When and where did democracy first emerge in Latin America? This chapter uses a plethora of data on elections and suffrage rights in South America to assess the degree of democracy in the region before 1930. As it shows, the first lengthy and meaningful experiences of democracy in the region did not occur until the early twentieth century, although there were brief episodes of democratization in the nineteenth century. Chapters 3–8 show how democracy arose during this period and why in some South American countries and not others.

Throughout the nineteenth century, Latin American countries held regular elections, but there is a significant debate about how democratic these elections were. Scholars traditionally portrayed elections in nineteenth-century Latin America as farces riddled with violence, fraud, and manipulation. In 1919, for example, Seymour and Frary (Reference Seymour and Frary1919, 267) reported that “it is probably not unfair to say that elections were usually a pure sham” in the region in the greater part of the nineteenth century.Footnote ¹ More recently, Nohlen (Reference Nohlen2005a, 4) argued that “[e]lections in Latin America were mostly characterized by fraud and were used more to provide political legitimacy to the incumbents rather than to control their right to govern a country.” The main democracy indexes exhibit similar skepticism, portraying Latin American elections during this period as unfree and unfair and the region’s regimes as consistently authoritarian.

In recent decades, however, scholars have challenged this view, showing that many nineteenth-century elections involved considerable electoral competition and participation (Annino Reference Annino1995; Aguilar Rivera, Posada-Carbó, and Zimmermann Reference Aguilar Rivera, Posada-Carbó and Zimmermann2022; Drake Reference Drake2009; Malamud Reference Malamud2000b; Posada-Carbó Reference Posada-Carbó and Posada-Carbó1996a; Reference Posada-Carbó2000a; Posada-Carbó and Valenzuela Reference Posada-Carbó2012; Sabato Reference Sabato2001b; Reference Sabato2018; Valenzuela Reference Valenzuela and Posada-Carbó1996; Reference Valenzuela2012). The revisionists acknowledge that there was significant electoral corruption in the nineteenth century and that many countries restricted the franchise, but they argue that democratic practices were much more widespread than the traditional literature recognized.

Until now, we have not had a comprehensive source of data on elections in the region during this period, which has made it difficult to arbitrate between these two conflicting views. Without such data, we cannot assess with any precision the degree of competition, participation, and democracy in elections in nineteenth-century Latin America.

This chapter takes a large step toward filling this gap by presenting and analyzing an original data set on presidential elections in South America from independence to 1929. LAHED is based on both archival research and wide-ranging scrutiny of the growing secondary literature on elections in the region. It contains information on all presidential elections during this period – 263 elections in South America alone – which is far more than existing data sets include.

The analysis of LAHED data in this chapter focuses on electoral competition and participation because they are widely viewed as the cornerstone of democracy, at least since Schumpeter (Reference Schumpeter2008 [1942]) and Dahl (Reference Dahl1971). I examine presidential elections in part because of the greater availability of data on these elections, but also because the president was clearly the most powerful actor in Latin American politics during the nineteenth century and had the greatest impact on democracy.

My analysis of LAHED data finds that the revisionist view is accurate in that democratic practices did exist in nineteenth-century Latin America. Some South American countries established broad suffrage rights during the nineteenth century and a few even enacted universal male suffrage. Voters of all social classes participated in elections, and in several countries voter turnout was relatively high. Most presidential elections were contested, some were competitive, and in a few cases the opposition won. A handful of nineteenth-century elections appear even to have been relatively free and fair.

Nevertheless, we must be careful not to exaggerate the degree of democratic practices in nineteenth-century South America. By focusing on the limited democratic trees that existed in the nineteenth century, the revisionist approach risks obscuring the authoritarian forest.Footnote ² Government intervention, fraud, and intimidation marred the vast majority of elections during the nineteenth century. More than 70 percent of presidential elections were uncompetitive and at least one-third were not even contested. Most South American countries also significantly restricted suffrage rights, maintaining literacy and/or economic restrictions throughout most or all of the nineteenth century as well as denying the vote to women. Even where suffrage rights were broad, voter turnout was often low. Indeed, valid votes typically represented less than three percent of the total population during the nineteenth century. To be sure, corrupt, uncompetitive, and exclusionary elections were also quite common in Europe, North America, and the Antipodes at that time, but, as we shall see, the level of competition and voter turnout was higher on average in these regions than in South America during the nineteenth century.

The few democratic episodes in South America during the nineteenth century proved ephemeral. In some cases, the governments elected in free and fair elections were quickly overthrown by their opponents. In other cases, the democratically elected presidents themselves subverted democracy, intervening in elections and repressing the opposition to ensure that their favored candidates triumphed. Moreover, the brief democratic episodes failed to leave a lasting mark.

It was not until the early twentieth century that South American countries experienced lengthy and meaningful episodes of democracy. At the outset of the twentieth century, a great divergence occurred in the region. A few nations, such as Argentina, Chile, Colombia, and Uruguay, established democratic regimes that would last for a dozen years or more. During this period, these nations regularly held reasonably free and fair elections with relatively broad voter participation. Three of these four countries remained democratic for most of the twentieth century (Argentina being the exception), although they all experienced coups at one point or another. By contrast, most of the other South American countries – Bolivia, Brazil, Ecuador, Paraguay, Peru, and Venezuela – became more authoritarian during the early twentieth century. Indeed, in Bolivia, Ecuador, Peru, and Venezuela, elections were less competitive and less democratic in the first three decades of the twentieth century than they had been on average during the late nineteenth century. Moreover, all six of the democratic laggards remained authoritarian for most of the twentieth century.

These findings indicate that democracy, as I defined it in the Introduction, did not emerge in South America until the early twentieth century. The nineteenth-century episodes correspond to what I refer to as ephemeral democratization: brief democratic openings that did not have enduring impact. By contrast, the democratic episodes of the early twentieth century were lengthy and had a lasting impact.

