A Governance Checklist for Institutions That Want to Avoid Unforced Errors

This checklist operationalizes a single question:

Does this system structurally reward restraint, or does it default to touchdowns?

If the answer is unclear, the system is already at risk.

I. Refusal & Restraint Legitimacy

Goal: Make “not acting” a valid, protected outcome.

An institution passes this section if:

  • ☐ Refusal is an explicit, documented outcome—not an exception
  • ☐ Saying “no” or “not yet” does not require higher justification than saying “yes”
  • ☐ Actors can delay or decline without career penalty
  • ☐ Refusal decisions are recorded with rationale, not buried
  • ☐ There is no cultural shorthand equating restraint with incompetence

Failure mode:

Institutions where restraint must be justified but action is assumed will always drift toward touchdowns.

II. Scope & Phase Discipline

Goal: Prevent silent expansion of authority and mandate.

An institution passes this section if:

  • ☐ Every system or role has a defined scope boundary
  • ☐ Phase distinctions exist (exploration vs. production vs. crisis)
  • ☐ Behavior that is acceptable in early phases is explicitly restricted later
  • ☐ Crossing scope or phase boundaries requires review or escalation
  • ☐ “Incremental change” is periodically re-evaluated for cumulative risk

Failure mode:

“It was just a small change” repeated until the system is no longer recognizable.

III. Credit–Risk Symmetry

Goal: Align rewards with long-term system health, not short-term visibility.

An institution passes this section if:

  • ☐ Credit for success is distributed, not concentrated
  • ☐ Downside risk is not fully externalized to future actors or the public
  • ☐ Performance metrics include risk avoided, not just outcomes achieved
  • ☐ Promotions do not depend solely on visible action or speed
  • ☐ Leaders are accountable for what they allowed, not just what they did

Failure mode:

Touchdowns produce personal upside; interceptions become institutional problems.

IV. Decision Attribution & Auditability

Goal: Ensure decisions can be reconstructed before failure occurs.

An institution passes this section if:

  • ☐ Decisions are attributable to roles, not abstractions (“the system decided”)
  • ☐ Rationale for action and inaction is recorded
  • ☐ Audit trails exist for how authority was exercised or withheld
  • ☐ Post-hoc reviews can identify where restraint failed or succeeded
  • ☐ There is no ambiguity about who owned the final call

Failure mode:

After failure, no one can identify the play that mattered.

V. Separation of Analysis and Execution

Goal: Prevent insight from becoming authority by default.

An institution passes this section if:

  • ☐ Analysts cannot execute without authorization
  • ☐ Tools that generate recommendations cannot finalize outcomes
  • ☐ Expertise informs decisions but does not replace governance
  • ☐ Dissent can exist without needing to “win”
  • ☐ No actor or system becomes a recurring decision hinge

Failure mode:

“Because it was smart, it was allowed.”

VI. Speed as a Governed Variable

Goal: Treat speed as a decision, not a virtue.

An institution passes this section if:

  • ☐ Speed is adjustable by context and phase
  • ☐ There are explicit slow-paths for high-risk decisions
  • ☐ Emergency authority is bounded and time-limited
  • ☐ Pauses and reviews are not framed as failures
  • ☐ No default assumption equates faster with better

Failure mode:

“We had to move quickly” becomes the universal excuse.

VII. Narrative Resistance

Goal: Reduce pressure to act for story resolution rather than system integrity.

An institution passes this section if:

  • ☐ Leaders are trained to explain non-events and delays
  • ☐ Communications do not require visible action to appear competent
  • ☐ Ambiguity is tolerated in public-facing messaging
  • ☐ Success is not defined solely by decisive moments
  • ☐ “Doing nothing” can be articulated as strategy

Failure mode:

Institutions act to end the story, not to preserve the system.

VIII. Hard Stops & Veto Power

Goal: Encode the ability to stop the play.

An institution passes this section if:

  • ☐ Independent veto authority exists
  • ☐ Vetoes do not require consensus or heroism
  • ☐ Stopping a process is treated as a skill, not an obstruction
  • ☐ Escalation paths are clear and protected
  • ☐ No one has unilateral authority over irreversible actions

Failure mode:

Everyone saw the risk, but no one could stop it.

IX. The Final Test: The Checkdown Question

Every major decision should be answerable to this prompt:

What is the safest version of not scoring right now?

If the institution cannot answer:

  • who would be blamed for restraint,
  • who would be rewarded for action,
  • and why the system prefers one over the other,

then governance is incomplete.

X. Interpreting Results

  • Mostly checked:
    The institution is structurally capable of restraint. Errors may still occur, but they are less likely to be catastrophic.
  • Mixed:
    Touchdown bias exists but may be containable with targeted reform.
  • Mostly unchecked:
    The institution will reliably select for high-variance action under pressure.

This is not a moral judgment.
It is a risk assessment.

Closing Note

Institutions do not fail because people try to win.
They fail because winning is misdefined.

The purpose of governance is not to stop action.
It is to make sure that when the moment comes, the system knows how to slide.