Monday, December 16, 2013

Detecting Voting Fraud in Honduras: El Paraíso, Copan

Since the Honduran election ended, we have busy here: we captured the Tribunal Supremo Electoral's election results, and placed them, along with the Voto Social vote count, in a database.

That means we are in a position to analyze apparent voting pattern across the country. And there are some glaring anomalies in the voting. 

None of these is more apparent than the case of voting in El Paraíso, Copan, where the voter turnout, according to the TSE numbers, was 85%. That would be exceptional participation, compared to previous Honduran elections, and is far above the average levels of participation in this election.

El Paraíso is an interesting place. Its Mayor, Alexander Ardon, is widely considered to be a member of the Sinaloa cartel in Honduras. In a report on organized crime in Guatemala, Honduras, and El Salvador, the Wilson Center reported:
Honduran police intelligence says that El Paraíso, Copan Mayor Alexander Ardon works with the Sinaloa Cartel.  Ardon has built a town hall that resembles the White House, complete with a heliport on the roof, and travels with 40 heavily armed bodyguards.  Cameras monitor the roads leading in and out of town, intelligence services say. And there are reports that the Mayor often closes the city to outsiders for big parties that include norteño music groups flown in from Mexico.

One correction: it's not just norteño music: it's narcocorridos that characterize the bands flown in to El Paraíso.

Voting was definitely distorted in El Paraíso on November 24th. 

One election monitoring group reported that 50 election workers from out of town staying in a hotel were locked in the night before the election, surrounded by over 100 armed men who threatened to shoot them if they tried to leave the hotel and assume their poll duties.  Another such group got out of the hotel, but were stopped along the way by armed men, who slashed their tires and told them if they continued onward, they'd be killed.  The group in the hotel was freed late in the day on Nov. 24 by Officer Erazo Mejia of the police, but were later locked up in another place and their election credentials stolen.

Adrienne Pine has an account from one of these poll workers whose tires were slashed.  The group waited until 4:40 AM for police to come and help them obtain new tires, then drove on to their polling places.  This group were representatives of LIBRE.  They were met at the polling place by a group of armed men who controlled the location, not the military, who they report just stood around.  The LIBRE group was told it was not welcome there.  Shortly thereafter there was an altercation between the local police and the armed group, which kicked out the police, then confiscated the identity cards and election credentials of all the LIBRE workers, before kicking them out.  The armed group tried to get the LIBRE representatives to sign the ballot tally forms, at gunpoint, but at least some still refused, and all were thrown out of the polling place.  At the doorway to the polling place they saw individuals questioning voters before they entered, and they report that if the prospective voters were not voting for Juan Orlando Hernández, they were not permitted in to vote.

LIBRE formally asked the TSE to invalidate the tally sheets from those precincts, but the TSE counted them.  You can see one such acta here

Here's how to read it.  The polling station was issued 320 ballots.  That means that there are 304 citizens eligible to vote there, plus up to 16 party representatives (which adds up to 320 ballots).  In the vote count, they reported 308 valid votes, plus 16 nullified ballots, which is 324 ballots, four more than they were issued or should have needed.

Not only did the TSE count this acta, signed by 4 poll workers and two alternates (none of them LIBRE members): it ignored the over-vote. 

This means that somehow, in the TSE's vote counting software, the check for more people voting than ballots issued either isn't implemented properly, isn't implemented at all, or someone at the TSE ignored the system to OK this acta

Given the OAS report of its audit of the vote counting software, it is not unlikely that an over-vote test either wasn't implemented, or didn't work. 

This tally sheet was not even subject to special scrutiny (escrutinio especial), or it would have been issued a new, less informative vote count sheet, looking like the one found here for acta 314. 

Another example from the same region is MER 2670.  They were issued 377 ballots to cover the 361 eligible votes plus 16 party representatives.  They report the total ballots cast as 348. But the individual candidate's votes, plus blank and null votes, add up to 404!  This would mean they had a voter turnout of 112%!

