Who is this committee working for— you or capital?

Not one committee of the 119th Congress has a stated objective that can be tested against anything. 32 of the 46 are never asked for one. The 17 that are asked filed 455 commitments between them, and not one of those 455 sets a date by which anything will be done.

And nothing is reported back. None of the 46 publishes a status against a commitment it made. None reports what changed as a result of its oversight. A commitment with no date and no status cannot be early, cannot be late, and cannot be missed.

This is the condition, not a lapse. The rules do not ask most committees for an objective, and ask the rest for nothing measurable. Every figure in this window is a count of that.

What is missing. Whether a committee pursued an intention it never wrote down. Nothing published would show it either way.

Who is counted, and who is not. 46 committees of the 119th: 36 standing, 4 select, 4 joint, plus Senate Aging and Senate Indian Affairs. congress.gov lists 53 bodies under a committee code. Five of them are commissions or caucuses — the Helsinki Commission, the Tom Lantos Human Rights Commission, the Congressional-Executive Commission on China, the U.S.-China Economic and Security Review Commission and the Senate Narcotics Caucus. None holds legislative jurisdiction and none can receive a referral, and between them they received none. Two more are the Joint Economic Committee emitted a second and third time under different system codes. Subcommittees are not counted; the unit is the parent committee. 53 − 5 − 2 = 46.

A correction. A version of this page published before 27 September 2026 counted 49, carrying the Helsinki Commission, the Senate Narcotics Caucus and one select subcommittee. Measuring a commission against questions asked of committees produces an absence that reads as a finding: of course it publishes no oversight plan, and of course no measure was referred to it. Seats move with the count, from 1,339 to 1,316.

This is not disobedience. It is the rule. Only 17 of the 46 bodies here — all of them House standing committees — are required to publish what they intend to do. The other 32 are asked for nothing, and nothing is what the public gets.

The Senate is the larger part of it. No Senate committee publishes a forward-looking plan at any point. Rule XXVI paragraph 8 asks only for a backward description, filed two years later, of what the committee already did. A senator is never on record in advance.

Who is excused, and it is not random. House Rule X clause 2(d)(1) names three: Appropriations, Ethics and Rules — the committee that writes the numbers, the committee that judges members, the committee that decides what the House may object to. The Senate has no forward-looking plan at all; its committees describe what they already did, two years later. Every select, special and joint body is required to say nothing, so the bodies standing up for the most contested subjects are the ones publishing least.

This is the condition, not the failure. The rules do not ask most committees for an objective. The failure is that nobody has to.

The 17 House committees required to file a plan filed 455 commitments between them. One states when. House Agriculture writes that its goal is to reauthorize the Commission on Security and Cooperation in Europe before the end of the 119th Congress. That is the only commitment in the 455 that names a point by which the committee will have acted. A commitment with no date cannot be early, cannot be late, and cannot be missed.

How the one was found. A date pattern — a named month or year, “no later than”, “within n days”, “before the end of” — matches 4 of the 455. Three of the four are dates inside background narrative rather than dates the committee sets for itself: a foreign military’s own target year of 2049, a past event in 2017, and a deadline a statute set for the Department of Veterans Affairs in 2023. Those three are excluded by name in ledger.py, so the exclusion can be checked rather than taken on trust.

55 of the 455 contain any verb that binds — reauthorise, report to, introduce, mark up, require. The rest describe attention, not action.

Every commitment appears on its committee’s card in the words it was filed in, so a member can be read their own sentence and asked what came of it.

A correction. A version of this page published before 27 September 2026 said none of the 455 set a date. One does.

386 of the 455 — 84% — are built on review, examine, monitor, consider, assess or continue to. These are the verbs of watching. None of them commits the committee to produce anything, and none can be failed.

A plan written this way satisfies the rule and promises nothing. That is not an accident of drafting: the rule asks what the committee will examine, and never asks what it will deliver.

Nowhere — not on a committee website, not in a filing, not on congress.gov — is there a statement of whether a commitment made in an oversight plan was acted on. The promise is filed in March of the first year and never mentioned again.