This chapter is organized as follows. The first section of the chapter describes LAHED and compares it to other databases on democracy and elections that cover this period. The second section uses LAHED to examine the evolution of suffrage rights and voter turnout in the region in the nineteenth century. The third section discusses the competitiveness and fairness of South American elections in the nineteenth century. It also describes the few episodes of democracy that occurred in the region during this period and demonstrates that they were short lived. The fourth section shows that sustained episodes of democracy arose in four South American nations during the early twentieth century, whereas the other countries in the region remained authoritarian. The concluding section summarizes the main findings of this chapter.

A New Database on Historical Elections

LAHED provides comprehensive coverage of presidential elections in Latin America from independence to 1929. It covers both direct and indirect elections for provisional as well as constitutional presidents, including those elections that were conducted in the legislature or constituent assembly. LAHED provides data on the electoral results, the candidates, suffrage rights, and the freeness and fairness of the elections, among other characteristics. The data were culled from more than 300 sources, including general histories, electoral compendia, and studies of individual elections, as well as archival sources such as newspapers, ambassadorial dispatches, presidential messages, and congressional minutes and reports.

LAHED counts as a presidential election only those elections that involve, directly or indirectly, the citizenry. This includes instances where the legislature, a constituent assembly, or an electoral college elects the president provided that those bodies are elected by the citizenry. LAHED does not count as elections instances where a body whose members were selected by the incumbent president elects the new president. Nor does LAHED cover instances where the resignation, death, forcible removal, or impeachment of the existing president leads to the ascension of the next-in-line for the presidency (e.g., the vice-president or the first designate), even if such ascension requires a vote by the legislature. Finally, LAHED does not cover instances where a plebiscite or legislative vote is held to decide whether a president should remain in office or extend his/her term.

LAHED has identified and coded 263 presidential elections in South America during this period: 182 in the nineteenth century and eighty-one between 1900 and 1929. It contains the actual vote totals for 151 elections in the nineteenth century and sixty-five elections in the first three decades of the twentieth century, including both direct and indirect elections.Footnote ³ In addition, it provides estimated results for thirty-four other presidential elections in which the vote totals are unknown but where only one candidate ran and/or the result was reported to be unanimous. Thus, it contains actual data on 82 percent of the presidential elections that took place in South America between independence and 1929 and actual or estimated data on 95 percent of all presidential elections in the region during this period.

The presidential elections covered include fifteen elections in Argentina, twenty-nine in Bolivia, eleven in Brazil, twenty-five in Chile, thirty-five in Colombia, thirty-five in Ecuador, twenty-four in Paraguay, thirty in Peru, twenty-five in Uruguay, and thirty-four in Venezuela. Some countries held more presidential elections than others, in part because they had shorter constitutional terms, or because the terms of some presidents were interrupted and new elections were held ahead of schedule. Brazil has the fewest elections in the data set because it did not hold presidential elections for most of the nineteenth century, since it was an empire. Argentina, meanwhile, was fragmented after independence and, with one exception, did not hold presidential elections during the first half of the nineteenth century.

LAHED goes significantly beyond any other source in its coverage of historical elections in the region and the data it provides. The most thorough and reliable source for twentieth-century elections in Latin America is Nohlen (Reference Nohlen2005a), but it does not include the nineteenth century and its coverage of early twentieth-century elections is spotty.Footnote ⁴ The main democracy databases also lack comprehensive data on elections during this period. Neither Polity, nor Boix, Miller, and Rosato, provide data on elections per se. The Political Institutions and Political Events (PIPE) data set developed by Przeworski (Reference Przeworski2013) provides data on suffrage rights and other nineteenth-century political institutions, but it does not report the results of the elections or assess their quality.Footnote ⁵ The Varieties of Democracy (V-Dem) project codes some elections during this period, but it omits presidents elected by the legislature or a constituent assembly, which accounted for 39 percent of presidential elections in South America during the nineteenth century and 28 percent in the first three decades of the twentieth century. V-Dem provides the results for only fifty-two presidential elections in South America during the nineteenth century and twenty-seven elections between 1900 and 1929, which represents only 36 percent of the presidential election results contained in LAHED (Coppedge et al. Reference Coppedge2022b).Footnote ⁶ V-Dem has even more limited data on the results of legislative elections during this period.

The most comprehensive source on nineteenth-century presidential elections in Latin America is Vanhanen (Reference Vanhanen2000), but he only provides data on the winner’s share of the total vote (competition) and the percentage of the total population that cast votes (participation).Footnote ⁷ Moreover, a significant portion of the data Vanhanen includes is misleading owing to some problematic coding rules.Footnote ⁸

As Table 2.1 indicates, LAHED’s variables measuring the freeness and fairness of elections and the vote share of the winner are only modestly correlated with variables from other databases that measure the level of democracy or the quality of elections in South America before 1930. The strongest correlations are between LAHED’s measure of free and fair elections and: (1) V-Dem’s measure of free and fair elections (0.58); (2) Boix, Miller, and Rosato’s dichotomous measure of democracy (0.57); and (3) Vanhanen’s Index of Democratization (0.51).Footnote ⁹ All of these variables, however, were much more weakly correlated in the nineteenth century than in the early twentieth century.Footnote ¹⁰ LAHED’s measure of votes cast as a share of the total population is highly correlated with Vanhanen’s equivalent measure of participation (0.82), but this high level of correlation should not be construed as validating Vanhanen’s data on participation. The vast majority of Vanhanen’s inaccurate data on participation come from indirect elections, but LAHED has very little data on voter turnout in these elections.

In sum, LAHED draws on extensive historical research to provide systematic data on South American presidential elections from independence to 1929. It contains considerably more data and many more elections than existing data sets, and the data it provides are only modestly correlated with that in existing databases. It therefore facilitates a more comprehensive and accurate assessment of the degree of democracy in the region during the nineteenth and early twentieth century.

