Asked and unanswered — the board that publishes how long a documented request to a public body has gone without an answer — went from ten rows to 29 over the past two days. It now covers 24 public bodies and carries 408 open priority recommendations, with 72 implemented since the previous letters.

The same day it grew, every row already on it was audited. Four were wrong.

What was added

The bulk of the growth is the 2026 round of GAO priority open recommendation letters. Once a year the Government Accountability Office writes to each agency head listing the recommendations it says warrant that person's personal attention, and — the part that makes it usable here — it publishes the arithmetic: how many it flagged last time, how many the agency implemented since, how many stand now.

Twenty-two agencies are now on the board, including Defense, Homeland Security, Health and Human Services, Veterans Affairs, Justice, Treasury, Labor, the Interior, Housing and Urban Development, Social Security, Energy and the Office of Management and Budget.

Three readings are worth stating plainly, because a bare count hides them:

Homeland Security implemented five over the year and had seven added, ending on 39 — exactly the number it started at.
Housing and Urban Development and Social Security each implemented none. Those are GAO's own words, not an inference from a count.
Defense took its list from 79 down to 53, the largest single reduction on the board. Progress belongs on a responsiveness ledger as much as backlog does, and this row exists partly to make that visible.

No 2026 letter could be located for the Department of State or for NASA. That is recorded as a gap in this board's coverage, not as a finding about either body.

What was corrected

A board about unanswered requests has one failure mode worse than all the others, and this desk hit it.

The row for the House Oversight Committee's 31 August 2025 request to the Treasury Department carried the status no published response located. Treasury responded twelve days later. Reporting on 12 September 2025 quotes a Treasury letter to the Committee saying it "plans to fully cooperate with the Committee and intends to provide all information responsive to your request."

The row was wrong for eleven months. It is now recorded as answered, and it states its own limit: the letter itself has not been published by Treasury or the Committee, so the response is known to this desk through reporting rather than from the document, and whether the records were subsequently produced is a separate question the board does not claim to have settled.

Three further corrections landed with it:

The Defense row recorded its prior letter as identifying 53 recommendations in May 2025. The letter's own footnote cites June 2025 and 79. The current total had been copied into the prior-count slot.
The New Mexico row attributed a quotation to a "9 July 2026 letter." 9 July was the press release; the letter is dated 30 June.
The Bank of America row named the asker as Senate Finance's ranking member. On the date of that letter, 4 April 2024, Senator Wyden was the committee's chairman — a 2026 title had been carried back onto a 2024 ask.

All four are on /corrections with the prior text kept verbatim, per this desk's standing rule that a record which was ever served to a reader gets a public correction rather than a quiet edit.

Why the audit ran at all

Because absence is the one status that decays.

Every other entry on this board is a fact about a document: a filing date, a published figure, a recommendation an agency disagreed with on the record. Those are wrong the same way forever, and re-reading the document catches them.

"No response located" is different. It is a fact about what this desk has been able to find, and it can be true the day it is published and false a week later without anything happening here at all. Somebody answers, somewhere nobody was looking, and the page goes on asserting a silence that has ended.

The board's standing rule is now explicit: publishing an absence creates a duty to re-check it. The method used yesterday — handing every live row to an independent reader whose only instruction is to assume it is wrong and go to the primary document — is the mechanism, and it is the same adversarial pass this desk has previously run against its case files.

One further note for anyone who audits records for a living. This desk pins published figures in automated tests so a later edit cannot quietly drift a federal number. There was such a pin on the Defense row. It asserted the wrong figure and passed every run, because it had been written from the row rather than from the letter — it was checking the claim against itself. A pin is only ever as good as the source it was read from.

What this entry does not claim

It does not claim any agency ignored anyone. A long count on that board is a measurement of elapsed time and nothing else; records requests are slow for many ordinary reasons, and priority recommendations are frequently difficult, expensive and multi-year. Nor does it characterise motive: where a body has declined on the record, the board says declined; where it answered, the board now says answered.

August 15: the responsiveness board goes federal, and four of its own rows are corrected.

What was added

What was corrected

Why the audit ran at all

What this entry does not claim

Source links