1doi: 10.5129/001041522X16222193902161
Democracy and the Adoption of
Electoral Gender Quotas Worldwide
Pär Zetterberg, Elin Bjarnegård, Melanie M. Hughes, and
Pamela Paxton
Keywords: gender quotas, democracy, reserved seats, candidate quotas, elections, electoral
autocracies.
Electoral gender quotas––policies requiring that a certain share of women be included
as legislators or candidates––have been adopted by more than 130 countries. As a
result, the literature on this rapidly spreading electoral reform has increased dramatically
during the past two decades.1
Adding to a rich literature that examines where and under
what circumstances quotas are adopted,2
this article theorizes and empirically investigates
the role of democracy in quota adoption.
Although quota adoption has been analyzed in all regime types, spanning advanced
democracies3
through transitional democracies4
to autocracies,5
scholars have suggested
competing accounts about its broader relationship to democracy.6
On the one hand,
democratic governance signifies normative commitments to inclusiveness and creates
space for women’s movements to exert pressure for change,7
which suggests that
democracy increases the chances of gender quota adoption. On the other hand, authoritarian
states may take steps to promote women in politics to appear more modern, to
attract foreign aid, or to distract from other forms of political exclusion.8
Furthermore,
in less democratic countries, national legislatures often wield less influence, which could
lower the cost of quota adoption. Between these positions, some scholars suggest that
quotas commonly are adopted when political institutions are under negotiation, for instance
after a conflict or in a democratic-transitional setting.9
Although empirical examinations of quota adoption have seldom been framed in
terms of democracy or regime type,10
empirical research on quota adoption has suggested
that level of democracy and gender quota adoption are correlated. One of the
few large-N analyses touching on the relationship between democracy and gender quota
adoption suggests that the relationship is curvilinear: quotas are most likely to be adopted
at the middle of the democracy scale.11
An emerging body of evidence further
2
Comparative Politics January 2022
suggests that democracy pushes countries into adopting distinct quota types—candidate
quotas (i.e., quotas that require parties to nominate a certain proportion of women for
office) versus reserved seats (i.e., quotas that set aside a certain number of seats in the
legislature for women). Scholars have suggested that democracies (mainly in Europe
and Latin America) mostly adopt candidate quotas, whereas more autocratic countries
(mainly in Africa and the Middle East) tend to adopt reserved seats.12
The two types
of quotas have been claimed to be rooted in different logics, target distinct parts of the
electoral process, and yield varied results.13
Yet, due to data limitations, these propositions
about democracy and quota adoption have been made mostly in small-N case study
analyses and have rarely been tested in large-scale statistical analyses.
To add even more complexity to this discussion, democracy as a concept is multi-dimensional
and its different dimensions may imply different incentives for how to include
women and push countries toward one trajectory of quota adoption versus another. We
follow in the tradition of democratic theorists (notably Robert Dahl) and many democratization
scholars who distinguish the democratic dimension of having free and fair elections
that include an authentically competitive political opposition from the dimension
of effective guarantees of civil liberties.14
Each dimension of democracy implies distinct
constraints on and incentives for the political actors who may push for quotas (e.g., civil
society activists) or ultimately debate and vote for quotas (e.g., political elites).
Using a new, global, and longitudinal dataset, Quota Adoption and Reform over
Time (QAROT),15
we analyze how the different dimensions of democracy relate to the
adoption of different types of gender quotas. These new data on quotas for the first time
distinguish reserved seats from candidate quotas cross-nationally and longitudinally.
Combining QAROT with a well-established measure of two dimensions of democracy
(political rights and civil liberties) provides the opportunity for novel hypothesis testing.
To examine the relationship, we make use of event history analyses.
Our results corroborate earlier findings that quotas are more likely to be adopted in
countries at the middle of the democracy scale, or that the relationship between democracy
and quota adoption is curvilinear.16
However, we demonstrate that considering only an
average effect masks two distinct processes: the adoption of reserved seats peaks at lower
levels of democracy, whereas the adoption of candidate quotas peaks at middle to upper
levels of democracy. Further, we demonstrate that separating dimensions of democracy
reveals further complexity in understanding quota adoption. Political rights (our measurement
of having free and fair elections that are fully contested) has a stronger effect on the
adoption of reserved seats. Countries with lower (but not the lowest) levels of political
rights are most likely to adopt reserved seats, whereas countries on the high end of political
rights approach a zero probability of introducing them. On the other hand, it is the
dimension of civil liberties (having freedoms of expression, assembly, and association,
etc.) that better predicts candidate quota adoption. The likelihood of adopting candidate
quotas is highest in countries with higher (but not the highest) civil liberties.
By showing this nuanced relationship between democracy and quota adoption, our
study makes several contributions. It provides, to our knowledge, the first global pieces
of evidence that countries, depending on their levels and dimensions of democracy,
3
Pär Zetterberg et al.
follow distinct trajectories in their adoption of different types of gender quotas. Understanding
the distinct routes to reserved seats and candidate quotas is important not
only for scholars, but also for policymakers, practitioners, and activists interested in
adopting new quotas across countries that vary in level and dimension of democracy. In
addition, and more broadly, the study speaks to researchers of authoritarian regimes and
democratization processes, suggesting that reserved seats, and gender equality policies
more broadly, may represent an unexplored legitimization strategy used by authoritarian
leaders engaging in elections.17
Gender Quotas and Their Adoption
Electoral gender quotas can take the form of national law (legal quotas) or be voluntarily
adopted by political parties (party quotas). Here, our focus is only on the adoption
of legal quotas and their two main types: reserved seats and candidate (or legislative)
quotas.18
These two quota types share the common purpose of increasing the number
of women in legislative institutions, and, because they are legal quotas, they also have
to be approved as changes to the electoral law by the legislature. The fact that reserved
seats and candidate quotas apply to different stages of the election process, however, also
makes them different in important ways. Applying to the stage of the elected, reserved
seats is the only quota type to guarantee that a minimum number of women are elected to
the legislature. In addition, reserved seats are often characterized as “add-on seats,” that
is, they are commonly adopted by enlarging the legislature and the election of women
to these seats usually runs parallel to the existing electoral system.19
Women elected
to reserved seats are elected directly (e.g., in elections with only women candidates)
or indirectly (e.g., after the election has taken place).20
In contrast, as candidate quotas
apply to the level of candidates, this quota type is integrated in political parties’ regular
selection process. Consequently, women elected through candidate quotas are selected
in the same way as men.21
A large number of studies have focused on understanding where quotas have been
adopted. They have noticed that quota adoption has taken place in all regions of the
world, in developed as well as in developing countries, and in contexts where gender
equality is more advanced as well as in countries where women have very few seats in
parliament.22
Quota types distribute differentially around the world: Reserved seats primarily
exist in Africa, Asia, and the Middle East, whereas candidate quotas have mainly
been adopted in Latin America, Europe, and Africa.23
Gender quota adoption is also seen as related to democracy, although scholars have
suggested that the relation between the two is complicated.24
Large-N analyses looking
at the relationship between level of democracy (or regime type) and gender quota adoption
show that countries at the middle of the democracy scale are the most likely quota
adopters. Using descriptive statistics, Dahlerup shows that 32 percent of “semi-democratic”
countries have adopted quota laws, whereas 16 percent of democratic countries
and 21 percent of authoritarian countries have done so.25
Hughes, Krook et al. were
4
Comparative Politics January 2022
the first to demonstrate a curvilinear relationship between democracy and gender quota
adoption using multivariate analysis.26
In line with Dahlerup, they find that countries at
the middle of the democracy scale are the most eager adopters of legal gender quotas,
whereas more advanced democracies and more full-fledged authoritarian states are less
likely to reform their electoral systems.27
One explanation of this pattern is that quota
adoption is more likely in countries where the political framework is under negotiation
in some way. They may be post-conflict settings or countries in which democratizing
reforms are underway. When constitutions and election laws are open to change, this
may be a window of opportunity for political gender equality gains, including the introduction
of gender quotas.28
However, one limitation of prior statistical analyses on the relationship between
level of democracy and quota adoption is that they consider quotas as singular, without
acknowledging differences between reserved seats and candidate quotas. Since the two
types of quotas are so different, the processes that lead to their adoption are likely to
be different as well. Further, by drawing on scholars that have employed a definition
of democracy that includes at least two dimensions––free and fair contested elections
and guarantees of civil liberties29
––we suggest that different dimensions of democracy
may also influence processes of quota adoption in different ways. We therefore suggest:
(1) that the influence of the level of democracy may vary across reserved seats and
candidate quotas and (2) that the different dimensions of democracy may differentially
affect quota adoption of different types. Towards the goal of getting a more comprehensive
picture of how different regimes adopt gender quotas, we next seek to theorize the
relationship between different dimensions of democracy and the adoption of quotas of
different types.
