THREAD: @OAS_official released its final audit of the Bolivian elections. In the 35 days since it released a preliminary report: Evo Morales was ousted in a coup; an opposition Senator swore herself in as president; 30+ have been killed, most by state security forces.
@OAS_official The Oct 20 results were annulled. The congress is working w/ the de facto gov’t to hold new elections. But can the OAS be an impartial election observer going fwd? All of the OAS’s Bolivia elections work has had serious problems. The final audit is no exception.
@OAS_official From day one, OAS claims have fueled an overarching fraud narrative. On Oct 20, the TSE (electoral authority) announced results of a non-binding preliminary vote count (TREP) w/ 84% of the vote counted. Morales led his nearest competitor, Carlos Mesa, by 7.9 percentage points.
@OAS_official The TREP was then suspended. The OAS expressed concern; the opposition cried foul. The following day, the TSE released an update of the TREP with 95% of the vote counted. Morales led Mesa by just over ten percentage points — a large enough margin to win in the first round.
@OAS_official The OAS issued a release expressing concern at the “inexplicable” change in the trend of the non-binding, preliminary results. They’ve repeated that ever since, including in the final audit. But was it inexplicable? No.
@OAS_official In a @ceprdc report, we found that when the TREP was suspended w/ 84% of votes processed, the votes that remained to be counted came from geographic areas that had expressed a clear trend in favor of Morales. cepr.net/images/stories…
@OAS_official @ceprdc More than 100 economists and statisticians agreed. Look at the example they provide — the gubernatorial race in Louisiana. Think of those CNN election maps where they zoom in on precincts and project how the remaining votes will break… 
@OAS_official @ceprdc So, what did the OAS find in their final audit? For starters, they identified significant security vulnerabilities, a lack of proper chain of custody with electoral materials. They found a mysterious third server that could apparently access the TREP (preliminary) results.
@OAS_official @ceprdc This raises many questions — most importantly, were these vulnerabilities taken advantage of in order to intentionally manipulate the results of the election? The OAS, in its press release, claims the answer is yes.
@OAS_official @ceprdc In order to target the audit, the OAS selected a sample of 4,692 tally sheets (aggregation of votes from each voting table across the country — more than 34,000 in all). Was it a random sample? No. These were the criteria used:
@OAS_official @ceprdc The OAS analyzed all those tally sheets looking for evidence of manipulation. They performed handwriting analysis. They identified 226 “irregular” tally sheets from 86 voting centers in 47 municipalities. This contains 38,001 valid votes — 91% went to Morales (34,718).
@OAS_official @ceprdc How does this fit with the narrative of fraud? Were these tally sheets plausibly connected to the interruption of the quick count? Do the vote totals on the 226 sheets look drastically different than those from nearby voting centers? The simple answer: no.
@OAS_official @ceprdc An example may help: the OAS flags 13 tally sheets in Sacaca, a small, indigenous-speaking municipality in Potosí. 10 were in the TREP pre-interruption. Are these tally sheets notably different from others in Sacaca? No, the entire area voted overwhelmingly for Morales.
@OAS_official @ceprdc The 13 tally sheets come from 5 voting centers. Is it surprising that, in a small center with only 2 or 3 tables, 1 person filled out the tally sheets? Is that evidence of a coordinated fraud? Or evidence of the type of irregularities we see in elections across the hemisphere?
@OAS_official @ceprdc This is just one example. But from the rest of the 226 tally sheets, two-thirds were counted *before* the TREP was interrupted. The flagged sheets are also consistent with those from the unflagged corresponding precincts, before and after the interruption.
@OAS_official @ceprdc In other words, the “fraudulent” tally sheets don’t appear abnormal in terms of the actual result. If the OAS is holding them up as evidence of intentional manipulation, then wouldn’t you expect them to look different than those from the same location that the OAS didn't flag?
@OAS_official @ceprdc It is important to point out that this appears to be the best evidence the OAS has that systemic vulnerabilities were exploited. The OAS lists evidence of “Deliberate Actions that Sought to Manipulate the Results of the Election.” These are the only tally sheets mentioned.
@OAS_official @ceprdc The implication of this is made clear later in the report, when it notes that the 34,789 Morales votes on these 226 tally sheets represent “almost” the entire margin above the ten percentage point threshold for winning the election in the first round.
@OAS_official @ceprdc Removing all these votes doesn’t actually reduce Morales’ margin below 10 pp. However, the OAS says that they’d find more “fraud” if they examined more votes, since the sample was only 13%. But again, it wasn’t a random sample (they already analyzed the most suspicious sheets).
@OAS_official @ceprdc Would they find additional cases of “fraud” in areas where Mesa did well? How might that impact the result? They don’t ask these questions. The OAS focuses only on the areas that most heavily favored MAS. An impartial audit wouldn’t look in only one direction.
@OAS_official @ceprdc And what about that “inexplicable” change-in-trend? In the final audit report, the OAS doubles down, and takes a shot at me and my colleagues at @ceprdc:
@OAS_official @ceprdc Ouch. But it’s not true. We address the “marked change” by showing it was predictable and explainable with a precinct-level analysis at the presidential and legislative levels. Based on this final audit report, it’s clear that it’s not @ceprdc trying to justify a certain result.
@OAS_official @ceprdc Again, the OAS claimed from day one that this entirely explainable phenomenon was inexplicable. @Almagro_OEA2015 went before the permanent council and claimed widespread fraud and that the real coup was Evo’s fraud.
@OAS_official @ceprdc @Almagro_OEA2015 This is not a comprehensive response to the OAS report (95 pages w/ hundreds more in annexes). This is not about the candidates, nor favoring an outcome. It is about the role of the OAS in the current situation, and how that impacts its credibility to serve as an honest observer.
@OAS_official @ceprdc @Almagro_OEA2015 It is about accountability for an organization that, under the leadership of @Almagro_OEA2015, and with the full backing of the Trump administration, has abandoned any pretense of neutrality. That matters in Bolivia, and it matters across the Hemisphere.
@OAS_official Edit: 25 days.
