Vacancies and the Philippine Senate’s Impeachment Threshold

Article XI, Section 3(6) of the 1987 Constitution sets what looks like a simple rule: no impeachable official may be convicted “without the concurrence of two-thirds of all the Members of the Senate.” In practice, no impeachment judgment has ever had to answer the question the clause quietly begs — two-thirds of what number? The Supreme Court has confronted materially identical language elsewhere, and the impeachment question itself was put to it directly during the Estrada trial — but neither episode forced a controlling answer. Does “all the Members” mean the Senate’s constitutionally prescribed complement of 24, or however many seats are actually occupied when the trial ends?

The distinction is not academic. If the base is fixed at 24, conviction requires 16 votes no matter how many seats are vacant. If the base tracks actual membership, the number needed can shrink as vacancies accumulate. Retired Justice Antonio Carpio has made the extreme case bluntly, in an ANC interview: “The denominator depends on who are qualified to vote. For example, if nine senators die or resign from office, you are left with 15. So… you can never convict.” He made essentially the same argument in independently verified later remarks: “If you interpret the Constitution as saying there must be 16 senators to convict, then you render the Constitution ineffective when there are senators who die or resign.” Under a fixed-24 reading, both statements are exactly right — even a unanimous 15–0 vote would fall one short of the 16 required, making conviction mathematically impossible despite unanimity among every senator remaining in office.

Not every reason a senator might be unable to participate presents the same constitutional problem, and the distinction matters throughout what follows. At least three categories must be kept separate: actual termination of membership (death, resignation, expulsion, final unseating); temporary legal incapacity while still holding the seat (preventive suspension); and mere physical unavailability while nothing has formally changed the senator’s status (detention without suspension, evading arrest, medical leave, ordinary absence). The constitutional case for reducing the denominator is strongest in the first category, genuinely contested in the second, and weakest in the third — and, as the sections below show, the current factual controversy has so far stayed entirely within the second and third categories.

A question the arithmetic keeps dodging

The ambiguity has surfaced repeatedly, but circumstances have repeatedly kept it from becoming outcome-determinative.

During President Estrada’s 2000 impeachment trial, the Senate was down to 22 incumbents (two vacancies), with only 21 having taken their oath as judges. San Beda law professor and Bar reviewer Jose Sundiang filed a petition for declaratory relief asking the Supreme Court directly: was the threshold two-thirds of 22, or of 24? He argued for 22. The Supreme Court never reached the merits — but not because EDSA II later mooted it. On January 16, 2001, four days before Estrada left office, the Court disposed of Sundiang v. Estrada and the Philippine Senate, G.R. No. 146131, on jurisdictional grounds: an action for declaratory relief did not fall within the Court’s original jurisdiction and belonged in the Regional Trial Court. The vacancy question had been squarely presented during an actual impeachment trial — the Court simply declined to answer it because it arrived through the wrong procedural door.

The 2012 Corona trial (23 sitting senators, one vacancy) and the 2018 Sereno proceedings (also 23 sitting, per then–Senate President Pimentel’s own “16 of 23” framing) both produced the same number either way: two-thirds of 23 requires 16 whole votes, same as two-thirds of 24. Officials could speak loosely of “16 of 23” without ever being forced to say which theory of “all the Members” they actually held.

The question resurfaced in 2026 during the Duterte impeachment trial, this time framed around detained, hiding, or otherwise unavailable senators rather than vacancies proper. UP Law’s Paolo Tamase staked out what he called the prevailing view: “all the Members” means all 24, full stop, regardless of who is present or seated.

