Lintel
File evidence-register.csv Rows 231 Maintained One claim, one ID, one row Rule Nothing enters a document that is not in the register Corrections log What the world publishes What two states publish Research Research documents Download the CSV

The spine of the project

The evidenceregister.

Every claim on this site has a row here. Thirty-two rows are gaps nobody has measured — they are the research agenda.

Seventy-two rows read Untested, and they are two different things.

Thirty-two are gaps nobody has measured anywhere; those are deliberate, and they are the agenda. The other forty rows whose citation named no document were regraded in August 2026 — they cited a publisher rather than a publication, so nobody could open the source or check the row against it, and the grade asserted nothing. That is a filing failure, not a research gap, and it is recorded rather than repaired quietly: see C-017. Nothing in that regrade says any of those claims is false. Each one carries the grade it lost and what it would take to earn it back.

Every row, in register order — one cell is one claim

Confirmed 109 Supported 45 Contested 5 Untested 72 The solid amber cells are the research agenda. Hover any cell for its row.

Confirmedindependent published evidence supports the claim Supportedevidence consistent; mechanism contested or partial Contestedcredible published evidence points in more than one direction Untestednobody has measured this — open research question
231 OF 231 ROWS
Value Source Retrieved Used in Notes
E-001 Regulation share of new single-family home price $131,734 / 26.4% of $499,500 avg Contested NAHB, Government Regulation in the Price of a New Home 2026-06 2026-08 HAP-02 S05/S13 2027-06 Two self-report surveys of NAHB’s own members: 54 of 2,125 land developers (~2.5% response) plus 337 builders via HMI special questions. Sole source, trade association arguing for deregulation. PROVENANCE: the figure is published on NAHB's own advocacy blueprint, by the party it benefits, and NAHB's separate member survey (E-099) does not list approval duration among builders' serious problems at all. The two NAHB instruments disagree about what constrains building. Regraded Contested on 2026-08-10 (C-013) because evidence pointing more than one way is what Contested means, and this row's provenance now belongs on the row rather than only in the prose around it.
E-002 Single-family authorization to completion, as two published averages 8.8 months for 2025, on the like-for-like Total 1-unit columns: 1.4 months authorization to start plus 7.4 months start to completion. Census has resumed the average tables and now publishes 2023, 2024 and 2025 (start to completion 8.6, 7.7, 7.4). The 9.7 figure this row previously carried was the 2022 vintage Supported US Census Bureau, Survey of Construction: "Average Number of Months from Authorization to Start of Buildings Started in Permit-Issuing Places" and "Average Number of Months from Start to Completion", both carrying annual data 1971-2025 2026-08 2026-08-17 HAP-02 S09; POSTER 2027-02 Corrected under C-016 from "9.1 months (1.4 to start + 7.6 building), 2024", which was wrong three ways. The 1.4 was the Total 1-unit column and the 7.6 was the Built-for-sale column of the other table, so the sum mixed two populations; the stated 9.1 equals neither like-for-like sum (Total+Total 9.7, built-for-sale+built-for-sale 9.0); and the series carries no 2024 figure at all. Graded Supported rather than Confirmed because the SUM is derived: the two averages are published on different cohorts — authorization-to-start covers units STARTED in the year, start-to-completion covers units COMPLETED in the year — so 9.7 describes no single population of houses. Each component is Confirmed; the addition is arithmetic across two cohorts and must be presented as such. The "eight months of building" the site quotes rounds from the 8.3 half and survives this correction; see E-114 for what 2024 actually publishes. UPDATED under C-038: the claim this row used to make about the years Census had published was true when written and is no longer true. The average tables now carry 2023, 2024 and 2025. The sum remains a cross-cohort addition and the grade stays Supported for that reason, not because of the vintage. Verified 2026-08-17 by parsing the source workbook directly: 2025 Total 1-unit start to completion 7.4, authorization to start 1.4.
E-003 Start to completion, buildings with 2 or more units 17.0 months for 2022, the latest year published Confirmed US Census Bureau, Survey of Construction: "Average Number of Months from Authorization to Start of Buildings Started in Permit-Issuing Places" and "Average Number of Months from Start to Completion", both last updated 2024-03-19 and carrying data through 2022 2024-03 2026-08-10 HAP-02 S09 2027-02 Corrected under C-016 from "16.5 months (2025)". 16.5 appears in no year of this series and no 2025 average is published — the last published year is 2022. The neighbouring years are 15.4 for 2019 through 2021 and 17.0 for 2022, so the old figure was not a stale reading of an earlier year either. Cohort: units completed in the year.
E-004 Subdivision entitlement duration GAP 6–18 months typical; 18–36 acquisition-to-entitled complex Untested Practitioner sources (multiple) 2026 2026-08 HAP-02 S09; POSTER 2027-02 CITATION UNRESOLVED (C-017) — regraded from Supported on 2026-08-10 because the source field names no document a reader can open. This is a statement about the grade, not about the claim: nothing here says the claim is false, only that it was never checkable. Restore the grade by naming the document — publication title, section, DOI or URL — and re-reading it. Practitioner estimates not a national dataset. See E-038.
E-005 California subdivision total duration GAP 3 to 7+ years common Untested Practitioner sources (CA) 2025-12 2026-08 HAP-02 S04/S09 2027-02 CITATION UNRESOLVED (C-017) — regraded from Supported on 2026-08-10 because the source field names no document a reader can open. This is a statement about the grade, not about the claim: nothing here says the claim is false, only that it was never checkable. Restore the grade by naming the document — publication title, section, DOI or URL — and re-reading it. Regional; not generalizable.
E-006 EIR requirement added to approval time in Los Angeles 484 days average Confirmed Gabriel & Kung, "Development approval times and new housing supply: Evidence from Los Angeles", Journal of Urban Economics 2026 2026-08 SITE landing; POSTER 2027-06 CORRECTED from "504 days" — see C-009. 504 came from the December 2023 working paper, whose title page is described in C-009 as carrying a preliminary do-not-cite stamp. The published Journal of Urban Economics version gives 484 days, using column (2) of Table 3 as the preferred specification, and the string "504" does not appear in it. HAP-02 r3.7 is closed and still carries 504; it is not edited, because a citation must resolve to what it cited. QUOTATION MARKS REMOVED 2026-08-10 under C-021: the stamp wording was quoted verbatim, including its typo, but the working paper has not been retrieved in this cycle and the exact string is unverified. The substance stands — the published version gives 484 and does not contain 504 — and the wording returns to quotation only when someone opens the paper.
E-007 By-right vs discretionary approval speed and variance GAP 28% faster with lower variance Untested LA Transit-Oriented Communities study, ~350 projects 2018–2020 2024 2026-08 HAP-02 S04/S10; Red Team 01 2027-06 CITATION UNRESOLVED (C-017) — regraded from Confirmed on 2026-08-10 because the source field names no document a reader can open. This is a statement about the grade, not about the claim: nothing here says the claim is false, only that it was never checkable. Restore the grade by naming the document — publication title, section, DOI or URL — and re-reading it. Pre-empts our variance claim — cite rather than claim.
E-008 Median multifamily entitlement Oakland vs San Francisco GAP 6 months vs 25+ months Untested Pew Charitable Trusts 2024 2026-08 HAP-02 S09/S13 2027-06 CITATION UNRESOLVED (C-017) — regraded from Confirmed on 2026-08-10 because the source field names no document a reader can open. This is a statement about the grade, not about the claim: nothing here says the claim is false, only that it was never checkable. Restore the grade by naming the document — publication title, section, DOI or URL — and re-reading it. Also: >80% of proposed CA multifamily required entitlement.
E-009 Manufactured-home purchase applications that result in an origination, against site-built 27% of manufactured-home purchase applications result in an origination against 74% for site-built (2019 HMDA). This is an ORIGINATION rate, not an approval rate: the report defines approval rate separately at n.32 as originations plus approved-not-accepted over that plus denials, and reports it only by credit band Confirmed CFPB, "Manufactured Housing Finance: New Insights from the Home Mortgage Disclosure Act Data", May 2021 (2019 HMDA data) 2021-05 2026-08 HAP-02 S08/S10; HAP-08 lending-penalty 2027-08 Core evidence for the appraisal lock. Corrected twice. On 2026-08-10 the value was corrected from the site's rounded "<30% vs >70%" to the report's own figures (C-013) — the digits were right and the noun was not. The rounded pair was a fair rounding of these same origination figures, so that was a de-rounding rather than a substitution; what survived it was the word "approval", which names a different measure the same report defines and publishes. Corrected again under C-014, which also split this row: the denial comparison is E-112 and the credit-banded approval rates are E-113. One row carrying three measures under one noun is how the noun stayed wrong through a correction.
E-010 Chattel refinance share vs site-built (2019 rate lows) GAP <4% vs 44% Untested CFPB 2021 2026-08 HAP-02 S08/S10; HAP-08 lending-penalty 2027-06 CITATION UNRESOLVED (C-017) — regraded from Confirmed on 2026-08-10 because the source field names no document a reader can open. This is a statement about the grade, not about the claim: nothing here says the claim is false, only that it was never checkable. Restore the grade by naming the document — publication title, section, DOI or URL — and re-reading it. MH mortgages were 31%.
E-011 Interest rate spread manufactured vs general lenders GAP 5.61% vs 1.20% Untested Urban Institute 2026-08 HAP-02 S08/S10; HAP-08 lending-penalty 2027-06 CITATION UNRESOLVED (C-017) — regraded from Confirmed on 2026-08-10 because the source field names no document a reader can open. This is a statement about the grade, not about the claim: nothing here says the claim is false, only that it was never checkable. Restore the grade by naming the document — publication title, section, DOI or URL — and re-reading it.
E-012 Manufactured-home shipments, peak to collapse 373,143 shipped in 1998, the peak. 130,748 in 2004 — a 65% fall reached BEFORE the financial crisis, while site-built housing was mid-boom. 49,717 in 2009, the trough, 87% below peak. Recovery since: 105,772 (2021), 112,882 (2022), 89,169 (2023), 103,314 (2024), 102,738 (2025) — back above 100,000, and still under a third of 1998 Confirmed US Census Bureau, Manufactured Housing Survey, "Annual Totals of Shipments to States: 1994-2023" (extended through 2026), US total row, https://www2.census.gov/programs-surveys/mhs/tables/time-series/annual_shipmentstostates.xlsx 2026 2026-08-10 HAP-02 S03/S10 2027-08 CORRECTED AGAIN under C-024, and the second correction is the instructive one. The first pass (C-022) derived this series by scraping the MONTHLY workbooks and summing them. Those files lay two year-blocks side by side, the parse mis-assigned years by one block, and three attempts produced three different answers — including "not above 100,000 since", which is false, and a 1977-1995 mean computed over a range missing eight of its years. This row now reads the ANNUAL shipments-to-states table, which carries an explicit US total row, so no reconstruction is involved. The peak, trough and 2020 figures agree with the monthly files to the hundred, which is what makes the rest of it trustworthy. THE LESSON: a derived series is a claim. Find the table that states the number instead of assembling one from parts. Census still warns that data before 2014 uses a different methodology and is not comparable to the current series; 1998, 2004 and 2009 all sit before that break, so the peak-to-trough and boom-era comparisons are internally clean. 2026 is excluded as a partial year.
E-013 Share of MH purchase loans that are chattel GAP 42% (2019 HMDA) Untested CFPB 2021 2026-08 2027-06 CITATION UNRESOLVED (C-017) — regraded from Confirmed on 2026-08-10 because the source field names no document a reader can open. This is a statement about the grade, not about the claim: nothing here says the claim is false, only that it was never checkable. Restore the grade by naming the document — publication title, section, DOI or URL — and re-reading it. Unused — available for expansion.
E-014 MH borrowers who own their land yet take chattel GAP 60% own land; 17% still chattel Untested CFPB 2021 2026-08 HAP-02 S08; HAP-08 lending-penalty 2027-06 CITATION UNRESOLVED (C-017) — regraded from Confirmed on 2026-08-10 because the source field names no document a reader can open. This is a statement about the grade, not about the claim: nothing here says the claim is false, only that it was never checkable. Restore the grade by naming the document — publication title, section, DOI or URL — and re-reading it. Complicates simple legal-classification story. Channel matters.
E-015 Construction productivity Kuznets curve Stagnant 1900–40, boom 1940–70, collapse after 1970; homes/worker at times outgrew cars/worker Confirmed D'Amico, Glaeser, Gyourko, Kerr & Ponzetto, NBER 33188 2024 2026-08 HAP-02 S10 2027-06 Best single fact for the assembly-line framing.
E-016 Productivity gap from firm size distribution ~60% more productive with manufacturing's size distribution Confirmed NBER 33188 2024 2026-08 HAP-02 S04/S10; Red Team 06 2027-06 Competing diagnosis — unresolved against our thesis.
E-017 Entry vs project regulation effect on firm size Opposite directions Confirmed NBER 33188 2024 2026-08 Red Team 06 2027-06 Implies type certification behaves as entry regulation.
E-018 Measured construction productivity decline Real but partly deflator artifact Contested Goolsbee & Syverson NBER 30845; Garcia & Molloy; Richmond Fed 2023–25 2026-08 HAP-02 S13 2027-06 Live dispute. Must be engaged not assumed.
E-019 First successful automated code compliance system GAP CORENET — proposed 1997 Untested BCA/URA Singapore; published analyses 2021 2026-08 HAP-02 S01/S03/S05 2027-06 CITATION UNRESOLVED (C-017) — regraded from Confirmed on 2026-08-10 because the source field names no document a reader can open. This is a statement about the grade, not about the claim: nothing here says the claim is false, only that it was never checkable. Restore the grade by naming the document — publication title, section, DOI or URL — and re-reading it.
E-020 CORENET X timeline GAP Initiated 2018, soft launch Dec 2023, mandatory Oct 2026 Untested BCA Singapore 2024–26 2026-08 HAP-02 S01/S03 2026-11 CITATION UNRESOLVED (C-017) — regraded from Confirmed on 2026-08-10 because the source field names no document a reader can open. This is a statement about the grade, not about the claim: nothing here says the claim is false, only that it was never checkable. Restore the grade by naming the document — publication title, section, DOI or URL — and re-reading it. Mandatory date is imminent — verify it took effect.
E-021 CORENET X multi-agency review GAP 8 agencies concurrent (BCA, URA, LTA, NEA, NParks, PUB, SCDF, SLA); consolidated response within 20 working days — a service-level maximum, not measured turnaround Untested BCA Singapore 2024 2026-08 HAP-02 S01/S03/S10 2027-06 CITATION UNRESOLVED (C-017) — regraded from Supported on 2026-08-10 because the source field names no document a reader can open. This is a statement about the grade, not about the claim: nothing here says the claim is false, only that it was never checkable. Restore the grade by naming the document — publication title, section, DOI or URL — and re-reading it. In production since Dec 2023. No achieved-turnaround figures are published; a Design Advisory Panel route extends the gateway to 30 working days. Rollout: mandatory Oct 2025 for projects ≥30,000 m², Oct 2026 all new projects, Oct 2027 ongoing projects.
E-022 Shared national data schema GAP IFC+SG (extends IFC4) Untested BCA Singapore 2024 2026-08 HAP-02 S01/S03 2027-06 CITATION UNRESOLVED (C-017) — regraded from Confirmed on 2026-08-10 because the source field names no document a reader can open. This is a statement about the grade, not about the claim: nothing here says the claim is false, only that it was never checkable. Restore the grade by naming the document — publication title, section, DOI or URL — and re-reading it. Production proof for L3.
E-023 Scaling limit of automated code checking GAP Rules hand-encoded; users cannot author new rules Untested Published analyses of CORENET/BIM Assure/Solibri/EDM 2021 2026-08 HAP-02 S01/S03/S05 2027-06 CITATION UNRESOLVED (C-017) — regraded from Confirmed on 2026-08-10 because the source field names no document a reader can open. This is a statement about the grade, not about the claim: nothing here says the claim is false, only that it was never checkable. Restore the grade by naming the document — publication title, section, DOI or URL — and re-reading it. The load-bearing citation for the whole threshold argument.
E-024 US building code edition fragmentation GAP 2024 adoptions spanned 2012–2021 editions; IN/TN/TX on 2012 Untested ICC adoption records; adoption-map analyses 2024–26 2026-08 HAP-02 S01/S10 2027-02 CITATION UNRESOLVED (C-017) — regraded from Confirmed on 2026-08-10 because the source field names no document a reader can open. This is a statement about the grade, not about the claim: nothing here says the claim is false, only that it was never checkable. Restore the grade by naming the document — publication title, section, DOI or URL — and re-reading it. Verify current-year state before publication.
E-025 State amendment practice GAP Very few adopt as-is; State Amended and State Edition tiers Untested ICC; state code databases 2026 2026-08 HAP-02 S01 2027-06 CITATION UNRESOLVED (C-017) — regraded from Confirmed on 2026-08-10 because the source field names no document a reader can open. This is a statement about the grade, not about the claim: nothing here says the claim is false, only that it was never checkable. Restore the grade by naming the document — publication title, section, DOI or URL — and re-reading it.
E-026 Cost of code changes since 2012 GAP 11% of new apartment cost Untested 2022 trade survey via Cato Institute 2024 2026-08 HAP-02 S02 2027-06 CITATION UNRESOLVED (C-017) — regraded from Supported on 2026-08-10 because the source field names no document a reader can open. This is a statement about the grade, not about the claim: nothing here says the claim is false, only that it was never checkable. Restore the grade by naming the document — publication title, section, DOI or URL — and re-reading it. Trade survey — advocacy-adjacent. Verify primary.
E-027 Arizona ADU preemption GAP HB 2720 (2024) municipalities; HB 2928 counties eff. 2026-01-01; automatic loss of local control on missed deadline Untested Arizona statute 2024–26 2026-08 HAP-02 S04/S10 2027-06 CITATION UNRESOLVED (C-017) — regraded from Confirmed on 2026-08-10 because the source field names no document a reader can open. This is a statement about the grade, not about the claim: nothing here says the claim is false, only that it was never checkable. Restore the grade by naming the document — publication title, section, DOI or URL — and re-reading it. The live natural experiment.
E-028 Arizona Starter Homes Act GAP Vetoed 2024; stalled subsequently Untested Legislative record 2024–26 2026-08 HAP-02 S04; Red Team 08 2027-02 CITATION UNRESOLVED (C-017) — regraded from Confirmed on 2026-08-10 because the source field names no document a reader can open. This is a statement about the grade, not about the claim: nothing here says the claim is false, only that it was never checkable. Restore the grade by naming the document — publication title, section, DOI or URL — and re-reading it. Contrast with E-027 predicts where pilots survive.
E-029 Auckland Unitary Plan scope GAP ~75% of residential land upzoned; min lot sizes abolished; ~300% capacity increase Untested Greenaway-McGrevy et al. 2021–25 2026-08 HAP-02 S04 2027-06 CITATION UNRESOLVED (C-017) — regraded from Confirmed on 2026-08-10 because the source field names no document a reader can open. This is a statement about the grade, not about the claim: nothing here says the claim is false, only that it was never checkable. Restore the grade by naming the document — publication title, section, DOI or URL — and re-reading it.
E-030 Auckland consent increase +21,808 consents at 5yr (~4.11% of stock); 52,200 of 112,300 permits attributable over 7yr (46%) Confirmed Greenaway-McGrevy & Phillips (JUE 136, 2023); Greenaway-McGrevy (Economic Modelling 160, 2026) 2023–26 2026-08 HAP-02 S04/S10 2027-06 Two methods, two magnitudes — cite both. The published 7yr figure supersedes the +43,500-at-6yr preprint this row quoted until Aug 2026.
E-031 Auckland rent effect 23% below counterfactual at 8yr (synthetic control) Supported Greenaway-McGrevy & So, Economic Inquiry, DOI 10.1111/ecin.70075 2026-07 2026-08 HAP-02 S04/S10; Red Team 01 2027-06 Strongest rebuttal to full land capitalization. Supported not Confirmed: single research group, and the estimate moved 22–35% → 26–33% → 28% → 23% across versions — quote the published 23% only. The 6yr “28%” (WP016) states the counterfactual base: rents would be 28% higher without reform, i.e. actual ~21.9% below — dropped. “Lower Hutt 21%” untraceable to any primary paper (that study reports construction effects, no rent effect) — removed.
E-032 Japanese national zoning 12–13 national zones; residential in all but exclusively industrial; nuisance ceilings; as-of-right permitting Confirmed Japan City Planning Law 1968; Building Standards Law 1968–2018 2026-08 HAP-02 S04/S10 2027-06 Municipalities may set stricter limits within national ranges.
E-033 Japanese housing depreciation GAP Depreciates to near zero; ~30-year rebuild cycle Untested Multiple secondary sources 2016–26 2026-08 HAP-02 S04/S10 2027-02 CITATION UNRESOLVED (C-017) — regraded from Supported on 2026-08-10 because the source field names no document a reader can open. This is a statement about the grade, not about the claim: nothing here says the claim is false, only that it was never checkable. Restore the grade by naming the document — publication title, section, DOI or URL — and re-reading it. Causal direction argued not established. Worth a dedicated piece.
E-034 Vienna social housing share GAP ~60% of residents in public or limited-profit; 7.4% unrestricted for-profit rental; new LPH rents ~27% below private; ~21% of income on rent Untested City of Vienna; CCI; EIB 2024–25 2026-08 HAP-02 S04 2027-06 CITATION UNRESOLVED (C-017) — regraded from Confirmed on 2026-08-10 because the source field names no document a reader can open. This is a statement about the grade, not about the claim: nothing here says the claim is false, only that it was never checkable. Restore the grade by naming the document — publication title, section, DOI or URL — and re-reading it.
E-035 Critique of Vienna model Marginal vs German cities once tax/utilities counted; ~17% newcomer premium; minimal means-testing Contested American Enterprise Institute 2025 2026-08 HAP-02 S04 2027-06 Engage not dismiss. Advocacy source on the other side.
E-036 Phoenix metro ADU cost GAP $120,000–$350,000+; $150–400/sqft Untested Regional builder sources 2025–26 2026-08 HAP-08 the-instrument 2027-02 CITATION UNRESOLVED (C-017) — regraded from Supported on 2026-08-10 because the source field names no document a reader can open. This is a statement about the grade, not about the claim: nothing here says the claim is false, only that it was never checkable. Restore the grade by naming the document — publication title, section, DOI or URL — and re-reading it. Vs ~$200/sqft national new-home average. Regional builder sources rather than primary budgets — replace with real project costs. The only concrete price of a built thing anywhere in the argument.
E-037 Arizona ADU permit volume since preemption GAP NO CITY PUBLISHES A SERIES — three cities’ counts were reported secondhand in August 2025 (E-210); Tempe publishes a dashboard whose figures could not be read Untested 2026-08-17 HAP-02 S10; Gate 1 GAP. Records requests. Fills via Gate 1. Value restated under C-048: the earlier NOT PUBLISHED was wrong as written — no city publishes the series, which is not the same as no figure existing. Three cities’ counts reached print in a newspaper a year before this row was checked, and reading is cheaper than a records request. The grade does not move: three cities in three different units, obtained by telephone, do not measure throughput.
E-038 Full idea-to-sale cycle time GAP NEVER MEASURED END TO END Untested 2026-08 HAP-02 S10; POSTER caveat GAP. Census starts at authorization. No dataset spans site ID to closing.
E-039 Objective vs discretionary split of US zoning GAP NEVER MEASURED Untested 2026-08 HAP-02 S10; Gate 0 GAP. Most load-bearing untested claim. Fills via Gate 0.
E-040 Symbium — computational law GAP Stanford-origin; statewide CA ADU coverage; address-in envelope-out; design marketplace; government portals Untested Symbium; GovTech coverage 2021–26 2026-08 HAP-06 S03 2027-02 CITATION UNRESOLVED (C-017) — regraded from Confirmed on 2026-08-10 because the source field names no document a reader can open. This is a statement about the grade, not about the claim: nothing here says the claim is false, only that it was never checkable. Restore the grade by naming the document — publication title, section, DOI or URL — and re-reading it. Refutes HAP-02 Red Team 04 as written.
E-041 Buildability — parcel screening scale GAP Claimed 50 states / 3,100+ counties / 485+ enhanced jurisdictions / 150M+ parcels; 20s vs $12k feasibility study Untested Buildability marketing 2026 2026-08 HAP-06 S03 2027-02 CITATION UNRESOLVED (C-017) — regraded from Supported on 2026-08-10 because the source field names no document a reader can open. This is a statement about the grade, not about the claim: nothing here says the claim is false, only that it was never checkable. Restore the grade by naming the document — publication title, section, DOI or URL — and re-reading it. Vendor claim — unverified independently.
E-042 Canibuild — builder-facing compliance GAP Site fit + automated compliance incl. topography/overlays/hazard Untested Canibuild; trade coverage 2021 2026-08 HAP-06 S03 2027-02 CITATION UNRESOLVED (C-017) — regraded from Confirmed on 2026-08-10 because the source field names no document a reader can open. This is a statement about the grade, not about the claim: nothing here says the claim is false, only that it was never checkable. Restore the grade by naming the document — publication title, section, DOI or URL — and re-reading it.
E-043 Housing as electoral issue GAP Top issue for 18–34 ahead of 2026 midterms Untested CNBC All-America Economic Survey 2026 2026-08 HAP-02 S13 2026-11 CITATION UNRESOLVED (C-017) — regraded from Confirmed on 2026-08-10 because the source field names no document a reader can open. This is a statement about the grade, not about the claim: nothing here says the claim is false, only that it was never checkable. Restore the grade by naming the document — publication title, section, DOI or URL — and re-reading it. Time-sensitive — expires after midterms.
E-044 Congressional housing caucus GAP Bipartisan; founding members include AZ-06 Untested Public reporting 2025–26 2026-08 HAP-02 S13; SITE about 2027-02 CITATION UNRESOLVED (C-017) — regraded from Confirmed on 2026-08-10 because the source field names no document a reader can open. This is a statement about the grade, not about the claim: nothing here says the claim is false, only that it was never checkable. Restore the grade by naming the document — publication title, section, DOI or URL — and re-reading it. Reachable contact for AZ-based work.
E-045 Ivory Prize Up to $300,000 a year across three focus areas; self-nomination permitted and free; the posted cycle opens September 2026 and closes 30 October 2026 at noon MST, not December Confirmed Ivory Innovations, "Ivory Prize for Housing Affordability" cycle timeline and Terms and Conditions — https://www.ivoryinnovations.org/ivory-prize and https://www.ivoryinnovations.org/ivory-prize/terms 2026 2026-08-17 SITE about 2026-10 Regraded and corrected 2026-08-17 — see C-041. The row said nominations run Sept-Dec. The Terms do say that, but the POSTED TIMELINE for this cycle closes 30 October 2026 at noon MST, and the prior cycle closed 31 October 2025. Acting on December would have missed it by two months. Nominations are not yet open; the form at /ivory-prize-nominate reads 'Content coming soon'. Review date set to 2026-10 because this row expires.
E-046 HUD Advancing Building Technology GAP $10M total; $500k–$1.5M awards; ~15 agreements; individuals and sole proprietorships ineligible Untested HUD NOFO 2026 2026-08 SITE about 2027-02 CITATION UNRESOLVED (C-017) — regraded from Confirmed on 2026-08-10 because the source field names no document a reader can open. This is a statement about the grade, not about the claim: nothing here says the claim is false, only that it was never checkable. Restore the grade by naming the document — publication title, section, DOI or URL — and re-reading it. Requires an entity. 2028+ target.
