A discussion paper on parallel counts and postponing elections

[You may download this discussion paper here.]

The erroneous counts reported during the final field testing of PCOS machines and traced to the gross misalignment of ovals on the local side of the ballots have shattered public confidence in the voting machines. As a result, two different proposals have been raised: one is to postpone the elections, and the other is to hold a manual count. However, the COMELEC seems to believe that the misalignment problem can still be remedied. Smartmatic is now trying to reconfigure memory cards to take into account the misalignment, and the COMELEC has made clear its intention to go ahead with the May 10 scheduled date of the automated elections, and to do so without any parallel manual counting.

Debate: Automated, manual or both? Postpone or not?

Thus, a debate now rages about the best option to take in the interest of a clean, honest and credible elections. As a contribution to this debate, let us clarify the various options and the pros and cons of each option:

  • Automated count is the process implemented by Smartmatic using PCOS machines to scan and count the votes cast as shaded ovals, with the results being transmitted electronically or transported physically to municipal/city canvassing centers
  • Manual count is the process managed by the board of election inspectors based on the old method of reading each vote aloud – now based on marked ovals beside candidate names – and recording the vote in tally sheets and the totals in manually filled-up election returns
  • Postponement means setting a new date for the elections one to four weeks after May 10 to give either the automated or the manual count or both enough time to complete the preparations, tie up loose ends, patch up problems, and generally ensure that the elections can be successfully held securely, accurately and credibly and a new set of elected officials proclaimed by June 30.

Here’s an exhaustive list of the different combinations of these three options:

No elections Manual Only Automated Only Parallel Count
No postponement No-El May 10 manual COMELEC NAMFREL
Postponement Postponed manual Macalintal Perlas
  1. No elections. Although some officials currently occupying elected positions or military junta advocates may wish for “No-El” as their dream option, it will not be considered an option here.
  2. Back to the manual count. This is the electoral process that cheats are now so familiar with that they can probably pad and shave votes in their sleep. While few will defend the old manual system of Philippine elections, it should be noted that votes in many other countries are counted manually through various processes that are generally acknowledged as clean, honest and credible. It should also be noted that even in the Philippines, a majority – perhaps up to 80% — of election jurisdictions conduct the counting, canvassing and consolidation of votes in a clean, honest and credible way, just as it is done in other countries. It is those few corrupt local and national election officials who chronically sell their services to equally corrupt candidates that give a bad name to Philippine elections. Because cheats are rarely prosecuted, much less get meted the punishment of imprisonment and perpetual disqualification from public office that await convicted election offenders, they will surely welcome a return to the familiar manual system, whether it is conducted on May 10 or a subsequent date.
  3. Automated only, no postponement. This is the COMELEC’s position. Their reasons have been well-reported in media: automation will solve the problems of delay and cheating; conducting a manual count will only confuse the public. delay the process, and provide an opening for the cheats to operate in familiar territory; it is still possible to provide the solution to the misalignment of local ovals, thus there is no need to postpone. Unfortunately for the COMELEC and Smartmatic, their credibility has been shattered by the gross errors of the PCOS machines. Winning back public confidence may take more than a few days. It is impossible now to look at a PCOS machine without asking the question: will it count accurately this time? Under these conditions, it is hard to imagine how can the machine results can be credible. Because a 10-ballot test set is inadequate for weeding out slightly inaccurate machines, and the voter verification feature of the machine has been disabled, this question about the true accuracy of the machine will remain unanswered.
  4. Automated only, with postponement. This is the Macalintal proposal. Many have reacted negatively to this proposal raised by the lawyer of Gloria Macapagal-Arroyo, who was herself caught in the act of micro-managing election fraud in 2004 with her field operator Comelec commissioner Virgilio Garcillano. Her party and the Ampatuans did it again in 2007, when they made Zubiri senator over the real winner Pimentel through statistically impossible results in six Maguindanao towns. That they got away with these crimes has hugely reinforced the sense of impunity among election cheats. Malacanang design is perceived to be behind the Macalintal proposal. Postponement is seen as a slippery slope that can lead to open-ended extra-constitutional scenarios that Malacanang is suspected of planning. The proposal also suffers from the same infirmity as the COMELEC position, with regards to its automated-only aspect. The public now knows that PCOS machines make mistakes too, and grievious ones at that. Once short-changed by an ATM machine, a depositor will want to carefully double-check every withdrawal made from these machines.
  5. Parallel count, no postponement. This is the NAMFREL proposal. The proposal for a parallel count has been raised earlier under a different name. Cong. Felicito Payumo was one of the earliest to propose a 100% manual audit. HALAL had likewise raised a version of the Payumo proposal. There have been variations, from HALAL’s president-only proposal, expanding the audit if significant discrepancies are found, to the business sector’s president, vice-president and mayor, to the Philippine Bar Association’s five-positions proposal. A parallel run is part of the usual industry methods of phasing in automation projects: pilot phase, dry run, parallel run, manual on standby, and so on. Given the shattered credibility of the PCOS machines when they gave grossly erroneous counts in the final 10-ballot testing, it now seems necessary for the machines to go through a parallel run before they can recover, if ever, this credibility. Unfortunately, the COMELEC remains adamant against a parallel count.
  6. Parallel count, with postponement. This is the Perlas proposal. Although Perlas wanted “up to three months”, variations that are being considered include a few days to a few weeks. Parallel run considerations have been covered above. The main concern about a postponement – especially since Perlas had said that “an extension of the Arroyo term was better than chaos and violence in the streets” – is that it would become open-ended and a prelude to the term extension of Arroyo, who seems ready and willing to do anything to hold on to power. This is the reason NAMFREL and others have called postponement “a dangerous option” and would not touch this option with a 10-foot pole.

