In recent decades, the percentage of women in cabinets has climbed around the world. This upward trajectory applies not only to democracies but also to autocracies, where approximately 68% of the world's population live (Alizada et al. Reference Alizada2021). While still far from reaching gender parity, women's share of cabinet posts among authoritarian countries grew from 0.84% in 1970 to 10.6% in 2018.Footnote 1 The trend of appointing more women to cabinets in autocracies is particularly strong in Africa. Figure 1 presents the 25th, 50th, and 75th percentiles of women's share of cabinet posts in Africa's autocracies from 1966 to 2016 using data from WhoGov (Nyrup and Bramwell Reference Nyrup and Bramwell2020). Although the proportion of women ministers has grown, with women constituting approximately 20% of ministers in 2016, patterns of growth within African autocracies, shown in light grey in Figure 1, have varied widely. What explains variation in the inclusion of women in authoritarian cabinets?
Percentage of Women Cabinet Ministers in African Autocracies, 1966–2016
The cabinet is a crucial site of power-sharing among political elites in authoritarian contexts. Much of the scholarship on gender, ministers and cabinets, however, focuses on the composition of cabinets in democracies (e.g. Annesley et al. Reference Annesley, Beckwith and Franceschet2019; Bauer and Darkwah Reference Bauer and Darkwah2022; Davis Reference Davis1997; Escobar-Lemmon and Taylor-Robinson Reference Escobar-Lemmon and Taylor-Robinson2016; Studlar and Moncrief Reference Studlar and Moncrief1997). Recent studies advance our understanding of women's representation in legislatures and the adoption of women's rights laws in autocracies (Bush and Zetterberg Reference Bush and Zetterberg2021; Donno et al. Reference Donno, Fox and Kaasik2022; Donno and Kreft Reference Donno and Kreft2019; Tripp Reference Tripp2019; Valdini Reference Valdini2019), but whether and how this scholarship applies to the gendered selection of ministers is unclear. The incorporation of women in authoritarian cabinets is not well understood, even though scholarly attention to power-sharing in the cabinet, particularly in autocracies in Africa, has grown in recent years (Arriola Reference Arriola2009; Francois et al. Reference Francois, Rainer and Trebbi2015; Kroeger Reference Kroeger2020; Meng Reference Meng2020; Raleigh and Wigmore-Shepherd Reference Raleigh and Wigmore-Shepherd2022; Ricart-Huguet Reference Ricart-Huguet2021; Woldense Reference Woldense2018). Less considered is the autocrat's logic of including women at the apex of power, the cabinet.
Our article argues that who comes to hold positions in authoritarian cabinets is gendered. Historically, autocrats have favoured the inclusion of men in their cabinets. We seek to explain departures from the status quo – that is, increases in the percentage of women ministers in autocracies over time. We propose that international norms of gender equity, which have become increasingly intertwined with understandings of democracy, create material and symbolic incentives for electoral autocrats to appoint more women to their cabinets. We identify two situations in which we expect electoral autocrats to strategically appoint more women to the cabinet. First, we expect increasing dependence on aid from democratic donors to increase women's share of cabinet posts in electoral autocracies as leaders seek to signal their commitment to democracy and gender-equality norms. Second, we expect electoral autocrats moving away from democracy (i.e. autocratizing) to appoint more women ministers in an effort to offset damage to their international and domestic reputation.
To test these hypotheses, we analyse variations in women's share of cabinet seats in 38 African countries under authoritarian rule between 1973 and 2013. We find that increases in aid from democratic donors are associated with increases in women's share of cabinet seats in electoral autocracies in more recent years. Nevertheless, the effects are modest with a 250% increase in aid from democracies being associated with a 1.8% increase in women's share of cabinet seats during the later years in our sample. As expected by our theory, this effect is not present when including both electoral and closed autocracies in the sample or during earlier time periods when international norms of gender equality were weaker. Conversely, we find no evidence that autocrats increase the share of women in the cabinet to offset the reputational costs of their autocratization.
Our article makes several contributions. First, existing research suggests that autocrats use gender quotas and women's representation strategically in the legislature (Bjarnegård and Zetterberg Reference Bjarnegård and Zetterberg2016; Bush and Gao Reference Bush and Gao2017; Donno and Kreft Reference Donno and Kreft2019; Edgell Reference Edgell2017; Tripp Reference Tripp2019; Valdini Reference Valdini2019). We show that, in certain circumstances, electoral autocrats signal commitments to democracy and gender-equality norms by appointing more women to their cabinets. Second, while scholars have highlighted the exclusion of women in the cabinets of military autocracies (Barnes and O'Brien Reference Barnes and O'Brien2018; Bauer and Okpotor Reference Bauer and Okpotor2013; Bauer and Tremblay Reference Bauer and Tremblay2011), we highlight the importance of incentives in increasing women's representation in electoral authoritarian cabinets. Finally, scholars of authoritarian cabinets in Africa have considered the strategic co-opting of members of ethnic and regional groups (Francois et al. Reference Francois, Rainer and Trebbi2015; Raleigh and Wigmore-Shepherd Reference Raleigh and Wigmore-Shepherd2022) and opposition members (Arriola et al. Reference Arriola, DeVaro and Meng2021). We add to that literature by showing that electoral autocrats, under certain circumstances, strategically increase the percentage of women in the cabinet.
Women in cabinets: insights from global and regional studies
A rich cross-national literature spanning regime types offers several explanations for the inclusion of women in cabinets. Existing explanations that may be relevant to understanding women's cabinet representation in African autocracies are patronage politics, gender egalitarianism and international norms.
