The Missing Machinery: Distilled

The Emergency

On 15 May 2024, the Scottish Parliament declared a national housing emergency. The vote passed 95-29. Mark Griffin moved the motion. When it passed, approximately 10,000 children were in temporary accommodation.¹ As at 30 September 2025, the figure was 10,480.² The declaration did not resolve the emergency.

The Scottish Government’s own target is 110,000 affordable homes by 2032, at least 70% of them for social rent. The target runs from 2022. By December 2024, 24,400 had been delivered, with annual completions well below the 11,000 required to close the gap. The constraint is not ambition. RSL borrowing capacity is at its structural limit, land costs absorb public subsidy before it reaches construction,³ and no mechanism compels build-out on consented sites. Existing Section 75 agreements tie affordable delivery to percentage thresholds against market completions, not calendar dates; the national target has no transmission into individual consents. The proposals below supply both the deadlines and the holding costs that make indefinite delay irrational.

None of the manifestos that contested the 7 May 2026 election contained the structural machinery to change this. This paper proposes it: four instruments, two legislative vehicles, two executive actions. Two of the four do not require primary legislation and can be committed before a Bill is introduced. The year-one deliverable is permanent tenancies for households currently in temporary accommodation: identify long-term empty properties through the vacancy register, acquire from willing sellers, refurbish to lettable standard, let as permanent tenancies. No construction lag.

One instrument did pass: the Housing (Scotland) Act 2025 introduced rent controls, capped at CPI+1% in designated areas. That protects affordability for sitting tenants in the private rented sector. The 18,092 households in temporary accommodation are not in the private rented sector; they are on waiting lists, in B&Bs, in emergency hostels. Rent controls without supply instruments redistribute scarcity; they do not end it. The four proposals below are the supply side. The two halves belong in the same programme.

Two Structural Causes

The 18,092 households and 10,480 children in temporary accommodation are not a misfortune that befell Scotland. The 24,400 homes delivered against a 110,000 target are not a timing problem. Both are the present output of two structural arrangements.

The first is a land market that converts a public planning decision into a private speculative windfall with no holding cost. When a Local Development Plan allocates land for housing, its value typically multiplies overnight. Nothing requires the landowner to build. The Scottish Land Commission (2020) found that on sites of 250 homes or more, only 12% of permissions granted over three years had reached completion, compared with 30% for smaller sites. On those larger sites, the build-out phase is “considerably longer,” driven by absorption rate constraints. That is not a construction timeline; it is a holding strategy. The rational behaviour, from every speculative holder’s perspective, is to wait while values rise.

What makes this a public emergency rather than a private choice is that most major residential sites already carry a legal obligation – accepted voluntarily as a condition of consent – to deliver at least 25% affordable housing under Section 75 agreements (the planning obligations that run with the land as a condition of permission). Glasgow’s Housing Land Audit (2025) records large housebuilders averaging 25 homes per site per year while build-to-rent achieves 220 on the same sites. The private drip-feed is the mechanism by which the affordable obligation gets deferred. Land banking that defers or eliminates the publicly mandated affordable component is where the emergency is made. The housing shortage is partly a planning enforcement failure: the social homes that planning law already requires are not being delivered. Volume builders have standing counsel. Local authorities have overworked planning officers. The instruments to enforce delivery at scale do not exist.

The second is a Registered Social Landlord (RSL) finance route at the edge of what the sector can carry. The Scottish Housing Regulator’s (SHR) 2024/25 review records the sector paying £247.3 million in interest charges on £7.18 billion in total loan facilities, consuming 13.9% of rent income. Of that total, 21% is capital market funding – approximately £1.5 billion. The remaining 79%, approximately £5.6 billion, is traditional bank lending priced at commercial rates, with no institutional guarantee behind it. That £5.6 billion in bank lending is secured against the value of the housing RSLs own. If land reform compresses residential property values – which is its purpose – the collateral behind those loans deteriorates everywhere at once. Solving one problem without solving the other creates a new crisis in the sector the first reform was supposed to help.

Four paired instruments address both causes.

Package One: Breaking the Land Market

Proposals 1 and 2 are a pair. Neither works without the other.

