How not to write a constitution: lessons from Chile
PMCID: PMC9944366
PMID: 36844030
Abstract
On September 4, 2022, Chilean voters massively turned down a constitutional proposal that responded to widely shared criticisms of the 1980 constitution and emerged from a consensual and participatory process. This result is paradoxical because ex ante, the odds seemed largely in favor of changing the status quo. We argue that three factors, which derived from the interaction between rules and political contingency, explain the outcome: a Convention under the control of party-less independents, the exceptional underrepresentation of the political right, and a highly decentralized and public writing process. We extract some lessons from the failed experience that can be useful for countries seeking to deepen democratization through constitutional change and for a future constitution-making process in Chile.
Full Text
The results of the convention election on May 15 and 16, 2021, did indeed lead to a socially and culturally plural assembly, with an equal proportion of seats for men and women, representation from indigenous groups, and delegates with a wide diversity of educational backgrounds and professions. From the perspective of the individuals that composed it, the Chilean Convention was reasonably endowed with technical capability. For instance, there were 59 lawyers among the 155 members. There were also teachers, engineers, journalists, psychologists, doctors, actors, social workers, entrepreneurs, an economist, a political scientist, a farmer, an accountant, a business administrator, a dentist, an architect, a geographer, and even a chess player. Only nine delegates lacked any form of completed tertiary education. From the point of view of institutional experience, there were six former members of parliament and nine former authorities of the executive branch.1
As shown in Table 1, the center-right and right coalition (Vamos por Chile) was the most voted list and obtained 37 seats (24%). The leftist coalition (Lista Apruebo Dignidad) outperformed the traditional center-left (Lista del Apruebo) that had governed Chile from 1990 to 2010; they obtained 28 and 25 seats, respectively. Among the independents and independent lists, which together obtained 48 seats, the vast majority was also located on the center-left or left of the ideological spectrum. The most salient list among them was the Lista del Pueblo, a leftist group linked to various social movements involved in the 2019 protests, which obtained 24 seats, followed by a more centrist group, Independientes no Neutrales, with 11 seats. While the ideological preferences of indigenous representatives were mixed, the majority had left of center or leftist views. At the time of the convention election, in May 2021, it was unclear whether this ideological balance signaled a lasting shift to the left in the distribution of political preferences among voters. As it turned out, however, the convention election was indeed exceptional and not part of a new trend.
At the time of the convention election, the performance of parties on the center-right and right was affected by the negative image of Piñera´s presidency, which they supported. Piñera was experiencing very low levels of approval and increasingly high levels of public disapproval since late 2019, first due to the government´s mishandling of social mobilizations and public order and then to the negative economic and social effects of the pandemic.2 In addition, the purpose of the convention election was to select delegates who would approve reforms to the existing constitution. Yet the center-right and right parties had been traditionally against changing the 1980 constitution and had no reform proposals to mobilize the electorate. Finally, the convention election, like the initial referendum to authorize the replacement of the constitution, was held under voluntary voting so that those most disaffected with the status quo were the most likely to vote. In total, only 41.5% of the electorate participated.
The confirmation that the convention election was exceptional was revealed in the legislative and presidential elections held in November 2021, six months after the Convention was formed. In the legislative elections, the center-right and right parties obtained 36.6% of the vote and 43.9% of the seats in the Chamber of Deputies, a proportion very similar to the 36.7% of the vote and the 46.4% of deputies they had obtained in the 2017 legislative election.3 In the presidential election, the main contender to the center-left candidate Gabriel Boric was the rightist Jose Antonio Kast, who ended up first in the first round but lost in the runoff.
With the composition, rules, and structure described in the previous section, the Convention started to work within each committee by the end of 2021. Between March and May 2022, the plenary voted on the different reports of each committee, in a complex iteration between plenary and committees. The draft was completed in June and published on July 4. It was a lengthy constitutional proposal (even by Latin American standards) of 388 provisions.4
As shown in Fig. 1, the mean level of public support for approval of the new constitution in the future ratification referendum was 56% in January 2022, before the first reports with provisions approved in the committees were known. However, between February and May 2022, when committee reports were submitted to the plenary and the first draft of the constitution was taking shape, support for the new constitution declined from 47 to 35%. Meanwhile, the percentage of those surveyed willing to reject the proposal increased steadily. A similar trend occurred with levels of public trust in the Convention, which went from 51 to 43% between January and May 2022.5 In spite of the fact that the proposal completed in June ended up responding to long-standing social demands in Chile, these trends anticipated the defeat of the proposal.