By the trial’s ninth week, the question was no longer academic. Four senator-judges were not participating in person: Sen. Loren Legarda is abroad on extended medical leave; Sen. Ronald “Bato” dela Rosa is in hiding amid an International Criminal Court warrant; Sen. Jinggoy Estrada is detained on plunder and graft charges; and Sen. Rodante Marcoleta is detained on a plunder charge and related criminal charges. Senator-Judge Erwin Tulfo formally appealed Presiding Officer Francis Escudero’s standing ruling — issued back on July 6, when Escudero first set the threshold at 16 votes based on the full 24-member Senate — asking the impeachment court to reconsider it in light of these absences. The court responded by scheduling independent legal experts to weigh in on September 16, with prosecution and defense oral arguments on September 23, 2026, followed by a vote of the senator-judges on the threshold itself. Four retired Supreme Court justices were named as amici curiae for the September 16 hearing: retired Chief Justices Artemio Panganiban and Reynato Puno and retired Associate Justice Adolfo Azcuna will appear in person; retired Chief Justice Hilario Davide Jr. — who presided over Estrada’s own impeachment trial a generation earlier — will submit a written legal memorandum instead.

Retired Justice Antonio Carpio has staked out a third position in this debate, distinct from both a flatly fixed 24 and a raw headcount of whoever shows up. His test turns on legal qualification rather than physical presence: the denominator, in his view, should include only senators legally qualified to exercise their senatorial functions and vote, rather than treating mere physical availability as controlling. On that test, dela Rosa and Legarda still count — neither has been suspended or removed, whatever their physical circumstances. Carpio argues that Estrada and Marcoleta are different: in his view, the pending plunder informations against both senators make suspension mandatory under the applicable law, and a suspended senator cannot exercise the functions of the office, including the impeachment vote. On Carpio’s further theory that such senators should therefore be excluded from “all the Members,” excluding both would reduce the denominator to 22, and two-thirds of 22 requires a clean 15 votes — no disputed treatment of a fractional remainder required.

That 22-to-15 arithmetic, though, rests on two separate propositions Carpio needs to hold, not one. The first is a constitutional theory — that a validly suspended senator should not count among “all the Members” — which the Senate impeachment court has not adopted. The second is a distinct legal and procedural proposition specific to Marcoleta: that the pending criminal proceedings already place him in the category of a suspended senator despite the absence, as of early September, of an issued preventive-suspension order. That proposition is actively disputed. Estrada is genuinely under an effective 90-day preventive suspension that took effect in June. Marcoleta, by contrast, was not under an issued preventive-suspension order as of early September: he expressly maintained that no such suspension had been imposed, and the Sandiganbayan had not yet resolved the pending suspension matter. Applied to Estrada alone, without Marcoleta, the arithmetic doesn’t move — two-thirds of 23 still requires 16 whole votes, same as 24. Retired Justice Adolfo Azcuna has floated a related but distinct theory: that a senator who becomes a fugitive might be treated as having effected a kind of “constructive resignation,” which could remove them from the count even absent formal suspension — a reading that would apply most directly to dela Rosa, not Estrada or Marcoleta.

None of this has been adopted by the Supreme Court, or even definitively by the Senate sitting as an impeachment court. As of this writing, it is precisely what September 23 is for.

The nearest doctrine on point

No Philippine case has squarely held what Article XI, Section 3(6) means when genuine vacancies make two-thirds of the Senate’s constitutional complement diverge from two-thirds of its actual membership. But one decision comes unusually close — and it isn’t the one most commonly cited.

In Bayan v. Zamora (G.R. No. 138570, Oct. 10, 2000), the Court interpreted “at least two-thirds of all the Members of the Senate” under Article VII, Section 21’s treaty-concurrence clause — materially the same formula Article XI uses for impeachment conviction. Only 23 senators were then incumbent, Gloria Macapagal-Arroyo having vacated her seat upon election as Vice President. The Court nonetheless began from Article VI’s declaration that “the Senate shall be composed of twenty-four (24) Senators,” and held that 16 votes “unquestionably” satisfied the requirement. But the Court also noted, in the same breath, that the result would hold even computed on the actual 23 — 16 clears two-thirds either way. Bayan therefore supplies the strongest existing judicial language for the fixed-24 reading, without ever being forced to decide the case where the two denominators actually diverge.

