Abstract
What Kind of an Outsider Was Marco Enríquez-Ominami in Chile in 2009? Was He an Outsider at All?
In consolidated democracies, political parties represent citizenry. 1 Parties nominate candidates to channel the ideas and preferences of citizens (Aldrich 1995; Dahl 1971; Lijphart 1984, 1990; Lipset 2000; Mair 1997; Przeworski 1991; Sartori 1976; Stokes 1999). In the United States, candidates from the Democratic and Republican parties take on average 98 percent of the votes and seats in Congress. In consolidated democracies, candidates who do not come from political parties experience a more difficult path to electoral success (Abramson et al. 1995).
We thank the anonymous reviewers and the attendants of the Midwest Political Science Association meeting of 2011 for their comments and suggestions. This paper was partially funded by Fondecyt Project #1120638 – How have electoral preferences, institutional incentives and internal party/coalition politics determined who wins and who loses in legislative and municipal elections in Chile, 1989–2009?
The generic term “third-party candidate” – coined in the context of two-party systems – means different things. Third-party candidates may be independents. They may belong to smaller, up-and-coming, circumstantial or short-lived parties. They may represent anti-systemic or reformist movements. Yet, third-party candidates, independent candidates, and other candidates that originate parallel to the established party system share at least one common characteristic: they are “outsiders.”
Not all outsiders are the same. Depending on the definition, different candidates will be labeled as outsiders. Barr distinguishes three types of candidates: outsiders, insiders and mavericks (Barr 2009):
An
a
Barr's definition assumes a stable party system. If a party system is unstable, there can be no insiders, as all candidates would be either outsiders or mavericks. In contrast, in a stable, but delegitimized party system, insiders would normally present themselves as outsiders or mavericks in order to persuade voters that they would, using Barr's wording, radically reshape the party system if elected. For example, in democracies with a crisis of representation (see Mainwaring 2006), most of the insider candidates have difficulty campaigning as party militants, and thus, as an electoral strategy, they label themselves as outsiders or mavericks.
Barr's definition also requires an ex post evaluation. Under many circumstances, we can only distinguish insiders from mavericks after the fact. In addition, outsiders inevitably turn into insiders if they win (whereas losing candidates can remain outsiders if they run again as independents). For example, Luiz Inácio Lula da Silva (“Lula”) can be described as having been an outsider in 1989. However, he would have been considered an insider long before he became president in 2002. Similarly, leaders like Hugo Chávez, Evo Morales and Rafael Correa were outsiders when they first reached office, but soon became insiders. Ollanta Humala, on the other hand, can be considered an outsider both in 2006, when he lost, and in 2011, when he won.
Carreras (2012) offers a similar categorization, identifying four different types of candidates: insiders, mavericks, amateurs, and full outsiders. He classifies candidates according to political experience and party ties. Those candidates who come from established parties and have prior political careers are insiders. Those who come from established parties but do not have prior political careers are amateurs. Those who come from new parties and have prior political careers are mavericks. Finally, those who come from new parties and are newcomers are full outsiders.
Carreras’ definition makes it almost impossible for candidates in countries with weak party systems not to be outsiders. Distinguishing between established parties and new parties depends on the party system and the specific context of each country. In Argentina, for example, the
Using Carreras's definition also requires one to judge what constitutes political experience. For instance, Rafael Correa, the Ecuadorian president, is coded as not having had a political career before he became a candidate (Carreras 2012). However, Correa served as minister of finance under Alfredo Palacio – admittedly, it was only for five months. Still, it was sufficient enough to make him known and launch a presidential bid. Failed 2003 Argentine presidential candidate Ricardo López Murphy is also classified as an outsider, but he had previously served as minister of defense (in 1999). Carreras also includes Evo Morales as having been an outsider in 2002, but not Lula in 1989, even though they both represented recently created, labor-based political parties and had plenty of political experience as labor union leaders. Carreras also considers 1993 Chilean presidential candidate José Piñera an outsider, even though Piñera had been a prominent cabinet minister in the Pinochet dictatorship. Similarly, prior political experience is difficult to assess for many candidates who ran in the first democratic election in their countries after a dictatorship.
Successful Latin American outsider candidates are normally lumped together. Sometimes they are automatically labeled antisystemic and/or populist. In recent years, because of the dominant Washington Consensus economic model, many of them have also been assigned the leftist and anticapitalist tag. The rise of Workers’ Party leader Lula in Brazil and the emergence of Evo Morales from the indigenous Movement for Socialism (MAS) in Bolivia have been grouped together with the rise of the independent, antisystemic, former military, populist leader Hugo Chávez as part of a “left turn” in Latin America (Cameron 2009; Castañeda 2006, 2006; Castaneda and Morales 2008; Roberts 2007; Weyland, Madrid, and Hunter 2010). Notwithstanding the fact that both Morales and Lula were leaders of established political parties and thus, strictly speaking, not antisystemic, they have been grouped together with other outsiders because of their left-wing credentials (Castaneda and Morales 2008; Cleary 2006; Leiras 2007; Lomnitz 2007; Lynch 2007; Mair 2007).
