Bureaucracy Times

All the news fit to be declassified.

Category: Federal Agencies & Oversight

  • The Fed Finally Hikes, and Conveniently Skips the Month Before Voters Do

    The Reversal Nobody at the Podium Wanted to Name

    On September 16, 2026, the Federal Open Market Committee did something it had not done since 2023. It raised interest rates. The vote was 12 to 0, a quarter point, lifting the target range to 3.75 to 4 percent.

    You would not know from the statement itself that this was a reversal of years of cutting. The Committee’s prose, as always, arrived sanded down to the texture of a hospital waiting room…

    “Economic activity is expanding at a solid pace.”

    “Productivity growth is strong.”

    “Job gains have kept pace with the workforce.”

    Four sentences later, almost as an afterthought: “Inflation remains elevated.” Today’s action, the statement assured us, “will support a timelier return to the Committee’s 2 percent goal.”

    Timelier than what, exactly, the statement does not say. Timelier than the last several years of the Committee insisting inflation was transitory, then persistent, then manageable, then, apparently, in need of an actual hike to fix.

    If you have whiplash, you are reading the sentence correctly. The institution that spent 2024 and 2025 cutting rates on the theory that inflation was heading obediently toward target has just told you, in the driest possible language, that the theory did not hold.

    “Geopolitical Developments” Is Doing an Enormous Amount of Work

    The statement attributes elevated uncertainty, in part, to “geopolitical developments.” This is a phrase built for a press release, not a diagnosis.

    Chair Kevin Warsh, at the podium, came slightly closer to naming the mechanism, telling reporters that the Fed “cannot affect any individual price” on items like oil or groceries but would work to ensure that any change in relative prices “don’t broaden out, don’t have second and third order effects on the economy.”

    That is a serviceable description of what a central bank actually does. It is also, notably, not a description of any specific tool the Fed is using to accomplish it. Raising the federal funds rate a quarter point does not touch the price of oil, does not touch the price of eggs, and does nothing to whichever geopolitical development is allegedly driving both.

    What it does is tighten financial conditions broadly and hope the broadening effect Warsh described never shows up. That is a bet, not a mechanism, and the statement’s studious refusal to name the underlying driver means the public is being asked to trust the bet without being told its odds.

    A Skip Calendared Suspiciously Close to an Election

    Here is where the institutional story gets more interesting than the monetary one. Most FOMC participants pencil in a total of two hikes this year in their updated projections, and the Committee is expected to skip its October meeting, a decision one asset manager tied explicitly, in reporting following the announcement, to the meeting’s proximity to the midterm elections.

    The Federal Reserve has spent decades building a reputation, however imperfect, on the idea that its calendar answers to data, not to ballots. A skipped meeting is not itself a scandal. But when the skip is reported in the same breath as the midterms, and nobody at the Fed rushes to correct the record, the silence tells you something.

    Either the independence is real and the timing is coincidental, in which case the Fed has an obligation to say so plainly; or the timing is not coincidental, in which case the public deserves to know that too.

    What the Institution Owes the Public It Did Not Deliver

    None of this means the rate decision was wrong on the merits. A quarter point hike after a multiyear cutting cycle, with inflation still running hot, is a defensible call, and 12 to 0 votes do not happen by accident. But defensible policy and adequate disclosure are two different things, and the September statement gave the public one while withholding the other.

    The Fed will tell you, correctly, that a terse statement is a feature, not a bug, designed to avoid spooking markets with unnecessary color. Fair enough. But if the data justified a reversal this stark, the institution owes the public more than four sentences of boilerplate.

    And if the October skip really is unrelated to the midterms, the burden is on the Fed to say so, not on financial reporters to speculate in its absence.

    The Committee reconvenes, by the current schedule, well after voters have had their say. Whether that is coincidence or caution dressed up as one, the Fed has not told us, and until it does, the question is not going away.

    Primary sources: Federal Reserve, September 16, 2026 FOMC statement.

    Fediverse Reactions

  • Washington’s Poverty Ledger: Millions Lifted, Billions Paid Improperly

    The federal government is remarkably good at keeping millions of people above the poverty line, and considerably less tidy when asked to account for every dollar it spends doing it.

    Those two facts are not contradictory.

    New data released by the U.S. Census Bureau show that Social Security alone kept an estimated 28.8 million people from falling below the Supplemental Poverty Measure in 2025. Refundable tax credits kept another 6.1 million above the line. The Supplemental Nutrition Assistance Program, better known as SNAP, accounted for 3.1 million. Supplemental Security Income accounted for 2.2 million, while housing subsidies accounted for another 2.1 million.

    That is the good side of the ledger.

    On the other side sits a federal payment system that reported an estimated $186 billion in improper payments during fiscal year 2025, including tens of billions associated with some of the same enormous programs intended to improve Americans’ economic security.

    It would be tempting to choose one number and tell a simple story.

    The numbers refuse to cooperate.

    Two Ways to Measure Poverty

    The Census Bureau reported Monday that the official U.S. poverty rate fell from 10.7% in 2024 to 10.2% in 2025, leaving about 34.5 million people below the official poverty threshold.

    But Census maintains another measurement called the Supplemental Poverty Measure, or SPM.

    Unlike the traditional measure, the SPM attempts to account for considerably more of the economic machinery surrounding a household. Government benefits such as SNAP, housing subsidies and refundable tax credits count as resources. Taxes, work expenses and medical expenses reduce those resources.

    Under that broader calculation, the poverty rate was 13.1% in 2025 and did not change significantly from 2024.

    The distinction allows Census to perform an unusually revealing exercise: remove individual benefits or expenses from the calculation and see how many people cross the poverty line.

    It amounts to something resembling an antipoverty ledger.

    And Social Security owns the largest entry.

    Without Social Security benefits, Census estimates another 28.8 million people would have been in SPM poverty in 2025.

    Refundable tax credits reduced the number by 6.1 million. Within that category, the Earned Income Tax Credit accounted for about 4 million. SNAP reduced it by 3.1 million, SSI by 2.2 million and housing subsidies by 2.1 million.

    Other noncash benefits—including energy assistance, WIC and school lunch—collectively kept about 1.5 million people above the threshold. Other cash benefits accounted for another 1 million.

    These estimates overlap and cannot simply be added together. A household can receive more than one benefit, and Census specifically warns that the individual components are not mutually exclusive.

    Still, the direction is unmistakable.

    Federal assistance changes measured poverty substantially.

    Then the Bills Arrive

    Census’s ledger also runs in reverse.

    Medical out-of-pocket expenses pushed approximately 7.7 million people into SPM poverty in 2025. Payroll taxes, or FICA, added roughly 4.5 million. Work expenses added about 4 million, and federal income taxes pushed another 1.4 million below the threshold.

    The medical number is particularly striking because it increased from approximately 7.5 million in 2024.

    At the same time, several government benefits became somewhat less consequential in the poverty calculation.

    Refundable tax credits kept 6.8 million people out of poverty in 2024, compared with 6.1 million in 2025. SNAP’s effect declined from 3.6 million to 3.1 million. SSI declined from 2.5 million to 2.2 million.

