Harrow's 2026 Local Plan: the small-site exemption survives

Harrow Council adopted the Harrow Local Plan 2021–2041 on 24 March 2026, retiring the Core Strategy that had been in force since February 2012 and the Development Management DPD that followed it in 2013. Most of the coverage went to the large-scheme details: a 35% fast-track affordable housing minimum, a 50% mandatory rate on a defined list of site categories, and a new 70/30 tenure split between low-cost rented and intermediate product. For anyone whose pipeline consists of schemes of one to nine units, however, the change that matters most is the one that did not make it into the discussion at all.

The exemption that survived

The 10-unit threshold has been left exactly where it was. Schemes of one to nine units fall outside Policy HO4 entirely, and the only policy that catches them is HO3, “Optimising the use of small housing sites”, which sets design and PTAL criteria but levies no affordable housing contribution. There is no in-lieu payment, no commuted sum, and no sub-threshold band in which, for instance, a seven-unit scheme would owe a proportionate share of what a twelve-unit scheme owes. The cliff between exemption and contribution sits at exactly the unit count it has always sat at.

This kind of clean threshold is unusual in outer London. Several of Harrow’s neighbouring boroughs have for years required some form of contribution on schemes between five and nine units, whether as a commuted sum that scales with floorspace or as a graduated percentage that begins to bite well before the headline threshold is crossed. Harrow’s binary cliff, with no affordable housing required up to nine units and the full HO4 target landing at ten, has remained in place. Developers who design schemes to stay within the small-site exemption are therefore unaffected by the new plan, and the principal commercial reason to be working in Harrow rather than in one of its neighbours is the same reason it was six weeks ago. That fact did not appear in the headlines because, in the strictest sense, it is not a change. It is the absence of one, and an absence is harder to write a press release about than a percentage that has moved.

What changed, for 10+ unit schemes

For schemes that do cross the threshold, two things have moved, and both of them matter.

The first is the on-site target itself, which has stepped down from 40% by habitable room under the old CS1.J to 35% as the new fast-track minimum under HO4 part B. Above that, 50% is set out in the plan as the Boroughwide aspirational target measured on habitable rooms or floorspace, and 50% becomes mandatory on public-sector land and on Strategic, Locally Significant, and Non-Designated Industrial Sites. Most residential small-site portfolios will not run into any of those categories in practice, so for most developers the figure that will actually shape the negotiation is the 35% fast-track minimum, not the 50% aspirational target.

The second, and the more consequential of the two for the underlying economics of any 10+ unit scheme, is the tenure split. Under HO4 part F, Harrow now requires 70% low-cost rented (social rent or its London Plan H6 equivalents) and 30% intermediate (London Living Rent or shared ownership) on the affordable component of any qualifying scheme. The old CS1.J target ran the other way, 60/40 in favour of intermediate. Paragraph 4.4.8 of the supporting text is blunt about why the shift happened: the borough has decided it needs more social-rent supply and less intermediate product, and KPI HO-I5 commits that intention to a number, with 70% of new units expected to be low-cost.

For a 10+ unit scheme delivering affordable housing on-site, the practical consequence of this inversion is that the affordable component will be a markedly more rented-heavy product than it was under CS1.J, and a more rented-heavy product produces a different commercial conversation at every link of the chain: a different appetite from the bridging lender that prices intermediate product more comfortably than rented, a different choice of registered provider partner, and a different GDV blend at the back end of the appraisal. Buying two adjacent five-unit plots and combining them into a single ten-unit scheme is therefore no longer the same decision it was six weeks ago, even though the headline percentage shifted by only five points.

The pattern in the data

Across the 2023–2026 analysis window, Harrow determined 264 small-site (1–9 unit) applications under the CS1.J framework, of which 53.0% were approved. The median proposed scheme size was two units, and the largest scheme to reach a determination was nine. The dataset is therefore overwhelmingly composed of the kind of micro-developments that the new plan continues to leave alone.

What the refusal data shows, more clearly than any other single pattern, is that the borough’s planning committees concentrate their objections on design quality. Of the 124 refusals for which we have extracted reasons, 85 cite design as the primary ground for refusal. That works out to 68.5% of refusals, comfortably above the 49.9% London-wide average. Policy non-compliance accounts for a further 9%, space standards for another 6%, and heritage, transport, and amenity grounds are minor by comparison.

The point is not that Harrow rejects small-site development in principle, since its approval rate sits close to the outer-London median, but rather that it rejects specific design proposals whose form, scale, and detailing do not read against the suburban character that the borough has, fairly clearly, set out to protect.

The dominant site type by volume is conversion, which accounts for 142 of the 264 decisions and achieves a 54% approval rate that is broadly in line with the borough-wide figure. Demolition-and-replacement sits at 50% across 46 decisions. The two outliers in opposite directions are extensions, which the borough approves at 88% across 25 decisions, and end-of-terrace schemes, which it refuses to a comparable degree, with only 33% of the 30 decisions in that category producing an approval. The reading for an acquiring developer is reasonably clear: Harrow is comfortable approving modest extensions, but it routinely refuses substantive end-of-terrace work whose form or scale does not sit easily against the local pattern.

Conservation status is, perhaps surprisingly, a smaller factor in the refusal pattern than the dominance of design grounds might at first suggest. Only 11 of the 264 decisions in our window (a little over 4%) sit either inside or adjacent to one of Harrow’s 30 designated conservation areas, and the small IN cohort of eight produced an 88% approval rate, running counter to the broader West London pattern in which conservation status tends to suppress approvals. The design refusals in Harrow are not, in other words, happening predominantly in the borough’s heritage-sensitive streets. They are happening on standard suburban streets, where the contested point is not heritage character but the more diffuse question of suburban character.

Why this matters now

HO4 is now in force for any application determined from 24 March onwards, while applications that were validated before that date continue to be assessed under the old framework. The first wave of small-site decisions to be determined under the new plan is therefore only just beginning to land in committee, and the practical effect for developers working on schemes of one to nine units is, for the moment, none at all. The cliff is in the same place, HO3 carries no contribution mechanism that was not already there, and the established CS1.J pattern of design-led refusals is very likely to carry over into the new regime, since the officer cohort writing the recommendations has not changed, the suburban character the borough is protecting has not changed, and the policy hooks now available to officers under HO3 and DM1 are recognisably the same as the ones they have been working with.

For anyone considering combining adjacent sites into a single scheme that crosses the 10-unit threshold, however, the new tenure mathematics is already live. The 70/30 split should be sitting inside the GDV model before the second site is optioned, not after it.

See the live Harrow planning dashboard