Mexico 2006

[ag]

The final decision is still not clear. It turns out, there are still some 3 mln. votes to be added to the tallies. The first one to publically note the discrepancy between the published turnout number (58.9%) and the number of votes tallied (around 54.5% of the register) was Lopez Obrador. IFE chief Ugalde has now explained that if there were errors in filling out the precinct protocols (such as lack of signatures of some party representatives, etc.) they added the overall number of votes to the turnout, but so far didn't add them to public tallies. They will only do that after the protocols are reviewed by all party representatives.

Thus, it is still too close to call: just over 400,000 vote gap with 3 mln. outstanding ballots. Admittedly, in a three-way race, it would be hard to overcome the existing advantage, but it is not impossible. And it could definitely get a lot closer.

The official count only STARTS tomorrow, and would last a few days. Still, suspense.

[ag]

And another news.

PRI governors have forced a very humiliating Madrazo climbdown. Madrazo publically announced that he lost, and that the election was clean, transparent and  fair.

This is likely to be the last we hear from Madrazo for a while. After a meeting with the governors and union leaders it has been made public that the party is conducting negotiations with the Calderon campaign. Neither Madrazo, nor the official party head, Mariano Palacios, will take part in these negotiations or do anything else, at least in the party name. Palacios will, presumably, be shortly replaced as the party leader.

Update: Madrazo has called Calderon, essentially to congratulate (officially they discussed "avoiding confrontation in the coming months"; Calderon thanked Madrazo for "showing maturity").

Also, Panal's candidate Campa has called for an official recognition of Calderon's victory.

[ag]

Indications are that PRI's leadership will be contested between Beatriz Paredes (who just came third the Mexico City Mayoralty, but still got the highest number of votes of any PRI candidate in the City since 1997) and the outgoing leader of PRI's faction in the senate, Enrique Jackson. The latter one would be viewed as clearly pro-PAN, Paredes is more ambiguous, though not pro-PRD.  Palacios term actually expires on Saturday, so replacing him will be easy.

What has it come to! They main discussion is to which of the "major parties" should PRI ally to

Also, it seems like the PVEM has run a great bargain with PRI.  Even though they seem to have done nothing to help PRI in the campaign, they should get about 12 seats in each chamber.

[angus]

http://blog.washingtonpost.com/postglobal/miklos_vamos/2006/07/the_free_trader_the_subsidiato.html

[ag]

Well, there is still no election outcome. Actually, they are still to count about 3.5 million votes, give or take. They goofed up announcing things - as people tend to, when they've been doing the same thing over and over and it seems obvious (to them). We won't know the outcome till Friday.
 
The preliminary count is really preliminary. If there are any problems with a protocol, they just put it aside, simply recording the number of votes for the turnout, but not distributing the votes. The preliminary reporting system shows it as "counted", but it really is not.
 
The real summing up of the protocols starts tomorrow, and it will resolve those "preliminary" problems - and it is the only thing that counts officially. But they goofed up announcing things appropriately. From what they said repeatedly, it seemed that 98.45% of all the precincts have reported, and with the gap between the top two of over 1% it looked like the election was done. It remained entirely unclear why they wouldn't declare the winner - as they consistently refused to do.
 
Of course, Lopez Obrador noticed the 3 mln. vote gap (in addition, there are probably another 500,000 votes in those precincts that haven't yet reported) between the declared turnout and the total number of votes counted, which wasn't hard to do, since it was on the same electoral commission web page (39 million votes counted or pending, turnout 58.9 of 71.3 million registered voters, you do your calculation). The IFE head, Ugalde, has now tried to clarify it, but with the  tin ear of an elections' official, he did this not very convincingly.

It took me several hours to confirm that what he is saying is really true, and that there is no problem, except for having still to count those votes.  I looked at the state-level elections that also hapened on Sunday, assuming that if the state electoral commissions do something different I will see different numbers of voters - but, subject to a minimal give or take, the numbers of voters were the same. Since local and state elections are run separately, by independent state-level agencies, this was surprising. In particular, I looked at the Mexico City results, since Mexico City electoral commission (which runs the mayoral election) was appointed by the PRD-controlled City Council, while the national IFE (which runs the presidential election everywhere, including the city) was appointed by the PAN and PRI factions in the Congress (PRD voted against on that occasion). The City and Federal precincts, though located at the same address, where separate, with separated officials, monetors and ballot boxes. The counting was done independently by two entirely independent bodies. No turnout difference (out of 4.7 mln. votes cast, the difference was 17 thousand, easily accounted for by the slightly different number of precincts declared as counted by the two commissions). Thus, clearly the two independent electoral bodies were doing, roughly, the same thing.

