Professional Documents
Culture Documents
cenSEI
: . +
8KVUXZ
CONTENTS BUSINESS NATION WORLD TECHNOLOGY HEALTH/LIFESTYLE
Fixing the Fix
A credible Random Manual Audit process is crucial
for an automated election system to deliver credible
results in the May 2013 elections
By Zandro G. Rapadas, M.D.C.
STRATEGY POINTS
The May 2010 automated national and local elections were regarded
as a landmark achievement in the countrys history, but while
the results were generally accepted, serious doubts have arisen
regarding the accuracy of the automated election system
Doubts center on the faiIure of the post-eIection audit procedure,
or the Random Manual Audit, to check accurately the results of
automated counting
The direct participation of the Commission on Elections and its
citizens arm in the audit process, combined with the exclusion
of other observers, has cast doubt on the integrity of the entire
automated counting process
AAA ccrreeddiibblee Randoom Mannuual AAuuuuddddit processss iiss ccrruucciiaal
for an aautomated election ssyyyysssstttem to deliver creddiibbllee
results iin tthhee MMaayy 22001133 eelleeeccccttiionns
By By ZZan andro G. Rapadas, M.D.C.
NATION
PREVIOUS PAGE NEXT PAGE
14
:NKcenSEI8KVUXZ - July 23-29, 2012
Fixing the fx
It is the policy of the State to ensure free, orderly, honest, peaceful, credible,
and informed elections, plebiscites, referenda, recall, and other similar
electoral exercises by improving on the election process and adopting
systems, which shall involve the use of automated election system that will
ensure the secrecy and sanctity of the ballot and all election, consolidation,
and transmission documents in order that the process shall be transparent
DQGFUHGLEOHDQGWKDWWKHUHVXOWVVKDOOEHIDVWDFFXUDWHDQGUHHFWLYHRIWKH
genuine will of the people.
- DecIuruLIon oI PoIIcy, SecLIon 1, RepubIIc AcL q6q or un AcL umendIng R.A. 8q6
entitled An Act Authorizing the Commission on Elections to use an Automated
Election System
T
he 2010 automated national and
local elections in the Philippines were
collectively lauded as a landmark achievement
in the nations history of democracy and
suffrage, but continue to be hounded with
unresolved issues and serious doubts on their
accuracy and credibility.
From technology and security concerns
involving the hardware, software, and process of
the automated election system (AES) provided
by vendor SmurLmuLIc-TM Lo LIe nILLy-grILLy
oI ImpIemenLuLIon, udequucy oI resources,
and preparation time for the Commission on
Elections (Comelec) and its partner institutions,
the issues raised by various local election
stakeholders, international observers, and IT
experts left more to be desired from what was
touted as the cure to our manual election woes.
The general acceptability of the results
notwithstanding, the 2010 automated elections
have turned out to fall short of expectations.
Checking automated election results.
JusL us LIere ure sysLem hxes Lo uddress
errors or data corruption in personal computers
-- unLIvIrus, dIsk cIeunup, und deIrugmenLuLIon
progrums -- LIere Is u buIIL-In IeuLure provIded
by law to check on the accuracy and reliability
of automated election results: the random
manual audit (RMA).
SecLIon zq In RepubIIc AcL 8q6 provides for a
posL-eIecLIon uudIL mecIunIsm In LIe AES:
Where the AES is used, there shall be
a random manual audit in one precinct
per congressional district [emphasis
ours] randomly chosen by the
Commission in each province and city.
Any difference between the automated
and manual count will result in the
determination of root cause and
initiate a manual count for those
precincts affected by the computer
or procedural error.
Kathy Dopp, in her 2010 dissertation at the
Rockefeller College of Public Affairs and Policy,
CIeckIng EIecLIon OuLcome Accurucy: PosL-
election Audit Sampling Methods, dehnes
posL-eIecLIon uudIL us u cIeck oI LIe uccurucy
of reported election results done by manually
counLIng uII LIe voLer-verIhubIe puper buIIoLs
associated with randomly sampled reported
initial vote counts, and checking such additional
records as necessary to ensure the integrity of
the electoral process.
Moreover, in its Final Report, the Carter Center
Limited Mission to the May 2010 Elections in
the Philippines noted that in an automated
system such as the Smartmatic AES, the burden
of building public trust in the results rests
PREVIOUS PAGE NEXT PAGE
15 cenSEI
: . +
8KVUXZ
CONTENTS BUSINESS NATION WORLD TECHNOLOGY HEALTH/LIFESTYLE
IurgeIy on LIe uudIL process. PosL-eIecLIon
audits must be both comprehensive and
weII-ImpIemenLed, In order Lo verIIy LIuL
the machines performed as intended and
were not subject to any manipulation or
programing error.
Too much, too late. On April 30, 2010,
just 10 days before the elections, the Comelec
promulgated the General Instructions for the
Conduct of a Random Manual Audit (RMA) for
the May 10, 2010 National and Local Elections.
Comelec Advisory Council member Henrietta
de Villa, who heads the Parish Pastoral Council
for Responsible Voting (PPCRV), which was the
Comelecs accredited citizens arm in 2010, was
also the chair of the Technical Working Group
on LIe Rundom MunuuI AudIL (TWG-RMA),
with a representative from Comelecs Internal
AudIL OIhce und LIe NuLIonuI SLuLIsLIcs OIhce
(NSO) as members. The NSO slot was originally
for the Commission on Audit (COA), but the
latter withdrew in March 2010, citing possible
conIcL oI InLeresL.
The general instructions for RMA also
increased the number of clustered polling
precincts to be audited from 1 per legislative
dIsLrIcL (us provIded Ior In R.A. q6q) Lo hve,
Ior u LoLuI oI 1,1q (Irom LIe orIgInuI zzq). AIso,
the elective positions to be audited included
onIy LIose Ior PresIdenL und VIce-PresIdenL on
the national level, and Governor, Congressman,
and Mayor on the local level.
The Provincial Election Supervisors (PES)
implemented the RMA in the legislative
districts, including designating members of the
audit team. The auditors were members of the
Board of Election Inspectors (BEI) from the
clustered polling precinct that was immediately
after the randomly selected precinct or before
it, should the latter be the last precinct in the
voting center, or from the next barangay should
the randomly selected precinct be the lone
precinct in the voting center.
AccordIng Lo TWG-RMA, seIecLIng unoLIer
set of BEI to do the manual audit in the
randomly selected precincts ensures a blind
audit, given that it precludes the possibility
of any attempt to reconcile the manual
count with the AES count, as noted in their
terminal report submitted to the Comelec on
July 20, 2010, a copy of which was obtained by
The CenSEI Report.
Prolonged selection process leads to
delays, errors, confusion. The selection of
LIe 1,1q poIIIng precIncLs wus done munuuIIy
by druwIng pIng-pong buIIs Irom u tambiolo (a
roLuLIng recepLucIe used In ruIe druws), wILI
one ball assigned to a clustered polling
precinct per legislative district, as witnessed
by representatives of various election
stakeholders. The draw was done from
10:30 a.m. to 7 p.m. on May 10 at the
Philippine International Convention Center
in Pasay City (where Comelec set up its
national canvassing). However, the
prolonged process delayed the transmission
of the selection results and instructions to
the audit teams, and there was confusion
us weII us severuI gIILcIes, us reecLed In
reports from Davao City and Quezon City,
the former being the largest city in the
country geographically, and the latter in
terms of voting population.