E-047 State industrialized building preemption GAP Georgia Act preempts local authority over approved buildings; Commissioner determination final Untested Georgia statute; equivalents in MD/PA/TX/CA/NY 2026-08 HAP-02 S03/S13 2027-06 CITATION UNRESOLVED (C-017) — regraded from Confirmed on 2026-08-10 because the source field names no document a reader can open. This is a statement about the grade, not about the claim: nothing here says the claim is false, only that it was never checkable. Restore the grade by naming the document — publication title, section, DOI or URL — and re-reading it. Existing legal precedent for L2.
E-048 Interstate reciprocity GAP IIBC streamlined certification recognized across state lines Untested Interstate Industrialized Buildings Commission 2026-08 HAP-02 S13 2027-06 CITATION UNRESOLVED (C-017) — regraded from Confirmed on 2026-08-10 because the source field names no document a reader can open. This is a statement about the grade, not about the claim: nothing here says the claim is false, only that it was never checkable. Restore the grade by naming the document — publication title, section, DOI or URL — and re-reading it.
E-049 Scope limit of modular preemption GAP Local retains zoning/site review/utilities/foundation/final inspection Untested State program documentation 2026-08 HAP-02 S03 2027-06 CITATION UNRESOLVED (C-017) — regraded from Confirmed on 2026-08-10 because the source field names no document a reader can open. This is a statement about the grade, not about the claim: nothing here says the claim is false, only that it was never checkable. Restore the grade by naming the document — publication title, section, DOI or URL — and re-reading it. Defines the gap L2 must close.
E-050 Private plan review as a statutory right, Indiana Seven business days for plan review and three for inspection; an applicant may use a private provider regardless of whether the unit can meet those times, and the unit shall accept the report without further inspection or approval Confirmed Indiana House Enrolled Act 1005 (2025), Pub. L. 146-2025 sec. 4, adding IC 36-7-2.5 secs. 17, 18, 19, 32; effective 2026-01-01 2026 2026-08-17 HAP-02 S10 2027-06 Regraded from Untested and corrected 2026-08-17 — see C-040. C-017 had put this row in Untested because "Indiana reform" named no openable document; the enrolled act has now been read verbatim. TWO SUBSTANTIVE CORRECTIONS: the old value said "or applicant may hire private inspector", which made private review a missed-deadline fallback. Sec. 19 makes it unconditional — the deadline triggers only a fee refund and a convenience fee capped at $100. The old value also implied a 2025 effective date; the act reads January 1, 2026.
E-051 Cost to compile one jurisdiction's ruleset GAP NEVER MEASURED — hand-encoding was the only method until ~2023 Untested 2026-08 HAP-02 S05; HAP-08 determination-specimen; HAP-08 crossing-point-unmeasured GAP. Mechanism is LLM-assisted translation of prose provisions into executable rules. Singapore's hand method is the only datum. RETIRED AS A THESIS, KEPT AS A QUESTION: on 2026-08-10 the claim that this cost was the binding constraint on replication was refuted (E-080, E-081, E-082) and withdrawn from the site. The cost is still unmeasured and still worth measuring; it is no longer claimed to be what closed the market. Fills via Gate 0, which records encoding hours per provision.
E-052 Housing share of US GDP GAP 15.9% of GDP (Q1 2026), 16.0% (Q4 2025) — residential fixed investment ~3.7% plus housing services ~12.3% Untested NAHB analysis of BEA national accounts 2026-06 2026-08 HAP-08 sector-scale; SITE landing 2026-11 CITATION UNRESOLVED (C-017) — regraded from Confirmed on 2026-08-10 because the source field names no document a reader can open. This is a statement about the grade, not about the claim: nothing here says the claim is false, only that it was never checkable. Restore the grade by naming the document — publication title, section, DOI or URL — and re-reading it. Computed from BEA national accounts but published by a trade association — same caveat as E-001. The SHARES are the sourced claim. Dollar figures shown on the site are arithmetic on those shares against a nominal GDP of roughly $30T for 2025, and are deliberately rounded because BEA news releases publish growth rates rather than levels — the level needs verifying against NIPA Table 1.1.5 before any precise dollar figure is quoted. Quarterly series; Q1 2026 already supersedes the Q4 2025 reading.
E-053 US housing shortage estimate 3.7M units (Freddie Mac, Q3 2024); 3.78M (Up for Growth, 2025) Supported Freddie Mac; Up for Growth; CRS IN12628 review of methods 2024–26 2026-08 HAP-08 sector-scale; HAP-08 supply-shortfall 2027-02 Estimates vary widely by method and definition and CRS reviews the spread. Cite as a range, never as one settled number. Freddie Mac separately attributes ~1M delayed household formations. Its decomposition, from the captured snapshot: a standing stock of 147.0M housing units against 150.7M needed, made up of 2.7M vacancy adequacy plus that 1.0M of delayed formation — so most of the headline is the slack a functioning market needs, not a count of families with nowhere to live. The same capture reports eleven studies of the question spanning 1.0M to 4.4M, mean about 2.7M.
E-054 Output cost of housing supply constraints DISPUTED — no figure quotable as settled Contested Hsieh & Moretti, AEJ:Macro 11(2) 2019; Greaney, "Housing Constraints and Spatial Misallocation: Comment", AEJ:Macro 18(2):409-28, April 2026 (doi:10.1257/mac.20230141); Caplan 2021 2019–26 2026-08 SITE landing (dispute only, no figure) 2027-06 Do not quote any single figure from this row. The published paper is internally inconsistent — the abstract states 36 percent growth while the body reports a 3.7 percent GDP level effect — and the widely repeated 36 / 14 percent pair comes from Caplan compounding the paper's annual rates in a 2021 blog post, which produces level effects rather than growth forgone. The authors conceded the inconsistency by email. Superseding all of that: Greaney (AEJ:Macro, April 2026, peer-reviewed) replicated the paper, documented errors in its code, found the counterfactual as coded LOWERS output, showed the result depends on an arbitrary choice of population unit, and reported a corrected effect roughly two orders of magnitude smaller. No published reply or corrigendum exists; the authors responded on a blog. Cite the dispute, not the number. Removed from the landing page figures in C-010.
E-055 Value of the US owner-occupied housing stock Just over $48 trillion (Q3 2025) Confirmed Federal Reserve Financial Accounts (Z.1) household real estate at market value 2025-12 2026-08 HAP-08 sector-scale; SITE landing 2026-12 The asset base the politics of this argument runs on. Federal Reserve Z.1 household real estate at market value. Note: the homeownership rate is NOT from this source — do not attach it to this row. Quarterly series; next release supersedes.
E-056 Aviation type certification and delegated verification 14 CFR 21.11: subpart B "prescribes— (a) Procedural requirements for the issue of type certificates for aircraft, aircraft engines, and propellers; and (b) Rules governing the holders of those certificates." 14 CFR 183.41(a): subpart D "contains the procedures required to obtain an Organization Designation Authorization, which allows an organization to perform specified functions ON BEHALF OF THE ADMINISTRATOR related to engineering, manufacturing, operations, airworthiness, or maintenance", with an ODA Unit defined as two or more individuals within the holder's organization performing the authorized functions Confirmed 14 CFR 21.11 and 14 CFR 183.41, section text from the Government Publishing Office, https://www.govinfo.gov/content/pkg/CFR-2024-title14-vol1/xml/CFR-2024-title14-vol1-part21.xml and https://www.govinfo.gov/content/pkg/CFR-2024-title14-vol3/xml/CFR-2024-title14-vol3-part183.xml 2026-08-10 HAP-02 S06; HAP-08 certify-the-checker 2027-08 SOURCE PASS, phase 1. Section text now read from GPO, upgrading the earlier check which reached only subpart headings via Cornell. The row's existing correction stands: type certification is Part 21, ODA is Part 183 Subpart D, and C-007 was right. ONE THING THE SECTION TEXT ADDS THAT THIS PROJECT HAD NOT RECORDED: an ODA performs its functions "on behalf of the Administrator". The delegation is explicit and the authority stays with the FAA — the ODA holder is exercising the agency's power, not its own. That is a sharper reading of the precedent than "delegated verification", and it cuts both ways for this project. It supports the structure being real; it also means the aviation model is delegation WITHIN an agency's authority rather than an outside determination the agency accepts, which is the thing E-089 and E-088 are cited for. Do not blur the two.
E-057 Abbreviated approval on demonstrated equivalence Hatch-Waxman (1984) permits approval on demonstrated equivalence to a reference product rather than repeating full trials Confirmed Drug Price Competition and Patent Term Restoration Act of 1984, Pub. L. 98-417 1984 2026-08 HAP-02 S07; HAP-08 certify-the-checker 2027-06 The template for the valuation lock: a new product reaching market on equivalence to a reference rather than on its own comparables. Statutory fact; the analogy to housing is argued, not established.
E-058 Auckland building permits per capita after upzoning GAP Doubled against the synthetic-control counterfactual within five years Untested Greenaway-McGrevy et al. (Auckland upzoning literature) 2021–26 2026-08 HAP-02 S04/S10; HAP-08 four-cities-matrix 2027-06 CITATION UNRESOLVED (C-017) — regraded from Confirmed on 2026-08-10 because the source field names no document a reader can open. This is a statement about the grade, not about the claim: nothing here says the claim is false, only that it was never checkable. Restore the grade by naming the document — publication title, section, DOI or URL — and re-reading it. Part of the same body of work as E-029 to E-031. Stated across four surfaces before it had a row — added retrospectively; verify against the specific paper before publication.
E-059 Operation Breakthrough scale GAP 22 housing system producers selected from over 200 competitors, 1969–74 Untested HUD USER program archive 1969–74 2026-08 HAP-02 S03; HAP-08 corenet-singapore 2027-06 CITATION UNRESOLVED (C-017) — regraded from Supported on 2026-08-10 because the source field names no document a reader can open. This is a statement about the grade, not about the claim: nothing here says the claim is false, only that it was never checkable. Restore the grade by naming the document — publication title, section, DOI or URL — and re-reading it. Widely repeated in secondary accounts of the program. Cited from the HUD archive rather than a primary selection record — verify before relying on the exact counts.
E-060 Cycle-time effect of the screening market GAP [UNMEASURED] — no published figure exists for what the screening market did to cycle time. The "roughly 2 of 32 months" this row carried was an internal estimate, never measured and never published by anyone Untested — (this project's own estimate) 2026-08-10 HAP-06 S03; HAP-08 the-instrument 2027-02 Corrected under C-019. The estimate was fine as an estimate; what was wrong is that two beats stated it as a finding the screening market had "published", which nobody has. An Untested row cannot support a sentence that says the experiment has been run and the answer is in. GAP. Nobody has measured what compiled screening tools do to end-to-end cycle time; this is an inference from what those products do rather than a published figure. It is load-bearing for the claim that information alone is insufficient, and it should be presented as an estimate, never as a measurement.
E-061 Auckland population GAP About 1.7 million Untested Stats NZ / standard reference 2024–26 2026-08 HAP-08 one-lock-turned 2027-06 CITATION UNRESOLVED (C-017) — regraded from Confirmed on 2026-08-10 because the source field names no document a reader can open. This is a statement about the grade, not about the claim: nothing here says the claim is false, only that it was never checkable. Restore the grade by naming the document — publication title, section, DOI or URL — and re-reading it. Scale anchor only. Used to keep the Auckland result from being read as nationally generalizable.
E-062 How American homeowners finance an ADU GAP NOT MEASURED Untested 2026-08 Gate 1 GAP. HELOC / cash-out refi / construction loan / cash, and what share qualifies at prevailing rates. Decides whether the valuation veto binds in the entry segment at all. Fills via Gate 1 lender interviews.
E-063 Appraised-value contribution of a permitted ADU to the parent parcel GAP NOT MEASURED Untested 2026-08 GAP. Phoenix metro. If an ADU adds little or nothing to the parent appraisal, equity-based build financing is constrained and the entry segment inherits the valuation lock earlier than assumed.
E-064 Post-approval ADU project failure rate, by refusing party GAP NOT MEASURED Untested 2026-08 Gate 1 GAP. Which consent actually kills projects after entitlement is granted — lender, appraiser, utility, fire, covenant holder, or homeowner withdrawal. Fills via Gate 1 builder interviews.
E-065 HOA / CC&R prevalence over Arizona parcels eligible under HB 2720 and HB 2928 GAP NOT MEASURED Untested 2026-08 GAP. Private covenants are a hard veto that state preemption does not reach. If prevalence over eligible parcels is high, the addressable segment is materially smaller than the statute implies. County assessor and HOA registry data.
E-066 Housing Affordability Breakthrough Challenge GAP $10M per cycle across five $2M grants; 2026 cycle applications closed 15 May 2026; three categories — design and construction, finance, service and delivery Untested Enterprise Community Partners program pages; Wells Fargo newsroom announcement 2026 2026-08 SITE about 2027-06 CITATION UNRESOLVED (C-017) — regraded from Confirmed on 2026-08-10 because the source field names no document a reader can open. This is a statement about the grade, not about the claim: nothing here says the claim is false, only that it was never checkable. Restore the grade by naming the document — publication title, section, DOI or URL — and re-reading it. Largest per-award route recorded — $2M against HUD's $1.5M ceiling (E-046) and the Ivory Prize's $300k pool (E-045). Funded by Wells Fargo, administered by Enterprise. Applicants must show existing results, an evidence base and a scaling strategy, so it carries the same entity constraint as E-046 and a demonstrated-traction constraint on top. The 2026 cycle closed before this work was ready; it is a future-cycle target, not a near-term action. This was the third cycle — the interval between cycles is irregular, so do not assume a date for the fourth. Review by 2027-06 to catch the announcement.
E-067 Florence Towne Apartments — one project, end to end 1,725 days (4y9m): 465 approval + 1,260 construction, for 51 units Confirmed Gabriel & Kung, "Development approval times and new housing supply: Evidence from Los Angeles", Journal of Urban Economics 2026 2026-08 SITE landing; SITE story 2027-06 51-unit affordable development at 410 E. Florence Ave., South L.A. Entitlement cases DIR-2017-4059-TOC and ENV-2017-4060-CE submitted 2017-10-10, approved 2018-02-06; DBS permit applied 2018-03-01, issued 2019-01-18; Certificate of Occupancy 2022-07-01. Of the 465 approval days, 119 were entitlement and 323 permitting. The only named project on this site with a published, dated, end-to-end timeline — cited because a sourced real project is worth more than an illustrative one, and because this project does not invent people.
E-068 Average total development time, LA multifamily 1,413 days (3.9 years) — 652 approval + 863 construction Confirmed Gabriel & Kung, "Development approval times and new housing supply: Evidence from Los Angeles", Journal of Urban Economics 2026 2026-08 SITE story 2027-06 All multifamily projects permitted by the City of Los Angeles, 2010–2022. Population figure that E-067 sits inside, so the named project is not read as an outlier. Multifamily and Los Angeles specific — do not generalize to single-family or to other jurisdictions.
E-069 Number of US jurisdictions writing their own land-use rules 19,491 municipal + 16,214 township + 3,031 county governments (2022) Supported US Census Bureau, 2022 Census of Governments (Organization) 2022 2026-08 SITE landing; SITE story; TITLE 2027-06 Sources the site title and its most-repeated figure, which until now carried no row at all — the least-evidenced number on the site was also the most repeated. "Roughly twenty thousand" tracks the municipal count (19,491) and is deliberately conservative: counting townships and zoning counties, the number of bodies writing their own rules is larger. Graded Supported rather than Confirmed because the government counts are Confirmed but the share exercising zoning authority is not separately published — not every township or county zones. That absence is consistent with the thesis rather than awkward for it, but it is an absence and is recorded as one.
E-070 Compiled-path cycle time GAP ~7 months (1 query+certify + 2 site + 4 factory build and set) — MODELED, NEVER RUN Untested 2026-08 POSTER caveat Assumption row. No compiled project has ever run, and the 32-month today-side baseline is itself assembled from separate sources (E-038). WITHDRAWN FROM THE SITE 2026-08-10: four of the seven months are factory build, so the 32-to-7 comparison sets site-built-today against factory-built-tomorrow and attributes to compiled approval a saving that partly belongs to a change of construction method. The estimate is not disproved; the comparison it was used to make is not like for like.
E-071 Developer margin and annualized-return arithmetic GAP 25% margin ≈ 9%/yr at 32mo vs ≈ 47%/yr at 7mo; 10% margin at 7mo ≈ 18%/yr — ASSUMED MARGINS Untested 2026-08 Assumption row. Margins are stated, not surveyed; annualization is arithmetic on E-070's modeled 7 months. Gross margin on total project cost; ignores leverage. WITHDRAWN FROM THE SITE 2026-08-10 with E-070, because it annualizes a comparison that is not like for like.
E-072 US median home price to median household income GAP ~3.5× (1985: $82,800 vs $23,620) → ~5.0× (2025: $416,900 vs $83,150) Untested NAR median existing-home price; Census median household income; ratio series via JCHS and secondary compilations 2025 2026-08 HAP-08 price-to-income-gap 2027-02 CITATION UNRESOLVED (C-017) — regraded from Supported on 2026-08-10 because the source field names no document a reader can open. This is a statement about the grade, not about the claim: nothing here says the claim is false, only that it was never checkable. Restore the grade by naming the document — publication title, section, DOI or URL — and re-reading it. The gap in one line, and the register's job is to say how the line is built: two primary series (NAR price, Census income) joined by secondary compilations, so the endpoints want verifying against the primary tables before print. JCHS separately records the price-to-income ratio at a record high. The direction and rough magnitude — from ~3.5× to ~5× over forty years — are robust across compilations; the second decimal is not. States the symptom only: no causal share is claimed, and the supply-constraint share of it is Contested (E-054).
E-073 Cost-burdened renter households 22.7M renter households (49%) cost-burdened in 2024 — a record; 12.1M (26%) severely burdened Confirmed Harvard JCHS, State of the Nation's Housing 2026 and America's Rental Housing 2026 2026-06 2026-08 HAP-08 price-to-income-gap 2027-06 Cost-burdened = more than 30% of income on housing; severe = more than 50%. Data year 2024, published June 2026. Annual series — the 2027 State of the Nation's Housing supersedes.
E-074 First-time homebuyer age and market share Median first-time buyer age 40 (record high); first-time share 21% of purchases (lowest since tracking began in 1981; ~40% before 2008) Supported NAR, 2025 Profile of Home Buyers and Sellers 2025-11 2026-08 HAP-08 price-to-income-gap 2026-12 Trade association survey — same posture as E-001, which has since been regraded Contested — but the series is its own, with consistent methodology since 1981, covering transactions July 2024–June 2025. Age and share are what convert the ratio in E-072 from an abstraction into a life outcome: the buyer who entered at 29 in 1981 now enters at 40, or not at all. The 2026 Profile (Nov 2026) supersedes.
E-075 Copyright status of an enacted model code Enacted model code enters the public domain as to the enacting jurisdiction; the same work retains protection as an unenacted model code Confirmed Veeck v. Southern Building Code Congress Int'l, 293 F.3d 791 (5th Cir. 2002) (en banc), 9-6; Georgia v. Public.Resource.Org, 590 U.S. 255 (2020) 2002; 2020 2026-08 2027-08 The two halves of the doctrine, and they do not point the same way. Veeck is the favorable authority and is the older one. Georgia narrowed the government-edicts test to who authored the work — a judge or a legislator — rather than whether it carries the force of law, which is the reading the ICC pressed against UpCodes. Confirmed as to what each court held. Neither addresses compiling law into executable rules; that question is E-077's note and is untested anywhere.
E-076 Unenacted model-code text is not free to copy S.D.N.Y. denied both sides summary judgment but ruled UpCodes had no right to copy model codes and related material that had not been enacted Confirmed Int'l Code Council, Inc. v. UpCodes, Inc., 17 Civ. 6261 (VM) (S.D.N.Y. 2020); appeal on Lanham Act claims, No. 21-826 (2d Cir. 2022) 2020 2026-08 2027-08 The one square holding against the republisher in a decade of litigation, and the source of the architectural constraint on the rule graph: compile per adopting jurisdiction from adopted text, never as one national model-code graph. Core copyright claims were never finally resolved — the 2020 order denied cross-motions rather than deciding. Cited for the unenacted-text holding only.
E-077 Republication of incorporated standards by a for-profit is likely fair use 3d Cir. affirmed denial of preliminary injunction: purpose of making the law accessible is transformative, commerciality is one consideration and not dispositive, full copying reasonable because a reader cannot comply with a law by reading part of it Confirmed American Society for Testing & Materials v. UpCodes, Inc., No. 24-2965 (3d Cir. Apr. 7, 2026), aff'g No. 2:24-cv-01895 (E.D. Pa.); accord ASTM v. Public.Resource.Org, 82 F.4th 1262 (D.C. Cir. 2023) 2026-04 2026-08 2027-04 The closest case that exists to what the rule graph does, decided in a for-profit defendant's favor. Read the posture precisely: preliminary injunction, likelihood of success on a fair-use defense, NOT a final merits judgment — do not quote it as settled law. Facts: ten ASTM steel and aluminum standards incorporated by reference into the IBC, which Philadelphia and other jurisdictions adopted; UpCodes publishes the historical version incorporated into law and does not publish unincorporated standards. Fourth factor was equivocal, the court declining to predict market harm. No decision anywhere addresses compiling law into executable rules rather than republishing it for a reader; that remains untested.
E-078 CORENET introduction and the scope of what e-PlanCheck checks CORENET introduced 1995 to check 2D plans; upgraded 2002 as e-PlanCheck for 3D IFC models; scope is two domains — architectural and building services, reported elsewhere as building control regulations, barrier-free access and fire code Confirmed Zou, Guo, Papadonikolaki, Dimyadi & Hou, "Lessons Learned on Adopting Automated Compliance Checking in AEC Industry: A Global Study", Journal of Management in Engineering 39(4), DOI 10.1061/JMENEA.MEENG-5051; Khemlani, "CORENET e-PlanCheck: Singapore's Automated Code Checking System", AECbytes, 2005 2023; 2005 2026-08-10 HAP-08 corenet-singapore; SITE revisions 2027-08 Two sources and they date the launch differently: the peer-reviewed paper puts the 3D upgrade at 2002, the 2005 trade article says the system officially started in September 2000. They agree on what matters here — the scope is part of a code, not a whole one. This is the row that retires the site's framing of Singapore as having compiled its building code.
E-079 e-PlanCheck reached commercial use Commissioned by the Building and Construction Authority and commercially used; described as the most successful implementation at government level Supported Zou, Guo, Papadonikolaki, Dimyadi & Hou, "Lessons Learned on Adopting Automated Compliance Checking in AEC Industry: A Global Study", Journal of Management in Engineering 39(4), DOI 10.1061/JMENEA.MEENG-5051; Khemlani, "CORENET e-PlanCheck: Singapore's Automated Code Checking System", AECbytes, 2005 2023 2026-08-10 HAP-08 corenet-singapore; SITE revisions 2027-08 Load-bearing and resting on one peer-reviewed source, which states the system reached commercial use without dating first industry use. Khemlani (2005) describes it operating three years after the 2002 upgrade, which corroborates the timing but is a trade publication, not research. Graded Supported: corroboration is outstanding and is owed before this is quoted as settled. Retires "twenty-eight years bought one city". QUOTATION MARKS REMOVED 2026-08-10 under C-021: "has been commercially used" was quoted verbatim from Zou et al. and that paper has not been retrieved in this cycle. It is the phrase the dormancy argument turns on, so it is paraphrased until someone reads it. Retrieval is already owed by Part B.
E-080 The Singapore checking engine was transplanted to other jurisdictions Pilot implementations completed in Norway (Norwegian rules on the e-PlanCheck base) and in New York City (ICC codes); testing under way with Japanese and Australian models as of 2005 Supported Khemlani, "CORENET e-PlanCheck: Singapore's Automated Code Checking System", AECbytes, 2005 2005 2026-08-10 HAP-08 crossing-point-unmeasured; SITE revisions 2027-08 Single contemporaneous trade-publication source. It matters because each transplant inherited the engine and the platform work, so whatever the encoding cost was, these projects did not pay it from zero. CHALLENGED AND HELD, 2026-08-10 (C-026). The New York City claim was challenged in review as a suspected conflation with the ICC SMARTcodes thread. It is not one: Khemlani (2005) states "A pilot implementation in New York City using the ICC (International Code Council) codes has also been completed" and that all these implementations were "developed based on novaCITYNETS' FORNAX platform". The claim traces correctly to its cited source and the chronology rules the conflation out — the NYC pilot is 2005, SMARTcodes is 2006. THE CORROBORATION GAP, recorded because holding is not the same as being corroborated: this rests on one trade article, from the platform vendor's ecosystem, and three later surveys of this exact field do not mention it — Dimyadi & Amor (2013) contains no reference to New York, Potter (2022) describes no implementation outside Singapore, and Zou et al. identify no production system outside Singapore. Also note "inherited it for free" is this project's inference, not Khemlani's wording, and a completed pilot is not a jurisdiction adopting.
E-081 No transplant of the Singapore engine has been identified reaching production NO PRODUCTION DEPLOYMENT IDENTIFIED — Norway, New York City, Australia, Japan Supported Zou, Guo, Papadonikolaki, Dimyadi & Hou, "Lessons Learned on Adopting Automated Compliance Checking in AEC Industry: A Global Study", Journal of Management in Engineering 39(4), DOI 10.1061/JMENEA.MEENG-5051, whose eight-country study identifies no such production system outside Singapore; Khemlani, "CORENET e-PlanCheck: Singapore's Automated Code Checking System", AECbytes, 2005 for the pilots themselves 2005; 2023 2026-08-10 HAP-08 adoption-variables; SITE revisions; HAP-08 crossing-point-unmeasured 2027-08 An argument from absence, and graded as one — a system could have reached production without appearing in either source. Stated because it is the observation that refutes this project's own encoding-cost thesis: four projects inherited the engine for free and none of them shipped, so encoding cost cannot have been the binding constraint. Closing this properly means asking the four agencies directly.
E-082 Variables governing adoption of automated compliance checking 12 identified: 4 technology-related, 8 non-technical Confirmed Zou, Guo, Papadonikolaki, Dimyadi & Hou, "Lessons Learned on Adopting Automated Compliance Checking in AEC Industry: A Global Study", Journal of Management in Engineering 39(4), DOI 10.1061/JMENEA.MEENG-5051 2023 2026-08-10 HAP-08 adoption-variables; SITE literature 2027-08 Technology-related: technology integration; BIM modelling standard; BIM information standard and requirements; standard of interpreting building code. The other eight are government support, human resistance, industry readiness, education and training, and four more. Cross-case analysis of eight countries. Directly contradicts a diagnosis in which the binding constraint is a unit cost.
