The Election in Mexico

On June 5, 2006, Mary Anastasia O'Grady proclaimed "Democracy Lives" in her opinion column in The Wall Street Journal:

Felipe Calderón of the National Action Party hasn't any doubt that he won Sunday's presidential election, and he says that his adversary, Andres Manuel López Obrador of the Democratic Revolutionary Party (PRD), knows it, too. In an interview with The Wall Street Journal at his campaign headquarters yesterday, Mr. Calderón appeared rested and confident. "I'm not going to get into personalities," he told me, "but all the parties have copies of the tally sheets showing the voting, and the PRD knows that it lost."

Mr. Calderón's numbers jibe with those of Mexico's Federal Electoral Institute (IFE) and it seems almost certain that he won. But as we go to press, an official announcement has not yet been made. A preliminary ruling is scheduled for today, with the official decision to come on Friday. And even then the country may be in for weeks of the Mexican equivalent of challenges to hanging chads.

... The Sunday vote was an amazing electoral symphony, featuring thousands of small polling stations all over this sprawling metropolis, in private homes, schools and community centers. I passed by or visited a mere fraction of them. Yet my impressions were confirmed by wider press reports and independent observers. Everywhere I went, from well-to-do districts in the north and south to the popular neighborhoods downtown, the scene was one of saintly patience and First World order. A summer downpour didn't faze voters one bit. They covered their heads with newspapers, shared umbrellas--and waited.

... As polls closed around the country on Sunday evening, poll workers from each party signed off on tally sheets and sent the totals to the IFE. There was a marked absence of complaints from any polling station about irregularities, but there was a surprise. There had been wide expectations that Mr. López Obrador would have the early lead, since this city--which is his stronghold--was expected to report first. Instead, the northern state of Coahuila seems to have reported ahead of the national capital, and so, from the earliest tallies, Mr. Calderón had the lead. It was a lead he would never lose.

The IFE had worked hard to make all this happen and it wasn't about to commit a misstep. IFE president Luis Carlos Ugalde announced at 11 p.m. that Mexicans would have to wait until Wednesday morning for a winner. Still, the transparency of the process was working against Mr. López Obrador, and as polling stations reported--and tallies were posted on the Web--it became clear that Mr. Calderón had won by a whisker. On Sunday evening, the head of the European Union observer team attested to the fairness and transparency of the election.

Mr. López Obrador smelled defeat and swung into action with Plan B. Just after Mr. Ugalde asked the candidates not to declare victory, the PRD candidate jumped in front of the cameras and told all of Mexico that he had won, adding that he had "information" from the "quick count" showing that he was up by 500,000 votes.

It is doubtful that he had any such information. Rather it looked like classic López Obrador, working to arouse suspicion among his supporters that the establishment had carried out "a plot" against them. Arturo Sarukhan, campaign adviser for the PAN (Mr. Calderón's party), told me that his side had no choice but to respond: "We were standing in the 'war room' at campaign headquarters, discussing how to calibrate Felipe Calderón's message in light of the request that Mr. Ugalde had made when we heard him [Mr. López Obrador] claiming a lead of 500,000 votes. That's when we decided we had to report the independent exit polls." Mr. Calderón appeared shortly thereafter on television, rattling off a string of independent exit polls that showed him in the lead. 

Mr. López Obrador now claims that there are three million "lost" votes, and that he senses all kinds of "irregularities," none of which are backed up with evidence. While all the votes are not yet in the official tallies because of a technicality in reporting, the Calderón team remains confident that they are accounted for in the totals that all parties now have, and that the outcome will not change.

Is this all over already?  Chuck Collins wrote in AlterNet on the same day:

On a morning interview I did on Texas statewide commercial radio, the host started the program by saying: "In a cliffhanger election, the conservative candidate Felipe Calderon has beaten the left-wing firebrand Lopez Obrador by 1 percent. But Lopez Obrador is demanding a recount and threatening street protests."  Later, CBS Radio called to set up an interview to discuss the "Calderon victory and Lopez Obrador's street protests."

Whoa. Time out! Before we allow Rupert Murdoch to call the Mexican election, let's cover the basics here:

Mexico's election is indeed a cliff-hanger, but there are no official results. Mexico's Federal Election Institute (IFE) has indicated that its preliminary computer tallies, which gave Calderon a 1 percent lead, were insufficient. On Tuesday, it admitted that 2.58 million additional ballots still need to be counted. So let's repeat: There are no official results or official count estimates.

