Before Judging Moraes, the Supreme Court Must Decide Who Can Judge
An extraordinary session on a possible investigation into Alexandre de Moraes shifts the debate to a prior and potentially deeper question: the lawful judge, impartiality, case assignment and the STF's own capacity to constitute a legitimate adjudicator.
Fernando Fagundes Ferreira ·
The extraordinary session of Brazil's Supreme Federal Court (STF) on Tuesday, September 15, began with an apparently narrow object: to decide the fate of PET 16.662, formed from information produced by the Federal Police in the course of the investigations into Banco Master and Daniel Vorcaro, which include references to Justice Alexandre de Moraes.
Little more than two hours later, the center of the proceeding had already shifted.
Before addressing any possible criminal liability, the full bench began to discuss a more elementary and institutionally more delicate question: who may judge, who must step aside, who is the legitimate rapporteur, and which procedures must be observed so that the Court's own decision can be recognized as valid and impartial.
The shift is not merely rhetorical. It turns a proceeding initially centered on Alexandre de Moraes into a debate about the institutional conditions of the STF itself to examine facts that reach members of the Court.
The bench shrinks for the first time
Justice Kassio Nunes Marques was the first to announce his withdrawal from the case. Before doing so, he publicly defended his position, stating that he had never ruled on any case related to Banco Master, that his son had not provided services to the bank, and that he exchanged no messages with Daniel Vorcaro.
Despite those statements, he declared himself disqualified. His justification was institutional: as president of the Superior Electoral Court, he said he did not feel comfortable taking part in a proceeding with potential repercussions on the 2026 electoral process.
Dias Toffoli followed. He recalled earlier questions about his role in the Master case and said that, although he did not recognize any legal impediment, he would maintain the position already adopted in other proceedings and declare himself recused on grounds of personal conscience.
Two justices were therefore leaving the deciding composition before the bench had even effectively reached the merits.
That matters in more than numerical terms. In a case involving references to justices, relatives of justices, decisions taken by different members of the Court and questions about the very conduct of the investigations, the composition of the adjudicating body becomes part of the problem.
The lawful judge moves to the center of the crisis
The decisive turn came with Flávio Dino and Gilmar Mendes.
Dino questioned the way PET 16.662 left André Mendonça's docket and passed to the STF's Presidency. Gilmar developed the same concern from the perspective of the principle of the lawful judge and of the prohibition of the old power of avocation.
The 1988 Constitution eliminated from the Brazilian system the possibility of an authority simply choosing to remove a given case from its lawful judge in order to submit it to another. Within collegiate courts, assignment by lot performs precisely that function: preventing anyone from choosing who will decide a given case.
It is exactly the legal validity of that move that has now been contested.
Fachin rejected the characterization of avocation. He argued that he had not arbitrarily taken the case from Mendonça, since Mendonça himself referred the records to the Presidency, and maintained that, in a matter concerning a member of the Supreme Court, the lawful judge would be the full bench.
Dino and Gilmar raised another question: recognizing the competence of the full bench does not necessarily settle who should act as rapporteur, nor does it eliminate the need for assignment under the Court's internal rules.
The discussion therefore ceased to be only about facts attributed to Moraes. It reached the procedural architecture used to judge them.
When the adjudicator is also part of the conflict
Gilmar Mendes added an even more sensitive layer.
If Fachin took procedural steps in PET 16.704, suspended decisions by other justices, received related cases and went on to conduct PET 16.662, it would be necessary to discuss to what extent he could simultaneously remain responsible for acts that would later be assessed by the collegiate body itself.
The argument meets clear resistance from the Presidency, which maintains that it acted precisely to prevent conflicting decisions among justices and to preserve procedural order.
But the controversy reveals the structural problem of the session: every attempt to identify an adjudicator apparently external to the conflict runs into another procedural connection, previous decision or institutional involvement capable of generating a new objection.
The case begins to produce a kind of infinite regress of impartiality.
From Moraes to the justice system
The point of order raised during the session widens the problem still further.
Gilmar Mendes argued that PET 16.662 should be joined to PET 16.704, that all judges and justices mentioned in the investigations related to Banco Master should be identified, and that only afterwards should the composition of the adjudicating body and the assignment of the cases be determined.
