Havering refuses 70.5% of small site applications; the highest rate of any London borough. No site type achieves better than 39% approval. Approved schemes average just 22 units per hectare, suggesting the borough is functionally hostile to small site intensification.
What makes Havering the hardest borough for small sites?
At what point does a borough’s refusal rate stop being a planning risk and start being a market signal? For developers working in outer London, Havering has a reputation. The data confirms it — and suggests the reputation may still understate the problem.
70.5% refusal rate with no safe harbour across site types
Across 383 decided small site applications between January 2023 and March 2026, Havering refused 270. That is a refusal rate of 70.5% — the highest of any London borough.
The London-wide average sits at 43.5%. Havering is 27 percentage points above it.
What makes this figure striking is not just the headline. It is the consistency across site types. In most boroughs, demolish-and-rebuild (DMR) schemes perform significantly better than conversions or extensions — London-wide, DMR achieves roughly a 60% approval rate. In Havering, DMR applications were refused at 60.8% (n=125). The type that succeeds most readily across the rest of London barely scrapes a 39% pass rate here.
Other types fare worse. Conversions (CNV) achieve only 26.5% approval from 155 applications. End-of-terrace schemes (END): 22.7% from 22 applications. Extensions (EXT): 23.8% from 21. Backland (BCK): 25.9% from 27. There is no safe harbour.
The refusal reasons follow a pattern that is recognisable but intensified. From 223 refusals with extractable reasons, design quality accounts for 55.6% — above the London average of 49.9% but not dramatically so. Space standards take 21.5%, and amenity impact 16.1%. The profile is not unusual in its composition. It is unusual in its severity. Havering applies broadly conventional grounds but reaches negative conclusions far more often.
Determination times add texture. Approved applications take a median of 86 days — 30 days longer than the 56-day median for refusals. In most boroughs, the gap between approval and refusal timelines is 5 to 15 days. A 30-day gap suggests that the schemes which do get through are subject to extended negotiation, multiple rounds of amendment, and considerable officer engagement. Approval in Havering is not just unlikely. It is slow.
Perhaps the most telling figure is density. The median approved DMR scheme in Havering achieves 22 dwellings per hectare. The London-wide median for DMR sits around 60 dwellings per hectare. Havering is permitting at roughly a third of the density that most London boroughs accept for the same site type. This is suburban housebuilding density — detached and semi-detached territory — applied to sites in a borough with Zone 4 transport links and London Plan density expectations.
Rethinking acquisition arithmetic in Havering
The acquisition arithmetic in Havering is punishing. Standard London small site underwriting assumes a 50–60% approval probability, factors in a 12-to-18-month planning timeline, and models density at 50–80 units per hectare depending on PTAL. None of those assumptions hold here.
Consider a site in Romford — a semi-detached property on a 400-square-metre plot, guide price £475,000. A developer might reasonably propose 4 units at a density of around 100 dwellings per hectare.
Havering’s data suggests the likely outcome is different. At a 29.5% approval rate and a median approved DMR density of 22 units per hectare, the scheme the borough is likely to accept is closer to 1 or 2 units — not 4. At 2 units with a local GDV of perhaps £350,000 each, the gross development value is £700,000. Against an acquisition cost of £475,000 and build costs that remain disproportionately high for a 2-unit scheme, the margin compresses to the point where the project may not justify the risk-adjusted capital.
The 30-day approval premium compounds this. If your planning timeline extends to 86 days median (and that is median — the upper quartile will be longer), finance costs on a £475,000 acquisition eat further into already thin margins.
This does not mean Havering is uninvestable. It means the viable strategy looks fundamentally different from inner London. The developers who succeed here appear to be those proposing low-density schemes — houses, not flats — that align with Havering’s suburban character from the outset. Trying to import a Lambeth-density scheme into a Havering streetscene is, on the evidence, a strategy with a 70% failure rate.
What the data cannot account for
Havering’s Aurora planning portal does not support automated document harvesting, so refusal reason data comes from a subset of applications (n=223) rather than the full refused population. The density figures rely on site area data, which is not available for all applications. The analysis also cannot account for pre-application engagement — it is possible that some developers avoid Havering entirely after informal advice, which would make the submitted population self-selected and the true underlying hostility even higher.
The full Havering Intelligence Report covers area-level variation, temporal trends, site type playbooks, and process intelligence across all 383 decided applications.