A single 2023 coordination recommendation sits open in three separate priority letters to financial regulators dated June 2026

What is happening?

The Government Accountability Office designates a subset of the open recommendations it holds against each federal agency as priority, and writes once a year to that agency head about them. The letters of the 2026 cycle read here are dated from 2026-06-09 to 2026-08-12, one agency at a time.

Three of those letters carry the same item. The letter to the Securities and Exchange Commission of 2026-06-10, the letter to the Comptroller of the Currency of 2026-06-23 and the letter to the Federal Reserve Board of 2026-06-25 each record a 2023 finding that financial regulators lacked an ongoing mechanism for coordinating on blockchain risk, and each records the recommendation that the regulators establish such a mechanism jointly. At the Securities and Exchange Commission it is the only priority recommendation on the list, and that letter records it as not yet fully implemented.

So the same open item is reported three times, in three documents, each addressed to one agency head, and the action it asks for is one that the addressees take together. None of the three letters names a lead agency, a sequence, or a test that one regulator acting alone could meet.

Around that item the wider mechanism is measured and published. GAO reported in January 2026 that 77 percent of the recommendations it made five years earlier had been implemented, against a target of 80 percent that GAO sets for its own performance measure, and it sent 31 priority letters in fiscal year 2025.

Whose problem is this?

The item that appears in three letters has three addressees and no owner named anywhere in the record read here. The population file for this document records not-derivable for the same reason the first row above is empty.

Where does this problem end?

A reader who thinks the recommendation is right and a reader who thinks it is wrong can both agree on the dates and the counts set out below.

What is the state now, and what should it be?

Now — the twelve letters read, in publication order.

Where a cell carries a dash, the letter read does not give that figure. The counts in this table are as of each letter date in the second column and not as of 2026-08-24.

What it should be

The recommendation states its own target, and it is not a number. The three letters record GAO as having recommended that the financial regulators jointly establish an ongoing mechanism for coordinating on blockchain risk. That is the condition the item names, and the letters do not add a date, a lead agency or a test that one regulator could satisfy alone.

The system around it has one published target, and it belongs to a different denominator. GAO reports the share of its recommendations implemented five years on against a target of 80 percent, and reported 77 percent in January 2026. That target is GAO measuring its own performance, government-wide and across all recommendations rather than priority ones. No source read here states a numeric target for priority recommendations specifically.

How big is it?

The system, government-wide and across all recommendations. The Performance and Accountability Report for fiscal year 2025, published 2026-01-29, reports $62.7 billion in financial benefits against a target of $50 billion, a five-year average of $64.5 billion, about $68 returned for every dollar invested and a ten-year average return of $116. In the same year GAO made 1,833 new recommendations, recorded 1,295 other benefits, issued 671 products and sent 31 priority letters. It estimates future benefits of $132 billion to $251 billion from implementing the open matters for congressional consideration and the open recommendations, as of October 2025.

None of those figures is apportioned to priority items. No source read here divides the benefits, the recommendation counts or the achievable-savings estimate between priority recommendations and the rest.

The twelve letters read. Two sums, on two different bases, side by side.

Those two sums are not a rate and they are not divided here. The first counts items standing open at twelve different dates and the second counts implementations over twelve different intervals of different lengths. The fiscal year 2025 cycle comprised 31 letters, so twelve is a partial sample and no cycle-wide figure follows from it.

Affected population. No count of people is derivable from these sources. They count recommendations, dollars and implementation rates, and none of them converts any of those into a number of persons.

Under what conditions does it arise?

1. The action asked for is joint and the mechanism that carries it is not. GAO found in 2023 that financial regulators lacked an ongoing mechanism for coordinating on blockchain risk, and recommended that the regulators establish one jointly. That single finding reaches each regulator through a separate letter addressed to that agency head.

2. At one addressee it is the whole list. The Securities and Exchange Commission letter of 2026-06-10 carries exactly one priority recommendation, and it is this one. A list of one has no second item that could move.

