04b — What the record showsNobody ever voted on the 16-bed limit
Every roll call of the 89th Congress was read, along with the Statutes at Large, the Federal Register and CBO's cost estimates. The rule was never debated on a floor, never amended on a floor, and never voted on by itself. Here is what the record does say.
First, precisely what the rule doesMedicaid does fund mental health care — outpatient treatment, and psychiatric units inside general hospitals. What the exclusion bars is federal Medicaid payment for an adult aged 21 to 64 who is a patient in a facility primarily engaged in treating mental disease with more than 16 beds. That is narrower than “Medicaid won’t pay for psychiatric care” — and devastating exactly where inpatient beds live, because almost no psychiatric hospital has 16 beds or fewer.
Nobody ever voted on it.
All 394 House and 497 Senate roll calls of the 89th Congress were enumerated. Exactly 16 touched H.R. 6675, and the House took only three recorded votes on the bill — none of them on an amendment. Every Title XIX provision dealing with institutions for mental diseases (Senate amendments 255, 262-263, 275, 276 and 279) was a committee amendment. Not one reached the floor.
Voteview roll-call data →Medicaid inherited the rule. It did not create it.
The exclusion dates to 1950, not 1965. P.L. 81-734 §303(a), 64 Stat. 549, created federal vendor payments and carved tuberculosis and mental-disease institutions back out of them in the same sentence. The 1935 Social Security Act rule was about an "inmate of a public institution" — an almshouse rule. The words "insane" and "asylum" appear zero times in the 1935 Ways and Means report or in the Committee on Economic Security report. In 1965 Congress actually narrowed the exclusion, by adding coverage for people 65 and over.
P.L. 81-734, 64 Stat. 477 →The Senate voted to abolish it in 1960. A conference committee killed it overnight.
On 23 August 1960 the Senate adopted an amendment by Sen. Russell Long (D-LA) permitting federal matching of vendor payments to public mental and tuberculosis hospitals. On 24 August the conference committee agreed to the Senate’s medical-care provisions "except for Senator Long’s amendment." That is the one moment in the record where the rule was genuinely contested, and it lasted a day.
SSA legislative chronology, 1960 →The number 16 was enacted as a technical correction.
It entered the statute at P.L. 100-360 §411(k)(14)(A), 102 Stat. 798-799, under the heading "SEC. 411. TECHNICAL CORRECTIONS TO CERTAIN HEALTH CARE PROVISIONS", subheading "(k) CORRECTIONS TO SUBTITLE B OF TITLE IV (RELATING TO MEDICAID)", paragraph "(14) CLARIFICATION OF TERM ‘INSTITUTION FOR MENTAL DISEASES’". It sits between a provision on educationally related services and a technical correction about eligibility verification.
P.L. 100-360, 102 Stat. 683 →The statute came first. The regulation copied it.
The intuitive story — that regulators invented the number and Congress ratified it — is the wrong way round. The 1978 rule at 43 FR 45204 defined an institution for mental diseases by its "overall character" and set no bed count at all. HCFA put Congress’s number into the Code of Federal Regulations at 56 FR 8854 on 1 March 1991, in a rule titled "OBRA ’87 Conforming Amendments" — two years and eight months after the statute.
43 FR 45204 (29 Sept 1978) →Congress fixed this for tuberculosis and not for mental illness.
Both conditions were excluded in the same 1965 clause. §2335(f) of the Deficit Reduction Act of 1984 (P.L. 98-369) struck the words "tuberculosis or" and left the psychiatric half of the sentence standing. Forty-two years later it has not been revisited.
P.L. 89-97, 79 Stat. 286 →In 2023 the House voted 386-37 to lift part of it. The Senate never took it up.
Section 304 of H.R. 4531 in the 118th Congress was titled "LIFTING THE IMD EXCLUSION FOR SUBSTANCE USE DISORDER". It cleared subcommittee 49-0 and full committee 29-3, then passed the House 386-37 on 12 December 2023 — Republicans 190-26, Democrats 196-11. It went to the Senate HELP Committee the next day and has not moved since: no hearing, no vote, and no member on record against it.
Voteview roll-call data →What keeps it alive is the price tag.
CBO scored the options in April 2023. Over 2024-2033 the narrow state-plan option for substance use disorder costs $155-560 million in net federal Medicaid spending — and that is the one Congress enacted. Repeal for substance use disorder costs $7.7 billion, repeal for mental illness $33.5 billion, and full repeal $38.4 billion. Everything ever enacted sits on the cheap side of that line. Every repeal bill sits on the expensive side, and none has left committee.
CBO Publication 59071 (April 2023) →
The real lineage
The chain runs thirty years before Medicaid. Each step was verified against the enacted statute, not against a summary of it.
