Cross-Collection Statistics: Approval Rates, Refusal Patterns, and Channel Differentials Across 14 Wards

Four Ward Groups, One Dataset

The South Manchester HMO Planning Intelligence series has documented planning activity across four geographic ward groups covering 14 wards. Articles 1 to 22 have examined individual wards, submission patterns, refusal reasoning, and appeal decisions within each ward group. This article brings the four ward groups together and sets out the cross-ward statistical record.

The analysis covers 101 applications validated between January 2024 and March 2026. Determined applications — those receiving a granted or refused decision — total 92. Withdrawn and restricted applications are excluded from approval-rate calculations throughout, consistent with the methodology applied across the series.

Ward Group Approval Rates

The four ward groups produced materially different approval rates. The table below sets out the headline figures for each ward group.

Rusholme, Moss Side, Hulme and Whalley Range recorded the highest approval rate of the three collections with sufficient volume for comparison. Burnage, Gorton & Abbey Hey, Levenshulme and Northenden recorded the lowest — 36.4%, compared with 69.0% across the dataset as a whole. The 40.7 percentage point gap between Rusholme, Moss Side, Hulme and Whalley Range and Burnage, Gorton & Abbey Hey, Levenshulme and Northenden is the widest ward group differential in the dataset.

Chorlton, Chorlton Park, Didsbury East and Didsbury West is excluded from percentage-rate comparisons throughout this article, as one determined application is not a sufficient basis for a rate. The single Chorlton, Chorlton Park, Didsbury East and Didsbury West determination — an approval of a conversion from HMO to self-contained apartments in Chorlton — is included in the combined totals.

Withington, Fallowfield and Old Moat and Rusholme, Moss Side, Hulme and Whalley Range: Similar Rates, Different Structures

Collections 1 and 2 produced approval rates within 6 percentage points of each other — 71.1% and 77.1% respectively. The surface similarity masks a structural difference in how those rates were produced.

Withington, Fallowfield and Old Moat was agent-dominated: 59.6% of submissions were agent-submitted. On a Certificate-only basis, agent-submitted applications in Withington, Fallowfield and Old Moat achieved a 90.5% approval rate, compared with 66.7% for self-submitted Certificates. The agent channel outperformed the self-submission channel by 23.8 percentage points.

Rusholme, Moss Side, Hulme and Whalley Range inverted this pattern. Self-submitted applications accounted for 83.8% of submissions. On a Certificate-only basis, self-submitted Certificates achieved a 92.0% approval rate, compared with 40.0% for agent-submitted Certificates — a differential of 52.0 percentage points running in the opposite direction. As documented in Articles 16 and 17, the agent-submitted underperformance in Rusholme, Moss Side, Hulme and Whalley Range was concentrated: two of the five agent-submitted Certificate refusals were the simultaneous Pentapura batch submissions at 67 Viscount Street and 338 Great Western Street.

The two collections produced similar headline rates through structurally different channel compositions. The dataset records this as an observable feature of the two collections. It does not record why submission channel behaviour differed between them.

Burnage, Gorton & Abbey Hey, Levenshulme and Northenden: A Structurally Different Outcome Profile

Burnage, Gorton & Abbey Hey, Levenshulme and Northenden is the most distinct collection in the dataset by approval rate. At 36.4% across 11 determined applications, it sits 32.7 percentage points below the combined rate and is the only collection where refusals outnumber approvals — 7 refused against 4 approved.

The channel split in Burnage, Gorton & Abbey Hey, Levenshulme and Northenden is the most extreme in the dataset. All four approvals were self-submitted Certificates. No agent-submitted application in Burnage, Gorton & Abbey Hey, Levenshulme and Northenden was approved within the study period. Agent-submitted applications in Burnage, Gorton & Abbey Hey, Levenshulme and Northenden recorded a 0.0% approval rate across 7 determined submissions. Self-submitted applications recorded 100.0% across 4 determined submissions.

This 0.0% / 100.0% split holds on both a Certificate-only basis and an all-applications basis. Burnage, Gorton & Abbey Hey, Levenshulme and Northenden contained no agent-submitted Certificates that were approved, and no self-submitted applications of any type that were refused. The figure is not a product of methodological framing — it is a feature of the Burnage, Gorton & Abbey Hey, Levenshulme and Northenden record within the study period.

