MANILA, Philippines—Automated elections are here to stay.

This was the pronouncement of Jose Melo, chair of the Commission on Elections (Comelec), after results of a random manual audit of ballots from the May 10 polls showed a 99.6-percent accuracy.

Henrietta de Villa, chair of the Parish Pastoral Council for Responsible Voting (PPCRV), yesterday presented a report on the audit of 1,145 clustered precincts, which declared that the results of the automated elections were accurate.

The report said that none of the presidential, vice presidential, congressional, gubernatorial and mayoral races audited in the selected precincts “failed the accuracy test of 99 percent.”

Melo said he was heartened by the findings, which compared the results of the voting machines with the manual count conducted by auditors.

“It is a very encouraging report. It means that the machines worked 99 percent,” the Comelec chief said.

That losing politicians conceded after preliminary results were known a day after the polls closed was a sign that the automated election system was trustworthy, Melo said.

“Automation is here to stay. As long as we have the money, we will automate,” Melo said.

Critic skeptical

But one of the information technology critics of the Comelec is skeptical of the audit results.

Augusto Lagman said the Comelec request for a proposal from the private sector for an automated election system had mandated that the voting machines have a 99.99-percent accuracy rate.

Lagman said the findings of the random manual audit “do not meet that requirement.”

Discretion

De Villa said the PPCRV could not use the accuracy threshold for the voting machines because the audit was subject to a host of external and human factors.

“The auditors used discretion in reading the votes. The machines did not have that,” she said.

In addition, the National Statistics Office (NSO) determined the allowable margin of variance for the audit, according to De Villa.

Delay questioned

Lagman also questioned the delay in the release of the results. “I’m not happy about it. Why only now? That itself is a failure,” he added.

De Villa said the audit team had several logistical and procedural problems that dragged the audit for several weeks.

The random manual audit only managed to count 1,046 clustered precincts, which translated to about 500,000 votes. De Villa said the remaining precincts were not able to send their results to the Comelec.

There were some 75,000 clustered precincts during the May 10 general elections.

Incomplete

Although the audit was incomplete and took longer than expected, De Villa and Melo said the results were more than enough to establish the integrity of the voting machines.

The automated elections law says that the accuracy rate of the count must be 99 percent, allowing a variance of 1 percent or a difference of 10 votes for every 1,000 valid ballots cast.

Variance refers to the difference in the machine count and in the audit of the results by teacher-auditors enlisted by the Comelec to examine the ballots from the participating precincts.

NSO verification

Counts with a difference of at least 10 votes were sent to the NSO for root-cause verification, which determined the errors. Some errors, officials said, were due to miscounting by the auditors.

NSO representative Florante Varona said the presidential, vice presidential, House, gubernatorial and mayoral races had an accuracy rate of 99.6 percent. “There was a variance of 0.4 percent,” he said.

Varona said the variance was lower than the 1-percent threshold.

The audit report said the variance for the count in the presidential race was 0.49 percent. In the vice presidential race, it was 0.40 percent. The combined variance for the local posts was 0.37 percent.

The audit report also showed certain problems in the counting that has to be raised with the Comelec.

Discrepancies in Manila

The report said there two-digit discrepancies in four precincts, high number of rejected ballots in two precincts, and “large” variances for the Manila mayoral race that was due to a voting-machine error.

“The validation team could not identify the reason for the large variances, especially for the position of mayor, even after the retrieval and opening of the ballot boxes,” the report said.

Upon inspection of the ballot images, it was revealed that there was a vertical line running through the local ballots that was not seen in the paper documents.

Foreign object in scanner

The report said Smartmatic TIM Corp., the supplier of the automated counting machines, told the team that a “foreign object” could have been lodged in the scanner, resulting in inaccurate reading.

When the machines were inspected in the presence of watchers, “a fiber-like substance was found inside the machine. This case appears to be machine, and not procedural, error,” it said.

Atienza protest

Former Manila City Mayor Lito Atienza has filed an electoral protest alleging errors, discrepancies, irregularities and fraud in the results generated by the voting machines in all 1,441 clustered precincts in Manila and their transmittal to the City Board of Canvassers.

Atienza yesterday expressed apprehension and dismay over the delay in the proceedings of his election protest against Mayor Alfredo Lim in light of the large discrepancies the manual audit uncovered in the city’s mayoral race.

Atienza asked the Comelec, where his protest is pending, to expedite the release of the order for the immediate retrieval of the boxes containing the ballots of certain precincts so the revision and recount of the votes could start.

Difference of votes

The audit team said it was up to the Comelec to decide on a handful of cases that it could not explain.

The random manual audit committee said there were a difference of 24 votes in a precinct in Pambujan, Northern Samar; 13 votes in Tuburan, Basilan; 14 votes in Gerona, Tarlac; and 18 votes in Tarlac City, Tarlac.

There were 99 rejected ballots in Lutayan, Sultan Kudarat, and very low valid votes in Maluso, Basilan.