Mexico June 7th 2015 elections

[jaichind]

I did notice something every Mexican election but never got to ask about.  It seems that the IFE is very thoughtful about how they let voters vote for when it comes in alliances.  So if 3 parties (A, B, C) run as allies it seems the ballot allows the voter to vote for

A
B
C
A-B
B-C
A-C
A-B-C

My question is what if there are 4 or 5 or even 6 parties that want to form an alliance.  The size of the ballot will grow out of control if taken to its logical mathematical conclusion.  Will IFE say at some stage "Look, this is getting to complex and will produce a ballot that is too long.  What we will do is to let voters vote for

A
B
C
D
E
F
A-B-C-D-E-F

And that is it"

[ag]

Big news from Morelos. The new mayor of Cuernavaca will be... Cuauhtemoc Blanco, the famous (well, for those who care about soccer) soccer player.  He ran for PSD - the party, which has lost its federal register long ago, still survives as a state party in Morelos. They got a star candidate to run - and it worked!

How does a party retain register as a state party?  Is it about winning more than 3% in the governor race or in the state Chamber of Deputies ?

Depends on the state law, but, generally, in this spirit.

[ag]

I did notice something every Mexican election but never got to ask about.  It seems that the IFE is very thoughtful about how they let voters vote for when it comes in alliances.  So if 3 parties (A, B, C) run as allies it seems the ballot allows the voter to vote for

A
B
C
A-B
B-C
A-C
A-B-C

My question is what if there are 4 or 5 or even 6 parties that want to form an alliance.  The size of the ballot will grow out of control if taken to its logical mathematical conclusion.  Will IFE say at some stage "Look, this is getting to complex and will produce a ballot that is too long.  What we will do is to let voters vote for

A
B
C
D
E
F
A-B-C-D-E-F

And that is it"

There is a difference between alliance and common candidate.

Alliance means there is only one line on the ballot, with the name of some sort of a coalition ("United for Puebla", "For the Benefit of All", "For everything good and against everything bad"), etc. The voter chooses the coalition, and constituent parties get allocated votes based on a formula written into the alliance agreement. I believe, at this point this is only possible in some state elections, not in the federal elections anymore (though it used to be common at federal level as well).

A common candidate means that parties choose to nominate the same person. It seems, current federal law specifies that this can only be done if there is a prior coalition agreement between parties (this, probably, varies in state laws). If that happens, parties retain separate lines, but the same name appears next to each party emblem. In this case the voter can choose the party for which he votes. The votes get aggregated from all the party lines to determine the victor in FPTP, but stay separate for the purposes of PR. The voter could also choose to divide the vote between parties, by marking more than one symbol. If the parties have a common candidate the vote remains valid, and the PR allocation is fractional: divided equally between the parties the voter chose. If it so happens that the parties nominated different candidates, the vote is considered spoiled, though (in 2012 this cost PRI/PVEM quite a few seats in districts where they ran separately - they campaigned on voting for Pena Nieto on either ballot, but in some districts that led to many spoiled congressional ballots).

In any case, the ballot size cannot be larger than the number of parties + the number of independents in the district.

[ag]

Now that PES is in, will PES and PAN form an alliance for 2018 ?

I would conjecture, PES would go for the highest bidder. If PAN has something to pay with, they will go with PAN. But do not count on them not joining PRI.

In any case, yeah, there is now a minor party on the right as well.

RIP PT.

Actually, is PT really out?  Is the 3% rule for 3% of the total vote or 3% of the non-null vote?  Because the PT vote share of 2.87%, for now, once we normalize against the non-null vote, is actually above 3%.

You may be right!

Article 54, Section II of the constitution states

TODO PARTIDO POLITICO QUE ALCANCE POR LO MENOS EL TRES POR CIENTO DEL TOTAL DE LA VOTACION VALIDA EMITIDA PARA LAS LISTAS REGIONALES DE LAS CIRCUNSCRIPCIONES PLURINOMINALES, TENDRA DERECHO A QUE LE SEAN ATRIBUIDOS DIPUTADOS SEGUN EL PRINCIPIO DE REPRESENTACION PROPORCIONAL;

Every political party that achieves at least three per cent of the total of the valid vote cast in the regional lists of the multi-member circumscriptions shall have the right to have Deputees assigned to it on the principle of proportional representation.