E-083 Automated compliance checking reaches only quantifiable provisions Scope is mainly quantifiable regulations; qualitative requirements relying on human judgement are named as the key barrier Confirmed Zou, Guo, Papadonikolaki, Dimyadi & Hou, "Lessons Learned on Adopting Automated Compliance Checking in AEC Industry: A Global Study", Journal of Management in Engineering 39(4), DOI 10.1061/JMENEA.MEENG-5051 2023 2026-08-10 SITE revisions 2027-08 The same distinction E-039 says nobody has measured at scale, observed qualitatively across eight countries. It raises the prior that the objective share is the binding number, and it is the reason Gate 0 measures a cross-tab of objective against binding rather than objective alone.
E-084 Closed rule engines are named as an adoption barrier Existing tools reported limited by their black-box nature; participants call for systems open enough that anyone can change the rules rather than having them hard-coded Confirmed Zou, Guo, Papadonikolaki, Dimyadi & Hou, "Lessons Learned on Adopting Automated Compliance Checking in AEC Industry: A Global Study", Journal of Management in Engineering 39(4), DOI 10.1061/JMENEA.MEENG-5051 2023 2026-08-10 2027-08 The black-box remark in the paper is made about Solibri Model Checker, not about the Singapore engine; the hard-coding remark is general. Recorded at that precision deliberately — E-023 carries the hand-encoding claim itself.
E-085 Agencies interpret the same regulations differently, and the remedy attempted is to change the rule Reported: officials from different agencies interpret rules differently; the current solution is to quantify the code as much as possible to avoid disagreement Supported Zou, Guo, Papadonikolaki, Dimyadi & Hou, "Lessons Learned on Adopting Automated Compliance Checking in AEC Industry: A Global Study", Journal of Management in Engineering 39(4), DOI 10.1061/JMENEA.MEENG-5051 2023 2026-08-10 SITE revisions 2027-08 From the paper's cross-case table. The remediation brief attributes this quotation to Singapore; the column attribution could not be verified from the published PDF, whose table structure does not survive text extraction. Graded Supported for that reason, and the attribution is owed. The substance is what matters and does not depend on the country: the remedy for an unmodellable rule was to change the law, not to model it better.
E-086 Regulations must change before they can be fully translated Reported: existing regulations are not precise enough to build an automated system on; the regulations need to change to achieve 100% code translation Supported Zou, Guo, Papadonikolaki, Dimyadi & Hou, "Lessons Learned on Adopting Automated Compliance Checking in AEC Industry: A Global Study", Journal of Management in Engineering 39(4), DOI 10.1061/JMENEA.MEENG-5051 2023 2026-08-10 SITE revisions 2027-08 Same table and same caveat as E-085 — the brief attributes this to Estonia and the column attribution is unverified. If it holds, it is the most important sentence in the paper for this project: the constraint is legislative, not computational.
E-087 Effort to hand-translate a page of code into executable rules GAP [UNMEASURED] Untested 2026-08-10 GAP. The remediation brief cites roughly one expert-day per page of code including QC, on a New Zealand datum. That figure is not in Zou et al. (2023) and no primary source for it was found. Opened rather than quoted. Fills from the Gate 0 encoding log, which measures hours to produce a candidate encoding directly.
E-088 Private organizations approve manufactured-home designs for HUD, and the approval travels to every state 24 CFR 3282.351(a): the subpart sets the requirements for "States or private organizations which wish to qualify as primary inspection agencies". 3282.351(d): accepted agencies are granted provisional then final acceptance, and "HUD accepted agencies can perform DAPIA functions for any manufacturer in any State". 3282.351(e): PIAs "may contract with manufactured home manufacturers" for those services. 3282.352 reserves exclusivity to a State for IPIA (production inspection) only; there is no exclusive-DAPIA provision Confirmed 24 CFR 3282.351, 3282.352 (acceptance of primary inspection agencies); 24 CFR 3282.361 (DAPIA); 24 CFR 3282.362 (IPIA); 41 FR 19852, 13 May 1976; National Manufactured Housing Construction and Safety Standards Act of 1974, 42 U.S.C. ch. 70 1976; current 2026-08-10 HAP-08 the-instrument; HAP-08 the-instrument; SITE literature 2027-08 Regraded Confirmed on 2026-08-10 from the regulation text — see C-015. The row previously flagged its own unread half: whether private firms are among the agencies HUD accepts. They are, expressly, in the first sentence of the subpart, and the section carries two things the row did not claim. A private DAPIA contracts with the manufacturer it reviews, and its design approval is good for any manufacturer in any state — a private determination that is portable by federal regulation, since 1976. SCOPE LIMIT, and it is the whole limit: the approval is against one federal standard that preempts local construction rules (E-091), not against 20,000 local rulesets. DAPIA is evidence that portability works where a single standard exists; it is an argument for preemption, not for a compiled multi-jurisdiction determination.
E-089 Private plan review substituting for the building department, with automated review permitted A fee owner or authorized contractor may use a licensed private provider for plans review and inspections in lieu of the local building department; the provider carries $1M/$2M professional liability below $5M project value and $2M/$4M above; officials must issue or identify deficiencies within 20 business days (5 for single-trade residential); subsection (6) permits a private provider to use an automated or software-based plans review system for single-trade review Confirmed Fla. Stat. §553.791 (2025), subsections (1)(n), (2)(a), (6), (7)(a), (14)(a), (18) 2025 2026-08-10 SITE literature 2027-06 The closest thing in American law to what this project proposes, and it is nineteen years old with an automated-review provision added in 2025. Read against the argument this site has been making: the mechanism is not unprecedented, it is already statutory in one state. What Florida does not do is make the determination itself authoritative across jurisdictions.
E-090 The permanent chassis is no longer required of a manufactured home The statutory definition now reads "with or without a permanent chassis"; HUD is to issue revised standards Confirmed 42 U.S.C. §5402(6) as amended; 21st Century ROAD to Housing Act, H.R. 6644, §301, 119th Congress, enacted 11 July 2026 2026-07 2026-08-10 SITE revisions 2027-02 Verified in the codified definition itself, which now carries the amended language. The enactment date and vote totals come from bill-tracking secondary sources; the codified text is primary. Provisions the brief also attributes to the Act — state certification deadlines, FHA Title I limits, a small-dollar mortgage pilot, a pattern-book grant program, investor-purchase restrictions — are NOT verified here and are not claimed.
E-091 The 1974 Act federalized construction and left property classification alone The Act preempts state and local construction and safety standards for the same aspect of performance (42 U.S.C. §5403(d)); it does not address whether a manufactured home is real or personal property Confirmed 42 U.S.C. §§5402, 5403(d); National Manufactured Housing Construction and Safety Standards Act of 1974 1974; current 2026-08-10 SITE revisions 2028-08 The silence is the finding. Federal law made the product uniform and left the question that determines how it is financed to fifty state legislatures — which is the shape of the failure E-009 to E-014 measure.
E-092 Manufactured homes are titled like vehicles States issue a certificate of title through a motor vehicle agency or a dedicated manufactured housing authority; conversion to real property requires meeting state requirements including permanent attachment to land Supported State titling statutes (e.g. Miss. Dep't of Revenue Motor Vehicle Services; Nev. Rev. Stat. ch. 489; Tex. Dep't of Housing and Community Affairs Statement of Ownership); Freddie Mac, "Titling Manufactured Housing as Real Property" 2026-08-10 2027-08 The mechanism is documented in several state statutes and in a GSE fact sheet; no single national source states it. The count of states still issuing certificates of title — the brief says 42 — is not verified and is not claimed here. Opened as E-093's neighbour rather than asserted. SEARCHED 2026-08-10 (C-022): the leading 50-state legal survey — the CFED and National Consumer Law Center review of titling law — contains no count of states issuing certificates of title. The number it does give is about a different question: "Approximately three-quarters of the states have statutes that set forth a procedure to convert a manufactured home from personal to real property", with the states listed in its Appendix A. So 42 is not merely unverified, it is unlocated: the obvious authority does not state it. Use the conversion finding, which is sourced, or say nothing.
E-093 Share of manufactured homes that ever move after initial placement GAP [UNMEASURED] Untested 2026-08-10 SEARCHED AND NOT FOUND, 2026-08-10 (C-022). The figure is stated by the Lincoln Institute of Land Policy — a serious research organisation — twice on one page, with no footnote, no hyperlink and no named study, and in two incompatible forms in adjacent sentences: "fewer than 5-7 percent of all manufactured homes are ever moved after initial placement" and "roughly 95 percent of homes are never moved". Pew repeats it. Nobody cites anybody. A number that varies between 5% and 7% in the same clause is a number without an origin. DO NOT WRITE THIS CLAIM. It is the most rhetorically useful figure available to this argument — the permanent chassis, and the vehicle-titling regime built on it, are justified by a transportability that supposedly almost never happens — and that is exactly why it must not go on the site unsourced. WHAT WOULD CLOSE IT: the Census Manufactured Housing Survey asks dealers about placement, and the American Housing Survey records whether a unit has moved; either could support a real estimate. Until then the argument makes the point structurally — a home titled as a vehicle is taxed, financed and depreciated as one whether or not it ever moves — which needs no figure at all.
E-094 Interest rate by secured property type on manufactured housing Median rate 8.6% on chattel loans, 4.9% on manufactured-home mortgages, 4.1% site-built; median rate spread 5.2 / 1.6 / 0.4; 93.8% of chattel originations are higher-priced mortgage loans against 11.1% site-built Confirmed CFPB, "Manufactured Housing Finance: New Insights from the Home Mortgage Disclosure Act Data", May 2021 (2019 HMDA data) 2021-05 2026-08-10 SITE revisions 2027-05 Table 6 of the report, 2019 HMDA data. Distinct from E-011, which is the Urban Institute lender-type spread; this one is by what secures the loan.
E-095 Chattel borrowers who own their land are not weaker credits Landowners who take chattel loans have similar median credit scores and incomes and better loan-to-value and debt-to-income ratios than landowners who take mortgages Confirmed CFPB, "Manufactured Housing Finance: New Insights from the Home Mortgage Disclosure Act Data", May 2021 (2019 HMDA data) 2021-05 2026-08-10 SITE revisions 2027-05 The report's own reading: this "suggests that chattel borrowers' credit profiles would not have prevented them from getting a mortgage". Denial is therefore not explained by credit quality — which is the finding that moves the diagnosis from borrower risk to product architecture. The report names small loan amounts as a likely mechanism.
E-096 Chattel lending is unevenly distributed by race and age Hispanic, Black and African American, American Indian and Alaska Native, and elderly borrowers are more likely to take out chattel loans even after controlling for land ownership Confirmed CFPB, "Manufactured Housing Finance: New Insights from the Home Mortgage Disclosure Act Data", May 2021 (2019 HMDA data) 2021-05 2026-08-10 SITE revisions 2027-05 The report calls determining the reasons a vital area of future research, and this register does not offer one.
E-097 Average sales price of a new manufactured home $137,500 US average, $90,700 single-section, $164,200 multi-section (March 2026); 2025 mean of the twelve monthly averages $127,367 / $85,433 / $157,342 Confirmed US Census Bureau, Manufactured Housing Survey, "Average Sales Price by Region and by Size of Home" time series 2026-03 2026-08-10 SITE literature 2026-12 Monthly figures are from the published table verbatim. The 2025 figures are this register's own arithmetic — an unweighted mean of twelve monthly averages, not the Census annual estimate, which is shipment-weighted and will differ. Recorded that way rather than passed off as published. The remediation brief quoted $115,557 average and $95,074 single-section for 2025; neither matches this series and neither is used.
E-098 Price per square foot, manufactured against site-built, like for like GAP [UNMEASURED] Untested 2026-08-10 SITE literature GAP, and it is the one that decides the strongest objection on the site. The claim that factory-built housing is cheaper per square foot needs a like-for-like comparison — same finish level, same region, same year, land excluded on both sides. The often-quoted ~$101/sq ft for manufactured homes could not be sourced to the Census series, which publishes price but not floor area in the same table. Until this row is filled the Potter objection stays conceded.
E-099 What builders themselves call a serious problem Cost/availability of developed lots 63% and cost/availability of labor 61% in 2025; high interest rates 84%, buyers expecting prices to fall 81%; approval or plan-review duration does not appear anywhere in the top ten Confirmed NAHB, builder survey reported in "Builders' Top Challenges for 2026", Eye on Housing 2026-02 2026-08-10 SITE literature 2027-02 A survey of NAHB's own members, with the same provenance problem as E-001 — and pointing the other way. E-001 is the trade association's advocacy blueprint arguing regulation is 26.4% of price; this is the trade association asking its members what actually hurt, and approval duration is absent from the answer. Both are cited on this site. The methodological note on /literature says which is used for what.
E-100 Separate sale of an accessory dwelling as a condominium in California AB 1033 (effective 2024) lets a jurisdiction opt in to permitting separate condominium conveyance of an ADU; three jurisdictions had opted in as of April 2026 — San Jose, Santa Monica, unincorporated San Diego County; conversion requires a condominium map, plan, CC&Rs and an HOA, estimated at $15,000-$30,000 in legal and filing fees Supported Cal. AB 1033 (2023, eff. 2024); County of San Diego, "Guidance for Separate Sale of ADUs under AB 1033", 3 April 2026; trade and practitioner coverage for the opt-in count and cost range 2024; 2026-04 2026-08-10 2027-04 The opt-in count and the cost range come from practitioner sources, not from a state registry, and no state registry of opt-ins is published. Graded Supported for that reason. A right granted statewide and exercised by three jurisdictions in two years is the same pattern as E-027 and E-037: the right exists, nothing exercises it.
E-101 Housing supply growth has fallen and converged across US metros If the housing stock had grown 2000-2020 at its 1980-2000 rate there would be about 15 million more units; growth rates have fallen and converged across large Sunbelt markets — Atlanta, Dallas, Miami, Phoenix — with prices rising most where supply fell most; the slowdown is suburban, not only central-city Confirmed Glaeser & Gyourko, "America's Housing Supply Problem: The Closing of the Suburban Frontier?", Brookings Papers on Economic Activity 56(1), Spring 2025, pp. 375-453; NBER Working Paper 33876 2025 2026-08-10 SITE literature 2028-03 The specific per-metro figures the remediation brief quoted — all four Sunbelt metros below 1% stock growth since 2020, a decline beginning in the 1970s — are not verified here and are not claimed; the summary findings above are. Relevant because it locates the constraint on the land side, where E-099 also puts it.
E-102 Prefabrication does not systematically cost less than conventional construction Prefabricated construction typically costs no less than conventional building; US construction productivity has risen roughly 10% since 1945 while manufacturing productivity rose 8.6x Supported Brian Potter, "The Ups and Downs of Automated Code Checking Software", Construction Physics, 27 May 2022, https://www.construction-physics.com/p/the-ups-and-downs-of-automated-code 2022-05-27 2026-08-10 SITE literature 2027-08 Citation resolved under C-026, clearing the CITATION UNRESOLVED flag C-017 put on it: the source was "Potter, Construction Physics", a person and a publication but no article. Restored to Supported rather than Confirmed — it is a well-argued industry essay, not peer-reviewed research. Potter is the source conceded against this project on prefabrication cost AND cited for it on CORENET (E-128), which is the right way round: a source used against yourself is worth more when used for yourself. A well-documented industry analyst writing outside peer review, and the strongest available challenge to any factory-cost claim on this site. Graded Supported, cited as the objection rather than as a settled result, and answered only to the extent E-098 can be filled — which it currently cannot.
E-103 The manufactured-housing finance agenda is established published work Pew's Housing Policy Initiative has worked since 2020 on small mortgages under $150,000, personal property lending for manufactured homes and alternative financing arrangements, extending to housing shortage and land use in 2023; "Overcoming Barriers to Manufactured Housing: Promising Approaches from Five Case Studies" (2024, with Harvard JCHS) finds a new manufactured home can save a buyer $50,000 to more than $100,000 against a comparable site-built house Confirmed The Pew Charitable Trusts, Housing Policy Initiative; Herbert & Reed, "Overcoming Barriers to Manufactured Housing", Harvard Joint Center for Housing Studies for Pew, 2024 2024 2026-08-10 SITE literature 2027-08 Recorded because this project claimed novelty it does not have. The financing half of the argument has been published by a major research organization for six years, and the site says so on /literature rather than leaving a reader to discover it.
E-104 HUD-code housing is the least expensive way to build, and its buyers earn less Testimony: "Manufactured housing built to the Department of Housing and Urban Development's code is the least expensive way to build a new home in the US today. The median income of the homebuyer of manufactured housing is $57,000 compared to $93,000 for site-built housing." The same testimony puts manufactured homes at about 20 percent of all new single-family houses before 2000 against 9 percent today Supported Emily Hamilton, Mercatus Center at George Mason University, testimony before the Subcommittee on Housing and Insurance, US House Committee on Financial Services, 4 March 2025, HHRG-119-BA04-Wstate-HamiltonE-20250304, https://docs.house.gov/meetings/BA/BA04/20250304/117970/HHRG-119-BA04-Wstate-HamiltonE-20250304.pdf 2025-03-04 2026-08-10 SITE literature 2027-08 Closed under C-023 from the committee's own posted PDF, quoted verbatim. Restored to Supported rather than Confirmed, and the distinction matters: this is an expert asserting a figure in a congressional record, not a measurement. Hamilton attributes the income comparison to her footnote 7 and the share figures to footnote 8; NEITHER FOOTNOTE HAS BEEN CHASED. Quote it as testimony — "told Congress" — not as a finding, until the underlying sources are read. The 20-percent-to-9-percent share is a useful second reading on E-012's shipment collapse because it is a share of a growing denominator rather than a count.
E-105 A research network already covers all four layers of this argument GAP The Innovations in Manufactured and Modular Homes (I'm HOME) network, founded 2005 and convened by the Lincoln Institute of Land Policy since March 2022, works on home quality, land tenure and security, financing, and land use and zoning Untested Lincoln Institute of Land Policy, I'm HOME network 2022; current 2026-08-10 SITE literature 2027-08 CITATION UNRESOLVED (C-017) — regraded from Confirmed on 2026-08-10 because the source field names no document a reader can open. This is a statement about the grade, not about the claim: nothing here says the claim is false, only that it was never checkable. Restore the grade by naming the document — publication title, section, DOI or URL — and re-reading it. Those four are the same four layers this project separates. Recorded for the same reason as E-103.
E-106 Manufactured housing is already a coded characteristic of the national zoning data The National Zoning Atlas codes upwards of 200 regulatory characteristics per district and its captured permissions include accessory dwelling units, planned residential developments and mobile homes; 33,295 jurisdictions identified and 102,000 districts logged; methodology published as "How To Make a Zoning Atlas 2.0" (2024) Confirmed National Zoning Atlas, "How the National Zoning Atlas is Made"; "How To Make a Zoning Atlas 2.0" (2024) 2024; current 2026-08-10 SITE literature 2027-02 This refutes the specific novelty claim the remediation brief expected to survive. Manufactured housing is not an absent characteristic in the national zoning data — it is coded. What remains open is narrower and is stated that way: not whether mobile-home permission is recorded, but whether the dimensional and materials standards that make a permission unusable are recorded, and whether any published atlas has tested real purchasable models against real parcels. That is E-107.
E-107 The placement gap GAP [UNMEASURED] — share of residential parcels in a jurisdiction where a dwelling is permitted but a HUD-code home that someone can actually order is not placeable Untested 2026-08-10 SITE literature; Gate 0 GAP, and the narrowed target of Gate 0. Distinct from E-106: the atlas records whether a district permits mobile homes; this measures whether a specific purchasable model clears district permission, dimensional standards, materials requirements, lot geometry and setbacks on a specific parcel. Distinct from E-039, which asks whether provisions are objective at all. Both must hold for the number to mean anything.
E-108 Uptake of Florida's private-provider plan review GAP [UNMEASURED] — share of Florida building permits reviewed by a private provider under Fla. Stat. §553.791 rather than by the local building department, since 2006 Untested 2026-08 SITE story GAP, and a nineteen-year natural experiment nobody here has read. E-089 records that the substitution is lawful, insured, fee-protected and clock-bound; nothing records whether anyone uses it. Either answer is load-bearing: high uptake makes the remaining question portability across jurisdictions, and near-zero uptake after nineteen years of availability means a mechanism that is lawful, faster and free of extra fee still did not get used — which is the most important negative result available to this project. Obtainable from public permit records rather than from a pilot, and cheaper than Gate 0. Tests what Gate 0 does not reach: not whether rules can be computed, but whether anyone accepts a computation they did not perform.
E-109 Variety under industrialized housing production in Japan GAP Prefabricated housing is roughly 13% of Japanese housing starts (64,035 of 495,737 detached starts in 2013; ~13.1% reported more recently), produced by manufacturers whose documented method is mass customization — per-buyer variation inside a standardized production system Untested Academic literature on Japanese prefabricated housing manufacturers (Daiwa House, Sekisui House, Misawa Homes); Matsumura, Japan Architectural Review (2019) 2013–2024 2026-08 SITE story 2027-06 CITATION UNRESOLVED (C-017) — regraded from Supported on 2026-08-10 because the source field names no document a reader can open. This is a statement about the grade, not about the claim: nothing here says the claim is false, only that it was never checkable. Restore the grade by naming the document — publication title, section, DOI or URL — and re-reading it. Cited for VARIETY only, never for cost — E-102 records that prefabrication does not systematically cost less, and the prefab-cost objection is conceded on this site. The claim here is narrower and is about what standardization does to choice: an industry can standardize the certified production system and still vary the product per buyer, at scale, for decades. The share figure is a share of DETACHED starts in the 2013 datum and wants restating against a current national series before it is quoted precisely. Mass customization is the literature's term, not this project's. Graded Supported: the manufacturers and the method are well documented, the share is from secondary compilations.
E-110 Motor vehicle conformance is certified by the manufacturer, with no agency approval before sale A vehicle may not be sold unless it complies with the applicable federal motor vehicle safety standard (49 U.S.C. 30112(a)(1)), and the manufacturer or distributor "shall certify to the distributor or dealer at delivery that the vehicle or equipment complies with applicable motor vehicle safety standards" (49 U.S.C. 30115(a)). Nothing in the chapter provides for agency approval or pre-sale testing; enforcement runs after sale — defect or noncompliance determination, owner notification, remedy (30118-30120) Confirmed 49 U.S.C. 30112, 30115, 30118-30120; 49 CFR 567.4 (certification label) 1966 Act; 1994 recodification; 2020 notice 2026-08-10 HAP-08 certify-the-checker 2027-08 Statutory text verified from 49 U.S.C. 30115(a) on 2026-08-10. QUOTATION REMOVED UNDER C-020: this row previously carried "NHTSA does not pre-approve vehicles, through testing or other means, before they can be sold or otherwise introduced into interstate commerce", attributed to 85 FR 83145. That notice could not be retrieved and the sentence could not be verified, so it is gone rather than hedged. The claim does not need it: the absence of pre-approval is what the statutory scheme says by containing no approval provision, which is a stronger basis than a quotation from a notice nobody on this project has read. The third donor structure in the certify-the-checker beat, and the one furthest from what this project proposes: cars are not type-approved at all. A claim about regulatory structure, not about outcomes — nothing here says self-certification made cars safer, and this row is not an argument for importing it. The disanalogy is load-bearing and is stated in the beat: the automotive check sits after sale and its remedy is the recall, and a house cannot be recalled. Cited beside E-056 so the precedent is bracketed at both ends — aviation approves the design in advance and delegates the verification, motor vehicles approve nothing in advance. What makes the determination portable is E-111.
E-111 A federal motor vehicle safety standard preempts differing state standards, so one conformance determination is valid in every state Where a federal motor vehicle safety standard is in effect, a State may prescribe a standard on the same aspect of performance only if it is identical to the federal one (49 U.S.C. §30103(b)(1)); compliance with the standard does not exempt a person from liability at common law (§30103(e)) Confirmed 49 U.S.C. §30103(b)(1), (e) 1966 Act; 1994 recodification; current 2026-08-10 HAP-08 certify-the-checker 2027-08 The half of the automotive model this argument actually needs: a certification is only worth having if it travels, and what makes it travel is a standard whose authority does not stop at the state line. Housing holds one instrument of that shape — the HUD code (E-088) — and it reaches one product class; state industrialized-building preemption (E-047) stops at the state border, and IIBC recognition (E-048) is voluntary rather than preemptive. Note what preemption did not displace: §30103(e) leaves common-law liability intact, so the trade is a uniform standard plus tort, not a uniform standard instead of it. Nobody here has priced that trade for housing.
E-112 Denial rate on manufactured-home purchase applications, against site-built 42% of all manufactured-housing home purchase applications were denied — 50% of chattel and 33% of MH mortgage applications — against 7% of site-built applications Confirmed CFPB, "Manufactured Housing Finance: New Insights from the Home Mortgage Disclosure Act Data", May 2021, s.4.1 p.15 (2019 HMDA data) 2021-05 2026-08-10 2027-08 The cleanest statement of the disparity in the source and the one the site has never used: one measure, one denominator, a six-fold gap, no rounding and no mixing. Added under C-014. Prefer this to E-009 wherever the point is that the loan is refused rather than that the buyer walked away — 42 against 7 is a denial, while 27 against 74 also carries applications withdrawn or closed for incompleteness.
E-113 Approval rate by secured property type, holding credit score constant In the super-prime band above 720, chattel 63% and MH mortgage 80% against site-built 95%. Sub-prime applicants for site-built homes were more likely to be approved than super-prime chattel applicants. Approval rate is defined at n.32 as originations plus approved-not-accepted, over that plus denials Confirmed CFPB, "Manufactured Housing Finance: New Insights from the Home Mortgage Disclosure Act Data", May 2021, Figure 3 and n.32 (2019 HMDA data) 2021-05 2026-08-10 2027-08 Split out of E-009 under C-014, where it sat in the same value field as an origination rate — so a reader comparing 27% to 63% was comparing different denominators and would have read it as a rise. This is the load-bearing half of the finding: the gap survives controlling for credit score, so it is not explained by borrower quality. See E-095, which shows the same for landowners who take chattel.
E-114 Distribution of single-family completion times, 2024 Of single-family units completed in 2024: 13% in 3 months or less, 46% in 4 to 6 months, 20% in 7 to 9, 9% in 10 to 12, and 13% in 13 months or more Confirmed US Census Bureau, Survey of Construction, "Number of Housing Units Completed by Number of Months from Start: Percent Distribution and Relative Standard Error, 2024", published 2026-07 2026-07 2026-08-10 2027-08 Added under C-016 as the answer to what 2024 actually publishes: a distribution, not an average. At the time of that correction Census had published no annual average past 2022. DERIVATION, SHOWN SO IT CAN BE CHECKED AND NOT QUOTED: taking band midpoints 2, 5, 8 and 11 months and assuming the open 13-months-or-more band sits at 15, 16 or 18 gives 7.0, 7.2 or 7.4 months. The open band is the whole uncertainty and it is 13% of the weight. This is arithmetic on a published distribution, it is NOT a published average, and it must not appear on the site or be used to update the eight-months figure. DIRECTION, AND WHY IT MATTERS: it points DOWN against the 8.3 of 2022. If the build share of the cycle is falling while the total is not, the paper share is rising — which strengthens the diagnosis this project makes rather than weakening it, and is the more interesting reading of the number. That makes it worth sourcing properly rather than dropping. DO NOT TIDY THIS ROW AWAY. It reads like a loose end and it is a lead: the thing to obtain is a published 2023-2025 average, or the SOC microdata, so the trend can be stated rather than estimated. SUPERSEDED IN PART under C-038: the sentence this row used to carry about Census having stopped is no longer true — 2023, 2024 and 2025 averages are now published. The derivation recorded here can now be checked against them: it produced 7.0-7.4 for 2024 and the published 2024 average is 7.7. The distribution figures in this row are unaffected and the row stays Confirmed.