Suffrage Rights and Electoral Participation in the Nineteenth Century

As noted, the traditional view was that elections in the nineteenth century were elite affairs in which only a small sector of the population participated. Indeed, many scholars have referred to Latin American regimes of this period as elite or oligarchic republics because of their limited mass participation. In recent decades, however, historians have challenged this view, emphasizing that elections in Latin America during this period often involved considerably more voter participation than was traditionally believed. They note that the 1812 Spanish constitution, the so-called Cádiz Constitution, provided a model for the extension of broad suffrage rights in Latin America after independence. The Cádiz Constitution extended suffrage to almost all free adult males in the Americas, including indigenous people. It excluded people of African origin, but they could apply for citizenship based on their talents, good behavior, and service to the country (Aguilar Rivera, Posada-Carbó, and Zimmermann Reference Aguilar Rivera, Posada-Carbó and Zimmermann2022, 12–13). The Cádiz Constitution imposed no property or income requirements, and it suspended literacy requirements until 1830, but it barred domestic servants, debtors, the unemployed, and criminals from voting.

My analysis of LAHED data suggests that the revisionists are correct in that nineteenth-century South American suffrage rights were sometimes broad and voter turnout was occasionally high. Half of the South American countries adopted nearly universal male suffrage at some point during the nineteenth century. Moreover, some of the countries that maintained economic restrictions, such as Brazil and Chile, set the requirements to vote at relatively low levels or allowed inflation to erode their real value over time, which reduced the impact of these restrictions. Social and economic progress also gradually increased the share of people who satisfied the economic and literacy requirements. In addition, the restrictions that did exist were not always enforced. For example, in some cases, citizens were asked only to sign their names to demonstrate that they could read or write. At different points, the electoral registries of various countries contained many people from the lower classes, including illiterates, artisans, farmhands, day laborers, and free men of color (Madrid Reference Madrid2019a, 8; República de Chile 1863; 1871; 1879; Gil Fortoul Reference Gil Fortoul1942, 272–273; Gilmore Reference Gilmore1964, 17; Graham Reference Graham1990; Klein Reference Klein1995; Navas Blanco Reference Navas Blanco1993; Sabato Reference Sabato2001a; Sabato and Palti Reference Sabato and Palti1990).Footnote ¹¹ Artisans not only voted frequently, they also played a key role in campaigns in Chile, Colombia, Peru, and elsewhere (Sanders Reference Sanders2004; Sowell Reference Sowell1992; Wood Reference Wood2011; García Bryce Reference García Bryce2004; Gazmuri Reference Gazmuri and Thomson2002; Posada-Carbó Reference Posada-Carbó and Thomson2003; Sobrevilla Perea Reference Sobrevilla Perea and Thomson2002).

Nevertheless, we should not exaggerate the degree of suffrage rights and voter participation in the nineteenth century. Suffrage restrictions in the region disenfranchised a large majority of the adult population for most of the century. Women were prohibited from voting throughout the nineteenth century in South America as elsewhere – Ecuador in 1929 was the first Latin American country to grant women the right to vote in national elections.Footnote ¹² Most South American countries also banned illiterates from voting for the bulk of the nineteenth century, which had a major impact since the vast majority of the population in the region was illiterate. In addition, South American countries imposed a host of economic restrictions on the franchise, from income and property requirements to prohibitions on voting by workers in dependency relationships, such as domestic servants, peons, and day laborers. The economic requirements typically disenfranchised fewer people than the literacy provisions, but they both significantly reduced the size of the electorate.

The economic and literacy restrictions were quite pervasive. One or the other (or both) were in place in 532 of the 768 country-years that South American nations were independent during the nineteenth century, or 69 percent of the time.Footnote ¹³ These data understate the percentage of years in which countries imposed economic or literacy restrictions by excluding the 1816–1853 period in Argentina and the 1853–1863 period in Colombia in which suffrage rules were set at the state or provincial level. (Some states or provinces in Argentina and Colombia imposed economic and literacy restrictions during these periods, but others did not.) Although South American countries did not typically ban members of marginalized ethnic or racial groups from voting, the economic and literacy restrictions disenfranchised most indigenous people and Afro-Latinos during the nineteenth century.

Other restrictions had a smaller impact. Slaves, for example, were prohibited from voting, but the enslaved population in South America declined rapidly over the course of the century.Footnote ¹⁴ Rank-and-file soldiers and the clergy were also frequently banned from voting, although neither of these groups typically represented a large sector of the population. In addition, Brazil and Ecuador restricted the suffrage to Catholics during parts of the nineteenth century, but it is unclear how strictly these religious restrictions were enforced.

As Table 2.2 indicates, suffrage rights varied considerably across countries and over time. On the whole, Argentina, Paraguay, and Venezuela granted the broadest suffrage rights, followed by Peru, Colombia, and Brazil. By contrast, Bolivia, Chile, Ecuador, and Uruguay had the most restrictions.

The majority of Latin American countries followed what Sabato (Reference Sabato2001b, 1297) referred to as a zig-zag path of suffrage rights, alternating between expansion and contraction over the course of the nineteenth century. During the independence and immediate post-independence periods, most South American countries extended relatively broad suffrage rights, imposing economic restrictions but eschewing or suspending literacy restrictions so as not to disenfranchise the many illiterate men who had fought in the wars of independence. Almost all countries subsequently tightened their voting restrictions only to later relax or eliminate them, but the timing of these suffrage contractions and expansions differed significantly from country to country. Bolivia, Chile, Ecuador, and Uruguay imposed literacy restrictions in the 1830s or 1840s and did not lift them until the twentieth century, whereas Brazil, Colombia, and Peru did not significantly tighten their voting restrictions until late in the nineteenth century.

