"According to our statistics, we have won the presidency," Andres M.L. Obrador roared Sunday night. "The triumph is irreversible." But just over half an hour later, Felipe Calderon of the ruling National Action Party (PAN) announced "there is no doubt we have won."
So both claimed victory. Pre-election polls had seemed to give it to Obrador but his advantage over Calderon kept shrinking as the election approached. The preliminary official figures gave it to Calderon by a 1.04% margin. Later, the IFE said Obrador was in the lead, then later [4:00am Mexico City time on 6 July] the IFE claimed that 97.70% of the votes had been counted and Calderon was again leading Obrador by 0.01%: 35.60% to 35.59%! Later on 6 July the IFE claimed Calderon had won by 243,000 votes (i.e. about a 0.6% margin) by 35.9% to 35.3%, with 99.6% of the votes counted so that the margin was irreversible (but counting continued). Obrador then announced he was going to demand a recount because there were "many irregularities" and there was a 100,000-strong demonstration in Mexico City on 8 July (then a million-strong one on 17 July) protesting his loss.
| Candidate | PREP% | final% |
|---|---|---|
| Calderon | 36.38 | 35.89 |
| Obrador | 35.34 | 35.31 |
| Madrazo | 21.57 | 22.26 |
| Mercado | 2.81 | 2.70 |
| Campa | 0.99 | 0.96 |
| other | 0.72 | 0.71 |
| spoiled | 2.14 | 2.16 |
Calderon is the conservative, Vicente Fox's picked successor, and openly supported by US president G.W. Bush. Obrador is the leftist. However, a more-leftist leftist (i.e. actually feminist and willing to say things against the Catholic church...) was Patricia Mercado, who got about 3% of the vote. So clearly Mercado was the "Nader" in this race who threw the election to Calderon when it "should" have gone to Obrador.
However, that assessment ignores the PRI candidate Roberto Madrazo, who was in third place with about 22%. The PRI was the party that held Mexico for decades in thrall of corrupt single-party rule, and which had previously defeated Obrador in a massively corruptly rigged elections in 1988 and 1994 (Tabasco Governor race). It had also defeated Calderon's NAP party in rigged races too. The PRI was supplanted by Vicente Fox who appears to have cleaned up Mexican corruption to a considerable extent and set up what appears to be a highly professional and independent election agency, the IFE. Madrazo was actually running for an unusual and interesting alliance of the PRI with the Mexican Green Ecologist Party (PVEM) and focused on water issues.
So the question – if we want to talk about who "should" have won – hinges on who the Madrazo PRI voters would have preferred: Obrador or Calderon?
That is unclear to everybody, far as I can tell.
But there is another hypothesis, which is that many leftists would have preferred Madrazo as the "lesser evil" versus the rightist Calderon, and similarly many Calderon voters would have preferred Madrazo over Obrador. In that case, quite possibly Madrazo was the "Condorcet winner," i.e. would have beaten every other candidate in a head-to-head contest.
So all in all, I conclude this election probably was an example of either the "spoiler" phenomenon (a fate warned about in both this and this other newspaper article) – or the "eliminated Condorcet winner" phenomenon, and in either case Calderon was the "wrong winner." Whether or not that was true, almost certainly range voting would have produced a far more clearly fair election result. I also think IRV, Condorcet, and plurality with separate top-2 runoff both would have handled it better than the plain plurality system that was used. And there is no doubt Mercado's 4% total was vastly smaller (due to plurality distortions) than what it would have been under range voting; it probably would have been in the same ballpark as Calderon and Obrador.
(This election also may have been swung by election manipulation and fraud: Obrador claims there were numerous statistical "anomalies" and suspicious signs were noted by investigative reporter Greg Palast in the Guardian – see also his followup article – triggered by the Mexicans arresting operatives of the notorious US contractors ChoicePoint for stealing private voter databases... and here Narco News, "reporting on the drug war and democracy from Latin America", documents that fraud happened... and the IFE's computer systems were partly designed by companies and partners of Calderon's brother-in-law Diego Hildebrando Zavala, as contracts show... and the IFE had to admit that the PAN had "somehow" gotten hold of private voter registration data [Jorge Herrera article in El Universal 28 June 2006]... Obrador sued alleging fraud, but the IFE, after conducting a very partial recount of 9% of the precincts, declared Calderon the victor. As J.C.McKinley put it in his NY Times article page A4, 7 sept 2006: "Obrador pointed out that more than half of the tally sheets from the nation's 130,000 election precincts contained errors in arithmetic, a sign of widespread incompetence among poll workers or of extra ballots magically appearing in some boxes and disappearing from others [and rather worrying since even a single vote change per precinct would have mathematically been enough to swing the election to Obrador]. But the seven-member electoral tribunal rejected that argument, chalking up the irregularities to human error, which they said had affected all parties, and so could not be fraud. They also denied [Obrador]'s request for a full recount based on the errors." Obrador, outraged by the IFE's ruling, then declared he was going to set up a shadow government.) Judging 1%-margin elections can be difficult (and 0.01%-margin elections are well beyond what present systems are capable of). That difficulty is illustrated by the following facts: