Myrtle Beach, SC
This Forum has asked for a published number in every one of the last seventeen questions. One was published. Eleven days ago the audit office reported on it and the report is the answer to this question.
The savings page claimed roughly a hundred and ten billion dollars from terminated contracts, grants and leases. The audit found that of the thirteen thousand four hundred and seventy-six contracts listed as terminated, nearly forty-five percent had no termination action at all — about twenty-seven billion dollars of claimed savings against contracts that were never cancelled. Another fourteen percent could not be checked because identifying information had been left off.
Only about twenty-seven percent of the reported savings could be traced to the methodology the page itself said it was using. Sixty percent used some other calculation that was not disclosed, or could not be explained.
Ninety-six percent of the claimed savings on grants could not be verified.
And the office did not respond to the auditors’ requests for information or interviews.
I work in marketing and I want to be exact about what that means, because it is worse than the thing this question was worried about.
The question asks whether replacing a slogan with numbers would cost the coalition. It turns out a number was published, and the number was a slogan with decimal places.
So my answer is yes, retire the phrase, and add the thing this file did not know to ask for. Not three published numbers — three published numbers produced by somebody other than the person being measured.
That is the whole lesson of eighteen months. We spent a year and a half demanding a scoreboard and never once specified who should keep it.
Novi, MI
A plant that posts its own scrap number and will not let the auditor see the method has not published a metric. It has published a claim. That is the finding of the last eleven days and it settles a question this Forum has been circling since December of 2024.
What the audit found, and the detail matters because the detail is the argument.
Of thirteen thousand four hundred and seventy-six contracts the page listed as terminated, the auditors could find no termination action for forty-four point nine percent — representing twenty-seven point four billion dollars in claimed savings. A further thirteen point eight percent could not be verified at all because the page omitted the identifying information needed to look them up.
Only twenty-seven point five percent of reported savings could be reconciled to the page’s own stated formula. Sixty point seven percent used an undisclosed method or could not be explained.
The methodology did not account for termination settlement costs, contract closeout costs, or work that was descoped from one contract and picked up on a follow-on — sometimes with the same contractor.
On leases, the General Services Administration told the auditors that a hundred and eight of the two hundred and sixty-four listed were already terminating before the office was established.
On grants, ninety-six percent of claimed savings could not be verified.
And the office declined to respond to requests for information or interviews.
In quality terms this is the outcome we predicted in March 2025 and got wrong about the mechanism.
I said then that the eight-billion-dollar entry that was really eight million was a repeatability failure and that the fix was publishing the contract identifier, the obligated value, and the termination date. That was the right fix and it was never adopted — the omitted identifying information is one of the audit’s explicit findings a year and a half later.
What I did not anticipate is the larger defect. Nearly half the terminations did not happen. That is not a measurement error. A measurement error means the quantity is wrong. This means the event is absent — the page reported an action that was never taken, at scale, for a year and a half, with no correction.
In a plant, a defect log listing repairs that were never performed is not a bad log. It is a fabricated record, and the discovery ends the inspector’s career regardless of whether the underlying defects were real.
So the answer to this question.
Retire the slogan, yes. Publish the three numbers, yes, and this file specified them correctly: improper payments as the outcome measure, unexecuted IT, and programs failing the screen that Congress will not repeal.
And add the specification we missed for eighteen months. The party publishing a metric may not be the party being measured by it. Improper payments work as a metric precisely because agencies report them under a statutory method and an independent office audits them. Nobody’s career depends on the number reading well, which is the only reason it can be believed.
A self-reported savings tally has none of that structure and was never going to survive an audit. It did not.
Will honesty cost the coalition? The audit has already answered. Anybody who needed the number to be right has just been told it was not, and anybody who did not need it to be right was never a measurement coalition to begin with.
Orange County, CA
Orange County has been asking for the reconciliation since March of last year. It arrived eleven days ago, produced by somebody else, and it does not reconcile.
Let me price what the audit actually establishes, because there is a distinction inside it that will get lost.
The strong finding. Nearly forty-five percent of the listed contract terminations had no termination action on the record — twenty-seven point four billion dollars in claimed savings attached to contracts that were not cancelled. One example: one point seven billion claimed on a Defense Health Agency IT contract that was identified for termination and never terminated or reduced.
That is not an estimate that came in high. That is a reported event that did not occur.
The methodological finding. Only about a quarter of the reported savings could be tied to the page’s own stated formula. The rest used an undisclosed calculation. And the formula, where it was used, ignored termination settlement costs — the amounts a contractor is legally entitled to when the government cancels for convenience — and ignored work moving to follow-on contracts.
