Abstract
Research has shown that countries with weak institutions are more likely to experience higher levels of protest as a means to achieve political objectives or express policy demands. A growing body of literature portrays Argentina as a case of widespread institutional weakness, and the country currently sustains the highest rates of protest participation in Latin America. However, existing literature has yet to explain why apparently similar subnational units within the same national democratic regime experience different levels of protest. By moving down to a subnational level of analysis, this article explores the political factors that shape protest activity across the country’s 23 provinces and the city of Buenos Aires for the period 1993–2005. It demonstrates that the electoral incentives created by shifting patterns of political competition and the nature of partisan opposition influence the spatial and temporal unevenness of subnational protest activity.
Following Latin America’s third democratic wave that began in 1978, much of the scholarly literature anticipated that conflicts involving society would be resolved through parliamentary institutions that evoke compliance among the relevant political forces. However, starting in the 1990s several Latin American countries began to experience a generalized rise in various new forms of political protest. These protests were effective in rolling back unpopular economic policies, such as privatizations of natural resources, government utilities, pension systems, and social services. Waves of street protests also forced presidents who supported market policies to leave office early.
Prior research on the rise of protest activity has emphasized the importance of democracy as an intervening variable affecting the relationship between economic liberalization and protest activity (e.g., Almeida, 2009; Arce & Kim, 2011) and further shown that the quality of institutional representation shapes collective action in the streets (e.g., Arce, 2010; Machado, Scartascini, & Tommasi, 2009). On average, citizens in countries with strong and well-institutionalized parties are less likely to resort to protest as a mechanism to achieve political objectives or express policy demands compared to citizens in countries with weak and poorly institutionalized parties. These explanations, however, largely dwell on cross-national comparisons and, as such, suffer from two related limitations. First, it is widely known that institutional structures vary far more across countries than they do within countries. Therefore, an institutional explanation is less theoretically compelling for countries stagnated by weak institutions. Second, variation in patterns of mobilization across subnational boundaries is abundant (Murillo & Ronconi, 2004; Wilkinson, 2004). This forces us to ask the following: What political factors account for the subnational unevenness of popular contention? Why do apparently similar subnational units within the same national democratic regime have different levels of protest?
Our article focuses on Argentina, which represents an ideal setting to explore these questions for two reasons. First, based on recent public opinion data (Lodola & Seligson, 2011, p. 179), Argentina currently sustains the highest rates of protest participation in Latin America. The incidence of protest also varies considerably within the country (Delamata, 2002; Favaro, Iuorno, & Cao, 2004). Second, the country has been recently portrayed as a case of widespread institutional weakness (Levitsky & Murillo, 2005), and parties and protest activity have become increasingly intertwined in the present democratic period. Hence, an analysis of the political factors that shape protest activity in Argentina sheds light on similar processes elsewhere in the region.
To explain Argentina’s cycle of popular contention, and seeking to correct the analytical bias of the existing literature in favor of cross-national comparisons, we draw attention to the electoral and partisan dynamics that make subnational provinces more protest prone. As Gibson and Suarez-Cao (2010, p. 29) write, the country is “a mosaic of distinctive provincial political regimes,” where political competition and the powers of provincial political actors vary considerably. Therefore, our sample includes Argentina’s 23 provinces and the city of Buenos Aires for the period from 1993 to 2005. This time frame under investigation begins with the violent unrest in the northwestern province of Santiago del Estero (also known as the Santiagazo riots), which marked the beginning of the spiral of protest in the Argentine interior provinces, and concludes with the first 3 years of the Néstor Kirchner presidency, whose successful government policies diminished social conflict following the collapse of the country’s Convertibility Plan.
Our main empirical results demonstrate that the electoral incentives created by shifting patterns of political competition and the nature of partisan opposition influence the spatial and temporal distribution of subnational protest activity. These results show that political competition in general and partisanship in particular still matter, even in a period when representative institutions have been discredited because of poor government stewardship and corruption scandals. After providing background information on Argentina’s most recent cycle of protest, we summarize the data and methods used in the article. We then present our empirical findings. The last section concludes by suggesting new areas of research on the repoliticization of protest activity after Latin America’s third democratic wave.
