Introduction
Over the last two decades, compulsory voting has acquired prominence in a few South American countries. In Chile, the legal obligation to vote regained the spotlight in 2022, ten years after its abolition, as it was used during the process of constitutional reform. Then, Law #21524 amended the Constitution to re-establish mandatory voting for elections and plebiscites.Footnote 1 Although discussions on compulsory voting have been especially salient in Chile, they have also come to the fore in Peru, where some governments have contemplated its abrogation while Law #28859-2006 updated its sanctions, and in Argentina, where it was extended to primaries and steps were taken to strengthen its enforcement through Law #26571-2009 and Law #26774-2012, respectively.Footnote 2
Social scientists’ interest in mandatory voting has surged in parallel with the foregoing and similar developments elsewhere, producing very informative work.Footnote 3 However, scholarship has been overwhelmingly concerned with the effects of this institution, while neglecting its causes. This article seeks to start filling the gap by studying the introduction of compulsory voting in Latin America during democratisation, an intriguing phenomenon from the viewpoint of mainstream theories in political economy.
Definitions
Compulsory voting (hereafter CV) is the legal obligation of citizens to attend the polls at election time, which carries specific penalties delineated in the legislation for those who fail to do so without a valid excuse.Footnote 4 It has a long history, going as far back as the Middle Ages, but it enjoyed a revival in nineteenth-century Europe, when it was instituted in Belgium, Liechtenstein and several Swiss cantons.Footnote 5 In Latin America, the pioneer was Peru, which introduced mandatory voting in 1824, within a system of corporatist representation with a restricted franchise and indirect elections.Footnote 6 Many years later, in 1912, Argentina became the first country in the region to adopt the legal obligation to vote in a democratising context.
Democratisation is here understood as the process of establishing a representative democracy, i.e. a political regime with high levels of contestation and inclusiveness (in practice, this means access to power through free, fair and competitive elections in which most of the adult population is allowed to vote).Footnote 7 Historically, democratisation has sometimes entailed the simultaneous introduction of all key components of democracy (e.g. some of the transitions from dictatorship in Southern Europe and Latin America in the last quarter of the twentieth century). At other times, progress along the two dimensions, contestation and inclusiveness, has been incremental, with the constituent parts of democracy being set up through separate episodes of institutional change over a longer time span (e.g. Great Britain and pre-1973 Chile).Footnote 8 In keeping with most of the scholarship in comparative politics, I use the term democratisation to refer to both of these scenarios: the wholesale replacement of autocracy by democracy, and the implementation of smaller-scale reforms that make elections fairer, more competitive and more inclusive (for instance, by protecting ballot secrecy; creating independent authorities to adjudicate disputes; lifting bans on parties and allowing them to field candidates and campaign; extending the suffrage; etc.).
In point of the research question at hand, the introduction of CV along with democratisation, it is worth noting that sometimes this was literally the case, insofar as mandatory voting was enacted at the same time as democratising institutional changes. However, it would be too restrictive to set simultaneity as a requirement for falling within the purview of this study. In the instances of wholesale regime transition, the establishment of the legal obligation to vote in the run-up to the founding elections (e.g. Venezuela in 1958), or somewhat earlier in the transitional phase, also qualifies as CV adoption in a democratising context. In the case of discrete democratising reforms, mandatory voting may be established at the same time (e.g. Argentina in 1912) or shortly afterwards.Footnote 9
Compulsory Voting in Latin America
Latin America hosts 31 per cent of the polities that have adopted CV since 1800, and in terms of the relationship between mandatory voting and political regime, it features three subsets of countries.Footnote 10 First, those that introduced CV along with democratisation, having no previous experience with this institution under autocracy (Argentina and Chile, whose case studies are presented below, Ecuador and Venezuela).Footnote 11 Second, those that adopted and practised mandatory voting only under autocracy (e.g. Costa Rica and El Salvador). And, third, the polities that established the legal obligation to vote under autocracy in the nineteenth or early twentieth century, used it extensively under authoritarian regimes of different stripes and kept it through their transition to democracy (e.g. Bolivia, Brazil, Paraguay and Honduras until 2004).Footnote 12 Although the puzzle posed by the introduction of mandatory voting along with democratisation that is described below might also apply to this last group of cases, they are qualitatively dissimilar to the first subset and their study would require a different research design (in particular, one probing their experiences with CV under autocracy and contrasting persistence with abolition at the juncture of the transition to democracy).
Puzzle
If, as held by influential theories in political economy, democracy is costly to the governing elites inasmuch as it transfers political power to the citizens and is also likely to redistribute income, and if these consequences are expected to be reinforced by effective turnout, why would incumbents ever adopt mandatory voting along with democratisation?Footnote 13
Below, I address this question through a comparative study of three South American countries. I begin by elaborating on my explanation and competing hypotheses. I subsequently depict my methodological approach and present the case studies. In closing, I briefly review current trends and offer directions for future research.
Argument
My explanation of CV adoption in democratising polities rests on two implications of what this institution can reasonably be expected to do, i.e. increase turnout. The first logic is pro-democratic, even when driven by strategic motivations, because it aims to enhance electoral probity. The second, in turn, seeks to offer incumbents some protection from the effects of democratisation.
There is little mandatory voting can do for electoral integrity in countries with sham elections (i.e. with massive fraud organised by the state). However, in those in which such systematic ballot rigging is absent or measures have been taken to tackle it, as is often the case in democratising polities, CV can help. In places with decentralised electoral corruption (vote buying, impersonation, multiple voting, etc.), a large(r) actual electorate can reduce the impact of such irregularities on election results and, if high turnout is sustained over time, may eventually discourage them.Footnote 14 Therefore, the legal obligation to vote may be established in such contexts out of a belief that, by bringing many people to the polls, it can undermine the cost-effectiveness and viability of electoral malfeasance. Mandatory voting may be adopted along with democratisation because this line of reasoning prompts the opposition to push for it in its negotiations with the government, or because the latter finds that this rationale is in its own interest.
