Will manual recount showmore votes than voters?

Comelec chief Sixto Brillantes will never allow a manual recount of even only the automated senatorial voting. He will do all to block it, from stunning his critics with the cost (“Pay up P200 million first”) to having presidential spokesmen speak for him (“We preferred to sweep the entire slate but we didn’t”). A manual recount is forbidden. For, it not only will confirm the statistically dubious 60-30-10 percent outcomes of winners and losers in all regions. It could also expose that there were more votes than voters.

The discrepancy of votes and voters is the reason why both the precinct counting and the official canvassing were never completed. No politician is questioning for fear perhaps of the powerful Comelec, critics say.

To recall, Brillantes on Election Day, May 13, declared a low 65-percent voter turnout, 33.8 million of the 52 million registered voters. It was only a midterm balloting, he said.

The next day the seven Comelec commissioners convened as the national board of canvassers. Sluggishly they started with the advanced overseas votes, since undisclosed kinks were delaying the transmissions of local results to the central server. Then suddenly on Thursday, May 16th, they proclaimed six senators, and on Friday the 17th three more.

The Comelec proclamations were on the basis of only 72 of the 304 certificates of canvass (CoCs), or 23.7 percent. The 72 CoCs represented only a little over 13 million ballots of the 52 million total voters and 33.8 million turnouts. Still, the Comelec insisted that the margins were too wide for the nine “winners” to be overtaken, based on “group canvasses” pouring in.

Experts decried the Comelec move as contrary to Constitution and law. Ignoring them, the Comelec proclaimed the last three senators on Sunday the 19th.

By then, the Comelec had canvassed 129, or 42 percent, of the 304 CoCs. On the night of Saturday the 18th, it posted on its website the partial tally of the 33 senatorial candidates (see Table 1). It said the 304 CoCs contained 39,898,992 votes. That turnout wasn’t only 65 percent, but a hefty 77 percent, like in a presidential election (the turnout in 2010 was 74.34 percent). That partial tally stayed in the Comelec website for three weeks, till June 6.

On June 7 the Comelec declared that the tallies of the 175 more CoCs, the 58-percent balance of the 304 CoCs, had finally arrived. So it posted on its website the new final votes (see Table 2).

And this is where the mystery begins:

The votes of the 13 frontrunners increased by mere tens of thousands, not millions of votes, from the last posting of May 18-June 6 to the new posting of June 7. Consisting of 58 percent of the CoCs, the increments should have been double the last tally of May 18-June 6, not loose change. Never mind who the biggest recipients were of additional votes. That the highest increase was only 195,473 and the lowest was 72,270 is very odd.

Given such extremely low increases, the logical explanation would be that the No. 13 to No. 33 candidates got tens of millions more votes by June 7. But they didn’t. The standings of the Top 12 stayed as is; Gordon remained in No. 13, where he had been since May 18-June 6.

It would appear from the wee increases that the Comelec never completed the canvassing of the 175, or 58 percent, remaining CoCs. Presumably it merely adjusted the figures to conform with recent past elections, where the senatorial top-placer got roughly 50 percent of the votes, and so on down the ladder.

(A source analyzed that the first batch of 129 CoCs comprised 62 percent of the turnouts, and the 175-balance 38 percent. He totaled the votes of top Ranks 1 to 12, second cluster Ranks 13 to 21, and “also ran” Ranks 22 to 33 [see bottom of Table 1, totals in parenthesis]. He multiplied these by 12, the maximum possible number of senatorial marks per ballot, and extrapolated the percentages from the Comelec’s announced 39,898,992 total turnouts. I have not obtained the source’s permission to identify him. Mathematicians might wish to make their own analyses.)

Earlier (Gotcha, 22 May 2013), I reported that the Comelec’s so-called “transparency server” also did not finish the precinct tallies. Manned by the poll watchdog PPCRV, the server on May 20 stopped the tallying at 59,667 of the total 78,166 precincts nationwide. It did not bother to complete the tallies of the 18,499 other precincts, or 23.4 percent, consisting of eight to 11 million voters. This oddity came after an equally mysterious spurt of 12 million – technically impossible – tallied votes on the first three hours after precincts closed on Election Day. This was attributed to a supposed “computer programming error.”

Based on both the (unfinished) precinct tallies and the official canvasses, critics of the Comelec noticed the 60-30-10 percent uniform outcomes of winners and losers in all regions. That is, 60 percent for the Top 12, 30 percent for the next 12, and 10 percent for the mostly independent also-rans. The AES (Automated Election System) Watch said the percentages incredibly disregarded cultural differences and political preferences, such as the bloc-voting Bicolanos, the tribes folk and the Muslims of the Cordillera and Mindanao autonomous regions, and the traditionally oppositionist metropolises. –Jarius Bondoc (The Philippine Star)