# Introduction ivil wars are a phenomenon with diverse and heterogeneous geographical distribution. The studies about them have multiplied in recent decades, but few have tried to assess its relation with the various institutional settings. As far as it is known, none of them dealt specifically with the high chambers, legislative houses fundamental for the political organization of the countries where they are located, especially those which are in the process of transition (or consolidation) towards democracy. This is the main purpose of the paper: assess the extent to which the high chambers serve as a tool of negotiation between elites, in order to reduce the occurrences of civil wars in the country. The proposal seems to make sense, if we consider the high chambers usually represent the upper classes and the privileged minority interests (UHR, 2008), thus making the change of the status quo more difficult (TSEBELIS; MONEY, 1997). In addition, they tend to be extremely flexible in their composition and operation, they are adapted to many contexts and situations and count with different forms of recruitment of their members (the majority is chosen indirectly). The idea is that they may be easily set up and configured to assist in the negotiation between privileged segments of the population, as well as serving as one more channel available to dictators, authoritarian regimes, to co-opt representatives of the elites. They are also justified by their own geographical distribution of such legislative houses. On the one hand, there is a reasonable concentration of strong and longlasting Senates in some countries of South America, a region with little incidence of civil war. Also, the creation of various high chambers in recent years in the countries of Africa and Eastern Europe, regions that are democratizing rapidly and are peppered with ethnic and religious conflicts. In the specific case of the region with the highest incidence of civil wars, Africa, it is seen that about 40% of the countries adopt a bicameral conformation, against 30% in the rest of the world. On the other hand, there is a low frequency of dual-chambered systems in two quite distinct groups of countries: 1) In the Middle East, a region with a high rate Lastly, the impression is that many Senates appear in contexts reasonably peaceful and/or with the of wars; 2) in the Scandinavian countries and suchlike 1 , which Author ? ?: University of Brasilia. e-mail:prneiva@gmail.com reached a high degree of social development and that experienced few military conflicts. prospect of peace; they are less frequent in situations where there is no conflict or where they are very sharp and/or with low perspective of change. Up to a certain point, within its limitations, this perception suggests a relation with the thesis that the incidence of civil war is greater in the socalled "anocracies" (or "semi-democracies" / "hybrid regimes") and lower in consolidated democracies and authoritarian systems. The relation between bicameral arrangements and civil wars is not evident. It is added to that the fact that this type of conflict is complex and involves many facets. Neither the high chambers present only one model or configuration, in the case of "a concept still in search of a theory" (SMITH, 2003, p. 3;UHR, 2008). Nevertheless, we believe it is possible to think about relations between the two variables. In the next section we began the discussion by making a brief review of the literature on civil wars and trying to identify their supposed explanatory variables. Next, we situate the discussion of the institutions in this context, with emphasis on the role and characteristics of the high chambers. In section III we do the analysis of our data by means of three logistic regression models, which have as dependent variable the incidence of civil war in a certain country, in a particular year. Finally, in section IV, we make some final considerations, drawing attention to the need for further study of the species, as a way to get a more complete understanding of this so important phenomenon to the contemporary world. # II. # Traditional Justifications for the Civil Wars The justifications for the outbreak, permanence and finalization of civil wars have followed differentiated criteria, with multiple approaches. One of the most addressed concerns has to do with degree of democratic development of the country. On the one hand, some authors have found that its impact was not significant or had ambiguous effects (Collier;Hoeffler, 2004); others have found a negative association between the two variables (GURR, 2000); and still others found this negative association when the anocratic regimes repress a considerable portion of the population (Zimerman And Alves, 2007). A more complex relation, involving other intervenient aspects was identified by some authors. Hegre (2001) found the effect of democracy on the incidence of civil wars also depended on the income level of the population; according to him, in the societies in which it was lower, the risk of war was greater. For some authors, more important than the degree of democracy itself is the type of representation: according to Reynal-Querol (2005), violence is less likely in proportional systems. When working with the concept of civil rights, Fearon (2004) identified its absence helped explain the outbreak of violent conflicts. But Pezzini and MacCulloch (2004) found they reduced the propensity for revolution. A