The second logic is inspired by the notion of differences in the intensity of political preferences.Footnote 15 Scholarship has shown that most people are not well informed about politics; that the probability of voting is lower among those without strong opinions about public policies or issues; and that citizens with weak partisanship are more likely to abstain.Footnote 16 What the introduction of a legal obligation to vote does is to add (many of) these individuals to the pool of voters. Hence, incumbents may adopt mandatory voting to meet the electoral challenge posed by an opposition capable of generating intense loyalties on the basis of a programme, an ideology or an identity. In the context of democratisation, citizens with such strong political commitments are very likely to vote, and, given low turnout, could significantly affect the outcome of elections under voluntary voting. Those in government can hedge against this possibility by legally requiring all eligible voters to attend the polls, anticipating that, by making the bulk of the citizenry vote, CV will mobilise an indifferent, uninformed, less politicised and/or moderate mass that could drown out the ballots of the highly motivated. According to this second logic, incumbents enact mandatory voting along with democratisation as an exercise in damage limitation.
Whether it is the first or the second logic, or a combination of both, that accounts for the establishment of the legal obligation to vote in a particular case depends mostly on the characteristics of the incumbents relative to their competitors – (more/less) clientelistic; (more/less) programmatic – as well as on the degree of concentration of decision-making. For some adopters of CV, one of the logics may be irrelevant (e.g. if decentralised electoral corruption is not significant or concerning to them), whereas for others, both may be aligned with their goals. Some actors facing large trade-offs between the two logics may choose to prioritise one over the other (especially if the structure of decision-making allows them to engage in horse trading) but for others the cons of one may outweigh the pros of the other, leading them to not support or to reject mandatory voting.
Competing Explanations
It has often been taken for granted that the legal obligation to vote was instituted to perfect the democratic franchise by building citizenship. Depending on how it is framed, this idea may be compatible with an efficiency interpretation of institutions, or with the sociological strand of new institutionalism, according to which the introduction of CV would have been a principled decision.Footnote 17
Alternatively, in cross-national work on the origins of mandatory voting, Gretchen Helmke and Bonnie Meguid propose that it was established to promote the electoral participation of a set of non-voters that formed the government’s natural base when that of the opposition was already voting and had reached its full potential or was smaller.Footnote 18 They explain that the legal requirement to vote was typically introduced when the lower classes were enfranchised and socialist parties were on the rise, so incumbents meant to use it to mobilise non-industrial workers in order to outdo the socialists. This is an interesting idea and may have been the background of CV adoption in some European countries, but mandatory voting is a blunt tool and such calculations of the relative sizes of natural bases and/or of full potential seem implausible in many cases due to a changing social environment and simultaneous reforms. Moreover, even in Europe, the legal obligation to vote was championed by both right- and left-wing parties, adopted in some instances with support from those that this hypothesis singles out as the intended losers, and also instituted in response to non-leftist challengers.Footnote 19
The second logic in my argument differs from the above in that it posits that, in the context of democratisation, rather than being introduced to target a well-defined set of non-voters that constituted the incumbents’ natural base, CV was adopted to make a large mass of citizens vote and thereby undercut the mobilisational advantage enjoyed by parties with intensely committed followers. Besides being a crude instrument for electoral engineering, the establishment of mandatory voting during democratisation was characterised by concurrent changes, strain and uncertainty. My explanation is thus not about incumbents identifying their as yet unmobilised constituencies and gauging their size vis-à-vis those of the opposition, but about bringing multitudes to the polls in order to drown out the votes of citizens with solid political preferences and loyalties.
Methodology
A combination of cross-case comparison of a small number of countries and within-case analysis is well suited to the close examination of the causal mechanisms that led to mandatory voting in democratising polities. I therefore rely on archival research to probe political actors’ beliefs, expectations and stated preferences with regard to CV, as well as to get a more accurate picture of the decision-making process.
I have chosen three South American countries for fieldwork on the basis of variation on the dependent variable: a polity that has never introduced the legal obligation to vote, Colombia, and two that did so when democratising, Argentina and Chile. The former constitutes an instance where the electoral challenge to the incumbent came from a non-leftist party. The latter has had a convoluted history with mandatory voting, but illustrates – at least on one occasion – its use against the Communist–Socialist bloc.
Empirical Analysis
Argentina
The period from 1880 to 1916 is frequently referred to as the Conservative Order. It pivoted around the Partido Autonomista Nacional (National Autonomist Party, PAN), an official political organisation that lacked a disciplined, hierarchical structure and was exceedingly personalistic. The so-called Regime has been characterised on the basis of three core elements: (1) de facto control of presidential succession by the federal executive and provincial governors; (2) ballot rigging amid universal suffrage for Argentine men; and (3) the restriction of effective competition to factions within the PAN or insiders. It follows from these and other features that, although citizens voted, it was the incumbents that actually chose (i.e. they were elector governments).Footnote 20
New political forces emerged in the 1890s, notably the Unión Cívica Radical (Radical Civic Union, UCR). In 1905, under Hipólito Yrigoyen’s leadership, it called for electoral abstention until free and fair elections were held and organised a failed armed revolt that sparked concern in the government because it showed that support for the Radicals was growing.
As for intra-PAN politics, President Figueroa Alcorta (1906–10) asserted his authority by ordering the federal intervention of several provinces to dislodge his opponents. In so doing, he weakened the old guard of Conservatism and paved the way for the reformists, whose leader he ended up endorsing as his successor. In April 1910, with the president’s backing, Roque Sáenz Peña (hereafter RSP) won an uncontested election.
President RSP and his fellow reformists pursued two goals: getting the Radicals to give up insurrections and election boycotts, and cementing their control of the Conservative camp in order to renew its political offer.Footnote 21 They therefore negotiated an electoral reform with the former without losing sight of the need to contain and co-opt their remaining adversaries among the Conservatives.
