Still no final numbers on Bolivia’s recall referendum, but results are already much clearer. W/ a near-final vote count, Oruro’s prefect is safe w/ just over 50% of the vote. That means that only 2 of 8 prefects up for recall lost their seats: Manfred Reyes Villa in Cochabamba & José Luis (“Pepelucho”) Paredes in La Paz. At least now Reyes Villa is also acknowledging his defeat.
As more numbers come in, it’s also clear that Evo is winning about two thirds of the vote (the current tally is 66.97%, a significant jump from 53.74% in December 2005). That certainly strengthens his position & gives him a mandate. But the Media Luna prefects have also strengthened their positions & received their own mandates. Take Santa Cruz: Ruben Costas won only 47.88% in 2005 when he was elected prefect; w/ most of the vote counted, he looks to have won more than two thirds of the vote (his current tally is 68.01%). The story is similar in Tarija: Mario Cossio won only 45.65% in 2005; w/ that department’s vote count final, he increased his vote to 58.06%.
But the results could be read w/ a (slight) grain of salt. W/ most of the vote counted, Pepelucho lost in La Paz w/ little more than a third of the vote (the current tally is 36.09%). But he looks to have received nearly as many votes as in 2005, when he won the prefecture w/ only 37.99%. The only prefect to have seen a real decline in support was Reyes Villa, who received only 35.19% of the vote, compared to 47.64% in 2005.
It’s possible the 2005 prefect victories in La Paz & Cochabamba were relative “flukes.” After all, these were multi-candidate contests decided by simple plurality. In Cochabamba, Reyes Villa (a former city of Cochabamba mayor who still controlled the urban political machine) ran against there challengers, including one from MIP (a predominantly Aymara-based indigenous party). In La Paz, Pepelucho faced seven challengers, including MIP (which has historically done well in La Paz) & FREPAB (whose candidate, former police major David Vargas, had led the La Paz police mutiny in February 2003). Had MAS been able to coordinate w/ these two parties, it could have won in La Paz (and perhaps also in Cochabamba).
From the start, Pepelucho was in a precarious position. Reyes Villa’s defeat is less expected (the vote was expected to be closer), but he had made a number of serious political blunders in the last two years & was increasingly unpopular.
The current scramble for candidates in La Paz suggests that MAS may again have difficulty coordinating a victory for its preferred candidate, though it’s unclear if any viable non-MAS candidate could win. It’s virtually impossible for an opposition candidate to win La Paz (as a relatively popular El Alto mayor, Pepelucho was perhaps the only “establishment” politician who could’ve won). But divisions w/in the MAS “big tent” (particularly between moderates & radicals) may surface in any upcoming multi-candidate contest.
Of course, this is only the first such election in the country’s democratic history. But results do suggest that removing incumbents is a tall order. The Reyes Villa case suggests that only a very unpopular & controversial incumbent is likely to lose votes relative to a previous election. The Pepelucho case suggests that an incumbent elected w/ less than 40% of the vote is incredibly vulnerable.
The final result is that everyone (except Pepelucho & Reyes Villa) are vindicated. So while Evo has a mandate, so do the Media Luna prefects. And though nearly two thirds of Bolivians support Evo’s mandate to govern, these are disproportionately concentrated in the Andean regions. More than anything else, the election highlighted the country’s growing regional polarization. Still, this kid of electoral solution is much preferable to using number in the streets.
But the polarization is within the eastern departments also, rural-urban. It's more complex than east vs. west. Better to characterize the divide as eastern cities vs. the rest of the country. Maybe the Media Luna does not exist.
Still, this kid of electoral solution is much preferable to using number in the streets.
Unfortunately, the former will not replace the latter.
Good analysis and round up. Thanks.
@John: Good point. And, yes, there is also a rural/urban split. But the rural/urban split is more pronounced in the east than in the west. I've not run the numbers for the most recent election, but in a paper (which I hope is published soon) statistical analysis of disaggregated data showed that east/west matters even when controlling for rural/urban.
Yes, there was definitely an east/west difference in the Cuellar results, but it was less what I would call polarization. I will not be surprised to see differences in Tarija and Santa Cruz too, in the recall referendums. But that's my point really, that Tarija and Santa Cruz are not going to be more easily governable, in the long run, than Bolivia.
I look forward to your paper.
A much older version of the paper (not including December 2005 data) was presented at the April 2005 MPSA conference. You can read that here:
http://users.dickinson.edu/~centellm/papers/mpsa2005.pdf
The newer version is under review. Though the findings aren't much different when adding the 2005 data (actually, they're stronger), they do include different model specifications (as well as more models).
OK. I'm not trying to say that there are no differences between east and west rural voting, The differences (looking at MABB's table from La Razon that breaks down Evo's support by rural vs. urban in each department, pretty early results) are pretty big, ranging from mid 50s up to 95%. That's actually a real big difference in my book. But Evo got majorities in all 9 rural populations, so I would not call that polarization. Since there's only one pole.
Table at http://mabb.blogspot.com/2008/08/morales-wins-and-so-do-prefects.html
What I'm really hoping you'll do (just joking, mostly) is look at rural voting by elevation or some better measure of altiplano-ness vs. media luna-ness.