The Federal Election Institute (IFE) had hoped to announce the winner on Sunday night using a sophisticated system of sampling from around the country, known in Mexico as "PREP." This system of compiling and releasing preliminary, unofficial election results has come under fire for causing confusion and potential unrest.

Jonathan Roeder reported Wednesday in Mexico's daily El Universal that "despite repeated reminders that PREP results are unofficial, Calderon has claimed they show he is the clear winner and that Lopez Obrador should step aside, while Lopez Obrador has highlighted the system's inaccuracies to suggest the contest was rigged against him."

Lopez Obrador has not yet suggested there is fraud, but expressed concern about uncounted ballots from 16,000 polling areas that were not included in the PREP calculations. The head of the IFE, Luis Ugalde, confirmed that 2.58 million votes were set aside because of "irregularities or inconsistencies" in addition to an estimated 1.5 million uncounted ballots.

Here's how things were supposed to work: 

First, on Sunday, after polls closed at 6:00 p.m., ballots would be counted on location.

I witnessed this remarkable event in the city of Miahautlan, Oaxaca, as election officials opened up ballot boxes on the city's central plaza. After separating the ballots into piles for each candidate, election observers from each party and election officials counted the ballots together, out loud. Visualize 20 voices together saying "Treinta y cinco, treinta y seis." Unfortunately, such transparency still doesn't exist in every town in Mexico, especially in regions controlled by the Institutional Revolutionary Party (PRI), which governed Mexico for 71 years prior to the 2000 election.

Second, the count from each polling area was put onto a tally sheet that all witnesses signed. IFE staffers in a sample of polling areas called their results into IFE headquarters for the PREP, a scientific sampling could have declared a possible winner. The IFE never released the results of their "quick count" because it was too close to call.

Third, the ballots were bundled and sealed in special packets and transported, usually under armed guard, to one of 300 IFE district offices. At this moment, that is where the sealed ballot packets are.

Over the ensuing hours, more tally sheets were reported into the central IFE office. But by 11 p.m. Sunday, the election was still too close to call using the PREP system. Both leading candidates, with their own exit polls and results called in by election workers, had numbers showing their lead.

The process of counting ballots begins today, Wednesday, at the local district offices. The winner, according to IFE officials, could be announced before Sunday. But the process could require more time, depending on the number of ballots counted by hand.

Regional officials will first review tally sheets, computer results and polling data from over 130,000 polling areas. If there are formal complaints about specific areas, they will open sealed ballot packets and do hand counts. The good news is Mexico's next president won't take office until Dec. 1, and the IFE is not required to legally certify the election until Sept. 6. Everyone is counseling calm.

The U.S. media, however, have seized upon the Calderon 1 percent lead number and begun coronation proceedings. They've implied that it's all over, except "firebrand" Lopez Obrador and his street mobs won't concede. This is highly irresponsible.

The U.S. media should butt out of the Mexican election until it can get its facts straight on the actual process and the historical context of Mexico's evolving democracy.

On June 6, 2006, Mark Stevenson reported in Associated Press:

Felipe Calderon's slim lead in Mexico's presidential race was called into question as an official recount of vote tallies showed his leftist opponent slightly ahead late Wednesday. However, Calderon insisted he had won and even offered to include his rival in his Cabinet.

With nearly 95 percent of tally sheets recounted at 300 district headquarters across the country, former Mexico City Mayor Andres Manuel Lopez Obrador had 35.9 percent, compared with 35.3 percent for Calderon. The preliminary count completed earlier in the week had Calderon winning by 1 percentage point.

Officials from Calderon's party said Lopez Obrador was only leading because more votes had been recounted in areas where he was strongest, and they insisted the trend would not hold.  They also accused the Lopez Obrador's party of stalling tactics in states where the conservative Calderon was strongest, saying it was deliberately trying to give the impression that Lopez Obrador was ahead as the count progressed.

Earlier on Wednesday, Calderon told the Associated Press in an exclusive interview he would be willing to include Lopez Obrador in his Cabinet — an effort to build a coalition government and avoid weeks of political impasse.  "Mexico needs us all," Calderon said. But he said he did not believe Lopez Obrador would accept, adding that the two men had not spoken to each other since Sunday's election.

Lopez Obrador insisted he was victorious and said there was "serious evidence of fraud."  Leonel Cota, president of Lopez Obrador's party, accused election officials of deliberately mishandling the preliminary count to confirm a win for Calderon, the ruling-party candidate. He said Lopez Obrador won Sunday's vote.  "We are not going to recognize an election that showed serious evidence of fraud, that was dirty from the start, manipulated from the start," he said.