Soon after his election, RSP initiated talks with Yrigoyen in order to avert new Radical uprisings. These negotiations bore fruit: the laws that made up the so-called Sáenz Peña reform adopted a new voters’ register based on the military roster, CV, provisions to protect ballot secrecy and curb electoral manipulation, and a less majoritarian electoral system (the incomplete list).Footnote 22
Minority representation was a Radical demand that, together with the new electoral register and measures to clean up elections, constituted a sine qua non for abandonment of their abstention and insurgency tactics.Footnote 23 By contrast, the UCR did not advocate for mandatory voting. Yrigoyen did not do so in his meetings with RSP, nor had the party done so in public speeches or internal documents: this institution was put forth by the incumbent.Footnote 24
Some of the changes the UCR pressed for were instrumental in realising the president’s programme and undermining rival Conservative factions that relied heavily on electoral malfeasance. The legal obligation to vote could also serve the reformists in this regard, via the mechanism described above as the first logic. Importantly, it had an additional benefit: through the second logic, mandatory voting could help mitigate the tough test the Radicals would pose in freer and fairer elections. It was precisely because the incumbents were playing on two fronts – against both Conservative adversaries and the UCR – and because of the characteristics of these two competitors – the one reliant on machine politics and the other more programmatic and modern in organisation – that CV made so much sense from their perspective. As a whole, then, the electoral legislation sponsored by the government sought to satisfy some of the UCR’s conditions while preserving the reformists’ chances against it in future elections, for which they aimed to build a new conservative party on the ashes of the Regime’s political machine.Footnote 25
In my description of the second logic (i.e. the adoption of mandatory voting due to its potential for damage limitation to the incumbent) I have underscored that a left-wing electoral contender is not a necessary condition. The Argentine case is instructive in this regard. The Sáenz Peña legislation was not passed in response to class conflict: it was meant to deal with a strictly political challenge launched by a non-leftist opposition. The Radicals’ demands had nothing to do with the social question or economic policy.Footnote 26 Certainly, Radical uprisings were called revolutions, but the term denoted attempts to restore legitimate institutions without any class-based vindications.Footnote 27 More generally, labour unrest in early twentieth-century Argentina was not associated with the UCR, or any political party for that matter, but with anarchist and syndicalist groups that were unconcerned with suffrage and electoral reforms.
That said, the UCR did embody a menace to the Conservatives. Its impugnation of the Regime was deemed so destabilising that it became apparent to many of them that concessions had to be made. Miguel Angel Cárcano recalled: ‘[a]t home, I listened to my father and a group of deputies, even some of the most distinguished exponents of the old regime … “The way to disarm the revolution … is to ensure the freedom and authenticity of elections,” some said. “The Radical Party has to be incorporated into civic life …”’Footnote 28 Moreover, Minister Ezequiel Ramos Mejía had written to RSP: ‘[i]t is necessary to put an end, once and for all, to these … revolutions that disturb the normal development of the country, and for this, it is essential to pass a law that guarantees freedom of suffrage for all …’Footnote 29
The Radicals also posed a challenge within the framework of the electoral rules upon which their renunciation of violence hinged. The ideas the UCR stood for engendered considerable loyalty among its followers (it has been remarked that there was something ‘sacred’ to the Radicals’ mottos ‘the Cause or the Regime’ and ‘National Reparation’), and its organisation was novel in Argentina.Footnote 30 Given guarantees of free and fair elections, Yrigoyen had pledged to give up the UCR’s policy of abstention. Hence, Radical supporters, who had shown their identification with the party and anti-incumbent sentiment by means of armed rebellion, would surely vote. In the context of low secular turnout, this would give the Conservatives a hard time; however, by bringing a great proportion of the citizenry to the polls, mandatory voting could dilute the Radicals’ ballots.Footnote 31
The reform agreed upon by the government and the UCR had to be enacted by Congress, which only comprised representatives of different Conservative factions. All had been elected by the mechanisms of the Regime, and about 60 per cent were anti-reformist or undecided.Footnote 32 The legislative debate and the way the electoral law was passed offer evidence that both the first and second logics were behind the adoption of the legal obligation to vote.
The RSP reform consisted of two separate bills, one on electoral registration based on mandatory military enlistment (enacted as Law #8130 in 1911) and another on voting procedures, the electoral system and CV (which became Law #8871, aka the Sáenz Peña Law, in 1912).Footnote 33 The message the executive sent to the legislature with the latter was quite clear about the aims of mandatory voting: ‘[t]he collectivity that in its institutions arms the citizen with the right to vote has, in turn, the right to demand that he does not leave [it] inactive … Without the selfish inertia of the largest number, the minorities would never manage to … endanger the institutional or political order, and even the foundations of the social order’.Footnote 34
Reformists were familiar with the use of CV in Europe – not only Interior Minister Indalecio Gómez, who had served as a diplomat and was said to have first-hand knowledge, but also legislators like Senator Joaquín V. González and Deputy Manuel A. Montes de Oca. The 1907 Spanish reform warrants special attention here, given that it was referred to throughout the congressional debate and a few Argentine historians have held that it inspired Law #8871.Footnote 35 Spain’s adoption of mandatory voting had been spearheaded by the Conservatives and endorsed by other parties to, as its proponents stated, make the ‘neutral masses’ turn out and thereby counteract the advance of the new republican forces, which was allegedly made possible by the ‘almost complete abstention of [the] electorate’.Footnote 36 In addition to the legal requirement to vote, the Spanish electoral system was said to be very similar to the one in the Argentine bill, and some detractors of the RSP reform claimed that it was just a copy of Spain’s 1907 legislation.Footnote 37 Even though the Argentine incumbents drew on the Spanish experience for mandatory voting, the latter assertion is unsubstantiated: the Spanish law did not do as much as the Sáenz Peña bill or its final version in law to fight electoral corruption; its CV penalties were different; its electoral system was not the same and it included a provision that allowed for proclamation instead of elections.Footnote 38
When enumerating the measures in the bills that would clean up elections, President RSP and reformist legislators (e.g. Deputy José Fonrouge) mentioned the electoral register based on the military roster, the secret ballot and the supervisory role of the federal judiciary.Footnote 39 Minister Gómez, one of the main architects of the reform, agreed but averred that the primary means to undermine machine politics was the executive’s commitment to non-interference in elections (i.e. the end of elector governments).Footnote 40 Importantly, he highlighted that there was a role for mandatory voting to play in limiting any electoral malfeasance that might remain: ‘In our elections there have been between 150,000 and 180,000 votes … [If] 500,000 electors were obliged to vote, their number would render ineffective the influence of … venal votes, [and] would in due course also make … venality useless’.Footnote 41
The reform had some uncompromising opponents in the Chamber of Deputies who took a non-conciliatory stance toward the Radicals and whose objections to CV were intertwined with the rejection of the incomplete list: they defended a majoritarian electoral system.Footnote 42 They knew this would torpedo the RSP reform by hindering efforts to incorporate the UCR into the institutionalised political arena. Moreover, the government had made it clear that the bill was indivisible, a position that was exploited by its detractors in order to kill it, which they almost did when a vote was called in the absence of many pro-reform lawmakers.Footnote 43 Interestingly, it was the Senate, the old stronghold of the Regime, that saved the day. It introduced only a handful of amendments, including a clause for stricter enforcement of mandatory voting that had been proposed by a leading member of the Conservatives’ old guard, Senator Benito Villanueva.Footnote 44 The bill was then sent back to the Chamber of Deputies, and was passed three days later.
