Four errors in officer reports under the 2026 NPPF: 243 reports from 27 councils scored
The Stratford note found four recurring errors in Stratford-on-Avon District Council's officer reports: the settlement tier used as the sustainable-location test, S4 or S5 applied where the Green Belt policies govern, the DP3(3) question not asked after a design conflict, and plan policies cut wholesale for the housing shortfall. This page asks whether the errors are Stratford's or everyone's. It scores 243 housing reports from 27 councils on one published rubric (99 read in full from 22 councils, then 144 from a sweep of 17 councils' registers), shows what the better reports do instead, and says which of the Stratford counts were too hard.
In short #
- The Stratford note found four recurring errors in Stratford-on-Avon District Council's officer reports under the National Planning Policy Framework (NPPF) of 17 August 2026. We scored 243 reports from 27 other councils with the same rubric (99 read in full from 22 councils, then 144 from a sweep of 17 councils' registers) to see whether the errors are Stratford's or everyone's.
- They are mostly Stratford's. Scored the same way, the other councils rest a location finding on a village's tier in 7% of their location reports against Stratford's 20%, take the wrong route in Green Belt villages in 8% against 23%, and recommend approval after an unaddressed design conflict in 10 reports between them against Stratford's 17. The paragraph that cuts the settlement hierarchy policies to very limited weight wholesale appears in 54% of Stratford reports and 10% of the others.
- One council repeats the pattern. Warwick District Council, which shares Stratford's Green Belt and plan structure, makes the tier error in 2 of its 3 location reports and applies S4 in a washed-over village at Norton Lindsey. Wychavon, Elmbridge and Dacorum, on comparable sites, do not, and Bromsgrove's one tier error in 12 location reports is a short notice with no report behind it.
- Two of the Stratford counts are re-presented. A text search for "TR3" counts small schemes that officers and inspectors often treat as below the TR3(1)(a) threshold, and reports written to the 2024 Framework that could not name the policy. The measure used here instead asks, of every report that has to decide whether the location is sustainable, whether it applies TR3. Stratford applies it in 14 of 35 such reports (40%) against 36% across the 27 other councils, and decides the location on the route, with facts, in 40% against 74%. On Green Belt sites, where GB7(1)(g)(iii) asks the question "with particular reference to policy TR3" whatever the scale, Stratford applies TR3 in 7 of 14 against 26 of 45. DP3(3) "not asked" on a refusal is a formal slip that cannot change the outcome; the finding that matters is the 17 Stratford recommendations to grant after a found design conflict, against 3 across 22 councils.
- One absence is shared. One Stratford report in 59 uses the Connectivity Tool, the Kineton refusal of 4 November 2026. Nor do 232 of the 243 reports elsewhere.
- The reports read in full say the same on their own. Among the 99 reports from 22 councils read cover to cover, the tier error runs at 7%, the Green Belt route error at 14%, the wholesale weight cut at 6%, and approval after an unaddressed design conflict 3 times. The sweep set of 144 reports, read through keyed passages, gives 8%, 3%, 13% and 7.
The four errors #
The Stratford note scored every officer report held for a decision made, or a committee item published, from 17 August 2026 against four errors. The same rubric, with the same codes, was applied to the other councils' reports. Each error is a step the Framework asks for and the report leaves out.
- A village's tier used as the location test. The report treats the village's place in the settlement hierarchy, or a list of its facilities, as showing that the location is sustainable, instead of assessing the route and the transport choice. The Framework's test is in TR3(1)(a): development that "could generate a significant amount of movement" should be in a location that will "limit the need to travel, particularly by private car, and offer a genuine choice of transport modes for residents and users". Service village does not mean sustainable sets out how inspectors apply it.
- TR3 not assessed. The report does not run the limbs of TR3(1), and does not use the evidence TR3(2) asks for: "The Connectivity Tool (Connectivity Tool - GOV.UK) should be used alongside other relevant quantitative or qualitative evidence in assessing the connectivity of particular locations proposed for development". The question arises wherever the report has to decide whether the location is sustainable: for a Green Belt site taken through GB7(1)(g), whose limb (iii) asks whether "The development would be in a sustainable location, with particular reference to policy TR3 of this Framework", whatever the scale; for a site outside a settlement, where S5(1)(j) asks whether it is "physically well-related to an existing settlement" and S5(4) weighs the effect "in relation to promoting sustainable patterns of movement"; and for any proposal that "could generate a significant amount of movement, in the context of the area within which they would be situated", which inspectors have found at one to five homes in a village or an open rural setting. How to assess a sustainable location explains the test and the Tool.
- Annex B and the Green Belt route. Annex B's definition of a settlement says settlements "do not include hamlets and scattered groups of houses located outside predominantly built-up areas" and "also exclude villages which lie within and are defined as part of the Green Belt in the development plan". S5(5) says S5 "does not apply to development proposals in the Green Belt". The error is applying S4 or S5(1)(j) in a washed-over village or a hamlet, or using a built-up area boundary or a tier as the GB7(1)(g)(iii) test. See What is a "settlement"?
- DP3(3) not asked. A design or character conflict is found and weighed in the balance, but the DP3(3) question is not asked: proposals "should be refused if, without clear justification, they conflict with paragraph 1 of this policy or relevant aspects of the principles in paragraph 2". S4(2)(c) and S5(2) make a failure under such a policy a case where the benefits are "likely to be substantially outweighed". See DP3: the design policy.
Two further columns record whether the report applied the 2024 Framework without conversion after 17 August 2026, which Annex A(1) does not allow (the 2026 policies "must be taken into account in decision-making from the day of its publication"), and whether it carries a template paragraph giving the plan's settlement hierarchy policies very limited weight wholesale, where Annex A(2) cuts weight only from policies "(or parts of those policies) which are materially inconsistent with national decision-making policies".
Codes: error; partial for a tier finding with some route fact beside it; cited-only and dodged for TR3 named but not run; ok; na where the question does not arise. Rates below are among the reports where the question arises. The table of the 99 hand-read other-council reports, the sweep table of 144 and the Stratford table carry a page reference and a verbatim quotation for every adverse code, and every "not mentioned" is a text search with no hits.
The comparison #
| Error | Stratford (59 reports) | 27 other councils (243 reports) | Reading |
|---|---|---|---|
| Tier used as the location test | error 7, partial 14, ok 14 of 35 error 20%; error or partial 60% | error 9, partial 23, ok 90 of 122 error 7%; error or partial 26% | A Stratford habit, shared by Bromsgrove (1 of 12), Cotswold (3 of 7), Maidstone (1 of 9), North Hertfordshire (2 of 4) and Warwick (2 of 3). Bromsgrove's one is a three-page notice with no report behind it. 15 councils with two or more location reports have no tier error at all. |
| TR3 not applied where a sustainable-location decision is made | 21 of 35 location decisions 60%; location decided on the route in 14 (40%) Green Belt sites: TR3 applied in 7 of 14 | 78 of 122 location decisions 64%; location decided on the route in 90 (74%) Green Belt sites: TR3 applied in 26 of 45 | The measure for the complaint's second head: where a report has to decide whether the location is sustainable, does it apply TR3? Naming the policy is uncommon everywhere, but the other councils decide the location on the route in 74% of those reports against Stratford's 40%, and apply TR3 in terms on 58% of Green Belt sites, where GB7(1)(g)(iii) asks for it, against Stratford's 50%. |
| TR3 not run, raw count of all reports | not mentioned 17, cited only 8, dodged 1, assessed 14 of 40 not run 65% | not mentioned 107, cited only 55, dodged 1, assessed 52 of 215 not run 76% | Common everywhere and not a fair measure on its own: it counts small schemes and reports written to the 2024 Framework. The row above is the measure used; the size split is below. |
| Connectivity Tool used | 1 of 59 | 11 of 243, plus 1 indirect Bath and North East Somerset (1), Cheshire East (3), Elmbridge (2), Hertsmere (1), Tonbridge and Malling (1), Wychavon (3) | Absent almost everywhere. Where it appears it is one team's template. |
| Annex B and the Green Belt route | error 6 of 26 Green Belt or hamlet cases 23% | error 9 of 111 8% | Stratford and Warwick apply S4 or S5(1)(j) in washed-over villages. Elsewhere most of the errors are belt-and-braces runs of S5 before S5(5) is applied; only Basildon and North Hertfordshire change the route. |
| DP3(3) not asked after a design conflict | error 23 of 56 41%; 17 of the errors recommend grant | error 35 of 233 15%; 10 of the errors recommend grant | The clearest Stratford outlier, partly a composition effect: other officers rarely find a design conflict and then recommend approval. |
| 2024 Framework applied without conversion | 4 of 59 | 26 of 243 | Rare at Stratford and in the hand-read set; the sweep found 23 more, most of them notices issued on 2024 or 2025 committee resolutions once a legal agreement was signed. The same defect as Stratford's Checketts Abattoir. |
| Template paragraph cutting the hierarchy policies to very limited weight | 32 of 59 54% | 24 of 243 10% Babergh and Mid Suffolk (5), Cotswold (4), East Devon (3), Maldon (2), North Hertfordshire, Swale, Tonbridge and Malling (7), West Northamptonshire | A Stratford peculiarity in form. The other 8 councils cut hierarchy policies on five-year-supply or age grounds, the 2024 logic that Annex A(2) also does not allow. |
The Stratford column counts all 59 reports. The table below keeps the 44 reports scored on 13 October 2026 apart from the 11 delegated reports on decisions issued 20 to 30 October, scored on 2 November, and the 4 delegated reports on decisions issued 30 October to 5 November, scored on 5 November, so that what each later batch changed can be seen. The eleven decide the location on the route more often (5 of 7 location cases against 7 of 25 before) and apply TR3 more often (6 of 10 TR3-scored reports against 6 of 27); none of them adds a tier error or an Annex B error; the template paragraph is commoner among them (9 of 11 against 19 of 44), and 4 of the eleven do not ask DP3(3) after a design conflict, 2 of them before a recommendation to grant. The four of early November add one tier error (a single self-build home granted at Napton-on-the-Hill on being "adjacent to the Built Up Area Boundary" of a Category 2 village, with no route fact), two route-based location findings, the first Stratford use of the Connectivity Tool (a score of 40 at Kineton) and one DP3(3) omission, on a refusal; the Kineton report asks the DP3(3) question in the design standard's own words, testing the SPD's "exceptional circumstances" exception for a forward garage and finding none, which the rubric counts as the test applied; all four carry the template paragraph. All fifteen later reports are delegated; the first set had 18 committee reports among its 44, so the batches are not like-for-like samples.
| Error | 44 reports scored 13 October | 11 decisions of 20 to 30 October | 4 decisions of 30 October to 5 November | All 59 |
|---|---|---|---|---|
| Tier used as the location test | error 6, partial 12, ok 7 of 25 | error 0, partial 2, ok 5 of 7 | error 1, partial 0, ok 2 of 3 | error 7, partial 14, ok 14 of 35 |
| TR3 applied where a location decision is made | 6 of 25; route-based 7 | 6 of 7; route-based 5 | 2 of 3; route-based 2 | 14 of 35; route-based 14 |
| TR3, raw count | not mentioned 13, cited only 7, dodged 1, assessed 6 of 27 | not mentioned 3, cited only 1, dodged 0, assessed 6 of 10 | not mentioned 1, cited only 0, dodged 0, assessed 2 of 3 | not mentioned 17, cited only 8, dodged 1, assessed 14 of 40 |
| Connectivity Tool used | 0 of 44 | 0 of 11 | 1 of 4 | 1 of 59 |
| Annex B and the Green Belt route | error 6 of 17 | error 0 of 7 | error 0 of 2 | error 6 of 26 |
| DP3(3) not asked after a design conflict | error 18 of 41; 15 recommend grant | error 4 of 11; 2 recommend grant | error 1 of 4; 0 recommend grant | error 23 of 56; 17 recommend grant |
| 2024 Framework applied without conversion | 4 of 44 | 0 of 11 | 0 of 4 | 4 of 59 |
| Template paragraph | 19 of 44 | 9 of 11 | 4 of 4 | 32 of 59 |
The three tables were scored with one written rubric; 99 of the other councils' reports were read in full and 144 through keyed passages, as the sweep section explains. The Stratford reports are mostly delegated reports on one to seven homes in villages; the other councils' reports are a mix of delegated reports and committee reports on larger schemes, which matters for TR3 and DP3(3) and is taken up below. Rates for the other councils are also pulled down by their urban reports, where the tier, Annex B and route questions do not arise and are coded na on both sides.