Leaders build coalitions to support their time in office, and this process affects the composition of cabinets in a gendered way. When leaders form a multiparty coalition government, they encounter demands by coalition partners to appoint party members to the cabinet; this process tends to favour the selection of men (Davis Reference Davis1997; Escobar-Lemmon and Taylor-Robinson Reference Escobar-Lemmon and Taylor-Robinson2005; Reynolds Reference Reynolds1999; Studlar and Moncrief Reference Studlar and Moncrief1997; Whitford et al. Reference Whitford, Wilkins and Ball2007). In contexts where ethnicity is politically salient, leaders reach out to ethnic patrons to construct multi-ethnic coalitions. Political appointments, of which ministerial positions are some of the most important, are often used to recruit and retain the support of ethnic patrons. In many African contexts, however, women are not considered ethnic patrons, in part because of colonial-era policies and practices that marginalized women from the formal economy and the state (Fallon Reference Fallon2008). Thus, similar to dynamics in multiparty governments, women are often crowded out of the cabinet when there are many ethnic patrons to be co-opted (Arriola and Johnson Reference Arriola and Johnson2014).
Gender egalitarianism in politics may explain why women hold a relatively high share of cabinet seats in some contexts and not others. The most common way the literature has considered gender equality in the political sphere is through women's representation in the national legislature. A higher percentage of women in legislative office suggests wider acceptance for women's presence in politics. Further, the pool of ministerial candidates typically includes those with legislative experience. In parliamentary contexts where ministers are selected from the members of parliament, the number of women in the national legislature is a crucial supply-side factor in enabling more women to join the cabinet (Whitford et al. Reference Whitford, Wilkins and Ball2007). This is also important in presidential systems where ministers sometimes have previous experience in the legislature. In global and regional statistical analyses, the percentage of women in the legislature correlates with the percentage of women in cabinets (Arriola and Johnson Reference Arriola and Johnson2014; Barnes and O'Brien Reference Barnes and O'Brien2018; Bego Reference Bego2014; Krook and O'Brien Reference Krook and O'Brien2012; Reynolds Reference Reynolds1999; Stockemer Reference Stockemer2018; Stockemer and Sundström Reference Stockemer and Sundström2019).
Previous studies posit that a country's exposure to international norms of gender equality affects the women's share of cabinet appointments (Bauer and Okpotor Reference Bauer and Okpotor2013; Bauer and Tremblay Reference Bauer and Tremblay2011; Russell and DeLancey Reference Russell and DeLancey2002). Norms of equality may diffuse in multiple ways, such as through a longer commitment to international women's rights law and by emulating the practices of neighbouring countries and international organizations (Jacob et al. Reference Jacob, Scherpereel and Adams2014). Previous studies, however, arrive at mixed results regarding the ratification of a key women's rights treaty, the Convention on the Elimination of All Forms of Discrimination against Women (CEDAW) (Arriola and Johnson Reference Arriola and Johnson2014; Krook and O'Brien Reference Krook and O'Brien2012; Stockemer and Sundström Reference Stockemer and Sundström2019).
While the existing literature on women in cabinets provides many valuable insights, scholars of cabinets have not examined whether autocrats respond strategically to international norms of gender equality by increasing women's representation in cabinets. Scholars have, however, examined the strategic inclusion of women in other autocratic institutions. In the next section, we build on that literature as well as that on authoritarian regimes and democracy promotion to theorize about the strategic incentives electoral autocrats have to include more women ministers in their cabinets.
International influences on the inclusion of women in authoritarian cabinets
We theorize that autocrats are influenced by the global context in which norms about democracy and women's rights evolve and spread. As Figure 1 shows, women have only been regularly included in the cabinets of African autocracies since the late 1980s. This coincides with important changes in the international environment that altered the landscape of authoritarian regimes. The end of the Cold War produced what Seva Gunitsky (Reference Gunitsky2017) calls a ‘hegemonic shock’. This shock, produced by the collapse of the Soviet Union, not only altered the international balance of power, it provided new incentives for domestic political reform in autocracies around the world. In a world without the Soviet Union, democratic aid donors in the West began emphasizing democratization and human rights as preconditions for continued aid flows (Bearce and Tirone Reference Bearce and Tirone2010; Bermeo Reference Bermeo2011; Dunning Reference Dunning2004). This international pressure, along with domestic movements for reform, led many autocratic regimes to adopt multiparty elections. These transitions to multiparty politics led to democratization in some instances, but more often they helped entrench electoral authoritarian regimes (Bratton and van de Walle Reference Bratton and van de Walle1997; Diamond Reference Diamond1997; Levitsky and Way Reference Levitsky and Way2010; Ottaway Reference Ottaway2003; Schedler Reference Schedler2002). Whether the outcome was democratization or electoral authoritarianism, however, democratic aid donors rewarded transitions away from closed autocracy (Kim and Kroeger Reference Kim and Kroeger2017).
As many closed autocracies transitioned to electoral autocracies, understandings of democracy were also changing. In the post-Cold War world, wealthy democracies and the international bodies they led increasingly ‘bundled’ gender-equality norms together with liberal and modern democratic principles (Bush Reference Bush2011; Donno et al. Reference Donno, Fox and Kaasik2022; Towns Reference Towns2010). This bundling of gender-equality norms with democracy created material and reputational incentives for autocrats to go beyond simply legalizing multiparty competition. For autocrats to present at least a minimal appearance of democracy, they must also improve women's representation within government institutions. Amina Mama (Reference Mama1995: 38) recognized this early among African autocracies, stating that ‘governments have found it expedient to exploit the gender question so as to receive economic aid in an international climate that has become increasingly sympathetic towards women's demands for greater equality’. Thus, many autocrats strategically implemented reforms to appear more democratic in response to shifting international priorities at the end of the Cold War. As gender equality has become increasingly intertwined with understandings of democracy, autocracies seeking to appear more democratic have also had incentives to improve women's representation in government institutions.
