That was a scene from the HBO documentary, “Hacking democracy” where Howard Dean’s shown how election results can be altered by making use of security flaws in counting software.
After the Bush versus Gore election, Americans vowed they’d modernize their machines. In this year’s midterms, the new, modernized machines have become mired in controversy. America’s experience is so messy, do we Filipinos really need voting machines?
I’m Manolo Quezon, the Explainer.
I. Politics is addition
If you visit the blog Comelec AKO, you will see that election time’s around the corner. It begins on January 14, voting day is May 14, and the election period winds up on June 13. So the first six months of next year will be all about election fever.
Elections of course involve three main parts. The campaign. The election itself. And then counting. We’re used to the secret ballot, which has you writing your candidates on a piece of paper that the government takes custody of and counts. It’s secretly filled out by you, so that you can vote according to your conscience. And it’s counted with great ceremony –and recounted over and over again- to prove to you and everyone else, that what you’ve written is counted properly.
The 1987 Constitution declares as its First Principle that, quote, “The Philippines is a democratic and republican State. Sovereignty resides in the people and all government authority emanates from them.” unquote.
Every Citizen possesses an individual and equal particle of that Sovereignty which “resides in the people”. In a democratic republic, REPRESENTATIVES of the PUBLIC are chosen in elections governed by the rule, “One Man One Vote”. Thus the government authority which emanates from the People actually comes in discrete lumps called VOTES, and whoever in an election receives the highest number of votes, that Citizen receives the authority to set the course of the Ship of State, to run its day to day operations, to be Captain or Officer or Crew Member of the ship of state.
The problem comes with the rudimentary operation of counting the votes. As the population grows getting it right about the highest number of votes has become a logistical nightmare. The polls are vulnerable to illegal forms of addition and subtraction, and even high tech forms of decimal multiplication and division during canvassing.
In 2004 over 32 million citizens cast ballots during the national elections. There could be as many as 45 million registered voters in 2007 casting their ballots at over 250,000 polling precincts belonging to over 1600 Municipalities or Cities in 80 to 81 provinces and one autonomous region.
Philippine national elections are a veritable blizzard of paper forms, ballots, statements and certificates, smudgy fingerprints and barely legible signatures. These are individually processed by boards and committees and canvassed for the votes of six or seven Presidential candidates, up to a hundred Senatorial candidates, up to a thousand Congress and party list candidates, and tens of thousands of other elected government officials. The amount of manual labor required is staggering.
Consider only what is officially required on Election Day: a Board of Election Inspectors composed of three public school teachers or “citizens of known probity” for each of the quarter of a million voting precincts. That’s 750,000 just at the first stage of the manual vote count, a number that increases directly with the number of voting precincts.
Why it should take six weeks to count and canvass the votes and proclaim a winner makes us wonder what century we are in, or if our goose is getting cooked in the interval.
Unfortunately, MODERNIZATION and AUTOMATION are easier said than done.
For example, in 1997 the Congress passed Republic Act 8436, the Election Modernization Act of 1997, “an Act which authorizes the Comelec to use an automated election system.” But nine years and billions upon billions of pesos later, there is no automated election system, no automated registration system, no automated counting machines system, no automated transmission system. Huge controversies, spectacular failures, shocking exonerations and stunning Supreme Court Decisions have marked, or marred, the course of election modernization of the last decade or so.
The historic January 2004 Supreme Court Decision, ITF vs. Comelec, which voided a billion peso contract with MegaPacific Consortium for Automated Counting Machines, records for posterity our collective failures in this regard.
The awesome task had fallen upon the Commission on Elections which gets little gratitude and plenty of derision, deservedly or not, from the citizens, many of whom do not actually vote, or if they bother to do so, they do not check before election day if they are registered and where their voting precinct might be located.
Now it doesn’t help very much that the Comelec cannot publish Voters Lists and precinct locations in time to be useful, and that it employs people like Virgilio Garcillano, or that the Supreme Court finds it to be a serial grave abuser of its discretion when it tries to implement naively crafted laws…
But you see how the blame game goes as old fashioned chaos and discord have trumped new fangled electoral reform concepts like automation modernization.
Is there a way out of this sorry mess? The House and Senate are working on new legislation to do just that. But what is it that we missed the last time? What are the questions we did not know enough about to even ask? What is the explanation for the failure to modernize our Ship of State’s operating system for selecting its Captain and Crew? We’ll try to answer these questions when we return.
II. Garbage in, garbage out
That was another scene from the HBO documentary, where they demonstrate how a voting machine can get the counting of votes wrong.
But WHY have WE failed to modernize our electoral system?
It is certainly not for a lack of wanting to or trying hard enough.
Perhaps it has something to do with this bit of graffiti in a Massachusetts Institute of Technology restroom purported to be an immutable Law of the Information Age:
IN ANY LARGE COMPLEX COMPUTER PROJECT, EVERYTHING THAT CAN GO WRONG WILL GO WRONG.
