LET'S DESIGN AN AUTOMATED ELECTION SYSTEM… TOGETHER

By Gus Lagman
Published, Manila Times Online

Introduction
While we are still almost three years away from the 2019 elections, it is, in fact, the best time to start planning for it.There are very valid reasons why it has to be started this early. Many observers, IT practitioners, mostly, have been commenting that the system used in 2010, 2013, and 2016 – i.e., Smartmatic’s Precinct Count Optical Scan, or PCOS – while fast, did not, however, meet the required accuracy level, was definitely not transparent, and was very vulnerable to tampering by an insider. All these have been proven to be true.

These failures could all be traced to the fact that the Commission on Elections (COMELEC), quite apparently, did not go through the standard project analysis and cost/benefits study before deciding on the system to be used in all three elections. This important step is a must, especially for major projects costing billions of pesos.

After the 2010 elections, the Chairman of the COMELEC Advisory Council (CAC) said in his report that we should not use PCOS again. The then Chairman of the Committee on Suffrage and Electoral Reforms in the Lower House, said in his report that if the loopholes in the PCOS are not plugged, “… a reversion to manual elections with heightened vigilance by organizations like PPCRV and NAMFREL would probably yield more credible and accurate results”. Why Smartmatic continued to win the contracts is something that we fail to understand. Private companies would have given them the boot long ago.

It is feared that if nothing different is done this time around, then the same decisions would be made by COMELEC; the same doubts, from those in the know, on the system’s unreliability would be generated, the same wasteful spending would occur.

This writer proposes to undertake this study, together with the public … and, hopefully, with the COMELEC as well, if it will so agree to participate. Through a series of articles in this paper, this writer will present to the readers the different aspects about elections and the options that we could take, in order to arrive at the “most appropriate” automated system for our elections, as required by law. Contributions to the discussion would of course be appreciated.

We will begin by reviewing the old, pure manual system, then discuss its defects, problems, and pitfalls. After that, we will scan the environment for automation election tools that are currently available, discuss the pros and cons of each, and then proceed to designing and formulating a workable system.

The pure manual elections

The pure manual elections (no automated component at all) that we used from our very first election in the 1940s up to 2007 followed a very simple system that required no training of voters at all. Common sense and the ability to read and write were almost all one needed. Those who cannot read and write and those with disabilities were allowed to be assisted by a close relative. Even the Board of Election Inspectors (BEI) hardly needed any training; most of them only received written instructions that would come with the usual election paraphernalia distributed to them a few days before the elections.Here’s how the system works:

Precinct voting and counting
The BEIs man a precinct each. A BEI is composed of three teachers, one of whom is assigned as Chairman and the other two as poll clerks. Approximately 200 voters (maximum of 250) are registered under each precinct. When this system was last used in 2007, there were a little more than 200,000 precincts that opened nationwide.

The election process consists of stages – the voting period, the vote-counting, the transport of the ballots and results of counting, referred to as Election Returns (ERs), and the three-level canvassing – municipal, provincial, and national consolidation of votes.

The voting period would start at 7:00 AM and close at 3:00 PM. The BEI, however, would start working as early as 5:00 AM in order to prepare the classroom and the election materials for the voters’ registration and actual voting. As many as 15-20 voters would usually be allowed to vote simultaneously or as many as could be comfortably accommodated inside the classroom. Even as the voting closes at 3:00 PM, all voters making up a queue of up to 30 meters long, would still be allowed to vote.

After a brief rest following the close of the voting period, the BEI would start counting the votes. The first step would be for the Chairman to count all the ballots inside the ballot box. Once done and the count recorded, the Chairman would start reading off aloud, the candidates’ names written on each ballot. One of the members would record the votes into the Election Returns using “taras” (or sticks), while the third member would record the same votes into the Tally Boards that are taped or stapled to the walls of the classroom.

When all the votes have been read and recorded, the BEI would then add up the total votes garnered by each candidate in both the Election Returns and the Tally Boards, making sure that both reports have the same totals. The BEI would then complete the Minutes of the proceedings, sign all the forms, then pack the documents for transport to the canvassing center.