Existing scholarship has emphasized these international incentives for increased women's representation in a variety of autocratic institutions, but not in cabinets. For instance, state-run women's organizations and first ladies' projects have been conceptualized as strategic responses by African autocrats to international norms of gender equality (Adams Reference Adams2007; Ibrahim Reference Ibrahim2004; Kah Reference Kah2014; Mba Reference Mba and Parpart1989; Okeke Reference Okeke1998). Other scholarship suggests that international norms play a role in women's representation in legislatures. It has been argued that authoritarian leaders adopt legislative gender quotas to signal commitment to international norms of gender equality and maintain flows of foreign aid (Bush Reference Bush2011; Edgell Reference Edgell2017). Such quotas are known to be an effective way of increasing women's representation in legislatures (Tripp and Kang Reference Tripp and Kang2008). Similarly, Melody Valdini (Reference Valdini2019) argues that greater women's representation in legislatures provides hybrid regimes with a mechanism to signal their commitment to democracy and boost their domestic and international reputation. Indeed, increasing the presence of women in legislatures does appear to boost the image of authoritarian regimes among citizens in democracies. In survey experiments conducted in Sweden and the US, Sarah Bush and Pär Zetterberg (Reference Bush and Zetterberg2021) find that citizens are more willing to support giving foreign aid to an electoral autocracy when women's representation in the legislature is high.
We build on this literature to theorize about the strategic use of women's representation in authoritarian cabinets. Cabinets are particularly important for understanding the extent to which autocrats respond strategically to changing international norms on gender equality. Cabinets are a visible political space that both democracies and international organizations monitor. When leaders create or reshuffle their cabinets, it garners national and international attention. Cabinets draw scrutiny because they come with substantial authority over state institutions. Depending on the portfolio, cabinet ministers command a large budget, oversee thousands of state officials, and in many contexts are responsible for drafting the majority of bills that eventually become law. Countries and international bodies keep track of who holds ministerial posts in other countries, making it relatively easy to see whether women are present (or absent). The Bureau of Democracy, Human Rights, and Labor in the US Department of State, for example, issues yearly Country Reports on Human Rights Practices that are widely used by government officials. In 2016, the reports mentioned the number of women in the cabinet for 14 out of 23 African autocracies (US Department of State 2016). Further, international organizations use women's presence in the cabinet as part of their world rankings of countries on gender equality. For instance, the more women in the cabinet, the higher countries are rated on the World Economic Forum's Global Gender Gap Index. Thus, appointing more women to the cabinet provides autocrats with a mechanism to signal publicly their commitment to gender equality and democracy to domestic and international audiences.
However, the mere presence of this signalling mechanism does not mean that all autocrats will find it useful. For electoral autocracies, appointing more women to the cabinet can help create a more credible façade of democracy. This façade can be helpful in bringing material and reputational benefits to the regime. And while electoral autocracies have become the modal authoritarian regime in the post-Cold War, they are not universal.Footnote 2 Closed autocracies in the post-Cold War have publicly demonstrated that they care little about constructing a veneer of democracy. In such contexts, increasing the number of women in the cabinet is unlikely to send a credible signal that the regime is democratizing or produce a shift in international attitudes towards the regime. As a result, we expect only electoral authoritarian regimes to increase women's representation in the cabinet strategically to boost international support for the regime.
In the sections below, we propose two hypotheses concerning the strategic use of women's appointments to the cabinet by electoral authoritarian leaders. The first suggests that electoral autocrats respond to material incentives to appear liberal and modern to donor countries that consider gender equity to be a feature of democracies. The second expects electoral autocrats who are backsliding away from democracy to compose more gender-inclusive cabinets to improve their reputation.
Autocracies, international signalling and foreign aid
In recent decades, wealthy democracies that provide foreign aid such as the United States have increasingly linked the ideals of women's equal political participation with that of democracy. This linking of ideals can be seen in the programming of the democracy establishment, which includes democracy-promoting NGOs funded by Western governments. It has become more common for Western governments and democracy-promoting NGOs, for instance, to implement programmes to enhance women's access to political office (Bush Reference Bush2011). Aid packages may specifically carve out funding for programmes to improve women's political representation. For example, Jordan received $25 million in aid from the Millennium Challenge Corporation Threshold Program, which included $16.5 million to increase women's participation in politics (David and Nanes Reference David and Nanes2011).
We hypothesize that foreign aid from democratic donors increases the appointment of women to cabinets in electoral autocracies in more recent years. Electoral autocracies have demonstrated that they care about projecting a façade of democracy to international and domestic audiences. As international norms of gender equality have strengthened and become intertwined with understanding of democracy, increasing women's representation in cabinets offers electoral autocrats an additional mechanism to signal their commitment to democracy and gender-equality norms. We argue that electoral autocrats deploy these signals to help maintain relationships with democratic donors. Closed autocracies, on the other hand, have not shown an interest in signalling their commitment to democratic norms. And while closed autocracies also receive foreign aid from democratic donors, such aid is often linked to the strategic political or economic interests of donors. Democratic aid donors are less likely to emphasize democratic norms and good governance when allocating aid where strategic interests are at play (Alesina and Dollar Reference Alesina and Dollar2000; Dunning Reference Dunning2004; Hoeffler and Outram Reference Hoeffler and Outram2011). Therefore, closed autocracies are both unlikely to use the appointment of women ministers as a signal of their commitments to democratic norms and such signals are also unlikely to be demanded by democratic donors.
