Regime Support and Gender Quotas in Autocracies

Abstract

Gender quotas are increasingly being adopted by autocrats in part to legitimize their rule. Yet, even in autocracies, these quotas increase women’s political representation. It thus stands to reason that public support for gender quotas in autocracies might be shaped by this trade-off between advancing women’s rights and granting the regime legitimacy. All else equal, regime opponents should be less supportive of gender quotas in autocracies, wary of legitimizing the regime. We uncover evidence of this proposition in an analysis of region-wide Arab Barometer surveys and a survey experiment in Algeria. We also find that evaluations of this trade-off are conditioned by other demographics, with women, gender egalitarians, and Islamists remaining more consistent in their support for/opposition to gender quotas regardless of regime gains. Overall, our findings suggest that gender quotas in autocracies are viewed through a political lens, creating a potential backlash toward women’s empowerment.

INTRODUCTION

Electoral quotas by gender have been adopted by over 130 countries (Hughes et al. 2019), becoming the main determinant of female legislative representation around the world (Jones 2009; Krook 2009; Tripp and Kang 2008). Gender quotas have been found to not only increase women’s descriptive representation, but also substantively shape policies in favor of the rights of women (Clayton 2021; Clayton and Zetterberg 2018). Quotas have also been shown to have a positive impact on women’s symbolic representation, inspiring women to run for and win public offices (Bhavnani 2009; Nanes 2015) and improving the public’s perception of women in politics (Ben Shitrit 2016).

Gender quotas have been adopted not just by democracies, but also by autocracies (Hughes, Krook, and Paxton 2015). While occasionally motivated by gender egalitarianism and women’s rights activism, autocrats’ calculations for adopting gender quotas are often more cynical: an attempt to garner international and domestic legitimacy (Bjarnegård and Zetterberg 2022; Bush 2011; Donno, Fox, and Kaasik 2022; Tripp 2019). Indeed, recent scholarship finds that autocrats often succeed in boosting their international reputations by adopting gender quotas (Bush and Zetterberg 2020).

Despite the motivations for their adoption, gender quotas even outside of democracies still produce a variety of benefits for women—descriptive, substantive, and symbolic (Dahlerup 2013). Therefore, unlike in democracies, gender quotas in autocracies generate a unique trade-off: advancing gender equality, but also legitimizing dictatorship. Does this trade-off in turn shape public support for gender quotas in autocracies? Are regime opponents less supportive of gender quotas, hesitant to grant the regime legitimacy? Or, do the potential benefits of empowering women outweigh these legitimacy costs?

In this article, we explore this trade-off by examining public opinion in the Arab world, a region where the trade-off is expected to be particularly pronounced. With the region historically having some of the worst records on women’s rights, quotas have had a dramatic effect on boosting female political representation (Kang 2009). At the same time, such gender-based reforms grant Arab dictators a progressive reputation both domestically and internationally, creating a stark contrast to the conservative Islamists widely viewed as their alternative (Tripp 2019). Accordingly, both the gains to women and to the regime are thought to be especially strong in the Arab world.

We pursue a two-pronged approach to evaluate Arab public opinion toward the autocratic gender quota trade-off. We first present a region-wide analysis using representative surveys from the Arab Barometer. We show that regime support/opposition is a major predictor of attitudes toward gender quotas in almost every country surveyed. Regime supporters are significantly more supportive of quotas than opponents, even when controlling for demographic and attitudinal variables that might otherwise confound this analysis. At the same time, we also find that the importance of regime support/opposition varies by subgroups, with the trade-off seemingly less salient for women and gender-egalitarian respondents, who remain consistently supportive of quotas regardless of their attitudes toward the regime. These groups thus appear to evaluate the trade-off differently, more highly valuing the gains to women.

Having established the broad patterns, we then move to experimentally test the mechanism: that regime supporters are more supportive of gender quotas because the quotas grant the regime legitimacy. We field an original survey in Algeria during the 2019–20 Hirak protests—a context of mass frustration with the regime where we would expect concerns over regime legitimacy to be highly salient, and thus for our mechanism to be especially clear. Embedded in the survey was an experiment that primed respondents to think of who gained from the gender quotas under President Abdelaziz Bouteflika (r. 1999–2019): women, the regime, or both, compelling respondents to evaluate the trade-off.

The experiment uncovers two major findings. First, reminding Algerians that the regime gained legitimacy from quotas produced a divergent reaction from regime supporters and opponents, with supporters becoming significantly more supportive of quotas. Regime opponents, by contrast, were affected by the trade-off, becoming significantly more supportive of quotas when primed that women gained, but not when primed that the regime also gained. For regime opponents, the gains to women thus did not outweigh the gains to the regime. Second, the experiment also reveals support for our subgroup hypotheses. Women, gender egalitarians, and Islamists were less moved by the trade-off, remaining consistently supportive/opposed to quotas regardless of who gained.

These results hold several important implications for the study of women’s rights in Arab autocracies and beyond. First, we theorize and empirically validate a potential source of backlash to gender quotas in dictatorships. Our findings suggest that people view quotas through a political lens, conscious of their dictator’s attempts to “genderwash” their image. Populations that might otherwise support gender quotas are wary of doing so when they might legitimize a dictator. Popular support for such gender-based reforms may thus be low, especially among regime opponents. In this way, our research adds to the literature on how quotas may cause a backlash against women’s empowerment (e.g., Batista Pereira and Porto 2020; Berry, Bouka, and Kamuru 2020; Clayton 2015; Meier 2008; Zetterberg 2009).

In turn, this dampened public support for quotas in autocracies may undermine their durability and effectiveness. Without widespread public buy-in into quotas, they may prove brittle, less likely to withstand a regime transition. In Algeria, for instance, once Bouteflika was toppled, the gender quotas were watered down, cutting women’s representation from 26% to 8% (Marwane 2021). If quotas become tainted by a particular dictator, they may not outlast his rule. More generally, quotas are less effective in producing policy reforms when they lack public support (Clayton 2021, 245).