Law #8871 was promulgated in early 1912, adopting the secret ballot and CV, detailing the role of the federal judiciary in the organisation and supervision of elections and changing the electoral system to the incomplete list. On 7 April 1912, congressional elections took place under the new legislation, with a turnout rate of 69 per cent of registered voters.Footnote 45 Fifteen opposition deputies took seats in the chamber, faced with the forty-three Conservatives. In a speech to Congress in June, President RSP stressed how pleased he was with the exceptional turnout ‘and resolute advance of a great inorganic mass’ of voters.Footnote 46
To recapitulate, the two logics I have laid out were simultaneously at play in the adoption of mandatory voting in Argentina. Both were aligned with reformists’ goals. In contrast, the first logic (limiting the effectiveness of electoral corruption) was not in the interest of other, non-reformist, Conservatives whose support was needed for the passage of the law, while the second (hedging against a Radical mobilisational advantage) was.
Moving on to the assessment of the alternative explanations presented above, the first posits that the sponsors of CV were driven by a normative commitment to creating citizenship. For Argentina, this is frequently framed in connection to its large proportion of non-naturalised immigrants.Footnote 47 Although there are no data indicating that their native children abstained at a higher rate, RSP’s likening of mandatory voting to compulsory schooling and military service in an early speech is often offered as evidence for this view.Footnote 48 However, in the lengthy legislative debate, there was hardly any discussion of the national question or of the purported potential of CV to impress upon voters a sense of civic responsibility and awaken their interest in the affairs of the polity. In fact, reformists had stressed that these were tasks for (renewed) political parties, and Minister Gómez stated that education would gradually take care thereof whereas mandatory voting would just increase turnout.Footnote 49
The main drawback of this hypothesis is its excessive focus on the president at the expense of the other reformists. The image of RSP as a statesman who was only driven by his vision for the nation’s future and uninterested in the struggle for power may well be accurate, but the reformist Conservatives who worked alongside him did have political ambitions.Footnote 50 These men participated in the design of the electoral legislation, pushed it through Congress and/or implemented it; all the while trying to negotiate with and check both their rivals within the Conservative ranks and the Radicals.Footnote 51 They ran in elections and, in 1914, some of them founded the Partido Demócrata Progresista (Democratic Progressive Party, PDP) to organise the Conservative forces and compete against the UCR.
As for the second alternative account, it is possible to tease out whether mandatory voting was meant to have a targeted effect by getting a sense of pre-1912 non-voting patterns and looking at the sanctions that were selected.
Argentine historiography has traditionally maintained that in the early 1900s it was the rich that did not vote, reproducing the complaints of contemporary politicians.Footnote 52 However, there is reason to doubt that the latter referred to a specific socioeconomic stratum such as the landed elite, as the second competing hypothesis would predict. First, those debating the Sáenz Peña reform sometimes characterised non-voters as ‘well-to-do people’, but also as ‘a large mass of the people’.Footnote 53 Even when they used the first expression, their language was too general to denote a particular socioeconomic category. Besides, with a mean abstention rate of about 80 per cent of eligible citizens, and universal male suffrage, it is implausible that non-voters constituted a homogenous, easily identifiable subset of the potential electorate.Footnote 54 And, even if many low-class men were regular voters, it does not follow that abstention was concentrated at the top of the wealth hierarchy.