    Social Security, meanwhile, barely moved: 28.7 million in 2024 and 28.8 million in 2025. Housing subsidies remained at approximately 2.1 million.

    So while the topline SPM poverty rate barely changed, some of the machinery underneath it did.

    That is one advantage of looking below the headline number.

    Another is discovering what happens to all the money.

    $186 Billion in Improper Payments

    Federal agencies reported approximately $186 billion in improper payments during fiscal year 2025, according to the Government Accountability Office.

    About $153 billion—roughly 82%—were overpayments.

    The total was $24 billion higher than the previous fiscal year. GAO says the $186 billion does not represent the entire universe of improper federal payments because some programs susceptible to significant errors were not included in the total.

    Several programs connected to the poverty ledger appear prominently in those numbers.

    Medicaid reported an estimated $37.39 billion in improper payments, representing a 6.12% error rate.

    The Earned Income Tax Credit reported approximately $21.1 billion in improper payments and an estimated error rate of 32.7%.

    SNAP’s fiscal 2025 payment-error rate was 10.62%, representing approximately $10.1 billion in improper payments.

    SSI’s latest available payment-integrity estimate, covering fiscal 2024, found approximately $7.33 billion in improper payments from $63.29 billion in outlays.

    Those are enormous numbers.

    They are not, however, enormous fraud numbers.

    Improper Does Not Mean Stolen

    An improper payment can involve fraud. It can also involve an innocent eligibility mistake, the wrong payment amount, missing paperwork, an underpayment or even a payment made to the correct person for the correct amount through a process that failed to satisfy federal requirements.

    Medicaid provides an almost absurdly useful example.

    Of the program’s $37.39 billion in estimated fiscal 2025 improper payments, CMS says 77.17% resulted from insufficient documentation, which the agency says is generally not indicative of fraud or abuse.

    That does not make the problem imaginary.

    It means the government could not demonstrate that those payments satisfied all applicable requirements.

    The distinction becomes even stranger in the Affordable Care Act premium-tax-credit program. CMS reported that approximately 43% of that program’s improper payments involved situations in which the government paid the right recipient the right amount—but violated a statutory or regulatory requirement while doing it.

    Financially correct.

    Procedurally improper.

    There may be no phrase more suited to the federal government.

    Actual fraud nevertheless exists, and treating every payment error as innocent would be just as misleading as calling every error theft.

    The Holes in the Fence

    SNAP provides a good illustration of the difference between a known improper payment and a potential control weakness.

    USDA reported $10.1 billion in SNAP payment errors for fiscal 2025. The agency says its payment-error rate measures whether states correctly determine household eligibility and benefit amounts and includes both overpayments and underpayments.

    A separate USDA program-integrity review found something different.

    Reviewing eligibility information from 29 state agencies, USDA identified records involving dummy or missing Social Security numbers, duplicate participation within states, interstate duplication and deceased individuals appearing as active participants.

    The agency estimated that dummy Social Security numbers and intrastate duplication alone represented more than $1.5 billion in annual financial exposure.

    “Exposure” is not the same as confirmed loss, and neither is synonymous with fraud.

    But it identifies holes in the fence.

    The distinction matters because a prosecution can tell taxpayers how much one fraudster stole. A control weakness raises the more consequential question of how many fraudsters the government might not have identified.

    And What About Noncitizens?

    Immigration status creates another place where the numbers require more care than the political shorthand normally permits.

    Some noncitizens are legally eligible for particular federal benefits. Others are not. Mixed-status households may contain an ineligible adult and eligible U.S.-citizen children.

    A benefit received by such a household therefore cannot automatically be described as money improperly paid to a noncitizen.

    Social Security Administration records provide an unusually clear example.

    In December 2025, 7,392,201 people received federally administered SSI payments.

    Of them, 7,082,134 were citizens and 310,067 were noncitizens—about 4.2% of recipients.

    Noncitizen recipients received a lower average monthly payment: $596.77 compared with $719.70 for citizens. Nearly 252,000 of the noncitizen recipients were 65 or older.

    Those figures establish that noncitizens receive a measurable share of SSI benefits.

    They do not establish that those payments were illegal.

    That requires a separate eligibility determination.

    And fraud cases involving noncitizens using false identities or otherwise defeating eligibility controls should likewise be categorized as fraud—not used to characterize every lawful noncitizen beneficiary.

    The distinction may be less satisfying than a sweeping conclusion, but accounting generally is.

    The Government Knows Some Things Better Than Others

    The federal government can estimate with remarkable specificity how individual programs affect poverty.

    Census can tell us that Social Security moved 28.8 million people across its Supplemental Poverty Measure threshold.

    It can estimate that SNAP moved 3.1 million.

    It can tell us that medical expenses pushed 7.7 million the other direction.

    Elsewhere in Washington, however, the accounting becomes considerably fuzzier.

    GAO says federal agencies identified $186 billion in improper payments during fiscal 2025—and also says that total does not capture the full government-wide universe.

    Treasury estimates that nearly one-third of EITC payments were improper. USDA reports a SNAP error rate above 10%. CMS reports tens of billions in Medicaid errors while explaining that most resulted from inadequate documentation rather than apparent fraud.

    And the Social Security Administration can tell us exactly how many SSI recipients are noncitizens, while that fact alone tells us almost nothing about whether their benefits were properly or improperly paid.

    The deeper one digs into the ledger, the harder it becomes to reduce it to “government programs work” or “government programs waste money.”

    Both propositions are too simple.

    What Taxpayers Are Buying

    Census’s new statistics establish something important.

    The American safety net has a measurable effect.

    Remove Social Security, SNAP, refundable tax credits, SSI, housing subsidies and other assistance from household resources and millions more people fall below the Supplemental Poverty Measure.

    That is not an ideological claim. It is what the government’s own statistical model measures.

    The government’s other records establish something equally important.

    Programs distributing hundreds of billions of dollars make mistakes. Agencies send money in incorrect amounts, fail to maintain required documentation, pay people who are not eligible, underpay people who are eligible and sometimes get defrauded.

    The scale is large enough that even small error rates produce numbers measured in billions.

    In fiscal 2025, the government estimated $186 billion in improper payments across 64 programs. GAO says federal improper-payment estimates since fiscal 2003 total roughly $3 trillion.

    Those numbers do not erase the 28.8 million people Social Security kept above the SPM threshold.

    And those 28.8 million people do not erase the payment errors.

    The more useful accountability question is what lies between them:

    How much poverty is the federal government preventing, how much does it spend doing so, how much money leaves the system improperly, how much of that represents actual fraud—and how confidently can Washington tell taxpayers the difference?

    Census has provided a surprisingly good answer to the first question.

    The federal government’s own auditors suggest there is considerably more work to do on the rest.

    Fediverse Reactions

  • Nashville’s Postal Honor System Just Failed an Audit

    The Postal Service Office of Inspector General does not typically announce itself. It shows up before the carriers do, counts what it finds, and compares that count to what management already told headquarters. In Nashville, over three days in July, that comparison broke down almost everywhere it was tried.

