國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Chapter One: Introduction
1.1 Puzzle
Elections are crucial for the proper functioning of the country’s political, social, and economic development, especially for democratic countries (Powell 1982). Elections allow citizens to accept or reject their representatives and their policy platforms (OSCE 2013), but it is common to observe null and blank ballots in every election. Those ballots are considered as invalid votes, and they reflect no support for parties or candidates. The presence of invalid voting has gained scholarly attentions (e.g., Cohen 2018; Lysek et al. 2019; Martinez I Coma and Werner 2018; Power and Garand 2007; Superti 2013; Uggla 2008). While some studies contend that invalid voting results from the absence of voters’ political skills or information about the elections (e.g., Pachon et al. 2017; Pierzgalski et al. 2019), other studies suggest that invalid voting is an expression of voter discontent or apathy (e.g., Cohen 2018; Power and Garand 2007; Uggla 2008).
Chilean Senator Navarro B. Alejandro argues that "The blank vote is a rejection against politicians and their programs, symbol of democracy, which gains adherents every day and is considered by many as an exemplary protest and as the best way to reject corruption and abuse of the oligarchic system of political parties" (Navarro 2010). Similarly, Malamud (2018, 5) mentioned, “The ‘anger vote’ is a cathartic act which response to the grave and widespread unrest and discontent with national institutions and representatives.”
Using invalid voting as a form of political protest, influential elites or social groups might encourage the supporters to spoil their vote during a democratic crisis and repetitive political disenchantment. For instance, during the Argentine Constitutional Assembly Election of 1957, Domingo Peron called on voters to cast blank ballots (Baeza 2016). Eventually, invalid votes topped with 25% of the total votes, considering it as a political signal of the Peronist Resistance (Baeza 2016). In Argentina's 2001 legislative election, a strong campaign encouraged people to cast the so-called "voto bronca" (anger vote). It was an action for "the anger of citizens with the political class, which many accuse of corrupt and inoperative to get the country out of the severe
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
economic recession" (Sabanes 2001). In this manner, the "voto bronca" became a "democratic option" for argentine voters (EL País 2001). During the second round of Argentina's 2015 presidential election, Nicolás del Caño (candidate for the Left and the Workers' Front) called to cast a blank ballot as an opposition to the political agenda of Mauricio Macri and Daniel Scioli (La Izquierda Diario 2015).
In Mexico, during the Lower Chamber election in 2009, campaigns for null votes through social networks such as #VeVotaAnula or "EllosNoNosRepresentan" ("They do not represent us") were regarded as flag of protests against the high rates of crimes, economic crises, and impunity in cases of corruption (Cisneros 2012). In Peru, during the second round of Presidential elections in 2001, Alvaro Mario Vargas Llosa and Jaime Bayly promoted the blank vote, arguing that "We are going to register a third candidate. That is the only clean candidate we can trust; the blank vote" (La Nación 2001). In the 2017 Ecuadorian presidential election, Jorge Cedeño López, the leader of the "Organization, Production and Development" movement, led a strong null vote campaign (El Universo 2017). In the 2018 Colombian presidential election, two influential politicians, Sergio Fajardo y Humberto de la Calle, stated their support for the blank vote, declaring that neither the right-wing candidate Ivan Duque nor the left-wing candidate Gustavo Petro was convincing them (Gonzales 2018).
Invalid voting varies widely across different countries and periods. Among different continents, Latin America has the highest invalid voting rates in Presidential and Lower Chamber elections in the world. On average, invalid ballots in Latin America represented more than 5 percent of all votes cast in Presidential elections and nearly reached to 9 percent of all ballots observed for Lower Chamber elections. In contrast, in regions such as Europe and Asia, rates of invalid voting are less than 3 percent on average. Hence, the high invalid votes in Latin America poses an interesting empirical puzzle.
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Figure 1: Invalid vote by World Region (1980-2019)
Source: International Institute for Democracy and Electoral Assistance (International IDEA 2019)
Amid a climate of political instability and citizens' support for democracy decreasing (Corporación Latinobarómetro 2018), Latin America faces a wave of significant increasing of invalid votes of the last four decades. As Figure 2 shows, the invalid voting rates have been generally higher for the lower chamber elections than that for the presidential elections, except Bolivia and Panama. The invalid voting rates in Mexico, Paraguay, and Uruguay are the lowest among all Latin American countries, while the invalid voting rates in Brazil, Ecuador, and Peru are the highest.
The empirical data suggest that focusing on only a few explanatory variables might not be sufficient to explain the variation of invalid voting in different countries. For instance, Ecuador, Peru and Brazil, where compulsory voting is strictly enforced by a system of sanctions displayed the highest percentages of invalid votes. However, using similar electoral regulation, Uruguay had the lowest level of invalid votes. In countries where voting is compulsory but without penalties, invalid voting rates were high in Guatemala and El Salvador but not in Costa Rica and
0 1 2 3 4 5 6 7 8 9 10
Latin America Asia Africa Europe Oceania
% Invalid vote
Presidential Lower Chamber
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Paraguay. In Chile (post-2012), Mexico, and Colombia, where voting is voluntary, the percentage of invalid votes is low for presidential elections only.
Figure 2: Percent Invalid vote in Presidential and Lower Chamber elections among Latin American countries (1916-2018)
Source: Author. Data are from Nohlen (2005a; 2005b) and International IDEA (2019).