At the same time, our results suggest that dictators might succeed in coopting certain segments of the population through gender quotas. In our survey, women and gender egalitarian respondents tended to support these reforms regardless of the potential regime gains. Prioritizing women’s empowerment, these groups may fear that if the regime falls, the quotas might fall as well. In the Arab world, for instance, secular feminists often feared the loss of women’s rights if Islamists were to come to power (Tripp 2019). Our survey results confirm that these segments of society might on average be more susceptible to cooptation by progressive autocrats. For regimes, these segments may also be particularly important for staying in power. Chenoweth and Marks (2022) argue that having more women on the front lines makes revolutions significantly more likely to succeed. For dictators, then, coopting women through gender quotas may have significant benefits for helping them stay in power and weather mass uprisings.

THE GENDER QUOTA TRADE-OFF

Existing literature suggests that when dictators adopt gender-based reforms, they often do so for reasons unrelated to empowering women. ¹ Gender quotas in particular have been a popular choice for dictators as they provide numerous benefits yet carry minimal risk. Scholars have shown that often, regimes simply create additional seats for women without hurting male incumbents (Bjarnegård and Zetterberg 2016; Goetz 2002; Tripp 2022). Moreover, quotas strengthen the dictator’s international and domestic legitimacy.

In the wake of the global push toward women’s rights such as the 1995 Beijing World Conference on Women, women’s rights promotion has become an international norm (Towns 2010). To signal norm compliance, authoritarian regimes have adopted gender-based reforms to gain international legitimacy and foreign support. Scholars find that autocrats who receive Western aid are more likely to adopt quotas (Bush 2011; Donno, Fox, and Kaasik 2022; Edgell 2017; Welbourne 2010) and that the adoption is associated with a boost to their international reputations (Bush and Zetterberg 2020). Previous research also finds that quota adoption helps autocrats fend off international pressure to democratize, aiding authoritarian survival (Abou-Zeid 2006; Bjarnegård and Zetterberg 2022).

In addition to increasing international prestige, gender reforms strengthen the dictator’s domestic prospects. First of all, gender quotas help dictators to coopt women and women’s rights activists, increasing their domestic support (Valdini 2019) and the perceived legitimacy of their institutions (Kao et al. 2023). Moreover, reforms like gender quotas are often unfairly designed to strengthen the dictator’s grip on power (e.g., Adams 2007; Donno and Kreft 2019; Tripp and Kang 2008). For example, Bjarnegård and Zetterberg (2016) show how quotas benefit the ruling party by facilitating the selection of loyalist women to the reserved seats. Finally, female legislators in autocracies tend to be more loyal than male members to their parties, further solidifying the ruling party (Clayton and Zetterberg 2021).

In the Arab world in particular, adopting gender reforms appears to reward dictators with a progressive reputation, creating a contrast with the conservative Islamists who are often their primary opponents. Tripp (2019) documents how dictators in Algeria, Morocco, and Tunisia exploited this contrast by strategically promoting women’s rights, winning support at home and abroad. In Tunisia, the regime “demanded women’s unconditional support and enlistment in the anti-Islamist struggle” in return for protection of women’s rights (Yacoubi 2016, 258). In Morocco, female legislators elected through quotas have attributed their success “thanks to the young King” (Sater 2007, 729). Likewise, Arab dictators across the region have adopted quotas to gain financial and moral support from the West (Abou-Zeid 2006; David and Nanes 2011; Sater 2007). Even Egypt’s brutal military dictator, Abdelfattah al-Sisi, has won praise from Western audiences for his adoption of gender quotas in 2020, with one prominent American journalist writing that Sisi is a “friendly” autocrat, under whose rule “[w]omen’s rights have advanced, too, with women now mandated to hold 25 percent of the seats in parliament” (Ignatius 2021).

Despite the strategic motives for adoption, gender quotas in authoritarian regimes still often lead to measurable gains in women’s empowerment beyond descriptive representation (Bauer 2012). ² For instance, quota-elected female representatives are more likely to be responsive to female constituents’ demands for services (Benstead 2016). Gender quotas can also lead to increased female representation in more prestigious committees in parliaments (Shalaby and Elimam 2020) and generate a significant symbolic effect changing entrenched attitudes toward women and increasing the acceptability of women in politics (Ben Shitrit 2016). Moreover, quotas lead to women being more respected in the community by gaining greater autonomy in family decision-making, eventually leading to increased female political engagement (Burnet 2011). Accordingly, gender quotas in autocracies produce a unique trade-off: empowering women, but at the same time, helping to legitimize dictatorship.

How does this trade-off affect the public’s attitudes toward gender quotas? There are at least two reasons to worry about a backlash. First, even in democracies, quotas sometimes produce a negative impact on women’s empowerment (e.g., Batista Pereira and Porto 2020; Brulé 2020; Htun, Lacalle, and Micozzi 2013). Second, if imposed from above by a dictator, there is more reason to suspect a backlash against policies that may not reflect the will of the public. If people perceive the quota as stemming from some ulterior motive, such as an attempt to genderwash, an otherwise positive reform may become tainted in their eyes.

Interviews suggest that Arabs are well aware of this trade-off. Tunisian parliamentarian Hela Omrane, reflecting on Zine El Abidine Ben Ali’s adoption of gender quotas in 2009, noted that “it was only a PR exercise for the regime” (quoted in Gouvy 2020). “The promotion of women’s rights was instrumentalized to legitimize the system [and] stifle dissent,” noted a Tunisian activist (quoted in Gray 2012, 290). Similarly, Egyptian activist Noralla (2022) argues that Sisi’s “‘feminist-washing’ propaganda” is simply designed “to cover up the regime’s widespread human rights abuses.” A Jordanian MP likewise told Bush and Jamal (2015, 38) that “the regime just takes the quota as a form of make-up to put on the face of the regime […] It’s just a façade because the international community cares about it.”