That certainly applies to Republic Act No. 8436, the Election Modernization Act of 1997, since every phase of the Comelec’s implementation of it—registration, counting and transmission—have basically come to NOUGHT. It is not entirely the Comelec’s fault as that enabling law itself was admittedly flawed in certain very critical ways.
So, what mistakes should be avoided in the future? What lessons ought to be learned from our abject failure to finally create a credible and reliable electoral system so we can try again and succeed?
One lesson is that trying to merely mechanize or roboticize the motions of the old system by replacing human beings and manual processes with computers and automation is not as straightforward as it would seem. In fact there are hidden and surprising pitfalls when we try to replace Man with Machine.
Take for example the matter of Automated Counting Machines to replace the function of tallying ballots usually done by a Board of Election Inspectors, composed of three public school teachers at every voting precinct.
It was but natural for Comelec to specify an Accuracy Rating for any Automated Counting Machine (ACM) to be qualified for the automation project. In this case, the ACM’s were required to have a 99.995% accuracy rating, a requirement that is not imposed on the Human Counting Machines – the three public school teachers that make up the Board of Election Inspectors– that currently perform the counting function at every voting precinct now.
The matter of the Accuracy Rating creates two new problems that don’t even exist in the Manual Election System:
(1) The ACMs have to be tested to show that they conform to the accuracy rating requirement, whatever that rating is. Comelec ran afoul of the Supreme Court on this matter in 2004, because of a major arithmetic problem with decimal points in its bidding specs. Later we shall discuss this with a proponent in the famous case of ITF vs. Comelec.
(2) But a second problem is this. With an accuracy rating is 99.995% (a fraction equal to 99,500/100,000) Comelec is willing to tolerate as many as 5 erroneously processed ballots per 100,000 so that in an election of 50 million voters there could be as many as 2500 erroneously counted votes. But under the manual system, what we expect is 100% accuracy.
So why this tradeoff? Because speed is supposed to be beneficial. And perhaps because people tend to think computers don’t make mistakes.
But the less than total accuracy that counting machines are expected to achieve presents a novel Constitutional issue. The issue’s with respect to the Principle of One Man One Vote. Because, if the Comelec relies on such ACMs to count the votes, Comelec cannot detect a TIE in any election that can only be broken by Congress and it cannot guarantee to declare victory for any candidate that wins by less than 2500 votes.
Although only the Constitutional fundamentalists will be happy about this, it appears that an Automated Counting Machine with an Accuracy Rating less than 100% probably violates the fundamental democratic principle of One Man One Vote.
The use of any ACM with an accuracy rating less than 100% is unconstitutional on its face. This may seem like a nasty technicality but remember that US Pres. George W. Bush became President because of an incredibly tiny 500 vote margin in the State of Florida during the 2000 US national elections. And many local races here at home can be decided by an equally small margin of votes.
The transcendental Constitutional issue of preserving One Man One Vote in any automated system cannot be ignored. It represents an important intellectual challenge to proponents of automated elections and lawmakers who may be tempted to see Automation as a Deus Ex Machina for Modernization.
This problem of modernizing the elections will clearly take more than throwing computers and cell phones at it because of another Bitter aw of the Information Age:
People are smarter than computers, but computers are smarter than programmers. Our guests will help us come to grips with this statement.
In 2004, the Social Weather Stations estimated that 900,000 Filipinos were disenfranchised, which means they couldn’t, or weren’t allowed, to vote. That’s as big a scandal as the controversies that have surrounded the national results since. It’s as important to be able to vote, as it is to have your vote accurately counted.
It’s also necessary for us to stop fearing our fellow voters.
In “The Vote of the Poor,” recently published by the Ateneo’s Institute of Philippine Culture, our humblest fellow citizens tell us something interesting. Most will take money to vote for a certain candidate, but as long as the voting involves the secret ballot, most intend to vote not for the person they’re paid to vote for, but the person they want to vote for. And this to me explains why the manipulation of voting results has to take place through padding and shaving the results after precinct results are tabulated.
And let’s accept that on a basic level, our democracy functions well enough.
Democracy works on the precinct level. That seems something everyone agrees on. It breaks down in the intervening counting and re-counting and reporting on the municipal, provincial, and national levels.
So where can we do better?
Our democratic ingenuity would be better devoted to finding a way to have our precinct voting results directly tabulated, and the precinct totals reported quickly. There is no reason why it should take more than 24 hours to know the result of any contest, whether local or national.
But until that happens, do ask yourself: are you registered to vote? Have you already started thinking of your criteria for voting? And will you vote? And what can you do, to help make sure election day and the counting days are properly held? Here’s a good place to start: it’s a website called “Botante Kami” (http://botantekami.wetpaint.com/). See you there.