    Four reports released September 17 (covering the Church Street Station, the Hendersonville Main Post Office, the Lebanon Main Post Office, and the Nashville Processing and Distribution Center that feeds all three) describe an operation running on self-certification, in a district where almost nobody was checking whether the certifications matched the floor.

    The Paperwork and the Reality

    Start with the number every field manager is supposed to report daily, undelivered mail sitting in the building after carriers leave for the street. The Postal Service has a system for this, the Delivery Condition Visualization tool, and a policy requiring managers to log it accurately before day’s end.

    At Church Street, auditors found 5,158 delayed pieces stacked at carrier cases, on the workroom floor, and in the throwback case on a single July morning. Management had reported none of it.

    At Hendersonville, auditors counted 7,276 delayed pieces; management’s own DCV entry captured 48 percent of that total.

    At Lebanon, the smaller rural unit, the shortfall was proportionally similar, 164 pieces found against 40 reported, a 24 percent capture rate isolated to one route overwhelmed by a botched reassignment.

    Three units, three different failure stories (an untrained acting supervisor at Church Street, an absentee postmaster at Hendersonville, a carrier case too small for its own route at Lebanon), and one identical result: the number that travels up to district, area, and headquarters bears little relationship to the number sitting in the building.

    The OIG’s own language is almost bureaucratic understatement here; inaccurate DCV reporting, it notes, “can result in improper actions taken to address issues.” That’s one way to put it. Another is that the agency has been managing its mail backlog using data it privately admits, in three separate audits, it cannot trust.

    Nobody’s Holding the Keys

    Arrow keys open street collection boxes and apartment mail panels, and the Postal Service treats them, on paper, as tightly controlled accountable property. In practice, Nashville treated them like spare change.

    Church Street had one key missing and nineteen more in active use that had never made it into the national RADAR inventory system. Hendersonville had two keys missing and one more found on-site despite being logged as lost. Lebanon had one missing and thirteen unlogged keys sitting in a vault behind a lockbox whose own key, per the report, was kept somewhere that had “not always” secured it.

    The detail that unifies all three units is not the sloppiness, which is almost expected at this point, but what happened after. Postal Service policy requires any missing or lost arrow key to be reported immediately to the U.S. Postal Inspection Service. None of the three stations did it.

    The Lebanon postmaster told auditors she knew the reporting requirement and skipped it anyway, on the theory that the keys were probably just in transit between facilities. That is not an oversight; it’s a judgment call to not escalate a security gap that, per the OIG’s own boilerplate repeated in all three reports, “can result in undelivered mail” and “increased risk of mail theft.”

    Three units, three postmasters, and not one referral to the agency’s own law enforcement arm.

    The Scan That Isn’t a Scan

    Every tracked package’s status, the “Delivered,” the “No Access,” the “Held at Post Office,” exists because a carrier or clerk scanned it. Customers, insurers, and USPS’s own performance metrics all treat that scan as ground truth. Nashville treated it as a formality.

    At Church Street, 21 of 30 sampled packages, seventy percent, had scans that didn’t match what actually happened to them, including thirteen marked “Delivered” that plainly weren’t.

    At Hendersonville the rate was worse, 23 of 30, with some “No Access” scans logged more than four miles from the delivery address.

    At Lebanon, all six sampled packages had integrity problems, with hold notices posted on parcels that had no accompanying hold authorization form.

    Scale that same failure up to the Nashville P&DC, the plant that processes and dispatches for all three stations, and it stops looking like a delivery-unit habit and starts looking like a systemic one. Load scan compliance there sat below both the FY25 and FY26 goals, with a documented downward trend management admitted it wasn’t monitoring.

    And when auditors went looking for the handheld scanners themselves, twenty of seventy Surface Visibility units and one of twenty-three PVS units hadn’t been used since mid-April and simply couldn’t be located, a shortfall the OIG valued at $12,075 in questioned costs. An inspector watched a driver try to walk out of the transportation office with a scanner without signing for it.

    Nobody stopped him until the OIG did.

    Why Nobody Caught It

    Every one of these findings traces back to the same structural cause, stated by the people responsible for catching it. The Hendersonville postmaster was on extended leave for two weeks before and during the audit, leaving three recently promoted, undertrained supervisors to run the unit’s closeout, route adjustments, and package handling with no meaningful oversight; staff described her, when she was present, as unapproachable.

    At Church Street, an acting PM supervisor closed out the unit alone because the redline process, the checklist meant to catch exactly this kind of gap, was never taught to him. At the P&DC, logistics management runs with a single supervisor per shift, splitting attention between dock activity, driver availability, and scanner distribution until, inevitably, something goes unwatched.

    None of this reads as concealment so much as capacity collapse. That is, the checks exist on paper, the people meant to run them are stretched across too many duties to run them consistently, and the gap only closes when an outside auditor counts what’s actually there.

    USPS agreed with every finding across all four reports, which is the expected and mostly empty gesture at this stage. The recommendations tied to the P&DC report carry implementation dates running out to May 2027, nearly two years after the conditions that triggered them. The three delivery unit reports, meanwhile, are interim.

    A capping report covering all three stations and the district as a whole is still pending, and based on how long the equivalent report took after a similar 2023 audit in Memphis, Nashville shouldn’t expect it before winter.

    Until then, the operative fact is simple. An agency that runs on self-reported compliance data, unaudited inventory logs, and scans nobody double-checks is only as accurate as its most honest, best-trained, least overworked field supervisor.

    In Nashville this summer, that bar turned out to be considerably lower than the number the Postal Service had been reporting to itself.

    For more USPS OIG coverage, see The Pantry That Held a Federal Problem — on a joint OIG-DEA investigation that dismantled a drug distribution network running through the postal system — and The Postal Service Passed Its Pay Audit. The Bill Just Got Postponed. — on how the same OIG found full compliance with executive salary caps while $1.1 million in deferred compensation quietly accrues interest on a $9 billion annual loss.

    Sources: Church Street Station, Nashville, TN: Delivery Operations (26-110-1-R26) | Hendersonville Main Post Office, Hendersonville, TN: Delivery Operations (26-110-2-R26) | Lebanon Main Post Office, Lebanon, TN: Delivery Operations (26-110-3-R26) | Efficiency of Operations at the Nashville Processing and Distribution Center, Nashville, TN (26-109-R26)

  • The Actuary’s Gift: How a $316 Billion SSI Expansion Got Pre-Scored While the Inspector General Was Still Counting the Last Round of Overpayments

    The Social Security Administration’s Office of the Chief Actuary keeps a very particular kind of stationery: dry, unfailingly polite, and capable of detonating a fiscal grenade. On September 15, 2026, Chief Actuary Karen P. Glenn sent one of these letters to Speaker Emerita Nancy Pelosi, scoring a Supplemental Security Income overhaul that would cost the federal government an additional $316 billion over the next decade.

    It arrived, as these things do, with a straight face and a clean conscience — and an awkward footnote: the same week the letter went out, the SSA’s own inspector general was still trying to explain why the agency cannot stop overpaying people whose bank accounts it never bothered to check twice.