A few South American countries, namely Argentina, Paraguay, and Venezuela, took a different route, institutionalizing virtually universal male suffrage during the mid-nineteenth century.Footnote ¹⁵ Suffrage rights in these countries went well beyond what most European countries had at the time. Colombia and Peru also instituted virtually universal male suffrage during the nineteenth century, but only for brief periods: Colombia had it between 1853 and 1863, although some provinces continued to maintain it until 1886; Peru adopted it between 1828 and 1834 and again in 1855–1856. Overall, however, universal male suffrage was the exception rather than the norm in the nineteenth century. In total, South American countries maintained universal male suffrage at the national level in 138 country-years during the nineteenth century, which represents 18 percent of the region’s 768 independent country-years during this century.Footnote ¹⁶

As Przeworski (Reference Przeworski2009b, 15) has shown, universal male suffrage arrived at about the same time in Latin America and western Europe. According to his estimates, the proportion of the population that was eligible to vote in each region was relatively similar in the first half of the nineteenth century – it was well under 10 percent of the population in both regions. Nevertheless, by 1900, approximately 20 percent of the population had suffrage rights in western Europe, whereas only about 12 percent did in Latin America (Przeworski Reference Przeworski2009b, 17).

The economic and literacy restrictions in South America dramatically reduced the size of the electorate since the vast majority of the population was poor and illiterate. In 1870, only 23.5 percent of the population aged fifteen years or older in Argentina could read and write and only 16 percent of this population could do so in Brazil (Thorp Reference Thorp1998, 354). By 1900, literacy rates in South America had risen somewhat, but they still only averaged 35.8 percent for the population over fifteen years of age (Thorp Reference Thorp1998, 354).

There are relatively little data available on the number of people registered to vote in South America during the nineteenth century, but the data that exist suggest that they represented a small percentage of the total population. During the nineteenth century, registered voters constituted only 4.8 percent of the population on average. This proportion grew somewhat over the course of the century but still remained low. In the first half of the nineteenth century, the number of registered voters averaged 4 percent of the total population, whereas in the second half it averaged 5.1 percent – even in the 1890s, registered voters only represented 6.2 percent of the total population on average.

The number of registered voters in the nineteenth century represented only a fraction of males of voting age, probably less than one-third on average. Precise data on the voting age population are lacking, but it seems reasonable to assume that during the nineteenth century, males of voting age typically represented between 15 and 25 percent of each country’s population, depending on the country’s demographics and its voting age requirement. For example, men who were at least twenty-five years of age represented only 18.8 percent of the total population in Chile in 1875 (Oficina Central de Estadística 1885, xlii). Latin America’s population was quite young in the nineteenth century because fertility rates were high and life expectancy was low. In 1900, for example, life expectancy at birth was only twenty-five years of age in Paraguay, twenty-six in Bolivia, twenty-nine in Brazil and Chile, and thirty-nine in Argentina (Thorp Reference Thorp1998, 356). As a result, the majority of the population in the region was below the minimum voting age.

Suffrage restrictions and the relative youth of the region’s population, along with citizen disinterest, limited the number of people who cast ballots. In the ninety-four nineteenth-century elections for which LAHED has data, the number of actual voters constituted 3.4 percent of the total population on average. These data include both presidential and legislative elections – the two types of election had similar levels of turnout.Footnote ¹⁷ Voter turnout fluctuated considerably over the course of the nineteenth century, but it followed a generally upward trend. During the first half of the nineteenth century, it was relatively low, averaging only 1.7 percent of the total population. In the 1850s, it soared to an average of 5.4 percent, but then declined steadily for several decades, falling to an average of 2.6 percent in the 1880s, before rising again in the 1890s, when it averaged 4.6 percent.

According to LAHED data, a few South American countries, namely Venezuela, Peru, and Brazil, had high levels of voter turnout, although data are available for only a few elections in each of these countries.Footnote ¹⁸ Voters averaged 10.5 percent of the total population in six elections in Venezuela and 10.6 percent of the population in four elections in Peru during the nineteenth century. In Brazil, the number of votes cast represented an average of only 3.4 percent of the population in six elections in the nineteenth century, but voter turnout was much higher prior to the enactment of the suffrage restrictions of 1881: Voters represented 5.1 percent of the population in the 1835 elections for regent, and 8.6 percent in the 1872 legislative, elections.

Nevertheless, low voter turnout was the norm in most South American countries. Colombia had very high voter turnout in the 1856 elections, when it reached 8.6 percent, but this election, which took place using universal male suffrage, was atypical. On average, voter turnout constituted 3.4 percent of the population in thirteen nineteenth-century elections in Colombia. Voter turnout was even lower in Chile and Ecuador, averaging 2.3 percent of the population in twenty-four nineteenth-century elections in Chile and 2.3 percent in eleven elections in Ecuador during the nineteenth century. In both countries, voter turnout fluctuated considerably in the nineteenth century, but it never exceeded 5 percent of the total population. Bolivia, Argentina, and Uruguay had the lowest levels of voter turnout in the nineteenth century, averaging 1.5 percent in thirteen elections in Bolivia, 1.4 percent of the total population in nine elections in Argentina, and 1.3 percent in seven elections in Uruguay.

Even those countries with relatively high levels of voter turnout saw their turnout levels drop considerably when they enacted suffrage restrictions. Only 0.8 percent of the total population voted in Brazil in 1881 after it implemented a literacy requirement, down from 8.6 percent of the population in 1872 (Castellucci Reference Castellucci2014, 194; Carvalho Reference Carvalho, Posada-Carbó and Valenzuela2012, 27–28; Graham Reference Graham1990, 202, 332). Similarly, only 1.6 percent of the total population voted in the 1899 presidential elections in Peru after it tightened suffrage requirements, down from 12 percent in 1890 (Aguilar Gil Reference Aguilar Gil2002, 25; Chiaramonti Reference Chiaramonti and Malamud2000, 249; Tuesta Soldevilla Reference Tuesta Soldevilla2001, 611). Conversely, the loosening of suffrage requirements could lead to a dramatic expansion in voter turnout, as it did in Venezuela after 1857 and in Colombia, briefly, after 1853. Nevertheless, even maintaining broad suffrage rights was no guarantee of high voter turnout, as the Argentine case attests.