3. Two of the three carry a second shared item. The Federal Reserve Board and the Comptroller of the Currency both carry a July 2024 finding on retrospective review of existing rules, so the overlap between those two lists is more than a single entry.

4. Nothing on the calendar. GAO designates these items and no statute sets a date by which an agency must implement one of them. The consequence of an item staying open is that it appears in the next letter.

5. The lever that exists routes through Congress. Footnote 5 of every letter read cites the James M. Inhofe National Defense Authorization Act for Fiscal Year 2023, Public Law 117-263, section 7211(a)(2), 136 Stat. 2395, 3668, codified at 31 U.S.C. 719 note, under which GAO helps Congress identify oversight actions including incorporating priority recommendations into legislation. GAO also names hearings and the withholding of funds.

6. The cycle publishes across months rather than at once. The twelve letters read here are dated from 2026-06-09 to 2026-08-12, so a reading taken at any point inside that window covers only the letters published up to it, and the earliest letters are the ones that have had the least time since the prior cycle.

7. GAO gives its own account of why implementation takes time. It moved its measurement window from four years to five, on the stated ground that the problems facing government have become increasingly complex and often involve multiple agencies, which increases the time agencies take to implement.

What has been tried?

What was found?

Why is it still unsolved?

Coordination failure — the action the recommendation asks for is one the regulators take together, and the mechanism that carries the recommendation addresses them one at a time.

The first part is the shape of the item itself. GAO found in 2023 that financial regulators lacked an ongoing mechanism for coordinating on blockchain risk and recommended that the regulators establish one jointly. Each regulator receives that finding in its own letter, addressed to its own head, and each letter reports the standing of that agency alone. The letters are addressed one agency at a time and the action they ask for is one that the regulators take together. None of the three names a lead, a sequence, or what implementation by one regulator acting alone would look like, so the item can stand open at three addressees at once, and no letter read records a decline by any of them.

The second part is that at one addressee the item is the entire list. The Securities and Exchange Commission letter of 2026-06-10 carries this recommendation and nothing else, so there is no second item on that list whose implementation could move the count.

The third part is that nothing on the calendar attaches. GAO designates and no statute sets an implementation date, so the consequence of an item staying open is that it appears in the next letter. The statutory provision GAO cites in footnote 5 of every letter routes oversight through Congress rather than through the agency, and no source read here records an action taken under it on this item.

The fourth part is that GAO publishes its own account of why implementation takes time, and that account is about how many agencies a problem spans. GAO widened its own measurement window from four years to five on the stated ground that the problems facing government have become increasingly complex and often involve multiple agencies, which increases the time agencies take to implement. On that account a jointly addressed item taking longer is the ordinary cost of the item spanning agencies, and the 77 percent measured at five years stands three points under the target GAO sets for itself rather than describing a system that has stopped moving.

The fifth part is a reading the record here does not support, and it is set down so that it is not drawn by accident. Across the twelve letters read the overall implementation rate and the count of priority items implemented do not move in the same direction. That is a description of twelve letters of a cycle whose total size is not established here, three of the five letters at zero belong to agencies holding one, three and five priority items, and the letters give no reason for either group. It is carried at medium grade as a property of this sample, and no account of agency behaviour is built on it.

What observation would mean it is solved?

Candidates — (a) the three letters stop carrying the coordination item and a mechanism exists that each addressee names (b) the item leaves the lists recorded as implemented rather than as removed or closed as no longer valid (c) the government-wide measure of recommendations implemented at five years reaches the 80 percent target.

(a) is the condition the recommendation itself names. Because no letter states a test for partial implementation, an observation that one regulator has done its share is not available from this record.

