
Western Asiaa −17.665** (6.332) −.572 (2.998) 
 x GDP/capita (2005 US$, ln) 2.643 (5.955) −14.634*** (2.689) 
Eastern Europea −6.060 (3.551) 3.900 (2.823) 
 x GDP/capita (2005 US$, ln) −.843 (3.354) −11.347*** (2.901) 
Western Europea −12.106 (14.518) 11.363 (12.778) 
 x GDP/capita (2005 US$, ln) 1.521 (8.761) −12.452 (7.275) 
Democracy/Autocracy .516*** (.124) −.283*** (.078) 
OPEC member (1 = yes) 3.251 (4.701) .347 (1.421) 
Trade (% GDP) .016 (.016) .007 (.011) 
International war (1 = yes) −1.781 (3.724) −.068 (1.405) 
Civil war (1 = yes) 2.719 (3.064) −.445 (.902) 
Fertility rate −6.453*** (1.134) −.055 (.483) 
Female population (%) −.828 (.589) .243 (.270) 
Religious fractionalization 6.450 (3.692) 2.274 (1.978) 
Communist (1 = yes) −1.719 (3.425) 10.568*** (1.908) 
Women's INGO linkages .088* (.044) .031 (.029) 
CEDAW member (1 = yes) 3.014 (1.589) .924 (.816) 
Proportional representation (1 = yes)   3.871** (1.217) 
Years of female suffrage   .007 (.027) 
Gendered educational attainment ratio   .015 (.036) 
Women's share of parliamentarians .026 (.069)   
Constant 145.122*** (33.163) −16.455 (15.940) 
F 142.539  90.262  
df 81  83  
Adjusted R-squared .873  .570  
Gendered educational attainment ratioWomen's share of parliamentarians
Coeff.(S.E.)Coeff.(S.E.)
GDP/capita (2005 US$, ln)b −4.446 (3.576) 13.232*** (2.252) 
Western Africaa −29.368*** (7.254) 6.239 (3.644) 
 x GDP/capita (2005 US$, ln) 6.719 (6.088) −10.762*** (2.879) 
Eastern Africaa −24.660*** (6.480) 2.733 (4.237) 
 x GDP/capita (2005 US$, ln) 1.304 (5.888) −17.081*** (3.294) 
Northern Africaa −27.699* (13.093) −.055 (3.617) 
 x GDP/capita (2005 US$, ln) 7.925 (16.169) −16.968*** (4.674) 
Middle Africaa −21.697*** (6.240) 3.337 (3.315) 
 x GDP/capita (2005 US$, ln) 11.812* (4.861) −12.405*** (2.641) 
Southern Africaa 15.171** (5.526) 7.117* (2.866) 
 x GDP/capita (2005 US$, ln) −30.413*** (6.038) −10.973*** (2.548) 
South Americaa −4.281 (5.046) 5.186 (3.166) 
 x GDP/capita (2005 US$, ln) 12.307* (5.357) −18.562*** (4.839) 
Central Americaa −.108 (5.746) 6.991* (3.055) 
 x GDP/capita (2005 US$, ln) −2.290 (4.951) −11.642** (4.275) 
Northern Americaa 1.605 (18.176) 27.168* (13.308) 
 x GDP/capita (2005 US$, ln) −3.393 (12.148) −25.189** (9.294) 
Caribbeana −5.518 (5.098) 7.099* (3.193) 
 x GDP/capita (2005 US$, ln) 8.930 (5.833) −3.355 (2.609) 
Eastern Asiaa −15.857 (8.641) 3.310 (3.214) 
 x GDP/capita (2005 US$, ln) 4.861 (6.507) −15.646*** (2.662) 
Southern Asiaa −16.509 (14.255) −2.317 (4.656) 
 x GDP/capita (2005 US$, ln) 30.459** (9.757) −19.974*** (3.690) 
South-Eastern Asiaa −21.407*** (5.106) 4.463 (3.002) 
 x GDP/capita (2005 US$, ln) .787 (5.498) −13.550*** (3.313) 
Southern Europea −10.930* (5.080) 6.919* (2.907) 
 x GDP/capita (2005 US$, ln) −.030 (3.964) −13.840*** (3.325) 
Australia and New Zealanda 16.855* (8.136) 9.120 (15.767) 
 x GDP/capita (2005 US$, ln) −15.284** (5.309) −9.305 (11.689) 
Melanesiaa −8.700 (5.188) 4.012 (3.210) 
 x GDP/capita (2005 US$, ln) 29.026*** (8.714) −6.901* (3.327) 
Central Asiaa 7.555 (4.112) 6.098 (3.676) 
 x GDP/capita (2005 US$, ln) 5.657 (4.534) −11.047** (3.609) 
Western Asiaa −17.665** (6.332) −.572 (2.998) 
 x GDP/capita (2005 US$, ln) 2.643 (5.955) −14.634*** (2.689) 
Eastern Europea −6.060 (3.551) 3.900 (2.823) 
 x GDP/capita (2005 US$, ln) −.843 (3.354) −11.347*** (2.901) 
Western Europea −12.106 (14.518) 11.363 (12.778) 
 x GDP/capita (2005 US$, ln) 1.521 (8.761) −12.452 (7.275) 
Democracy/Autocracy .516*** (.124) −.283*** (.078) 
OPEC member (1 = yes) 3.251 (4.701) .347 (1.421) 
Trade (% GDP) .016 (.016) .007 (.011) 
International war (1 = yes) −1.781 (3.724) −.068 (1.405) 
Civil war (1 = yes) 2.719 (3.064) −.445 (.902) 
Fertility rate −6.453*** (1.134) −.055 (.483) 
Female population (%) −.828 (.589) .243 (.270) 
Religious fractionalization 6.450 (3.692) 2.274 (1.978) 
Communist (1 = yes) −1.719 (3.425) 10.568*** (1.908) 
Women's INGO linkages .088* (.044) .031 (.029) 
CEDAW member (1 = yes) 3.014 (1.589) .924 (.816) 
Proportional representation (1 = yes)   3.871** (1.217) 
Years of female suffrage   .007 (.027) 
Gendered educational attainment ratio   .015 (.036) 
Women's share of parliamentarians .026 (.069)   
Constant 145.122*** (33.163) −16.455 (15.940) 
F 142.539  90.262  
df 81  83  
Adjusted R-squared .873  .570  
a

Northern European countries are the omitted reference group.

b

Standardized to have mean = 0 and standard deviation = 1. Standard errors, adjusted for first-order autocorrelation, in parentheses.

Fixed effects for years included in the analyses but not reported.

*p < .05, **p < .01, ***p < .001 (two-tailed).

N(country-year observations) = 2,234; N(countries) = 85.

TABLE A10.
Heteroskedastic-robust OLS estimates for the effect of consolidated U.N. region indicators, economic development, and control variables on gendered educational attainment ratios, 1980 to 2010
Culture onlyNet of GDP/capitaFully specified
Coeff.(S.E.)Coeff.(S.E.)Coeff.(S.E.)
Africaa −41.412*** (3.363) −23.371*** (5.294) −3.440 (5.710) 
Latin Americaa −3.344 (2.144) 6.575* (2.893) 14.541*** (3.024) 
North Americaa 5.565*** (.922) .795 (1.722) 3.204 (2.980) 
Asiaa −24.422*** (3.776) −12.363** (3.978) −2.733 (3.836) 
Oceaniaa −12.001 (8.534) −6.274 (5.042) −1.021 (3.955) 
GDP/capita (2005 US$, ln  6.038*** (1.208) 2.489 (1.389) 
Democracy/Autocracy     .306 (.224) 
OPEC member (1 = yes)     5.474 (4.173) 
Trade (% GDP)     .034 (.042) 
International war (1 = yes)     2.574 (3.052) 
Civil war (1 = yes)     .894 (4.312) 
Fertility rate     −7.413*** (1.208) 
Female population (%)     .885 (.647) 
Religious fractionalization     14.858** (5.034) 
Communist (1 = yes)     −5.035 (3.881) 
Women's INGO linkages     −.059 (.056) 
CEDAW member (1 = yes)     .705 (2.287) 
Women's share of parliamentarians     .180* (.080) 
Constant 84.770*** (1.401) 26.802* (11.572) 23.433 (40.237) 
F 24.130  29.224  46.539  
df 35  36  48  
Adjusted R-squared .504  .579  .712  
Culture onlyNet of GDP/capitaFully specified
Coeff.(S.E.)Coeff.(S.E.)Coeff.(S.E.)
Africaa −41.412*** (3.363) −23.371*** (5.294) −3.440 (5.710) 
Latin Americaa −3.344 (2.144) 6.575* (2.893) 14.541*** (3.024) 
North Americaa 5.565*** (.922) .795 (1.722) 3.204 (2.980) 
Asiaa −24.422*** (3.776) −12.363** (3.978) −2.733 (3.836) 
Oceaniaa −12.001 (8.534) −6.274 (5.042) −1.021 (3.955) 
GDP/capita (2005 US$, ln  6.038*** (1.208) 2.489 (1.389) 
Democracy/Autocracy     .306 (.224) 
OPEC member (1 = yes)     5.474 (4.173) 
Trade (% GDP)     .034 (.042) 
International war (1 = yes)     2.574 (3.052) 
Civil war (1 = yes)     .894 (4.312) 
Fertility rate     −7.413*** (1.208) 
Female population (%)     .885 (.647) 
Religious fractionalization     14.858** (5.034) 
Communist (1 = yes)     −5.035 (3.881) 
Women's INGO linkages     −.059 (.056) 
CEDAW member (1 = yes)     .705 (2.287) 
Women's share of parliamentarians     .180* (.080) 
Constant 84.770*** (1.401) 26.802* (11.572) 23.433 (40.237) 
F 24.130  29.224  46.539  
df 35  36  48  
Adjusted R-squared .504  .579  .712  
a

Europe is the omitted reference group.

Standard errors, adjusted for first-order autocorrelation, in parentheses.

Fixed effects for years included in the analyses but not reported.

*p < .05, **p < .01, ***p < .001 (two-tailed).

N(country-year observations) = 3,878; N(countries) = 154.

TABLE A11.
Heteroskedastic-robust OLS estimates for the effect of U.N. consolidated indicators, economic development, and control variables on women's share of parliamentarians, 1980 to 2010
Culture onlyNet of GDP/capitaFully specified
Coeff.(S.E.)Coeff.(S.E.)Coeff.(S.E.)
Africaa −8.163*** (1.778) −7.208*** (1.787) −3.786* (1.745) 
Latin Americaa −5.329* (2.096) −4.804* (2.002) −3.420 (1.957) 
North Americaa −5.431 (3.264) −5.683 (3.353) −5.951 (3.798) 
Asiaa −9.843*** (1.841) −9.205*** (1.789) −5.419*** (1.555) 
Oceaniaa −5.974 (4.521) −5.671 (4.292) −3.219 (2.380) 
GDP/capita (2005 US$, ln  .320 (.471) .689 (.559) 
Democracy/Autocracy     −.204* (.082) 
OPEC member (1 = yes)     −2.143 (1.547) 
Trade (% GDP)     .009 (.009) 
International war (1 = yes)     −1.007 (1.881) 
Civil war (1 = yes)     −.570 (.904) 
Fertility rate     1.105* (.516) 
Female population (%)     .209 (.223) 
Religious fractionalization     1.913 (2.102) 
Communist (1 = yes)     12.735*** (1.947) 
Women's INGO linkages     .075*** (.022) 
CEDAW member (1 = yes)     1.681* (.852) 
Proportional representation (1 = yes)     3.282** (1.212) 
Years of female suffrage     .064* (.030) 
Gendered educational attainment ratio     .047 (.024) 
Constant 13.553*** (1.656) 10.485* (4.273) −20.233 (12.686) 
F 9.957  10.083  12.214  
df 35  36  50  
Adjusted R-squared .261  .263  .425  
Culture onlyNet of GDP/capitaFully specified
Coeff.(S.E.)Coeff.(S.E.)Coeff.(S.E.)
Africaa −8.163*** (1.778) −7.208*** (1.787) −3.786* (1.745) 
Latin Americaa −5.329* (2.096) −4.804* (2.002) −3.420 (1.957) 
North Americaa −5.431 (3.264) −5.683 (3.353) −5.951 (3.798) 
Asiaa −9.843*** (1.841) −9.205*** (1.789) −5.419*** (1.555) 
Oceaniaa −5.974 (4.521) −5.671 (4.292) −3.219 (2.380) 
GDP/capita (2005 US$, ln  .320 (.471) .689 (.559) 
Democracy/Autocracy     −.204* (.082) 
OPEC member (1 = yes)     −2.143 (1.547) 
Trade (% GDP)     .009 (.009) 
International war (1 = yes)     −1.007 (1.881) 
Civil war (1 = yes)     −.570 (.904) 
Fertility rate     1.105* (.516) 
Female population (%)     .209 (.223) 
Religious fractionalization     1.913 (2.102) 
Communist (1 = yes)     12.735*** (1.947) 
Women's INGO linkages     .075*** (.022) 
CEDAW member (1 = yes)     1.681* (.852) 
Proportional representation (1 = yes)     3.282** (1.212) 
Years of female suffrage     .064* (.030) 
Gendered educational attainment ratio     .047 (.024) 
Constant 13.553*** (1.656) 10.485* (4.273) −20.233 (12.686) 
F 9.957  10.083  12.214  
df 35  36  50  
Adjusted R-squared .261  .263  .425  
a

Europe is the omitted reference group.

Standard errors, adjusted for first-order autocorrelation, in parentheses.

Fixed effects for years included in the analyses but not reported.

*p < .05, **p < .01, ***p < .001 (two-tailed).

N(country-year observations) = 3,878; N(countries) = 154.

