Every Emergency Power Needs an Expiration Date
Every emergency power should be born with a death certificate. Not because emergencies are unreal, or because governments should be slow when speed matters, but because exceptional authority has a structural tendency to survive the condition that justified it.
The familiar argument happens at the front door: Is the threat serious enough? Is this surveillance authority, information-sharing arrangement, emergency procurement rule, temporary command structure or restriction actually necessary? Those are essential questions. But there is a second decision that receives less attention: Who has to prove that the exception should still exist thirty, sixty or one hundred and eighty days later?
If nobody does, the default changes. What began as a temporary response becomes infrastructure. A crisis office acquires routines. A special data-sharing channel becomes convenient. An extraordinary procurement path develops suppliers and expectations. A surveillance permission remains technically available because removing it now looks like an additional decision. The emergency ends, but the architecture stays.
That’s why expiration isn’t administrative housekeeping. It’s a security control.
The clock is already part of a serious emergency law
Some legal systems already encode versions of this principle. Under Article 352 of India’s Constitution, a proclamation of emergency ordinarily requires approval by both Houses of Parliament within one month or ceases to operate; the article provides a separate timetable when the House of the People is dissolved. An approved proclamation is itself time-limited, with further continuation requiring parliamentary approval. Canada’s Emergencies Act gives a public-order emergency a thirty-day life unless it’s revoked or continued. A continuation may run for no more than another thirty days, and before continuing it the government must review the orders and regulations already in force and revoke or amend those it no longer considers necessary. When the declaration expires, the measures made under it expire with it. The Act also requires a post-emergency inquiry.
The United Kingdom’s Civil Contingencies Act takes a similar approach to emergency regulations: they lapse no later than thirty days after being made, and sooner if the regulations specify. They must also be laid before Parliament as soon as reasonably practicable and lapse seven days after being laid unless both Houses approve them. Fresh regulations can be made, but a new set doesn’t escape the scrutiny requirement. An expiry date and a renewal mechanism do different work; neither substitutes for a decision about whether the power should exist.
These systems are not identical, and a sunset clause does not make an emergency measure wise, lawful or proportionate by itself. What the clock changes is the default. Continuation becomes a new decision instead of institutional inertia.
Five controls should exist on day one
An emergency authority should not be created with a vague promise to review it later. Its exit architecture should be specified at the same time as its operational authority.
First, define the triggering condition. The power should answer a particular state of affairs, not an atmosphere. A government may need latitude in a rapidly changing crisis, but the trigger should still be concrete enough that a later reviewer can ask whether the condition continues to exist.
Second, name the authorizing office and the owner of renewal. Exceptional powers often become difficult to unwind because responsibility becomes distributed. One unit uses the data, another administers the contract, another owns the legal authority, and nobody owns the decision to stop. Renewal should have a named institutional owner.
Third, fix the expiry or review date at creation. The date should be close enough to force reconsideration while leaving enough operational stability to manage the emergency. Different powers can have different clocks. A temporary data-retention rule may deserve a different review period from a crisis procurement authority. What matters is that silence cannot extend either one.
Fourth, define the evidence threshold for renewal. The original crisis cannot be treated as permanent evidence. Renewal should require current facts showing both that the relevant threat or disruption continues and that the specific exceptional measure remains necessary. This is where Canada’s requirement to review existing emergency orders before continuing a declaration is especially important: the question is not simply whether an emergency persists, but whether each extraordinary tool still earns its place.
Fifth, document the route back to ordinary governance. An emergency power can formally expire while its effects remain embedded in contracts, databases, inter-agency workflows or command practices. The exit plan should say what happens to emergency contracts, temporary delegations, special access permissions, retained data, exceptional reporting lines and any records created under the regime.
Renewal should be harder than adoption
A crisis can justify making an initial decision under uncertainty. That does not justify renewing the same decision under the same evidentiary standard after institutions have had time to learn.
The burden should therefore change over time. At adoption, decision-makers may reasonably ask whether the risk of doing too little is intolerable. At renewal, the state should have to show what the measure actually did, what harms or costs it produced, what less exceptional alternatives now exist, and why the ordinary legal or administrative system still cannot carry the load.
That approach does not require automatic abolition. A power can be renewed repeatedly when the evidence supports it. The important point is that renewal must be affirmative. Somebody must put a name, date and evidentiary reason on the decision to continue.
The principle extends beyond formal states of emergency
The most durable exceptional systems are not always created by a dramatic declaration. They may appear as temporary procurement flexibilities, accelerated information-sharing, emergency access to databases, provisional surveillance tools, special crisis teams or temporary waivers that never quite receive a final accounting.
These arrangements can be operationally sensible. In fact, that is precisely why they need clocks. Useful emergency machinery is the machinery most likely to become ordinary through convenience rather than conscious choice.
The dangerous emergency power is not only the obviously abusive one. It is also the one nobody remembers is exceptional.
A mature security institution should be able to move quickly into an emergency posture and just as deliberately move out of it. That requires more than legal authority to act. It requires stable identifiers for exceptional measures, reviewable records of who authorized them, current evidence for continuation, and a controlled path back to ordinary rules.
Expiration is therefore not an anti-state constraint. It is continuity engineering for the state itself. It prevents crisis architecture from accumulating until nobody can distinguish temporary authority from permanent capacity. It forces institutions to remember why a power exists, who owns it and what evidence still supports it.
When the next emergency arrives, speed will matter again. The government that already knows how its exceptional powers end will be better positioned to act decisively without turning each crisis into another permanent layer of unreviewed authority.
References:
Constitution of India, art. 352(4)–(6). Government edition: https://www.legislative.gov.in/constitution-of-india/ ; article text: https://www.constitutionofindia.net/articles/article-352-proclamation-of-emergency/
Canada, Emergencies Act, R.S.C. 1985, c. 22 (4th Supp.), especially ss. 18, 23, 26 and 63, Justice Laws Website: https://laws-lois.justice.gc.ca/eng/acts/E-4.5/
United Kingdom, Civil Contingencies Act 2004, especially ss. 26–27: https://www.legislation.gov.uk/ukpga/2004/36/contents
UK Cabinet Office, Civil Contingencies Act 2004: Call for Views, 14 July 2026, emergency-powers explanation: https://www.gov.uk/government/calls-for-evidence/civil-contingencies-act-2004-call-for-views/civil-contingencies-act-2004-call-for-views-html
The views expressed are those of the authors and do not represent the views of CNAWS.
Image Source: ChatGPT