The House activity report filed at the end of the Congress describes what the committee did, but it is not written against the plan and does not refer to it. The two documents never meet.

So a constituent asking “what happened to the thing you said you would look at” has no place to look. That is the gap this instrument exists to make visible.

No committee publishes an outcome. Not a problem found and fixed, not money recovered, not a programme changed, not a rule withdrawn. The record shows hearings held and reports filed; it does not show what any of it produced.

What can be seen from outside is thin and indirect: 1,373 hearings held in the 119th, 238 of 832 statutorily required reports delivered late, and 512 programmes still funded after their authorisation of appropriations expired. Those are inputs and failures, not results.

An audit that cannot state an effect is not finished. Neither is Congress’s account of its own oversight.

Congress authorises a programme, then appropriates money for it. The two are supposed to match. The Congressional Budget Office lists 1,447 authorisations that have already expired. On 512 it can name the appropriation. Against the other 935 it records one word: unidentifiable.

This is not missing money. It is a broken trail. The spending exists, inside an appropriations account. What cannot be done is tie it back to the expired authorisation of appropriations that was supposed to justify it. Appropriations are made to accounts in lump sums; an authorisation is for a named programme inside one.

Nor is it a gap in the effort. Those 935 rows carry a public law and a named appropriations subcommittee in every case, a U.S. Code citation in 71% of them, and a stated authorisation ceiling on 742 of them worth $52,395,955,000. Every field was filled in except the one that would complete the chain. A further 193 were authorised in indefinite amounts — such sums as may be necessary, no ceiling at all.

721 more are expiring rather than expired — still live, still fundable, and the ones a committee could still act on.

Everything this page says about the purse concerns the 512 where the connection can be made, which is 35% of the expired ones. Those figures are exact. For the rest the chain does not close.

The report is required; the format is not. House Rule XI clause 1(d) makes every committee file an account of the Congress just ended. It does not say what the account must contain, and the 20 reports for the 118th show what follows from that.

Ten state no hearing total at all. Of the ten that do, no two count the same thing: Armed Services reports 228 meetings and hearings together; Rules reports 70 days of hearings; Homeland Security reports 22 for the full committee and leaves its subcommittees out; Appropriations reports 71 and 61 for its two sessions and never adds them; House Administration’s 13 counts only elections hearings; Natural Resources’ 86 only the oversight ones. Financial Services is the one that can be checked — it states 114 and tabulates all 114, which recounts exactly.

Twenty mandatory reports that cannot be set against one another is itself the finding. Each is linked on its committee’s card with whatever figure that committee stated and the scope it stated it in. They cover the 118th. Activity reports are filed at the end of a Congress.

The rule asks for exactly this. Element (A) of clause 2(d)(2) tells a committee to list the programmes in its jurisdiction whose authorisation of appropriations has already run out. It is the one element with money attached.

Ten plans list none. 162 of the 512 expired authorisations of appropriations sit in the jurisdictions of those ten committees, carrying $226,406,199,000 in fiscal 2026. The programmes are funded; the document meant to name them does not.

And no plan is complete. Not one of the 17 contains language responsive to all five elements. The rule qualifies the duty with “to the maximum extent practicable”, so this is a reading of filed text, not a finding of breach.

The definition, in full. A measure advanced if any committee it was referred to reported it or was discharged of it, or if it became law. 1,618 did: 1,324 were reported, 260 were discharged without being reported, and 34 became law without either. Everything else did not advance. 1,618 + 17,631 = 19,249.

How to reproduce it. Take bill_committee_referrals_119.csv, the congress.gov BILLSTATUS referral harvest for the 119th — 24,288 rows, one per bill-committee referral, as at 2026-09-26. Group by bill type and number to get 19,249 distinct measures; 3,295 of them went to more than one committee. Mark a measure advanced if any of its rows has reported, discharged or became_law set. The script is build_disposition.py and it writes every figure in this tile.

