SciELO - Scientific Electronic Library Online

vol.26 número1Evaluating Electoral SystemsLas virtudes del sistema binominal índice de autoresíndice de materiabúsqueda de artículos
Home Pagelista alfabética de revistas  

Servicios Personalizados




Links relacionados


Revista de ciencia política (Santiago)

versión On-line ISSN 0718-090X

Rev. cienc. polít. (Santiago) v.26 n.1 Santiago  2006 


revista de ciencia politica / volumen 26 / Nº 1 / 2006 / 216 - 225

Electoral Reform Doesn't Matter-or Does It? A Moderate Proportional Representation System for Chile

Peter M. Siavelis

Wake Forest University, Estados Unidos


En contra de las intenciones de los ingenieros electorales en Chile, el sistema binominal ha tenido poco impacto sobre el sistema de partidos. Pareciera entonces que el debate sobre la reforma electoral es poco relevante, porque con cualquier tipo de reforma los partidos políticos simplemente adoptarán estrategias electorales para asegurar su sobrevivencia. Este artículo sostiene que es poco probable que la reforma electoral pueda transformar el sistema de partidos en términos del número de partidos. No obstante, este también plantea que el enfoque sobre el número de partidos en el debate sobre la reforma electoral oculta otros beneficios importantes que la reforma electoral pueda producir, incluyendo mejor representación política para todos los partidos de peso, más democracia interna en los partidos, procesos de selección de candidatos más transparentes, un nivel más alto de accountability entre los legisladores, competencia real entre los partidos y gobernabilidad a largo plazo que no se trata solamente de la capacidad de los presidentes para legislar. El artículo argumenta que un sistema de representación proporcional moderada con pequeños magnitudes es el mejor sistema para lograr estas metas.


Contrary to the intentions of electoral engineers, Chile's binomial system has failed to fundamentally transform the party system. It may seem that the debate on electoral reform is, therefore, irrelevant, because whatever election system is implemented, parties will simply adopt electoral strategies to ensure their survival. This article acknowledges that electoral reform, whatever its nature, is unlikely to significantly affect the number of parties in Chile. However, it argues that by focusing on notions of stability based in the number of parties, the debate on electoral reform has overlooked other important benefits of that electoral reform would provide, including the representation for all significant parties in parliament, enhanced internal party democracy, open candidate selection procedures, more accountable members of congress, real competition between parties, and a long-term governability that goes beyond the simple ability of presidents to pass legislation. The article makes the case that a moderate proportional representation system with small district magnitudes can best achieve these goals.

PALABRAS CLAVE • ChileSistema electoralSistema binominalPartidos políticos Democratización


Contrary to the intentions of military reformers and the assumptions of many electoral systems theorists, Chile's radical electoral reform seems to have had little effect. After fifteen years of the binomial election system the number of significant political parties in Chile is not that different from the historical norm. In addition, though a two-coalition pattern has prevailed since the return of democracy, the underlying ideological divisions of the party system are still characterized by the historic tres tercios, though it is undeniable that the ideological spectrum of the political system has narrowed remarkably. This continuity in terms of numbers and ideological orientation is testament to the perseverance of Chile's strong and institutionalized parties. In a remarkable, but not completely unexpected, way parties have successfully negotiated to ensure their survival, challenging assumptions about the ability to transform party systems through electoral engineering.

The natural conclusion is that the current debate on electoral reform in Chile does not matter. Whichever election system is ultimately adopted (or if the binomial system is maintained) it would seem that Chile's strong parties will continue to adapt, evolve and negotiate to ensure their survival, and the party system will retain the same general format. On this basis, one could conclude that the debate on electoral reform is largely irrelevant to Chile's democratic future.

However, I argue that electoral reform does matter, but it matters in a different way than the debate has been framed up until now. Traditionally most pose the question of electoral reform as a classic tradeoff between representation and governability1, where proportional systems provide benefits with respect to representation, but at cost in terms of governability, and majoritarian systems do the opposite. These arguments are based on the presumption that the most important characteristics of election systems are their potential transformative effect on the number of parties. A change in electoral systems, in theory, can produce the desired change in party systems to build democracies that are either more governable or more representative. There are two ways the Chilean case is instructive in this regard. It clearly shows that the stark tradeoff between representation and stability is really a false dilemma—multiparty presidential systems can provide both representation and stability. It also shows that it is not so easy to engage in electoral engineering in countries with strong and institutionalized parties that can confound the intentions of electoral engineers.

