The gift that costs nothing to keep

Planning consent multiplies the value of a field overnight and hands the gain to whoever owns the ground. Holding that gain costs the owner nothing, so holding it is the rational thing to do. Until delay carries a price, the land waits, and the homes wait with it.

The grant of planning consent is a gift. The moment permission is given, the state manufactures a leap in the land’s value and hands it, unbilled, to whoever happens to own the ground. No one has dug a trench or laid a pipe; the value appears because the state’s signature appears, and it lands in a private account. In England, where the figures are published, consent can lift a field’s value tenfold at the conservative floor and a hundredfold or more where demand is fiercest. Scotland keeps no equivalent survey of residential land value, only a low agricultural baseline, so the precise multiple here is read across from the English data, and the missing survey is its own admission.¹

The consent creates the value, and the owner merely holds the ground it lands on. The principle is old, and it has a name: betterment. The uplift is made by the public decision to permit, not earned by the person who owns the field, and a century of land reform argument has tried and mostly failed to recapture it. England claws back roughly a third of development value through planning obligations; Scotland recovers a small fraction of that.² The gift is real money, publicly created and privately kept.

The cost of holding that gain is nothing. Council tax applies to homes that exist, non-domestic rates to occupied commercial property; land declared for housing with no home on it carries neither. The owner holds an asset the state has just enriched, watches it appreciate, and pays for the privilege with patience. Releasing it slowly is rational, the correct answer to the incentives as they stand: an appreciating asset with no carrying cost is one you can meter out at whatever pace protects its price, because nothing charges you for the wait.

The wait is a choice

Glasgow’s own land audit puts the most a large housebuilder manages at around twenty-five market homes per site a year, and calls market housing far slower and more variable to deliver than other tenures.³ Across Scotland the Land Commission found only twelve per cent of large-site permissions completed within three years, against thirty per cent of small ones.⁴ The pattern is metering: a volume builder releases homes at the rate the local market will take at the price it wants; sell any faster and the price softens, and the margin with it. The build-out rate is set by the margin the builder holds out for, a real constraint, but one of the builder’s own choosing.

In an ordinary market a high price would not hold. A rival would undercut it, and the price would fall to meet demand. Housing is sold by location, and no rival can undercut you on the plot you alone hold the consent for. The price holds, and the wait does the rest. The handful of large builders need no agreement to keep it there: each knows that flooding its local market would soften the price for everyone, itself included, so each meters, and the restraint emerges with no word exchanged. The competition regulator looked for coordination and found none, and none is needed: the super-profit is defended by deferral, not by a pact.

Where the money is made shows the rate is chosen. The three largest builders by volume held gross margins between seventeen and thirty-two per cent, year after year, and the margin came from the land rather than the building, from buying strategic land, carrying it through the planning system, and capturing the uplift when permission lands.⁵ A Taylor Wimpey chief executive called it an oligopoly in land buying. The margin being protected is, in large part, the gift itself, the uplift the state created at consent; the metered build-out is that gift being defended, the public’s handed-over value guarded by the slow release that defers the homes behind it. Slow is not a failure of the model. Slow is the model.

Holding the consented land costs nothing; building on it costs money. That asymmetry catches both kinds of builder. The volume builder with the balance sheet to carry a site meters it to protect the price on every home already sold; the small builder, its capital tied up in the ground it is already turning, builds only as fast as it can fund. Different reasons produce one result: nothing charges either for the time the land waits, so delay is the frictionless default, and nothing in the system pulls the land toward the homes. A holding cost is the missing measure, the competitor a location-bound market never sends. It has a name: the Designated Residential Land Tax, a local charge on declared residential land that contains no residence, lifted only by building one.

The regulator’s objection

The competition regulator blames the planning system. The Competition and Markets Authority studied housebuilding across Great Britain and found the size of land banks to be “primarily a symptom” of problems in the market, “mainly driven by the time and uncertainty associated with obtaining planning permission”. It counted the largest builders holding some 107,000 plots in Scotland, cautioned that artificially shrinking land banks without fixing the planning system underneath would miss the point, and declined to refer the market for investigation.⁶ If the regulator is right, a holding cost punishes builders for a queue they did not create: the land is banked because the system upstream is slow, and charging for the wait blames the symptom for the disease.

The charge does not need the objection won on its terms. Whether a site sits for price protection, thin capital, a slow planning queue or plain inertia, none of those reasons meets a cost. The charge prices the holding, not the motive: it puts a price on the wait itself, whatever lies behind it.

The regulator’s own analysis points the same way, separating the questions the objection runs together. The size of a land bank is one question, and the CMA puts that down to planning delay. How fast consented land then gets built is another, and there the CMA names the for-sale model itself, “the limited amount of housing being built outside the speculative approach”, with build-out rates a further factor.⁷ The tenures built outside that model, affordable housing and build-to-rent among them, carry no per-unit sale price to protect, and deliver differently for it: Glasgow’s build-to-rent ran well ahead of its programme in the same year that for-sale build-out sat at its assumed rate.⁸

A holding cost does not attack the size of anyone’s land bank, the part the CMA ties to the planning queue. It charges non-delivery on declared land that could be built and is not, the build-out question the regulator puts on the speculative model itself. The site genuinely stuck in that queue is spared by a suspension built into the charge for exactly that case. The charge bites wherever declared housing land sits undeveloped, and lifts off only where building is genuinely blocked.

