|

Do the Venezuelan Opposition’s Cries of Fraud Stand Up to Scrutiny?

Venezuelanalysis took a deep dive into the data behind the Venezuelan opposition's allegations of electoral fraud. This is what we found.

02076a9345774c2846cd44f338ba27ef

Crying fraud is an extremely familiar routine to the Venezuelan opposition, and one that it has pursued at virtually every election since 1999, with the exception of the 2015 National Assembly elections when the opposition won by a landslide and suddenly found itself a born again believer in Venezuela’s democratic system.

And so it is no surprise then that since the shock results of last Sunday’s regional elections were announced — handing over an admirable win of 18 states out of 23 to the Great Patriotic Pole Chavista coalition — the MUD has rapidly been making use of all of the social media and international media connections at its disposal to circulate accusations that it was robbed of last Sunday’s elections due to fraud. Yet a cursory look at these latest allegations exposes them to have very little credibility.

Venezuela’s Electoral System

Firstly, it is important to highlight that with the notable exception of a controversial case in Bolivar state, the MUD has so far failed to produce any evidence to show that the final tallies circulated by the National Electoral Council differ from votes cast by the electorate – a fact that even some of their own leaders have recognised. Nor did any of the 70 international or 1240 national observers (of which 50% were opposition supporters) in electoral centres report any irregularities during the vote the vote itself.

As former BBC journalist and international electoral observer Javier Farje noted last Sunday, it is very difficult to rig Venezuela’s “water-tight” automated voting system.

In this system, both ID cards and fingerprints are used to establish the voter’s identity (preventing double voting), while the voter chooses their preferred candidate from a selection on the screen of an electronic voting machine. Once they have made their choice, they are immediately presented with a paper confirmation of their vote which they can check and then deposit in the ballot box. The voter then proceeds to sign and leave a fingerprint in the centre’s record book. When voting closes, 54% of the paper ballots are cross-referenced against the electronic tallies printed off by each voting machine to ensure the veracity of the result, while electoral witnesses are given the opportunity to confirm that the electronic vote corresponds to the paper vote. A full 100% audit of paper votes can be carried out on request, and both pro-government and opposition witnesses are able to retain their voting centre tallies.

All in all, it’s pretty bullet-proof. As the economist Francisco Rodriguez at the financial think tank Torino Capital states: “Under this system, it is straightforward to demonstrate that an alteration of the vote totals has taken place as inconsistencies between the national vote total and witnesses’ records of voting center tallies would emerge”.

While no such inconsistencies have been highlighted in 22 out of Venezuela’s 23 states, on Thursday the Roundtable of Democratic Unity [MUD] coalition published alleged copies of eleven tallies printed from different voting machines in Bolivar, claiming that they differed from the CNE’s published results.

Following a revision, the tallies do in fact differ from those on the CNE website, with the votes obtained by the GPP candidate Justo Noguera Petri inflated by several hundred. This suggests two possible scenarios: either the tallies print by the MUD are forgeries, or there has somehow been some kind of tampering with results.

According to the opposition, these particular voting machines failed to transmit the vote tallies electronically to CNE headquarters, and as such the votes for Petri were allegedly inflated by the people on the ground responsible for sending them manually.

Though there are a number of holes in this version of events – not least because each of the eleven voting machines was just one out of several used across ten different voting centers, meaning that at least eleven different people in ten different voting centers would have had to be in communication and simultaneously partake in this fraud undetected and beneath the noses of opposition witnesses – it is clear that a full audit of paper votes needs to be conducted to clarify what happened.

This is particularly important given the closeness of results in Bolivar – the tightest race of this year’s regional elections – where Petri beat his opponent Andres Velasquez by 276,655 votes to 275,184 – a difference of just 1471.

In conclusion, Venezuela’s regional elections last Sunday were clean, with the exception of some alleged irregularities in Bolivar state which need to be urgently resolved.

Other Allegations: Candidate Substitutions

Two other claims put forward by the MUD to add weight to its allegations of fraud are that 1) the CNE did not accept the coalition’s candidate substitutions in some states, meaning that two MUD candidates appeared side by side on the ballot when there should have been a unified ticket and 2) that the National Electoral Council relocated 273 electoral centres in the run-up to the elections.

Both of these issues should be taken with a pinch of salt, given that the MUD knew about both of these points well in advance of the elections and decided to proceed anyway, making it particularly hard to take their complaints seriously – a bit like a team that takes part in a football match in full knowledge of the rules, only then to complain about those rules once they lose the game.

