On-line version ISSN 1981-3821
Braz. political sci. rev. (Online) vol.3 no.se Rio de Janeiro 2008
Women in upper houses: a global perspective*
Brazilian Centre for Analysis and Planning (CEBRAP), Brazil
Replicated from Brazilian Political Science Review (Online), Rio de Janeiro, v.2, n.1, Jan/Jun. 2008.
In analyses of female representation in lower houses, the adoption of quotas, the electoral system and the religious aspect have been identified as the main explanatory variables. In the case of upper houses,1 I see a relationship between their political strength and women's presence in them: when they are weak, the presence of female representatives tends to be larger; when they are strong, women's presence is smaller. Furthermore, the article shows that an analysis based solely on the number of seats held by women is insufficient for one to gather the true dimension of their participation in politics. Evaluating the role and expressiveness of the institutions in which they are present is also necessary.
Keywords: Gender and electoral participation; Upper houses; Senate; Legislative; Women representation; Quotas.
In spite of being half the world's population, women occupy a very small space in politics: 3.7% of heads of government (president or prime minister), 9% of UN ambassadors, 7% of ministers and 8% of mayors throughout the world. Since 1990, only 30 women have become their country's main political leader (Paxton and Hughes 2007, 1, 80).
On average, they hold 17% of parliamentary seats all over the world. Although this percentage appears considerably low, women have been increasing their participation in parliament, to the tune of 50% compared with 1996. In recent years, this has led to advances in the literature with respect to this theme. One problem is that authors have devoted themselves entirely to the study of lower houses, leaving aside the study of upper houses. This seems to be an error of judgment, since upper houses have their peculiarities and display great variations, both in comparison with lower houses and among themselves. This article aims to contribute to filling this gap, by evaluating the variation in the number of women in the upper houses of bicameral countries considered minimally democratic by Freedom House.2
Often, upper houses differ in their roles, in relation to their powers, with respect to the number of members, the minimum age for members to hold office, the duration of their term, their form of selection. This diversity gives them a special capacity to adapt to different political systems. Hence, this allows these institutions to be characterised as "protean," a term used by Tsebelis and Money (1997), which refers to the Greek god Proteus, an old, prophetic man who lived in the sea and constantly changed his shape, and also had the capacity to foresee the future. Upper houses present similar numbers to those of lower houses with regard to women's representation. This happens, for example, in South Africa, Belgium, Burundi, Namibia, Grenada etc. In other countries, we find much lower percentages than those found for lower houses, as in the case of Tunisia, Poland and the Dominican Republic. In others still, the percentage of women is much higher in the upper houses: St. Lucia, Belize, Malaysia, Swaziland, Barbados and the Bahamas.
When comparing upper houses among themselves, we find cases ranging from female representation at or near zero Palau (0%), Yemen (1.8%), Morocco (1.1%), Algeria (3.1%) all the way to near-parity Argentina (43.1%), the Bahamas (43.8%). We find relatively high percentages both in developed countries (Australia, Canada, Austria) and in developing countries (Rwanda, Belarus, Namibia). There are also developed countries such as the United States, Japan and France with percentages below the world average of 18.3%.
In light of such meaningful variations, it is probable that the determinants of the number of women in upper houses are not the same as those for lower houses. For the latter, the literature has presented as the main explanatory variables the adoption of quotas, the electoral system, the level of economic development and historical, cultural and religious characteristics. In this article, I seek to evaluate the importance of the institution's political strength in defining the number of female senators controlling for the variables traditionally used. My main hypothesis is that the percentage of women in upper houses is inversely related to the extent of their powers: in weak upper houses, one finds a larger proportion of women; in strong ones, the proportion is smaller. In other words, the greater the power of the upper house, the lower is the percentage of women in it.
In the following section, I present descriptive statistics on female participation in the two legislative houses in groups of specific countries. Then, in section III, I comment on the independent variables that will be used. In section IV, I present the results of the regression model (Heckman), considering not only bicameral but also unicameral countries. Lastly, I conclude by suggesting new research agendas.
Distribution of Women in Legislatures
In the table below, I show the average percentage of women and the respective standard deviation in the two legislative houses per group of countries.