Democracy and Quota Adoption
While the final decision for legal quota adoption often lies with members of parliament
(MPs), the literature identifies other actors as critical for pushing for this reform, including
women’s and international organizations. Various analyses have emphasized
women’s collective mobilization—both in national30
and transnational31
organizations—as
fundamental for quota adoption. Women in civil society, political parties, and
legislative institutions have pushed for reforms to increase their presence in political
decision-making bodies. Women have, in some cases, been backed up by international
organizations, pushing for quota adoption as part of their democracy promotion ef-
forts.32
International organizations have been important actors whenever quotas have
been one of many components in post-conflict peace operations. In some cases, the
adoption of gender quotas has also been integrated in international organizations’development
aid programs.33
Civil society actors like women’s movements and international
organizations both intersect with political elites within a given political (democratic)
context. The democratic context provides them with either opportunities or limitations
to push for quota adoption in general and of specific types. That is, the different levels
5
Pär Zetterberg et al.
and dimensions of democracy provide incentives and constraints that create unique
trajectories for quota adoption. Importantly, the constellation of level of democracy and
its dimensions produces trajectories that more easily lead to a candidate quota or to a
reserved seat quota.
Let us begin with the adoption of candidate quotas, the more common type of gender
quota that is pushed by both women’s movements and international actors. A country
at the upper levels of democracy, where citizens are granted individual liberties and
freedoms, is more open to pressure from both above and below. This pressure is likely
to be for candidate quotas rather than reserved seats because where quota advocates are
able to engage in large coalitions, the nature of quota debate tends to be fairly transformative.
The emphasis is often on moving away from women playing a secondary role in
politics and putting them on par with men in their political parties. Women’s ambition is
to challenge and replace men, including sitting incumbents, to level out inequalities in
political power. With such a focus, attention is naturally paid to the reshuffling of parties’
candidate lists. As a consequence, the adoption of candidate quotas becomes the plausible
policy choice for countries at a higher level of democracy.34
This argument makes clear that guarantees of civil liberties are the key dimension
of democracy operating in the adoption of candidate quotas. For coalitions that include
both women party activists and autonomous women’s movements to emerge, freedom of
expression, association, and assembly are needed. Wherever governments repress civil
society actors, they also restrict the possibilities for quota advocates to push for policy
adoption. Thus, we expect civil society actors such as women’s movements to be most
successful in pushing for candidate quotas in countries where individual liberties and
freedoms are relatively protected. In such countries, citizens are more often able to mobilize
politically between elections, and civil society actors have more space to advocate
for changes in the electoral system.35
To give an example, Argentina’s pioneering candidate quota law (in 1991) has
been attributed to the country’s return to democratic rule. With democracy reinstalled,
women party activists increased their contacts with women in other organizations in
the Argentine civil society as well as with politically active women in other countries.
The broad coalition of women, with different ideological positions, focused on transforming
the political parties and their candidate selection practices. Activists met with
women from countries where women’s representation was larger and where women had
a stronger position in political parties (e.g., Germany and Sweden). When discussing
what strategies to use, they decided both to put pressure on political parties to change
their party statutes and to push for an electoral reform that included the adoption of
gender quotas. From the first bill proposal presented by a woman senator in 1989 and
forward, the quota debate concentrated on candidate quotas. The possibility to adopt
reserved seats rather than candidate quotas never reached the government’s agenda.36
After Argentina became the first country in the world to adopt a sizeable candidate
quota law, quota advocates in other countries learned from Argentina, which resulted in
similar reform processes and the adoption of the same quota type in many other Latin
American countries.37
6
Comparative Politics January 2022
In contrast, the adoption of reserved seat quotas generally comes from a different
set of incentives for political actors. Although elite women and women’s groups have
pushed for reforms in all sorts of political contexts (including in authoritarian regimes),
adopting a quota as a signal of a democratic reform fits the agenda of authoritarian
regimes, in particular the large group of electoral authoritarian countries that hold, but
often manipulate, multiparty elections.38
In an effort to appear democratic, gain international
recognition, and attract foreign aid and loans, (electoral) authoritarian rulers have
often considered the adoption of reserved seats as a low-cost strategy.39
Remember that international organizations push democratization, as well as quota
adoption. We suggest that their influence is most present and most efficient in less democratic
countries, such as in electoral autocracies.40
Of course, such a suggestion does
not imply that these organizations (such as the UN and the EU) lack influence on quota
campaigns in other contexts. For example, the UN’s Beijing Declaration and Platform
for Action (1995) put pressure on most governments across the globe to adopt quotas.
However, international organizations have primarily actively pushed for quotas in less
democratic countries where they have also engaged in democracy-promoting activi-
ties.41
International organizations have often pushed for quotas as part of peacebuilding
operations and post-conflict constitution-making, as well as in developing aid programs.
In the latter case, they have encouraged aid-dependent countries to signal their commitment
to democracy by adopting quotas.42
Following a “low-cost strategy,” therefore, largely autocratic leaders responding
to pressure from international organizations will most commonly prefer reserved seats
quotas. This quota type is less likely to challenge incumbents and the existing political
system than candidate quotas. In other words, it is a relatively “safe” quota for political
leaders to adopt, and they have the maneuverability to do so because of low contestation
in the electoral arena. A reserved seats strategy quickly increases the number of women
in parliament (and thus meets the goal of international organizations) without posing a
threat to incumbents.
Thus, free and fair elections that include an authentically competitive political
opposition are the key dimension of democracy operating in the adoption of reserved
seats. A nondemocratic regime wishes to signal some normative compliance without
threatening the existing power structure. The regular election system can continue to
work in parallel to the reserved seats, and thus sitting (men) incumbents do not run
the risk of being replaced by women as a consequence of the quota policy. This feature
of reserved seats makes this quota type attractive for those leaders who want to
strategically use quotas to increase legitimacy in the eyes of the international community,
or to electorally boost the ruling party.43
We would mainly expect the adoption
of reserved seats in cases where the regime’s power is secure and where contestation
in the electoral arena is low (such as in electoral authoritarian regimes that include a
dominant party).
