Permitted but Caught Out: Six Policy Changes Hit 170,000 Permitted Homes Mid-Flight

Between 2019 and 2021, Ireland granted 22,769 residential new-build permissions. Unit counts are recorded for 59% of these, totalling 100,513 units — implying roughly 170,000 when the gaps are filled. Six policy changes took effect between grant and build. Not all six hit every site — NZEB and the Section 42B expiry are universal, Part V's increase applies to applications filed after September 2021, the Defective Blocks Act only bites in pyrite and mica zones, and the RZLT only charges on zoned residential land. The 2019 cohort is the most exposed: 84% of those grants pre-date NZEB, 28% are in defective-blocks zones, and 16% sit on RZLT-liable land. For 2020-2021 grants, NZEB was already priced in at application, but the remaining changes still stack — 613 grants covering 6,099 units face both the RZLT charge and defective-blocks requirements simultaneously.

The changes: NZEB building standards (November 2019), Part V doubling from 10% to 20% (September 2021), the abolition of the SHD fast-track pathway with no extension mechanism for existing holders (February 2022), Defective Concrete Blocks Act foundation requirements (July 2022), Section 42B Covid relief expiry (December 2023), and the Residential Zoned Land Tax at 3% annually (first charges 2025).

No single change kills a scheme. But a developer who priced a scheme in 2019 budgeted for none of the ones that apply to their site.

How many permissions sit in the window?

Those 22,769 grants (unit counts recorded for 59% of applications) are the cohort most exposed to post-grant policy changes: granted before most of the six changes took effect, now approaching or past their five-year statutory expiry.

The 2020 pandemic-year cohort was the worst affected. For large schemes of 20 or more units on the standard planning pathway, 69.4% have no linked building control commencement record. That figure overstates real non-commencement because building control data matching is incomplete, particularly in Dublin where the linkage rate is approximately 23% versus 64% elsewhere. But even outside Dublin, the pattern holds: a substantial share of permissions granted during this window have not converted into construction starts.

What happened to the SHD pathway?

The Strategic Housing Development process was designed to fast-track large residential schemes by routing them directly to An Bord Pleanala, bypassing the local authority. It ran from 2017 to February 2022. In the council planning registers where SHD applications are visible, 105 grants appear covering approximately 27,652 units.

SHD permissions had one structural vulnerability that standard permissions did not: no extension of duration mechanism. A standard planning permission can apply for an extension under Section 42. SHD permissions had no equivalent. When the five-year clock ran out, the permission died.

The data confirms this mattered. For SHD grants visible in council registers, commencement rates dropped from roughly 64-72% for 2018-2020 grants to about 40% for 2021 grants. The pathway was abolished before many holders could start. They could not extend.

What replaced it tells its own story. Since 2022, 63 Large-scale Residential Development applications have been filed. Of those, 26 (41%) explicitly reference an existing SHD permission in their proposal text. They are amendments to existing SHD permissions rerouted through LRD because the original pathway no longer exists. Planners treat the SHD grant as the "parent permission" and the LRD as a modification vehicle. A Galway City planner report contains a side-by-side comparison table headed "COMPARISON WITH EXTANT SHD PERMISSION."

The LRD pathway was supposed to replace SHD. In practice, it is substantially being used to keep SHD permissions alive.

Who has no margin for error?

Lapse rates are high across the board. Portfolio developers let 95% of expiring uncommenced permissions lapse. Small portfolios: 91%. One-site developers: 83%. The spread is narrow — the story is not that one group lapses more, but that lapsing is the norm regardless of developer size. Portfolio developers can afford it because they hold options elsewhere.

The difference is what lapsing costs you. A portfolio developer who lets one of five permissions expire loses 20% of their pipeline. A one-site developer who loses their single permission loses everything. No other sites, no other revenue, no portfolio to absorb the hit.

Of 107 one-site developers with uncommenced permissions expiring in 2024-2025, 89 (83%) had no new application at the same site in the year before expiry. Their average scheme size was 54 units. Those 89 lapsed permissions represent roughly 4,800 units that left the pipeline of the developers who originally fought for them.

Some of these sites will be repermitted by other developers. Some will have been sold. Irish developers routinely use special purpose vehicles for individual sites, which can make portfolio developers look like one-site operators in the data. The 83% figure is the rate at which permissions lapse from their original holder, not the rate at which sites are permanently abandoned.

But one-site developers who do fight to reapply typically come back with smaller schemes. Where these developers reduced their unit count, the average cut was 47%, from 59 units down to 27. A 27-unit scheme is marginal for most developers. It may not be financially viable even if planning is secured.

Do developers come back with smaller schemes?

Between 2019 and 2021, reapplications at the same site reduced unit counts almost twice as often as they increased them: 160 reductions versus 97 increases. The average reduction was 52% of the original scheme, from 67 units down to 33.