MER 2687 should have qualified for special scrutiny as well.  Actas with similar mistakes were scrutinized and replaced elsewhere, but not in El Paraíso for some reason.  The acta for MER 2687 reports 320 ballots issued for its 304 eligible votes plus up to sixteen party representatives.  Six people signed the acta and on the line where they report how many people voted, they reported six. This acta, in reality, reports on 304 votes, including blank and null votes. Six of those votes were presumably by poll workers, so the effective turnout is 98%.

The TSE accepted "6" as the count of voters who voted and used it in counting how many citizens voted, which means the numbers it reports for this MER are false.  Apparently the software does not check that the number of votes is equal to the number of people who voted, a required sanity check for any such vote counting system.

MER 2691 also shows anomalies. They were issued 222 ballots for 206 eligible voters plus up to 16 poll workers.  In total, 210 votes were reported cast, and six of those were by poll workers.  That means that only 2 of the precinct's eligible voters supposedly didn't vote. That would be extraordinary.

MER 2693 shows the same issue.  They were issued 179 ballots for 163 eligible voters plus up to 16 poll workers.  They report 166 votes, and only 5 poll workers voted, meaning that only 2 of their eligible voters didn't vote.  The same can be said about MERs 2711, 2712, and 2713.

Voter turnouts in the 80-100% range should be suspect.

No poll watchers reported such extremely high turnouts anywhere in Honduras.

But in the 47 MER that comprise El Paraíso, 51% (n=24) reported voter turnout above 90%.

Only 36.1% (n=17) MER in El Paraíso reported a voter turnout of less than 85%.  If we set the threshold lower, to 75%-- still a great turnout level for Honduras-- then only 27% (n = 13) had voter turnouts less than 75%.

These are extraordinary levels of voter participation.

Extraordinary-- and suspect under any circumstances. But not troubling to the TSE in Honduras.


RAJ said...

Thanks to Jake Johnston for pointing us to his December 6 post noting the same suspicious pattern of exceptional voting in El Paraíso. The project of assembling the election database-- undertaken in part to ensure that the data won't disappear-- has occupied us since the election and we missed his excellent post, which hits the same points as ours.

The polling place by polling place information shows that the aggregate vote for El Paraíso that Johnston reported-- already suspicious-- derives from individual, incredibly unlikely voting results that should not have passed sanity checks, had the TSE been paying as much attention as we, and Johnston, were.

Ardegas said...

Is Juan Hernández a legitimate presidente-elect or is he not? What do you think?

RAJ said...

Legitimacy is an interesting concept. We can either take it in terms of official procedures-- in which case, the TSE, backed by other governments' recognition of Juan Orlando Hernández, has the last word; or we can ask what the populace in general thinks. Here, for Hernández, the issue is that even by the TSE's numbers, almost two thirds of the country voted against him.

Note that this would be the case no matter who was declared the winner in this election. Because of the combination of the fragmented political landscape, unequalled in Honduran electoral history, and the structure of the electoral system (predicated on a two party system in which the person with the higher vote total would normally be somewhere closer to a majority), winning this election never promised widespread popular support. There is no "mandate" here.

But I suspect you mean "legitimate" here in a slightly different way: did JOH actually gain the number of votes the TSE claims, in a free and fair election? Our answer to that will be forthcoming, as we continue the laborious but necessary task of tracing what actually happened in the vote count. Right now, we would say that there are indications of many irregularities in voting. These may not have been enough to throw the election, although they are remarkably consistent in the direction they lean. But they do demonstrate that Honduras did not experience the kind of electoral transparency that some parts of the international community have claimed.

Nor is this surprising, since the election came in the midst of a series of violent attacks on party members and candidates from the new opposition, attacks that continued through the evening before the election, and resumed immediately after the international observers left.

If we want to return to the concept of "legitimacy", and this time treat it as a moral claim-- to be the leader of a population that has the freedom to make informed choices without fear or favor-- then there is no way that any candidate who would have won in Honduras could be considered to be entirely legitimate. Buying votes openly, or through false policy commitments; intimidating voters and opposing parties; spreading propaganda instead of engaging in debate-- these are highly relevant factors in the election just passed. And none of these are absent in the elections of countries that act as if they can judge Honduras.