Let us take a deeper look at both the parallel-run and postponement options:

Parallel count

The obvious role of a parallel run in any automation project is to mitigate the impacts of potential technology failures. In addition, a parallel-run will also make the work of cheats doubly hard, because they have to doctor both the machine and the manual counts. If the fraud creates significant discrepancies, public demand for explanation can trigger a cascade of investigations that may eventually uncover the cheating and expose the cheats. Because the technology is new, the cheats may not have mastered the fine art of electronic cheating, as they have manual cheating, hence keeping discrepancies to the minimum may be harder the first time around. For this reason, a parallel run will add to, not detract from, the credibility of an election outcome. With the shattered credibility of the PCOS machines (and Smartmatic/COMELEC as well), the COMELEC may have no choice but to agree to parallel run – unless it allows sheer stubbornness and unwillingness to admit mistakes color its decision-making. But then, it cannot anymore sweep under the rug any inaccuracy that the PCOS will probably still show, after their memory cards are recogfigured.

A credible parallel count needs more time

Unfortunately, the automated and the manual processes required for a parallel count are both not ready, less than a week before election day. The automated process has been under terrible time pressure from the beginning of the project. Today, that pressure has reached incredible proportions. Under this pressure, people will tend to make more wrong decisions, commit more mistakes, forget minor details (like the side-effect of changing ballot designs), and so on. Haste makes waste. The automated process has to work to turn the parallel run into reality. For this to happen, the technicians and engineers who are trying to make it work must be given enough breathing space and the incredible time pressure they are now working under must be eased up. HALAL has already pointed out from the beginning that the chance of success of the automation project was unacceptably low. Given the recent partial failures in the project, it is even lower today. Our analysis of the SysTest report shows that there are other problems that might have been masked by the gross errors displayed by the PCOS at the last minute. Most of these subtler problems cannot be corrected anymore. Their effect is to make the PCOS machine less accurate and more open to insider attacks, which, however, can be detected with a parallel run. At best, a postponement can give us PCOS machines that are only somewhat accurate, possibly 98-99% or even lower, that can still serve to flag the gross effects of pad-and-shave operations by cheats who have mastered the manual system.

Even a manual count, the other side of a parallel run, needs time to make it work. HALAL has a former COMELEC commissioner among its convenors, and Comm. Sadain estimates that at least three months is needed to fully prepare for a manual election. It is true that no additional ballots need to be printed, and 30% of the canvassing paraphernalia have been printed (if COMELEC statements are to be believed), but put the harried COMELEC personnel under more time pressure by making them produce and deliver the remaining 70% within a few days invites its own set of problems.

Thus insisting on a parallel run on May 10 may lead to having automated voting only in some areas, manual-only in other areas, both in still other areas, and perhaps none at all in a few areas – a formula for confusion and chaos.