When polls closed, citizens staffing the 130,488 polling places opened the ballot boxes and counted the votes, then sealed them into packages with their tallies attached and reported unofficial totals to the Federal Electoral Institute, or IFE. The institute then posted preliminary results on its Web site from about 41 million ballots cast.  The sealed packages were delivered to district headquarters, where elections workers used the tallies Wednesday to add up the formal, legal vote totals.

Workers were not reviewing individual ballots except when the packages appeared tampered with or their tallies were missing, illegible or inconsistent — including at least 2.6 million ballots likely to shrink Calderon's lead to 0.64 percent if included, election officials said Tuesday.

At one electoral office in Mexico City, officials opened a ballot box because the vote tally was missing. The votes were then re-counted out loud while 10 party representatives stood by with tape recorders and video cameras.  "I'm exhausted. I'm still tired from election day," said counter Rocio Sanchez, 41, an IFE employee. "But this is something we have to do by law."

Cota said Democratic Revolution would not recognize the results without a ballot-by-ballot recount. But IFE President Luis Carlos Ugalde said that was not possible.  "Mexican law is very clear on when a ballot box can be opened: only when there are problems with the vote tallies, when the tally sheet has obviously been changed, or when the box has been tampered with," Ugalde said.

Once the count is complete, the seven-judge Federal Electoral Tribunal hears any complaints and can overturn elections. By law, it must certify a winner by Sept. 6, and its decision is final.

Where is the "serious evidence of fraud" that Lopez Obrador insists that he has but O'Grady says does not exist?  Look at the nice chart from Narco News.  This is selective release of results, and not necessarily fraud as yet.  But it may be construed as intended to mislead public opinion.

The first maneuver was transparent enough: IFE began its online preliminary count on Sunday by selecting result estimates mainly from the Northern Mexican states where Calderón won the vote. This was evident on the IFE PREP results because they were listed state-by-state and also by the five electoral regions of Mexico. Less than an hour after polls closed – at 6:57 p.m. – the earliest IFE results claimed Calderón had 40.87 percent to just 33.69 percent for López Obrador; a difference of more than seven percentage points. But look at this “photograph” of that moment in time.

“Circunscripción 2,” circled in red, represents the northern region of Mexico – where all polls showed to be Calderón’s strongest support and López Obrador’s weakest – and although it represents only 20 percent of the country’s population, IFE led with its numbers as 40 percent of its preliminary vote total. The region includes three states in the pacific time zone where polls had yet to close and the rest of it is far from IFE headquarters in Mexico City. It is the most geographically disperse region, too, making it a slower process to get the results in to Mexico City. But IFE doubled its statistical influence from this region in the first hour to simulate a false impression that Calderón was far in the lead.

And so it went, all night long. With each and every update, IFE selectively released the vote tallies in a manner that kept Calderón in the lead. This is statistically impossible to do with a final tally of 0.6 percent difference between the two candidates (such a process with a close vote would, if reported as results came in, always show the tally tipping back and forth from one candidate to the other) – unless the results were being rationed selectively.

As Sunday night marched into Monday morning, López Obrador closed the gap. Until 70 percent of the preliminary results were tallied, the López Obrador vote rose in a straight and steady line (see red line on chart). But suddenly, with between 70 and 80 percent of the preliminary results tallied, the trajectory that would have put López Obrador in the lead when less than 90 percent of the votes had been counted, took a swift downturn, exactly corresponding to a swift upturn by the third-place candidate Roberto Madrazo of the Institutional Revolutionary Party (PRI), represented by the green line on the chart.

Defenders of IFE explain this suspicious turn of events as one of “the rural precincts coming in last,” where PRI was expected to do better. But rural district voters constituted the weakest vote for Calderón across the country; he came in a distant third among rural farmers. As the line graph shows, in that latter stage of the posting of PREP results, Calderón’s vote trajectory continues to run in the same downward straight line that it had traveled since early in the night. Had these truly been numbers from rural districts, his trajectory (the blue line) would have dipped significantly farther below.

On Monday, IFE closed its preliminary results, claiming that it had counted 98.5 percent of the precincts. With Calderon up by 377,000 votes (about 1.4 percent) it seemed to the casual observer that his lead was insurmountable. The problem is, the IFE did not tell the truth. Only about 92 percent of the preliminary tallies had been included in that count, leaving 3.3 million votes out of the count. The claim – posted with the PREP results – that 98.5 percent had been counted was knowingly false. It was intended, as has every step taken by IFE in the vote counting, to create the false impression of a clear lead by the candidate of the State, Felipe Calderón.

This was a blatant act of tampering with the PREP results by Ugalde and IFE officials. As IFE’s own website says: “La alteración de estos resultados es delito federal.” That means, “Alteration of these results is a federal crime.” The selective withholding of those results on Sunday and since – again, and again, and again – constitutes multiple counts of what ought to be a criminal charge against those IFE officials responsible for withholding the tallies and also for falsely claiming that 98.5 percent had been counted and included in the final PREP tally when they knew it to be false.

Additional information at: A Full Recount Would Show that López Obrador Won Mexico’s Presidency by More than One Million Votes by Al Giordano, NarcoNews, July 8, 2006; Death by Video: Mexico’s Election Fraud Is Coming Undone by Al Giordano, NarcoNews, July 8, 2006.

(Salon.com)  Mexico 2006: Florida all over again?  Eliza Barclay.  June 11, 2006.

"Ciberfraude," or cyberfraud, is not a word in the average Mexican's vocabulary. But most Mexicans have heard of the extraordinary electoral debacle that befell their neighbors to the north in 2000, and Martí Batres, the head of the Democratic Revolutionary Party's Mexico City chapter, was going to capitalize on that knowledge. At a press conference Friday afternoon at PRD headquarters, the close advisor to Andres Manuel Lopez Obrador, the leftist politician who lost an exceedingly close election here this week, nodded to an aide to turn on his laptop. "And now I'm going to show you a video," he told the roomful of reporters.

The lights went out, and on a pull-down screen, computer programmer Clinton Curtis explained in English to an American audience how he had allegedly been hired by Tom Feeney, speaker of Florida's House of Representatives and a Republican, to create a computer code that fixed that state's vote in favor of George W. Bush six years ago. Batres, or someone on YouTube, had added Spanish subtitles to the 3 and a half minute clip. "This video shows that cyberfraud is possible," Batres insisted when the lights came up. "There may have been a source code used to manipulate our elections just as with the Florida elections in 2000." 

Rumors of fraud were swirling in Mexico's streets, on TV and on blogs long before Thursday's official count confirmed that Felipe Calderon of the conservative PAN party had beaten Lopez Obrador of PRD by .58 percent, a difference of just 236,000 votes out of 42 million cast. "A black hand was at work, I believe," says Jorge Ortiz, a taxi driver who voted for Lopez Obrador. "The numbers just don't make sense."

Some of the rumors were reminiscent of the bad old days when the Institutional Revolutionary Party, or PRI, ran Mexico without challenge and elections were rife with fraud and ballots were burned in bonfires. There were tales of 3 million votes missing from the preliminary tally issued early in the week, which had Calderon winning by more than a million votes. There were allegations of vote buying and ballots buried in a Mexico City garbage dump.

But the fraud claims made at the PRD press conference were decidedly 21st century and very American. Batres, head of the PRD's Mexico City chapter, detailed several instances where the votes reported by the government's preliminary tabulation system, called the PREP, did not match the actual voting record, always to the deficit of Lopez Obrador and the benefit of Calderon, in one case by as many as 3,828 votes. 

According to Batres, the inconsistencies cannot be chalked up to human error or deliberate destroying of paper votes, but to conspiracy, to a source code like the one Clinton Curtis claimed to have designed in Florida that systematically moved votes from the PRD to the PAN.

"In many different states around the country, we have seen what we believe to be cybernetic manipulation," Batres insisted. "We are going to enlist the help of information crime experts to look for a code inside the [electronic tabulation] system and then we need a recount, vote by vote."

Batres did not specify whether the PRD believes the entire election was manipulated or if the numbers were tweaked for only a select number of polling sites. The party will make a formal presentation to the Federal Electoral Tribunal contesting the validity of the election as soon as Saturday morning. As of Friday night, the party had not yet announced which tack it would take, but there were three likely scenarios.

1. The PRD could claim that certain polling sites were manipulated and call for the reopening of the ballot boxes at those sites for a vote-by-vote count.

2. The PRD could claim that the entire election should be annulled because cyberfraud was widespread and affected every polling site.

3. The PRD could claim that the entire election should be annulled because of cyberfraud and because of other factors. 

Chief among those "other factors" is the annulment of more than 904,000 votes, 2.16 percent of all votes cast. While discussion of cyberfraud sounds as much Ohio 2004 as Florida 2000, it's annulment that really inspires a Tallahassee flashback.

Explains University of Texas political science professor Kenneth Greene, who has been in Mexico as an observer, "This election had the 'overvote' factor, similar to the butterfly ballot issue in Florida where people punched more than one option. Here people sometimes marked more than one box and so the IFE didn't know who to count it for and annulled it."