It seems, PT is at 3.02% right now! Unless, of course, spoiled ballots are somehow considered valid.

[ag]

Yep!

Article 15, Section 1 of the electoral law

Se entiende por votación total emitida, la suma de todos los votos depositados en las urnas. Para
los efectos de la aplicación de la fracción II del artículo 54 de la Constitución, se entiende por votación
válida emitida la que resulte de deducir de la suma de todos los votos depositados en las urnas, los votos
nulos y los correspondientes a los candidatos no registrados.

One should understand as the total vote cast the sum of all the votes deposited in the ballot boxes. For the effects of application of the section III of the article 54 of the Constiution, one should understand as the total valid vote cast what results from deducting from the sum of all the votes deposited in the ballot boxes the null votes and those that correspond to candidates not registered.

[jaichind]

I did notice something every Mexican election but never got to ask about.  It seems that the IFE is very thoughtful about how they let voters vote for when it comes in alliances.  So if 3 parties (A, B, C) run as allies it seems the ballot allows the voter to vote for

A
B
C
A-B
B-C
A-C
A-B-C

My question is what if there are 4 or 5 or even 6 parties that want to form an alliance.  The size of the ballot will grow out of control if taken to its logical mathematical conclusion.  Will IFE say at some stage "Look, this is getting to complex and will produce a ballot that is too long.  What we will do is to let voters vote for

A
B
C
D
E
F
A-B-C-D-E-F

And that is it"

There is a difference between alliance and common candidate.

Alliance means there is only one line on the ballot, with the name of some sort of a coalition ("United for Puebla", "For the Benefit of All", "For everything good and against everything bad"), etc. The voter chooses the coalition, and constituent parties get allocated votes based on a formula written into the alliance agreement. I believe, at this point this is only possible in some state elections, not in the federal elections anymore (though it used to be common at federal level as well).

A common candidate means that parties choose to nominate the same person. It seems, current federal law specifies that this can only be done if there is a prior coalition agreement between parties (this, probably, varies in state laws). If that happens, parties retain separate lines, but the same name appears next to each party emblem. In this case the voter can choose the party for which he votes. The votes get aggregated from all the party lines to determine the victor in FPTP, but stay separate for the purposes of PR. The voter could also choose to divide the vote between parties, by marking more than one symbol. If the parties have a common candidate the vote remains valid, and the PR allocation is fractional: divided equally between the parties the voter chose. If it so happens that the parties nominated different candidates, the vote is considered spoiled, though (in 2012 this cost PRI/PVEM quite a few seats in districts where they ran separately - they campaigned on voting for Pena Nieto on either ballot, but in some districts that led to many spoiled congressional ballots).

In any case, the ballot size cannot be larger than the number of parties + the number of independents in the district.

Ah, I see.  So when I see result on PREP that lists Party A <X votes>, Party B <Y votes> Party A/Party B <Z votes>, the Z votes came for someone marking both Party A and Party B and not there is a separate line on the ballot called Party A/Party B.  Thanks for clearing this up.

[ag]

To sum up: PT might still live!

Would, actually be good - fewer seats for the PRI/PVEM combine. And, though it means more money for the left, it also means more parties AMLO would have to buy and/or bully to get the nomination next time

[jaichind]

Yep!

Article 15, Section 1 of the electoral law

Se entiende por votación total emitida, la suma de todos los votos depositados en las urnas. Para
los efectos de la aplicación de la fracción II del artículo 54 de la Constitución, se entiende por votación
válida emitida la que resulte de deducir de la suma de todos los votos depositados en las urnas, los votos
nulos y los correspondientes a los candidatos no registrados.

One should understand as the total vote cast the sum of all the votes deposited in the ballot boxes. For the effects of application of the section III of the article 54 of the Constiution, one should understand as the total valid vote cast what results from deducting from the sum of all the votes deposited in the ballot boxes the null votes and those that correspond to candidates not registered.

Interesting.  Well this is based on PT getting 2.87% of the total vote (including nulls) based on PREP.  We will see what the real number is tomorrow.  Of course this will sink further the PRI/PVEM seat share since this will push up the number of parties/vote to share in the PR seat allocation and increase  effective vote to apply the 8% rule.

[ag]

Yep!