It seems obvious, from a technical perspective at least, that a little extension will do both the automated and the manual aspects of a parallel run a lot of good, if only to ease up the terrible time pressure that election workers have been working under for several months now.

Is postponement constitutional?

HALAL considers the constitutionality of any postponement another important factor to consider. If legal luminaries all agree that any postponement will be unconstitutional, then it is pointless to postpone. Any winner proclaimed in a postponed election can be questioned before the Supreme Court and we are back in political limbo.

This is what Article VII Section 4 of the Philippine Constitution says: “Unless otherwise provided by law, the regular election for President and Vice-President shall be held on the second Monday of May.” While the date leaves no room for flexibility, initial qualifying phrase seems to create an loophole that can make postponement constitutional if a law provides for it. Is there such a law?

Section 5 of the Omnibus Election lists the grounds for postponing elections: “any serious cause such as violence, terrorism, loss or destruction of election paraphernalia or records, force majeure, and other analogous causes of such a nature that the holding of a free, orderly and honest election should become impossible in any political subdivision.” Again, the reference to “other analogous causes” might be interpreted as another loophole that can justify postponement.

Postponement: slippery slope

Although this matter can probably be settled with finality only in the Supreme Court, the larger concern against postponement is not its constitutionality but that it may create the opening for Malacanang term-extensionists, junta advocates, and other plotters and conspirators. A fifteen-day postponement (as proposed by Villanueva and Estrada), can easily become a month or more. Beyond a month, issues of constitutional succession kick in. In addition, volatile public mood primed for elections can turn frustrated and angry, spill over into the streets, and provide perfect excuse for a declaration of emergency and even martial law. Public anger, a mailed-fist response, and the chaos may be interpreted by the Left as a new revolutionary upsurge, triggering a new vicious cycle of confrontation politics. Indeed, postponement may turn out to be a slippery slope into a political blackhole.

Overall, then, we have been pushed into a corner, with no viable option left.

This was all triggered by a suggestion of changing a single-spaced layout to a double-spaced layout. Several questions come to mind: did the suggestion come from the Smartmatic or the Comelec side? Why did the Comelec approve the suggestion? When Smartmatic requested a second round of PCOS tests to double-check the new layout, why did the Comelec disapprove the request?

The above questions will have to remain unanswered for the moment. The next question may be easier to answer: who benefits from the shattered credibility of the automated elections? In fact, since the Comelec refuses to consider a parallel run or a 100% manual audit, then the credibility of the electoral process itself has been shattered.

Can the 2010 elections still be credible?

The answer may lie on the irreconcilably antagonistic relationship between Arroyo and survey frontrunner Aquino. The latter’s campaign promise to investigate the former’s misdeeds has led Arroyo to reportedly adopt an “anyone but Noynoy” policy. Since Arroyo was herself caught cheating in 2004, and again led her party in the 2007 cheating for Zubiri through the Ampatuans and other ARMM warlords, can we be faulted for expecting that she will use – perhaps even improve on – these tactics again in 2010? Threatened with prosecution and possibly a similar fate as Estrada, or even Marcos, of course we can expect Arroyo to use every means fair or foul to make sure frontrunner Aquino does not become president.

But cheating has its limits too. Consider Estrada’s case in 1998. The administration party had prepared the time-tested methods of election fraud against Estrada. But because he led by a landslide, they did not dare. The cheating would been too obvious and still would have failed to accomplish its goal anyway. Consider again Obama’s case in 2008. The Republicans were ready to cheat through the voting machines, most of which were controlled by Republican vendors/supporters. But because of Obama’s landslide lead, the Republican cheats didn’t dare. But the close contests of 2000 and 2004 are now seen by an increasing number of people in the U.S. as stolen elections.

If Malacanang believes in the surveys and anticipates a landslide win for Noynoy, the cheating option is not viable anymore.

But another option is left: to disrupt and discredit the electoral process itself, robbing of victory what Malacanang anticipates to be the most probable winner of the 2010 elections, without taking efforts anymore to help the runner up overcome the frontrunner’s lead through fraud. That is exactly what the innocuous ballot layout change has accomplished.

This analysis is speculative, but it explains the events quite well. All we need to do now to test the speculation is to trace where the suggestion to change the ballot layout came from.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: