The Clearance Is a Veto

Government runs on information it collects from the public: surveys, forms, returns, applications. Collecting it costs the public time, and that cost is real.

The clearance requirement responds to it by making an agency obtain approval, from a central office, before it may collect information at all.

The stated purpose is burden reduction, and burden reduction is a legitimate aim that had been neglected for a long time before the requirement existed.

What the mechanism actually does is subject every question government wants to ask to review by an office that is not the agency asking it.

A review that can refuse a collection is a review that can decide what an agency is permitted to find out, which is a different and much larger power.

The burden estimate itself deserves scrutiny, because it is the number the entire framework turns on and it is produced by the agency proposing the collection.

Nothing in the framing acknowledges this. Approval is described in the language of forms and hours, and the substance of what would have been learned does not appear.

Hours per response multiplied by expected responses is a defensible method and a highly manipulable one, and there is no independent check on either input.

The clearance record is public and searchable, which is more transparency than most of what this publication examines, and it is transparency about process rather than about consequence.

An agency that wants a collection approved has an obvious incentive to estimate low, and one that has been told informally to expect resistance has an incentive to estimate lower still.

You can find out that a collection was approved, modified or withdrawn. You cannot find out what the agency would have known had it proceeded as proposed.

So the ledger of burden saved rests on figures supplied by the parties whose proposals the ledger is scoring, which nobody would accept in any other accounting context.

The timeline is the part that operates as a constraint without ever refusing anything. Clearance takes months, and a question that takes months to authorise is not a question you ask about a fast-moving situation.

None of that is fraud. It is what happens when a single unverified number becomes the currency in which a decision is denominated.

So the process systematically favours routine recurring collections over new ones, which is another way of saying it favours measuring what was already measured.

The register carries the notices that each collection requires, and reading a run of them shows how much of the traffic is renewal of existing instruments.

An agency wanting to understand something new must justify the burden in advance, against a standard that weighs a certain cost in hours against an uncertain benefit in knowledge.

That comparison is structurally unfair to the benefit, because the hours are calculable and the value of information you do not yet have is not.

The predictable result is that agencies propose fewer novel collections, and the reduction shows up in the statistics as a success.

The administrative framework sets out the standards, and they are reasonable standards applied by people doing a job they were given.

The problem is not the office or its staff. It is that a burden-reduction instrument has no counterweight: nobody's job is to argue for the collection that was not proposed.

There is a further effect on independence. An agency whose ability to gather evidence depends on approval from the centre is an agency whose evidence base can be shaped from the centre.

That is available to any administration, requires no instruction to anybody, and leaves no trace beyond a slower queue and a few withdrawn proposals.

The safeguards are procedural and thin. There are deadlines, and a collection can proceed if the review lapses, which happens rarely enough not to matter.

What would balance it is not weaker review but a published account on the other side: what each refused or withdrawn collection was intended to establish.

A ledger of burden saved is only half an accounting. The other half is the ignorance purchased with it, and at present nobody is required to record that at all.