well-established thinking argues that the relation is not linear. Hegre et al. (2001), for example, found the anocracies (or semi-democracies/hybrid regimes) were more prone to conflict than largely democratic or authoritarian regimes. Other authors have pointed in the same direction, that is, they found that authoritarian states, as well as institutionally consistent democracies, experienced less civil wars than regimes that were in intermediate stages (DE NARDO, 1985;FRANCISCO, 1995). One of the explanations given is that, in semi-democracies there is the possibility of the population to get organized, but its capacity for collective action towards peace is still limited (VREELAND, 2008). It is worth adding that the level of repression in those regimes is not enough to prevent rebellions, at the same time the institutions and mechanisms for resolving conflicts are not sufficiently developed yet. Another aspect widely considered in the justification of the civil wars is the ethnic identity. Here there is also no consensus among the authors. Some claim these variables are strongly related (FEARON; LAITIN, 2003). Others, that this relation is not as important, unless there is a dominant ethnic group (CAREY, 2007;Collier, 2003;Collier;Hoeffler, 2002Hoeffler, , 2004)). But Sambanis (2001) considers it has effect only for ethnic wars. In addition to ethnic issues, the religious ones also have a share in the explanation of civil wars. According to Reynal-Querol (2002), societies divided by religion are more prone to experiencing intense conflicts than those where the main cleavage is linked to social class or the language. Because the religious identities are fixed and particularly difficult to be negotiated, thus increasing the chance of violence. The author shows empirically that the religious polarization is one of the most important factors in the explanation of the incidence of civil war. A third factor widely considered as a predictor of civil war has to do with geographical features, either in its physical or human aspects. De Soysa (2002) found the population density was positively associated with the probability of conflict. However, Collier and Hoeffler (2004) found no evidence in that direction. With regard to location, Sambanis (2001) found the fact of a civil war happening in the previous year in a neighboring country increases the risk of its outbreak, once guerrilla groups of the country may support rebels in the other, with weapons, expertise and human resources, thus enabling them to a rebellion. However, part of the literature found no support for these hypotheses (CAREY, 2007; J. FEARON;LAITIN, 2003;HEGRE, et al., 2001). Whereas a civil war involves armed conflict between the State and rebel groups, some authors have also taken into account the location of the capital, that is, the place where the power is established (BUHAUG; GATES, 2002). The topography of the country seems to be especially important. Mountainous terrain and forest cover were used as predictors of conflicts (FEARON; LAITIN, 1999LAITIN, , 2003;;COLLIER;HOEFFLER, 2001;GURR, 2000). According to Buhaug and Gates (2002), in areas of mountains and forests, in addition to greater dispersion of the population, the rebel troops have greater freedom of movement and transport of arms, in addition to having greater possibility to hide and avoid attacks of the enemies. Economic factors also have been presented as important in the explanation of civil wars. According to some authors, poor countries which have low per capita income and low level of growth, are more prone to experiencing them (COLLIER; HOEFFLER, 1998; DE SOYSA, 2002). The great proportion of primary products in export tariff, which generate resources for the financing of weapons by rebel groups, is also highlighted (CAREY, 2007; DE SOYSA, 2002; FEARON; LAITIN, 2003;FEARON, 2004;HUMPHREYS, 2005;ROSS, 2004). This is what has happened in several countries, from the extraction of oil, timber and precious stones. There also happens to be associated with the cultivation and smuggling of illegal drugs (FEARON, 2004; ROSS, 2004). # a) The Role of the Institutions In recent decades, literature in political science has emphasized the role of the institutions in the pacification of conflicts, either in the consolidated or consolidating democracies 2 . In the first ones, Lijphart (1977) affirms the institutions promote the division of power, facilitate the accommodation and the cooperation between the elites, thus making them more prone to stability and good governance. According to the author, the elites -in their various segments -form a kind of "cartel", with the goal of resolving the inter-ethnic differences and avoiding self-destruction. In countries undergoing democratic transition, Snyder (2000) also states that the consolidation of democracy involves the accommodation of the elites, even those which may have been excluded from the electoral process. In a non-democratic system, there are those who say the institutions are " fake" and it makes no sense to analyze them (FRIEDRICH; BRZEZINSKI 1961, BROOKER 2000, PACKENHAM 1970). We must consider, however, that even dictators need to maintain a political support to continue in power. After all, the use of force has a cost and may not always be used efficiently. In that sense, they try to build coalitions from the co-opting of groups that do not participate in it (BUENO DE MESQUITA ET AL., 2003; LINZ, 1973; O'DONNELL, 1979), making