What this does not say. It does not say a committee killed anything. A referred measure that does not advance has, in most cases, simply received no further action, and the record does not distinguish inaction from a decision. Nor is this a comparison across Congresses: the 119th is still sitting at 2026-09-26, so a measure still live is counted here as not having advanced.

A denominator, not an indictment. Almost everything introduced dies, in every Congress, and that is what a filter does. It is here so the other numbers have something to be measured against.

Counted as distinct measures. A bill referred to three committees is one bill, not three.

A correction. A version of this page published before 27 September 2026 gave this pair as 17,665 and 1,584. Those two figures came from different populations — the denominator from the whole harvest, the numerator from the rows naming one of the 46 committees on this page, and the numerator left out the 34 measures that became law without being reported or discharged. Both figures above now come from the same population.

113 measures became law in the 119th. On 104 the record names a recipient. On 19 of those, delivery is verified. Not a reading of the bill’s language — a delivery test applied to each enacted law, asking whether anybody actually received the thing it describes. Each is listed on its committee’s card with the recipient named: veterans on a disability cost-of-living rise, trafficking survivors able to vacate a conviction, disaster-hit taxpayers given a deadline, victims of intimate-image abuse given a removal right.

14 name the person directly; 5 arrive through a government body first.

This does not mean the value arrived whole. The test asks whether a recipient can be named. It does not measure what share of the money reaches them, or what an administrator, contractor or county takes on the way. Nothing published measures that.

The record says so itself in places. P.L. 119-97 names its recipients as veterans and contractors. Secure Rural Schools pays counties, which then spend it. Voc-rehab runs through contracted providers. And the largest expired-authority line on this instrument — veterans’ medical care — reaches veterans through Optum and TriWest, private administrators earning fees on the way.

What the other 85 did. 44 delivered to an asset holder. 16 delivered nothing at all. 13 produced transparency or awareness but nothing anyone receives. 8 left no delivery evidence either way. 2 imposed a reporting duty on an agency. 1 was a funding vehicle carrying no delivery of its own. 1 changed a federal workforce term. 44 + 16 + 13 + 8 + 2 + 1 + 1 = 85.

9 enacted laws have not been through the delivery test yet. Public Laws 119-103 through 119-111 — H.R. 6500, S. 858, H.R. 1276, H.R. 2069, H.R. 2196, H.R. 5366, S. 32, S. 307 and H.R. 5334 — became law after the test was last run, and carry no recipient and no delivery finding. They are the difference between 113 and 104. Each is a law of the 119th in its own right, not a provision riding inside another measure. Nothing published answers the question for them, which is not the same as an answer of no one.

5 laws counted as citizen-serving fail their own delivery test and are recorded as going to an asset holder — among them the Whole Milk for Healthy Kids Act, a milk-in-schools measure sought by the dairy industry.

A correction. A version of this page published before 28 September 2026 gave this pair as 19 of 104 and described 104 as the count of measures that became law. 104 is the number of enacted laws where a recipient can be named; the law count is 113, as recorded in this instrument’s own disposition file and in the Layer 1 enactment table at the 26 September as-of (111 public laws and 2 private). 19 is the number where delivery is verified, not the number naming a recipient. The three figures nest: 113 laws, 104 naming a recipient, 19 verified.

A blank is unverified, never disproved. Where no recipient can be named, the record does not show that nobody received anything — it shows that the question cannot be answered from what is published.

A Congressional Review Act resolution does not suspend a rule. It ends it. Under 5 U.S.C. §801(b)(2) a disapproved rule may never be reissued, nor one substantially similar, unless a later statute allows it. No further vote keeps it dead.

23 were passed in the 119th. Among what they ended: the CFPB overdraft-fee cap, the IRS reporting rule for decentralised finance, energy-efficiency standards for appliances, commercial refrigeration and gas water heaters, and California’s vehicle-emissions waivers.

That is out of 51 stopping actions in all — the other 28 being 23 appropriations riders defunding implementation and 5 statutory moratoria. Those two lapse: a rider with the fiscal year, a moratorium on its own date. The CRA repeal does not. The full list with vehicle, agency and duration is in the table below this section.

