Explaining Women’s (Low) Political Representation

The rich literature on women’s political representation internationally shows that a complex array of factors account for differing levels of women’s representation. Barriers to women’s greater political representation are typically classified as structural or socioeconomic, institutional or political, and cultural or ideological. Under the “supply and demand” model (Norris and Lovenduski Reference Norris and Lovenduski1995) used to analyze which phases of the political process are most critical for women’s representation, structural factors are seen as largely having an impact on the “supply” of women candidates (i.e., the store of qualified women willing to run) and institutional factors on the “demand” for women candidates (i.e., willingness of parties to nominate women in winnable positions). But, as Krook (Reference Krook2010) reminds us, ideological or cultural factors underlie both supply and demand. They also potentially explain the responses of voters to female candidates.

While research on institutional and socioeconomic determinants is well developed and shows that PR systems, gender quotas, strong left-wing parties, and high levels of democracy and economic development are all associated with higher female political representation, when it comes to the role of culture or ideology, there is considerable debate. Part of the problem is the greater difficulty of measuring variation in culture or gender attitudes. Most scholars have used population surveys for this purpose: for example, Inglehart and Norris (Reference Inglehart and Norris2003) and Paxton and Kunovich (Reference Paxton and Kunovich2003) drew on cross-national results from the World Values Survey to calculate support for gender equality, using a range of questions concerning gender equality in politics, employment, university education and child-rearing. Both studies concluded that gender-egalitarian values were more important than institutional and structural factors in explaining varied levels of women’s political representation (see also Ruedin Reference Ruedin2012).

That there is a link between attitudes toward women in politics and women’s representation is clear from previous research. However, this still leaves open the question of causality. Inglehart and Norris (Reference Inglehart and Norris2003, 136–39) hypothesized that gender attitudes drove women’s representation and not the reverse, but they were unable to provide conclusive evidence for this proposition. Alexander (Reference Alexander2012) addressed causality directly, and found a complex, mutually reinforcing reciprocal relationship between the proportion of women in political office and women’s belief in women’s ability to lead.

One controversial aspect of the literature concerns the influence of religion, particularly Islam. Inglehart and Norris (Reference Inglehart and Norris2003, 67–68), in their investigation of the sources of traditional, patriarchal views on social role differentiation between men and women (and hence opposition to women’s involvement in politics), famously declared that “type of religion matters for beliefs about gender equality far more than the strength of religiosity” (emphasis in original). They argued that “an Islamic religious heritage is one of the most powerful barriers to the rising tide of gender equality” (Inglehart and Norris Reference Inglehart and Norris2003, 71). Ross (Reference Ross2008) responded by finding that it is not Islam that underpins patriarchal views in Muslim societies but structural features associated with dependence on oil. Seguino (Reference Seguino2011) proposed that religiosity, not one particular religion, is associated with negative views on gender equality. Alexander and Welzel (Reference Alexander and Welzel2011), in contrast, found that among people with equally high levels of religiosity, Muslims hold more patriarchal views than non-Muslims and that context is also important: living in a majority-Muslim society increases the patriarchal views of Muslims even more than individual identification as Muslim across countries, with non-Muslims in such societies also holding more patriarchal views than non-Muslims living in non-Muslim societies. Lussier and Fish (Reference Lussier and Fish2016, 55) agreed that both self-identification as Muslim and living in a majority-Muslim society increase the likelihood an individual will oppose gender equality, but also found that gender has a stronger impact than religious identification: “Non-Muslim men have a greater likelihood of favoring gender inequality than do Muslim women, and the strength of sex as an independent predictor of attitudes is robust even when controlling for religion.”

More generally, the link between gender attitudes and voter support for women at the ballot box is not straightforward. As noted earlier, cross-national research reveals a general trend toward greater women’s representation in countries where gender egalitarianism, including support for women as political leaders, is stronger. Yet individuals make voting decisions on the basis of numerous considerations, not simply cultural, religious, or ideological values concerning women as leaders. Research in established European, North American, and some other democracies has identified the presence of widespread gender stereotypes when voters evaluate candidates, yet studies have rarely found that these constitute a significant obstacle to women at the ballot box. One recent meta-analysis of 67 candidate choice survey experiments found that, on average, women candidates experienced a slight (2%) gain in voter support compared to men (Schwarz and Coppock Reference Schwarz and Coppock2022). Scholars generally conclude that party affiliation, incumbency, and other contextual factors are more important than gender stereotypes when voters decide whether to support a particular candidate (e.g., Dolan Reference Dolan2014; Dolan and Lynch Reference Dolan and Lynch2014; Setzler and Yanus Reference Setzler and Yanus2015).