To give just one example, in electoral authoritarian Tanzania the dominant party,
Chama Cha Mapinduzi (CCM), has been able to design reserved seats quotas in a manner
that does not threaten incumbents’ power and that reinforces the party’s electoral
7
Pär Zetterberg et al.
power vis-à-vis weak opposition parties. In this aid-dependent country, where the CCM
has upheld an uneven electoral playing field for decades, women are elected indirectly to
reserved seats, after the election has taken place. Thus, the election of women to reserved
seats does not interfere with the regular election process or threaten the CCM’s electoral
dominance. While a broad range of politically active women support quotas, the group
of women that have benefitted most from the reserved seats are those women who are
already coopted and strongly controlled by the ruling party.44
Summing up the argument, we suggest that making a distinction between quota
types, as well as the relevant dimensions (or attributes) of democracy, qualifies our understanding
of the relationship between democracy and quota adoption. Redefining the
curvilinear relationship identified by Hughes, Krook et al.,45
we expect the curves to
look different for reserved seats and candidate quotas. The adoption of the former quota
type will peak at the lower end of the democracy scale whereas the latter quota type will
be preferred by countries at the mid to high end of the democracy scale. The underlying
arguments backing up these expectations relate to different dimensions of democracy.
Whereas international organizations’ targets are mainly governments of countries
where electoral institutions are poorly developed or manipulated by the governing party,
broad and effective coalitions of women activists are more likely where citizens have
effectively enforced liberties and freedoms. Consequently, we expect candidate quota
adoption to be most likely where citizens are granted a set of civil liberties, whereas we
expect reserved seats adoption to peak where (multiparty) elections are not free and fair
and where there is rarely an authentic political opposition to the regime.46
These arguments are, of course, not without exceptions. For instance, there are
examples of women’s groups and elite women in less democratic or semi-authoritarian
regimes (e.g., Algeria and some Arab states) who have put strong pressure on the ruling
men to adopt gender quotas, sometimes in coalitions with international organizations.
In addition, in democracies men elites have also adopted quotas for strategic reasons:
to attract new groups of voters by signaling to constituents that their party is progres-
sive,47
to give party gatekeepers more control over candidate selection,48
and because
men incumbents believe that it will increase their reelection possibilities if they face
inexperienced women candidates.49
Still, we believe our general argument—that particular
constellations of international organizations, coalitions of women activists, political
elites, and democratic institutions will result in differential adoption of candidate
versus reserved seat quotas—will hold globally across a range of countries, as we now
proceed to test.
Data and Methods
Our analysis includes 150 countries at all levels of democracy from 1989 to 2014.
Our selection criteria include sovereignty and population size. States not independent
in 1989 enter the analysis in their first year of sovereignty. This means, for
instance, that the constituent parts of Communist states like the USSR (e.g. Russia
8
Comparative Politics January 2022
and Ukraine) and SFR Yugoslavia (e.g. Serbia and Croatia) enter the analysis in their
first year as sovereign states, while the Communist states themselves (i.e. the USSR
and SFR Yugoslavia) are excluded from the analysis. Countries that reached formal
independence after 2000 are excluded from the analysis. Following cross-national
research conventions, we analyze only countries with a population greater than one
million in the year 2000.
Modeling quota adoption requires that we consider when, in historical time, the
“risk” of quota adoption begins (e.g., when we start the clock). Expectations of widespread
policy adoption are not appropriate prior to being demonstrated by a pioneering
state. Therefore, we consider the risk of quota introduction to begin in 1989, when
Uganda became the first state to adopt a reserved seat quota guaranteeing women at
least a 10 percent share of seats (Uganda’s quota was 18 percent). Argentina shortly
thereafter, in 1991, became the first country to adopt a sizeable candidate quota with an
electoral law mandating that parties nominate at least 30 percent women. Once a quota
law has been introduced in at least one country, it is reasonable to assume that all sovereign
countries are also at “risk” of passing one.
Given our interest in explaining quota adoption across countries over time, duration
or event history analysis is the appropriate method.50
Because parliamentary information
is typically recorded yearly, we use discrete time logistic regression models. In this
analysis, time represents historical time.51
That is, a country’s “risk” of adoption begins
at particular historical dates rather than in relation to a country’s internal clock. We are
interested in the first attainment of a quota law, so we do not treat quota adoption as a
repeatable event, even though some states have revised or repealed their laws. Countries
that have not yet adopted a quota by 2014 are right censored.52
Dependent Variable  Drawing from the new, global, and longitudinal Quota Adoption
and Reform Over Time (QAROT) dataset,53
we predict the adoption of gender
quotas affecting the lower or single house of the national legislature. Adoption is coded
as the year that a gender quota first becomes part of a country’s constitution or
secondary law such as the electoral law and is coded whether the law was ever implemented
in an election. Consistent with other research on quota adoption,54
we exclude
quotas that regulate less than 10 percent of candidates or seats and those that do not
specify a quota threshold. We see these lower-threshold quotas, which emerged in the
1950s, typically as very small reservations, as distinct from the newer wave of quotas
that took hold in the 1990s and have transformed women’s descriptive representation.
Practically, this exclusion rule affects ten countries that adopted quotas after 1989, and
all but one country (Romania) went on to adopt a more substantial quota during our
study period.
We model quota adoption looking first at all national quotas and then separately
at reserved seats and candidate quotas. To reiterate, reserved seats set aside seats in the
legislature for women, whereas candidate quotas require all parties to field a certain
percentage of women candidates or nominees. A handful of countries use a combination
9
Pär Zetterberg et al.
of reserved seats and party list requirements for candidates (e.g., Rwanda) and therefore
do not fit neatly into either category. In the models that distinguish reserved seats from
candidate quotas, we model these cases as having reserved seats.55
For both practical and theoretical reasons, we do not analyze voluntary party quota
adoption. Because these measures are introduced by individual political parties, large
and small, information on exact years of their adoption is not available for all cases.56
Furthermore, given that their adoption often stems from considerations of party ideology
and electoral competition, pressures on parties to pursue quotas are different than those
operating on a national scale.57
Independent Variables  To measure our two-dimensional conceptualization of democracy,
we make use of data on democracy from Freedom House (2010) that reports
annual ratings from 1972 to the present. Freedom House breaks its scales into political
rights and civil liberties, which resonate well with our two dimensions: free and fair
elections that are fully contested (political rights) and guarantees of civil liberties, discussed
above. According to Freedom House, political rights exist to the extent that a
meaningful process elects the national government and parties compete for political
power. In particular, Freedom House assesses political rights as “Electoral Process,
Political Pluralism and Participation, and Functioning of Government.”58
Civil liberties,
on the other hand, exist when the people of a country have the freedom to express
political opinions in any media and the freedom to organize and to participate in any
political group. Freedom House assesses civil liberties as “Freedom of Expression and
Belief, Associational and Organizational Rights, Rule of Law, and Personal Autonomy
and Individual Rights.”59
We estimated the models using the full Freedom House democracy
score (civil liberties and political rights) as well as political rights and civil
liberties separately.60
We model all three as both linear and nonlinear. Below we also
discuss the results from auxiliary models substituting another frequently used measure
of democracy, Polity IV’s composite democracy measure (polity2), for Freedom
House’s full measure of democracy.61
Understanding the relationship between democracy and quota adoption requires
placing both in a reasonable baseline model of quota adoption. Based on the set of existing
cross-national analyses of quota adoption, we include control variables of two
types: those that assess international influences on quota adoption and those that address
the national characteristics that may affect whether a country is willing or easily able to
adopt a national quota measure.
International Influences  Researchers have increasingly considered global and regional
forces on quota adoption, focusing especially on the impacts of foreign aid and regional
diffusion.62
We first account for the importance of foreign aid on a country’s decisions to
adopt gender quotas with a logged measure of official development assistance.63
Second,
following research demonstrating that growth in quotas in a region increases the chances
10
Comparative Politics January 2022
of quota adoption by other countries in that region,64
we construct a measure of the density
of regional adoption. We measure the share of countries within a given region that have
adopted a gender quota by a given year. For example, by 2000, 27 percent of Asian countries
had adopted some form of national quota, while only 7 percent of Eastern European
countries had done so. The regional density variable is therefore coded twenty-seven for
all Asian countries in 2000 and seven for all Eastern European countries for that year.