Article 15, Section 1 of the electoral law

Se entiende por votación total emitida, la suma de todos los votos depositados en las urnas. Para
los efectos de la aplicación de la fracción II del artículo 54 de la Constitución, se entiende por votación
válida emitida la que resulte de deducir de la suma de todos los votos depositados en las urnas, los votos
nulos y los correspondientes a los candidatos no registrados.

One should understand as the total vote cast the sum of all the votes deposited in the ballot boxes. For the effects of application of the section III of the article 54 of the Constiution, one should understand as the total valid vote cast what results from deducting from the sum of all the votes deposited in the ballot boxes the null votes and those that correspond to candidates not registered.

Interesting.  Well this is based on PT getting 2.87% of the total vote (including nulls) based on PREP.  We will see what the real number is tomorrow.  Of course this will sink further the PRI/PVEM seat share since this will push up the number of parties/vote to share in the PR seat allocation and increase  effective vote to apply the 8% rule.

So far, this PT has 3.029% of the total valid vote as defined by law (write-ins are also invalid). Of course, the PREP, really, is missing nearly 7% of the precincts (between those that never made it and those that had "inconsistencies"). So, we shall have to see tomorrow's official count. And, possibly, a recount. And, probably, a TRIFE ruling.

[ag]

BTW, the fact that INE designed PREP to report percentages the way they reported is really stupid. This has fulled not only us here - but the media as well (headlines are out announcing PT demise). 

[jaichind]

To sum up: PT might still live!

Would, actually be good - fewer seats for the PRI/PVEM combine. And, though it means more money for the left, it also means more parties AMLO would have to buy and/or bully to get the nomination next time

Thinking big picture about 2018, the fact that MORENA did fairly well means AMLO will run in 2018 which implies that unless PAN comes up with something credible this is another scenario for PRI-PVEM to win again.  AMLO running would inhibit someone else on the Left from running since they will have no chance of winning with AMLO getting at least half of the Left vote share.  AMLO is the main alternative to PRI-PVEM also would mean PRI-PVEM can always win as the least bad choice.  I am not even sure what is the point of ALMO trying to buy PT.  He must know he cannot win no matter what with or without the PT line.  Perhaps AMLO thinks/hopes there is a complete policy meltdown under Nieto between now and 2018 ?  He just has to intimidate anyone else from the Left from running by running a credible campaign and retain is role as dominate force on the Left in Mexico.  Only way to avoid this would be be for PAN to have a dynamic candidate that can somehow get up to above 33% support to become the main alternative to PRI-PVEM.   Or else something like El Bronco running a grand anti-PRI anti-AMLO alliance candidate.

[ag]

To sum up: PT might still live!

Would, actually be good - fewer seats for the PRI/PVEM combine. And, though it means more money for the left, it also means more parties AMLO would have to buy and/or bully to get the nomination next time

Thinking big picture about 2018, the fact that MORENA did fairly well means AMLO will run in 2018 which implies that unless PAN comes up with something credible this is another scenario for PRI-PVEM to win again.  AMLO running would inhibit someone else on the Left from running since they will have no chance of winning with AMLO getting at least half of the Left vote share.  AMLO is the main alternative to PRI-PVEM also would mean PRI-PVEM can always win as the least bad choice.  I am not even sure what is the point of ALMO trying to buy PT.  He must know he cannot win no matter what with or without the PT line.  Perhaps AMLO thinks/hopes there is a complete policy meltdown under Nieto between now and 2018 ?  He just has to intimidate anyone else from the Left from running by running a credible campaign and retain is role as dominate force on the Left in Mexico.  Only way to avoid this would be be for PAN to have a dynamic candidate that can somehow get up to above 33% support to become the main alternative to PRI-PVEM.   Or else something like El Bronco running a grand anti-PRI anti-AMLO alliance candidate.

AMLO is convinced that the only reason he is losing is that elections are being stolen from him - because, of course, he is the only hope and the self-evident choice of all Mexicans, opposed only by the small cabal of corrupt plotters (which, of course, includes everyone who does not support him wholeheartedly). The guy is completely devoid of anything resembling introspection or self-criticism: and is truly, sincerely messianic about himself.

[jaichind]

Wait, now Nieto said that the evaluation assessments of teachers across the country will go ahead and will not be cancelled “for any reason”.  This after the government suspended the testing plan right before the election so the radical teachers union will not disrupt elections.  Now that the elections are over it is right back.

