POLLSTERS SAY HOLY WEEK POLLS MAY PREDICT NEXT

UNCLAS MEXICO 001002

SIPDIS

SENSITIVE
SIPDIS

E.O. 12958: N/A
TAGS: PGOV MX
SUBJECT: POLLSTERS SAY HOLY WEEK POLLS MAY PREDICT NEXT
PRESIDENT



1. (SBU) Summary: The Ambassador met on 2/21 with four
leading Mexican pollsters who, in a lively discussion found a
few things to agree on -- including the fact that Democratic
Revolution Party (PRD) candidate Andres Manuel Lopez Obrador
(AMLO) continues to lead the race. They also agreed that if
he holds his lead through Holy Week (April 10-14),it's a
good indicator the race will be his. The pollsters uniformly
expressed confidence in exit polling and quick counts on
election day, saying they will be accurate indicators of the
victor. But they disagreed on nearly everything else.
End Summary


2. (SBU) The pollsters included PRD pollster Ana Cristina
Covarrubias, National Action Party (PAN) pollster Rafael
Jimenez, Televisa pollster Roy Campos and TV Azteca pollster
Francisco Abundis, who with some argument agreed that AMLO
leads the polls, PAN candidate Felipe Calderon is second, and
PRI candidate Madrazo is third (although several argued that
the latter two are nearing a virtual tie.) AMLO's poll
numbers couldn't go up much more, they agreed, but certainly
had room to fall. If he held his current lead (at least five
points across the various polls) through Semana Santa -- the
race was virtually his, they concurred.


3. (SBU) But they didn't all think that was a likely
contingency. In a campaign already studded with a number of
scandals, Roy Campos suggested that there could yet be
revelations that could affect any of the candidates. Rafael
Jimenez stressed that, if the PRI imploded, the race could
become a more traditional two-way contest, and PRI voters
might look closely at the PAN. He also suggested that
candidate Calderon, who had been a little lackluster in his
campaigning recently, might regain the dynamism he had shown
in December. Abundis cautioned that many voters were now
choosing by candidate rather than by party, and decisions on
congressional lists (which will be finalized in March and
April) could impact the presidential race as well. Jimenez
agreed, noting that the PRI in Morelos had divided sharply
over candidate selection, and this would undeniably affect
the presidential race.


4. (SBU) Asked whether we should expect the numbers to come
closer together or spread farther apart, Campos cautioned
against modelling this presidential race against any previous
elections. There are a number of firsts in this race that
make it unique in Mexican experience, he added. There are
three candidates who, for the first time, each have a chance
of winning. It is the first time a non-PRI president defines
his role in the process -- and that may affect the outcomes.
It is the first election after a dramatic -- but perhaps
unsatisfying -- transition in power. These and other factors
have an impact on the way voters think. Jimenez agreed, but
suggested that last July's Mexico State election reflected
the campaign tactics each campaign was using -- the PRI
relied on its reliable vote, the PRD spent little money on
advertising, the PAN candidate started strong then fell
dramatically. The decisive factors there, Jimenez said, were
the PRI's reliable vote combined with low voter turnout
overall delivered a PRI victory. In Veracruz, Jimenez
continued, there were three strong candidates until the very
end -- and different polls favored different candidates. The
PRI finally won, but by a small (and controversial) margin.


5. (SBU) The pollsters also debated the role and size of
the independent vote. While some estimated as much as 40
percent of the electorate was independent, others suggested
it was as low as 20 percent. These voters, Covarrubias said,
largely belonged to AMLO. Campos agreed, but noted a
percentage were more likely to switch votes and were clearly
the PAN target as well. They came around again to the idea
that by mid-April the voters will have settled to a larger
degree on their candidates, and the polls will better reflect
what might happen on July 2.


6. (SBU) All four pollsters expressed absolute confidence
in exit polling and quick count procedures that will be used
on July 2. They said that in Mexico the pollsters shared
information across party lines and company lines in order to
validate one another's findings. This had worked effectively
in 2000, Abundis said, when it was already clear by early
afternoon that Fox was the winner. Exit polls and quick
counts in 2000 were entirely consistent, and the pollsters
expected the same in 2006. There is no uncertainty with exit
polling, Campos said, because people know how they voted.


Visit Mexico City's Classified Web Site at
http://www.state.sgov.gov/p/wha/mexicocity

KELLY