This is the fourteenth article in a series of blog posts that examines Diego Escobari and Gary Hoover covering the 2019 presidential election in Bolivia. Their conclusions do not hold up to scrutiny, as we note in our Nickels Before Dimes report. Here, we expand on the different claims and conclusions made by Escobari and Hoover in their paper. Publication links: Part One, Part Two, Part Three, Part Four, Part Five, Part Six, Part Seven, Part Eight, Part Nine, Part Ten, Part ElevenPart twelve and part thirteen.
This blog series examined these assumptions and biases in Escobari and Gary A. Hoover’s 2020 report, “Evo Morales and Electoral Fraud in Bolivia: Natural Experiment and Evidence for Discontinuity.” We looked at counting biases and noticed that support for a particular candidate doesn’t have to be uniform on count. A candidate’s vote share may trend upwards and even swing rapidly upwards near the end of the count.
We have seen that Bolivia’s elections in 2019 as well as the referendum in 2016 showed similar trends, with arguably “left” voters tending to announce their ballots later than others. Of course, the rural districts – where voters strongly favored the referendum – came later. Constituencies in capitals — where Mesa was a particular favourite, even among opposition candidates — tend to report fully early in the count. Voting patterns at the time of the TSE announcement, combined with knowledge about polling stations that remained undisclosed, suggested that Morales would win the first round of the 2019 presidential election.
This clearly disproves the OAS claim that the late results and first round victory was “inexplicable”. An election of any size can in no way be explained until the last vote. However, it appears, based on the data provided by the TSE at the time of the announcement, that a Morales victory in the first round was not only possible, but highly likely. As of this writing, the Organization of American States has not been adequately held accountable for its role in undermining the credibility of Bolivia’s electoral system. Thus, we reiterate our call to the OAS to engage openly and transparently regarding its activities surrounding the 2019 elections.
We have seen that naive ‘divergence model’ estimates are misleading. Because late polling stations were suspected of being tainted by fraud, these models measure the effects of confounding factors such as volatility.
We’ve seen that adding small amounts of information about elections, and district-level results from 2016, for example, goes a long way in explaining the results of elections in 2019, but care must be taken not to over-build our model by making very strict assumptions. Benign differences between 2016 and 2019 can violate the divergence model assumption of parallel trends and lead to a biased estimation.
We have seen that once Escobari and Hoover begin to allow for these benign differences, very little of the election is left unexplained. However, Escobari and Hoover reinterpreted their results by identifying the same non-parallel directions as a sign of fraud. We stress that this reinterpretation is completely inconsistent with previous modeling assumptions. In any case, the reinterpretation re-imposes overarching assumptions that exclude the very important benign causes of those differences in directions.
We are highly skeptical of the results of the “triple” distinctions of Escobari and Hoover. The common baseline for the 2016 results reduces the model to a difference in variances that contrasts major end margins (MAS-CC) to minor end margins (MTS-21F) although the direction of these different margins can in no way be expected to parallel. We wonder how Escobari and Hoover could find such young “Triple” differences regardless of interpretation.
Even accepting the results presented by Escobari and Hoover, their reinterpretation that the measured difference in trends indicates 2.5 percentage points of fraud again requires unjustifiably strong modeling assumptions.
We reiterate our call to Escobari and Hoover to engage openly and transparently on the sensitive political issue in which they have chosen to insert themselves.