However, less present in the debate on Chile is how electoral systems have effects well beyond simple numbers and the stabilizing or destabilizing consequences of many or few parties. Indeed, election systems help determine the quality of representation, the incentives of politicians, the ability of citizens to choose and hold accountable their representatives, the level of competition between parties, and the capacity of voters to understand electoral processes and make informed choices—each of these may have little to do with the number of parties, and in Chile's case are probably more important to governability than the number of parties. Therefore, debate in Chile should center not on the usual broad brush strokes of what electoral reform can or cannot do with respect to the simple number of parties or ideological blocs, because the country is characterized by a well-institutionalized party system that shows little propensity for change. Debate should focus, instead, on finding an electoral system that will shape the behavior and incentives of politicians to enhance the performance of an existing, functional, and well-institutionalized multiparty system. I present those arguments here, proposing that a moderate, small magnitude, open list proportional representation system will best enhance the performance of Chile's multi-party democracy2. Before arguing for such a system, I analyze the capacity of each the other reform proposals to similarly enhance the quality of democracy in the terms set out here.


There appears consensus that the binomial system is in dire need of reform, though some continue to advocate keeping the binomial option on the table. The arguments against the system as it has functioned since the return of democracy are familiar and do not need to be recounted in depth here. In short, the system effectively provides a lock on power for the two major coalitions in the country. It provides a bizarre combination of sensitivity and insensitivity to partisan shifts in the electorate, depending on the proximity of parties and coalitions to the election system's thresholds. The strategic complexity of the system shifts the power candidate nomination into the hands of party elites who must make strategic calculations on candidate placement in the interests of coalition maintenance. Voters have little input into the selection of candidates. More seriously, the outcome of elections is usually a forgone conclusion: except in few cases one member of the Concertación and one member of the Alianza are likely to win in each district, and the deck is stacked in the favor of incumbents.

However, for its proponents, the binomial system provided strong incentives for the creation of a two-coalition pattern that has been at the root of post-authoritarian stability. While true, these arguments tend to ignore how the balance between the binomial system's benefits and tradeoffs are time dependent. In other work on Chile, I have argued that many aspects of the Pinochet constitution, while useful in the immediate post-authoritarian period, would prove less conducive to democratic governability for the long-term. I argue something similar concerning the binomial system. It certainly did buy democracy a measure of security by providing strong incentives for coalition formation and maintenance that were functional and useful in the immediate post-authoritarian period. However, these benefits came with the understanding that the binomial system exacted a cost in terms of representation, competition, party turnover, internal democracy, and the impact of voters on outcomes. This cost-benefit analysis has now changed. In a transformed environment where democracy is now secure, the costs of the security and predictability created by the binomial system outweigh the benefits it provides. There are compelling reasons to make reform of the binomial system the final in a series of reforms to the Pinochet constitution, signaling the end of Chile's democratic transition and its arrival as a full-fledged modern democracy. These reasons will become clearer with further analysis of the other options on the table.


No electoral reform is better than a bad electoral reform. The adoption of a single member district (SMD), first-past-the-post system in Chile would be a less optimal choice than maintaining the binomial system. The debate on the SMD system for Chile plays into a long and unfortunate tendency of Chileans to think that "Made in the USA" means quality. Single member district systems are a dying breed and it would be absurd, just as the system is being abandoned world-wide, for Chile to step backward by adopting one. Across Latin America these systems have been discarded, and for good reasons. Indeed, among the two best known cases of SMD systems, one is experiencing deep problems including polarization, disgust with the two major parties and a crisis of legitimate representation (the United States), and the other is undergoing a process of rethinking and possible electoral reform (the UK).