The data display important differences in relation to those presented by the Inter-Parliamentary Union (IPU) and followed by several authors, like Paxton and Hughes (2007), Norris and Inglehart (2000) because I take into consideration not only geographic/territorial criteria, but also historical, political and institutional influences and characteristics.3 For example, I do not think it makes sense to group together the whole of the American continent, mixing historical experiences as different as those of the Caribbean and Latin America. It is also important to group together the countries of Central and Eastern Europe that belonged to the former communist bloc, the most generalized and long-running experience in the application of quotas to reserve parliamentary seats (Araújo 2001).4
As may be observed, women's representation is low in all the groups of countries, with slightly higher figures in the industrialised countries and lower figures in Asian and Middle Eastern countries, in both houses. If we look exclusively at the numbers for lower houses, we see that African, Caribbean and Central and Eastern European countries are closer to the world average, which is 16.2%.5
Based on a simultaneously vertical and horizontal reading, we note that the variation does not take place along the same lines: while in the industrialised countries there is little difference between the two legislative houses, in the Caribbean, upper houses have a higher percentage of women and lower houses display a percentage below the overall average. In Latin American, African and Asian countries, it is the lower houses that have higher percentages of women.
It is worth reminding ourselves that Caribbean countries followed the institutional design of their colonizers (Great Britain), with parliamentary regimes and upper houses with few powers. On the other hand, Latin American countries firmly adopted the US model, i.e., the presidential system and upper houses with wide-ranging powers (Neiva 2006).
I am not aware of any specific study about women's participation in upper houses. Rush (2001, 33) deals with the theme superficially, suggesting that the male presence in them is greater, in comparison with lower houses. In one of the first studies on the US Senate, Matthews (1959) spoke of a men's club, whose members had close relations and were in tune with one another, where consensus-building and decision-making were made easier and commitments tended to last.6 Although many changes have taken place in the world's most studied Senate (Sinclair 1989), until 1991 women had never occupied more than 2% of its seats, and they remain a clear minority in this legislative house (16% of its members). It seems that the influence of the US Senate on its Latin American congeners was not limited to the institutional configuration, but extended to this aspect as well. With the exception of Argentina, Latin American senates also contain low percentages of women: Bolivia (4%), Chile (5%), Paraguay (9%), Uruguay (10%), Colombia (12%), Brazil (12%).
Figure 1 shows more clearly the supposed relationship between powers of upper houses and the respective percentages of women in each of them .
Though rather dispersed, there seems to be an inverse relationship between the two variables: the percentage of women in the upper house tends to rise as its power diminishes.7 This notwithstanding, there is no single cause that can sustain itself in all places and at all times. For one to gain more accurate information regarding this element, it is important to control for other characteristics that may influence women's participation.
Presentation of Explanatory Variables
Several studies have demonstrated that demographic, cultural and socioeconomic factors interact with politico-institutional characteristics in defining women's participation in politics. Of the latter, the adoption of quotas is one among those displaying most association with the percentage of women in lower houses (Escobar-Lemmon and Taylor-Robinson 2006; Norris 1996; Jones 1998; Dahlerup 2003; Caul 2001; Jones 1996, 1998; Steininger 2000; Norris 2006; Avelar 2006). They are being used by more than half the countries, as a direct and immediate means of increasing this participation, and take three main forms: 1) constitutional rules that reserve a certain number of seats for women;8 2) legal demands for parties to select a given percentage of women candidates;9 and 3) measures taken voluntarily by parties along these lines.
Another important institutional variable in the representation of minorities is the electoral system. There exists a reasonable consensus that systems based on proportionality are more effective at electing women than those based on majority rules (Reynolds 1999; Norris 2004, 2006; Paxton and Hughes 2007; Matland 1998; Kenworth and Malami 1999; Kittilson 2006). One of the explanations for this is that systems with proportional representation display greater district magnitude, which leads to greater party magnitude. Both are important because they affect the party's strategy when choosing its candidates; when the district magnitude is equal to one, as is the case in almost all plurality/majority systems, women candidates must compete directly with men, who often control the party structure. As district magnitude increases, the party obtains the possibility of dividing the candidate list among various interests, including women's. Besides attracting more votes, it is a way of maintaining good relations inside the party and ensuring support for different factions.10
In a study on the socioeconomic status of parliamentarians, Rush (2001, 29-31) found that the system of proportional representation results in the election of more women, both for lower and upper houses. However, the author warns that the situation is less clear for the latter, since some do not have their members elected directly.