The Social Security Act — and it is not about mental illness
Title I bars aid to anyone who “is not an inmate of a public institution.” That is a poorhouse rule, aimed at almshouses. The words insane and asylum appear zero times in the House Ways and Means report or in the Committee on Economic Security’s report to the President. Aid to Dependent Children carried no such clause, which makes the exclusion an adult-category rule from birth.
P.L. 74-271, 49 Stat. 620 →The actual origin point
P.L. 81-734 §303(a), 64 Stat. 549, creates federal “vendor payments” to medical providers — and then carves two categories back out of them. Mental institutions enter federal law here, as an exception to a liberalisation.
…but does not include any such payments to or care in behalf of any individual who is an inmate of a public institution (except as a patient in a medical institution) or any individual (a) who is a patient in an institution for tuberculosis or mental diseases…
P.L. 81-734 §303(a), 64 Stat. 549
House conference report 373–1, with Republicans 137–1 in favour. Senate 82–2. Signed by Truman.
P.L. 81-734, 64 Stat. 477 →The one time it was nearly killed
From the Social Security Administration’s own legislative chronology, two consecutive days:
August 23, 1960 — “An amendment introduced by Senator Long to permit Federal matching of vendor payments… to public mental and tuberculosis hospitals was adopted.” · August 24, 1960 — “The Conference Committee agreed to the medical care provisions added by the Senate, except for Senator Long’s amendment.”
SSA legislative chronology, 1960
The Senate voted to end this rule, and a conference committee reversed it in a single day. Corroborated by reading the enacted statute: no provision authorising such matching survives in P.L. 86-778, only a 42-day allowance for general-hospital care. If you want one identifiable decision point in the whole history, this is it.
SSA legislative chronology, 1960 →Medicaid inherits it — and narrows it
P.L. 89-97 §121(a), 79 Stat. 351–352, carries the 1950 sentence into the new Title XIX, now bounded by age, while §1905(a)(14) affirmatively covers people 65 and over in an IMD. SSA’s own contemporaneous account describes this as removing the exclusion with respect to payments for aged individuals.
“(B) any such payments with respect to care or services for any individual who has not attained 65 years of age and who is a patient in an institution for tuberculosis or mental diseases.”
P.L. 89-97 §121(a), 79 Stat. 351–352
Sixty-one years later those words are unchanged, except that “tuberculosis or” is gone. In 1965 there was no under-21 carve-out either — the exclusion hit everyone under 65. The 21–64 band was created in 1972.
P.L. 89-97, 79 Stat. 286 →
Where the number 16 came from
A group-home rule
SSI law exempts “a publicly operated community residence which serves no more than 16 residents.” This is the earliest 16 in this corner of federal law.
The IMD rule has no number
43 FR 45204 defines an institution for mental diseases by its “overall character” — no bed count anywhere in it. The figure 16 appears in that same rule only for community residences.
16 becomes a wall
OBRA ’87 had just emptied §1905(i). A 1988 “clarification” fills it with “more than 16 beds.”
The regulation catches up
56 FR 8854 edits the rule to match: “‘an institution’ is changed to ‘a hospital, nursing facility, or other institution of more than 16 beds’.” Two years and eight months after the statute.
A hypothesis this research killed
The intuitive story — that career regulators invented 16 and Congress later rubber-stamped it — is wrong, and the paper trail runs the other way. Congress wrote the number in July 1988. HCFA copied it into the Code of Federal Regulations in March 1991, in a rule expressly titled “OBRA ’87 Conforming Amendments.” The agency was conforming to Congress, not the reverse. So the “technical correction” that carried the number was not correcting anything: there was nothing yet to correct.
Where it sits in the 1988 law- SEC. 411. TECHNICAL CORRECTIONS TO CERTAIN HEALTH CARE PROVISIONS
- (k) CORRECTIONS TO SUBTITLE B OF TITLE IV (RELATING TO MEDICAID)
- (13) TREATMENT OF EDUCATIONALLY-RELATED SERVICES
- (14) CLARIFICATION OF TERM “INSTITUTION FOR MENTAL DISEASES”
- (15) ELIGIBILITY VERIFICATION TECHNICAL CORRECTION
Sandwiched between a schools provision and a paperwork fix. The 1976 → 1988 link is a documented structural parallel, not a documented causal chain — no paper trail says Congress borrowed the SSI figure. What is established: the number did not come from the IMD regulation, and no clinical or empirical basis for it appears anywhere in the record searched.