Approval and rejection rates according to opinion polls. Source: Authors, based on Aylwin et al. (2022)
Writing a constitution is an act of cooperative decision-making that involves several mechanisms to aggregate or transform individual preferences into a collective choice (see Elster 1995; Voigt, 1997). The traditional organizations that act as vehicles for the articulation of preferences of constitution makers are political parties. Parties provide positions across different issues and dimensions. Party leaders, in turn, forge compromises and organize voting. As party representatives are likely to be present not only in a special body temporarily responsible for drafting the constitution but also in institutions at the post-constitutional stage that will implement the new constitution over time, they have a stake in both the design and enforcement of the new constitution.6 In addition, if the draft constitution must be ratified in a referendum, parties could be central for voter mobilization in support of the draft version.
After democracy was recovered, political parties in Chile enjoyed massive support. According to the World Values Survey, between 1989 and 1993, half of the population had confidence in political parties. This support has suffered steadily since then. When the survey was conducted for the 7th Wave between 2017 and 2022, support had declined to 17.6%.7 Several rules for the election of the Convention aggravated this situation. The most important was the rule enabling the vote for lists of independents. Groups of independents lacked organization and consistent reform programs. Most independents were single-issue activists seeking to participate in the Convention as environmentalists, feminists, or traditionalists rather than as agents responsible for negotiating across multiple, complex dimensions. For this reason, groups of independents tended to fragment further in the convention, even when they were initially elected in a single list.
In theory, there are arguments in favor of electing delegates to a constitutional convention on an individual basis. In his treatise on constitutional democracy, Mueller (1996), for example, argues that electing individuals rather than parties is likely to lead to better results as parties would have particular interests after the convention which might influence their behavior during the convention’s deliberation. In practice, however, electing non-partisan delegates does not guarantee their impartiality and may reinforce a tendency of independents to focus on the single issues they advocated during the electoral campaign. Agreeing on a constitution necessitates many negotiations and compromises which are difficult to attain if delegates only care about isolated, discrete reforms.
The most immediate consequence of the weak representation of the right was that it was unable to block any reform and thus induce negotiations because it did not reach by itself the required 33% + 1 votes. Had the right achieved that threshold the constitutional proposal would have been less radical in content. In addition, divisions within parties on the right and potential allies in the traditional center-left limited its capacity to build a coalition such that it would reach the needed threshold to veto proposals. The alternatives for the right were restricted to negotiating with a portion of moderates of the traditional center-left gathered in the Lista del Apruebo and eventually some independents. However, in order to agree on something with them, the most moderate within the right-wing coalition would have lost the support of their far-right colleagues. Indeed, the right was divided in two subgroups, one that showed interest in starting negotiations with the center left, and the other with no intention to bargain at all.8 On the other hand, the most important group within the traditional center-left, the Socialist Party, decided early on to negotiate reforms with the left parties in the list Apruebo Dignidad because together they were in a better position to win allies to approve reforms by two-thirds in the plenary.
The two-thirds voting rule to pass any constitutional provision was adopted to secure consensus. This is consistent with the idea that constitutions are the most important formal component of social contracts. Ideally, all citizens who would be subject to a new constitution should consent to it. Unanimous agreement to a constitutional draft is, however, extremely unlikely.9 If every delegate has veto power, the temptation to misuse that veto power strategically is very large, which is why supermajorities for the passing of draft constitutions have been suggested (Wicksell, 1896, e.g., proposes a 5/6 majority). Although supermajority rule is not frequently required, it has been used in several recent constitution-making processes, such as those of South Africa, Tunisia, and Nepal (see Negretto, 2017).10 In the Chilean case, different voting and approval rules were used. For the final plebiscite, a 50% of valid votes was sufficient. If one thinks of the constitution as a social contract binding all citizens, making the draft pass if a simple majority of citizens approves it appears questionable (Michel & Cofone, 2017) yet it has become somewhat of a convention and is part of the Chilean constitution-making tradition.