TABLE A12.
Heteroskedastic-robust OLS estimates for the effect of economic development on gendered educational and political outcomes, as conditioned by consolidated U.N. region indicators, 1980 to 2010
Gendered educational attainment ratioWomen's share of parliamentarians
Coeff.(S.E.)Coeff.(S.E.)
GDP/capita (2005 US$, ln)b 6.906*** (1.763) .789 (2.826) 
Africaa .653 (2.065) −5.564 (6.026) 
 x GDP/capita (2005 US$, ln−7.958*** (2.131) 6.155 (4.858) 
Latin Americaa 2.388 (2.098) 10.811** (3.704) 
 x GDP/capita (2005 US$, ln−5.640 (3.355) 2.318 (5.542) 
North Americaa 19.987 (11.519) −5.388 (21.872) 
 x GDP/capita (2005 US$, ln−17.331* (7.094) 6.098 (14.629) 
Asiaa −.957 (1.819) −6.009 (3.531) 
 x GDP/capita (2005 US$, ln−9.259*** (1.869) 3.344 (3.867) 
Oceaniaa .147 (2.103) −5.417 (5.563) 
 x GDP/capita (2005 US$, ln−.216 (2.319) 4.418 (4.102) 
Democracy/Autocracy −.235** (.075) .333 (.223) 
OPEC member (1 = yes) −1.154 (1.380) 4.729 (4.360) 
Trade (% GDP) .012 (.009) .034 (.042) 
International war (1 = yes) −.630 (1.416) 2.468 (2.915) 
Civil war (1 = yes) −1.108 (.958) .909 (4.220) 
Fertility rate .493 (.465) −6.868*** (1.365) 
Female population (%) .120 (.211) .803 (.693) 
Religious fractionalization 3.669 (1.936) 14.480** (5.080) 
Communist (1 = yes) 11.681*** (1.958) −5.301 (3.951) 
Women's INGO linkages .034 (.022) −.040 (.065) 
CEDAW member (1 = yes) 1.021 (.844) .796 (2.270) 
Proportional representation (1 = yes) 3.511** (1.163)   
Years of female suffrage .022 (.033)   
Gendered educational attainment ratio .056* (.025)   
Women's share of parliamentarians   .212** (.082) 
Constant −13.206 (10.480) 48.306 (37.421) 
F 114.538  55.376  
df 55  53  
Adjusted R-squared .482  .714  
Gendered educational attainment ratioWomen's share of parliamentarians
Coeff.(S.E.)Coeff.(S.E.)
GDP/capita (2005 US$, ln)b 6.906*** (1.763) .789 (2.826) 
Africaa .653 (2.065) −5.564 (6.026) 
 x GDP/capita (2005 US$, ln−7.958*** (2.131) 6.155 (4.858) 
Latin Americaa 2.388 (2.098) 10.811** (3.704) 
 x GDP/capita (2005 US$, ln−5.640 (3.355) 2.318 (5.542) 
North Americaa 19.987 (11.519) −5.388 (21.872) 
 x GDP/capita (2005 US$, ln−17.331* (7.094) 6.098 (14.629) 
Asiaa −.957 (1.819) −6.009 (3.531) 
 x GDP/capita (2005 US$, ln−9.259*** (1.869) 3.344 (3.867) 
Oceaniaa .147 (2.103) −5.417 (5.563) 
 x GDP/capita (2005 US$, ln−.216 (2.319) 4.418 (4.102) 
Democracy/Autocracy −.235** (.075) .333 (.223) 
OPEC member (1 = yes) −1.154 (1.380) 4.729 (4.360) 
Trade (% GDP) .012 (.009) .034 (.042) 
International war (1 = yes) −.630 (1.416) 2.468 (2.915) 
Civil war (1 = yes) −1.108 (.958) .909 (4.220) 
Fertility rate .493 (.465) −6.868*** (1.365) 
Female population (%) .120 (.211) .803 (.693) 
Religious fractionalization 3.669 (1.936) 14.480** (5.080) 
Communist (1 = yes) 11.681*** (1.958) −5.301 (3.951) 
Women's INGO linkages .034 (.022) −.040 (.065) 
CEDAW member (1 = yes) 1.021 (.844) .796 (2.270) 
Proportional representation (1 = yes) 3.511** (1.163)   
Years of female suffrage .022 (.033)   
Gendered educational attainment ratio .056* (.025)   
Women's share of parliamentarians   .212** (.082) 
Constant −13.206 (10.480) 48.306 (37.421) 
F 114.538  55.376  
df 55  53  
Adjusted R-squared .482  .714  
a

Europe is the omitted reference group.

b

Standardized to have mean = 0 and standard deviation = 1.

Standard errors, adjusted for first-order autocorrelation, in parentheses.

Fixed effects for years included in the analyses but not reported.

*p < .05, **p < .01, ***p < .001 (two-tailed).

N(country-year observations) = 3,878; N(countries) = 154.

TABLE A13.
Heteroskedastic-robust OLS estimates for the effect of colonial legacy indicators, economic development, and control variables on gendered educational attainment ratios, 1980 to 2010
Culture onlyNet of GDP/capitaFully specified
Coeff.(S.E.)Coeff.(S.E.)Coeff.(S.E.)
Dutch colonya −13.064 (7.007) −1.551 (4.835) −3.953 (4.222) 
Spanish colonya 1.324 (3.802) 8.563* (3.899) 16.955*** (3.155) 
US colonya 10.956*** (2.448) 25.211*** (3.582) 40.109*** (4.593) 
British colonya −18.825*** (4.692) −7.058 (4.435) 2.790 (4.219) 
French colonya −40.904*** (3.836) −24.607*** (4.728) −8.652 (4.672) 
Portuguese colonya −34.049** (11.618) −18.918* (9.564) −7.769 (8.266) 
Belgian colonya −41.753*** (4.210) −15.671* (6.124) 1.824 (5.930) 
Australian colonya −37.048*** (2.518) −21.289*** (3.808) −11.079** (4.235) 
GDP/capita (2005 US$, ln  7.204*** (1.110) 3.542* (1.449) 
Democracy/Autocracy     .301 (.188) 
OPEC member (1 = yes)     6.745 (5.120) 
Trade (% GDP)     .012 (.040) 
International war (1 = yes)     3.344 (3.587) 
Civil war (1 = yes)     −2.393 (2.665) 
Fertility rate     −7.204*** (1.038) 
Female population (%)     1.611* (.744) 
Religious fractionalization     17.463*** (4.372) 
Communist (1 = yes)     −2.648 (3.972) 
Women's INGO linkages     −.124* (.050) 
CEDAW member (1 = yes)     −.332 (1.944) 
Women's share of parliamentarians     .161* (.081) 
Constant 78.615*** (2.975) 13.765 (11.065) −21.040 (45.124) 
F 54.637  65.108  82.795  
df 38  39  51  
Adjusted R-squared .454  .592  .749  
Culture onlyNet of GDP/capitaFully specified
Coeff.(S.E.)Coeff.(S.E.)Coeff.(S.E.)
Dutch colonya −13.064 (7.007) −1.551 (4.835) −3.953 (4.222) 
Spanish colonya 1.324 (3.802) 8.563* (3.899) 16.955*** (3.155) 
US colonya 10.956*** (2.448) 25.211*** (3.582) 40.109*** (4.593) 
British colonya −18.825*** (4.692) −7.058 (4.435) 2.790 (4.219) 
French colonya −40.904*** (3.836) −24.607*** (4.728) −8.652 (4.672) 
Portuguese colonya −34.049** (11.618) −18.918* (9.564) −7.769 (8.266) 
Belgian colonya −41.753*** (4.210) −15.671* (6.124) 1.824 (5.930) 
Australian colonya −37.048*** (2.518) −21.289*** (3.808) −11.079** (4.235) 
GDP/capita (2005 US$, ln  7.204*** (1.110) 3.542* (1.449) 
Democracy/Autocracy     .301 (.188) 
OPEC member (1 = yes)     6.745 (5.120) 
Trade (% GDP)     .012 (.040) 
International war (1 = yes)     3.344 (3.587) 
Civil war (1 = yes)     −2.393 (2.665) 
Fertility rate     −7.204*** (1.038) 
Female population (%)     1.611* (.744) 
Religious fractionalization     17.463*** (4.372) 
Communist (1 = yes)     −2.648 (3.972) 
Women's INGO linkages     −.124* (.050) 
CEDAW member (1 = yes)     −.332 (1.944) 
Women's share of parliamentarians     .161* (.081) 
Constant 78.615*** (2.975) 13.765 (11.065) −21.040 (45.124) 
F 54.637  65.108  82.795  
df 38  39  51  
Adjusted R-squared .454  .592  .749  
a

Never colonized is the omitted reference group.

Standard errors, adjusted for first-order autocorrelation, in parentheses.

Fixed effects for years included in the analyses but not reported.

*p < .05, **p < .01, ***p < .001 (two-tailed).

N(country-year observations) = 3,878; N(countries) = 154.

TABLE A14.
Heteroskedastic-robust OLS estimates for the effect of colonial legacy indicators, economic development, and control variables on women's share of parliamentarians, 1980 to 2010
Culture onlyNet of GDP/capitaFully specified
Coeff.(S.E.)Coeff.(S.E.)Coeff.(S.E.)
Dutch colonya −3.155* (1.415) −1.477 (1.082) 1.556 (1.419) 
Spanish colonya −2.576 (1.826) −1.521 (1.696) −1.964 (1.620) 
US colonya −2.426 (1.254) −.348 (1.098) −1.203 (2.088) 
British colonya −5.897*** (1.580) −4.181** (1.524) −.667 (1.598) 
French colonya −6.760*** (1.544) −4.384** (1.441) −3.486* (1.522) 
Portuguese colonya −1.391 (3.173) .815 (3.541) 1.214 (4.391) 
Belgian colonya 1.178 (5.152) 4.980 (5.196) 5.672 (4.635) 
Australian colonya −12.374*** (1.281) −10.076*** (1.170) −3.865* (1.756) 
GDP/capita (2005 US$, ln  1.050* (.460) 1.335* (.612) 
Democracy/Autocracy     −.178* (.075) 
OPEC member (1 = yes)     −2.787 (1.579) 
Trade (% GDP)     .010 (.010) 
International war (1 = yes)     −2.226 (2.353) 
Civil war (1 = yes)     −1.692 (1.271) 
Fertility rate     1.417** (.518) 
Female population (%)     .466 (.241) 
Religious fractionalization     .433 (2.380) 
Communist (1 = yes)     14.111*** (2.023) 
Women's INGO linkages     .070** (.026) 
CEDAW member (1 = yes)     1.778 (.989) 
Proportional representation (1 = yes)     3.896*** (1.152) 
Years of female suffrage     .066* (.033) 
Gendered educational attainment ratio     .042 (.029) 
Constant 10.686*** (1.361) 1.231 (3.841) −41.223** (13.570) 
F 13.560  104.490  397.584  
df 38  39  53  
Adjusted R-squared .214  .234  .428  
Culture onlyNet of GDP/capitaFully specified
Coeff.(S.E.)Coeff.(S.E.)Coeff.(S.E.)
Dutch colonya −3.155* (1.415) −1.477 (1.082) 1.556 (1.419) 
Spanish colonya −2.576 (1.826) −1.521 (1.696) −1.964 (1.620) 
US colonya −2.426 (1.254) −.348 (1.098) −1.203 (2.088) 
British colonya −5.897*** (1.580) −4.181** (1.524) −.667 (1.598) 
French colonya −6.760*** (1.544) −4.384** (1.441) −3.486* (1.522) 
Portuguese colonya −1.391 (3.173) .815 (3.541) 1.214 (4.391) 
Belgian colonya 1.178 (5.152) 4.980 (5.196) 5.672 (4.635) 
Australian colonya −12.374*** (1.281) −10.076*** (1.170) −3.865* (1.756) 
GDP/capita (2005 US$, ln  1.050* (.460) 1.335* (.612) 
Democracy/Autocracy     −.178* (.075) 
OPEC member (1 = yes)     −2.787 (1.579) 
Trade (% GDP)     .010 (.010) 
International war (1 = yes)     −2.226 (2.353) 
Civil war (1 = yes)     −1.692 (1.271) 
Fertility rate     1.417** (.518) 
Female population (%)     .466 (.241) 
Religious fractionalization     .433 (2.380) 
Communist (1 = yes)     14.111*** (2.023) 
Women's INGO linkages     .070** (.026) 
CEDAW member (1 = yes)     1.778 (.989) 
Proportional representation (1 = yes)     3.896*** (1.152) 
Years of female suffrage     .066* (.033) 
Gendered educational attainment ratio     .042 (.029) 
Constant 10.686*** (1.361) 1.231 (3.841) −41.223** (13.570) 
F 13.560  104.490  397.584  
df 38  39  53  
Adjusted R-squared .214  .234  .428  
a

Never colonized is the omitted reference group.

Standard errors, adjusted for first-order autocorrelation, in parentheses.

Fixed effects for years included in the analyses but not reported.

*p < .05, **p < .01, ***p < .001 (two-tailed).

N(country-year observations) = 3,878; N(countries) = 154.