Democratization process might matter for explaining the variation in invalid voting. As Mainwaring and Pérez- Liñán (2005) argue, the third wave of democratization has been one of the most extensive around the world, involving a high number of countries and lasted for longer than any past waves. The overall trend towards democratization started in different years for each Latin American country. The first countries to celebrate them were the Dominican Republic and Ecuador in 1978; Peru in 1980, Honduras in 1981, followed by Bolivia (1982), Argentina (1983), Uruguay, and Brazil (1985); Paraguay, Panama, and Chile in 1989, Nicaragua (1990), El
0 5 10 15 20 25 30
% Invalid vote
Presidential Parliament
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Salvador (1994), Guatemala (1996). Finally, Mexico incorporated institutional mechanisms to enable the holding of competitive elections since 1997. Before the Third Wave of Democratization, Costa Rica reached its democratic regime in 1949, followed by Colombia and Venezuela in 1958 (Alcántara and Freidenberg 2006).
As Power and Garand (2007) contend, elections held in more democratic countries should generate lower levels of invalid voting. Did the third wave democratization affect the variation of invalid votes in Latin America? Figure 3 compares the average rates of invalid voting in presidential elections in sixteen Latin American countries before and after the third wave democratization. In eight of the sixteen Latin countries that held presidential elections such as Argentina, Bolivia, Dominican Republic, El Salvador, Guatemala, Mexico, Nicaragua, and Panama, the variation of invalid votes was lower after the third wave. Nevertheless, contrary to what Power and Garand (2007) have theorized, in countries such as Brazil, Chile, Colombia, Ecuador, Honduras, Paraguay, Peru, and Uruguay, the average rates of invalid voting were higher after the third wave. Moreover, Figure 4 compares the average rates of invalid voting in lower chamber elections in thirteen Latin American countries before and after the third wave democratization.
Again, contrary to Power and Garand (2007), the invalid voting rates are higher in the post-third wave period for most Latin American countries.
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Figure 3: Invalid vote rates in presidential elections by Latin American countries Pre vs Post the Third Wave of Democratization
Notes: The availability of information limits the range of years of each country: Argentina (1916 to 2015); Bolivia (1951- 2014); Dominican Republic (1962 - 2016); El Salvador (1984 - 2014); Guatemala (1958 - 2015); Mexico (1982 - 2018); Nicaragua (1984 - 2016); Panama (1948 - 2014); Brazil (1945- 2018); Chile (1938-2017), Colombia (1946 - 2018); Ecuador (19682017); Honduras (19482017); Paraguay (19582018); Peru (19562016) and Uruguay (1984 -2014). Likewise, due to the limited information, Costa Rica and Venezuela were not included in this Figure.
Source: Author. Data are from Nohlen (2005a; 2005b) and International IDEA (2019).
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
% INVALID VOTE
Before 3rd WD After 3rd WD
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Figure 4: Invalid vote rates in Lower Chamber elections by Latin American countries Pre vs Post the Third Wave of Democratization
Notes: The availability of information limits the range of years of each country: Argentina (1916 -2015); Bolivia (1956-2014); Brazil (1950-2018); Colombia (1951-2018); Dominican Republic (1962-2016); Ecuador (1962-2017); El Salvador (1985-2018); Guatemala (1959-2015); Mexico (1970-2018); Panama (1978-2014); Paraguay (1963-2018);
Uruguay (1971-2014) and Venezuela (1947-2005). Due to the limited information, Chile, Costa Rica, Honduras, Nicaragua and Peru were not included in this Figure.
Source: Author. Data are from Nohlen (2005a; 2005b) and International IDEA (2019).
The growing invalid votes in Latin America elections has been a remarkable phenomenon.
The above discussion shows that the level of Latin America’s invalid vote is higher than that of other regions, and there is a great variation of invalid voting rates within the region. Moreover, it seems that electoral systems and the democratization process are insufficient to explain the variation in invalid voting rates. Although, it is expected a different situation when this variation of invalid voting comes from the voter discontent or apathy against corruption committed by politicians or the incumbent government and its economic measures as it is seen in the context of mentioned invalid vote campaigns.
0 5 10 15 20 25
Before 3rd WD After 3rd WD
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Therefore, the invalid voting in Latin America poses an empirical puzzle. To address this puzzle, this thesis aims to systematically examine the factors that impact voters’ propensity to cast invalid votes. Specifically, I will test hypotheses regarding compulsory voting system, concurrent elections, level of economic performance, level of education, level of democracy, and corruption on invalid voting rates. Based on data of presidential elections and lower chamber elections in 18 Latin American countries from 1916 to 2018, I conduct mixed-method analyses to analyze the impacts of different factors. I conduct quantitative analyses to show the relative importance of each variable on invalid vote. I also conduct qualitative analyses to illustrate how invalid voting campaigns work in Latin America. In the aggregate, invalid voting responds predictably to certain features of institutional, socioeconomic, and political such as compulsory voting system, concurrent elections, economic development, and corruption, in a different way, for the presidential and lower chamber election. And while I find evidence that a country with an enforced compulsory voting system and a high level of corruption affect more the increasing of invalid vote rates for lower chamber elections than presidential elections like it happened in Argentina 2001 and Ecuador 2006 elections. Concurrent elections tend to increase invalid vote only for presidential elections; and, a higher level of economic development tends to increase invalid vote just for lower chamber elections. Overall, this study aims to fulfill the gap in the literature and provide policy implications for voting behavior and democratization.