Finally, at the web page of Nuevo Leon Electoral Commission (when in doubt about anything in Mexico, check in Nuevo Leon - they are the only ones here to know how to do things right) I found an example of complete reporting, which exactly said how many protocols have been postponed in each local municipal or state legislative election (they didn't have a governor's race on Sunday, so this was all there was for the state Commission to run).  I believe, this sort of postponing the count in disputed cases must be a common pracitce accross the country, but only Nuevo Leon bothered to report this (most states had even less info than the Feds). But, though information is actually available on the web, I doubt if anybody else on earth bothered to find it. Thus, all the suporters of Lopez Obrador will believe a massive fraud is being perpetrated, even though it is a minor bureaucratic misreporting.
 
In any case, we still have 3.5 million votes to count. With an effort I managed to get enough information to think that it is pretty randomly distributed accross the country, so things are unlikely to change that much.  If anything, judging by Nuevo Leon numbers, the missing protocols tend to be concentrated in the areas where PRI is the strongest (poor rural areas with a lot of badly educated people, where local polling oficials might be counted on not filling things correctly). This might account for why the exit polls tended to give Madrazo about 1% more than the data so far. Thus, I would expect the biggest adjustment to be in PRI doing slightly less miserable. Still, we have to wait.
 

[ag]

IFE has just come out clean.

The  11,184 "missing" (containing errors) protocols account for 2 581 226 votes.

These are distributed as follows:

743 795 for Calderon.

809 003 for Madrazo.

888 971 for Lopez Obrador.

13 946 for Campa.

28 040 for Mercado

15  019 for write-ins.

82 452 invalid.

Calderon's lead is thus cut to 257,000 votes.

[ATFFL]

Has anyone hired Dean Logan yet?

[True Democrat]

IFE has just come out clean.

The  11,184 "missing" (containing errors) protocols account for 2 581 226 votes.

These are distributed as follows:

743 795 for Calderon.

809 003 for Madrazo.

888 971 for Lopez Obrador.

13 946 for Campa.

28 040 for Mercado

15  019 for write-ins.

82 452 invalid.

Calderon's lead is thus cut to 257,000 votes.

So there were really were 2 million missing votes. . .weird.  Why so many for the PRI?

[ag]

So there were really were 2 million missing votes. . .weird.  Why so many for the PRI?

These were the precinct protocols that had errors in them.  Election-night officials are drafted (jury-duty style) from the local population. The less literate the population, the likelier it is they make mistakes in filling out the paperwork.  The less educated are more likely to vote both Madrazo and Lopez Obrador.

Another potential scandal, this time in the other direction. In a garbage dump of the (desperately poor and PRD-dominated) Mexico City suburb called the City of Nezahualcoyotl they found large amounts of voting material (ballot boxes, protocols, etc.).  El Universal mentioned the precinct numbers on some of the protocols, and for one of them gave the vote distribution. Coming from Nezahualcoyotl, it is unsurprising that these had Lopez leading, with Madrazo distant second and Calderon even more distant third. However, when I punched in the precinct number in the preliminary results information system, it gave me even more votes for Lopez, and even fewer votes for both Madrazo and Calderon (something like a shift of about 20 votes each from the other two to Lopez Obrador - in a precinct with about 400 votes, this is non-negligible). This is sure to give cause for Calderon's campaign to file an official protest.

PRD is calling for the full recount.  IFE is calling for calm.

In other news, and in other sign of PRI's intentions, the losing PRI candidate for the governor of Jalisco has accepted his defeat.