TABLE A15.
Heteroskedastic-robust OLS estimates for the effect of economic development on gendered educational and political outcomes, as conditioned by colonial legacy, 1980 to 2010
Gendered educational attainment ratioWomen's share of parliamentarians
Coeff.(S.E.)Coeff.(S.E.)
GDP/capita (2005 US$, ln)b 4.907 (3.001) 3.221** (1.116) 
Dutch colonya −1.391 (2.728) 1.192 (1.548) 
 x GDP/capita (2005 US$, ln13.249 (9.040) −5.190 (3.212) 
Spanish colonya 16.847*** (3.985) −1.033 (1.602) 
 x GDP/capita (2005 US$, ln−5.565 (5.034) −1.774 (2.497) 
US colonya 31.719*** (8.526) −14.953** (5.582) 
 x GDP/capita (2005 US$, ln−12.439 (13.864) −28.066** (9.466) 
British colonya 3.458 (4.234) −.593 (1.779) 
 x GDP/capita (2005 US$, ln2.888 (4.770) −3.900* (1.584) 
French colonya −12.350* (5.284) −1.082 (1.606) 
 x GDP/capita (2005 US$, ln−3.137 (5.557) −.544 (1.568) 
Portuguese colonya 2.462 (6.143) −3.096 (2.827) 
 x GDP/capita (2005 US$, ln19.069** (6.136) −11.883*** (2.517) 
Belgian colonya −25.625 (16.844) 29.096 (26.363) 
 x GDP/capita (2005 US$, ln−15.433 (9.431) 11.960 (14.380) 
Australian colonya −4.149 (4.682) −13.232*** (2.396) 
 x GDP/capita (2005 US$, ln9.551 (7.905) −16.265*** (3.978) 
Democracy/Autocracy .233 (.191) −.172* (.074) 
OPEC member (1 = yes) 8.704 (5.688) −2.982 (1.742) 
Trade (% GDP) .008 (.041) .014 (.010) 
International war (1 = yes) 1.821 (3.862) −.794 (2.355) 
Civil war (1 = yes) −1.586 (2.490) −1.918 (1.253) 
Fertility rate −7.016*** (1.022) 1.165* (.481) 
Female population (%) 1.971** (.749) .123 (.258) 
Religious fractionalization 16.904*** (4.547) .392 (2.275) 
Communist (1 = yes) −3.005 (4.033) 13.575*** (1.889) 
Women's INGO linkages −.111 (.058) .049* (.024) 
CEDAW member (1 = yes) −.783 (1.960) 1.796 (.952) 
Proportional representation (1 = yes)   3.751*** (1.104) 
Years of female suffrage   .078* (.032) 
Gendered educational attainment ratio   .064* (.028) 
Women's share of parliamentarians .225** (.072)   
Constant −11.381 (39.480) −15.568 (12.281) 
F 460.898  220.524  
df 59  61  
Adjusted R-squared .761  .462  
Gendered educational attainment ratioWomen's share of parliamentarians
Coeff.(S.E.)Coeff.(S.E.)
GDP/capita (2005 US$, ln)b 4.907 (3.001) 3.221** (1.116) 
Dutch colonya −1.391 (2.728) 1.192 (1.548) 
 x GDP/capita (2005 US$, ln13.249 (9.040) −5.190 (3.212) 
Spanish colonya 16.847*** (3.985) −1.033 (1.602) 
 x GDP/capita (2005 US$, ln−5.565 (5.034) −1.774 (2.497) 
US colonya 31.719*** (8.526) −14.953** (5.582) 
 x GDP/capita (2005 US$, ln−12.439 (13.864) −28.066** (9.466) 
British colonya 3.458 (4.234) −.593 (1.779) 
 x GDP/capita (2005 US$, ln2.888 (4.770) −3.900* (1.584) 
French colonya −12.350* (5.284) −1.082 (1.606) 
 x GDP/capita (2005 US$, ln−3.137 (5.557) −.544 (1.568) 
Portuguese colonya 2.462 (6.143) −3.096 (2.827) 
 x GDP/capita (2005 US$, ln19.069** (6.136) −11.883*** (2.517) 
Belgian colonya −25.625 (16.844) 29.096 (26.363) 
 x GDP/capita (2005 US$, ln−15.433 (9.431) 11.960 (14.380) 
Australian colonya −4.149 (4.682) −13.232*** (2.396) 
 x GDP/capita (2005 US$, ln9.551 (7.905) −16.265*** (3.978) 
Democracy/Autocracy .233 (.191) −.172* (.074) 
OPEC member (1 = yes) 8.704 (5.688) −2.982 (1.742) 
Trade (% GDP) .008 (.041) .014 (.010) 
International war (1 = yes) 1.821 (3.862) −.794 (2.355) 
Civil war (1 = yes) −1.586 (2.490) −1.918 (1.253) 
Fertility rate −7.016*** (1.022) 1.165* (.481) 
Female population (%) 1.971** (.749) .123 (.258) 
Religious fractionalization 16.904*** (4.547) .392 (2.275) 
Communist (1 = yes) −3.005 (4.033) 13.575*** (1.889) 
Women's INGO linkages −.111 (.058) .049* (.024) 
CEDAW member (1 = yes) −.783 (1.960) 1.796 (.952) 
Proportional representation (1 = yes)   3.751*** (1.104) 
Years of female suffrage   .078* (.032) 
Gendered educational attainment ratio   .064* (.028) 
Women's share of parliamentarians .225** (.072)   
Constant −11.381 (39.480) −15.568 (12.281) 
F 460.898  220.524  
df 59  61  
Adjusted R-squared .761  .462  
a

Never colonized the omitted reference group.

b

Standardized to have mean = 0 and standard deviation = 1.

Standard errors, adjusted for first-order autocorrelation, in parentheses.

Fixed effects for years included in the analyses but not reported.

*p < .05, **p < .01, ***p < .001 (two-tailed).

N(country-year observations) = 3,878; N(countries) = 154.

FIGURE A1.

Coefficient estimates for the effect of Huntington's civilizational indicators on women's educational and political outcomes

FIGURE A1.

Coefficient estimates for the effect of Huntington's civilizational indicators on women's educational and political outcomes

FIGURE A2.

Coefficient estimates for the effect of Welzel's cultural zone indicators on women's educational and political outcomes

FIGURE A2.

Coefficient estimates for the effect of Welzel's cultural zone indicators on women's educational and political outcomes

FIGURE A3.

Coefficient estimates for the effect of U.N. region indicators on women's educational and political outcomes

FIGURE A3.

Coefficient estimates for the effect of U.N. region indicators on women's educational and political outcomes

FIGURE A4.

Coefficient estimates for the effect of consolidated U.N. region indicators on women's educational and political outcomes

FIGURE A4.

Coefficient estimates for the effect of consolidated U.N. region indicators on women's educational and political outcomes

FIGURE A5.

Coefficient estimates for the effect of colonial legacy indicators on women's educational and political outcomes

FIGURE A5.

Coefficient estimates for the effect of colonial legacy indicators on women's educational and political outcomes

FIGURE A6.

Marginal effects of economic development on women's educational and political outcomes, by Huntington's civilizational indicators and Welzel's cultural zones

FIGURE A6.

Marginal effects of economic development on women's educational and political outcomes, by Huntington's civilizational indicators and Welzel's cultural zones

FIGURE A7.

Marginal effects of economic development on women's educational and political outcomes, by U.N. region indicators

FIGURE A7.

Marginal effects of economic development on women's educational and political outcomes, by U.N. region indicators

FIGURE A8.

Marginal effects of economic development on women's educational and political outcomes, by colonial legacy indicators

FIGURE A8.

Marginal effects of economic development on women's educational and political outcomes, by colonial legacy indicators

Overall, patterns of results using these alternative coding schemes and indicators are similar to those we report using our cultural zone classifications. In general, cross-national differences in women's educational attainment relative to men are much smaller than cross-national differences in women's share of parliamentarians. In all but a handful of cases, women's share of parliamentarians is higher in the “West”—broadly or narrowly defined, depending on the set of indicators in question—than it is in other cultural zones or regions, even in the presence of control variables. Conversely, much of the cross-national variation in women's educational attainment traces to other political, economic, social, and demographic factors encapsulated by our control variables. In the fully specified models, many non-Western cultural zones or regions have gendered education ratios that are statistically equivalent to or even higher than the “West.” Moreover, each set of cultural zone and regional indicators conditions the relationship between GDP per capita and women's outcomes in similar ways. In particular, the marginal effects of economic development across cultural zones are generally stronger for educational outcomes than for political outcomes, with development improving the former, on balance, but exerting little effect on the latter.

Consider the first set of results, based on Huntington's (1996) civilizational framework. The fully specified model in Table A1 shows that only two civilizational groupings—Islamic and Sinic—have gendered educational attainment ratios that are significantly lower than the West, net of control variables. In contrast, Table A2 demonstrates that women's share of parliamentarians is significantly higher in the West than in six of nine non-Western civilizations: Buddhist, Islamic, Japanese, Latin American, Orthodox, and Sinic. Analyses that condition cross-cultural variation on economic development (Table A3, with marginal effects plotted in the top panel of Figure A6) suggest that women's share of parliamentarians will continue to diverge across civilizations as countries develop. In Western countries, women's share of parliamentarians increases with per capita GDP, whereas in Hindu and Islamic countries, as well as in Japan (a single-country civilization in Huntington's framework), it declines with development. Conversely, women's educational attainment increases with development in Hindu and “other” countries, but otherwise civilizational affiliation does not significantly condition the relationship between GDP per capita and gendered educational outcomes. All told, these patterns suggest inter-civilizational convergence at best or stable differences at worst in women's educational outcomes, compared with cross-cultural divergence in gendered political representation.

These overall patterns hold when using Welzel's cultural zone indicators. Indeed, if anything, the results presented in Tables A4 through A6 reinforce our main conclusions with even greater force. According to the fully specified model in Table A4, none of the cultural zones have gendered educational attainment ratios that are significantly lower relative to the Reformed West, the omitted reference category. In fact, women's educational outcomes in three zones—the New West, Orthodox East, and Latin America—exceed those of the Reformed West, all else being equal. Table A5, which analyzes cross-cultural variation in women's share of parliamentarians, tells a much different story. In that analysis, each and every cultural zone indicator is negative and statistically significant, indicating that women's legislative representation is lower in these countries than in the countries of the Reformed West. When conditioned by GDP per capita, Table A6 and the bottom panel of Figure A6 show cross-cultural convergence in gendered educational attainment ratios, with four sets of countries—sub-Saharan African, Latin American, Indic Eastern, and Sinic Eastern—catching up with Western nations as they develop economically.

The next set of analyses addresses regional variation in women's outcomes, rather than cross-cultural differences per se. According to the results in Table A7, women's educational attainment relative to men is significantly lower in roughly half of the 20 fine-grained regions than in Northern European countries (the omitted reference group), once other factors are taken into account. Even so, comparing effect sizes across the baseline, culture “plus” development, and fully specified models demonstrates that the presence of control variables greatly attenuates the magnitude of this variation. An even closer look shows that large parts of Africa and Asia account for most of the significant differences. All African regions except for Southern Africa have lower gendered educational attainment ratios than Northern Europe; Southern African countries’ ratios, conversely, are higher relative to Northern Europe. Ratios are also significantly lower in Western, Southern and South-Eastern Asia, but not in Central and Eastern Asia. More surprisingly perhaps, women's educational attainment in the countries of Western and Southern Europe lags behind their Northern European counterparts. In short, although there is evident regional variation in gendered educational outcomes, it appears to affect countries in Africa, Asia, and, to a somewhat lesser extent, Europe.

Patterns of regional variation in women's political outcomes, presented in Table A8, can be described much more easily: in every region except Australia and New Zealand, women's share of parliamentarians is lower vis-à-vis the countries of Northern Europe. Moreover, regional effect sizes are remarkably durable in the presence of control variables. In conjunction, these results lend continued support to the pattern of findings we reported in our article.

Table A9 and the top panel of Figure A7 analyze the effects of these fine-grained regional indicators when conditioned by economic development. As per capita GDP increases, women's educational attainment relative to men increases in four regions (Southern Asia, Melanesia, South America, Middle Africa), but it declines with development in three regions (Southern Africa, Australia and New Zealand, and Eastern Europe). For all other regions, the marginal effect of per capita GDP overlaps with zero, denoting statistical insignificance. Women's share of parliamentarians increases with per capita GDP in three regions (Northern Europe, Caribbean, Melanesia) and declines in one (Southern Asia).

Analyses that use the consolidated continental region indicators tell a similar story. In the fully specified model of Table A10, no large-scale region has significantly lower educational attainment ratios relative to Europe, the reference category; in fact, the Latin American effect is positive. Conversely, women's political representation in Africa and Asia is lower than it is in Europe, but the largely Western cognate countries of the Americas and Oceania do not differ from Europe in this regard. Economic development does not significantly condition regional variation in gendered educational attainment ratios, as indicated in the bottom panel of Figure A7; however, all marginal effects are positive, suggesting improvement. Women's share of parliamentarians, in stark contrast, shows inter-regional divergence. In the countries of Europe and Oceania, where women's political representation is already higher than average, the share of female parliamentarians increases with economic development. In Asia, however, rising GDP per capita correlates with a slight but statistically significant decline in women's political representation.

Finally, we conducted analyses using colonial legacies as another indicator of cross-national variation. As shown in Table A13, large baseline differences in women's educational attainment across countries clustered by colonial legacy weaken once other factors are included in the model. Gendered educational attainment ratios are significantly lower in the lone former Australian colony in the sample—Papua New Guinea—than in never-colonized countries. Former U.S. and, especially, British colonies have higher levels of female educational attainment. Otherwise, effect sizes in the fully specified model are relatively small and fail to achieve statistical significance. As for women's share of parliamentarians (Table A14), only two sets of countries—former French and Australian colonies—trail behind non-colonized countries. The majority of former French colonies are classified in our analysis as Islamic (54 percent of French colonies) or Animist—that is, sub-Saharan African (20 percent). Overall, these analyses suggest that colonial legacies, by themselves, have relatively little explanatory purchase with respect to our outcomes.

Colonial legacies do appear to moderate the relationship between economic development and women's outcomes, however. Table A15 analyzes and Figure A8 illustrates the marginal effects of economic development on women's outcomes by colonial legacy. Women's educational outcomes improve with per capita GDP in former Portuguese, Dutch, Australian, and British colonies. Marginal effects are otherwise statistically unreliable. Women's share of parliamentarians increases as a function of development only among never-colonized countries, and even then the substantive effect is infinitesimal. Conversely, as former Portuguese, Australian, and U.S. colonies develop economically, women's parliamentary representation declines. Again, these results point to either convergence or stability in women's educational attainment but increasing divergence in women's political outcomes.

SECTION 2: ANALYSES ESTIMATED ON SAMPLES SPLIT BY OCED MEMBERSHIP

In our article we explored whether economic development, measured as GDP per capita, moderates cross-cultural variation in women's educational and political outcomes. Here we assess a similar issue with a different measure of “economic development.” Rather than treat development as a continuum, in these analyses we compare “highly developed” and “less developed” countries based on membership in the Organization for Economic Co-operation and Development (OECD). Qualitatively, we compared women's outcomes in more and less developed countries; quantitatively, we continue to analyze the effect of GDP per capita within each subsample.

Table A16 presents heteroskedastic-robust OLS estimates for the effect of cultural zones, GDP per capita, and control variables on gendered educational attainment ratios for 33 members of the OECD. Note the reduced set of cultural zone indicators in this analysis: not all zones are represented among OECD members. Moreover, many of the represented zones contain only one or two member countries. Cross-cultural differences are consequently somewhat more pronounced. For example, the large and statistically significant negative coefficient on the Middle Eastern Islamic indicator in the fully specified model describes the effect for only one country, Turkey. (Turkey is the lone Islamic country represented in this subsample). Likewise, Greece is the sole Orthodox member of the OECD, while Chile and Mexico are the only member states from Latin America. For these reasons, the significantly negative coefficients on these zones must be interpreted with caution.