Under a different categorization, Colombia's right-wing independent candidate Álvaro Uribe was labeled a populist, along with Humala in Peru in 2006 and Hugo Chávez (Barr 2003; Dugas 2003; Ellner 2003; Levitsky and Cameron 2003; Roberts 2007; Shifter and Jawahar 2005). Independent antineoliberal candidate Rafael Correa won the presidential election in Ecuador in 2006 without any support in the concurrently elected Congress. Yet Correa is grouped with Evo Morales as part of the antineoliberal “pink tide” and considered evidence of weak party systems (Castañeda 2006, 2006; Castaneda and Morales 2008; Corrales 2006). Finally, despite her having been a member of the moderate, market-friendly and ruling center-left Concertación coalition, Michelle Bachelet's victory in 2006 in Chile was also cited as further evidence of a left turn in Latin America. Because Bachelet's electoral platform focused on reducing inequality, she was categorized as another manifestation of an antisystemic shift in the region (Morales 2008; Navia 2009; Siavelis 2006). For us, all these candidates shared one component: they were in some way outsiders. They can also be defined as mavericks or amateurs, depending on the criteria used.
Admittedly, many party and systemic candidates attempt to define themselves as outsiders – a common feature in industrialized democracies as well. However, the cases we have described in Latin America were outsiders beyond a convenient electoral claim. They had verifiable characteristics that made them outsiders, such as (1) being either third-party candidates or independents or (2) emerging through small, up-and-coming, circumstantial or short-lived parties; some candidates also represented antisystemic or reformist movements.
The success of outsider candidates has set off a debate on the reasons for their emergence. Five competing explanations can be derived from the literature: First, the consolidation of democratic institutions has allowed countries to move beyond the political and social cleavages that characterized their transitions to democracy. Second, the decreasing ideological identification along a left-right scale has allowed candidates to attract popular support. Third, dissatisfaction with the neoliberal Washington Consensus economic model has invigorated left-leaning candidates. Fourth, unfavorable social and economic conditions have fostered the rise of leaders who promise swift action. Fifth, the weakening of the party system has strengthened candidates who do not belong to traditional political parties.
The remainder of this article is organized as follows: The second part discusses the empirical and theoretical literature on the five specific explanations that have been advanced to account for the success of outsiders in Latin America. The third part focuses on the rise of ME-O in Chile. The fourth part derives testable hypotheses, from each of the five explanations, to determine whether ME-O was an outsider presidential candidate in 2009, as recent literature has suggested. In the fourth part we use polling data to see whether his electoral success corresponds to any of the five explanations. We report our results. The final part presents the conclusions.
Explaining the Rise of Outsiders
Explanation 1: Consolidation of Democracy
After the transitions to democracy from the late 1980s to the early 1990s in Latin America, scholars began to focus on the quality of democracy (Dominguez and Shifter 2003; Mainwaring and Scully 2010; O'Donnell 1996; Seligson and Smith 2010). Public opinion surveys reflect the consolidation of democracy. Support for democracy has moderately increased over the past 10 years. Although there are fluctuations in some countries, a majority of Latin Americans remain committed to democracy. In 2011, average support for democracy in the 18 countries included in the Latinobarómetro was 58 percent. Though that represented a 3 percent drop from 2010, it was still above the 2000–2010 average of 56.4 percent (Latinobarómetro 2011). The Latin American Public Opinion Project (LAPOP) has found similar stable support for democracy in Latin America (Seligson and Smith 2010). Notwithstanding challenges to widely studied democratic consolidation (Dominguez and Shifter 2003; Mainwaring and Scully 2010), democracy – at least in the minimalist sense of holding elections – is the only game in town.
As the electoral process has normalized, democratic consolidation challenges have emerged in Latin America. Countries have left behind the political and social cleavages that characterized their transitions. No longer defined by the divide between supporters of authoritarianism and advocates of democracy, people are considering new alternatives and presidential candidates who do not conform to the bipartisan transitional divide. Presidential candidates that were not key players in the transition would represent evidence of democratic consolidation rather than a threat to the political system.
