Five Systems, No Feedback: Why Ireland's Planning Complexity Advantages Scale

Eighty-two percent of developers who received residential grants for ten or more units since 2018 hold exactly one permission. Outside Dublin, 45% of those one-grant developers have a linked commencement notice, a formal building control record that construction began. Developers with six or more grants commence at 58%.

The absolute rates understate true commencement because building control data matching is incomplete. But the same method is applied to every developer, so the gap between groups is informative. That 13-point gap is not explained by site location, scheme size, or regulatory exposure. It holds within every scheme size band and across every region. Restrict the comparison to standard planning permissions only, stripping out extensions of duration and compliance submissions, and the gap widens to 25 points: 48% for single-grant developers, 73% for the largest operators.

On this evidence, the rules are not heavier on some sites than others. The system advantages scale.

What does the planner see?

Five state systems govern any residential site of ten or more units in Ireland: planning (the permission itself), Revenue (the Residential Zoned Land Tax), building control (NZEB energy standards), social housing (Part V obligations), and infrastructure (wastewater treatment, water supply). Each system appears in planner reports. None references the others.

We searched 500 recent planner reports across all 31 councils (98% document extraction coverage). What the planners write, and what they don't, tells the story.

The RZLT is virtually invisible. Two genuine mentions in 500 reports. One Meath planner cited the RZLT as a policy objective: "delivery of housing units on this site will realise the aims and objectives of the RZLT." One Wicklow planner noted the problem: "raised issue of RZLT i.e. zoned lands but can't develop." Every other match was a false positive, generic zoning descriptions where "residential" and "zoned" and "land" happened to appear near each other.

The planning system processes applications on RZLT-liable land without knowing the site carries a 3% annual tax.

NZEB is a compliance checkbox. Nearly Zero Energy Building standards appear in planner reports frequently, but only as a compliance statement. "The proposed building will meet or exceed the NZEB requirements." "Designed to comply fully with NZEB requirements." Not once was NZEB discussed as a cost factor. Planners note the building regulations exist. They do not assess what they cost.

Part V is the most visible of the five systems, with its own section in virtually every residential planner report. But it is a procedural entry: "Part V obligation to a total of 22 units." The number of social housing units is recorded, the agreement type noted, and the planner moves on. The obligation is never connected to scheme viability.

Then there is viability itself. In extension of duration applications, planners routinely document financial difficulty. "Making it unfeasible to carry out some projects." "The applicant is unable to complete the development within the lifespan of the permission due to financial circumstances." "Considerable financial outlay associated with securing permission is acknowledged."

These phrases appear in reports for schemes on RZLT-liable land, in pyrite zones, subject to Part V at 20%, built to NZEB standards. But no planner report connects the financial difficulty to any specific regulatory layer. The planner sees the symptom. The diagnosis never happens, because no single report has visibility into all five systems.

The D18A/0799/E case in Dun Laoghaire-Rathdown is typical. The planner documents that the development is "unfeasible" due to cost inflation in one section and cites the Part V obligation in another. The two sections do not reference each other.

Has a sixth system emerged since?

This analysis used data through mid-2025. Since then, a sixth system has become material: rent regulation.

Rent Pressure Zones were extended nationwide in June 2025, applying rent-growth caps to areas that were previously uncapped. March 2026's Residential Tenancies Amendment Act added a CPI-only carve-out for apartments commenced after 9 June 2025, creating a bifurcation between old-regime and new-regime commencements. Delivering Homes, Building Communities (DHBC), which replaced Housing for All in November 2025, removed the explicit annual cost-rental target that had anchored AHB and LDA acquisition planning.

Unlike the five supply-side systems, the rent regime does not appear in planning files. The Residential Tenancies Board administers tenancy law. The Department administers the caps. A planner assessing a scheme cannot see whether it will clear the new rent regime; the RTB cannot see which permitted schemes it is rate-capping. A PRS-oriented permission granted in 2023 may now face a fundamentally different revenue side than the one it was priced against. Part 4 covers the full rule-change catalogue.

Six systems, one site, still no feedback.

What does this look like on one site?

In May 2020, a developer applied directly to An Bord Pleanala under the SHD fast-track pathway for 239 residential units on zoned land west of Leixlip, Co. Kildare. The site is zoned R1 (new residential), RZLT-liable, in a designated pyrite risk county, within the 60-64 dB noise band from the M4 motorway, and served by a wastewater treatment plant. All five systems, layered on a single field.

ABP granted the scheme in September 2020 with 30 conditions.

Seventeen months passed before the first commencement notice was filed in February 2022. The developer spent that time navigating compliance submissions for the 30 conditions, many requiring sign-off from different council departments or external bodies. NZEB compliance for the energy standards. Part V agreement for the social housing units. Pyrite precautions for the foundations. Roads and infrastructure for the site access. Water and drainage for Irish Water. No single body tracked the overall timeline.

The unit count grew from 239 to 243 through subsequent amendments. The SHD permission had no extension of duration mechanism. The clock ran from September 2020. By October 2025, it expired with 236 of the 243 units complete but seven houses only partially constructed. The engineering report on the subsequent retention application states that "permission lapsed in October 2025, resulting in these 7 units only partially constructed." The planner's report notes that "when the developer became aware that permission had expired, works were already substantially underway."

On the adjacent estate, Block D was reduced from 28 apartments across five storeys to 24 across four. The top floor was removed entirely. A Phase 2 LRD application for 229 additional units was appealed by third parties, adding three months. It was later amended to 227 units to accommodate a creche relocation and ESB undergrounding.

This is one site. A well-capitalised national developer, not a one-grant operator. They still ran out of clock.

Who actually builds?

The developer size gradient is the most consistent finding in this research. It holds within every scheme size band we tested, across every region.

Outside Dublin, where building control data is more reliable, the pattern is monotonic. One-grant developers: 45%. Two-to-five grants: 54%. Six or more: 58%.

We cannot say from the data whether larger developers succeed because they push through the complexity or because they have the resources to avoid it, choosing cleaner sites, pre-consulting with councils, structuring applications to minimise appeals. The data shows the outcome. The mechanism is harder to pin down.

Scheme size matters too, though not the way you might expect. The highest commencement rate outside Dublin is in the 51-to-100-unit band: 53%. Below that, 10-to-20-unit schemes commence at 47% and 21-to-50 at 50%. Above 100 units, commencement drops sharply: 34% for 101-to-200 units, 28% for schemes over 200.

The 51-to-100-unit range is where the economics seem to work: large enough to absorb the overhead of five systems, small enough to finance without institutional backing.

The gradient above reflects 2018-2024 data. Since that window closed, the top of the scale has substantially withdrawn from new filings. Of the fifteen operating groups that dominated Dublin's 2022-2025 apartment-led bulk-PPR, nine have stopped filing new residential permits against an industry baseline rising 9.7%. The system isn't only selecting who commences. Since 2024 it is also selecting who continues to apply.

Local authority housing stands apart. South Dublin County Council's own housing applications commence at 93%. The council builds what it permits. It is the applicant, the planning authority, the Part V administrator, and the infrastructure provider. In a system of silos, it is the one entity that can see through the walls. Since 2024, the Land Development Agency and the AHB sector have grown into a similar coordinating role at the Dublin-region scale.

What we expected to find but didn't

RZLT does not predict commencement. Outside Dublin, 32% of commenced schemes are on RZLT-liable land, compared with 31% of uncommenced schemes. A 1.9-point difference. Noise. 571 commenced schemes carry RZLT exposure. 183 have the double overlay of RZLT and defective blocks risk. The tax makes development harder. It does not distinguish which schemes get built.

The regulatory overlay penalty is also weaker than expected. We counted how many state system flags each site carries: RZLT, flood zone, defective blocks, ABP case, failing wastewater plant. Sites with zero overlays commence at 48%. Sites with four: 42%. Six points over four additional system layers. The number of systems hitting a site matters less than who is holding the permission.

Demand does not differentiate either. Census 2022 metrics for population, vacancy, unemployment, and deprivation are nearly identical between commenced and uncommenced schemes. Both groups are in similar areas. The difference is on the supply side.

What this means before you submit

If you are a one-site developer scoping a residential scheme of ten or more units, the data says your baseline commencement probability outside Dublin is 45%. The system does not actively block you. 571 schemes with RZLT exposure have commenced. The regulatory overlay penalty is modest. But the 13-point gap between you and a multi-site developer is real, and it reflects an advantage in either navigating or avoiding the complexity of five independent systems.

If you are advising on site acquisition, the scheme size sweet spot is 51-to-100 units. Below 50, the compliance overhead per unit is higher. Above 100, the financing and coordination requirements steepen. Above 200, fewer than three in ten commence.

If you are a council planner, the silo pattern is not a criticism of individual decisions. No planner should be expected to administer the RZLT or assess NZEB costs. But the cumulative effect of five systems operating independently on the same land is visible in the data. It filters by developer capacity.

Methodology

Population. 3,855 residential grants of 10 or more units received since 1 January 2018 across all 31 Irish local authorities, covering 267,883 units.

Commencement. A grant is classified as "commenced" if it has a linked record in the National Building Control Office register. This understates true commencement: building control data matching is incomplete, particularly in Dublin where the linkage baseline is approximately 23% versus 64% elsewhere. For this reason, all headline commencement comparisons use outside-Dublin figures unless stated otherwise. Absolute rates are lower bounds; relative comparisons between groups using the same methodology are informative.

Developer size. Based on distinct applicant name counts in the planning register. This overstates the number of independent developers because portfolio operators use Special Purpose Vehicles — separate legal entities for each site. Glenveagh Homes appears under three name variants totalling 41 grants; Cairn Homes under two variants totalling 27. The 82% one-grant figure is an upper bound on true single-site developers.

Document searches. 500 recent residential applications (10+ units, received since January 2024) with planner report extraction at 96-98% coverage. Searches used BM25 full-text matching for "RZLT", "Residential Zoned Land Tax", "NZEB", "Nearly Zero Energy Building", "Part V", and viability-related terms. All reported match counts are from the full 500-application scope.

System overlays. RZLT liability from GeoHive mapping (defaults to false for unmapped sites). Flood zone from OPW CFRAM. Defective blocks risk is a county-level flag covering Donegal, Mayo, Clare, Limerick, Sligo (mica) and Fingal, Meath, Offaly, Kildare (pyrite) — not all properties in these counties are affected. ABP case linkage is a boolean flag on the planning register. Wastewater treatment plant compliance from the EPA.

Leixlip case study. Ref. ABP-307223-20 (KCC 20/307223). Document IDs cited: 138004778 (ABP decision), 304332230/237/238 (retention application 26/60406). All dates from council register and building control records.

Figures deliberately excluded. The 71% increase in documents per granted scheme (2018-2023) was dropped from the body text after the critical review found it substantially inflated by growing average scheme sizes and improved council register digitisation. The finding that commenced schemes have more documents than uncommenced schemes was dropped after controlling for application type revealed it to be a composition artifact. Further information request rates were downgraded to illustrative status because seven major councils (representing 47.5% of grants) do not report FI data separately.

Sources