Feedback Authority

I. A rule that publishes its own breach

In the family of fiscal comply-or-explain rules, a government breaches a target. The rule does not automatically reverse an appropriation, invalidate a budget, or remove an official. It requires an explanation, a correction plan, or exposure to an independent monitor. Everyone calls this a rule. What kind of feedback authority does it actually have?

Comply-or-explain converts breach into a published explanation. The rule’s language is statutory — must comply, must explain — and what it compels in practice is a document. The institutional consequence of non-response is reputational: the explanation enters the public record, opposition parties and the press can cite it, the next government can claim its predecessor failed. None of that is nothing. But the rule’s text appears to promise more — automatic correction, sanction, structural consequence — and the mechanism does not deliver it.

The family of comply-or-explain regimes ranges from pure reputational exposure to statutory response duties with patchy enforcement. The UK Corporate Governance Code does not impose binding compliance with its governance provisions; listed companies are instead subject to disclosure obligations to report compliance or explain departures, with the adequacy of explanations primarily disciplined by investors, proxy advisers, and market scrutiny. The OECD Principles of Corporate Governance use the same architecture. The EU Stability and Growth Pact has procedural and sanction architecture on paper, but in political practice many episodes cash out closer to reputational or procedural pressure than automatic consequence. The Wales Future Generations Commissioner publishes statutory recommendations that public bodies must respond to — a statutory response duty, not a binding implementation duty.

The common failure across the family is selling the practical bite one grade higher than the mechanism can carry. The architecture sincerely uses graded language; the mechanism delivers below the authority profile its language suggests.

The invoked/delivered gap is the object of the essay.

II. Authority is a vector, not a ladder

A feedback channel is any architecture by which signals reach an institution: a consultation portal, an audit office, a regulatory comment, a court filing, a press question, a parliamentary motion, a referendum, a constitutional challenge. The question this essay asks is not whether the channel exists, nor whether the institution is legitimate to receive it. The Legitimacy Gate covers the first question; the second is the broader problem of state architecture. The question here is narrower: once feedback reaches the institution, what does ignoring it cost?

The cost is produced by a vector of mechanisms. A feedback channel can impose ignore-cost through any combination of:

1. Acknowledgement — the institution must receive or log the input.
2. Reason-giving — the institution must explain acceptance or rejection.
3. Visibility — the answer (or non-answer) is publicly exposed.
4. Reviewability — another actor can evaluate the adequacy of the response.
5. Sanctionability — penalty follows inadequate response.
6. Resource consequence — money, staff, eligibility, or budget room changes.
7. Automaticity — the consequence follows without fresh discretion.
8. Reversibility — the decision can be reopened, remanded, corrected, or invalidated.
9. Time limit — the response must occur by a deadline.
10. Delegation — a third party has authority to enforce or certify.

Bindingness is not one-dimensional. Abbott and Snidal made the same point for international legalisation along obligation × precision × delegation; Bovens formalised accountability as a relation involving information × debate × judgment × consequence; the answerability/enforcement split was Schedler’s. The profile vocabulary that follows is a heuristic projection of this vector onto a single scalar, useful for diagnostic comparison but not a substitute for it.

The grade is operational. It is determined by what the institution must actually do when feedback arrives, regardless of how the architecture is described, marketed, or codified in title. The line is not hard ontology — rhetoric can be part of the mechanism when it changes the cost of non-response. Public reasons, ministerial commitments, comply-or-explain statements, guidance documents, and consultation promises generate ignore-cost through reliance, inconsistency exposure, and review invitation. Classify by enforceable institutional consequences, including rhetorical acts where they change the cost of non-response.