Avelino v. Cuenco (83 Phil. 17, 1949) is older, arose under the 1935 Constitution, and concerned a Senate quorum rather than impeachment — but its fractured opinions read like an early draft of the entire 2026 debate. Following Senate President Avelino’s walkout, four justices (Paras, Feria, Pablo, and Bengzon) reasoned in the Court’s original March 4 resolution that Senator Confesor’s absence abroad reduced the operative Senate to 23, making the 12 remaining senators a majority. Justice Perfecto dissented sharply: the Senate “was and actually is composed of 24 Senators,” a majority could not be fewer than 13, and membership “cannot be reduced upon any artificial or imaginary basis not authorized by the context of the Constitution itself or by the sound processes of reason.” Justice Briones, dissenting separately, rejected the 23-member theory in similar terms. Justice Tuason went further, proposing that the denominator should track senators “elected and duly qualified… who have not ceased to be senators by death or legal disqualification” — explicitly rejecting mere absence as grounds for exclusion.

On reconsideration ten days later, a seven-justice majority declared that a quorum had existed after all, resting “partly… upon the grounds stated by Mr. Justice Feria, Mr. Justice Perfecto, and Mr. Justice Briones” — three justices whose underlying theories were not the same. Feria’s separate opinion is the clearest statement of the actual-membership view: the deliberate 1935-to-1939 shift from “majority of all the members” to “majority of each House” showed an intent to base the count on actual incumbents capable of serving, excluding those incapacitated by “death, incapacity, or absence from the jurisdiction.” Perfecto, for his part, joined the result without abandoning his 24-member premise — pointing to a pragmatic, case-specific basis for treating the 12 as sufficient, distinct from Feria’s general theory. Tuason dissented to the end on the quorum question.

No single “Avelino rule” emerges. What the case shows instead is that Philippine jurisprudence has entertained at least three competing denominator theories — fixed constitutional complement, functionally available membership, and legally qualified membership — since 1949, without ever settling on one for a case where the choice would decide the outcome.

Defensor-Santiago v. Guingona (1998) is worth mentioning only to retire it from this argument. The case concerned a Senate minority-leader dispute; that only 23 of 24 seats were filled at the time (Arroyo’s election as VP) is background color, not a holding. The vote at issue, 20–2, cleared a majority under either a 23 or 24 base, so the Court never had to choose between them.

Taken together: Bayan gives the fixed-24 reading its strongest language, Avelino gives every competing theory a pedigree, and none of them was ever outcome-determinative. September 2026 may be the first time the choice of denominator actually decides who wins.

How other constitutions closed this gap

The Philippines is not the only democracy to condition removal from high office on a legislative supermajority — but other constitutional systems define the relevant denominator in different ways.

The United States chose different words entirely. Article I, Section 3, Clause 6 requires conviction by “the Concurrence of two thirds of the Members present.” Not “all the Members” — present. That single word closes the loophole for any conviction vote that validly takes place: the denominator is the senators present, not the chamber’s full authorized membership, so genuinely vacant seats and absent senators are not embedded in the denominator as automatic votes against conviction. A separate constitutional quorum requirement still governs whether the Senate can conduct business at all — but that is a different kind of obstacle from the one this piece examines, where a vote could take place yet the required number would remain mathematically unreachable even with unanimous support from every senator participating.

India went further, addressing the competing concerns through a dual threshold — though for judicial removal, not executive impeachment. Article 124(4), governing removal of a Supreme Court judge, requires both a majority of the total membership of each House and two-thirds of the members present and voting. The structure guards against conviction by a small participating rump, since the “total membership” prong sets a floor no depleted quorum can satisfy on its own, while the “present and voting” prong still imposes a real supermajority on whoever actually shows up. It doesn’t fully eliminate the vacancy-driven impossibility at the heart of this piece — a majority of total membership is a much lower bar than two-thirds, but in a sufficiently depleted chamber even that floor could become consequential. What India’s design reliably solves is the rump-chamber manipulation problem; the extreme-vacancy problem, it merely mitigates.

