UPDATED. In a tiny little bit of a Colombia syndrome, Dilma Roussef, Lula's handpicked successor will have to fight it out for a second round vote in a few weeks. In fact, with 70% of the votes counted there is the real possibility that she will reach barely 45%, thus making it possible for Serra to beat her in the second round, though by no more than a point or two even if the surprising Marina Silva scored much better than expected as the third runner.
True, due to some recent scandals in the Brazil government polls were suggesting that Roussef would miss by a few votes a first round victory. But 45% is not "a few missing" and thus polls got it wrong like they got wrong the first round in Colombia's election. Not as bad of course, by far, but bad enough to start wondering about polling techniques in countries as broken physically as Colombia or Brazil.
Overall Lula has had a good administration and one should not be surprised that whomever he supported as the best successor for his polices should be leading in the polls. But arrogance has a price.
First, Lula did not want to hear anything about primaries and what not. It was Dilma or else, and surely this affected the voters, explaining in part why Marina Silva did much better than expected as many disgruntled Lula supporters turned to his former ally (no PPT debacle here, as in Venezuela).
Second the scandalous involvement of Lula in the campaign was a clear strategy at turning the election into his "third term" plebiscite as well as an attempt to secure a first round victory for Dilma and thus leaving him in a good position to keep pulling a few strings in Brasilia. Now Dilma is going to need to win on her own the second ballot...
I understand that Marina Silva favors Serra but her voters come at least in large part from past Lula preferences. No matter what agreement she may reach with Serra it is doubtful that she will be able to transfer her 20% to Serra and thus we can expect at least a 5% to go Dilma or stay home. The corruption scandals in the entourage of Lula and Dilma are only starting to have an effect and are certain to become an issue in the campaign for the second round. Thus Serra has a real chance but I still give Dilma ahead, by no more than 2-3 points. That is, unless she manages a deal with Marina Silva and clears up the mess in her offices. All is possible. What is certain is that the Brazil vote has suddenly become a real election....
Still, we need to meditate some here. It is amazing that what is perceived here and in Brazil as a good administration gets a surprising low 45% whereas what is now considered an hyper-corrupt failed administration, even by many of its supporters that still did vote for it last Sunday, gets a 48% for Chavez. Makes one wonder about Messianic leadership when this one combines with electoral fraud.
UPDATE
Two things.
Observe that as I type this, at 8:15PM, most of Brazil's results are up on the web, including senators and governors. Were they in Venezuela, well, they woudl still have at least 5 hours more to wait. Look at the O'Globo page which even gives you the color scheme of won or lost state, something unthinkable from our local CNE....
I did not realize that for some reason Nordeste votes are slower in coming than for the rest of Brazil. And the Nordeste is a Lula bastion now. So it looks like Dilma will cross the 46% mark which not only helps her for the second round but also psychologically help her some by making her look that she "barely" missed the first round win.... Below 46 she cannot pull that one. Silly but that is the way electoral psychology works, the automatic "rounding up".