December 12, 2019
December 12, 2019, Jake Johnston
Analysis by Jake Johnston, Senior Research Associate, Center for Economic and Policy Research to the OAS Final Audit on the Bolivian Elections.
December 12, 2019
Introduction
The OAS final audit of the Bolivian elections, published on December 4, concludes that there was “intentional manipulation” and “serious irregularities” that make it impossible to verify the results of the election.[1] Members of the de facto government and other critics of former president Evo Morales have used the findings of the OAS audit to corroborate allegations of electoral fraud leveled in the hours and days following the vote. But the following analysis shows that the report itself presents a biased and misleading account of the audit’s findings, presenting serious inaccuracies and downplaying or ignoring altogether any evidence that runs counter to the fraud narrative that the OAS has promoted since the day after the election.
This analysis is not meant to serve as a validation of the electoral results themselves. Rather, it is an analysis of the OAS’s actual findings and of the neutrality and rigor of the audit itself. This is not about supporting one political party or candidate over another. Nor is it solely about Bolivia. This is about the need for independent electoral observation in the hemisphere, and about accountability for an organization that has abandoned any semblance of neutrality under the leadership of Secretary General Luis Almagro.
A “Drastic” and “Inexplicable” Change in Vote Count Trend
-
The final audit repeats the false claim, which the OAS alleged the day after the vote,[2] that there was a “drastic” and “inexplicable” change in the trend of the results after the interruption of the preliminary TREP system on October 20.
-
As we explained in our November report, “What Happened in Bolivia’s 2019 Vote Count,” the trend change was both explainable, and predictable, based on the first 84 percent of the results processed before the TREP interruption.[3] Tally sheets counted later were more likely to come from areas that had already expressed a clear pattern in favor of Morales. In none of the OAS press releases or reports is this explanation seriously considered even though geographic differences in voting behavior are a common feature of elections.[4]
-
In the final audit report, the OAS slightly changes its analysis, alleging the real disruption in the trend occurred over the final 5 percent of the vote processed. However, the data provided on pages 93–94 in the final audit reveal that Morales’s share of the vote actually decreased in the final 5 percent of the votes counted as compared to the 5 percent counted directly before, discrediting the OAS’s own analysis purporting to show a change in the trend.
-
Moreover, the trend seen in the final 5 percent of the vote count is again entirely predictable based on the prior trends seen in the geographic areas from where these final votes came.
Security Vulnerabilities and the Manipulation of Electoral Results
-
Nearly 50 percent of the final audit is dedicated to information technology and security vulnerabilities in the electoral software. This analysis concludes by finding: “It is not possible for the audit team to guarantee the integrity of the data and give certainty of the results.” The OAS auditors appear to be particularly concerned by the discovery of a “hidden” server that was set up after the suspension of the preliminary TREP system on the day of the vote and that the report authors suggest would have allowed for the direct manipulation of the results of the election.
-
In performing an audit of the elections, the OAS could have undertaken a number of different verification exercises to check the legitimacy of the vote, i.e., to confirm whether the “hidden” server or other database vulnerabilities had been taken advantage of to actually manipulate results. These tests are simple and obvious to anyone familiar with electoral observation.
-
More troubling is the fact that the audit report conceals or fails to provide information on the basic verifications of voting material that the auditors did perform and that could provide evidence — if evidence exists — of potential manipulation of the voting results.
-
The auditors analyzed tally sheet images from the TREP and Cómputo (official vote count) websites. The OAS report notes, on page 82, that in 99.8 percent of cases, the data entered into the Cómputo (official system) matched what was reported on the tally sheets themselves.
-
The final audit report discounts this finding by noting that the image analysis does not confirm the authenticity of the tally sheets themselves. But nowhere in the report do the auditors note that there are simple processes to verify tally sheets.
-
There is no evidence in the report that the OAS auditors cross-checked the tally sheet images posted publicly online with any of the copies of the physical tally sheets that are provided to political parties, notaries, and local electoral officials on the day of the vote.
-
On page 84 of the report, the auditors note that 894 tally sheets, out of a statistical sample of 2,863, were analyzed in order to cross-check with original electoral material. The sample size is not explained and only mentioned in a footnote. The OAS reports that auditors were sent to five departments in order to analyze the 894 tally sheets and that, for 230 of them, original electoral material had been burned, preventing their verification.
-
Nowhere in the audit are the findings of the verification of the rest of the sample provided. To be clear: the OAS performed a verification exercise meant to check the validity of the tally sheets themselves, and did not report the results of that exercise anywhere in the final audit report.
-
To the extent that OAS auditors were unable to verify the tally sheets, it was largely due to the destruction of electoral materials. The audit includes the destruction of electoral materials in its findings of “deliberate actions that sought to manipulate the results of the election.” Nowhere in the report does it provide contextual information about the destruction of electoral material.
-
The Inter-American Commission on Human Rights, an autonomous body of the OAS, found that, in the days after the election, protesters burned departmental electoral buildings, where sensitive material is stored, in the departments of Chuquisaca, Potosí, Beni, and Santa Cruz.[5] Local news reports indicate that offices in the department of Pando were also attacked during protests.[6] According to information provided on page 78 of the OAS audit report, 99.98 percent of the burned or destroyed tally sheets and 100 percent of the burned or destroyed electoral lists were located in these five departments.
