The May 10, 2010 elections is perhaps information technology’s biggest moment in
Philippine politics. The automation of the counting of ballots is intended to speed up the
counting at the precinct level. It should be made clear though that the automation is
only up to the counting at precinct level. Everything else is the same process of manual
consolidation of ballot counts per precinct to municipal, provincial and then national levels.
Some sectors have suggested that a parallel manual count must be done but this is
unnecessary and very costly, not only to the government but to the political parties who
have to monitor the manual count per precinct. The random manual count to audit the
results of a randomly selected precinct is sufficient for the purpose of verifying the level of
error occuring which we hope should be close to zero. Also, let’s not forget that manual
counting is most likely going to be more error prone due to the biases of the persons doing
the counting.
The candidates can be confident they are elected, even with some level of counting error,
by making sure they have a significant lead. A lead of 10% should be quite secure even
with a 2% counting error. A lead of just 1% is just subject to naturally occuring error. It
is the responsibility of the candidate to make sure he leads by enough votes and that he
makes sure the votes he gets are protected from fraud after the automated counting.
Can the computer that counts the votes on the PCOS machine be programmed to favor a
certain candidate? Unfortunately the answer to this is yes. However, that is the reason
for having a random manual count to audit the results. Organizations watching the
elections should make sure the random selection of precincts where manual
counting will be done is truly random and that random selection is done only after
the official results are printed on all PCOS machines in the area. A candidate with
enough influence and financing can implement a dagdag-bawas scheme if the manual count
audit is not random.
Also, the results of the precincts with manual counts should be compared to other precincts
nearby to make sure the results are similar.
If the random manual count for audit is done properly, it would be impossible not to detect
an intentional rigging of the program doing the counting on the PCOS machine.
As Ronald Reagan used to say, “Trust, but verify”. We should trust COMELEC and the
technology they have implemented to speed up our election process. However, let us
verify the results and not be complacent. And lastly, let us hope candidates not elected will
concede gracefully and not blame technology for their situation after the elections.