Costa Rican Free-Trade Beancounting: Preliminary Cases

“Workers required by employers to record themselves voting yes, despite ban on photographic equipment in voting booths,” says NO campaign. That is, “prove you voted yes, or lose your job.” Documented cases, but no investigation into exactly how frequent or common this practice might have been that I have found yet. In Brazil, this is known as the voto do cabresto. Nullification petitions in such cases, and TSE resolutions here. The TSE has now officially ratified the result. Studying.

TSE has virtually eliminated fraud with citizen identification cards that incorporate electronic images and biometric information. –Unisys customer case study for Costa Rica’s TSE

The political and black-propaganda aspects of the affair aside, the Costa Rican free trade referendum presents us with some beancounting lore that has some intrinsic interest, qua beancounting, I think.

The Costa Rican elections tribunal, the TSE, has now ruled on cases of alleged fraud and irregularities and certified the results of the referendum. All complaints were filed by election observers of the PAC, the party of Otton Sólis, a leading NO proponent.

NO proponents are saying that fraud, vote-buying, coercion, and other corruptions of the process were rampant, while oversight was nonexistent and the TSE intransigent in its refusal to properly investigate such allegations.

On which more later. The University of Costa Rica statistics department’s report on the 2006 election concluded — phrased more diplomatically, of course — that the nation’s elections system had failed once again to produce fully credible results. Although, weirdly, it seemed to want to praise it for making progress toward generating less implausible results.

The publication of widely disparate poll results prior to the voting — one or more of them indicating a “technical tie” — is often correlated with, and show be taken as one risk indicator for, election fraud, according to experts I have been reading. And that is exactly what happened in this case, throwing a flag for a more detailed audit of the process.

The poll that showed a technical tie was prepared by a firm that lists the presidency of Costa Rica — and Chiquita Brands — as market research clients. The poll that showed a 12% advantage for NO does not list its clientele, but says it is part of the Research International network.

So anyway, I am starting to read some of the cases published by the Costa Rican TSE to see if there is anything in there that might advance the the global state of knowledge on how to count things — such as cookies, ducks or votes — beyond the traditional “see and say” approach propounded by Count von Count.

Observers of the 2006 national elections in Mexico may experience a sense of deja vu, I think.

I wish we could convince UNAM’s John Ackerman, who authored some of the clearest analysis of the abject failure of beancounting controls in the Mexican election, and was a visible spokesman for the “vote by vote” movement there, to turn his attention to this case.

En escritos recibidos en la Secretaría de este Tribunal el 18 de octubre del 2007, el señor Edgar Ruiz Cordero, en su condición de fiscal del Partido Acción Ciudadana, interpuso demandas de nulidad de las Juntas Receptoras de Votos número 3815, 3945, 3999, 4000, 4034, 4040, 4077, 4136, 4378, 4543, 4545, 4548, 4558, 4559, 4691, 4733, 4803, 4886, 4890, 4904 y 4910, en cuyas actas de escrutinio se consignó que el Padrón-Registro aparecía en blanco, y contra las Juntas Receptoras de Votos número 2956, 3183, 3190, 3284, 3434, 3779, 3833, 3867, 3895, 3913, 3940, 4006, 4290, 4351, 4368, 4449, 4451, 4484, 4525, 4553, 4555, 4563, 4592, 4602, 4628, 4648, 4678, 4688, 4812, 4871, 4862 y 4920, en cuyas actas de escrutinio se consignó que el acta de cierre del Padrón-Registro aparecía en blanco.

In papers received by the clerk of this elections tribunal on October 18, 2007, Mr. Edgar Ruíz Cordero, in his capacity as elections monitor for the Citizen Action Party (PAC), demanded the nullification of Vote Reception Boards Nos. 3815, 3945, 3999, 4000, 4034, 4040, 4077, 4136, 4378, 4543, 4545, 4548, 4558, 4559, 4691, 4733, 4803, 4886, 4890, 4904 and 4910, in whose documentation it was observed that the voter-registration rolls were missing [or blank]; and against Vote Reception Boards Nos. 2956, 3183, 3190, 3284, 3434, 3779, 3833, 3867, 3895, 3913, 3940, 4006, 4290, 4351, 4368, 4449, 4451, 4484, 4525, 4553, 4555, 4563, 4592, 4602, 4628, 4648, 4678, 4688, 4812, 4871, 4862 and 4920, in whose documentation it was noted that the certification of the closure of the voter-registration rolls was missing or blank.

Alega que es imposible verificar que el número de votos escrutados corresponden al total de votantes que se presentaron, ya que ese aspecto se verifica en el Padrón-Registro lo que, en su criterio, provoca un vicio insubsanable, pues las papeletas escrutadas no pueden ser prueba de la manifestación de la voluntad del elector, en tanto no existe constancia documental de que los ciudadanos inscritos en el Padrón se presentaron a ejercer el voto, pues esa circunstancia se consigna en el Padrón Registro con la firma de cada ciudadano. Agrega que, una situación como la denunciada, provoca la nulidad de toda la junta, al no poder verificarse cual fue la manifestación de la voluntad popular del elector.

He alleges that it is impossible to verify that the number of votes counted corresponds to the total number of voters who presented themselves to vote, given that the voter-registration rolls are required in order to perform this check, and that in his judgement this creates an incurable defect in the process, because the ballots counted cannot be taken as reflecting the will of the voter unless it be documented that the citizens registered to vote actually presented themselves to vote, a fact that is reflected by the signature of each voter in the voter-registration rolls. He adds that a situation like the one charged must lead to the nullification of the entire results of that board, since it cannot be verified that the vote count corresponds the popular will.