A SMD system would leave Chile with most of the drawbacks of the binomial system, and add some new ones. It would provide few if any additional remedies for the most widely criticized aspects of the binomial system. Some of these drawbacks are associated with the essential traits of SMD systems, and others are associated with the adoption of this type of system in the Chilean context. There are two strains of argument for adopting a SMD system in Chile (and interestingly these arguments are contradictory). The first set of arguments asserts that a SMD system would lead to the formation of two parties by definitively fusing the parties of the Alianza and the Concertación. There is no reason to think that if the binomial system failed to lead to the fusion of parties that single member district system would do so. Parties continue to have strong individual identities, organizations and cultures, which would quite likely mean that the incentive to continue to simply strike coalition bargains rather than to fuse would remain even under a SMD system.

A second group of arguments addresses the criticism that the binomial system limits competition by providing the two major coalitions a lock on power. It rests on the assumption that a SMD system would lead to an opening of the party system because parties (or each of the three blocs of the tres tercios) would individually compete in single-member districts. This assumption is problematic on a number of counts. First, the simple adoption of a SMD system will not necessarily result in the dissolution of the pattern of two-alliance competition. Indeed, the irony in the basic argument of proponents of the SMD system is that some rely on Duvergerian arguments with respect to party consolidation (which would suggest that SMDs would solidify a two-party system), while others simultaneously contend that a SMD system would introduce real competition by prompting an electoral shake-up and competition between the three traditional blocs. There is little evidence to suggest that the SMD system will necessarily enhance the competitiveness of the system in this way. Parties will still have strong incentives to maintain their coalitions and continue to negotiate and to pass out seats, because a breakup of the current coalitional pattern would continue to present risks for parties that decide to go it alone. The major assumptions behind the rationale for the adoption of a SMD system are, therefore, contradictory, and at least one must be flawed.

SMD advocates should also recognize that problems of competitiveness have the potential to become worse with the adoption of a SMD system, given certain of its intrinsic characteristics.

SMD systems are very prone to gerrymandering. Indeed the US system has been so gerrymandered that it is estimated that of the 435 lower house seats only about 30 of them have been truly competitive in recent elections. Widespread gerrymandering provides a disincentive for voting because of the perception that participation makes no difference in outcomes. The current lock on power that the binomial system provides for certain parties has the potential to become much worse.

There are also other compelling reasons to avoid a SMD system. In terms of internal democracy and candidate selection, parties will continue to have an incentive to control nominations. While primaries can be used in SMD systems (and are in the United States), there is no guarantee that parties will move to these types of internal party contests. Indeed, the left and the center (and RN and UDI in the Alianza) can just as easily pass out seats and reach agreements of exclusion across districts in order to assure representation for all parties in their respective coalitions. In this situation, candidate selection will continue to be centrally directed and elitist.

Finally, and perhaps most importantly when it comes to competitiveness, if the center and left do, indeed, stay together the SMD system has the potential to completely shut the right out of the electoral game. That is precisely the reason military authorities did not adopt a SMD system, and this remains a compelling reason to avoid adopting one now. A SMD system will provide a tempting option and strong rational for the center and the left to remain allied. No good can come by excluding entire political sectors from representation. All in all, a SMD system is simply a bad idea for Chile and would represent a step backward for its democracy.


While single-member-district systems are decidedly out of fashion, mixed-member systems (MMS) are becoming increasingly popular for their purported ability to balance the benefits and drawbacks of SMD and PR systems. A mixed system would no doubt represent an improvement over the binomial system, and is a better choice than a SMD system. However, it is questionable how much of an additional benefit it provides over a well designed PR system, given the complexity it introduces for voters and parties and other problems it would create.

A mixed system would improve upon the binomial system by enhancing representation of Chile's smaller parties. While these parties have largely been represented under the binomial system (with the exception of those that do not form alliances like the Communist Party), they achieve representation after a very complex strategic negotiation to extract candidacies from the larger anchor parties of the two major coalitions. A mixed system would effectively eliminate this strategic dance and enhance representation by allowing Chile's smaller parties a voice in congress for those elected in the proportional seats.