The Federal Electoral Institute, known by its acronym IFE, is controlled by appointees from Calderon's party, PAN, just as the chief vote-counter in Florida, Katharine Harris, was a member of George W. Bush's party. In fact in Mexico, the IFE has even tighter control than Harris had six years ago. "It's parallel to what happened in Palm Beach except that in Mexico only the IFE can count the votes, unlike in Florida where the lawyers and the party members got involved with counting."

Once Lopez Obrador presents his formal complaint, he’s not likely to have any better luck. The Federal Electoral Tribunal, which will review the complaint, is also packed with PAN and PRI affiliates.

PRD officials are unlikely, therefore, to persuade anyone in power of their arguments. They seem confident, however, that they can convince the public that something untoward happened in this election. Expert observers are skeptical. 

Victor Manuel Alarcón, head of the sociology department at the Metropolitan Autonomous University-Iztapalapa, says the errors were likely human, a common occurrence with any non-digital voting system. "A lot of the people tallying the votes were poorly educated -- remember this country is poor -- so it's very possible the discrepancies between the PREP and the official count were their fault, not the result of fraud."

Despite a history of vote fraud, it's also difficult to reconcile the claims of manipulation with Mexico's revamped electoral institutions and voting procedures. While Mexico's electoral system has remained technologically simple, with voters using fat black crayons and paper instead of touch screens, it is also one of the most expensive and labor-intensive in the world. The government spent $1.2 billion this year in part to ensure that no polling site would have more than 300 voters assigned to it. In the 1990s, procedures were changed so that no party officials would work at polling sites. Parties are allowed to have representatives observing the procedures, but the sites themselves must be manned by ordinary citizens to create accountability.

According to Ulises Beltran, a professor of political science at CIDE, a leading graduate research institution in Mexico City, organized fraud is virtually impossible in contemporary Mexican elections because of all the safeguards. "To find enough evidence of fraud for Lopez Obrador to win," explains Beltran, "there would have to be 50,000 citizens involved with the conspiracy. The small size of our polling sites and the large number of citizens working in them should really prevent it."

Still, no matter how plausible or implausible the PRD's case turns out to be -- and enlisting the questionable Clinton Curtis on their behalf is hardly convincing -- 65 percent of the public voted for someone besides the election's declared winner. Some, at least, are likely to remain dubious about Calderon's victory. Cuahtemoc Cardenas, the founder of the PRD, who could make a much more compelling case that he was robbed of the presidency in the controversial election of 1988, argues that all questions must be answered for the sake of Mexico. "Without even being at fault," wrote Cardenas in a Friday editorial, "those who resist and oppose the clearing up of doubts awake unnecessary suspicions." 

Mexicans are especially sensitive and alert to tampering in elections, because they have probably more experience than any other people in the world.  The best book on the subject is Opening Mexico: The Making Of A Democracy by Julia Preston and Samuel Dillon.  This book might be characterized as a 600-page manual on how to commit vote fraud as practiced by the best in the world for an unbeaten string over half a century.  Here is what happened in Chihuahua state in July 1986:

July 6 was a Sunday.  Ortiz Pinchetti and Aziz met early at the Hotel San Francisco, guzzled some coffee, and headed out to watch the balloting.

They took a cab to a proleterian neighborhood with dirty streets on the outskirts of the capital.  There, at polling place 95-B, a piece of political theater unfolded.  The PAN's poll watcher, Lourdes Cerecero, twenty-one years old, had arrived minutes before voting was scheduled to being at 8 a.m.  To her surprise, the other poll watches and authorities -- all from the PRI -- told her that voting had begun half an hour earlier, before the legal voting hour.  They shrugged off her protests and refused to let her examine the corrugated-cardboard ballot boxes, even though a document on the polling table indicated, falsely, that she had helped inspect them before the balloting began.

At 8:15 a.m. a bricklayer who was one of the first voters made a ruckus.  He had marked his ballot, but then had trouble inserting it into the slot atop a ballot box.  He soon discovered why: all the boxes were already filled with ballots.  Indignant, he shouted out his discovery, and an infuriated crowd gathered around the polling place, chanting, "Magic! Magic!" -- a reference to alchemy, Mexican slang for the PRI's mysterious powers to transform the lead of electoral loss into the gold of triumph.

Soldiers arrived, and the voting came to a halt.  Only forty-five voters had cast ballots.