Council by council #
Tier used as the location test
No error, 20 councils: Wychavon (11), East Devon (10), Basildon (9), Malvern Hills (9), Tonbridge and Malling (8), Bath and North East Somerset (5), Cheshire East (5), Solihull (5), Uttlesford (5), Babergh and Mid Suffolk (4), Elmbridge (3), Chorley (2), Dacorum (2), Lichfield (2), Maldon (2), Hertsmere (1), Sefton (1), Swale (1), Three Rivers (1), West Northamptonshire (1).
DP3(3) not asked after a found design conflict
No error, 12 councils: Wychavon (17), Basildon (9), Lichfield (9), Sefton (8), Warwick (7), Bath and North East Somerset (6), Cheshire East (6), Hertsmere (6), North Hertfordshire (4), Chorley (3), Maldon (2), Three Rivers (2).
Counts exclude na rows. Small counts are small: 3 of the 27 councils are represented by one or two reports, because that is what their committees decided in the period.
| Council | Reports | Tier error / partial / ok | TR3 not mentioned / cited only / dodged / assessed | Tool | Annex B error / ok | DP3(3) error / ok | 2024 unconverted | Template |
|---|---|---|---|---|---|---|---|---|
| Babergh and Mid Suffolk | 18 | 0/0/4 | 2/13/0/2 | 0 | 0/8 | 2/14 | 1 | 5 |
| Basildon | 9 | 0/2/7 | 7/0/0/2 | 0 | 2/5 | 0/9 | 0 | 0 |
| Bath and North East Somerset | 6 | 0/0/5 | 1/1/0/3 | 1 | 1/3 | 0/6 | 0 | 0 |
| Bromsgrove | 13 | 1/4/7 | 5/0/1/6 | 0 | 2/9 | 1/11 | 1 | 0 |
| Cheshire East | 6 | 0/2/3 | 2/1/0/3 | 3 | 0/2 | 0/6 | 1 | 0 |
| Chorley | 3 | 0/1/1 | 1/1/0/0 | 0 | 0/1 | 0/3 | 0 | 0 |
| Cotswold | 21 | 3/2/2 | 9/9/0/1 | 0 | 0/13 | 6/15 | 1 | 4 |
| Dacorum | 4 | 0/0/2 | 1/2/0/1 | 0 | 0/3 | 2/2 | 0 | 0 |
| East Devon | 13 | 0/1/9 | 11/1/0/1 | 0 | 0/5 | 5/7 | 5 | 3 |
| Elmbridge | 4 | 0/0/3 | 0/0/0/3 | 2 | 0/2 | 1/3 | 0 | 0 |
| Hertsmere | 6 | 0/0/1 | 4/0/0/2 | 1 | 0/3 | 0/6 | 0 | 0 |
| Lichfield | 9 | 0/1/1 | 5/0/0/1 | 0 | 0/3 | 0/9 | 1 | 0 |
| Maidstone | 21 | 1/0/8 | 3/15/0/1 | 0 | 1/10 | 5/16 | 1 | 0 |
| Maldon | 2 | 0/0/2 | 0/1/0/1 | 0 | 0/0 | 0/2 | 0 | 2 |
| Malvern Hills | 16 | 0/1/8 | 10/0/0/2 | 0 | 0/5 | 4/12 | 3 | 0 |
| North Hertfordshire | 4 | 2/0/2 | 1/3/0/0 | 0 | 1/0 | 0/4 | 0 | 1 |
| Nuneaton and Bedworth | 1 | 0/0/0 | 0/0/0/0 | 0 | 0/0 | 0/0 | 0 | 0 |
| Sefton | 10 | 0/1/0 | 7/0/0/1 | 0 | 0/1 | 0/8 | 0 | 0 |
| Solihull | 8 | 0/1/4 | 2/0/0/6 | 0 | 0/3 | 1/7 | 2 | 0 |
| Swale | 6 | 0/0/1 | 5/1/0/0 | 0 | 0/2 | 3/3 | 1 | 1 |
| Three Rivers | 3 | 0/0/1 | 0/1/0/1 | 0 | 0/3 | 0/2 | 0 | 0 |
| Tonbridge and Malling | 21 | 0/3/5 | 9/4/0/8 | 1 | 1/10 | 1/20 | 2 | 7 |
| Uttlesford | 6 | 0/1/4 | 6/0/0/0 | 0 | 0/1 | 1/5 | 5 | 0 |
| Warwick | 9 | 2/1/0 | 3/1/0/0 | 0 | 1/3 | 0/7 | 1 | 0 |
| West Northamptonshire | 1 | 0/1/0 | 0/0/0/1 | 0 | 0/0 | 1/0 | 0 | 1 |
| Worcester City | 6 | 0/0/0 | 5/0/0/0 | 0 | 0/0 | 2/4 | 0 | 0 |
| Wychavon | 17 | 0/1/10 | 8/1/0/6 | 3 | 0/7 | 0/17 | 1 | 0 |
| Stratford-on-Avon | 59 | 7/14/14 | 17/8/1/14 | 1 | 6/20 | 23/33 | 4 | 32 |
The enlarged set: 144 more reports from the council sweep #
On 19 and 20 October 2026 a sweep of 48 councils' planning registers and committee pages fetched every officer report it could reach for decisions and committee items from 17 August 2026; 17 of them had a housing report in the period that was not already scored. Of the 1,095 reports listed, 144 were housing proposals not already in the hand-scored table (new homes, conversions and replacements; not householder works, prior approvals, section 73 variations, traveller sites or rural workers' dwellings), and all 144 were scored on 28 and 29 October 2026 with the rubric below. The sweep adds 5 councils (Babergh and Mid Suffolk, Hertsmere, Solihull, Swale and Uttlesford) and takes the comparison to 243 reports from 27 councils. 27 of the sweep reports are committee reports and 117 delegated; 20 are Green Belt sites.