Region is coded into six categories: the West, Eastern Europe and former Soviet Asia,
Latin America and the Caribbean, Asia and the Pacific, the Middle East and North Africa,
and Sub-Saharan Africa. In auxiliary models we test effects of a sub-regional measure of
diffusion (see our results below). We also considered the robustness of our models with
the addition of a measure of the global institutionalization of the international women’s
movement,65
but it does not substantively change our main results.66
Domestic Factors  Case-study, comparative, and cross-national researchers have identified
a wide range of national characteristics that may affect quota adoption.67
We include
four as controls. First, electoral systems are coded into four categories: PR, plurality-majority,
mixed-PR, and an “other” category that includes periods of one-party rule, coup
years, and other legislative interruptions. Plurality-majority systems are the reference
category. Second, we also differentiate all countries currently or historically MarxistLeninist
from those never under Marxist rule.68
Third, to test whether quota adoption is
more likely during post-conflict reconstruction,69
we include a time-varying measure of
armed conflict that begins three years prior to the first election after conflict has reached
high intensity (more than 1,000 battle deaths) and runs through the first post-conflict election.
Finally, we include women’s national legislative presence measured as the percentage
of women in the lower house of the national legislature, lagged one year. Note, however,
that many countries experience periods of legislative interruption due to coups or armed
conflict. We thus carry forward prior levels of women’s representation to fill gaps in the
data. Countries never electing a woman parliamentarian are assigned values of zero.
In auxiliary models, we also assess effects of economic development by including
measures of GDP per capita in constant 1990 USD, logged to reduce skew. Our primary
measure is from Maddison, who offers the best coverage of countries and time
points.70
However, economic development has no significant effects on quota adoption,
and, by including it, we lose a considerable number of observations. Therefore,
consistent with other research predicting quota adoption,71
we exclude it from our
models presented below.
Results
We begin by presenting histograms that visualize the percent of countries that adopted a
gender quota at each level of political rights and civil liberties. These figures help us to
identify whether countries tend to adopt gender quotas when they are at a particular level
of civil liberties or political rights.
11
Pär Zetterberg et al.
The upper histogram in Figure 1 focuses on the fifty-two countries that adopted candidate
quotas over our period of study, and the lower histogram focuses on the sixteen countries
that adopted reserved seats. The dimensions of democracy are coded such that large numbers
mean that civil liberties and political rights are well protected. Looking first at candidate
quotas, we see that countries with moderate levels of civil liberties are frequently adopting
quotas. Half of the countries that adopted a candidate quota scored either a four or a five on
the civil liberties scale, and nearly 80 percent of countries that adopted a candidate quota
Figure 1  The Distribution of Civil Liberties and Political Rights across Countries in
the Year Prior to Quota Adoption, by Quota Type
12
Comparative Politics January 2022
scored between three and six. However, we observe a different pattern for political rights,
where countries at low, medium, and high levels of political rights are adopting candidate
quotas. Turning to reserved seats in the lower histogram, what is most striking is the lack of
adoption of reserved seats in countries with even moderately high levels of democracy. We
also see differences when we compare civil liberties and political rights. Reserved seats are
especially common when political rights are low (i.e., poorly protected)—a score of two is
the most common—and when civil liberties are moderately low—a score of three is the most
common. Overall, these histograms illustrate clearly that even though civil liberties and political
rights are highly correlated with one another (r=.98), a country’s levels of civil liberties
and political rights during the run-up to quota adoption can be dissimilar.
Turning to multivariate analysis, Table 1 displays the first set of results, the baseline
model, in which we include the average measures of democracy and make no distinction
between quota types. Starting with Model 1, our discrete-time event history model
Table 1  Discrete Time Event History Models Predicting National Quota Adoption
with Average Level of Democracy
Model 1 Model 2
Average Democracy 20.08 0.68
(0.09) (0.43)
Average Democracy Squared 20.09t
(0.05)
Marx-Leninism 20.17 20.19
(0.32) (0.32)
PR Electoral System 1.08** 1.02**
(0.37) (0.37)
Mixed-PR Electoral System 0.73t
0.70
(0.43) (0.43)
Other Electoral System 0.55 0.78
(0.48) (0.50)
Logged Foreign Aid 0.58 0.31
(0.36) (0.40)
Recent Civil War 1.24*** 1.32***
(0.36) (0.36)
Regional Diffusion 0.04*** 0.04***
(0.01) (0.01)
% Women in Legislature 0.00 0.01
(0.01) (0.02)
Year 0.01 0.02
(0.03) (0.03)
Intercept 245.92 251.67***
(55.86) (56.12)
N-Country-Years 3014 3014
***p<0.001, **p<0.01, *p<0.05, tp<0.10, two-tailed tests; standard errors in parentheses.
13
Pär Zetterberg et al.
shows no linear relationship between democracy and national quota adoption. The average
democracy coefficient is substantively non-existent and statistically insignificant.
However, when we take the possibility of a non-linear relationship into consideration,
we find a significant relationship, at p<.10 level of significance (Model 2). Countries
at the middle of the democracy scale are more likely to adopt national quotas than advanced
democracies (highest numbers on the scale) or more full-fledged authoritarian
states (lowest numbers on the scale). Moving from one to four on the democracy scale
doubles the predicted probability that a country would adopt a quota in any given year
(from .50 percent to 1.00 percent; see Figure 2).72
Next, we evaluate whether the relationship between democracy and quota adoption
varies by dimension of democracy and across quota types. Models predicting the adoption
of candidate quotas appear in white, and those predicting the adoption of reserved
seats are shaded light grey. Models 3 and 4 in Table 2 demonstrate that the curvilinear
relationship between democracy and quota adoption generally holds across both quota
types. Reserved seats as well as candidate quotas are most frequently adopted in countries
that are neither fully authoritarian nor fully democratic. However, the relationship
with candidate quotas is only significant at p<.10 level of significance.
Disaggregating democracy into separate dimensions reveals important differences
in the factors that predict candidate quotas and reserved seats. Most notably, only
civil liberties significantly predict the adoption of candidate quotas (Models 5 and 6 in
Table 2). Political rights are not close to being statistically significant. For the adoption
of reserved seats, we observe (in Models 7 and 8 in Table 2) a slightly (but not entirely)
reversed pattern: mainly political rights have an effect on this quota type. The effect of
civil liberties on the adoption of reserved seats is substantively smaller, and, in contrast
to our other main findings, it is sensitive to model specification.73
Figure 2 helps us to visualize these different effects. Here, we convert coefficients
into the predicted probability of quota adoption in a given year. Whereas the solid line
shows the nonlinear, inverted U-effect of average democracy on the adoption of all
types of quotas (Model 2 in Table 1), here we concentrate on the dashed and dotted line.
Looking first at the dashed line, it shows the effect of the civil liberties dimension of
democracy on the probability of adopting a candidate quota (Model 6 in Table 2). While
still an inverted U, the peak shifts toward countries with higher levels of civil liberties.
The adoption of a candidate quota is more likely when a country has a moderately high
level of civil liberties. Moving to the dotted line, it displays the effect of the political
rights dimension of democracy on the probability of adopting a reserved seat quota
(Model 7 in Table 2). This graph shows a sharply different pattern. Lower levels of political
rights increase the probability of adopting reserved seats for women. Countries
in the midrange of political rights have a lower probability of adopting a reserved seat,
and there is zero probability of a country at a high level of political rights adopting a
reserved seat quota.
