Nine independent agents searched the web to verify every major claim in this site. Here's what they found.
Nine separate fact-checking agents ran concurrently, each assigned one page of the site. Each agent:
A ninth agent separately assessed the analytical methodology of the overview page. All errors identified as LIKELY WRONG or WRONG have been corrected in the live pages.
The site's core statistical claims — cost figures, land price data, legal citations, case outcomes — held up well under scrutiny. The errors that were found fall into three categories: minor dating mistakes, one wrong section number, and a few cases where nuance was lost (the Supreme Court/Kenny Report claim being the most significant).
All of the following have been fixed in the live pages.
Highly accurate. The SHD 217/167 application count is in the right ballpark but the exact cut-off date varies by source (183–312+ across different reporting periods). The Section 47 vs 247 error is a minor internal inconsistency — the page body uses the correct number.
Strong overall. One confirmed error corrected: Part E impact sound 62 dB → 58 dB L'nT,w. Two further errors flagged by the first fact-checker (S.I. 365 Part L attribution; nZEB primary energy ≤25 kWh/m²/yr for apartments) could not be found in the current page — the page uses EPC ≤0.30 as its primary metric and never states a kWh threshold for apartments specifically. Phantom findings removed from scorecard.
Priory Hall and Celtic Tiger defects data is largely correct. Three corrections in this pass: US fire death rate updated to 13.1/million (USFA/FEMA 2023, not ~9-10 as stated); Netherlands building control updated to reflect partial privatisation under Omgevingswet (Jan 2024); Fiachra Daly death corrected to reflect coroner's open verdict rather than stated suicide.
Strong page overall. The treatment of who bears the Part V cost (developer vs. landowner vs. buyer) was found to be fair and balanced. The Cherrywood and Luas levy figures confirmed. DLR S.49 rates are indexed annually — caveats added to live page noting rates should be verified with DLR directly. Part V cumulative figure updated from ~26,700 (unverified estimate) to 25,631 (official gov.ie dataset to Q3 2025). €91m cash buy-out figure retains unverified caveat.
Strong on EU law citations and case law. Two corrections made: heritage value corrected to €1.5bn GVA (the Ecorys source figure, not €2bn); Lismullin Henge case study updated to clearly state the site was demolished under "preservation by record" — it is a cautionary tale, not a conservation success story. No systematic bias found.
The CSO land price data is exact. Four errors corrected: the Supreme Court 2018 Hargey case does not exist; the LVS rate is 25% (not 30%) as enacted; Section 34(6) material contravention threshold is three-quarters (not two-thirds); and the interim review under PDA 2024 is at year 5 (not 4). The core strategy "44,000 hectares" figure is cited differently across sources (40,000–44,000).
Very strong on utility charges and ESB investment figures, all verified. Three additional corrections: EECC attribution replaced with correct Directive 2014/61/EU; SuDS framing softened (no single national statutory mandate); development bond confidence level downgraded C1→C3. GDD date error (2018→2019) corrected in previous pass.
Strong overall. Three corrections: O'Grianna heading removed (a wind energy case, not heights); SPPR constitutionality updated from "pending" to resolved (Supreme Court dismissed July 2024); land cost arithmetic fixed (€59k at 11%, not €70.7k). SCSI and TCD/SCSI cost data verified. July 2025 relaxations fully confirmed.
The methodology checker rated the overview page's analytical framework on a scale of 1 (strongly pro-developer/anti-regulation) to 10 (strongly pro-regulation).
Score: 3.5/10 — Modestly leans toward supply-reform / pro-developer-cost-relief framing, but more balanced than typical advocacy. The three-lens structure, Murray critique inclusion, and honest "fine print" section all demonstrate genuine epistemic care.
The additive approach is a legitimate analytical tool but imagines a zero-regulation baseline. The marginal regulatory cost — above what competent builders would voluntarily do — is lower than the totals suggest. The page acknowledges this limitation. The €115,000–133,000 range is internally consistent with the components but relies on several inferred cost allocations (per-Part building reg costs, cost-of-delay model).
The CSO land price data is exact. The 23x/27x multipliers are correct arithmetic. The framing that planning "creates" the zoning uplift is broadly correct but slightly oversimplified — infrastructure investment, transport links, and economic agglomeration also contribute. The Kenny Report framing (as a missed opportunity) reflects a specific policy position that is mainstream in Irish housing discourse but not universally held.
The Lyons (2024) elasticity figures (-1.9 cost, +0.9 price) are verified and correctly cited. The viability gap analysis is sound. The 39,000 uncommenced permissions is verified. The supply-restriction framing is well-supported by the Irish-specific evidence base (Lyons, ESRI, IMF).
Murray critique included substantively. Mulheirn/credit school given real treatment. Norris & Shiels (2007) included as counterevidence. The Hsieh-Moretti 36% GDP claim — verified as correct numerically — should note that the paper has been subject to formal academic criticism and acknowledged computational errors; a caveat has been added to the live page. It is also a US estimate that may not translate directly to Ireland. The Louie/Mondragon/Wieland 2025 finding is correctly described as contested.
The methodology checker identified these perspectives as underrepresented or absent:
Fact-checked by nine independent Claude agents running concurrently on 26 February 2026. Each agent used live web search rather than training data knowledge. Full reports for each page are available at /workspace/planning-permissions/verification/. Errors identified as LIKELY WRONG or WRONG have been corrected in the live pages. The remaining minor errors are documented here but are lower-priority corrections that do not affect the site's core analysis.
This verification exercise itself has limitations: the fact-checkers used web search, which may miss paywalled academic sources, and the agents' own rating judgements carry uncertainty. Confidence in the ratings is highest for statutory/legislative claims and lowest for industry cost estimates.