| Error | Stratford (59) | Hand-scored set, 22 councils (99) | Council sweep, 17 councils (144) | All other councils, 27 (243) |
|---|---|---|---|---|
| Tier used as the location test | 7 of 35 (20%) | 5 of 71 (7%) | 4 of 51 (8%) | 9 of 122 (7%) |
| Tier used, counting "partial" | 21 of 35 (60%) | 23 of 71 (32%) | 9 of 51 (18%) | 32 of 122 (26%) |
| TR3 not applied where a sustainable-location decision is made | 21 of 35 (60%) | 41 of 71 (58%) | 37 of 51 (73%) | 78 of 122 (64%) |
| Location decided on the route, with facts, where a decision is made | 14 of 35 (40%) | 48 of 71 (68%) | 42 of 51 (82%) | 90 of 122 (74%) |
| TR3 applied on Green Belt sites, where GB7(1)(g)(iii) asks for it | 7 of 14 (50%) | 17 of 32 (53%) | 9 of 13 (69%) | 26 of 45 (58%) |
| TR3 not run, raw count of all reports | 26 of 40 (65%) | 50 of 81 (62%) | 113 of 134 (84%) | 163 of 215 (76%) |
| of which six homes or more | 6 of 12 (50%) | 22 of 41 (54%) | 23 of 30 (77%) | 45 of 71 (63%) |
| Connectivity Tool used | 1 of 59 | 8 of 99 | 3 of 144 | 11 of 243 |
| Green Belt route: S4 or S5 run where it does not apply | 6 of 26 (23%) | 7 of 50 (14%) | 2 of 61 (3%) | 9 of 111 (8%) |
| DP3(3) not asked after a found conflict | 23 of 56 (41%) | 8 of 96 (8%) | 27 of 137 (20%) | 35 of 233 (15%) |
| of which the officer recommended grant | 17 | 3 | 7 | 10 |
| 2024 Framework applied without conversion | 4 of 59 | 3 of 99 | 23 of 144 | 26 of 243 |
| Template paragraph cutting plan policies wholesale | 32 of 59 (54%) | 6 of 99 (6%) | 18 of 144 (13%) | 24 of 243 (10%) |
The enlarged set moves the headline figures a little and the conclusion not at all. On the four errors the complaint is about, Stratford's rates are 20% against 7% for the tier, 23% against 8% for the Green Belt route, 54% against 10% for the template paragraph, and 17 recommendations to grant after an unaddressed design conflict against 10 across the 27 councils together. The sweep's own DP3(3) rate (20%) is higher than the hand-scored set's (8%) because 20 of its 27 errors are refusals where the design or landscape conflict was itself a refusal reason, and because the sweep applies the broad landscape-character reading set out in the rubric; the 7 that recommended grant (Cotswold 26/02006/PLP, Cotswold 26/02330/PLP, East Devon 22/2781/MOUT, East Devon 25/1788/MOUT, Maidstone 25/504992/FULL, Swale 24/504121/HYBRID, Solihull PL/2025/01502/PPOL) are the ones that could have changed an outcome.
What the sweep adds to the picture:
- The tier error is a Cotswold habit too. 4 sweep reports rest a location finding on a village's tier or boundary with no route fact: Cotswold 26/02006/PLP, Cotswold 26/01577/FUL, Cotswold 26/02330/PLP, Maidstone 26/502459/FULL. Cotswold's Non-Principal Settlement label does the work at Upper Oddington ("The Local Plan examination process determined that Non-Principal Settlements constitute sustainable locations for small scale residential development even if they have limited facilities") and Maidstone's hierarchy does it in the refusal direction at Langley Heath. The other 15 councils' location findings give a distance, a footway, a bus frequency or a station, as the Framework asks.
- The Green Belt route is mostly stated correctly. 2 of 61 sweep reports where the question arises run S5 on a Green Belt site (Maidstone 26/502713/PIP, Tonbridge and Malling 26/00049/FL), both beside a full GB7 assessment and without changing the result. Hertsmere, Solihull, Three Rivers and Tonbridge and Malling state S5(5) in terms, as in "Policy S5 does not apply to the development as the site is located within the Green Belt". No sweep report applies S4 in a washed-over village.
- The template paragraph has cousins. 18 sweep reports (13%) cut a plan policy wholesale, in 5 councils (Babergh and Mid Suffolk, Cotswold, East Devon, Swale and Tonbridge and Malling): Babergh says that "reduced weight is afforded to those policies relating to the spatial strategy and housing distribution" for its supply position, Cotswold says its hierarchy policies "can currently only be given very limited weight", and Tonbridge and Malling gives "no weight is given to development plan Policy CP14, as it is materially inconsistent with Policy S5 of the NPPF". None reaches Stratford's 54%, and Solihull's cut of one policy, P7, for being "prescriptive on expected distances to local services and is therefore inconsistent with the Framework", is the reasoned Annex A(2) judgement the rubric does not count.
- Unconverted 2024 reports are commoner than the hand-scored set suggested. 23 sweep reports (16%) apply the 2024 Framework with no conversion, most of them committee resolutions from 2024 or 2025 issued as notices after 17 August 2026 once a legal agreement was signed: Babergh and Mid Suffolk DC/25/02757, Cheshire East 26/0581/FUL, Cotswold 25/02443/FUL, East Devon 24/1721/FUL, East Devon 23/1247/MOUT, East Devon 25/2614/FUL, East Devon 22/2781/MOUT, East Devon 25/1788/MOUT and 15 more. Stratford's 4 are the same defect.
- The Connectivity Tool stays rare. 3 sweep reports use it (Cheshire East, Hertsmere and Tonbridge and Malling), and Solihull sets it aside at Peartree Farm because "it appears to contain outdated public transport information". 232 of 243 reports from 27 councils do not use it; nor does any Stratford report.
The sweep table, with a page reference and a verbatim quotation for every adverse code, is other-councils-four-errors-sweep-2026-10.tsv; the inventory lists every report the sweep fetched with the reason each excluded one was left out, and the analysis note records how the set was read. The sweep reaches 17 of the 48 councils it swept with a housing report in the period; registers behind a login or a bot wall, and councils that publish no delegated reports, are listed in the note.