Taken together, our findings suggest that distinguishing among the dimensions of
democracy helps us better understand how quotas are adopted. Civil liberties generally
increase the probability of adopting candidate quotas, in a nonlinear manner, with a peak
14
Comparative Politics January 2022
Table 2  Disaggregating Dimensions of Democracy and Gender Quota Types
(Discrete Time Event History Models)
Model 3 Model 4 Model 5 Model 6 Model 7 Model 8
Candidate
Quotas
Reserved
Seats
Candidate
Quotas
Candidate
Quotas
Reserved
Seats
Reserved
Seats
Average
Democracy
0.85 2.87*
(0.52) (1.38)
Average
Democracy
Squared
0.10t
0.56*
(0.06) (0.24)
Political Rights 0.23 2.56*
(0.42) (1.16)
Political Rights
Squared
20.03 20.50*
(0.05) (0.21)
Civil Liberties 1.22* 2.06t
(0.59) (1.24)
Civil Liberties
Squared
20.12* 20.41*
(0.06) (0.21)
Controls
Marxist-Leninism 0.21 22.50* 0.19 0.25 22.48* 2.61t
(0.34) (1.18) (0.34) (0.34) (1.18) (1.19)
PR Electoral
System
1.55** 20.49 1.63*** 1.49** 20.40 0.34
(0.48) (0.88) (0.48) (0.48) (0.84) (0.87)
Mixed PR Electoral
System
1.37** [NCA] 1.41** 1.28* [NCA] [NCA]
(0.53) (0.53) (0.53)
Other Electoral
System
0.94 0.34 0.62 1.07 0.46 0.56
(0.71) (0.70) (0.70) (0.68) (0.73) (0.70)
Logged Foreign Aid 20.03 1.13* 20.01 20.03 1.10* 1.19*
(0.20) (0.55) (0.25) (0.19) (0.56) (0.56)
Recent Civil War 1.09* 2.39*** 1.00* 1.21** 2.34*** 2.11**
(0.45) (0.71) (0.45) (0.45) (0.68) (0.69)
Regional Diffusion 0.05*** 0.04 0.05*** 0.05*** 0.04 0.03
(0.01) (0.03) (0.01) (0.01) (0.03) (0.03)
% Women in
Legislature
0.00 0.06* 0.00 0.00 0.07* 0.06*
(0.02) (0.03) (0.02) (0.02) (0.03) (0.03)
Year 0.01 0.04 0.00 0.01 0.04 0.06
(0.03) (0.07) (0.02) (0.03) (0.07) (0.07)
Intercept 221.50 2117.21 12.05 223.60 2105.20 2151.07**
(63.09) (147.63) (34.70) (62.85) (149.00) (145.13)
N-Country-Years 3014 3014 3014 3014 3014 3014
***p<0.001, **p<0.01, *p<0.05, tp<0.10, two tailed tests; standard errors in parentheses;
“NCA”=No countries achieved (coefficient will not estimate).
15
Pär Zetterberg et al.
at higher levels of civil liberties. Increasing political rights suppress adoption of reserved
seat quotas. Countries with lower levels of political rights have the greatest probability
of adopting a reserved seat quota. Building on our theoretical expectations, these findings
suggest that the trajectories for (or forces behind) the adoption of candidate quotas
and reserved seats are different.
Although our focus is on effects of democracy, it is also important to look briefly
at our controls, which provide additional evidence of different trajectories for quota
adoption. Perhaps most important for our theoretical argument is the impact of foreign
aid, which is an important driver of reserved seats, but not of candidate quotas (see
Table 2, Models 5–8). This result is in line with our suggestion that countries that want
to secure international benefits are more likely to adopt reserved seats. Similarly, the
finding displayed in Table 2 that regional diffusion matters for candidate quotas but not
for reserved seats also resonates well with our theoretical expectations. As mentioned
above, women’s movements in many Latin American countries— where civil liberties
such as freedom of association are protected—learned from the experiences ofArgentina
and its pioneering candidate quota law. They pushed their governments to adopt similar
quota policies. In addition to these two findings, the results for the other controls show
that electoral systems are important drivers of candidate quotas,74
Marxist-Leninist history
and women’s share of seats in the legislature matter for reserved seats, and recent
civil war increases the likelihood of adoption for both quotas (although the effects for
reserved seats are much stronger).75
The significant effect of women’s representation
in the legislature on reserved seats suggests that women MPs can play a role in quota
adoption in semi-authoritarian regimes.
Figure 2  Relationship between Democracy and Quota Adoption (from Discrete Time
Event History Models)
16
Comparative Politics January 2022
Conclusion
This article has contributed to scholarship on gender quota adoption by theorizing and
empirically investigating how a country’s level of democracy relates to the adoption of
different quota types. Our results corroborate previous empirical research that there is
a curvilinear, inverted U-shaped, relationship between democracy and quota adoption.
However, taking a closer look at the group of countries at the middle of the democracy
scale, this analysis qualifies these findings in several ways. Most importantly, we show
that different quota types are related to different dimensions of democracy: the adoption
of a candidate quota is more likely when a country’s citizens are guaranteed certain civil
liberties such as freedoms of association, whereas reserved seats are most commonly
adopted by countries with relatively low levels of electoral contestation.
These findings are in line with our theoretical expectations, which suggested that
different levels and dimensions of democracy produce different incentives and constraints
on key actors in processes of quota reforms. For instance, civil liberties such as freedom
of organization are crucial for the “bottom-up” processes that research commonly has
shown characterize the adoption of candidate quotas. Case studies on countries and regions
such as Mexico76
and the Balkans77
(e.g., North Macedonia) have highlighted the
connections between domestic grassroots movements and politicians as important for
candidate quota adoption. These studies bolster our findings by demonstrating that a fair
amount of civil liberties are needed for activists to openly meet, learn from each other,
and criticize, cooperate, and influence those in power. This latter point is important because
the candidate quota type is likely to pose more of a challenge to the existing system,
including to incumbents. Thus, a certain level of democracy measured as civil liberties
creates the space needed for these kinds of reform demands to emerge and be successful.
There are also several case studies, on countries such as Morocco,78
Cameroon,79
Rwanda,80
Uganda,81
and Jordan,82
that support our other main finding and illustrate
why countries where electoral contestation is low commonly adopt reserved seats. Such
countries often have a dominant political actor that is unlikely to pass any form of legislation
that could threaten the regime’s electoral dominance. Therefore, in countries
with low political rights, dominant parties set the limits for women’s rights advocacy
(including for quota adoption advocacy). These parties have large leeway to respond to
demands (e.g., by international organizations) and adopt quotas if and only if the quota
reform serves their purposes, either to increase international legitimacy83
or by designing
them in a manner that electorally favors the ruling party.84
Thus, our study demonstrates that the curvilinear relationship found between level
of democracy and quota adoption hides some important differences. With a more nuanced
analysis, distinguishing both between quota types and between dimensions of democracy,
we add to those who have pointed at the multiple causes of quota adoption by
proposing two substantially different processes of quota adoption.85
Pushing our argument
further, our analysis raises the question whether quotas really should be perceived
of as one type of electoral reform. As already noted, candidate quotas and reserved seats
are fundamentally different in design, apply to different stages of the election process,
17
Pär Zetterberg et al.
and affect political parties and parliament in different ways. They also seem to be pushed
through by partly different actors, with different agendas, and, as our results show, in
different political systems. Therefore, it is important to separate reserved seats and candidate
quotas in large-scale, global assessments of gender quotas, for instance, their
effects on women’s substantive and symbolic representation.86
One implication of this work is that scholars should break down both quotas and
democracy more than has been typical. A lot of prior quantitative cross-national research
has focused on gender quotas generally.87
Differentiating candidate and reserved seats
quotas bore fruit for our analysis. Future research might similarly benefit by making
distinctions within reserved seat quotas or between candidate quotas with and without
strengthening features like placement mandates. Consider too the distinction between
the adoption of a gender quota and its implementation. Adopted quotas have sometimes
been challenged and removed, or simply abandoned.