Research shows that women's political representation is also influenced by religion, being smaller in Islamic countries, though Catholic countries are not far behind (Reynolds 1999; Matland and Taylor 1997; Paxton 1997; Norris and Inglehart 2000; Paxton and Hughes 2007, 53). In some Arab countries, where almost all the population is of the Muslim faith, women are forbidden from occupying parliamentary seats.11 On the other hand, specialists state that predominantly Protestant countries make women's education easier, promote non-hierarchical religious practices and accept the religious leadership of women (Paxton and Hughes 2007, 111). I will test the individual impact of each of these three main religions, which will assume the value "1" when it is the country's predominant religion, i.e., when over 50% of the population profess that faith, and the value "0" when the opposite happens.
The level of socioeconomic development has been mentioned as an element that influences women's participation and the promotion of cultural values favourable to gender equality (Norris and Inglehart 2000; Avelar 2001; Matland 1998; Paxton and Hughes 2007, 131). On the other hand, Hassim (2006, 2) states that the link between socioeconomic development and political representation is tenuous, since institutional barriers and the political culture cancel out the gains obtained on the basis of socioeconomic status.12 Along the same lines, Miguel (2006) found that female candidates' performance in three Brazilian municipal elections was significantly better in more backward, less industrialised regions with lower schooling levels and worse social indicators.
In order to measure socioeconomic development, I will use the Human Development Index (HDI). It considers life expectancy, literacy levels, educational levels and Gross Domestic Product (GDP).13
There is disagreement within the literature as to the importance of the variable democracy level in the country. Some authors state that its relationship with women's representation is insignificant, weak or negative (Kenworthy and Malami, 1999; Paxton and Kunovich 2003; Paxton 1997). Others find a positive effect (Reynolds 1999; Lindberg 2004; Sweeney 2004). As a measure of the democracy level, I will use the indexes made available by Freedom House, which are widely used by political scientists.
In the regression model that has as the dependent variable the percentage of women in the lower house, all the variables displayed the expected sign. The variables which refer to the constitutional quota and to the adoption of the proportional electoral system displayed 1% statistical significance. The fact that the country is mostly Muslim impacts negatively on the number of representatives (with 5% significance) and the fact of being Protestant has a positive impact (with 1% significance). The coefficient of the HDI variable displayed a plus sign and 5% statistical significance. Such results largely confirm the position of most authors. I have abstained from presenting them in detailed fashion with the intention of avoiding a shift in focus away from this article's main purpose, which is to identify the determinants of women's participation in upper houses rather than in lower houses.
For that purpose, beyond the variables mentioned, two others are included. The main one refers to the powers of the upper house, which is an index made available by Neiva (2006) based on the attributes set out in the constitutions of bicameral countries. Beyond the legislative powers (proposition, amendment, veto), this includes attributes related to the monitoring and control of the Executive, and to the appointment and approval of authorities. The list of these countries with the respective powers of their "second houses" can be found in Appendix 1.
Another variable that expresses (indirectly) the powers of the upper house has to do with the method of selecting its members. Differently from lower houses, all of which are elected directly, we find greater variation among upper houses with regards to members' recruitment: direct election, indirect election (by the lower house or by state/provincial assemblies), a mix of direct and indirect election and appointment (total or partial). So as to simplify the analysis, I will use just two dichotomic variables: one reflects the fact that it is totally elected and the other, the fact that it is totally appointed. In the first case, one expects the coefficient to display a plus sign and, in the second, a minus sign.
Presentation of the Model and Interpretation of the Results
Even though one is analysing only upper houses, limiting oneself to an analysis of bicameral countries may yield distorted results. This is because women's participation may be related to characteristics (measured and not measured) present in such countries. The Heckman model seems to be a more adequate technique than a simple linear regression model, since it checks for the existence of a selection bias and controls its effects in case it indeed exists.14
The model is executed in two steps, expressed by two equations, whose errors are related. In the selection equation, through which I try to explain the existence or not of an upper house, I adopt the same independent variables used in Neiva (2006), with updated values: the adoption of a federal system; the system of government;15 the size of the population and territory; the ethnic and religious fragmentation; the British colonial heritage; and the state of political rights in the country. Table 2 presents the model's results; it must be read starting from the selection equation on the bottom half, which has "Bicameralism" as the dependent variable. After the evaluation of a potential bias, we may proceed to the main model, which has "Percentage of Female Senators" as the variable to be explained.