    An Actuary’s Gift, Itemized

    The letter to Pelosi scores a seven-provision proposal drafted by Dr. Wendell Primus of the Brookings Institution, and it is, structurally, a marvel of quiet generosity.

    Provision A raises the SSI federal benefit rate by $70 for individuals and $105 for couples.

    Provision B triples the general income exclusion.

    Provision C triples the earned income exclusion.

    Provision D excludes 40 percent of Social Security benefits from the SSI income calculation entirely.

    Provision E quintuples the countable resource limit, from $2,000 to $10,000 for individuals, while sheltering up to $50,000 in retirement savings from counting against eligibility at all.

    Provision F eliminates the consideration of in kind support (food, shelter, a couch to sleep on) as income.

    Provision G has the IRS hand SSA a list of people who might qualify but haven’t applied.

    Total combined cost through 2034: $316 billion, layered on top of a baseline SSI program the actuaries already project will cost $722 billion over that same window. The single largest line item is not the benefit increase everyone will read about in the press release; it is Provision G, the outreach provision, at $135 billion, more than the benefit hike and the resource limit changes combined.

    Glenn is admirably honest about the uncertainty here, noting flatly that “the actual effect that this provision would have on participation is uncertain.” One might call that the actuarial equivalent of a shrug, delivered at a cost of $135 billion.

    A Proposal With a Long, Familiar Résumé

    None of this is new, and that is precisely the point worth dwelling on. Provision D, the 40 percent Social Security exclusion, is lifted nearly verbatim from a 2024 Brookings paper by Primus and coauthors. Primus, notably, already has a different SSA-scored proposal circulating through Congress; a broader solvency package he sent to then Minority Whip Steny Hoyer produced its own actuarial letter in January 2025, one dealing with the OASDI trust funds rather than SSI alone.

    Provision E, the resource limit increase, is essentially the text of the SSI Savings Penalty Elimination Act, and the broader package overlaps substantially with the SSI Restoration Act, reintroduced this past March by Rep. Adelita Grijalva, Sen. Elizabeth Warren, Rep. Jan Schakowsky, and Rep. James Moylan, with 30 cosponsors.

    What is missing from the Pelosi letter is a bill number. This is not legislation being scored, rather an idea being pre-scored, run through the actuary’s office by a private economist working through a member’s office, presumably so that whoever eventually introduces the combined package can wave around a $316 billion figure with an official SSA letterhead attached to it.

    It is a perfectly normal way to legislate in Washington, and it is also, if you squint, a way of credentialing an idea through an agency whose job is arithmetic, not politics, so that the arithmetic can later be cited as though it settled the politics.

    The Resource Limit and the Inconvenient Auditor

    Here is where the story gets uncomfortable for everyone involved. The SSA’s own Office of the Inspector General published its Fiscal Year 2024 compliance audit under the Payment Integrity Information Act in May 2025, and it found that SSA failed two of the ten reporting requirements, both in the SSI program specifically.

    The SSI improper payment rate climbed from 9.41 percent (roughly $5.3 billion) in 2019 to 10.62 percent (roughly $6.5 billion) in 2023, moving in exactly the wrong direction despite years of corrective action plans that read, in retrospect, like a to do list nobody finished.

    The leading cause, according to the OIG, is unreported resources sitting in financial accounts above the current limit, and the agency’s own data shows that 89 percent of those overpayments occurred because a recipient’s circumstances changed after their initial approval, in the gap between reviews where nobody was looking.

    The OIG estimates $2 billion in overpayments could have been prevented in 2023 alone had SSA simply run its existing verification tool more often. It has not implemented that recommendation. A separate OIG audit released the same month documents the mechanism in detail: 63 unresolved SSI earnings alerts sat for an average of 733 days, long enough for the agency’s own finality rules to erase $664 million in overpayments from collectability entirely.

    Into this specific, well documented gap, Provision E proposes to quintuple the resource limit and exempt up to $75,000 in retirement accounts for couples from ever being counted.

    But it is worth noticing that Congress is being asked to widen precisely the category of asset that its own watchdog just identified as the single largest driver of a persistent, unresolved overpayment problem, at the same moment SSA is finally rolling out zero dollar tolerance verification for the old, much smaller limits.

    A Program on Autopilot, Redirected

    The 2025 SSI Annual Report, which supplies the baseline assumptions for Glenn’s letter, tells a story of a program quietly shrinking relative to everything around it. SSI expenditures were 0.22 percent of GDP in 2024 and are projected to fall to 0.17 percent by 2049, a decline attributed to a shrinking pool of eligible citizens and benefits growing slower than average incomes. Recipient numbers actually dropped slightly between January 2024 and January 2025, from 7.30 million to 7.26 million.

    Against that backdrop, the Primus proposal does not merely add to the program; it reverses its trajectory. The letter’s own table shows $104 billion in new costs against a $339 billion baseline for 2025 through 2029, a roughly 30 percent acceleration over five years, applied to a program the actuaries themselves describe as coasting toward relative irrelevance.

    That may well be defensible policy. Poverty among elderly and disabled Americans is a real and measurable thing, and $688 a month, the current average SSI payment, buys considerably less than it did when the resource limits were last touched in 1989. But it is not the same claim as “modest technical fix,” which is how proposals like this tend to be marketed once the actuarial letter gets attached to a press release.

    Somewhere between the Brookings paper, the Grijalva bill, and Karen Glenn’s very polite letter, the size of the ask got lost in translation, and it is the sort of thing a $316 billion program deserves to have said about it plainly, in public, before anyone votes on it.

    SSA Chief Actuary Letter to Speaker Emerita Pelosi (September 15, 2026)Download

  • GSA Has a $17 Billion Building Maintenance Problem. Now It’s Inspecting Buildings Less Often.

    The federal government has a building maintenance problem measured in billions of dollars, fewer people managing those buildings than it did two years ago, and a new policy allowing as much as five years to pass between the physical assessments it uses to estimate what many of those buildings need.

    That combination sits at the center of a new audit by GSA’s Office of Inspector General, which found that the agency’s Public Buildings Service was already understating repair and maintenance liabilities in some of its building assessments. This is Part 2 of our coverage; Part 1 covers how hundreds of those assessments were completed and approved by the same employee, defeating the agency’s own quality-control requirement.

    The September audit examined 196 buildings and found that 79 surveys underreported repair and alteration costs by a combined $73.2 million.

    The OIG did not extrapolate that figure across GSA’s entire portfolio, and neither should anyone else. The audit does not establish that every federal building assessment is understated by the same percentage.

    What it does establish is more consequential for the future: the system GSA relies upon to understand the condition and repair costs of its buildings was capable of missing known liabilities before the agency substantially lengthened the time between required physical assessments.

    And the government already has quite a maintenance bill.

    A $17 Billion Backlog

    GSA’s Public Buildings Service serves as the federal government’s civilian landlord, providing workspace for hundreds of thousands of federal employees across a portfolio that requires continuous attention — roofs, elevators, boilers, chillers, electrical systems, fire protection equipment. The backlog has been growing for years.

