LATEST APOYO POLLING INDICATES GARCIA'S LEAD OVER

UNCLAS LIMA 002146 SIPDIS SENSITIVE SIPDIS E.O. 12958: N/A TAGS: PGOV PINR PE SUBJECT: LATEST APOYO POLLING INDICATES GARCIA'S LEAD OVER HUMALA NARROWING TO FOUR POINTS WITH UNDER A WEEK TO GO REF: A. LIMA 2096 ¶B. LIMA 1887 Sensitive but Unclassified. Please protect accordingly. ¶1. (U) SUMMARY: According to a simulated secret vote carried out by the Apoyo consultancy, APRA's Alan Garcia's lead over Union por el Peru's (UPP) Ollanta Humala in the presidential race is down to four points (52-48 percent). Simultaneous polling based on face-to-face interviews found the difference to be 10 points (55-45 percent),which tends to confirm our suspicions that Humala has a large "hidden" vote that has not/not been reflected in the polls. Consequently, the outcome of the presidential elections remains uncertain as the campaign enters its final week. END SUMMARY. ¶2. (U) Apoyo carried out its simulated vote and poll from 5/24-26 in 175 districts located in 77 different provinces nationwide. Apoyo's surveys do not/not claim to represent approximately 19 percent of the population that lives in isolated rural areas. ¶3. (SBU) The results of the simulated vote, in which eligible voters cast secret ballots, differed markedly from the responses to the normal face-to-face interviews in several respects: -- The percentage of blank/null/no response votes was only 13 percent in the simulation, as opposed to 20 percent in the face-to-face poll, demonstrating that at least seven percent of respondents had determined who they will vote for, but were reluctant to identify their preferences to a pollster. -- In the face-to-face poll, Garcia posted a slim 51-49 percent lead in Arequipa and the southern coast, an area in which he lost heavily to Humala in the 4/9 first round. In the simulated vote, however, Humala trounced the APRA candidate by 56-44 percent (COMMENT: a much more believable result given the first round figures - also see Ref B for our prior doubts about polling results from this region. END COMMENT). -- In the simulated vote, Humala posted large leads in the central sierra (57-43 percent) and the Amazonian region (56-44 percent),whereas in the face-to-face poll Humala's advantage was just eight percent in the central sierra and Amazonian region. (COMMENT: Again, the simulated vote is more in line with first round voting patterns. END COMMENT.) ¶4. (SBU) Garcia's lead in the face-to-face poll fell from 12 to 10 percent in the "valid vote" count (excluding blank/spoiled/no preference responses) over the past week. Nonetheless, he and his APRA followers can take heart from the poll's finding that his total support rose during this period, from 41 percent to 44 percent of those polled. Humala's total support also rose by three percent (from 33 to 36 percent). Consequently, the reduction in the level of Garcia's "valid vote" lead did not represent a gain in Humala's vote count, but rather reflected a roughly equal shift of formerly undecided voters to the two candidates that had a larger statistical impact on Humala's totals. Garcia should also be pleased that 50 percent of respondents said that they "definitively would not" vote for Humala, while 40 percent had a similar negative view of the APRA candidate (up from a questionable six percent last week - Ref A) . ¶5. (SBU) COMMENT: Apoyo's simulated vote tends to confirm our suspicions that there is a sizable "hidden" vote for Humala in his Arequipan, Andean and Amazonian strongholds, which has not/not been reflected in the normal face-to-face polls (some of which have Garcia ahead by nearly 20 points or more!). Consequently, the presidential contest appears to remain tight going into the final days of campaigning. Garcia's chances of holding off Humala in the race to the wire are bolstered by Apoyo's finding that he is benefiting equally with Humala from the shift in undecided voters, as well as by the UPP candidate's continued higher negative rating. END COMMENT. STRUBLE