During the nineteenth century, voter turnout in South America was on the whole low in comparison to the more developed Western countries. According to V-Dem data, voters represented an average of 10.4 percent of the population in fifty elections in the United States, 12.4 percent in eight elections in Canada, 4.9 percent in eighteen elections in Australia, and 22.2 percent in fourteen elections in New Zealand during the nineteenth century. Voter turnout was also higher in Europe than in Latin America on average: Voters represented an average of 7 percent of the population in the seventeen European countries for which V-Dem has data on nineteenth-century elections. Nevertheless, there were some notable exceptions: Belgium, Greece, Italy, the Netherlands, Norway, and Sweden all had voter turnout that averaged less than four percent of the population during the nineteenth century.

Thus, a comprehensive analysis of suffrage rights and voter participation in South America during the nineteenth century shows that suffrage restrictions were significant and voter turnout tended to be low, particularly in comparison to more developed countries in other regions. A few South American countries enjoyed universal male suffrage and/or high voter turnout for parts of the nineteenth century, but the norm was to maintain significant economic and/or literacy restrictions as well as gender restrictions, all of which depressed voter turnout.

Electoral Competition and Democracy in the Nineteenth Century

How competitive or democratic were elections in South America during the nineteenth century? As revisionist scholars have pointed out, Latin American countries had numerous competitive elections during the nineteenth century. The opposition frequently won representation in the legislature, and on a few occasions even prevailed in presidential elections. Nevertheless, comprehensive data from LAHED indicate that most presidential elections were uncompetitive, and many elections were not even contested. Fraud, intimidation, and government manipulation deprived the opposition of any chance of winning most presidential contests. Although a few elections were relatively democratic, the openings they produced did not last long and left no enduring influence.

Throughout the nineteenth century, governments, parties, and candidates employed numerous strategies to manipulate elections. These included:

It is difficult to assess the true extent of fraud and intimidation because the losing sides often denounced abuses without providing specific evidence of them, but it is clear that abuses were widespread. There was considerable variation in the degree of electoral manipulation, however. Moreover, electoral fraud and intimidation were far from unique to Latin America. Contemporary elections in the United States and Europe suffered from similar problems.

All sides engaged in electoral abuses, but the national government typically had the most opportunities to manipulate elections since it usually controlled the electoral authorities and could direct the military and the police to intervene. In some countries, the executive branch became known as the great elector since it decided the fate of elections. Nevertheless, the opposition was also guilty of electoral abuses. The opposition, for example, sometimes controlled the local electoral authorities or the police, which intervened on its behalf. In addition, the opposition at times resorted to violence to intimidate government supporters or to protest electoral abuses. Indeed, many of Latin America’s civil wars in the nineteenth century stemmed at least in part from electoral disputes.

The prevailing electoral regulations and laws facilitated the electoral abuses. In the nineteenth century, Latin American countries, with few exceptions, did not have or did not enforce the secret ballot. This made it easy for the electoral authorities and others to identify, intimidate, and disenfranchise opposition supporters. Similarly, the counting of the ballots at the polling places, and the initial absence of voter registries and party representation at the polls, provided numerous opportunities for cheating (Sabato Reference Sabato2018, 77). In addition, for much of the nineteenth century, Latin American countries used the complete-list electoral system, which awarded all the legislative seats in a district to the party or candidate list that finished first, thereby making it more difficult for the opposition to win legislative representation.

The electoral abuses led the opposition to frequently boycott elections in the nineteenth century. According to LAHED data, 36 percent of all presidential elections during the nineteenth century were uncontested, although in some of these elections more than one candidate pursued the ruling party’s nomination. In some cases, opposition candidates initially mounted campaigns but withdrew from the race when they saw that they had no chance of winning. Presidential elections that took place in the legislature or in constituent assemblies were particularly likely to be uncontested, as were elections that occurred after the previously elected constitutional presidents had been overthrown by force. Forty-nine percent of presidential elections that took place in Congress or in a constituent assembly were uncontested, as opposed to only 27 percent that took place via other means (i.e., via a direct popular election or an electoral college). Similarly, 46 percent of elections where the previously elected president had been overthrown were uncontested, as opposed to only 29 percent of other types of elections.Footnote ¹⁹

The percentage of elections that were uncontested increased over the course of the nineteenth century: 45 percent of the elections in the last three decades of the nineteenth century were uncontested, as opposed to only 28 percent before 1870. As Table 2.3 indicates, uncontested elections were most common in Paraguay, Uruguay, Venezuela, Chile, Peru, and Bolivia (in descending order). In these countries, uncontested elections represented at least one-third of all presidential elections during the nineteenth century. In Paraguay and Uruguay, they represented a majority of the elections in this period.

Even where elections were contested, they were usually uncompetitive. During the nineteenth century, the winning candidate in South American presidential elections won an average of 81 percent of the vote. These data include both contested and uncontested elections, but if we only count the former, the winner still captured an average of 72 percent of the vote. Direct elections were somewhat more competitive than indirect elections, but the winning candidate nevertheless won an average of 75 percent of the vote in direct presidential elections during the nineteenth century. Perhaps the least competitive types of election were those that took place after the previously elected president had been overthrown. In these elections, the winning presidential candidates typically won 87 percent of the vote.