    The Government Accountability Office reported in 2025 that deferred maintenance and repair across Defense Department and civilian federal properties had more than doubled, rising from approximately $171 billion in fiscal year 2017 to $370 billion in fiscal year 2024.

    GSA’s portion exceeded $17 billion by March 2025, and GAO warned that allowing the trend to continue could cause federal assets to deteriorate further and require premature replacement — potentially costing substantially more than performing maintenance when originally scheduled.

    The problem was serious enough that GAO added the condition of federal buildings as a new concern within federal real property’s longstanding designation on its High-Risk List.

    GSA was therefore already trying to manage an aging portfolio carrying billions of dollars in deferred work when its workforce shrank dramatically.

    PBS Lost 45 Percent of Its Workforce

    Between September 2024 and November 2025, the Public Buildings Service went from 5,683 employees to 3,126 — a reduction of approximately 45 percent.

    GAO later found that PBS had not conducted strategic workforce planning before making those reductions.

    The consequences were not purely theoretical: officials from federal agencies served by PBS told GAO about concerns regarding the organization’s capacity, including project delays associated with having fewer employees available to perform essential real-estate work.

    PBS officials subsequently began trying to determine where workforce and skills gaps existed, even as the schedule for assessing federal buildings changed.

    Two Years Became Five

    GSA uses its Building Assessment Tool, or BAT, to estimate the cost of correcting major deficiencies in buildings for which PBS has maintenance responsibilities.

    A BAT survey is not simply an accounting exercise.

    It is a 40-section assessment based on a physical inspection of an individual building. The information produced through that process feeds GSA’s financial reporting, long-term capital planning and decisions about which repairs receive priority.

    Previously, PBS required those surveys every two years; beginning in fiscal year 2025, GSA moved to a five-year assessment cycle.

    The agency’s FY2025 financial reporting attributed the change to resource constraints and alignment with the minimum assessment frequency established by the Federal Real Property Council.

    GSA is not violating the federal minimum by inspecting buildings every five years. It has moved from its previous, more frequent schedule to the minimum interval permitted under the applicable government-wide standard.

    Nor does a building’s maintenance estimate simply sit untouched until someone returns with a clipboard five years later.

    GSA updates BAT estimates annually to account for changes in labor and material costs and other circumstances. Better project-specific information can also be incorporated when it becomes available.

    Those procedures can help keep the estimated price of known repairs current.

    But updating the estimated cost of a problem isn’t the same thing as discovering a new problem.

    A spreadsheet can tell GSA that the roof repair it identified three years ago is more expensive today.

    The spreadsheet cannot look at the roof.

    That distinction is precisely why the timing of the new OIG audit matters.

    The Inspector General Found Missing Repairs

    The GSA OIG concluded that PBS was not consistently following its own procedures for ensuring that BAT estimates were complete and accurate.

    Auditors found 79 surveys among their 196-building sample that underreported repair and alteration liabilities by approximately $73.2 million.

    The problem was not always that GSA lacked the information; in some cases, relevant records existed elsewhere within the agency, but surveyors and approvers did not consistently review them when preparing assessments.

    That means the agency could possess records showing that work was needed while the building assessment used to calculate maintenance liabilities failed to incorporate it.

    The OIG concluded that the deficiencies limited assurance that PBS was consistently using and reporting complete, accurate and reliable information about the repair and maintenance needs of its portfolio.

    Because PBS underwent significant staffing reductions and a major reorganization in 2025, the OIG noted, BAT surveys are now required every five years instead of every two, explicitly connecting that problem to the longer inspection schedule.

    The longer interval, auditors warned, increases the risk that GSA will be working with outdated maintenance-liability information.

    In other words, the five-year cycle isn’t the reason auditors found $73 million missing from the sampled surveys.

    But discovering that the system can miss liabilities makes the additional three years between required physical assessments considerably more important.

    Avoiding a Future Television Career

    There is, admittedly, an entertaining way to contemplate the eventual destination of buildings whose maintenance problems are allowed to accumulate.

    Fans of Mysteries of the Abandoned know the formula.

    Somebody builds an impressive structure. Somebody stops taking care of it. Nature begins submitting its own change orders. Eventually, a television crew arrives to explain why there are trees growing through what used to be the conference room.

    A federal building does not become Discovery Channel material merely because GSA inspects it every five years.

    But the joke works because deterioration isn’t particularly interested in the government’s inspection calendar.

    Roofs leak. Pipes corrode. Boilers age. Concrete cracks. Electrical equipment becomes obsolete. Water intrusion produces new problems while existing problems become more expensive.

    And the federal government is already carrying a maintenance backlog large enough that preventing deterioration ought to be considerably cheaper than documenting the ruins afterward.

    The Minimum Standard Meets a Maximum-Sized Problem

    There is nothing inherently improper about a five-year assessment cycle — it meets the federal minimum. That may actually be the more important point.

    GSA did not move to a five-year cycle while managing a pristine portfolio with abundant staff and a maintenance system demonstrating exceptional accuracy.

    It made the change amid resource constraints, a major reorganization, a 45-percent reduction in PBS staffing and a deferred-maintenance backlog exceeding $17 billion.

    The federal government’s broader maintenance backlog, meanwhile, has more than doubled since 2017.

    Now the GSA inspector general has demonstrated that the information system helping the agency understand those liabilities can omit repair costs already documented elsewhere inside the government.

    The OIG has recommended tighter controls over how BAT surveys incorporate maintenance information and stronger quality-control procedures. PBS agreed with the recommendations.

    Those changes may improve the accuracy of the information GSA collects, but they do not change the arithmetic of time. Under the old schedule, a required physical assessment returned every two years. Under the new one, it may not return for five.

    For a government already trying to decide which of billions of dollars in building repairs it can afford to make, three additional years is a long time for something to start leaking.

  • GSA Required a Second Set of Eyes on Building Inspections. Sometimes They Belonged to the Same Person.

    The federal government’s civilian landlord, the General Services Administration, has a quality-control system for assessing the condition of federal buildings: one employee conducts the survey, and another approves it, at least in principle.

    In practice, a new audit by the General Services Administration’s Office of Inspector General found that hundreds of building assessments were both completed and approved by the same employee — effectively allowing the person conducting the assessment to serve as their own second set of eyes.

    GSA’s Public Buildings Service uses its Building Assessment Tool, or BAT, to document building conditions and estimate repair and maintenance liabilities across the federal real-estate portfolio. The OIG examined 2,949 BAT surveys completed between fiscal years 2022 and 2025 and found that 524 — nearly 18 percent — were both completed and approved by the same employee. Those assessments were not financially trivial: together, they represented approximately $5.27 billion in reported repair and maintenance liabilities.

    Independent Review, With an Asterisk

    Under the BAT process, GSA assigns separate roles to a surveyor and an approver. The surveyor assesses building systems and identifies deficiencies, while the approver is supposed to review that work for completeness and accuracy before the assessment becomes final. On paper, that division provides an obvious internal control: somebody checks the work. In the system as implemented, however, that somebody did not always have to be somebody else.

    According to the OIG, PBS guidance did not explicitly prohibit an employee from serving as both surveyor and approver. A GSA frequently asked questions document even addressed the possibility directly.

    “Can I be a Surveyor as well as an Approver for a BAT survey?”

    The guidance acknowledged that employees could perform both roles, although doing so was “not recommended.” The OIG concluded that allowing employees to approve their own surveys undermined the independence of the review and increased the risk that errors or omissions would go undetected.

    The Fix Had a Loophole

    PBS eventually attempted to close the loophole. On June 30, 2025, GSA implemented a system control intended to prevent the same employee from being assigned as both surveyor and approver on a BAT assessment. The control, however, came with an exception: administrative and management users could override the restriction.

    The OIG subsequently identified 44 surveys in which the same employee completed and approved the assessment even after the new control had been implemented. Auditors also found broader inconsistencies in how survey responsibilities were recorded. Some surveys had no approver assigned; others lacked both an assigned surveyor and approver; and in 106 instances, according to the OIG, the employee who actually completed and approved the survey was not listed as either the assigned surveyor or approver.

    The OIG recommended that PBS establish policies explicitly prohibiting employees from approving their own BAT surveys and strengthen system controls to enforce the separation. GSA agreed with the recommendations.

    A Much Larger Ledger

    The procedural flaw matters because BAT is not merely an inspection checklist. The assessments feed into GSA’s accounting for billions of dollars in federal building repair and maintenance needs, so errors in those surveys can affect the government’s picture of the condition — and cost — of its own real estate.

    The same audit concluded that GSA understated repair and maintenance liabilities in dozens of building assessments and identified broader weaknesses in how the agency gathers information about the condition of federal properties. Those findings raise a larger question: how accurately does the federal government’s landlord know what its buildings actually need, and what that will cost?

    That question, and the understated liabilities behind it, are explored in Part 2: GSA Has a $17 Billion Building Maintenance Problem. Now It’s Inspecting Buildings Less Often. For now, the quality-control problem is more straightforward: the federal government required someone to check the building inspector’s work, and for hundreds of assessments, it was prepared to let the building inspector do it.

    Fediverse Reactions

  • Washington Finally Measured Your Unpaid Labor. It Used Data Older Than Your Car.

    The Government Accountability Office wants Congress to know that unpaid household work, the cooking, the cleaning, the eldercare, the endless negotiation with a toddler over pants, was worth somewhere between $5.6 and $7.4 trillion in 2024 (GAO-26-107871). That is up to a quarter of gross domestic product, generated by nobody the Bureau of Economic Analysis has bothered to formally price since 2020.

    Six years is not a long time in most contexts. It is an eternity in economic statistics, where the whole point is to tell you what is happening now, not what was happening when “Encanto” was still in theaters. GAO’s new report, delivered to the Senate Special Committee on Aging, is framed as a straightforward exercise in economic measurement. Read past the tables and it becomes something else, a case study in two federal statistical agencies quietly deciding that a trillion-dollar blind spot in how America understands its own economy is somebody else’s problem.

    The Bureau That Waited to Be Asked

    The Bureau of Economic Analysis has known the answer to this question for a while. Its household production satellite account, the official government attempt to price unpaid domestic labor, has been running in some form since the 1960s. As of May 2026, per GAO’s own accounting, BEA’s most recent published estimate covered the years 1965 through 2020. Four years of a rapidly reshaping post-pandemic labor market went unmeasured by the agency built to measure exactly this.

    Then, conveniently, something happened. GAO circulated a draft of this report to the Department of Commerce for comment. Commerce’s response, reproduced in the report’s own Agency Comments section, notes that BEA “issued a staff working paper in June 2026 that included updated household production estimates.” June 2026. Three months before GAO’s report went public. The timing invites an uncharitable but not unreasonable reading; an agency that had gone half a decade without updating a core national account found the motivation to do so only once its homework was about to be checked by outside auditors.

    To be fair to BEA, updating this account is genuinely hard. It requires linking American Time Use Survey data to Current Population Survey wage data, choosing a valuation method that will be second guessed regardless of which one you pick, and defending the whole exercise to an audience that half suspects the entire project is ideological. None of that explains why the fix showed up on a GAO deadline instead of a public one.

    The Bureau That Didn’t Bother

    BEA’s lag is embarrassing. The Bureau of Labor Statistics’ position is worse, because at least BEA has an account to be late on. Buried in a footnote of GAO’s report is the admission that BLS “has explored developing estimates of the value of unpaid household work but has not yet produced them.” The agency describes this as part of a larger effort to improve its Consumer Expenditure Surveys, the survey that underpins how the government understands what American households actually consume.

    As of March 2026, GAO writes, BLS officials “were unsure if the agency will produce a consumption measure that includes valuing unpaid household work due to resource constraints.” Read that sentence twice. This is not a promise of future data with a vague timeline attached, the customary bureaucratic dodge. It is a federal statistical agency telling its own auditor that it might simply never finish work it has already acknowledged would improve its data, because the money isn’t there and nobody has made it a priority.

    For an agency whose entire mandate is producing the numbers policymakers use to make decisions, “we might not get to it” is a strange thing to say out loud, even in a footnote. It is stranger still when the number in question, by GAO’s own estimate, represents up to a quarter of GDP. BLS has effectively conceded that a measurement problem of that scale sits below the line on its priority list, indefinitely, and saw no reason to soften that admission for the record.

    Why This Isn’t Just an Economist’s Argument

    It would be easy to file this under abstract methodology, the kind of dispute that stays contained to journal articles and BEA working papers. It doesn’t stay contained. GAO’s report specifically flags that better data on unpaid caregiving could help set stipend levels for family caregivers of disabled veterans, citing a 2024 study in Social Science & Medicine that found VA caregiver stipends were largely offset by reduced healthcare spending for the veterans in question (Jacobs et al., 2024).

    That is not a new insight for GAO. This publication has covered agencies that keep rediscovering their own prior recommendations, and this is another entry in that file. GAO flagged the retirement security angle of unpaid caregiving back in 2019, in a report on parental and spousal caregivers facing financial risk in retirement (GAO-19-382), and flagged VA’s family caregiver program specifically that same year (GAO-19-618). Seven years later, the underlying data those recommendations depended on is still six years stale at one agency and nonexistent at the other. The policy questions did not go away while the measurement infrastructure sat idle; they just kept getting answered with worse information than they deserved.

    There is a version of this story where BEA and BLS are simply under resourced agencies doing their best with limited budgets, and there is real truth in that version. But “under resourced” is a description of a condition, not an excuse that dissolves the consequence. Congress asked a fair question about how the economy actually works, and the honest answer required an outside audit, six years of catch up work quietly rushed out ahead of publication, and a footnote confessing that the other half of the problem may never get solved at all.

    The Ledger Nobody Kept

    GAO’s report closes, as these reports do, with the polite fiction that agency comments were received and considered. The Department of Labor had nothing to say. The Department of Commerce pointed to a working paper that happened to appear right on schedule. Nobody at either agency was asked, on the record, why it took a congressional audit to produce data that should have existed already.