The Indonesian Case

When Indonesia democratized after 1998, it adopted a PR system and, later, a mandated quota of 30% female candidates at all three legislative levels (district, provincial, and national), with a one-in-three zipper system for candidate lists. Although recent studies have raised questions about the efficacy of PR systems in increasing women’s representation in developing countries, studies show that mandated quotas for women candidates have a positive impact (Hughes et al. Reference Hughes, Paxton, Clayton and Zetterberg2019; Krook Reference Krook2010; Rosen Reference Rosen2017). Combining the two should improve women candidates’ electoral chances (Reyes Reference Reyes and Hutchcroft2019). In adopting both of these institutions, Indonesia has introduced an electoral system designed to boost the number of women entering legislatures.

Yet the proportion of women elected to office has consistently lagged government targets, and in July 2023, Indonesia ranked 108th globally in the proportion of national legislators who are female (Inter-Parliamentary Union 2023). In the national legislature, the first democratic election of 1999 saw the proportion of women drop to 8.8%, from 11.4% in the final authoritarian-era election of 1997. Over the subsequent two decades, women’s representation tracked slowly upward: in 2019, it reached 21% of seats in the national legislature (in provincial and district legislatures, the numbers were lower, at 15% and 18%, respectively). This figure is still well below the candidate quota level of 30% and below the international average at that time of 24.3%.

The dearth of female representatives in Indonesia has been the focus of several studies (Bessell Reference Bessell, Aspinall and Mietzner2010; Hillman Reference Hillman2017, Reference Hillman2018; Perdana and Hillman Reference Perdana and Hillman2020; Prihatini Reference Prihatini2018, 2019b); several scholars have also analyzed the pathways of women who have managed to break through the barriers and win elective office, sometimes by using connections to male elites, sometimes by mobilizing support from grassroots or religious women’s networks (Aspinall, White, and Savirani Reference Aspinall, White and Savirani2021; Bayo Reference Bayo2021; Choi Reference Nankyung2018; Kabullah and Fajri Reference Kabullah and Fajri2021; Mahsun, Elizabeth, and Mufrikhah Reference Mahsun, Elizabeth and Mufrikhah2021; Prihatini Reference Prihatini2019c; Shair-Rosenfield Reference Shair-Rosenfield2012; Wardani and Subekti Reference Wardani and Subekti2021). Much of this literature focuses on supply and demand barriers associated with political recruitment, highlighting obstacles such as reluctance by political parties to endorse strong female candidates (e.g., Hillman Reference Hillman2018, 331–32; Wardani and Subekti Reference Wardani and Subekti2021) and the limited material resources and political networks accessible to female candidates compared to their male competitors (Hillman Reference Hillman2018, 328–31).

Most scholars acknowledge that patriarchal values likely undermine attempts to increase Indonesian women’s political representation (e.g., Bessell Reference Bessell, Aspinall and Mietzner2010, 223–25; Hillman Reference Hillman2018, 326–27; Prihatini Reference Prihatini2019b, 88–89), though some have suggested or implied that cultural barriers are declining in importance relative to institutional barriers (Hillman Reference Hillman2017, 42; Hillman Reference Hillman2018, 323; Shair-Rosenfield Reference Shair-Rosenfield2012, 586). To the extent that they have addressed the issue, most researchers have discounted the importance of Islam in determining the political fortunes of women. Fattore, Scotto, and Sitasari (Reference Fattore, Scotto and Sitasari2010, 272), for example, drew on a 2003 national survey to find that “Islam and hostility to gender equity in their legislatures were not inexorably linked.” Prihatini (Reference Prihatini2018, 19) compared Islamic parties with religiously pluralist parties and found that “Islam, as a party’s ideology, is not more gender inequitable” given that “Islamist parties tend to nominate more women than pluralist ones.” Overall, however, with the partial exception of Fattore, Scotto, and Sitasari (Reference Fattore, Scotto and Sitasari2010), the effect of patriarchal attitudes on women’s political representation in Indonesia has not been subject to rigorous survey-based analysis.