The high level of significance of the Mills Lambda is an indication that the selection bias is present and that the decision to apply the Heckman model of correction was appropriate (Greene 2003). Without it, it would not be possible to make safe conclusions. The selection equation (at the bottom of the table) confirms the conclusions of Neiva (2006): the federal arrangement and the system of government are important variables to explain the existence of an upper house.16
In the substantive analysis that which has as the dependent variable the percentage of female senators we note that, in line with what the literature says, the establishment of quotas leads to an increase in the number of women in the Legislative. In this study, only constitutional quotas are considered. It would not make sense to evaluate the effects of electoral or party quotas, due to the fact that many upper houses have appointed members.
As for religion, one notes that the mainly Protestant countries tend to have a higher percentage of women senators in their upper houses, whilst the fact that a country is mainly Catholic does not seem to make a difference. Unlike what the literature says, one cannot state for sure that there is a lower percentage of women in the parliaments of Muslim countries, at least with regard to their upper houses. Although this does happen in some countries such as Yemen, Bahrain and Kuwait, for example the high standard deviation shows that this is not regular. Note that such a variable displayed statistical significance in the regression involving the percentage of women members of lower houses described in the previous section , which reinforces the difference between the two houses.
We also find that the variables related to development and human rights do not influence the definition of the number of female senators, a finding in line with Krook (2003). The idea that more socially advanced countries have more women in their parliaments is not confirmed, at least as far as upper houses are concerned. The fact that Sweden, Denmark and New Zealand, all of which are well ranked in indexes of this type, have abolished their upper houses contributes towards this situation. Another contributing factor is the fact that African countries with very low HDIs significantly increased the percentage of women in their parliaments through the adoption of quotas.17
The main finding is demonstrated by the high statistical significance of the variable "powers of the upper house". Every time there is an increase of one unit in its scale (that goes from 1 to 32), the percentage of women in the upper house falls by 0.65; in Model 2, this reduction is of 0.55. In both models, this relation is significant to the level of 1%, thus confirming the hypothesis that in countries where upper houses are strong, there is a lower percentage of women; in those where upper houses are weak, women's participation is greater.
These conclusions are strengthened by the variables referent to the recruitment of parliamentarians: in elected upper houses that therefore possess greater legitimacy and political strength there tends to be a higher percentage of women. However, the statistical non-significance does not allow us to go far with this conclusion. For their part, appointed/indirectly elected upper houses and, for this reason, weaker ones possess a lower percentage of women, with the test having displayed a level of statistical significance of 5%.
Another way to evaluate the role of women in the legislative houses is by means of the posts they hold in the hierarchy of the house. Even though there are a significant number of them, it may be that women do not exert major influence over decisions. On the other hand, even if they have a low percentage of seats, their influence may be increased if the posts they hold are relevant.
Information on posts held by women within parliaments is difficult to obtain, especially when it involves a large number of cases. When it relates solely to the post of speaker, it is easier to come by. According to the IPU, on 31 May 2007, 35 women presided one of the houses of parliament, which means 13.4% of them. In the specific group of upper houses, this figure reaches 21%. With the intention of checking what type of upper house is being presided by women, I compared their average of powers with the average of those presided by men. With the same aim in mind, I also evaluated the method of selection of their members.
As can be seen on table 3, upper houses presided by women are not the strongest ones. On average, they have fewer powers and most are not elected, which, consequently, means they are weaker. Therefore, the discrimination against women is manifested beyond the figures. Women not only possess less political representation, but they also occupy less important posts.
I am not aware of any comparative research involving a large number of countries that seeks to check the relationship between the powers of legislative houses and female representation. However, research into state legislatures in the United States reach similar conclusions to those of this study. It shows that the number of women tends to be smaller in the more attractive assemblies, i.e., those that pay the best salaries, those that are more active, and that have more and better qualified employees (Diamond 1977; Arceneaux 2001). In the comparison between the state lower houses and senates (Norrander and Wilcox 2005 cited in Paxton and Hughes 2007) found that women hold a higher percentage of seats in those that meet only part-time.