Furthermore, in consonance with the argument put forth in this article, the penalties for non-compliance with mandatory voting included in the 1912 legislation indicate that it was not intended to have a targeted effect, but rather to mobilise a heterogenous mass of citizens. Tellingly, legislators rejected a graduated fine, even though they acknowledged that the amount they selected, ten pesos (m$n10), was not at all onerous for the well-off; and they included a jail penalty for those who could not even afford that fine, which further calls into question the notion that CV was introduced to put an end to abstention by the rich.Footnote 55
Law #8871 established that failure to vote would be punished by the publication of the non-complier’s name and a fine of m$n10 – to be doubled each time for repeat offenders. Article 94 added that, if the non-voter could not pay the fine, he would spend one day in jail for every m$n5. Finally, Article 96, as clarified in the executive’s decree of 29 February 1912, stipulated that in order to enter public office or state employment, it would be compulsory to present proof of having voted. A similar requirement that applied to citizens’ dealings with the state administration was included in another decree issued around that time.Footnote 56 The exemptions (Article 7) and waivers of penalties (Article 84) were standard. As in other countries where illiterate people had voting rights when CV was enacted, they were mandated to vote but, if they failed to comply, they could have their sanctions waived (Article 84).Footnote 57
In closing, it is important to address the question of whether the 1912 reform was responsible for the Conservatives’ defeat in the 1916 presidential election. At first glance, it seems possible that the incumbents miscalculated and that, given their resources and institutional safeguards, they overestimated their chances of success under the new rules.Footnote 58 In fact, this is what some studies have argued.Footnote 59 However, others have disputed the idea that the Sáenz Peña Law brought about the collapse of the Conservatives, pointing instead to their inability to overcome factionalism and the deep fissure that emerged between the newly founded PDP and the followers of Marcelino Ugarte, then governor of the province of Buenos Aires.Footnote 60
As mentioned above, the PDP was set up by some of President RSP’s closest collaborators and other reformists seeking to renew the Conservative political offer. Under the leadership of Lisandro de la Torre, it tried to bring the provincial Conservative forces into its fold but was only partially successful as some allied with Ugarte, an old foe of the reformists. The PDP and the ugartistas vied for control over the Conservative camp and, despite their shared animosity toward Yrigoyen and/or the UCR, they would not compromise. Far from heeding the late RSP’s advice to build a cohesive party, the Conservatives not only fielded separate candidates in the 1916 presidential election but did not even cooperate in the electoral college.Footnote 61
That said, at the time, some Conservatives blamed their defeat on components of the electoral reform, albeit not on CV.Footnote 62 Furthermore, when they got the upper hand after the 1930 military coup that toppled Yrigoyen’s second government, they introduced changes to the electoral system but retained mandatory voting.Footnote 63
Chile
In the first half of the twentieth century, Chile’s political regime was more competitive than Argentina’s. It featured a multiparty system and some elections that led to coalitions of different stripes taking power in alternation. Chile’s experience in those years was characterised by phases of incremental democratisation interspersed with instances of outright autocratisation and periods of partial democratic reversal.
In the early 1900s, a class cleavage became salient in the Chilean party system. It did not immediately supersede the clerical–anticlerical historical divide; instead, the two interacted for some time. After Carlos Ibáñez’s dictatorship and a period of political instability plagued by coups and countercoups, the Liberal Arturo Alessandri was elected to the presidency on the basis of an alliance of his party with the centrist Radicals and Democrats. Then, in an extremely close race in 1938, Pedro Aguirre Cerdá became president with the backing of a coalition called the Popular Front, which encompassed his Radical Party, the Democrats, and an assortment of political groups, including the Marxist Left.Footnote 64
In the 1940s, the right-wing Conservative and Liberal parties experienced a steady decline in their vote share, which they tried to reverse with a few strategic moves, such as female enfranchisement.Footnote 65 At the time, though, the Right was still electorally strong in the countryside –especially in the populous Central Valley, where it operated through the clergy and landowners who controlled rural voting – and was thus at least assured of enough seats in Congress to block any motions from the Popular Front that might threaten its interests.Footnote 66 Furthermore, in 1948 the Right successfully pressed for a broad ban on the Communist Party, which had made some impressive electoral gains the year before.Footnote 67 More fundamentally, Chilean parties had come to an understanding known as ‘state of compromise’ by which centre-left governing coalitions pledged not to foster agrarian reform and peasant unionisation.Footnote 68
The above situation changed in the 1950s. First, the state of compromise began to crumble as Catholic reformists activated the rural labour movement and the Church shifted its allegiance from the Conservative Party to the forerunner of the Christian Democrats, developments that were followed later in the decade by leftist forces making notable inroads into right-wing rural strongholds.Footnote 69 Second, after years of intense rivalry, the parties of the Marxist Left forged an electoral alliance, the Frente de Acción Popular (Popular Action Front, FRAP). Third, in 1958, a majoritarian centre-left bloc by the name Bloque de Saneamiento Democrático (Coalition for the Healing of Democracy) formed in Congress with the backing of populist President Carlos Ibáñez, and both repealed the legislation that had outlawed the Communist Party and adopted the Australian ballot (cédula única).Footnote 70 The rightist parties forcefully opposed these initiatives. Their objections to the Australian ballot revolved around its design, which they claimed would confuse citizens, but their main concern was obvious: by making it difficult to monitor voters’ choices, it would hinder their long-standing clientelistic practices.Footnote 71
The aforementioned reforms were democratising, making elections cleaner and allowing communists to openly participate again. Importantly, the law that adopted the Australian ballot featured a clause on mandatory registration and voting, sponsored by that same centre-left bloc. Proponents of the bill contended that an increase in the electorate would ‘significantly weaken the influence of money in elections’ and, with time, disincentivise vote buying.Footnote 72 In Congress, it was lawmakers affiliated with the ibañista Partido Agrario Laborista (Agrarian Labour Party, PAL) and the centrist Radical Party who spoke about these measures on behalf of the coalition. For example, Deputy Ramón Espinoza remarked: ‘There are countries that are more advanced than ours … in which suffrage is … compulsory … We believe the same must happen in Chile. The electoral reform currently being discussed must include sanctions for all [those] who do not fulfil their civic duties’.Footnote 73 Likewise, Deputy Hermes Ahumada declared that they wanted to make both registration and voting mandatory because few eligible citizens registered and there was considerable abstention among those that did, and subsequently reiterated that everything in the bill was aimed at fighting electoral malfeasance.Footnote 74
If compulsory registration and voting were part of the reform introduced by the centre-left bloc in 1958, why is it that 1962 is regarded as the year of adoption of these institutions? One common explanation is that the relevant article lacked mechanisms for implementation, but there was also a change in legal tender in 1959 that rendered the fines for non-compliance inapplicable.Footnote 75 The legislation passed in 1962 fixed this.
In September 1958, a conjunction of events enabled the Right to get Jorge Alessandri elected to the presidency in a very narrow win over the leftist candidate, Salvador Allende.Footnote 76 In mid-1961, the Alessandri government formalised its alliance with the Radical party, the long-time pragmatic Centre, giving rise to the Frente Democrático (Democratic Front, FD) and getting a slim majority in Congress.