The BEIs of the precints in a polling center (usually, a school) would normally share a jeepney, or jeepneys, when transporting the ballot boxes and other documents to the city/municipal canvassing center. It is also safer for them to wait for each other, as most of them would finish the counting in the very early hours of the following day.

In the first installment of this series of articles, I proposed to re-design an automated election system for the Philippines, hopefully, with some participation from the readers. The first step in the process is a discussion of the pure manual system, starting with (in the first article) the precinct voting, precinct counting, the transport of the results of such counting, and in this article, the fourth and final step, which is the three-level canvassing, or consolidation of votes. (Some people refer to this as “ladderized” canvassing.)

City/Municipal canvassing

In order to arrive at the winning candidates for mayor, vice-mayor, and councilors in each city or municipality, the votes garnered by each of the candidates for those positions in all the precincts in that city or municipality, would have to be consolidated. This is what is referred to as city/municipal canvassing and this first level of consolidation is performed by the City/Municipal Board of Canvassers (CMBOC).

This Board is composed of the Election Officer or a representative of the Commission as Chairman; the Municipal Treasurer as Vice-Chairman; and the District School Supervisor, or in his absence, the most senior Principal of the school district, as Member. The Board will normally wait for a sufficient number of Election Returns (ERs) to arrive before it starts its work. This is so the canvassing work would be continuous and would have a minimum of interruptions caused by the non-arrival of ERs.

Aside from the city/municipal positions, the votes of the candidates for national, provincial, and district level positions are canvassed, as well, to prepare them for further consolidation at the next two levels of canvassing.

Two forms are used at each level of canvassing:

1) The Statement of Votes (SOV) – where the total votes garnered by the candidates from each precinct are entered. The precinct numbers are recorded at the top of the columns. The columns in the form are filled up as ERs arrive from the precincts. When all the precinct votes have been recorded into the SOV, totals by candidate are computed and recorded in figures and in words on the rightmost columns of the form.

2) The Certificate of Canvass (COC) – is a summary of the SOV. The total votes per candidate appearing on the rightmost columns of the SOV (figures and words) are entered into the COC.

The COC is then transported to the Provincial Board of Canvassers (PBOC) for further consolidation.

Provincial canvassing

As in city/municipal canvassing, in order to arrive at the winning candidates for governor, vice-governor, members of the provincial board, and district congressmen,the votes garnered by each of the candidates for those positions in all cities and municipalities in each province, would have to be consolidated. This is what is referred to as provincial canvassing and this second level of consolidation is performed by the Provincial Board of Canvassers (PBOC). (There are some variations to the procedures on account of multi-district provinces and chartered cities.)

The PBOC is composed of the Provincial Election Supervisor or a lawyer of the Commission, as Chairman; the Provincial Prosecutor, as Vice- Chairman; and the District School Superintendent, as Member-Secretary.

At this level of canvassing, an SOV (provincial version) is used to facilitate the consolidation, except that this time, the columns represent the votes in the COCs coming from all CMBOCs in the province, instead of, in ERs from precincts in a city/municipality. Totals by candidate are computed and recorded in figures and in words on the rightmost columns of the form. A Provincial Certificate of Canvass (PCOC), which is a summary of the Provincial SOV, is prepared.

Aside from the provincial positions, the votes of the candidates for national level positions are canvassed as well, to prepare them for further consolidation at the third and last level of canvassing. The PCOCs are then transported to the National Board of Canvassers (NBOC) for the third and final consolidation of votes for national candidates.

National canvassing

To arrive at the winning candidates for president, vice-president, senators, and party-list, the votes garnered by each of the candidates for those positions in all provinces and chartered cities (as appearing in the PCOCs), would have to be consolidated. This is what is referred to as national canvassing and this third and final level of consolidation is performed by the National Board of Canvassers (NBOC).