The first point regarding the substitute of candidates is fairly easy to discard as a reason for the opposition’s heavy losses on a national level. The controversy between the opposition and the CNE over the candidate substitutions in the run up to elections was centred around the legal timeline during which parties can make changes to who will stand on their behalf: the MUD argued that it should have been able to substitute candidates up to ten days before the elections, while the CNE argued that August 16 was the cut-off date and this had been agreed with opposition representatives.

Either way, in all but one state where the GPP won, the vote was not sufficiently split between the opposition candidates on the ballot for official MUD parties in order to make a difference to the overall results. In fact, most “secondary” opposition candidates won on average between 3,000-10,000 votes – coming well below the GPP’s margin of victory in almost all states where it won.

Take Lara for instance: both MUD candidates Henri Falcon and Luis Florido appeared on the ballot, but Florido won 3,283 votes – even coming in behind Salas Tony, the alleged neutral “NUVIPA” candidate. In comparison, the official candidate Falcon won 325,109 votes, but still lost by 145,746 votes to GPP candidate Carmen Melendez. This is pretty much the picture across states where a secondary opposition candidate appeared on the ballot.

Results in Lara

The exception to this pattern was once again to be found in the state of Bolivar. As in other states, the secondary opposition candidate Francisco Sucre, who still appeared on the ballot, received a much lower number of votes. However, in this case Sucre received 3,787 votes – enough to tip the vote over to Velasquez if they had been intended for the official candidate of the opposition.

Nonetheless, there are a series of reasons to believe that these votes could have been a protest vote against the MUD’s official candidate, as was surely the case in other states. Given that the opposition campaign focussed on the issue of the candidate substitutions for weeks prior to the elections: filing two appeals with the Supreme Court and constantly reminding their supporters to vote for the official candidate featured on the MUD website, it is difficult to believe that the message did not reach the 3,787 Sucre voters.

In addition, while the opposition primaries in early September were marred by infighting and even violent street brawls, the competition for the candidacy of Bolivar was particularly intense. Velasquez’s candidacy stands out as unique, as he did not represent any of the main MUD parties, but rather the centre-left party La Causa Radical.

Rising to prominence through worker organisation in the factories of industrial Guayana City in the 1980s, the LCR commands significant support from trade-unions and workers’ organisations in Bolivar, but has less muscle on a national level – especially since the left-wing of the party split to join the emerging Chavista movement in the late 1990s.

Though Velasquez may have been the favourite of opposition aligned-trade unions in Bolivar, Francisco Sucre was the preferred candidate for the opposition’s biggest three parties in Caracas: Justice First, Democratic Action and Popular Will. This was not the only time when MUD leaders in Caracas tried to impose their will over local conditions during the primaries. Following Sunday’s elections, former opposition governor for Amazonas Liborio Guarulla accused the MUD of putting an end to a 17 year indigenous people’s project in the state “by imposing a candidate according to their interests in Caracas”.

After Velasquez’s victory, the MUD leadership never really threw its full national weight behind his campaign and he fell back on local support. It is perfectly possible that several hard-right Popular Will supporters simply found it too unpalatable to vote for Velasquez, whose support base has more working class origins, and instead opted for Sucre – a VP member himself.

However, even if we contextualise Bolivar’s outcome within broader results across other states, we can see that the votes for the secondary candidate were still one of the lowest – even though Bolivar is a mid-level state. Meanwhile, according to results released by the CNE, the GPP increased its vote by 33,806 relative to 2015, while the MUD plummeted by 112,587, demonstrating a strong general tendency for a Chavista win. This tendency would remain the case, even if Pietri’s votes had been inflated by several hundred.

In other words, the effect of the MUD’s inability to substitute its candidates had no bearing on the final election results in twenty-two out of Venezuela’s twenty three states. While it may have had an impact on the overall results in Bolivar, it is just one of many possible factors leading to a MUD loss there.

Nonetheless, given the issues mentioned previously in relation to allegations of vote-rigging in Bolivar, it is clear that this state once again needs to be examined by electoral authorities and a full audit carried out. The MUD could also argue for a re-run in this state based on the controversy surrounding candidate substitutions, given that it cannot be stated with absolute certainty that the 3,787 people who voted for Sucre did so intentionally.