political concessions by means of an institutionalized environment. On the same line, Gandhi and Pzerworski (2007) claim the autocrats must be able to realize the strength of the threats and comply an institutional response to them, under penalty of having their permanence in power either reduced or compromised. In the opinion of Gandhi and Pzerworski (2007), the legislative are the locus ideal for that, once they are capable of revealing the demands, control access, organize the commitments and draw up the legal and public agreements. On the same line, Gandhi and Vreeland (2004) found that dictatorships where there was a legislative power 3 were less inclined to try a civil war than those which did not have one. According to the two authors, that happens because the first ones combine a mixture of optimal repression to divergent insurgencies and accommodation/co-opting of different social groups. With that, they suggested the key institutional element to explain the existence of civil war is not the fact of being a dictatorship or a democracy, but of there being or not a legislative power.More specifically, Leonard (2013) suggests that a Senate, with members of the professional, religious, business elites and representatives of regional assemblies and civil associations would be a good instrument for the control of other government bodies. Our assumption is that the Senate is an instrument in the hands of the rulers for the division of power and for the construction of agreements. And that happens independently of the fact of being an autocratic country or being in the process of transition to democracy. It also depends on the extension of formal legislative powers, once even the high chambers politically weak exert some influence (HELLER, 1997; Tsebelis; Money, 1997).Our point is that, just like it happens in democratic systems, bicameralism system is also used in authoritarian and semi-democratic regimes as an element promoter of its stability and survival, as well as inhibitor of violence. The main hypothesis is that the high chambers are used as a tool for building consensus among elites. Consequently, they reduce the occurrence of civil wars, as they help solve the problem of political instability generated by the majority rule, thus creating a set of stable results (a "core"). When the legislature is divided into two chambers, it is more difficult to form a coalition consisting of the majority of each of them (HAMMOND; Miller, 1987). This assumption is justified by two main reasons. First of all, by the fact that the high chambers are taken as elitist, often chosen on the basis of class criteria and with severe restrictions with regard to the minimum level of income of their members. This is the case, for example, of the House of Lords in the British parliament, the French Senate, the Brazilian Senate during the Empire and the extinct Senates of Denmark and Sweden (CONGLETON, 2002). Case studies and also comparative studies offer support for the idea that the high chambers concentrate specific groups of elites, often favoring some of them. In the United States, Weingast (1991) showed the over-representation of slave states in the Senate, at the beginning of the Republic, ensured that no liberating legislation was approved, even if a majority of the population (or the House of Representatives) would be favorable. According to the author, that way, the bicameralism system became a key mechanism for the protection of a minority of owners of slaves 4 . In Latin America, Samuels and Snyder (2001) suggest the disproportionality of representation -much greater in the high chambers -tends to encourage politically conservative districts (and their elites), at the expenses of urban and politically progressive districts. On the same line, Stepan (2000) says the over-representation, especially in the Senate, happens in the states with the worst income distribution and with strong oligarchic traditions. The second justification for our hypothesis is the adapting ability of the high chambers to different contexts. In such a way that Tsebelis and Money called them "protean", that is, they are able to take different forms and characteristics and easily adapt to various situations. In fact, this seems to have been a frequent mark in many high chambers in the world, especially in periods of accelerated transformation. In Brazil, it is worth mentioning the so called Pacote de Abril -Package of April, which went into force in 1977 by the former President Geisel, by means of which he arrogated to himself the indication of half the senators 5 . In the period of democratic resume, the best example of rapid transformation of the Senate took place when the former territories became states, thus increasing the number of senators of a more poor and politically backward region (SOARES, 1973). It is also worth of mention the privilege granted to former presidents of several countries (Chile, Paraguay, Rwanda, Burundi, etc.) to occupy the office of senator after the end of its mandate. Our assumption is that the high chambers play, in various countries, the function of accommodating some elites in power. In the regimes under transition, it allows the process to be done in a negotiated settlement. In the specific case of this work, our expectation is that its existence in a given country is associated with the occurrence of civil wars, and their reduction more specifically A suggestive example of this last situation is the