Protest in Argentina After Economic Liberalization
Popular contention has played a major role in key political developments in Argentina. Prior to the country’s transition to democracy, which began in 1983, popular mobilizations ushered in coups and brought cohesiveness to military regimes, as in the junta of Jorge Rafael Videla in 1976. Protest also led to the collapse of these regimes, as in the mobilizations during the early 1970s that restored civilian rule and brought Argentina’s historical leader Juan Domingo Perón back into power. During the present democratic period, the implementation of economic reform policies under President Carlos Menem (1989–1999) triggered a sustained wave of antimarket contention (Silva, 2009). Most observers agree on three founding episodes of contention: the 1993 Santiagazo riots, which initiated the wave of protest in the Argentine interior provinces; the 1996 mobilizations in the province of Neuquén that led to emergence of the unemployed piqueteros (or picketers); and finally the massive protests that began in the province of Corrientes in 1999 and culminated in the city of Buenos Aires in December 2001, against the then president Fernando De la Rúa (1999–2001). Persistent protests triggered a string of presidential resignations, a total of four starting with De la Rúa, and popularized the slogan que se vayan todos (“throw everyone out”).
As a result of these episodes of contention, Argentine democracy has been widely recognized as a true landscape of collective insurgency and groundbreaking repertoires of popular contention (Auyero, 2001). Attacks on governmental buildings and politicians’ houses, national and provincial road blockades, banging pots and pans, and setting up camps at civic squares, soup kitchens, and popular assemblies have all epitomized what Peruzzotti (2001, p. 141) characterized as the productive “politicization” of the country’s civil society. This cycle or wave of popular contention, which originated as a consequence of economic liberalization policies, has also introduced new patterns of claim making and produced formal and informal arenas of political bargaining in a widespread context of democratic political competition and economic change. The enduring contentious phenomenon in Argentina is particularly striking in view of the literature that formerly regarded the nation as a paradigm of party system institutionalization (Mainwaring & Scully, 1995) and democratic consolidation (Schamis, 2002).
In this cycle of protest, road and street blockades are considered the most salient form of protest activity (Svampa & Pereyra, 2003). Starting with the 1993 Santiagazo riots, well-known episodes of roadblocks have occurred in the provinces of Río Negro (1995), Neuquén (1996), Jujuy (1997), Corrientes (1999), Salta (1997 and 2000), and Buenos Aires (2002). 1 Roadblocks continued to expand throughout the Argentine interior provinces (Favaro et al., 2004), even after the collapse of the promarket government of De la Rúa and numerous transitional presidents in the biennium 2001–2002. For instance, although there were 184 roadblocks during the period 1993–1997, the number of roadblocks swiftly increased to 4,536 in the period 1998–2002. Mounting protests in the provinces of Santa Cruz in 2004 and Entre Ríos in 2005 have reaffirmed this rising contentious tendency. 2
Figure 1 compares the country’s national level of strikes with the number of roadblocks. These figures are measured according to the number of mobilizations that took place on any given day and represent the sum of the absolute number of mobilizations for all provinces for the period of our study. The graph reveals the changing basis of antigovernment mobilizations following Argentina’s transition to a market economy, paralleling trends in other Latin American countries and showing organized labor’s decline in political clout, thus making room for other actors and forms of popular resistance. Although strikes typically disrupt commerce at the point of production, roadblocks do the same but at the point of exchange (Silva, 2009). Yet protest activity has not been uniformly spread over the country. As Auyero (2001, pp. 34-35) makes it clear, it is central to underscore that “contention might occur more frequently in some regions and during certain periods of time,” and in Argentina “there is an increasing body of evidence that points to the spatial and time patterning of . . . protest.” For example, the numbers of roadblocks in northern provinces such as Tucumán, Salta, and Jujuy as well as more populous provinces such as Buenos Aires and Santa Fe have surpassed the sample mean of 25.72 roadblocks for the period 1993–2005, whereas the remaining 19 provinces (including Córdoba, the second most populous province after Buenos Aires) have on average experienced fewer roadblocks than the sample mean.
Comparison of strikes and roadblocks, 1993–2005.
To further illustrate the variation of roadblocks across provinces and across time, Figure 2 presents the number of roadblocks for four provinces (Corrientes, Entre Ríos, Tucumán, and Salta), which are very similar to each other when it comes to population size (between 1 million and 1.5 million people). Geographically, these provinces occupy the country’s central northeastern corridor, which suffers the greatest concentration of roadblocks. Yet across the sample period, the mean levels of roadblocks for the provinces of Corrientes and Entre Ríos are 5.38 and 8.23, respectively, a good distance from the sample mean. In contrast, the mean levels of roadblocks for the provinces of Tucumán and Salta are 29.23 and 32.77, respectively, which again surpass the sample mean. What, then, accounts for this subnational variation of popular contention? Why do apparently similar provinces within the same national democratic regime have different levels of roadblocks?