TABLE A16.
Heteroskedastic-robust OLS estimates for the effect of cultural zones, economic development, and control variables on gendered educational attainment ratios, OECD countries only, 1980 to 2010
Culture onlyNet of GDP/capitaFully specified
Coeff.(S.E.)Coeff.(S.E.)Coeff.(S.E.)
Catholic (Western)a −5.264*** (1.211) −5.257*** (1.327) −3.843* (1.728) 
Catholic (Latin American)a −11.331*** (.962) −11.288*** (2.431) −12.048*** (3.584) 
Islamic (Middle Eastern)a −37.367*** (.635) −37.317*** (2.583) −40.993*** (4.463) 
Orthodoxa −10.169*** (.626) −10.151*** (1.082) −11.250*** (2.220) 
Buddhista −4.560 (4.191) −4.551 (4.141) −3.214 (3.889) 
Non-religiousa −4.581*** (.976) −4.549* (1.930) −1.107 (2.879) 
GDP/capita (2005 US$, ln  .028 (1.417) −1.213 (1.758) 
Democracy/Autocracy     1.323 (.936) 
Trade (% GDP)     −.013 (.014) 
International war (1 = yes)     −3.974 (2.344) 
Fertility rate     3.461 (3.136) 
Female population (%)     −1.584 (1.295) 
Religious fractionalization     −3.457 (2.351) 
Women's INGO linkages     .025 (.022) 
CEDAW member (1 = yes)     .464 (1.382) 
Women's share of parliamentarians     −.050 (.052) 
Constant 93.084*** (.973) 92.799*** (14.342) 168.587* (76.108) 
F 2851.280  2.2e+05  8477.857  
df 36  37  46  
Adjusted R-squared .786  .786  .815  
Culture onlyNet of GDP/capitaFully specified
Coeff.(S.E.)Coeff.(S.E.)Coeff.(S.E.)
Catholic (Western)a −5.264*** (1.211) −5.257*** (1.327) −3.843* (1.728) 
Catholic (Latin American)a −11.331*** (.962) −11.288*** (2.431) −12.048*** (3.584) 
Islamic (Middle Eastern)a −37.367*** (.635) −37.317*** (2.583) −40.993*** (4.463) 
Orthodoxa −10.169*** (.626) −10.151*** (1.082) −11.250*** (2.220) 
Buddhista −4.560 (4.191) −4.551 (4.141) −3.214 (3.889) 
Non-religiousa −4.581*** (.976) −4.549* (1.930) −1.107 (2.879) 
GDP/capita (2005 US$, ln  .028 (1.417) −1.213 (1.758) 
Democracy/Autocracy     1.323 (.936) 
Trade (% GDP)     −.013 (.014) 
International war (1 = yes)     −3.974 (2.344) 
Fertility rate     3.461 (3.136) 
Female population (%)     −1.584 (1.295) 
Religious fractionalization     −3.457 (2.351) 
Women's INGO linkages     .025 (.022) 
CEDAW member (1 = yes)     .464 (1.382) 
Women's share of parliamentarians     −.050 (.052) 
Constant 93.084*** (.973) 92.799*** (14.342) 168.587* (76.108) 
F 2851.280  2.2e+05  8477.857  
df 36  37  46  
Adjusted R-squared .786  .786  .815  
a

Western Protestant countries are the omitted reference group.

Standard errors, adjusted for first-order autocorrelation, in parentheses.

Fixed effects for years included in the analyses but not reported.

*p < .05, **p < .01, ***p < .001 (two-tailed).

N(country-year observations) = 796; N(countries) = 33.

Table A17 conducts similar analyses on a sample of 126 non-OECD member countries. Although more cultural zones are represented in this analysis, no Western Protestant nations are included; as such, the reference category for these models is Western Catholic. The fully specified model demonstrates that cross-cultural variation in women's educational attainment among developing countries is low. Only three cultural zone coefficients achieve statistical significance in this model, and one—for Latin American countries—is significantly positive (indicating more equitable educational outcomes relative to the Catholic West).

TABLE A17.
Heteroskedastic-robust OLS estimates for the effect of cultural zones, economic development, and control variables on gendered educational attainment ratios, non-OECD countries only, 1980 to 2010
Culture onlyNet of GDP/capitaFully specified
Coeff.(S.E.)Coeff.(S.E.)Coeff.(S.E.)
Protestant (non-Western)a −25.952*** (6.762) −6.300 (5.668) 9.964 (6.106) 
Catholic (Latin American)a −1.707 (3.175) 8.255* (4.064) 19.670*** (3.729) 
Catholic (Other)a −31.878*** (8.707) −9.695 (10.271) 16.279 (8.561) 
Islamic (Middle Eastern)a −30.947*** (4.856) −26.012*** (5.835) 1.695 (6.187) 
Islamic (non-Middle Eastern)a −39.909*** (4.708) −16.906** (6.366) .676 (5.635) 
Orthodoxa −7.521 (6.326) 5.971 (6.444) −.076 (3.856) 
Buddhista −18.898* (8.065) 1.777 (7.648) 2.045 (6.388) 
Hindua −32.491** (12.288) −12.020 (9.396) −16.446* (7.059) 
Animista −39.626*** (7.887) −14.372 (8.639) 4.135 (8.533) 
Non-religiousa −2.285 (5.954) 9.443* (4.014) 1.921 (3.841) 
Confuciana −24.719*** (2.958) −16.793 (9.730) −28.339* (11.647) 
GDP/capita (2005 US$, ln  9.537*** (1.582) .015 (2.099) 
Democracy/Autocracy     .400 (.245) 
OPEC member (1 = yes)     5.284 (4.570) 
Trade (% GDP)     .074* (.037) 
International war (1 = yes)     5.742 (2.994) 
Civil war (1 = yes)     −1.676 (3.476) 
Fertility rate     −10.907*** (1.332) 
Female population (%)     .405 (.659) 
Religious fractionalization     12.512 (7.320) 
Communist (1 = yes)     −3.573 (4.397) 
Women's INGO linkages     −.049 (.072) 
CEDAW member (1 = yes)     −1.415 (2.609) 
Women's share of parliamentarians     .163 (.108) 
Constant 83.164*** (2.818) −3.017 (14.404) 78.308 (43.548) 
F 16.810  26.471  43.630  
df 41  42  54  
Adjusted R-squared .403  .536  .735  
Culture onlyNet of GDP/capitaFully specified
Coeff.(S.E.)Coeff.(S.E.)Coeff.(S.E.)
Protestant (non-Western)a −25.952*** (6.762) −6.300 (5.668) 9.964 (6.106) 
Catholic (Latin American)a −1.707 (3.175) 8.255* (4.064) 19.670*** (3.729) 
Catholic (Other)a −31.878*** (8.707) −9.695 (10.271) 16.279 (8.561) 
Islamic (Middle Eastern)a −30.947*** (4.856) −26.012*** (5.835) 1.695 (6.187) 
Islamic (non-Middle Eastern)a −39.909*** (4.708) −16.906** (6.366) .676 (5.635) 
Orthodoxa −7.521 (6.326) 5.971 (6.444) −.076 (3.856) 
Buddhista −18.898* (8.065) 1.777 (7.648) 2.045 (6.388) 
Hindua −32.491** (12.288) −12.020 (9.396) −16.446* (7.059) 
Animista −39.626*** (7.887) −14.372 (8.639) 4.135 (8.533) 
Non-religiousa −2.285 (5.954) 9.443* (4.014) 1.921 (3.841) 
Confuciana −24.719*** (2.958) −16.793 (9.730) −28.339* (11.647) 
GDP/capita (2005 US$, ln  9.537*** (1.582) .015 (2.099) 
Democracy/Autocracy     .400 (.245) 
OPEC member (1 = yes)     5.284 (4.570) 
Trade (% GDP)     .074* (.037) 
International war (1 = yes)     5.742 (2.994) 
Civil war (1 = yes)     −1.676 (3.476) 
Fertility rate     −10.907*** (1.332) 
Female population (%)     .405 (.659) 
Religious fractionalization     12.512 (7.320) 
Communist (1 = yes)     −3.573 (4.397) 
Women's INGO linkages     −.049 (.072) 
CEDAW member (1 = yes)     −1.415 (2.609) 
Women's share of parliamentarians     .163 (.108) 
Constant 83.164*** (2.818) −3.017 (14.404) 78.308 (43.548) 
F 16.810  26.471  43.630  
df 41  42  54  
Adjusted R-squared .403  .536  .735  
a

Western Catholic countries are the omitted reference group.

Standard errors, adjusted for first-order autocorrelation, in parentheses.

Fixed effects for years included in the analyses but not reported.

*p < .05, **p < .01, ***p < .001 (two-tailed).

N(country-year observations) = 3,016; N(countries) = 126.

The analyses reported in Tables A18 and A19 show that cross-cultural differences in women's political outcomes for both developed and developing countries is much more pronounced, relative to variation in educational outcomes. Women's share of parliamentarians among the Western Catholic, Orthodox, Buddhist, and “secular” countries of the OECD is lower when compared with the Protestant West, all else being equal. The coefficients on the Latin American Catholic and Middle Eastern Islamic indicators are also negatively signed but statistically insignificant. Again, it bears repeating at only two Latin American Catholic nations and one Islamic country appear in this sample, which reduces the statistical power of these estimates. Turning to the non-OECD subsample, seven of 11 cultural zones have significantly negative coefficients in the fully specified model of Table A19, indicating that these countries have fewer women in parliament as a share of total members than countries in the Catholic West do. It is noteworthy that two of the four non-significant coefficient estimates pertain to Catholic nations in Latin America and elsewhere, while a third is associated with predominantly secular societies.

TABLE A18.
Heteroskedastic-robust OLS estimates for the effect of cultural zones, economic development, and control variables on women's share of parliamentarians, OECD countries only, 1980 to 2010
Culture onlyNet of GDP/capitaFully specified
Coeff.(S.E.)Coeff.(S.E.)Coeff.(S.E.)
Catholic (Western)a −9.387** (3.549) −8.107* (3.555) −6.259* (3.148) 
Catholic (Latin American)a −11.026*** (3.139) −3.261 (4.190) −.367 (5.021) 
Islamic (Middle Eastern)a −22.358*** (3.164) −13.387** (4.496) −7.185 (8.714) 
Orthodoxa −17.561*** (3.162) −14.239*** (3.300) −14.715*** (3.767) 
Buddhista −17.783*** (3.509) −16.172*** (3.116) −15.400*** (2.817) 
Non-religiousa −14.044*** (3.085) −8.303* (3.686) −6.902* (3.519) 
GDP/capita (2005 US$, ln  4.982** (1.823) 4.397* (1.848) 
Democracy/Autocracy     −.977 (.526) 
Trade (% GDP)     −.005 (.028) 
International war (1 = yes)     −8.161* (4.093) 
Fertility rate     −2.183 (3.428) 
Female population (%)     −1.419 (1.726) 
Religious fractionalization     −5.526 (5.292) 
Women's INGO linkages     .113* (.044) 
CEDAW member (1 = yes)     4.970** (1.566) 
Proportional representation (1 = yes)     13.133*** (3.179) 
Years of female suffrage     .107* (.049) 
Gendered educational attainment ratio     .230 (.229) 
Constant 16.694*** (3.331) −34.139 (19.171) 19.013 (87.790) 
F 3899.905  212.526  90.432  
df 36  37  48  
Adjusted R-squared .474  .505  .730  
Culture onlyNet of GDP/capitaFully specified
Coeff.(S.E.)Coeff.(S.E.)Coeff.(S.E.)
Catholic (Western)a −9.387** (3.549) −8.107* (3.555) −6.259* (3.148) 
Catholic (Latin American)a −11.026*** (3.139) −3.261 (4.190) −.367 (5.021) 
Islamic (Middle Eastern)a −22.358*** (3.164) −13.387** (4.496) −7.185 (8.714) 
Orthodoxa −17.561*** (3.162) −14.239*** (3.300) −14.715*** (3.767) 
Buddhista −17.783*** (3.509) −16.172*** (3.116) −15.400*** (2.817) 
Non-religiousa −14.044*** (3.085) −8.303* (3.686) −6.902* (3.519) 
GDP/capita (2005 US$, ln  4.982** (1.823) 4.397* (1.848) 
Democracy/Autocracy     −.977 (.526) 
Trade (% GDP)     −.005 (.028) 
International war (1 = yes)     −8.161* (4.093) 
Fertility rate     −2.183 (3.428) 
Female population (%)     −1.419 (1.726) 
Religious fractionalization     −5.526 (5.292) 
Women's INGO linkages     .113* (.044) 
CEDAW member (1 = yes)     4.970** (1.566) 
Proportional representation (1 = yes)     13.133*** (3.179) 
Years of female suffrage     .107* (.049) 
Gendered educational attainment ratio     .230 (.229) 
Constant 16.694*** (3.331) −34.139 (19.171) 19.013 (87.790) 
F 3899.905  212.526  90.432  
df 36  37  48  
Adjusted R-squared .474  .505  .730  
a

Western Protestant countries are the omitted reference group.

Standard errors, adjusted for first-order autocorrelation, in parentheses.

Fixed effects for years included in the analyses but not reported.

*p < .05, **p < .01, ***p < .001 (two-tailed).

N(country-year observations) = 796; N(countries) = 33.