Hypothesis 1: Increases in foreign aid from democracies have increased women's representation in electoral authoritarian cabinets in more recent years.
Autocratization and reputation
As discussed above, autocrats have incentives to construct a façade of democracy in the post-Cold War period. Transitioning from closed autocracy to electoral autocracy is perhaps the most important part of constructing that façade. However, there is more to establishing the appearance of democracy than simply holding multiparty elections. It requires presenting at least the appearance of a wide variety of political and civil rights. As such, transitions to electoral autocracy often involve liberalizing reforms such as expanding freedoms of association and the press.
Electoral autocrats must work carefully to manage political and civil rights to prevent challenges to their rule. As Anna Lührmann and Staffan Lindberg (Reference Lührmann and Lindberg2019: 1098) state, ‘Electoral autocrats secure their competitive advantage through subtler tactics such as censoring and harassing the media, restricting civil society and political parties and undermining the autonomy of election management bodies.’ So long as the undermining of civil and political rights remains subtle, autocrats can avoid severe reputational costs. Yet, the subtle management of rights cannot always be maintained. When autocrats face direct challenges to their rule, they ‘more aggressively and openly attack remaining democratic space’ (Mechkova et al. Reference Mechkova, Lührmann and Lindberg2017: 167). These moments of escalating restrictions on the ‘democratic space’, often referred to as democratic backsliding or autocratization (e.g. Bermeo Reference Bermeo2016; Lührmann and Lindberg Reference Lührmann and Lindberg2019; Waldner and Lust Reference Waldner and Lust2018), produce an image problem for electoral autocrats. As electoral autocrats more severely and openly restrict political and civil rights, they are more likely to tarnish their democratic façade.
We hypothesize that periods of autocratization create incentives for electoral autocrats to appoint more women to their cabinets, at least in more recent years when norms of gender equality and democracy have become more important. Electoral autocrats who orchestrate or condone anti-democratic actions – such as imprisoning journalists and opposition leaders, changing election laws in their favour and banning political parties – expect pro-democracy actors and organizations to criticize them. Appointing more women to the cabinet provides electoral autocrats with one mechanism to divert attention away from their autocratization. Putting more women in the cabinet may draw praise from both international observers and domestic women's movements that advocate for the inclusion of women in politics. Above and beyond foreign aid inducements, electoral autocrats who are moving their countries further away from democracy are interested in their image and see the appointment of women to the cabinet as a potential reputation enhancer. Again, we expect only electoral autocrats to strategically appoint more women to the cabinet while engaging in autocratization.
Hypothesis 2: Autocratization has increased women's representation in electoral authoritarian cabinets in more recent years.
Data and methods
Hypotheses 1 and 2 are conditional upon both the type of authoritarian regime and the time period. We expect that only electoral authoritarian regimes increase the appointment of women to the cabinet to signal their commitment to democracy and gender-equality norms, and only in more recent years when norms of gender equality as a component of democracy have strengthened. Since testing these doubly conditional hypotheses is complex, our empirical analyses follow a two-step process. First, we examine the influence of aid from democracies and periods of autocratization on women's representation in cabinets using a sample of only electoral authoritarian regimes. Second, we examine whether our findings hold when including closed autocracies in the sample. If our theoretical argument is correct, including closed autocracies in the sample should result in diminished or null results for the relationship between aid from democracies and periods of autocratization and women's representation in cabinets.
We test our hypotheses using data from 38 African countries under authoritarian rule between 1973 and 2013.Footnote 3 Our unit of analysis is the country-year. We code country-years as being under authoritarian rule using the Lexical Index of Electoral Democracy (LIED) dataset (Skaaning et al. Reference Skaaning, Gerring and Bartusevičius2015). LIED classifies regime type using six binary indicators identifying the presence of: (1) elections for the legislature; (2) elections for the national executive; (3) multiparty competition; (4) male suffrage; (5) female suffrage; and (6) the competitiveness of elections. We distinguish between democracies and authoritarian regimes using component 6, the competitiveness of elections. Svend-Erik Skaaning et al. (Reference Skaaning, Gerring and Bartusevičius2015: 1501) define competitive elections as those ‘sufficiently free to enable the opposition to gain power if they were to attract sufficient support from the electorate’. In our full sample of African autocracies, we include only country-years where elections, if they exist, are considered non-competitive. From this full sample, we also create a sample of electoral authoritarian regimes. Electoral authoritarian regimes are defined as those with non-competitive multiparty elections for both the legislature and executive.
Dependent variable
Our dependent variable is the percentage of cabinet ministers that are women. Information on cabinet members comes from the WhoGov dataset (Nyrup and Bramwell Reference Nyrup and Bramwell2020). WhoGov records the names and titles of individual ministers in July each year. It also provides a binary coding of minister gender (man or woman). Since we theorize that the cabinet is a visible space where leaders can show their commitment to gender equality, we use ministers coded by WhoGov as full-rank or core to calculate both the numerator (number of women ministers) and denominator (total number of ministers). In general, full-rank incorporates ministers with titles such as minister of finance, minister of education and minister of women's affairs. Core ministers include positions in the executive branch that are seen to be integral to the cabinet and highly visible, such as vice president, second vice president, prime minister and members of a military junta. WhoGov does not code ambassadors and permanent representatives to the United Nations as core ministers. We collapse the individual minister data from WhoGov into a country-year format recording the percentage of ministers that are women.