Among the candidates who have emerged with the normalization of electoral politics, outsiders are among those who have obtained more benefits. Murillo, Vaishnav and Oliveros (2009: 87) suggest that the resurgence of leftist leaders is evidence of the “normalization of democratic politics” rather than a backlash or a revolution. Similarly, Greene and Baker (2011) find a strong presence of issue voting in elections, which also indicates the normalization of democracy. Outsiders who campaign on common-sense issues, even if on a leftist platform, are an attractive alternative for those voters who no longer identify with authoritarian/democratic-aligned candidates.
Explanation 2: Weak Ideological Identification
In multiparty systems, presidential candidates compete to win a simple or absolute majority of votes. When electoral rules call for a simple majority, candidates from large parties tend to position themselves closer to their ideological niche. When the rules require an absolute majority, candidates compete according to the distribution of voters’ ideology. If the distribution of ideological preferences is bell-shaped – with the majority being centrist – the two largest candidates will strategically moderate their campaigns to attract the median voter.
In Latin America, using ideological labels has been a predominant way to approach the electorate. Many parties can be typified according to traditional left, right and center etiquettes. Yet in recent elections, preferences have been increasingly explained by a utility function informed by short-term economic concerns rather than by long- and medium-range determinants, (Greene and Baker 2011). As in industrialized countries, voters are more likely to punish incumbent governments when economic conditions are harsh, unemployment rates grow, and inflation increases. When conditions are good, incumbents are rewarded with reelection (Duch and Stevenson 2008).
With the decrease of ideological identification reported across the region (Dugas 2003; Luna and Altman 2011; Luna and Mardones 2010; Mainwaring 2006; Seligson and Smith 2010), electoral strategies have run into important challenges. Candidates nominated by large parties face the challenge of running campaigns ideologically in tune with their parties but not with the electorate. In some cases, this conflict has even led candidates to resign from their parties and compete as independents to adopt more flexible positions. Thus, partisan candidates must not only compete against candidates that come from outside the system, but also against their ex-fellow partisans. In an ideologically decaying system, candidates who resign from parties to compete as independents have become increasingly attractive as they share the ideological label of their ex-party, but can also focus on issues important for nonpartisan voters.
Outsiders are more attractive when voters have a weaker ideological identification – or at least their ideology does not fully explain their electoral behavior. Under these conditions, outsiders can quickly rise as they attract popular support from voters who are not ideologically aligned. Outsiders that campaign on a nonideological platform can offer to solve issues that partisan candidates cannot – namely, because it is more difficult for party candidates to depart from partisan guidelines. Outsiders do not need to be consistent across the board. Thus, they are an attractive alternative for issue voters less inclined to believe ideologically aligned candidates.
Explanation 3: Resurgence of the Left
The recent success of left-of-center presidential candidates in Latin America has sparked a burgeoning body of literature explaining the causes and effects of the Left's resurgence. Some of this research has interpreted these successful candidacies as isolated events. For example, the Left's triumph in Brazil was attributed to Lula's personality, charisma and commitment to democracy (Cason and Power 2009); in Chile, to the unique context of long-lasting authoritarian rule (Bresnahan 2003; Garretón 2000; Siavelis 2006); and in Venezuela, to the endemic failure of the mainstream parties (Ramirez 2005).
A second approach has explained the rise of the Left as a wave of occurrences connected by more than chance alone. Castañeda makes the distinction between the social democratic Left and the populist Left. He argues that the
social democrats are open-minded and modern, while populists are closed-minded and strident; the former respect democracy, the latter are irresponsible and abusive; the former operate within an orthodox market framework, the latter is statist (Castañeda 2006: 29).
Chávez and Correa should be categorized in the former category, while Bachelet and Dilma Rouseff should be categorized in the latter category.
Left-wing candidates generally come from parties that are ideologically left of center. However, even those who do not campaign on a traditional ideological label can still be considered part of a leftist wave. In fact, many leftists come from outside the party system. In some cases, such as in Chile, this might be because the traditional Left has shifted to the Right (Navia 2010), whereas in Bolivia, the traditional Left has systematically ignored minority groups (Rice and Van Cott 2006; Van Cott 2009; Yashar 2005). This has resulted in the appearance of new left-wing candidates who are seeking to restore traditional and inclusive left-wing principles.
Left-wing candidates may be independents or party militants who have resigned from their respective parties. Those from traditional parties may be nominated presidential candidates by party elites (such as Dilma Rousseff in Brazil) or may win their party primaries (such as Ricardo Lagos in Chile). Alternatively, they may be independents trying to restore original leftist principles or recently resigned militants hoping to build a larger constituency for the Left (such as Morales in Bolivia).
