Analysis of the Final OAS Report on Bolivia’s Elections

The Organization of American States (OAS) again issues a report with clear intentional bias. The document clearly aims to certify its own preliminary version and, with it, the argumentative basis for the coup d’etat in Bolivia. However, the reading of this report, like that of its preliminary version, yields simple conclusions about the intent – intent to manipulate – with which they were prepared. For the Spanish PDF version click here.

General analysis
The OAS report focuses on arguing alleged irregularities in two phases of the electoral process. On the one hand, the TREP (preliminary election results transmission system), and on the other an alleged fraud by forgery of signatures.
The OAS divides its report into “evidence” of three types: those in which there is intentionality or intent (deliberate actions), serious non-malicious irregularities, and errors and findings. Attachments to the report are done detailing the work carried out.
To facilitate the understanding of this counter-report, we will largely follow the sections of the document issued by the OAS.

  • On the analysis of the TREP

The OAS focuses the report on the suggestion of fraud in the TREP system:

  • 10 of the 12 deliberate actions refer to the TREP
  • 10 of the 13 serious irregularities refer to the TREP
  • 4 of the 7 errors refer to the TREP

It is necessary to indicate, in addition, that the bulk of the arguments contained in these elements refers, on the other hand, to the events that occurred after the end of the TREP count with 83.7% of the recorded voting ballots.
However, even if irregularities are admitted or, in the extreme, deliberate actions in the TREP, the OAS omits the central facts about this system and that is that, according to Bolivian electoral regulations, the TREP is not the system of official counting, and therefore not a binding counting system.
It does not seem reasonable to base an accusation of fraud on an unofficial counting system, but it is also that the OAS – the main promoter of the introduction of this counting system in Bolivia in this electoral process – omits crucial information about it, namely:

  • That, according to the statements of a member of opposition affiliation in the TSE (Supreme Electoral Tribunal), Antonio Costas, made to the Xinhua news agency on October 10, prior to the elections in Bolivia: “For the first time in its democratic history, Bolivia may know (fast about) 80 to 90 percent of the election count.”(1)

OMISSION 1st: The TREP was designed to deliver data up to 80% to 90%, and no more. The Bolivian TSE delivered data up to 83.76%, therefore attending to its commitment. The OAS, by deliberately omitting this information on the limited role of the TREP in the design of the electoral structure, promotes the erroneous conclusion that the TREP was the counting system that should yield a definitive result in the elections counting up to 100%.

  • The general manager of the company Neotec, in charge of administering the TREP, Marcel Guzmán de Rojas, confirmed this information, indicating “that 34,000 records will be verified in two hours to get between 80 and 90 percent of the results in the same election day. ”

OMISSION 2nd: According to the TSE, a total of 34,558 voting tables worked in the country and abroad during the day of the general elections. The TREP was designed in such a way that it would NEVER count 4,558 tables. Nothing is said about the criteria that defined which tables were counted and which were not, much less the probable political orientation of these according to their rural or urban location, with their consequent bias.

  • Given that the OAS prepares its report, gravitating primarily on the TREP, it is strange that it dedicates only a few lines to the crucial role of the company in charge of it. It is also surprising that the OAS obviates the fact that this system was installed in Bolivia under its [OAS] recommendations. About the aforementioned company, the principal in charge of the TREP, the OAS deliberately conceals at least three crucial facts:

OMISSION 3rd: That, as indicated by the member Costas, it was the first time that this technology was used and that, therefore, inexperience could be behind the imprudence detected. The OAS recognizes it on page 46 of its report, but deliberately omits it from the conclusions or executive summary section: “There was a lack of maturity of the process in relation to software, on the one hand, due to the absence of cases of use and several software tests (unit test, integration test and regression test) and, on the other, because the tests performed lacked a formal software acceptance process with formal test cases”. (2)

OMISSION 4th: That Neotec was primarily responsible for the computer configuration reported by the OAS. Again, the declarations on dates prior to the election of the opposition witness itself, proves it: “the TSE made a great investment in hiring a company to have the results of the elections as soon as possible and thus provide certainty to the population. It is the system of Transmission of Preliminary Electoral Results (TREP), with programmatic technology that had not been used before in Bolivia”.

OMISSION 5th: Finally, the OAS obviates that Neotec’s manager, Marcel Guzmán de Rojas, had expressed his preference for Carlos Mesa, with whom he is known to have a close relationship. This fact does constitute a serious violation of the appearance of impartiality by the principal manager of the counting system.

Conclusion: The OAS bases its conclusions on the alleged finding of 12 malicious (intentional) irregularities. Of the 12 irregularities listed by the OAS, 10 refer to the TREP, an auxiliary system, a non-binding system and a system that, as the OAS itself and the parties participating in the electoral process recognize, could NOT determine the electoral result since NONE of them were designed to count more than 34,000 records.

2) On the irregularities in the official count
Regarding the incidents in the official count – remember, the only valid procedure because it scrutinizes 100% of the voting records and, therefore, the only one that could reveal any evidence of real fraud, is to suppose that the OAS – the agency – should have been especially cautious in its conclusions. However, what does the OAS say about the official count? Let’s look at it point by point:

2.1) On deliberate actions

  1. “In an exercise that sought to analyze possible adulterations or manipulations, a sample of 4,692 records was reviewed. In this analysis, 226 voting rolls were identified in which two or more ballots at the same voting center were filled out by the same person, denoting an intentional and systematic action to manipulate the electoral results in violation of the powers of the Board Jury determined by law. The records correspond to 86 voting centers in 47 municipalities of the country. The sum of its valid votes is 38,001, of which 91% (34,718) were awarded to the Movement to Socialism (MAS).”

This point is especially serious, since it constitutes the only argument to justify that fraud was done in the system of the official count, the manual one. However, the OAS omits in its initial relationship crucial information that it later reveals in the report itself, namely:

  • OMISSION 1st: the OAS points to 34,718 votes from 47 municipalities in which, allegedly, it found irregularities. It omits that only 4.6% (as referred to on page 9 of the report) could become compromised and, therefore, even if we adjudicated all the records analyzed to Carlos Mesa, of the Citizen Community (CC) 85.4% (91% minus 4.6%) of the votes in those areas would indeed be for Evo Morales, a figure that would have given him the victory in the first round.
  • OMISSION 2nd: The OAS conceals from its conclusions that the 4,692 ballots recorded correspond to a partial sample of ballots that did NOT pass through the TREP (page 9). They are, therefore, a sample of 5% of votes scrutinized last, that is, rural votes in more remote areas and where literacy levels are lower.
  • OMISSION 3rd: that the ballots identified as “irregular” make up 4.8% of this universe (226 of 4,692), which represents only 0.25% of the entire universe. In other words, even if Carlos Mesa was assigned 100% (unlikely, as we have seen), the difference for Evo Morales would not fall below 10.41%.
  • OMISSION 4th: Finally, it is necessary to indicate that all the records objected by the OAS are in areas where there was NO presence of opposing panel juries, a clear indication of their low presence in those places. On the other hand, the OAS itself recognizes that it was NOT able to make additional checks because of the burning of records (at the hands of the opposition, something it omits). (3)
  • OMISSION 5th: At no time does the OAS make a simple random sample of the proceedings of the electoral process for validation and comparison. Only such a sampling would have statistical validity, allow a projection of its figures to the entire universe and, therefore, identify the magnitude of the irregularity.
  • OMISSION 6th: “In spite of being sensitive material, records were burned (the number is uncertain) and more than 13,100 lists of authorized voters (or index lists), which does not allow the verification of the information recorded in the examination and counting of records.” It’s powerfully notable that the OAS imputes as proof of fraud the burning of records at the hands of the opposition, a burning of records that were carried out after its own claim of fraud.