Independent variables
Our hypotheses focus on two primary independent variables. First, Hypothesis 1 expects that increases in aid from democratic donors has increased women's share of cabinet seats in electoral autocracies in more recent years. The variable ln(DAC aid per capita) measures the natural log of official development assistance per capita from the 30 members of the Organization of Economic Cooperation and Development (OECD) Development Assistance Committee (DAC).Footnote 4 We use the natural log of DAC aid per capita because we expect the relationship between aid and women's share of cabinet seats to be subject to diminishing returns. In other words, we expect a DAC aid per capita increase from $0 to $50 to have a larger effect than an increase from $400 to $450. Data on DAC aid come from AidData 3.1 (Tierney et al. Reference Tierney2011).
Second, Hypothesis 2 expects that electoral autocrats engaged in autocratization will increase women's share of cabinet posts to boost their reputation. We code the variable Autocratize using data on civil and political rights from Freedom House (Freedom House 1973–2013). Freedom House codes both civil and political rights on a 0–7 scale, with 7 indicating the lowest level of rights and 0 indicating the highest level of rights. We invert the civil and political rights scales and add them together, producing a scale of civil and political rights (CPR) ranging from 0 (lowest level of rights) to 14 (highest level of rights). Autocratization periods begin when the CPR score declines by one point relative to the previous year. Autocratization periods continue as long as there are further yearly declines in CPR or if CPR remains stable for up to four years. Potential autocratization periods end when the CPR has remained stable for at least five years beyond the initial decline in CPR or if the CPR increases by at least one point. This coding procedure is loosely based on the coding of autocratization by Lührmann and Lindberg (Reference Lührmann and Lindberg2019). We depart from their approach by using Freedom House rather than the Electoral Democracy Index from the Varieties of Democracy dataset (Coppedge et al. Reference Coppedge2021) to identify autocratization. We find that, at least in African autocracies, Freedom House better captures periods when leaders escalate their attacks on political and civil rights.
Lastly, we expect our main independent variables to be associated with women's share of cabinet seats only in more recent years as international norms of gender equality have strengthened. Accordingly, we interact both variables with a dichotomous variable, 2000–2013, that equals 1 between 2000 and 2013 (the last year in our sample) and 0 otherwise.
Controls
We include several control variables discussed in the literature on women's representation in cabinets and legislatures. First, the United Nations Security Council adopted a resolution in 2000 that helped solidify the norm of women's inclusion in governance in post-conflict contexts. Before 2000, international and regional expectations of improved women's political representation in post-war countries were in their nascence (Hughes and Tripp Reference Hughes and Tripp2015; Tripp Reference Tripp2015). We code a binary Post-conflict variable following Alice Kang and Aili Tripp (Reference Kang and Tripp2018).Footnote 5 Post-conflict equals 1 if a country experienced the end of major intrastate armed conflict in 2000 or thereafter, and 0 otherwise. Once a country is coded as post-conflict, it stays as such unless a new major armed conflict begins.
Second, the adoption of the United Nations CEDAW in 1979 provided states the opportunity to demonstrate their commitment to gender equality. We control for the natural log of years since CEDAW ratification to account for the possibility that women's representation in cabinets increases following the ratification of CEDAW.
Third, women's share of cabinet seats may be shaped by the politics of multi-ethnic coalition-building. Leonardo Arriola and Martha Johnson (Reference Arriola and Johnson2014) find that a higher number of politicized ethnic groups discourages the appointment of women ministers because leaders prioritize the appointment of ethnic patrons, who tend to be men, to cabinet posts. We include the number of politicized ethnic groups as measured by the Ethnic Power Relations dataset (Vogt et al. Reference Vogt2015).
Fourth, to be appointed to cabinet posts, women must be part of the broader pool of potential ministerial candidates. In many cases, serving in the legislature provides a stepping stone to the cabinet. Accordingly, we include the percentage of women in the unicameral or lower-house legislature. Data for the variable Women's share legislature come from the Varieties of Democracy dataset (V11) (Coppedge et al. Reference Coppedge2021).
Fifth, leaders may expand women's inclusion in the cabinet by increasing the number of cabinet-level positions. Women are often appointed to low-prestige cabinet portfolios. Leaders could proliferate the number of low-prestige portfolios to increase women's representation rather than including them in long-existing portfolios. We control for cabinet size to address this possibility.
Finally, we control for the influence of economic development on women's representation in the cabinet by including the natural log of GDP per capita. Data on GDP per capita are taken from the Maddison Project (Bolt and van Zanden Reference Bolt and van Zanden2020).
Certain explanatory factors discussed in the global literature on women in cabinets are omitted from our analyses due to lack of variation among African dictatorships. Previous studies have examined whether parliamentary systems have higher levels of women's cabinet representation than presidential ones (Jacob et al. Reference Jacob, Scherpereel and Adams2014; Stockemer Reference Stockemer2018). Among the autocracies in our study, however, only one has a parliamentary system (Ethiopia). Other studies examine whether women presidents and prime ministers assemble more inclusive cabinets than do male leaders and find mixed evidence (O'Brien Reference O'Brien2015; Stockemer Reference Stockemer2018). None of the autocracies in our study has had a woman president or prime minister. Therefore, we do not include controls for parliamentary systems and women leaders, as previous studies have done.