Outsiders that campaign on a left-wing platform are an attractive alternative for voters who identify with the Left, yet do not wish to vote for an establishment candidate. In this scenario, there is not a shift in electoral preferences, but rather a reinvigoration of the Left following the adoption of more centrist or conservative positions by traditional left-wing parties.
Explanation 4: Populism
Populism is a personalist style of politics, which is generally characterized by a charismatic leader, a multiclass social base and an urban setting (Di Tella 1965, 1997). The golden era for Latin American populism is usually cited as being from the 1930s to the 1960s and identified with preeminent figures such as Cárdenas in Mexico, Perón in Argentina and Vargas in Brazil (Conniff 1999; Laclau 1977; Panizza 2005; Weyland 2001).
A second wave of populism occurred in the 1990s, when right-leaning presidents with short-term economic promises were voted into office. Among them was Alberto Fujimori in Peru, Carlos Menem in Argentina and Fernando Collor de Mello in Brazil (Barr 2003; Ellner 2003; Roberts 1995; Weyland 1999, 2001). The two waves of populism in Latin American have been attributed to the economic crises of the 1930s and 1980s (Remmer 2012).
In more recent years, the presence of populism has been related to the Left (Castañeda 1994). Populists have been accused of pursuing an irresponsible, less-than-modern, and sometimes-authoritarian path (Castañeda 2006). Castañeda argues that these populists have no interest in ideological coherence, solid economic diversification or democracy. Rather, their interest is popularity, which they maintain by handing out cash to supporters, squeezing money from the state and taxing the traditional agricultural and mining sectors.
Outsiders can become an attractive alternative for voters who sense problems are not being solved fast enough. In contrast to establishment candidates, outsiders can be strategically more effective in delivering quick verbal responses to the priorities of voters, especially in countries that have been dominated in recent years by traditional parties. Thus, we here operationalize the concept of populism as candidates who primarily campaign against inequality.
Explanation 5: Weak Party Systems
The inability of the multiparty systems to handle the political turbulence resulting from severe economic disruptions and demands for political inclusion was behind the breakdown of democracy in several Latin American countries – Brazil in 1964, Argentina in 1966 and 1976, Uruguay in 1973 and Chile in 1973 (Linz and Stepan 1978; O'Donnell 1979; Skidmore 1988; Valenzuela 1978).
After the transitions to democracy in the 1970s and 1980s, some countries sought to avoid the endemic problems of multiparty systems by introducing reforms that reduced the number of parties, but strengthened remaining parties. Despite these changes, there are still signs that party systems remain weak (Mainwaring 2006). Instead of creating more stable systems, electoral reforms have resulted in large groups of like-minded voters without legislative representation (Bunker and Navia 2010).
When the party system is weak, voters turn to nontraditional alternatives (i.e., outsiders) that claim to be better able to offer solutions to their demands. This “defection” of voters as well as sudden changes in electoral rules (which reduce entry barriers for outsiders) leads to an increase in outsider candidates (or, depending on the definition, mavericks or amateurs) such as Abdalah Bucaram in Ecuador in 1996. In fact, the rise of candidates who are not formally affiliated with political parties – including those who resign from parties – is an indication of the further weakening of a party system.
The Presidential Bid by Enríquez-Ominami in Chile in 2009
In Chile, the electoral rules and a history of strong parties in a stable party system discourage the emergence of independent and third-party candidates. Given its presidential election system with runoff provisions and its restrictive legislative electoral system, with two-seat legislative districts allocated by proportional representation, independents and third-party candidates find it difficult to get elected. Consequently, more than 90 percent of seats go to candidates from the center-right Alianza and center-left Concertación coalitions.
In presidential elections, where a runoff is required if no candidate wins an absolute majority in the first round vote, the Concertación and Alianza have been dominant. Four consecutive Concertación presidents ruled from 1990 to 2010. The first two presidents won an absolute majority in the first round. In 1999 and 2005, the Concertación candidates won in a runoff. In 2010, Alianza candidate Sebastián Piñera won in a runoff, becoming the first opposition candidate to win the presidency.
Third-party presidential candidates have been present in every election since 1989. They have run seeking to force a runoff, rather than credibly aspiring to win the election (Angell 2007; Soto and San Francisco 2005). In 1989, the anticipated victory by the Concertación facilitated the rise of a right-wing independent aspirant (Angell and Pollack 2000). In 1993, given the electoral strength of the Concertación, a number of issue-oriented candidates entered the race (Godoy 1994). In 1999 and 2005, third-party candidates prevented the Concertación from winning the presidency in the first round vote.