E-115 Regulator oversight of a delegated-approval programme in which the reviewer is paid by the reviewed party DOT OIG on the FAA Organization Designation Authorization programme: "management and oversight weaknesses limit FAA's ability to assess and mitigate risks with the Boeing ODA"; "more than 5 years after our recommendation, FAA has not yet implemented a risk-based approach to ODA oversight"; ODA unit members reported "pressure from Boeing company management to approve items or affirm compliance with regulations without sufficient time to perform a review"; 14 recommendations made Confirmed US DOT Office of Inspector General, Report AV2021020, "Weaknesses in FAA's Certification and Delegation Processes Hindered Its Oversight of the 737 MAX 8", 23 February 2021 2021-02 2026-08-10 2027-08 Read from the report PDF, not from a summary. This is the adverse case for delegated conformance review and it belongs to this project rather than to a critic: the structure E-088 cites as the strongest American precedent is the same structure whose aviation counterpart the regulator's own inspector general found inadequately overseen. Note the limit of what this row supports — it is a finding about FAA oversight of one ODA, not a finding about HUD, and no equivalent audit of DAPIA performance has been identified. It is cited as the objection, not as proof the objection holds here. See E-116 for the mitigations HUD's regulation actually provides.
E-116 What the HUD scheme does to control a reviewer paid by the manufacturer it reviews 24 CFR 3282.453: each primary inspection agency is monitored four times per year, reducible to a minimum of one where performance is deemed superior and increased where performance is suspect, with at least one records review annually. 3282.451: monitoring is carried out by joint teams of state agency and HUD personnel, feeding final acceptance, continued acceptance and disqualification decisions. 3282.356: the Secretary may disqualify an agency, and interested persons may petition for disqualification. 3282.351(e): fees excessive in relation to the services rendered are themselves grounds for disqualification Confirmed 24 CFR 3282.451, 3282.453 (subpart J, monitoring); 24 CFR 3282.356 (disqualification and requalification); 24 CFR 3282.351(e) 1976-05 2026-08-10 2027-08 Read from the regulation text. These are the mitigations AS WRITTEN and nothing here says they work. No published finding has been identified on DAPIA or IPIA disqualification rates, monitoring outcomes, or whether the four-visits-a-year cadence is met — which is the gap that would have to be closed before this row could be offered as an answer to E-115 rather than as the text it is. Do not present a written control as a demonstrated one.
E-117 A jurisdiction mandating automated pre-application review Honolulu: "Starting September 1, 2026, eligible residential permit applications must complete the CivCheck pre-application process before submitting through HNL Build" — automated review made a mandatory step by the permitting authority itself Confirmed City and County of Honolulu, Department of Planning and Permitting, "Permit Process Improvements", https://www.honolulu.gov/dpp/permitting/building-permits-home/building-permits-permit-process-improvements/ 2026-08 2026-08-10 2027-02 Read from the city page. Note precisely what this is and is not: a PRE-APPLICATION screen the city requires before submission, not a determination that replaces the department's decision. The city still decides. It matters to this project because it is the competing path — a jurisdiction buying automation directly, needing no outside determination, no professional of record, no carrier and no portability.
E-118 Effect of municipal automated plan review on permit decision time GAP [UNMEASURED] — no figure from a permitting authority has been located. Vendor-published material and trade coverage report Honolulu decision times falling from about 73 days to about 32.5 days, and per-application review time from 60-90 minutes to 15-20 minutes; the city itself publishes no before-and-after figures Untested CivCheck, "City and County of Honolulu — Pilot Case Study (2024)", https://www.civcheck.ai/blog/honolulu-pilot-case-study-2024 (vendor-published); trade coverage in HousingWire, "How Honolulu leveraged AI to cut permit review times in half" 2026 2026-08-10 2027-02 THE FIGURE IS THE VENDOR'S. Added under C-019 and graded Untested for that reason. The only sources located are the company selling the software and a trade publication; the Honolulu DPP page announcing the mandate publishes no figures at all, and the city auditor's permit-processing report predates the deployment. This is the same shape as C-009 and C-010 — a headline number whose newest source is not independent — and this project is not going to make that mistake a third time by quoting it. What would settle it: DPP decision-time data by quarter, from the city, spanning the deployment.
E-119 Whether HUD's controls on a manufacturer-paid reviewer have ever bound GAP [UNMEASURED] — how many Design Approval or Production Inspection Primary Inspection Agencies have been disqualified under 24 CFR 3282.356 since 1976; how many for fees excessive in relation to services under 3282.351(e); how many third-party petitions for disqualification have been filed under 3282.156(b) and with what outcome; and whether the monitoring cadence required by 3282.453 — four visits a year, minimum one records review — has been met Untested 24 CFR 3282.356, 3282.351(e), 3282.156(b), 3282.453 name the controls; no compliance or enforcement record for them has been located 2026-08-10 2027-02 THE TEST FOR THE REVIEWER-CAPTURE OBJECTION, which until now was the only conceded objection on this site with nothing named that would settle it. E-116 records what HUD's regulation requires; this row records whether any of it has ever happened. A written control and a binding one are different claims, and only the second answers the objection. HOW TO CLOSE IT: this is a records request, not a research programme. HUD Office of Manufactured Housing Programs holds the acceptance, monitoring and disqualification records; the ask is a count by year since 1976 of disqualifications, fee-based removals and third-party petitions, with outcomes, plus the monitoring-visit completion rate. FOIA if it is not published. A null result is a finding either way: zero disqualifications in fifty years reads as a control that has never bitten, and a steady rate reads as one that has. Do not treat the absence of records as the absence of a problem.
E-120 Conversion statutes evidence that the default classification is personal property "Approximately three-quarters of the states have statutes that set forth a procedure to convert a manufactured home from personal to real property and document that conversion." The procedure generally involves surrendering the certificate of title or manufacturer's certificate of origin and filing an affidavit in the local land records, with the home permanently affixed. Appendix A lists the states with such statutes Confirmed Corporation for Enterprise Development and National Consumer Law Center, "Titling Homes as Real Property", 50-state review, Appendix A, https://www.nclc.org/wp-content/uploads/2022/09/cfed-titling-homes.pdf 2022 2026-08-10 2027-08 Opened under C-023 as the second leg of the titling argument, because statutory silence does not establish state practice. 42 U.S.C. 5403(d) preempts construction standards and says nothing about property classification (E-091) — that shows federal law left the question alone, not what the states then did. This row is what the states did, from the source that actually says something. THE INFERENCE, STATED SO IT CAN BE CHALLENGED: a statute providing a procedure to CONVERT a home from personal to real property only has work to do where the default is personal property. Three-quarters of states writing one is evidence of the default. It is not a count of states issuing certificates of title — see E-092, where that count could not be located at all.
E-121 A load-bearing housing statistic that circulates with no citation anywhere in its chain The claim that few manufactured homes are ever moved after placement is stated by the Lincoln Institute of Land Policy twice on a single page with no footnote, hyperlink or named study, and in two incompatible forms in adjacent sentences: "fewer than 5-7 percent of all manufactured homes are ever moved after initial placement" and "roughly 95 percent of homes are never moved after their initial installation". Pew repeats the figure. No primary source has been located at any point in the chain Confirmed Lincoln Institute of Land Policy, "Removing the Permanent Chassis Requirement for Manufactured Homes", https://www.lincolninst.edu/land-wise/removing-permanent-chassis-requirement-manufactured-homes/; Pew Charitable Trusts, "Proposal Could Lower Manufactured Home Costs, Expand Housing Supply", 10 November 2025 2025-11 2026-08-10 SITE literature 2027-08 This row records a FINDING ABOUT THE LITERATURE, not a housing fact, and it is graded Confirmed on that basis: what is confirmed is that these organisations publish the figure without a citation, which was read directly from the pages. The absent number itself stays [UNMEASURED] at E-093. Why it is worth a row: the figure is load-bearing for chassis reform across the policy literature — the permanent chassis is justified by transportability, and this is the statistic used to say transportability does not happen — and it varies between 5 and 7 percent inside one clause. A number that cannot hold its own value across a sentence has no measurement behind it. Belongs on /literature next to the Auckland and CORENET entries as an example of what the field has and has not established.
E-122 The chattel credit event that coincides with the 1998-2004 shipment collapse Securitization of manufactured-housing chattel loans "dropped to half the previous year's levels" in 2000 (JCHS/Ford Foundation). Against the Census shipment series for the same year — 348,102 in 1999 to 250,419 in 2000, a fall of 28.1% — the financing contracted about 1.8 times as fast as the volume it financed. The report also records an estimated 75,000 homes repossessed in 2000 (citing Stringer 2001) and two percent of outstanding manufactured-housing loans in repossession proceedings in 2001 (citing Walker Guido 2001) Supported William Apgar, Allegra Calder, Michael Collins and Mark Duda, "An Examination of Manufactured Housing as a Community- and Asset-Building Strategy", Report to the Ford Foundation by the Neighborhood Reinvestment Corporation in collaboration with the Joint Center for Housing Studies of Harvard University, September 2002, W02-11, https://www.jchs.harvard.edu/sites/default/files/w02-11_apgar_et_al.pdf; shipment denominator from US Census Bureau, Manufactured Housing Survey, annual shipments-to-states table (E-012) 2002-09 2026-08-10 HAP-08 shipments-collapse 2027-08 Corrected under C-025: the disproportion is stated on OUR OWN Census denominator, not the report's. The report says shipments fell about 20 percent in 2000, which would make the ratio 2.5x; Census gives 28.1 percent, which makes it 1.8x. Using their looser number would have flattered the argument by forty percent, so the figure we read directly is the one that counts and the report is cited only for the securitization half. Opened so the site states the competing explanation rather than waiting for a reviewer to supply it. THE OBJECTION: the 1998-2004 collapse was the manufactured-housing subprime bust — Green Tree, acquired by Conseco in 1998, and the securitized chattel lending that failed with it — not the property classification. That event is real and it maps onto the window exactly. WHY IT DOES NOT DISPLACE THE MECHANISM: it was a CHATTEL credit event, and chattel lending is what the vehicle title produces. The bust runs through the classification rather than around it. The reply concedes the trigger and keeps the cause. GRADE: Supported, not Confirmed. This is a serious research report, but the repossession figures are its citations of trade press (Stringer 2001, Walker Guido 2001), which have not been chased.
E-123 Whether repossession resale supply is separable from origination collapse in the shipment data PARTLY CLOSED. Census excludes the channel by construction: "The Manufactured Housing Survey does not include data on repossessed homes or resale units." No market-statistics series exists. But a lender filing carries the industry figure — Oakwood Homes' FY2002 10-K reports that "annual repossessions have increased to an estimated 90,000 units in 2002, more than 50% of new home shipments", and gives its own counts: 8,666, 11,727 and 13,423 homes repossessed or foreclosed in fiscal 2000, 2001 and 2002, with unsold repossession inventory rising from 2,603 to 5,940 to 7,063 Supported Oakwood Homes Corporation, Form 10-K for the fiscal year ended 30 September 2002, filed 14 January 2003, SEC accession 0000950144-03-000409, https://www.sec.gov/Archives/edgar/data/73609/000095014403000409/g79926e10vk.htm; US Census Bureau, Manufactured Housing Survey FAQ (exclusion of repossessed and resale units), https://www.census.gov/programs-surveys/mhs/about/faq.html; US GAO, GAO-07-879, 29 August 2007 2003-01-14 2026-08-10 HAP-08 shipments-collapse 2027-02 The lender-filings route worked where market statistics failed, which is the finding worth keeping: the channel is invisible in Census by design and visible in the accounts of the companies doing the repossessing. C-026. GRADE AND ITS LIMIT: the 90,000 is the COMPANY REPORTING AN INDUSTRY ESTIMATE, not a count it made — "the industry estimates" is the filing's own framing — so it is Supported, not Confirmed. Oakwood's own repossession counts are its own numbers and are firmer. One lender is not the market: Conseco Finance was the larger originator and its filings have not been read. WHAT IT DOES TO THE ARGUMENT: it makes the competing-supply objection BIGGER, not smaller. Repossessions at more than half of new shipments in 2002 is an enormous overhang. That is why the reply is E-124 rather than a denial — the wave's size and its speed are both consequences of the title.
E-124 What the title changes about repossession is the cost and time per unit, not the size of a wave Personal-property collateral is repossessed under UCC 9-609, which permits a secured party after default to take possession "(1) pursuant to judicial process; or (2) without judicial process, if it proceeds without breach of the peace" — self-help, no court. Real property requires foreclosure: judicial process, or a non-judicial procedure with statutory notice and redemption. GAO states the asymmetry: delinquent borrowers "can be subject to repossession (if the loan was for personal property) or foreclosure (if for real property), but the consumer protections for repossession are often less extensive than those for foreclosure" Confirmed UCC 9-609(a)-(b) (possession after default), https://www.law.cornell.edu/ucc/9/9-609; US GAO, "Federal Housing Administration: Agency Should Assess the Effects of Proposed Changes to the Manufactured Home Loan Program", GAO-07-879, 29 August 2007 2007-08-29 2026-08-10 HAP-08 shipments-collapse 2027-08 CLAIM NARROWED under C-028, and the earlier wording was wrong. This row was used to say no mortgage market could dump inventory at that speed. That does not survive 2008 to 2011, when foreclosure moved millions of site-built homes: real property can absolutely produce an enormous wave, and aggregate capacity was never the constraint. WHAT THE TITLE ACTUALLY CHANGES is the per-unit path. Section 9-609 removes the court from each individual repossession; foreclosure does not. That is a claim about time and cost to clear ONE unit, and it is what the statute supports. STILL UNMEASURED: no source has been located comparing repossession-to-resale duration or cost against foreclosure-to-resale for manufactured housing in this period. The mechanism is documented; the magnitude of the per-unit difference is not. Do not put a number on it.
E-125 The AutoCodes proof of concept and its finding ICC and Fiatech completed the proof-of-concept phase in January 2012, covering accessibility and egress in Chapters 10-11 of the 2009 International Building Code with 13 authorities having jurisdiction. Announced finding: "Review inconsistency is the norm, not the exception." Fiatech's own Phase II report states that "In Phase I, the Fiatech AutoCodes Project Team verified and documented the inconsistency in code reviews by different jurisdictions using the same base model code", and that those inconsistencies redirected Phase II toward outreach to code enforcement Confirmed International Code Council, "Code Council, Fiatech Partner to Complete First Phase of Automated Code Checking Technology", 16 May 2012, https://www.iccsafe.org/building-safety-journal/press-releases/code-council-fiatech-partner-to-complete-first-phase-of-automated-code-checking-technology/; Fiatech, "AutoCodes Project Team Phase II Report", September 2015, retrieved via the Internet Archive, https://web.archive.org/web/20170319231955/http://www.fiatech.org/images/stories/projects/FiatechAutoCodesPh2-Report-Sept2015.pdf 2012-05-16 2026-08-10 HAP-08 adoption-variables 2027-08 Both sources retrieved and read. This is the confirmed half; the quantified spread is E-126 and was not retrievable. Bears directly on Gate 0: these are the most objective chapters in the code, and the body that writes the code found review of them inconsistent enough to redirect its own project.
E-126 How far apart thirteen jurisdictions were on the same plan set Reported: a 2-D set of 137 plan sheets for a Target Corp. retail store prototype was submitted to 13 jurisdictions, with review limited to access and egress; responses ranged from one issue flagged in one jurisdiction to 43 in another, and the proof of concept was scaled back because the manual review process proved erratic Supported Engineering News-Record, "Making the Case for Automated BIM Review", 28 May 2012, https://www.enr.com/articles/8456-making-the-case-for-automated-bim-review (reporting Fiatech AutoCodes committee member Bill Gould) 2012-05-28 2026-08-10 HAP-08 adoption-variables 2027-02 REPORTED, NOT RETRIEVED — and the site says so wherever the figure appears. ENR returns HTTP 403; the Fiatech Phase 1 proof-of-concept report is 404 at source and archived only as a 404. Six routes were tried: the original fiatech.org URL, a direct Wayback fetch, the Wayback availability API, a CDX search across the whole fiatech.org domain, the Fiatech Phase II report (retrieved — confirms the finding, carries no numbers) and the AutoCodes funding prospectus (retrieved — no numbers). The figures come from two independent search-engine renderings of the ENR article that agree with each other. That is better than nothing and worse than reading it. WHAT WOULD CLOSE IT: the Fiatech AutoCodes Phase 1 Proof-of-Concept Final Report, or ENR access.
E-127 The American attempt to make model codes machine-readable, and why it stopped The ICC produced SMARTCodes in 2006, "containing official representations of a few important standards" and providing an authoring tool to manage code amendments. "Unfortunately, SMARTCodes development ended in 2010 due to a lack of funding" Confirmed Johannes Dimyadi and Robert Amor, "Automated Building Code Compliance Checking - Where is it at?", CIB World Building Congress 2013, https://wbc2013.apps.qut.edu.au/papers/cibwbc2013_submission_241.pdf; corroborated by Brian Potter, "The Ups and Downs of Automated Code Checking Software", Construction Physics, 27 May 2022, https://www.construction-physics.com/p/the-ups-and-downs-of-automated-code 2013 2026-08-10 HAP-08 adoption-variables 2027-08 Peer-reviewed conference paper, retrieved and read; Dimyadi is also a co-author of the Zou et al. study behind E-078 to E-084. Potter corroborates independently: the effort "lost funding post-financial crisis, and the project died". WHY IT COMPLICATES THIS PROJECT'S ARGUMENT rather than supporting it: a named budget cause is CONTINGENT. It is not encoding cost and not assent cost, so it cannot be counted as evidence for a structural diagnosis. Recorded because the beat that cites it must say so.
E-128 A third-party assessment of how CORENET ended Of Singapore's e-PlanCheck: it "didn't succeed (like many failures, it died silently, so the reasons are unclear)". The effort has since been revived as CORENET X on a different engine Confirmed Brian Potter, "The Ups and Downs of Automated Code Checking Software", Construction Physics, 27 May 2022, https://www.construction-physics.com/p/the-ups-and-downs-of-automated-code 2022-05-27 2026-08-10 HAP-08 adoption-variables 2027-08 What is Confirmed is that Potter assesses it this way, not that the assessment is correct — an important distinction on a row whose content is somebody's judgement. Cited because he is the same source this project concedes to on prefabrication cost (E-102): a source used against yourself is worth more when it is used for yourself. It is also the honest label for CORENET in the failure taxonomy — no named cause, an inference from silence, unlike SMARTcodes at E-127.
E-129 A second lender balance sheet on the repossession wave, from the larger originator Conseco Finance (formerly Green Tree) reports 24,131 unsold properties in repossession or foreclosure at 31 December 2001, against 20,110 a year earlier. On its managed manufactured housing portfolio it "liquidated 25,750 units at an average loss severity rate ... of 57 percent in 2001 compared to 23,861 units at an average loss severity rate of 54 percent in 2000", and liquidated about 70 percent of repossessed units through the retail channel Confirmed Conseco Finance Corp (formerly Green Tree Financial Corp), Form 10-K for the year ended 31 December 2001, filed 1 April 2002, SEC accession 0000890175-02-000002, https://www.sec.gov/Archives/edgar/data/890175/000089017502000002/cfcbody.txt 2002-04-01 2026-08-10 HAP-08 shipments-collapse 2027-08 The second direct observation C-027 said was owed, and it is roughly four times Oakwood's: 24,131 unsold units against Oakwood's 5,940 at a comparable date. One lender liquidated 25,750 units in 2001 — about 13 percent of that year's 193,120 new shipments, from a single balance sheet. Two lenders together account for something near 37,000 units of flow, which makes the 90,000 industry estimate on E-123 look conservative rather than inflated. THESE ARE THE COMPANY'S OWN NUMBERS, not an industry estimate, which is why this row is Confirmed where E-123 is Supported. IT ALSO CARRIES CONTEMPORANEOUS EVIDENCE ON SPEED: other lenders "have acted to more quickly dispose of repossessed manufactured housing inventory", and GreenPoint announced an objective to "quickly liquidate its repossessed inventory at below market prices in the wholesale market". Loss severity of 57 percent is what dumping looks like in an income statement.
E-130 Manufactured homes sold, 1996 against 2005 (a sold series, not a shipped one) GAP 332,000 sold in 1996 against 118,000 in 2005 — a 64.5% fall on a sold series, which tracks demand more closely than shipments do Untested US Government Accountability Office, report on manufactured housing — exact report number and title not yet retrieved HAP-08 shipments-collapse 2026-09-30 CITATION UNRESOLVED. Opened 2026-08-11 by the figure trace, which found the sentence on the story page and the deck with no row behind it. The figures are quoted from a GAO report this project has not opened; until it is retrieved and the numbers read from it, the grade asserts nothing. The claim is corroboration for E-012 and the argument does not rest on it. Retrieve or cut.
E-131 No single standard governs where a certified housing type may be sited GAP [UNMEASURED] — the SIZE of the siting fragmentation: how many distinct siting regimes govern where a certified home may go, and how far their binding provisions diverge. The existence of the construction/siting split is carried by E-091 and E-132, read from the statutes directly; this row is only its magnitude. Untested HAP-08 jurisdictional-fork 2026-09-30 Opened 2026-08-11 ahead of the restructure. WHAT WOULD SETTLE IT: an enumeration, for one state, of which siting provisions are set by a preempting standard and which are set locally — the same instrument as Gate 0 and answerable alongside it. Until then the beat must say that standard fragmentation is asserted and not yet evidenced. This row exists so that the reframe cannot ship as if it were sourced. 2026-08-13, C-034: narrowed to magnitude only — E-132 (A.R.S. §41-4006(D), Confirmed) now carries the existence half alongside E-091.
E-132 Arizona preempts manufactured-home construction standards and names siting as local, in one section A.R.S. §41-4006: no local building code may subject a unit certified under the chapter to local inspection "to determine compliance with any standard covering any aspect of the unit that is inspected pursuant to this article"; then subsection (D): "Nothing in subsection A, B or C of this section shall prevent the application of local codes and ordinances governing zoning requirements, fire zones, building setback, maximum area and fire separation requirements, site development and property line requirements and requirements for on-site utility terminals for factory-built buildings, manufactured homes and mobile homes." Confirmed Arizona Revised Statutes §41-4006, "Preemption of local building codes; responsibility for maintenance of utility connections" — https://www.azleg.gov/ars/41/04006.htm current statute 2026-08-13 HAP-08 jurisdictional-fork 2027-08-01 Read from the legislature’s own site and quoted verbatim. The construction/siting boundary drawn in a single section: subsections A–C preempt local inspection of anything the certification already covers, and (D) names what stays local — zoning, fire zones, setback, maximum area, fire separation, site development, property lines, on-site utility terminals. The federal counterpart is E-091 (42 U.S.C. §5403(d)). Together they carry the EXISTENCE of the construction/siting split; E-131 carries its unmeasured MAGNITUDE. E-049 asserts the same scope limit for modular programs generically and stays Untested for want of a named document — this row is the named document, for one state.
E-133 The model residential code adds far more than it deletes, and grew in both revision cycles measured 2015 IRC: 58 deletion indicators against 2,993 sections (1.94%). 2018 IRC: 42 against 3,064 (1.37%), a net gain of 71 sections over 2015. 2021 IRC: 78 against 3,182 (2.45%), a net gain of 118 over 2018. Section count rose 6.3% across six years without reversing; both measured cycles ended larger. Confirmed International Residential Code, 2015, 2018 and 2021 editions — counted from the publisher’s own marginal deletion indicators, whose meaning is stated in each edition’s "Marginal Markings" front matter. Full-text PDFs: 2015 https://cumming.iowa.gov/wordpress/wp-content/uploads/2023/07/2015-IRC.pdf ; 2018 https://archive.org/details/2018intlresidentialcode ; 2021 https://core-docs.s3.amazonaws.com/documents/asset/uploaded_file/3836/Lennox/3042182/2021_International_Residential_Code.pdf 2015, 2018, 2021 editions 2026-08-13 HAP-08 jurisdictional-fork 2027-08-01 METHOD: each edition states that a margin arrow marks where an entire section, paragraph, exception or table was deleted since the previous edition, so the code self-reports removals and no cross-edition diff is needed. The arrows are not text — the legend’s own arrow extracts as an empty parenthesis. In the body the mark is a ZapfDingbats glyph at exactly 10pt placed only in the outer margin; counted on that signature and confirmed by rendering image crops in two editions. Sections counted by heading pattern from the same text extraction. LIMITS: this is volume, not stringency — a code can add sections that liberalize, so the row supports "bigger", not "harder". Model residential code only; excludes local zoning and land use. Three editions, two intervals: free pre-2015 Archive copies are image scans with no text layer and no free 2024 full text exists. The arrow is generous (fires for a deleted list item as for a section), so removal counts are upper bounds against addition floors. Section denominators are approximate — an alternative regex gave 3,345/3,419/3,567; the growth is robust across both methods, the absolute counts are not. Change-bar counts were also collected but are a geometric heuristic and are deliberately not cited here. Code text was counted, never reproduced (ICC copyright).