Roadblocks in four provinces, 1999–2005.
Political Competitiveness and Peronism
Existing perspectives on the rise of protest activity make important contributions to the study of popular contention, yet they do not provide a complete explanation for the unevenness of subnational protest within a democratic regime. For instance, an analysis centered on the importance of institutions in countering collective action in the streets helps understand why in some countries discontent toward economic liberalization proceeds through the legislative arena, whereas in others this dissatisfaction explodes onto the streets. Based on this approach, institutions operate as a safety valve for citizen complaint and satisfaction, but this safety valve malfunctions when institutional structures become irresponsive to societal demands, making “street protest” the relief mechanism for built-up societal pressure. Countries with weaker political institutions thus breed a higher intensity to use protest as an alternative means to achieve political goals or express policy demands (Machado et al., 2009). This “hydraulic” understanding of protest, however, is theoretically incomplete given the potential static nature of institutions. For several Latin American countries stagnated by weak institutions (Kitschelt, Hawkins, Luna, Rosas, & Zechmeister, 2010), an institutional explanation is limited because institutions do not appear to be evolving much, especially in Argentina.
Another perspective on the rise of protest activity makes note of the “normalization” of contentious politics in Latin America, including Argentina (Moseley & Moreno, 2010). This approach suggests that protest politics has become a legitimate venue of political expression, rather than a threat to the democratic system as a whole. Supportive of this argument is the finding that citizens who participate in protest are equally, if not more, interested in participating in politics more generally. However, the characterization of protest as normal is not very convincing, particularly in the face of ample variation in patterns of mobilizations across subnational boundaries and across time. It is simply unclear at what levels protest becomes normalized, or ceases to be normal. All in all, the central idea presented by the existing literature may well be summarized by stating that protest is normal in countries with weak institutions.
To understand Argentina’s subnational variation of popular contention, considering the insufficiency of existing explanations, we rely on two other conjectures: first, the electoral incentives associated with competition and, second, partisan dynamics. The first conjecture revolves around the uneven pace of democratization throughout the interior provinces, which has resulted in shifting patterns of political competition and struggle; and the second conjecture relates to the comparative advantage of the country’s main political actor—the Peronist Party—in building informal but strong partisan linkages with popular sector organizations. This advantage has allowed Peronists to utilize various political strategies to interact with and compete against emerging contentious forces. We address these developments and their effects on mobilizations in turn.
First, since the country’s transition to democracy, a growing number of scholars have examined the presence of subnational electoral regimes. O’Donnell (1993), for instance, argued that the nation’s uneven pace of democratization established territorial enclaves that were seized by patronage networks riddled with nepotism, where political parties were weakly institutionalized and worked as meso-level machines, producing a “democracy of low-intensity citizenship.” Echoing O’Donnell (1993), Gibson (2005, p. 103) pointed out that Argentina’s transition has been shaded by an uneven “territorial distribution of the practices and institutions of democracy,” which in turn affected the “transparency of electoral procedures” in distinct jurisdictions. Remmer and Wibbels (2000) noted that the “structure of political competition at the provincial level varie[d] considerably” (p. 429), supporting the idea that “each province has its own political system” (Calvo & Escolar, 2005, p. 99). Gibson and Suarez-Cao (2010) point out that the level of political engagement across several provincial party systems embodies competition, fragmentation, and even the dominance of hegemonic provincial parties.
Several authors have also shown that the varying patterns of political competition produced by the uneven nature of subnational political regimes affect a number of outcomes. For instance, in their analysis of the country’s market transition, Remmer and Wibbels (2000) show that political competition facilitates economic adjustment by undercutting the incentives for the construction and maintenance of large public sectors. Gervasoni (2010) finds that oversized federal transfers to provincial regimes are associated with lower levels of political competition, thus diminishing the quality of subnational democracy. More generally, other studies have demonstrated that electoral competition leads to higher national and subnational government responsiveness (e.g., Griffin, 2006; Hobolt & Klemmensen, 2008).
Turning to mobilization, we argue that varying levels of political competition can also help us understand the spatial and temporal distribution of popular contention. Similar to Wilkinson (2004), we argue that party competition induces greater responsiveness among elected officials, so when local officials face significant electoral challenges, they are more likely to enact policies that can potentially dampen protest activity. 3 Therefore, provinces with evenly balanced political competition should be less prone to protest compared to uncompetitive provinces, given the former’s greater incentives to make government more responsive to constituents’ needs.