(b) is observable because the letters publish the reconciliation. The letters set out designated, implemented, priority status removed, closed as no longer valid and added, and in the six letters carrying a full reconciliation the arithmetic closes exactly: Housing and Urban Development 9 designated, 0 implemented, 1 priority status removed, 2 added, 10 open · Defense 79, 17 implemented, 16 removed, 7 added, 53 open · Homeland Security 39, 5 implemented, 2 closed as no longer valid, 7 added, 39 open · Office of Management and Budget 37, 11 implemented, 5 removed, 8 closed as no longer valid, 7 added, 20 open · Office of Personnel Management 14, 3 implemented, 3 removed, 8 open · NASA 3, 0 implemented, 1 added, 4 open. An implementation and a removal are therefore separately observable in the published counts.

(c) counts the whole system rather than this item. The 80 percent target is stated for all recommendations at five years and government-wide, so movement in it would not show whether this recommendation had been implemented, and no source read here states what implementing this recommendation would contribute to that measure.

What is it connected to?

Needs a new measurementthe letters and the annual report read here do not describe how this problem stands in relation to others. The one adjacency they carry is a citation. The letters point to a separate GAO product, the High-Risk Series report titled Heightened Attention Could Save Billions More and Improve Government Efficiency and Effectiveness, GAO-25-107743, of 2025-02-25. No source read here states how an item on that list relates to a priority designation, whether membership of one bears on the other, or whether the coordination item sits inside any high-risk area.

What these sources do not say

• No government-wide count of priority recommendations. The Performance and Accountability Report gives the number of letters sent, 31 in fiscal year 2025, and the aggregate implementation rate, and it does not sum priority items across agencies. The recommendations database that carries per-item records returns an automated refusal to unattended requests, so no total was obtained there. That is an access refusal and not a statement that no total exists.
• No cycle-wide figure. The full index of priority recommendation letters and the search endpoints return the same automated refusal, and the landing page exposes ten report numbers at a time. The twelve letters read here therefore sit against a cycle whose size is not established, and no share of agencies at zero can be stated for the cycle.
• No position after the letter dates. Every per-agency count above is as of the letter date in the second column of that table, and those dates run from 2026-06-09 to 2026-08-12. The database described as updated daily returns an automated refusal, so nothing here establishes whether any of those counts changed afterwards. Access was refused, and the counts are not recorded as unchanged.
• No agency response. These letters are one-way correspondence from GAO to an agency head and carry no agency comments section. Across the twelve read there is no passage in which an agency sets out why an item is still open, so the position of the addressees on this recommendation does not appear in this record at all.
• No target specific to priority recommendations. The 80 percent figure is the target GAO reports against for recommendations implemented five years on, government-wide and across all recommendations. No source read here names a target for the priority subset.
• No apportionment of the benefit figures. The $62.7 billion in financial benefits for fiscal year 2025 and the estimate of $132 billion to $251 billion in future benefits are stated across all recommendations and all open matters, and no source read here divides either between priority items and the rest.
• No test for partial implementation. The three letters record the coordination item without naming a lead agency, a sequence or a standard that one regulator acting alone could meet, and none of them states that one regulator could not meet it.

See the evidence

The load-bearing figures here come from the letters themselves and from the annual report they cite, rather than from the summaries on the landing page. Twelve agency letters carry the per-agency counts, the reconciliations and the repeated government-wide rate, the three letters to financial regulators carry the coordination item in near-identical terms, and the Performance and Accountability Report carries the scale figures and the stated reason for widening the measurement window. Where the record is unsettled it is left visible rather than resolved: the cycle contains more letters than the twelve read, the government-wide count of priority recommendations is not published in any source read, the position after each letter date is unestablished because the surface that would carry it refuses unattended requests, and the addressees have no voice in this correspondence because the letters carry no response section. The totals of 181 and 44 are arithmetic over twelve letters and are graded medium for that reason, and they are not divided into a rate because they rest on different bases. This is a research-based definition, so observation_refs is empty and provenance_mode is press-derived.

People affected

Estimated range Not derivable

The reason and what is missing are listed under “What is missing” below