BLOGS

London needs roughly 52,000 new homes a year, according to MHCLG targets. It has not come close to hitting that figure in years. The reasons are well rehearsed — land scarcity, infrastructure constraints, community opposition — but in 2025, the bottleneck that developers flag most consistently is the planning system itself, and the viability gap that makes many consented schemes undeliverable even when permission is eventually granted.

REalyse planning data paints a stark picture of a pipeline under sustained pressure. Understanding where the blockages sit — and how they interact with cost inflation and affordability dynamics — is essential for investors, lenders and developers trying to underwrite London residential schemes today.

A shrinking approval pipeline

The headline trend in London's residential planning system is a significant and sustained fall in the number of schemes being granted consent.

REalyse data, tracking new residential planning applications across all 32 London boroughs and the City, shows granted applications falling from 1,106 in 2020 to just 662 in 2024 — a decline of approximately 40% over four years. Total residential units approved across granted schemes followed a similar path, dropping from around 50,200 in 2020 to approximately 41,100 in 2024.

Perhaps more revealing is what is stuck in the system. In 2025 alone, 312 applications are currently recorded as in progress, with over 101,000 residential units proposed across those schemes — an average of more than 350 units per application. These are disproportionately large, complex schemes: precisely the kind of development London needs most, and precisely the kind most exposed to protracted decision-making.

The national picture aligns. MHCLG data consistently shows that fewer than a third of major planning applications in England are decided within the statutory 13-week timeframe. For EIA (Environmental Impact Assessment) schemes — the category that captures most significant London residential developments — average decision times have stretched well beyond six months and frequently past a year. Every month a scheme sits undecided is a month of holding costs accumulating against a development appraisal that may already be marginal.

Refusal rates remain elevated

Approval volumes are down, but refusal rates have stayed stubbornly high.

Across the 2020–2024 period, the ratio of refused to granted applications in London has ranged from around 50% to over 60% in peak years. In 2022 and 2023, for every 878 and 730 schemes respectively that received consent, between 449 and 547 were refused outright — before accounting for schemes withdrawn by applicants who anticipated a negative outcome.

Withdrawals are themselves a meaningful signal. An application withdrawn is typically a developer who has read the room: who has had pre-application discussions, absorbed officer feedback, and concluded that the commercial and reputational cost of a refusal outweighs the sunk cost of pulling the scheme. In 2020, 213 applications were withdrawn across London. By 2022 that figure had risen to 145 before falling — but the consistent presence of a withdrawal cohort representing 10–15% of total submissions reflects ongoing uncertainty about what local authorities will and won't accept.

The viability trap

Even when consent is granted, delivery is not guaranteed. Developers across London have faced a compounding viability squeeze since 2021 that has made a material proportion of consented schemes economically unviable to start.

Build cost inflation is a primary driver. BCIS (Building Cost Information Service) data recorded construction tender price inflation running at double-digit annual rates through 2022 and into 2023. Labour shortages, energy costs, and supply chain disruption following the pandemic pushed typical London residential build costs from roughly £250–£300 per square foot in 2019 to £350–£450 per square foot or more by 2023, depending on specification and location.

The Building Safety Act 2022 added another layer of cost, particularly for residential buildings over 11 metres. Gateway compliance, cladding remediation obligations, and the higher-risk building regime have added meaningful upfront costs and programme risk to medium and high-rise schemes — precisely the typology London relies on for density.

Against this backdrop, GDV (gross development value) assumptions have been pressured from both sides. REalyse transaction data shows London new-build average prices around £761,000 in 2024, reflecting a market where buyers remain sensitive to price and mortgage affordability. New-build premiums are harder to sustain when purchaser financing costs are elevated and second-hand stock offers relative value. For developers trying to price schemes off land purchased at 2019–2021 values, the maths frequently does not work.

Affordable housing obligations compound the pressure

Affordable housing requirements — typically 35–50% on-site affordable in London, depending on the borough and scheme type — are central to the viability debate.

In principle, viability assessments can be used to negotiate affordable housing obligations downward where development economics genuinely cannot support the policy requirement. In practice, these negotiations are time-consuming, contested, and increasingly public following transparency requirements introduced by the GLA. Developers report that even where viability cases are clear-cut, the process of agreeing a reduced affordable housing package with borough planners and the GLA can add six to twelve months to the pre-commencement timeline.

The result is a paradox: boroughs that hold firm on affordable housing targets may see fewer schemes come forward at all. Boroughs that flex on viability grounds risk reputational and political pushback. Neither outcome reliably delivers the supply London needs.

What the pipeline data tells investors and lenders

For institutional investors and development lenders, the planning and viability environment has direct implications for underwriting.

REalyse data showing large schemes sitting in progress — with an average of 350+ proposed units per application currently awaiting decision — points to significant latent supply that may or may not materialise depending on whether consented projects can clear the viability bar. Lenders underwriting development finance against London residential schemes should stress-test GDV and build cost assumptions carefully; the gap between a consented scheme on paper and a deliverable scheme in practice has rarely been wider.

Investors looking at land or early-stage development opportunities should pay close attention to planning risk pricing. Schemes with extant consent in boroughs with faster decision tracks and lower refusal rates represent meaningfully lower-risk entry points than unconsented land in high-scrutiny boroughs, even where headline land values suggest otherwise. REalyse planning data can help identify which London boroughs are running lean pipelines and which have meaningful consented but unimplemented stock — a signal of either opportunity or systemic delivery failure, depending on the reason.

Outlook: reform on the horizon, delivery the test

The government's Planning and Infrastructure Bill, introduced in 2025, signals an intention to accelerate decision-making and expand planning capacity in local authorities. Mandatory housing targets, streamlined environmental assessment processes, and new infrastructure levy mechanisms are all in scope. Whether these reforms translate into faster, more predictable decisions for London's major residential schemes remains to be seen.

For now, the data tells a consistent story: fewer schemes coming forward, slower decisions, elevated refusal and withdrawal rates, and a viability environment that is testing even well-capitalised developers. London's new-build pipeline is not broken, but it is operating well below the capacity the capital needs. Until planning certainty and build cost economics align more favourably, private new-build supply in London will remain a constrained and carefully rationed resource.