Wednesday, June 1, 2011

Eurasia on Peru's election

Political risk people Eurasia are calling it all the following way. FWIW (not much) I don't agree with their call (and notably they've been calling it for Ollanta for weeks now, when most people were sticking with "too close to call") but WTFDIK and they could be right. Here's the mailer sent out to its private clients that was published today just before 1pm EST:

The latest polls suggest the 5 June second round runoff between nationalist candidate Ollanta Humala and center-right candidate Keiko Fujimori will be very tight. However, we still believe Humala holds a slight edge and is thus favored to win the presidency. While Fujimori is ahead in most polls, her lead has narrowed, which suggests negative campaign could be finally hurting her. In addition, Humala looks likely to win a larger share of undecided voters, who are mostly poor and therefore tend to be less concerned about the risks commonly associated with Humala. Despite the tight race, the risk of serious post-election political instability is low, even though a slow counting of votes could generate significant noise in the weeks following the election.
Most of the polls released over the weekend show center-right candidate Keiko Fujimori slightly ahead of nationalist candidate Ollanta Humala, but her lead has narrowed from last week. The latest Ipsos/Apoyo poll shows support for Fujimori at 41% (down from 43% in the previous poll) and support for Humala stable at 39%. In an Ipsos/Apoyo vote simulation, Fujimori had 50.5% of valid votes (down from 51.4% last week) and Humala 49.5% (up from 48.6%). A Datum vote simulation shows Fujimori with an almost five-point lead (52.3% vs 47.7%), but Fujimori's lead has narrowed slightly from previous polls. Another vote simulation conducted by the Catholic University of Lima (PUCP) shows the two candidates practically tied; Fujimori has 44.2% support and Humala has 43.7% (50.3% vs 49.75 in terms of valid votes). And an Imasen vote simulation shows Humala leading 50.8% versus 49.2%. The average of these polls (expressed in valid votes) shows Fujimori leading 50.6% versus 49.4%. Sunday May 29 was the last day pollsters could release their results before the 5 June run-off, although polls may be leaked later this week.

These results suggest the race will be very tight, but we remain of the view that Humala has an edge. Fujimori's narrower lead in polls suggests that negative associations related to her father's track record for corruption and human right violations, which have intensified in the past few weeks, could be hurting her bid. The increase in Fujimori's rejection rate over the course of May is further evidence that that this is occurring. In Ipsos/Apoyo's 9 May poll, Fujimori's rejection rate stood at 34%; it has increased is subsequent polls and reached 39% in the 29 May poll. In the meantime, Humala's rejection rate has fluctuated around 40%. Earlier in the month, his rate was five points higher than Fujimori's, but the difference has narrowed to just one point - the 29 May Ipsos/Apoyo poll shows Humala's rejection rate at 40%. The latest polls did not capture the impact on voters of last Sunday's televised debate. The impact will most likely be very limited, but if anything the debate will probably help Humala at the margin. He was able to expose clearly the negative aspects of Fujimori's father government and tie her to his administration. While these issues constitute a significant liability for Fujimori, they haven't received much media attention yet. From the perspective of undecided voters, this is what probably stood out in the debate.

Such campaign dynamics will help Humala gain a higher share of support from undecided voters, whose profile suggests many will lean towards Humala. Undecided voters are concentrated among poor, who tend to be less concerned about the risks commonly associated with Humala (that he will be a radical) and more sensitive to accusations of corruption. According to the latest Ipsos/Apoyo poll, the highest rate of undecided voters (9%) is in the "D" class which represents about a third of the electorate. Rates are just slightly lower in the other socioeconomic classes, but this indicates that poor voters constitute a majority of undecided voters. According to the PUCP poll, which we view as credible, the rates of undecided voters are comparatively higher among the poor. There are 8% of undecided voters in the "D" and "E" social classes, while only 4% of voters in the "C" class and 5% in the "A" and "B" classes, respectively, remain undecided. Additionally, the number of undecided voters could be higher that polls suggest. A poll conducted after the 2006 shows that 15% of voters chose their candidate on Election Day. With undecided voters more likely to lean towards Humala, this could tilt the election in his favor.

Despite the tight race, the risk of instability stemming from potential questioning of the electoral result is probably low. The more worrisome scenario here is one in which Humala loses by a very narrow margin. Some of his advisors expressed concern about the risk of fraud and could therefore raise doubts about the result. But Peru's electoral authorities are fairly independent, so the risk of widespread fraud that could justify serious questioning of the results looks low. In addition, Humala's willingness to question the results probably won't be strong. He has made the strategic decision to move closer to the center of the political spectrum, so serious questioning of the result could undermine the credibility of his move and thus his chances of winning future elections. And if he does decide to question the results, his ability to conduct nation-wide protests that paralyses large cities will be limited given that his supporters are concentrated in rural areas and probably lack the party structure and cohesion to sustain widespread protests. Fujimori has affirmed emphatically that she will accept the result. Despite the low risk of serious questioning, a slow counting of votes and/or potential requests of recounting could generate significant noise and doubts about the final result during the weeks that will follow the vote.


Erasto Almeida
Analyst, Latin America