Proposal 1: Designated Residential Land Tax

Scotland’s housing emergency framing rests on 164,000 homes with planning consent that remain unbuilt.⁴ The incoming Housing Minister’s first action should be an evidence step: consolidate and publish site-level data on the 164,000 sites within 60 days, drawing on data already held across Scottish Government Housing Statistics, Homes for Scotland member returns, and planning authority records. Five fields per site – local authority, consent date, infrastructure status, developer status, years elapsed without commencement – turn an undifferentiated aggregate into a deployable enforcement target. DRLT applied uniformly to all 164,000 hits the wrong sites; the speculative holder absorbs a charge calibrated against a fraction of a portfolio while the developer blocked by a Scottish Water connection queue is penalised for an infrastructure failure that is not theirs.

The financial logic of that holding is straightforward. Scottish agricultural land averages around £16,500 per hectare.⁵ In comparable Northern English local authorities, residential land ranges from £750,000 to £1.9 million per hectare – a ten to over a hundred-fold uplift on the same agricultural base.⁶ No Scottish residential land value survey exists; the cross-border comparison is the best available.

Carrying costs approach zero: no Council Tax (no dwelling), no Non-Domestic Rates (bare land without buildings), agricultural exemption persisting on an active-use test until the owner chooses to act. Empty commercial property eventually faces NDR; bare allocated residential land has no equivalent. On large Scottish sites, the build-out phase stretches years beyond the national average of 2.7 to 3.2 years.⁷ Years of compounding on a base with no annual drain. The rational choice is to hold.

An annual charge on LDP-allocated residential land that carries no residence, from the date of allocation, not from the date of planning application. Application timing is the owner’s decision; a DRLT that attaches at application can be deferred indefinitely. The windfall is created by the allocation. The rate starts nominal while infrastructure is being delivered, escalates once the infrastructure window closes and the excuses are gone, and must satisfy proportionality under Article 1 of Protocol 1 to the ECHR (A1P1 – the right to peaceful enjoyment of property) on every exercise: the instrument changes behaviour; it does not confiscate the asset.

DRLT is non-rebatable. There is no deferral during construction and no abatement on completion. Each completed hectare exits the taxable base; the charge falls on undeveloped land for as long as it remains undeveloped. At a 6% illustrative ceiling on a per-hectare value of £1.5 million, a speculative holder faces £90,000 a year for the privilege of not building. A builder who completes faster pays less. The instrument responds to output, not intent. The specific rate schedule and step sizes are matters for the Scottish Fiscal Commission, Law Officers, and Bill team.

Ireland’s Residential Zoned Land Tax, live from February 2025, provides the empirical anchor. RZLT was enacted in 2021 and deferred twice. In 2024 (three years after passage and still before the first charge date) residentially zoned land sales rose almost 30% and the median price per acre fell by 6%.⁸ The announcement moved the market; the charge had not yet cost anyone a euro. By September 2025, over 1,800 returns had been filed with approximately €40 million in liabilities paid.

An infrastructure viability gateway runs in parallel: DRLT suspended where a local authority confirms a site is absent from a named statutory capital programme of Scottish Water, SSEN (electricity grid), Transport Scotland, or SEPA (flood risk and drainage). Where a local authority fails to determine a gateway application within 60 days, the application is treated as refused and the DRLT runs. Passive non-determination cannot shelter a site. Once the infrastructure window closes, the escalator runs on schedule. A statutory review by the Scottish Fiscal Commission before the year-four rate step gives Parliament a structured intervention point.

A rural and island suspension applies where no planning application has been made within five years: a behavioural instrument is disproportionate where the holder has no realistic path to the behaviour the charge is designed to produce. If an allocation repeatedly triggers suspension, the allocation was wrong; the remedy is de-allocation in the next LDP review, not a rolling suspension.

The success metric is homes built, not revenue raised. The tobacco-duty parallel applies: the instrument works when the yield falls. The revenue profile is rise-then-fall, not decline from day one: the DRLT motivates accelerated disposals in early years, and the Scottish Housing First Foundation (Proposal 3) draws its seed capital from that peak. Declining receipts later confirm land is entering development.

Primary legislation required.

Proposal 2: Statutory Pre-Emption at EUV Plus

Pressure to dispose is meaningless without a companion instrument capping the return on disposal. Proposal 2 is that companion.

A statutory right for More Homes Scotland and local authorities to purchase LDP-allocated land at EUV plus (existing use value plus a statutory premium; the rate set by order within a ceiling in the Act, subject to affirmative procedure and three-year SFC review) on any voluntary disposal. The seller may decline; if they do, the DRLT continues to run and the pre-emption right remains attached to any future disposal.

