Housing · scarcity · political power

The city with
two housing policies

One policy shields a limited number of households from prices they cannot afford. The other decides how many homes may exist, often making that rescue necessary in the first place.

After Gabriel Metcalf, 2018 Later evidence + a London coda July 2026
Enter the queue

01

Los Angeles, 2017

One voucher, up to 250 applicants

When Los Angeles reopened its waiting list for federal rent support after thirteen years, up to 600,000 people were expected to apply for 2,400 vouchers.1

Every voucher could change a life. At the upper estimate, the arithmetic still describes a lottery: 249 people remain outside for every person who gets through the door.

Upper estimate · 250 applicants

1 receives help

249 remain in the market that produced the queue

A field of 250 dots. One is highlighted to represent one voucher recipient; 249 are not highlighted.

The queue exposes a distinction that housing debates often blur. A city can decide who pays less for an existing home. It can also change how many people the housing system can accommodate. The first task is relief. The second is capacity.

Gabriel Metcalf’s paper followed both policies through expensive American cities. He found energetic programmes for subsidised homes, rent support and tenant protection sitting beside a much larger system of zoning, approvals, standards and fees that restrained the supply of ordinary housing. The relief was real. Its scale was tiny beside the scarcity around it.

02

Four levers, four different jobs

Four different jobs hide inside “affordable housing”

The four main instruments act on different parts of the system. They can complement one another. They can also cancel one another out.

KEEP

Social housing creates a permanent foothold

Public, cooperative or nonprofit ownership can keep homes affordable beyond one tenancy and remove some urban land from the race for speculative gains. It can serve people whom falling market rents would still not reach.

The constraint is scale. Metcalf’s deliberately simple illustration used a subsidy of $300,000 per home: helping one million households would require $300 billion. In the most expensive cities, the subsidy per new home can be higher.1

Best at: durable affordability and protection for households the market will not serve.

03

The policy beneath the policy

Scarcity can swallow the subsidy

A programme may create a few permanently affordable homes while the rules surrounding it make every market home harder and more expensive to produce.

Inclusionary housing shows the tension neatly. A city requires a share of below-market homes in a private development, or charges a fee to fund them elsewhere. The affordable homes are valuable. Yet the obligation is supported by the market-rate homes that remain viable after land, construction, finance, delay and all other charges have been paid.

Metcalf reported San Francisco fees and exactions of roughly $60,000 to $150,000 per market-rate home at the time.1 A cost known in advance may reduce what a builder bids for land. A cost that is uncertain raises risk. Either can leave marginal sites undeveloped. The city then collects affordable-housing revenue from a smaller stream of projects.

1Every project becomes a negotiationMore studies, meetings, concessions and legal risk
2Time and capital become more expensiveLarge insiders cope better than smaller entrants
3Fewer sites can carry the total costThe tax base for affordable housing also shrinks

Discretion is a housing cost

A rule can be strict and still predictable. A discretionary process adds a second burden: nobody knows the final price of permission. Developers pay for years of holding land, specialist reports, lobbying and the risk of rejection. Investors demand compensation for uncertainty. Smaller builders are less able to survive the wait.

Each concession can be defensible on its own. The system-wide result may still be fewer homes, less competition and higher prices. This is one of the paper’s most portable insights: project-by-project virtue can produce city-wide scarcity.

04

One market, many vetoes

The shortage is regional. The decision is local.

People cross municipal boundaries to work, study, care and socialise. Housing and labour markets therefore operate across a metropolitan region. Planning power is usually divided among the local governments inside it.

I

Every municipality can hope another one builds

New residents require schools, services and infrastructure. Jobs and commercial activity often strengthen the tax base. Each local government can therefore prefer employment growth inside its boundary and housing growth somewhere else. Across the region, everyone free-rides on everyone else.

II

Homeowners vote on the value of their largest asset

For many households, a home is shelter, savings, inheritance and retirement security at once. Restricting nearby supply can protect scarcity and prices. Even without a conscious cartel, elected officials hear repeatedly from people who already own a place in the city.

III

Future residents are missing from the hearing

The people most helped by additional homes may live two hours away, share overcrowded rooms or have abandoned the move altogether. They cannot vote in the jurisdiction that excluded them. Their absence is treated as consent.

IV

Renters can fear building for rational reasons

A market-rate project may not make today’s low-income tenant’s rent affordable. Construction can coincide with neighbourhood change, landlord pressure and visible wealth. A credible pro-building coalition therefore needs direct anti-displacement measures, not lectures about the regional average.

The missing constituency

The person who would live here if a home existed

Local democracy counts current residents. Housing scarcity is also experienced by absent residents: the nurse who declines the job, the adult child who leaves, the family pushed beyond the rail map, the migrant barred from a productive city by its admission price.

05

The argument after publication

What later evidence has added

The broad diagnosis has gained empirical support. Some trade-offs have also become sharper than the original paper suggested.

15%less rental supply

Peer-reviewed · San Francisco

Rent control protected tenants and changed landlord behaviour

A 2019 quasi-experimental study found that covered tenants moved less and were less likely to leave San Francisco. Landlords reduced the rental supply of treated properties by 15%, mainly through conversion and sale. Protection now and affordability later pulled in opposite directions.2

45–70moves reach below-median areas

Peer-reviewed · 12 US cities

New homes start moving chains

Following address histories from new apartment buildings, Evan Mast estimated that 100 new market-rate homes set off 45–70 moves out of below-median-income neighbourhoods within five years. The new units were expensive; the vacancies they released did not remain at the top of the market.3

2%lower nearby rents

Working paper · San Francisco

Nearby supply and neighbourhood change can occur together

Kate Pennington estimated that rents within 100 metres of new construction fell by 2% and the risk of displacement to a poorer neighbourhood fell by 17%. The same immediate area also showed more renovation, business turnover and richer residents. Supply relief and gentrification signals were both present.4

¾of residential land upzoned

Working paper · Auckland

Large-scale reform produced enough building to test the theory

Auckland allowed more housing on roughly three-quarters of its residential land in 2016. Six years later, a synthetic-control study estimated three-bedroom rents at 26–33% below the counterfactual without reform; estimates for two-bedroom homes were smaller and less consistently significant.5

The case has strengthened. Building is still hard.

