The Housing Dividend

The problem

America is short somewhere between 3.8 and 5.5 million homes, depending on whose count you trust. Freddie Mac, the National Association of Realtors, and Up for Growth have used different methods and arrived at the same neighborhood, which is usually a sign the number is real.

The shortage is not a construction problem. American builders put up more than two million homes a year in 2005. It is a permission problem: in the metros where jobs actually are, it is illegal to build the kind of housing people can afford on most of the residential land.

Here is the part that this catalog cares about, and that almost nobody talks about.

When a city council votes to allow six units where one was allowed before, the value of that parcel can double or triple overnight. No one dug a foundation. No one poured concrete. A body of elected officials changed a line in a code book, and tens or hundreds of thousands of dollars appeared.

Where did that value come from? It came from the public. From the road, the water main, the transit stop, the school, the police force, and the permission itself. And it goes, entirely, to whoever happened to hold the deed on the morning of the vote.

This is the largest routine transfer of publicly created wealth to private hands in American life, and it is invisible because it never appears in a budget.

The proposal

Legalize abundant housing, and capture a modest share of the land-value uplift that legalization creates. Homes get built. The public keeps a stake in the value its own decision made.

How it would work

1. Permission first, and generously. By-right approval for multifamily housing near transit and along commercial corridors in any metro above a population floor, with a fixed shot clock for review. This proposal is worthless without the supply half, and the supply half is worth doing even without the rest. Nothing here is a reason to build less.

1. A land-value increment charge. When a rezoning raises the permitted development capacity of a parcel, an assessed share of the resulting uplift is due, payable at permit issue or at sale. The rate is a minority share, not a majority. The owner must keep the larger part or the mechanism kills the very projects it is meant to unlock.

1. Payable in homes. An owner may discharge the charge by deeding permanently affordable units instead of cash, at an exchange rate published in advance. Most will choose cash. The option matters because it lets a city take the dividend in the form it actually needs.

1. Split the proceeds. A majority to the local jurisdiction for the infrastructure that new density requires, which is the only way councils will ever vote for this. The remainder to the American Permanent Fund as principal under the locks of № 023.

1. Federal land, released. The federal government holds roughly 640 million acres, about 28 percent of the country. Almost none of it is anywhere near a job. But parcels of surplus federal land inside and adjacent to Western metros are real, identified, and idle. Release them for housing on long ground leases rather than sale, so the freehold stays public and the ground rent runs to the Fund forever.

The numbers

Start with the uplift, because it is the number nobody publishes.

Land is roughly 35 to 55 percent of home value in supply-constrained coastal metros, against 15 to 20 percent in Texas and the Southeast. That gap is the price of permission, and it is the base this proposal taxes.

Suppose upzoning across the constrained metros unlocks 500,000 additional units a year, and the average parcel-level uplift attributable to the rezoning is $60,000 per unit. That is $30 billion of publicly created value a year. A 25 percent increment charge yields $7.5 billion annually, of which a 60/40 local split sends $3 billion a year to the Fund.

At a 5 percent real return, $3 billion a year compounds to roughly $40 billion of corpus in a decade, throwing off $2 billion a year permanently. Per citizen, about $6 a year.

Six dollars. Say it plainly, as with № 031.

The dividend is not the point of this proposal and we will not pretend otherwise. The point is the 500,000 homes. The uplift capture is what makes the 500,000 homes politically survivable, because it gives the neighbors and the city a visible, funded answer to "who pays for the sewer, the school, the bus." A rezoning that arrives with money attached passes; one that arrives as a pure imposition on incumbents does not.

The Fund's share is the tail, not the dog. It is real, it compounds, it stacks with every other stream in this catalog, and it establishes that publicly created value is publicly owned. But the housing is the product.

The honest objections

"A charge on development reduces development. You are taxing the thing you say you want." The most serious objection, and the reason for the design choices above. The answer is that this charge falls on land, not on structures, and land does not flee. The classical result, running from Ricardo through George to modern public finance, is that a tax on unimproved land value is close to non-distortionary because the supply of land is fixed. A parcel worth $500,000 after rezoning and $300,000 before is worth $450,000 after a 25 percent increment charge, which is still far better than $300,000. Nobody declines that trade. The design fails only if the rate creeps toward the full uplift, which is why the statute must cap it well below.

"Cities will treat this as a revenue source and start selling permission." They will be tempted, and this is a genuine hazard. Permission that can be bought is corruption with a receipt. Two guardrails: the charge attaches to by-right upzoning applied to whole districts, never to parcel-by-parcel negotiation, and the local share is restricted to infrastructure, not general funds. Discretionary, deal-by-deal value capture is exactly the disease. This is the vaccine, and it only works if it stays formulaic.

"This is a federal proposal about a local power. Zoning is not Washington's business." Correct, and the proposal does not pretend otherwise. Washington cannot rezone Palo Alto and should not try. What it can do is condition the federal transportation and housing dollars it already sends on adopting the framework, which is the same lever used for the drinking age and highway funds. That is a real constitutional argument and reasonable people land on both sides of it. We would rather use the lever than pretend the shortage is not national.

"Existing homeowners see their equity fall. That is most of the country's net worth." Some will, and honesty requires saying it. About 65 percent of American households own their home, and for most of them the house is the balance sheet. A serious supply expansion slows price growth, and slowed growth feels like loss. Our position is that a country cannot price its own children out of its cities to protect the paper wealth of its parents, and that the same parents are, in the main, the ones housing those children into their thirties. But it is a real cost borne by real people and we are not going to describe it as a win for everyone.

"Six dollars a year." As with № 031: the compounding corpus and the principle, not the first cheque. And here, unusually, the dividend is explicitly the secondary benefit. If this proposal delivered zero dollars to the Fund and 500,000 homes a year, we would still publish it.

Sources

• Housing shortfall estimates: Freddie Mac (~3.8M units), National Association of Realtors (~5.5M), Up for Growth (~3.9M) — differing methodologies, converging range
• Federal land holdings: approximately 640 million acres, ~28% of the US land area, managed principally by BLM, USFS, FWS and NPS (blm.gov)
• Land share of property value in constrained versus unconstrained metros; Lincoln Institute of Land Policy land-price series (lincolninst.edu)
• Value-capture precedents: Hong Kong MTR "rail plus property"; São Paulo CEPAC auctions; Texas tax increment reinvestment zones
• Henry George, Progress and Poverty (1879), on the taxation of unimproved land value
• Proposals № 001 (The American Dividend); № 008 (Baby Shares); № 023 (The Debt Covenant, which governs payout); № 045 (The Permitting Clock)