The previous four installments described a mechanism that operated for a century and produced an unknown number of convictions. This one is about why the number is unknown, and why it is likely to stay that way.
MCL 767.4 restricts disclosure of information presented to a judge conducting a one-person grand jury inquiry, with misdemeanor exposure for violation. That restriction attaches to the material rather than to the fate of any prosecution. Michigan therefore holds a body of proceedings spanning roughly a century, including the Flint water evidence, that produced charges through a mechanism later held unauthorized and that cannot be examined by anyone outside them.
Is this series arguing grand jury secrecy should be abolished?
No. The rationales are real, and witness protection in particular is not negotiable. The argument concerns duration, and what happens to the record once the proceeding it protected has concluded.
Does secrecy work differently for a one-person grand jury?
The practical effect differs sharply. A citizen grand jury involves thirteen to seventeen jurors and an indictment requires the concurrence of at least nine. A one-person grand jury involves a single judge. The same rules produce very different levels of internal check depending on how many people are in the room.
Could the Flint material ever become public?
Any release would require a legal basis, and this article does not assert one exists. What can be said is that no automatic mechanism unsealed it when the prosecutions ended.
What would producing a count actually require?
A field in case management systems recording that a prosecution began by one-person grand jury indictment, or a manual review of charging documents across counties over decades. Neither exists, and the first was never created.
The Question I Could Not Answer
Every installment in this series has run into the same wall, and I want to describe it directly rather than working around it again.
I cannot tell you how many people were charged by one-person grand jury indictment in Michigan. Neither can the Attorney General, whose office has described the figure in litigation as thousands over a century. Neither can the Supreme Court, which said in Robinson that the exact number is unknown. Neither can the Legislature, which is considering repeal without one.
That is unusual. Most questions about the criminal system are answerable with enough patience and enough records requests. This one is not, and the reason is structural rather than administrative.
Two Kinds of Invisibility
The unavailability of this information has two independent causes, and conflating them obscures what fixing either would take.
The first is secrecy. The proceedings occurred in closed session, and MCL 767.4 restricts disclosure of information presented, with misdemeanor exposure attached. That governs what happened inside the room.
The second is the absence of coding. Setting the contents aside entirely, nothing in Michigan’s case management infrastructure records that a given prosecution began this way rather than by complaint. That is not a secrecy problem. It is a records design problem, and it would exist even if every proceeding had been held in public.
Lifting the statutory restriction tomorrow would not produce a count. It would make the contents of individual proceedings available to whoever could identify them, and identification is the part with no solution, because the charging route was never recorded as a data field anywhere.
What Secrecy Is For
Grand jury secrecy is not an aberration and I do not want to treat it as one. The rationales are long-standing and serious.
Witnesses testify more candidly when they are not exposed, and in cases involving violence or organized activity, exposure carries genuine danger. Targets who learn of an investigation may flee, destroy evidence, or approach witnesses. And people who are investigated and never charged have a real interest in that fact not becoming public, which is among the strongest arguments in the set.
Each of those justifications concerns an active investigation or an uncharged person. None concerns a prosecution that has concluded.
That gap between what secrecy is for and how long it lasts is the subject of this installment.
The Field Kit includes the Clutch Justice records request templates with jurisdiction guidance and exemption anticipation language, built for situations where the first response is a denial.
Open the Field KitThe Defendant’s Position
Consider what the arrangement meant for a person charged this way between 2001 and 2022.
A judge sitting alone reviewed evidence in closed session and issued an indictment. Under Glass, no preliminary examination followed. Under Green, the one-person grand jury procedure was understood to serve the same function as one. The material presented was restricted by statute.
So the defendant faced charges arising from a proceeding they did not attend, based on evidence they had not seen, with no hearing at which the prosecution would be required to produce witnesses subject to cross-examination before trial.
The Peeler court described the ordinary Michigan route by contrast: a complaint issued by a prosecutor followed by a preliminary examination in open court at which the accused can hear and challenge the prosecution’s evidence. The distance between those two descriptions is the whole of what the mechanism did.
A prosecutor issues a complaint. A preliminary examination follows in open court. The defense hears witnesses under oath and cross-examines them before trial.
How almost everyone in Michigan is charged, in the Peeler opinion’s own framing.
A judge reviews evidence in closed session and issues an indictment. No preliminary examination follows. The material presented is restricted by statute.
Held in 2022 to have been unauthorized as to the indictment, and held in 2026 not to disturb anyone already final.
Route one produces a transcript of an open hearing. Route two produces a restricted file and no public proceeding of any kind before trial.
The difference determines whether anything can be reconstructed afterward.
What Happened to the Flint Evidence
The Flint prosecutions produced an extensive investigative record. Witnesses were subpoenaed and testified before a judge sitting as a one-person grand jury. Indictments issued against nine former officials.
Then the charging mechanism was held unauthorized, the indictments were treated as void from inception, the charges were dismissed, and the Supreme Court declined to revive the cases.
The material gathered in those proceedings did not become public as a consequence. Nothing in the statutory scheme provides that the restriction lifts when a prosecution ends, and no separate mechanism accomplished it.
What that means concretely is worth stating without ornament. Testimony was taken about how a city’s water became contaminated. That testimony exists. No court has determined whether it establishes anything, because no trial occurred. And the people who drank the water cannot read it.
A prosecution that ends on a charging defect produces no findings. A statutory restriction that survives the prosecution produces no record. The combination leaves the underlying question permanently open and permanently unexaminable, which is a stronger form of closure than either mechanism achieves alone.
Why the Single-Judge Design Matters Here
One further point about secrecy in this context, offered as an observation about design rather than about any individual.
Secrecy rules developed around bodies of thirteen to seventeen citizens. Under Michigan law a citizen grand jury indictment requires the concurrence of at least nine grand jurors. Whatever happens in that room happens in front of a group, and requires agreement among a majority of them.
A one-person grand jury has one participant. The same secrecy rules apply, but the internal check that made secrecy tolerable in the citizen model is absent by construction. There is no deliberation, no dissent, and no other participant who could later describe the proceeding.
That is not an accusation about how any Michigan judge conducted an inquiry. It is a statement about what the design permits and what it forecloses. A record only one person can describe is a record with no redundancy in it.
There is no auditor of one-person grand jury proceedings. No reviewing body examines whether the evidence presented supported the indictment issued, no transcript enters a public docket, and the Supreme Court’s 2022 conclusion that these indictments were never authorized was reached without any institution ever having looked inside a single such proceeding.
Why This Matters
The four preceding installments were about how a defect gets remedied. This one is about what happens when a defect cannot be measured.
Robinson’s outcome turned in part on a factor asking about the effect on the administration of justice, and the court noted the number of affected defendants is unknown. That unknown is not a gap in anyone’s research. It is a product of the mechanism’s own design, which held proceedings in closed session, produced no public hearing, and left no coded trace in any searchable system.
So the mechanism produced a defect, the same mechanism made the defect’s scope unmeasurable, and the unmeasurability then entered the analysis of whether anyone should get relief.
I do not think anyone designed it that way. The 1917 Legislature was not planning for a 2026 retroactivity decision. But the effect is a closed loop, and closed loops do not open on their own.
Part VI takes up House Bill 4434, which would repeal the one-person grand jury provisions entirely. It passed the Michigan House in June 2025 with bipartisan support and thirty-three Democrats opposed, and moved to the Senate. The installment examines what repeal fixes, what it does not, and why a Detroit representative’s objection deserves to be taken seriously rather than dismissed.
Continue Your Investigation
If this reporting raised more questions, use the Clutch Justice ecosystem to keep going.