What the best reports elsewhere do #
The Green Belt route stated and followed #
Bromsgrove's delegated template quotes the S3 split and then says that "S5 does not apply to development proposals in the Green Belt and that these proposals should be determined in accordance with NDMP GB6 and GB7", runs GB7(1)(a) to (g) as a list, and in the Green Belt conclusion applies the S5(5) balance in terms. Dacorum's update report on Chesham Road, Berkhamsted, puts it in one line:
Policy S5 does not provide an alternative route for approval. S5(5) provides that S5 does not apply to development proposals in the Green Belt
Dacorum, 24/02868/MOA, officer report p.7
Elmbridge's report on the former Moore Place golf course at Esher sets up the same route, "policy S5(5) of the NPPF is of relevance", runs GB7(1)(g)(i) to (iv), and returns to S5(5) and S5(2) at the balance. Compare Stratford at Butts Lane, Tanworth-in-Arden, where the S5(1)(j) test for a site "well-related to an existing settlement" is run in a village washed over by the Green Belt before S5(5) is reached (Tanworth-in-Arden, 26/00918/PIP, p.9).
GB7(1)(g)(iii) answered with the route, in TR3's words #
At Beoley, 270 metres outside the village envelope, the Bromsgrove officer gives the footway side, the speed limit, the lighting and the distance to each facility, notes "there is a not a continuous footway to the identified facilities. There is no street lighting for much of the route", cites an appeal on the movement threshold, and concludes:
the site is not considered to offer a genuine choice of transport modes for future residents who are likely to be reliant
Bromsgrove, 26/00889/PIP, officer report p.4
At Dodford the same council says "Criteria iii requires that the development would be in a sustainable location having regard to policy TR3 of the 2026 Framework", finds a bus "one bus in each direction to/from Bromsgrove on a Tuesday and Friday only", and holds that "the site does not lie within a sustainable location", so the scheme is inappropriate and needs very special circumstances. Stratford's counterpart at Tithe Barn Lane, Earlswood, passes the same limb because the dwelling is inside the built-up area boundary of a Category 3 Local Service Village, with no route assessed (Earlswood (Hockley Heath), 26/01458/FUL, p.9).
The Connectivity Tool in an ordinary delegated report #
Wychavon's officers quote TR3(1)(a) in full and give the score. At New End, Astwood Bank: "Policy TR3 Paragraph 2 of the NPPF requires local planning authorities to use the GOV.UK Connectivity Tool", and "The application site has a Connectivity Score of 23". At Offenham: "The application site has an overall connectivity score of approximately 37", then the footway, the speed limit and a bus every two hours, with the Category 1 label as context and the conclusion that "connectivity could be better, on this occasion the proposal is considered to be acceptable". Cheshire East at Shavington: "the site is rated at Local Authority Band B". Bath and North East Somerset at the former Bath Press: "a Department for Transport connectivity score of 80-82, categorised as 'very high' nationally". 8 reports in 99 is not many, but it shows the Tool fits in a delegated report.
DP3(3) quoted as a refusal rule #
Wychavon's template says it by page number:
NPPF page 67 is clear that development proposals should be refused, if, without clear justification, they conflict with Paragraph 1 of Policy DP3
Wychavon, W/25/02384/FUL, officer report p.5
Elmbridge at Hersham finds that "the proposal is considered to have an adverse effect on the character of the area and so conflicts with national decision-making policy DP3", quotes DP3(3), and says that under S4 "it is necessary to consider whether the benefits of approving development are likely to be substantially outweighed by adverse effects": the trigger Stratford's reports leave out. Cheshire East at Bollington asks the S4(2)(c) question in terms: "There is no conflict with any of the national decision-making policies which state that development proposals should be refused in specific circumstances". Stratford's one report that names DP3(3), Snitterfield, uses it only as a source of weight:
In light of NDMP DP3 (3), I afford this harm substantial weight.
Snitterfield, 26/00617/PIP, officer report p.15
A Stratford report that does the same #
One Stratford report decided in late October does most of this. The delegated report refusing 13 homes in a listed farm building at Castle Farm, Studley (25/00989/FUL, 30 October 2026) opens with the transition in terms, "The presumption in paragraph 11(d) of the December 2024 Framework no longer exists. It has been replaced by policies S3 to S5. Policy S5 does not apply in the Green Belt", applies TR3(1)(a) to the applicant's own trip rates, "that is a significant amount of movement for the purposes of TR3(1)(a)", finds a character conflict and asks the DP3(3) question, "there is no clear justification for that conflict", and closes the balance with "Applying S5(2), the proposal conflicts with policy DP3(3), a national decision-making policy which states that development should be refused in these circumstances". It is the first Stratford report in the table to apply DP3(3) and the S5(2) trigger in terms, and it has no template weight paragraph: Annex A(2) is applied limb by limb to AS.10(d). It does not use the Connectivity Tool. The report is read step by step, in its own order, on its own page.
Annex A(2) applied policy by policy #
Dacorum at Polehanger declines to cut a saved policy because it "is not considered to be so materially inconsistent with the NPPF to the extent it should be afforded very limited weight". Worcester City: "No such material inconsistency has been identified in respect of the policies relied upon". Tonbridge and Malling at Town Hill, West Malling, ties its cut to a part of a policy: "Policy CP14 attracts reduced weight insofar as it seeks to apply a restrictive approach to development outside settlements". Stratford's template gives CS.15, CS.16 and AS.10 very limited weight in a paragraph that does not say which part of which policy conflicts; see which Core Strategy policies keep their weight.
The Annex B settlement test applied to a cluster #
Wychavon at Pinvin quotes the definition and concludes that the site "is not considered within a settlement as defined within the NPPF. Therefore, in this case the proposal needs to be assessed against Policy S5". Malvern Hills at Shoulton Lane, Hallow: "The fact that the site is associated with an existing residential cluster does not, in itself, mean that it forms part of the settlement". Every South Worcestershire officer facing a cluster with no boundary held it not to be a settlement and ran S5. Stratford at Ardens Grafton held a hamlet with no boundary to be a settlement so that S4 applied.