The central argument is simple: it would not be institutionally acceptable to concentrate hundreds of pages of investigation on references to a single justice while possible mentions of other judges, justices or relatives were subjected to different criteria.
Gilmar framed the question in terms that go beyond the concrete case: the rigor of criminal prosecution cannot vary according to the identity of the person being investigated.
This is where the session definitively ceases to be only about Alexandre de Moraes.
If the bench adopts that approach, the real object becomes much larger: the relationships possibly identified between members of the justice system and the network of interests built around Banco Master.
This does not mean presuming wrongdoing by any judge mentioned. A mention in a document, a professional relationship, an institutional contact or a family tie does not constitute proof of a crime.
It means establishing uniform criteria to decide which facts deserve investigation, who may conduct it, and which guarantees must apply regardless of the name involved.
The risk of selective prosecution
This is a particularly relevant question from the perspective of lawfare.
The problem does not arise only when someone is unjustly held liable. It also appears when methods, guarantees, evidentiary requirements or investigative intensity change according to the chosen target.
If a given procedure is considered legitimate against an adversary but inadmissible when it reaches an ally, the institutional question precedes even the discussion of guilt or innocence.
Likewise, if a given investigative measure is legally invalid, its invalidity cannot depend on who will benefit from the annulment.
Consistency of method is one of the boundaries between adjudication and the instrumentalization of law.
From disagreement to confrontation
The temperature of the session showed, however, that this discussion is far from taking place in a neutral environment.
Gilmar Mendes used extremely harsh language in criticizing the handling of the case. André Mendonça reacted to accusations of political and electoral motivation. Alexandre de Moraes challenged the investigation produced under Mendonça's orders and again asserted that he had been the target of a directed procedure.
These accusations must be treated as allegations made by the justices themselves, not as facts already demonstrated.
Accusations of fraudulent investigation, political-electoral action, external interference, deliberate leaks or manipulation of evidence require independent scrutiny and cannot be automatically converted into truth merely because they were voiced from the Supreme Court bench.
The court that must constitute its own adjudicator
Perhaps this is the most important aspect of this first part of the session.
The STF began the day trying to decide whether there was a legal basis to allow an investigation involving one of its members to proceed. Before reaching that answer, it discovered that it had to resolve another question:
who has the legitimacy to give that answer?
That problem is not solved merely by the formal existence of eleven seats on the bench.
A constitutional court also depends on the perception that its judges are subject to the same rules they demand of all other citizens: the lawful judge, due process, adversarial proceedings, impartiality and reasoned decisions.
The more members of the Court are directly or indirectly reached by the facts under discussion, the harder it becomes to separate three roles that should remain clearly distinct: the investigated, the investigator and the adjudicator.
The Master crisis has therefore reached a particularly delicate stage.
The Supreme Court is no longer examining only the conduct of one of its justices.
It is examining the capacity of its own rules to prevent personal, institutional and political conflicts from determining who will be investigated, who will investigate and who will judge.
Autophagy takes on another dimension
The image of institutional autophagy helps to understand what happened during the morning, but it must be used precisely.
This is not simply a matter of justices attacking justices.
The deeper phenomenon appears when every attempt to resolve a suspicion creates a new suspicion about the procedure used to resolve it.
Moraes is questioned. Mendonça is questioned. Fachin's decision is questioned. The disqualification of other members is debated. It is proposed that references to further judges be examined. And, successively, it becomes necessary to ask who remains sufficiently distant from the facts to decide about all the others.
It is at this point that the crisis ceases to be interpersonal.
It becomes systemic.
Opening the session, Fachin recalled Supreme Court justices struck down by authoritarianism during the dictatorship and said that the Court's current composition would have no right to hand future generations a court smaller than the one it received.
The sentence acquires additional meaning after what occurred in the following hours.
Preserving an institution does not mean shielding its members from investigation.
Nor does it mean allowing limitless investigations because the target is politically convenient.
It means preserving the rules capable of surviving a change of characters.
At the end of this first stage, the question may perhaps be put even more simply:
it is not enough for the Supreme Court to reach a decision.
It must demonstrate that it can still constitute an adjudicator.
Update note
Article published on the basis of the first part of the STF's extraordinary session of September 15, 2026, before proceedings resumed at 2 p.m. Lawfare4All will continue to follow the deliberation on the point of order and will update this coverage as the session develops.