However, after eliminating this particular brand of strategic dance, a MMS would introduce another one, and this time both at the level of the voter and elites. The problem with mixed systems is the complexity that they introduce on both levels. Simplicity is a virtue when it comes to election systems. Indeed, simplicity is perhaps one of the few virtues of the SMD system. Mixed election systems, though simple for electoral theorists, tend to be perceived by voters as complex, and even in long-standing democracies many typical voters have trouble explaining exactly how they work. When voters do not understand an election system it diminishes its legitimacy and that of the politicians elected using it.

MMSs are also strategically complex, and tend to encourage strategic vote-splitting (among those who do understand the system). This often produces difficult to predict outcomes and elicits strategic behavior by parties in campaigning and nominating candidates. Unexpected outcomes contribute to voters' perception that these systems somehow subvert the popular will (which subverts the perceived legitimacy of democracy). In turn, strategic nominating and campaigning behavior by their very nature tend to push candidate recruitment and nomination into the hands of party elites who seek to better manage electoral outcomes. Internal democracy is unlikely to be enhanced.

In essence, proponents of mixed systems fail to make a convincing argument that a MMS would be any better than a moderate magnitude PR system. In order to provide a convincing rationale for the adoption of a MMS, proponents must make a case that the introduction of such a complex system is justified by its payoffs. They usually contend that a MMS would provide a more direct connection between the electors and the elected than PR (at least in the SMD districts). However, mixed systems are more defensible in federal states like Germany, where proportional seats are still based on a defined and relatively small geographical unit, i.e. they are distributed at the Länd level. Chile's unitary system means that proportional seats would be distributed in single national constituency. Therefore, there will be little or no identification with any defined geographical area below the level of the country as a whole for these legislators. In proposals like the one elaborated by David Altman this would be a full 1/2 of the Chamber of Deputies. This raises a number of problems.

Two types of legislators are effectively created in such systems (PR legislators and SMD legislators). It also creates something of an electoral apartheid between candidates with and without districts, as national level deputies have no defined geographic electoral base to represent, which by definition means that small parties that can pass the national electoral thresholds will have no districts to represent, while large parties will. However, we must also remember that constitutionally Chilean legislators have a very limited capacity to secure particularistic benefits for their districts. Therefore, in the end, it is difficult to sustain that the creation of two types of legislators, one with strong constituent ties (and the inability to deliver any particularistic goods) and another with no ties to any district, represents an improvement over the balanced representative capacity of a PR system. A small magnitude PR system would produce a small number of legislators of equal status in each district, with ties to their constituency and reduced incentives to engage in the fruitless game of extracting particularistic resources. Therefore, most proposals fail to make convincing arguments that a MMS would provide substantial de-facto benefits over a moderate proportional system and certainly not a strong enough argument to justify the increased complexity that a MMS would introduce.


The adoption of a proportional system in Chile would signal the abandonment of fruitless electoral engineering aimed at reducing the number of parties, and simply recognize that Chile is a multiparty system. However, not all PR systems are the same, and herein lies the potential for electoral engineering aimed at inducing the optimal behavior of party leaders, candidates and politicians and producing clear electoral choices for voters.

The adoption of a moderate proportional system in Chile would provide the balance of representativeness, governability, accountability, and competition that any electoral system should guarantee. At the same time it would introduce many other of the desirable outcomes stressed here that the binomial system fails to promote. Opponents of PR often argue that it would bring about party system fragmentation and disincentives for party discipline as it has in Brazil. Let me dispense with the argument concerning hyper-fractionalization of the party system at the outset. PR is often rejected out of hand without recognizing that all PR systems are not the same. Indeed, the binomial system is a PR system—it just happens to be one with the lowest district magnitude possible. If the new Chilean PR system is designed correctly it can avoid the problems that exist in Brazil, and provide myriad additional benefits. I propose for Chile a moderate proportional system with open lists and magnitudes ranging from 4-6.

If the adoption of a proportional system would do nothing to fundamentally change the number of parties and nature of competition, why bother adopting it? A proportional system would enhance the functioning of Chile's multiparty system relative to the binomial systems in a number of ways. First, in terms of representativeness, a well designed PR system would provide representation in parliament for all of Chile's significant parties. Granted, these parties for the most part currently have parliamentary representation. However, they are represented precisely because they engage in a very elaborate process of negotiation in both alliances (that has sometimes come close to failing) in order to provide joint electoral lists that include all of their constituent parties. A moderate PR system would allow parties to simply present their lists of candidates without negotiation, and let voters decide.