Our argument and findings imply similar benefits to considering more fine-grained
understandings of the dimensions of democracy. Scholars of democracy conceptualize a
variety of dimensions of democracy that, in some cases, go beyond the two we have considered
here (see endnote 14). For example, the Varieties of Democracy project (V-Dem)
distinguishes between multiple “principles” of democracy including electoral, participatory,
deliberative, and egalitarian. Since our findings suggest that distinguishing among
two dimensions of democracy helps us better understand quotas, future research could
consider whether quota types relate to other dimensions of democracy in even more nuanced
ways. So too might additional research consider democracy as moments of transition,
rather than the continuous, dimensional construct we have theorized and modeled
here. It is possible that specific democratic transition points open unique opportunities for
the adoption of quotas, something that scholars can assess in the future.88
Another possible extension of our research is to analyze the exceptions to the broader
patterns we observe. For instance, although we show that it is less likely that countries
that are more democratic adopt reserved seats, we know that a few of them have done
so (e.g., India, at the local level). We theorize that acknowledging multiple actors, such
as political elites, civil society, and international actors, as well as the particular constellation
they take in a country, will yield clues as to the types of quotas considered for
adoption. Therefore, we suggest that case studies using process-tracing would further
understanding of the motives for selecting certain quota types and contribute to a more
complete picture of the broader relationship between democracy and quota adoption.
Finally, on a broader note, we call on scholars of (electoral) authoritarian regimes
to increasingly include gender dynamics in their analyses.89
Our global analysis
of a wide-reaching contemporary electoral reform provides support to studies
that have suggested that reserved seat adoption fits the agenda of authoritarian lead-
ers.90
In other words, there is increasing evidence that authoritarian regimes that
hold (multiparty) elections not only respond to pressures for quota adoption but
also make strategic use of certain gender equality policies to “mimic the appearance
of liberal democracy”91
in order to avoid receiving pressures for democratization.
The analysis of gender quota adoption suggests that an increased focus on these
18
Comparative Politics January 2022
strategies is warranted. They deserve the attention not only of scholars but also of
practitioners working with democracy promotion and with the reduction of gender
inequalities in politics.
NOTES
Pär Zetterberg’s and Elin Bjarnegård’s contributions to this article have been possible thanks to support from
the Swedish Research Council (grant no. 2015–00955). The authors thank participants at the 2017 European
Conference on Politics and Gender (ECPG), the 2017 annual meeting of the American Political Science Association
(APSA), and the Uppsala Gender and Politics seminar (@uppgap) for constructive feedback on this
project. They offer a special thanks to Amanda Clayton, Alice Kang, and the anonymous reviewers for very
helpful comments and suggestions.
1. Melanie M. Hughes, Pamela Paxton, and Mona Lena Krook, “Gender Quotas for Legislatures and Corporate
Boards,” Annual Review of Sociology, 43 (July 2017), 331–52.
2. Drude Dahlerup, ed., Women, Quotas and Politics (New York: Routledge, 2006); Mona Lena Krook,
Quotas for Women in Politics: Gender and Candidate Selection Reform Worldwide (New York: Oxford University
Press, 2009).
3. Rainbow Murray, Mona Lena Krook, and Katherine A. R. Opello, “Why Are Gender Quotas Adopted?
Party Pragmatism and Parity in France, “ Political Research Quarterly, 65 (September 2012), 529–43; Tania
Verge, “Institutionalising Gender Equality in Spain: From Party Quotas to Electoral Gender Quotas,” West
European Politics, 35 (February 2012), 395–414.
4. Lisa Baldez, “Elected Bodies: The Gender Quota Law for Legislative Candidates in Mexico,” Legislative
Studies Quarterly, 29 (May 2004), 231–58.
5. Gihan Abou-Zeid, “The Arab Region: Women’s Access to the Decision-Making Process Across the Arab
Nation,” in Dahlerup, ed., 2006, 168–93.
6. Melanie M Hughes, Mona Lena Krook, and Pamela Paxton, “Transnational Women’s Activism and the
Global Diffusion of Gender Quotas,” International Studies Quarterly, 59 (June 2015), 357–72.
7. Dahlerup, 2006.
8. Sarah S. Bush, “International Politics and the Spread of Quotas for Women in Legislatures,” International
Organization, 65 (Winter 2011), 103–37; Sarah S. Bush and Eleanor Gao, “Small Tribes, Big Gains: The Strategic
Uses of Gender Quotas in the Middle East,” Comparative Politics, 49 (January 2017), 149–67; Melinda
Adams, “‘National Machineries’and Authoritarian Politics,” International Feminist Journal of Politics, 9 (May
2007), 176–97; Amanda B. Edgell, “Foreign Aid, Democracy, and Gender Quota Laws,” Democratization, 24
(September 2017), 1103–41; Melody E. Valdini, The Inclusion Calculation: Why Men Appropriate Women’s
Representation (New York: Oxford University Press, 2019); Sarah S. Bush and Pär Zetterberg, “Gender Quotas
and International Reputation,” American Journal of Political Science, 65 (April 2021), 326–41; Abou-Zeid.
9. Aili Mari Tripp, Women and Power in Postconflict Africa (Cambridge: Cambridge University Press,
2015); Melanie M. Hughes and Aili Mari Tripp, “Civil War and Trajectories of Change in Women’s Political
Representation in Africa, 1985–2020,” Social Forces, 93 (June 2015), 1415–540.
10. For exceptions, see e.g. Gretchen Bauer and Jennie E. Burnet, “Gender Quotas, Democracy, and Women’s
Representation in Africa: Some Insights from Democratic Botswana and Autocratic Rwanda,” Women’s
Studies International Forum, 41 (November-December 2013), 103–12; Mi Yung Yoon, “Special Seats for
Women in Parliament and Democratization: The Case of Tanzania,” Women’s Studies International Forum, 41
(November-December 2013), 143–49.
11. Hughes, Krook, and Paxton.
12. Dahlerup, 2006.
13. Dahlerup, 2006; Krook, 2009; Susan Franceschet, Mona Lena Krook, and Jennifer M. Piscopo, eds.,
The Impact of Gender Quotas (New York: Oxford University Press, 2012).