Within 24 months of the Acts taking effect, every holder must either lodge a binding build-out commitment preserving market-value rights on any future sale, or accept that prospective EUV-plus pre-emption attaches to any future disposal. Infrastructure-blocked holders who commit are not penalised for delays caused by Scottish Water, SSEN, or other statutory bodies; their schedule is conditioned on verified delivery from those bodies.

The S75 obligation provides the A1P1 foundation. The developer accepted the affordable housing commitment voluntarily at consent. Enforcing that commitment is specific performance of a bargain, not a new imposition, and it is the stronger Article 1 Protocol 1 position, not the weaker one. James v United Kingdom (1986) and Lithgow v United Kingdom (1986) confirmed that statutory transfer at below-market value is compatible with A1P1 where public interest is sufficient and procedural protections exist. The 2012 Inner House judgment in Pairc Crofters v Scottish Ministers requires a published proportionality decision open to appeal for each exercise of the right; the legislation must discharge that requirement on every use.

The pre-emption provisions reference a national S75 register as the evidence base for non-delivery. The register aggregates obligation data already held by 32 planning authorities and Registers of Scotland into a single monitoring dataset. It does not require primary legislation: ministerial direction and a planning circular establish the reporting format; More Homes Scotland hosts the aggregation. The Bill defines what constitutes actionable non-delivery and the register provides the trigger data.

For new consents, the S75 agreement specifies an absolute completion date for the affordable component, derived from the AHSP programme requirement. Miss the date, pre-emption activates. This is not a retrospective imposition on existing rights: it is a condition accepted at consent, priced into the land purchase as Circular 4/2025 already requires. Existing consents without calendar dates are addressed by a deemed delivery deadline: the end of the LDP plan period under which the allocation was made (Annex 1). Where the plan period has already expired, the crystallisation trigger is available from the end of the declaration window. DRLT’s escalating holding cost operates in parallel, making indefinite delay irrational regardless of the crystallisation route.

The operational mechanism is whole-site acquisition at EUV plus, not pre-emption of the S75 portion alone. If pre-emption applied only to the affordable component, a developer could let the obligations lapse while retaining hope-value uplift on the private portion. Partial pre-emption is partial enforcement; the incentive to default survives. Whole-site acquisition at EUV plus removes it.

DRLT prices the holding cost. Pre-emption caps the disposal return. Together they remove the financial logic of speculative holding without removing owner agency. A holder who declined the declaration window holds an asset worth its productive value and pays an escalating annual tax on it. Either instrument alone leaves a gap the other fills.

Primary legislation required.

Package Two: Building the Delivery Machine

Proposals 3 and 4 are a pair. Proposal 4 must be committed before Proposals 1 and 2 commence.

Proposal 3: Scottish Housing First Foundation

Scotland has no national institution to acquire housing at scale, house people immediately, and pass stock to RSLs as they build the capacity to absorb it. It has 138 Registered Social Landlords operating locally, with local registers and no statutory national remit. That gap explains a specific part of the emergency: 18,092 households distributed across the country, concentrated in urban areas, requiring permanent tenancies that no existing organisation has the mandate or scale to deliver at pace. The 110,000 affordable homes target requires 11,000 completions a year. No institution in Scotland is configured to deliver at that rate.

Finland reduced its homeless population by roughly 80% between 1985 and 2023, from approximately 20,000 to 3,429.⁹ The trend has since reversed: 3,806 in 2024, the first increase since 2012, driven by welfare cuts. The Y-Foundation’s approach was Housing First: permanent housing as the starting point, support services alongside rather than as prerequisites. The Foundation converted emergency shelter and temporary accommodation into permanent tenancies first, with a supply pipeline built behind that. It now holds over 19,000 apartments across 58 municipalities,¹⁰ financed through state-guaranteed loans at preferential rates, recycling surpluses into new development. Finland was housing people before building for them. Scotland should do the same.

The Scottish Housing First Foundation (SHFF) should be constituted as a co-founding coalition including Scottish Government, COSLA (the local authority convention), SFHA (the RSL sector body), and civil society organisations with operational Housing First delivery capacity. Year-one task: acquire long-term empty residential properties, refurbish to lettable standard, and let as permanent tenancies to households currently in temporary accommodation. From day one, SHFF draws on existing local authority council-tax empty-property data to begin voluntary acquisition negotiations; the national register built on cross-agency occupation signals is assembled behind that, not ahead of it. Over 44,000 privately owned homes have sat empty for more than six months;¹¹ SHFF lists them by estimated value, identifies owners through Registers of Scotland, and makes voluntary acquisition offers at condition-reflective prices.