E-134 No state or local construction-duration statistic exists in the United States Census tabulates the Survey of Construction for the United States and four Census Regions only. There is no state figure, and no records request produces one Confirmed US Census Bureau, Survey of Construction methodology: "Statistics from the SOC are tabulated only for the United States and four Census Regions." and "The SOC does not have a large enough sample size to make state or local area estimates." 2026-08 2026-08-17 2027-08 Both sentences quoted verbatim from the methodology page and verified 2026-08-17. Load-bearing for this project in a way that cuts both directions: it supports the claim that American housing delivery is unmeasured at the level jurisdictions actually operate at, and it forecloses any state-by-state comparison this site might otherwise be asked for. The sample is roughly 900 permit-issuing places drawn from 169 sampling units out of about 19,900. SOURCE: https://www.census.gov/construction/soc/methodology.html
E-135 Authorization-to-start gap, United States single-family 1.4 months mean for 2025 (all 1-unit buildings). The series runs 1.0 in 2018 to 1.5 in 2023 Confirmed US Census Bureau, Survey of Construction, "Average Number of Months from Authorization to Start" 2026-08 2026-08-17 2027-08 EVIDENCE AGAINST THE SITE'S OWN ATTRIBUTION, RECORDED AS SUCH. The gap between a permit being issued and a shovel entering the ground is the one post-authorization delay Census does measure, and in the United States it is small — 1.4 months, not years. Any account of the months around the build has to place them BEFORE authorization, because this row closes off the window after it. Verified 2026-08-17 by parsing the source workbook. SOURCE: https://www.census.gov/construction/nrc/xls/avg_authtostart_cust.xlsx
E-136 Permit-to-start idle gap, Netherlands, and its trend 8.1 months median in 2025, against 5.0 months in 2015 — a 62% increase. Construction over the same period went 9.4 to 13.3 months, a 41% increase Confirmed CBS (Statistics Netherlands) maatwerk, "Uitsplitsing doorlooptijden nieuwbouwwoningen 2015 - 2e kwartaal 2026" 2026-08 2026-08-17 2027-08 A full BAG register count, not a survey: 56,824 dwellings with a registered start in 2025, 82% of 69,189 completions. The Netherlands is the only country found that publishes the delivery period decomposed into a permit phase and a construction phase, so it is the only place this third phase is directly visible. It matters here because the waiting is neither approval nor construction, and it sits downstream of the Dutch statutory permit deadline — no permit-office reform reaches it. Verified 2026-08-17 by parsing the source workbook. SOURCE: https://www.cbs.nl/nl-nl/maatwerk/2026/33/uitsplitsing-doorlooptijden-nieuwbouwwoningen-2015-2e-kwartaal-2026
E-137 Outline planning determination time for major housing, England Mean 710 days in 2024, against 284 days in 2014 — an increase of 426 days, or 150%. Approvals take longer than refusals: 857 days against 453. Only 4% of 2024 applications were decided inside the statutory 13 weeks Supported Lichfields, "How long is a piece of string? The timescales for securing outline planning permission for housing between 2014 and 2024" 2024 2026-08-17 2027-08 Graded Supported rather than Confirmed because the analyst is a planning consultancy rather than a statistical agency, though it works from MHCLG application records. England is the clearest case found anywhere of approval time exceeding construction time by a wide margin, and the 857-versus-453 split says the delay is concentrated in getting to yes rather than in refusing. MHCLG holds the decision date for every application in England and publishes only percentages against statutory and renegotiated targets, which is why a consultancy is the best available source for a duration. Figures verified against the source PDF 2026-08-17. SOURCE: https://lichfields.uk/media/b2rewgxv/how-long-is-a-piece-of-string_the-timescales-for-securing-outline-planning-permission-for-housing-between-2014-and-2024.pdf
E-138 Construction duration by who commissions the house, United States Owner-built 12.8 months against 6.1 months built for sale in 2025 — 2.1x — while the owner-built house is the SMALLER of the two at 1,922 sq ft median against 2,216. Contractor-built sits between at 10.5 months Confirmed US Census Bureau, Survey of Construction (2025): "Average Length of Time from Start to Completion" by purpose of construction, and Characteristics of New Housing (2025): "Median and Average Square Feet of Floor Area in New Single-Family Houses Completed by Purpose of Construction" — https://www.census.gov/construction/nrc/xls/avg_starttocomp_cust.xlsx and https://www.census.gov/construction/chars/xls/squarefeet_cust.xls 2026-08 2026-08-17 2027-08 The size figures are what make this row worth carrying: the slowest category is the smallest, so the gap is not explained by building more house. It points at financing, part-time self-management and staged cash flow. Relevant to this project because it is a two-fold spread in CONSTRUCTION time inside one regulatory environment, which sets a floor on how much of any national average can be attributed to approval regimes at all. Verified 2026-08-17 by parsing the source workbook. SOURCE: https://www.census.gov/construction/nrc/xls/avg_starttocomp_cust.xlsx
E-139 Construction duration for an American single-family house, on a five-year basis 7.84 months, the 2021-2025 mean of the Census annual averages (7.2, 8.3, 8.6, 7.7, 7.4). This is what the site's "eight months of building" rounds from. Single years in the same series run 6.8 to 8.6 Confirmed US Census Bureau, Survey of Construction: "Average Length of Time from Start to Completion", Total 1-unit buildings, annual 1971-2025 — https://www.census.gov/construction/nrc/xls/avg_starttocomp_cust.xlsx 2026-08 2026-08-17 SITE landing (month strip) 2028-08 DERIVATION, SHOWN SO IT CAN BE CHECKED: (7.2 + 8.3 + 8.6 + 7.7 + 7.4) / 5 = 39.2 / 5 = 7.84. Added under C-039. The reason this row exists rather than a single-year one: the annual series is noisier than the claim built on it. 2023 peaked at 8.6 on pandemic supply chains and 2020 bottomed at 6.8, so a headline anchored to whichever year Census published last would have moved from eight to nine to eight to seven across four consecutive vintages without anything changing about how houses are built. Every window from three to seven years rounds to eight: 3-year 7.90, 5-year 7.84, 7-year 7.57. Census publishes MEANS only for this series and no median; E-002 carries the current-vintage figure and this row carries the basis the site quotes.
E-140 Countries publishing approval, waiting and construction as one regular series None of the 38 countries examined. 13 of 183 records carry all three phases at once, and every one of them splices two publishers, splices two years, or puts a statutory deadline where a measured approval time should be Supported Computed over the P1, idle-gap and P2 columns of "House build times dataset", 183 country-by-segment records across 38 countries, published at https://lintelengine.netlify.app/downloads/house-build-times-dataset.csv 2026-08 2026-08-17 SITE /international/ 2027-08 THE STRONGEST FINDING IN EITHER STUDY, AND IT IS ABOUT MEASUREMENT RATHER THAN CONSTRUCTION. The Netherlands comes closest and publishes two of the three — CBS decomposes the delivery period into a permit phase and a construction phase, which is E-136 — but its approval stage is a statutory deadline, not an observed duration. Graded Supported rather than Confirmed because this is an absence established across a compilation of 38 countries and not an exhaustive search of every statistical agency on earth: it says nobody found one, not that none exists. The compilation ships beside the page so the count can be re-run against the same columns.
E-141 Measured coverage of the 38-country build-time compilation 60 of 183 records carry a construction figure and 76 of 183 carry an approval figure. 26 of 38 countries have any construction duration at all; 12 have none Confirmed Counted from "House build times dataset", published at https://lintelengine.netlify.app/downloads/house-build-times-dataset.csv 2026-08 2026-08-17 SITE /international/ 2027-08 A 38-country comparison is true of the scope and not of the measurements, and a row count flatters an evidence base. Carried as its own row because the page states it: the number of countries examined and the number actually measured are different numbers, and quoting the first while resting on the second is the ordinary way a comparison of this kind misleads. The twelve with no construction figure anywhere this compilation looked are Argentina, Austria, Denmark, Finland, Indonesia, Italy, Norway, Philippines, Spain, Switzerland, Thailand and Vietnam.
E-142 Elapsed time from first application to final inspection, Chilean housing projects 2,001 days in 2025 against 1,298 days in 2018 — an increase of 703 days, or 54%. Municipal processing accounts for 471 days of the 2025 total Supported Colliers, "Ranking Pro Vivienda" 5th edition, reported 2026-05-28 — https://www.biobiochile.cl/noticias/economia/actualidad-economica/2026/05/28/tramitacion-y-ejecucion-de-proyectos-de-vivienda-toma-700-dias-mas-que-hace-siete-anos-segun-analisis.shtml 2026-08 2026-08-17 SITE /international/ 2027-08 Verified against the source 2026-08-17. Panel of more than 1,250 projects and more than 2,400 municipal transactions. Graded Supported because the analyst is a property consultancy rather than a statistical agency. DO NOT READ THE REMAINDER AS A CONSTRUCTION DURATION: subtracting the municipal stages from the total leaves a residual that also absorbs the idle gap, financing waits and staged sales conditions, and the compilation flags this in capitals on its own row. Chile is carried here for its approval stage, which is measured, and not for its construction stage, which is not.
E-143 Observed building-permit stage against the statutory deadline, Chile 223 days observed for the building permit alone against a statutory deadline of 30 days — 7.4 times the legal ceiling. The full pre-construction sequence runs 372 days Supported Observed durations from Colliers, "Ranking Pro Vivienda" 5th edition — https://www.biobiochile.cl/noticias/economia/actualidad-economica/2026/05/28/tramitacion-y-ejecucion-de-proyectos-de-vivienda-toma-700-dias-mas-que-hace-siete-anos-segun-analisis.shtml ; statutory deadline from Chile Law 21.718 (2024), summarised at https://www.dlapiper.com/es-pe/insights/publications/2024/12/chile-law-no-21718-regarding-the-expediting-of-construction-permits 2026-08 2026-08-17 SITE /international/ 2027-08 Chile has the shortest statutory permit deadline found anywhere in this compilation and one of the longest observed permit stages. ANY CROSS-COUNTRY COMPARISON BUILT ON LEGAL DEADLINES RATHER THAN OBSERVED TIMES GETS CHILE EXACTLY BACKWARDS, which is the most useful single fact in the compilation and the reason this row exists. Law 21.718 sets 30 days for a standard building permit and 60 days where occupancy is 1,000 or more, each halved when an independent reviewer files a favourable report, and introduces negative administrative silence: a missed deadline lets the applicant treat the application as rejected and escalate to the regional housing secretariat, which is a remedy against inaction rather than a grant. Both halves verified 2026-08-17.
E-144 Municipal approval time for multi-unit residential, Canada 11.2 months mean across 5,232 applications, 2022 to 2024. Saskatoon 2.0 months against Hamilton 31.0 months. Ontario averages 18.8 months and the Prairies 4.6 months Supported Canadian Home Builders' Association, "Municipal Benchmarking Study" 3rd edition, March 2025 — https://www.chba.ca/wp-content/uploads/2025/03/chba-municipal-benchmarking-study-3rd-edition-2024_compressed.pdf 2026-08 2026-08-17 SITE /international/ 2027-08 Verified against the source PDF 2026-08-17. Graded Supported because the analyst is an industry association rather than a statistical agency; CMHC independently describes municipal review for multi-unit as averaging about a year, which corroborates the level. The spread matters more than the mean — more than fifteen times between the fastest and slowest municipality, inside one country, one building type and one three-year window, so it is not explained by climate, technology or product. The five slowest municipalities are all in Ontario and the five fastest are in the Prairies or Atlantic Canada.
E-145 Construction duration by building type, Poland 51 months for a single-family house against 24.8 months for a multi-family building in 2025. The statistical office's own all-buildings average of 42.2 months is weighted by building rather than by dwelling Confirmed Glowny Urzad Statystyczny (Statistics Poland), "Efekty dzialalnosci budowlanej w 2025 r." — https://stat.gov.pl/obszary-tematyczne/przemysl-budownictwo-srodki-trwale/budownictwo/efekty-dzialalnosci-budowlanej-w-2025-r-,3,21.html 2026-08 2026-08-17 SITE /international/ 2027-08 Quoted from the publication 2026-08-17: single-family buildings took more than twice as long as multi-family, 51 months against 24.8, and the all-buildings average was 42.2 months, down 1.4 months on the previous year. THE COUNTING METHOD IS THE POINT. 42.2 is weighted by building, so one detached house counts the same as a hundred-dwelling block, and the overwhelming majority of Polish residential buildings by count are single-family — a national average build time can be dominated by how the average is taken rather than by how anything is built. Poland measures construction start to handover only, so this is not comparable to a permit-to-completion total.
E-146 Private against public building-confirmation review times, Aichi Prefecture, Japan On the routine class that needs no structural conformity review, designated private bodies take 7 days against 11 days for the administrative authority. On the class that does need it the order reverses: 38 days private against 34 days public Supported Aichi Prefecture Building Guidance Division, "average building confirmation review days" — https://www.pref.aichi.jp/soshiki/kenchikushido/kenchiku-shinsanissuu.html 2026-08 2026-08-17 SITE /international/ 2027-08 Read from the prefectural page 2026-08-17. The page publishes a rolling current figure and these values move, so the retrieval date is load-bearing. Same statute, same drawings, two processors — which is why it is worth a row: the delegation gain is real and it is BOUNDED, and it shrinks to nothing or reverses exactly where the review becomes substantive. The published counts include the applicant's own correction time and the fire-service consent, not only the examiner's. Graded Supported: one prefecture, and a live administrative page rather than a statistical release. The compilation this row came from quoted 10 days against 35 for financial year 2025; that pair could not be reproduced from the page as it now stands, so the figures carried here are the ones actually on it.
E-147 Published permit service standard, Saudi Arabia 1 to 10 days end to end for a building permit, as the Balady platform publishes it Supported Balady Platform (MOMRAH), "Issuing Building Permit" service card — https://balady.gov.sa/en/services/issuing-building-permit 2026-08 2026-08-17 SITE /international/ 2027-08 A PUBLISHED SERVICE STANDARD AND NOT AN OBSERVED DURATION, which is why this is Supported rather than Confirmed. Verified on the service card 2026-08-17. It is carried for the mechanism rather than the number: the published workflow moves substantive code review onto accredited private engineering offices and technical inspection companies that check building-code compliance and carry liability, so the figure measures what is left at the counter after the review has been moved rather than the review itself. How long the private stage takes is not published anywhere found, which makes the headline unfalsifiable as a delivery figure and is the reason it should not be used as one.
E-148 Schedule effect of mandated volumetric modular construction, Singapore 1 to 2 months saved on average against conventional methods, with some projects saving 4 months or more. Mandatory since 1 November 2014 for selected non-landed residential Government Land Sales sites Confirmed Building and Construction Authority, "Prefabricated Prefinished Volumetric Construction" — https://www1.bca.gov.sg/buildsg/productivity/design-for-manufacturing-and-assembly-dfma/prefabricated-prefinished-volumetric-construction-ppvc 2026-08 2026-08-17 SITE /international/ 2027-08 Quoted from the authority's own page 2026-08-17. THE CORRECTIVE MATTERS MORE THAN THE FIGURE. The widely repeated "up to 40%" is a productivity claim about manpower and time at activity level, not an end-to-end schedule saving; the end-to-end saving the authority itself publishes is 1 to 2 months. This is the closest thing in the compilation to a controlled test of construction technology under a mandate, in the one jurisdiction that made it compulsory, and it is a large part of why this project does not treat building faster as the lever. No percentage of a total cycle is computed here, because the cycle length that figure would be drawn against is not published beside it.
E-149 Share of Mexican owner-occupied dwellings that were self-built rather than bought 57.3% were built by their occupants against 35.4% purchased, 2020 Confirmed INEGI, "Encuesta Nacional de Vivienda (ENVI) 2020" press release, 2021 — https://www.inegi.org.mx/contenidos/saladeprensa/boletines/2021/envi/ENVI2020.pdf 2026-08 2026-08-17 SITE /international/ 2027-08 Quoted from the release 2026-08-17. Sample of 55,147 dwellings representing 35.3 million occupied private dwellings. LOAD-BEARING FOR EVERY PERMIT COMPARISON ON THE PAGE: permit statistics only see permitted housing, and here the majority of the owner-occupied stock was not produced through the channel a permit duration measures. A country can hold the fastest formal permitting in its region and still deliver most of its housing outside that system, so a fast permit statistic is not evidence that housing is arriving quickly. This cuts against this project as much as for it — it limits what any national permitting figure, including an American one, can be said to describe.
E-150 Completion time for self-built housing in suburban Nigerian cities 8.2% finish inside 1 to 2 years, 56.59% take 2 to 4 years, 19.92% take 4 to 6, 8.59% take 6 to 8, 5.08% take 8 to 10 and 1.17% take more than 10. The interpolated median is 41.7 months Supported Fagbohun, Shotunde, Oladiboye and Ajaegbo, "Reliability of Homeownership Through Self-Build Approach in The Suburban Area of Nigeria Cities", International Journal of Research and Innovation in Social Science 6(8), 2022, pp. 214-220 — https://rsisinternational.org/journals/ijriss/Digital-Library/volume-6-issue-8/214-220.pdf 2026-08 2026-08-17 SITE /international/ 2027-08 DERIVATION, SHOWN SO IT CAN BE CHECKED: the median falls in the 2-to-4-year class, so 2 + ((50 - 8.2) / 56.59) x 2 = 3.48 years = 41.7 months. PROVENANCE DEFECT, RECORDED RATHER THAN SMOOTHED: the compilation that supplied this row dated the paper 2020 and described it as a survey of Lagos-metropolitan self-builders. The paper is dated 2022, is a review article relaying a distribution from work it does not fully identify, and covers suburban areas of Nigerian cities rather than Lagos specifically. The distribution above was read from the paper 2026-08-17 and is what this row rests on. Graded Supported and no higher for exactly that reason.
E-151 Building consent processing against a statutory deadline, New Zealand Median 13.0 working days across 16,532 applications in the first quarter of 2026, with 94.5% decided inside the 20 working day statutory period. Code compliance certificates run a median of 4.0 working days across 9,902 applications, 97.2% inside the statutory period Confirmed New Zealand Ministry of Business, Innovation and Employment, "Building Consent System: Performance Monitoring", Q1 January to March 2026 — https://www.mbie.govt.nz/dmsdocument/32005-building-consent-system-performance-monitoring-q1-january-to-march-2026 2026-08 2026-08-17 SITE /international/ 2027-08 Verified against the document 2026-08-17. The fastest regulatory stage found anywhere in this compilation, and it is carried as COUNTER-EVIDENCE: approval is not slow everywhere, and where a statutory clock is short, published and monitored, it is met. THE CLOCK STOPS during a request for further information, so the published figure measures the authority's own time and not the applicant's elapsed wait; the same document reports a median applicant response of 9.0 working days where a request is issued. The compilation this row came from also carried a 64.3% figure for the share of applications receiving such a request; that figure is not in the document and is not carried here.
E-152 How Ireland defines a housing completion A new dwelling is counted when an ESB Networks domestic electricity connection is energised, filtered on the interval between authorisation and energisation and validated against building energy ratings and census records Confirmed Central Statistics Office Ireland, "New Dwelling Completions" background notes — https://www.cso.ie/en/methods/surveybackgroundnotes/newdwellingcompletions/ 2026-08 2026-08-17 SITE /international/ 2027-08 Verified 2026-08-17. The most objective completion definition found in the compilation, and carried for what it exposes rather than for what Ireland is doing well: a completion here is a physical event recorded by a utility, not a form filed by the party whose schedule is being measured. Every duration in this compilation inherits the definition of its endpoint, and most countries define completion administratively. A cross-country build-time comparison is therefore partly a comparison of when different countries decide to stop the clock, which is a limit on this page and on every other comparison of its kind.
E-153 Status of the only cross-country construction-permit duration series Discontinued on 16 September 2021, after the World Bank found data irregularities in the 2018 and 2020 editions Confirmed World Bank Group, "World Bank Group to Discontinue Doing Business Report", statement of 16 September 2021 — https://www.worldbank.org/en/news/statement/2021/09/16/world-bank-group-to-discontinue-doing-business-report 2026-08 2026-08-17 SITE /international/ 2027-08 Verified against the statement 2026-08-17. Doing Business is the reason a cross-country permitting comparison feels like something that already exists: it was the only dataset that ever published a comparable permitting duration across nearly every economy, and almost every secondary "how long does a permit take" comparison still standing is built on it. It has been gone for five years. Its construction-permit measure was also scoped to a standardised warehouse rather than to housing, which is a second and independent objection to using it for this question; that scope is stated in its own methodology and was not re-verified here.
E-154 Project approval delay by local regulatory stringency, United States 3.7 months in the least-regulated quartile of municipalities, 5.0 months across the interquartile range and 8.4 months in the most-regulated quartile, with a tail of communities reporting 18 to 24 months Supported Gyourko, Hartley and Krimmel, "The Local Residential Land Use Regulatory Environment Across U.S. Housing Markets: Evidence from a New Wharton Index", NBER working paper 26573 — https://www.nber.org/papers/w26573 2026-08 2026-08-17 SITE /international/ 2027-08 Figures read from the working paper 2026-08-17, including its own statement that the 18 to 24 month tail was checked and is not survey misreporting. Graded Supported because it is a survey of local officials rather than an administrative measurement of decided applications, and because the survey has not been re-run since 2018. Set beside E-139 it says something sharp: in the most regulated quartile, getting permission takes longer than building the house. The index publishes no state table at this wave — its finest geography is the metropolitan area — so it cannot support a ranking of states, which is the use it is most often put to.
E-155 Entitlement and building-permit time, San Francisco 523 days on average to entitle a housing project, and a further 605 days to issue a building permit to an already entitled project, calendar year 2022 Confirmed California Department of Housing and Community Development, "San Francisco Housing Policy and Practice Review", October 2023 — https://www.hcd.ca.gov/sites/default/files/docs/policy-and-research/plan-report/sf-housing-policy-and-practice-review.pdf 2026-08 2026-08-17 SITE /international/ 2027-08 Quoted from the review 2026-08-17: an average of 523 days for a housing project to be entitled, against 385 days for the next slowest jurisdiction in the state, and an average of 605 days to issue a building permit to an already entitled project, against 418 days in the next slowest. Confirmed because it is a state housing agency measuring one of its own cities from that city's statutory filings. THE COMPARATORS MATTER AS MUCH AS THE HEADLINE: the next slowest jurisdiction in California sits at 385 days and 418 days, so San Francisco is an outlier inside its own state and must never be read as a Californian or an American figure. It is the most-quoted permitting number in the country and the least representative.
E-156 Share of housing delivery time spent getting permission, Los Angeles County 4.2 years on average from permit submission to occupancy for a standardised 30-unit building, roughly twice Raleigh and Fort Worth, with time-to-permit about 40% of total time-to-build Supported Soltas and Gruber, "How Costly Is Permitting in Housing Development?", February 2026 — https://evansoltas.com/papers/Permitting_SoltasGruber2026.pdf 2026-08 2026-08-17 SITE /international/ 2027-08 Quoted from the paper 2026-08-17. Graded Supported because it is an unrefereed working paper. It is the closest thing in the American literature to a like-for-like cross-city duration comparison, because it holds the building constant instead of comparing whatever each city happened to build — which is the defect in every ranking of cities by permit time this project has examined. NOTE WHAT THE SHARE DOES NOT SAY: it is a share of one metropolitan area's total, not a claim that abolishing permitting would return that share of delivery time anywhere else.
E-157 What England publishes about how long a planning decision takes No duration. 91% of major applications were decided within 13 weeks or within an agreed extended time in the first quarter of 2026, against 19% decided within the statutory 13 weeks. 77% of major decisions involved a performance agreement Confirmed Ministry of Housing, Communities and Local Government, "Planning applications in England: January to March 2026" statistical release — https://www.gov.uk/government/statistics/planning-applications-in-england-january-to-march-2026/planning-applications-in-england-january-to-march-2026-statistical-release 2026-08 2026-08-17 SITE /international/ 2027-08 Verified against the release 2026-08-17. THE CLOSING ARGUMENT OF THE INTERNATIONAL PAGE RESTS ON THIS ROW. The department holds the application date and the decision date for every planning application in England and publishes neither a mean nor a median — only compliance percentages against a target that can be renegotiated after the clock has started. The gap between the two percentages above is what that renegotiation is worth. It is why the best available England duration is a planning consultancy's analysis of the department's own records at E-137 rather than the department's own statistic, and it is the same shape as the American gap recorded at E-134: the join exists, it is held by a public body, and it is not published.
E-158 Depth of the market that would have to write this cover Seventeen named carriers participate in the annual A/E professional liability market survey, including AIG/Lexington, Beazley, Berkley Design Professional, Travelers and Victor Confirmed Insurance Journal, "Survey: Most Architects and Engineers Professional Liability Rates to Rise", 2024-04-01 — https://www.insurancejournal.com/magazines/mag-features/2024/04/01/766865.htm 2026-08 2026-08-17 SITE /solution/ 2027-08 Verified 2026-08-17 by opening the article. The line of coverage a conformance determination would need is architects and engineers professional liability, and it is a deep admitted market rather than a novel one.
E-159 Statutory insurance floor for private plan reviewers, Florida $1M per occurrence / $2M aggregate below $5M construction cost; $2M / $4M above it; A.M. Best A minimum; five-year tail on claims-made; certificate filed with the building official before work Confirmed Fla. Stat. 553.791(18) — https://www.flsenate.gov/Laws/Statutes/2025/553.791 2026-08 2026-08-17 SITE /solution/ 2027-08 Read in the statute 2026-08-17 and cross-checked against leg.state.fl.us. A legislature has already priced this exposure and named the form.
E-160 Whether a jurisdiction accepting substituted review keeps its own exposure, Florida It does not. The local government, the building official and code enforcement personnel are immune from liability for any action or inaction by a private provider Confirmed Fla. Stat. 553.791(21) — https://www.flsenate.gov/Laws/Statutes/2025/553.791 2026-08 2026-08-17 SITE /solution/ 2027-08 Read in the statute 2026-08-17. LOAD-BEARING AND IT CUTS BOTH WAYS: Florida did not solve acceptance and insurability separately. The same section immunises the jurisdiction and mandates rated cover, so the immunity is what buys acceptance and the policy is what makes the immunity survivable. This site currently proposes only the second half.
E-161 Whether a professional liability policy covers a document that certifies rather than opines UNRESOLVED. The standard A/E form excludes express warranties and guarantees, carving back only that services meet the applicable standard of care Supported Professional Underwriters, "Common Exclusions on the Architects/Engineers Professional Liability Policy" — https://www.profunderwriters.com/common-exclusions-on-the-architectsengineers-professional-liability-policy/ 2026-08 2026-08-17 SITE /solution/ 2027-08 Read 2026-08-17. Graded Supported rather than Confirmed because this is a broker description of policy language, not a filed form. THE THREAT TO THE THESIS: a determination reads as a finding of fact. If it is insurable only when worded as a professional opinion, the instrument is weaker than this site describes. Obtain an actual form to settle it.
E-162 What the largest A/E liability programme tells its own insureds about certifying Certifications must clearly express a professional opinion and be qualified as limited in use; certifications beyond the scope of services or available knowledge are listed under accepting exposures Supported Musica, Victor O. Schinnerer & Co., "Managing Professional Exposure as a Small Project Practitioner", The AIA Trust, 2015-10-28 — https://theaiatrust.com/wp-content/uploads/2021/03/Managing-Professional-Exposure-as-a-Small-Project-Practitioner.pdf 2026-08 2026-08-17 SITE /solution/ 2027-08 Slides 9 and 13, read directly 2026-08-17. The manager of the AIA-endorsed programme warns against exactly the posture a conformance determination takes.
E-163 Severity of professional liability claims against design professionals 23% of seventeen surveyed insurers paid a claim of $5M or more in 2023; 12% paid $10M or more; structural engineering ranked highest severity at 82% Supported Insurance Journal, "Survey: Most Architects and Engineers Professional Liability Rates to Rise", 2024-04-01 — https://www.insurancejournal.com/magazines/mag-features/2024/04/01/766865.htm 2026-08 2026-08-17 SITE /solution/ 2027-08 Read 2026-08-17. Supported rather than Confirmed: trade press reporting a private survey whose underlying data is not public.
E-164 What a design professional's liability claims are actually about By severity, project owner or client 71%, general contractor 11%, third-party property damage 6%, non-worker bodily injury 4%, worker bodily injury 3% Supported Victor O. Schinnerer & Co. / CNA programme data 2004-2013, in "Managing Professional Exposure as a Small Project Practitioner", The AIA Trust, 2015-10-28 — https://theaiatrust.com/wp-content/uploads/2021/03/Managing-Professional-Exposure-as-a-Small-Project-Practitioner.pdf 2026-08 2026-08-17 SITE /solution/ 2027-08 Percentages read off labelled charts on slides 4 and 5, 2026-08-17. CUTS AGAINST AN ASSUMPTION ON THIS SITE: bodily injury is roughly 7% of severity, so the dominant loss from a wrong determination is owner economic loss, not the egress miss the specimen implies.