What is missing. Rules an agency withdrew under committee pressure with no instrument at all. That leaves no document, so it cannot be counted.

Three of every four people sent to these 13 Senate committees were confirmed. Any account of nominations reporting only the failures is not describing the institution.

The rest. 484 were never confirmed: 183 died in committee, 90 returned to the President, 57 withdrawn, 36 heard and then dropped, 87 still pending. 212 more cleared committee and were parked on the Executive Calendar. The source records no rejection at all — a committee that refuses in the open leaves a record; one that never schedules the vote leaves none.

House committees confirm nobody — the power is the Senate’s alone, so House cards show a dash here rather than a zero.

Congress writes a reporting duty into law and the law sets a date. The agency files with the Government Publishing Office, which records the date due and the date it arrived. For the 119th: 832 reports owed to these committees, 594 on time, 238 late. 31 of the 46 committees receive them, and each count is on its card.

What is missing. The repository holds only what agencies filed. A report never delivered does not appear as late — it does not appear. This counts lateness among reports that arrived, and so understates the failure.

One definition applied to all 46. Every committee meeting of the 119th is recorded as an event with a type, a date and a named committee: 1,373 hearings, 276 markups, 1,039 other business meetings. 57 cancelled or postponed events excluded; 35 hearings were closed to the public. Eight committees held no hearing at all, among them Rules, the Joint Committee on Taxation and both House select committees.

Not from the committees’ own accounts. Those cannot be compared to one another — see the first window. Each is still linked on its committee’s card with the figure that committee stated and the scope it stated it in.

Witnesses and subpoenas are not counted anywhere. Witnesses exist only inside printed transcripts; subpoenas are recorded nowhere systematically. The subpoena is the difference between a request and a requirement.

Start with the question nobody asks: where does the money end up? Of the 512 expired-authority lines, an end operator and its owner could be established on 16. Not that the other 496 are clean. No one publishes it, and each one has to be traced by hand.

Of those 16, ten run to private investment funds, carrying $103,576,607,000: NIH research, Head Start, Title I schools, workforce training, diplomatic security and three NASA programmes. The funds named in the record are New Mountain Capital, Francisco Partners, Blackstone, BlackRock Impact Opportunities, The Builders Fund, HPS, OneIM, BC Partners, Veritas Capital and Madison Dearborn Partners.

The largest line is not a fund. Veterans’ medical care, $115,086,000,000, is operated by Optum — UnitedHealth Group, a public company. Three more lines worth $65,233,302,000 — Section 8 housing, Pell grants, public housing — stop at “owner not established”, which is where the trail actually ends for most federal money.

What this is not. 16 of 512 is not a sample and nothing here should be read across the other 496. Two further lines name a fund-owned operator in the sector without a verified payment to it, and are excluded from the 10. What the 16 establish is narrower and harder to answer: when somebody does the tracing by hand, this is what is found — and Congress does not do the tracing.

Every one of these programmes is running on an authorisation that expired. The committee that was meant to revisit it has not, and the operator collecting the money changed hands without that committee examining the change.

Passing something on is not the same as reporting it. 186 measures had their text enacted inside a public law carried by a different bill. 146 were never reported or discharged by any committee at all. 74 vehicles carried the 186, and three of those carried 95.

A provision that rides appears on neither record — not the committee that wrote it, which never reported it, and not the one that carried it, which never claimed it.

Why the cards show more. The per-committee column totals 208, because a provision touching two jurisdictions is recorded against both. 186 is the distinct-measure figure and the one to quote.

Why the number does not exist. A detailee is paid by the lending agency, so no committee payroll shows them, and neither chamber publishes a central staff listing that identifies them.

This is the gap that matters most in this window: the authority to borrow staff from the agencies you supervise is on the record, and the headcount is not. The door is demonstrably open. How many walked through it is not recorded anywhere.

What would close it. Committee-level disclosure of detailee counts, or agency reporting of outbound details to Congress. Neither is currently required.

