Skip to content

Bolivia election fraud fraud update

Last November there was a disputed presidential election in Bolivia. The Organization of American States wrote, “Given all the irregularities observed, it is impossible to guarantee the integrity of the data and certify the accuracy of the result. . . . In all likelihood, given more time to process documentation, even more irregularities would surface.”

But an outside team looked at the data and didn’t find evidence of fraud; rather, they found that the OAS report had problems.

I took a look and, without trying to judge the integrity of the election as a whole (I know nothing about Bolivia other than seeing this movie many years ago), I agreed that the OAS report was flawed: as I put it, one of their analyses was “a joke, maybe suitable for publication in the Proceedings of the National Academy of Sciences but I wouldn’t expect to see it any serious report.”

Since then, more analyses have come out.

Election fraud expert Walter Mebane wrote that “fraudulent votes in the election were not decisive for the result.”

And today John Williams points us to this news report that states, “a study by independent researchers, using data obtained by The New York Times from the Bolivian electoral authorities, has found that the Organization of American States’ statistical analysis was itself flawed.”

The new report is by Nicolás Idrobo, Dorothy Kronick, Francisco Rodríguez, and here’s the abstract:

Surprising trends in late-counted votes can spark conflict. In Bolivia, electoral ob- servers recently sounded alarms about trends in late-counted votes—with dramatic political consequences. We revisit the quantitative evidence, finding that (a) an apparent jump in the incumbent’s vote share was actually an artifact of the analysts’ error; (b) analysis of within-precinct variation mistakenly ignored a strong secular trend; and (c) nearly identical patterns appear in data from the previous election, which was not contested. In short, we examine the patterns that the observers deemed “inexplicable,” finding that we can explain them without invoking fraud.

So, more evidence that the OAS was jumping the gun in its conclusions.

P.S. regarding the title: I haven’t actually seen evidence of election fraud fraud (i.e. fraud in the claim of fraud). My guess is that they were convinced going into the election that the election wasn’t fair, and then they interpreted all the evidence with that in mind. Not literally fraud—but remember Clarke’s law.


  1. Alberto says:

    Very plausible.

    Morales wasn’t supposed to run for re-election: Bolivians had denied him the right to a fourth term in a referendum. When Morales had the referendum result overturned by the Constitutional Court, the opposition concluded that Morales was ready to rig the system and wanted to president for life.

    When seemingly suspicious results came in on election night, the opposition unsurprisingly thought more rigging was going on and took the streets. The rest is history. A case of prior convictions indeed.

    • Carlos says:

      He became president in 2006, under a constitution that required him to leave at the end of 2009. Personally, I think that’s already illegitimate to change the Constitution and make it apply to yourself.

      Anyway, he changed it, allowing for 2 terms of 5 years, and sat for another term.

      Then, he went back on his commitment, decided that his first term didn’t count because it was before the new Constitution, and ran yet again. AGAIN, he was announcing his retirement.

      Only then, referendum, lost, declared his own constitution as unconstitutional (violating human right to be elected indefinitely under international treaties or something), and ran again.

      Votes aren’t all that make up a democracy, fraud or not. Imho, he should’ve been kicked out of the race from 2010.

  2. Luis says:

    I don’t know about the other two authors of the newest paper, but I would hardly characterize Francisco Rodríguez as an “independent researcher” given his position in Venezuelan politics. He might have incentives to undermine the OAS.

  3. Joshua says:

    Andrew –

    > My guess is that they were convinced going into the election that the election wasn’t fair, and then they interpreted all the evidence with that in mind.

    Glenn Greenwald makes the case that there was more than just confirmation bias going on.

  4. Not Trampis says:

    So you are saying the report was not boliviable.

  5. Michael Nelson says:

    Frankly, election data fraud seems to be less of a threat than election “anecdote” fraud. Data, as we’ve seen, can be checked independently. But if an entire political party just decides to tell the same lie over and over again, many people will just assume there must be evidence, while others will proclaim they don’t need evidence to know that the other party is a bunch of cheating cheaters. And still others will know it’s a lie but stay silent because it’s to the benefit of their party, class or race.

  6. John Williams says:

    I don’t know what happened in this election, and I was hoping that someone would comment on the statistical analysis in question, but I think the situation should be viewed in the context of the long history of government by oligarchies in Latin America.

    All I know about Francisco Rodríguez is what I just read on Wikipedia and the Tulane website, but that does not make me think his analysis would be biased toward Morales.

  7. steven t johnson says:

    The implicit idea that the coup—which it was—was a response seems to be entirely unjustified. Any arguments about errors in the analysis seem to be have the effect of distracting from the actual course of events.

    The moral, i think is the old one: Figures don’t lie…but liars can figure.

  8. Another Joshua says:

    What changed is that the latest group of researchers, prominent academics from top universities who had their research summarized in the New York Times, couldn’t be casually dismissed quite as easily as the ones before and, crucially, they got their hands on the non-public electoral data. It wasn’t hard to tell that the OAS was wrong based on the less-detailed public data, but in order to reproduce their analysis, it was essential to have this non-public data. Previous researchers, such as John Curiel and Jack Williams, requested the OAS’s data and methods, but were stonewalled. I personally sent multiple emails to the author asking for this information, but he would not respond to me or to anyone else who sent him similar requests. I’m relieved that the analysis has finally been laid bare, but it’s dispiriting to see how long it took to get to this point. What if the Bolivian electoral authorities hadn’t been willing to share their data? Perhaps the issue would’ve simply remained unresolved for the public.

    Some optimists might think that this was all an honest mistake by the OAS, but it absolutely was not. It was part of a clear effort to deceive the public and preserve their own reputation. Jake Johnston of CEPR has claimed that a high-level official from the OAS admitted to him outright that they knew the statistics were bad. If the OAS truly believed in their analysis, then why did they obstinately refuse to share their data and code? Because they knew they were wrong but that simply did not matter to them. Their analysis was believe merely because it came from them, and it never occurred to the many credulous academics and journalists out there to check their work. The OAS spent many months cynically defending this defective analysis, and may continue to defend it in the future. In fact, it was an earlier of these defenses which seemed to provoke Francisco Rodríguez into studying the issue, as he called their non-response to Curiel and Williams “the research equivalent of pleading the fifth” and wrote perhaps the only perceptive summary of the debate back in March.

    Also, the statistics aren’t the only part of the OAS’s report that’s misleading. Many have wishfully concluded that the problems are only limited to that one section, but the rot goes deeper than that. The IT system the OAS is always on about? They don’t actually have any evidence of fraud there either, just lots of evidence of incompetence and bad practices. The account from their report omits many key facts that cut against the simple story of fraud they present. If you’ve only read that report, then you would have little idea of what actually happened. I don’t have much confidence that a major media organization will stick its neck out on that one, though, as it’s murkier than the clear-cut statistical evidence, which the mainstream media should have settled back in November or December. Instead, it took them eight months to finally drive a stake through its heart, and even then, the Times article which presented the refutation attempted to downplay its significance whenever possible.

    And lest you think that none of this matters, it must be noted that the OAS report is the central piece of evidence in legal proceedings against officials of the former government, many of whom are currently in pre-trial detention after having been charged with electoral fraud by the same people the OAS helped bring to power. Ordinarily one might expect journalists to be interested in helping to clear the names of the wrongly accused, but their general hostility toward the former government seems to have precluded such interest.

  9. Viggo says:

    What is wrong with PNAS?

Leave a Reply