The proposal to eliminate the Senate by the committee on the political system provides an appropriate example. When this committee approved the creation of a unicameral legislature by 13 to 12 votes in mid-February 2022, it was apparent that such a proposal would not be approved in the plenary.11 Moreover, the first report submitted to the plenary had already abandoned the unicameral legislature and included instead a very weak, almost symbolic second chamber representative of the regions. This provision, along with many others, was rejected by the plenary and sent back to the committee for further negotiations. After subsequent iterations of voting, a stronger second chamber was eventually included in the draft constitution. But the initial unicameral proposal polarized the debate from the very beginning and set the agenda for discussion in future negotiations. The frequent rejection in the plenary of provisions approved in the committees also created an image of disorder and conflict in the eyes of the public, aggravated by the fact that integration of the different committee reports into a single draft did not take place until the very end of the process.
However, if all positions ever taken by all members of the constitutional assembly become public knowledge, this makes it more difficult for assembly members to change their mind. In other words, achieving compromise and consensus on a draft could become more difficult if deliberations are public. Elster (1999) claims that publicity explains some of the failures of the post-1789 French constitutional assembly.
Yet, it would be misleading to think that a majority of voters in the ratification referendum opted to maintain the 1980 constitution. Although formally, this was the choice and the result of the rejection, it turns out that Chilean voters perceived a third option. Polls taken after the defeat of the proposed constitution showed that 79% were still in favor of changing the constitution.12 This situation is markedly different from a setting in which a country has no valid constitution or one in which there is a shared understanding that the existing constitution is clearly dysfunctional. In these cases, the ex-ante likelihood of a proposal being accepted is very high. But it is also different from a setting in which voters opt between a proposal of dubious quality and an existing functional constitution.13 Chileans still want another constitution but not just any other one. For a majority of the voters, the far-reaching changes of the proposal were unpersuasive and generated deep doubts about their implementation. The third option was to be neither in favor of keeping the 1980 constitution nor of the draft constitution but in favor of yet another constitutional assembly.
Procedural rules should guarantee that important minority positions cannot simply be outvoted. Voting on the constitution by two-thirds of the Convention was a reasonable rule in Chile. Yet due to its weak electoral performance, the center-right and right list did not command veto power despite being the single largest group. This situation, however, could have been improved by allocating authority positions within the Convention and the committees according to the relative political weight of each group. This would have given representatives on the right the incentives and opportunity to present their proposals, shape the agenda for discussion, and eventually seek agreements. Yet an assembly biased to the left was unwilling to provide this group with such an influence.14
Sections
"[{\"pmc\": \"PMC9944366\", \"pmid\": \"36844030\", \"reference_ids\": [\"Fn1\"], \"section\": \"The convention: composition and rules\", \"text\": \"The results of the convention election on May 15 and 16, 2021, did indeed lead to a socially and culturally plural assembly, with an equal proportion of seats for men and women, representation from indigenous groups, and delegates with a wide diversity of educational backgrounds and professions. From the perspective of the individuals that composed it, the Chilean Convention was reasonably endowed with technical capability. For instance, there were 59 lawyers among the 155 members. There were also teachers, engineers, journalists, psychologists, doctors, actors, social workers, entrepreneurs, an economist, a political scientist, a farmer, an accountant, a business administrator, a dentist, an architect, a geographer, and even a chess player. Only nine delegates lacked any form of completed tertiary education. From the point of view of institutional experience, there were six former members of parliament and nine former authorities of the executive branch.1\"}, {\"pmc\": \"PMC9944366\", \"pmid\": \"36844030\", \"reference_ids\": [\"Tab1\"], \"section\": \"The convention: composition and rules\", \"text\": \"As shown in Table 1, the center-right and right coalition (Vamos por Chile) was the most voted list and obtained 37 seats (24%). The leftist coalition (Lista Apruebo Dignidad) outperformed the traditional