A possible explanation for the findings above, as pointed out by Neiva (2006), is that beyond the function of legislating and monitoring, strong upper houses play an important role in controlling the Executive and in the "affairs of State", including those related to security, war, justice, currency, tax issues and international policy the so-called high politics. There is a degree of consensus in the feminist literature that these are themes generally dealt with by men, leaving to women a leading role in social themes, such as problems related to children, the elderly, the unemployed, the environment, public education, housing, family planning and minorities (Chodorow 1978; Ruddick 1989; Cook and Wilcox 1991; Paxton and Hughes 2007; Finamore and Carvalho 2006, 352; Grossi and Malheiros 2001).19 Therefore, when dealing more intimately with the so-called "high politics", upper houses would tend to attract more men.
As for military action and the use of force, some authors have found that, in fact, women tend to be less supportive (Wilcox et al. 1996; Smith 1984). However, as Miguel (2001, 260-61) warns us, "the experience of women in power reveals that the relationship between gender and the so-called 'politics of tenderness' is not at all automatic. Examples to the contrary are numerous and all the evidence indicates that, for better or for worse, women and men can wield power in the same way." According to the author, the fact that women devote themselves more to "social" themes only occurs because it is the only niche available to them in the political field. This is a polemical discussion, which I do not intend to deepen at the moment, but that should be better discussed on empirical bases.
This article evaluates the participation of women in politics, more specifically in the upper houses of the 58 minimally democratic bicameral countries. Unlike what happens with lower houses, I found that the upper houses of Muslim countries do not always display lower percentages of women. As for quotas set out in the constitution, they actually do help to increase such participation. The main explanatory variable is the political strength of the upper house: the data show that often the greater participation of women is associated with weak upper houses.
This leads us to question to what extent the number by itself is enough to reach a conclusion regarding the occupation of political space by women. For one to be able to say that the level of their participation is low or high, or that it is increasing or decreasing, it is also important to evaluate the role and political power of the bodies where they are present. A relatively high percentage of women in the lower house of a country like Rwanda or Cuba, where the Congress has little or no strength, is very different from what happens in the Scandinavian countries. Equally, the high percentage of women in the Argentinian Senate one of the world's strongest generates a much greater impact than in countries such as the Bahamas, Grenada and Trinidad and Tobago, whose upper houses are almost insignificant.
At least in the case of upper houses, the higher percentage of women does not necessarily mean that they are gaining ground. One cannot discard the hypothesis that the apparent strengthening of women's role in politics is a palliative created by the elites to overcome a possible crisis of representation, to win an election campaign or to support a regime. The adoption of quotas has in fact speeded up the process of women's representation, but has not necessarily improved the quality of democracy. On the contrary, it may even generate a negative effect, when applied in contexts of little democracy.
This study's conclusions reinforce the criticism of other authors (Avelar 2006; Goetz and Hassim 2003; Paxton and Hughes 2007), in the sense that the number of women on its own is neither enough to ensure an effective female participation in political decisions, nor that their interests are being seen to. The reason is that women who are elected will not necessarily defend women's interests. This is what one would expect from the Brazilian Congress, for example, where 40% of female deputies (representatives) and 50% of female senators are part of the so-called relatives caucus (Costa and Queiroz 2007, 43), having attained their positions on the strength of the political capital of their fathers and husbands.
In the case of legislative studies, it is important that future research investigate the posts women are occupying in the parliamentary hierarchy, the proposals they are putting forward, the success level they are having, the committees they are sitting on, the decisions they are making etc. In short, it is necessary to check whether the work of female representatives and senators is effective, whether they have a different style of politicking, whether they have a special eye for social problems and whether they legislate differently from their male peers. In the cases of Argentina and Chile, it would also be worth evaluating whether the recent presidential election victories of women have influenced the policies adopted, and the relationship with Congress.
1 Upper houses are also called "second houses", "territorial houses" and "revisory houses". In 67% of countries, they are termed "Senate" (Neiva 2004, 6). Although these denominations do not always correspond to the role and characteristics of these legislative houses, they will be used indistinctly in this text.
2 Included in the sample were countries considered "free" or "partially free" by Freedom House, i.e., those that received up to grade 5 in a scale from 1 to 7.