[jaichind]

This election cycle I am really interested in the dynamics of election alliances in the Lower House elections. I know that there are a lot of egos, personalities, and local historical rivalries involved but given I do not know that much about Mexican politics, I have no great insight into that.  I can only look at it from a mechanical game theory point of view and then look at what should the right alliance behavior have been and what actually took place.  This is interesting because Mexico uses one vote to convey both the FPTP victor AND the PR seat allocation.  This is fairly unique in the world although my own ROC where I am from used a similar system in its legislative election system from 1992-2004 with the difference that the constituencies are multi-member and not single member districts.

To come up with a set of "rules" on if Party A and Party B should form an alliance in a state or even a district, we have to assume

a) Party A and Party B has a primary objective of maximizing its own seat count and the secondary objective of maximizing its ally's seat count.
b) Party A, Party B and the voters roughly know the relative balance of power in terms of size of vote base of all parties contesting.

What that said, the set of rules on if Party A and Party B should form an alliance should be

a) If one of the two parties is sure to win the seat no matter what then they should NOT got into an alliance since the FPTP seat is locked up, the goal should to be to maximize the vote for PR allocation. A alliance in this case could lose votes from some core voters from either party that dislike the other party and may be motivated to not vote or vote for a third party.

b) If even in an alliance where the joint vote share of the two parties is not enough to be in contention to win the FPTP seat then there should be no alliance for the same logic a)

c) If being in an alliance the fusion of the two vote shares puts the united vote share to be in contention (top 2 finishers or top 3 if the top 3 are close in terms of vote share) AND not being in an alliance buts the two parties out of contention, then there SHOULD be an alliance between the two parties.  Besides the obvious point of the fact that the alliance could generate a FPTP win, there is also the danger that having the two parties running separately but neither in contention would mean that supporters of both parities might tactically vote for a third party that is in contention and as a result both parties lose out in the PR section.

d) The hardest is when one of the two parties is in contention but does not have a lock while the other one is not in contention.   Here the goal of winning the FPTP seat conflicts the goal of maximizing the PR vote.  All things equal I feel in cases like this forming an alliance makes sense since you rather try to lock in a win in FPTP.  Also the individual candidate  is more likely to try to push harder if he/she feels that party is doing everything to get him/her elected which in turn will also push up the vote share for the benefit of PR.

[jaichind]

So with these "rules" set out, I can now see how PRI-PVEM alliance making strategy worked out relative to my set of "rules."  I will look at it some of the interesting states.

1)  AGUASCALIENTES - here PRI-PVEM failed for form an alliance.  Here PRI does not have a lock on the seats so an alliance should have been formed.  PRI won 1 of the 3 by very tiny margin while PAN won 2 of the 3. Had an alliance been formed PRI-PVEM would have won 2 of the 3 and also make sure that the one that PRI did win did not risk being lost.  So here PRI-PVEM did not follow optimal behavior and paid for it.

2)  CHIAPAS - Here PRI-PVEM had an alliance and won all 12 seats.  But since in every seat the vote share of PRI and PVEM separately was bigger than every other party, the more optimal choice would have been no alliance and PRI PVEM have friendly fights.  This would have created tactical voting by supporters of other parties toward either PRI or PVEM and help push up PR vote levels.

3) DF - Here PRI-PVEM has an alliance in 16 of the 27 seats and no alliance in 11 of the 27 seats.  This seems fairly smart as in the 11 of the 27 seats, even a united PRI-PVEM was not in contention.  Only exception was DF 16th district where there was no alliance between PRI and PVEM but combining the vote share of the two parties did put them in contention but still would have lost.  PRI-PVEM actually won 3 seats in the 16 where they had an alliance.    So on the whole PRI-PVEM chose fairly optimally.

4) MORELOS - Here PRI-PVEM did not have an alliance.  The result was a disaster for PRI.  The Left parties are strong in MORELOS but the PRD Morena split is giving the PRI-PVEM a chance to break through.  PRI running by itself ended up with only 1 out of 5 seats.  Had PRI-PVEM formed an alliance then PRI-PVEM would have won 4 out of 5.  Instead PAN and PANAL each won a seat taking advantage of both the PRD-PT Morena split as well as the PRI PVEM split.