Where other councils repeat Stratford #
Warwick is Stratford's twin. At Norton Lindsey the committee report says "The site is washed over by Green Belt and is located within the Infill Village" boundary, then opens the principle section with "Policy S4 states that development proposals within settlements should be approved unless" the benefits are substantially outweighed. It never mentions S5(5) or the Annex B definition, and it passes GB7(1)(g)(iii) in one sentence:
The application site is located within the settlement boundary of Norton Lindsey, which includes a number of facilities for day to day living and therefore the proposal is considered to be in compliance with Policy TR3 of the Framework
Warwick, W/26/0066, officer report p.5
No footway, distance, bus or cycling fact appears anywhere in the report. Warwick's own Local Plan has a policy numbered TR3, on parking, so every Warwick report mentions "TR3" while the Framework policy goes unreasoned. Cubbington W/25/1088 was issued on 7 October 2026 on a report written to the 2024 Framework in April, with no re-anchoring.
North Hertfordshire at Pirton (50 homes, granted) is the tier shorthand at scale: the site adjoins a "Category A village" and "The development is considered a logical extension of the village", with no bus stop distance, frequency, footway or walking distance in the report, and the conversion update maps the 2024 paragraphs "Paragraphs 110, 115, 116" to TR6 alone, dropping TR3. At New Pound Farm, Hitchin, a Green Belt plot is supported because "the provision of a dwelling on a windfall site outside of a settlement would be supported by Policy S4 (j. i.) of the NPPF", which is S5(1)(j) under the wrong name on a site where neither applies.
Basildon at Pound Lane opens with "The principle of development within settlements is set out in policy S4 of the NPPF" for a washed-over Plotlands site, although GB7(1)(g) is then run in full.
Malvern Hills assesses routes candidly and then approves against its own finding. At Brookfield, Shoulton Lane, the officer records the Highway Authority's view that "the site occupies an inherently unsustainable location", finds that "The application site has no established public footpath directly serving it and is not provided with street lighting", and grants permission in principle on "a reasonable case that it constitutes previously developed land". The rubric scores that ok on the tier column because it tests method, not result; a column for "approved against the officer's own route finding" would catch it.
What the comparison does to the Stratford findings #
This section works from the 99 reports from 22 councils that were read in full, where every passage could be checked; the sweep section runs the same checks on the enlarged set.
Readings that hold #
- Annex B and washed-over villages. The government's response to the consultation on the Framework says the definition was amended "to address how the definition applies to villages within the Green Belt". No officer outside Stratford and Warwick applied S4 in a washed-over village. The sample is thin, because washed-over villages are rare outside the two districts (4 rows elsewhere), but the reading stands.
- S5(5) and the Green Belt route. Bromsgrove, Dacorum, Elmbridge, Three Rivers and Bath and North East Somerset reports all state it.
- DP3(3) as a "should be refused" policy. Wychavon, Elmbridge, Cheshire East and Tonbridge and Malling officers quote or paraphrase it as one; so do the inspectors on the DP3 page.
- Annex A(1), no transition. Every council in the sample converted or re-anchored its reports. The three unconverted notices elsewhere (Bromsgrove 26/00744/FUL, Warwick W/25/1088, Wychavon W/23/01358/FUL) are the same defect the Stratford table records four times.
- Annex A(2) cuts parts, not policies. The government's response says the wording was changed "to address concerns that entire policies could be given limited weight based on partial inconsistency". The template paragraph remains a real point, though the other reading, that S5(1)(j) displaces settlement-based restraint policies wholesale, has reputable support and the paragraph is defensible as a position.
Counts that were too hard #
TR3 "not mentioned" includes small schemes, so the count is re-based. Officers and inspectors often treat a single dwelling or a handful of homes as below the TR3(1)(a) threshold. Bromsgrove at Fairfield: "The development would not generate significant amounts of movement being one dwelling. The site is therefore considered sustainable". Inspectors say the same for small schemes and the opposite for larger ones:
Having examined the proposal, as a single dwelling, I do not consider that it would result in a significant amount of movement, given its location at the edge of the village of Burnham.
Burnham, appeal 6005162 ¶23, one self-build home
In this case, however, the development would not generate a significant amount of movement, given its small scale and the relatively short travel distances to a suitable range of services and facilities.
Washington, appeal 6009593 ¶19
An indicative forecast for vehicle trip generation has been prepared, with the proposed 28 dwellings expected to generate around 12 vehicle movements in the morning peak hour, nine in the evening peak hour, and around 112 movements per day overall.
Hatton Station, appeal 6006637 ¶28, TR3 applied and the appeal dismissed
The threshold is set "in the context of the area within which they would be situated", not by a number, and inspectors have found it met by very small schemes in villages and open country:
set against the size of Bournheath up to 5 dwellings would be a significant amount of movement
Bournheath, appeal 6011301 ¶10, a Green Belt village with three pubs
that increase would be noticeable in the context of this lightly developed rural location
Halsall, appeal 6007428 ¶14, one to three homes
must be read as a whole
South Nutfield, appeal 6009966 ¶27, up to five homes; TR3(1)(a) weakly engaged, the policy still breached
So "not mentioned" is a weak finding on its own, and the comparison above does not rest on it. It asks instead, of every report that has to decide whether the location is sustainable (every row scored on the tier column), whether TR3 is applied in terms, that is a limb of TR3(1) run with facts or GB7(1)(g)(iii) answered in its words, and whether the location is decided on the route: Stratford 14 of 35 and 14 of 35, the 27 other councils 44 of 122 and 90 of 122. Split by scheme size, the raw count looks like this (the count of homes is the largest number the application proposes; reports creating no new homes sit in the second row):
| Scheme size | Stratford | Other councils |
|---|---|---|
| Six homes or more | not run 6 of 12 (50%); assessed 6 | not run 22 of 41 (54%); assessed 19 |
| Five homes or fewer, or no new homes | not run 20 of 28 (71%); assessed 8 | not run 28 of 40 (70%); assessed 12 |
For five homes or fewer the two rates are close (71% against 70%), and the threshold point explains much of both. For six homes or more, where the threshold is not in doubt, Stratford's rate is 50% against 54%: Home Farm, Southam (217 homes), Alcester Road (65), Checketts Abattoir (7), the Pillerton Hersey showpeople site (7 plots) and Kineton (6) do not mention the policy, Harbury (38) names it without running it, and Stockton (75) names it and answers a different question. The three six-plus-home reports decided between 20 and 30 October 2026 (Castle Farm, Studley, 13 homes; Wixford Grange, 32; Park Farm, Aspley Heath, 6) and the one decided on 3 November (Forshaw Heath Lane, Earlswood, up to 6) all apply it, which is why the Stratford rate for larger schemes has fallen since the note of 13 October. The Stratford table's dodged code for Ravenstone, three homes, labels as an evasion a position that is mainstream for a scheme of that size outside the Green Belt: "In respect of NDMP TR3(a), the development would not generate a significant amount of movement" is the view inspectors often take of schemes that size. Ravenstone is a grey belt case under GB7(1)(g), though, and limb (iii) asks whether the site is in a sustainable location with particular reference to TR3 however small the scheme. Saying the scheme is too small for TR3 does not answer that question; the report still has to assess the route and the transport choice.
DP3(3) "not asked" on a refusal is formal. Where the design conflict is itself a refusal reason, the missing step could not change the outcome. Split by what the officer recommended:
| DP3(3) not asked after a found conflict | Stratford | Other councils |
|---|---|---|
| Officer recommended grant | 17 | 3 |
| Officer recommended refusal | 6 | 5 |
The finding that survives is the first row. Stratford officers recommended approval after a found design conflict, without asking whether it was clearly justified, in 17 reports; across 22 other councils the count is 3 (Malvern Hills at Leigh Sinton, Elmbridge at Moore Place and Dacorum at Polehanger). Part of that gap is composition: other officers more often find no conflict, or refuse. Two readings also need stating. The Stratford table treats a conflict with a landscape or character policy as a DP3 conflict; Elmbridge and Dacorum treat it as harm to weigh under N2, and either reading is arguable. And a weighing can itself be a clear-justification finding where the harm is slight; the cases where the omission could change the result are those where the harm was found significant and the scheme was still recommended, as at Snitterfield, Ladbroke and Pillerton Priors.
The Connectivity Tool is a shared absence. The other councils used it in 8 of 99 hand-read reports, all from 4 councils' templates. Used in 1 of the 59 is a fair fact about Stratford, but not a Stratford fault; the one use, at Oaks Holt, Kineton on 4 November 2026, gives a score of 40 "within the 10-20% percentile nationally" beside a bus stop opposite the site and paths into the centre, which is the Tool doing what TR3(2) asks of it. TR3(2) is one input to the location judgement, not a gate.
partial is a wide band. It covers Tredington, where the only transport fact is that occupants "would have access to a local bus service allowing residents to travel to nearby settlements", and Town Hill, West Malling, with "bus stops located immediately adjacent to the site, West Malling High Street approximately 700m away, and West Malling railway station around 1.4km distant". Most of Stratford's 14 are the first kind; most of the other councils' 18 are the second.
The tier code measures drafting, not the inspectors' rule. Inspectors have cited a tier inside an allowed location finding, always alongside a daily bus, lit footways or an urban setting. The reports coded error give no route fact at all, which is what the code should be read as: tier or facilities list as the whole location finding.
One finding is strengthened #
Approvals recommended after a found design conflict with no clear-justification question: 17 at Stratford, 3 across 22 councils. Snitterfield, Ladbroke, Pillerton Priors, Stockton, and from the late-October decisions Evenlode Close and Bayberry Meadows, have no counterpart elsewhere except Leigh Sinton and Moore Place.
The rubric #
Every report in the three tables is scored on the same columns with the same codes. This is the rubric in full; it is also published as four-errors-rubric.md.
Codes #
- Tier used as the location test.
error: the report's finding that the site is (or is not) a sustainable location rests on the village's tier in the settlement hierarchy, its development boundary, or a recital of its facilities, with no route fact.partial: a tier or boundary finding with at least one transport fact beside it (a bus stop distance, a footway remark, a frequency), however thin.ok: the location is decided on the route and the transport choice, with facts, whether or not a tier is mentioned and whatever the decision.na: no location finding is made or needed: householder works, listed building consent, replacement dwellings and reconstructions, conversions decided under a reuse category, employment, and sites inside a town or city. - TR3 not assessed.
not-mentioned: the text search for "TR3" finds nothing (a Local Plan policy that happens to be numbered TR3 does not count).cited-only: TR3 is named, listed or quoted but no limb of TR3(1) is applied to the site.dodged: TR3(1)(a) is said not to apply because the scheme is small, and nothing else is run.assessed: at least one limb of TR3(1) is applied to the site with facts, or GB7(1)(g)(iii) is answered in TR3's terms.na: as for the tier, except that a report inside a town is still scored because TR3 applies there. The split by scheme size (six homes or more, against five or fewer) is reported because officers and inspectors treat small schemes as below the TR3(1)(a) threshold. The headline figure is taken from the rows where a sustainable-location decision is made (every row scored on the tier column): the share in which TR3 isassessed, and the share in which the location is decided on the route (tierok), with Green Belt sites shown separately because GB7(1)(g)(iii) asks the question whatever the scale. - Connectivity Tool.
yeswhere a score or band from the Tool is given;indirectwhere the applicant's score is relied on;nootherwise, including where the Tool is mentioned and set aside. It is one input to TR3(2), not a gate, and it is reported as a shared absence, not as an error. - Annex B and the Green Belt route. The question arises for a Green Belt site, a hamlet or a cluster without a boundary, or a village whose status as a settlement the report has to decide.
error: S4 or S5(1)(j) is applied in a village washed over by the Green Belt or in a hamlet, S5 is run on a Green Belt site that S5(5) sends to GB6 to GB8, or a boundary or a tier is used as the GB7(1)(g)(iii) test; a belt-and-braces run of S5 beside a full GB7 assessment still scoreserror, and the note says so.ok: the Green Belt route is taken through GB6 and GB7 with S5(5) stated or implied, or a cluster is tested against the Annex B definition and S5 applied.na: a site inside a defined boundary that is not Green Belt, a replacement dwelling, or a 2024 report that runs no S4 or S5 route. A separate column records whether Annex B is cited. - DP3(3) not asked. The question arises for every proposal with design content.