A different kind of third-party candidate entered the race in 2009. That election was characterized by the high approval for outgoing president Concertación Michelle Bachelet, the weak support for the Concertación coalition candidate and the surprising rise of an independent candidate, Marco Enríquez-Ominami (ME-O). As Table 1 shows, the Alianza candidate, Sebastián Piñera received 44.1 percent of the vote. The Concertación candidate, Eduardo Frei of the Christian Democratic Party (PDC) and former president, received 29.6 percent. ME-O received 20.1 percent, the highest ever for an independent candidate. In the run off, Piñera defeated Frei by 51.6 percent to 48.4 percent.
Presidential Election Results in Chile, 1989–2009
Note:
Alianza had two presidential candidates in 2005, transforming the first round vote in a de facto primary.
Source: Authors’ calculation with data from ≤www.elecciones.gov.cl≥.
ME-O was elected to the Chamber of Deputies in 2005 as a member of the PS, one of the Concertación parties. A filmmaker, well-known polemist and public intellectual, ME-O came from a family historically tied to the political elite. His mother was the daughter of a former senator and founder of the centrist PDC. His biological father, Miguel Enríquez, was a left-wing guerilla fighter in the 1960s, killed by the military dictatorship in 1974. On his biological father's side, ME-O's relatives have been centrist Radical Party legislators since the mid-1950s. His adoptive father, Carlos Ominami, was a well-known PS leader, a minister in the first Concertación administration and the incumbent senator in the region where ME-O was elected deputy in 2005.
ME-O's rise to power can be understood by the internal dynamics of the Concertación. After announcing that open primaries would be held to nominate their presidential candidate, the Concertación reversed its decision in order to devise a mechanism to guarantee the nomination of former President Frei. ME-O, who had complied with the initial requirements to run in the Concertación primaries, thus resigned from the PS and ran as an independent.
In Chile, registration requirements for independent presidential candidates are difficult to meet. Candidates need a notary public to certify individual endorsements from thousands of voters. The actual number has to exceed 0.5 percent of all votes cast in the previous presidential election. Endorsers cannot be registered as party militants. The candidate is not allowed to run if more than 30 percent of the endorsers are deemed ineligible, even if he or she has more than the required signatures. Thus, independent candidates must be able to generate the legal support and funding necessary to ensure that there is a notary public seal on each of the signatures needed (more than thirty-five thousand in 2009) to pass the threshold. Those requirements are designed to reduce the number of presidential candidates, but they also discourage independents from entering the race. Because registered parties have the power to nominate candidates directly, nominees have incentives to stay within parties or join existing parties.
ME-O ran a presidential campaign that sought to be different things to different people. Because he had been elected to the Chamber of Deputies on the PS ticket, he sought to attract support from disenchanted Concertación voters. Since he was significantly younger (he turned 36 in 2009) than the other candidates, he also targeted the younger vote. As he was not a part of the political elite that led the transition to democracy, his campaign platform left behind the discourse that had characterized Chilean politics since the late 1980s. Figure 1 shows selected media coverage of ME-O's campaign. The four images reflect the different campaign qualities that ME-O underlined to cater to different constituencies.
Selected Media Coverage of ME-O Campaign in 2009
The “Ominami show” headline in
ME-O aimed to be a catch-all candidate. He positioned himself as a moderate, but also raised contentious issues. He discussed abortion, an issue opposed by the Alianza and divisive within the Concertación. He also suggested that the state copper giant CODELCO should be partially privatized, a position adamantly opposed by the Left but quietly favored by the Alianza. He also promised institutional reforms to introduce more participatory mechanisms to Chile's rigid and top–down representative democracy.
As Table 2 shows, ME-O led the other candidates on personal attributes in pre-electoral polls. However, Piñera was more trusted to address government issues, which – according to 63 percent of those asked – was one of the three key attributes needed to be a good president. Being sincere came in sixth, with only 23 percent mentioning as one of the three key attributes for a good president.
Presidential Candidate Attributes in Pre-electoral Poll, October 2009
Source: Authors’ calculation with data from October 2009 Centro de Estudios Públicos poll.
The fact that he was trusted less than Frei and Piñera in the polls represented a major challenge for ME-O. A bigger problem was that ME-O was significantly more popular among younger voters, who were less likely to be registered to vote. Table 3 shows the breakdown in vote intention by registration status (using the Instituto de Investigación en Ciencias Sociales at the Universidad Diego Portales [ICSO-UDP] poll). ME-O was more competitive when nonregistered Chileans were included. Given that only 65 percent of eligible Chileans were registered, ME-O would have done better if more Chileans were incorporated. In 2009, voting was mandatory, but registration was optional. Many people – mostly those who turned 18 after 1990 – opted not to register. ME-O's vote intention among registered Chileans was 17.3 percent, but his vote intention among those not registered was significantly higher. Considering both the registered and unregistered population, ME-O was ahead of Frei in September of 2009.