[ag]

So, assuming the "unclear" precincts have been cleared up (perhaps there were a few more, where the protocols were so illegible, that a recount is necessary to determine anything, but we won't know this till later), we are left with 2027 precincts that are still truly not reported. Assuming the average number of votes per not reported precinct is the same as nationwide, we would account for almost 41.8 mln. votes; the declared turnout of 58.9% of the 71.3 mln. voters with valid voter ID's, is just under 42 mln. Either, once these precincts report, the turnout drops a bit to 58.6%, or there are another 200 thousand votes to be accounted for (probably, those most severely damaged protocols hypothesised about above), but the ballpark is there. In any case, the biggest chunk of "missing" votes now is in the remaining precincts.

I have divided the precincts not reported by the identity of the presidential candidate leading in the congressional district to which they belong.

1226 precincts in districts led by Calderon
782 precincts in districts led by Lopez
19 precincts in districts led by Madrazo (in 14 of these Lopez is far ahead of Calderon, 5 are in a district with a close race for the second spot).

At first glance this should help Calderon, but it is more ambiguous. In particular, the biggest single cluster of precincts missing from the same state comes from Chihuahua (484), of which humongous 283 (nearly 15% of the nationwide total) comes from the largely rural Hidalgo Del Parral district, where the average precinct is likely to be smallish (say, half the average size). Since all of these are counted above as Calderonista, this will reduce his advantage. In fact, a big chunk of these seem to come from the area of the district where Madrazo did well, reducing Calderon's advantage even further. In contrast, the 251 missing precincts from Mexico state, of which 227 come from Lopez-led districts (the largest single block is the 82 precincts in the solidly Lopezian Valle de Chalco Solidaridad) might be larger than average. The only other state with more than 100 precincts missing is Mexico City, where there are 101 of these, of which a 100 come from Lopez-led districts (though, there aren't big clusters here; the largest is the 12 precincts missing in the 2nd district, and that one is rather evenly split between Lopez and Calderon).

So, on balance it would seem that the remaining precincts should split the votes evenly. Though, as we've just seen, in areas where Calderon does well, local officials are less likely to make mistakes (the education factor). Thus, there might be a slight lean to Lopez.

Still, the  lead would have to be substantial. If the average precinct size among the unreported precincts is no different from the countrywide, we are talking about 650,000 votes. Even if another 200,000 votes are found to justify the 58.9% turnout figure, we are talking about 850,000 votes, give or take a few. In a three-way race, to wipe out Calderon's advantage of over 250,000 votes, Lopez would have to win a huge proportion of these 850,000. There are enough precincts left in solidly Calderonista states (even if we ignore Chihuahua) to make this unlikely. Thus, it is doubtful if there are enough votes left to count for the first time to change the leader of the race.

Of course, if the gap narrows even further, partial or full (if PRD demand is accepted) recounts might change things even more. And, even if that doesn't do enough, the narrower the gap, the easier it is for the loosing side to have enough precincts annulled for irregularites to win. After the final count is announced, probably Friday or some time this weekend, both sides have 4 days to file claims in courts. If these are accepted, they have to be resolved by the end of August, with final results to be certified before the new congressional session in September.  The president's inauguration is not due till Dec. 1.

It's going to be a hot summer.

[minionofmidas]

IFE has just come out clean.

The  11,184 "missing" (containing errors) protocols account for 2 581 226 votes.

These are distributed as follows:

743 795 for Calderon.

809 003 for Madrazo.

888 971 for Lopez Obrador.

13 946 for Campa.

28 040 for Mercado

15  019 for write-ins.

82 452 invalid.

Calderon's lead is thus cut to 257,000 votes.

Now all they need to do is declare any ballots but these invalid!

[jaichind]

AMLO have reasons to complain.  Lets take a look at Guerrero which went for AMLO 51%, and 17% for Calderon.  But turnout in this PRD stronghold is an unusually low 47% when the national turnout was 59%. 

[Sam Spade]

Actually, this whole election is reminding me why the American creation of an electoral college is such an extremely good idea.

Imagine what would happen here if the vote was within 1% and we had a national popular vote.  Talk about madness.  And it's happened here three times in the last 50 years.

[jaichind]

I guess in an EC style system PAN wins since the PRD vote are concentraded in the Greater Mexico City area where they have massive leads.