The peak was 2020-2021, when 61-67% of reapplications reduced unit counts. It has since reversed. By 2024, only 30% reduced their scheme size and the average unit change turned positive. The reversal likely reflects survivorship: by 2024, only the most viable sites were still being reapplied for. The marginal sites had already been abandoned.

The commuter belt stands out. In Meath, Kildare, Wicklow, and Louth, 75% of reapplications reduced unit counts, versus 41-48% in other regions. The sample is modest (28 reapplications), and the commuter belt's larger average scheme sizes may account for some of this, but the pattern is consistent with the economics described in Part 1: the cost-to-price gap is widest in areas outside Dublin where new-build sale prices are below the cost of delivering them.

When rural schemes do shrink, the cuts are severe. The average rural reduction took schemes from 44 units down to 18, a 56% cut. A scheme of 18 units is below the threshold where most developers can achieve economies of scale.

Does the state's own delivery mechanism escape?

If this were a story about developer behaviour, the state's own housing delivery would be immune. Councils and Approved Housing Bodies have no land-banking incentive. Their sole purpose is to build as fast as funding allows.

AHBs received 28 residential grants of 10 or more units between 2019 and 2021 (covering 3,201 units). Seven AHB extension of duration applications appear in the planning register since 2016, roughly 90 units across seven cases. That is a small number, and it may mean AHBs overwhelmingly deliver on time. But even among the few that don't, the reasons documented in planning files are the same ones private developers cite.

Cabhru Housing Association applied for an extension on a 6-unit scheme in Dun Laoghaire-Rathdown in January 2024. The original permission was granted in 2018. The planner's report records "delays and significant cost increases and Covid 19." CAS funding, the Capital Assistance Scheme administered by the Department of Housing, had been submitted and was "awaiting approval."

Six years after permission was granted, the state's own funding mechanism had not kept pace with the planning permission's lifespan.

Respond Housing Association, Ireland's largest AHB, sought an extension in Waterford for 10 sheltered housing units for the elderly. The planner noted that "substantial works were not carried out due to delays associated with accessing funding." In Louth, Respond filed another extension for a shop-to-dwelling conversion. The planner recorded that "funding and housing approvals were not forthcoming."

Tramore Voluntary Housing Association sought an extension for 18 sheltered elderly units in Waterford. A watermain diversion had caused "significant and unanticipated delay to the project which has been designed up to tender" stage. The infrastructure was the council's own responsibility.

These are non-profit organisations whose sole purpose is delivering social housing. On this evidence, when they cannot start within the five-year permission window, the constraint is not behaviour. It is the system.

What do developers and planners say about compounding costs?

The planning documents from extension of duration applications describe the same pattern in different language across different councils.

The planner assessing this extension summarised the applicant's case as "Covid-19, shortage of building materials, and costs." The extension was granted.

The Murphystown applicant had entered a development agreement to bridge the "viability gap between the cost of building apartments and the market sale price." This is the gap described by the cost data in Part 1 of this series, stated in a planning document.

A Kildare planner flagging that the development may fall short of the new 20% Part V requirement depending on when the land was purchased. This is the regulatory transition in action: a scheme designed around a 10% obligation now assessed against 20%.

This language was largely absent from planning documents filed in 2019. In a sample of 500 residential new-build applications from 2024-2025, three of the top five cost-related document matches contained explicit inflation disclaimers referencing external benchmarks like the Line Sight Average Irish Construction Costs index. "No provision for future inflation" appears as a standard caveat in Part V cost submissions. It did not appear in the equivalent 2019 sample, though the 2019 search had lower document extraction coverage (18% for submissions versus 96% in 2024-25).

What does this mean for the pipeline?

NZEB improves building quality. Part V delivers social housing. The Defective Blocks Act protects homeowners. None of the six policy changes listed at the top of this article are mistakes.

But every improvement is a cost to someone who already has permission. The developer who priced a scheme in 2019 did not budget for NZEB compliance or a 3% annual RZLT charge — and if they need to modify or reapply, the Part V obligation has doubled. The housing association that secured permission in 2018 did not anticipate that CAS funding would still be awaiting approval six years later.

The planning system grants permissions. The regulatory system changes what those permissions require. The tax system charges for non-delivery. None of them references the others.

If you hold a 2019-2021 permission approaching expiry, the extension data from The Silent Lapse applies with additional dimensions: the Part V regime may have changed since your grant, NZEB compliance costs were not in your original costings, and the RZLT clock is running. The planning conditions on your extension may differ from the conditions on your original grant.

If you are reapplying at a previously permitted site, the 2020-2021 data shows developers typically reducing scheme sizes by around half. The commuter belt pattern is most pronounced. Whether the scheme remains viable at a reduced unit count depends on the site-specific cost structure — which the planning data can quantify.

If you advise on regulatory exposure, the audit for a specific site is: which of the six changes post-dates the grant? Is the site RZLT-liable? In a pyrite or mica zone? Was Part V at 10% or 20% when permission was granted? Did the permission come through SHD (no extension) or the standard pathway? The answers are in the planning register.