Zoning permission does not pour concrete. Interest rates, infrastructure, construction capacity, land assembly and demand still matter. City-level results travel imperfectly. The strongest conclusion is more modest and more useful: new supply can relieve pressure beyond the new buildings themselves, while tenant protection and social housing remain necessary for people exposed during the transition.

06

A London coda · figures dated

The institutions differ. The structural problem is familiar.

London is one labour market spread across many planning authorities, with severe affordability, large public ambitions and a delivery system that repeatedly falls short.

10.6×average London house price relative to average earnings, 2025ONS
41.6%average private rent as a share of median private-renter household income, 2024ONS
88,000homes a year in London’s current government target; historic delivery about 30,000–45,000London Assembly

The London Assembly reported that the capital added 7,674 affordable homes net in 2023–24, around 18% of the GLA’s previously estimated annual need of 42,841.8 The precise programmes and targets will change. The scale gap is the durable fact.

Britain’s Competition and Markets Authority reached a diagnosis that echoes Metcalf’s account of American approvals: lack of predictability, long and complex processes, high costs and inconsistent incentives were key drivers of under-delivery. It also warned that planning reform alone might not produce the quantity or affordability policymakers want.9

That last sentence matters. Private construction, social and council housing, tenant security, infrastructure and planning reform each answer a different question. London needs a programme large enough to hold all of them at once.

07

A programme equal to the problem

Build more. Protect people. Keep homes affordable.

Across the system, the same concrete test returns: does this policy widen access to a successful city, or ration it more fiercely?

A credible housing settlement has three simultaneous jobs. It must allow many more homes in places connected to work and public life. It must protect people whose homes and communities are at risk during change. It must create a growing stock of homes whose affordability does not disappear with the next market cycle.

1

Expand capacity

Legalise more homes, smaller homes and different forms of shared living; make compliant approval predictable; align infrastructure with growth.

2

Protect people

Use tenant safeguards, relocation support, legal aid and targeted rent assistance so regional gains are not purchased with avoidable individual harm.

3

Build a nonmarket stock

Fund social, council, cooperative and community-owned housing from a broad tax base, preserving affordability across generations.

Metcalf’s seven proposals included upzoning, rethinking minimum standards, better transport links, new towns, regional tax sharing, higher-level responsibility for housing and more social housing. The list matters less than the architecture behind it: match the scale of government to the scale of the market, remove local rewards for exclusion, and spend public money where it creates durable access.

08

A portable policy test

Six questions for the next housing promise

A policy can sound compassionate, ambitious or pro-growth and still fail at the level of the whole city. These questions expose the missing mechanism.

  1. 01
    Who receives help now?

    Current tenants, applicants on a list, first-time buyers, landowners, developers, or a wider public?

  2. 02
    Does it add homes?

    Some policies redistribute access to a fixed stock. Others expand the stock. A serious programme says which job each measure performs.

  3. 03
    What happens outside the protected group?

    Follow rents, vacancies, conversions, waiting lists and displacement beyond the immediate beneficiaries.

  4. 04
    Who has veto power?

    Compare the people in the hearing with the people who would live closer, move in or stay if housing were available.

  5. 05
    Is the decision made at the scale of the housing market?

    A regional shortage cannot be solved by municipalities each volunteering the next municipality.

  6. 06
    How is affordability funded?

    A broad tax base can support a durable social stock. Loading every public goal onto each new home may reduce the number that gets built.

09

Reading and evidence

Sources

Historical figures retain their original dates and dollar values. Later studies are presented as additions to the evidence, not retroactive proof that every claim in a 2018 essay was correct.

  1. 1
    Gabriel Metcalf (2018), “Sand Castles before the Tide? Affordable Housing in Expensive Cities,” Journal of Economic Perspectives 32(1), 59–80. Article and free PDF.
  2. 2
    Rebecca Diamond, Tim McQuade and Franklin Qian (2019), “The Effects of Rent Control Expansion on Tenants, Landlords, and Inequality,” American Economic Review 109(9). Article and free PDF.
  3. 3
    Evan Mast (2019; JUE publication 2023), “The Effect of New Market-Rate Housing Construction on the Low-Income Housing Market.” Open working paper and publication details. See also Bratu, Harjunen and Saarimaa’s Helsinki moving-chain study, open access.
  4. 4
    Kate Pennington (2021), “Does Building New Housing Cause Displacement?” Working paper. Abstract and paper.
  5. 5
    Ryan Greenaway-McGrevy (2023), “Can Zoning Reform Reduce Housing Costs? Evidence from Rents in Auckland.” University of Auckland Economic Policy Centre Working Paper 016. PDF.
  6. 6
    Office for National Statistics (2026 release, 2025 data), “Housing affordability in England and Wales: 2025.” Bulletin.
  7. 7
    Office for National Statistics (2025 release, 2024 data), “Private rental affordability, England, Wales and Northern Ireland: 2024.” Bulletin.
  8. 8
    London Assembly (5 January 2026), “How will London fund its housing needs?” Briefing.
  9. 9
    Competition and Markets Authority (2024), Housebuilding market study, final report and summary. Report.

Editorial note

This page is an interpretive guide to a policy argument. It distinguishes direct findings, the original author’s judgments and later evidence. It does not estimate the effect of a specific proposal in a specific neighbourhood.