[ag]

AMLO have reasons to complain.  Lets take a look at Guerrero which went for AMLO 51%, and 17% for Calderon.  But turnout in this PRD stronghold is an unusually low 47% when the national turnout was 59%. 

Actually, using your logic:

"Calderon has reasons to complain.  Lets take a look at Baja California which went for Calderon 47%, and 23% for AMLO.  But turnout in this PAN stronghold is an unusually low 47% when the national turnout was 59%."

Actually, the highest turnouts (68%) were in Mexico City, where Lopez got 58% (to Calderon's 27%) and in Tabasco, which gave Calderon the worst result (Lopez 56%, Calderon 3.6%).  And Tabasco didn't even have a  concurrent state election.

Now, Guerrero (like Mexico City) is controlled lock, stock, and barrel by PRD. PRD controls the state governorship, legislative majority and most important mayoralties (including the largest city, Acapulco).  If anybody could "discourage" voters from showing up it was the PRD - no other  party can as much as fart there anymore, without PRD letting this happen. PRD inherited the PRI machine in the state. Perhaps, they can blame themselves: when no trash had been picked up in residential areas of Acapulco for over a month (the PRD mayor cancelled the garbage-picking contract without signing a new one), perhaps it is not that surprising that a lot of people were not very motivated to vote. In any case, the lowest turnouts were in rural districts - much of the population there is north of the border, so those people couldn't possibly show up.

[Sam Spade]

I guess in an EC style system PAN wins since the PRD vote are concentraded in the Greater Mexico City area where they have massive leads.

This might be true, though most of these areas have much less population than the Obrazador center of Mexico City.  Obrazador also won a good amount of Southern rural Mexico, which is fairly unpopulated as well.

I was simply speaking to the logic of possibly only having to count (and recount and recount) one or two states of 300,000 votes as opposed to a national 30,000,000 votes.

[ag]

On US EC system (# of congressional districts+2= number of electoral votes) Lopez wins 199:165. Though, if Veracruz (23 EV) would switch (the gap is about 1%), Calderon would win.

There are several sates where Lopez won by slim margins (in Campeche - 4 EV - the difference is about 2 thousand votes).  With one or two exceptions, every state where Calderon won, he won by at least 10%, mostly by at least 2:1 ratio. If winning a state had meaning, the campaign would be very different, of course, so Calderon would have likely won more states.

[ag]

For those who would like to play with US-style electoral scenarios, here are the states arranged by the difference between the Calderon and Lopez vote shares (ignoring Madrazo). First column - state, second column - Calderon's share, 4th column Lopez's share, 5the column difference between the previous two, 6th column - the number of congressional districts +2 , which I call "EV". The numbers don't take into account the latest batch of 2.6 mln. votes, though.

Overwhelming (>40% difference) Lopez

Tabasco   3.63   56.11   -52.48   8 EV

Very strong Lopez (20-40% difference)

Guerrero   17.22   51.06   -33.84   11 EV
Mexico C. 27.33   58.13   -30.8   29 EV
Oaxaca   17.26   46.51   -29.25   13 EV
Chiapas   17.66   43.68   -26.02   14 EV
Nayarit   19.19   42.05   -22.86   5 EV

Strong Lopez (10-20% difference)

Hidalgo   26.86   40.97   -14.11   9 EV
Morelos   31.93   44.09   -12.16   7 EV
Mexico S.   31.19   43.2   -12.01   42 EV

Weak Lopez (5-10 % difference)

Tlaxcala                   34.16   43.96   -9.8   5 EV
Quintana Roo   29.52   37.91   -8.39   5 EV
Baja California Sur   34.39   42.69   -8.3   4 EV

Swing Lopez (5% or less difference)

Michoacan   35.57   40.57   -5   14 EV
Zacatecas   32.52   35.5   -2.98   6 EV
Veracruz                   34.32   35.36   -1.04   23 EV
Campeche   31.89   32.7   -0.81   4 EV

Swing Calderon (5% or less difference)

none

WeakCalderon (5-10% difference)

Sinaloa   37.25   31.76   5.49   10 EV
Puebla   37.82   32.31   5.51   18 EV

Strong Calderon (10-20% difference)