With respect to partisanship, the existing literature has shown that parties that are sympathetic to social movements often yield political opportunities for elected officials to shape partisan alignments and thus exercise discretionary control over protest (Goodwin, 2001; Van Dyke, 2003). These partisan linkages enhance exchanges between elected officials and mobilized groups, and thus affect organized actors’ expectations about the use of protest (Murillo & Ronconi, 2004). In the Argentine case, its provinces have been traditionally led by two major national political parties: the Radical Civic Union (Unión Cívica Radical, or UCR) and the Peronist Party, officially known as the Justicialist Party (Partido Justicialista, or PJ). 4 These parties are institutionally weak regarding their internal structure, yet they maintain deep roots in society. The Radicals are regarded as an electoral, professional party representing the middle class and with considerable presence in the country’s capital city, whereas the Peronists are seen as a mass political organization with informal yet strong historical alignments with popular sectors. The Peronist Party has often had a substantial comparative advantage over non-Peronists parties by establishing all-encompassing ties with subaltern and popular sectors such as labor unions (Levitsky, 2003), low-income constituencies (Calvo & Murillo, 2004) and, more recently, federations of the unemployed (Garay, 2007).
Recent studies have drawn attention to the harmonizing ways by which Peronists take advantage of their societal linkages and effectively interact with collective actors, depending on whether the PJ is in office or in opposition. On one hand, Peronist officials turn to contain and co-opt social movements when the party controls the executive power (governorships in our case). This perspective supports the idea that strong and informal links with popular organizations allowed the PJ to loosen the union-based ties that had historically sustained local Peronist organizations and, consequently, to demobilize its old labor militancy in the aftermath of Menem’s economic stabilization plan in the 1990s (Levitsky, 2003). In addition, after retaking the government in 2002, this partisan advantage helped the PJ’s provisional president Eduardo Duhalde to contain the types of large-scale social protests that eventually toppled the preceding Radical Party government of De la Rúa (Roberts, 2007). More recently, Garay (2007) has examined Peronist Kirchner’s efforts to incorporate the unemployed federations that surfaced in the late 1990s. Some of these federations received important state funding and subsequently modified their initial contentious tactics against the Kirchner government. As Garay (2007, p. 318) writes, Néstor Kirchner’s goal was to “end protest [and] mobilize votes” in support of his political faction. Hence, in provinces governed by strong Peronist officials, collective and social actors tended to get neutralized (Gibson, 2005), and political opposition to incumbents is equally discouraged (Gervasoni, 2010). As an illustration, provinces ruled by strong Peronist incumbents, such as Formosa, La Rioja, San Luis, Santa Cruz, and Santiago del Estero, have lower levels of protests in comparison to the sample mean.
On the other hand, when the PJ is no longer in office and becomes an opposition party, its privileged capacity to deal with popular sectors incentivizes Peronist Party officials and brokers to actively organize and encourage protest activity as an “alternative political technology” (Scartascini & Tommasi, 2010) to offset Radicals or non-Peronists officeholders. In this current cycle of protest, Peronist factions, politicians, and brokers have had a major presence in supporting street mobilizations and recruiting rioters in many provinces. During the 2001–2002 food riots, for instance, Auyero (2007) documents that Peronist “brokers from the party and top officials from the Peronist administration were . . . the masterminds behind the lootings” (p. 113) by circulating flyers and instructions throughout urban poor neighborhoods. Thus, when the Radical opposition party or a non-Peronist party coalition was in office, the unemployed piqueteros and other mass protestors were notably more inclined to demolish the non-Peronist side of the party system (Roberts, 2007). All in all, this partisan perspective on protest is consistent with the historical tradition of Peronism in carrying out strategies to “destabilize [any] government in power, be it military or civilian,” particularly more so when the PJ became a political challenger (Farinetti, 2010, p. 110). 5
To summarize, the uneven pattern of democratization throughout the Argentine interior provinces as well as the Peronist Party’s well-built linkages with popular sectors can help us understand the varying nature of popular contention in time and space. In the first place, different levels of political competition and struggle as a result of uneven subnational democratization boost responsiveness to constituents’ needs, thus discouraging protest activity. Therefore, we hypothesize that evenly balanced party competition should decrease the level of subnational protest. Second, if Peronists exercise their comparative advantage in neutralizing collective actors when they are in office, provinces ruled by Peronist governors should be less prone to mass mobilization. In contrast, when the Peronist Party is in the opposition, provinces ruled by the Radical Party or a non-Peronist coalition should be more protest prone. We thus hypothesize that the level of subnational protest is higher when Peronists are in opposition (and lower when they are in office).