The competitiveness of elections varied somewhat across countries and over time. Paraguay had the least competitive elections since all but one of its elections in the nineteenth century were uncontested. It was followed by Venezuela, Chile, Peru, and Uruguay, where the winner averaged between 83 and 86 percent of the vote. Elections in Brazil, Ecuador, Argentina, and Bolivia were not typically competitive either: The winner earned between 75 and 80 percent of the vote on average in these countries during the nineteenth century. Only in Colombia, where the winner averaged 68 percent of the vote, were elections usually competitive, and even there, competitive elections represented a narrow majority of all presidential elections if we define competitiveness as the winner earning less than 70 percent of the vote. The winner’s share of the presidential vote fluctuated somewhat from decade to decade, reaching a peak of 89 percent in the 1880s. Elections generally became less competitive as the nineteenth century progressed. During the late nineteenth century, the winner earned an average of 85 percent of the presidential vote, as opposed to 77 percent during the early nineteenth century.

Latin American presidential elections were much less competitive than elections in the more developed Western countries during the nineteenth century. In twenty-five nineteenth-century presidential elections in the United States, the winner earned only 53.9 percent of the vote on average, according to V-Dem data. In Canada, the largest parliamentary party won an average of 41 percent of the vote during the nineteenth century, in Australia it captured a mean of 43 percent, and in New Zealand it captured an average of 55 percent. European parliamentary elections also tended to be much more competitive during the nineteenth century than South American presidential elections. For example, the largest legislative party in Great Britain only won an average of 54.7 percent of the vote in sixteen parliamentary elections in the nineteenth century, and in France, the largest party only won 53.2 percent of the vote.

The lack of competitiveness of most elections in nineteenth-century South America becomes even more clear if we use another measure of competitiveness: the winner’s margin of victory. In all nineteenth-century presidential elections, the winner’s share of the vote was sixty-eight percentage points more than that of the runner-up on average, and in contested elections, the victor still won by an average of fifty-three percentage points. The winner’s margin of victory in South America was lower before 1870 than in the last three decades of the nineteenth century, but in no decade did it drop below an average of seventy percentage points. The winner’s margin of victory was also lower in direct elections than in indirect ones, although winners still won by an average of fifty-six percentage points in direct elections.

The winner’s average margin of victory was much greater in South America than in the industrializing countries during the nineteenth century. In the United States, for example, the winner’s average margin of victory was only 14.5 percentage points in twenty-five nineteenth-century presidential elections, according to V-Dem data.Footnote ²⁰ In western Europe, the largest parliamentary party’s share of the vote was on average only eighteen percentage points higher than that of the second-largest party.

Although most elections were uncompetitive, a sizable number of elections did involve significant competition. In fifty-one presidential elections in South America during the nineteenth century, the winner won less than 70 percent of the vote. This represents more than one-quarter of the nineteenth-century presidential elections for which we have data. In thirty-four elections, the winner won less than 60 percent of the vote, and in fifteen elections the margin separating the winner and the runner-up was fewer than ten percentage points. Colombia had the largest number of competitive elections: Fifteen out of its twenty-eight elections in the nineteenth century were competitive, meaning that the winner earned less than 70 percent of the vote. Ecuador had the second highest number of competitive elections in the nineteenth century – in seven of its twenty-three elections the winner won less than 70 percent of the vote – followed by Bolivia (six out of twenty-one), Peru (five out of twenty-one), Venezuela (five out of twenty-six), Argentina (four out of ten), Chile (four out of eighteen), and Uruguay (four out of eighteen). Brazil had one competitive presidential election in the nineteenth century and Paraguay had none, but Brazil only had three presidential elections during the nineteenth century, whereas Paraguay had thirteen.

Nine presidential elections in the nineteenth century appear to have been relatively free and fair. I coded elections as free and fair when they did not have systematic violations that appeared to have affected the outcome of the election, even if there were some minor irregularities or incidents of fraud or manipulation. The coding scheme considered all aspects of the electoral process, including voter registration, but did not take suffrage rights into account in evaluating whether an election was free or fair. It is not easy to assess how free and fair elections were during this period, given the limited data available and the potential biases of the sources.Footnote ²¹ My assessment is necessarily uncertain, but it is based on a critical scrutiny of the sources available, granting more weight to those sources deemed more reliable.

The elections that appear to meet the criteria for free and fair elections during this period are: the 1834 election in Venezuela; the 1836, 1848, and 1856 elections in Colombia; the 1872 election in Peru; the 1873 and 1884 elections in Bolivia; the 1875 election in Ecuador; and the 1896 elections in Chile.Footnote ²² Most of these elections were competitive: on average, the winner won 45.4 percent of the vote in these elections, and a margin of 22.3 percentage points separated the winner from the runner-up. In five of the elections, the winner won by fewer than ten percentage points. More importantly, in four of these nine elections, the opposition candidate managed to defeat the candidate supported by the incumbent president. This is a strong sign that the election was free and fair since the government clearly did not intervene so heavily in favor of the official candidate as to prevent the opposition from winning. In the other five elections, there was no clear official candidate.

Unfortunately, only one of these elections led to an extended period of democracy and to the institutionalization of democratic practices. As Table 2.4 indicates, revolts overthrew presidents who had been elected freely and fairly in Venezuela in 1834, Bolivia in 1873, and Ecuador in 1875. The presidents who were democratically elected in Peru in 1871, and in Colombia in 1836, 1848, and 1856, also faced revolts, but they managed to survive them. However, these leaders responded to the revolts by repressing the opposition and manipulating elections in ways that brought an end to the short-lived democratic openings. In Chile, however, the 1896 election paved the way for a lengthy democratic episode that lasted until 1924. This extended experience with democracy helped to establish democratic practices that would make Chile, along with Uruguay, the most democratic country in South America for most of the twentieth century.

Thus, except for Chile’s 1896 election, the few democratic episodes that the South American countries experienced during the nineteenth century proved quite short lived. The governments that were elected democratically were either overthrown or themselves undermined democracy to retain power. As we shall see, it was not until the early twentieth century that some South American countries had prolonged experiences with democracy.

