Predicting the unpredictable: the primaries result

If in previous elections looking at historic trends could allow me to do some electoral predictions, for these primaries it is basically a useless task.  And yet, who can resist?

In the past elections I kept doing better and better at predicting results getting to a rather amazing 69 seats for the opposition in 2010.  But then the trends existed even if the CNE figures were not quite fully reliable (after all we are still waiting for the 2007 referendum final and complete result).  This time around, since primaries are for president, governor and mayor, and since there is no consistency in who supports whom, it is simply impossible to do a prediction based on previous returns since almost at every circuit the parameters will be slightly different from next door circuit.  Still, it is a worthy exercise in that I can decide from such a study that the only ones on paper that can win are Capriles or Perez.  True, with one month of campaign left, implosion is still possible.  But after almost two month of candidate vetting it seems unlikely that Capriles or Perez will collapse, though it is possible that Lopez will sneak into second place, beating Perez by a hairbreadth.

How did I reach such a conclusion.  I will not worry you with the tables I built on Excel (not to mention that this being such a speculative venture I am in no mood to have a table done in good faith be slammed at me in a month from now).  Thus you will only see a summary of my method and the results I got.

First I assumed that that the results of 2010 Latin american Parliament vote as given by the CNE are legit.  Since those are the only ones nation wide to carry the political party symbols it is the best suggestion we have.  And yet not complete since for example Convergencia did not post a card and thus trumped the Yaracuy results for my prediction (but Yaracuy being small, who cares really).

Second I divided the votes by region:
- Oriente: Nueva Esparta, Sucre, Anzoategui, Monagas and Bolivar (this last one because I think that for primaries it will behave roughly like the other states of the list)
- Llanos: Apure, Barinas, Portuguesa, Guarico and Cojedes (though I even questioned worrying about them because they are heavy chavista and the most likely to have the highest abstention).
- Occidente: Zulia, Tachira, Merida and Trujillo (Trujillo really does not matter much and the other ones together because in these three states UNT and COPEI will behave more predictably, simplifying my calculations).
- Centro: Falcon, Lara, Yaracuy and Carabobo (because Carabobo is the only state where Proyecto Venezuela is strong enough to make the difference and it goes for Lopez in theory)
- Caracas at large: Vargas, Libertador, Miranda and Aragua (because it is the stronghold of Primero Justicia)

In third I started giving abstention coefficients to each party according to each region.  For example AD will have in general more voters absent from voting stations because it does not have its own candidate.  However where AD has a strong candidate for a governor primary its abstention would be lower, benefiting Perez.  Or PPT who I doubt that more than half of its voters will go and vote in a primary.  These coefficients were at least 30% going  up to 50%.

Then, the fourth parameter is an absolutely subjective coefficient on which party is more able to convince all of its voters to vote for the supported candidate.  For example I have strong doubts that COPEI voters will follow massively Perez when their more natural fit would be Capriles, Machado or even Lopez.  However UNT and PJ voters should be more consequent.  Still, I gave every party a desertion coefficient of at least 10%, up to 60% in some states.

Finally I added all in four groups, Capriles, Perez, Lopez and "uncommitted" meaning a sum of small parties in the CNE result, the "unity" cards in those ballots (MIN and UNAPARVE) and other subjective criteria that I pleased to have (my blog, my criteria).

The results are as follow (all rounded up, of course)

Participation: in 2010: 5,1 million (5,6 with PPT and OPINA)
Participation in one month from today: 3,2 million

This is quite good if you ask me.  In previous posts I wrote that anything above 2 million would be good and I think that we can make it to 3 million.  Of course chavismo will consider anything that does not reach 50% of registered voters a disaster for the Unidad, but let them eat cake.  Anything that reaches half the vote of normal elections is considered anywhere as a primary huge success. Note: in 2010 there were 17.5 million registered voters and an abstention of 34% with 11,3 million valid votes.  So 3 million votes represent almost 30% of the voters and any number above that can only mean that chavistas have gone to vote in the primaries too..........  hence the campaign of discouragement that the regime is undertaking to make sure that participation will not go above the 2 million that they could still sort of explain.

I then give:
Capriles 993.000
Perez 1.195.000
Lopez 279.000
Uncommitted 695.000

On paper Perez is winning but all polls so far give Capriles ahead.  Thus I am wrong in estimating the loyalty to Perez of AD and Copei voters (even some UNT who would have preferred Rosales).  Also my uncommitted number might be too high as many may have migrated to Capriles already.  Thus in all subjectivity I am passing 100.000 to Capriles from Perez and 50.000 from uncommitted.

Capriles 1.143.000
Perez 1.095.000
Lopez 279.000
Uncommitted 645.000

And lo'and behold, we do sort of get the latest complete poll we had in October!  30% Capriles, 25% Perez and the rest, the rest, with Lopez ahead of the pack.

From this exercise we can see that Lopez is too far behind on paper, and political parties loyalty, to overcome the lead or Perez and Capriles.  Even if the 645.000 were to go all to him he would still not reach the million votes and at best could only challenge for second position.  And in the 645.000 we can be certain that Medina will get at the very  least 50.000, Arria 100.000 and Machado 150.000 (though each one is supposed to get at least 200.000 from the people that signed for their candidature).

However my gut feeling does not give me these numbers.  I think that Capriles and Perez will indeed not go much above 1.000.000 each because their campaign is not attracting the hard core opposition and chavista will vote for them in October but not necessarily this time around, too afraid of being caught voting in the primaries at work or at the mision.  In other words they are at their peak already and their campaign playing it safe, I do not think they can grow much (though the Perez campaign seems suddenly more reactive, having endorsed today the no-re-election policy and the Unity card for October, very popular issues among hard core opposition).

The ones that are motivating the hard core opposition voters are Lopez, Machado and Arria and they will bring possibly an extra 500.000 votes to the 3 million I calculated, and split it among themselves.  So these three may have as much as 1,4 million votes to share, making a surprise victory by one of them a possibility if either Capriles or Perez make some significant faux-pas in the next 4 weeks.  Though in all conscience I doubt that at this point Capriles can lose unless Perez manages to revive dramatically.

AGAIN: this is not an official prediction, it is at best an educated gut feeling on how the electoral base for the campaign moves.  And with 4 more weeks to go.  It is a mere starting point for the trends to come, illustrating the challenge for each candidate.  I will refine this model and include polls as they come available and as I know more about some local governor races of influence (Miranda and Anzoategui in particular).  Then, and only then, may I dare make a prediction.