The authority is on the record. Each Senate committee’s biennial funding resolution authorises it to use, on a non-reimbursable basis, the personnel of departments and agencies. The committee gets the staff; the agency keeps paying them.

Five do not carry it: Senate Committee on Appropriations, Senate Committee on Environment and Public Works, Senate Committee on Indian Affairs, Senate Select Committee on Ethics, United States Senate Caucus on International Narcotics Control.

What is missing. The House side. House committee staffing is authorised differently and the equivalent provision has not been read across the 20 House committees.

This is a floor, and a biased one. The only systematic place a committee detail shows up is the covered-position disclosure a person files when they later register to lobby. It therefore catches only those who went on to lobby, and misses everyone who did not.

Each name appears on its committee’s card, as filed.

What is missing. Everyone currently on loan, everyone who returned to their agency, and everyone who left government. There is no reason to think 17 is close to the real number.

Finance leads, then Homeland Security and Governmental Affairs and Commerce with two each. Read against the floor above, this ranks disclosure, not exposure: it identifies the committees whose borrowed staff most often went on to lobby.

Direction is the point. The literature on institutional capacity describes the executive branch loading work onto agencies. This is the reverse — the committee taking staff from the body it is supposed to supervise, and not paying for them.

People who left and now lobby are a separate record and appear here only so the two reconcile.

What a federal fund holds is not money. When a fund takes in more than it pays out, the surplus is not banked. It is lent to the general fund and spent on whatever the government is spending on that year. The fund is issued a non-marketable certificate in exchange — a Government Account Series security. Treasury reports the total as intragovernmental holdings.

The size of it. $8.12 trillion at 30 June 2026. Treasury’s own Bulletin lists every one of them in Table FD-3, and the Monthly Statement of the Public Debt is the underlying record.

Who is owed the most. Old-Age and Survivors Insurance $2,268,181 million. Federal employees’ retirement funds $1,117,814 million. Hospital Insurance $277,926 million. Supplementary Medical Insurance $191,373 million. Unemployment $100,139 million. The Highway Trust Fund $57,328 million. The Airport and Airway Trust Fund $21,998 million.

Which way it is moving. Old-Age and Survivors Insurance held $2,755,785 million at the end of fiscal 2021 and $2,400,808 million at the end of fiscal 2025. Redeeming a certificate takes cash the general fund has to find that year.

Two fifths of it sits in a line called Other. Treasury itemises sixteen funds by name and puts the rest — $3,408,783 million, 42 per cent of the total — in a single unnamed line. The table does not say which accounts are owed it.

What is missing. The certificates are recorded by fund. They are not recorded against the appropriation the borrowed cash paid for. The trail stops at the loan.

Each fund’s authorising committees are named in the table below.

A repeal under this authority does not expire. 23 joint resolutions of disapproval became law in the 119th. Under 5 U.S.C. 801(b)(2) an agency may not issue a substantially similar rule again unless a later statute says it may. There is no expiry on any of them.

What they were aimed at. 13 were rules published in the Federal Register as rules. 4 were published as Notices. 6 never appeared in the Federal Register at all: they were Bureau of Land Management records of decision and resource management plans, treated as rules on the strength of a letter of opinion from the Government Accountability Office, printed in the Congressional Record. Every opinion is dated and its Record pages are named in the table below.

Why that is the control question. The Congressional Review Act is the change-control mechanism over agency rulemaking. On 10 of these 23 it was used on something that was never filed as a rule.

Who carried them. House Energy and Commerce 9, House Natural Resources 7, Senate Banking 3, Senate Energy and Natural Resources 2, House Ways and Means 1, Senate Environment and Public Works 1.

What is missing. Whether any member raised the classification on the floor. That is in the Congressional Record, line by line.

The two that were scored. H.J. Res. 35, the waste emissions charge on petroleum and natural gas systems: the Congressional Budget Office put federal revenues down $7,475 million over 2025–2035. H.J. Res. 25, broker reporting on digital asset sales: the Joint Committee on Taxation put revenues down $4,500 million over the same period. $11,975 million between the two.

