Who Pays for a New Neighbourhood?
How the French ZAC captures rising land value to fund infrastructure and public amenities
A zone d’aménagement concerté (ZAC), coordinated development zone, is the primary framework in France for planned urban development at the scale of a district.
The whole (re)development operation, from land assembly to residents moving in, is managed in a single account. Strategic land value capture is a fundamental component of ZACs. The integration in a single account captures value to fund roads, networks and public amenities before a single plot is sold.
First steps in a ZAC
A collectivité (local authority) sets the zone’s boundary, establishes the mix of housing, offices and public amenities that it wants, and appoints an operator (aménageur) to administer the project from land assembly to occupants moving in.
The aménageur runs one account across the lifetime of the zone. Fifteen to thirty years is normal. The account is the bilan d’aménagement (development account). It’s a multi-decade cash-flow projection and balancing account. The development account includes every expected cost and receipt, from the first land purchase to the final handover of buildings and public spaces.
In the early years the development account mainly contains expenses: buying land (usually by negotiation, with compulsory purchase available where needed), then preparing and servicing sites. These are then sold to promoteurs (property developers), who construct the buildings.
The land under schools, parks and other amenities is eventually transferred to the local authority. The receipts that pay for all of this come, overwhelmingly, from the sale of the serviced plots sold to developers.
Land charges and value capture
On the income side, the dominant item in the development account is charges foncières (land charges). These are prices paid by developers to the aménageur for building plots. The price set by the aménageur is differentiated by programme type (free-market housing, social housing, offices, retail, facilities).
In the finished buildings, end-user prices in open-market units correspond to the prevailing market price. An apartment buyer or commercial tenant will weigh up options within the ZAC against others across the city.
However the land charge - the developer’s site cost - is not based on an open market value of the plot itself. It’s based on a reverse calculation starting with the expected sale value of the completed scheme. From that value is subtracted construction costs, professional fees, finance and the developer’s margin. The resulting sum is charged to the developer; it’s the highest price they can pay for the land while maintaining the scheme’s viability.
By charging the residual of the sale price minus allowable costs and margin, the aménageur captures the planning and infrastructure premium the ZAC itself creates. That value contributes significantly to the budget for public amenities and social housing.
Ability to Pay
I mentioned the reverse calculation starting with the expected sale value of the completed scheme. What about non-market-based units like subsidised or social housing? The land charge is adjusted in these cases. This adjustment is built into the ZAC approach - referred to as péréquation (cross-subsidisation).
The plots that can bear a high charge are priced above their share of the infrastructure cost. The surplus covers the plots that cannot. Open-market housing, in effect, carries the social housing and the public amenities around it.
Two observations
Two noteworthy findings emerged from a 2024 study by the Institute of Advanced Studies for Action on Housing (IDHEAL), “Transparence sur les ZAC”. The first concerns which building use carries the lion’s share of ZAC development costs. The second relates to which cost category dominates the development account.
The common belief regarding ZAC operations is that profits from office development fund the housing. Across the thirteen operations IDHEAL examined, it was open-market housing that carried the highest land charge. Housing subsidised the offices, not the reverse. The land charge for social housing came in at around 65% of the open-market housing charge.
As for the dominant cost in the development account: across IDHEAL’s sample, land averaged 21% of total spend, ranging from 3% on a former military site in Brest to 46% on dense private land in Montreuil. The largest single line was infrastructure: roads and networks averaged 40%. Based on the study, opening up an enclosed inner-city site can cost as much as servicing a green field, sometimes more.
How the UK compares, and why the difference matters
As outlined above, France fixes the cost framework for a whole neighbourhood upfront, in one account, before plots are sold. The UK settles infrastructure costs site by site, after the value has already been created by the urban plan zoning. In England, two mechanisms are used.
The first is Section 106 (S106) agreements. These are obligations negotiated site by site as a condition of planning permission: for contributions to affordable housing, local infrastructure, and so on. There is no national tariff. What a developer provides depends on what the authority can extract in negotiation.
The second mechanism is a Community Infrastructure Levy. This is a fixed charge per square metre of development. It is set by each local planning authority, but only around half of planning authorities have adopted one. The rest rely on S106 negotiation alone, which is known to be slow and costly.
The cost of the UK approach extends beyond inefficiency. A 2026 report for the UK Collaborative Centre for Housing Evidence described the negotiation system as a “complexity industry,” where expertise is unevenly distributed and the better-resourced side tends to win. The report found that developers paid on average 2,450% more for sites than the land values they themselves declared when later arguing, at the planning stage, that the scheme could not afford its full affordable proportion. In reality the developer paid a high price, betting the social obligations could be negotiated away. When they are, the original landowner has already pocketed the uplift, and the public loses the affordable housing, infrastructure and amenities the land value could have funded.
The French ZAC works the other way round. The site cost (land charge) is worked back from end values and fixed in the concession agreement before the developer bids. There is no later negotiation stage at which the obligations can be reopened and bargained down. The obligations come out of the land value uplift, as intended.
Most US tools capture value after the fact. Tax increment financing freezes the property-tax take in a district, then ring-fences the future rise in tax revenue as property values climb to repay the cost of the infrastructure. Developer impact fees are fixed, published charges levied per unit when the building permit is issued, set to cover the roads, utilities and schools that new development requires.
Both tools wait for the market: neither involves assembling land or funding infrastructure before the development gets the green light, which is what sets the French model apart.
California’s Mello-Roos districts come nearest to the French approach, funding infrastructure upfront through bonds, but the bill falls on future residents as an annual levy over decades. It doesn’t adjust the land price at the point the developer buys it.
The common thread of most Anglophone systems is that authorities collect after the fact. They wait for permission to be granted or values to rise, then take a share of an uplift that has already landed in private hands. France reverses the order: it commits the spending first and recovers it from the land price charged to the developer.
When the music stops
As the land charges in a ZAC are linked to projected open-market sale costs, the ZAC timeline exposes the operation, and the development account, to the highs and lows of the market cycle.
A ZAC planned in a rising market reaches its plot sales years later, possibly in a downturn. Site preparation and infrastructure is front-loaded and land charges and other receipts come later. Any delay in selling plots inflates the aménageur’s financing costs and the development account balance gets closer to negative . Who absorbs a deficit depends on the contractual link between the local authority and the aménageur.
Where the collectivité runs the zone directly, in régie (direct management), it carries the whole risk on its own budget. Where the zone is handed to an operator under concession, the contract decides. Real risk only passes to the operator if the risk is contractually transferred. Where the authority guarantees the operator’s debt, or agrees to buy back unsold plots, the risk stays in the public domain.
The Paris Rive Gauche ZAC was created in 1991 on 130 hectares of former railway land. It shows where public exposure leads. The development was meant to be complete in 2006, but the regional audit court now puts the realistic date at 2032. The City of Paris has put in at least €1.42 billion into the project, once off-account public works are included, and guarantees the aménageur’s debt on preferential terms. The court’s conclusion was blunt: the city underwrites the commercial risk.
Smaller ZACs in deficit simply get cancelled. In 2024 the commune of Villemomble closed a ZAC operation that was facing a forecast €8m deficit, swallowing the sunk study costs. And the present downturn is testing the model across the board, with new-home reservations roughly halved against 2019 and unsold stock at around 18 months’ worth of sales.