As households leave B&Bs and emergency placements, freed temporary stock with conversion potential can be acquired and enter the renovation pipeline. Faster and cheaper than new build; the same sequencing logic underpins the Finnish model. Capital recycles through the portfolio: acquisition cost recovered from rental income, surplus reinvested into the next tranche of properties, institutional capacity built as the stock grows. Y-Foundation financed its 19,000 apartments on this basis. From year two: act as RSL of last resort for the statutory affordable housing floor in Proposal 5, and begin pipeline development through the finance structure in Proposal 4.

The long-term model transfers acquired stock to RSL partners as their balance-sheet capacity develops; SHFF holds and manages in the interim. The anti-alienation clause prevents privatisation, not transfer to registered social landlords within the non-profit framework.

Capital comes from three sources: DRLT seed receipts hypothecated through year five, direct state lending from year three, and guarantee-backed bond issuance under Proposal 4 from year two. Capital and revenue mechanics are a design decision for the founding coalition. The founding articles carry the anti-alienation clause from day one; the Housing Emergency Bill adds the statutory backstop. A two-tier allocation architecture must prevent COSLA holding a structural veto over SHFF’s operating mandate.

Finland’s 80% reduction between 1985 and 2023 is the precedent of what the architecture can achieve over decades, though the trend reversed in 2024 following welfare and income-support cuts. Scotland does not start from Helsinki’s 70% municipal land ownership; the realistic first-decade working assumption is 30 to 50% reduction in temporary accommodation placements. The mechanism is the same. The starting position is not.

No fresh primary legislation is required to establish SHFF. Section 2(2)(o) of the Housing (Scotland) Act 1988 empowers Scottish Ministers to form companies under the Companies Act 2006 for their general housing functions.

Proposal 4: Scottish Housing Finance Guarantee

The Guarantee is not an optional improvement to the cost of capital. It is the structural precondition for Proposals 1 and 2 being safely implementable.

The land-market reform at Proposals 1 and 2 is designed to compress the speculative residential land premium and, through that, moderate new-build and existing-stock price appreciation over time. The bank-lending tranche of the RSL sector, £5.6 billion of the £7.18 billion total, is secured against property values the reform is designed to moderate. If the reform succeeds, that collateral position deteriorates across the sector at once. The Guarantee substitutes sovereign credit support for asset-value collateral. Without it, the land-market reform creates conditions for a sector-wide credit stress event.

The Guarantee-first sequencing instruction is therefore structural necessity, not operational preference. “Committed” has a measurable meaning: initial issuances executed and market pricing confirmed within a defined spread of sovereign-backed comparators, demonstrating that RSLs are borrowing through the Guarantee at reduced rates, not that a company has been registered.

A Scottish Housing Finance Guarantee modelled on the Dutch Waarborgfonds Sociale Woningbouw (WSW, which guarantees €94.9 billion in housing association loans across 264 member associations; the government backstop has never been reached, including through the fund’s first-ever guarantee claims in 2018, thirty-five years after establishment), operated by a Scottish Housing Guarantee Fund (SHGF) constituted as an RSL-owned mutual body classified as private sector for ONS purposes and institutionally separate from SHFF. SHFF is the SHGF’s largest borrower; the guarantor cannot underwrite its own debt. RSL members incorporate the SHGF as a company limited by guarantee under the Companies Act 2006; no new legislation is required. Government backstops 80% of principal on eligible RSL bond issuances, with the SHGF holding a first-loss tranche of approximately 10% before the backstop is reached. Greater Manchester’s Housing Investment Loans Fund committed £983 million across nine years with zero defaults,¹² operating without a guarantee and within a jurisdiction where Right to Buy still erodes the collateral base. Scotland adds the guarantee on top and revoked Right to Buy in 2016.

Three moral hazard controls: total guaranteed issuance capped at three times the SHGF’s net asset value; RSL eligibility restricted to those holding SHR Compliant status with no Under Review flag; and a derivatives and treasury risk management code as a supplementary eligibility condition.