Yet there are strong prima facie reasons for suggesting that patriarchal political attitudes might have a strong effect. First, the public sphere contains many expressions of overt support for patriarchal arrangements. Partly this is a matter of historical legacy: under the authoritarian Suharto regime (1966–98), the state promoted a gender ideology sometimes called “state ibuism” (ibu is the Indonesian word for “mother”) that valorized the domestic role of women as wives and mothers and limited women’s public role to the pursuit of state-sanctioned ideological and development goals (Blackburn Reference Blackburn2004; Suryakusuma Reference Suryakusuma2011). Since the end of the Suharto regime, Indonesian Islam has undergone a “conservative turn” (Bruinessen Reference Bruinessen2013) characterized by many public expressions of conservative readings of gender roles and opposition to feminism (e.g., Kartika Reference Kartika2019).

Second, Indonesia’s party system provides space for voters to act out their gender biases even if they express strong partisan loyalties. Under Indonesia’s open-list PR system, parties nominate a list of candidates to compete for seats in each constituency; voters can choose to vote for a party or an individual candidate on the party’s list. Scholars have pointed out that in determining their lists, party elites give the coveted number-one position to men more often than to women (Perdana and Hillman Reference Perdana and Hillman2020; Prihatini Reference Prihatini2019a, Reference Prihatini2019c), providing evidence of demand obstacles to women’s political advancement. The open-list system also allows voters to determine which candidates within a party win seats, allowing them to express personal gender prejudice without compromising partisan loyalty. Suggestively, Dettman, Pepinsky, and Pierskalla (Reference Dettman, Pepinsky and Pierskalla2017, 117) showed that in the 2014 elections, women had a lower probability than men of winning a seat, even when they were incumbents.

Third, while the numerical representation of women in Indonesia’s national legislature has tracked upward over the past two decades—albeit slowly—a series of public opinion surveys indicate that support in Indonesia for women in political leadership positions has actually declined. Notably, International Foundation for Electoral System surveys conducted in 2003, 2005, and 2010 asked respondents whether they thought the current level of female representation in parliament was too low, about right, or too high (Fattore, Scotto, and Sitasari Reference Fattore, Scotto and Sitasari2010, 266–67; IFES 2011, 30). The proportion of respondents believing the proportion was too low dropped from 61% in 2003 to 41% in 2010. We asked the same question in 2019 and found that the figure had fallen further, to 32%. Given that there was only a 10% increase in the proportion of women in the national legislature between the early 2000s and 2019, the change in attitudes picked up in surveys is unlikely to reflect growing satisfaction with women’s increasing political representation.Footnote ¹ A regressive shift in attitudes toward women’s political representation is a more plausible explanation.

Circumstantial evidence supporting such an interpretation is provided by World Values Survey results between 2000 and 2018, which reveal that the proportion of Indonesian respondents who strongly agreed or agreed with the statement that “in general, men make better political leaders than women,” rose from 57.5% to 71.9% (Haerpfer et al. Reference Haerpfer, Inglehart, Moreno, Welzel, Kizilova, Diez-Medrano, Lagos, Norris, Ponarin and Puranen2022; Inglehart et al. Reference Inglehart, Haerpfer, Moreno, Welzel, Kizilova, Diez-Medrano, Norris, Ponarin and Puranen2014a, Reference Inglehart, Haerpfer, Moreno, Welzel, Kizilova, Diez-Medrano, Lagos, Norris, Ponarin and Puranen2014b). Remarkably, the fall in support for women’s leadership has been driven by a fall in support among women. Such results indicate that Indonesia is an exception to the general pattern identified by Alexander (Reference Alexander2012) of a reciprocal relationship between rising women’s representation and popular belief in women’s ability to govern—particularly among women. In Indonesia, there does not appear to be a strong role model effect of increased numbers of women in parliament (cf. Shair-Rosenfield Reference Shair-Rosenfield2012).Footnote ² These attitudinal trends point toward the urgency of understanding the extent and social roots of patriarchal political attitudes in Indonesia and their effects on women’s electoral success.

