Administrative Panic and Pre-Decisional Authority Collapse in Complex Institutions

Institutions increasingly incur irreversible harm before adjudication, legal review or factual determination. Programmes are cancelled, staff removed and precedents set not through findings but through rapid pre-decisional action under asymmetric pressure. The usual explanations — ideological capture, moral conflict, procedural error — do not account for the timing. Prior to any of it is a question about which institutions break, since the same compressed timeline reaches many and only some do. Survival--legitimacy decoupling — the two governed by different actors on different clocks — answers it. Pressure campaigns work because they target survival rather than persuasion, and under decoupling an institution experiences pressure as threat rather than contestation. This paper names administrative panic as a governance failure mode arising from a rate mismatch: pressure escalates faster than authority can legitimately be exercised. Under reputational asymmetry, low-cost coordination and impact-based governance, an institution rationally substitutes speed for judgement unless an architectural constraint prevents it. Courts and post-hoc review cannot repair the result, because legitimacy loss and precedent formation occur before adjudication begins. It further separates experienced harm from manufactured unsafety — a pressure-induced condition in which danger is produced upstream of conduct by coordinated escalation aimed at the institution. Confusing the two collapses process into reaction and converts pressure into de facto authority. The remedy is the Panic Buffer: a pre-committed rule that prohibits irreversible action during asymmetric pressure until minimum process conditions are met. It governs when an institution may act, never what it must decide, and it triggers without assessing the credibility or intent of the pressure — which is the point, since those assessments are least reliable exactly when they are being made. Two conditions make it an instrument rather than a slogan, and both are easy to state wrongly. It must be adopted prospectively and generally, binding matters the institution cannot yet see — a test of generality, not of timing, since institutions adopt governance after their first crisis and a timing rule would disqualify everyone who has learned anything. And it must terminate: an indefinite pause is not the absence of a decision but the same decision taken without being announced, because what is paused disperses while the pause runs. The contribution is architectural rather than normative. Naming the panic window as a distinct phase is what makes it governable.