Later that year, the FD tabled a reform that regulated parties’ legal personality, overhauled registration boards and procedures, broadened the scope of the 1960 prohibition of electoral pacts and spelled out an array of penalties for non-compliance with the obligation to register. As far as CV is concerned, though, it only updated the currency of the pecuniary sanction in the 1958 article.
Law #14851 was published in the Diario Oficial (official state journal) on 30 April 1962. It stipulated a jail penalty of intermediate to maximum level, exchangeable for a fine of 0.50 escudos per day of prison, for non-compliance with mandatory voting. It laid out additional sanctions for failure to register, insofar as the certificate of electoral registration would henceforth be required for multiple types of transactions (e.g. to apply for a loan) and banks, notaries and public or private offices that did not demand such proof would be fined.
In sum, the first logic suggested in my argument underpinned the efforts to introduce mandatory registration and voting in 1958. The 1962 legislation, passed under a right-leaning administration, did not really innovate in these respects, but just made the adjustments necessary to finally put both into practice. As shown below, decades later, in preparation for a potential democratisation, a different right-wing government – a dictatorial one – would adopt CV on the basis of the second logic.
While mandatory registration and voting should be credited for part of the growth in electoral participation after 1962, political mobilisation peaked in those years for a variety of reasons.Footnote 77 Chief among them were the Christian Democrats’ intensive proselytising, and their policies after winning the 1964 presidential race, for which they had received the explicit support of the right-wing parties.
In 1970, despite intense scheming, the Right was unable to prevent Socialist Salvador Allende from taking office after his razor-thin victory in the presidential election.Footnote 78 Following the coup against President Allende in September 1973, the military unleashed brutal repression and embarked on a series of far-reaching socio-economic and political transformations. The ruling junta established CV in 1988, when it set out its penalties in Law #18700, and had included a clause about suffrage being mandatory in the Constitution it had issued in 1980.
The 1980 Charter is best understood as two constitutions in one: ‘the transitory’ and ‘the permanent’.Footnote 79 The former would rule for eight years, with General Augusto Pinochet as president, after which a plebiscite would be held on a presidential nominee selected by the military junta. If the latter was rejected, elections would be called. The permanent constitution, which contained the institutional framework for a potential democratisation, would apply thereafter.
The junta crafted the new constitution and the related electoral laws in such a way as to limit its loss of influence in case of a defeat in the 1988 plebiscite and to benefit its political allies in the aftermath. The process of institutional design within the military government has been well documented. Perusal of these primary sources leaves little doubt that the Chilean dictatorship considered mandatory voting a means to offset the mobilisation of radicals or extremists when elections were reinstated.
The Comisión Ortúzar, an advisory body set up by the junta, devoted one of its early meetings to the issue of the legal obligation to vote.Footnote 80 Commissioner Jorge Ovalle emphasised that mandatory voting had to be included in the Constitution so that future legislators would have no leeway on this score. Tellingly, Ovalle averred that if voting were not compulsory, ‘the usual activists’ would impose their will, given the inertia and disinterest of those who were contented with society; ‘one way to make [the latter] express that satisfaction is precisely not just to induce them to vote, but to compel them to do so’.Footnote 81 In a similar vein, Commissioner Alejandro Silva warned that voluntary voting would be threatening to social order because it would ultimately strengthen the influence of extremists, who had such intense preferences and beliefs that they would never refrain from voting.Footnote 82
The commissioners were aware that Chile had had mandatory voting before 1973, but reasoned that there was no link between that and Allende’s presidential victory. In fact, their view was the opposite, as Enrique Evans and Ovalle explained: ‘… experience shows that … the inefficiency of the old legislation in enforcing compulsory voting led many Chileans to neglect to uphold their duty to defend democracy … trusting that others would do so in adequate numbers … These Chileans preferred recreation and rest to attending [the polls] to manifest their rejection of a group that … would come to power in the country, with the disastrous consequences known to all’.Footnote 83
The members of the Ortúzar Commission agreed on mandatory voting and drafted the corresponding constitutional article.Footnote 84 Their work on other sections of the Charter continued for a few more years, and it was not until December 1978 that the junta’s Council of State voted on their CV proposal. As revealed by the transcripts of their meetings, the councillors approved this clause unanimously and without much deliberation.Footnote 85 It was then included in the 1980 Constitution. In 1987, a year ahead of the scheduled plebiscite, General Pinochet presented the bill for the law on elections, Bulletin #867-06. Article 119 delineated the sanction for non-compliers with the obligation to vote: a fine of between one half and three unidades tributarias mensuales (monthly tax units).Footnote 86 This bill was promulgated as Law #18700 in April 1988, effectively adopting mandatory voting. That same year, the dictatorship was defeated in the plebiscite and Chile re-democratised soon afterwards.
The evidence thus points to my posited second logic being behind the institution of CV in this instance. The first logic, on the other hand, was irrelevant: the dictatorship’s decision-makers and advisors did not appear concerned about vote buying or other forms of electoral corruption that could be limited by the high(er) turnout that mandatory voting would bring about.Footnote 87 A different type of electoral manipulation had attracted their attention early on, though. Shortly after the coup, they had accused the Allende administration of tampering with the voter rolls, an issue they claimed to have solved by destroying the electoral register.
The Chilean dictatorship ended up establishing voluntary registration, even though the Ortúzar Commission had selected an automatic procedure.Footnote 88 A special group, the subcomisión (sub-commission), had been tasked with working out the details but, after a while, it began expressing doubts about the feasibility of automatic registration, given technical issues and data constraints. Still, most commissioners took for granted that such problems would be soon overcome, and their constitutional blueprint mentioned automatic registration in Article 19.Footnote 89 However, a report signed by the minister of the economy in January 1986 explained that automatic registration had to be ruled out due to insurmountable practical obstacles, and the law promulgated on 11 September 1986 established optional registration.Footnote 90 Since the registration rate for the 1988 plebiscite was so high – 92 per cent of eligible citizens – the issue of self-selection into CV that might arise from voluntary registration was negligible for over a decade. Eventually, though, it resulted in the aging of the effective electorate, which contributed to the 2009/2012 reform that rescinded mandatory voting.Footnote 91
Turning to the evaluation of the first alternative hypothesis about the introduction of CV, the fact that it was embraced by the Pinochet dictatorship calls into question the idea that this institution was adopted out of a normative commitment to bolster democratic engagement. Indeed, the Ortúzar Commission and the Council of State, which backed the legal obligation to vote, voiced qualms about democracy.Footnote 92 Even the 1958 reform, which was certainly pro-democratic, did not justify mandatory voting on the grounds of building citizenship. As discussed above, its rationale was different: expanding the effective electorate in order to curb vote buying and related electoral malfeasance.