TABLE A19.
Heteroskedastic-robust OLS estimates for the effect of cultural zones, economic development, and control variables on women's share of parliamentarians, non-OECD countries only, 1980 to 2010
Culture onlyNet of GDP/capitaFully specified
Coeff.(S.E.)Coeff.(S.E.)Coeff.(S.E.)
Protestant (non-Western)a −6.512* (2.777) −8.024* (3.153) −5.544* (2.686) 
Catholic (Latin American)a −2.849 (2.341) −3.616 (2.384) −1.687 (1.672) 
Catholic (Other)a −1.967 (2.479) −3.674 (2.362) −1.546 (2.478) 
Islamic (Middle Eastern)a −12.526*** (2.228) −12.906*** (2.096) −7.360** (2.385) 
Islamic (non-Middle Eastern)a −5.972** (2.069) −7.742*** (2.301) −3.652* (1.746) 
Orthodoxa −5.572* (2.633) −6.610* (2.710) −5.921** (1.801) 
Buddhista −6.345* (2.612) −7.937** (2.606) −7.436*** (1.803) 
Hindua −2.459 (3.539) −4.035 (3.600) −1.053 (2.286) 
Animista −5.333* (2.443) −7.277** (2.442) −5.080* (2.197) 
Non-religiousa .829 (3.096) −.074 (2.856) −3.199 (1.795) 
Confuciana −.105 (4.901) −.715 (4.134) −6.296** (2.013) 
GDP/capita (2005 US$, ln  −.734 (.419) −.994 (.512) 
Democracy/Autocracy     −.222** (.073) 
OPEC member (1 = yes)     −.641 (1.198) 
Trade (% GDP)     .010 (.008) 
International war (1 = yes)     .916 (1.582) 
Civil war (1 = yes)     −1.683 (1.110) 
Fertility rate     −.405 (.425) 
Female population (%)     .438* (.204) 
Religious fractionalization     4.361* (2.183) 
Communist (1 = yes)     11.697*** (2.103) 
Women's INGO linkages     .046 (.027) 
CEDAW member (1 = yes)     .179 (.746) 
Proportional representation (1 = yes)     2.289* (1.031) 
Years of female suffrage     −.034 (.028) 
Gendered educational attainment ratio     .038 (.027) 
Constant 11.616*** (2.203) 18.250*** (4.274) −8.889 (12.678) 
F 8.633  9.294  19.655  
df 41  42  56  
Adjusted R-squared .267  .275  .436  
Culture onlyNet of GDP/capitaFully specified
Coeff.(S.E.)Coeff.(S.E.)Coeff.(S.E.)
Protestant (non-Western)a −6.512* (2.777) −8.024* (3.153) −5.544* (2.686) 
Catholic (Latin American)a −2.849 (2.341) −3.616 (2.384) −1.687 (1.672) 
Catholic (Other)a −1.967 (2.479) −3.674 (2.362) −1.546 (2.478) 
Islamic (Middle Eastern)a −12.526*** (2.228) −12.906*** (2.096) −7.360** (2.385) 
Islamic (non-Middle Eastern)a −5.972** (2.069) −7.742*** (2.301) −3.652* (1.746) 
Orthodoxa −5.572* (2.633) −6.610* (2.710) −5.921** (1.801) 
Buddhista −6.345* (2.612) −7.937** (2.606) −7.436*** (1.803) 
Hindua −2.459 (3.539) −4.035 (3.600) −1.053 (2.286) 
Animista −5.333* (2.443) −7.277** (2.442) −5.080* (2.197) 
Non-religiousa .829 (3.096) −.074 (2.856) −3.199 (1.795) 
Confuciana −.105 (4.901) −.715 (4.134) −6.296** (2.013) 
GDP/capita (2005 US$, ln  −.734 (.419) −.994 (.512) 
Democracy/Autocracy     −.222** (.073) 
OPEC member (1 = yes)     −.641 (1.198) 
Trade (% GDP)     .010 (.008) 
International war (1 = yes)     .916 (1.582) 
Civil war (1 = yes)     −1.683 (1.110) 
Fertility rate     −.405 (.425) 
Female population (%)     .438* (.204) 
Religious fractionalization     4.361* (2.183) 
Communist (1 = yes)     11.697*** (2.103) 
Women's INGO linkages     .046 (.027) 
CEDAW member (1 = yes)     .179 (.746) 
Proportional representation (1 = yes)     2.289* (1.031) 
Years of female suffrage     −.034 (.028) 
Gendered educational attainment ratio     .038 (.027) 
Constant 11.616*** (2.203) 18.250*** (4.274) −8.889 (12.678) 
F 8.633  9.294  19.655  
df 41  42  56  
Adjusted R-squared .267  .275  .436  
a

Western Catholic countries are the omitted reference group.

Standard errors, adjusted for first-order autocorrelation, in parentheses.

Fixed effects for years included in the analyses but not reported.

*p < .05, **p < .01, ***p < .001 (two-tailed).

N(country-year observations) = 3,016; N(countries) = 126.

Estimating models that predict women's share of parliamentarians on subsamples defined by broad levels of economic development is important for another reason. Past research suggests that standard theoretical predictors of women's legislative representation, including “quantitative” levels of economic development, work differently in developed and developing countries (Viterna, Fallon, and Beckfield 2008). Standard explanations for variation in women's share of parliamentarians seem to fare better for developed than developing nations. Although the core variables of theoretical interest in our analysis—cultural zone indicators—generally perform as expected in analyses conducted on subsamples defined by OECD membership, control variables operationalizing other factors of theoretical import do show variable effects, in line with Viterna, Fallon, and Beckfield's analyses. Patterns of explained variation also differ. For instance, compared with the OCED sample, R-squared values are lower in the analyses of non-OCED countries. The fully specified model accounts for 73 percent of the explained variation in women's share of parliamentarians among OCED members but only 44 percent among non-OECD members. As for individual predictors, developing countries account for the negative relationship between democracy and women's legislative representation observed in the complete sample, whereas per capita GDP and linkages to women's INGOs increase women's representation only in developed countries.

In sum, these analyses confirm that the overall patterns of cross-cultural variation in women's educational and political outcomes are robust even when estimated using subsamples defined by OECD membership. More specifically, variation in women's outcomes across cultural zones is greater in the political domain than in educational outcomes.

SECTION 3: DIAGNOSING THE POTENTIAL ENDOGENEITY OF ECONOMIC DEVELOPMENT

In our article we analyzed whether the relationship between economic development, on the one hand, and women's educational and political outcomes, on the other hand, varies cross-culturally by estimating models that include interaction terms between GDP per capita and each cultural zone indicator. We found that the effect of economic development in most cultural zones is generally stronger (and more often positive) for women's educational attainment than for women's representation in national parliaments.

In these analyses, we assumed that economic development is causally prior to women's outcomes. This assumption is problematic, however, to the extent that increased levels of gender equality is a cause rather than a consequence of economic development. To be sure, much research demonstrates that gender equity in educational outcomes, in particular, contributes to economic growth and development (e.g., Berik, Rodgers, and Seguio 2009; Kabeer and Natali 2013; Klasen 2002; Klasen and Lamanna 2009).

To rule out the possibility that reverse causality introduces endogeneity bias in our analyses, we estimate two-stage limited-information maximum-likelihood (LIML) models with instrumental variables to isolate causal directionality. Compared with conventional two-stage least-squares estimators, models estimated by LIML are less biased, more efficient, and perform better in the presence of weak instruments (Angrist and Pischke 2009; Hahn, Hauman, and Kuersteiner 2004; Stock and Yogo 2005).

Two-stage models are only as good as the measures used to instrument potentially endogenous variables. Suitable instruments must satisfy two conditions. First, instruments must be sufficiently correlated with the potentially endogenous variable of interest (in our case, GDP per capita). Second, they must be uncorrelated with the error term of the structural equation (here, the equations modeling women's educational and political equality). The first condition, known as instrument relevance, can be directly observed, whereas the second condition, instrument exogeneity, must be inferred because the error term is unobserved.

We selected carbon emissions (in metric tons per capita, logged) as an instrument for GDP per capita. Diagnostic tests, discussed below, suggest that this instrument is both relevant and exogenous. Some of our analyses estimate just-identified models whereas others are overidentified. In the just-identified models, carbon emissions serve as the lone instrument. Angrist and Pischke (2009:213) encourage practitioners to select the “single best instrument and report just-identified estimates using this one only.” However, to obtain diagnostics for instrument exogeneity models must be overidentified—that is, the number of instruments must exceed the number of endogenous variables instrumented. To estimate overidentified models we add a squared term for carbon emissions. This specification is inspired by research demonstrating the existence of an “environmental Kuznets curve,” whereby carbon emissions increase with per capita GDP up to a threshold but then decline thereafter (York, Rosa, and Dietz 2003; Jorgenson and Birkholz 2010; Jorgenson and Clark 2011).

Before proceeding with our two-stage analyses, we first present models designed to assess the relevance and exogeneity of our instruments. Table A20 reports heteroskedastic-robust OLS estimates for the effect of carbon dioxide emissions and its squared term on (1) GDP per capita and (2) women's educational and political outcomes, net of cultural zones and control variables. These instruments are relevant if the estimated effect of carbon emissions (linear and squared) on GDP per capita is statistically significant. Exogeneity is suggested if the estimated effect of carbon emissions on women's educational and political outcomes is not statistically significant. The results indicate that both conditions are satisfied. In the analyses of GDP per capita, the linear coefficient on carbon emissions is positive and statistically significant while the quadratic coefficient is negative and statistically significant, a pattern that comports with the environmental Kuznets curve hypothesis. Conversely, carbon emissions do not reliably predict gendered educational attainment ratios or women's share of parliamentarians.

TABLE A20.
Heteroskedastic-robust OLS estimates for the effect of carbon dioxide emissions, cultural zones, and control variables on GDP per capita and women's outcomes, 1980 to 2010
Gendered educational attainment ratioWomen's share of parliamentarians
Relevancy: Outcome = GDP/capita (2005 US$, ln)Exogeneity: Outcome = Women's share of parliamentariansRelevancy: Outcome = GDP/capita (2005 US$, ln)Exogeneity: Outcome = Gendered educational attainment ratio
Coeff.(S.E.)Coeff.(S.E.)Coeff.(S.E.)Coeff.(S.E.)
Carbon emissions per capita .140*** (.024) .051 (.339) .142*** (.022) .107 (.180) 
Carbon emissions per capita, squared −.002** (.001) .005 (.007) −.002*** (.001) −.003 (.003) 
Protestant (non-Western) −1.045** (.330) 3.379 (5.295) −1.211*** (.300) −14.635*** (3.562) 
Catholic (Western) −.407* (.207) −7.145** (2.403) −.505** (.192) −10.412*** (3.154) 
Catholic (Latin American) −.839*** (.249) 12.212*** (3.695) −1.047*** (.239) −13.668*** (3.252) 
Catholic (Other) −.713 (.422) 3.219 (7.033) −.856* (.415) −11.839** (3.660) 
Islamic (Middle Eastern) −1.016** (.367) −9.520 (5.510) −1.049** (.361) −18.481*** (3.307) 
Islamic (non-Middle Eastern) −1.475*** (.320) −8.124 (5.441) −1.490*** (.298) −14.173*** (3.088) 
Orthodox −1.358*** (.361) −5.639 (3.095) −1.498*** (.322) −16.375*** (3.059) 
Buddhist −1.414** (.468) −4.955 (5.035) −1.450*** (.398) −18.109*** (2.938) 
Hindu −1.913*** (.420) −22.106** (6.789) −1.968*** (.425) −10.071** (3.388) 
Animist −1.175** (.368) −3.490 (7.912) −1.309*** (.355) −14.166*** (3.435) 
Nonreligious −1.405*** (.301) −5.090 (3.909) −1.416*** (.283) −14.131*** (3.091) 
Confucian −1.821*** (.354) −31.901*** (8.382) −1.772*** (.382) −16.125*** (3.707) 
Democracy/Autocracy .009 (.009) .234 (.230) .006 (.009) −.268*** (.075) 
OPEC member (1 = yes) .463* (.201) 4.778 (4.043) .353 (.202) −1.738 (1.370) 
Trade (% GDP) .004*** (.001) .050* (.022) .004*** (.001) .015 (.009) 
International war (1 = yes) −.498 (.254) 5.113* (2.454) −.450 (.242) −1.877 (2.776) 
Civil war (1 = yes) −.280 (.161) −.092 (3.241) −.267 (.158) −1.417 (1.138) 
Fertility rate −.319*** (.042) −10.216*** (.956) −.286*** (.052) .472 (.398) 
Female population (%) −.050 (.034) .430 (.554) −.038 (.035) .338 (.316) 
Religious fractionalization −.403 (.241) 9.102 (5.062) −.324 (.238) .320 (2.191) 
Communist (1 = yes) −.144 (.188) −3.673 (3.950) −.089 (.178) 13.216*** (2.052) 
Women's INGO linkages .010*** (.002) −.019 (.042) .011*** (.002) .042* (.021) 
CEDAW member (1 = yes) −.162 (.092) .014 (1.829) −.179* (.083) 1.550 (1.010) 
Proportional representation (1 = yes)     .224* (.110) 4.027*** (1.121) 
Years of female suffrage     −.008* (.003) −.006 (.025) 
Gendered educational attainment ratio     .004 (.004) .053* (.027) 
Women's share of parliamentarians −.002 (.005) .153* (.075)     
Constant 12.006*** (1.773) 82.025** (30.231) 11.179*** (1.803) −6.986 (15.839) 
F 42.444  47.940  44.748  14.311  
df 56  56  58  58  
Adjusted R-squared .864  .788  .871  .526  
Gendered educational attainment ratioWomen's share of parliamentarians
Relevancy: Outcome = GDP/capita (2005 US$, ln)Exogeneity: Outcome = Women's share of parliamentariansRelevancy: Outcome = GDP/capita (2005 US$, ln)Exogeneity: Outcome = Gendered educational attainment ratio
Coeff.(S.E.)Coeff.(S.E.)Coeff.(S.E.)Coeff.(S.E.)
Carbon emissions per capita .140*** (.024) .051 (.339) .142*** (.022) .107 (.180) 
Carbon emissions per capita, squared −.002** (.001) .005 (.007) −.002*** (.001) −.003 (.003) 
Protestant (non-Western) −1.045** (.330) 3.379 (5.295) −1.211*** (.300) −14.635*** (3.562) 
Catholic (Western) −.407* (.207) −7.145** (2.403) −.505** (.192) −10.412*** (3.154) 
Catholic (Latin American) −.839*** (.249) 12.212*** (3.695) −1.047*** (.239) −13.668*** (3.252) 
Catholic (Other) −.713 (.422) 3.219 (7.033) −.856* (.415) −11.839** (3.660) 
Islamic (Middle Eastern) −1.016** (.367) −9.520 (5.510) −1.049** (.361) −18.481*** (3.307) 
Islamic (non-Middle Eastern) −1.475*** (.320) −8.124 (5.441) −1.490*** (.298) −14.173*** (3.088) 
Orthodox −1.358*** (.361) −5.639 (3.095) −1.498*** (.322) −16.375*** (3.059) 
Buddhist −1.414** (.468) −4.955 (5.035) −1.450*** (.398) −18.109*** (2.938) 
Hindu −1.913*** (.420) −22.106** (6.789) −1.968*** (.425) −10.071** (3.388) 
Animist −1.175** (.368) −3.490 (7.912) −1.309*** (.355) −14.166*** (3.435) 
Nonreligious −1.405*** (.301) −5.090 (3.909) −1.416*** (.283) −14.131*** (3.091) 
Confucian −1.821*** (.354) −31.901*** (8.382) −1.772*** (.382) −16.125*** (3.707) 
Democracy/Autocracy .009 (.009) .234 (.230) .006 (.009) −.268*** (.075) 
OPEC member (1 = yes) .463* (.201) 4.778 (4.043) .353 (.202) −1.738 (1.370) 
Trade (% GDP) .004*** (.001) .050* (.022) .004*** (.001) .015 (.009) 
International war (1 = yes) −.498 (.254) 5.113* (2.454) −.450 (.242) −1.877 (2.776) 
Civil war (1 = yes) −.280 (.161) −.092 (3.241) −.267 (.158) −1.417 (1.138) 
Fertility rate −.319*** (.042) −10.216*** (.956) −.286*** (.052) .472 (.398) 
Female population (%) −.050 (.034) .430 (.554) −.038 (.035) .338 (.316) 
Religious fractionalization −.403 (.241) 9.102 (5.062) −.324 (.238) .320 (2.191) 
Communist (1 = yes) −.144 (.188) −3.673 (3.950) −.089 (.178) 13.216*** (2.052) 
Women's INGO linkages .010*** (.002) −.019 (.042) .011*** (.002) .042* (.021) 
CEDAW member (1 = yes) −.162 (.092) .014 (1.829) −.179* (.083) 1.550 (1.010) 
Proportional representation (1 = yes)     .224* (.110) 4.027*** (1.121) 
Years of female suffrage     −.008* (.003) −.006 (.025) 
Gendered educational attainment ratio     .004 (.004) .053* (.027) 
Women's share of parliamentarians −.002 (.005) .153* (.075)     
Constant 12.006*** (1.773) 82.025** (30.231) 11.179*** (1.803) −6.986 (15.839) 
F 42.444  47.940  44.748  14.311  
df 56  56  58  58  
Adjusted R-squared .864  .788  .871  .526  