Vote Intention by Registered and Nonregistered Voters
Note:
Total includes 6 cases of registration status unknown.
Source: Authors’ calculation with data from the ICSO-UDP poll, September 2009.
As Figure 2 shows, ME-O experienced a rapid rise as a candidate. His vote intention in April of 2009 was less than 10 percent; by August, it was almost 20 percent. Figure 2 presents all polls published in Chilean newspapers in 2009. Given their different methodology and reliability, the polls present noticeable fluctuations. Nevertheless, ME-O's rise was constant throughout the campaign.
Presidential Vote Intention, All Published Polls, January–December 2009
Figure 3 shows that when distributed on the left-right ideological scale, Chileans are mostly moderate. A bell-shaped distribution underlines the concentration of political preferences around the center. Piñera's strongest support was among right-leaning voters, whereas Frei and Arrate were most popular among left-of-center voters. ME-O had the widest dispersion in the ideological self-positioning of supporters.
Electoral Preferences by Left-Right Ideological Scale
Explaining the Rise of Enríquez-Ominami
In this paper, we seek to define whether ME-O was an outsider at all and, if so, which type of outsider candidate he was. Each of the 5 explanations discussed above is stated below in the form of a hypothesis:
Support for ME-O is explained by the consolidation of democracy, reflected by the disposition of voters to disregard the authoritarian/democratic-aligned candidates.
Support for ME-O is explained by the decline of ideological identification, reflected by the disposition of voters to prefer nontraditional candidates.
Support for ME-O is explained by the resurgence of the Left, reflected by the disposition of voters to identify with anti–Washington Consensus candidates.
Support for ME-O is explained by a demand for quick government action, reflected in the predisposition of voters to support candidates who solve problems fast, even if they don't ask voters for their opinions.
Support for ME-O is explained by the declining support for established parties, reflected by the predisposition of voters to favor antisystemic candidates.
The five hypotheses combine individual-level determinants with structural-level determinants. Normally, this would lead to the use of the micro- and macrolevel data to test each respectively. However, because we are testing whether the candidate campaigned as an outsider, we only need to use microlevel data. Pre-electoral polls capture political strategies well. In this case, we expected that if ME-O were indeed an outsider, it would be reflected in voter opinion. Furthermore, structural level data (such as national levels of inequality to measure populism) may introduce bias. It could be the case that high levels of inequality are so internalized that politicians no longer campaign on the issue. Thus, the more accurate manner of measuring populism is to directly ask the question.
To test these hypotheses, we use an ICSO-UDP pre-electoral national poll and a poll by the United Nations Development Programme (UNDP) Human Development team in Chile. The questionnaires, datasets and methodological information are publicly available at ≤www.icso.cl≥ and ≤www.desarrollohumano.cl≥, respectively. The fieldwork for the ICSO-UDP poll was done in late September and early October 2009, two months before the election. The poll was representative of 85 percent of the urban population (excluding cities in the far north and far south) and 70 percent of the entire national population (excluding the rural population). The fieldwork for the UNDP poll took place in July and August and was a probabilistic sample of the entire country.
The polls’ findings coincided well with the results from the first round vote on 16 December 2009 (see Table 4). The ICSO-UDP poll overrepresented the vote intention for ME-O as it included all voters (registered and nonregistered). Since we are interested in evaluating the causes for ME-O support, and not just his vote share, we included the entire population (registered and nonregistered). As support for ME-O was also stronger in urban areas, the representativeness of the ICSO-UDP poll slightly tilted results in his favor. The UNDP poll also over-represents ME-O. It did not inquire about the registration condition of respondents. In addition, it was conducted at the time the ME-O campaign was at its strongest (as shown in Figure 2). In both polls, ME-O and Frei were in a statistical tie. Polls conducted later during the campaign reflected an advantage for Frei, especially among registered voters.
ICSO-UDP Pre-electoral Polls and Actual Results from 2009 Election
Note:
Only includes valid votes for the 4 candidates.
Source: ICSO-UDP poll and ≤www.elecciones.gov.cl≥.
As Table 5 shows, ME-O's vote intention was higher among younger voters. In terms of socioeconomic status, support for ME-O was U-shaped, with higher support among the upper class and the low-middle class. Frei did better among working-class voters, and Piñera did well at the extreme ends of the socioeconomic distribution. Piñera also did better than Frei among younger voters, while Frei found his strongest support among women. Compared to Frei, ME-O was stronger among upper-class women and among lower-middle-class men and women.
Vote Intention in 2009 Presidential Election by Different Categories
Source: Authors’ calculation with data from ICSO-UDP poll.