Hard cases follow the compelled consequence, not the legal source. A FOIA request is statutorily authorised but produces a procedural response duty (the agency must respond within statutory deadlines; the response can be denial; appeal is available); the grade is procedural. An NTSB safety recommendation is issued by a statutorily independent body but produces only a tracked public response; the FAA can decline to adopt; the grade is a procedural-reputational hybrid, not consequential. A statute that creates an enforcement agency without funding it operates at whatever grade the budget supports, not at the grade the statute names. A constitutional court ruling on incompatibility produces consequential force in systems that install automatic invalidity; in constitutional systems where a ruling requires legislative implementation rather than automatic invalidation, the practical grade may remain procedural until implementing action occurs. The same rule applies in each case: the grade follows the verb the mechanism compels, not the document it appears in.

III. Five common authority profiles

The mechanism vector projects onto five recurring authority profiles. They are common combinations, not metaphysical rungs. A feedback channel can combine them, skip them, or counterfeit one with another.

The consequential profile wraps several distinct consequence channels — budgetary, statutory, constitutional — which are different kinds of force, not stronger versions of one force. The collapse is deliberate but should be named: a budget score can reshape feasibility without creating a duty to answer; a statutory duty can create an answer obligation without money or implementation capacity. The five-profile field guide deliberately ignores those distinctions to support reform-debate comparison; for design purposes, return to the vector in §II.

IV. Conflation failures

The common failure is profile mismatch: reform rhetoric invokes one authority profile while the mechanism implements another. In the simplest cases the mismatch is adjacent — advisory sold as reputational, reputational sold as procedural, procedural sold as consequential. The novel claim is not that feedback needs teeth (the corpus already says that). It is that reform energy disappears at the joints between profiles, because the rhetoric of the higher profile is installed without the mechanism that compels it.

Decorative ↔ advisory. A consultation channel publishes submissions without an answer duty to individual contributors. The line between decorative and advisory turns on whether aggregation and summary are required and reliable, not merely permitted. Large-scale public-consultation portals such as the European Commission’s normally deliver factual summaries and synopsis reports under Better Regulation discipline; the architecture sits at advisory, with contributors receiving no individualised answer and the depth of aggregation varying with the initiative. The mechanism slips toward decorative only where the summary step is in fact unenforced.

Advisory ↔ reputational. Expert commissions deliver reports that go to “noted” without becoming audit-triggers. The UK Office for Budget Responsibility illustrates advisory-with-statutory-publication and reputational force: it produces the official independent economic and fiscal forecasts used for Budgets and fiscal events; the Treasury typically uses those forecasts as the basis for fiscal-policy decisions; but they do not themselves impose automatic correction, and the Chancellor is not obliged to accept them. The rhetoric of “fiscal watchdog” invites reputational expectation about wider fiscal policy that the Charter does not actually compel.

Reputational ↔ procedural. This is the canonical conflation. Comply-or-explain regimes are a family of mechanisms, not one mechanism, but they share the same failure: the rhetoric anchors at procedural or higher; the mechanism cashes out at reputational. The EU Stability and Growth Pact has recommendations, deadlines, surveillance, and possible sanctions in its legal architecture for euro-area countries, but in many episodes the realised bite is closer to procedural or reputational pressure than to automatic consequence; the UK Corporate Governance Code has no binding compliance power, only disclosure obligations evaluated by investors, proxy advisers, and market scrutiny; the OECD Principles of Corporate Governance use the same architecture. The same phrase — comply or explain — hides different authority profiles. The common failure is selling the practical bite one profile higher than the mechanism can carry.

The corpus’s repair-or-override primitive (from Mechanism Analysis and the Mandate Trap) is the clean contrast. The architecture is procedural: when a mechanism analysis identifies a structural flaw, the legislature must either repair the named flaw or publicly override the analysis. The override is itself a procedural act — visible, on the record, citable in the next election cycle. Override is always available; legislative sovereignty is preserved. But the choice is forced, and the institution cannot make the diagnostic disappear by ignoring it. The rhetoric is restrained (“repair or override”); the mechanism is procedural; profile and mechanism match.