In our own research, interviewees were cognizant not only of how the regime gained legitimacy, but even how they strategically used the quotas for their own benefit. Opposition politicians interviewed in Algeria, for instance, noted that the quotas uniquely helped the regime, as it alone enjoyed the resources to recruit qualified female candidates. ³ An Algerian women’s rights activist observed that women elected through quotas would “go on TV and say whatever they were told to say by the [regime party’s] leadership…They were pretty much scapegoats.” ⁴ Other Algerian activists told Lorch and Bunk (2016, 14) that women elected through quotas for regime parties tended to focus on defending the party line. The abundance of qualitative evidence thus suggests that at least some citizens are aware of their dictator’s genderwashing schemes. If so, we would expect regime support/opposition to be a major predictor of support for gender quotas, with regime supporters eager to grant the regime legitimacy, and opponents wary of doing so.

H1: All else equal, regime supporters are more supportive of gender quotas than regime opponents, because quotas grant the regime legitimacy.

Bush and Jamal (2015) provide initial support for this hypothesis, finding in Jordan that regime opponents became less supportive of women’s representation when primed that U.S.-funded organizations and domestic religious leaders supported gender quotas. However, the authors acknowledge that they cannot conclusively determine whether support decreased because they opposed the legitimacy it would grant the regime or because they were simply more opposed to those particular endorsers. ⁵

In addition to explicitly priming regime legitimacy, it is also worth explicitly priming the gains to women, examining directly the trade-off produced through gender quotas in autocracies. We expect that while regime opponents will become more supportive of quotas when primed only about the gains to women, they will not when also primed about the gains to the regime. If our expectations are validated, they would demonstrate the strong influence that the trade-off plays in shaping public opinion.

Mitigating Factors

At the same time, we also expect evaluations of the gender quota trade-off to be mitigated or conditioned by a number of other factors. For one, prior research identifies a visible gender gap: women, compared to men, are more likely to support various gender equality policies (Elder and Green 2012; Morgan and Buice 2013) such as gender quotas (Barnes and Córdova 2016; Beauregard 2018; Gidengil 1996; Keenan and McElroy 2016). One possible explanation is that policies like gender quotas by nature are more favorable toward women themselves (Deckman and McTeague 2015; Meltzer and Richard 1981). Scholars have also argued that there are inherent differences and diverging socialization experiences between the sexes, generating different political preferences for women and men (Gilligan 1982). Another explanation is men’s discontent against the advancement of women; men may see women as competition and threats to their traditional status. Morgan and Buice (2013) argue that status discontent is a key factor that leads to backlash against female advancement by men who feel threatened by women’s progress. They find that men, especially those with relatively lower status, are less likely to be supportive of women in politics. Building off of this literature on the gender gap, we hypothesize that women, compared with men, will be more consistently supportive of gender quotas, even when it may grant the regime legitimacy.

H2: Women are more consistently supportive of quotas, regardless of regime gains.

Similarly, we expect individual-level gender egalitarian attitudes to influence evaluations of the authoritarian gender quota trade-off. Studies have shown that gender inegalitarianism is associated with bias against female politicians, including those elected through quotas (Blackman and Jackson 2019). Gender egalitarian values are also linked to support for gender equality policies such as equal pay (Cassese, Barnes, and Branton 2015) and electoral gender quotas (Barnes and Córdova 2016). Thus, we expect respondents with more gender egalitarian attitudes to consistently support quotas and consider the trade-off less relevant.

H3: Gender egalitarian respondents are more consistently supportive of quotas, regardless of regime gains.

Across the Arab world, scholars have shown that Islamist movements and parties are the most vocal critics of gender quotas (e.g., Tripp 2019). Islamists, such as the founder of the Muslim Brotherhood, Hassan al-Banna, have historically prioritized women’s family roles above others. Al-Banna ([1947] 2006, 147) advocated for male and female students to be segregated, for boys’ curricula to be made distinct from girls’, for women to be instructed in proper dress, and be encouraged to marry and procreate. Ben Shitrit (2016) extensively documents how three Islamist parties today—Hamas in Palestine, the Muslim Brotherhood in Egypt, and the Islamic Movement in Israel—have been the most critical opponents of legislated gender quotas, which they treat as “foreign.” The Palestinian Islamist movement, Hamas, publicly opposed the quota before its adoption in 2005. One leader cited a Hadith from Sahih al-Bukhari (4425): “a people that places their affairs at the hand of a woman will never prosper” (Lahlouh 2010). Similarly, the Islamic Constitutional Movement, the political wing of the Kuwaiti Muslim Brotherhood, opposed gender quotas because they would be “against Islamic principles” and not necessarily produce benefits for women. ⁶ In Jordan, the Islamic Action Front likewise condemned the quotas, arguing they were unconstitutional and would not empower women (al-Batayneh, quoted in David and Nanes 2011). In Algeria, even 5 years after quota adoption, several Islamist parties refused to print their female candidates’ faces on campaign posters. Given this ideological opposition, we expect Islamists’ attitudes to be consistently opposed to quotas, even if they support the regime.

H4: Islamists are more consistently opposed to quotas, regardless of regime gains.

We test these hypotheses in two ways. We first establish region-wide correlations using survey data from the Arab Barometer. We show that throughout the region, regime support/opposition powerfully shapes attitudes toward quotas, and that this varies in line with our secondary hypotheses. Second, we move to test the mechanism using an original survey experiment in Algeria. The experiment shows how the gender quota trade-off has a causal effect shaping support for quotas and that this effect is less salient for our theorized subgroups.

ARAB BAROMETER

We first explore survey data from the Arab Barometer, which has conducted nationally representative face-to-face surveys across the region. In Wave 5, the Arab Barometer featured a new question examining support for gender quotas: “Some people think that in order to achieve fairer representation, a certain percentage of elected positions should be set aside for women. To what extent do you agree with this statement?” Respondents could answer on a 4-point scale from strongly agree or strongly disagree (see Supplementary Figure A.1).