error: a conflict with DP3(1), a DP3(2) principle, a development plan design standard, or the character or landscape of the area is found and weighed in the balance, and the report does not ask whether it is clearly justified or run the S4(2)(c) or S5(2) trigger.ok: no such conflict is found, or the conflict is found and the DP3(3) test or the trigger is applied in terms (a refusal that quotes the test counts).na: the only harm is heritage harm under HE6, the consent is listed building consent, or there are no external works. Two readings are stated and applied the same way across the tables: a conflict with a landscape-character policy counts as a DP3 conflict (the broad reading; the narrow reading would weigh it under N2 instead), and an error on a refusal where the conflict is itself a refusal reason is a formal slip, so errors are also split by what the officer recommended. The finding that can change an outcome is a recommendation to grant after a found conflict with no clear-justification question. - Framework.
2026;2024+notewhere a 2024 report carries a conversion note or update;mixedwhere 2026 policies are applied beside 2024 paragraph numbers or the 2024 "tilted balance" wording;2024-unconvertedwhere a decision made after 17 August 2026 rests on a 2024 report with no conversion, including notices issued after a legal agreement on an older resolution. - Template paragraph.
yeswhere the report cuts a settlement-hierarchy or countryside policy wholesale to limited, very limited or no weight on five-year-supply or age grounds, or as "materially inconsistent" without saying which part conflicts.nowhere Annex A(2) is quoted and applied to a named part of a policy with a reason, or no cut is made. The Stratford table's column is the template paragraph itself; the other tables record any wholesale cut.
Rules that apply to every row #
- Every
error,partial,cited-onlyanddodgedcode carries a page number and a verbatim quotation from the officer report, and every quotation in the three tables is checked by machine against the text extracted from the PDF before the page is built; a quotation that does not match stops the build. - Every
not-mentionedis a text search for "TR3" with no hits, checked the same way. - The rubric tests method, not result: a report that assesses the route and approves against its own finding scores
ok; a report that reaches the right answer by the wrong route scoreserror. - Rates are given among the reports where the question arises;
narows are excluded from the denominator. - Committee reports and delegated reports are scored alike. A committee report written before 17 August 2026 is scored on what it says, and its framework code records whether it was converted before the decision.
How the sweep set was read #
The hand-scored set was read report by report in full. The sweep set was first screened by machine for the mechanical facts (the TR3 text search, the Connectivity Tool, Annex B, the Framework markers and the weight phrases), then each report was read through its keyed passages: the site description, every paragraph that bears on location and transport, on the S4, S5 and Green Belt route, on design and character, on weight, and the conclusion, with page numbers; where those passages left a judgement open, the report was searched again for the point. The codes that need judgement (tier, Annex B and DP3(3)) were given on that reading, and the quotations were taken from the collapsed text and checked back against the pages.
Method and limits #
The reports were chosen as sets, not as examples, so that the comparison is not a selection of the best or worst reports elsewhere.
| Set | Reports | How chosen |
|---|---|---|
| Held, Midlands | 27 | Every officer report or reasoned decision notice already held for Wychavon, Malvern Hills, Worcester, Bromsgrove, Warwick, Lichfield, West Northamptonshire, Nuneaton and Bedworth and Cheshire East (decisions 17 August to 23 September 2026) |
| Held, south and north | 23 | Every committee report already held for Basildon, Bath and North East Somerset, Chorley, Cotswold, Dacorum, East Devon, Elmbridge, Maidstone, Maldon, North Hertfordshire, Sefton, Three Rivers and Tonbridge and Malling |
| Fetched: Warwick and Bromsgrove | 13 | Every housing-type decision on the two councils' planning portals issued 1 September to 17 October 2026 |
| Fetched: South Worcestershire | 15 | Housing-type decisions issued by Wychavon, Malvern Hills and Worcester City 20 September to 17 October 2026 |
| Fetched: committee agendas | 21 | Every housing or Green Belt item on the planning committee agendas of Basildon, Elmbridge, North Hertfordshire, Dacorum, Chorley, Cheshire East, Tonbridge and Malling and Bath and North East Somerset, 24 September to 17 October 2026 |
| Council sweep | 144 | Every housing report reached by the 19 and 20 October 2026 sweep of 48 councils' registers and committee pages for decisions and committee items from 17 August 2026 and not already in the sets above (17 councils); read through keyed passages, see the enlarged set |
99 hand-read reports from 22 councils: 56 committee and 43 delegated, 44 on Green Belt sites. Stratford: 59 reports, 18 committee and 41 delegated, 21 Green Belt. The closest analogues to Stratford are Warwick and Bromsgrove (the same Green Belt arc, villages washed over by the Green Belt, delegated reports) and the South Worcestershire councils (a Category 1 to 3 village hierarchy, no Green Belt).
- The rubric tests method, not result. A report that assesses the route and then approves against its own finding scores
ok. A report that reaches the right answer by the wrong route scoreserror. - Committee reports are larger schemes. Design is often reserved and location argued from a transport assessment, which lowers the other councils' TR3 and DP3(3) error rates. The fair comparators for Stratford's delegated habit are Warwick, Bromsgrove and South Worcestershire: Warwick repeats the pattern, Bromsgrove and Wychavon do not.
- Washed-over villages are rare in the sample outside Stratford and Warwick, so the Annex B comparison rests on few cases.
- The template-weight comparison is partly a supply artefact. The South Worcestershire councils claimed a five-year supply until October 2026 and give the hierarchy full weight; the councils that cut weight are short of supply.
- Outcomes are missing for the Elmbridge and Dacorum October committees, whose minutes were not published when the reports were scored, and three Warwick and Bromsgrove files had no notice or no report.
- Checking. Each set was scored by a separate pass with the shared rubric, and each pass checked its own quotations against the text extracted from the PDF. Every quotation on this page is checked again, with its page number, when the page is built, and every count on the page is computed from the two tables when it is built.