center-left (Lista del Apruebo) that had governed Chile from 1990 to 2010;\\u00a0they obtained 28 and 25 seats, respectively. Among the independents and independent lists, which together obtained 48 seats, the vast majority was also located on the center-left or left of the ideological spectrum. The most salient list among them was the Lista del Pueblo, a leftist group linked to various social movements involved in the 2019 protests, which obtained 24 seats, followed by a more centrist group, Independientes no Neutrales, with 11 seats. While the ideological preferences of indigenous representatives were mixed, the majority had left of center or leftist views. At the time of the convention election, in May 2021, it was unclear whether this ideological balance signaled a lasting shift to the left in the distribution of political preferences among voters. As it turned out, however, the convention election was indeed exceptional and not part of a new trend.\"}, {\"pmc\": \"PMC9944366\", \"pmid\": \"36844030\", \"reference_ids\": [\"Fn2\"], \"section\": \"The convention: composition and rules\", \"text\": \"At the time of the convention election, the performance of parties on the center-right and right was affected by the negative image of Pi\\u00f1era\\u00b4s presidency, which they supported. Pi\\u00f1era was experiencing very low levels of approval and increasingly high levels of public disapproval since late 2019, first due to the government\\u00b4s mishandling of social mobilizations and public order and then to the negative economic and social effects of the pandemic.2 In addition, the purpose of the convention election was to select delegates who would approve reforms to the existing constitution. Yet the center-right and right parties had been traditionally against changing the 1980 constitution and had no reform proposals to mobilize the electorate. Finally, the convention election, like the initial referendum to authorize the replacement of the constitution, was held under voluntary voting so that those most disaffected with the status quo were the most likely to vote. In total, only 41.5% of the electorate participated.\"}, {\"pmc\": \"PMC9944366\", \"pmid\": \"36844030\", \"reference_ids\": [\"Fn3\"], \"section\": \"The convention: composition and rules\", \"text\": \"The confirmation that the convention election was exceptional was revealed in the legislative and presidential elections held in November 2021, six months after the Convention was formed. In the legislative elections, the center-right and right parties obtained 36.6% of the vote and 43.9% of the seats in the Chamber of Deputies, a proportion very similar to the 36.7% of the vote and the 46.4% of deputies they had obtained in the 2017 legislative election.3 In the presidential election, the main contender to the center-left candidate Gabriel Boric was the rightist Jose Antonio Kast, who ended up first in the first round but lost in the runoff.\"}, {\"pmc\": \"PMC9944366\", \"pmid\": \"36844030\", \"reference_ids\": [\"Fn4\"], \"section\": \"Rules and politics interacting\", \"text\": \"With the composition, rules, and structure described in the previous section, the Convention started to work within each committee by the end of 2021. Between March and May 2022, the plenary voted on the different reports of each committee, in a complex iteration between plenary and committees. The draft was completed in June and published on July 4. It was a lengthy constitutional proposal (even by Latin American standards) of 388 provisions.4\"}, {\"pmc\": \"PMC9944366\", \"pmid\": \"36844030\", \"reference_ids\": [\"Fig1\", \"Fn5\"], \"section\": \"Rules and politics interacting\", \"text\": \"As shown in Fig.\\u00a01, the mean level of public support for approval of the new constitution in the future ratification referendum was 56% in January 2022, before the first reports with provisions approved in the committees were known. However, between February and May 2022, when committee reports were submitted to the plenary and the first draft of the constitution was taking shape, support for the new constitution declined from 47 to 35%. Meanwhile, the percentage of those surveyed willing to reject the proposal increased steadily. A similar trend occurred with levels of public trust in the Convention, which went from 51 to 43% between January and May 2022.5 In spite of the fact that the proposal completed\\u00a0in June ended up responding to long-standing social demands in Chile, these trends anticipated the defeat of the proposal.