That third control addresses the failure mode SHR status alone does not catch. In 2012, the Dutch housing corporation Vestia accumulated approximately €2 billion in derivative losses without adequate treasury controls; it would have passed any standard credit filter. The sector absorbed the hit through the Centraal Fonds Volkshuisvesting (CFV) bailout and Vestia’s own repayment schedule; WSW’s loan guarantee mechanism was not triggered, and the government backstop was not called.

On central-case loss rates derived from WSW’s operating record (cumulative guarantee claim liabilities of approximately 0.27% of the guaranteed book at end-2024), net government exposure is approximately zero; the SHGF’s first-loss buffer absorbs expected claims before the backstop layer is reached. In a stress scenario at approximately fifty-five times the WSW historical rate (fifteen per cent cumulative) on a £2 billion guarantee book, the tail-risk exposure is approximately £80 million net. The full contingent liability methodology is in the Technical Annex.

The SHGF is company formation by RSL members under the Companies Act 2006; no legislation required. The government’s backstop commitment is executive action, partly deliverable through the 2026/27 bond programme. No Holyrood Bill slot required. Structured as a UK-wide instrument, a Sewel motion is the vehicle.

Proposals 5 to 9

Proposal 5: Statutory affordable housing floor. Replace negotiated Section 75 agreements on sites above 20 homes with a statutory 25% affordable floor, not dischargeable by commuted sum. At least 15% social rent as defined by the Housing (Scotland) Act 2010, with the remaining affordable tenure mix set by local Housing Need and Demand Assessment projections using the Scottish Government’s standardised HNDA tool. The 15% floor operates within private-developer lots; on publicly assembled SHFF land, the tenure target is the national 70% social rent commitment, producing the two-track delivery Annex 9 of the Technical Annex describes. A floor that permits commuted sums or accepts mid-market rent as social housing does not address the emergency the 18,092 figure describes. The 110,000 target requires 11,000 affordable completions a year. A statutory floor that cannot be bought out is the pipeline that feeds it.

If no RSL is identified within 12 months of consent, SHFF steps in as RSL of last resort. This proposal requires the finance machinery in Proposals 3 and 4 to be operational; without SHFF as a capitalised buyer and the Guarantee reducing RSL borrowing costs, the sub-floor triggers viability failures on marginal sites.

Proposal 6: More Homes Scotland as active land agency. More Homes Scotland was announced by the First Minister in January 2026, with a mandate the Cabinet Secretary for Housing described as acquiring land “at peppercorn value” (nominal cost), “carrying out necessary infrastructure and remediation work and then passing it on for development.”¹³ Proposals 1, 2, and 6 are the legislative architecture for that body. EUV plus via the Proposal 2 pre-emption right is the mechanism by which nominal-cost acquisition becomes operationally possible at scale; without the legislation, the Cabinet Secretary’s commitment is an aspiration. Infrastructure commissioning and cost-plus tendering of serviced plots to SME builders and RSLs sit inside the agency. SME builders who cannot compete for sites at hope value can compete for serviced plots at cost-plus; the architecture opens market entry that the current system forecloses.

Because MHS controls the land, tender conditions carry binding completion dates derived from the AHSP programme. Prolonged build-out is a product of private land assembly and absorption rate constraints; on publicly assembled sites, the agency sets the clock. MHS will be the most important near-term user of the national land valuation survey proposed in Proposal 9; the survey should be operational before MHS begins pricing acquisitions.

Proposal 7: Collective housing development pathway. A statutory planning pathway for collective self-build, with automatic eligibility for More Homes Scotland serviced plots and a structural anti-privatisation device preventing resale at market value.

Proposal 8: Scottish Vacant Residential Properties Register. Over 44,000 long-term empty residential properties in Scotland, though the policy-relevant subset (structurally sound, in housing pressure zones, held vacant for financial rather than operational reasons) is materially smaller. The register’s primary function is operational: it identifies the acquisition pool for SHFF’s year-one pipeline.

Ireland’s Vacant Homes Tax collected a fraction of expected returns; self-declaration does not work when every financial incentive runs against declaring. The register solves the identification problem directly.

The register is built on occupation signals the Scottish Government already holds: council tax, Scottish Water, Registers of Scotland, etc. The analytical approach identifies properties absent from all occupation signal datasets above a threshold period. Each source dataset is person-attributed at input; the query returns only a boolean per address. No personal data crosses system boundaries. The output is structurally incapable of identifying individuals. The analytical work is days. The bottleneck is data-sharing governance. The minister’s first question to officials should be: who has been asked to do this?