The other 21 carry no estimate at all. No cost estimate was published for them.

What the agencies had said about their own rules. Eight of the 23 repealed instruments state a figure of their own:

• P.L. 119-2 — -$7,475 million, on federal revenues, CBO. CBO cost estimate for H.J. Res. 35, 26 February 2025
• P.L. 119-5 — -$4,500 million, on federal revenues, Joint Committee on Taxation. CBO cost estimate for H.J. Res. 25, incorporating JCT
• P.L. 119-6 — $0.87bn (7%) to $3.06bn (3%), on net present value of consumer benefits, as the rule states it. the repealed rule itself, 89 FR 105188
• P.L. 119-7 — $2.00bn (7%) to $4.74bn (3%), on net present value of consumer benefits, as the rule states it. the repealed rule itself, 89 FR 104616
• P.L. 119-9 — $1.32bn (7%) to $3.43bn (3%), on net present value of consumer benefits, as the rule states it. the repealed rule itself, 90 FR 7464
• P.L. 119-10 — $5,200 million a year, on consumer savings per year, as the rule states it. the repealed rule itself, 89 FR 106768
• P.L. 119-11 — up to $1.4 million a year, on cost of supervision to industry, as the rule states it. the repealed rule itself, 89 FR 99582
• P.L. 119-14 — $13.3 million a year, on annualised compliance cost to industry, as the rule states it. the repealed rule itself, 89 FR 94886

What cannot be added up. Those figures sit on different bases — federal revenue, consumer benefit, industry cost. A single total across them would mean nothing, so none is given.

What is missing. The remaining 15 state no figure, and the six that were never published as rules carry no economic analysis to state one in.

Miles City, Montana. 1,745,040 acres of Bureau of Land Management coal made unavailable for further coal leasing, stated in the notice at 89 FR 93650. Voided by P.L. 119-48 on 11 December 2025.

Buffalo, Wyoming. 413,250 acres, in the same notice. Voided by P.L. 119-51 the same day.

And 225,504 acres in Minnesota. National Forest System land in Cook, Lake and Saint Louis counties, withdrawn for 20 years from the mineral and geothermal leasing laws — naming the Mineral Leasing Act of 1920 — by Public Land Order 7917 at 88 FR 6308. Voided by P.L. 119-85 on 27 April 2026.

And 3,601,000 acres in Alaska designated as Areas of Critical Environmental Concern or Research Natural Areas in the Central Yukon plan, at 89 FR 92716. Voided by P.L. 119-50.

Two of the seven state no acreage anywhere Congress could read it. The Coastal Plain record of decision and the National Petroleum Reserve plan. Their notices give none, and the records themselves were never published in the Federal Register.

Separately, the minimums written into statute. P.L. 119-21 requires 30 Gulf of America sales offering at least 80,000,000 acres each, 6 Cook Inlet sales of at least 1,000,000 acres, 4 Coastal Plain sales of at least 400,000 acres, 5 National Petroleum Reserve sales of at least 4,000,000 acres, and at least 4,000,000 additional acres of coal made available within 90 days of enactment. Those are acres that must be offered, not acres leased.

And two more land disapprovals are still sitting. S.J. Res. 109 in Senate Energy and Natural Resources and H.J. Res. 151 in House Natural Resources, both aimed at the Bureau of Land Management’s Grand Staircase-Escalante management plan, both referred on 4 March 2026 and neither reported since.

The cut. P.L. 119-21 section 50202(a) amends the Mineral Leasing Act: 12½ per cent becomes not more than 7 per cent from 4 July 2025 to 30 September 2034.

It reaches backwards. Section 50202(b) applies the lower rate to any coal lease issued before, on or after enactment that has not been terminated.

And it refunds. Section 50202(c) directs the Secretary to credit a lessee the difference between advance royalties already paid and what would have been owed at the lower rate.