E-165 Whether HUD requires its production inspection agencies to carry insurance No such requirement found. The sections governing acceptance, contracts, background and duties impose no insurance, bond, surety or indemnification obligation Supported 24 CFR 3282.202, 3282.355, 3282.357, 3282.362 — https://www.law.cornell.edu/cfr/text/24/part-3282/subpart-H 2026-08 2026-08-17 SITE /solution/ 2027-08 Five sections plus the subpart index read 2026-08-17. SCOPED HONESTLY: this is not a full-text sweep of Part 3282, which was blocked. Graded Supported for that reason. The contrast with Florida is the finding — the two substituted-review regimes in American law solve the same problem oppositely, one by mandating cover, the other by retaining federal revocation power.
E-166 Adoption of the interstate modular building compact Three member states after thirty-four years: New Jersey, North Dakota and Minnesota. The compact took effect on enactment by three states and its commission was created in 1992 Confirmed Interstate Industrialized Buildings Commission — https://interstateibc.org/faq/ and https://interstateibc.org/about-us/ 2026-08 2026-08-17 SITE /solution/ 2027-08 Membership as stated by the commission itself, retrieved 2026-08-17. NARROWS A CLAIM THIS SITE MAKES: interstate reciprocity exists in law and is barely adopted, which is a weaker and more interesting fact than the mechanism simply existing.
E-167 Whether the interstate reciprocity compact allocates liability It does not. No provision on insurance, bonding, indemnification or liability appears in any of the thirteen articles, and Article 12 expressly preserves every state court's jurisdiction Confirmed N.D.C.C. ch. 54-21.4, Interstate Compact on Industrialized/Modular Buildings — https://ndlegis.gov/cencode/t54c21-4.pdf 2026-08 2026-08-17 SITE /solution/ 2027-08 All thirteen articles read in full 2026-08-17. The compact moves regulatory acceptance across state lines and leaves every state's tort exposure where it was.
E-168 Why interstate certification reciprocity was not widely adopted GAP UNMEASURED. No report, hearing record or study attributing low adoption to insurance, liability or any other cause was found Untested GAP — no document located 2026-08 2026-08-17 SITE /solution/ 2027-08 GAP. This site has implied that insurance killed certification-reciprocity schemes more often than politics did. Nothing found on 2026-08-17 supports that. The honest claim is that the mechanism exists, is barely adopted, contains no liability provision, and the cause is unknown. Do not restate the stronger version until this row is filled.
E-169 Arizona third-party review of overdue single-family permits A.R.S. 9-470.01: if a municipality of 30,000 or more misses 15 working days on a single-family permit, review may be performed by a qualified third party selected by the municipality, which must maintain a list of at least three eligible reviewers Confirmed SB 1353, Chapter 187, Laws 2025 (Arizona, 57th Leg. 1st Reg. Sess.), signed 2025-06-03 — https://www.azleg.gov/legtext/57leg/1R/laws/0187.pdf 2026-08 2026-08-17 SITE /solution/ 2027-08 Enacted text read 2026-08-17. TWO DETAILS THAT MATTER: the section does not define qualified third party — the Senate fact sheet's ICC/PE/architect language did not survive into it — and the municipality selects the reviewer, not the applicant. The jurisdiction is immunised under A.R.S. 12-820.01 and 12-820.02.
E-170 Whether any Arizona municipality has used the third-party review fallback GAP UNMEASURED. No count of referrals under A.R.S. 9-470.01, and no published list of eligible third-party reviewers, was located Untested GAP — no document located 2026-08 2026-08-17 SITE /solution/ 2027-08 GAP, and cheap to close: it is a records request or a phone call to one building official. Fourteen months after the statute took effect, whether the mechanism is live or dormant is unknown, and either answer is a finding.
E-171 Whether a legislature anticipated the accepting jurisdiction's exposure Florida's immunity clause was in the original 2002 enactment as subsection (18), renumbered to (19) in 2011, (20) in 2023 and (21) in 2025 Confirmed Fla. Stat. 553.791(21); ch. 2002-293, Laws of Florida — https://www.flsenate.gov/Laws/Statutes/2025/553.791 2026-08 2026-08-17 SITE /solution/ 2027-08 Verified 2026-08-17. The immunity was not bolted on after a loss — it was written into the enabling act on day one. That is the answer to the question a city attorney asks first.
E-172 What a jurisdiction gives up in exchange for the immunity The right to look again. The local building official may not replicate the plan review or inspection being performed by the private provider; audits are capped at four per provider per year and may not delay the work Confirmed Fla. Stat. 553.791(1)(b), (14), (15), (20) — https://www.flsenate.gov/Laws/Statutes/2025/553.791 2026-08 2026-08-17 SITE /solution/ 2027-08 Read 2026-08-17. THE COST IS REAL AND THIS SITE HAS NOT STATED IT: a jurisdiction cannot buy the immunity and keep the second review. The certificate of occupancy is also automatically granted if the official misses ten business days, or two for one- and two-family dwellings.
E-173 Indiana's private plan review regime Seven business days for plan review, three for inspection, private provider use unconditional; unit shall accept the report without further inspection or approval; unit and its agents immune for the provider's acts Confirmed Indiana House Enrolled Act 1005 (2025), Pub. L. 146-2025 sec. 4, adding IC 36-7-2.5 secs. 17, 18, 26, 32, 34; effective 2026-01-01 2026-08 2026-08-17 SITE /solution/ 2027-08 Enrolled act read verbatim 2026-08-17. Insurance floor is $1M per claim / $2M aggregate with prior-acts coverage. The immunity is also indexed at IC 34-30-2.1-583.7. Supersedes the description previously carried by E-050 — see C-040.
E-174 Whether the statutory immunity covers a jurisdiction's own negligence GAP UNMEASURED. Florida and Indiana both immunise the unit for acts of the owner or the private provider. Neither text reaches the jurisdiction's own failure to verify licensure or insurance, its own certificate decision, or its own safety judgement Untested GAP — no construing decision located 2026-08 2026-08-17 SITE /solution/ 2027-08 GAP. Read from the statutes 2026-08-17, and the residual is on the face of them. No court has construed it either way. This is the honest answer to the city attorney and it is not reassurance.
E-175 Whether any court has construed a private-provider immunity clause No reported appellate decision located in twenty-four years of Florida operation. Full-text search returned two Florida Attorney General opinions from 2006, both on notice timing under 553.791(4), neither on liability Confirmed CourtListener full-text search of 553.791 and related terms, 2026-08-17 — https://www.courtlistener.com 2026-08 2026-08-17 SITE /solution/ 2027-08 SCOPED DELIBERATELY: this is an absence in reported appellate decisions to the depth of one index. CourtListener does not cover state trial courts, and settlements are unreported. The publishable claim is that no reported decision construes the immunity, NOT that no jurisdiction has ever been sued.
E-176 Whether HUD's manufactured-housing framework is precedent for a jurisdiction accepting an outside determination It is not. HUD preempts rather than delegates: no state may require its own inspection for anything the federal standards cover, and the regulations are the exclusive enforcement system Confirmed 24 CFR 3282.11(b), 3282.11(c) — https://www.ecfr.gov/current/title-24/part-3282 2026-08 2026-08-17 SITE /solution/ 2027-08 Full part text retrieved and searched 2026-08-17 for liability, indemnity and immunity terms; every hit concerns private parties. CORRECTS A CATEGORY ERROR THIS SITE RISKS MAKING: under HUD the local authority is displaced, not delegating, so there is no acceptance decision to get wrong and no liability provision is needed.
E-177 A third statutory model, where the jurisdiction keeps the duty Washington DC: third-party review shall not relieve the District of its obligation to review all construction documents in the manner otherwise prescribed by law. No immunity clause and no insurance requirement in the section Confirmed D.C. Code 6-1405.02(i) — https://code.dccouncil.gov/us/dc/council/code/sections/6-1405.02 2026-08 2026-08-17 SITE /solution/ 2027-08 Read verbatim 2026-08-17. Three models exist in American law and this site has described only one: immunity (FL, IN), preemption (HUD), and retained duty (DC). Which one a state adopts decides whether the instrument is adoptable there.
E-178 Arizona's immunity for acts of a public entity's contractor Qualified immunity unless intent to injure or gross negligence, extending to injury caused by a contractor of a public entity acting within the scope of the contract, and expressly not to the contractor Confirmed Ariz. Rev. Stat. 12-820.02(A)(5), (A)(6), (B) — https://www.azleg.gov/ars/12/00820-02.htm 2026-08 2026-08-17 SITE /solution/ 2027-08 Read verbatim 2026-08-17. Relevant because A.R.S. 9-470.01 has the MUNICIPALITY select the third-party reviewer, so the reviewer is the entity's contractor rather than the applicant's — a different posture from Florida and Indiana, and one this immunity appears to reach.
E-179 The background rule where no private-provider statute exists Split, and it decides adoptability. Washington holds no duty is owed for negligent building-code enforcement; Wisconsin, Alaska and Louisiana imposed one; Colorado, Arizona and Oregon abolished the public duty doctrine outright Confirmed Taylor v. Stevens County, 111 Wn.2d 159, 759 P.2d 447 (1988); Coffey v. City of Milwaukee, 74 Wis. 2d 526, 247 N.W.2d 132 (1976); Adams v. State, 555 P.2d 235 (Alaska 1976); Stewart v. Schmieder, 386 So. 2d 1351 (La. 1980); Leake v. Cain, 720 P.2d 152 (Colo. 1986); Ryan v. State, 134 Ariz. 308, 656 P.2d 597 (1982); Brennen v. City of Eugene, 285 Or. 401, 591 P.2d 719 (1979) 2026-08 2026-08-17 SITE /solution/ 2027-08 Every opinion above was retrieved in full and read 2026-08-17 via the Harvard Caselaw Access Project, with name, citation, court and year cross-checked against CourtListener. Taylor: 'no duty is owed by local government to a claimant alleging negligent issuance of a building permit or negligent inspection'. Stewart is the one squarely on plan review — Baton Rouge issued on unreviewed structural plans, the building collapsed, three died, and the city-parish was held liable. NOTE: Stewart declined to apply the doctrine on special-duty grounds rather than abrogating it, so Louisiana is not an abrogation state on this record.
E-180 What a jurisdiction's exposure is capped at where immunity is waived Florida: $200,000 per person and $300,000 per incident Confirmed Fla. Stat. 768.28(5)(a), 2025 edition — https://www.flsenate.gov/Laws/Statutes/2025/768.28 2026-08 2026-08-17 SITE /solution/ 2027-02 Read 2026-08-17. Caps are periodically amended; re-check the figure before it is quoted anywhere that matters. Review set early for that reason.
E-181 Whether a legislature can restore inspection immunity a court removed It can, and three have. California immunises negligent inspection absolutely; Nevada codifies the public duty doctrine by name and forecloses the undertaking theory; Alaska immunises municipalities only, leaving the State exposed Confirmed Cal. Gov't Code 818.6 and 821.4 (Stats. 1963, ch. 1681); Nev. Rev. Stat. 41.033; Alaska Stat. 09.65.070(d)(1) (ch. 37 sec. 3, SLA 1977) 2026-08 2026-08-17 SITE /solution/ 2027-08 All read verbatim 2026-08-17. Nevada is the strongest: it bars an action for failure to inspect 'whether or not there is a duty to inspect', which forecloses the voluntary-undertaking theory that won in Coffey and Adams. Arizona is the clearest sequence — Ryan removed the public/private duty distinction in 1982 and the legislature restored permit and inspection immunity by statute, subject to a gross-negligence override the California and Nevada statutes do not carry.
E-182 The federal test for whether a discretionary act is immune Two parts: whether the act involved judgement or choice at all, and whether that judgement is of the kind the exception was designed to shield. Where policy allows discretion, the act is presumed grounded in policy Confirmed 28 U.S.C. 2680(a); Berkovitz v. United States, 486 U.S. 531 (1988); United States v. Gaubert, 499 U.S. 315 (1991); Commercial Carrier Corp. v. Indian River County, 371 So. 2d 1010 (Fla. 1979) 2026-08 2026-08-17 SITE /solution/ 2027-08 Opinions read in full 2026-08-17. Berkovitz: the exception 'will not apply when a federal statute, regulation, or policy specifically prescribes a course of action'. RELEVANT AND UNCOMFORTABLE: a compiled determination is precisely a specifically prescribed course of action. The more mechanical the rule, the less discretionary the act, and the less immunity attaches to getting it wrong.
E-183 Whether an individual rather than an organization may win the Ivory Prize CONTESTED. The Terms and Conditions say the contest is open to individuals over the age of 18 and organizations of all kinds; the FAQ on the prize page says organizations within the United States are eligible Contested Ivory Innovations, "Ivory Prize" FAQ and "2026 Ivory Prize Terms and Conditions" sec. 2.1 — https://www.ivoryinnovations.org/ivory-prize and https://www.ivoryinnovations.org/ivory-prize/terms 2026-08 2026-08-17 SITE about 2026-10 Both read 2026-08-17 and they disagree. Complicating it: the Terms page is titled for the 2026 cycle and may not have been refreshed, and sec. 3.4 refers to organizations selected as finalists. THIS IS DISPOSITIVE FOR A SOLO RESEARCHER AND MUST NOT BE ASSUMED EITHER WAY — email ivoryprize@ivoryinnovations.org before the form opens. This site previously stated the individual-eligible reading as fact.
E-184 Match rate joining California APR Table A to Table A2 260,811 of 330,855 projects, 78.8%. 133,088 of those carry a usable application-to-issuance duration in the cohort applied on or before 2022-12-31 Confirmed California Department of Housing and Community Development, Housing Element Annual Progress Report data by jurisdiction and year, Table A and Table A2, reporting years 2018-2025, retrieved 2026-08-17 from https://data.ca.gov/dataset/housing-element-annual-progress-report-apr-data-by-jurisdiction-and-year 2026 2026-08-17 SITE /permitting/ 2027-08 THE MATCH RATE IS NOT THE BINDING CONSTRAINT HERE, WHICH IS THE OPPOSITE OF WHAT WAS EXPECTED. Terner Center matched roughly 29,000 of about 60,000 projects doing this for ADUs, and a rate near 48% was the anticipated ceiling. Joining on jurisdiction plus tracking id, falling back to jurisdiction plus parcel number, reaches 78.8% because the two keys fail on different projects. Both tables are first collapsed to one record per project — earliest application date, earliest permit issue date — because Table A repeats a project in every year it is still pending and Table A2 repeats it in every year it books activity. Skipping that step inflates the count roughly threefold. The constraint turned out to be what the matched dates MEAN, which is the next row.
E-185 California jurisdictions whose own APR filing can produce a permit duration 190 of 529 jurisdictions. 339 cannot: 234 have fewer than 30 matched projects in the cohort, and 105 fail the identical-date screen Confirmed California Department of Housing and Community Development, Housing Element Annual Progress Report data by jurisdiction and year, Table A and Table A2, reporting years 2018-2025, retrieved 2026-08-17 from https://data.ca.gov/dataset/housing-element-annual-progress-report-apr-data-by-jurisdiction-and-year 2026 2026-08-17 SITE /permitting/ 2027-08 A jurisdiction that files the form and cannot produce a duration from it is a finding, not a gap in this measurement. Every city and county in California is compelled to file these dates and roughly two in three of them file something a duration cannot be computed from. The screen is stated so it can be argued with: at least 30 matched projects in the cohort, and at most 5% of them carrying an identical application and issuance date. Both thresholds are judgment and both are published as columns in the shipped extract, so a reader who prefers different ones can re-run the screen rather than take this on trust. The same screen runs on every series independently, which is why some jurisdictions publish a headline median and no entitlement figure.
E-186 California projects whose application and issuance dates are identical 13.7% of all matched projects. Concentrated rather than spread: Los Angeles is at 1.4% across 29,938 projects, Bakersfield at 81.5% Confirmed California Department of Housing and Community Development, Housing Element Annual Progress Report data by jurisdiction and year, Table A and Table A2, reporting years 2018-2025, retrieved 2026-08-17 from https://data.ca.gov/dataset/housing-element-annual-progress-report-apr-data-by-jurisdiction-and-year 2026 2026-08-17 SITE /permitting/ 2027-08 A nought-day elapsed time from application to building permit is a jurisdiction entering one date in two fields, not a permit issued the day it was applied for. It is recorded rather than filtered because it is the single most useful thing this exercise learned about the record: the data quality is a property of the JURISDICTION, not of the state programme, so a statewide median computed over all filings is contaminated in a way that a per-jurisdiction one is not. The 8,992 matched projects whose permit date precedes their application date are dropped and counted separately in the extract.
E-187 Median application-to-issuance in California, by product class, across jurisdictions that pass the screen Single-family detached 130 days across 82 jurisdictions; ADU 196 days across 138 jurisdictions; multifamily of 5 units or more 440 days across 10 jurisdictions; application to entitlement 102 days across 48 jurisdictions. Each is a median of jurisdiction medians Supported California Department of Housing and Community Development, Housing Element Annual Progress Report data by jurisdiction and year, Table A and Table A2, reporting years 2018-2025, retrieved 2026-08-17 from https://data.ca.gov/dataset/housing-element-annual-progress-report-apr-data-by-jurisdiction-and-year 2026 2026-08-17 SITE /permitting/ 2027-08 Graded Supported and not Confirmed, for two reasons that both push the figures DOWN. First, right-censoring: a project that applied and has not been permitted cannot appear in a join, so every figure here describes projects that finished, and the slow ones from recent years have not. The cohort is restricted to applications on or before 2022-12-31 to give three years of follow-up, which reduces the bias and does not remove it. Second, a median of medians weights a jurisdiction with 30 projects equally with one carrying 29,938. Both choices are stated rather than corrected, because correcting either would require assuming something about the projects that are missing. Only 10 jurisdictions clear the screen for multifamily, so the 440 is the weakest figure in this row and the range behind it runs 48 to 891 days.
E-188 Application to building-permit issuance in the largest California jurisdictions Los Angeles 189 days overall and 633 for multifamily of 5 units or more; Los Angeles County 417 overall and 891 for multifamily; San Diego 236; San Francisco 484. Medians, cohort applied on or before 2022-12-31 Supported California Department of Housing and Community Development, Housing Element Annual Progress Report data by jurisdiction and year, Table A and Table A2, reporting years 2018-2025, retrieved 2026-08-17 from https://data.ca.gov/dataset/housing-element-annual-progress-report-apr-data-by-jurisdiction-and-year 2026 2026-08-17 SITE /permitting/ 2027-08 Carried as its own row because the page names these four and a reader is entitled to see which jurisdiction each figure belongs to rather than a distribution. Los Angeles County at 891 days for multifamily is the largest well-populated figure in the set — two and a half years from application to a building permit, before anything is built. The San Francisco figure is independent corroboration rather than a new claim: E-155 carries 605 days to permit for CY2022 from the city's own accounting, and this measure, on a different population over a different window, lands at 484. Same order, different instrument, and neither is evidence for the other.
E-189 Washington jurisdictions required to report permit timeliness, and participation 51 jurisdictions required for 2025 — seven counties and 44 cities over 20,000 population. 50 submitted; Bellingham did not. Grant County was not required to report and filed anyway, which its own entry states Confirmed Washington State Department of Commerce, "Annual Permitting Performance Report - 2025", published under RCW 36.70B.080(2)(d), retrieved 2026-08-17 from https://deptofcommerce.box.com/s/cwnxwnxwdjyhzdrxu6jsxpuwb000x252 2026 2026-08-17 SITE /permitting/ 2027-08 The reporting universe is larger than a count of the seven Review and Evaluation counties suggests, because RCW 36.70B.080 reaches the cities over 20,000 within them as well. Appendix A carries 48 jurisdiction tables: the 50 submitters less the three that reported no permits at all. Grant County is in the appendix and outside the requirement, so it is in the shipped extract and excluded from any count of what the mandate produced.
E-190 Applicant hold time as a share of elapsed permit time, Washington Elapsed time exceeds active review time in 43 of the 48 jurisdiction tables. Seattle 696 calendar days elapsed against 231 active; Shoreline 804 against 349; Bellevue 390 against 85. Across tables the medians are 108 days elapsed and 66 days active Confirmed Washington State Department of Commerce, "Annual Permitting Performance Report - 2025", published under RCW 36.70B.080(2)(d), retrieved 2026-08-17 from https://deptofcommerce.box.com/s/cwnxwnxwdjyhzdrxu6jsxpuwb000x252 2026 2026-08-17 SITE /permitting/ 2027-08 THE MOST CONSEQUENTIAL ROW IN THIS SET, AND IT IS ABOUT A DEFINITION RATHER THAN A DELAY. Washington is one of the few authorities anywhere that publishes both the elapsed wait and the agency's own clock with pauses removed, and the two differ by a factor of three in the worst cases. Almost every permitting figure in public circulation is the second kind, because it is the number an authority measures itself against — and it is not the wait a builder experiences. This row is why the measurement spec includes applicant hold time in the primary measure: the quantity this project cares about is delivery time, not agency throughput. Reporting Seattle at 231 days would have described a real quantity and called it something it is not.
E-191 Labelling of the two duration columns in Washington's 2025 permitting report 45 of 48 jurisdiction tables print column 5 as "(with pauses)" and column 6 as "(without pauses)". Arlington and Battle Ground print the two reversed. Column 5 is the larger figure in all 288 data rows in the document Confirmed Washington State Department of Commerce, "Annual Permitting Performance Report - 2025", published under RCW 36.70B.080(2)(d), retrieved 2026-08-17 from https://deptofcommerce.box.com/s/cwnxwnxwdjyhzdrxu6jsxpuwb000x252 2026 2026-08-17 SITE /permitting/ 2027-08 Recorded because anyone repeating this work will hit it, and because reading the labels rather than the positions silently swaps elapsed time for active review time in two jurisdictions — the one distinction that matters most here. The ordering is established from the data and confirmed against the report's own deadline-compliance column: Arlington records nought decisions exceeding a 170-day deadline while its column 5 shows 286 days, which is only consistent if column 6 is the figure compliance is measured on. This is a defect in the presentation and not in the data, and it is not a criticism of the return: Washington publishes a split almost nobody else does.
E-192 What Washington's headline permitting table actually measures Table 5, the report's comparative ranking of all jurisdictions, is median review days COMPARED TO the statutory deadline — a signed difference, not a duration. Most of its bars are negative Confirmed Washington State Department of Commerce, "Annual Permitting Performance Report - 2025", published under RCW 36.70B.080(2)(d), retrieved 2026-08-17 from https://deptofcommerce.box.com/s/cwnxwnxwdjyhzdrxu6jsxpuwb000x252 2026 2026-08-17 SITE /permitting/ 2027-08 Filed as a standing example of the rule that statutory deadlines are recorded as such and never converted into durations. The most prominent table in the document, and the one a journalist would quote, cannot be read as a permit duration at all: a jurisdiction 50 days inside a 170-day deadline and one 50 days inside a 65-day deadline appear identical. Chile is the standing proof of the same rule from the other direction — the shortest statutory permit deadline found anywhere in the world beside one of the longest observed permit stages.
E-193 Whether the two states that compel permit-timing reports publish a permit duration Neither does. California publishes the dates and no duration — the figures in these rows had to be computed by joining two tables it publishes separately. Washington publishes a duration, for a permit class that excludes building permits since HB 1935 (2025), on a clock that starts at the notice of complete application rather than at submittal Supported California Department of Housing and Community Development, Housing Element Annual Progress Report data by jurisdiction and year, Table A and Table A2, reporting years 2018-2025, retrieved 2026-08-17 from https://data.ca.gov/dataset/housing-element-annual-progress-report-apr-data-by-jurisdiction-and-year. And: Washington State Department of Commerce, "Annual Permitting Performance Report - 2025", published under RCW 36.70B.080(2)(d), retrieved 2026-08-17 from https://deptofcommerce.box.com/s/cwnxwnxwdjyhzdrxu6jsxpuwb000x252 2026 2026-08-17 SITE /permitting/ 2027-08 THE ABSENCE THIS EXERCISE WAS BUILT TO TEST, AND IT SURVIVED. It is a stronger version of E-134 rather than a restatement: E-134 records that no construction-duration statistic exists, and this records that the two states which compel the reporting of permit timing still do not produce a permit duration a reader can look up. Graded Supported and not Confirmed because it is an absence across two state programmes examined in full rather than an exhaustive search of every publication either state issues — it says none was found in the mandated returns, not that none exists anywhere. What each state is missing is different, and that is the useful part: California has the dates and no statistic, Washington has a statistic measuring a narrower thing than the phrase "permit duration" implies. CHECKED SPECIFICALLY, because an absence claim of this shape is most often refuted by a dashboard nobody counts as a publication: HCD publishes an APR Data Dashboard over the same Table A and Table A2 submissions, and it reports counts and RHNA progress rather than any elapsed time. Berkeley's Possibility Lab, writing about this same dataset, records that better data is needed to know how long approval and permitting take, which is corroboration from a party that would have said otherwise if a duration existed.
E-194 Applicant hold time as a share of elapsed permit time, by permit type, Washington Median 37% of elapsed time, across 118 jurisdiction-by-permit-type cells. By class: construction permits 54%, multifamily housing permits 44%, preliminary subdivisions 22%, binding site plans 19%, final subdivisions 13%. Hold time exceeds half of elapsed time in 39 of the 118 cells Confirmed Washington State Department of Commerce, "Annual Permitting Performance Report - 2025", published under RCW 36.70B.080(2)(d), Appendix A, retrieved 2026-08-17 from https://deptofcommerce.box.com/s/cwnxwnxwdjyhzdrxu6jsxpuwb000x252 2026 2026-08-17 SITE /permitting/ 2027-08 THE GRADIENT IS THE FINDING, NOT THE MAGNITUDE. E-190 records that elapsed and active review time diverge in Washington; this records WHERE. The gap is widest on construction permits and multifamily housing — the two classes that involve the most applicant-side design iteration — and narrowest on final subdivisions, which are largely a paperwork step. That is a mechanism rather than a level, and it means the understatement is worst for exactly the permits that put housing on the ground. Computed over the two duration columns of Appendix A, read by position rather than by label for the reason at E-191. Clark County is excluded because its figures are a first review cycle by its own note and belong in no column with elapsed times. 28 of the 118 cells record no hold time at all, which is carried rather than smoothed: those are real zeroes in the return and some of them are jurisdictions whose systems cannot track a pause, which the report says in its own list of implementation challenges.
E-195 What Portland, Oregon publishes as its permit timeline, and where that timeline ends Median and average business days taken by City staff and by applicants, reported separately for each review type. The headline dashboard measures time to "approved to issue" status, which the city defines as the point at which review teams finish their work and await payment of final permit fees by the customer Confirmed City of Portland, Oregon, "Check Permit Timelines" and "Permitting Metrics", retrieved 2026-08-17 from https://www.portland.gov/ppd/how-get-building-permit/check-permit-timelines 2026 2026-08-17 SITE /permitting/ 2027-08 THE BEST-DESIGNED PERMIT METRIC FOUND ANYWHERE, AND ITS HEADLINE STILL STOPS BEFORE THE PERMIT EXISTS. Portland is the only authority found that splits staff time from applicant time as a published series, which is why the measurement spec names it. Recorded here for the endpoint rather than the split: "approved to issue" precedes fee payment and issuance, so the figure ends where the city's work ends and not where the applicant's wait ends. Reported in BUSINESS days against Washington's calendar days, so the two are never placed in one column and no conversion is attempted here. The underlying numbers sit in a Power BI embed that was not extracted; this row is about what Portland measures, not how long Portland takes, and no Portland duration is asserted anywhere on this site.