\"}, {\"pmc\": \"PMC9944366\", \"pmid\": \"36844030\", \"reference_ids\": [\"CR1\"], \"section\": \"\", \"text\": \"Approval and rejection rates according to opinion polls. Source: Authors, based on Aylwin et al. (2022)\"}, {\"pmc\": \"PMC9944366\", \"pmid\": \"36844030\", \"reference_ids\": [\"CR13\", \"CR11\", \"Fn6\"], \"section\": \"A non-partisan convention\", \"text\": \"Writing a constitution is an act of cooperative decision-making that involves several mechanisms to aggregate or transform individual preferences into a collective choice (see Elster 1995; Voigt, 1997). The traditional organizations that act as vehicles for the articulation of preferences of constitution makers are political parties. Parties provide positions across different issues and dimensions. Party leaders, in turn, forge compromises and organize voting. As party representatives are likely to be present not only in a special body temporarily responsible for drafting the constitution but also in institutions at the post-constitutional stage that will implement the new constitution over time, they have a stake in both the design and enforcement of the new constitution.6 In addition, if the draft constitution must be ratified in a referendum, parties could be central for voter mobilization in support of the draft version.\"}, {\"pmc\": \"PMC9944366\", \"pmid\": \"36844030\", \"reference_ids\": [\"Fn7\"], \"section\": \"A non-partisan convention\", \"text\": \"After democracy was recovered, political parties in Chile enjoyed massive support. According to the World Values Survey, between 1989 and 1993, half of the population had confidence in political parties. This support has suffered steadily since then. When the survey was conducted for the 7th Wave between 2017 and 2022, support had declined to 17.6%.7 Several rules for the election of the Convention aggravated this situation. The most important was the rule enabling the vote for lists of independents. Groups of independents lacked organization and consistent reform programs. Most independents were single-issue activists seeking to participate in the Convention as environmentalists, feminists, or traditionalists rather than as agents responsible for negotiating across multiple, complex dimensions. For this reason, groups of independents tended to fragment further in the convention, even when they were initially elected in a single list.\"}, {\"pmc\": \"PMC9944366\", \"pmid\": \"36844030\", \"reference_ids\": [\"CR6\"], \"section\": \"A non-partisan convention\", \"text\": \"In theory, there are arguments in favor of electing delegates to a constitutional convention on an individual basis. In his treatise on constitutional democracy, Mueller (1996), for example, argues that electing individuals rather than parties is likely to lead to better results as parties would have particular interests after the convention which might influence their behavior during the convention\\u2019s deliberation. In practice, however, electing non-partisan delegates does not guarantee their impartiality and may reinforce a tendency of independents to focus on the single issues they advocated during the electoral campaign. Agreeing on a constitution necessitates many negotiations and compromises which are difficult to attain if delegates only care about isolated, discrete reforms.\"}, {\"pmc\": \"PMC9944366\", \"pmid\": \"36844030\", \"reference_ids\": [\"Fn8\"], \"section\": \"Underrepresentation and exclusion of the right\", \"text\": \" The most immediate consequence of the weak representation of the right was that it was unable to block any reform and thus induce negotiations because it did not reach by itself the required 33%\\u2009+\\u20091 votes. Had the right achieved that threshold the constitutional proposal would have been less radical in content. In addition, divisions within parties on the right and potential allies in the traditional center-left limited its capacity to build a coalition such that it would reach the needed threshold to veto proposals. The alternatives for the right were restricted to negotiating with a portion of moderates of the traditional center-left gathered in the Lista del Apruebo and eventually some independents. However, in order to agree on something with them, the most moderate within the right-wing coalition would have lost the support of their far-right colleagues. Indeed, the right was divided in two subgroups, one that showed interest in starting negotiations with the center left, and the other with no intention to bargain at all.8 On the other hand, the most important group within the traditional center-left, the Socialist Party, decided early on to negotiate reforms with the left parties in the list Apruebo Dignidad because together they\\u00a0were in a\\u00a0better position to win allies to approve reforms by two-thirds in the plenary.