Oil and gas, in the same law. Section 50101(a)(1) repealed the 2022 act’s onshore royalty rate. Section 50101(d) repealed the offshore rate and set a ceiling of 16⅔ per cent. Section 50103 repealed the charge on extracted methane.

What Congress was told. The Congressional Budget Office scored the Natural Resources committee print: onshore leasing up $12.8 billion in receipts over 2026–2034, offshore up $6.3 billion, the coal sections down $237 million, the title reducing deficits by $20.2 billion. The receipts are bonus bids from sales the statute mandates. The estimate stops in 2034. The coal rate cut stops on 30 September 2034. Production does not stop at either.

Two estimates, pointing opposite ways, measuring different windows. The Congressional Budget Office window closes in 2034 and counts bonus bids up front. The Resources for the Future DOGMA model, run by Brian C. Prest, follows the royalty over the life of production: federal royalty revenue down $6 billion over 2026–2035 and $42 billion over 2036–2050, averaging close to $3 billion a year in the later period. Onshore falls from 16.67 per cent and offshore from 18.75 per cent, both to 12.5 per cent — a cut of a quarter to a third.

States lose too. The federal government keeps a little over half of royalty revenue. States and the water and conservation funds that receive the rest face losses of similar size, for as long as the cut runs.

Which of the two is official matters less than what each one counts. One is the scorekeeper working to a ten-year window. The other is an outside model, named, with its author and method stated. Neither is hidden. The difference between them is the finding, so both are here.

What is missing. What federal royalties actually come to, by commodity and by year — the base both estimates are modelled against. It is published through a portal that serves no file a reader can set against a committee.

The prohibition. Clause 2(a)(1) of House Rule XXI: an appropriation “may not be reported in a general appropriation bill, and may not be in order as an amendment thereto, for an expenditure not previously authorized by law.”

Why it does not bite. It is not self-enforcing. It works by point of order, so a member has to be on the floor, recognised, and willing to raise it. Five special rules brought these 477 lines to the floor, and each of them provided that all points of order against consideration of the bill and against provisions in it were waived. A point of order that is waived cannot be made.

Five rules, named. Every one of the 477 lines came up under one of five resolutions. Each is set out below the tiles — the five rules that waived the objection — with the waiver in the resolution’s own words, the bills it brought up, and whether it named the clause carrying the prohibition.

Senate Rule XVI is narrower and does not create the same bill-level objection to unauthorised appropriations already carried in a reported measure.

What this does not say. It does not say the point of order would have been sustained. That is the Parliamentarian’s call, it is not recorded in this data, and CBO states that its list is not definitive for applying House or Senate rules about appropriating funds. It also does not say nobody objected: whether any member rose is not in this data, and the Congressional Record would have to be read line by line.

What is missing. 35 of the 512 lines carry no special rule in the record. That is silence, not a finding that they came to the floor unprotected.

Three groups, and they do not overlap. On 3 lines the resolution named clause 2 of Rule XXI — the clause that carries the prohibition (H.Res.530, 2 lines; H.Res.580, 1 line). On 71 lines the resolution named Rule XXI but a different clause: H.Res.992 disapplied clause 2(e), the emergency-designation restriction, which has nothing to do with unauthorised appropriations. On 403 lines the resolution never mentioned Rule XXI at all. 3 + 71 + 403 = 477.

So the count depends on what you are asking. 403 lines came up under a rule that never named Rule XXI. 474 came up under a rule that never named the clause containing the prohibition. The second is the figure that answers the question, and it is the one in the headline.

At the rule level: 80 special rules were adopted in the 119th, 29 of them governing appropriations. 18 of those 29 waived all points of order at once, and 9 of those 18 never named Rule XXI.

The line count and the rule count are computed from the filed resolutions and stated separately. They answer different questions: one counts money lines, the other counts resolutions.

A correction. A version of this page published before 27 September 2026 gave this figure as 438. That number could not be reproduced from the filed resolutions. The figures above are derived from the five resolutions named in the table, and they reconcile to 477.