3 This being so, despite the fact that Japan is an industrialised country, it was placed in the group of Asian countries due to the greater cultural proximity with its neighbours. On the other hand, Australia was included in the group of industrialised countries because of its level of economic development and cultural proximity with Western Europe and the USA.
4 The high percentage of women in these countries' parliaments was a consequence of Lenin's conviction that no revolution would be possible without their participation. This does not mean, however, that they had actual power. The central committee of the Communist Party called the shots; the Legislative merely rubber-stamped the decisions made by that group (Paxton and Hughes 2007, 106, 225).
5 In the case of countries from Central and Eastern Europe, the reduction in the number of women after the fall of the "iron curtain" was meaningful: it went from a level around 30% (Matland and Montgomery 2003) to an average of 16.4%. This is not surprising if one considers the fact that many women were chosen for not having political experience: they were citizens who held honorary titles for outstanding services to the arts, textile workers, seamstresses and others with low-status jobs (Kostova 1998; Waylen 1994).
6 Another characteristic that suggests that women occupy a less important position in strong upper houses (in presidentialist countries) refers to the date when the first female senator took office, generally after the first female representative: in the USA, the difference is of 6 years; in Chile, 3; in Mexico and Uruguay, 12; and in Brazil, 57.
7 A discrepant case is Argentina, whose upper house, in spite of being strong, has 43.1% of women, behind only that of the Bahamas. A probable explanation is linked to the establishment of quotas, associated with the closed list system (Miguel 2006). Argentina is one of the world's three countries that adopt three kinds of quota at the same time. They started being used in 1951 by the Peronist party, which resulted in the lower house having 15% of seats taken by women in the 1952 election, and 22% in the 1955 election then the world's fourth highest percentage (Jones 1998). In 1990, Argentina became the first country to adopt quotas based on the electoral law, the so-called Ley de Cupos ("Law of Quotas"). Also worth recalling is the country's tradition in women's mobilisation and participation in politics: the first International Feminist Congress was held in Buenos Aires in 1910; the "Mothers of the Plaza de Mayo" became known worldwide for their struggle in search of their children, who were "disappeared" during the military dictatorship; before 1980, only six women had taken up positions of substantial leadership in their respective countries, one of them being Isabel Perón; recently, a woman, Cristina Kirchner, was elected president of the country.
8 According to Paxton and Hughes (2007, 154), countries that have constitutional quotas also have, on average, 21.3% of women in parliament about 5% higher than the world average.
9 Paxton and Hughes (2007, 158) warn that, even if the law defines that parties must put up a certain percentage of women candidates, there is no guarantee that the parties will support them or that they will be elected. The authors cite the example of Brazil, where there is a nationwide 30% quota, but only 8.6% of parliamentary seats are held by women.
10 Another advantage of proportional systems is that when a party nominates women to positions of prominence, it compels other parties to do the same (contagion-effect). The cost of this is much higher in systems of plurality/majority, and the benefits much smaller, since to make way for a woman, the party has to exclude a male candidate, thus excluding a faction that traditionally received the nomination.
11 This is the case of Kuwait, Qatar, Saudi Arabia, Oman and the United Arab Emirates.
12 In fact, there are poor countries such as Mozambique and Rwanda that present much higher percentage of women in their parliaments than developed countries such as the United States, France and Japan.
13 In principle, the Gender-related Development Index, that takes into account the difference between men and women, would be a better measure for this study. However, it is only available for a much smaller group of countries, which would lead to the loss of many cases in my sample. Furthermore, the very high correlation between the two indexes (0.98) ensures that there will not be any meaningful distortion.
14 The model is attributed to James Heckman, winner of the Nobel Economics Prize in 2000. When analysing a group of 2,000 women, Heckman found that the 400 who were not on the labour market were in this situation not only because of the conditions of the labour market, but also for other individual reasons (because they had children, because their husbands had sufficient income etc) that influenced the pay of those who were on the labour market. By means of his model, Heckman demonstrated that leaving them outside the equation would lead to a serious selection bias.
15 Countries considered presidentialist are those where the president is both head of state and head of government and where there is no prime minister. The semi-presidentialist countries are those in which the president has executive authority but a smaller role as head of state. Parliamentary countries are those in which the prime minister is the active head of the Executive and also the leader of the Legislative, even though there is a president.
16 This was already expected, as the only difference in relation to that study is the updating of the data and the i nclusion and exclusion of a few countries.