The president of the polling station, Emma Ramirez, a fifty-year-old neighborhood housewife recruited by the PRI, acknowledged to Ortiz Pinchetti that the ballot boxes had come sealed with tape.  No one had inspected them to ensure that they were empty before voting began, she said.

Ramirez's neighbors began to shout at her to cancel the balloting.  "Don't preside over the fraud," they said.  "You'll regret it your whole life."

Ramirez was torn.  PRI poll watchers surrounded her, insisting that she ignore her neighbors.  Finally the crowd prevailed, and she ordered the voting cancelled and the ballot boxes annulled.  Just then, however, a delegate from the PRI-controlled election agency arrived and took command.

"Nothing will be canceled!" he proclaimed.  "If the señorita representing National Action wants to lodge a protest, let her do it in writing, and we'll attach it to our electoral report.  Now let's get on with the voting."

"Why should we vote if our ballot is going to be wasted?" someone shouted.  And the embittered crowd, surrounding the polling place, prevented anyone else from voting that day.

After the polls closed, Ortiz Pinchetti retrieved the official results from the precinct.  They showed that the PRI received 660 votes, the PAN 29.  And soon it became clear that precinct 95-B was representative of hundreds of precincts across Chihuahua.  The polling results from many working-class neighborhoods were equally lopsided.  Ortiz Pinchetti phoned fellow journalists in Juarez and other cities.  All told of voting abuses.  At scores of the state's eighteen hundred polling places, authorities had simply ejected the PAN representatives, often violently.  At one precinct a PAN poll watcher who had been shoved bodily from the balloting area returned with a lawyer to draft a notarized document reporting his expulsion.  But PRI hooligans abducted the lawyer.

At 7 p.m., Aziz and Ortiz Pinchetti attended a news conference at which Gurria, the PRI leader, announced triumphantly that his party had won almost every office at stake that day.  Reports, including many foreign correspondents, shouted questions.  "Would you dare to characterize these elections as clean?" one journalist asked.

Gurria paused.  "I'd call them legitimate," he responded.

The Chihuahua election would actually become a globally known incident.  This began with an advertisement taken out in several newspapers and magazines by a number of public intellectuals:

The official results of the recent balloting in Chihuahua showed victories for the PRI in 98 percent of the posts up for election.  From a distance, we who have no ties to any of the parties think these figures demonstrate a dangerous obsession for unanimity.  A broad and diverse sector of Chihuahua society, being closer and having more information with which to make a reasoned judgment, believes that its vote was not respected.  ... Citizens as well as the national and international press have documented sufficient irregularities to sow reasonable doubts about the legality of the entire process.  To clear away these doubts, which touch the very fiber of the credibility of politics in Mexico, we think the authorities, acting in good faith, should reestablish public harmony and annul the Chihuahua elections.

Several days later, Government Secretary Manuel Bartlett invited several of the signers to dinner.  Here is where the book picks up:

"There was no fraud," Bartlett said.  "What do you mean, 'reasonable doubts about the legality of the entire process?' That's absurd!  Where's your evidence?"

Bartlett's assertiveness put his dinner guests against the wall.  "We didn't really know any more about the election than we'd read in the press," Aguilar Camín recalled.

In contrast, Barlett spoke with authority, claiming to have studied the complete official election file, precinct by precinct, document by document.  "You say in your manifesto that you're judging from a distance," Bartlett went on with just a hint of disdain.  "Well, I'm not!  I know what I'm talking about!  The balloting was perfectly legal."

Then, perhaps growing cocky, he went further.  "If you want, I'll make the election documents available to you," he said.  "Analyze them.  Draw your own conclusions."

...

"We'd like to look at the documents," Aguilar Camín said.

He knew that Bartlett was no fool; the Secretary had based his offer on the assumption that anyone not directly involved in the organization of the Chihuahua voting would find it virtually impossible to make sense of the official papers.

But Aguilar Camín knew somebody who might put them to very good use.

...

The morning after the Zona Rosa dinner, Aguilar Camín phoned Monlinar and described Secretary Bartlett's offer.

"You think you could find evidence of fraud?"  Aguilar Camín asked.

"They have to leave tracks," Molinar answered.  "It's just a question of study."

"What will you need?"  Aguilar Camín asked.

Molinar said he wanted the voter lists and electoral results, precinct by precinct, for all the elections in Chihuahua from 1982 throught 1986.  Aguilar Camín phoned the request to one of Bartlett's deputies.

A week later Aguilar Camín summoned Molinar to the Nexos office.  There Molinar found a stack of cardboard boxes that reached nearly to the ceiling.  There were thousands of pages of electoral data, so much that digesting it all seemed impossible.  But Molinar gamely began reading and classifying documents.