Wave 5 of the Arab Barometer was conducted in 12 countries in 2018–19. We examine whether support for quotas varies by support for the regime, both pooling across all countries and showing each individually. Eight of the 12 countries had gender quotas in place at the time of the survey and in all eight, the regime in power was the one that had introduced the current version of those quotas (see Supplementary Table A.1). Our theory should apply most readily to these eight cases, where populations should associate the current quota with the regime. However, even in the four countries without quotas at the time (Egypt, Kuwait, Lebanon, and Yemen), were they to introduce one, the current regime would reap the benefits, as Egypt’s subsequent adoption of quotas in 2020 showed. Thus, calculations of regime support/opposition should still be salient here. Moreover, the countries without quotas also help us rule out an alternative explanation that regime supporters might just be more in favor of all regime policies. Instead, since quotas were not a regime policy, respondents may be thinking about who would gain from such a policy. Our main results below accordingly pool all countries (Figure 1), but we show in Supplementary Figure A.2 that results hold in both the countries with and without quotas.

Figure 1. Predicted Probabilities for Support for Quota

Note: Figures created from Supplementary Table A.2.

Our theory might also apply less well to the three flawed democracies (Lebanon, Iraq, and Tunisia), compared with autocracies. However, each of these countries around the time of the survey (2018–19) saw mass protests calling for the “fall of the regime” (isqat an-nizam), suggesting that populations still viewed politics in a regime-opposition lens. We, therefore, continue to include these countries in the analysis, though we show in Supplementary Figure A.3 that results are still significant, yet slightly weaker in the three flawed democracies, a topic we revisit in the conclusion.

To capture support for the regime, we create a government performance index based on four questions that measure citizens’ evaluations of the government in four issue areas: creating employment opportunities, narrowing the gap between rich and poor, providing security, and keeping prices down. The composite forms our primary measure of support for the regime. For robustness checks, we also examine two other independent variables: general satisfaction with the government, and agreement that “Citizens must support the government’s decisions, even if they disagree with them.” Both of these alternative independent variables produce similar results. ⁷

We examine the relationship between regime support and support for gender quotas while controlling for numerous demographic and attitudinal variables. Following the literature on the link between social structural indicators and female political representation (e.g., Norris 1987), we control for age, gender, education, unemployment, marriage, and having children. We also control for several attitudinal variables that might correlate with regime support and/or support for quotas, such as generalized trust, interest in politics, gender egalitarianism, and support for Islamism. ⁸ In the pooled analysis, we also control for Polity score, regime type (monarchy vs. republic), the presence of quotas in each country at the time of the survey, and country fixed effects.

Figure 1 presents the results pooled across the region (left) and then by individual country (right). Region-wide, there is a significant, positive correlation between evaluations of government performance and support for quotas. In line with H1, regime supporters are significantly more likely to support quotas than regime opponents. The right-hand side of the plot shows the predicted support for each country in the analysis. We see the same significant result in 10 of 12 countries, with only Iraq and Palestine showing no relationship. The correlations are particularly strong in Algeria and Kuwait, where regime support moves respondents from roughly opposing to supporting the quotas (from 2 to 3, respectively, on the response scale). These results from the Arab Barometer thus provide strong support for H1: regime support/opposition is a major determinant of support for quotas.

Having established the overall association, we now turn to our secondary hypotheses (H2–H4), exploring whether regime support/opposition matters less for women, gender egalitarians, and Islamists. Figure 2 plots each interaction.

Figure 2. Predicted Probabilities for Support for Quota: Subgroup Analysis

Note: Figures created from models 1–3 in Supplementary Table A.3.

We first examine how gender shapes support for quotas (H2). As seen in the plot on the left, women (gray) are on average more supportive of gender quotas than men (black), as expected. Notably, women’s support for quotas only modestly changes by regime support. On the other hand, the slope of the effect for men is significantly greater: their support for quotas is considerably more moved by their attitudes toward the regime ( $ p<0.05 $ ). In short, while women are more supportive of quotas across the board, men tend to be more shaped by the regime trade-off.

We see a similar story when interacting regime support with gender egalitarianism (H3), using an index based on questions about women’s role in society. ⁹ As seen in the middle plot, gender egalitarian respondents (top line) support quotas across the board: their line is flat, unshaped by their attitudes toward the regime. However, among the most gender inegalitarian respondents (bottom line), regime support matters to a much greater extent, considerably shaping their support for the quotas. In line with H3, gender inegalitarian respondents are more likely to be influenced by the regime trade-off ( $ p<0.001 $ ).

Finally, the plot on the right plots the Islamism interaction (H4). Here, we use a question about whether respondents prefer a religious or nonreligious party. Although not significant, respondents who prefer a religious party (gray) are leaning less supportive of gender quotas. They are also slightly, but not significantly, less moved by regime support/opposition, more consistently opposing quotas across the board.

According to the results by country (Supplementary Tables A.5), the interaction between Islamism and regime support is significant in our hypothesized direction only in Algeria, where Islamists consistently oppose quotas, while non-Islamists are shaped by their degree of support for the regime. The interaction is not statistically significant elsewhere. One explanation could be that many of the pro-regime Islamist parties elsewhere, such as (at the time) Morocco’s PJD and Tunisia’s Ennahda, had come to embrace gender quotas, strategically using them to win seats in parliament. Ben Shitrit (2016), for instance, notes that although Islamists initially viewed quotas as “foreign,” some later justified their fielding of women as an “Islamic” decision. Those elite cues might be shaping their followers’ views to become more supportive of gender quotas. In short, this recent variation in Islamist party behavior across the region may be undercutting support for this particular hypothesis.

Overall, our analysis from the Arab Barometer uncovers significant support for a gender quota trade-off. Across almost every country in the region, regime support/opposition strongly correlates with support for gender quotas, in line with the trade-off quotas pose in autocracies. Moreover, we find that the trade-off matters less among women and gender egalitarian respondents, who tend to support quotas across the board.

Why does regime support/opposition shape attitudes toward gender quotas? This correlation is somewhat puzzling, as we might expect the opposition in authoritarian contexts to be more supportive of reforms, and particularly of progressive ones. We theorize that respondents are cognizant that the regime will gain from gender quotas, leading regime supporters to support quotas, and opponents to oppose them. To test if that mechanism might indeed be driving the results, we now turn to a survey experiment conducted in Algeria.