5) NAYARIT - No PRI-PVEM alliance but it seems PRI won all 3 seats easily by itself so the decision not to have an alliance was optimal.

6) OAXACA - No PRI-PVEM alliance.  PRI was able to win (or at least leading) in 7 out of 11 seats taking advantage of the PRD-PT Morena split.  But had PRI-PVEM formed an alliance PRI-PVEM would have won 9 out of 11 seats.  So optimal behavior implied that PRI-PVEM should have had an alliance in OAXACA all things equal.

7) TABASCO - No PRI-PVEM alliance.  Here PRI by itself won 2 out 6 seats taking advantage of the PRD-PT Morena split.  But had PRI-PVEM formed an alliance PRI-PVEM would have won 4 out of 6 seats.  So here PRI-PVEM should have formed an alliance.

8 ) TAMAULIPAS - No PRI-PVEM alliance.  Here PRI running by itself easily won all 8 seats.  So no alliance between PRI-PVEM was optimal.

9) TLAXCALA - No PRI-PVEM alliance.  Here PRI running by itself easily won all 3 seats.  So no alliance between PRI-PVEM was optimal.

So overall, the behavior for PRI-PVEM in AGUASCALIENTES, CHIAPAS,  MORELOS, OAXACA, and TABASCO was non-optimal.  While the behavior in DF,  NAYARIT, TAMAULIPAS, TLAXCALA  was.  
Most other states where PRI-PVEM had an alliance all things equal were optimal since anything else would have risked losing FPTP seats.  I am sure on a district by district basis perhaps a lack of PRI-PVEM alliance would have been better but that would be splitting hairs.  

[jaichind]

BTW, the fact that INE designed PREP to report percentages the way they reported is really stupid. This has fulled not only us here - but the media as well (headlines are out announcing PT demise). 

Sounds like you can give  Lorenzo Cordova  this feedback since it seems you do know him.

[ag]

BTW, the fact that INE designed PREP to report percentages the way they reported is really stupid. This has fulled not only us here - but the media as well (headlines are out announcing PT demise). 

Sounds like you can give  Lorenzo Cordova  this feedback since it seems you do know him.

I will But after the results are finished with. He is far too busy now.

[ag]

So, the oficial count is on.

There are supposed to be, in theory, 149,270 ballot boxes (casillas). Of these 211 were never installed and another 341 did not send the packages to the district headquarters (presumably, because something happened during the election day), leaving us with 148,718 boxes to be counted. So far, we have a count from 52.202 boxes, or 34.97% of the total of the theoretical count and 35.10% of what we should expect.

Vote shares (of the total/of the valid votes)

PRI  30.38% 31.87%
PAN 21.24%  22.27%
PRD 10.02% 10.52%
Morena 7.85% 8.24%
PVEM 7.26% 7.62%
MC 6.15% 6.46%
Panal 3.86% 4.06%
PES 3.09% 3.25%
PT 2.86% 3.01%
PH 2.06% 2.16%
Ind 0.52% 0.55%
write-ins 0.12%
null 4.51%

[jaichind]

BTW, the fact that INE designed PREP to report percentages the way they reported is really stupid. This has fulled not only us here - but the media as well (headlines are out announcing PT demise). 

Sounds like you can give  Lorenzo Cordova  this feedback since it seems you do know him.

I will But after the results are finished with. He is far too busy now.

You can point out to him that in India they also have NULL with NOTA.  But ECI reports all percentages with NOTA stripped out. Vote share should be in terms of vote support of all voter that wishes to express a support to a particular/party.  Russia does this "against all" although I am not sure they include "against all" in vote share calculations.

[jaichind]

Real count 40.88% in


Vote share

PAN          21.21
PRI           30.33
PRD          10.02
PVEM          7.24
PT               2.88
MC              6.21
PANAL         3.86
Morena        7.85
PH              2.07
PES            3.10
Ind             0.55
Other          0.12
Null            4.52

I am actually hoping PT gets below of 3% in terms of total votes but greater than 3% in terms of non-null votes so we get a ruling once and for all exactly what is the rule on this 3% threshold as far what should count and what should not.  Should NULLs count? Most likely not.  Should write-ins count ? Should independents count ?  This is the problem you have by having one ballot serve to purposes (FTPT and PR.)  Mexico is better of doing what Germany, Japan and ROC does.  Have 2 separate ballots.