To sum up the expectations arising from our review of the comparative and country literature in the preceding two sections, we expect factors associated with modernization, such as education, middle-class occupations, and urban status to be associated with gender-egalitarian views, and men to hold more patriarchal attitudes than women. We have less strong expectations regarding the effects of religious, especially Islamic, beliefs, with the comparative literature generating expectations that Muslims will be more patriarchal than non-Muslims, but the Indonesia literature more equivocal on this score. We also expect to find politically patriarchal attitudes to be associated with less support for women at the ballot box and less support for measures designed to increase women’s political representation, such as gender quotas.

Measuring Patriarchal Attitudes in the Political Arena

What does “political patriarchy” mean, and how should we measure it? “Patriarchy” as a term has been in use for several centuries and as tool of feminist analysis since the 1960s. Most basically, patriarchy, in Hunnicutt’s (Reference Hunnicutt2009, 557) words, “means social arrangements that privilege males, where men as a group dominate women as a group, both structurally and ideologically.” Patriarchy is thus a multidimensional concept (Benstead Reference Benstead2021), operating both through structures and institutions and through ideology and norms (Higgins Reference Higgins2018; Hunnicutt Reference Hunnicutt2009). Following Walby (Reference Walby1989), we differentiate between patriarchy in the private and public spheres, and we further distinguish that element of the public sphere that relates directly to the exercise of political power by women. We are particularly interested in how patriarchal norms or values held by society at large (i.e., by both women and men) may impact women’s ability to run for political office.

In our understanding of the term, a patriarchal political system is one in which men dominate positions of power in the executive and legislature. Those who support such a system favor men being the primary political decision makers in society because they believe that men make better leaders and that women’s involvement in politics threatens and destabilizes the natural social order. These beliefs are underpinned by a set of values regarding the desirability of women becoming political leaders and a set of assumptions about women’s capability to play leadership roles. People who hold patriarchal views support a clear division of gender roles, such that men’s role is in the public sphere and women remain primarily in the domestic sphere, and they believe that society functions best when this division is upheld. These values inform a set of assumptions about women’s capabilities, such that women are perceived to lack the intellectual or emotional capacity to enter politics, while men are believed to have a natural ability to lead and to make decisions for the good of a community or country.

To measure the depth and spread of such patriarchal attitudes within Indonesian society, we use responses to a nationally representative survey conducted by Lembaga Survei Indonesia (Indonesian Survey Institute; LSI) during May 11–16, 2019, following legislative and presidential elections in April that year. The interviews were conducted face-to-face with 1,210 voting-age adults who were selected using multistage random sampling, proportionally distributed over Indonesia’s 34 provinces.Footnote ³ The survey asked respondents a series of questions about gender equality and women’s involvement in various types of activity in both the private and public spheres. Some of the questions are commonly asked in international surveys, while others we designed for the Indonesian context.

A series of questions focused specifically on politics, and it is these that we rely upon in making our analysis.Footnote ⁴ Table 1 displays the wording and the results for nine questions that we use to construct an index on patriarchal political attitudes. The questions capture attitudes along three dimensions: (1) the desirability of having women in political positions, (2) perceptions of women’s capability as politicians, and (3) voters’ opposition to or support for increasing women’s representation at different levels of government.Footnote ⁵

Table 1. Attitudes toward political roles of women

The results reveal significant variation. For example, a strong majority of respondents believed that, if given the opportunity to hold office, a woman would have the same capacity to do the job as a man. Yet just under half agreed with increasing the number of females in political positions. Fewer than half the respondents felt that being a leader conflicts with women’s inherent nature, yet over 60% believed that men should be the community decision makers.

We used responses to these nine questions to generate a political patriarchy index rather than creating a more general index on attitudes toward women’s engagement in politics as well as in the workforce, education, and the household. We chose this path for two reasons. First, while other analyses of the influence of culture/ideology on women’s political representation have analyzed attitudes toward gender inequality across all spheres, they do so in part because they rely on large cross-national surveys that have few questions on gender inequality. In contrast, our survey asked a large number of questions across a range of domains, allowing us to zero in on questions focused on women’s political engagement, our primary topic of interest.

Second, as Price (Reference Price2014) points out, cross-national studies using combined measures of attitudes toward women in both social and political realms reveal that there are significant differences in attitudes toward women’s participation in each realm. While support for women’s participation in higher education and the workforce has risen across the globe as societies have modernized, support for women’s political participation has not done so at the same rate (World Economic Forum 2020).Footnote ⁶ Indonesia has one of the largest gaps between support for women undertaking higher education and support for women as political leaders; according to Price’s (Reference Price2014, 354–55) calculations based on the World Values Survey, the gap is around 40% (in our survey, the gap was somewhat less extreme, depending on which question on political leadership was used). Including questions on acceptance of women in spheres such as education and employment would thus blunt the analysis and shift the focus away from our primary focus on women in politics.