The second competing explanation postulates that the establishment of CV may have aimed to diminish the abstention of a well-defined subset of the potential electorate that constituted its proponents’ natural base. However, there are no references to significant social biases in turnout in any of the primary sources considered in the previous section. Moreover, the generic penalties for non-compliance that were selected and the foregoing summaries of the deliberations in the relevant fora, as well as the depiction of the political context, indicate that mandatory voting was not meant to have a targeted mobilising effect but a general one. This is consistent with both my first logic, articulated by Chilean lawmakers in 1958, and my second logic, i.e. the view of CV as a tool for damage limitation to the incumbents, which became manifest three decades later.
Colombia
Throughout the 1800s, Colombia’s traditional parties, Conservador (Conservative) and Liberal, generated strong allegiances among the citizenry on the basis of community ties, family tradition and patronage. The latter fed into, and were reinforced by, frequent electioneering and inter-party violent strife. In the 1850s Conservatives and Liberals had already developed remarkable mobilisation capabilities and continued dominating the political arena thereafter.Footnote 93
On 9 April 1948, Liberal politician Jorge Eliécer Gaitán was assassinated in Bogotá and massive riots broke out. Violence soon escalated across Colombia, as Liberal and Conservative militias fought for territorial control. The upshot was a phase of all-out interparty warfare known as La Violencia (The Violence), and a military coup led by General Gustavo Rojas Pinilla with the backing of moderate Conservatives and Liberals. Before long, however, Rojas Pinilla’s attempt to perpetuate himself in office alarmed Colombian politicians, who overthrew him in May 1957. The leaders of the two traditional parties had begun pursuing an accommodation in 1956, when they had agreed on bipartisan action to bring about democratic rule. A year later, they forged a compact stipulating that they would alternate in the presidency and that there would be parity in other public positions as well as a set of mutual guarantees that would constitute the cornerstones of the Frente Nacional (National Front).Footnote 94
The Front’s years featured increasing abstention, which might have raised concerns about its legitimacy, but Conservatives and Liberals did not introduce CV. Both parties were electorally safe, alternating in the presidency and dividing public offices among themselves. In the 1970 presidential election, the official candidate narrowly defeated Rojas Pinilla, who ran under a Conservative factional label and led the populist organisation Alianza Nacional Popular (National Popular Alliance, ANAPO). However, at the time, the traditional parties were still protected by their power-sharing framework.Footnote 95 It was only in the last quarter of the twentieth century that the latter was gradually dismantled.
Electoral absenteeism remained high in the post-National Front era; the mean participation rate in presidential contests was 47.5 per cent of registered voters.Footnote 96 Notably, the Conservatives and Liberals continued to prevail in elections, mostly relying on extensive clientelistic networks, as citizen partisanship had waned. From the 1970s to the early 1990s the two traditional parties won, on average, over 85 per cent of the presidential vote and more than 90 per cent of the seats in the national legislature.Footnote 97 No significant electoral challenger rose until the Alianza Democrática-M19 (M-19 Democratic Alliance, AD-M19) was founded in 1990.
The Virgilio Barco administration (1986–90) signed a peace accord with the Movimiento del 19 de Abril (19th of April Movement, M19) guerrilla organisation that required amendments to the 1886 Charter. The M19 turned in its weapons in March 1990, and in that year’s presidential contest, its party, the AD-M19, garnered more votes than the official candidate of the Conservatives. The AD-M19’s performance would be even more impressive in the election of the Constituent Assembly in December. Poll attendance was extremely weak (around 30 per cent of registered voters) and the AD-M19’s was the most voted list.Footnote 98 The combination of very low turnout with the exceptional showing of the more programmatic AD-M19 catalysed the discussion of mandatory voting in the Constituent Assembly under the auspices of Conservatives and Liberals.
The legal obligation to vote figured in the Conservative Party’s blueprint for constitutional reform, as well as in the proposals of Liberals such as Jaime Arias López, Horacio Serpa Uribe, Antonio Galán Sarmiento, Iván Marulanda Gómez, Hernando Herrera Vergara and Juan Fernández Renowitzky, among others.Footnote 99 By contrast, this institution was not only absent from the AD-M19’s reform project but also faced objection from the party’s representatives in the Constituent Assembly.Footnote 100 Instead, AD-M19 delegates insisted on measures to enhance electoral integrity and on the inclusion of the recently established Australian ballot (tarjetón) in the constitutional text.Footnote 101
During the meetings of the relevant commissions, a compromise on mandatory voting was eventually reached: the constitutional text would state that voting was both a right and a duty while leaving the adoption of penalties and exemptions to Congress.Footnote 102 The latter did not happen, though: following the promulgation of the Charter, the traditional parties lost interest in CV, as they maintained their electoral dominance and the AD-M19 dwindled.Footnote 103 It finally dissolved in the mid-1990s after getting less than 4 per cent of the presidential vote and only one seat in the legislature.Footnote 104
Dissidents from the Conservative and Liberal parties had long been a fixture in Colombian politics, but in the late 1990s and early 2000s, the party system experienced extreme fragmentation.Footnote 105 There were then some attempts at political reform, a few of which featured CV. Motions tabled to enact this institution were peculiar, though, as they proposed it would apply to just two consecutive elections, and mixed penalties for electoral absenteeism (mandatory voting) with rewards for voters.Footnote 106 CV resurfaced during the Uribe administration (2002–10). President Alvaro Uribe’s planned referenda faced such intense opposition that there were calls for abstention to invalidate them. In order to counteract those campaigns, the executive and uribista legislators submitted mandatory voting bills. These were swiftly rejected by the Constitutional Court, which upheld the juridical efficacy of non-participation in referenda in a series of rulings.Footnote 107 Nonetheless, toward the end of President Uribe’s tenure, an even more controversial idea was floated: a one-time implementation of CV coupled with the removal from the electoral register of those who had not voted in the most recent presidential contest. This proposition allegedly sought to update the voter rolls, but such a retroactive imposition of penalties was challenged by the head of the National Civil Register Bureau, and the Constitutional Court ruled against it, effectively settling the issue.Footnote 108
In short, CV was seriously contemplated in Colombia when its traditional parties faced a credible electoral threat from a programmatic force in the context of low turnout and the absence of both the protections of the National Front and the mobilising attachments they had enjoyed in an earlier era. When that electoral scare proved short-lived, both parties’ interest in this institution dissipated. A few initiatives emerged afterwards, but these contained unusual features, looking nothing like mandatory voting as it is known in Latin America (and beyond).