Empirical model
We test our hypotheses by estimating linear regression models with country-fixed effects. The inclusion of country-fixed effects directs our analysis at within-country variation in women's share of cabinet posts. It also adjusts our estimates for time-invariant country-specific confounders that may influence women's share of cabinet posts. In each model, we cluster standard errors by country and lag all independent variables by one year.
Results
As discussed above, the first step in our analyses focuses on a sample of electoral authoritarian regimes in Africa. The results from our tests of Hypotheses 1 and 2 are presented in Table 1. Table 1 contains four model specifications of increasing complexity. Models 1 and 2 provide baseline estimates for our variables of interest without including interaction terms with the 2000–2013 time period variable. Model 2 differs from Model 1 by including the full set of control variables described above along with our variables of interest. Models 3 and 4 interact both ln(DAC aid per capita) and Autocratize with 2000–2013. Model 4 differs from Model 3 by including the full set of control variables. In our discussion of the results below, we focus on our preferred specification presented in Table 1, Model 4.
Fixed Effects Models of Women's Share of Cabinet Posts in Electoral Autocracies in Africa
Note: Linear regressions with country-fixed effects. Standard errors clustered by country are in parentheses. All independent variables are lagged one year. * p < 0.05, ** p < 0.01, *** p < 0.001.
Table 1 provides some support for Hypothesis 1, but not for Hypothesis 2. Since the interpretation of the results in Table 1 is complicated by interaction terms, we present the marginal effects of ln(DAC aid per capita) and Autocratize graphically in Figure 2. As we expect, aid from democratic donors is positively and significantly associated with women's share of cabinet seats during the 2000–2013 period. The marginal effect of aid is similar for years before 2000, but is not significant.Footnote 6 To further interpret the association between ln(DAC aid per capita), we plot the relationship in terms of percentage changes in DAC aid per capita in Figure 3. It shows, for example, that a 250% increase in DAC aid per capita is associated with a modest 1.8% increase in women's share of cabinet seats. We also include a rug plot on the bottom of Figure 3 to show the distribution of actual aid increases in our data. As the rug plot shows, most aid increases in our data were below 250%. These findings suggest that the impact of any individual aid increase is slight, but countries experiencing large increases in aid across multiple years could see more substantial increases in women ministers.
Marginal Effects of Aid and Autocratization from Table 1, Model 4
Examining the Effects of Aid during the 2000–2013 Period in Electoral Authoritarian Regimes
Figure 2 does not support Hypothesis 2. As expected, the marginal effect of Autocratize is positive in the 2000–2013 period, but the estimate is imprecise and far from statistically significant. Electoral autocrats in Africa do not appear to have consistently used appointments of women to the cabinet to offset the reputational costs of their autocratization. While previous research from Valdini (Reference Valdini2019) suggests that women's representation in legislatures increases during more recent periods of autocratization in hybrid regimes, we do not find support for this dynamic in electoral authoritarian cabinets in Africa.
Next, we proceed to the second step in our analysis by examining whether our findings from Table 1 are sensitive to the inclusion of closed authoritarian regimes in the sample. The results are presented in Table 2 and Figure 4. Consistent with our expectations, the marginal effects for ln(DAC aid per capita) plotted in Figure 4 suggest that increases in aid from democracies are only associated with increases in women's share of cabinet seats in electoral authoritarian regimes. When including both electoral and closed autocracies in the sample, the marginal effects of aid from democracies, shown in Figure 4, are not significant in either time period. Including closed autocracies in the sample also decreases the size of these marginal effects. This increases our confidence in the theory behind Hypothesis 1. Increases in women's share of cabinet seats in response to aid increases from democratic donors appear to be unique to electoral autocracies. We argue that this pattern can be explained by the incentives electoral autocrats have to signal their commitment to the increasingly intertwined norms of democracy and gender equality.
Marginal Effects of Aid and Autocratization from Table 2, Model 4
Fixed Effects Models of Women's Share of Cabinet Posts in African Autocracies
Note: Linear regressions with country-fixed effects. Standard errors clustered by country are in parentheses. All independent variables are lagged one year. * p < 0.05, ** p < 0.01, *** p < 0.001.
The results for Hypothesis 2 remain consistent when including both electoral and closed autocracies in the sample. As Figure 4 shows, the marginal effects for Autocratize remain statistically insignificant when using the broader sample of African autocracies. While African autocrats occasionally engage in autocratization, they do not appear to regularly increase women's representation in cabinets to offset the reputational costs of their attacks on political and civil rights.
Examining the indirect effects of aid and autocratization
One potential reason for the relatively modest findings for Hypotheses 1 and the null findings for Hypothesis 2 is that aid from democratic donors and periods of autocratization have an indirect effect on women's share of cabinet seats. An indirect relationship may operate through women's share of legislative seats, which has a robust positive association with women's share of cabinet seats. For example, previous scholarship shows that decreases in the level of democracy in hybrid regimes are associated with increases in women's share of legislative seats (Valdini Reference Valdini2019). Therefore, the inclusion of women's share of legislative seats as a control in our models may be masking indirect effects that aid from democracies and autocratization have on women's share of cabinet seats.