Brazil is a different kind of comparison — not a textual fix, but an institutional one. Its Constitution (Article 52) requires conviction by “two-thirds of the votes of the Federal Senate,” language open-textured enough to raise the same present-versus-total question the Philippines faces. But the Senate’s own internal rules close the gap the constitutional text leaves open: Article 378 of its Regimento Interno provides that a conviction may be entered only by “dois terços dos membros do Senado” — two-thirds of the members of the Senate, not of however many happen to be voting. Institutional practice is consistent with that member-based formulation: when the Senate convicted Dilma Rousseff in 2016, all 81 senators voted, 61–20, and 54 votes constituted the required two-thirds threshold. But because all 81 seats were represented in that vote, the proceeding did not test what happens when senators are absent, much less when seats are genuinely vacant. Article 378 establishes a member-based denominator rather than an expressly present-and-voting denominator; it does not, by itself, answer whether “membros do Senado” means the full authorized composition or only actual incumbents when genuine vacancies exist. Brazil therefore offers a useful contrast while leaving the specific vacancy problem at the heart of this piece unresolved.

The comparative lesson is not that the Philippines got this uniquely wrong. It’s that other constitutional systems have addressed the denominator more explicitly — whether in constitutional text, as in the United States and India, or through institutional rules and practice, as in Brazil — while the 1987 Philippine Constitution leaves “all the Members” without an express vacancy rule.

The precedent argument — and its obvious rebuttal

There’s a strong case that insisting on a rigid two-thirds of 24 could create a precedent with serious structural consequences. An interpretation under which enough genuine vacancies reduce the Senate below 16 incumbents could make conviction mathematically impossible for as long as that depletion persists — regardless of the strength of the evidence or even unanimity among every senator remaining in office. A supermajority safeguard would then operate, temporarily at least, as an absolute numerical barrier. The Constitution permits Senate vacancies to be filled through special elections, but it does not guarantee that every vacancy will immediately be filled.

The obvious rebuttal is that the danger cuts both ways. If actual membership or legally qualified membership controls, the definition of which senators cease to count becomes constitutionally consequential in its own right. But that danger is more institutionally constrained than it might first appear, because different forms of suspension proceed through different legal channels. The Senate may suspend or expel one of its own members only with the concurrence of two-thirds of all its Members; if the penalty imposed is suspension, it may not exceed sixty days. On the ordinary 24-member reading, a bloc numerically capable of imposing such internal discipline would already be large enough to supply the 16 votes required for impeachment conviction under the fixed-complement theory, so internal Senate suspension would not provide an obvious shortcut around that threshold. Judicial preventive suspension is different. Under statutes such as the Plunder Law and the Anti-Graft and Corrupt Practices Act, it proceeds through a criminal case, a valid information, and judicial process rather than through an ordinary Senate majority. It therefore is not simply available to a ruling bloc on demand. What remains real is subtler: a variable denominator could create strategic incentives to accelerate, resist, or exploit prosecutorial and judicial decisions affecting a rival senator’s legal status, even without direct control over the result. The danger on that side is real, but it takes a different and more indirect form from the danger created by genuine vacancies under a fixed-24 rule.

The honest version of the “bad precedent” argument has to hold both risks at once rather than treating fixed-24 as the only danger. But naming both dangers does not by itself answer the question this piece is actually about: what the Constitution now in force means by “all the Members of the Senate” — not what a better-designed constitution might have said instead. An India-style combined threshold addresses the competing risks more explicitly by pairing a total-membership floor with a supermajority of those present and voting, but the Philippine Constitution establishes only one threshold, not two. Adopting a second one by judicial or Senate construction would not be interpretation but amendment. If the existing text cannot deal cleanly with genuine vacancies and potentially manipulable exclusions at the same time, that is itself an argument for future constitutional reform — not license to read a second threshold into Article XI, Section 3(6) that is not there.