Early Twentieth-Century Democratization

A great divide opened in South America during the early twentieth century that would last, with a few interruptions, into the twenty-first century. Some countries, specifically Argentina, Chile, Colombia, and Uruguay, took important steps toward democracy during this period, enacting major reforms that expanded suffrage rights and helped make elections relatively free and fair. The initial democratic episodes in each of these countries lasted at least a dozen years, and they had even longer implications. Beginning in the early twentieth century, democratic rule became the norm in these countries, except for Argentina. Although all four countries experienced democratic breakdowns at some point in the twentieth century, most of the breakdowns in Colombia, Chile, and Uruguay were short lived.Footnote ²³ Indeed, over the course of the twentieth century, these countries collectively experienced far more years of democracy than of authoritarian rule.

Chile was the first country to experience a prolonged democratic episode in South America, which was facilitated by the enactment of an 1890 law that established safeguards to ensure the secrecy of the ballot. Beginning in 1896, Chile enjoyed a long period of relatively free and fair elections that was only interrupted in 1924. During this period, the executive branch ceased to impose its preferred candidates, although some electoral abuses, especially vote buying, continued to take place, especially at the local level. Elections were quite competitive during this time and an alternation in power occurred on several occasions. The 1896, 1915, and 1920 elections were particularly close, with the winner triumphing by only a few electoral votes. Voter turnout also increased significantly during this period, averaging 5.9 percent of the total population between 1900 and 1929, as opposed to 2.3 percent in the nineteenth century. Although Chile experienced a few military interventions between 1924 and 1932, it subsequently developed into one of the region’s most vibrant democracies, with high levels of political participation and contestation.

Uruguay established a strong democracy after the passage of the 1918 constitution, which mandated the secret ballot, universal adult male suffrage, and proportional representation. In the wake of this reform, electoral fraud and manipulation declined significantly, and voter turnout soared, typically exceeding 15 percent of the population, as opposed to 1.3 percent in the nineteenth century. National elections became highly competitive. Between 1919 and 1929, the winner in presidential elections won an average of 55 percent of the vote, as opposed to 83 percent in the nineteenth century. Presidential elections were often decided by a narrow margin: For example, in the 1922 election, the winner won by only 2.1 percent and in the 1926 election the victory margin was only 0.5 percent. More importantly, Uruguay managed to maintain its vibrant democracy for most of the twentieth century.

Colombia democratized in 1910 when it enacted a series of constitutional reforms that ensured the representation of minority parties, expanded suffrage rights, and strengthened the powers of the legislature and the judiciary. For the next few decades, Colombian governments generally conducted electoral processes fairly and respected political and civil rights. Minority parties won significant representation in the legislature and opposition candidates even won the presidency in 1914 and 1930. Voter turnout also rose significantly, averaging 7.1 percent of the population in presidential elections, as opposed to 3.4 percent in the nineteenth century. Not all presidential elections were competitive, but the ones held in 1910, 1918, and 1922 were. To be sure, some electoral manipulation and violence continued, most notably in the 1922 elections. Nevertheless, Colombia after 1910 became one of the more democratic countries in the region, and it remained democratic for most of the twentieth century, although it was plagued by intermittent violence.

Argentina began to democratize in 1912 following the passage of a sweeping electoral reform that established the secret ballot, made voting compulsory, and mandated minority representation. In the wake of this reform, voter turnout soared, averaging 8.6 percent of the population between 1912 and 1929, as opposed to only 1.4 percent in the nineteenth century. Electoral abuses also diminished considerably since political bosses could no longer easily monitor or control voter behavior. Elections became more competitive: Between 1912 and 1929, the winner won only 53 percent of the vote on average in presidential elections, as opposed to 79 percent in the nineteenth century. Minority parties gained greater representation in the legislature, and in 1916, the opposition UCR won control of the presidency for the first time. Democracy, however, proved to be less stable in Argentina than in the other three democratic pioneers. A military coup brought democracy to an end in 1930 and over the next five decades the country was only intermittently democratic. Nevertheless, from 1916 until 1930, Argentina had a vibrant, if imperfect, democracy.

By contrast, authoritarian rule persisted in the other South American countries during the early twentieth century, and elections became less competitive on average. In Bolivia, Ecuador, Peru, and Venezuela, electoral fraud and manipulation worsened in the early twentieth century, and in Brazil and Paraguay, elections showed meager, if any, improvement. Between 1900 and 1929, the winner averaged 92.4 percent of the vote in the six democratic laggards, as opposed to 83.9 percent in the nineteenth century. Although these countries experienced democratic episodes after 1930, until the 1980s the democratic openings tended to be short lived.Footnote ²⁴ Authoritarianism, not democracy, remained the norm in these countries for most of the twentieth century, and even today, these nations typically have shallower and more fragile democracies than the democratic pioneers.

Table 2.5 presents a simple comparison of electoral statistics for the democratic pioneers and the democratic laggards during the first three decades of the twentieth century. As the table indicates, from 1900 to 1929, voter turnout was much higher among the democratic pioneers, averaging 7.6 percent of the population, as opposed to 4.9 percent among the democratic laggards. A similar gap emerged with respect to the competitiveness of presidential elections during this period. Between 1900 and 1929, the winner of presidential elections in Argentina, Chile, Colombia, and Uruguay captured 69.9 percent of the vote on average, whereas in the democratic laggards, the victor won an average of 92.4 percent of the vote. In addition, the frequency of free and fair elections was much higher in Argentina, Chile, Colombia, and Uruguay than in the other countries during this period: sixteen of the twenty-six presidential elections held between 1900 and 1929 in the pioneer countries were relatively free and fair, as opposed to one out of forty-nine elections in the laggards. A series of t-tests indicate that the difference in the means of the democratic pioneers and laggards is statistically significant for all three variables (voter turnout, the winner’s share of the vote, and free and fair elections) at the 0.0001 level. Moreover, the gaps between the democratic pioneers and laggards on these indicators are even greater if we focus on the period after the pioneers enacted key democratic reforms – that is, after 1890 in Chile, 1910 in Colombia, 1912 in Argentina, and 1918 in Uruguay.