ALGERIA SURVEY EXPERIMENT

Algeria is a useful venue for testing the mechanism for several reasons. First, in the Arab Barometer analysis, Algeria was the country where the relationship between regime support and gender quotas was strongest (see Figure 1). Where the relationship is strongest, the mechanism should be clearest (Seawright and Gerring 2008).

Qualitative, contextual knowledge helps us understand why the relationship might have been strongest in Algeria. Women’s issues have historically been a salient part of the country’s public discourse. Algerian women were active participants in the War of Independence. While the postcolonial regime initially placed women’s rights issues on the back burner (Lazreg 1990), women’s activism reemerged in the 1980s in response to regressive laws like the 1984 Family Code. During the Algerian Civil War, activists organized large demonstrations demanding an end to violence against women, who were often the targets (Moghadam 2001; Turshen 2002).

More recently, Algeria has been among the Arab countries that have adopted gender quotas and where those quotas have considerably advanced women’s representation. In 2011, former President Abdelaziz Bouteflika introduced a gender quota for legislative elections mandating women comprise approximately 30% of the candidates on parliamentary lists. ¹⁰ Subsequently, Algeria achieved one of the highest rates of female legislative representation in the Arab world, between 25% and 31% (second only to Tunisia). Previously, the rate had only been 6%–8%. Thus, the gains to women’s representation, as a result of the quota adoption, have been particularly salient in Algeria.

In addition, bringing more women into Algeria’s political sphere had substantive and symbolic effects. Most notably, in 2015, the parliament amended the family code to criminalize domestic violence and sexual harassment. Female MPs also played a key role in the 2016 constitutional reform which included an article promoting gender equality in the labor market (Article 36). While these legislative and constitutional amendments were part of Bouteflika’s broader genderwashing strategy, the new female MPs helped push them through and detail the specifics. Moreover, male politicians were increasingly working together with female MPs and important parliamentary committees became more likely to include women. The increased presence of female politicians as a result of quotas likely also transformed attitudes and made people more accepting of women in politics. While the gender quotas did not elevate women into the highest echelons of power (Hamadouche 2016), they did produce a wide variety of benefits for Algerian women.

On the other hand, a second reason why the relationship may have been strongest in Algeria is that the gains to the regime have also been highly salient. The regime adopted the gender quota, among other reforms, in the wake of the 2011 Arab Spring protests in an effort to demonstrate Bouteflika’s willingness to make political reforms. While the quota increased women’s descriptive representation, the new female MPs tended to come from the ruling coalition, and, as mentioned above, often served as regime mouthpieces.

Moreover, the adoption of the gender quota brought Bouteflika considerable praise, increasing the regime’s legitimacy. Domestically, Nadia Aït Zaï, the founder of the women’s rights group Centre d’Information et de Documentation sur les Droits de l’Enfant et de la Femme (CIDDEF), praised Bouteflika for his “courage” in implementing gender quotas. ¹¹ “This is a victory” for Algerian women, she said. ¹² This praise was no surprise since Aït Zaï acknowledged they had worked with the president to make this happen. ¹³ Nouria Hafsi, the secretary-general of the Union Nationale des Femmes Algériennes (UNFA), endorsed Bouteflika for reelection in 2014, saying “we are convinced that Bouteflika will do everything to realize the rights of women, as he has always done.” ¹⁴

Internationally, Reuters in 2012 hailed Bouteflika as the Arab world’s “new trailblazer for women in politics,” noting he was praised by both United Nations (UN) Secretary-General Ban Ki-moon and U.S. Secretary of State Hillary Clinton for adopting the quota (Ahmed 2012). UN Women’s executive director Michelle Bachelet likewise commended him, calling the quotas “a welcome step in Algeria’s progress toward democratic reform and gender equality” (UN Women 2012; UNDP 2013). “The selection of seven females as ministers, along with 145 women entering parliament, is something that has never happened before in the Arab world,” said a spokeswoman for UN Women. “What Algeria has reached so far is very impressive.” ¹⁵

The salience of regime gains was further heightened by the particular timing of our survey. Conducted in February 2020, our survey captured the tail end of the 2019–20 Hirak protests, a mass uprising that toppled Bouteflika and then continued on against the remnants of his regime. In this context of mass frustration with the regime, we would expect regime opponents to be especially wary of granting the regime legitimacy, and supporters especially eager to.

With both the gains to the regime and to women salient, Algeria in 2020 is thus a uniquely powerful case for examining public opinion toward the trade-off. Our primes, discussed below, are likely to tap into real and strongly held beliefs. Algerians’ support or opposition for gender quotas in our survey is thus likely not an idle opinion, but a highly salient position with visible implications.

A final reason for studying Algeria is to better understand why the gender quotas were watered down after Bouteflika was toppled. Recognizing that Bouteflika’s regime had lost popularity, his successor, Abdelmadjid Tebboune, attempted to distance his autocracy from Bouteflika’s, inaugurating a “New Republic.” Within this effort, Tebboune announced that the “ignorant” era of gender quotas had ended (Marwane 2021) and repealed them to break with the “bad practices of the past” (Algérie Presse Service 2020) and better reflect the “will of the people” (Algérie Presse Service 2021). The 2021 electoral law thus stopped enforcing the candidate quotas, and even opened up party lists to allow voters to select particular candidates. By doing so, women’s representation in the new parliament fell back down to just 8%, on par with the level prior to the 2012 quota adoption. Notably, this repeal elicited minimal public outcry, even from women’s rights activists. Our hypotheses provide one potential explanation for why: that Algerians who might otherwise have valued the gains to women understood that quotas had helped legitimize Bouteflika, and accordingly were less committed to keeping these tainted quotas in place.

Survey Methodology

Although Algeria is a useful case to test the mechanism, it is also a difficult environment for survey research, due to both state repression and animosity toward foreign researchers. For instance, the local partner of the Arab Barometer and Afrobarometer was placed under house arrest for over a year for conducting a survey prior to the Hirak protests. Even a French MP, who met with protesters in October 2019, was arrested and deported (Latrous 2019). These concerns made in-person survey research infeasible.