N(country-year observations) = 3,789; N(countries) = 150.

Standard errors, adjusted for first-order autocorrelation, in parentheses.

Fixed effects for years included in the analyses but not reported.

*p < .05, **p < .01, ***p < .001 (two-tailed).

Using these instruments, Table A21 presents the results of our two-stage models that endogenize GDP per capita. Coefficient estimates for the effect of (instrumented) GDP per capita on each measure of gender equality in both the just-identified and overidentified models are statistically insignificant, net of the cultural zone indicators and control variables. (For comparison, recall that the estimated effect of GDP per capita in the fully specified models of our analyses was also statistically insignificant.) The bottom of the table reports three diagnostic tests for assessing instrument relevance, instrument exogeneity, and endogeneity. The Kleibergen-Papp F tests clearly demonstrate that the models are strongly identified (i.e., relevant), and the Hansen J statistics (available in the overidentified models only) suggest that the exogeneity condition is also satisfied. The final test statistic is based on a Hausman-like test of endogeneity. These results suggest that the GDP per capita can indeed be treated as exogenous in our analyses, and that standard OLS models are preferred to two-stage estimates.

TABLE A21.
Two-stage limited-information maximum-likelihood (LIML) estimates for the effect of cultural zones, economic development, and control variables on women's educational and political outcomes, 1980 to 2010
Gendered educational attainment ratioWomen's share of parliamentarians
Just-identified modelaOveridentified modelbJust-identified modelaOveridentified modelb
Coeff.(S.E.)Coeff.(S.E.)Coeff.(S.E.)Coeff.(S.E.)
GDP/capita (2005 US$, ln3.183 (3.326) 1.215 (2.529) .125 (1.830) .568 (1.395) 
Protestant (non-Western) 7.796 (7.296) 4.228 (6.335) −14.747** (4.777) −13.868** (4.312) 
Catholic (Western) −5.368 (3.152) −6.483* (2.777) −10.475** (3.209) −10.186** (3.201) 
Catholic (Latin American) 15.883** (5.810) 13.058** (4.807) −13.769** (4.466) −13.037** (4.008) 
Catholic (Other) 6.484 (8.360) 3.824 (7.498) −11.970** (4.482) −11.306** (4.181) 
Islamic (Middle Eastern) −6.559 (6.551) −9.076 (6.239) −18.321*** (3.995) −17.739*** (3.744) 
Islamic (non-Middle Eastern) −2.658 (8.180) −6.777 (7.078) −14.187** (4.776) −13.252** (4.161) 
Orthodox −.579 (6.573) −4.176 (5.104) −16.378*** (4.420) −15.509*** (3.970) 
Buddhist .360 (8.285) −3.459 (7.081) −18.139*** (4.488) −17.261*** (3.938) 
Hindu −14.771 (10.958) −20.306* (9.516) −10.136 (5.761) −8.874 (4.849) 
Animist 1.549 (9.624) −2.288 (8.680) −14.317** (4.885) −13.395** (4.352) 
Nonreligious .218 (6.532) −3.017 (5.255) −14.163*** (4.074) −13.429*** (3.763) 
Confucian −25.499* (11.665) −30.469** (10.346) −16.087** (5.478) −14.998** (4.841) 
Democracy/Autocracy .205 (.228) .218 (.228) −.269*** (.073) −.270*** (.074) 
OPEC member (1 = yes) 2.817 (4.557) 4.266 (4.254) −1.669 (1.713) −1.949 (1.576) 
Trade (% GDP) .034 (.029) .046 (.026) .015 (.015) .013 (.012) 
International war (1 = yes) 6.959* (3.032) 5.868* (2.744) −1.880 (3.170) −1.650 (2.979) 
Civil war (1 = yes) 1.056 (3.571) .300 (3.341) −1.443 (1.291) −1.277 (1.255) 
Fertility rate −8.986*** (1.708) −9.823*** (1.469) .451 (.706) .621 (.589) 
Female population (%) .362 (.591) .154 (.643) .395 (.307) .437 (.261) 
Religious fractionalization 9.663 (5.193) 9.769 (5.039) .528 (2.247) .476 (2.289) 
Communist (1 = yes) −3.081 (4.155) −3.675 (4.070) 13.207*** (2.096) 13.315*** (2.105) 
Women's INGO linkages −.059 (.065) −.035 (.057) .043 (.032) .037 (.028) 
CEDAW member (1 = yes) .488 (1.885) −.075 (1.947) 1.585 (1.002) 1.718 (1.013) 
Proportional representation (1 = yes)     3.994** (1.296) 3.906** (1.248) 
Years of female suffrage     −.005 (.027) −.002 (.027) 
Gendered educational attainment ratio     .051 (.030) .049 (.029) 
Women's share of parliamentarians .156* (.077) .153* (.075)     
Constant 53.109 (56.106) 84.410 (53.562) −10.430 (31.270) −17.102 (23.750) 
F 60.00  52.36  14.27  14.12  
df 55  55  57  57  
Weak identification testc 25.07 (.00)  21.37 (.00)  29.90 (.00)  26.84 (.00)  
Overidentification testd –  .78 (.38)  –  .50 (.48)  
Endogeneity teste .34 (.56)  .00 (.99)  .11 (.74)  2.50 (.11)  
Gendered educational attainment ratioWomen's share of parliamentarians
Just-identified modelaOveridentified modelbJust-identified modelaOveridentified modelb
Coeff.(S.E.)Coeff.(S.E.)Coeff.(S.E.)Coeff.(S.E.)
GDP/capita (2005 US$, ln3.183 (3.326) 1.215 (2.529) .125 (1.830) .568 (1.395) 
Protestant (non-Western) 7.796 (7.296) 4.228 (6.335) −14.747** (4.777) −13.868** (4.312) 
Catholic (Western) −5.368 (3.152) −6.483* (2.777) −10.475** (3.209) −10.186** (3.201) 
Catholic (Latin American) 15.883** (5.810) 13.058** (4.807) −13.769** (4.466) −13.037** (4.008) 
Catholic (Other) 6.484 (8.360) 3.824 (7.498) −11.970** (4.482) −11.306** (4.181) 
Islamic (Middle Eastern) −6.559 (6.551) −9.076 (6.239) −18.321*** (3.995) −17.739*** (3.744) 
Islamic (non-Middle Eastern) −2.658 (8.180) −6.777 (7.078) −14.187** (4.776) −13.252** (4.161) 
Orthodox −.579 (6.573) −4.176 (5.104) −16.378*** (4.420) −15.509*** (3.970) 
Buddhist .360 (8.285) −3.459 (7.081) −18.139*** (4.488) −17.261*** (3.938) 
Hindu −14.771 (10.958) −20.306* (9.516) −10.136 (5.761) −8.874 (4.849) 
Animist 1.549 (9.624) −2.288 (8.680) −14.317** (4.885) −13.395** (4.352) 
Nonreligious .218 (6.532) −3.017 (5.255) −14.163*** (4.074) −13.429*** (3.763) 
Confucian −25.499* (11.665) −30.469** (10.346) −16.087** (5.478) −14.998** (4.841) 
Democracy/Autocracy .205 (.228) .218 (.228) −.269*** (.073) −.270*** (.074) 
OPEC member (1 = yes) 2.817 (4.557) 4.266 (4.254) −1.669 (1.713) −1.949 (1.576) 
Trade (% GDP) .034 (.029) .046 (.026) .015 (.015) .013 (.012) 
International war (1 = yes) 6.959* (3.032) 5.868* (2.744) −1.880 (3.170) −1.650 (2.979) 
Civil war (1 = yes) 1.056 (3.571) .300 (3.341) −1.443 (1.291) −1.277 (1.255) 
Fertility rate −8.986*** (1.708) −9.823*** (1.469) .451 (.706) .621 (.589) 
Female population (%) .362 (.591) .154 (.643) .395 (.307) .437 (.261) 
Religious fractionalization 9.663 (5.193) 9.769 (5.039) .528 (2.247) .476 (2.289) 
Communist (1 = yes) −3.081 (4.155) −3.675 (4.070) 13.207*** (2.096) 13.315*** (2.105) 
Women's INGO linkages −.059 (.065) −.035 (.057) .043 (.032) .037 (.028) 
CEDAW member (1 = yes) .488 (1.885) −.075 (1.947) 1.585 (1.002) 1.718 (1.013) 
Proportional representation (1 = yes)     3.994** (1.296) 3.906** (1.248) 
Years of female suffrage     −.005 (.027) −.002 (.027) 
Gendered educational attainment ratio     .051 (.030) .049 (.029) 
Women's share of parliamentarians .156* (.077) .153* (.075)     
Constant 53.109 (56.106) 84.410 (53.562) −10.430 (31.270) −17.102 (23.750) 
F 60.00  52.36  14.27  14.12  
df 55  55  57  57  
Weak identification testc 25.07 (.00)  21.37 (.00)  29.90 (.00)  26.84 (.00)  
Overidentification testd –  .78 (.38)  –  .50 (.48)  
Endogeneity teste .34 (.56)  .00 (.99)  .11 (.74)  2.50 (.11)  
a

Instrument: carbon dioxide emissions per capita.

b

Instruments: carbon dioxide emissions per capita and its squared term.

c

Kleibergen-Paap F test—H0: equation is weakly identified, i.e., the instruments are only weakly correlated with the endogenous regressor. Evaluated against Stock and Yogo (2005) critical value, 10% maximal IV size: 16.38 for just-identified models; 8.68 for overidentified models.

d

Hansen J statistic—H0: instruments are uncorrelated with the structural equation error term and are correctly excluded. Estimable for overidentified models only.

e

Test of endogenous regressors—H0: the suspected endogenous regressor can be treated as exogenous, indicating that the model estimated by OLS is preferred.

N(country-year observations) = 3,789; N(countries) = 150.

Standard errors, adjusted for first-order autocorrelation, in parentheses.

Fixed effects for years included in the analyses but not reported.

*p < .05, **p < .01, ***p < .001 (two-tailed).

The analyses in Table A22 present one final test of endogeneity. Following a procedure recommended by Wooldridge (2003:506–507), this test proceeds in three steps. First, we analyze first-stage equations by regressing GDP per capita on cultural zone indicators, control variables, and the instrumental variables (i.e., carbon emissions and its squared term). We then include the residuals from these equations as predictors in the structural equations of women's outcomes. If the coefficient on the residuals is statistically different from zero, we conclude that GDP per capita is endogenous. Statistically insignificant coefficient estimates suggest that GDP per capita is exogenous to women's outcomes. With respect to both gendered educational attainment ratios and women's share of parliamentarians, the estimated coefficients on the first-stage residuals are statistically insignificant.