This leads to the second benefit of moderate PR. It would further enhance representativeness by allowing voters rather than elites the most important voice in choosing candidates. Democracy is based on the principle that citizens choose their leaders. In Chile, a selectorate of party elites still exercises de facto control over who is nominated and, in turn, who wins. The binomial system privileges party elites because voters only choose between the two candidates put forth by a coalition in the district, and it is often a forgone conclusion who is going to win. Candidate slates are largely negotiated by party elites rather than responding to the popular will of voters. If a closed-list PR system were introduced parties would still dominate in the selection and placement of candidates. An open list system, on the other hand, would enhance the power of voters relative to the party. With an open list system, parties would nominate a number of candidates and voters would decide who among them ultimately would win. Indeed, with respect to internal party democracy a moderate PR system with open lists provides a good deal more input for voters into who ultimately is chosen to serve. Open lists also eliminate the need for primaries by effectively providing a simultaneous general election and primary, because voters decide who among a party's slate of candidates is elected.

The mention of open lists may raise another red flag for some. Open list systems are purported to undermine governability by decreasing party discipline. Once again district magnitude is the key. With small magnitude systems parties retain a good deal more control over who gets on the ballot, and incentives for party discipline will remain. However, this will be balanced by the ability of voters to choose among those candidates, or to sanction those who they feel are not representing their interests.

For those who oppose a PR option for Chile, governability is often the purported Achilles heel of this type of election system. Proponents of the binomial system point to the strong incentives the election system provides for coalitions, and how these coalitions have been at the core of Chilean presidents' ability to successfully legislate. The argument goes that without the binomial system, the very roots of the coalitions which have successfully competed and governed would be undermined. While I acknowledge that this is a strong argument, it is also somewhat reductionist, and more seriously, it takes too narrow a view of what governability means.

While Chile's impressive record of governability is due to a certain extent to the election system, it really goes a lot deeper. Presidents, ministers, legislators and party leaders made purposive efforts to counteract the potential disincentives for cooperation characterized by the uncomfortable combination between mulitpartism and presidentialism. Elites crafted a series of informal institutions and agreements that simultaneously kept the multiparty system alive and counteracted the purported negative features of multiparty presidentialism. These informal institutions include el cuoteo, el partido transversal, la democracia de los acuerdos and a series of others that facilitated coalition formation and successful democratic governance3. There is no reason to believe that these types of norms and informal institutions cannot be similarly and purposively constructed within the context of a more proportional election system.

Arguments concerning how the binomial system facilitates governability also rely on very short-term and constrained perceptions of what constitutes governability. Governability was extraordinary in Mexico during the period of PRI hegemony and during the experiences with pacted democracy in Venezuela and Colombia. However, a wiser view is a long-term one that recognizes that the simple ability to decisively pass laws is too narrow a view of what governability means. A system that is governable for the long-term takes into account the quality of representation, takes seriously the idea that voter preferences should matter, and provides for real competition between parties. With respect to long-term governability the Venezuelan and Colombian experiences with pact-making are telling. When parties are guaranteed a lock on power, short-term governability may indeed be enhanced, but long-term governability may be imperiled. In essence, Chile's two major coalitions have such a lock on power.

In terms of accountability, proponents of the binomial system contend that, unlike PR systems, the binomial system provides accountability because voters can clearly identify their representative. While this is true, it ignores another key aspect of accountability: the ability to remove a representative who is not performing up to task. By providing an effective lock on power for the two coalitions, rather than a boon to accountability, the binomial system is actually an impediment. Incumbents are very difficult to defeat in Chile. In rare cases can both candidates on a coalition's electoral list win, which makes it very difficult to single out a particular incumbent for defeat. Because votes are pooled, a vote for one candidate on a list is in many respects a vote for both. Suppose voters want to sanction the incumbent of their own coalition. They can, of course, opt for the candidate running with the incumbent on the same list. However, in doing so they may actually contribute votes to the incumbent because votes are pooled for the whole list. In this case voters must either abandon their ideological convictions and vote for another list, or grudgingly provide support for the entire list, potentially contributing support to the incumbent they would like to defeat. With a moderate PR system (with open lists) voters would be better able to hold their representatives accountable by removing them. Even if party elites insisted an incumbent should run, because open list systems allow preference voting, voters can more easily remove a badly performing deputy.