Tamaulipas      41.53   26.7   14.83   10 EV
Colima          41.75   23.91   17.84   4 EV
Coahuila          43.53   24.12   19.41   9 EV

Very strong Calderon (20-40% difference0

Durango                   45.25   22.94   22.31   6 EV
Baja California   47.34   23.57   23.77   10 EV
Sonora                   50.47   25.54   24.93   9 EV
Aguascalientes   46.75   21.76   24.99   5 EV
Queretaro        49.25   24.26   24.99   6 EV
San Luis Potosi   49.1   21.85   27.25   9 EV
Chihuahua   46.3   18.31   27.99   11 EV
Yucatan                   46.3   16.23   30.07   7 EV
Jalisco                   49.39   19.25   30.14   21 EV
Nuevo Leon   49.23   15.97   33.26   14 EV

Overwhelming Calderon (>40% difference)

Guanajuato   59.09   15.3   43.79   16 EV

[Sarnstrom]

The dream still lives:

Mexico leftist has slim early lead in vote recount

MEXICO CITY (Reuters) - Mexico's leftist presidential candidate Andres Manuel Lopez Obrador had an early, narrow lead over his conservative rival on Wednesday in a recount of a contested election.

Results on display at the Federal Electoral Institute showed Lopez Obrador had 37.05 percent of the vote with results in from 36.6 percent of polling stations. Ruling party candidate Felipe Calderon was second with 34.38 percent.

It was too soon to say whether the trend would hold. Preliminary results earlier this week from Sunday's election gave Calderon a lead of about 0.6 percentage points over Lopez Obrador.

http://news.yahoo.com/s/nm/20060705/ts_nm/mexico_election_dc_35;_ylt=AkT0xln_UuvMYF6MutLkaNUdl.0A;_ylu=X3oDMTBiMW04NW9mBHNlYwMlJVRPUCUl

[ag]

Actually, Lopez has had a fairly constant lead in the official count today. It might come from the fact that PRD is more active challenging booths for a full recount, in which case the recount of Calderonista areas is going slower. Or it might be something real.

So far, the official count of 60.07% of the booths shows

Lopez Obrador 37.09%
Calderon 34.56%
Madrazo 21.79%
Mercado 2.75%
Campa 0.92%

Still, the basic party disposition hasn't changed. PAN says that so far the count is going in accordance with the preliminary count (ie, it is just a bad selection of booths for them), while PRD threatens to reject the count, unless a full vote by vote recount is ordered (according to the law, unless a particular booth is challenged for a cause - eg., discrepancy in diferent copies of the protocols - it does not get recounted, just the protocol numbers verified).

[jaichind]

with 62.6 of votes counted

AMLO       37.1%
Calderon  34.5%

[○∙◄☻¥tπ[╪AV┼cVê└]

Wow, this is a dramatic election.

[ag]

With 66% of precincts counted

Lopez Obrador 36.9%
Calderon 34.6%

Well, the gap has started shrinking in the last half hour.

By the way, these are not votes counted, but precincts counted. Not all precincts are the same size. A small precinct, of course, means that it is in an isolated area, which in the preliminary count suggested that it would report late. But now all the ballot boxes are at the district headquarters, but, if challenged, larger precincts would take longer to recount. So districts with lots of small precincts are likely to report a lot of these, and areas with larger precincts (more densely populated parts) are likely to report relatively fewer precincts, though more votes. So, we might actually still have less than 60% of the votes, even though we have 66% of precincts. Unfortunately, they do not this time provide numbers of votes counted, only the precincts.

Update: with 69.01% of precincts reporting

Andrés Manuel López Obrador: 36.81 %
Felipe Calderón: 34.64 %
Roberto Madrazo: 21.98%
Patricia Mercado: 2.73 %
Roberto Campa: 0.94 %

[ag]

My forecast:

both main candidates will have between 35% and 36% of the vote and then simultaneously sue.

[Sarnstrom]

72.86% of the precincts in and Lopez Obrador has slightly increased his lead. He now leads 36.86% to Calderon's 34.55%. Calderon would have to win the remaining precincts by 4.26% to erase Obrador's lead. However, the rumor is that the remaining precincts are from western Mexico which is PAN friendly territory so it is still possible Calderon could take the lead.