In finance those are not footnotes. A termination settlement can consume a large share of the notional saving, and descoped work that reappears on another vehicle is not a saving at all. Any competent valuation nets both, and this one netted neither.
The distinction that matters. None of this establishes that there was no waste. The improper-payment series still exists, still runs into the hundreds of billions annually, and is still published every year by the audit office regardless of who is in government. That problem is real and was real before any of this started.
What the audit establishes is that the claim was not a measurement of that problem. It was a tally of the office’s own asserted activity, unaudited, self-published, and now substantially unsupported.
Those are different failures and conflating them will be the temptation for the next eighteen months. One side will say the audit proves there was no waste. The other will say the audit is a bureaucratic counterattack. Both will be wrong, and the correct sentence is narrow: a specific claim about specific contracts did not survive a check against the contracting record.
So on the question. Retire the phrase. Publish the three numbers. And specify the auditor in the same breath, because we spent eighteen months asking for a scoreboard without ever saying who keeps it, and that omission is now the most expensive thing in this file.
If publishing audited numbers costs a coalition, the coalition was buying something other than numbers. That was always true and it is now demonstrated rather than asserted.
Bismarck, ND
Run it the other way, one last time.
A Democratic administration stands up a cost-cutting office, claims something over two hundred billion in savings, publishes a page listing thirteen thousand terminated contracts, and an audit eighteen months later finds that nearly half of them were never terminated, that only a quarter of the claimed savings match the office’s own stated method, that ninety-six percent of grant savings cannot be verified, and that the office refused to speak to the auditors.
Would our side accept “the direction was right”? We would not. We would say the number was invented, and we would be entitled to say it, and we would say it for a decade.
Same answer here. I have run this test eighteen times in this file and this is the one where it costs the most, and it is the one where it matters.
In compliance work the rule is simple and it does not bend. An examiner who publishes an unverified figure has done more damage to his own program than any evasion by the examined party could. The examined party has an interest in the finding being wrong. The examiner has no excuse, and when his figures fail an audit, every accurate finding he ever made goes with them.
That is now the position. The improper-payment problem is real — hundreds of billions a year, two point eight trillion cumulative since 2003, published by the audit office every year and never seriously disputed by anybody. It was real before January 2025 and it is real today.
And the next person who raises it will be answered with a page listing six thousand contracts that were never cancelled.
So: retire the slogan. Publish the three numbers. And specify that they are audited by an office that does not report to the person being graded.
Would honesty have ended the coalition? Some of it. The part that wanted the number to be large would have left when the number came in small and dull.
The part that wanted the leak fixed would have stayed, and that part would still exist today with something to point at.
Clermont, FL
If the ride posts a wait time and a safety check you can argue about the ride. If it is a rope and a speech it is junk. That was the test this Forum wrote in March of last year and it now applies to us.
Eleven days ago the audit came out. Nearly half the terminations on the page never happened. A quarter of the savings match the stated method. Ninety-six percent of the grant claims cannot be verified. Leases counted that were already ending before the office existed. And the office would not talk to the auditors.
I consult with small businesses and I have watched this exact thing destroy owners who were right about the underlying problem.
The pattern is always the same. The owner is correct that money is leaking. He is impatient with the bookkeeper. He brings in somebody who produces a dramatic number very fast. The number is checked, badly, or not at all. Eighteen months later an accountant goes through it and finds the number was assembled rather than measured.
And here is the part that matters: the leak was still real. It was real the whole time. But the owner has now spent every bit of credibility he had, the staff has a permanent answer to any future finding, and the leak goes on being real for another decade because nobody can raise it without being handed the old number.
That is where this ends up. The improper-payment figure has been published every year since 2003 and runs into the hundreds of billions. It did not need any of this. It was sitting there the whole time and this Forum spent seven questions before anybody cited it.
So: retire the slogan, yes. Publish the three numbers, yes. And publish them from an office that has nothing to gain from how they read.
Would that have thinned the crowd? Yes, and it should have. A crowd that leaves when the number turns out to be dull was never coming to fix anything.
Jacksonville, FL
Ports grade cargo against a manifest, not against a speech on the dock. Eighteen months ago this Forum wrote the manifest. Eleven days ago somebody checked it.
The audit, in the fields that matter.
Thirteen thousand four hundred and seventy-six contracts listed as terminated. No termination action found for forty-four point nine percent — twenty-seven point four billion in claimed savings against contracts still in force. Thirteen point eight percent unverifiable because identifying information was omitted from the listing.
Only twenty-seven point five percent of reported savings reconciled to the page’s own stated method. Sixty point seven percent used an undisclosed calculation or could not be explained at all.
No accounting for termination settlement costs, closeout costs, or descoped work moving to follow-on contracts with the same contractor.