P.L. 100-360, 102 Stat. 683 →The votes that did happen
Reconstructed from the Voteview congressional-votes database and SSA's own tallies. The House Clerk publishes no roll calls before 1990, so these are the authoritative machine-readable record.
| Year | Chamber | Measure | Tally | Dem | Rep |
|---|
| 1950 | House | Conference report, P.L. 81-734 — origin of the carve-out | 373-1 | 233-0 | 137-1 |
| 1960 | Senate | Passage, Kerr-Mills (P.L. 86-778) | 74-11 | 43-10 | 31-1 |
| 1965 | House | Passage, H.R. 6675 | 313-115 | 248-42 | 65-73 |
| 1965 | Senate | Passage, H.R. 6675 | 68-21 | 55-7 | 13-14 |
| 1972 | House | Conference report, P.L. 92-603 — psych under 21 | 305-1 | 176-0 | 129-1 |
| 1984 | House | Conference report, P.L. 98-369 | 268-155 | 192-69 | 76-86 |
| 1988 | House | Conference report, P.L. 100-360 | 328-72 | 230-9 | 98-63 |
| 1988 | Senate | Conference report, P.L. 100-360 | 86-11 | 52-0 | 34-11 |
| 2023 | House | Passage, H.R. 4531 | 386-37 | 196-11 | 190-26 |
Why none of these is "a vote for the 16-bed limit"Every row is a vote on an omnibus. A member voting yes in April 1965 was voting to create Medicare; a member voting no was voting against Medicare, not for the exclusion — the Republican substitute would not have removed it either. In sixty-one years the exclusion has never once been put to a standalone recorded vote. That is the finding, not a hole in it.
How that negative was measured
“There was no vote” is a claim about a complete set, so the set was enumerated. All 394 House and 497 Senate roll calls of the 89th Congress were pulled and filtered to H.R. 6675: exactly 16 roll calls, every description read. The House took only three recorded votes on the bill in its entire life — recommit, passage, conference report — and none on any amendment at all. Then the amendments themselves: all 513 Senate amendments were read from the official conference print. Every Title XIX provision touching institutions for mental diseases — amendments 255, 262–263, 275, 276 and 279 — is labelled a Committee amendment. Not one reached the floor.
Positive control: the same method surfaced 13 Senate roll calls on other specific 1965 provisions, including a 26–64 vote to strike Medicare Parts A and B outright, and it did flag Senator McCarthy’s psychiatric-exclusion amendment (which concerned Medicare and drew no roll call). An instrument that found those found zero for the Medicaid IMD exclusion.
Document gate: that 32-page conference print is a pure image scan with no text layer. A keyword search would have returned zero and meant nothing, so it was read page by page.
Who the record names
- Rep. Wilbur D. Mills (D-AR) — Sponsored H.R. 6675 and chaired Ways and Means, where the Medicaid title was written.
- Sen. Harry F. Byrd Sr. (D-VA) — Chaired Senate Finance and reported the bill out 12-5, while joining four Republicans in voting against it.
- Sen. Russell Long (D-LA) — The only member on record who tried to end the rule. His 1960 amendment passed the Senate and died in conference.
- Rep. Fortney "Pete" Stark (D-CA) — Sponsored the 1988 bill that carried the 16-bed paragraph. There is no evidence he wrote that paragraph.
- Rep. Brett Guthrie (R-KY) — Sponsored H.R. 4531, the 2023 bill the House passed 386-37.
And who it does not- The 1988 drafter (Author of the 16-bed paragraph) — Not identifiable in the published record.
- The 1960 conferees (Killed the Long amendment) — Not individually recorded in the chronology.
Who drafted the 16-bed paragraph in 1988 is not recoverable from the published record, and neither is any rationale for that particular figure. The conferees who struck Senator Long’s amendment in 1960 are not named in the chronology either. These are gaps in the record, not gaps in the research, and they are stated here rather than filled in.
Why it is still here
This is the part that actually answers the question. Congressional Budget Office, April 2023, Publication 59071 — net increase in federal Medicaid spending, 2024 through 2033.
$0.2–0.6B
Extend the narrow addiction option
✓ Congress passed this
$7.7B
Repeal for addiction
Passed House, died in Senate
$33.5B
Repeal for mental illness
Never voted on
$38.4B
Full repeal
Never scored in a bill
Every IMD change that has ever become law sits on the cheap side of that line. Every repeal bill sits on the expensive side, and not one has ever been reported out of committee, scored, or given a vote in any Congress. The gap between what passes and what does not is roughly 70× to 250×.
2023: the closest it has ever comeH.R. 4531 in the 118th Congress carried a section titled, verbatim, LIFTING THE IMD EXCLUSION FOR SUBSTANCE USE DISORDER. It cleared subcommittee 49–0 and full committee 29–3, then passed the House 386–37 on 12 December 2023 — Republicans 190–26, Democrats 196–11. It went to the Senate HELP Committee the next day and never moved again: no hearing, no markup, no vote, and no recorded opponent. The substance was rescued three months later inside an appropriations bill, which made the narrow 30-day addiction option permanent. The broader relief was not.