A recent survey by the elections authority in Brazil showed a number of municipalities with a number of registered voters equal to, or exceeding, their current population. Voters who have moved but failed to update their credential is the leading explanation given, but there are others that need to be audited for before being ruled out.

The elections in the state of Alagoas last year produced a very signficant discrepancy between the total number of votes counted and the number of people recorded as having voted.

So what did the TSE rule?

En el caso concreto, a pesar que el señor Ruiz Cordero estima que consignar en el acta de escrutinio que el Padrón-Registro se encuentra en blanco (tratándose, según se indicó, del acta de cierre de dicho documento electoral) es motivo de nulidad de la respectiva Junta Receptora de Votos, se debe indicar que el artículo 32 del Código Electoral establece que el Padrón-Registro es plena prueba del resultado de la votación, pero dicha norma también establece una excepción en cuanto a su valor probatorio, quedando sujeto a que no aparezca otro documento de igual valor.

In the present case, though Mr. Ruiz believes that the attested fact that the voter-registration list was blank [or missing] (dealing, as he notes, with the certification that this document has been closed) provides grounds for nullifying these Vote-Reception Boards, it must be noted that while Article 32 of the Elections Codes does provide that the voter-registration roll is full proof of the result of voting, it also establishes an exception with respect to its probative value, subject to the absence of another document with equal probative value.

Incluso, el párrafo segundo del numeral 32 antes citado, confiere a las certificaciones emitidas por los miembros de la Junta “igual valor probatorio”, de suerte tal que el hecho que el espacio correspondiente del Padrón Registro se encuentre en blanco o incompleto, por sí mismo no supone un vicio que provoque la nulidad de la totalidad de la Junta en los términos del artículo 142 del Código Electoral, toda vez que en los restantes documentos que complementan el material electoral de cada Junta Receptora de Votos, existen otros instrumentos con igual valor probatorio que el Padrón-Registro, por ejemplo la certificación del resultado emitido por la Junta (artículo 121, inciso k) del Código Electoral) y el resto de la documentación electoral (mensaje de transmisión de datos, constancias de los sobres que contienen los votos y, desde luego, las papeletas utilizadas y sobrantes) que permiten verificar cuál fue el resultado de la votación y, por ende, la voluntad popular.

In the second paragraph of Article 32, the certifications issued by members of the Junta are afforded “equal probative value,” such that in the event that a blank or incomplete voter-registration roll in itself does not vitiate the process to the extent that the entire result from that vote-reception board must be nullified under Article 142 of the Code, given that other documents, complementing the documents issued by each Vote-Reception Board, exist, with equal probative value to that of the voter-registration roll — for example the certification of the result issued by each Board (Article 121, Paragraph (k) of the Code) and the rest of the documentation (message of transmission of data, list of contents of the envelopes containing votes, and, later, the ballots used and unused) to permit the verification of the results of voting and, therefore, the popular will.

Petition denied.

Maybe this is an ingenous question, but how can the precinct officials credibly certify that the number of voters who voted equals the number of votes cast if they lack the very information, required to verify that fact?

You give me $20 to buy you pencils.

I return with 10 pencils and $10 in change, but without a receipt to show that each pencil cost $1.

But I do give you a document certifying that each pencil did cost $1. Bearing my signature.

Does that document have the same probative value as a store receipt?

My word is my bond.

Why should you believe me?

Just because I say so.

Is that not what we are talking about here? If the results are certified, but the person who certified them does not have to give adequate reasons for the decision to certify, is the process adequately transparent?

Somebody explain this to me.

As I noted before, between the electronic quick count and the hand-counting of the ballots in the FTA referendum lay a process of several weeks, during which the physical ballots had to be transported to the capital and counted by hand.

As a sign of the transparency of the process, the TSE broadcast the vote-counting live.

More relevant would have been to track the chain of custody of every single ballot between the time it left the polling station to its arrival at the national vote-counting center in San Jose.

I wonder what the expected, and the actual, margin of error in that process was?

The stated differential between preliminary count and final count: 0.00%.

Preliminary count indicates 51.6% YES.

Final count indicates 51.6% YES, though with a different (higher) number of votes counted.

I wonder what risk management and security procedures were employed to minimize inaccuracies to that astonishing degree? I mean, it is astonishing, isn’t it?

Does anyone have any cases for comparison in which a quick count was 100.0000000% accurate?
Process handoffs in Costa Rican electoral tribunal quick-count system. Votes (1) transmitted by telephone to regional centers, where the data is (2) manually entered for transmission to the TSE, via both private network and Internet, and from there to (3) the news media, via Internet. Network and telephone lines monitored by two private firms, ICE (telephony) and RACSA (ISP). A manual count is taken to confirm the quick-count, once (5) the ballots are physically delivered to the TSE in San Jose. White-hat hackers: Devise the simplest possible plan for defrauding this system. What occurs to me is that you (1) mess with the preliminary results during data entry or transmission, and then (2) mess with the ballots during transportation to make the results agree with the quick count. This is how the mapaches did it in Mexico. A very rigorously documented chain of custody on the ballot documents — from voter sign-in to the final vote-count — would therefore be paramount. In Mexico, cases of sealed ballots tampered with while in IFE custody — the classic mapache scenario — were documented.

The image “” cannot be displayed, because it contains errors.
Sesame Street’s Count “The Count” von Count counts snowflakes. Source: Wikipedia.

Just some first thoughts on this. Still studying. Our books have just arrived from Santos and need to be shelved now.


Leave a Reply

Please log in using one of these methods to post your comment: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s