The Burnage, Gorton & Abbey Hey, Levenshulme and Northenden refusal pattern was examined in Articles 19 to 21. The predominant refusal ground across the collection was failure to demonstrate continuous 10-year HMO use — the same evidentiary basis as Certificate refusals in Collections 1 and 2. What differs in Burnage, Gorton & Abbey Hey, Levenshulme and Northenden is the rate at which agent-submitted applications encountered that ground, and the absence of any agent-submitted approval to offset it.

Submission Channel Comparison — Certificate-Only Basis

The following figures set out the Certificate-only channel comparison across the three collections with sufficient volume for analysis. Chorlton, Chorlton Park, Didsbury East and Didsbury West is excluded (one determined application, self-submitted).

The direction of the channel differential reverses between Withington, Fallowfield and Old Moat and Collections 2 and 4. In Withington, Fallowfield and Old Moat, agent-submitted Certificates outperformed self-submitted Certificates. In Collections 2 and 4, self-submitted Certificates outperformed agent-submitted Certificates — by increasing margins. The dataset does not record why the direction of this differential varies between collections, or whether the reversal reflects differences in the evidentiary quality of submissions, the characteristics of the applicant pool, or other factors not captured in the public planning record.

The agent-submitted totals in Collections 2 and 4 are small — 5 and 7 determined applications respectively. The differentials should be read in that context. The Rusholme, Moss Side, Hulme and Whalley Range agent-submitted figure is further shaped by the Pentapura simultaneous batch refusals, which account for two of the five agent-submitted determinations. The Burnage, Gorton & Abbey Hey, Levenshulme and Northenden agent-submitted figure includes the two McLoughlin Planning refusals documented in Articles 20 and 21. The cross-collection agent performance record, including named agent outcomes, is examined in Article 24.

Refusal Patterns Across Collections

The dominant refusal ground across the dataset is evidentiary: failure to demonstrate continuous HMO use for the required 10-year period. This ground applies to Certificate of Lawful Development applications and accounts for the majority of refusals in Collections 1, 2, and 4.

Policy-based refusals — citing H11 density, amenity harm, or loss of family housing — are concentrated in Full Application refusals. Full Applications account for a minority of determinations across the dataset but a disproportionate share of policy-ground refusals. The three Planning Inspectorate appeals documented in Article 22 all originated as Full Application refusals on policy grounds.

Burnage, Gorton & Abbey Hey, Levenshulme and Northenden produced the highest proportion of refusals in the dataset — 7 from 11 determined applications. All seven refusals cited evidentiary insufficiency. No Burnage, Gorton & Abbey Hey, Levenshulme and Northenden refusal cited a policy ground. The collection’s low approval rate is therefore a product of evidentiary failure rather than policy-based restriction — a distinction the dataset records but does not explain.

Combined Dataset: Summary Figures

The 69.0% combined approval rate masks the ward group variation set out above. A single combined figure does not capture the structural differences between a collection producing a 77.1% rate and one producing 36.4%. The ward group breakdown is the more informative unit of analysis for understanding outcome distribution across the dataset.

Conclusion

The four-collection dataset records 100 applications, 84 determinations, 58 approvals, and 26 refusals across 14 South Manchester wards between January 2024 and March 2026. The combined approval rate is 69.0%. Ward group rates range from 36.4% to 77.1%, with the ward group differential being the most structurally significant feature of the cross-ward record.

The submission channel differential runs in opposite directions in Withington, Fallowfield and Old Moat versus Collections 2 and 4 — agent-submitted Certificates outperforming in Withington, Fallowfield and Old Moat, self-submitted Certificates outperforming substantially in Collections 2 and 4. The dataset records this reversal without explaining it.

Burnage, Gorton & Abbey Hey, Levenshulme and Northenden’s approval rate of 36.4% — the lowest in the dataset — is driven entirely by evidentiary refusals of agent-submitted Certificates. Its 0.0% agent approval rate and 100.0% self-submission approval rate represent the most extreme channel differential recorded across any collection in the dataset.

The cross-collection agent performance record, including individual named agent outcomes across all four ward groups, is examined in Article 24.