We test each of the five hypotheses. In addition to our variables of interest, we use common control variables (age, sex and socioeconomic status) (Dockendorff, Figueroa, and López 2012; Izquierdo, Morales, and Navia 2008; Morales 2008) to draw a more accurate picture of how people felt about ME-O in comparison to the other candidates.
We used a multinomial logistic regression to identify voting preference patterns for each of the candidates. Following an established statistical methodology (Alvarez and Nagler 1995; Alvarez, Nagler, and Willette 2000; Dow and Endersby 2004), we compared the probability of each candidate of receiving a vote according to a predetermined set of predictors. Although it has been suggested that the binary choice model is restricted to simply comparing pairs of parties and making assumptions on the independence of irrelevant alternative properties on voters (Alvarez and Katz 2009), the use of this statistical model effectively fits the nature of the Chilean political landscape (Bunker 2010; Morales 2008; Ortega Frei 2003; Torcal and Mainwaring 2003).
Table 6 shows the results for the first hypothesis (the consolidation of democracy). The poll questioned people's preferences for a democratic or authoritarian government. We used that question as a proxy for the normalization of politics and democratic consolidation. Our expectation was that those less likely to identify with the authoritarian/democratic divide would be more likely to vote for ME-O. Table 6 confirms our predictions. According to the odds ratios, Piñera supporters more strongly back an authoritarian government or are indifferent to an authoritarian or democratic government. ME-O voters are more inclined than Piñera voters to believe that democracy is preferable to any other form of government. In terms of those who are indifferent about a democratic or authoritarian government and those who believe that an authoritarian government is sometimes better, ME-O supporters are situated between Frei and Piñera supporters. Those who are more opposed to authoritarianism were more likely to support Frei than ME-O. If Chilean politics in the 1990s was divided along the authoritarian-democratic cleavage that characterized the transition to democracy (Drake and Jaksic 1999; Menéndez-Carrión and Joignant 1999; Tironi and Agüero 1999), the rise of ME-O signals the weakening of that cleavage. MEO supporters were in between Frei's opposed-to-authoritarianism supporters and Piñera's inclined-to-authoritarianism voter base. Still, because Frei supporters were more inclined to support democracy than ME-O's, the young candidate's strong showing cannot be seen as definitive evidence of democratic consolidation blurring the alignments that characterized Chilean politics in the 1990s and the first decade of the twenty-first century.
Multinominal Logistic Regression of Democratic Consolidation on Presidential Vote Intention in Chile in 2009
Note:
Significant at 0.01;
Significant at 0.05. Results are presented in odds ratios. The empty rows are the reference category for the independent variables and the empty column is the reference category for the dependent variable.
Source: Authors’ calculation with data from 2010 UNDP Human Development Report poll (conducted in July/August 2009).
Table 7 shows the results for the second hypothesis (the decline of ideological identification). The poll asked people to identify themselves along a left-right continuum. We used that question as a proxy for weak ideological identification. Those who did not identify on the left-right scale (1–10) reflect weak ideological identification. In the ICSO-UDP and UNDP polls, 39.7 percent and 46.7 percent, respectively, did not identify on the left-right scale. Table 7 shows the difference in vote intention between those that align on a left-right scale and those who reject such alignment.
Multinominal Logistic Regression of Self-Identification on Ideological Scale on Presidential Vote Intention in Chile in 2009
Note:
Significant at 0.01;
Significant at 0.05. Results are presented in odds ratios. The empty rows are the reference category for the independent variables and the empty column is the reference category for the dependent variable.
Source: Authors’ calculation with data from 2010 UNDP Human Development Report poll (conducted in July/August 2009).
Compared to ME-O supporters, Frei supporters were less probable to identify on the left-right scale. This questions the perception that ME-O backers were less ideologically inclined than Frei's. To separate those who identified themselves anywhere on the left-right scale from those who did not, we created a dichotomous variable. We found that, in fact, Frei voters were less likely than ME-O supporters to be ideologically inclined. However, there was no statistically significant difference between ME-O and Piñera backers. Thus, we cannot claim that ME-O's electoral performance reflects a weakening of Chileans’ ideological identifications. ME-O's supporters were more likely to ideologically identify themselves than Frei's and were as equally likely as Piñera's. Even when controlling for the fact that ME-O sympathizers were younger and of higher socioeconomic status, there is no evidence that ME-O support was stronger among those who did not identify on the ideological scale.