Data and Measurement
The central task of our study is to examine the political factors that influence the current round of societal protest across Argentina’s 23 provinces and the city of Buenos Aires for the period 1993–2005. 6 This subnational comparative analysis corrects for the national bias of the existing literature when studying the rise of popular contention after Latin America’s third democratic wave. It also allows us to coalesce changing patterns of mobilization and nonmobilization within a single national unit of analysis, all of which helps us to formulate valid causal inferences on the factors that shape protest activity locally and over time.
The dependent variable PROTEST is the annual number of roadblocks compiled by Argentina’s Centro de Estudios Nueva Mayoría (CENM) and the Programa de Investigación sobre el Movimiento de la Sociedad Argentina (PIMSA), two well-known Argentine research institutes for the study of contentious events and social movements, both based on police reports and national newspapers. 7 To our knowledge, these are the only data currently available that measure the variation of protest activity across Argentina’s 23 provinces and the city of Buenos Aires, and across time in each province.
Our main explanatory variable of interest is the level of political competitiveness. Similar to Remmer and Wibbels (2000), COMPETITIVENESS is the percentage vote received in the first round of the election by provincial governors elected for the 1993–2005 period, measured in consistently negative terms as distance from the 50th percentile, where higher values indicate more competitiveness. 8 We theorize that evenly balanced party competition induces greater responsiveness among elected officials. Thus, when provincial governors face significant electoral challenges, they are more likely to enact policies that can potentially dampen protest activity (similar to Wilkinson, 2004). By the same logic, provincial governors elected with fragile pluralities or overwhelming majorities are expected to be equally unlikely to respond to constituents’ needs, thus propelling protest activity. The COMPETITIVENESS variable ranges from a high of −0.01% for the province of Entre Ríos in the early 1990s to −40.08% for San Luis during the early 2000s. Similar to Gervasoni (2010), we argue that the effective number of parties (ENP)—a commonly used indicator of competition—is invalid because uncompetitive party systems might produce higher and distorted scores than competitive ones. 9
Turning to partisanship, we hypothesize that when the Peronist Party is an opposition party, the level of subnational protest activity should be higher. PERONIST OPPOSITION is a dummy variable with a value of 1 for governors from the Radical Party or a provincial non-Peronist coalition (e.g., Cruzada Renovadora in Chaco, Partido Nuevo in Corrientes, or Fuerza Porteña in the city of Buenos Aires) and 0 for Peronist candidates. 10 As noted earlier, Peronists’ comparative advantage in establishing broad ties with popular sectors has led to a twofold strategy by which Peronists interact with contentious forces, depending on whether the PJ is in office or in the opposition. Therefore, this dummy variable addresses the duality of Peronism as it moves from government to opposition insofar as Peronists co-opt and demobilize collective actors when they hold office, yet encourage mass mobilizations when they are in opposition. 11
In addition to competitiveness and partisanship, we consider a number of alternative political explanations. These alternative explanations have been identified in prior research as important for understanding subnational politics in Argentina (e.g., Murillo & Ronconi, 2004; Remmer & Wibbels, 2000) and aim to take into account the strong Argentine tradition of linkages between parties and society, which again are said to affect organized actors’ expectations about the utility of protest. We hypothesize that shared political alignments across levels of government are more likely to lead to lower levels of protest activity. Shared partisanship across levels of government, as Murillo and Ronconi (2004) write, “increases the value of informal channels of communication, trust based on previous experiences, and credibility based on multiple interactions” (p. 80). Thus, the levels of protest activity should be lower in provinces with shared partisanship than in provinces without it. We take into account two possible variations of shared political alignments, one at the federal level and the other at the provincial level. PRESIDENTIAL ALIGNMENT is a dummy variable indicating presidential party control of provincial governorships, and GUBERNATORIAL ALIGNMENT is a dummy variable recording gubernatorial party control of provincial legislatures. 12
Our control variables are PER CAPITA GSP (LOG)t-1, UNEMPLOY-MENTt-1, and POPULATION (LOG) in each province. Higher levels of PER CAPITA GSP (LOG)t-1 should reduce socioeconomic grievances and thus decrease the level of societal protests. 13 The variable UNEMPLOYMENTt-1 is a second measure of socioeconomic grievances that should be positively correlated with the level of protests. These two variables are lagged 1 year to properly capture their realized economic effects. Finally, the variable POPULATION (LOG) seeks to control for the possibility that larger provinces would experience higher levels of mobilization compared to smaller provinces. 14 All of the variables in the analysis are measured annually for each province.