[ag]

BTW, the fact that INE designed PREP to report percentages the way they reported is really stupid. This has fulled not only us here - but the media as well (headlines are out announcing PT demise). 

Sounds like you can give  Lorenzo Cordova  this feedback since it seems you do know him.

I will But after the results are finished with. He is far too busy now.

You can point out to him that in India they also have NULL with NOTA.  But ECI reports all percentages with NOTA stripped out. Vote share should be in terms of vote support of all voter that wishes to express a support to a particular/party.  Russia does this "against all" although I am not sure they include "against all" in vote share calculations.

Well, Russian reporting is hardly model

Mexico's problem with reporting relevant info is nothing new. If you recall, in 2006 PREP there was no report on the number of problem casillas. Therefore, PREP was saying 98%, but the numbers really came from 93%. This was picked up by AMLO as evidence of some sort of fraud. Only by comparing with the PREP results reported by the Nuevo Leon electoral institutes one could figure this out. These days they have the annotation in the federal reporting as well. But most of the time federal reporting is, actually, better. 3 years ago electoral institute of Baja Calirofnia was simply reporting ridiculous percentages throughout the night because they were unable to figure out the formula for computing percentages in Excel right I am not joking: the vote numbers were correct, but percentages went haywire. In a semi-numerate country this is what happens (and, unlike in India, in Mexico problems with numeracy extend to the elite). Then, again, in other countries the journalists would have calculated PT share before setting the headlines. At least, INE itself has its numbers straight.

[ag]

Real count 40.88% in


Vote share

PAN          21.21
PRI           30.33
PRD          10.02
PVEM          7.24
PT               2.88
MC              6.21
PANAL         3.86
Morena        7.85
PH              2.07
PES            3.10
Ind             0.55
Other          0.12
Null            4.52

I am actually hoping PT gets below of 3% in terms of total votes but greater than 3% in terms of non-null votes so we get a ruling once and for all exactly what is the rule on this 3% threshold as far what should count and what should not.  Should NULLs count? Most likely not.  Should write-ins count ? Should independents count ?  This is the problem you have by having one ballot serve to purposes (FTPT and PR.)  Mexico is better of doing what Germany, Japan and ROC does.  Have 2 separate ballots.

The law is very clear: nulls and write-ins do not count. Independents do. There is no ruling necessary. INE has not made any statements about PT losing register - it is just the media making the conclusion based on misreading the numbers. Anyway, you ARE going to get your wish, it seems

BTW, just to make things even more interesting, you realize that the totals accross districts and the nationwide PR totals are not the same? Guess why

[jaichind]

The law is very clear: nulls and write-ins do not count. Independents do. There is no ruling necessary. INE has not made any statements about PT losing register - it is just the media making the conclusion based on misreading the numbers. Anyway, you ARE going to get your wish, it seems

BTW, just to make things even more interesting, you realize that the totals accross districts and the nationwide PR totals are not the same? Guess why

My guess is the way the voter marked the ballot.  The voter might have market the ballot in a way to indicate support for a party but not the candidate.  So for FPTP purposes the vote should NOT count in the district but should count in the national PR count.

[jaichind]

Even as these results are coming in there is very little results from OAXACA and this is almost 3 days after the voting is done.  There must be major problems over there.  The only place worse than this that I am aware of, and I am sure OAXACA does not want to be lump in with this place, is Westchester County where I live where one can go weeks before the vote count is done and results concluded.

[ag]

The law is very clear: nulls and write-ins do not count. Independents do. There is no ruling necessary. INE has not made any statements about PT losing register - it is just the media making the conclusion based on misreading the numbers. Anyway, you ARE going to get your wish, it seems

BTW, just to make things even more interesting, you realize that the totals accross districts and the nationwide PR totals are not the same? Guess why

My guess is the way the voter marked the ballot.  The voter might have market the ballot in a way to indicate support for a party but not the candidate.  So for FPTP purposes the vote should NOT count in the district but should count in the national PR count.

Nope. You cannot mark the ballot without voting for a candidate. It is WHERE you vote. There is a small number of precincts for people traveling on election day to vote out of their district (e.g., in airports, bus stations, etc.). Votes from those precincts do not get added to district FPTP totals, but are included in the PR allocation.