We use the nine questions listed in Table 1 to generate an index that provides a thorough and accurate measure of respondents’ overall attitudes toward women in politics. The inter-item correlation test produced an alpha of 0.68, so we can be confident that the index is loading upon a single underlying dimension and that it constitutes a useful tool for measuring people’s patriarchal preferences in the political sphere.

We recode these questions in Table 1 so that higher values signify a more patriarchal response. We drop all “don’t know/no answer” responses. Several questions in our patriarchy index allow respondents to choose a “neutral” response, while other questions do not. This means our index is made up of questions that have different numerical scales. To manage these differences, we recode our nine questions so that all are measured on a 5-point scale (1–5). Each respondent then receives an average score out of 5. Figure 1 reports the distribution of respondents along the political patriarchy index.

Figure 1. Political patriarchy index.

For analytical simplicity, we categorize our respondents in four groups that indicate the consistency of their views on women and politics: egalitarian (scores between 1 and < 2.5), leaning egalitarian (2.5 and < 3), leaning patriarchal (3 and < 3.5), and patriarchal (scores > 3.5). Figure 2 shows the distribution of respondents among these four categories. Only around 37% of our sample fall into the egalitarian categories, with most Indonesians classified as leaning patriarchal or patriarchal.

Figure 2. Political patriarchy index (groups).

Who Are the Patriarchs? Mapping Patriarchal and Egalitarian Gender Attitudes

Who are the supporters of patriarchal politics in Indonesia? As noted earlier, we come to this task with expectations derived from the comparative literature. We anticipate factors associated with modernization, such as education, middle-class occupations, and urban status, to be associated with gender-egalitarian views and men to be more patriarchal than women; our expectations about the effects of religious belief are less strong. In addition, drawing on the Indonesia literature, we include a measure for residence on Java, given that there is a significant body of country research suggesting that attitudes toward gender equality may be somewhat more egalitarian in Java because of the concentration of economic development on this island, the presence of strong women’s organizations associated with the Islamic organization Nahdlatul Ulama (Mahsun, Elizabeth, and Mufrikhah Reference Mahsun, Elizabeth and Mufrikhah2021), and a history of more gender-egalitarian attitudes within Javanese society compared to some other parts of the archipelago (e.g., Geertz Reference Geertz1961).

Our findings mostly, though far from completely, line up with our expectations. Beginning with gender and religion, the cross tabulations (Table 2) provide a starting point. They show that women generally have more gender-egalitarian views than men, though not dramatically so: around 43% of women fall into the two egalitarian categories, compared to 30% of men. Religious differences are more striking: over 60% of Muslims are classified as patriarchal or leaning patriarchal, compared to 39% of non-Muslims. The effects of religion and gender are additive, with the largest gap in attitudes arising between Muslim men and non-Muslim women. We considered the effect of religiosity on these results, but such a large majority of both Muslim and non-Muslim respondents in our survey are classified as “highly religious” that we could observe no meaningful variation or correlations.Footnote ⁷

Table 2. Patriarchal attitudes by religion, gender, location, and social class (percent)

Among social class indicators, professional middle-class occupations, and living in an urban location are associated with more egalitarian views—but the differences are not large. There is, however, a remarkable difference in the views of respondents who have a university education. Over 60% of these highly educated citizens received egalitarian or leaning egalitarian scores, compared to just 29% for those with primary education only.

The cross tabulations thus suggest that two demographic characteristics are driving differences in patriarchal attitudes: religion and education. These associations are confirmed by our ordinary least squares (OLS) regression (Table 3). We include in the model four independent variables in binary forms: gender (male/female), religion (Muslim/non-Muslim), location (urban/ rural), and island of residence (Java/outside Java). Our age variable is measured in years, and for education and income, we divide respondents into categories for which “no education” and “< 1 million rupiah per month” are the respective base categories.

Table 3. Social and economic determinants of political patriarchy

Note: Standard errors in parentheses.

*** p < .001;

** p < .01;

* p < .05.

As displayed in Table 3, we find that, indeed, religion has the largest substantive effect: Muslims are significantly more likely to hold politically patriarchal views than non-Muslims, even once we control for a range of other factors such as income, education, and gender.Footnote ⁸ Among our measures of social class, only education has a significant association with the likelihood of expressing patriarchal attitudes: the effect is large and increases for each level of education. For example, compared to having no education, having a tertiary education is associated with a decrease of 0.375 on the patriarchy scale. Gender also matters. Taking all other demographic characteristics into account, men are still, on average, more patriarchal than woman, and this effect is significant and substantively large as well (around 0.17).

In the Indonesian context, the significant association between patriarchal views and educational attainment is a dramatic finding, given that others have noted the persistence of traditional attitudes about gender roles among university graduates (Utomo Reference Utomo2012) or pointed to the growth of Islamist political movements with conservative social attitudes among university-educated Muslims (e.g., Rinaldo Reference Rinaldo2013). While there are reservoirs of conservative attitudes among university graduates, we find little evidence of broad support for patriarchal attitudes among the university educated. On the contrary, it seems that university education is a generator of gender-egalitarian attitudes. Presumably, this outcome is a result of students’ exposure to progressive ideas as part of their education at the university level, though it might also result from exposure to female lecturers and other women in positions of authority on campuses and the generally egalitarian social spaces that students are exposed to there. We also note the potential endogeneity of these education effects as there may be unobservable variables that jointly affect both patriarchal views and education attainment.

Do Patriarchal Attitudes Matter? Support for the Quota

Do the patriarchal attitudes that we capture in our index predict citizens’ opposition to real-world policies that affect women’s political representation, and do they help explain why, despite a 30% candidate quota, Indonesian women get elected in such low numbers?

We look first at support for quotas. As noted earlier, Indonesia has a legislated candidate quota of 30% with a placement mandate of one woman in every three candidates and sanctions for noncompliance, thus meeting the requirements for an effective quota (Hughes et al. Reference Hughes, Paxton, Clayton and Zetterberg2019). We asked respondents two questions regarding the quota (Table 4).

Table 4. Support for quotas

The results reveal that a strong majority (77%) of Indonesians support the current 30% quota (Table 4), which had existed for 15 years at the time of the survey. This measure of support is stable, with a 2009 survey revealing 75% support (LSI 2009), and high by international standards (for figures on Latin America, see Seligson, Smith, and Zechmeister Reference Seligson, Smith and Zechmeister2012). There is even support for the quota among Indonesians holding patriarchal values (Figure 3). However, such strong support does not extend to increasing the quota to 50%. As Figure 4 illustrates, over 70% of egalitarians support such an increase, while over 70% of those holding patriarchal views oppose it.

Figure 3. Support for 30% quota.

Figure 4. Support for 50% quota.

We test these associations using logistic regressions, as the dependent variables are binary (Table 5). We examine the odds of a person supporting each quota when the index is included as continuous variable (Models 1 and 3) and at each level of the patriarchy index categories (Models 2 and 4). Model 1 in Table 5 reveals that the patriarchy index has a significant and negative effect on the likelihood of supporting the 30% quota, as does being male. In Model 2, we can observe a substantial difference between egalitarians (the base category) and those in the most patriarchal category, with this group having the largest and most significant negative effect on supporting the quota.

Table 5. Determinants of support for female candidate quotas

Notes: Base category for the Political Patriarchy Index is egalitarian; base category for education is no schooling; base category for income is < 1 million rupiah; age is measured in years. Standard errors in parentheses.

*** p < .001;

** p < .01;

* p < .05.

The results are more striking when it comes to the question of supporting an increase in the candidate quota to 50%. As Model 3 illustrates, the likelihood of a person supporting an increased quota is substantially and significantly reduced as the patriarchy index increases (–1.457). Model 4 again reveals that a person in the patriarchal category is significantly and substantially less likely to support an increased quota compared to egalitarians (–1.8850). Model 4 also shows that being a man, being Muslim, and living in an urban area all reduce the odds of supporting an increased quota. Age also has a significant and negative effect in Models 3 and 4, but the coefficients are small.