Legislative vehicle: operational infrastructure established by SHFF and More Homes Scotland once constituted, alongside the national S75 register; not in the Bill.

Proposal 9: National Land Valuation Survey. Scotland has no national land valuation survey. The SLC publishes agricultural and rural prices annually; no equivalent exists for residential, commercial, brownfield, or any other non-agricultural use class. The institutional capacity half-exists: Scottish Assessors already value every non-domestic property on a rolling revaluation cycle, producing composite rental values – land and improvements together. Extending their mandate to separate the land component and to cover residential land would produce the survey from within an institution that already does most of the underlying analysis. Where distributional consequences require phasing, transitional relief – capping annual increases at 5% until the new valuation is reached – is the standard mechanism; every business rates revaluation already works this way.

MHS will be the most important user in the short term, but the survey serves every land-based fiscal instrument: DRLT rate-setting, pre-emption valuations, and any future reform of local taxation. Its wider uses are a matter for Parliament.

The Package Logic

The four core proposals are not four separate policies that happen to share a topic. Each is a precondition for at least one other.

DRLT without pre-emption imposes a holding cost without closing the hope-value exit; speculative holders absorb the charge if values are rising fast enough. Pre-emption without DRLT compresses the disposal return without creating urgency around the disposal decision. SHFF without the Guarantee cannot reach bond capital at the rates the revolving fund requires; it becomes a modestly capitalised agency limited to what the Scottish capital budget directly provides. The Guarantee without the SHGF has no first-loss tranche, no leverage cap anchor, no layered loss-absorption architecture; it is the open-ended government backstop the Dutch explicitly rejected.

The sequencing is therefore load-bearing:

1. SHGF incorporated by RSL members as CLG under Companies Act 2006; requires minimum membership threshold by share of sector borrowing.
2. Government backstop committed via 2026/27 bond programme or Sewel motion; requires SHGF operational.
3. SHFF formed under s.2(2)(o) Housing (Scotland) Act 1988; requires Guarantee committed.
4. Housing Emergency (Structural Measures) Bill: Part 1 DRLT, Part 2 Pre-Emption; requires Guarantee committed.

One Bill, two executive actions. The government backstop is executive action through the bond programme. SHFF is company formation under existing Housing Act powers. The SHGF is not an executive action; RSL members incorporate it as a private-sector mutual. The vacancy register and national land valuation survey (Proposals 8 and 9) are operational infrastructure, not legislative provisions. Proposals 5 and 7 are planning framework amendments deliverable in the same session without fresh primary legislation.

The Comparators

Five primary comparators underpin the design. None is a clean success story.

Andy Burnham’s Greater Manchester is the closest operational precedent under UK fiscal rules. The Combined Authority’s Housing Investment Loans Fund started at £300 million in 2015; by mid-2024 it had committed £983 million from that revolving base, with zero defaults across 56 fully repaid loans and 10,974 homes delivered or contracted.¹⁴ Without the fund, only 46% of the 7,100 units completed between 2017 and 2023 would have been completed by 2023. The Housing First pilot housed 350 people at a 78% tenancy sustainment rate; rough sleeping fell 42% between 2017 and 2024. Three constraints prevented Greater Manchester from building the full architecture: Right to Buy remained in force, a Combined Authority cannot legislate the instruments this paper proposes, and the fund’s lending concentrated on a single developer group to a degree no commercial credit committee would approve. Scotland revoked RTB in 2016, holds the legislative competence, and the SHFF architecture distributes lending across a regulated membership base with portfolio limits that prevent the same failure.

Finland’s 80% homelessness reduction between 1985 and 2023 is real, achieved on top of Helsinki’s 70% municipal land ownership, though the trend reversed in 2024 following welfare cuts. The conversion mechanism transfers to Scotland; the starting position does not. The Finnish 80% is the architecture’s ceiling under favourable conditions, not its first-decade target.

The Dutch WSW’s backstop record is real; the moral hazard controls in Proposal 4 are built around the Vestia failure described there, not despite it.

Norway’s Husbanken has provided below-market housing loans since 1946. It survived multiple changes of government and was periodically narrowed in mandate on crowding-out arguments. A Scottish equivalent should expect the same political weather and build institutional durability against it from day one.

Ireland’s RZLT is the closest design comparator. The first-year behavioural signal is promising: land sales up 30%, median price per acre down 6%, three years after enactment and before a single liability was due. One year of data from a different jurisdiction with different planning and tax structures is indicative, not definitive.