Instead, we conducted our survey online, recruiting Algerians through Facebook advertisements. Facebook ads have become an increasingly common survey recruitment device, in both the United States and the developing world (e.g., Cassese et al. 2013; Guiler 2020; Samuels and Zucco 2014). Following this scholarship, we purchased advertisements on Facebook shown to all adult Algerians. ¹⁶ Clicking on the advertisement took users out of Facebook and into Qualtrics, where they first agreed to a consent form before viewing the survey. ¹⁷ Between February 5 and 21, 2020, 1,119 Algerians clicked on the advertisement and completed the survey experiment (for more details on survey methodology, see Appendix B.1 of the Supplementary Material). For all analyses, we subset to the 911 survey respondents who correctly answered an attention check question, but results are similar when using the full sample (see Appendix B.7 of the Supplementary Material).

Naturally, given the survey methodology, our sample is not nationally representative. Only 45% of Algerians (19 million) are active monthly users of Facebook. These 45% skew younger, more male, and likely also more urban, wealthy, and educated. We followed Zhang et al. (2018) in implementing age and gender quota sampling during recruitment, bringing our sample more in line with the population on observable demographics (see Appendix B.1 of the Supplementary Material for more details on the survey sample). Still, there are likely other unobservable differences between individuals who are on Facebook and those who are not, so we cannot and do not claim that our sample is nationally representative. However, what is most important for our purposes is that demographics are balanced across treatment groups (see Supplementary Figure B.4), permitting a valid comparison for our survey experiment.

Survey Experiment

Embedded in the survey was an experiment priming respondents about the various effects of gender quotas. ¹⁸ Respondents were randomly sorted into either a control group or one of three treatment groups (Table 1). All respondents, including those in the control group, were told that Algeria adopted gender quotas starting in the 2012 elections requiring about 30% of party lists be women. The treatment groups then received additional text priming them to think about how these quotas led to gains for women (Treatment 1), led to gains for the regime (Treatment 2), or led to gains for both, compelling them to evaluate the trade-off (Treatment 3). The text of each prime was derived from the discussion above.

Table 1. Algeria Survey Experiment ( $ N=911 $ )

The experimental design thus helps us determine how the public views the trade-off. In both the women and trade-off primes, respondents learn about how quotas benefited women. If respondents do not mind the regime also gaining legitimacy, then they should evaluate the trade-off prime similarly to the women prime: in both cases producing a significant increase relative to the control. If, however, as we hypothesize, respondents do worry about regime gains, then they should be turned off by those gains, showing a significant increase in support for quotas in the women treatment but not also in the trade-off treatment. The regime prime, meanwhile, provides a clean test of our mechanism. We expect that reminding respondents that the regime gained legitimacy should produce a divergent reaction among regime supporters and opponents.

Notably, the design also helps us rule out alternative explanations for the overall Arab Barometer results. All groups, including the control, are told that the regime has implemented this policy, thus holding constant any endorsement effect by the regime. The primes instead single out how the regime has gained. Likewise, all groups, including the control, might be primed to think about parliament, controlling, therefore, for their thoughts about parliament and its strength or weakness.

To ensure respondents paid attention to the experiment, we followed up with an attention check: “Approximately what percent of party lists are currently reserved for women?” with answer options of 0%, 30%, or 50%. For all analyses, we subset to the 911 respondents (81%) who correctly answered 30%. ¹⁹

After these primes, respondents were asked for their level of support for gender quotas, worded the same as in the Arab Barometer. ²⁰ Respondents could answer on a 5-point scale from strongly disagree to strongly agree. Figure 3 presents the results. Overall, only about 28% of respondents agreed or strongly agreed with the gender quota, 36% were neutral, and 36% disagreed or strongly disagreed. In line with our expectations, overall support for the quotas was low. Experimentally, we examine how support varied by the primes listed above.

Figure 3. Support for Gender Quotas

Experimental Results

Figure 4 presents the base results from our experiment. ²¹ First, when primed just about the gains to women (Treatment 1), Algerians became significantly more supportive of (and/or less opposed to) gender quotas than in the control group (2.99 vs. 2.74, $ p=0.023 $ ). Substantively, this effect is quite large: if we dichotomize the scale, about 24% of respondents in the control group supported gender quotas, compared to 32% in the women treatment, an 8 percentage-point increase. Given that few respondents supported quotas in the control group, the women treatment increased support by 33% from baseline. A short, informational prime highlighting the benefits to women thus considerably increased support for quotas.

Figure 4. Support for Quotas by Treatment Group ( $ N=911 $ )

Note: Figure created from model 2 in Table 2.

Supplementary analyses (Appendix B.5 of the Supplementary Material) shed light on why the women prime increased support for quotas. Post-experiment, we asked whether respondents believed “gender quotas advance women’s rights.” Respondents who received the women prime were significantly more likely to agree. Moreover, mediation analyses show that the only reason why the women prime increased support for quotas was by increasing this belief that quotas advance women’s rights (Supplementary Figure B.5).

However, neither of the other two treatments showed a significant effect on quota support relative to the control. When primed about the legitimacy quotas may grant to the regime (Treatment 2), respondents were no more supportive than in the control (2.81 vs. 2.74, $ p=0.49 $ ). As we show in the subgroup analyses below, this null effect masks important variation, with regime opponents becoming slightly less supportive of quotas, and regime supporters becoming significantly more supportive.

Finally, when presented with the trade-off (Treatment 3), respondents were likewise no more supportive than those in the control (2.84 vs. 2.74, $ p=0.34 $ ). The combination of a significant effect for Treatment 1, but not also for Treatment 3, shows that while learning about the gains to women increased support for quotas, learning about the gains to women and the regime did not. This suggests that while Algerians may value the gains to women, on average the cost of legitimizing the regime appears to outweigh these gains, depressing support in the trade-off. ²² The results thus provide strong evidence for the existence of a trade-off, showing that Algerians on average evaluate the trade-off negatively, more concerned about (not) granting the regime legitimacy than about the gains to women.