\"}, {\"pmc\": \"PMC9944366\", \"pmid\": \"36844030\", \"reference_ids\": [\"Fn9\", \"CR12\", \"CR8\", \"Fn10\", \"CR5\"], \"section\": \"A decentralized and public writing process\", \"text\": \"The two-thirds voting rule to pass any constitutional provision was adopted to secure consensus. This is consistent with the idea that constitutions are the most important formal component of social contracts. Ideally, all citizens who would be subject to a new constitution should consent to it. Unanimous agreement to a constitutional draft is, however, extremely unlikely.9 If every delegate has veto power, the temptation to misuse that veto power strategically is very large, which is why supermajorities for the passing of draft constitutions have been suggested (Wicksell, 1896, e.g., proposes a 5/6 majority). Although supermajority rule is not frequently required, it has been used in several recent constitution-making processes, such as those of South Africa, Tunisia, and Nepal (see Negretto, 2017).10 In the Chilean case, different voting and approval rules were used. For the final plebiscite, a 50% of valid votes was sufficient. If one thinks of the constitution as a social contract binding all citizens, making the draft pass if a simple majority of citizens approves it appears questionable (Michel & Cofone, 2017) yet it has become somewhat of a convention and is part of the Chilean constitution-making tradition.\"}, {\"pmc\": \"PMC9944366\", \"pmid\": \"36844030\", \"reference_ids\": [\"Fn11\"], \"section\": \"A decentralized and public writing process\", \"text\": \"The proposal to eliminate the Senate by the committee on the political system provides an appropriate example. When this committee approved the creation of a unicameral legislature by 13 to 12 votes in mid-February 2022, it was apparent that such a proposal would not be approved in the plenary.11 Moreover, the first report submitted to the plenary had already abandoned the unicameral legislature and included instead a very weak, almost symbolic second chamber representative of the regions. This provision, along with many others, was rejected by the plenary and sent back to the committee for further negotiations. After subsequent iterations of voting, a stronger second chamber was eventually included in the draft constitution. But the initial unicameral proposal polarized the debate from the very beginning and set the agenda for discussion in future negotiations. The frequent rejection in the plenary of provisions approved in the committees also created an image of disorder and conflict in the eyes of the public, aggravated by the fact that integration of the different committee reports into a single draft did not take place until the very end of the process.\"}, {\"pmc\": \"PMC9944366\", \"pmid\": \"36844030\", \"reference_ids\": [\"CR4\"], \"section\": \"A decentralized and public writing process\", \"text\": \"However, if all positions ever taken by all members of the constitutional assembly become public knowledge, this makes it more difficult for assembly members to change their mind. In other words, achieving compromise and consensus on a draft could become more difficult if deliberations are public. Elster (1999) claims that publicity explains some of the failures of the post-1789 French constitutional assembly.\"}, {\"pmc\": \"PMC9944366\", \"pmid\": \"36844030\", \"reference_ids\": [\"Fn12\", \"Fn13\"], \"section\": \"A decentralized and public writing process\", \"text\": \"Yet, it would be misleading to think that a majority of voters in the ratification referendum opted to maintain the 1980 constitution. Although formally, this was the choice and the result of the rejection, it turns out that Chilean voters perceived a third option. Polls taken after the defeat of the proposed constitution showed that 79% were still in favor of changing the constitution.12 This situation is markedly different from a setting in which a country has no valid constitution or one in which there is a shared understanding that the existing constitution is clearly dysfunctional. In these cases, the ex-ante likelihood of a\\u00a0proposal being accepted is very high. But it is also different from a setting in which voters opt between a proposal of dubious quality and an existing functional constitution.13 Chileans still want another constitution but not just any other one. For a majority of the voters, the far-reaching changes of the proposal were unpersuasive and generated deep doubts about their implementation. The third option was to be neither in favor of keeping the 1980 constitution nor of the draft constitution but in favor of yet another constitutional assembly.\"}, {\"pmc\": \"PMC9944366\", \"pmid\": \"36844030\", \"reference_ids\": [\"Fn14\"], \"section\": \"\", \"text\": \"Procedural rules should guarantee that important minority positions cannot simply be outvoted. Voting on the constitution by two-thirds of the Convention was a reasonable rule in Chile. Yet due to its weak electoral performance, the center-right and right list did not command veto power despite being the single largest group. This situation, however, could have been improved by allocating authority positions within the Convention and the committees according to the relative political weight of each group. This would have given representatives on the right the incentives and opportunity to present their proposals, shape the agenda for discussion, and eventually seek agreements. Yet an assembly biased to the left was unwilling to provide this group with such an influence.14\"}]"
Metadata
"{\"issue-copyright-statement\": \"\\u00a9 Springer Science+Business Media, LLC, part of Springer Nature 2023\"}"