The reason comply-or-explain and repair-or-override differ so sharply, despite surface similarity, is the answer-duty structure. Comply-or-explain treats the explanation as the consequence — once the explanation is published, the procedure is complete. Repair-or-override treats the explanation as input to a binary forced choice. The mechanism’s grade is what the institution does after the explanation arrives, not what it has to do to produce the explanation.

Procedural ↔ consequential. Procedural force is reasoned response in the record; consequential force is automatic operational consequence. The joint is where reform debates over “binding authority” most often arrive. The EU AI Act’s Article 56 code-of-practice architecture uses advisory and procedural verbs — encourage and facilitate, monitor and evaluate, publish assessments of adequacy — backed by a consequential statutory fallback (implementing acts can impose common rules if codes are inadequate). The political marketing of the package is consequential; the front-loaded mechanism is advisory-to-procedural with a contingent fallback that may or may not fire. GDPR is a mixed case rather than a clean one: the regulation is statutorily consequential and has produced major fines, but DPA resource limits and uneven cross-border enforcement mean many controllers experience it day-to-day as documentation, complaint handling, and reputational exposure, with consequential sanctions concentrated in selected cases. The profile an institution invokes (consequential) is the profile the architecture promises; the profile the mechanism delivers (procedural with patchy enforcement) is one step down.

The diagnostic across all four joints: ask what the institution must actually do when feedback arrives. The compelled verb reveals the profile. Receive or display is decorative. Acknowledge or summarise is advisory. Expose to an audience is reputational. Give a reasoned answer in a record another actor can review is procedural. Change legal, fiscal, or operational state is consequential. The phrasing of the architecture is not the grade; the verb the mechanism compels is.

V. Prior art

The nearest existing map is legalisation theory. Abbott and Snidal (“Hard and Soft Law in International Governance,” International Organization 54:3, 2000) decompose legalisation along three dimensions — obligation, precision, delegation — and argue that “soft” and “hard” law are combinations across all three, not endpoints on one axis. The five authority profiles narrow that framework into a scalar diagnostic for a broader class of feedback channels (consultations, audits, court judgments, expert reports, referenda, regulatory comments). Shaffer and Pollack (Minnesota Law Review 94, 2010) supply the conflation argument: hard and soft law can substitute for, complement, or undermine each other, and reform debates routinely treat them as interchangeable when they are not. The conflation table in §IV is an application of that insight to receiver-side response duties.

The split between answerability and enforcement was made by Schedler (“Conceptualizing Accountability,” 1999); Bovens (“Analysing and Assessing Accountability,” European Law Journal 13:4, 2007) formalised accountability as an actor–forum relation involving information, debate, judgment, and consequence. The profiles turn that split into a field diagnostic — the lower three (decorative, advisory, reputational) are answerability without enforcement; the upper two (procedural, consequential) add review, legal process, or enforcement channels. Mashaw’s regimes (political, legal, bureaucratic/managerial, market) classify the source of accountability; the profiles classify the grade of consequence once the source operates.

The doctrinal infrastructure for the procedural profile comes from US administrative law (Nova Scotia Food Products 1977, State Farm 1983, Ohio v. EPA 2024), already engaged in The Procedural Object. A comment is procedurally binding when an agency must produce a reasoned response under judicial review for arbitrary-and-capricious action — the textbook procedural mechanism.

VI. Relation to the corpus

The procedural object essay names the sender-side discipline: what an artifact must become before bureaucracy can process it. Its seventh field, traceable consequence, ends with the artifact entering the response space. This essay starts there. Once an artifact has entered the response space, what grade of response is the institution obligated to produce? The five profiles answer that.

Powerless Intelligence names the conversion-capacity primitive: authority × resource × answerability. As feedback profiles become more forceful, failure in any element pushes the realised grade downward. Feedback Authority classifies the grade of cost once some conversion path exists. The triad asks whether conversion capacity is present; the profiles ask what conversion grade has actually been built.