These and no others are the only two arguments the OAS makes to announce fraud in Bolivia. Clearly, the OAS incurs, in manifest incapacity and in serious intentional interference, in the process since neither argument sustains an accusation of such magnitude.
Let’s look at them in detail:

2.2) About the statistical “findings”

  • The OAS, surrounded by the certainty that it cannot demonstrate any fraud in the official count, uses the argument that the official counting system was fed to a measure that is not determined strictly (“more than 5%”), by the TREP. However, on page 45 of its own report it confesses that the transfer between TREP and official counting was reduced by foreign records (which would take time to be sent otherwise) and by the records burned by the opposition in Chuquisaca, Potosí and Santa Cruz. Note that 1) said burn or delay is not attributable to the TSE and 2) these ballots are not among those challenged or questioned, even among those that make up the final 5% to be charged, so they do not imply any reduction in validity of the procedure.
  • The OAS argues as a final argument that the trend observed in the count of the last 5% of minutes is not statistically consistent with that of the general process. However, the OAS – as already happened in the preliminary report – obviates the obvious: that the remaining 5% of records were obtained from voting centers in areas of overwhelming MAS majority, areas in which even the vote by prescription is community practiced, a phenomenon fully known among the different forms of participation of the Bolivian indigenous rural population.
  • If we go into the detail of this 5%, the OAS itself is forced to recognize in its report that only “59 (5.5%) with serious irregularities from the expert point of view” (4) were found among the universe of the last 5% of ballots entered in the manual official computer system.
  • Extending from this point and on the final ballots of rural areas, it is said: “Subsequently, the universe of analysis was expanded, taking a new sample of 3,618 ballots. Of these, 167 (4.6%) with irregularities of expert interest were identified. Different ballots from the same voting center were found again that had been completed by a single person, which clearly constitutes an illegal transgression of the powers of the panel jurors and sows doubts about the reported results. In total, 4,692 ballots were analyzed, of which 226 (4.8%) presented the irregularities described.”
  • As the report explains, the suspicious acts arise from a sample – not random but from those in which the MAS obtained its best voting, more than 77% of the votes – of 4,692 electoral records that did not pass through the TREP. Of these, the OAS affirms that in 226 – less than 5% – it happened that two or more ballots of the same voting center were completed by the same person.
  • Now, what happens in these 226 tables? Is there an anomalous electoral behavior that is exponentially opposed to the rest of the tables of their respective municipalities? To answer these questions, we have prepared a table, in which, on the one hand, the percentages obtained by the MAS and the Citizen Community are recorded at the tables detailed by the OAS on pages 56 to 58 of the aforementioned report, while that in the columns on the right the percentages obtained by both forces are registered in all the other tables of said municipalities –that is, discounting the disputed tables of the municipality:

Screenshot at 18-00-36.jpg

Screenshot at 18-03-01.jpg

  • As we can see in the table, in the vast majority of the cases the percentages do not vary substantially. In 22 of the 47 municipalities – the OAS in its table consigned 48 municipalities but in reality tables 35723 and 35724 do not belong to the municipality of Totora but to that of Mizque, we will take it as a simple error on their part and not as a malicious action- the difference between the advantage that Evo Morales obtained in the tables in question exceeds the advantage obtained in the rest of the tables of the respective municipality by less than 10 percentage points. In fact, to speak more rigorously, among these 22 municipalities there are even cases such as Ansaldo, Vila Vila, Punata and Mocomoco, in which the MAS obtained a greater difference in the rest of the municipality than in the questioned records.
  • The OAS seems to have drawn attention to the fact that in many of these ballot, MAS obtained percentages of votes greater than 90%. However, it is obvious to say that these voting levels are recorded in municipalities that are tremendously favorable to MAS. Of the 47 municipalities, still discounting the questioned ballots, it is observed that in 39 – almost 85% – the difference for Evo Morales was more than 50 percentage points. Even if we look closely at these 39 municipalities without the questioned minutes, we will see that MAS exceeds in 70 of them 70% of valid votes, while CC only exceeds 10% in 11 municipalities and in none reaches even 20% of the votes.
  • For all the above, we can only qualify as tendentious the attempt to challenge the almost 35,000 votes of advantage that Morales obtained in these 226 records since most of them do not report significant differences with respect to their municipal averages.
  • Moreover, it seems that the proof of the OAS complaint is that the values in these records are slightly higher in favor of the MAS, when in practice this is but the logical result of an arbitrary selection precisely because of that condition. According to the OAS, Morales did “suspiciously” well in municipalities where the vast majority of unquestioned records reflect a difference of more than 50 points in favor of MAS, an otherwise repeated difference in successive elections over time.
  • Finally, and even if we accept all the arguments of the OAS, and that the alleged “fraud” detected in 4.8% of its strongly skewed sample could be extended to the total universe of 5% of the final records of the whole of the scrutinized ballots that did not go through the TREP, the result is that doubts could only be affirmed in 0.24% of the total records. If for greater security we assign 100% of the votes of that 0.24% of the minutes to Carlos Mesa, the result of the electoral process would have been 46.83% for Evo Morales and 36.75% for Carlos Mesa, which would have kept the difference between the two at 10.08%, giving Evo Morales the victory in the first round.

The OAS directs all its analyses and conclusions to support an alleged fraud in the Bolivian elections. To do so, it omits crucial legal-normative and technical-analytical information. Indeed, its construed argument on the alleged irregularities in the TREP omits three transcendental facts: 1) this is not the official counting system, 2) the TREP was not designed to yield data above 90% in any case and 3) that the alleged irregularities in the data loading procedure, if true, would justify a procedural irregularity, since nothing is proven about how and why the alleged additional servers would have been used.
With regard to the technical-analytical issue, the OAS bases its entire argumentation of fraud on the existence of 226 records of 4,692 ballots (4.8% of this universe) in which supposedly there could be irregularities (although it confesses not be able to access the original records). These questioned records, which did not go through the TREP and correspond to the final 5% of the scrutiny questioned by the OAS, represent 0.24% of the total of the recorded ballots.
Finally, in regards to finding 5, in which the OAS speculates about a usual or unusual result in which the last 9% votes counted clearly favors MAS, you must perform the following considerations:

  • That the OAS arbitrarily estimated what is usual or unusual
  • The OAS again ignores the very clear gap in terms of the vote that has systematically taken place in Bolivia between rural and urban areas.
  • That an essential piece of information for all subsequent arguments regarding the robustness of the TREP sample is obvious in the report: how was it set up, according to what criteria, privileged rural areas over urban areas or vice versa?
  • Ultimately and fundamentally, that even in the event that we admit all the arguments of the OAS and that, in effect, the alleged “fraud” detected in 4.8% of its highly skewed sample could be assigned 100% to the opposition and extended to the total universe of 5% of the final ballots of the whole of the scrutiny that did not go through the TREP, the result of the electoral process would be 46.83% for Evo Morales and 36.75% for Carlos Mesa, which would have maintained the difference between both at 10.08%, giving Evo Morales the victory in the first round.

In short, the OAS report, whether due to technical disability or intent, incurs manifest weaknesses. Taking into account its importance, it is at least questionable that such conclusions issue from it, such as those repeatedly made by the OAS Secretary General, Luis Almagro, and even much less to question, giving wings to the subsequent actions of the armed forces, the general integrity of the electoral process in Bolivia.


  2. Page 46 of the OAS report.
  3. Page 59 of the report: “The audit team could not fully deepen this analysis and carry out additional checks because part of the electoral material of the departments of Potosí, Chuquisaca and Santa Cruz were burned down. Yes, it was possible to obtain and analyze original material from the department of Cochabamba ”.
  4. Page 9 of the report.


Translated by JRE/EF

Analysis of the Final OAS Report on Bolivia's Elections
Website | + posts