As the foregoing analysis unveils, the first logic put forth in my explanation of CV adoption was not offered as a rationale for the proposals of a legal obligation to vote in Colombia. This is perhaps not surprising, given the heavy reliance of both traditional parties on electoral clientelism.Footnote 109 Why was mandatory voting then not favoured by representatives of the new and more ideological political forces that demanded guarantees against electoral irregularities in the Constituent Assembly? One possible reason may be the trade-off between CV’s potential to curb such malfeasance and its likely detrimental effect on the electoral influence of these forces through the second logic.
What does the Colombian experience tell us regarding the alternative hypotheses about the enactment of mandatory voting? First, if the latter were intended to create citizenship, it should have been at least entertained in the nineteenth or early twentieth century, given Colombia’s tortuous nation-building process; but it was not.Footnote 110 And, despite the decline in turnout under the National Front, the incumbents did not make use of the legal obligation to vote to foster civic engagement. Second, this case does not appear to support the notion of CV introduction as meant to mobilise a particular subset of the citizenry that forms part of the natural base of those championing this institution. There is nothing in the proposals made to, or the debates held in, the Constituent Assembly (linked in footnotes above) to substantiate this hypothesis.
Conclusion
Mandatory voting has been common in Latin America, and its adoption along with democratisation poses a puzzle from the standpoint of influential theories in political economy. In this article, I have put forth an explanation that rests on two implications of what this institution can reasonably be predicted to do, i.e. increase turnout. The first logic suggests that the legal obligation to vote was established out of a belief that a large(r) actual electorate could undermine the cost-effectiveness and viability of vote buying and related electoral malfeasance. In turn, the second posits that incumbents facing challengers who had a mobilisational advantage introduced CV with the expectation that, by making the bulk of the citizenry vote, it would bring to the polls an indifferent, less politicised and/or moderate mass that would drown out the ballots of the strongly committed opposition supporters. In the first logic, the adoption of mandatory voting, even when driven by strategic motivations, reinforces democratisation. In the second, it aims to limit the damage that the latter can inflict on those in power.
Whether one or the other (or both jointly) of these logics account for the establishment of the legal obligation to vote in a particular instance depends on the characteristics of the incumbents and their main competitors, as well as on the degree of concentration of decision-making. As shown in the empirical section, both logics operated simultaneously in Argentina in 1912, whereas in Chile, the first drove the 1958 reform (implemented in 1962) and the second underpinned the enactment of CV three decades later.
Here, I have tested my explanation against competing hypotheses through a comparative historical study of three South American polities that cover the range of variation on the dependent variable. In addition, plausibility probes I have presented elsewhere indicate that this argument also holds for the two remaining cases of CV adoption along with democratisation in South America, as well as for some beyond the region.Footnote 111
Further research is needed on the origins of mandatory voting, especially on its introduction and use under autocracy and on the subsequent decision to keep or abolish it when the polity transitioned to democracy. As mentioned in the first part of this article, Latin America offers an excellent setting to explore these questions.
Over the past two decades, the legal obligation to vote has been floated in some democracies as a solution to declining turnout (and Chile has recently reinstated it).Footnote 112 These advocates of CV may find it disheartening to learn that, as noted above, in Latin America, the instances of adoption and practice under autocracy have outnumbered those of introduction in democratising contexts. However, as with other institutions, the establishment and implementation of mandatory voting under dictatorship does not mean it is intrinsically authoritarian. Moreover, even though the second logic in my explanation of CV introduction along with democratisation portrays it as one of the incumbents’ safeguards against the latter, it does not imply that it impaired democracy. Indeed, given this view of mandatory voting among those in power, its adoption may have helped prevent the reversal of democratising reforms and may have contributed to regime transition.
That being said, there are reasons for caution regarding contemporary CV proposals in democracies. First, it is critical to understand what is driving the increase in electoral absenteeism in each case, as mandatory voting may not be the best (or even an adequate) intervention for all causes. Second, the crises of representation that plague several democratic regimes these days and the global uptick in citizen dissatisfaction with democracy should give pause to proponents of CV as a quick fix.Footnote 113
Acknowledgements
This article has benefited from discussions with colleagues at both Yale University and the University of Oxford, as well as from the comments of three anonymous reviewers. For assistance in locating archival materials, I am grateful to the staff at Dirección de Referencia Legislativa, Biblioteca del Congreso de la Nación Argentina; Biblioteca del Congreso Nacional de Chile; Unidad de Atención Ciudadana del Congreso de la República de Colombia and Biblioteca Luis Angel Arango in Bogotá. I am particularly indebted to my interviewees at the Servicio Electoral de Chile for their time and hospitality.
Funding
Financial support for this project was granted by the Georg Leitner Program in International and Comparative Political Economy at Yale University, the Yale Council for Latin American and Iberian Studies and Merton College at the University of Oxford.