    The caregivers, the cooks, the people quietly propping up a quarter of the American economy in the background of their own lives, were never actually the ones being audited here. The agencies were. They passed, technically, in the way a student passes an exam by cramming the material the night before because they knew it was coming. Whether that counts as understanding the subject is a separate question, and it’s the one Congress should be asking next.

    Fediverse Reactions

  • Federal Register Watch: September 18, 2026

    The day’s most sweeping action came out of federal procurement policy, where four separate proposed rules published together would rewrite large stretches of the Federal Acquisition Regulation under an executive order aimed at loosening how the government buys goods and services.

    • OFPP proposes a four-part rewrite of the Federal Acquisition Regulation. The Office of Federal Procurement Policy, joined by the Defense Department, the General Services Administration, and NASA, published four separate proposed rules that together touch parts 8, 9, 12, 13, 14, 15, 16, 17, 27, 28, 35, 36, 38, 44, 47, 51, and 52 of the FAR, implementing Executive Order 14275, “Restoring Common Sense to Federal Procurement.” Splitting one rewrite into four dockets makes each filing easier to process, but it also means no single comment period covers the full scope of the change. That window closes October 19, just 31 days out, a tight timeline for contractors, watchdog groups, and other agencies to weigh in on a rewrite that reaches into commercial-item buying, indefinite-delivery contracts, construction, and dozens of other procurement mechanics. Read the lead proposal.
    • Justice Department settles part of its RealPage rent-pricing case, while the core suit continues. DOJ and ten state attorneys general published a proposed final judgment resolving their claims against Pinnacle Property Management Services, one of several large landlords sued alongside software vendor RealPage over allegations that its pricing tools let competing landlords coordinate rents through shared, competitively sensitive data rather than competing on price. Pinnacle would have to stop licensing that kind of revenue-management software and put antitrust compliance training in place, but the case against RealPage itself, which the complaint says controls roughly 80 percent of the commercial revenue-management software market, is unaffected. Settlements like this one go through the Tunney Act, which gives the public 60 days to comment before a judge decides whether the deal serves the public interest, so renters and tenant groups still have a formal chance to object before it becomes final. Read the filing.
    • FTC settles a rare interlocking-directorate case between two rival gunmakers. The FTC published a proposed consent order over Beretta’s right to appoint two directors to the board of competitor Sturm, Ruger & Co., which the agency says violated the Clayton Act’s ban on competitors sharing board members, a rule meant to stop rivals from trading sensitive business information through shared leadership. The order requires any Beretta-appointed director to be financially independent of Beretta, bars Beretta from seeking Ruger’s nonpublic information, and requires 15 days’ notice before any future board appointment. Cases enforcing that particular provision are uncommon, so one naming two of the country’s largest firearms makers stands out on its own. Comments are due October 19. Read the notice.
    • State Department eases export controls on underwater drones without the usual advance comment period. State issued an interim final rule removing certain uncrewed underwater vehicles from Category XX(a) of the U.S. Munitions List, loosening export restrictions on a defense-relevant undersea drone technology. Interim final rules take legal effect, October 19 here, on their own schedule rather than waiting for the public input an ordinary proposed rule would collect first; State says it will still take comments on refining the change, but those comments won’t hold up the rule from becoming binding. Reaching for that faster track to loosen, rather than tighten, export controls on militarily relevant hardware is worth flagging on its own. Read the rule.
    • A 2022 DOE energy rule was erased by Congress, not by DOE. The Energy Department published a notice stating that its 2022 final rule setting energy-conservation standards for manufactured housing “has no legal effect,” because a new law, the 21st Century ROAD to Housing Act, overrode it. The notice took effect immediately with no comment period, because there was nothing left for DOE to decide: a rule that took years of standard rulemaking to finish was wiped out by lawmakers rather than through the notice-and-comment process agencies normally have to use to repeal their own regulations. Read the notice.
    • OPM proposes rewriting who counts as an “excepted service” federal employee. OPM proposed changes to the rules governing the excepted service, the Pathways Programs for students and recent graduates, and how administrative law judges get appointed. Excepted-service positions are filled outside the competitive process that applies to most federal jobs and often carry fewer job protections, so redrawing which jobs fall into that category can shift both how much hiring flexibility agencies have and how much insulation from ordinary civil-service rules. Comments are due November 17. Read the proposal.

    This edition draws on the Federal Register’s full docket of 114 documents published on September 18, 2026. Set aside as routine: antidumping and countervailing-duty determinations and orders, a batch of roughly 18 SEC self-regulatory-organization filings tied largely to exchanges’ move toward a 23-hour trading day, FERC combined notices and utility and pipeline filings, routine OMB/PRA information-collection notices and corrections, advisory-committee and agency meeting notices, and other standard administrative filings that didn’t trip an institutional-significance or anomaly signal.

  • Born in Bailout, Raised in Denial: The Fannie and Freddie Story Washington Still Can’t Finish

    A Subsidiary Trying to Forget Its Parent

    Every family has one relative nobody wants to bring up at dinner. For Fannie Mae, that relative is the Reconstruction Finance Corporation, the New Deal lending machine this series covered two weeks ago, the one that spiraled into a corruption scandal so severe it forced Congress to write the Government Corporation Control Act of 1945 just to contain the damage.

    Fannie Mae was not inspired by the RFC. It was chartered on February 10, 1938, as a wholly owned subsidiary of the RFC itself, capitalized with roughly ten million dollars in RFC funds and tasked with buying up FHA insured mortgages so banks would keep lending during the Depression, according to a detailed institutional history of the entity compiled from congressional records. Seventeen years after its parent got shut down for operating as an unaccountable financial power center, Fannie Mae is still around, still backed by an implicit federal promise nobody in Washington will fully admit to, and still, as of this writing, technically a ward of the state. Its sibling, Freddie Mac, has been along for the ride since 1970. Between them they now guarantee or hold something like 7.7 trillion dollars in mortgage debt, more than the debt of every company and government on earth except the United States and Japan, according to researchers at the University of Florida’s Warrington College of Business.

    The Vietnam War Era Accounting Trick That Still Runs the Mortgage Market

    Here is where the story gets genuinely absurd. In 1968, the Johnson administration needed to make the federal budget look smaller while paying for the Vietnam War and the Great Society simultaneously. Fannie Mae, still a line item on the government’s books, was an inconvenient liability. So Congress passed the Housing and Urban Development Act of 1968, split Fannie Mae into two pieces, spun off the fully government owned mortgage insurance function into a new entity called Ginnie Mae, and turned the rest into a nominally private, shareholder owned corporation. Fannie Mae’s debt vanished from the federal balance sheet essentially overnight, a maneuver policy analysts have bluntly described as a clever accounting solution driven by budget politics rather than any free market conviction that Fannie Mae belonged in private hands.