E-196 Whether any authority examined publishes the elapsed applicant wait as its headline permitting figure None of the four whose definitions could be verified. Washington publishes both clocks and is the only case where the gap can be measured; Portland splits both and leads with a figure ending before issuance; England reports 91% within 13 weeks or an agreed extended time against 19% within the statutory 13 weeks; California publishes the dates and no duration Supported Synthesis over E-157, E-189, E-190, E-193, E-194 and E-195. Sources named there. Portland: City of Portland, Oregon, "Check Permit Timelines" and "Permitting Metrics", retrieved 2026-08-17 from https://www.portland.gov/ppd/how-get-building-permit/check-permit-timelines 2026 2026-08-17 SITE /permitting/ 2027-08 Graded Supported, and the hedge is doing real work: four authorities is not a survey and this row says only what was checked. It does NOT say that most published permitting figures measure agency time — that is probably true, it is the reason this row exists, and nothing here establishes it. What the four cases do show is that the shortfall takes a different structural form each time: the clock stops during applicant hold, the endpoint sits before issuance, the deadline itself moves by agreement, or no duration is computed at all. A reader looking for a defect will look for one of those and there are at least four. DROPPED FROM THIS ROW: New Zealand, whose 20-working-day consent clock is suspended on a request for information and which would have been the cleanest statutory instance of the first form. mbie.govt.nz and building.govt.nz are both behind bot protection and legislation.govt.nz returns 403, so it could not be read from here — the same wall that dropped the Israeli and Spanish series. The 64.3% request-for-information share that a search engine attributes to MBIE is also not carried: a previous pass looked for it in the MBIE document and did not find it, and a search summary is not a source.
E-197 Every state must certify that its own law treats a chassis-free home identically, and the deadline is dated 42 U.S.C. 5403(i)(1)(A): "not later than 1 year after July 11, 2026, a State shall submit to the Secretary an initial certification that the laws and regulations of the State - (i) treat any manufactured home in parity with a manufactured home (as defined and regulated by the State); and (ii) subject a manufactured home without a permanent chassis to the same laws and regulations of the State as a manufactured home built on a permanent chassis, including with respect to financing, title, insurance, manufacture, sale, taxes, transportation, installation". Initial deadline 11 July 2027; (i)(1)(C) extends it to 2 years for a State "with a legislature that meets biennially". (i)(3) requires annual recertification. (i)(4): the Secretary "shall publish and maintain in the Federal Register and on the website of the Department of Housing and Urban Development a list of States that are up to date with the submission of initial and subsequent certifications" Confirmed 42 U.S.C. §5403(i) as added by Pub. L. 119-101, title III, §301(c), 11 July 2026, 140 Stat. 888; codified text and amendment note read at https://uscode.house.gov/view.xhtml?req=granuleid:USC-prelim-title42-section5403 2026-07 2026-08-17 HAP-08 certification-window 2027-07 Read from the codified text, subsection by subsection. E-090 listed "state certification deadlines" among the provisions it declined to claim because they were unverified; this row is that provision, opened. WHY IT MATTERS TO THIS ARGUMENT: the property classification the 1974 Act left alone (E-091) is now the subject of a dated federal obligation that names title and financing expressly. What the certification requires is parity, not reclassification, and the difference is the whole question - see E-199. (i)(4) also makes the compliance record a published series rather than a records request, which is the first instrumented before-and-after this project has been able to name.
E-198 A state that does not certify must prohibit the chassis-free home 42 U.S.C. 5403(i)(5)(B): "If a State does not submit a certification under paragraph (1)(A) or (3) by the date on which those certifications are required to be submitted - (i) with respect to a State in which the State administers the installation of manufactured homes, the State shall prohibit the manufacture, installation, or sale of a covered manufactured home within the State; and (ii) with respect to a State in which the Secretary administers the installation of manufactured homes, the State and the Secretary shall prohibit the manufacture, installation, or sale of a covered manufactured home within the State". (5)(A) defines a covered manufactured home as one "not considered a manufactured home under the laws and regulations of a State because the home is constructed without a permanent chassis", that is a manufactured home under 42 U.S.C. 5402, and is "constructed after July 11, 2026". (1)(D)(i): the Secretary "may not waive the prohibition described in paragraph (5)(B)" for a late certification "unless the Secretary approves the late certification" Confirmed 42 U.S.C. §5403(i) as added by Pub. L. 119-101, title III, §301(c), 11 July 2026, 140 Stat. 888; codified text and amendment note read at https://uscode.house.gov/view.xhtml?req=granuleid:USC-prelim-title42-section5403 2026-07 2026-08-17 HAP-08 certification-penalty 2027-07 The enforcement runs against the home, not against the state. Read plainly: a statute enacted to widen housing supply forecloses the new product in any state that misses its own deadline, and the only relief is the Secretary approving a late certification. This is the one provision on this site that could make the addressable market smaller than it was before the Act. How many states will miss it is E-200, and it is empty.
E-199 The parity a state must certify runs to that state’s own definition of a manufactured home 42 U.S.C. 5403(i)(1)(A)(i) requires a state to certify that its laws "treat any manufactured home in parity with a manufactured home (as defined and regulated by the State)". Where the certification is not submitted with a state plan, (i)(2) requires "an attestation by an official that the State has taken the steps necessary to ensure the veracity of the certification", "including, as necessary, by - (A) amending the definition of 'manufactured home' in the laws and regulations of the State; and (B) directing State agencies to amend the definition of 'manufactured home' in regulations" Confirmed 42 U.S.C. §5403(i) as added by Pub. L. 119-101, title III, §301(c), 11 July 2026, 140 Stat. 888; codified text and amendment note read at https://uscode.house.gov/view.xhtml?req=granuleid:USC-prelim-title42-section5403 2026-07 2026-08-17 HAP-08 parity-to-what 2027-07 Both quotations are the statute. THE INFERENCE, STATED SO IT CAN BE CHALLENGED: parity with a thing titled as personal property is parity into the chattel regime, not out of it, and the compliance route the statute itself names first is amending the state definition - which for most states means adding six words to a motor-vehicle titling statute. On that route the certification is satisfied and the classification E-091, E-092 and E-120 describe is not merely preserved but re-enacted, which is harder to unwind than silence. WHAT WOULD REFUTE IT: certifications that reclassify rather than extend. Nothing here says states will take the cheap route; the row records that the statute permits it and suggests it. The distribution of routes actually taken is E-200.
E-200 How many states certify, and whether they certify out of chattel or into it GAP [UNMEASURED] — no certification has been published, and no analysis of the form certifications take exists. Observable from the list 42 U.S.C. 5403(i)(4) requires the Secretary to publish, and from state session law, on a known schedule: initial certifications due 11 July 2027, or 11 July 2028 for a state whose legislature meets biennially Untested 2026-08-17 HAP-08 certification-penalty; HAP-08 parity-to-what; SITE deadline 2027-07 GAP, and the one this project can actually close. Unlike E-037 and E-039 it needs no records request and no fieldwork: the denominator is fifty, the deadline is in the statute, and the Secretary is required to publish the list. Two readings are wanted and they are different measurements - how many states certified, and how each one did it. The second is the load-bearing one (E-199).
E-201 The chassis-free construction standard carries no deadline 42 U.S.C. 5403(a)(7)(A): "The Secretary, in consultation with the consensus committee, shall issue revised standards for manufactured homes built without a permanent chassis using the process described in paragraph (4)." No date appears in the paragraph. Paragraph (4) is the consensus-committee process. The first Manufactured Housing Consensus Committee meeting after enactment, held 23 July 2026, carried one substantive agenda item: "(7) Review the Energy Conservation Standards for Manufactured Housing". Chassis-free standards do not appear on that agenda Confirmed 42 U.S.C. §5403(a)(7) as added by Pub. L. 119-101, title III, §301(b), 11 July 2026, 140 Stat. 888; MHCC meeting notice, 91 FR 42210, 8 July 2026, doc. 2026-13760, docket FR-6549-N-03 2026-07 2026-08-17 HAP-08 undated-standard; SITE deadline 2027-02 The asymmetry is the finding, and it is a comparison of two provisions of one Act: states are given twelve months (E-197), and the agency that must define the product is given no date at all. So fifty legislatures will write law about a home that cannot yet be built. Nothing here says HUD will be slow; the row records that nothing in the statute requires it to be fast, and that the committee whose process the statute mandates had not taken the subject up twelve days after enactment.
E-202 A chassis-free home will carry a different federal label from a chassis home 42 U.S.C. 5403(a)(7)(B) requires standards ensuring that manufactured homes without a permanent chassis have "(i) a distinct label, with revenue generated to be deposited into the Manufactured Housing Fees Trust Fund established under section 5419(e)(1) of this title, to be issued by the Secretary distinguishing manufactured home built without a permanent chassis from manufactured homes built on a permanent chassis; (ii) a data plate, as described in section 3280.5 of title 24, Code of Federal Regulations (or any successor regulation), distinguishing manufactured homes without a permanent chassis from manufactured homes built on a permanent chassis; and (iii) a notation on any invoice produced by the manufacturer of a manufactured home that is distinguishable from the invoice for a manufactured home constructed with a permanent chassis" Confirmed 42 U.S.C. §5403(a)(7)(B) as added by Pub. L. 119-101, title III, §301(b), 11 July 2026, 140 Stat. 888 2026-07 2026-08-17 HAP-08 undated-standard 2027-02 Read alongside E-204 and this is a paperwork detail that decides a market. Both Enterprise guides identify an eligible manufactured home by the HUD Certification Label or Data Plate and by a permanent chassis; the statute now requires the chassis-free home to carry a label and data plate that are expressly different. Whether the guides read the new label as an equivalent is E-206, and nobody in the public record has yet asked.
E-203 The only chassis rule HUD has proposed is narrower than the statute it now has to implement The proposed rule would amend the definition at 24 CFR 3280.2 to read "Nothing in this definition requires each transportable section to be built on a permanent chassis except for the transportable sections designed to be used as the lowest floor of a manufactured home" - upper floors only, lowest floor still on a chassis. HUD: "The proposed rule would reduce the production, transportation, and installation costs of a multistory manufactured home by anywhere from $4,800 to $6,700 per unit", and "HUD assumes that 20 to 25 multi-story manufactured homes could annually benefit from the proposed change." Comments closed 11 August 2026. As of 17 August 2026 no final rule, supplemental proposal or withdrawal has been published, and the 2026 Unified Agenda entry for RIN 2502-AJ80 carries one timetable row, "NPRM 07/00/2026", with no projected final action Confirmed "Revising the Definition of 'Manufactured Home' to Lower Housing Costs", 91 FR 35632, 12 June 2026, doc. 2026-11851, RIN 2502-AJ80, docket FR-6537-P-01; Unified Agenda 2026 edition entry for RIN 2502-AJ80 (announced at 91 FR 52792, 14 August 2026, doc. 2026-16603) 2026-06 2026-08-17 HAP-08 undated-standard; SITE deadline 2027-02 The proposal predates the Act and does not mention it - searched for "ROAD to Housing" and "ROAD Act" in the full text, absent on two passes. So the rule on the table is not an implementation of 5403(a)(7); it is a narrower change HUD argued it already had authority to make, priced on 20 to 25 homes a year. SCOPE LIMIT: the $4,800-$6,700 figure is for a multistory home under this proposal and is not a saving available to an ordinary single-storey home. The Act struck the chassis from the definition; the standard a factory could build to does not exist yet, and E-201 is why nothing dates its arrival.
E-204 Both Enterprise selling guides still require a permanent chassis Fannie Mae Selling Guide B5-2-01 (11/06/2024): "Any dwelling unit built on a permanent chassis that is attached to a permanent foundation system and evidenced by a HUD Data Plate or HUD Certification Label(s) (for each section of the home) is a manufactured home for purposes of Fannie Mae's guidelines", and "The manufactured home and the land on which it is situated must be legally classified as real property." Freddie Mac Single-Family Seller/Servicer Guide 5703.2 (eff. 02/09/2026) requires that the home "Be built on a permanent chassis in compliance with the HUD Codes in effect as of the date the Manufactured Home was constructed" and "Not have wheels, axles or towing hitches" Confirmed Fannie Mae Selling Guide B5-2-01, Manufactured Housing (11/06/2024), https://selling-guide.fanniemae.com/sel/b5-2-01/manufactured-housing; Freddie Mac Single-Family Seller/Servicer Guide Section 5703.2, General requirements related to Manufactured Homes, effective 02/09/2026, https://guide.freddiemac.com/app/guide/section/5703.2 2024-11; 2026-02 2026-08-17 HAP-08 undated-standard 2027-02 Read as written, a chassis-free HUD-code home is ineligible for purchase by either Enterprise, whatever HUD does to the construction standard - the guides are contracts with sellers, not restatements of federal law, and they were not amended by the Act. Neither text is exotic or buried: the chassis is in the first sentence of the Fannie definition and the first bullet of the Freddie list. This is the most tractable blocker on this page and the least discussed - a guide amendment, not a research question.
E-205 The agency that insures the loans has said in writing that it has not acted on the Act FHA INFO 2026-18, announcing Handbook 4000.1 Update 18 on 12 August 2026: "This update does not include provisions related to the 21st Century ROAD to Housing Act (P.L. 119-101) which are under evaluation." Confirmed FHA INFO 2026-18, 12 August 2026, U.S. Department of Housing and Urban Development, https://www.hud.gov/hud-partners/single-family-fha-info 2026-08 2026-08-17 SITE deadline 2027-02 One dated sentence in the agencies’ own voice does the work a survey of silence cannot. A parallel sweep of Fannie Mae selling policy communications, Freddie Mac bulletins, FHFA news releases, FHA mortgagee letters and Ginnie Mae APMs from 11 July to 17 August 2026 found nothing addressing chassis-free homes - but that is an absence claim resting on the completeness of the search, and it is stated here as the weaker half of the row rather than as a finding. The FHA sentence is the finding.
E-206 Whether a chassis-free home can be financed as real property at all GAP [UNMEASURED] — no Enterprise, FHFA, FHA or Ginnie Mae guidance states whether a manufactured home built without a permanent chassis, carrying the distinct label required by 42 U.S.C. 5403(a)(7)(B), is eligible for purchase or insurance, or whether it is treated as a CrossMod-class home for appraisal Untested 2026-08-17 SITE deadline 2027-02 GAP. This is the row on which the financing half of the argument now turns, and it is not a research question - it is a question four named institutions could each answer in a paragraph. Distinct from E-207, which records that the appraisal problem is already solved for one class of factory-built home. The chassis repeal converts into buyer financing through this row or not at all.
E-207 Site-built comparables are already required, not merely permitted, for one class of factory-built home Fannie Mae B4-1.4-01 (02/04/2026), for MH Advantage: "if fewer than three MH Advantage sales or their equivalent (CHOICEHome) are available, then the appraiser must supplement those comparable sales with the best and most appropriate sales available. Such sales must include a minimum of two site-built homes in recognition of the design standards for MH Advantage." Freddie Mac Guide 5703.12 (eff. 07/01/2026), for CHOICEHome: "the appraisal report should contain at least one comparable CHOICEHome sale. If none are available, the appraisal report must contain site-built homes as comparable sales." FHA Mortgagee Letter 2023-18: "If fewer than two comparable MH Advantage or CHOICEHome sales are available, the Appraiser must use the most appropriate site-built comparable sales available and must provide detailed justification to support the Appraiser's selection" Confirmed Fannie Mae Selling Guide B4-1.4-01, Factory-Built Housing: Manufactured Housing (02/04/2026); Freddie Mac Guide Section 5703.12, Mortgages secured by a CHOICEHome, effective 07/01/2026, https://guide.freddiemac.com/app/guide/section/5703.12; FHA Mortgagee Letter 2023-18, Update to the Sales Comparison Approach for Manufactured Housing, 2 November 2023 2023-11; 2026-02; 2026-07 2026-08-17 HAP-08 undated-standard 2027-06 THIS ROW CUTS AGAINST THIS PROJECT’S OWN FRAMING and is kept for that reason. The site has argued that comparables-based valuation is a lock nothing has opened and that type-based valuation would be a first - see the road link "A lender values the type, not the comp", classified still standing in audit/absence-claims-2026-08.md on the strength of E-009, E-095 and E-063. For CrossMod homes the comp lock is already open, in all three rulebooks, one of them since 2023, and Freddie makes site-built comps mandatory rather than permitted. What is still absent is valuation by TYPE - a certified design carrying its own value independent of location - which is not what these provisions do: they widen the comparable pool for a home meeting design standards. The narrower claim survives; the broad one does not, and the road label should be reread against this row.
E-208 A city may grant a by-right casita and exclude the cheapest house from it in the same ordinance Flagstaff City Code §10-40.60.030: "A mobile home, recreational vehicle, or other movable habitable space shall not be used as an ADU." The same section: "A manufactured or modular unit placed and secured on a permanent foundation in conformance with the Building Code may be used as an ADU in compliance with this section and in zoning districts that permit these units" Confirmed Flagstaff City Code §10-40.60.030, Accessory Dwelling Units, as amended by Ord. 2025-26, 2 December 2025 (Res. 2025-61), https://www.codepublishing.com/AZ/Flagstaff/html/Flagstaff10/Flagstaff1040060.html 2025-12 2026-08-17 HAP-08 both-halves-missing 2027-06 One ordinance, read in full, doing both things at once - and the second sentence is the operative one. A factory-built unit is admitted only where the district already permits such units, which is the authority A.R.S. 41-4006(D) expressly reserves (E-132). So the statewide accessory-dwelling right and the least expensive way to build a home in America do not necessarily intersect on any given parcel, which is what E-107 measures and Gate 0 was rewritten to test. ONE ORDINANCE IS NOT A DISTRIBUTION: this row establishes that the exclusion is lawful and drafted, not how much land it covers. Similar language was read on a Glendale permit handout and a Cochise County summary page; neither is codified text and neither is cited here.
E-209 Whether any Arizona statute bars zoning discrimination against manufactured housing GAP [NOT LOCATED] — no such statute was found. A.R.S. 9-462.01 (municipal zoning) and A.R.S. 11-811 (county zoning) were read in full and neither mentions manufactured, mobile, modular or factory-built homes. A.R.S. 41-4006(D) runs the other way, expressly preserving local zoning over "factory-built buildings, manufactured homes and mobile homes". The permissive treatment that exists in Arizona is local rather than state-mandated Untested 2026-08-17 HAP-08 both-halves-missing 2027-06 AN ABSENCE CLAIM, GRADED AS ONE. Three sections were read directly and keyword sweeps for the standard formulations returned other states - Washington, Utah, Montana, Vermont, Oregon - and Arizona city and county ordinances, not an Arizona statute. What was NOT done: an exhaustive section-by-section reading of A.R.S. Title 41 Chapter 16, because the chapter index could not be opened. So this is a well-supported negative, not a demonstrated one, and it stays Untested until the chapter is enumerated. Naming the gap is the point: a row that claimed the negative on this evidence would be the same defect as C-017.
E-210 The first readable accessory-dwelling permit counts under Arizona preemption, and they are flat Mesa: "As of August 2025, the City has issued 29 ADU permits, nearly matching the total of 30 issued during all of 2024." Chandler: "For 2024, the city received 10 permit requests and this year to date, 18". Scottsdale: "Scottsdale has only approved one application for an ADU so far in 2025, though a building permit has not been issued for that yet." City staff figures, obtained by a reporter Supported Cecilia Chan, "Casita law has had little impact on Mesa", East Valley Tribune, 19 August 2025, https://www.eastvalleytribune.com/casita-law-has-had-little-impact-on-mesa/article_063cfc3c-30f6-5ed0-bafc-1ea6a74b2189.html 2025-08 2026-08-17 HAP-08 both-halves-missing 2027-06 Graded Supported, not Confirmed: these are city staff figures given to a reporter by telephone, not a published dataset, and no city publishes the series. THE THREE NUMBERS ARE IN THREE DIFFERENT UNITS - permits issued, permit requests received, and one application approved with no permit - so they must not be summed, ranked or averaged, and the temptation to write a single total for the metro is the error this note exists to prevent. What they do support is direction in one city: Mesa's first eight months of 2025 ran level with all of 2024. Partially fills E-037, which said NOT PUBLISHED - see C-048. Tempe publishes an accessory-dwelling dashboard whose figures could not be read; that is the next place to look and it is not cited here.
E-211 Arizona already requires the plan library this project said was missing A.R.S. 9-461.20(A): "A municipality shall establish standard preapproved housing design plans or a preapproved housing design plan program." Class 4: "Beginning July 1, 2026, in areas zoned as single-family residential that allow for the development of accessory dwelling units, the municipality shall create at least three standard preapproved housing design plans for accessory dwelling units", sized at "(a) Two hundred square feet. (b) Six hundred square feet. (c) One thousand square feet." Subsection (B)(9): a municipality "shall administratively approve or deny an application for a class 1, class 2, class 3 or class 4 preapproved housing design plan submitted pursuant to the preapproved housing design plan program without discretionary review." No population threshold applies Confirmed A.R.S. §9-461.20, Municipal housing; standard preapproved housing design plans; indemnification; definition, added by Laws 2025, Ch. 259 (SB 1529) as §9-461.19 and renumbered; chaptered text at https://www.azleg.gov/legtext/57leg/1R/laws/0259.htm 2025-07 2026-08-17 HAP-08 both-halves-missing 2027-06 Opened because this site asserted Arizona had granted the right and built "no plan library" - see C-047. It is statutory, it is untied to population unlike A.R.S. 9-461.18, and the accessory-dwelling class was due to begin on 1 July 2026, six weeks before this row was written. TWO LIMITS, BOTH IN THE ENACTED TEXT AND NEITHER TO BE SMOOTHED OVER: subsection (A) is disjunctive - plans OR a program - while (A)(1)-(A)(4) say the municipality "shall create at least three"; and the "without discretionary review" duty in (B)(9) attaches to deciding third-party applications under the program, not to a permit for a house built from a plan. Whether any municipality has actually published the class 4 plans is unmeasured, and it is a cheaper question than E-037.
E-212 Federal home-improvement lending now names the accessory dwelling unit 42 U.S.C. note - Title I of the National Housing Act, 12 U.S.C. 1703(a), as amended: Pub. L. 119-101, §303(a)(1)(A) "inserted 'construction of additional or accessory dwelling units, as defined by the Secretary,' after 'energy conserving improvements,' in fourth undesignated par." Confirmed 12 U.S.C. §1703(a) as amended by Pub. L. 119-101, title III, §303(a)(1)(A), 11 July 2026; 2026 amendment note read at https://uscode.house.gov/view.xhtml?req=granuleid:USC-prelim-title12-section1703 2026-07 2026-08-17 HAP-08 certification-window 2027-06 The second provision E-090 declined to claim, opened. It matters to the Arizona leg specifically: the gap in accessory-dwelling delivery is construction finance on a lot the owner already occupies, and a Title I property-improvement loan is the one federal instrument now written to cover it. SCOPE LIMIT, and it is large: this row is the amendment note, not a measurement of lending. Whether any Title I accessory-dwelling loan is originated depends on lender participation in a programme with thin volumes, and no volume is claimed here.
E-213 Median new single-family home price fell below median existing home price $403,200 new vs $404,600 existing, Q1 2026 — new is $1,400 lower Supported NAHB Eye on Housing, Onnah Dereski, "New Home vs. Existing Home Prices in Q1 2026", https://eyeonhousing.org/2026/05/new-home-vs-existing-home-prices-in-q1-2026/ 2026-05-29 2026-08-17 HAP-08 new-below-existing 2027-05 NAHB analysis of US Census Bureau new-home and NAR existing-home medians, not seasonally adjusted. Fourth consecutive quarter in which existing exceeded new, and six of the past eight quarters since the relationship reversed in Q2 2024 — so this is a sustained inversion rather than a single print. Regional split is not uniform: the Northeast still carries a $309,200 new-home premium while the West and South run the other way by $55,500 and $700. Graded Supported rather than Confirmed because it rests on one trade-association analysis of two agency series rather than on the agency series read directly; the underlying Census and NAR releases have not been pulled independently by this project.
E-214 The new-home price premium has gone negative for the first time in a five-decade series GAP reported as first negative reading since at least 1974 Untested Fortune, Catherina Gioino, reporting John Burns Research & Consulting, https://fortune.com/2026/07/15/cheaper-buy-new-home-old-resale-generation-boomers-sellers/ 2026-07-15 2026-08-17 HAP-08 new-below-existing 2027-02 UNTESTED, and the reason is the point: the fifty-year claim rests on a John Burns Research & Consulting series that is proprietary, reported at second hand by Fortune, and has not been examined by this project. The NAHB analysis in E-213 supports the inversion itself but makes no fifty-year claim at all — it dates the reversal to Q2 2024 and goes no further back. Anything on this site that says "first time in five decades" is repeating a number nobody here has checked. What would settle it: the Census new-home median and NAR existing-home median read directly back to 1974 and differenced.
E-215 Offered explanations for the new-build discount: incentives, smaller homes, locked-in resale inventory GAP incentives ~7-8% of new sale price; median new home ~2,400 sq ft vs ~2,700 mid-2010s; outstanding rates ~4.3% vs market ~6.5% Untested Fortune, Catherina Gioino, citing John Burns Research & Consulting, Realtor.com and Redfin, https://fortune.com/2026/07/15/cheaper-buy-new-home-old-resale-generation-boomers-sellers/ 2026-07-15 2026-08-17 HAP-08 new-below-existing 2027-02 Three mechanisms offered by trade reporting, none of them measured here, and they are not mutually exclusive. Builder rate buydowns and closing-cost credits are estimated at roughly 7-8% of new-home sale prices; the median new home sold has contracted to about 2,400 square feet from roughly 2,700 in the mid-2010s, so part of the discount is a smaller house rather than a cheaper one; and resale supply is held back by a spread between average outstanding mortgage rates near 4.3% and market rates near 6.5%. The middle mechanism matters most to this argument and cuts against the cheerful reading: a price that falls because the product shrank is not a productivity gain. Untested — no attempt has been made here to decompose the gap into incentive, size and mix.
E-216 Commercial real estate grades buildings A/B/C; the system is explicitly subjective and not a formal certification BOMA: "a subjective quality rating of buildings which indicates the competitive ability of each building to attract similar types of tenants" Supported BOMA International building class definitions, as documented in industry guidance, https://www.sanalifeenergy.com/blog/an-overview-of-bomas-commercial-building-classes 2026-08 2026-08-18 HAP-08 class-grades 2027-08 The point of this row is the caveat, not the classes. BOMA describes the A/B/C system as a guideline and a "standardization of discussion" rather than a standard, calls it somewhat subjective, and it is market-relative — a Class A building in Manhattan is not a Class A building in Winnipeg. So the commercial system is a shared vocabulary anchored to income and competitive position, not a measurement. Graded Supported rather than Confirmed because BOMA is quoted here through industry guidance rather than from BOMA’s own publication, which has not been pulled directly.