The NBOC is split into two groups of canvassers. The first – a joint Congressional committee composed of representatives coming from both Houses of Congress – canvasses the positions of president and vice-president, while the second, the Commission on Elections (COMELEC) canvasses the positions of senators and party list.

In each canvassing center, an SOV (national version) is again used to facilitate the consolidation, except that this time, the columns represent the votes in the PCOCs coming from all provinces and chartered cities in the country. Totals by candidate are computed and recorded in figures and in words on the rightmost columns of the form. A National Certificate of Canvass (NCOC), which is a summary of the national SOV, is prepared.

After the NCOCs are completed, the winning national candidates are proclaimed, thus ending the election period.

The pure manual system that I have just described is what the Philippines used from the first election in 1946 up to the elections in 2007.

Problems in the ‘pure manual system’

In the first two installments of this series, I discussed how the pure manual system that we used in all elections from 1946 to 2007, worked. This third installment will discuss the problems associated with that system, eventually becoming the reason for automating the process in order to mitigate them.

The biggest problem in the pure manual system was the many weeks it took to complete the process and proclaim the winners of the elections. In the last three decades, it would take more than 20 days before the winners would be proclaimed. Once, it even took all of 42 days to complete!

It was not a problem during the country’s election exercises prior to martial law because the number of registered voters then, even at its highest, was less than a fourth of what it is today. But the population grew during the 14 years of martial law, and consequently, so also did the number of registered voters. Let’s take a look at some historical figures (on Table A are the election dates while those on the right are the corresponding registered voters during those years):

The election period became longer and longer, as the voting population grew. The increase in the time period was however not caused by the precinct level activities, as more precincts were simply opened in order to accommodate the growth in the number of voters. The longer time period was caused by the longer canvassing activities.

To explain this further: let’s say that there are one million voters registered and 5,000 precincts opened to accommodate 200 voters per precinct. Let’s also assume that it took the precincts an average of eight hours to count the votes and produce the Election Returns. If the number of registered voters increased by half-a-million, and the number of precincts was increased by another 2,500, then presumably, it would still take just eight hours to count all the votes in all the precincts.

On the other hand, the increase in the number of precincts by another 2,500 will proportionately increase the canvassing time at the City/Municipal Board of Canvassers by 50 percent as there will still be only one Board that will do the consolidation.

But there were other reasons that caused the slowness of the process. Before 1992, local elections used to be run separately from national elections. As such, in each election, there were fewer positions/candidate names in the ballot that needed to be counted and canvassed. In 1992, however, the national and local elections were synchronized and that naturally resulted in an increase in the number of positions and names in the ballot that needed to be counted and canvassed. Counting at the precincts and the eventual canvassing therefore took a much longer time.

And then, there’s also the inclusion of the party-list candidates. Averaging almost a hundred candidates in every election, counting and canvassing became even more cumbersome.

The long election process caused other problems, too. Big-time cheating became easier as the delay provided the cheats with sufficient time to do their evil deeds. Thus, “dagdag-bawas” (literally, add-subtract) came into being. This occurred at the canvassing levels, particularly at the city/municipal canvassing and at the provincial canvassing.

The teachers, who composed the Board of Election Inspectors (BEI), were put at risk as they would end up going home in the early hours of the following morning. At that time, they would have put in almost 24 hours of back-breaking work, as they would have been up since 5 a.m. or earlier on election day.

During the transport of the ballot boxes and the election returns from the precincts to the City/Municipal Board of Canvassers, there had been occasions when the boxes would be hijacked and stuffed with fake ballots, or replaced with pre-stuffed boxes. Or, simply destroyed.

There had also been reports that the BEI chairman would sometimes intentionally misread the ballots, thus favoring certain candidates. This problem was substantially reduced when political party watchers were allowed to position themselves behind the BEI as the latter conducts the counting.

Truly, because of the enormity of the operations, individual problems had become very difficult to solve. Thus, the automation of the election process became the “end-all and be-all” solution. Expectations were high. And the general public believed that such expectations were met. But alas! Those in the know believe otherwise.