Table 2 shows that all results hold when controlling for covariates. ²³ Model 1 first presents the relationship without controls. In model 2, we add demographic variables: gender, age, education, income, urban, unemployment, student, and marriage. We also control for whether they are Amazigh, and for two knowledge questions: whether they had heard of Algeria’s gender quotas, and whether they knew when they were first implemented (asked prior to the primes). Model 3 then adds several attitudinal variables, including respondents’ evaluation of the economy; their level of support for democracy and for the Hirak protests; their level of agreement with the statement, “In general, social and economic problems would improve if there were more women in office” (as a proxy for gender egalitarianism); and finally, whether they self-identify as Islamist/Salafist. Despite these controls, the results remain: Algerians became more supportive of quotas when primed about the gains to women, but not when also primed about the gains to the regime. While Algerians, therefore, value the gains to women, these gains appear to be outweighed by concerns over granting the regime legitimacy, at least in this context of a mass uprising against the regime.

Table 2. Support for Gender Quotas by Treatment Group (OLS)

Note: $ {}^{*}p<0.1 $ ; $ {}^{**}p<0.05 $ ; $ {}^{***}p<0.01 $ .

Many of the covariates have their expected effects. Women and more gender egalitarian respondents were significantly more supportive of quotas, whereas Islamists were less supportive. Respondents who supported democracy were likewise more supportive of quotas. Curiously, respondents who had heard of the quotas prior to our survey were less supportive of them, perhaps more attuned to how they benefited the Bouteflika regime. We delve further into these considerations next in the subgroup analyses.

Subgroup Analysis

While Algerians overall tend to evaluate the trade-off negatively, there are also important subgroup differences. First, as expected (H1), quotas are evaluated differently by regime supporters and opponents. Figure 5 splits the sample by those who support the ongoing mass protests (the Hirak) against the regime ( $ N=451 $ , “Regime Opponents”) and those who do not ( $ N=386 $ , “Regime Supporters”). Regime opponents (left) mimic the general population: significantly moved by the gains to women (Women vs. Control, $ p=0.017 $ ), but not when also primed about the gains to the regime (Trade-Off vs. Control, $ p=0.198 $ ). Protesting the regime, they are highly concerned about granting it legitimacy. ²⁴

Figure 5. Treatment Effects by Regime Support/Opposition

Note: Figures created from models 1 and 2 in Table 3.

The regime supporters, on the other hand, want to grant the regime legitimacy. They are accordingly moved by the regime treatment: becoming significantly more supportive of quotas when primed about how it will bring the regime legitimacy (Regime vs. Control, $ p=0.030 $ ). ²⁵ Regime supporters appear less concerned about empowering women, and are not moved by the women treatment (Women vs. Control, $ p=0.2 $ ). They likewise are not moved by the trade-off (Trade-Off vs. Control, $ p=0.46 $ ), which to them perhaps muddies the gains to the regime by also discussing the gains to women. ²⁶ In short, while regime opponents face a major trade-off that outweighs their desire for women’s empowerment, regime supporters are instead eager to grant the regime legitimacy.

Table 3 presents each of the above results. Models 1 and 2 show the experiment when subsetting to regime opponents and regime supporters, respectively. Model 3 shows that results hold when instead interacting each treatment with regime opposition: priming regime gains increases support for quotas among regime supporters ( $ p=0.044 $ ), but this reverses among regime opponents ( $ p=0.046 $ ). Finally, most intuitively, model 4 subsets just to respondents who received the regime prime. When cognizant of the regime gains, regime supporters were significantly more supportive of quotas than opponents ( $ p=0.023 $ ). These results help explain why in the Arab Barometer regime supporters support quotas more than opponents.

Table 3. Treatment Effects by Regime Support/Opposition (OLS)

Note: $ {}^{*}p<0.1 $ ; $ {}^{**}p<0.05 $ ; $ {}^{***}p<0.01 $ .

Secondary Hypotheses

There are also important subgroup differences by our hypothesized demographic variables. First, Figure 6 splits the sample by gender (H2). Men ( $ N=417 $ ) mimic the general population, becoming more supportive of quotas when primed about the gains to women ( $ p=0.022 $ ), but no longer more supportive when presented with the trade-off. Women ( $ N=420 $ ), meanwhile, show high support for quotas across the board, regardless of treatment condition, in line with H2. More committed to quotas, women’s views are not shaped by the primes or the trade-off.

Figure 6. Treatment Effects by Gender

Note: Figures created from models 1 and 2 in Table 4.

Likewise, Algerians who were more gender egalitarian were more committed to quotas regardless of the regime benefiting. Figure 7 plots those who agree that “in general, social and economic problems would improve if there were more women in office” ( $ N=188 $ , left), compared to those who disagree, are neutral, or don’t know ( $ N=679 $ , right). In line with H3, gender egalitarian respondents show high support across the board, unmoved by treatment group. Gender inegalitarian respondents, meanwhile, mimic the general population, moved by the gains to women ( $ p=0.007 $ ) by not so much that they outweigh the trade-off. ²⁷

Figure 7. Treatment Effects by Gender Egalitarianism

Note: Figures created from models 3 and 4 in Table 4.

However, we acknowledge that the sample size for gender egalitarian respondents is quite small (188), raising the possibility of false negatives. In particular, the regime treatment is close to significance ( $ p=0.17 $ ), and might have been significant with a larger sample. One interpretation is that gender egalitarian respondents likely held a very favorable view of the women’s rights group CIDDEF, and were accordingly moved by CIDDEF’s praise for the regime in this treatment group.

Finally, there are also differences by political ideology (H4). Figure 8 shows that respondents who self-describe as Islamists or Salafists ( $ N=173 $ , left) tend to oppose gender quotas across the board (average of 2.4 on the 5-point scale), unaffected by any of the treatments. Meanwhile, respondents who do not identify as Islamists ( $ N=694 $ , right) mimic the general trend, significantly more supportive of quotas when told women would benefit ( $ p=0.003 $ ), but no longer so when told that the regime would, too. In other words, in line with H4, concerns over regime legitimacy appear to shape non-Islamists more than Islamists, who tend to be more consistently opposed to quotas.

Figure 8. Treatment Effects by Ideology

*Note: Figures created from models 5 and 6 in Table 4.

For the Islamists as well, we acknowledge that the small sample size might have increased the likelihood of a false negative. The regime treatment ( $ p=0.24 $ in our sample) might have been significant in a larger sample. One interpretation is that the Islamists in our survey, who tend to support various Islamist parties coopted by the regime, might have been relatively more pro-regime and accordingly swayed by the regime treatment.

Table 4 shows that each of these subgroup results is statistically significant in the presence of demographic and attitudinal covariates. While the overall sample thus showed evidence of the trade-off, this supplementary analysis shows that the salience of this trade-off varies by our theorized subgroups.

Table 4. Support for Gender Quotas among Various Subsets (OLS)

Note: $ {}^{*}p<0.1 $ ; $ {}^{**}p<0.05 $ ; $ {}^{***}p<0.01 $ .

In short, the survey experiment in Algeria reveals strong support for the theory: gender quotas present a salient trade-off in autocracies between empowering women and legitimizing the regime. This trade-off leads regime supporters to embrace quotas and regime opponents to be wary of them, even when otherwise valuing the gains to women. However, the salience of this trade-off also varies by a number of mitigating factors. Groups that are generally supportive of quotas (women, gender egalitarians) or that are generally opposed to quotas (Islamists) tend to be less concerned about the trade-off, being more ideologically committed for or against quotas. The trade-off thus appears to matter most among subgroups whose attitudes toward quotas are more indeterminate, and thus potentially movable by considerations about who is to gain.

While the survey experiment thus provides support for each of our hypotheses, it is also important to acknowledge two limitations. First, the results are attitudinal, and future research could examine whether such primes also shape behaviors, such as signing a petition in favor of quotas. Second, the women prime was not quite significant relative to the trade-off prime; we would encourage future research to leverage a larger sample size to tease this out further.

CONCLUSION

Our findings carry several important implications. First, we uncover the first causal evidence of a trade-off in public evaluations of gender quotas in autocracies. While quotas empower women, they also grant the regime legitimacy, producing a trade-off that shapes public opinion in a number of theorized ways. Adding to the literature on backlash effects of gender quotas (Berry, Bouka, and Kamuru 2020; Clayton 2015), we theorize and provide initial empirical support for this additional source of backlash in autocracies.

Second, our study implies that this trade-off might undermine the long-term durability of gender quotas, and potentially, women’s rights activism writ large. If quotas are widely viewed as a vestige of authoritarianism, tainted by the dictator’s schemes, then they may enjoy lukewarm support even after the dictator is toppled. Such tainted quotas may therefore be unlikely to stand the test of time, as seen in Algeria where Bouteflika’s quotas were watered down after he fell. Moreover, genderwashing may not only taint the reforms but also the activists who worked with the dictator. For instance, Tunisian women’s rights activists had become tainted by their affiliation with the previous regime of Ben Ali, making it more difficult for them to advance women’s rights after the revolution (Tripp 2019). As one activist noted, “the dictatorship was pro-women…[so] the hatred against the dictatorship is expressed through action against women” (quoted in Alami 2013). If future scholarship shows that these trends generalize, it would suggest that authoritarian genderwashing might provide short-term gains but long-term costs to women’s empowerment.

Third, for actors interested in increasing public support for gender quotas, our results demonstrate the power of even short, informational messages (solely) about the gains to women. More generally, depoliticizing gender reforms—disassociating them from a particular dictator or regime—might be crucial for securing widespread, long-term buy-in into such reforms.

Fourth, while our results focus on gender reforms, they likely speak more broadly to other liberalizing reforms undertaken by authoritarian regimes such as competitive elections or human rights councils. While these reforms are potentially more risky to regimes (Donno and Kreft 2019), they similarly earn them praise and help them survive (Blaydes 2010; Gandhi and Przeworski 2007; Magaloni 2006). If our results are any guide, such reforms may produce similar trade-offs in public opinion, with regime opponents skeptical of such reforms even when they support them in substance, wary of the legitimacy they may grant to the regime.

Our theory likely travels beyond the Arab world and beyond dictatorships. On the one hand, the trade-off we identify should be most salient in the Arab world, where the gains to women through quotas have been largest; in regimes with Islamists in the opposition, where the threat to those gains are clearest; and in dictatorships, where the costs of legitimizing the regime are highest. At the same time, it is possible that similar, if weaker, trade-offs might be present elsewhere, such as in democracies with strong polarization or populism. There as well, citizens might associate a gender quota with the particular party or political class that adopted it, and not want to give that group credit or legitimacy.

Democratic Tunisia (2011–21), for instance, recently saw a similar backlash dynamic as in autocratic Algeria. The democratically elected governments had strengthened Tunisia’s gender quotas, earning domestic and international praise (Yerkes and McKeown 2018). Yet, as Tunisians grew disillusioned with those governments, they also appeared to grow disillusioned with their gender quotas, as if they had become tainted by that “dark decade.” After staging his coup in 2021, President Kais Saied abandoned those quotas with minimal public outcry, and women’s representation in the 2023 parliament fell to just 15.6%. Thus, even in democracies, quotas may similarly become tainted and thus brittle. Indeed, in the Arab Barometer surveys, regime support still exhibited a significant, albeit weaker, effect in the flawed democracies than autocracies.

In Latin America, Barnes and Córdova (2016) show that perceptions of government quality similarly correlate with support for gender quotas. Our reanalysis of the Latin American Public Opinion Project survey data ²⁸ show that while the trade-off is stronger in autocracies, regime support still correlates with support for gender quotas even in the democracies, and likewise varies according to our secondary hypotheses regarding women, gender egalitarians, and political ideology. These results suggest that our findings might travel to other regions and regime types.

Finally, the trade-off we identify might also find cognates in established European democracies. For instance, when far-right parties in France and Italy fielded Marine Le Pen and Giorgia Meloni, liberals who might otherwise have praised women’s empowerment were more hesitant, wary of legitimizing the far-right (Chira 2017; Snipes and Mudde 2020). In sum, while these trade-offs should be especially salient in Arab dictatorships, we encourage future studies to explore similar logics more globally.