The Irish data also carries a warning. In England, Barratt, Persimmon, and Taylor Wimpey filed 183 planning appeals between 2012 and 2019, with success rates of 59-71%.¹⁵ The same firms operate in Scotland. Speculation by appeal is as systematic as speculation by holding. Over 500 first-year RZLT returns in Ireland claimed deferral on the basis of a planning appeal in progress. Scotland’s DRLT is non-rebatable: liability accrues regardless of any pending legal challenge, with payment deferrable during litigation but the clock running and the full liability payable on resolution. EUV as the mandatory land cost input in any viability assessment and DRLT explicitly excluded from viability calculations are statutory requirements, not monitoring commitments. The Working Paper sets out the design.

Spain built approximately five million new homes between 2000 and 2009.¹⁶ The recovery arc remains incomplete in 2026, concentrated around investment fund asset disposal rather than social housing absorption. Planning liberalisation without land market discipline produces the wrong homes in the wrong places. That is the failure mode Scotland is one planning reform away from repeating.

The Fiscal Reality

Every proposal in this paper interacts with the Scottish Fiscal Framework.

DRLT does not require Fiscal Framework renegotiation.

Whether an annual holding charge falls within existing devolved powers or requires a Section 80B Order in Council under the Scotland Act 1998 before primary legislation can follow is a question for the Scottish Law Officers to resolve before drafting begins. The Scotland Act 1998 (Specification of Devolved Tax) (Building Safety) Order 2024 is the direct procedural precedent for the latter route.

DRLT must be structured as a national tax. Local DRLT revenues default to local authority expenditure by statute; hypothecating them to SHFF requires a Non-Domestic Rates-style pooling mechanism with no guarantee of uniform rates across Scotland’s thirty-two councils. Per-site scope determinations and value base disputes are each individually challengeable, which places DRLT in the same category as Land and Buildings Transaction Tax (LBTT). Revenue Scotland administers LBTT for precisely that reason: site-specific application criteria cannot be left to councils whose planning departments cannot sustain multi-year tax appeals against volume builders’ legal teams.

Revenue stays in Scotland, hypothecated to SHFF capitalisation until year five or until SHFF’s net asset value reaches £500 million, whichever is earlier, then reverting to general taxation. Those are illustrative parameters; the precise thresholds are a Bill team decision. What matters architecturally is that the bridge is bounded and the sunset is named. DRLT is a behavioural instrument. The objective is homes, not revenue. DRLT receipts must not be written into baseline budgets; a government that treats declining receipts as a budget failure has given itself an institutional interest in the problem continuing.

The Housing Finance Guarantee interacts with HM Treasury classification. For the envelope not to count against Scotland’s capital borrowing cap, Treasury must classify the Guarantee as contingent rather than drawn liability. That is a negotiation point, not a given. The 2026/27 bond programme is the immediate vehicle; put the classification question to Treasury counterparts during programme design rather than discovering the answer at implementation. If Treasury declines, the Guarantee book and SHFF capitalisation compete for the same cap headroom; SHFF alone already sits at the edge of the current envelope. In that scenario, year-one SHFF capitalisation comes primarily from DRLT receipts and phased direct state lending, the Guarantee is sized below its full design envelope, and both scale as the bond programme track record develops. The architecture is slower, not broken.

The SHFF capitalisation envelope, illustratively scaled to £1 billion over five years, sits at the edge of Scotland’s current capital borrowing envelope if funded through direct state lending alone. The specific capital and revenue model is a design decision for the founding coalition. Bond market access through the Guarantee shifts lending capacity from the state balance sheet to private institutional capital. MHS acquires land at EUV plus under Proposal 2 and disposes serviced plots at cost-plus prices reflecting EUV acquisition, infrastructure investment, and the planning consent attached to the land; the margin between EUV plus and disposal price is the revenue that capitalises SHGF’s first-loss tranche and SHFF’s acquisition pipeline, replacing DRLT bridge funding as the disposal programme matures.

The architecture is self-financing in steady state. The ramp-up requires bridge funding whose full envelope depends on the HM Treasury classification outcome. Year-one sources without Treasury agreement are DRLT seed receipts and phased direct state lending; the Guarantee-backed bond issuance that unlocks institutional capital at scale requires the classification question to be resolved. The Treasury negotiation determines how fast the architecture scales, not whether it works.

