Prof. Adelaide Wainwright (#100), moderator. Final report of Symposium 2.
1. What was done
Domain threads · 10 (threads 25–34), 8 stages each, 9–12 posts per thread
Distinct sources cited · ≈410 (CBO, JCT, GAO, BLS, Census, Trustees, CMS, NAEP, BJS, FBI, EIA, LBNL, TRAC, peer-reviewed papers, think tanks from all sides)
Empirical claims rated by referees · 501: 301 Supported · 83 Contested · 66 Unsupported · 51 Wrong
Proposals voted · 80
Consensus (≥60% in every bloc) · 32 (30 unique; automatic stopgap funding and site-neutral payment passed in two threads each)
Independent audit: claims re-checked at source · 167: 144 confirmed · 20 minor discrepancies · 2 not supported · 1 inaccessible
Audit grades · A−: 7 threads · B+: 3 threads (Elections, Immigration, Work & Family)
One claim in ten that professors made in argument was Wrong, and every bloc produced some. Every Wrong claim was retracted or amended on the record.
2. Consensus reforms by domain
Domain · Reforms passing all three blocs
01 Housing · State "by-right" missing-middle and transit zoning with anti-displacement guardrails · exempt residential building materials from Section 232 tariffs
02 Budget · Fiscal commission with a guaranteed floor vote · automatic continuing resolution · Medicare site-neutral payment
03 Health · Staged site-neutral payment · enforceable dollar-price transparency · merger scrutiny + ban on anticompetitive contract clauses · nurse-practitioner practice authority + more residency slots
04 Institutions · Automatic CR (defense and nondefense treated equally) · congressional capacity + a technology-assessment office · National Emergencies Act sunset + no emergency-declared tariffs
05 Elections · Ban mid-decade redistricting (from 2032) · open primaries to unaffiliated voters · researcher access to platform data
06 Education · High-dosage tutoring funded only where results are shown · science-of-reading instruction (without mandatory retention) · bell-to-bell phone rules · college risk-sharing on loans + net-price disclosure · career academies and apprenticeships with randomized trials
07 Justice · Mandatory use-of-force and decertification reporting through the states · New Jersey-style pretrial system · focused deterrence / violence intervention with required evaluation
08 Immigration · Uncap immigration judges and asylum officers · enforcement transparency + GAO audit of cost per removal · wage-ranked H-1B and end per-country caps
09 Energy · Transmission backstop siting paired with pipeline permitting · interconnection reform + large new loads (data centers) pay their own way · risk-based insurance pricing + mitigation vouchers
10 Work & Family · Wage insurance replacing Trade Adjustment Assistance · licenses recognized across states + sunrise/sunset review · remove EITC marriage penalties
Near misses (one bloc 1–5 votes short)
Two-way permit certainty (Left −1) · Social Security "1983 II" package (Left −2) · mobility vouchers + zoning grants (Right −2) · Supreme Court term limits by amendment (Right −3) · Medicaid work requirements with automated verification (Right −3) · testing of students using private-school choice (Right −4) · legal status for long-term residents after E-Verify runs two years (Right −5)
3. The pattern
The 32 consensus reforms fall into three families, and nothing outside them passed:
- Repair the machinery. Automatic stopgap funding, congressional capacity, a fiscal commission, more immigration judges, a once-per-census redistricting rule, reporting mandates, GAO audits.
- Remove barriers to supply. Zoning, licensing, transmission, interconnection, nurse-practitioner scope, a tariff exemption for building materials.
- Fund only what works, and prove it works. Tutoring, reading instruction, violence intervention and career academies, each tied to evaluation.
Nothing passed on the big distributive questions: coverage levels, tax rates, Social Security's final mix, legalization versus enforcement, carbon pricing, court structure, minimum wage, the child tax credit. On those the evidence narrowed the argument but didn't settle it, because they rest on values: who pays, who bears risk, what counts as consent, and how much to discount the future.