17 This is what happens in Burundi, Ethiopia, Lesotho, Mauritania, Namibia, Rwanda and Zimbabwe.
18 The upper houses currently presided by women are those of the following countries: Antigua and Barbuda, Austria, the Bahamas, Belgium, Colombia, Jamaica, Japan, the Netherlands, the United Kingdom, St. Lucia, Swaziland and Trinidad and Tobago.
19 According to these authors, women act in this way because they were brought up to play the role of the mother and to take care of others, especially the defenceless. Upon entering the spheres of power, they might reduce the aggressive character of politics, valuing solidarity, compassion and peace.
Araújo, Clara. 2001. As cotas por sexo para a competição legislativa: O caso brasileiro em comparação com experiências internacionais. Dados 44 (1).
Arceneaux, Kevin. 2001. The gender gap in state legislative representation: New data to tackle an old question. Political Research Quarterly 54 (1): 143-60.
Avelar, Lúcia. 2001. Mulheres na elite política brasileira. São Paulo: Konrad Adenauer Stiftung/ Editora UNESP.
______. 2006. Mulher e política: O mito da igualdade. Democracia Viva, no.32:98-102.
Caul, Miki. 2001. Political parties and the adoption of candidate gender quotas: A cross-national analysis. Journal of Politics 63 (4): 1214-29.
Chodorow, Nancy. 1978. The reproduction of mothering. Berkeley: University of California Press.
Cook, Elizabeth, and Clyde Wilcox. 1991. Feminism and the gender gap A second look. The Journal of Politics 53 (4): 1111-12
Costa, Sylvio, and Antônio Augusto Queiroz, eds. 2007. O que esperar do novo Congresso: perfil e agenda da legislatura 2007/2011. Brasília: DIAP/Congresso em Foco.
Daherlup, Drude. 2003. Quotas are changing the history of women. Paper presented at the International IDEA/Electoral Institute of Southern Africa/Southern African Development Community Parliamentary Forum Conference on "The Implementation of Quotas: African Experiences", November 1112, in Pretoria, South Africa.
Diamond, Irene. 1977. Sex roles in the Statehouse. New Haven, CT: Yale University Press.
Escobar-Lemmon, Maria, and Michelle Taylor-Robinson. 2006. How electoral laws and development affect the election of women in Latin American legislatures: A test of 20 years into the third wave of democracy. Paper presented at the Annual Meeting of the American Political Science Association, August 31September 2, in Philadelphia, PA.
Finamore, Cláudia, and João Eduardo de Carvalho. 2006. Revista de Estudos Feministas 14 (2): 347-62.
Goetz, Anne Marie, and Shireen Hassim. 2003. No shortcuts to power: African women in politics and policymaking. London: Zed Books.
Greene, William. 2003. Econometric analysis. 5th ed. Prentice Hall.
Grossi, Miriam, and Sônia Malheiros. 2001. Transformando a diferença: As mulheres na política. Revista de Estudos Feministas 9 (1): 167-206.
Hassim, Shireen. 2006. Paradoxes of representation: The impact of quotas for women on democratization. Paper presented at the Yale Conference on Representation and Popular Rule, October 2728, in New Haven, Connecticut.
IPU - INTER-PARLIAMENTARY UNION. Parliament and democracy in the twenty-first century: a guide to good practice. Geneva: Inter-Parliamentary Union, 2007. http://www.ipu.org/dem-e/guide.htm
Jones, Mark P. 1996. Increasing women's representation via gender quotas: The Argentine ley de cupos. Women and Politics 16 (4): 75-98.
______. 1998. Gender quotas, electoral laws and the election of women: Lessons from the Argentine provinces. Comparative Political Studies 31 (1): 3-21.
Kenworthy, Lane, and Melissa Malami. 1999. Gender inequality in political representation: A worldwide comparative analysis. Social Forces 78 (1): 235-69.
Kittilson, Miki Caul. 2006. Challenging parties, changing parliaments: Women and elected office in contemporary Western Europe. Columbus: Ohio State University Press.
Kostova, Dobrinka. 1998. Women in Bulgaria: Changes in employment and political involvement. In Women and democracy: Latin America and Central and Eastern Europe, ed. J. S. Jaquette and S. L. Wolchik. Baltimore: John Hopkins University Press.