Do Santa Cruz (58%-37% for Evo, rural vs. urban, not final results) and Tarija (56%-36%) show rural vs. urban polarization? Well, in any case, these are big differences and should make Costas and Cossio uncomfortable if they take the hard line and then go out to the campo (kind of like Evo flying into El Trompillo or Tarija). Of course, I'm guessing that there are east-west differences within rural Santa Cruz and Tarija too. Maybe they should drive east.
Another way that I don't think this is polarization: I see differences in the Media Luna "pole" between Santa Cruz and Tarija. (Probably with Beni and Pando and anti-Evo Chuqisaca too.) Differences in details of autonomy, in political speech, in dependence on gas & oil and so willingness for negotiation on IDH, and on resistance to land reform. For example, Wednesday's El Pais (http://www.elpaisonline.com/) has several statements that Tarija does it differently than Santa Cruz. These opinions are not uncommon.
I was doing a similar comparison to 2005 and obviously came up with simliar conclusions. Pepelucho actually got 1,780 (at last count)more "Yes" votes in the referendum than he got votes in 2005. There was a 7.26% increase in valid votes this time around though. As to morales, he's currently showing a 14.03% increase placing him a solid 4th in improvement behind Potosí (up 38.08%), Beni (up 19.63%), and Santa Cruz (up 19.22% with more than 28,000 fewer valid votes!). Tarija, Oruro and Pando showed 12.41%, 9.91%, and 8.18% improvement respectively.
This doesn't completely discount morales advances, but as I said on brand X, 2005 was a multiple choice test and this was true-false. Of course the candidates showed a percent increase. It appears that the stalemate will continue. It does give morales a chance to appear presidential and extend an olive branch(which he has).
mcentellas: what do you think of the charges of vote fraud: specifically the charges that multiple voting cards were printed up in some mas regions and that this fraud swelled the rolls by over a million voters?
On a separate thought, do you think that the polarization could have been exaggerated by the fact that this was a recall election featuring Morales and autonomists for the most part.
While a voter may have preferred a different politician, he may not feel he has done so a bad job that his term should be ended early. Thus most incumbants get a "mandate".
@John: The east/west polarization isn't strictly directional, but regional. Better to use Andean/lowland than east/west. While there is a powerful rural/urban split in the electorate (as my statistical analysis showed), it wasn't as strong or as significant as the Andean/lowland split. In large part, perhaps, because El Alto is very urban & tended to vote anti-system (and currently, pro-Evo) in higher rates than the countryside. And Evo may be only one pole, but polarization often tends to be one dimension (you either are for or against X).
@DV: Accusations of vote fraud are actually not uncommon in Bolivia. Even during the 1980s & '90s, the systemic parties accused each other of manipulating the vote. But I see no reason to discount this vote, even if a few irregularities might have occurred. Also, the issue of multiple voting cards is a bit overblown. Voters have to dip their finger in permanent ink before casting their ballot, and polls are only open a few hours. Additionally, despite the swelling in voter rolls (something that's bound to happen in a country w/ a low median age & where rural voter registration becomes highly promoted), the total number of voters registered was on par w/ the numbers for previous years.
Interesting slicing of recall results by Inca Kola. (Just look at tables if you don't like analysis.) http://incakolanews.blogspot.com/2008/08/bolivia-recall-referendum-final-numbers.html
especially
http://4.bp.blogspot.com/_7Se7iswAanA/SKdH1JOOA_I/AAAAAAAACIQ/bg-s6gm5XWI/s1600-h/bolrecall2.gif
and
http://2.bp.blogspot.com/_7Se7iswAanA/SKdcss06A6I/AAAAAAAACJI/rX3Jh4Qvij8/s1600-h/bolrecall9.gif
(I'm a different crazy John from the commenter there.)
Thanks, John. Yes, overall Incakola does provide an interesting (and solid) analysis. Although I'm not entirely sure the evidence supports the harsh conclusion (which seems based on a straw man argument).
Apparently, Inkacola simply deleted my brief comment, then called me a racist. I'm not sure what (if anything) I wrote could've been construed that way, but since he deleted the comment, you have to take my word for it.
My comment (which also praised his overall solid analysis) merely raised two brief criticisms:
1) It seemed odd to argue that Santa Cruz votes didn't matter because they were fewer than the national total (that is, aggregate numbers matter more), then argue that inside Santa Cruz the rural/urban distinction meant that the city didn't matter (because geographic distribution, not aggregate numbers matter more).
2) It also wasn't clear that his conclusion followed from his evidence. He suggests that Santa Cruz autonomy is entirely driven by a small racist elite, but his own evidence suggests that a considerable number of voters supported *both* Evo & Costas, meaning that there is something else going on. In other words, some may want Evo's national policies *and* autonomy. I think the conclusion that based on voting results there is no "ostensible" basis for a case for autonomy is stronger than the evidence supporting it. I also don't see how an otherwise solid analysis of voting data leads to conclusions about the nature of autonomy elites (such an argument can be made, of course, but it would require additional evidence).
I think he stopped short of calling you a racist, but he obviously won't make your Blogroll.
Cossío talks of negotiation and peace. Here's hoping. Could happen.