TABLE A22.
Diagnosing the endogeneity of per capita GDP
Gendered educational attainment ratioWomen's share of parliamentarians
Coeff.(S.E.)Coeff.(S.E.)
First-stage residual .114 (3.087) −1.373 (1.615) 
GDP/capita (2005 US$, ln1.196 (2.499) .588 (1.337) 
Protestant (non-Western) 4.192 (6.322) −13.720** (4.335) 
Catholic (Western) −6.494* (2.773) −10.189** (3.237) 
Catholic (Latin American) 13.029** (4.791) −12.954** (3.954) 
Catholic (Other) 3.795 (7.521) −11.078** (4.180) 
Islamic (Middle Eastern) −9.104 (6.268) −17.821*** (3.677) 
Islamic (non-Middle Eastern) −6.820 (7.079) −13.163** (4.136) 
Orthodox −4.214 (5.110) −15.462*** (3.892) 
Buddhist −3.495 (7.012) −17.202*** (3.903) 
Hindu −20.361* (9.461) −8.841 (4.708) 
Animist −2.327 (8.661) −13.199** (4.390) 
Nonreligious −3.050 (5.247) −13.396*** (3.794) 
Confucian −30.518** (10.332) −14.963** (4.689) 
Democracy/Autocracy .219 (.228) −.270*** (.070) 
OPEC member (1 = yes) 4.281 (4.254) −1.847 (1.491) 
Trade (% GDP) .046 (.026) .013 (.012) 
International war (1 = yes) 5.855* (2.760) −1.574 (3.235) 
Civil war (1 = yes) .293 (3.345) −1.370 (1.186) 
Fertility rate −9.831*** (1.455) .586 (.527) 
Female population (%) .152 (.643) .442 (.250) 
Religious fractionalization 9.769 (5.041) .438 (2.212) 
Communist (1 = yes) −3.682 (4.079) 13.325*** (2.053) 
Women's INGO linkages −.035 (.057) .037 (.027) 
CEDAW member (1 = yes) −.080 (1.930) 1.758 (1.020) 
Proportional representation (1 = yes)   3.901** (1.206) 
Years of female suffrage   −.001 (.027) 
Gendered educational attainment ratio   .047 (.029) 
Women's share of parliamentarians .153* (.075)   
Constant 84.725 (53.370) −17.300 (22.345) 
F 52.091  13.582  
df 56  58  
Adjusted R-squared .788  .528  
Gendered educational attainment ratioWomen's share of parliamentarians
Coeff.(S.E.)Coeff.(S.E.)
First-stage residual .114 (3.087) −1.373 (1.615) 
GDP/capita (2005 US$, ln1.196 (2.499) .588 (1.337) 
Protestant (non-Western) 4.192 (6.322) −13.720** (4.335) 
Catholic (Western) −6.494* (2.773) −10.189** (3.237) 
Catholic (Latin American) 13.029** (4.791) −12.954** (3.954) 
Catholic (Other) 3.795 (7.521) −11.078** (4.180) 
Islamic (Middle Eastern) −9.104 (6.268) −17.821*** (3.677) 
Islamic (non-Middle Eastern) −6.820 (7.079) −13.163** (4.136) 
Orthodox −4.214 (5.110) −15.462*** (3.892) 
Buddhist −3.495 (7.012) −17.202*** (3.903) 
Hindu −20.361* (9.461) −8.841 (4.708) 
Animist −2.327 (8.661) −13.199** (4.390) 
Nonreligious −3.050 (5.247) −13.396*** (3.794) 
Confucian −30.518** (10.332) −14.963** (4.689) 
Democracy/Autocracy .219 (.228) −.270*** (.070) 
OPEC member (1 = yes) 4.281 (4.254) −1.847 (1.491) 
Trade (% GDP) .046 (.026) .013 (.012) 
International war (1 = yes) 5.855* (2.760) −1.574 (3.235) 
Civil war (1 = yes) .293 (3.345) −1.370 (1.186) 
Fertility rate −9.831*** (1.455) .586 (.527) 
Female population (%) .152 (.643) .442 (.250) 
Religious fractionalization 9.769 (5.041) .438 (2.212) 
Communist (1 = yes) −3.682 (4.079) 13.325*** (2.053) 
Women's INGO linkages −.035 (.057) .037 (.027) 
CEDAW member (1 = yes) −.080 (1.930) 1.758 (1.020) 
Proportional representation (1 = yes)   3.901** (1.206) 
Years of female suffrage   −.001 (.027) 
Gendered educational attainment ratio   .047 (.029) 
Women's share of parliamentarians .153* (.075)   
Constant 84.725 (53.370) −17.300 (22.345) 
F 52.091  13.582  
df 56  58  
Adjusted R-squared .788  .528  

N(country-year observations) = 3,789; N(countries) = 150.

Standard errors, adjusted for first-order autocorrelation, in parentheses.

Fixed effects for years included in the analyses but not reported.

*p < .05, **p < .01, ***p < .001 (two-tailed).

All told, our two-stage models, in conjunction with an extensive battery of diagnostic tests, inspire confidence that economic development is indeed exogenous to the measures of gender equality we use in our analyses.

WORKS CITED IN APPENDIX B

WORKS CITED IN APPENDIX B
Angrist, Joshua D., and Jörn-Steffen Pischke.
2009
.
Mostly Harmless Econometrics: An Empiricist's Companion
.
Princeton, NJ
:
Princeton University Press
.
Berik, Günseli, Yana von der Meulen Rodgers, and Stephanie Seguio.
2009
. “
Feminist Economics of Inequality, Development, and Growth
.”
Feminist Economics
15
(
3
):
1
33
.
Hahn, Jinyong, Jerry Hausman, and Guido Kuersteiner.
2004
. “
Estimation with Weak Instruments: Accuracy of Higher-order Bias and MSE Approximations
.”
Econometrics Journal
7
(
1
):
272
306
.
Huntington, Samuel P.
1996
.
The Clash of Civilizations and the Remaking of World Order
.
New York
:
Simon & Schuster
.
Jorgenson, Andrew K., and Ryan Birkholz.
2010
. “
Assessing the Causes of Anthropogenic Methane Emissions in Comparative Perspective, 1990–2005
.”
Ecological Economics
69
:
2634
2643
.
Jorgenson, Andrew K., and Brett Clark.
2011
. “
Are the Economy and the Environment Decoupling? A Comparative International Study, 1960–2005
.”
American Journal of Sociology
118
(
1
):
1–44
.
Kabeer, Naila, and Luisa Natali.
2013
. “
Gender Equality and Economic Growth: Is there a Win-Win?
” IDS Working Paper No, 417.
Brighton, UK
:
Institute of Development Studies
.
Klasen, Stephan.
2002
. “
Low Schooling for Girls, Slower Growth for All? Cross-Country Evidence on the Effect of Gender Equality in Education on Economic Development
.”
World Bank Economic Review
16
(
3
):
345
373
.
Klasen, Stephan, and Francesca Lamanna.
2009
. “
The Impact of Gender Inequality in Education and Employment on Economic Growth: New Evidence for a Panel of Countries
.”
Feminist Economics
15
(
3
):
91
132
.
Stock, James H., and Motohiro Yogo.
2005
. “
Asymptotic Distributions of Instrumental Variables Statistics with Many Weak Instruments
.” Pp.
109
120
in
Identification and Inference for Econometric Models
, edited by D. W. K. Andrews and J. H. Stock.
Cambridge, UK
:
Cambridge University Press
.
Viterna, Jocelyn, Kathleen M. Fallon, and Jason Beckfield.
2008
. “
How Development Matters: A Research Note on the Relationship between Development, Democracy and Women's Political Representation
.”
International Journal of Comparative Sociology
49
(
6
):
455
477
.
Wahman, Michael, Jan Teorell, and Axel Hadenius.
2013
. “
Authoritarian Regime Types Revisited: Updated Data in Comparative Perspective
.”
Contemporary Politics
19
(
1
):
19
34
.
Welzel, Christian.
2013
.
Freedom Rising: Human Empowerment and the Quest for Emancipation
.
Cambridge, UK
:
Cambridge University Press
.
Wooldridge, Jeffrey M.
2003
.
Introductory Econometrics: A Modern Approach
, 2nd edition.
Mason, OH
:
Thompson South-Western
.
York, Richard, Eugene Rosa, and Thomas Dietz.
2003
. “
Footprints on the Earth: The Environmental Consequences of Modernity
.”
American Sociological Review
68
:
279
300
.