Krook, Mona Lena. 2003. Candidate gender quotas: A framework for analysis. Paper presented at the 2003 Annual Meeting of the American Political Science Association, August 2831.
Lindberg, Staffan. 2004. Democratization and women's empowerment: The effects of electoral systems, participation and repetition in Africa. Studies in Comparative International Development 39 (1): 28-53.
Matland, Richard E. 1998. Women's representation in national legislatures: Developed and developing countries. Legislative Studies Quarterly 23 (1): 109-125.
______, and Kathleen Montgomery. 2003. Women's access to political power in post-communist Europe. Oxford, UK: Oxford University Press.
Matland, Richard E., and Michele Taylor. 1997. Electoral system effects on women representation. Comparative Political Studies 30 (2): 186-210.
Matthews, Donald. 1959. The folkways of the United States Senate: Conformity to group norms and legislative effectiveness. American Political Science Review 53:1064-89.
Miguel, Luis Felipe. 2001. Política de interesses, política do desvelo: Representação e "singularidade feminina". Revista de Estudos Feministas 9 (1): 253-67.
________. 2006. Diferenças regionais e o êxito relativo de mulheres em eleições municipais no Brasil. Revista de Estudos Feministas 14 (2): 363-85.
Neiva, Pedro. 2004. Estudo comparado de câmaras altas: Os poderes e o papel dos Senados nos sistemas presidencialistas. Ph.D. diss., Instituto Universitário de Pesquisas do Rio de Janeiro (IUPERJ).
________. 2006. Os determinantes da existência e dos poderes das câmaras altas: federalismo ou presidencialismo? Dados 49 (2): 269-99.
Norris, Pippa. 1996. Legislative recruitment. In Comparing democracies: Elections and voting in global perspective, ed. L. Leduc, R. Niemi and Pippa Norris. Thousand Oaks: Sage.
______. 2004. Electoral engineering: Voting rules and political behavior. Cambridge: Cambridge University Press.
______. 2006. The impact of electoral reform on women's representation. Special issue, Acta Politica 41 (2): 197-213.
______, and Ronald Inglehart. 2000. Social cleavages and elections. Paper presented at the International Political Science Association World Congress, August 3, in Quebec.
Paxton, Pamela. 1997. Women in national legislatures: A cross-national analysis. Social Science Research 26 (4): 442-64.
_____, and Melanie Hughes. 2007. Women, politics, and power: A global perspective. Los Angeles: Pine Forge Press.
_____, and Sheri Kunovich. 2003. Women's political representation: The importance of ideology. Social Forces 81 (5): 87-114.
Reynolds, Andrew. 1999. Women in the legislatures and executives of the world? Knocking at the highest glass ceiling. World Politics 51 (4): 547-72.
Ruddick, Sara. 1989. Maternal thinking: Towards a politics of peace. Boston: Beacon Press.
Rush, Michael. 2001. Socio-Economic composition and pay and resources in Second Chambers. Special issue, The Journal of Legislative Studies 7 (1).
Sinclair, Barbara. 1989. The transformation of the U.S. Senate. Baltimore: John Hopkins University Press.
Smith, Tom. 1984. The polls: Gender and attitudes toward violence. Public Opinion Quarterly 48 (1): 384-96.
Steininger, Barbara. 2000. Representation of women in the Austrian political system 1945-1998: From a token female politician towards an equal ratio? Women and Politics 21 (2): 81-106.
Sweeney, Shawna. 2004. Global economic transformations, national institutions, and women's rights: A cross-national comparative analysis. Paper presented at the Annual Meeting of the APSA, September 2-5, in Chicago.
Tsebelis, George and Jeannette Money. 1997. Bicameralism. Cambridge: Cambridge University Press.
Waylen, Georgina. 1994. Women and democratization: Conceptualizing gender relations in transition politics. World Politics 46:327-54.
Wilcox, Clyde, Lara Hewitt, and Dee Allsop. 1996. The gender gap in attitudes toward the Gulf War: A cross-national perspective. Journal of Peace Research 33 (1): 67-82.
Submitted in July, 2007.
Accepted in December, 2007.
* I wish to thank Argelina Figueiredo and two Brazilian Political Science Review (BPSR) anonymous peer reviewers for their comments.