Why it matters now

That this question has never been definitively resolved is less a sign of settled law than of fortunate arithmetic. In earlier impeachment episodes, the competing denominators either produced the same required number or the proceedings ended before the disagreement could become outcome-determinative. The constitutional ambiguity therefore survived without requiring an authoritative answer.

The Duterte trial may be the first proceeding in which the competing approaches could produce different numbers under circumstances serious enough to force the Senate sitting as an impeachment court to confront the underlying question directly. Escudero has already ruled that 16 votes are required on the basis of the full 24-member Senate; that ruling has now been formally challenged, and the impeachment court is scheduled to hear the parties and consider the threshold question on September 23. Whether the Senate retains the fixed-24 interpretation, adopts some conception of actual incumbent membership, accepts some version of Carpio’s legally qualified-membership approach, or finds another interpretation within the existing constitutional text, its decision could become the most consequential Senate interpretation yet of the meaning of “two-thirds of all the Members of the Senate.”

But the present controversy also exposes a deeper distinction that should survive whatever happens to this particular impeachment. Genuine vacancies created by death, resignation, expulsion, or final unseating are constitutionally different from temporary suspension, detention, flight, illness, or ordinary absence. None of the four senator-judges whose present circumstances triggered the 2026 dispute has actually ceased to be a senator. The extreme case therefore remains unanswered: if enough seats were genuinely vacant that fewer than 16 senators remained in office, would the Constitution still demand 16 affirmative votes and thereby make conviction mathematically impossible?

That is why the constitutional text matters more than the familiar shorthand. Article XI does not say that “sixteen senators” are required to convict. It requires “the concurrence of two-thirds of all the Members of the Senate.” Sixteen follows only after the denominator has been identified. Until Philippine constitutional law establishes what “all the Members” means when seats are genuinely vacant — and separately determines what legal incapacity short of vacancy does to that denominator — treating sixteen as though it were itself the constitutional command assumes the answer to the very question now in dispute.

Whatever rule ultimately emerges will reach far beyond Sara Duterte. A fixed-complement rule could preserve a high and stable barrier against politically manipulated exclusions, but extreme vacancies could make conviction impossible. A variable-denominator rule could prevent genuine vacancies from producing that result, but depending on how broadly it defines exclusion, could create incentives to manipulate who counts as a member. A legally qualified-membership approach might distinguish between those situations, but it would itself require difficult rules about suspension, disqualification, detention, and other temporary disabilities.

The present controversy therefore presents two separate constitutional tasks. The immediate one is interpretive: determine what the 1987 Constitution already means by “two-thirds of all the Members of the Senate.” The longer-term one is structural: if that language proves incapable of dealing cleanly with genuine vacancies and temporary incapacity at the same time, constitutional amendment — not judicial invention — is the proper means of adopting a more explicit formula, whether modeled on another constitutional system or designed specifically for Philippine institutions.

For now, the starting point should remain the words actually ratified by the Filipino people:

“Two-thirds” — not “sixteen.”

—

Holy Spirit of Emperor–Rescuer Jesus Messiah, be merciful to us and help us.

—

This piece deliberately avoids attributing a position to secondary treatises (e.g., Bernas or Cruz) that could not be verified against primary text; where the record is unconfirmed, it says so rather than guessing. Carpio’s ‘render the Constitution ineffective’ quote is drawn from reported September 2026 remarks and has been cross-checked against contemporaneous reporting. The ANC quote is transcribed directly from the broadcast.

—

Glem Melo is an imperfect but repentant evangelical missionary, based in Oakland, California, U.S.A.

—

With research and composition assistance from AI tools.