E-217 A residential grading vocabulary would make explicit the standards discretionary review applies implicitly GAP no letter-grade convention in residential; descriptive traits used instead Untested No source. This is a proposal recorded as a gap, not a finding. 2026-08 2026-08-18 HAP-08 class-grades 2027-02 UNTESTED and deliberately so. The argument: discretionary prose lets a reviewer apply a standard without stating it, so a project is refused for not fitting a community whose criteria were never written down. A published A/B/C grade with explicit features would force the criteria into the open, where they can be met, argued with, or compiled. Two things are unestablished. First, whether discretion is used this way as a matter of record rather than of impression — E-039, the objective-versus-discretionary split of American zoning, has never been measured anywhere, which is exactly the row this proposal would need. Second, whether the commercial analogy transfers: BOMA’s classes are not a formal certification (E-216) and are anchored to income and tenant competition, and residential has no equivalent anchor, so importing the letters without the anchor may import nothing. Recorded as a proposal so it cannot be read as a finding.
E-218 Monthly median sales price, new versus existing single-family homes, June 2025 to June 2026 new $409,200 (Jun 2025) to $398,300 (Jun 2026), shown as $398.3k; existing $438,600 to $446,400, shown as $446.4k. June 2026 gap: existing $48,100 above new Supported NAHB Table 15, "New and Existing Single Family Home Prices, U.S.", last updated 2026-07-24, compiled from US Census Bureau Construction Reports C-25 (new) and National Association of Realtors Home Sales (existing), https://www.nahb.org/-/media/NAHB/news-and-economics/docs/housing-economics/sales/median-prices.pdf 2026-07-24 2026-08-18 HAP-08 new-below-existing 2027-01 Read directly from the table rather than off anyone's chart. Thirteen monthly points, both series, not seasonally adjusted, same basis throughout. The annual rows in the same table show the premium closing first: new exceeded existing by $26.4k in 2021, $41.7k in 2022, $34.5k in 2023 and $7.8k in 2024. The monthly series is where it inverts and it does not invert cleanly — new was still above existing in December 2025 (429.1 against 409.5) and January 2026 (416.2 against 398.2). Anyone reading the figure as a smooth crossover is reading it wrong, and the chart plots every point rather than the endpoints for that reason. Graded Supported, not Confirmed: NAHB compiles two agency series and this project has not pulled the Census and NAR releases independently.
E-219 Annual median sales price, new versus existing single-family homes, across four years to 2024 new $383,500 / $434,500 / $428,600 / $420,300 (2021-2024); existing $357,100 / $392,800 / $394,100 / $412,500 (2021-2024) and $419,300 (2025). New-home premium: +$26,400, +$41,700, +$34,500, +$7,800 Supported NAHB Table 15, "New and Existing Single Family Home Prices, U.S.", last updated 2026-07-24, compiled from US Census Bureau Construction Reports C-25 (new) and National Association of Realtors Home Sales (existing), https://www.nahb.org/-/media/NAHB/news-and-economics/docs/housing-economics/sales/median-prices.pdf 2026-07-24 2026-08-18 HAP-08 new-below-existing 2027-01 The annual half of E-218, read from the same table. It is the span that shows the premium collapsing rather than the month-to-month noise: new stood $26,400 above existing in 2021, peaked at $41,700 in 2022, and had fallen to $7,800 by 2024 before inverting. THE 2025 NEW-HOME ANNUAL MEDIAN IS NOT PUBLISHED IN THIS TABLE — that cell is blank in the source, and the chart shows the break rather than bridging it. A 2025 figure could be averaged from the monthly rows, but only seven months of 2025 appear in the table, so any annual number computed from them would be invented. The existing-home 2025 annual is published ($419,300) and is plotted. Supported rather than Confirmed for the same reason as E-218: NAHB compiles two agency series and neither the Census nor the NAR release has been pulled independently here.
E-220 NAHB's published schedule of what a regulatory dollar becomes by closing, by the stage at which it is incurred 0.00% imposed directly on the buyer at closing; 14.94% incurred by the builder during construction; 16.35% at start of construction; 16.82% when applying for the building permit; 27.63% during site development; 30.17% when applying for development approval Supported NAHB, "Government Regulation in the Price of a New Home: 2021", Appendix I "Assumptions Used in the Calculations", section "Impact of Costs on House Price", table "Additional Charges Passed on to the Home Buyer Depending on When a Cost is Incurred", https://www.nahb.org/-/media/NAHB/news-and-economics/docs/housing-economics-plus/special-studies/2021/special-study-government-regulation-in-the-price-of-a-new-home-may-2021.pdf 2021-05-05 2026-08-18 HAP-08 duration-not-cost; thirty-two-to-seven 2027-08 Read from the source PDF directly, not from any secondary rendering of it. THE DOCUMENT IS EXPLICIT THAT THESE ARE DERIVED: the sentence introducing the table reads "The above assumptions imply the following mark-up percentages that vary depending on when a particular cost is imposed on a developer, builder or home buyer". They are arithmetic on financing assumptions stated on the two preceding pages - 6.96% interest and 75% loan-to-cost on acquisition and development loans with a 90 basis point initial fee; 6.65% interest and 86% loan-to-cost on construction loans with a 76 basis point fee; a 9.6% gross profit rate for builders and developers; a 2.9% broker fee; and the time lags at E-221 - not observations of markups anyone was charged. The prime rate underneath the loan rates is itself a projection: 5.5%, stated as "300 basis points above the median appropriate longer term policy path for the federal funds rate projected by Federal Reserve Board members". Graded Supported, and the grade is doing specific work. What is Confirmed is that NAHB publishes this schedule with its assumptions fully disclosed - which is more than this project's own E-070 and E-071 arithmetic offers. What is NOT established is that a cost incurred at development approval is in fact marked up 30.17%. It is a third party's model with visible inputs and must never be presented as a measurement. Same instrument family as E-001, which is graded Contested; that row's provenance caveat applies here too - NAHB advocates deregulation, and the two longest durations feeding this table come from its own member survey.
E-221 A published end-to-end decomposition from zoning application to closing, and what it is assembled from 16.6 months zoning application to start of site work + 10.0 start of site work to sale of lot to builder + 1.0 sale of lot to start of construction + 5.9 start of construction to completion + 0.8 completion to closing = 34.3 months Supported NAHB, "Government Regulation in the Price of a New Home: 2021", Appendix I, section "Average Time Lags", https://www.nahb.org/-/media/NAHB/news-and-economics/docs/housing-economics-plus/special-studies/2021/special-study-government-regulation-in-the-price-of-a-new-home-may-2021.pdf 2021-05-05 2026-08-18 HAP-08 delivery-clock 2027-08 The five components are attributed on the page to four different instruments covering different populations, and the source labels each one. The 16.6 and the 10.0 are NAHB's own 2021 Land Developer Survey of Regulatory Costs - developer self-report. The 1.0 and the 5.9 are Census Survey of Construction averages for single-family homes built for sale in 2020, and the 1.0 additionally rests on a stated assumption, that "the builder seeks authorization for construction at the same time as purchasing the lot". The 0.8 is an NAHB tabulation of SOC microdata combined with a Zillow resale figure. So the 34.3 is an ASSEMBLY, not a measurement of any cohort of houses - the same defect E-002 records for the 8.8, one level up. THIS DOES NOT CLOSE E-038. E-038 says no dataset spans site identification to closing; that remains true, and this is a second independent assembly rather than the measurement E-038 is waiting for. What it does is corroborate the order of magnitude of the site's ~32-month figure from a source with no connection to how this project built it, and it is carried for that reason and no other.
E-222 Builder gross margin and sales incentives, quarter by quarter, read from the filings Gross margin on home sales by fiscal quarter: FY2022 26.9%, 29.5%, 29.2%; FY2023 21.2%, 22.5%, 24.4%; FY2024 21.8%, 22.6%, 22.5%; FY2025 18.7%, 17.8%, 17.5%; FY2026 15.2%, 15.6% (Q1 and Q2 only). Sales incentives as a share of revenue: 14.3% in the quarter ended 31 Aug 2025 against 10.2% a year earlier, and 9.4% in the quarter ended 31 May 2024. Latest reading 15.6% Confirmed Lennar Corporation Forms 10-Q, quarterly periods ended 2023-02-28 through 2026-05-31, Management Discussion ("Gross margins on home sales were ... or N%") and the "Sales Incentives (1)" tables. Series assembled from fourteen quarterly readings across eleven filings, each filing stating the quarter and its prior-year comparative. Most recent: https://www.sec.gov/Archives/edgar/data/920760/000162828026046019/len-20260531.htm 2026-06-29 2026-08-18 HAP-08 builder-margin; new-below-existing 2027-06 Every reading is quoted from a filing, not from a chart about the filings. The FY2025 Q3 sentence reads: "Gross margins on home sales were $1.4 billion, or 17.5%, in the third quarter of 2025, compared to $2.0 billion, or 22.5%, in the third quarter of 2024." The company attributes that fall to "a lower revenue per square foot and higher land costs year over year, which were partially offset by a decrease in construction costs". THE TREND IS THE FINDING, not any single quarter. Fourteen consecutive readings fall from 29.5% to 15.6% - roughly halved across four fiscal years, declining in every year of the series. A secondary report prompted this row by quoting 17.5% alone as a "16 year low"; that framing is NOT carried, because this project has not pulled the pre-2022 quarters that would test it, and because the figure is already stale - the latest filing reads 15.6%. Quote the series or the latest number, never the 17.5% on its own. Fourth quarters are absent by construction: Lennar reports them in the 10-K, not a 10-Q, so this is a Q1-Q3 series and is not a full-year average. DENOMINATOR WARNING on the incentive figures: this is incentives as a percentage of REVENUE, where E-215 carries roughly 7-8% of new-home SALE PRICE from trade reporting - not interchangeable, and both can be right. Use the total: 14.3% nationally, where the South Central division reads 20.5%. The secondary report quotes a division figure (Texas 15.9%, 2024) without the 9.4% total beside it, which reads as though the outlier were the norm. The division called Texas in 2024 is reported as South Central by 2025, so divisions are not a clean like-for-like series either. One builder. The second largest in the country, disclosing under securities law, but one - this is not an industry measurement, and nothing here decomposes the new-versus-existing gap that E-215 is Untested about.
E-223 What a builder pays to buy a mortgage rate down, against what the same monthly payment costs in price About $16,000 for a 1 percentage point reduction on a $400,000 loan, against $60,000 to $65,000 of price adjustment to reach the same monthly payment - roughly a quarter of the cost Supported Taylor Morrison Home Corp Q4 2024 earnings call, 12 February 2025, Sheryl Palmer (Chairman, President and CEO), transcript published by Insider Monkey, https://www.insidermonkey.com/blog/taylor-morrison-home-corporation-nysetmhc-q4-2024-earnings-call-transcript-1451570/ 2025-02-12 2026-08-18 HAP-08 new-below-existing 2027-02 Quoted as spoken: "instead of having to just discount the house or do a very expensive forward commitment where we're assisting a consumer, guaranteeing that they're going to have a 1% reduction to the market rate. Let's say on an average - let me take a $500,000 house with a 20% down payment, a 400,000 loan, to do that it's costing me, let's say approximately $16,000. To get that consumer to the same monthly payment, it would cost me $60,000 to $65,000 price adjustment." Graded Supported and not Confirmed for three separate reasons, all of which stay on the row: the text is a THIRD-PARTY transcript rather than the company's own or an SEC exhibit; it is an executive characterising her own programme, not an audited disclosure; and the figures are an illustrative example on a hypothetical she chooses ($500,000 house, 20% down). Restore toward Confirmed only against the company-published transcript or the 8-K exhibit. Why it is carried: it supplies the MECHANISM behind the incentive share at E-222. The builder buys the rate rather than cutting the price because the buydown costs roughly a quarter as much for the same effect on the buyer's monthly payment. That is a reason to expect the new-home discount to be bought rather than earned, from the party doing the buying.
E-224 The long-run gross profit rate for American builders and developers, as assumed by the trade association's own cost model 9.6 percent, described as the average rate across NAHB Construction Cost surveys conducted between 2002 and 2019 Supported NAHB, "Government Regulation in the Price of a New Home: 2021", Appendix I "Assumptions Used in the Calculations", section "Construction of the Single-family Structure", https://www.nahb.org/-/media/NAHB/news-and-economics/docs/housing-economics-plus/special-studies/2021/special-study-government-regulation-in-the-price-of-a-new-home-may-2021.pdf 2021-05-05 2026-08-18 HAP-08 builder-margin 2027-08 Quoted verbatim: "A 9.6 percent gross profit rate for builders and developers, based on the average rate on NAHB Construction Cost surveys conducted between 2002 and 2019. Without a competitive return above costs, builders will typically be unable to get construction loans underwritten to build the homes." Read from the same appendix as E-220 and E-221, and it carries the same provenance caveat: a trade association's own member surveys, and the figure is an input to a model the association uses when arguing for deregulation. A low builder margin is a number that helps that argument, which is exactly why it is graded Supported and not Confirmed. It is nonetheless a seventeen-year average from the association best placed to collect it, and it is carried for one narrow purpose: to establish an order of magnitude for what the builder line actually takes, against a price-to-income gap of the size at E-072. It does not measure any individual builder, does not cover land developers separately from builders despite naming both, and says nothing about return on capital, which is the number an investor would actually want.
E-225 Whether a longer-duration product type systematically fails to convert starts into completions NO. 1968-2008 means: single-family 1.01, buildings of 5 or more units 1.05. 2015-2025 means: 1.04 and 1.08. The longer-duration type is not systematically lower and is slightly higher. What differs is VARIANCE: the 5-plus series runs 0.38 (2009) to 1.48 (2022) while single-family stays near 1.0, range 0.75 (2008) to 1.16 (2021) Confirmed Computed from US Census Bureau new residential construction series (Survey of Construction), annual means of monthly seasonally adjusted annual rates 1968-2025, retrieved as CSV from FRED: https://fred.stlouisfed.org/series/HOUST1F , https://fred.stlouisfed.org/series/COMPU1USA , https://fred.stlouisfed.org/series/HOUST5F , https://fred.stlouisfed.org/series/COMPU5MUSA . Census source tables: https://www.census.gov/construction/nrc/index.html 2026-08 2026-08-18 - 2027-08 A REFUTATION, recorded because a refuted line of argument is a finding and this one was about to reach a beat. Tested because a secondary report claimed that "when project duration doubles, the completion ratio cuts in nearly half", citing single-family 1.03 against multifamily 0.56 in 2024 alongside the durations at E-002 and E-003. The 2024 pairing is real - this computation returns 1.00 and 0.57 for that year - but it is one year, and over the full series the claimed relationship does not exist. Multifamily's long-run ratio is slightly ABOVE single-family's, not half of it. The 2024 reading is a pipeline artefact: the 5-plus ratio ran 1.28 in 2021 and 1.48 in 2022, and those starts were still completing in 2024, which raises the denominator without anything failing to be delivered. A ratio of starts to completions within one year compares two different cohorts of buildings and cannot carry a duration claim on its own. WHAT SURVIVES is a claim about variance rather than level, and it is not the claim that was made: the longer-duration type delivers far less predictably, which is a real cost to anyone financing it, but the series does not show it delivering systematically less. Nothing on the site may say that it does. The computation is reproducible from the named series: annual mean of the twelve monthly values, then the ratio or the peak-to-trough comparison as stated in the value field.
E-226 Recovery from the 2008 shock, American housing against American motor vehicles Against 2007: housing completions fell 61.4% to a 2011 trough and regained the 2007 level in 2024, 13 years after the trough; motor vehicle assemblies fell 46.6% to a 2009 trough and regained the 2007 level in 2013, 4 years after. Against each series own all-time peak: housing -72.2% from 1973 and still at 71% of that peak in 2025; vehicles -56.0% from 1999 and at 79% in 2025. Neither has regained its peak Confirmed Computed from US Census Bureau new residential construction (total housing completions) and Federal Reserve Board G.17 motor vehicle assemblies, annual means of monthly seasonally adjusted annual rates, retrieved as CSV from FRED: https://fred.stlouisfed.org/series/COMPUTSA and https://fred.stlouisfed.org/series/MVATOTASSS . Underlying releases: https://www.census.gov/construction/nrc/index.html and https://www.federalreserve.gov/releases/g17/current/ 2026-08 2026-08-18 HAP-08 productivity-plateau 2027-08 The arithmetic is Confirmed. NO CAUSAL CLAIM IS MADE OR SUPPORTED, and the reason is on the row: the American motor vehicle industry received a federal rescue in 2009 that housing did not, so the faster recovery has at least one large explanation with nothing to do with standardised technical protocols or type certification. Any beat quoting the asymmetry must state that confound in the same breath. BASELINE CHOICE CHANGES THE ANSWER AND MUST BE STATED. The secondary report that prompted this row mixes the two: it takes its magnitudes from all-time peaks (-70.4% and -58.3%, close to the -72.2% and -56.0% computed here) but its "auto returned to a higher rate within 6 years" only holds against 2007, and its claim that housing "still remains underneath where it was before, over 17 years later" is FALSE against 2007 on completions - housing regained its 2007 level in 2024. Against the 1973 peak it is true. What is robust across both framings is the recovery-speed asymmetry: four years against thirteen, measured from each series' own trough. The computation is reproducible from the named series: annual mean of the twelve monthly values, then the ratio or the peak-to-trough comparison as stated in the value field.
E-227 Selling, general and administrative expense as a share of home sales revenue, and its direction 8.2% in the quarter ended 31 Aug 2025, up from 6.7% a year earlier; 7.5% in the quarter ended 31 May 2024, up from 6.7% a year earlier. Set against gross margin at E-222, this leaves roughly 7% on the operating line in the most recent quarter measured Confirmed Lennar Corporation, Form 10-Q for the quarterly period ended August 31, 2025, Management Discussion, https://www.sec.gov/Archives/edgar/data/920760/000162828025044086/len-20250831.htm ; and Form 10-Q for the quarterly period ended May 31, 2024, https://www.sec.gov/Archives/edgar/data/920760/000162828024030508/len-20240531.htm 2025-10-03 2026-08-18 HAP-08 regulatory-share-of-price; builder-margin 2027-06 Quoted from the 2025 filing: "As a percentage of revenues from home sales, selling, general and administrative expenses increased to 8.2% in the third quarter of 2025, from 6.7% in the third quarter of 2024, primarily due to less leverage as a result of lower revenues and an increase in marketing and selling expenses." The 2024 filing carries the same construction for 7.5% against 6.7%. WHY THIS ROW EXISTS, stated plainly because the record is the point: on 2026-08-18 this session asserted to a peer session that SG&A "ran about 6.7% of revenue in the quarters I pulled". It had not been pulled. The figure came from a WebSearch result summary, in a session that had already refused a summariser's claim about a NAHB table and gone to the PDF instead. It was also stale in the exact shape this project had just finished criticising elsewhere: 6.7% is the 2023-24 reading, and SG&A has been rising. It reached no page, because the peer declined to publish a figure a reader could not check. See L-020. WHAT IT IS FOR: gross margin at E-222 is struck before SG&A, so the margin comparison on regulatory-share-of-price is conservative. 15.6% gross less 8.2% leaves roughly 7% operating, against regulation at 26.4% of price on E-001. The comparison does not depend on this row - "struck before SG&A" makes it conservative for any SG&A above zero - but this row lets a note say how much lower rather than only that it is lower. One builder, and an operating margin is still not a return on capital.
E-228 Whether anyone has measured the counter-cyclical gap in housing production - building least when land is cheapest GAP [UNMEASURED] - no study located that quantifies, for US housing, the return forgone by producing at the trough of land and construction costs rather than at the peak, or the share of builders financially able to do it Untested 2026-08-18 2027-08 A GAP, recorded rather than argued, because the argument around it is currently better than the evidence for it. THE OBSERVATION: US housing production reached its post-1968 low in the same years land was least expensive, and had recovered only after land costs rose substantially. E-226 carries the adjacent measured fact - completions took thirteen years to regain their 2007 level while motor vehicle assemblies took four. WHAT IS NOT TRUE, and was the framing that prompted this row: that this pattern is unique to housing. Pro-cyclical capital commitment is the norm in capital-intensive industry - shipbuilding, oil and gas, semiconductors, mining all commit capacity at cost peaks and withhold it at troughs. Nothing on this site may call the pattern unique. WHAT MAY BE DISTINCTIVE, and is the open question: in those industries a trough can reflect genuine demand destruction, so waiting is rational. Housing demand is demographic and the shortfall is a stock that accumulates (E-053), so the demand present at the trough is largely still present later. If that is right, the counter-cyclical trade has a higher expected value in housing than in a commodity cycle, and the binding constraint on taking it is the carry - entitlement duration plus closed acquisition-and-development lending at exactly the trough - rather than uncertainty about demand. WHAT WOULD SETTLE IT: land basis and realised return for builders who acquired through 2009-2012 against those who acquired 2015-2019, and the share of builders whose capital structure permitted acquisition without bank underwriting. Neither has been assembled here. THEORY NOT CITED: the real-options literature on irreversible investment (Titman 1985 on urban land; Dixit and Pindyck 1994) is the obvious frame and is deliberately absent from the source field - neither was obtainable in full text on 2026-08-18, and this project does not cite work it has not read.
E-229 The missing instrument is a preempting siting standard GAP [UNTESTED] — the load-bearing sentence of the precedent argument, opened as a row so it cannot ship ungraded. The inference: for a siting determination to travel the way a construction determination travels, a standard with preemptive effect must stand behind it. What is verified around the sentence: every portable determination in the register rides on a preempting standard — HUD-code design approval on 42 U.S.C. §5403(d) (E-088, E-091), a vehicle maker's certificate on 49 U.S.C. §30103(b)(1) (E-111) — and Florida's private plan review, with no preempting standard behind it, does not travel (E-089). The sentence itself is an inference from that pattern, not a finding. Untested HAP-08 what-must-give; HAP-08 certify-the-checker; HAP-08 three-things 2027-07 Opened 2026-08-24. The restructure map required this row before the three-things promotion: the statutory clauses around the sentence are Confirmed, the sentence is an inference, and the two must not share a grade. THE INFERENCE, STATED SO IT CAN BE CHALLENGED: the record holds no case of a determination crossing a boundary where the receiving jurisdiction kept its own rules; absence of a case is not evidence of impossibility, and whether compilation substitutes for preemption is exactly the open question. WHAT WOULD SETTLE IT: against — one documented siting or placement determination accepted across jurisdictions with no preempting standard behind it, which would refute necessity. For — a state adopting a preempting siting standard for certified types, with placement volume measured before and after, which would bear on sufficiency. Alongside — Gate 0's enumeration of which siting provisions are parameters and which are judgements (E-131), which prices what such a standard would have to preempt. The nearest dated observation point is the certification window: 42 U.S.C. §5403(i) puts every state on a one-to-two-year clock, most ending 11 July 2027 (E-197), and (i)(4) makes the compliance record a published series. HAP-13 is this project's drafted contribution to that window.
E-230 Whether committed demand aggregated ahead of production — a standing order book — is a third binding constraint on housing production GAP [UNMEASURED] — no measurement located of the share of US housing production built against committed orders rather than against forecast, nor of the utilization floor a plant needs to survive a demand trough Untested No source. This is a proposal recorded as a gap, not a finding. 2026-08 2026-08-27 2027-02 A PROPOSAL RECORDED AS A GAP, on the E-217 pattern, so that it cannot be read as a finding and cannot be quietly forgotten. It is filed as a row rather than left in a planning document precisely because a to-do rots and a graded row is counted. THE ARGUMENT: this project names three locks — politics, information, capital — and states the capital lock as financing cost and duration (E-070, E-071) rather than as demand commitment. The claim under test is that an industrialized producer dies of utilization rather than of unit cost: it must build to forecast, a demand trough idles the line, and the fixed cost that made the factory attractive is what then kills it. WHAT THE RECORD ALREADY HOLDS, pointing this way without establishing it: Operation Breakthrough bought its order book with an appropriation and the appropriation was withdrawn; Katerra built to forecast. E-228 records that nobody has measured the counter-cyclical gap, and carries the warning that pro-cyclical capital commitment is the norm across capital-intensive industry — nothing on this site may call the pattern unique to housing, and this row does not. WHAT IS UNESTABLISHED, and is why this is a row and not yet a beat: no measurement of the share of US housing built against committed orders; no measurement of the utilization floor; and no case of a standing book assembled at scale outside a subsidy. ORIGIN, recorded because provenance is the point here: raised by an external model asked to rebuild American housing production from a clean sheet, read against this register on 2026-08-27 and mostly declined — DECISIONS D-01 to D-04. Of everything in that reading, this was the only gap the register did not already hold. Its accompanying figures were refused (D-02). WHAT WOULD SETTLE IT: a producer’s committed-order-to-start ratio across a full cycle, or a documented committed-book programme with utilization tracked through a trough.
E-231 A published construction-to-delivery duration for a high-end HUD-code home, and why it cannot be differenced against this site’s own build figure GAP "The typical construction-to-delivery timeline for a high-end HUD-code home ranges from three to four months, compared with five to six months (or longer) for site-built equivalents." And: "From order to move-in, a high-end manufactured home can be completed faster than site-built. It’s typically about 60 days faster, depending on the backlog." Untested Harold D. Hunt, "Forget Everything You Think You Know About Manufactured Homes", Texas Real Estate Research Center, Texas A&M University, https://trerc.tamu.edu/article/forget-everything-you-think-you-know-about-manufactured-homes/ 2026-01-21 2026-08-27 2027-02 Both figures quoted verbatim, read from the article directly rather than from a search summary — the practice E-227 exists to enforce. GRADED UNTESTED DESPITE BEING PUBLISHED, and that is the value of the row. The article names no instrument, no sample and no period for either duration, and the hedges are the source’s own words: "typically", "depending on the backlog". THE CONFLICT, which is the thing to carry forward: the article’s site-built comparator is five to six months. The Census Survey of Construction average this site quotes is 7.4 months for 2025 and 7.84 months on the five-year basis (E-002, E-139). Those are different bases, so the 60 days MUST NOT be subtracted from this site’s eight-month build figure to manufacture a factory saving. Doing that would repeat, one level up, the cross-cohort arithmetic E-002 records under C-016. DIRECTIONALLY it points the way the queued beat `not-the-factory` argues — factory construction compresses the build by a couple of months rather than by an order of magnitude — but it supports the direction only, and any beat citing it must state the grade on the face of the beat. DO NOT TIDY THIS AWAY, it is a lead: the same article cites a Harvard Joint Center for Housing Studies figure that construction costs for modern manufactured structures are "73 percent or less than those for site-built homes". That does NOT close E-098. E-098 asks for a like-for-like price per square foot at matched finish; this is a structure-cost ratio quoted at second hand, and the JCHS report has not been pulled by this project. Pulling it is the next step on E-098.

Grades live in the register; documents display them. Review dates are set on entry; no number is silently updated — a changed value gets a changed row and a bumped document revision. Source snapshots exist for the most recent rows only, which is an outstanding gap recorded in the corrections log. Rows E-037, E-038, E-039, E-051 and E-060 are recorded gaps: things nobody has measured, kept as rows precisely so they cannot be quietly forgotten. Rows E-134 and E-140 are the same kind of gap at national scale, and what the world publishes is where they are argued: no country in a 38-country compilation publishes the three phases of housing delivery as one regular series, so the absence is the finding rather than a caveat on one. The same absence at home is argued on what two states actually publish: California and Washington are the only states that compel a jurisdiction to report its permitting time, and neither publishes a duration computed from what it collects.