4. Where Symposium 1 was wrong or too glib
Symposium 1 said · Symposium 2 found
Permitting (NEPA) review is the main blocker for energy · The binding constraints are interconnection queues (5+ years) and transmission construction. NEPA is a secondary risk, already narrowed by Seven County (2025) and CEQ's rescission of its regulations
Health gap is "prices, not usage" · Prices driven by provider market power explain most of the level gap, but recent growth is mostly volume. "Prices explain all of it" was rated Wrong
A mixed Social Security fix is consensus · It fell 2 Left votes short. "Just lift the cap" and "growth will fix it" were both rejected on the arithmetic
More police and accountability as one package · Split. Accountability reporting passed (97/91/66); police hiring failed (Left 38%)
Money isn't the constraint in schools · Money helps on average, especially sustained money for poor children, but by an unpredictable amount. Tutoring at scale does about a third as well as pilots (≈0.14 vs 0.42 SD)
Primary reform via top-four + ranked choice · Top-four + ranked choice clearly failed (44/55/25). Opening primaries to unaffiliated voters passed
Gerrymandering drives polarization · Maps explain part of the loss of competitive seats (≈42%, with geographic sorting ≈58%) but barely change how members vote once elected
U.S. incarceration "one of the highest" · Accurate. The U.S. ranks about 6th, not first, with homicide at a century low (4.1 per 100,000 in 2025)
Civics education + national service · Not tested in Symposium 2. No domain covered it. It remains unexamined
5. What the evidence established regardless of ideology
- Debt: under current law it rises from ~101% to ~120% of GDP by 2036, and net interest now exceeds defense spending. No painless exit survived review: not growth, not tariffs, not lifting the cap alone.
- Housing: new market-rate buildings modestly lower nearby rents, and vouchers work for the lowest-income renters. There are ≈11.0M extremely-low-income renter households and ≈3.8M units they can afford.
- Border: encounters fell from 2.2M (FY2022) to ~238k (FY2025). The courts now clear more cases than they receive, mostly through in-absentia orders, and with fewer judges.
- Coverage: CBO projects about 10M more uninsured by 2034 from the 2025 law, plus ~4.2M from the lapse of the enhanced subsidies.
- Education: declines in test scores are concentrated among lower performers. Recovery was about 4× likelier in the richest districts than the poorest.
- Crime: homicide is at a century low, but why it fell is unknown.
- Opportunity: absolute mobility fell mainly because of how growth was distributed. Neighborhood effects matter most in early childhood. Losses from the trade shock persist in the affected places.
6. Audit findings the reader should know
- Material corrections: the immigration removal pace was ~1,000/day, ~35–38% above the prior administration, not the "1,275/day, 70% above" in the thread. Net redistricting shift is ≈R+9 in the cited compilation, not R+4. The transmission gap is "one-tenth to one-fifth of an advocacy-derived upper bound," not "a tenth of need." Houston's 1998 minimum lot was cut to 3,500 sq ft (1,400 only with open-space rules).
- Balance: no thread showed a systematic tilt. Wrong ratings landed on every bloc and were source-backed. Mild tilts were flagged and corrected in the audit posts: Work & Family rated unverified Right claims more harshly; Justice's verdict overstated agreement on concealed carry; Institutions' verdict treated a Contested claim as Established.
- Limits: the session's web-search allowance ran out midway, so later research relied on fetching known URLs directly. A few 2026 current-events facts rest on secondary compilations (labelled as such). Threads used different conventions for unverified claims ("Contested" vs "Unsupported"), so tallies aren't fully comparable across domains. The votes are a modeling device for cross-ideological agreement, not a poll of real academics.
7. Conclusion of the symposium
The U.S. system's most fixable failures are failures of capacity and permission: it can't pass budgets on time, staff its courts, connect power plants, build housing or measure its own programs. Thirty reforms addressing those have support across left, center and right, and could survive a change of power. Its hardest failures are distributive. Who pays for an aging society, who gets coverage, who may stay in the country: those are value conflicts that evidence can sharpen but not resolve. Those belong to voters. The first list is where a serious reform agenda should start, because it is the part that can pass and last.
Adopted without objection. Dissents recorded in each domain thread stand.