“Data analyst Phil Daily presented historical trends and ‘suspicious ratios’ in the recent election results”
THE group U.S. Filipinos for Good Governance (USFGG) held an online press conference yesterday and presented the analyses of two experts probing “possible irregularities” in the May 2022 Philippine national elections.
The speakers were forensic election fraud expert and University of Michigan political science and statistics professor Walter Mebane and seasoned big data analyst and engineer Phil Daily.
The rationale of the USFGG for their press conference is that “fair elections are the foundation of any democracy,” adding that “when the legitimacy of elections is called into question, democracy is threatened.”
While the presscon was more of an “academic exercise,” according to Erik Lachica of USFGG, “we are preparing for the midterm elections in 2025 and the [presidential elections in] 2028.”
It’s “lessons learned,” he added, that will enable the taking of steps to ensure “a fair and accurate elections in the midterms.”
Lachica also said that based on the analyses and recommendations, there should be a move from fully automated vote counting to a “hybrid” type of vote counting “where we do manual counting at the precinct level to ensure accuracy and transparency.”
Due to lack of space, I’ll be reporting on Daily’s analysis first, and Mebane’s next week.
Historical trends in senator vs. president ratios
Daily discussed his analysis of the correlation between presidential votes and senatorial votes from the recent election and past elections, and the “improbability” of the 2022 election results count.
Using “all published data and published numbers,” Daily presented historical trends and “suspicious ratios” in the recent election results.
Historically, he said senators tend to get more votes than the presidential candidate, and this was the trend in the 1992 (Ramos), 1998 (Estrada), 2004 (Arroyo), 2010 (Aquino), and 2016 (Duterte) elections, where the top 15, top three, top four, top five, and top three senators beat the winning presidential candidate in terms of number of votes.
However, Daily said that in 2022, Marcos Jr.’s votes were around 15 percent more than Robin Padilla’s, the top winning senatorial bet.
This has never happened before, he said, adding, “We vote for 12 senators and just one president. This is why, historically, the senators have more votes than the president, not the other way around.” He can’t explain, he says, why this time around, Marcos Jr. got more votes than Padilla.
Daily also presented a slide on “suspicious ratios” in terms of Padilla votes versus Marcos Jr. votes.
These numbers show that Padilla had more votes than Marcos Jr. in only 20 percent of the provinces. Historically, Daily said, “it’s usually the other way around – it’s usually 80 percent of the provinces, the senators will have more votes than the winning president.”
However, in 73 percent of the provinces that Leni Robredo won, Padilla had more votes than Marcos Jr., in line with the historical trend.
It was only in 10 percent of the provinces that Marcos Jr. won that Padilla had more votes. So in 90 percent of the provinces, Marcos Jr. garnered a huge margin of votes versus the top senatorial bet, bucking the historical trend.
“It’s possible that this can happen,” Daily said, “but it’s highly improbable.”
In a slide labeled “more suspicious ratios,” Daily showed that in the 2016 elections, the top senatorial bet, Franklin Drilon, beat Marcos Jr. in over 80 percent of the provinces, while in 2022, Padilla beat the latter in only 10 percent of the provinces. The ratio in 2022, Daily said, is half that in 2016.
However, when compared to Robredo’s ratio, in 2016 and 2022, the ratio of senator votes to her votes were very similar, with the ratio in 2022 being in fact higher than in 2016.
Daily says the historical ratio of votes of top senators to winning president is between 1.1 to 2.23. In 2022, the ratio was 0.85.
He added if the ratio were adjusted to historical trends, the adjusted votes for Marcos Jr. would be dramatically different – it would not be 31 million.
If the ratio were 1.25, more in line with the historical trend, Marcos Jr. would have 20 million votes; in other words, he would lose 11 million votes.
Now, if half those 11 million votes were actually Robredo votes, “then the election results would have been, Leni, 20.5 million, and [Marcos Jr.], 20.1 million,” Daily explained.
In sum, Daily said, the question he sought to answer was, “Was the 2022 Philippine presidential election an example of a perfectly executed campaign, or the greatest scam ever?”
His answer is, “Unfortunately, we will never really know.” Short of, perhaps, Robredo asking for a recount, which is costly at P500 or so per voter.
In any case, Daily said the reason for the presentation is “not to overturn the election results” – “it’s too late for that” – but rather, to “mention election integrity and democracy.”
In my column next week, I’ll be reporting Professor Mebane’s “e-forensics” analysis of the 2022 elections, as well as the implications of such analyses in the future. (To be continued)
*** FB and Twitter: @DrJennyO