We include two protest controls to account for the potential spatial-temporal dependence of protest activity. Tarrow (1998), among others, argues that protests follow a cyclical pattern in which waves of protest spread rapidly across regions (provinces in our case), and then recede in the same manner. To control for the spatial dependence of protest activity, we include the variable LEVEL OF PROVINCIAL PROTEST, the mean level of protest throughout the sample in a given year. To control for the temporal dependence of protest activity, we include the lagged dependent variable PROTESTt-1.
We proceed by estimating an unconditional, fixed effects negative binomial event-count model. Event-count models use maximum likelihood estimation to assess the probability of event occurrences. As event counts always take on nonnegative integer values, the distribution of events is skewed and discrete, producing errors that are not normally distributed or homoscedastic (Long, 1997). In addition, overdispersion and goodness-of-fit tests indicated that a negative binomial model is the best method of estimation for our data. 15 We estimate an unconditional fixed effects version of the negative binomial model to account for unobserved (or unobservable) unit heterogeneity in the data. 16 Simply put, we need to control for all of the idiosyncratic factors that may make a particular province more or less protest prone. We accomplish this by including a set of province dummy variables in the regression model—one for each province in the sample, minus one. 17 Because we are concerned about the nonindependence of observations within provinces over time, we present the models below with robust standard errors clustered by province.
Empirical Results
Table 1 presents our empirical findings. Model 1 tests the effects of our main explanatory variable COMPETITIVENESS on the level of subnational protest. Model 2 tests the effects of partisanship based on the dummy variable PERONIST OPPOSITION. Models 3 and 4 add the alternative political explanations PRESIDENTIAL ALIGNMENT and GUBERNATORIAL ALIGNMENT.
Competitiveness and Subnational Protest in Argentina.
Robust standard errors in parentheses. All models are unconditional fixed effects negative binomial regressions.
p < .05. ***p < .01.
As previously theorized, political competition improves responsiveness to constituents’ needs, thus thwarting protest activity. Political competition in the Argentine provinces also varies considerably because of the uneven pattern of democratic practices. Our results demonstrate a strong statistical regularity between the variable COMPETITIVENESS and the level of subnational PROTEST (Models 1 thru 4), suggesting that the more equally votes are distributed between two major parties, the lower the levels of subnational protest activity.
Turning to partisanship, PERONIST OPPOSITION also covaries statistically with the level of protest (Models 2 thru 4). When the Peronist Party is an opposition party and the Radicals or a non-Peronist party coalition are in control of governorships, the level of subnational protests is higher. These results provide support for the double-sided effect of Peronism given its comparative advantage to simultaneously contain and co-opt social groups when the party holds office as well as incentivize mobilizations when the party is in the opposition. In contrast, the coefficients for the alternative political explanations PRESIDENTIAL ALIGNMENT and GUBERNATORIAL ALIGNMENT emerge with the expected sign, but are not statistically significant at the conventional levels (Models 3 and 4).
To give more substantive interpretations around the quantities of interest, we now provide the combined effect of COMPETITIVENESS and PERONIST OPPOSITION using a statistical simulation. Based on King, Tomz, and Wittenberg (2000), Table 2 presents the predicted event counts of PROTEST with statistical uncertainty using Model 2. The predicted event counts of PROTEST are computed for three different levels of COMPETITIVENESS (high, mean, and low) and with and without PERONIST OPPOSITION. The levels of COMPETITIVENESS are measured around the minimum, mean, and maximum values in the sample. All the estimates are statistically significant at the conventional level. First, the predicted event counts of PROTEST for any level of COMPETITIVENESS (high, mean, and low) clearly demonstrate the strong statistical regularity that there is a higher level of roadblocks with PERONIST OPPOSITION than without it. The predicted number of roadblocks, in fact, almost doubles when the Radical Party or a non-Peronist are in control of provincial governorships. Moreover, the predicted number of PROTEST increases from 6.260 at the high level of COMPETITIVENESS to 26.053 at the low level of COMPETITIVENESS with PERONIST OPPOSITION, and from 3.371 at the high level of COMPETITIVENESS to 14.327 at the low level of COMPETITIVENESS without PERONIST OPPOSITION.
Estimated Effects of Competitiveness and Opposition on Subnational Protest.