The Legitimacy Gate covers the binary question of whether procedural attachment exists. The profiles grade what procedural attachment does once it exists. Mandate Gap names a specific instance of the conflation diagnostic — formal closure at a higher grade while operational capacity sits at a lower one. Theatrical Accountability lives at the bottom of the field guide; the ritual is what feedback authority looks like when the grade is decorative or advisory and the architecture is sold as something stronger.

The repair-or-override primitive in Mechanism Analysis is the procedural profile instantiated correctly. The lifecycle-ownership organ described in The Mandate Trap requires graded response duty to function; the profiles give the typology that organ would deploy.

The receiver-side architecture now reads: gate (Legitimacy Gate) → response-duty preconditions (Powerless Intelligence) → output classification (this essay) → operational mandate (Mandate Gap).

VII. Wrong-repair warning

The reform anti-pattern: pass a law that names a stronger profile without installing the stronger mechanism. Some post-2008 financial-regulation packages exhibited this pattern — naming consequential authority while installing procedural mechanism, naming procedural mechanism while installing reputational. The GDPR enforcement-budget gap is a current example: statutory text at the consequential profile, enforcement resourcing at the procedural, mechanism delivery at reputational-with-occasional-fine.

Rhetoric without mechanism is not free. It tends to lower the actual feedback authority over time, as actors learn the gap. Once stakeholders know that comply-or-explain produces only an explanation, the explanation itself becomes formulaic, the reputational cost decays, and the institution drifts from reputational down toward advisory. Once an encourage and facilitate code-of-practice regime is known to deliver no enforcement under the statutory fallback, the targets adjust their behaviour to the realised mechanism, not the advertised one. The architecture’s eventual grade is the lower of what was promised and what the mechanism can actually carry.

Adding more feedback channels does not raise the grade. If the existing channels deliver advisory or reputational and the diagnosed defect is that consequential force is missing, opening a new advisory channel adds noise. The diagnostic for proposed reforms is the inverse: if the architecture names grade N+1, what mechanism compels the institution to act at N+1 rather than N? If no such mechanism is named, the reform will deliver at grade N regardless of rhetoric. The verb the architecture compels is the grade. Everything else is performance.

VIII. Close

The operational diagnostic: for any feedback channel an institution promises, ask what it must do when feedback arrives. The compelled verb reveals the profile.

• Display or receive it, no acknowledgement required. Decorative.
• Acknowledge and summarise. Advisory.
• Expose the response or non-response to an audience. Reputational.
• Give a reasoned answer in a record another actor can review. Procedural.
• Change legal, fiscal, or operational state. Consequential.

Escalation above the grade the verb compels requires a mechanism, not a phrase. The repair-or-override primitive installs procedural force; statutory text without budget or sanction installs reputational; reputational consequence under weakening enforcement drifts toward advisory. The architecture’s eventual grade is what it can compel.

A feedback channel is defined by what non-response costs. Truth alone does not bind; changed institutional cost does.

Related:

• The Procedural Object — sender-side admissibility and triggering; the seventh field (traceable consequence) ends where this essay begins.
• Implementation Ledger — execution-side companion; once a feedback profile forces a response, the ledger names the record structure that converts accepted response into traceable execution under conditional closure.
• Powerless Intelligence — authority × resource × answerability as the conversion-capacity primitive; failure in any element pushes the realised grade downward.
• The Legitimacy Gate — binary procedural attachment as gate; the five profiles grade what the gate produces.
• Mandate Gap — formal closure at one grade while operational capacity sits at a lower one; a specific instance of the conflation diagnostic.
• The Mandate Trap — lifecycle-ownership organ requires graded response duty.
• Mechanism Analysis — the repair-or-override primitive; procedural force installed without overclaiming statutory.
• Theatrical Accountability — what the decorative-and-advisory end of the field guide looks like in operation.
• The Stack — Feedback (Layer 10); this essay fills the open slot the Stack §IV identified.
• What Bureaucracy Is — the positive synthesis: bureaucracy as the runtime of traceable discretion, of which feedback authority supplies the contest-and-close function.