We explore the potential for indirect effects in two ways. First, we drop women's share of legislative seats from the Model 4 specifications in Tables 1 and 2.Footnote 7 The results for ln(DAC aid per capita) differ from those in Tables 1 and 2, but are not suggestive of an indirect effect operating through women's share of legislative seats. Interestingly, we find that the marginal effect of ln(DAC aid per capita) is only significant for the pre-2000 period in the sample restricted to electoral authoritarian regimes. These findings are the opposite of what we find in Figure 2 above. We also do not find evidence that the effect of Autocratize is being masked by the inclusion of women's share of legislative seats as an independent variable in our main specifications. When dropping women's share of legislative seats from our main specifications, the marginal effect of Autocratize remains insignificant for the pre-2000 and 2000–2013 periods when using both the electoral and all-authoritarian regime samples.
Second, we estimate additional models predicting women's share of legislative seats with aid from democracies and periods of autocratization.Footnote 8 We use a model specification similar to those in Tables 1 and 2 above, and interact both ln(DAC aid per capita) and Autocratize with the 2000–2013 period variable. As in our main analyses, we estimate models using both the electoral authoritarian and all-authoritarian samples. These models also do not suggest that our main findings are confounded by indirect relationships between aid, autocratization and women's share of cabinet seats that operate through women's share of legislative seats. We find that the marginal effect of ln(DAC aid per capita) on women's share of legislative seats is only significant (and positive) during the pre-2000 period in the all-authoritarian regime sample. This goes against our expectation that the effect of aid from democracies on women's representation should be strongest in electoral authoritarian regimes in more recent years. Similarly, we find no evidence that periods of autocratization significantly affect women's share of legislative seats. These null findings hold during both time periods (pre-2000 and 2000–2013) when using either the electoral authoritarian or all-authoritarian regime samples. Taken together, these analyses do not suggest our main findings are confounded by indirect effects through women's share of legislative seats.
Robustness checks
We conduct a variety of robustness checks to examine how sensitive our main findings are to alternative model specifications and alternative measures of our main independent variables.
We consider several alternative measures of aid to examine the robustness of our main findings. First, DAC aid per capita may not adequately capture a country's aid dependence. We replicate Model 4 from Tables 1 and 2 using a measure of DAC aid as a percentage of GDP from the World Bank (World Bank 1973–2013). The results are very similar to those of the specifications in Tables 1 and 2 using ln(DAC aid per capita) as the measure of aid.Footnote 9 As with our main analyses, we find that the marginal effect of ln(DAC aid % GDP) is significant and positive only during the 2000–2013 period in the electoral authoritarian regime sample.
Second, patterns of aid delivery are far from consistent across years for most countries. While a country may receive a robust aid package in year t, years t − 1 and t + 1 may draw little to no aid. It may be the case that aid from democracies has a larger impact on women's share of cabinet seats when aid delivery is consistently high across multiple years. We capture longer-term aid commitments using a three-year moving average of ln(DAC aid per capita). We find that the marginal effect for the three-year moving average of ln(DAC aid per capita) is significant and positive for the pre-2000 and 2000–2013 periods in the electoral authoritarian regime sample.Footnote 10 During the 2000–2013 period in electoral authoritarian regimes, for example, a 250% increase in the three-year moving average of ln(DAC aid per capita) is associated with a 3.2% increase in women's share of cabinet seats.
Third, it may be the case that women's share of cabinet seats responds to DAC aid, but does so more slowly that our main specification using a one-year lag of aid allows. We thus estimate our Model 4 specification from Tables 1 and 2 using three-year lags of our independent variables. The results suggest that the effects of DAC aid are more immediate, at least in electoral authoritarian regimes during the 2000–2013 period.Footnote 11 The marginal effect for the three-year lag of ln(DAC aid per capita) is only positive and significant in the electoral authoritarian regime sample during the pre-2000 period.
We also consider the robustness of our findings to an alternative coding of autocratization periods. In our main measure of Autocratize, autocratization periods begin when there is a drop in the combined Freedom House civil and political rights (CPR) score of at least one point. Autocratization periods end when the CPR score increases or remains constant for four years. It may be the case that this measure includes too many cases experiencing only minimal autocratization. Therefore, we code a stricter version of Autocratize that requires greater backsliding. To code this variable, we start with the list of cases where our main measure of Autocratize equals 1. Then, we code the variable Strict Autocratize which equals 1 for autocratization periods where the total decline in the combined Freedom House score was at least two points.Footnote 12 This reduces the country-years coded as autocratization periods by 36% in the electoral authoritarian regime sample and 30% in the sample of all-authoritarian regimes. Using this stricter measure of autocratization does not change our main findings.Footnote 13 The marginal effect of Strict Autocratize is not significant in either time period when using either the electoral authoritarian or all-authoritarian regime sample. However, our findings for ln(DAC aid per capita) remain consistent with those from Tables 1 and 2.
Our main analyses test whether aid and autocratization influence women's share of cabinet seats in the later years in our sample (2000–2013). It may be the case that international norms on gender equality had already begun shifting prior to 2000. Thus, we replicate our main analyses using a dichotomous variable covering the period from 1995 to 2013. This produces results similar to those in Tables 1 and 2.Footnote 14 The marginal effect for Autocratize is not significant in either time period or sample. Similar to our main analyses, the marginal effect for ln(DAC aid per capita) is significant and positive for the 1995–2013 period using the electoral authoritarian sample. The effect size is also similar. Additionally, the marginal effect for ln(DAC aid per capita) during the 1995–2013 period is also significant and positive in the sample of all-authoritarian regimes.