Studying other elections, he concluded that the PRI inflated the number of registered voters in regions the party controlled completely, thus providing a margin for ballot-box stuffing, while eliminating registered voters in opposition areas, enabling PRI authorities to turn them away from the polls on Election Day.

But how to test the hypothesis?  Molinar pored over the documents for weeks, and then hit on a solution . He began to compare the official voter lists for Chihuahua with the national census, which was carried out by a separate government agency.  The process took him several months, but in January 1987 his demographic comparisons had produced fascinating results.

In fifty-three of Chihuahua's sixty-seven towns the number of citizens registered to vote was greater than the total number of adults in the population.  One town with 1,128 adults had 2,627 people registered to vote; another with 987 adults had 2,112 registered voters.  The rolls had swelled rigth after the PAN's upset victory in the 1983 municipal elections, rising by more than a third from 1983 to 1986, eight times the rate of growth of the population.  Virtually all of the growth had occurred in PRI strongholds.

The 1986 balloting results were also at odds with the census figures.  The number of votes cast was far larger than the total adult population in many Chihuahua towns in remote corners of the Sierra Madre, where none of the candidates for governor had campaigned and the PAN had no poll watchers.  In those mountain towns more than 80 percent of the votes cast had been for the official party.

Molinar presented his findings to Aguilar Camín, who took a prudent step, characteristic of those authoritarian times: he contacted Bartlett.  "Mr. Secretary, we're going to publish an article with some strong criticisms," Aguilar Camín said.  "I'd like you to see it before it goes to print."  Bartlett told Aguilar Camín to send the article to José Newman, the Government Secretary's election expert at the National Voter Registry, and then phoned Newman himself.

"I want you to read this article and destroy it," Bartlett told Newman.

"How do you mean, destroy it?"  Newman asked.

"Well, convince this guy that his arguments are off base," Bartlett said.  "Tell him that he's an idiot, that he's badly mistaken."

When Newman read Molinar's work, however, he saw that he had a tough job.  No study had ever laid bare such authoritative evidence of PRI chicanery.

Newman phoned Molinar.  The two had met years earlier at the meetings of the Federal Electoral Commission (as the national agency was then called), and although they were political adversaries, they had a cordial relationship.  Newman proposed that they review the article together line by line.  Molinar agreed, and during the process Newman pointed out a few errors, the correction of which actually strengthened Molinar's arguments.

...

When the article, "Return to Chihuahua," was published in March 1987, it caused a stir.  Virtually everyone in the Mexican elite studied it, and thousands of international leaders read an account of Molinar's findings in the British magazine The Economist.

...

The turmoil of the 1986 election in Chihuahua was felt long after, and far beyond the state.  But it proved to be only a prelude to the events of the presidential vote of 1988.

That 1988 presidential vote indeed reached unimaginable heights in terms of electoral fraud.  I am going to skip over the horrendous story about the UNIVAC supercomputer that 'failed' on election night and go directly to the review of the election results:

In a vast hall on the second floor of one wing of the Government Secretariat, officials set up tables and laid out in three hundred piles the electoral materials from each district.  These included the approximately 180 tally sheets for the precincts within that district; a sábana, which was a paper spreadsheet the size of a large desktop, showing the results from each of the presidential, Senate, and congressional races in each precinct along with their combined vote totals; and the minutes of the July 10 vote tabulation sessions at the district committees.

Cárdenas's representatives at the electoral commission, Leonardo Valdés and Jorge Alcocer, were among the most vigorous in their scrutiny of these documents during the commission's proceedings in late July and early August.  Valdés and Alcocer worked from 8 a.m. to 4 p.m. each day reviewing the tallies, then rushed out for dinner, returning by 6 p.m. and remaining until midnight.

Even working every minutes at this exhausting pace, Valdés and Alcocer found a review of all the country's precinct tally sheets an impossible challenge.  With one tally for each of the presidential, Senate, and congressional races in the fifty-four thousand polling sites, there were 162,000 tallies.  All the hours available during the twenty-four days that the commission was in session left the opposition parties, on average, less than five seconds to look at any one tally sheet.  Still, their review allowed Valdés and Alcocer to learn a great deal about how the vote had been tabulated on July 10 at the country's three hundred district committees.

The tabulations followed the same procedure in the district committees all across Mexico.  An official would page through the pile of precinct tallies one by one, calling out in a loud voice -- in Spanish, cantando -- the votes for each candidate as a secretary wrote the totals onto the district spreadsheet.  The scribes who recorded the minutes of these meetings often clearly described the tactics used to inflate Salinas's vote totals.