Nevertheless, a case for running a repeat of the elections in Bolivar based on this argument would struggle, in view of the fact that the MUD was aware of this issue before it headed into elections. Likewise, any potential re-run would also have to take into account how the effects of the last few days, and in particular the protests called by Velasquez, might impact the holding of free and fair elections: in particular the intimidation of non-Velasquez voters.

Voting Station Relocations

The second point of the relocated electoral centres is slightly more complex to address and requires further scrutiny of voting trends. According to information provided by the CNE and CNE Rector Luis Rondon, 273 voting centres were relocated in the run-up to the elections – 72 across 15 states due to issues related to infrastructure, changes which are pretty routine before elections, and 201 across 7 states (Miranda, Carabobo, Merida, Tachira, Anzoategui, Lara and Aragua). The latter voting centres were moved as a result of the incredibly high levels of opposition violence that took place during the National Constituent Assembly elections on July 30.

As the CNE reported at the time, there were violent attacks against 200 voting centres, including with guns and grenades, while 181 voting machines were set on fire. This forced CNE officials to relocate the electoral centres, quite literally as voting was meant to be taking place.

Though international and Venezuelan national media have said that these changes were carried “without warning” and on the eve of elections, they were in fact made public well in advance. Take this article from Venezuelan newspaper El Nacional, which reported on September 15 that the CNE would relocate 200 centres – almost a month before elections. Even the economist Francisco Rodriguez of the center-right think tank and investment bank Torino Capital concedes that news of the relocations was made official three days before the elections.

Likewise, claims that voters were unable to find out the location of their new voting centres are also misleading. At any time, Venezuelan voters can find out in a matter of seconds which centre corresponds to them simply by entering their national ID number under the “consult your details” section on the CNE homepage. Voters were also able to text their ID numbers to a mobile number provided by the CNE and receive a text message with their newly allocated voting centre.

An example of CNE electoral-centre location service

In addition, it should also be noted that the average distance between former electoral centres and their replacements was hardly insurmountable. In the vast majority of cases this average was 2.2 km and ranged between 655 metres and 3.5km. Meanwhile, both the CNE and the opposition provided transport for voters to travel from their original centres to the new ones, and reports on the day also clearly show opposition voters lining up outside their new voting centres. Opposition spokespeople equally confirmed that their supporters still went out to cast their votes in spite of the changes to centres.

screen_shot_2017-10-17_at_6

screen_shot_2017-10-17_at_6

In short, though the relocation may have been an inconvenience for some, people were not impeded from voting. In fact, three out of the opposition’s five gubernatorial wins (Anzoategui, Merida and Tachira) were amongst the seven states where voting centres were relocated due to violent protest. Two of these states (Merida and Tachira) were where voting center relocation was highest, with fifty-eight and forty-two relocations respectively.

In a report published following elections, Torino Capital also estimated that the voting centre changes would have had only a minor effect in the states of Merida, Miranda, Aragua, Carabobo, and Lara, and a negligible to zero effect in remaining states. Conversely in Amazonas and Apure the changes actually worked against the government.

They conclude: “The decision to reallocate voting centers had only a minor quantitative effect on most races, and does not appear to be a key determinant of the results.”

However, if we look at the figures in the states that Torino Capital estimates could have been impacted by the centre changes, we can see that there is no way they were to blame for the opposition’s losses.

For instance, in the state of Carabobo where the opposition lost, 75,104 or just 4.8% of voters were affected by the relocation of 15 centres due to violence – out of a voting population of 1,561,777. But the MUD lost a total of 223,768 votes (down from 644,642 to 420,874) in comparison to its results from 2015 National Assembly elections – more than three times the amount of voters affected by the changes. Meanwhile, the GPP beat the opposition by 48,294 votes [452.081 (51.96%) vs. 403.787 (46.41%).

For the relocations to have changed the overall results in Carabobo, we would have to believe that almost 75% of the voters affected by the relocations across the state were opposition supporters that had intended to vote, but for some unknown reason were prevented from finding out the location of their new voting centre and traveling the short distance to get there.

We would also have to believe that these voters bucked the general trend of abstention amongst opposition voters in Carabobo, as well as national turnout levels of 61%, and that the election came down to their participation and not the other 175,474 former opposition voters that didn’t turn out.

You might be surprised to find that this version of events doesn’t stand up to further scrutiny. VA looked at the parishes most affected by centre changes in Carabobo, including Naguanagua (three centres), San Diego (four changes), Rafael Urdaneta (four changes), and San Jose (four changes), and found that abstention levels at these centres were in line with parish-wide abstention levels.