Finally, it may be the case that our findings are confounded by dynamics in leader tenure. For instance, new leaders may have incentives to appoint many women to their inaugural cabinets, but dismiss them soon after. Controlling for a leader's first cabinet and its interaction with the 2000–2013 period variable does not alter our main findings in Tables 1 and 2.Footnote 15
Conclusion
While not close to achieving gender parity, women's share of cabinet posts has been on the rise in autocracies. The goal of our article has been to highlight variation and explain change in women's cabinet share in autocracies. Focusing on increases in the percentage of women in cabinets, we theorized that exposure to international norms of democracy and women's rights induces electoral autocrats to appoint more women to the cabinet. Using data on 38 African autocracies between 1973 and 2013, we evaluated the relationship between women's cabinet share, foreign aid dependence and autocratization. In contrast to previous works, our study has put the inclusion of women in authoritarian cabinets front and centre in the theory and empirical analysis.
We find some evidence that Africa's electoral autocrats strategically appoint more women ministers to signal their commitment to democracy and gender-equality norms. In support of Hypothesis 1, our findings suggest that Africa's electoral autocrats have systematically increased the appointment of women ministers in response to increases in aid from democracies. Regarding Hypothesis 2, we do not find that autocrats, whether in electoral or closed regimes, increased the share of women ministers to improve their international reputation during periods of autocratization. One potential explanation for this null finding is that periods of autocratization tend to occur when autocrats face threats to their rule. In such situations, it may be too costly to increase the share of women ministers in the cabinet as a signalling mechanism. Cabinet appointments are among the most important rewards that authoritarian leaders can provide to societal elites. If leaders use these finite appointments to signal their commitment to democracy and gender-equality norms during times of heightened uncertainty, it risks creating even more threats from powerful domestic elites who are overlooked.
There may be less costly avenues than women's cabinet appointments for signalling commitment to international norms of gender equality and women's rights. Autocrats may more easily increase women's representation in legislatures by adopting gender quotas. As Assaf David and Stefanie Nanes (Reference David and Nanes2011) describe, this appears to be the case with the adoption of a gender quota for municipal councils in Jordan. Increasing women's representation in other areas of governance provides a way for autocrats to signal their commitment to international norms without disrupting the most important power structures within their regime. In some cases, the adoption of women's rights laws may be a relatively cost-free way for dictators to show that they are in line with international norms (Donno et al. Reference Donno, Fox and Kaasik2022).
The appointment of more women to authoritarian cabinets could be less costly if there is concerted domestic pressure for the inclusion of women in high-level government posts. Further analyses should inspect the relationship between international inducements and domestic incentives for autocrats to place more women in the cabinet. In some cases, such as Burundi, women's movements were successful in demanding a constitutionally enshrined gender quota for government positions that, at least on paper, locked selectors into appointing women ministers (Burke et al. Reference Burke, Klot and Bunting2001). In other cases, autocrats may see women voters or women in ruling party structures as an important base of support to placate, but it remains unclear how and when this calculus would affect the composition of the cabinet.
The findings from our statistical analyses suggest that a combination of factors encourage the appointment of more women to authoritarian cabinets. We found that among authoritarian countries, those with a higher percentage of women in the national legislature have higher percentages of women in cabinet. The descriptive representation of women in legislatures is used in the scholarship to capture two related ideas that may be part of the same causal process. Gender-egalitarian societies are amenable to having more gender-equal legislatures, and having more women with legislative experience increases the number of women in the pool of potential ministers. Our results provide general empirical support for this claim.
Our article has examined the overall gender composition of the cabinet, setting the stage for scholarship on gender and authoritarian cabinets. Existing studies show that autocrats, particularly in personalist regimes, regularly dismiss ministers from the cabinet or reshuffle them to new portfolios (Kroeger Reference Kroeger2020). However, few studies examine whether particular types of ministers in autocracies – such as men versus women – are more likely to be dismissed or reshuffled. Existing research spanning regime types shows that gains in women's share of cabinets seats are often subject to backsliding (Scherpereel et al. Reference Scherpereel, Adams and Jacob2018). However, explanations for decreases in women's cabinet representation are underdeveloped. Future work investigating the gendered dimensions of cabinet reshuffles has the potential to shed new light on the strategy behind coalition building in autocracies.
Additionally, future work should theorize about the appointment of women to high-visibility or high-prestige positions in authoritarian cabinets. Autocrats may strategically encourage the appointment of women to powerful posts such as prime minister, vice president or minister of foreign affairs to manage their country's reputation. Uganda's President Yoweri Museveni's appointment of Jessica Alupo as vice president and Robinah Nabbanja as prime minister in 2021 could be an example of an autocrat using highly visible changes in the government for strategic purposes (Tripp Reference Tripp2021). Cross-national studies suggest that the appointment of women to high-prestige posts in the government such as minister of finance follow major crises (Armstrong et al. Reference Armstrong2022). Understanding the conditions in which women join the inner cabinet in autocracies, and whether such high-level appointments yield the expected benefits to autocrats, is an area for further scholarship.
The number of women holding positions of political power has climbed around the world over the past several decades. Authoritarian countries have been a part of this trend, as indicated by women's rising share of ministerial posts in non-democracies. Understanding the inclusion of women at the highest level of government in autocracies is important for understanding how autocrats manage the challenge of staying in power and having a fuller picture of the conditions under which women obtain a seat at the table.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/gov.2022.32.
Acknowledgements
The authors thank Ludger Helms and Michelangelo Vercesi for their constructive feedback on an earlier draft of this article. Attendees at the Department of Political Science IR/CP Brownbag at the University of Nebraska–Lincoln provided helpful comments as well, especially discussant Makayla Wendland.
Open access