These results present a complex picture of the relationship between patriarchal attitudes and support for gender quotas in the political system. People who hold values that place them in the most patriarchal category of our index have substantially and significantly reduced odds of supporting the quota at either 30% or 50%. The proposition of increasing that quota motivates even more dramatic opposition among those with patriarchal attitudes. Importantly, even when we account for these ideological effects, we find that men and Muslims are less likely to support an increased quota. Our findings therefore align with comparative research suggesting that gender is an important determinant of support for candidate quotas (Barnes and Córdova Reference Barnes and Córdova2016; Bolzendahl and Coffé Reference Bolzendahl and Coffé2020; Gidengil Reference Gidengil1996; Keenan and McElroy Reference Keenan and McElroy2017; Smith, Warming, and Hennings Reference Smith, Warming and Hennings2017).

The finding that being Muslim reduced support for a 50% quota also highlights the impact of religion on support for gender-egalitarian measures. Most of the research on the impact of ideology on quota support focuses on the left/right political divide and shows a link between leftist ideology and support for quotas (Beauregard Reference Beauregard2018). Dubrow (Reference Dubrow2011) found a link between traditional Catholicism and rejection of gender quotas, but the study focused on elite attitudes. To our knowledge, our study is the first to examine the impact of religion on support for a quota among the general population.

Although our findings confirm the importance of gender-egalitarian attitudes as a determinant of support for the quota (Barnes and Córdova Reference Barnes and Córdova2016; Keenan and McElroy Reference Keenan and McElroy2017), it is intriguing that we still see significant support, especially for the existing 30% quota, among citizens holding relatively patriarchal views. Recent research on Brazil (Pereira and Porto Reference Pereira and Nathália2020) and Australia (Beauregard and Sheppard Reference Beauregard and Sheppard2021) demonstrates that people holding gender-inegalitarian views regarding women in politics may support a gender quota. Drawing on the concept of benevolent sexism (Glick and Fiske Reference Glick and Fiske1997), these authors show that citizens who hold stereotypical views about women’s “nature” and competencies may back a quota in the belief that women will otherwise be incapable of competing for office. Although we did not specifically test for benevolent sexism, it is plausible that at least some of the high support for the 30% quota in Indonesia is linked to such attitudes among those holding patriarchal views.

The effect of tertiary education is not significant in these models, most likely because we capture the effects of education through the patriarchy index. But it is worth noting that the coefficient is negative in our analysis of support for an increase to a 50% quota. Keenan and McElroy (Reference Keenan and McElroy2017) similarly found in Ireland that more highly educated voters were less likely to support a quota for women candidates, and they were unable to fully account for this. One possible explanation is that those with low levels of education may be most prone to benevolent sexism. More educated voters might also more strongly support gender-blind merit-based selection for office. It is thus possible that opposition to the quota being raised to 50% does not reflect a lack of support for having greater numbers of women in politics (Allen and Cutts Reference Allen and Cutts2018).

Finally, we analyzed the effect of partisanship and political ideology. It is conventional in comparative studies of attitudes toward women’s political representation to control for factors such as respondents’ positions on a left/right spectrum or their attitudes toward redistribution, with left-wing respondents typically more supportive of gender equality in politics. Indonesian parties do not align on a left/right spectrum, with the major division instead being on the political role of Islam (Fossati et al. Reference Fossati, Aspinall, Muhtadi and Warburton2020). At first sight, it would appear such affiliations could be relevant for support for the quota, with supporters of Islamic parties and of the 2019 presidential challenger, Prabowo Subianto (who was supported by a coalition containing Islamists), holding generally more patriarchal political attitudes than supporters of nationalist or pluralist parties and of incumbent Joko Widodo, respectively (see Tables 3 and 4 in the Appendix in the Supplementary Material). However, when we add support for presidential candidates and three major parties (an Islamist party, a pluralist party, and the party of Prabowo Subianto) to the regression, we found no significant effect, nor do the results change (see Tables 5 and 6 in the Appendix).

Do Patriarchal Views Matter? Voting for Women

We are especially interested in investigating the extent to which these patriarchal attitudes present a barrier to female candidates being elected to office in Indonesia. While it makes intuitive sense that voters who express skepticism or hostility to women’s political leadership will be less likely to support female candidates, there are plausible alternative explanations for the relatively low rate at which women are elected, including those derived from the supply and demand model. It may be, for example, that women lack access to the material or network resources that help candidates win elections in Indonesia, or that parties do not nominate well-resourced and capable female political leaders.