A hundred and eight of two hundred and sixty-four leases already in termination before the office existed.
Ninety-six percent of grant savings unverifiable.
No response to the auditors’ requests for information or interviews.
Now the answer to the question, and I want to give it in the order the evidence supports rather than the order that feels good.
First, the demand in this question was right and it was right from the beginning. Retire the phrase, publish improper payments, unexecuted IT, and the programs Congress will not repeal. Every one of those three is an outcome measure produced by somebody who does not benefit from how it reads. That is the whole design and this file identified it correctly in March of last year.
Second, the demand was partially met in form and the form failed. A number was published. It was the wrong kind of number — a tally of the office’s own asserted activity rather than a measure of the condition — and when checked against the contracting record it did not stand up.
Third, and this is the part I would put at the top of the manifest: the underlying problem is untouched and undisputed. Improper payments run into the hundreds of billions a year. Two point eight trillion cumulative since 2003. Federal IT still spends the large majority of its budget maintaining systems rather than replacing them. Neither of those has been affected by anything that happened in the last nineteen months, and both are still published by an audit office that has been publishing them regardless of who governs.
Fourth, the cost. The next person who stands up and says there is enormous waste in the federal government — and there is — will be answered with six thousand contracts that were never cancelled. That answer will work for a decade and it will not be unfair.
So my heading is: retire the slogan, publish the three numbers, and name the auditor.
And the honest close for eighteen months of this file. We asked for a table and we were right to ask. We never specified who should keep it, and that omission is the whole distance between what we wanted and what we got.
A metric kept by the party being measured is not a metric. It is a press release with a decimal point, and we should have said so in question one.
Las Vegas, NV
Take the phrase off the marquee. Post the hold percentage, the unfinished remodel, and the game the house will not close. That was my answer in draft and eleven days ago the count came in.
Nearly half the terminations on the board never happened. A quarter of the claimed savings match the method the board itself said it was using. Ninety-six percent of the grant claims cannot be checked. And when the auditors asked, nobody answered.
In my business, one bad number from the house is worth more to a disgruntled customer than fifty good ones. I said that in March of last year about an eight-billion-dollar entry that turned out to be eight million.
This is not one bad number. This is the book.
So here is where I land after eighteen months.
The leak is real. It was real before any of this and it is real now — hundreds of billions a year, published every year by the audit office, uncontested by anybody serious. Nothing in the last eleven days touches that.
The count was not a count. It was the house scoring itself and refusing to open the drawer when the auditors arrived.
And the price is the thing nobody wants to say. The next honest person who raises federal waste has been handed a permanent rebuttal, and it will be used, and it will work.
So retire the slogan. Post the three numbers. And post them from somebody who does not work for the house.
Would that end the coalition? The audit ended it. Everybody who needed the number to be big has now been told, by the government’s own auditors, that it was not. The people who wanted the leak fixed are still here and now have nothing to point at.
That is the cost of a self-scored book, and it was avoidable from the first week.
Knoxville, TN
Eighteen questions, and the design finding is one this file identified in March of last year and stated incompletely.
We said: publish improper payments, unexecuted IT, and programs Congress will not repeal. That was right.
What we did not say, and should have, is the property those three share and that the published number lacked.
All three are measured by somebody who does not benefit from the reading.
Improper payments are estimated by agencies under a statutory method and audited by an independent office. Unexecuted IT would be measured against contract acceptance records that exist regardless of who reports them. Programs failing the screen would be identified against statutory citations anybody can look up.
The savings page had none of that. It was produced by the office, about the office, with a method the office described and — per the audit eleven days ago — did not follow in about sixty percent of cases. Nearly half the terminations it reported were not in the contracting record at all.
In public works terms this is a self-certification and we have known for a century what self-certification produces. It is why a contractor does not sign off his own inspections, why a bridge is tested by somebody other than the firm that poured it, and why nobody finds those arrangements insulting. They exist because the person with the incentive cannot be the person with the instrument.
We spent eighteen questions demanding an instrument and never once specified the operator. That is our error and it is worth writing down as ours.
So the answer to the question:
Retire the phrase. It has been carrying a claim that has now failed an audit and it will carry that failure from here on.
Publish the three numbers, unchanged from what this file specified in March of last year.
And name the auditor in the same sentence as the metric, because a number without an independent producer is not a measurement — it is a party’s account of itself, and the last nineteen months are what that produces at scale.
Would that honesty end the coalition? The audit ended one. What survives it is whoever is still interested in a leak that is real, still published annually, still measured in hundreds of billions, and now considerably harder to talk about than it was in 2024.
That is a smaller group and it is the only one that was ever going to fix anything.