For instance, in Rafael Urdaneta, abstention levels were 39.33% while abstention levels across the relocated voting centres were 40.40%. In San Diego it was 42.19% across the parish vs. 40.8% in relocated voting centres. In Naguanagua it was 42.02% vs. 40.42% in the relocated centres and in San Jose it was 49.6% vs. 51.48%. Overall abstention levels in Carabobo were 40.93%.

In Miranda, where the opposition also lost, we can see a similar picture. In this state the opposition dropped from 838,292 to 555,347 – a total loss of 282,945 votes. But just 216,920 voters were affected by the electoral centre relocations, and presumably not 100% of them were opposition voters.

On Twitter, the losing candidate Carlos Ocariz said that 114,000 people were directly affected by the relocation of centres, and that they “couldn’t vote” due to this, while accusing the government of “stealing 88,000 votes” from the opposition.

But again, a cursory analysis of a sample of the 51 relocated voting centres in Miranda state gives a different picture.

In the Municipality of Sucre, Petare Parish, the average level of voter abstention was 47.15%. In this same parish, the CNE relocated 8 voting centres to 7 replacement centres. Voter abstention in these replacement centres ranged from 35.66% to 46.68%, with an average abstention level of 42.39%. In other words, in this municipality relocations actually correlated with higher voter participation, not less, as the opposition would claim.

In the Baruta parish of Baruta municipality, eight centres were moved. Their averages abstention levels were 55.5% versus 51.85% for the rest of the parish. In Las Minas, Baruta, three electoral centres were relocated. Their average abstention levels were lower at 51.41% than the rest of the parish at 54.19%. In Los Salias, San Antonio de Los Altos, seven centres were moved. Their average abstention levels were 47.4%, just 1.41% higher than the 45.99% in the rest of San Antonio.

What we see in Miranda is significant abstention amongst the opposition’s traditional rank and file, combined with the mobilization of working class communities in the state’s semi-rural outskirts such as Yare, Filas de Mariche and Cua, where the GPP candidate Hector Rodriguez almost took a clean sweep with participation levels reaching 77%.

VA also looked at the impact of fourteen relocated centres in Lara, as well as eighteen in Aragua, and the findings were the same: abstention levels in relocated centres were broadly in line with the rest of the parish. Though some of them were a couple of percentage points higher, others were significantly lower.

The numbers to support the claim that the MUD lost these particular elections because of the relocated centres just do not add up. But the statistics paint a clear picture: while the government somewhat rallied its forces, making gains in all states apart from Apure, the MUD suffered generalised and heavy losses, both where relocations took place and where they did not, going way above and beyond the potential knock-backs of electoral centre relocation. This includes states where the MUD won.

In Zulia, where the MUD emerged victorious, it still lost 329,289 votes – falling from 1,030,044 to 700,755. Just five centres were relocated in this state due to infrastructure issues. Though we do not have the specific numbers for these centres, even if each one corresponded to 5000 registered voters (centres tend to range between 1000-5000), then potentially affected voters would still be approximately 25,000 – a fraction of what the MUD lost. In Anzoategui – another 22,775 voters were affected by the relocation of six electoral centres, but the MUD lost 101,300 votes.

Even if we look at the states where the opposition won, but in which no voting centres were relocated, the pattern is still the same. In Nueva Esparta, the MUD went from 151,122 votes in 2015 to 117,430 on October 15 – a loss of 33,692.

In short: the data shows that electoral centre relocations had little to no impact on the outcome of the 2017 regional elections. The reasons for the MUD’s losses are likely complex, but if we were to speculate, the most obvious interpretation from these statistics is that the MUD hemorrhaged votes because of divisions amongst its ranks, poor handling of the political acumen gained from its 2015 win in legislative elections, and a simultaneous rejection from both moderate and hardline opposition supporters of the leadership’s 180-degree turn from violent insurrection to elections in a matter of days.

Meanwhile, Chavismo fought a grassroots campaign, and rallied voters into action across the country, increasing its share of the vote by 652,967 relative to 2015. These voters rejected the opposition and threw their weight behind Chavismo in spite of an ongoing economic crisis, demonstrating that Chavismo remains a key political player, if not the dominant force in Venezuelan politics.

But the mainstream media won’t say that, because any explanation that doesn’t prominently feature “fraud” would reveal that they have been consistently and grossly misrepresenting Venezuela’s reality and Chavismo’s popularity for years.

For Reference:

screen_shot_2017-10-20_at_10

With input and contributions from Lucas Koerner.