Michigan Judicial Tenure Commission Audit: What the Phase III Scope Quietly Admits About Systemic Risk

The Audit Is Framed Around Bias. The Scope Reads Like a Systems Warning.

The Michigan Judicial Tenure Commission announced on April 30, 2026, that the National Center for State Courts will conduct Phase III of its review of possible vulnerabilities to implicit bias in Michigan’s judicial disciplinary system.

That public description is accurate. It is also incomplete.

The Phase III scope is not just asking whether race correlates with outcomes. It is asking where the process itself may create unequal outcomes through filing patterns, transcript availability, allegation coding, investigative framing, and discretionary decision-making.

That matters because judicial discipline is not only about whether a judge did something wrong. It is also about whether the system receives the complaint, classifies it consistently, gathers the same categories of evidence, interprets the allegations through the same lens, and applies the same procedural safeguards at each decision point.

What Phase III Is Actually Studying

The Phase III scope says NCSC will conduct a comprehensive analysis of judicial disciplinary processes to examine possible racial disparities in how Requests for Investigation are filed, processed, and resolved against Black judges compared to White judges.

The work is organized around four tasks. Those tasks examine disproportionality in total filings, the role of transcript evidence in investigation outcomes, differences in severity of outcomes by allegation severity, and potential implicit bias inside organizational procedures.

That structure is important. It separates the discipline process into decision points. Complaint filing is one decision point. Evidence collection is another. Full investigation is another. Allegation coding is another. Outcome severity is another. Each point can introduce variation.

The Complaint Volume Problem

Task One focuses on the disproportionality of total Requests for Investigation filed against Black judges.

The scope states that Phase I found race was not associated with an increased likelihood of a judicial officer being a respondent to at least one grievance, but race was significantly associated with the number of grievances filed.

That distinction is not cosmetic. It means the problem may not be whether Black judges are more likely to be touched by the system once. The issue may be whether they are pulled back into the system more often.

Repeated filings can matter even when individual complaints are dismissed. Complaint volume creates exposure. Exposure creates review burden. Review burden creates institutional memory. Institutional memory can influence how later filings are perceived.

The Transcript Evidence Problem

Task Two is where the scope becomes especially important.

According to the Phase III document, Phase II showed that Black respondents were more likely to proceed to full investigation if transcript-only evidence approvals were included in the analysis. When transcript-only evidence approvals were excluded, there was no statistically significant difference in the likelihood of Black and White respondents proceeding to full investigation.

That is a flashing systems light.

It means transcript evidence may not be neutral in practice. It may function as a gatekeeping variable. A complaint with transcript access can travel differently from a complaint without transcript access. A complaint approved for transcript collection can become a stronger investigative file than one that never receives the same evidentiary development.

This does not prove intentional bias. It does not need to. The audit is studying whether the process creates the conditions for unequal movement through the system.

The Hidden Question: Who Controls the Record?

Transcript evidence sounds technical. It is not.

A transcript is not just a document. It is an official version of what happened. Once a transcript enters a disciplinary file, the system may treat the complaint differently because the allegation now appears more concrete, more reviewable, and more administratively usable.

That makes record control central to judicial discipline. The key question becomes whether similarly situated complaints receive similar evidentiary development.

If one complaint receives transcript collection and another does not, the difference may affect whether a full investigation occurs. If transcript-only approvals change the statistical picture, then transcript approvals are not background noise. They are part of the mechanism.

Allegation Severity Is Not Self-Executing

Task Three focuses on differences in severity of outcomes by allegation severity. The scope says NCSC will hand-code original Requests for Investigation for cases that proceeded to full investigation. Coding will capture Brown factors appearing in allegations, alleged Judicial Canon violations, mitigating factors, aggravating factors, and outcomes.

This sounds like technical research design. It is also a recognition that allegations are interpreted before they are resolved.

The same conduct can be framed as intemperance, poor demeanor, dishonesty, abuse of authority, misunderstanding, misstatement, or misremembering. Those labels matter. They shape the perceived seriousness of the conduct and the institutional response that follows.

The scope explicitly notes that Michigan JTC findings may include variables such as whether a respondent is described as lying rather than misstating or misremembering.

That is not a small wording issue. In discipline systems, language can become outcome architecture.

The Discretion Problem Is Stated Plainly

The most important part of the Phase III scope appears in Task Four.

The document says Phase III will explicitly address the grievance investigation process within the Michigan JTC and assess the strength of organizational policies and procedures. It further states that a lack of strong, clear organizational procedures and policies may create opportunities for discretion by individuals and lead to inequity in outcomes for otherwise similar situations.

That sentence is the center of the article.

It acknowledges the core institutional risk. If procedures are not strong enough, individual discretion can fill the gap. When discretion fills the gap, similar cases can move differently. When similar cases move differently, the system can produce inequity without ever announcing that inequity as policy.

Why This Case Matters

This matters because the Michigan Judicial Tenure Commission sits at a critical public trust point. It is one of the few institutions with authority to investigate judicial misconduct. If the discipline system itself has weak process controls, the public is left with a harder question than whether any single judge was treated fairly.

The question becomes whether the oversight system can prove its own fairness.

Phase III is important because it moves toward that question. It does not simply ask whether outcomes differ. It asks whether procedures and policies may allow implicit bias to influence outcomes at each decision point.

That is the correct frame. Bias does not only live in individual attitudes. It can live in workflows, defaults, evidence thresholds, classification systems, staff summaries, escalation practices, and unwritten norms.

What the Final Report Should Answer

The final Phase III report should answer whether similar complaints receive similar evidence development, whether transcript collection is governed by clear and consistently applied criteria, whether allegation severity is coded in a way that separates allegation from finding, and whether internal characterizations such as lying, misstating, or misremembering are applied consistently across respondents.

It should also identify where sample sizes prevent firm conclusions. The scope already anticipates that possibility. In those instances, descriptive statistics will matter. Small samples do not erase institutional risk. They simply limit the claims that can responsibly be made from the available data.

The Real Test

The real test of Phase III will not be whether the final report uses the right institutional language.

The test will be whether it identifies the decision points where process control is weak and recommends safeguards specific enough to change how cases are handled.

Judicial discipline cannot depend on vibes. It cannot depend on who writes the staff summary, who requests the transcript, who approves evidence collection, or who decides whether a judge lied or merely misremembered.

A credible oversight system needs visible rules, consistent thresholds, documented decision logic, and a record that can be audited after the fact.

That is what Phase III has the opportunity to expose.

Not just whether the system is biased.

Whether the system is controlled.