creation of the Senate of Lebanon in 1989, from the explicit forecast in the so called "Agreement of Ta'if", which ended the conflict in the country 6 . It also seems to have been what happened in Tajikistan, where the introduction of the bicameralism system responded to the concerns of stabilization of the political situation and consolidation of the new democratic regime in the period after the civil war (ZAVRABYAN, n/d). In the case of the United States, it was related to the greater stability of public policy in the twentieth century, in contrast with the instability in the United Kingdom: according to Riker (1992), in the context of the "almost" British single cameralism, the two most important changes made in the economy in the period after World War II returned to the status quo ; as for the United States, the two major changes made in the same period (the New Deal and the law of civil rights) were not reversed. Our supposition is supported by the empirical observation that the adoption of a bicameral system was related to the process of democratization in various countries in recent years. In addition to Tajikistan, that's the case for Kazakhstan, Uzbekistan, in 2005, and the Czech Republic and Poland after 1992 (ZAVRABYAN, n/d). In accordance with the thesis of Tsebelis and Money (1997), states where the bicameralism provides more sustainable political results, reduce the conflicts for a single main dimension. It is also worth noting the statement of Dahl (1958) that the political efficiency of a group depends on both its potential to control and its unity potential. According to this author, a group with a relatively small potential control, but with a high potential for unity, may be politically more effective than a group with high control potential, but with low potential for unity. As the high chambers are in general smaller than the low chambers, they tend to rely on greater unity between its members, thus facilitating the commitments and collective actions (OLSON, 1971). In a study on the bicameralism system in Africa, Neiva (2010) found a strong association between the existence of an upper chamber and the occurrence of civil wars.The author did not identify, however, the direction of this relation, that is, whether the bicameralism system was the cause or consequence of the existence of this type of conflict. It is possible for both situations to coexist. Works on the theme are rare and do not inform in this respect. The absence is justified by the very lack of studies on such institutions; although their number has increased substantially in recent years, the high chambers are still little known. Beforehand, it is important to remember these are fairly heterogeneous institutions, even with respect to their functions and political force. One of the perceptions on this was provided by Lijphart (1980;1999), who ranked the bicameral systems in four main types, from two specific dimensions: congruence and symmetry. The first one concerns to the preferences of each one of the Houses; the second one, to their powers. Despite the different bicameral arrangements, it is important to mention that also high chambers that seem weak or insignificant perform some type of influence (TSEBELIS; MONEY, 1997), even if they does not give the final word and have only an advisory role to the acting ruler. # III. # Data Analysis Our research involves all countries of the world with population greater than one million inhabitants in each year that is being investigated, during a period of 57 years (1950 to 2006). A total of 148 countries were researched, making a total of 6,618 observations. It is worth remembering that many countries were created in recent decades, only by entering the data bank from its creation. The analysis will be made by using a logistic regression model, which has as dependent variable the incidence (or not) of civil war in a certain country in a given year. If so, it will be coded with the number "one"; otherwise, with "zero". Our main explanatory variable concerns to the existence of the bicameralism in the country. We want to know up to what extent it impacts on the occurrence of civil war; in other words, we assessed whether the existence of a high chamber is not associated with the existence of internal conflicts. Our assumption is that its presence reduces the probability of such disputes, based on previously presented theoretical foundation. To assure that the effect of bicameralism is effectively what was observed, we also consider the possible impact that other important variables may have. With this aim, we use control variables relating to geographic, political, economic and social aspects of the respective countries. As a result of the variables 'population' and 'neighbor at war" relying on few observations, we decided to included them only in the second and third models. The results are presented in the following table . A detailed description of each of the variables, and their respective sources, is presented in appendix 1. In general, the three models confirm our expectations: almost all variables showed statistical significance at the level of 1 %. The most important results for our purposes relate to the "bicameralism" variable. It presented negative signal in all three models, proving our initial hypothesis. As seen in the first of them, the fact of having a high chamber in the country reduces by 27% the chance of incidence of a civil war in a given year. In the