Six bills in the 119th Congress would change this rule, and all six are sitting in the same committee. Not one of them has a Senate companion — the Michelle Alyssa Go Act has now been introduced in three consecutive Congresses without ever getting one, and its cosponsor count and its bipartisanship have both gone down since the 118th. No CBO score exists for any of the six; a bill has to leave committee to get one.
So which party did this?
The voting record does not support a partisan answer. Every enactment carrying this exclusion passed with majority Republican support, under Democratic and Republican presidents alike.
- 1950, when mental institutions were first carved out: House Republicans voted 137–1 in favour.
- 1960, the one contested moment: the amendment to abolish it was killed by a conference committee, not a party. More Democrats than Republicans opposed the underlying bill.
- 1984: a Democratic House and a Republican Senate together deleted the tuberculosis half and kept the psychiatric half. Nobody in either party moved to delete both.
- 1988: the number 16 rode in on a conference report backed by 98 Republicans and 230 Democrats, and was signed by Ronald Reagan.
- 2023: the House voted to lift part of it 386–37, with overwhelming majorities of both parties. It still died.
If you are looking for someone who destroyed this, the record refuses to give you one. What it gives you is worse and more useful: a 1950 cost-allocation decision between the states and Washington that nobody has ever had to defend on a floor, because nobody has ever been made to vote on it — protected now by a $38.4 billion price tag and a constituency too sick and too poor to hire lobbyists. It survives on procedure and arithmetic, not ideology. That is why it has outlasted every majority of the last sixty-one years.
What the record does not contain
Measured absences, not assumptions. Every row was searched with a method proven to find comparable material first — the control column is what makes the zero mean something.
| Searched | For | Result | Control |
|---|
| 1935 Ways and Means report + CES report to the President | Any mention of insanity or asylums | Zero | All 10 “institution” hits found |
| 1935, 1950 and 1965 statutes + contemporaneous SSA accounts | Any stated reason for the exclusion | None at any step | Rule itself located in each |
| 43 FR 45204, first 400 pages | “16 beds” in the 1978 IMD definition | Zero | Page 45204 confirmed present in the extract |
| Every CFR amendment to the IMD definition, 1978–1996 | When the regulation acquired the number | 1991 — after the statute | Amendment chain read in full |
| Pre-1978 regulation (45 CFR 248.60) | Whether any earlier bed count existed | Not searched — no CFR before 1996 is online | — |
| All 394 House + 497 Senate roll calls, 89th Congress; all 513 Senate amendments | Any vote on the IMD exclusion | Zero | 13 other amendment roll calls found |
| MACPAC Report to Congress on IMDs (128 pages) | Any rationale for the figure 16 | 2 mentions, both descriptive | “beds” → 15 hits |
| 2025 SUPPORT Act reauthorisation, all six text versions | Any IMD or Medicaid provision | Zero | “opioid” → 36 hits |
| CRS report IF10222 · congress.gov | Congressional intent for the threshold | Unmeasured — bot wall, not bypassed | — |
One correction worth stating plainlyThe familiar story — that Congress excluded asylums because President Pierce established in 1854 that they are a state responsibility — is a reconstruction by later scholars. Pierce’s veto is real and says what it is quoted as saying. But no 1935, 1950 or 1965 primary document reached here connects the two, and the 1935 record actively cuts against it: those drafters were legislating about almshouses and never mentioned insanity at all.
Caveats. Stated so you can discount them yourself. Vote sourcing: GovTrack draws its pre-1990 roll calls from Voteview, so they are one lineage, not two. The 1965 totals and the two conference-report party splits are independently confirmed by SSA’s official tally page; the 1965 passage-vote party splits rest on the Voteview lineage alone. Voteview’s member-level file and its own summary field also disagree by one vote on several 1965 and 1972 roll calls — official tallies are shown throughout. Bill status: congress.gov is bot-walled, so this came from GovInfo BILLSTATUS, the same Library of Congress feed, with as-of dates between November 2025 and June 2026; very recent committee action would not appear. Not established: no document connects the 1976 SSI 16-resident figure to the 1988 Medicaid number — that remains a hypothesis. The pre-1978 regulation was never searched, because no CFR earlier than 1996 is online; nor were the 1966 Handbook of Public Assistance Administration, the 1975–76 field instructions, or the 1982 State Medicaid Manual, each of which MACPAC records as carrying an IMD definition. So “16 had no regulatory predecessor” is proven for 1978–1991 and unmeasured before that. Sponsorship of the 1950 and 1972 bills is single-sourced.
The same research as a documentEverything above is also published as a 15-page report with the statutory text set as it appears in the Statutes at Large. Download the PDF (806 KB). The page you are reading is the whole of it — the PDF is for citing and printing, not for the parts the page left out.