D_003 — Council profiling: a purposive 5-council sample, not a national catalogue
Decision
SQ4’s planning limb is answered by profiling a purposive sample of 5 councils, not by cataloguing New Zealand’s district plans. The deliverable characterises the distribution of council treatment; it does not enumerate it.
The sample (confirmed 2026-07-14):
| Council | Stratum it covers |
|---|---|
| Auckland Council | Unitary — district and regional rules in one plan, so both consent limbs are tested at once. Real cases: earthsong_ranui, Awaawaroa. Open ArcGIS. Urban + rural. |
| Tasman District | Strongest prior data — the district plan is already ingested ([[reg_003_tasman-district-plan-dwelling |
| Queenstown Lakes | Open ArcGIS. Heavily landscape-constrained rural land — the physically “hard” end. High lifestyle-block demand. |
| Selwyn or Waimakariri (Canterbury) | Tests a known-binding water constraint: the Canterbury LWRP Rule 5.8 on-site wastewater regime (≤2,000 L/day and site ≥4 ha) already surfaced in [[cr_046_nz-offgrid-permitting-register |
| Gisborne (or Far North) | Māori purpose zone / papakāinga provisions — the dimension none of the other four cover, and the NES-P is the only pathway that natively contemplates a community. Open ArcGIS. |
The artifact carries a three-tier answer: national default (held on primary authority) → council overlay where we have data → an explicit “council-specific — verify locally” flag everywhere else.
Rationale
1. The trigger — reg 9 makes the national default a floor, not a ceiling
REG_008 reg 5(1) makes one DMRU per site a permitted activity; reg 9 expressly provides that “a district plan rule that applies to a DMRU may be more lenient than the rule in regulation 5(1)”. So the community pathway is not nationally closed — it is council-dependent. That raised the obvious question: must we now read every council?
2. The RMA replacement makes an exhaustive scrape actively unwise
The Planning Bill + Natural Environment Bill consolidate district and regional plans into 17 regional combined plans, with Royal Assent targeted mid-2026 — against an August 2026 submission. A council-by-council catalogue would survey a landscape that is being dismantled while it is surveyed, and would invite the fair question of why plans that no longer exist were enumerated. This is the strongest argument against going wide.
3. The National Planning Standards are the generalisation lever
The NPS (RMA s58E, notified 5 April 2019) standardises zone names and definitions across all district plans — it is where “minor residential unit”, “Māori purpose zone” and “mixed use zone” are actually defined; the NES-DMRU merely points at it. Reading one instrument lets us reason about zones generically instead of reading seventy plans to learn seventy names for the same thing. → RT_342 (high).
4. The variance IS the finding
A complete, defensible answer to SQ4’s planning limb does not require a census: “There is no national permitted-activity pathway for a community of dwellings; the national default is one per site; district plans may be more lenient; therefore feasibility is council-dependent and must be verified locally.” Enumerating 70 plans would add precision we cannot maintain, do not need, and that the reform will invalidate.
The sampling frame — and one thing we deliberately do NOT sample on
Do not stratify on permissiveness — that is the outcome variable
The instinct is to pick “a lenient council and a strict one”. That is circular. Whether a district plan is more lenient than reg 5(1) is precisely what we are trying to measure. Selecting councils on it would assume the answer and the sample would tell us nothing.
We stratify only on criteria knowable in advance, and let permissiveness be the discovered outcome:
- A real NZ intentional community exists there — external validity; we can test the finding against a build that actually happened.
- Rural / rural-residential zone dominance — the off-grid case; an urban-only plan does not test it.
- Open ArcGIS / machine-readable plan — feasibility; this is what makes the GIS overlay possible at all.
- Māori purpose zone / papakāinga provisions — SQ4’s planning limb and the Te Tiriti dimension.
- Constraint intensity (hazard, landscape, water over-allocation) — spans easy → hard on physical grounds, not on planning generosity.
What each council profile must cover — THREE layers, not one
Each of the 5 sampled councils is profiled on three council-specific layers, because all three vary and all three bind:
- The district plan — zoning, setbacks, coverage, subdivision rules.
- NES-DMRU reg 9 leniency — whether the plan permits more than one DMRU per site (the trigger for this decision). REG_008
- The engineering Code of Practice — 🆕 NZS 4404 is a voluntary guideline, and councils adopt it “in whole or in part” with local amendments. Larger authorities have their own Codes of Practice outright; QLDC, New Plymouth and Southland all publish “based on NZS 4404:2010 with local amendments”. So the operative subdivision-engineering standard is the COUNCIL’s, never NZS 4404 itself — and since the RMA 1991 repealed the LGA 1974 provisions, there is no statutory requirement for a council to have one at all. OT_103
This third layer is what actually prices a community’s earthworks, roads, stormwater, wastewater and water-supply infrastructure — the cost lines beyond the dwellings.
Implications
- Scope: the per-council component of RT_279 is bounded to these 5. The residual (a national zoning/overlay scrape) is retired as impractical and reform-obsolete, not merely deferred.
- Deliverable: the SQ4 document reports a distribution and a method, plus a per-council profile for the 5 — not a national table. The council-dependence is presented as a finding, not a gap.
- Artifact: the sampler’s 🏛️ Planning panel shows the national default with a “verify locally” flag; councils in the sample get a real overlay. A tool that says “the national default is one per site; your council may permit more; here is the question to ask them” is more useful — and more honest — than one that pretends to know all 70.
- Reform: the RMA transition is written into the deliverable as a structural section, with the regime-independent requirements kept separable from the current instruments.
- Thames-Coromandel (wilderland) is not in the sample: it adds a fourth “rural, constrained” case without adding a new stratum. Revisit only if Wilderland becomes a worked case study.
Supporting sources
- REG_008 — reg 5(1) one-per-site; reg 9 district plans may be more lenient (the trigger for this decision).
- CR_046 — the Canterbury LWRP Rule 5.8 constraint that motivates the Selwyn/Waimakariri pick; the “no NZ plan contains a universal eco-village permitted-activity category” finding.
- REG_003 — the one district plan already held, making Tasman the cheapest and best-evidenced sample member.
Connections
Links to
Sources (4): CR_046 · OT_103 · REG_003 · REG_008
Cases (4): Earthsong Eco-Neighbourhood · Riverside Community · Tui Community · Wilderland
Referenced by
Sources (6): OT_103 · REG_010 · REG_020 · REG_021 · REG_022 · REG_023
Concepts (2): Council Planning Profiles (D_003 five-council … · Regulatory Consenting for Off-Grid Communities…