Entries are the estimated event counts of PROTEST, calculated using Clarify 2.1. The level of competitiveness is measured around its minimum (−40.08) mean (−7.21), and maximum (−0.01) values in the sample. The minimum value indicates low competitiveness; the maximum value indicates high competitiveness. Standard errors are in parentheses.
Our control variables—UNEMPLOYMENTt-1, POPULATION (LOG), and the LEVEL OF PROVINCIAL PROTEST—also have a positive and statistically significant effect on subnational protests. Larger provinces and those provinces with a higher share of unemployment experience higher levels of roadblocks. These findings are consistent with our expectations and corroborate the salience of economic conditions in fomenting collective resistance (Silva, 2009). Setting the other variables to their mean values, the predicted annual number of provincial protests rises from 3.24 at the lowest share of unemployment (1.2%) to 9.72 at the highest share of unemployment (21.8%). In contrast, the control variable PER CAPITA GSP (LOG)t-1 is not statistically significant at the conventional level in any of the models.
Overall, our results reveal that the electoral incentives created by shifting patterns of political competition and the nature of partisan opposition shape the subnational unevenness of popular contention. The results help us understand why apparently similar provinces within the same national democratic regime have different levels of conflicts. For instance, the provinces of Córdoba and Santa Fe are very similar with respect to population size (approximately 8% of the national population), yet across the sample period their mean levels of COMPETITIVENESS are −1.67 and −4.34, respectively. Being more politically competitive, the province of Córdoba experiences fewer protests (23.69) than Santa Fe (29.15). Similarly, the provinces of Chubut and Río Negro have populations of approximately half a million people each, but their mean levels of COMPETITIVENESS are different: −4.51 and −6.73, respectively. The more politically competitive province of Chubut experiences far fewer protests (4.69) than Río Negro (12.23).
Returning to Figure 2, the provinces of Entre Ríos, Corrientes, Tucumán, and Salta are also very similar with regard to geographic location and population size (each approximately with 3% of the national population). However, their mean level of COMPETITIVENESS across the sample period varies. For instance, Entre Ríos (−1.89) and to a lesser degree Corrientes (−4.42) are more politically competitive compared to Salta (−5.30) and Tucumán (−6.02). Thus, Entre Ríos and Corrientes experience far fewer protests compared to Salta and Tucumán, which have higher levels of protest in comparison to the sample mean. As we have argued, when provincial politicians face significant electoral challenges, they are more likely to enhance government responsiveness and enact policies that can potentially reduce protest activity. It goes without saying that compared to the provinces of Salta or Tucumán, Entre Ríos and Corrientes are also known for their better political and socioeconomic performance across a wide range of areas, including lower levels of electoral volatility and party fragmentation; more independent judicial courts; cleaner public budgets; better fiscal performance, health, and housing policies; and lower income inequality, infant mortality, and crime rates. 18
Sensitivity Analysis
Beyond the results presented above, we have conducted extensive sensitivity analysis to ensure the stability of our main findings. First, concerning our main explanatory variable of interest, COMPETITIVENESS, we estimated models using alternative measures of electoral competition, such as Gervasoni’s (2010) measure of EXECUTIVE CONTESTATION (their correlation was .55 in our sample). This indicator “measures the extent to which there are real chances for the opposition to defeat the governor’s party” (Gervasoni, 2010, p. 316). It is calculated as one minus the proportion of the valid vote won by the incumbent party or coalition in the first round of the elections for governor, whereby the higher the incumbent’s share of the vote, the lower the level of competition. As shown in Table 3, our main findings remain qualitatively unchanged.
Subnational Protest in Argentina: Alternative Measure of COMPETITIVENESS.
Robust standard errors in parentheses. All models are unconditional fixed effects negative binomial regressions.
p < .1. **p < .05. ***p < .01.
Second, although our theory posits (and our results confirm) that COMPETITIVENESS and PERONIST OPPOSITION each have an independent effect on subnational protest (their correlation was .2 in our sample), we considered whether these two factors may have an interactive effect as well. Specifically, we estimated the effects of political competition on the level of subnational protest conditional on opposition control of provincial governorships. The interaction term COMPETITIVENESS × PERONIST OPPOSITION captures this conditional effect of political competition, and opposition control of subnational government is considered as an intervening variable that affects the relationship between political competition and subnational protests. The interaction term was not statistically significant. Moreover, we calculated the conditional standard errors to see if the marginal effect of the interaction term was significant. These estimates support the results in Tables 1 and 3. In brief, the results show that when Peronists are in control of subnational governments, they respond to increased political competition by reducing protest. Conversely, Peronists’ use of protest is not conditioned by political competition when they are out of government. Overall, these results (not shown) reconfirm the duality of Peronism as it moves from government to opposition insofar as Peronists co-opt and demobilize collective actors when they hold office, yet encourage mass mobilizations when they turn to opposition.
