On February 21st, Bolivians go to the polls to vote on a referendum to allow President Evo Morales to run for yet another term, his fourth, in 2019. First elected in December 2005, Morales was re-elected to the presidency in 2009 and again in 2014. Like many countries in the region, Bolivia removed the constitutional prohibition against immediate re-election with the adoption of a new constitution in 2009. Although the new constitution only allows presidents to serve two consecutive terms, Morales successfully argued that his 2009 election was the first under the new constitution, making him eligible to run for re-election in 2014. He won that 2014 election handily, with 61.4 percent of the popular vote.
There’s no question that Morales remains generally popular in Bolivia, despite deep distrust among many sectors that identify with the political opposition. Although closely identified with Hugo Chávez and Rafael Correa as part of the “radical left” in Latin America, Morales’ government has been less polarizing and more fiscally disciplined than that of Venezuela’s Hugo Chávez and his successor Nicolas Maduro. Unlike Venezuela, Bolivia retains a relatively healthy economy and has made significant (and hopefully lasting) improvements in reducing poverty and improving health and education indicators during Morales’ presidency.
Still, whether the referendum to amend the constitution allowing Morales to run for re-election in 2019 will pass remains uncertain. Polling data—which is frequently sketchy and never very clear in Bolivia—suggests a close vote. Much will depend, of course, on voter turnout and whether the opposition can convince a significant number of Morales supporters to reject re-election on principle, rather than being seen as a referendum on Morales. It’s tempting to make the vote a referendum on Morales, particularly since his government is in the midst of several corruption scandals. But Morales remains personally popular—certainly more popular than his own party.
Beyond the question of whether re-election is problematic for presidential democracies, however, the reality of Morales’s central and hegemonic position within his party (Movimiento as Socialismo or MAS) reveals some important dilemmas and paradoxes. One dilemma is the question of future leadership, and whether there is any plan or mechanism for an eventual successor to Morales. Bolivia has a long history of political parties dominated by individual caudillos, and a weak track record of such parties surviving after them. This also makes it difficult to determine whether voters are voting for the MAS political platform or its charismatic leader. The latter is more probable, and would explain why MAS candidates do less well in regional and municipal elections, when Morales isn’t on the ballot.
But, after more than a decade in power, the electoral trajectory of MAS and Morales is starting to reveal an interesting disjuncture. Looking closely at municipal-level data for the 2009 and 2014 presidential elections and the 2010 and 2015 mayoral elections, alongside census data, reveals some counter-intuitive patterns. We should be cautious, of course, about reading too much into findings from municipal-level data (this is known as the “ecological inference” fallacy—extrapolating from aggregate data to explain individual-level behavior), but that doesn’t mean that the story the data tells can be completely disregarded.
Because I wanted to look at elections between 2009 and 2015, I used the most recent census data of 2012 to build profiles of each of the country’s 339 municipalities, then merged this with municipal-level election data. My goal was to test two different propositions: first, whether the municipal-level vote share for MAS reflected differences in ethnic composition in those municipalities, and second, whether the municipal-level vote share of the MAS reflected socioeconomic differences across municipalities. Conventional wisdom is that Evo Morales and his MAS party receive most of their support from Bolivia’s indigenous majority and from the poor, and comparing census data municipality by municipality with local election returns would allow me to test this.
I calculated both the level of ethnic diversity and socioeconomic differences in all the municipalities based on the 2012 census. (For a more detailed description of the methods I used to do this please visit my methodology appendix.) I then used that data to estimate the electoral support for MAS in the 2009 and 2014 presidential elections, as well as the 2010 and 2015 municipal elections. Not surprisingly, the MAS vote in municipal elections tracked closely with that municipality’s MAS vote in the presidential election, so using it as a control variable helped distinguish between local support for national MAS leadership.
The differences were between the pairs of elections (2009 and 2010 and the 2014 and 2015), and between presidential and municipal elections. In both the 2009 and 2014 presidential elections, municipal-level vote share for MAS decreased as the non-indigenous population share of the population increased. This is consistent with the view that MAS attracts support mainly from indigenous voters.
However, a closer look at the two elections suggest a difference between Aymara and Quechua communities. In 2009, although MAS municipal-level vote share increased as the share of Aymara population increased, there was no similar difference for the Quechua population. That reversed in 2014, with municipal-level MAS vote share increasing as share of Quechua population increased, but with no significant difference for the Aymara population. In other words, in 2014 municipalities with high Aymara populations were not more or less likely to support MAS (when controlling for other factors) than the general population.