Though these explanations are obviously not mutually exclusive, our survey allows us to examine the extent to which patriarchal attitudes contribute to women candidates’ travails. We asked respondents who recalled voting for an individual candidate in the 2019 election whether they voted for a male or female candidate at (1) the district-level DPRD (Dewan Perwakilan Rakyat Daerah or Regional People’s Representative Council, i.e., the third tier of government) or (2) any of Indonesia’s three legislative levels. We asked specifically about district DPRDs because women find it especially hard to win office at this level of government.

Among those who indicated voting for an individual candidate at the DPRD level, 14.7% said they voted for a woman; at all levels, the figure was 38.5%. Respondents who indicated they did not vote for a woman were then asked to choose from a list of reasons for their decision. The responses (Table 6) point us, at least prima facie, toward supply and demand explanations: by far the most popular explanation provided by voters (56.7%) was that they “did not know” the female candidates in their area, a response that might suggest such candidates lacked the material wealth, networks, and other resources needed to promote themselves as effectively as their male counterparts. Notably, only 6% of respondents gave responses (“women are not generally suited to being leaders” and “religious reasons”) that point explicitly toward patriarchal attitudes; the remainder gave responses amenable to multiple interpretations.

Table 6. Reason for vote choice

However, when we explore how those who stated that they voted for a woman scored on our political patriarchy index, a clear pattern emerges. For example, while 23.3% of respondents in the gender-egalitarian category reported voting for a woman at the district level, only 12.5% of voters holding patriarchal views did so (Figure 5). A similar pattern is visible for voting for a woman at any of the three levels (Figure 6): approximately 50% of voters holding egalitarian views reported voting for a woman in at least one of these levels, while around 30% of patriarchal voters did so.

Figure 5. Share of respondents who voted for female candidate (DPRD).

Figure 6. Share of respondents who voted for female candidate (all legislative levels).

On the one hand, these figures do suggest that factors other than gender attitudes impede the success of women candidates. After all, only around a quarter of voters holding the most gender-egalitarian views voted for a woman at the DPRD level, and only 50% of this group voted for at least one woman at any level, falling considerably short of the 30% of candidates mandated by the quota (if voters randomly voted for candidates, and assuming 30% at each level were women, we would expect just under two-thirds of respondents would have voted for a female candidate at least at one level). By the same token, a significant minority (30%) of the most patriarchal voters reported voting for a woman candidate.

We again test these associations in several logistic regressions to examine how patriarchal attitudes might reduce the likelihood of voting for a female candidate (Table 7). Again we include the index in Models 1 and 3 as a continuous variable and in Models 2 and 4 as four different categories with egalitarian as the reference category for comparison. In Model 1, none of our variables has a significant effect on the odds of voting for a women in the DPRD—except our patriarchy index, which, as expected, reduces the likelihood (by 0.513). In Model 2, again as expected given the generally low numbers of respondents who voted for a woman at this level, the only significant effect comes from being in the most patriarchal category compared to the most egalitarian.

Table 7. Political patriarchy and voting for women candidates

Note: Standard errors in parentheses.

*** p < .001;

** p < .01;

* p < .05.

When it comes to voting for a woman at any level, Model 3 in Table 7 shows patriarchal attitudes again having a significant and negative effect. But we also find that respondents who live on Java are significantly more likely to have voted for a woman that those living off Java. Unexpectedly, those in the highest income group are less likely to support a female candidate compared to the poorest of our respondents.Footnote ⁹ The results are largely the same in Model 4—compared to the most egalitarian voters, patriarchal respondents are significantly and substantially less likely to vote for a female, as are those living off Java and those in the highest income category. Education, gender, and religion have insignificant associations with the odds of voting for a woman at any level. Finally, we again tested for political affiliations (see Tables 5 and 7 in the Appendix), but only voting for Joko Widodo is positively and significantly associated with voting for a woman at any level.

The picture that emerges is one in which patriarchal attitudes are a stronger predictor of voting for a female candidate than most of the demographic variables we examine, even when these variables do have an effect in the expected direction. When presented with a choice between female and male candidates at the ballot box, Indonesian voters are apparently directed by the prejudices they hold toward women in political life. Even when we account for other factors like religion and social class, our measure of patriarchal attitudes has a substantive and significant effect on the odds that a person will choose to vote for a woman.Footnote ¹⁰ Education also has relatively little effect, being already captured through the patriarchy index, with egalitarians most likely to come from those with higher educational backgrounds.