14. There is, of course, no scholarly agreement on a single definition or measure of democracy, neither
among theorists nor among comparative politics scholars working on democracy and democratization. However,
Dahl’s distiction between contestation and inclusiveness has been hugely influential (Robert A. Dahl,
Democracy and Its Critics (New Haven: Yale University Press, 1989), ch. 15). The two dimensions we include,
fully contested free and fair elections and guarantees of civil liberties, hew closely to this original conception-
19
Pär Zetterberg et al.
alization and are commonly used as baseline attributes in definitions of democracy (David Collier and Steven
Levitsky, “Democracy with Adjectives: Conceptual Innovation in Comparative Research,” World Politics, 49
(April 1997), 434. See also Steven Levitsky and Lucan A. Way, Competitive Authoritarianism: Hybrid Regimes
After the Cold War (New York: Cambridge University Press, 2010) and Michael Coppedge, Angel Alvarez, and
Claudia Maldonado, “Two Persistent Dimensions of Democracy: Contestation and Inclusiveness,” The Journal
of Politics, 70 (July 2008), 632–47). Other definitions may exclude a dimension (e.g., Adam Przeworski’s
minimalist focus on contestation alone, in Adam Przeworski, R. Michael Alvarez, Michael E. Alvarez, Jose
Antonio Cheibub, Fernando Limongi, and Fernando Papaterra Limongi Neto, Democracy and Development:
Political Institutions and Well-Being in the World, 1950–1990 (Cambridge: Cambridge University Press, 2000))
or increase dimensions (e.g., Michael Coppedge, John Gerring, David Altman, Michael Bernhard, Steven
Fish, Allen Hicken, Matthew Kroenig, Staffan I. Lindberg, Kelly McMann, Pamela Paxton, Holli A. Semetko,
Svend-Erik Skaaning, Jeffrey Staton, and Jan Teorell, “Conceptualizing and Measuring Democracy: A New
Approach,” Perspectives on Politics, 9 (June 2011), 247–67) but all must ultimately reckon with the fundamental
Dahlian differentiation between contestation and inclusiveness.
15. Melanie M. Hughes, Pamela Paxton, Amanda B. Clayton, and Pär Zetterberg, “Global Gender Quota
Adoption, Implementation, and Reform,” Comparative Politics, 51 (January 2019), 219–38.
16. Hughes, Krook, and Paxton.
17. C.f. Jennier Gandhi and Ellen Lust-Okar, “Elections Under Authoritarianism,” Annual Review of
Political Science, 12 (August 2009), 403–22.
18. In some cases, quotas apply to the level of aspirants. However, this quota type is quite rare and is not
included in our analysis. Dahlerup, 2006.
19. Elin Bjarnegård and Pär Zetterberg, “Removing Quotas, Maintaining Representation: Overcoming
Gender Inequalities in Political Party Recruitment,” Representation, 47 (July 2011), 187–99.
20. Hughes, Paxton, Clayton, and Zetterberg.
21. Krook, 2009.
22. Dahlerup, 2006; Krook, 2009; Ragnhild Muriaas, Liv Tønnessen, and Vibeke Wang, “Exploring the
Relationship between Democratization and Quota Policies in Africa,” Women’s Studies International Forum,
41 (November-December 2013), 89–93.
23. Franceschet, Krook, and Piscopo; Hughes, Paxton, Clayton, and Zetterberg.
24. Drude Dahlerup, “Electoral Gender Quotas: Between Equality of Opportunity and Equality Of Result,”
Representation, 43 (July 2007), 73–92; Bauer and Burnet; Mona Lena Krook, “Gender Quotas and Democracy:
Insights from Africa and Beyond,” Women’s Studies International Forum, 41 (November-December
2013), 160–63.
25. As for another quota type, voluntary quotas, Dahlerup shows that they are more common in democratic
countries: in 39 percent of these countries at least one of the political parties had included quotas in their party
statutes. In semi-democratic and authoritarian countries, the number was substantially lower: 16 percent and
18 percent, respectively. Dahlerup, 2007.
26. Hughes, Krook. and Paxton.
27. Ibid.
28. Bush; Hughes and Tripp; Tripp.
29. See e.g. Collier and Levitsky; Levitsky and Way; Dahl.
30. Alice J. Kang and Aili Mari Tripp, “Coalitions Matter: Citizenship, Women, and Quota Adoption in
Africa,” Perspectives on Politics, 16 (March 2018), 73–91.
31. Hughes, Krook, and Paxton.
32. Ann Towns, Women and States: Norms and Hierarchies in International Society (Cambridge: Cambridge
University Press, 2010).
33. Bush.
34. Bjarnegård and Zetterberg, 2011.
35. Mala Htun and Mark P. Jones, “Engendering the Right to Participate in Decision-Making: Electoral
Quotas and Women’s Leadership in Latin America,” in Nikki Craske and Maxine Molyneux, eds., Gender and
the Politics of Rights and Democracy in Latin America (Basingstoke: Palgrave, 2002), 32–56; Mona Lena
Krook, “Gender Quotas as a Global Phenomenon: Actors and Strategies in Quota Adoption,” European Political
Science, 3 (June 2004), 59–65; Karen Celis, Mona Lena Krook, and Petra Meier, “The Rise of Gender
Quota Laws: Expanding the Spectrum of Determinants for Electoral Reform,” West European Politics, 34
(May 2011), 514–30; Bauer and Burnet.
36. Maria José Lubertino, “Pioneering Quotas: The Argentine Experience and Beyond,” The Implementation
of Quotas: Latin American Experiences, Lima, Peru, February 23–24, 2003, International IDEA.
20
Comparative Politics January 2022
37. Adriana Piatti-Crocker, ed., Diffusion of Gender Quotas in Latin America and Beyond: Advances and
Setbacks in the Last Two Decades (New York: Peter Lang, 2011).
38. For an introduction to electoral––or competitive––authoritarianism, see e.g. Lucan A. Way and Steven
Levitsky, “Elections Without Democracy: The Rise of Competitive Authoritarianism,” Journal of Democracy,
13 (April 2002), 51–65; Andreas Schedler, ed., Electoral Authoritarianism. The Dynamics of Unfree Competition
(Boulder: Lynne Rienner, 2006); Gandhi and Lust-Okar.
39. See e.g. Abou-Zeid; Bush and Gao; Adams.
40. See also Bush; Towns.
41. Towns.
42. Bush.
43. See e.g. Bush; Daniela Donno and Anne-Kathrin Kreft, “Authoritarian Institutions and Women’s
Rights,” Comparative Political Studies, 52 (April 2019), 720–53.
44. Elin Bjarnegård and Pär Zetterberg, “Gender Equality Reforms on an Uneven Playing Field: Candidate
Selection and Quota Implementation in Electoral Authoritarian Tanzania,” Government and Opposition, 51
(July 2016), 464–86.
45. Hughes, Krook, and Paxton.
46. Our argument does not preclude the possibility of exceptions from our suggested broader pattern. Individual
cases, with unique histories, constellations of political actors, etc. could behave differently than what the theory
suggests. Exceptions and outliers to any patterns we observe are fruitful places to focus additional research.
47. See e.g. Krook, 2009.
48. See e.g. Ana Catalano Weeks, “Why are Gender Quota Laws Adopted by Men? The Role of Inter- and
Intra-Party Competition,” Comparative Political Studies, 62 (December 2018), 1935–73.
49. Guillaume Fréchette, Francois Maniquet, and Massimo Morelli, “Incumbents’ Interests and Gender
Quotas,” American Journal of Political Science, 52 (October 2008), 891–909.
50. Janet Box-Steffensmeier and Bradford S. Jones, Event History Modeling: A Guide for Social Scientists
(New York: Cambridge University Press, 2004).
51. Nathaniel Beck, Jonathan N. Katz, and Richard Tucker, “Taking Time Seriously: Time-Series––CrossSection
Analysis with a Binary Dependent Variable,” American Journal of Political Science, 42 (October
1998), 1260–88.
52. Kazuo Yamaguchi, Event History Analysis (Newbury Park: Sage, 1991).
53. Hughes, Paxton, Clayton, and Zetterberg.
54. Hughes, Krook, and Paxton.
55. In robustness tests, we also excluded these hybrid cases. Removing these cases from the analysis does
not substantively change our results.