I also found that while occupational concentration matters (the more concentrated the population is in a single occupation, the higher the vote share for MAS), wealth differences hardly mattered—and only then when ethnicity variables were excluded. This suggests that ethnicity, not wealth, is driving support for MAS. But even there, we see a slight paradox: municipal-level support for MAS decreased as municipalities got wealthier, but support for MAS increased in municipalities that saw greater improvements in wealth measures (although to a smaller, and less significant extent). In other words, when ethnicity is taken out of the mix, MAS does better in municipalities that are poorer. Yet as poverty levels fall, MAS receives some benefit, but perhaps not enough to offset loss in support as wealth increases. Taken as a whole, however, it looks as if the socioeconomic improvements championed by MAS may be slowly eroding the socioeconomic divide that has both fueled the MAS movement and divided the country for a decade.
Looking at municipal elections, the data show a rift between Aymara and Quechua communities. When controlling for other factors, in both elections the effect of share of Aymara or Quechua communities differed starkly: municipal-level support for MAS decreased as Aymara population increased, and increased as Quechua population decreased. Another significant difference is that, in municipal elections, support for MAS also increases as the share of non-indigenous population goes up—but only in models that do not control for Aymara population. This suggests that MAS is making gains among non-indigenous population, but simultaneously losing support in heavily Aymara regions.
Again, one should be skeptical about reading too much into municipal-level analysis. But these data suggest an interesting problem on the horizon for Morales and his MAS party. The evidence suggests a significant drop in support among Aymara communities in the span between the 2009/2010 and 2014/2015 elections. The drop is particularly surprising given the government’s emphasis on Aymara symbolism (for example, with the elaborate presidential inaugural ceremonies held at Tiahuanaco). These results also show that there may not be link between support for MAS and inequality, indicating that the MAS’s class-based political agenda and rhetoric may produce diminishing returns, especially as socioeconomic differences become less salient. At the same time, in recent years, Evo Morales and his MAS party have expanded their support within the middle classes, offsetting declining support among the indigenous and poor.
Evo Morales and his MAS party have been in power for more than a decade. If he runs for re-election in 2019, Morales will be campaigning in a very different electorate than the one that first brought him to power in 2005. By 2019, a significant portion of the electorate will have been born after the October 2003 “gas war” that brought him to national fame and eventually power, and will have no living memory of any president other than Evo Morales. For an ever-increasing share of the electorate, Morales represents not an “outsider” to the political system, but the central figure of the political establishment. And increasingly, MAS has become a nationalist, catch-all party, not too dissimilar from the catch-all parties of the 1980s and 1990s. In short, perhaps MAS is becoming the contemporary incarnation of the Nationalist Revolutionary Movement (MNR), but in reverse. Whereas the MNR emerged in the 1950s as a radical middle-class populist party that, during its time in government, expanded its reach to the rural indigenous peasantry, MAS has emerged as a rural-based, indigenous-oriented political movement that, during its time in government, has expanded its support within the urban middle-class sectors.
In the end, the looming political and democratic issue for Bolivia goes deeper than whether Morales can run again in 2019, but rather: Can MAS develop an institutional foundation that can survive beyond Morales? And perhaps more important: Can Bolivia develop a multiparty system that reflects real social cleavages that transcend ethnic differences? Currently, the socioeconomic cleavages are blurring as MAS gains support across social classes. The lack of a clear socioeconomic divide within the partidocracia system of the 1990s was itself a major factor in the political and social crisis that led to the October 2003 “gas war.”
Similar shifts are occurring in MAS’s support among the indigenous. Bolivia does not have a single, monolithic indigenous majority. Rather, it has two major ethnic groups and numerous other smaller ones. As growing differences between the Aymara and Quechua voters described above show, ethnic political mobilization is becoming more complex.
All of this leaves a series of unanswered questions for the future of MAS as a political project and the future of Bolivian democracy, well beyond the constitutional referendum.
Miguel Centellas teaches political sociology at the Croft Institute for International Studies at the University of Mississippi. He is also co-director of a summer methods field school in Bolivia in partnership between the University of Mississippi and the Universidad Católica de Bolivia.
