Philippines 2022 Presidential Election

ANFREL's interim statement recorded the failure of at least 1,867 vote-counting machines or related SD cards on election day, affecting about 1.76% of clustered precincts, with some voters in the National Capital Region waiting over 12 hours. COMELEC told a Senate hearing on 31 May 2022 that 1,310 machines (1,295 domestic and 15 overseas) and 618 SD cards had to be replaced, a figure COMELEC also gave as its final count of machines that broke down, under 1% of the 106,174 deployed. ANFREL judged that up to one million voters had to queue for hours and that some were likely disenfranchised as a result, attributing the failures to the age of the equipment and to poor contingency planning and communication by COMELEC, while finding that they did not compromise the integrity of the election process.

ANFREL's interim statement records that concerns about the secrecy of the ballot were raised in over half of the polling stations its observers visited. Its final report describes secrecy folders too short to cover the ballot and often disregarded, desks placed close to poll watchers, machines sited where ballots were fed in plain view, and cases where party agents in the room could see how each voter had marked their ballot; the Teresa district jail in Rizal province was the only precinct its observers visited that offered privacy screens. The mission concluded that the Philippines does not meet international standards on the secret ballot and that the gap can facilitate vote buying and intimidation.

ANFREL called the prevalence of vote buying the biggest flaw in Philippine elections, reported sums of 100 to 2,000 pesos per voter, and recorded the rampant use of money and goods to woo voters alongside the common abuse of local and national government resources for campaign purposes, with incumbents more often seen to benefit. It also found campaign-finance spending limits outdated and routinely ignored, with rare repercussions. ANFREL notes that vote buying is difficult to investigate and presents its observers' accounts as context rather than as accusations against named candidates.

Meta's Q2 2022 threat report describes manual removal of about 10,000 accounts for inflating the distribution of election-related posts, automated action against over 15,000 more in hundreds of Philippine clusters, and more than 50,000 accounts removed globally of which about 10% originated in or targeted the Philippines. In January 2022 Twitter suspended over 300 accounts from Marcos's supporter base for platform manipulation and spam after reviewing hashtags identified in a Rappler investigation.

Tsek.ph partners published ten fact checks on fabricated partial tallies showing Marcos leading, two showing Marcos and Sara Duterte winning, two showing Duterte dominating, five showing Robredo leading and one claiming a Robredo-Pangilinan landslide in overseas votes, all circulated while polls were still open or before counting could lawfully start under COMELEC Resolution 10695; one post alleging preprogrammed results drew 145,000 shares. ANFREL, citing Tsek.ph figures reported by GMA News on 7 May 2022, records that 92% of false or misleading claims about Marcos were favourable to him while 96% of the disinformation targeting Robredo was negative.

Petitions to cancel Marcos's certificate of candidacy over his 1997 tax conviction were filed at COMELEC from 2 November 2021; the COMELEC Second Division dismissed one on 17 January 2022, the Former First Division dismissed the other on 10 February 2022, and the en banc denied reconsideration in both on 10 May 2022, the day after the election. The Supreme Court dismissed the consolidated certiorari petitions on 28 June 2022 in Buenafe v. COMELEC. ANFREL called this an “inappropriate delay in addressing urgent election disputes” and said that, whenever possible, urgent eligibility questions should be settled before the election and ideally before the campaign period starts.

NAMFREL stated that the reception logs and raw files COMELEC published show no proof of a secret local area network acting as a man-in-the-middle device, of pre-programmed machines, or of fabricated returns, and that it observed no anomalies in the automated count or transmission. PPCRV explained that the recurrence of one private address reflects Network Address Translation, a legitimate technique that the carriers Globe and Smart confirmed using to connect the machines, and that the test of a man-in-the-middle attack is a mismatch between printed and transmitted returns, which its parallel count did not show. The random manual audit led by the Legal Network for Truthful Elections found 99% of audited ballots read accurately.

COMELEC's transmission logs, published on 23 March 2023, show the first machine transmitting at 7:08 pm on 9 May 2022. PPCRV's own log-timing analysis matched machine timestamps against the transparency server and concluded that printing the first returns and transmitting them can be done in under 30 minutes, against the claim that printing the first eight returns must itself take at least 30 minutes; COMELEC and NAMFREL's chairman both put that printing time at about ten minutes. Rappler notes that a comparable early surge occurred in 2016, when the transparency server carried about 60% of votes by 8 pm on election night. The stable ratio between the leading candidates as returns accumulated is the law of large numbers and was also observed in 2013 and 2016.

Tsek.ph documented fabricated result graphics for both leading candidates appearing from 8:27 am on 9 May, hours before polls closed, some mimicking the interfaces of news organisations and one reusing a graphic from Robredo's 2016 vice-presidential race. Counting and consolidation could not lawfully begin until after polls closed at 7 pm.

Meta classified its Philippine removals under its Inauthentic Behavior policy, describing spam and distribution inflation that in some cases used politics as a lure, and attributed none of it to a candidate or campaign; the Philippines does not appear in the coordinated-inauthentic-behaviour section of the same report. Twitter cited platform manipulation and spam, and said it publishes campaigns it can confidently attribute to state-linked activity in a separate archive, which this network was not placed in. Rappler's investigation, which prompted Twitter's review, described the accounts as an emerging network drawn from Marcos's supporter base; Marcos's spokesperson said there was “no certainty” that the suspended accounts were supporters and accused Rappler of being party to disinformation, an allegation Rappler rejected.

In its December 2022 report on the 2022 elections, NAMFREL recorded that the vote-counting machine system hash code shown at a demonstration did not match the one published at the second final trusted build, a difference the testing firm Pro V&V attributed to a typographical error made when the hash value was entered manually into the certification documents. NAMFREL wrote to COMELEC asking for the underlying documentation and an independent verification of all hash codes; Rappler reported in March 2023 that the request had not been granted. NAMFREL's chairman and the group of IT specialists pressing the fraud allegations also both said the files COMELEC uploaded on 23 March 2023 were reception logs rather than transmission logs, because they carried no carrier, IP address or SIM data, a characterisation COMELEC rejected; in its statement of 7 September 2023 NAMFREL was still calling on COMELEC to publish the transmission logs of the three telecommunications providers. NAMFREL nonetheless stated that it observed no anomalies in the count on the data available to it, and its chairman noted that those citing its report omitted that conclusion.