The verb is the measure. Reading the authorising statute behind each of the 512 lines and recording which verb it uses.

The four readings, and what each one means. Every one of the 512 lines falls in exactly one group. 164 shall — the authorising statute directs the agency to act. 55 may — the statute permits the agency to act and does not require it. 172 neither — the statute establishes the programme or the office and states no obligation and no discretion either way. 121 unknown — the authorising statute behind the line is not on the record here, so no verb can be reported for it. 164 + 55 + 172 + 121 = 512. The last group is an absence of evidence, not evidence of absence: nothing is known about the verb on those 121 lines, and none of them is counted in the other three.

$30,516,944,000 sits behind the discretionary ones. A programme funded on a discretion, with its authorisation of appropriations expired and the objection waived in advance, has nothing holding it in place at any of the three stages: not the statute, not the authorisation, not the floor.

What is missing. The statute behind 121 of the lines. Until it is on the record this split is provisional: any one of them could turn out to say shall.

Neither shall nor may. The statute was read. It establishes the programme or the office and states no obligation to act and no permission to act. These are the hardest to argue about: there is no promise in the text to hold anyone to.

This is not the same as the 121 unread lines. These 172 statutes were read and found to state no duty. On the other 121 the statute has not been read at all, so nothing is known about the verb. A reader who adds the two together is counting a finding and a gap as if they were the same thing.

The four readings, and what each one means. Every one of the 512 lines falls in exactly one group. 164 shall — the authorising statute directs the agency to act. 55 may — the statute permits the agency to act and does not require it. 172 neither — the statute establishes the programme or the office and states no obligation and no discretion either way. 121 unknown — the authorising statute behind the line is not on the record here, so no verb can be reported for it. 164 + 55 + 172 + 121 = 512. The last group is an absence of evidence, not evidence of absence: nothing is known about the verb on those 121 lines, and none of them is counted in the other three.

At 172 they are the largest of the three groups that have been read — larger than the 164 mandatory ones.

What is missing. Whether the duty sits elsewhere, in a different section or a later amendment. The read is one statute per line.

This is the opposite of what you would guess. Classifying all 512 expired lines by what the programme does: 91 pay a direct benefit to a person, carrying $272,891,553,000 — 55% of all the money on lines whose authorisation of appropriations has expired.

Citizen-facing against everything else. Direct benefit 91 · protective 33 · enforcement 12 · accountability 15 — 151 lines carrying $315,858,853,000, 63% of the total. Everything else is 361 lines and $177,745,941,000.

So the lapsed money is not mostly corporate. It is mostly benefits. The programmes Congress has stopped reauthorising are disproportionately the ones that hand something to a person — and they are the ones funded each year with the objection waived in advance.

What is missing. The classification comes from the programme description in the CBO file, applied by rule to all 512. Descriptions are short, and a programme serving both a person and an industry is coded by its primary function.

The oldest is 46 years — an authorisation that lapsed in 1980 and has been funded every year since.

Median rather than average: a handful of very old lines would otherwise carry the figure. Each committee’s own median is on its card.

What this is not. Expired authority does not mean the spending is unlawful. The appropriation is what makes it lawful. It means the authorising committee has not revisited the programme in the time shown.

The rate is in this window, and so is what it is estimated to cost. The collections are not. What is charged is in statute and what it is modelled to lose is published. What is actually received, by commodity and by year, is reported by the Office of Natural Resources Revenue in a form that cannot be set against a committee.

And the rest of that side of the purse is unreported. Grazing and lease rates unchanged for decades, user fees never indexed, penalties assessed and never collected. Money the government is entitled to and does not take is as much a committee’s business as money it spends.

Eight of the 512 expired lines name a fee or collection authority on a read of the programme titles. The expired-authorisation-of-appropriations file records what is spent, not what is left uncollected.

What it would take. Interior and royalty collection reporting set against the lease, General Services Administration and Bureau of Land Management lease schedules, agency fee-setting authorities, and Treasury reporting of uncollected receivables. This headline says so rather than leaving a reader to assume the purse has one side.