Can Comelec counting machines really count?

Benjamin Abalos is a cautious chief of the Commission on Elections. When he couldn’t get at once the needed P1 billion to prepare for next year’s absentee voting, he said he’ll just drop it. He got congressmen and senators so scared of a backlash from seven million overseas workers that they gave him the money in a flash. Before the recent bidding for poll automation, he told prospective suppliers he wants only the best at the least cost. "There will be no experimenting," he warned, "Comelec cannot risk a revolution because of a wrong choice."

And so he got the best. Sagem, a French defense and FBI contractor, won Phase 1 to fingerprint and photograph each voter for the verification system. It outdid equally qualified data-capturing and validation expert Unisys of America. PMSI, with its satellite network, won Phase 3 over MegaPacific to transmit precinct tallies to canvassing centers in a snap.

But what about the intervening Phase 2, the automated quick-count of ballots? Shh, don’t look now, Mr. Abalos, but do you notice that long-time Comelec officer snickering behind you? Ask him what happened to the bidding in which MegaPacific this time beat Technology Information Management (TIM). No, don’t bother. Ask instead the Dept. of Science and Technology, which you tasked with testing the counting machines.

It appears from records that both bidders failed the technical tests – grounds to disqualify them both and declare a failed bidding. Not all the tests using US-Federal Election Commission standards were conducted. After a few, Comelec bidding managers declared them qualified anyway, then hurriedly opened their bids. TIM’s was P1.298 billion for 2,272 counting machines, but only for Metro Manila and Mindanao since it admittedly can’t manufacture more in time for May 2004. MegaPacific’s was P1.249 billion – P49 million less – but for only 1,769 machines to be put in "strategic polling centers" nationwide. If provinces and cities want more, they’d have to buy the machines themselves. Yet MegaPacific won because it is cheaper – seemingly.

Can it really do a bigger job with 503 fewer machines? MegaPacific’s papers show that it will be financed and supplied by SK Group of South Korea. That same SK Group was discovered recently to be bankrupt. Its majority-owning chairman Chey Tae-Won was jailed for illegal stock trading. He and ten SK Group executives are presently indicted for Enron-style doctoring of books to inflate earnings. Creditors have moved in on SK’s many manufacturing subsidiaries. The latest to be choked, going by a report in The STAR business section of Apr. 25, is SK Global, the trading arm that presumably will supply MegaPacific’s counting machines. Like its mother firm, SK Global inflated earnings by $1.2 billion to hide losses.

DOST technical testers saw a sample of MegaPacific’s machine – and it was slow. In its bid papers MegaPacific presented a prototype of a faster model. This, despite Abalos’s guideline against experimenting. The bidding rules specified that the supplier should have been contracted and its machine should have been used in at least one election of at least 20 million voters. It even asked for ISO-9000 certification for a unit that is in production for the past five years. No proof of use in an election or ISO certification was submitted. For, no prototype can be in mass production for the past five years.

The slow model did not go through all the required tests. Weeks before Election Day the machines would be transported by land, sea and air. The unit thus should have gone through a vibration test, but didn’t. The machines would be operated in decrepit precincts by strangers. So the model should’ve gone through a drop test, but again didn’t. The machines would be stored between elections under harsh weather conditions. The unit should’ve been tested for durability from heat, rain, humidity and dust, but wasn’t.

The sample was tested for actual counting, though. And what a test it was. MegaPacific’s machine conked out thrice, and was restarted only after six Korean technicians tinkered with it. Six times 1,769 means the Comelec would need an army of 10,614 Korean technicians deployed nationwide for breakdowns. (Is Korea in the SARS list?)

The testing was inconsistent. MegaPacific was allowed to prepare its own ballots. Its technicians thus carefully shaded the blocks. TIM, which also includes Unisys and used the Election Systems & Softwares model for US polls, was not afforded the same opportunity but merely given ballots filled out by Comelec personnel. That was a headstart for MegaPacific, for on Election Day more than 35 million voters would shade their ballots in their own varying ways. In short, there would be no two ballots alike. MegaPacific’s machine counted the ballots only when fed frontside-up and topside-first, never inverted or rotated, unlike TIM’s which read the votes any which way. MegaPacific’s unit didn’t detect if a ballot was fed the wrong way; it just read the votes of the frontside-up, topside-first ballots.

This means that MegaPacific’s machine needs outside intervention. Poll clerks would have to read each ballot and stack them up the right way before feeding to the unit. That would be risky. A clerk may invert or rotate a ballot if he doesn’t like the vote, and the machine won’t detect a thing. Even if all poll clerks were angels, there’s still the possibility of error in feeding, and again the machine won’t notice.

TIM’s expensive ES&S is in use in all US polling precincts. (The Florida snafu was due to faulty ballot design, not the machine.) Unlike its stand-alone machine – that is, the hardware and software are rolled into one system – MegaPacific’s unit can only run with a PC. That means it needs yet another outside intervention - one that fraudulently can be manipulated - to consolidate the tallied ballots. Sort of hi-tech dagdag-bawas (adding-shaving) of votes.

If MegaPacific’s machines fail on Election Day, Sagem’s elaborate voter verification system and PMSI’s instant transmission would be for naught. The NPA would have a grand time, though. Abalos’s feared revolution could come true.

* * *

Catch Linawin Natin, Mondays at 11 p.m., on IBC-13.

* * *

You can e-mail comments to [email protected]