There is a question this paper does not resolve. The 110,000 target is already funded through the AHSP; the 18,092 households in temporary accommodation are already being paid for, in emergency services, social care, and deteriorating outcomes for 10,480 children. This architecture redirects existing expenditure flows. The question is not whether this is affordable. It is whether the incoming Parliament treats the emergency as one.

The Legislative Ask

The programme reduces to one Bill, two executive actions, and one workforce plan.

The Housing Emergency (Structural Measures) Bill carries two Parts: DRLT and statutory pre-emption as the paired land-market instruments. They belong in the same Bill for the same reason they belong in the same analysis; neither works without the other. The vacancy register (Proposal 8) and national land valuation survey (Proposal 9) are operational infrastructure; neither requires primary legislation.

SHFF does not need a Bill slot. Section 2(2)(o) of the Housing (Scotland) Act 1988 empowers Scottish Ministers to form companies under the Companies Act 2006 for their general housing functions. Company formation is faster than secondary legislation, consumes no parliamentary time, and can unlock year-one tenancy conversions before a single new home is built.

The Housing Finance Guarantee is partly deliverable through the 2026/27 bond programme as executive action. This should happen first, before the Bill receives Royal Assent, because SHFF cannot access bond capital at viable rates without the Guarantee already operating, and because the land-market reform must not commence before the Guarantee is committed.

A Construction Workforce Plan – further education apprenticeship pipelines, Modern Methods of Construction incentives, SHFF procurement clauses requiring minimum apprenticeship shares in trades positions – can be commissioned before the first housing Bill. Scotland needs 17,950 additional construction workers by 2029.¹⁷ Targets that outrun workforce capacity miss regardless of planning or finance reform.

Three objections arise predictably. On A1P1 compatibility: DRLT is a property tax, as all property taxes are, and the holder retains the choice to build, hold and pay, or sell throughout. Where pre-emption targets S75 non-delivery, it crystallises on failure to perform an obligation the developer accepted at consent; that is specific performance, which is the stronger A1P1 position, not the weaker one. On RSL sector stress: the Guarantee-first sequencing instruction is the structural answer; it must be committed before Proposals 1 and 2 commence, and the sequencing table is explicit on that. On fiscal headroom: the architecture scales from a confirmed year-one base; the Treasury classification outcome determines pace, not viability.

Who Loses, Who Gains

Any reform that produces real results creates real winners and real losers. A paper that does not name both is selling something.

Speculative landowners holding LDP-allocated stock for hope-value uplift lose the largest share of their expected return. That is the purpose of the reform. The speculative premium is the policy target. The rational response to DRLT and pre-emption depends on the holder’s capacity to build. Holders without it face a disposal at EUV plus.

Volume housebuilders face a business model change, not a margin trim. Foye and Shepherd (2023) found gross profit margins of 17-32% through 2014-2021 at Barratt, Persimmon, and Taylor Wimpey, with margins correlated to planning-permission value uplift rather than construction output: as house prices rose, the price paid for land stayed flat. Removing that model shifts return to construction efficiency. Adapters operate viably and gain: planning certainty and access to serviced plots they can build immediately, rather than years of land assembly at rising cost. Non-adapters lose market share.

Temporary accommodation providers, B&B operators, hostel chains, and private landlords holding dedicated TA units on local authority lease, lose a revenue stream that has grown to record levels on the back of Scotland’s housing emergency.

Against these: SME housebuilders benefit directly. The speculative premium is the moat that keeps them out. They can build; they cannot assemble land at hope value. Fill in the moat. RSL tenants benefit from faster routes to permanent tenancies and lower borrowing costs passed through as reduced rent pressure. The Scottish construction workforce gains from anchored multi-year demand at SHFF procurement scale. The 18,092 households and 10,480 children in temporary accommodation gain permanent housing faster. The 110,000 affordable homes target contains the 77,000 social homes where they and those who follow them will live.

The losers’ position rests on a market structure that produces 18,092 households in temporary accommodation. The status quo is not neutral. It is a distributional choice that allocates the cost of speculative land values to families waiting for tenancies and to public budgets absorbing the consequences. This reform changes that allocation. It does not invent it. Both outcomes produce homes.

The full policy paper The Missing Machinery and its Technical Annex are available on request. The Technical Annex carries the ECHR case law and proportionality treatment, DRLT revenue projections, SHFF capitalisation model, Guarantee contingent liability methodology and moral hazard controls, COSLA governance architecture, and the wider international comparator scan.