second and third models, that number was 30% and 25 %, respectively 7 . These results suggest that, in fact, the Senates serve as a stabilizing mechanism of the political system. Possibly, because they help promote an agreement among the elite of the country, thus including those that may have been away from other political disputes, either electoral or non-electoral ones. As a House lower than the lower chamber and which is often chosen indirectly, the second House usually has a structure much more flexible than the first one, thus allowing it to act in the negotiation and reduction of conflicts between distinct segments. Moreover, in several of the examined countries, they rely on the representation of different groups: ethnic, religious, regional and/or professionals 8 . Such differentiating characteristics confirm the suggestion of Tsebelis and Money (1997) that the high chambers are "protean", that is, they are easily transformed and adapt to many situations. The model also confirms expectations prevailing in the literature. The fact that the country is an anocracy (in contrast with being a consolidated democracy or to be driven by an authoritarian regime) increases the tendency to have a civil war. In model 1, 85%,in model 2, 70%, and in model 3, 139% in this type of regime. It remains, therefore, the idea that the differences are resolved in a peaceful and negotiated way in democratic regimes and by the dictator in command in the authoritarian regimes. The three models reinforce the idea that a rough terrain (mountain) and of difficult movement in and out favors the incidence of civil war, once the rebel groups would have movement advantages, without being noticed. In model 1, the existence of mountains increases in 18% the chance of a civil war taking place in a given year. The impact of the variable "GDP per capita (logarithmically) also confirms what the literature has stated on the topic: the income increase reduces the possibility of civil war. In model 1, this reduction is nothing less than 14%. The fact of having a 'neighbor at war" also appears to influence our dependent variable: model 2 shows it gets to increase this chance at 124%. Although model 3 may count with a smaller number of observations, it confirms the previous results and shows the great impact exerted by the "population size" variable. The only variable that has not confirmed the expected results was the one that informs about the fact that the country is a diamond producer. In the first and second models it presented a positive signal and a negative signal in the third one. However, none of them was statistically significant. The ideal would be for this variable to also cover other natural resources that generate exacerbated greed, such as gold and oil. However, such information is not readily available for a large number of countries and for a long period of time. # IV. # Final Remarks The literature on the outbreak and occurrence of civil wars has improved itself in recent years. However, only recently it started to deal with the impact exerted by political institutions. In this work, we suggest the fact that the country adopts a bicameral system reduces the chance of civil war incidence, even after controlling the possible influences of other explanatory factors. In some way, not yet clearly perceived, it seems the high chambers play an appeasing role, either because it serves as a flexible decision center for decision making, or because it presents itself as one more possibility of power-sharing between representatives of the elites. These, however, are not exhaustive or definitive explanations. We once more remind there is no single model of bicameralism nor an explanatory theory consolidated in its respect. As advises Uhr (2008), we must be careful when we present generalizations on the role of high chambers, even because their powers and practices are often renewed. On the same line, Norton (2007) talks about the difficulty in defining the number of bicameral countries in the world, once it is not always easy to identify to what extent the legislature of a country is effectively bicameral. Nevertheless, despite the difference in the format of the bicameralism adopted by the countries, it is important to continue unraveling the purposes, functions and justifications for the high chambers.According to the proposal of the 'Founding Fathers' in the North American original model, it would be the one to control and serve as a counterweight to other powers (checks and balances). In this article, we see that, even if the control function is not present, the counterweight seems to be more frequent, especially in the sense of giving greater overall balance to the political system. 1(1950 to 2006)OddsOddsOddsRatioRatioRatio(Z test)(Z test)(Z test)Bicameralism0,73***0,70***0,75***(-4,41)(-4,47)(-2,92)Anocracy1,85***1,70***2,39***(7,27)(5,82)(7,89)Mountains1,18***1,09***1,13***(5,41)(2,62)(3,00)Diamonds1,191,050,79(1,17)(0,35)(-1,17)GDP per capita0,86***0,90***0,91***(log)(-5,01)(-3,36)(-2,56)Neighbor at war-2,24***1,73***(9,33)(5,19)Populationsize--1,58***(log)(12,63)N4,7834,0973,667Pseudo R 20.0370.0580.12Qui-square164.07223.32364.07 ## Appendix 1 Some of these countries closed their high chambers: New Zealand (1950), Denmark (1953), Sweden (1970), Iceland (1991), Scotland (1999). Others count with a little evident bicameral arrangement (Norway) or with high chambers with few powers (UK, Canada). 