Table 8 shows the results for the third hypothesis (the resurgence of the Left). As his background would place him to the left of the Concertación (though his campaign strategy and message might have been less openly leftist), we test for the effect of ideological self-identification on the vote for ME-O. We also test for the vote for Arrate, Frei and Piñera. As expected, ME-O's backers were more to the left than Piñera's, but less so than Frei's (although the latter findings were not statistically significant). Although we do not have sufficient evidence to assert that ME-O's and Frei's supporters were ideologically different, the indications are that Frei voters were to the left of ME-O's.
Multinominal Logistic Regression of Identification with the Left on Presidential Vote Intention in Chile in 2009
Note:
Significant at 0.01;
Significant at 0.05. Results are presented in odds ratios. The empty rows are the reference category for the independent variables and the empty column is the reference category for the dependent variable.
Source: Authors’ calculation with data from 2010 UNDP Human Development Report poll (conducted in July/August 2009).
Table 9 shows the results for the fourth hypothesis (the demand for quick government action). As a proxy for populism support, we used a question in the ICSO-UDP poll. Voters were asked “what do you prefer, a government that solves problems fast, without asking people for their opinion, or a government that takes longer to solve problems, but asks people for their opinion?” A total of 32.8 percent preferred a government that solves problems fast, while 67.2 percent favored a government that takes longer to solve problems.
Multinominal Logistic Regression of Populism on Presidential Vote Intention in Chile in 2009
Note:
Significant at 0.01;
Significant at 0.05. Results are presented in odds ratios. The empty rows are the reference category for the independent variables and the empty column is the reference category for the dependent variable.
Source: Authors’ calculation with data from the ICSO-UDP poll, 2009, ≤www.icso.cl≥.
There are two possible readings of populism in the question. The first reading is that a context in which voters want authorities to act fast, without asking people for their opinions, would represent an extreme case of trustee government or delegative democracy (O'Donnell 1994). Though their discourse often calls for popular participation, populist leaders normally weaken institutions by concentrating power in their own hands (Weyland 2003). Thus, those who want their presidents to act swiftly can be considered more likely to support populist candidates.
A second reading highlights the notion of participation as central to the populist leader's message. Those who prefer presidents who consult with people, even if the process delays decisions, would be more inclined to support populist candidates. Occasionally, populism is associated with a view of presidents as agents rather than trustees (Conniff 1999).
The table shows that Chileans who want leaders who solve problems fast were more likely to vote for Piñera than for ME-O; however, there was no statistically significant difference between Frei and ME-O voters. Thus according to our proxy, there is no evidence that support for ME-O was associated with populist inclinations. Admittedly, there is no perfect proxy for a concept as contested as populism (Weyland 2001). Yet, using our proxy, ME-O was less associated than Piñera with the authoritarian version of populism.
Finally, Table 10 shows the results for the fifth hypothesis (the declining support for established parties). If voters did not identify with any of 10 existing political parties, it would reflect a weak party system. In the ICSO-UDP poll, 56.2 percent indicated no party identification, while 43.8 percent selected one of the 10 political parties. The data fails to confirm the expected results. People who identified with parties were more likely to vote for ME-O than for Frei, but there was no statistically significant difference in the vote for Piñera and ME-O among this category. Although there was the perception that ME-O built his support among those less likely to identify with parties, the results show the opposite. In fact, Frei was more likely than ME-O to receive support among those who did not identify with parties. Therefore, despite the well-recorded decline in identification with parties in Chile (Luna and Altman 2011; Morales 2012), there is no statistically significant evidence that links support for ME-O to this falling identification with parties.
Multinominal Logistic Regression of Populism on Presidential Vote Intention in Chile in 2009
Note:
Significant at 0.01. Results are presented in odds ratios. The empty rows are the reference category for the independent variables and the empty column is the reference category for the dependent variable.
Source: Authors’ calculation with data from the ICSO-UDP poll, 2009, ≤www.icso.cl≥.
Conclusion
The rise of outsider candidates in Latin American elections has been explained by references to (a) democratic consolidation processes that have left behind the transition-to-democracy equilibrium, (b) weakening ideological identification, (c) the resurgence of the Left, (d) the rise of populist leaders, and (e) a weak party system. In analyzing the independent Chilean presidential bid by ME-O in 2009, we found that none of the aforementioned explanations could account for the support (20 percent) he received. To the contrary, it seems that ME-O's candidacy simply reflected an internal breakdown within the ruling Concertación coalition. ME-O did manage to attract more support among younger and more-educated voters than Frei. However, his electoral base did not comprise the economically marginalized, populist-prone voters or ideologically disenchanted or nonideological Chileans. Though ME-O became an outsider following his resignation from his former coalition, he should not be equated with antisystemic candidates who elsewhere threaten to redefine the political-party structure or who generate support among voters disenchanted with existing parties. In the truest sense of the words, ME-O was neither a populist nor an outsider.