The minutes of the tabulation session in one district in Puebla were particularly vivid, Alcocer discovered, and he studied them, fascinated.  Soon after the July 10 session in Puebla began, the PAN representative interrupted to complain, the minutes indicated.  Each time Salinas's votes from a precinct were read out loud, the PAN representative complained, the district committee secretary was adding a zero to Salinas's total on the spreadsheet, changing 73 votes for Salinas to 730 votes, for instance.

"That can't be," the president of the district committee replied.

"But that's what he's doing," the PAN representative insisted.

"Let's vote to see if the complaint from the PAN representative is accepted," the committee president said.  The PRI had a majority on this as on all other district committees, so the PAN complaint was rejected.  The tabulation continued.  Soon a representative of one of the parties in Cárdenas's coalition interrupted to complain that the secretary was still adding zeros to Salinas's totals.

"We've already voted on that complaint, and since everything is democratic here, please stop interrupting!" the commission president replied.  "Please respect our democracy!"

At that, all the opposition party representatives walked out, and the PRI officials continued the tabulation without opposition interference.

Alcocer carefully noted the scribe's account of the July 10 proceedings in the Puebla district and protested the abuses before the Federal Electoral Commission later that evening in late July around the U-shaped table.  Bartlett ruled Alcocer out of order.

Valdés and Alcocer found that Salinas's votes ended in a zero in thousands of precincts, many times more than was statistically probably had the digits appeared following the laws of random probability.

In thousands of precincts the tally sheets showed that Salinas had won with 100 percent of the votes.  In a town where 990 voters had cast ballots, for instance, not a single soul would have voted for Cárdenas or Clouthier.  Many precincts where Salinas won every last vote bordered precincts in which he received barely 30 percent, a pattern that was inexplicable in terms of political demographics.

Valdés and Alcocer identified another anomaly.  In each of the three hundred districts, Mexicans had voted for the Congress as well as for President, and the number of votes cast in the congressional election should have been roughly equal to those cast for President.  But the tallies showed far more votes for President than for deputies almost everywhere.  In some states the differences were extreme, with Salinas receiving hundreds of thousands more votes than the PRI congressional candidates.  There was only one logical explanation: as PRI fraudsters had altered thousands of tally sheets after the balloting, adding votes for Salinas, they had not wasted time touching up the congressional tallies as well.

Alcocer, Valdés, and other opposition representatives sought to have the elections annulled.  They delivered impassioned speeches about the anomalies, held up charts, and occasionally beat the table in frustration.  The PRI representatives would listen impassively, then proceed to certify the election.

While all this sounds terribly depressing and discouraging, the opposition never gave up.  In year 2000, the PAN candidate Vicente Fox won the presidential election.  There may have been some amount of cheating, but not enough to change the results.  What did the Mexicans learn since?  Let us pick up with Preston and Dillon again in the epilogue of their book:

Certainly Fox had aroused unrealistic hopes and offered only mediocre presidential leadership.  It was also true that amid the malaise, a few voices, as elsewhere in Latin America, were complaining that democratic rule itself was too cumbersome to deal with region's economic and social challenges.  But a majority of Mexicans expressed pride in the democracy they had achieved and were willing to give it more change time.

Democratic government in the United States was born in the last decades of the eighteenth century with a brilliant constitution and tremendous grassroots vigor.  Yet it took U.S. democracy eight decades to abolish slavery, and after a full century urban corruption as bad as any seen under PRI was proliferating in the New York slums controlled by Tammany Hall.  Did U.S. democracy produce efficient economic policies?  The United States had experienced a century and a half of democratic government when the Great Depression put a third of its workers onto soup lines.

Democracy does not guarantee good government, but it is a set of rules and a culture for resolving differences that allow citizens to limit misgovernment.  Mexicans had been building their democracy in earnest for more than three decades, and it was a work in progress.  They were now directing a drama in which Vicente Fox, López Obrador, and the lawmakers of the newly independent Congress were temporary players.  If these politicians faltered, Mexicans now had the power to call forth other leaders to more boldly carry forward the nation's agenda.

Mexico's future, like its past, would undoubtedly include moments of turbulence.  But it was a nation whose citizens had created an effective division of powers and a well-tested electoral system, and had achieved complete liberty of expression.  Vigorous democratic institutions would endure the tests to come.

Mexico had seemed the perfect dictatorship.  Now it has an imperfect democracy.