Finally, some argue that moving toward a proportional system is untenable because it would involve re-drawing Chile's entire electoral map. The map has, of course, already been re-drawn once. What is more, contrary to the assertions of opponents of proportional formulae, there have been several tenable proposals for redrawing the existing electoral map through combining districts already established under the binomial system that are entirely workable and would not involve extensive re-districting.


If Chile retains the binomial system democracy will survive. If it adopts a single member district or a proportional representation system, democracy will survive. In this sense, electoral reform does not matter because the good news is that the fundamentals of Chile's democratic success are based on much more than the coalitional configuration produced by the binomial system. Indeed, the military government's experiment in electoral engineering failed to reduce the number of political parties. One could even argue that democracy survived despite the binomial system rather than because of it. While coalition formation was the key to success, it was coalition formation aimed at providing representation for Chile's many parties by counteracting the intended reductive effects of the binomial system, which were supposed to provide stability and governability. In this sense, the debate on electoral reform becomes much less pressing because of the ability of parties to counteract the aims of electoral engineers.

However, this does not mean that electoral engineering has no effects. The debate on electoral reform is significant because electoral system change can go a long way toward enhancing the quality of Chilean democracy. Representation and stability are multidimensional and depend on much more than just the number of parties. Many argue that the two coalition configuration is the root of Chilean success, relying on traditional assumptions about the number of parties and stability. One could just as easily argue that the inclusion of Chile's many parties in the political game is just as central to that success, putting to rest the notion that the participation of many parties threatens governability. Electoral reform should, therefore, begin with the recognition that Chile is a multiparty system, and then be aimed at producing incentives to encourage other types of desirable outcomes: the inclusion of all significant parties in parliament without arduous negotiations, enhanced internal party democracy, open candidate selection, accountable members of congress, real competition between parties, and a long-term governability that goes beyond the simple ability of presidents to pass legislation. A small magnitude moderate proportional system with open lists will better provide the political incentives to achieve these goals than any other option on the table.


1 "Governability" (for right or wrong) is often used interchangeably with "stability" or "effectiveness" in this debate.

2 Magnitude refers to the number of seats per electoral district. Open list refers to ballot structure. Open lists allow the voter to opt for a particular candidate. Votes for all candidates are then pooled and winners are determined by the preference vote received by each candidate. Closed lists require the voter to opt for an entire party list, and then winners are determined by the order that candidates appear on the list. This order is usually determined by political parties.

3 The cuoteo refers to two things: the distribution of executive-appointed positions based on partisan colors especially within the ministries, and the quota of parliamentary candidacies allotted to each individual party within the Concertación for parliamentary elections. The partido transversal refers to a shifting but consistent informal group of leaders from the separate parties of the Concertación who define themselves more as "leaders of the Concertación" rather than leaders of their respective parties. They played key roles in the first democratic governments. Finally, the model of democracia de los acuerdos is a pattern of informal negotiations with the congressional opposition and with powerful social groups outside of congress to reach agreements on controversial, potentially destabilizing issues.

Peter M. Siavelis es profesor asociado de Ciencia Política en Wake Forest University, Estados Unidos. Es autor de The President and Congress in Post-authoritarian Chile: Institutional Constraints to Democratic Consolidation (Penn State Press, 2000), y de varios capítulos de libros y artículos en revistas profesionales. Su trabajo ha aparecido en Revista de Ciencia Política, Estudios Públicos, Comparative Politics y Comparative Political Studies. Su agenda de investigación actual se centra en el reclutamiento político y la selección de candidatos en América Latina.



Creative Commons License Todo el contenido de esta revista, excepto dónde está identificado, está bajo una Licencia Creative Commons