2 However, there is no agreement on which institutions are most appropriate for that. As regards the decentralization, for example, some authors affirm the larger it is, the greater will be the incentives for the violence of peripheral groups (HECHTER, 2000; A. LIJPHART, 1980). But Tucker (2012) found no effect of federal arrangement on the propensity of conflicts of this type. 3 The authors cite the example of Poland, where the communists entered into an agreement with Catholic groups. It also happened in King Hussein's Jordan, where the Muslim community reduced the resistance to the Hashemite dynasty, after obtaining legislative office, ministerial posts and the influence on educational and religious policies (SCHWEDLER, 2000). 4 The Civil War broke out when the number of free states and slave owners decreased, thus allowing the formation of a new anti-slavery coalition (MILLER; HAMMOND; KILE, 1996). 5 These senators were known as "bionic", an allusion to the TV series "The man of six million dollars", also known as the "bionic man". 6 The agreement provided for the election of a lower chamber on non-confessional basis and the creation of a Senate, which would represent "all spiritual families". 7 For the reader unfamiliar with statistics, we explain the origin of this number. In order to express the result in percentage terms, we reduced in one unit the amount expressed by the odds ratio and we multiplied the result by 100. When the final product is less than "one", the impact will be negative; when larger than one, it will be positive. In the example above the "bicameral" variable, (1 -0.73) * 100 = 27 %. 8 Some representative examples of specific groups in high chamber are: a) ethnic groups -Burundi, Burkina Faso, Madagascar; b) professional groups -Ireland, * The geography of civil war HBuhaug SGates Journal of Peace Research 39 4 2002 * Rebellion in Africa: Disaggregating the effect of political regimes SCCarey Journal of Peace Research 44 1 2007 * Breaking the conflict trap: Civil war and development policy PCollier 2003 World Bank * On economic causes of civil war PCollier AHoeffler Oxford economic papers 50 4 1998 * On the incidence of civil war in Africa PCollier AHoeffler Journal of Conflict Resolution 46 1 2002 * Greed and grievance in civil war PCollier AHoeffler Oxford economic papers 56 4 2004 * Power in numbers De JNardo 1985 Princeton University Press Princeton, NJ * Paradise is a bazaar? Greed, creed, and governance in civil war IDe Soysa Journal of Peace Research 39 4 2002 * Ethnicity, insurgency, and civil war JFearon DLaitin American Political Science Review 97 01 2003 * Why do some civil wars last so much longer than others JDFearon Journal of Peace Research 41 3 2004 * The Relationship between Coercion and Protest An Empirical Evaluation in Three Coercive States RAFrancisco Journal of Conflict Resolution 39 2 1995 * Ethnic warfare on the wane TRGurr Foreign Affairs 2000 * The core of the constitution THammond GMiller The American Political Science Review 81 4 1987 * MHechter Containing nationalism USA Oxford University Press 2000 * Toward a democratic civil peace? Democracy, political change, and civil war HHegre TEllingsen SGates NPGleditsch American Political Science Review 95 1 2001 * Bicameralism and Budget Deficits: The Effect of Parliamentary Structure on Government Spending WHeller Legislative Studies Quarterly 22 4 1997 * Natural Resources, Conflict, and Conflict Resolution Uncovering the Mechanisms MHumphreys Journal of Conflict Resolution 49 4 2005 * Democracy, Liberty and Montesquieu: Constructing Accountable Order in African Conflict States DKLeonard IDS Bulletin 44 1 2013 * Democracy in plural societies: A comparative exploration ALijphart 1980 Yale University Press Yale * Patterns of Democracy: Government Forms and Performance in Thirty-Six Countries ALijphart 1999 Yale University Press New Haven * The role of freedom, growth, and religion in the taste for revolution RMacculloch SPezzini 2004 London School of Economics * Bicameralism and the Core: An Experimental Test GMiller THammond CKile Legislative Studies Quarterly 21 1 1996 * Bicameralismo na África PNeiva Revista Política Hoje 19 1 2010 * Adding value: The role of second chambers PNorton Asia Pacific Law Review 15 2007 * The logic of collective action: Public goods and the theory of groups MOlson 1971 Harvard University Press Boston * Ethnicity, political systems, and civil wars MReynal-Querol Journal of Conflict Resolution 46 1 2002 * The Justi fication of Bicameralism WilliamHRiker International Political 28 1992 Science Review * How do natural resources influence civil war? Evidence from thirteen cases MLRoss International Organization 2004 * Do ethnic and nonethnic civil wars have the same causes? A theoretical and empirical inquiry (Part 1) NSambanis Journal of Conflict Resolution 45 3 2001 * The value of a vote: malapportionment in comparative perspective DSamuels RSnyder British Journal of Political Science 31 04 2001 * From voting to violence: Democratization and nationalist conflict JLSnyder 2000 New York; Norton * Sociedade e política no Brasil GSoares 1973 Difel São Paulo * Bicameralism GTsebelis JMoney 1997 Cambridge University Press * Can peace be engineered? Institutions, political inclusion, and ethnic conflict JTucker 2012 * The Oxford Handbook of Political Institutions JUhr K. Pepperday and J. Wanna 2008 Oxford University Press Bicameralism * The effect of political regime on civil war unpacking anocracy JRVreeland Journal 2008