    The new “private” Fannie Mae kept a two billion dollar line of credit with the Treasury, stayed exempt from state and local income taxes, and never had to register its securities with the SEC, the kind of privileges no actual private mortgage company enjoys. Congress created a second entity, Freddie Mac, in 1970 specifically to give Fannie some competition, using the identical hybrid structure. The Competitive Enterprise Institute summarized the resulting arrangement about as bluntly as anyone has: the model privatized the profits to shareholders while socializing the losses onto taxpayers, and it stayed that way for four decades because nobody in either party had much incentive to change it while the mortgages kept performing.

    They stopped performing in 2007. By the time the subprime crisis hit full force, Fannie and Freddie together held or guaranteed 5.3 trillion dollars in mortgage backed securities and debt, an amount roughly equal to the entire publicly held debt of the United States government at the time, according to congressional testimony from then FHFA Director James Lockhart. On September 6, 2008, with both companies unable to raise capital and staring down eighty nine billion dollars in debt coming due within months, their own boards agreed to hand control to the newly created Federal Housing Finance Agency under the Housing and Economic Recovery Act, according to FHFA’s own history of the conservatorships. The federal government had spent forty years pretending Fannie Mae was not its problem. It became the government’s problem again in a single weekend, and Treasury ultimately pumped in 191.5 billion dollars to keep both entities solvent.

    The Entities Built to Dodge the Law Written to Catch Them

    This is the part that should genuinely bother anyone who followed the Fourth Branch piece in this series. Congress wrote the 1945 Government Corporation Control Act specifically to make sure entities like the RFC could never again grow into unaccountable, off books financial power centers without mandatory budgets, GAO audits, and borrowing limits. Fannie Mae and Freddie Mac sidestepped that entire accountability regime by design, not by accident. Because the 1968 restructuring made them nominally private, shareholder owned companies, they were never classified as government corporations under Title 31 at all. They are government sponsored enterprises, a separate legal category invented specifically to capture the benefits of a federal charter, the implicit backstop, the tax breaks, the cheap funding, while dodging the exact oversight machinery Congress built after the RFC scandal.

    In other words, Washington did not just fail to learn the lesson of the RFC. It engineered a structure explicitly designed to route around that lesson, and the result was an entity that grew larger and more systemically dangerous than the RFC ever was, precisely because nobody was legally required to watch it the way the 1945 Act required watching TVA or the FDIC. The GSE model is the RFC’s most consequential offspring not because it repeated the RFC’s mistakes, but because it corrected for the wrong ones.

    Eighteen Years of “Temporary”

    Conservatorship was supposed to be a stabilization measure. As of this writing it has run eighteen years, longer than the RFC itself lasted from birth to liquidation. Fannie and Freddie returned to profitability by 2012 and have since paid the government back many times over in dividends, which has only intensified the political fight over what to do with them next. Treasury Secretary Scott Bessent said in 2025 that ending the conservatorships was an explicit administration goal, and FHFA Director Bill Pulte has spent 2026 signaling that the arrangement should not continue indefinitely. In late June, Representative Scott Fitzgerald introduced the Sustainable Homeownership Act, a bill aimed at establishing statutory guardrails for a formal release, according to National Mortgage Professional’s coverage of the legislation. Hedge fund manager Bill Ackman has been buying the common stock of both companies for years on the bet that release is coming, and named it his best investment idea heading into 2026.

    Nobody involved seems eager to just let the conservatorship expire on its own terms, because nobody can agree on what happens to the implicit government guarantee once it does. Release Fannie and Freddie without an explicit backstop from Congress and investors may demand higher returns to compensate for the added risk, which likely means higher mortgage rates for every homebuyer in the country, a concern serious enough that both Bessent and Pulte have said publicly they will not move forward if the data shows rates going up.

    That is the real inheritance the RFC left behind. Not a single scandal or a single bailout, but a durable federal habit of building financial entities that look market disciplined on paper while carrying government risk in practice, then discovering decades later that unwinding the arrangement is far harder than creating it ever was. The RFC took twenty five years from founding to liquidation. Fannie Mae and Freddie Mac have already outlasted that timeline twice over, and the exit ramp is still being drafted in real time, in Congress, this year, while the rest of us wait to see who actually ends up holding the risk when the paperwork finally clears.

    This is Part 3 of our series on U.S. government corporations. Read Part 2: The Slush Fund That Ate Washington: How the RFC Invented the Modern Government Bailout.

    Next in the series: the entities that never broke this way, TVA and the FDIC, and what it actually takes for a government corporation to do the one thing it was designed to do without becoming somebody’s eighteen year headache.

  • Energy Department Confirms July and August Successfully Occurred

    At 4:05 p.m. on September 16, an email arrived from the U.S. Department of Energy’s Office of Cybersecurity, Energy Security, and Emergency Response bearing important news.

    July and August happened.

    The subject line, “CESER News & Updates – July & August 2026,” was refreshingly honest about the situation. There would be no September news in this September 16 news email. Instead, CESER had assembled a collection of developments stretching all the way back to July 2, presumably after allowing sufficient time for the news to mature.

    This might be unremarkable coming from an office devoted to historical preservation. CESER, however, is responsible for cybersecurity, energy security and emergency response, fields in which the phrase “we’ll get back to you next month” generally lacks a certain operational appeal.

    The newsletter opened with a collection of grid reliability actions. DOE issued an emergency order for the Carolinas on July 2 amid hot weather. Another followed for the Southwest Power Pool on July 21. There was another on July 26 covering 17 states, followed by August actions involving Guam, Puerto Rico, the Midwest and the Mid-Atlantic.

    These are not trivial subjects. Grid emergencies, blackout risks and extreme weather are exactly the sort of matters the public might reasonably associate with timely information.

    The cybersecurity section was similarly urgent in subject matter. CESER highlighted Stormbreaker, a testbed developed with Lawrence Livermore National Laboratory to evaluate large language models and agentic artificial intelligence in power systems and operational technology environments. It also pointed readers toward cybersecurity threat guidance, operational technology security, software bills of materials and reports of cyber incidents affecting water and wastewater systems.

    All useful material. All arriving in an email on September 16.

    The newest dated CESER event in the email was an August 27 appearance by Principal Deputy Director Emily Burdick at a National Governors Association summer meeting. That means even the freshest item in the “News & Updates” package had enjoyed nearly three weeks to contemplate its journey to our inbox.

    The newsletter even contains a particularly delightful sentence from an August 5 event at the National Lab of the Rockies. CESER Director Andrew McClure emphasized the office’s guiding principle of providing “timely and actional information to the energy sector.”

    We assume “timely” is being used here in the federal sense.

    To be fair, there is nothing inherently wrong with a periodic retrospective newsletter. Agencies routinely compile activity over weeks or months, and CESER did not disguise what it was sending. The words “July & August 2026” were right there in the subject line. Nobody promised breaking news.

    That may actually make it funnier.

    Somewhere in the machinery of the federal government, an office dedicated to responding to rapidly developing threats compiled its July and August news, shepherded it through whatever process was required, and successfully delivered it halfway through September.

    The grid remained standing. The email arrived. And Bureaucracy Times can now independently confirm that July and August 2026 successfully occurred.

    Fediverse Reactions