Other Alleged Irregularities
-
The OAS analyzed 4,692 of the most heavily pro-Morales tally sheets from across the country and performed handwriting analysis and other methods to determine if they had been manipulated. The analysis resulted in the identification of 226 tally sheets where the same person had filled out two or more tally sheets within the same voting center. The OAS final audit includes these tally sheets in its findings of “intentional manipulation.”
-
In only one of those cases does the OAS allege that the signatures on the tally sheets were forged or manipulated.
-
78 percent of the impacted voting centers are very small, with four or fewer voting tables (tally sheets correspond to these voting tables). Far from indicating fraud or the “deliberate” manipulation of results, as the OAS claims, this is indicative of a well-known phenomenon: in rural areas and smaller voting centers, it is not uncommon for one person to fill in the tally sheet, and then have the individuals each sign it. This is especially common where illiteracy rates are high. According to Article 49 of the Rules of the Election and the official guides issued by the TSE, it is the signatures that are required from the rest of the electoral workers and delegates.[7]
-
Further, the actual results expressed in the questioned 226 tally sheets are consistent with the results shown on tally sheets from the same voting centers and those in nearby geographical locations — many of which, because they are all in overwhelming pro-Morales areas, were included in the sample of tally sheets analyzed by the OAS auditors for manipulation. Two-thirds of the 226 tally sheets were in the TREP system before its interruption. Simply put: there is nothing abnormal about the results shown on most of these tally sheets.
-
The OAS final audit notes the presence of TREP (preliminary) tally sheets entered into the Cómputo (official) results. “The link between the TREP, an openly manipulated system, and the Official Computing affects the credibility of the latter,” the final audit states. However, the final audit reveals that 91 percent of the TREP tally sheets in the official results come from voting centers abroad, which is consistent with the established procedure, as the OAS final audit itself recognizes.
-
In the remaining 9 percent of cases, TREP tally sheets were used in the Cómputo (official) because the physical copies of the tally sheets had been burned in postelection violence. This is noted on page 45 of the final audit report.
Conclusion
In conclusion, the OAS audit identifies irregularities and vulnerabilities, many of which impact elections across the hemisphere — and the world — but then concludes that they amount to the “deliberate” manipulation of electoral results. By decontextualizing the evidence that they do provide, burying results that run contrary to their stated findings, and concealing evidence that they almost certainly have but do not report, the OAS final audit presents a misleading, biased, and incomplete picture of the Bolivian elections of October 20. The final audit report is an attempt to mislead the public and justify the organization’s — and its outspoken secretary general, Luis Almagro’s — unsubstantiated claims of fraud in the October 20 election in Bolivia.
What the OAS audit does identify are electoral irregularities, which the organization and its observation missions would typically provide recommendations for addressing in the future. Certainly, there are aspects of the Bolivian electoral system that should be improved. But there is an important distinction between irregularities and fraud, as the OAS election department is surely aware. The OAS final audit does not prove — or even convincingly argue — that the results of the election were manipulated. On the contrary, a close reading of the OAS final audit reveals the opposite: that despite the existence of vulnerabilities and irregularities, there is no clear evidence that the results of the election were systematically altered or manipulated.
[1] Organization of American States (OAS). 2019. “Final Report of the Audit of the Elections in Bolivia: Intentional Manipulation and Serious Irregularities Made It Impossible to Validate the Results.” December 4. https://www.oas.org/en/media_center/press_release.asp?sCodigo=E-109/19.
[2] OAS. 2019. “Statement of the OAS Electoral Observation Mission in Bolivia.” October 21. https://www.oas.org/en/media_center/press_release.asp?sCodigo=E-085/19.
[3] Long, Guillaume, David Rosnick, Cavan Kharrazian, and Kevin Cashman. 2019. “What Happened in Bolivia’s 2019 Vote Count?: The Role of the OAS Electoral Observation Mission.” Washington, DC: Center for Economic and Policy Research, November. https://cepr.net/images/stories/reports/bolivia-elections-2019-11.pdf?v=3.
[4] Chang, Ha-Joon, et. al. 2019. “The OAS has to answer for its role in the Bolivian coup.” The Guardian, December 2. https://www.theguardian.com/commentisfree/2019/dec/02/the-oas-has-to-answer-for-its-role-in-the-bolivian-coup.
[5] OAS. 2019. “CIDH presenta sus observaciones preliminares tras su visita a Bolivia, y urge una investigación internacional para las graves violaciones de derechos humanos ocurridas en el marco del proceso electoral desde octubre de 2019.” December 10. https://www.oas.org/es/cidh/prensa/comunicados/2019/320.asp.
[6] Corz, Carlos. 2019. “Justicia envía a la cárcel a seis personas por saqueos e incendio del TED de Pando.” La Razón, October 23. http://www.la-razon.com/nacional/animal_electoral/Justicia-personas-incendio-TED-Pando_0_3244475590.html.
[7] Órgano Electoral Plurinacional (OEP). 2019. “Reglamento para las Elecciones Generales 2019.” May 24. https://www.oep.org.bo/wp-content/uploads/2019/05/Reglamento_Elecciones_Generales_2019.pdf.
______. 2019. “Capacitación Electoral: Guía para Juradas y Jurados Electorales.” https://www.oep.org.bo/wp-content/uploads/2019/10/Guia_Jurados_EG_2019.pdf.