56. Hughes, Paxton, and Krook.
57. Unfortunately, the QAROT data set does not include quotas that specific political parties have voluntarily
adopted. Arguably, the adoption of party quotas by the dominant party in a one-party-dominant state (e.g.,
the African National Congress in South Africa) may produce a similar effect as a national quota law. However,
such cases are relatively rare. “Gender Quotas Database,” International IDEA (2000), https://www.idea.int/
data-tools/data/gender-quotas/. See also Mona Lena Krook, “Candidate Gender Quotas: A Framework for
Analysis,” European Journal of Political Science, 46 (May 2007), 367–94.
58. Freedom House (2020), https://freedomhouse.org/reports/freedom-world/freedom-world-research-
methodology
59. Ibid.
60.AlthoughwebelievethatFreedomHouse’smeasurementsofdemocracybestcaptureourdefinitionofdemocracy,
we acknowledge that this operationalization is not free from criticism. For instance, Freedom House’s measurements
include also non-procedural aspects of democracy, such as liberties related to socio-economic issues. Thus, its
index is more maximalist than Dahl’s definition of democracy. Gerardo L. Munck and JayVerkuilen, “Conceptualizing
and Measuring Democracy: EvaluatingAlternative Indices,” Comparative Political Studies, 35 (February 2002),
5–34. In addition, scholars have suggested that “political rights” and “civil liberties” are––empirically speaking––
too similar and should not be used as indicators of distinct aspects of democracy. Michael Coppedge,AngelAlvarez,
and Claudia Maldonado, “Two Persistent Dimensions of Democracy: Contestation and Inclusiveness,” Journal of
Politics, 70 (July 2008), 632–47. While we take criticisms of Freedom House seriously, we argue that it mainly
makes our test harder: if the critique holds, we should not expect to find any differences across the two dimensions of
democracy. It should also be added that other scholars have found these multiple dimensions to be appropriate. KennethA.
Bollen and Pamela Paxton, “Subjective Measures of Liberal Democracy,” Comparative Political Studies, 33
(February 2000), 58–86.
21
Pär Zetterberg et al.
61. To run such models is important because some scholars have found that different measures of democracy
do not correlate as much as we would expect. Thus, there is a risk that different operationalizations of democracy
yield different results. See e.g. Edward McMahon and Emilie Kornheiser, “Assessing the Assessors:
Correlating Democracy Methodologies,” Social Indicators Research, 97 (2010), 269–77.
62. Bush; Edgell; Liam Swiss and Kathleen M. Fallon, “Women’s Transnational Activism, Norm
Cascades, and Quota Adoption in The Developing World,” Politics & Gender, 13 (September 2017), 458–87;
Miriam J. Anderson and Liam Swiss, “Peace Accords and the Adoption of Electoral Quotas for Women in the
Developing World, 1990–2006,” Politics & Gender, 10 (March 2014), 33–61.
63. World Bank (2017), https://data.worldbank.org/indicator/DT.ODA.ODAT.CD.
64. Bush; Swiss and Fallon.
65. Hughes, Krook, and Paxton.
66. Our measure of the institutionalization of the international women’s movement combines three worldlevel
indicators: (1) cumulative foundings of WINGOs; (2) the cumulative count of international conferences,
UN treaties, and UN groups related to women; and (3) yearly resources of United Nations Development Fund
for Women (UNIFEM). We combine the three indicators by estimating values (factor scores) of the underlying
latent variable—the global institutionalization of the women’s movement—using confirmatory factor analysis.
67. See Hughes, Paxton and Krook for a review.
68. We update data from Pamela Paxton, Melanie M. Hughes, and Jennifer L. Green, “The International
Women’s Movement and Women’s Political Representation, 1893–2003,” American Sociological Review, 71
(December 2006), 898–920.
69. Mona Tajali, “Gender Quota Adoption in Postconflict Contexts: An Analysis of Actors and Factors Involved,”
Journal of Women, Politics & Policy, 34 (July 2013), 261–85; Gretchen Bauer and Hannah E. Britton,
Women in African Parliaments (Boulder: Lynne Rienner Publishers, 2006); Tripp.
70. Angus Maddison, “Statistics on World Population, GDP, and Per Capita GDP, 1–2008 AD,” (2010),
retrieved November 17, 2011, from http://ggdc.net/MADDISON/oriindex.htm.
71. Bush.
72. Auxiliary models with Polity’s democracy measure shows a similar curvilinear pattern. However, in
contrast to Freedom House’s average democracy score, the Polity democracy index is not statistically significant
with the adoption of electoral gender quotas. It is only in some auxiliary models that the nonlinear effect
reached statistical significance (at p<.10, i.e. the same level of significance as the one presented in Model 2 in
Table 1). Plots of the curve show that quota adoption peaks when the Polity democracy index is 3 (on a -10
to 10 scale, where 7 and above usually is characterized as “democratic”), and then declines, although not as
sharply as we observe using the Freedom House average democracy score. Full results from these auxiliary
models are available upon request. Polity does not provide separate measures of civil liberties or political
rights, so we are unable to replicate our remaining models.
73. For example, adding GDP per capita to the model or removing major armed conflict or Marxist-Leninist
history both produce results with no statistically significant effects of civil liberties on reserved seat
adoption. These same changes do not disrupt the statistical significance of our other main findings (auxiliary
models are available upon request).
74. See also Mark P. Jones, “Gender Quotas, Electoral Laws, and the Election of Women: Evidence from
the Latin American Vanguard,” Comparative Political Studies, 42 (January 2009), 56–81.
75. Notably, auxiliary models (available upon request) show that diffusion has a positive and statistically
significant relationship with quota adoption for both types when diffusion is measured at the subregional level,
suggesting that the diffusion of reserved seat quotas may not reach as far as diffusion effects for candidate
quotas.
76. Baldez.
77. Milica G. Antic and Sonja Lokar, “The Balkans: From Total Rejection to Gradual Acceptance of Gender
Quotas,” in Dahlerup, ed., 2006, 138–67.
78. James N. Sater, “Reserved Seats, Patriarchy, and Patronage in Morocco,” in Franceschet, Krook, and
Piscopo, eds., 72–86.
79. Adams.
80. Jennie E. Burnet, “Gender Balance and the Meanings of Women in Governance in Post-Genocide
Rwanda,” African Affairs, 107 (May 2008), 361–86.
81. Anne Marie Goetz, “No Shortcuts to Power: Constraints on Women’s Political Effectiveness in
Uganda,” The Journal of Modern African Studies, 40 (December 2002), 549–75.
82. Bush and Gao.
83. E.g. Abou-Zeid; Bush.
22
Comparative Politics January 2022
84. Ragnhild Muriaas and Vibeke Wang, “Executive Dominance and the Politics of Quota Representation
in Uganda,” The Journal of Modern African Studies, 50 (June 2012), 309–38; Donno and Kreft.
85. See e.g. Krook, 2009.
86. C.f. Amanda Clayton and Pär Zetterberg, “Quota Shocks: Electoral Gender Quotas and Government
Spending Priorities Worldwide,” Journal of Politics, 80 (July 2018), 916–32.
87. E.g. Hughes, Krook, and Paxton; Clayton and Zetterberg; Edgell.
88. As an additional extension to such a discussion, future empirical work could look closer at what Krook,
Lovenduski, and Squires refer to as “political citizenship models,” which they argue help explain variation in
the type of (if any) quota adopted. Mona Lena Krook, Joni Lovenduski, and Judith Squires, “Gender Quotas
and Models of Political Citizenship,” British Journal of Political Science, 39 (October 2009), 781–803.
89. See also Gandhi and Lust-Okar.
90. See e.g. Abou-Zeid; Adams; Muriaas and Wang; Sater; Bjarnegård and Zetterberg, 2016; Bush and
Gao; Donno and Kreft.
91. Bush and Gao, 2017, 150.