Third, concerning alternative explanations, we have estimated models taking into account other institutional explanations, such as ELECTORAL VOLATILITY, which is a common measure of party system institutionalization. We used the Pedersen (1979) index, which provides a measure of the net aggregate vote shifts from one election to the next (legislative elections in our case). We have also estimated models controlling for the effects of international moments, such as the years in which negotiations between Argentina and the International Monetary Fund (IMF) were taking place. We created a dummy variable SIGNED IMF ARRANGEMENT coded as 1 for every year in which the country signed an agreement with the IMF (1996, 1998, 2000, 2001, and 2003 in our case). These variables were not statistically significant at the conventional levels, and these models (not shown) produced results nearly identical to those presented in Tables 1 and 3.
Finally, concerning estimation techniques, our results hold when we estimate random effects negative binomial models as opposed to the fixed effects models presented throughout the text, when we scale the standard errors by the Pearson or the deviance dispersion statistic, and when we use robust standard errors not clustered by province. Our main results were also robust to the inclusion of time fixed effects (year dummies).
Conclusion
To date, much of the existing literature focusing on the rise of popular contention in contemporary Latin America has emphasized the national-level conditions and institutions that associate democracy with protest. Although this literature has made important contributions to the understanding of protest activity amid widespread democratization, very few studies have systematically explored the subnational patterns of protest in time and space. By taking into account the subnational unevenness of popular mobilization, our comparative study has sketched a “more general theory of the politics of provincial protests” in Argentina (Murillo & Ronconi, 2004, p. 89). Insofar as the incidence of protest in the country is very high compared to the rest of the region, the theoretical significance of the factors that affect protest in Argentina has broad implications.
Substantively, our article has shown that both political competition and partisanship affect the level of roadblocks in a manner that is consistent with numerous studies on Argentine politics (e.g., Auyero, 2007) and beyond (e.g., Wilkinson, 2004). Our results indicate that the Peronist Party holds a comparative advantage with regard to both the mobilization and demobilization of popular subjects, even after the radical reconfiguration of its traditional labor linkages following the implementation of economic liberalization policies during the early 1990s (Levitsky, 2003). Our findings also reveal that evenly balanced party competition is associated with lower levels of protest, supporting the idea that politicians are likely to make subnational governments more responsive to voters when they face vigorous party competition. In a period where political actors, in particular, and representative institutions, more generally, have been somewhat discredited in the public view, our results demonstrate that the different levels of party competition in which politicians operate influence their responses to rising subnational protests.
Methodologically, our subnational comparative analysis allows us to pinpoint how different levels of political struggle and bargaining as well as the nature of partisan opposition shape the spatial and temporal distribution of popular contention. It also allows us to control for a number of socioeconomic conditions. All of this enhances our ability to formulate valid causal inferences regarding the effects of political competition and partisanship. The contribution of these explanations to the larger literature on contentious politics cannot easily be teased out using the aggregate, macro-level indicators of the existing literature on the consequences of economic liberalization.
Existing research outside established advanced industrial democracies has yet to adequately explain the relationship between protest activity and democratization. The bulk of existing research, in fact, frames popular mobilizations as critical to the breakdown of authoritarian regimes, and assumes that once countries embrace democracy, parliamentary institutions and relevant parties take center stage in resolving societal conflicts. However, countries such as Argentina, Bolivia, and Peru, among others, which currently maintain high levels of “street politics,” suggest that this expectation involving routinized parliamentary behavior has fallen short. Our article also highlights the overlap between parliamentary and extraparliamentary politics as parties rely on protest to affect their electoral chances. These strategies cannot be neatly separated and should not be examined independently from each other since they often go hand in hand. In fact, as our comparative analysis of Argentina suggests, parties such as the Peronist with deeper roots in society remain critical actors in resolving these conflicts, even when Peronists themselves utilize mobilizations as an alternative source of political empowerment.
Footnotes
Acknowledgements
The authors thank Kenneth Roberts, Gustavo A. Flores-Macías, Carlos Gervasoni, Paul Bellinger, Gabriel Ondetti, and the anonymous reviewers for their valuable comments.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Moisés Arce’s research was supported financially by a research board grant from the University of Missouri.