REFERENCES

REFERENCES
Alesina, Alberto, Arnaud Devleeschauwer, William Easterly, Sergio Kurlat, and Romain Wacziarg.
2003
. “
Fractionalization
.”
Journal of Economic Growth
8
:
155
94
.
Asian Development Bank
.
2015
.
Women in the Workforce: An Unmet Potential in Asia and the Pacific
.
Mandaluyong City, Philippines
:
Asian Development Bank
.
Barro, Robert J.
1996
. “
Democracy and Growth
.”
Journal of Economic Growth
1
:
1
27
.
Barro, Robert J.
2002
. “
Education as a Determinant of Economic Growth
.” Pp.
9
24
in
Education in the Twenty-first Century
, edited by E. P. Lazear.
Stanford, CA
:
Hoover Institution Press
.
Bashir, Abdel-Hameed M.
2004
. “
Financing Basic Education in IDB Member Countries
.” Occasional Paper No. 10.
Jeddah, Saudi Arabia
:
Islamic Development Bank
.
Beck, Thorsten, George Clarke, Alberto Groff, Philip Keefer, and Patrick Walsh.
2001
. “
New Tools in Comparative Political Economy: The Database of Political Institutions
.”
World Bank Economic Review
15
(
1
):
165
76
.
Benavot, Aaron.
1989
. “
Education, Gender, and Economic Development: A Cross-National Study
.”
Sociology of Education
62
(
1
):
14
32
.
Berkovitch, Nitza.
1999
.
From Motherhood to Citizenship: Women's Rights and International Organizations
.
Baltimore, MD
:
Johns Hopkins University Press
.
Berkovitch, Nitza, and Karen Bradley.
1999
. “
The Globalization of Women's Status: Consensus/Dissensus in the World Polity
.”
Sociological Perspectives
42
(
3
):
481
98
.
Bolzendahl, Catherine.
2009
. “
Making the Implicit Explicit: Gender Influences on Social Spending in Twelve Industrialized Democracies, 1980–99
.”
Social Politics: International Studies in Gender, State & Society
16
(
1
):
40
81
.
Bolzendahl, Catherine, and Clem Brooks.
2007
. “
Women's Political Representation and Welfare State Spending in 12 Capitalist Democracies
.”
Social Forces
85
(
4
):
1509
34
.
Bradley, Karen.
2000
. “
The Incorporation of Women into Higher Education: Paradoxical Outcomes?
Sociology of Education
73
(
1
):
1
18
.
Bradley, Karen, and Francisco O. Ramirez.
1996
. “
World Polity and Gender Parity: Women's Share of Higher Education: 1965–1985
.”
Research in Sociology of Education and Socialization
11
:
63
91
.
Charles, Maria, and Karen Bradley.
2002
. “
Equal but Separate? A Cross-National Study of Sex Segregation in Higher Education
.”
American Sociological Review
67
(
4
):
573
99
.
Charles, Maria, and Karen Bradley.
2009
. “
Indulging our Gendered Selves? Sex Segregation by Field of Study in 44 Countries
.”
American Journal of Sociology
114
(
4
):
924
76
.
Cole, Wade M.
2016
. “
Human Rights and the Individual: Cross-Cultural Variation in Human Rights Scores, 1980 to 2010
.”
Social Forces
95
(
2
):
721
52
.
Cooray, Arusha, and Niklas Potrafke.
2011
. “
Gender Inequality in Education: Political Institutions or Culture and Religion?
European Journal of Political Economy
27
:
268
80
.
Cooray, Arusha.
2012
. “
Suffrage, Democracy and Gender Equality in Education
.”
Oxford Development Studies
40
(
1
):
21
47
.
Dauda, Risikat Oladoyin S.
2013
. “
Does Female Education Promote Economic Performance? Evidence from Nigeria
.”
International Journal of Economics and Finance
5
(
1
):
201
09
.
Dollar, David, Raymond Fisman, and Roberta Gatti.
2001
. “
Are Women Really the ‘Fairer’ Sex? Corruption and Women in Government
.”
Journal of Economic Behavior & Organization
46
(
4
):
423
29
.
Dollar, David, and Roberta Gatti.
1999
.
Gender Inequality, Income and Growth: Are Good Times Good for Women?
Washington, DC
:
World Bank
.
Dorius, Shawn F., and Glenn Firebaugh.
2010
. “
Trends in Global Gender Equality
.”
Social Forces
88
(
5
):
1941
68
.
Duflo, Esther.
2005
. “
Women's Empowerment and Economic Development
.” BREAD Policy Paper 11.
Eastin, Joshua, and Aseem Prakash.
2013
. “
Economic Development and Gender Equality: Is There a Gender Kuznets Curve?
World Politics
65
(
1
):
156
86
.
Eckel, Catherine C., and Philip J. Grossman.
2008
. “
Men, Women, and Risk Aversion: Experimental Evidence
.”
Handbook of Experimental Economics Results
1
:
1061
73
.
Eisenstadt, Shmuel Noah.
1999
. “
Multiple Modernities in an Age of Globalization
.”
Canadian Journal of Sociology
24
(
2
):
283
95
.
Eisenstadt, Shmuel Noah.
2000
. “
Multiple Modernities
.”
Daedalus
129
(
1
):
1
29
.
Eisenstadt, Shmuel Noah.
2001
. “
The Civilizational Dimension of Modernity: Modernity as a Distinct Civilization
.”
International Sociology
16
(
3
):
320
40
.
Fallon, Kathleen M., Liam Swiss, and Jocelyn Viterna.
2012
. “
Resolving the Democracy Paradox: Democratization and Women's Legislative Representation in Developing Nations, 1975 to 2009
.”
American Sociological Review
77
(
3
):
380
408
.
Forsythe, Nancy, Roberto Patricio Korzeniewicz, and Valerie Durrant.
2000
. “
Gender Inequalities and Economic Growth: A Longitudinal Evaluation
.”
Economic Development and Cultural Change
48
(
3
):
573
617
.
Frank, David John, Bayliss J. Camp, and Steven A. Boutcher.
2010
. “
Worldwide Trends in the Criminal Regulation of Sex, 1945 to 2005
.”
American Sociological Review
75
(
6
):
867
93
.
Gray, Mark M., Miki Caul Kittilson, and Wayne Sandholtz.
2006
. “
Women and Globalization: A Study of 180 Countries, 1975-2000
.”
International Organization
60
:
293
333
.
Hamdan, Amani.
2005
. “
Women and Education in Saudi Arabia: Challenges and Achievements
.”
International Education Journal
6
(
1
):
42
64
.
Hausmann, Ricardo, Laura D. Tyson, and Saadia Zahidi.
2012
.
The Global Gender Gap Report 2012
.
Geneva
:
World Economic Forum
.
Hill, M. Anne, and Elizabeth King.
1995
. “
Women's Education and Economic Well-Being
.”
Feminist Economics
1
(
2
):
21
46
.
Huntington, Samuel P.
1996
.
The Clash of Civilizations and the Remaking of World Order
.
New York
:
Simon & Schuster
.
Inglehart, Ronald.
2007
. “
Mapping Global Values
.” Pp.
11
32
in
Measuring and Mapping Cultures: 25 Years of Comparative Values Surveys
, edited by T. Pettersson and Y. Esmer.
Leiden
:
Brill
.
Inglehart, Ronald, and Wayne E. Baker.
2000
. “
Modernization, Cultural Change, and the Persistence of Traditional Values
.”
American Sociological Review
65
:
19
51
.
Inglehart, Ronald, and Pippa Norris.
2003a
.
Rising Tide: Gender Equality and Cultural Change around the World
.
New York
:
Cambridge University Press
.
Inglehart, Ronald, and Pippa Norris.
2003b
. “
The True Clash of Civilizations
.”
Foreign Policy
135
(March-April):
62
70
.
Inglehart, Ronald, and Christian Welzel.
2005
.
Modernization, Cultural Change, and Democracy: The Human Development Sequence
.
Cambridge
:
Cambridge University Press
.
Inglehart, Ronald, and Christian Welzel.
2009
. “
How Development Leads to Democracy: What We Know about Modernization
.”
Foreign Affairs
88
(
2
):
33
48
.
Inkeles, Alex, and David H. Smith.
1974
.
Becoming Modern: Individual Change in Six Developing Countries
.
Cambridge, MA
:
Harvard University Press
.
Jacobs, Jerry A.
1996
. “
Gender Equality and Higher Education
.”
Annual Review of Sociology
22
:
153
85
.
Jayasuriya, Dinuk S., and Paul J. Burke.
2013
. “
Female Parliamentarians and Economic Growth: Evidence from a Large Panel
.”
Applied Economics Letters
20
:
304
07
.
Jenkins, Philip.
2006
.
The New Faces of Christianity: Believing the Bible in the Global South
.
New York
:
Oxford University Press
.
Kaufmann, Daniel.
1997
. “
Corruption: The Facts
.”
Foreign Policy
107
:
114
31
.
Keefer, Philip.
2012
.
Database of Political Institutions
.
Washington, DC
:
World Bank Group
. Retrieved March 19, 2014 (http://go.worldbank.org/2EAGGLRZ40).
Kenworthy, Lane, and Melissa Malami.
1999
. “
Gender Inequality in Political Representation: A Worldwide Comparative Analysis
.”
Social Forces
78
(
1
):
235
68
.
Kifner, John.
1980
. “
Iran TV, Shaped by Ghotbzadeh, Now Has Led to His Imprisonment
.”
New York Times
, November 9, p. 1.
Klasen, Stephan.
2002
. “
Low Schooling for Girls, Slower Growth for All? Cross-Country Evidence on the Effect of Gender Inequality in Education on Economic Development
.”
World Bank Economic Review
16
(
3
):
345
73
.
Klasen, Stephan, and Francesca Lamanna.
2009
. “
The Impact of Gender Inequality in Education and Employment on Economic Growth: New Evidence for a Panel of Countries
.”
Feminist Economics
3
:
91
132
.
Leite, Carlos A., and Jens Weidmann.
1999
.
Does Mother Nature Corrupt? Natural Resources, Corruption, and Economic Growth
. Working Paper No. 99/85.
Washington, DC
:
International Monetary Fund
.
Lesthaeghe, Ron.
2010
. “
The Unfolding Story of the Second Demographic Transition
.”
Population and Development Review
,
36
(
2
),
211
51
.
Lipset, Seymour Martin.
1959
. “
Some Social Requisites of Democracy: Economic Development and Political Legitimacy
.”
American Political Science Review
53
(
1
):
69
105
.
Lovenduski, Joni.
2012
. “
Women Should Not Have to Justify Their Political Presence on Any Other Basis Than Justice
.” London School of Economics and Political Science blogs. Retrieved September 14, 2016 (http://blogs.lse.ac.uk/politicsandpolicy/28727/).
Maoz, Zeev, and Errol A. Henderson.
2013
. “
The World Religion Dataset, 1945-2010: Logic, Estimates, and Trends
.”
International Interactions
39
:
265
91
.
March, James G., and Johan P. Olsen.
1998
. “
The Institutional Dynamics of International Political Orders
.”
International Organization
52
(
4
):
943
69
.
Marshall, Monty.
2006
.
Major Episodes of Political Violence and Conflict Regions, 1946-2006
. Retrieved June 24, 2010 (http://www.systemicpeace.org).
Marshall, Monty G., Ted Robert Gurr, and Keith Jaggers.
2013
.
Polity IV Project: Dataset Users’ Manual
.
Vienna, VA
:
Center for Systemic Peace
. Retrieved February 16, 2017 (http://www.systemicpeace.org/inscrdata.html).
Matland, Richard E.
1998
. “
Women's Representation in National Legislatures: Developed and Developing Countries
.”
Legislative Studies Quarterly
23
(
1
):
109
25
.
Melander, Erik.
2005
. “
Political Gender Equality and State Human Rights Abuse
.”
Journal of Peace Research
42
:
149
66
.
Meyer, John W., John Boli, and George M. Thomas.
1987
. “
Ontology and Rationalization in the Western Cultural Account
.” Pp.
12
37
in
Institutional Structure: Constituting State, Society, and the Individual
, edited by George M. Thomas, John W. Meyer, Francisco O. Ramirez, and John Boli.
Newbury Park, CA
:
Sage
.
Meyer, John W., John Boli, George M. Thomas, and Francisco O. Ramirez.
1997
. “
World Society and the Nation-State
.”
American Journal of Sociology
103
(
1
):
144
81
.
Milliet-Einbinder, Martine.
2000
. “
Writing Off Tax Deductibility
.”
OECD Observer
. Retrieved December 12, 2016 (http://oecdobserver.org/news/archivestory.php/aid/245/Writing_off_tax_deductibility_.html).
Møller, Jørgen, and Svend-Erik Skaaning.
2013
.
Democracy and Democratization in Comparative Perspective
.
London
:
Routledge
.
Norris, Pippa, and Ronald Inglehart.
2001
. “
Cultural Obstacles to Equal Representation
.”
Journal of Democracy
12
(
3
):
126
40
.
Norris, Pippa, and Ronald Inglehart.
2002
. “
Islamic Culture and Democracy: Testing the ‘Clash of Civilizations’ Thesis
.”
Comparative Sociology
1
(
3-4
):
235
63
.
Norton, Seth W., and Annette Tomal.
2009
. “
Religion and Female Educational Attainment
.”
Journal of Money, Credit and Banking
41
(
5
):
961
86
.
Organization for Economic Co-operation and Development
.
2012
.
Gender Equality in Education, Employment, and Entrepreneurship
.
Paris
:
OECD
.
Paxton, Pamela, Melanie M. Hughes, and Jennifer L. Green.
2006
. “
The International Women's Movement and Women's Political Representation, 1893-2003
.”
American Sociological Review
71
(
6
):
898
920
.
Potrafke, Niklas, and Heinrich W. Ursprung.
2012
. “
Globalization and Gender Equality in the Course of Development
.”
European Journal of Political Economy
28
:
399
413
.
Ramirez, Francisco O., Yasemin Soysal, and Susanne Shanahan.
1997
. “
The Changing Logic of Political Citizenship: Cross-National Acquisition of Women's Suffrage Rights, 1890 to 1990
.”
American Sociological Review
62
:
735
45
.
Ramirez, Francisco O., and Christine Min Wotipka.
2001
. “
Slowly but Surely? The Global Expansion of Women's Share of Science and Engineering Fields of Study
.”
Sociology of Education
74
:
231
51
.
Reynolds, Andrew.
1999
. “
Women in Legislatures and Executives of the World: Knocking at the Highest Glass Ceiling
.”
World Politics
51
(
4
):
547
72
.
Richards, David L., and Ronald Gelleny.
2007
. “
Women's Status and Economic Globalization
.”
International Studies Quarterly
51
:
855
76
.
Rivas, M. Fernanda.
2012
. “
An Experiment on Corruption and Gender
.”
Bulletin of Economic Research
65
(
1
):
10
42
.
Ross, Michael L.
2008
. “
Oil, Islam, and Women
.”
American Political Science Review
102
:
107
23
.
Sala-i-Martin, Xavier X.
1996
. “
Regional Cohesion: Evidence and Theories of Regional Growth and Convergence
.”
European Economic Review
40
(
6
):
1325
52
.
Schmidt, Volker H.
2010
. “
Modernity and Diversity: Reflections on the Controversy between Modernization Theory and Multiple Modernists
.”
Social Science Information
49
(
4
):
511
38
.
Seguino, Stephanie.
2000
. “
Gender Inequality and Economic Growth: A Cross-country Analysis
.”
World Development
28
(
7
):
1211
30
.
Seguino, Stephanie.
2010
. “
The Global Economic Crisis, Its Gender and Ethnic Implications, and Policy Responses
.”
Gender & Development
18
(
2
):
179
99
.
Stepan, Alfred C., and Graeme B. Robertson.
2003
. “
An ‘Arab’ More Than a ‘Muslim’ Democracy Gap
.”
Journal of Democracy
14
(
3
):
30
44
.
Spohn, Willfried.
2003
. “
Multiple Modernity, Nationalism and Religion: A Global Perspective
.”
Current Sociology
51
(
3-4
):
265
86
.
Teorell, Jan.
2010
.
Determinants of Democratization: Explaining Regime Change in the World, 1972–2006
.
Cambridge
:
Cambridge University Press
.
Teorell, Jan, Nicholas Charron, Stefan Dahlberg, Sören Holmberg, Bo Rothstein, Petrus Sundin, and Richard Svensson.
2013
.
The Quality of Government Dataset
. Quality of Government Institute, University of Gothenburg. Retrieved March 19, 2014 (http://www.qog.pol.gu.se).
Thornton, Arland, Shawn F. Dorius, and Jeffrey Swindle.
2015
. “
Developmental Idealism: The Cultural Foundations of World Development Programs
.”
Sociology of Development
1
(
2
):
277
320
.
Tipps, Dean C.
1973
. “
Modernization Theory and the Comparative Study of Societies: A Critical Perspective
.”
Comparative Studies in Society and History
15
(
2
):
199
226
.
United Nations
.
1995
.
Beijing Declaration and Platform for Action
. UN Docs. A/CONF.177/20 (1995) and A/CONF.177/20/Add.1 (1995).
United Nations Population Fund
.
2005
.
The Promise of Equality: Gender Equity, Reproductive Health and the Millennium Development Goals
.
New York
:
United Nations Development Fund
.
Viterna, Jocely, Kathleen M. Fallon, and Jason Beckfield.
2008
. “
How Development Matters: A Research Note on the Relationship between Development, Democracy and Women's Political Representation
.”
International Journal of Comparative Sociology
49
(
6
):
455
77
.
Welzel, Christian.
2013
.
Freedom Rising: Human Empowerment and the Quest for Emancipation
.
Cambridge
:
Cambridge University Press
.
Welzel, Christian, Ronald Inglehart, and Hans-Dieter Klingemann.
2003
. “
The Theory of Human Development: A Cross-Cultural Analysis
.”
European Journal of Political Science
42
:
341
79
.
World Bank
.
2013
.
World Development Indicators
.
Washington, DC
:
World Bank
.

NOTES

NOTES
1.
See e.g. USAID, “Gender Equality and Women's Empowerment,” retrieved September 14, 2016 (https://www.usaid.gov/what-we-do/gender-equality-and-womens-empowerment); World Bank, “Gender: Overview,” retrieved December 12, 2016 (http://www.worldbank.org/en/topic/gender/overview).
2.
Extensive research in feminist economics finds that economic growth does not automatically lead to greater economic equality for women (Seguino 2000, 2010). Because our paper focuses on political and educational outcomes, we do not examine the types of growth or labor-market structures that are most relevant for women's economic equality.
3.
World Bank, “Girls’ Education,” retrieved October 13, 2015 (http://www.worldbank.org/en/topic/education/brief/girls-education).
4.
Other lines of research suggest the continued importance of cultural differences for gendered educational outcomes. Studies often conclude that gender equality in education is strongest in Protestant societies (Cooray and Potrafke 2011; Norton and Tomal 2009; Potrafke and Ursprung 2012).
5.
We do not argue that cultural resistance to girls’ and women's education no longer exists, nor do we suggest that the relationship between women's education and economic growth is universally positive. Dauda (2013), for example, recently showed that male educational attainment increased economic growth in Nigeria between 1975 and 2008 whereas female educational attainment did not.
6.
One recent exception found that economies grew faster in countries with higher shares of women in parliament over a 40-year period (Jayasuriya and Burke 2013).
7.
Estimates from the World Religion Dataset are less reliable in earlier years than in later years, making comparisons over time difficult (Maoz and Henderson 2013). In any event, we contend that deep-seated religious traditions should have enduring effects, over and above any short-term fluctuations in professed adherence.
8.
We coded dates of ratification from the United Nations “Status of Ratification Interactive Dashboard,” retrieved November 22, 2016 (http://indicators.ohchr.org/).
9.
We use the ivreg2 command in Stata, specifying robust standard errors that are clustered on countries.
10.
Analyses using the Prais-Winsten estimator with panel-corrected standard errors and an AR1 correlation structure produced substantively similar results.
11.
Additional analyses revealed that the addition of fertility rates to Model 3a accounts for these effects. That is, fertility rates suppressed the otherwise negative association between each of these cultural zones and gendered educational attainment gaps. Countries in the Confucian, Hindu, and Western Catholic zones have lower fertility rates, on average, than other countries; in turn, fertility rates correlate negatively with gender equality in educational attainment. Adjusting for these confounding relationships altered the insignificant estimates reported in Model 2a.
12.
Supplemental analyses suggest that fertility is a particularly strong predictor of gender equality in education. Adding fertility rates to a baseline model that includes only per capita GDP and year fixed effects increases the percentage of explained variation considerably, from 44 to 61 percent. Analyses also show, however, that GDP per capita absorbs more cross-cultural variability in gendered educational ratios than does fertility. When we substitute fertility rates for GDP per capita in Model 2a of Table 2, seven cultural-zone indicators remain negative and statistically significant.
13.
In supplemental analyses of women's political representation we stratify countries by OECD membership and explore the role of economic development within each of those developmental strata, as recommended by Viterna, Fallon, and Beckfield (2008). These researchers found that standard explanations for variation in women's share of parliament fare better for developed than for developing nations. Although the core variables of theoretical interest in our analysis—cultural-zone indicators—generally perform as expected in analyses of subsamples defined by OECD membership, control variables operationalizing other factors of theoretical import do show variable effects, in line with Viterna, Fallon, and Beckfield's analyses.