Scotland’s build-out is quicker than England’s, far short of the decade-plus build-outs on England’s very largest sites.⁹ The land banks, too, are in part genuinely needed: a builder must hold some forward consented land to keep working. Both are real, and neither costs the case. What is not needed is the surplus held back to protect price, and the charge tells the two apart, because land in active build-out sheds it hectare by hectare while the bank held slow is what pays.

The instrument

The charge applies each year to declared residential land with no residence, and the only way off it is completion, hectare by hectare, so that each finished parcel leaves the charge as the homes go up. Holding is what costs; building is what stops the cost. The pressure runs in one direction only, toward the outcome the consent was granted for in the first place.

The charge has to be non-rebatable to work, and that is the part most easily lost. Offer relief for a promise to build, or a rebate once building finishes, and a metered build-out simply reclassifies itself as a committed one: the builder pledges a schedule, pays little or nothing while the schedule runs, and draws the same homes out over the same decade with the charge deferred into a liability that never bites. A holding cost with an escape hatch for those who choose patience is no holding cost at all. Until its infrastructure is delivered, a hectare sits outside the charge; after that, the only relief is the finished home, because the finished home is the only thing the charge exists to produce.

The case the regulator worried about is handled by suspension, not by a loophole. A hectare held back by a verified infrastructure constraint, a water connection or a grid reinforcement the owner cannot supply alone, is suspended from the charge against a committed delivery date from the body responsible; if that date slips, the slippage sits with the water company or the grid operator, not with the owner of land no one can yet build on. The charge concentrates on the site that can be built and is not, and lifts off the site that cannot. The test is administrable because the information already exists: every local authority records each site’s constraints, effectiveness and programming in its housing land audit, amalgamated nationally to a single schema.¹⁰ The valve reads off data the system already collects.

The charge is not aimed at derelict or vacant ground, the absentee owner’s neglect that other instruments already chase; it applies to declared housing land released a little at a time.

One case the charge deliberately does not shelter is the site with no buyers. If demand at a location is genuinely absent, the answer is not to suspend the charge indefinitely while the owner waits for a market that may never arrive; it is to take the allocation back at the next plan review. An allocation nobody will build on is a planning misjudgement to correct, not a holding to protect. Suspension is for the site that cannot yet be built, never the site that should not have been allocated, and keeping those two apart is what stops a relief valve becoming a parking space.

The plea that a site can never be built is a confession. If the land truly cannot hold a home, it can hold a park, and the public may take it at what it was worth before consent dressed it up, the value the owner has just conceded. The plea that lifts the charge forfeits the gift.

A charge on holding presses on the builder who cannot fund a faster pace as hard as the one who is choosing to wait, the real cost of pricing the holding rather than the motive. Exempting the small builder would only open a gap the strategic holder would pour straight through; the fix belongs in the terms on which the land changes hands.

One thing more the charge shrugs off: where the owner lives. Income and ownership can be domiciled in a regime that does not return calls; the hectare cannot. It stays on the register with its coordinates, and the charge attaches to the land, not the name behind it. The annual bill accrues against the ground all the same; anything left unpaid becomes a lien on it, cleared when the land sells or develops, the way Ireland secures its own zoned-land tax. Where Scotland’s register of who controls the land strains to unmask the owners and largely fails, the charge need not ask: it bills the field.¹¹ That backstop leans on the Land Register being complete, which it is not yet quite. You can hide the owner; you cannot hide the hectare.

The point of the charge

What the charge raises is beside the point. A holding cost works by not being paid: faced with an annual bill that only building or selling will stop, the rational owner builds or sells, and a charge that empties the holdings it was aimed at collects very little by design. You would no more judge it by its yield than call a carbon tax a failure for the emissions it never priced because they never happened. Low revenue is the instrument working.

And the state is not flush. It is always short, and it freed that record £926 million for affordable housing in 2026-27 by effort, against every competing claim.¹² Then it spends the money down a channel that hands a share of every pound back to whoever owns the consented land, the grant buying land at its gifted value, so part of a budget it strained to assemble funds the gift instead of the homes. The cure was never more revenue, and the charge’s pennies are least of all. Local is deliberate, not a loose end. A local charge keeps the company of the council tax and rates, levied within devolved competence, where a national land tax would reopen the contested question of whether Holyrood can raise one without Westminster’s consent.¹³ What it raises and who spends it is the one question worth leaving open. The charge earns its place as pressure, not as income.

The pressure is not hypothetical. Ireland brought in a residential zoned land tax, an annual charge on zoned, serviced residential land, and in its first window Irish land sales rose by about thirty per cent while the median price of residential land fell about six per cent.¹⁴ Owners faced with a holding cost did what the cost was meant to make them do: they transacted, and the price of land softened toward the value of what could actually be built on it. The Irish version is deferrable while development is under way, and its first year drew a wave of deferral claims, the very loophole through which a metered build-out escapes; a charge with no relief but completion closes it. The behaviour is the evidence. The yield is not.

A holding cost changes the cost of delay. It does not change where the land ends up. An expensive wait still leaves the rational owner answers short of building: pay the charge out of a gift still appreciating faster than the charge bites, or sell to another owner who runs the same arithmetic and waits in turn. Nothing in an annual bill routes the land toward the homes the consent was granted for. The gift can still be kept; keeping it now simply costs a little.

So the pressure needs somewhere for the land to go when the pressure works, a destination for the site whose owner will not build it and will not, under financial weight alone, let it pass to anyone who will. A charge can make the wait expensive. It cannot make the homes. For that, the obligation the owner already accepted at consent has to carry a consequence when the deadline comes and the homes have not.