By contrast, no such democratic gap existed between the two groups of countries in the nineteenth century. Argentina, Chile, Colombia, and Uruguay did not experience more frequent free and fair presidential elections than the other South American countries during the nineteenth century. Approximately 5 percent of nineteenth-century presidential elections were free and fair in both groups, and a t-test of the differences in the frequency of free and fair elections does not approach statistical significance. The democratic pioneers did have slightly more competitive elections on average in the nineteenth century: The winner’s share of the presidential vote averaged 76.8 percent in the democratic pioneers compared with 83.9 percent in the laggards, a difference which is significant at the 0.05 level. However, this is mostly due to Colombia, which had the most competitive presidential elections in the nineteenth century, and Paraguay, which had the least competitive elections during this period. The difference in the means of the two groups loses statistical significance if we omit either Colombia or Paraguay from the sample. There was also a gap with respect to voter turnout in the nineteenth century, but it favored the democratic laggards. Whereas votes represented an average of 4.9 percent of the total population in Bolivia, Brazil, Ecuador, Paraguay, Peru, and Venezuela during the nineteenth century, they constituted only 2.3 percent in Argentina, Chile, Colombia, and Uruguay, a difference that was statistically significant at the 0.01 level.

As Table 2.5 indicates, the leading democracy indexes also show a democratic divide occurring in South America during the early twentieth century. Moreover, the differences in means of the democratic pioneers and democratic laggards are highly statistically significant for both the Mainwaring, Brinks, and Pérez-Liñán (MBP) and the V-Dem indexes. The MBP index does not cover the nineteenth century, but it lists Chile as semi-democratic beginning in 1900, Colombia as semi-democratic starting in 1910, and Argentina and Uruguay as democratic or semi-democratic beginning in 1916 (Mainwaring and Pérez-Liñán Reference Mainwaring and Pérez-Liñán2013, 67–68). By contrast, none of the democratic laggards are listed as becoming democratic or semi-democratic before 1930, except for Peru, which is coded as semi-democratic from 1912 to 1913 and again from 1915 to 1918.Footnote ²⁵

V-Dem reports dramatic increases in the Polyarchy index for Colombia beginning in 1910, Argentina in 1912, and Uruguay as of 1916, but the scores of the remaining countries either stagnate (Chile, Paraguay, and Venezuela) or decline (Bolivia, Brazil, Ecuador, and Peru) between 1900 and 1929 (Coppedge et al. Reference Coppedge2023). On average, V-Dem reports significantly higher Polyarchy scores for the democratic pioneers, especially Argentina and Uruguay, than for the democratic laggards during this period. As Table 2.5 make clears, however, Colombia scores slightly lower on V-Dem’s Polyarchy index than Ecuador and nearly the same as Brazil, Paraguay, and Peru, all of which I classify as democratic laggards during this period. Colombia’s low score on the Polyarchy index during this period presumably stems in part from the country’s continued suffrage restrictions. Indeed, Colombia scores well above all the democratic laggards on V-Dem’s measure of free and fair elections (v2eltype_6) during this period.

As the Conclusion shows, the gap between the democratic pioneers and the democratic laggards persisted for most of the twentieth century. Although the pioneer countries suffered military coups during the twentieth century, all of them except for Argentina experienced far fewer coups than the democratic laggards. With the exception of Argentina, the pioneer countries also enjoyed many more years of democracy during the twentieth century as a whole.

Thus, South America divided into two groups of countries during the first few decades of the twentieth century: those which democratized and those which did not. This division would persist throughout most of the twentieth century, albeit with some ups and downs. As the ensuing chapters show, the emergence of professionalized militaries and strong parties played a key role in bringing democracy to Argentina, Chile, Colombia, and Uruguay in the early twentieth century. Professionalized militaries helped bring an end to the revolts that undermined constitutional rule, disrupted elections, and led to state repression. Strong opposition parties, meanwhile, helped enact, implement, and enforce the electoral reforms that proved crucial to democratic progress. The South American countries that lacked strong parties and/or professionalized militaries, however, continued to struggle with political violence, personalistic rule, government repression, and electoral manipulation.

Conclusion

This chapter began by posing the question of when and where democracy first emerged in South America. Traditionally, scholars depicted nineteenth-century Latin American elections as authoritarian regimes with highly fraudulent elections, but in recent decades revisionist historians have pointed out that many democratic practices existed in the region during this period. This chapter, however, has shown that these democratic practices were the exception rather than the norm. Some South American countries adopted broad suffrage rights during the nineteenth century and enjoyed relatively high levels of voter turnout in elections, but the majority of South American countries maintained significant restrictions on the franchise, and voter turnout on the whole tended to be low. Although there were numerous competitive elections during the nineteenth century, most presidential elections during this period were uncompetitive and many were not even contested. Moreover, the few democratic episodes that occurred in the nineteenth century did not last long: The governments that were elected through relatively free and fair elections during this period were either overthrown or subverted democracy to maintain themselves and their allies in power.

It was not until the early twentieth century that South American countries enjoyed sustained periods of democracy, and even then, only some countries in the region democratized. Whereas Chile, Uruguay, Colombia, and Argentina adopted democratic reforms and repeatedly held free and fair elections with high voter turnout between 1900 and 1929, the other South American nations mostly moved in the opposite direction, deepening government intervention in elections, repressing the opposition, and clamping down on civil and political liberties. Except for Argentina, the countries that joined the ranks of democracies in the early twentieth century would remain democratic for most of the century. By contrast, authoritarian rule would remain the norm in the other South American countries until the 1980s.

The remaining chapters explore what led democracy to emerge in some South American countries and not others during the early twentieth century. Why were some countries able to enact democratic reforms and experience long periods of democratic governance after a century of almost uninterrupted authoritarian rule? And why did other countries remain firmly in the grip of authoritarianism?