In coastal cities, a home costs twice what it takes to build one

The wedge between price and cost

Take the price of a house and subtract what it cost to build. In most of the country the remainder is small. Across the metropolitan areas Glaeser and Gyourko studied, roughly three-quarters of homes in 2013 sold near or below what they call the minimum profitable production cost, the sum of land, materials, labor, and a builder's normal profit.¹ Building an economy-quality house runs between about $72 and $86 a square foot.¹ Where a finished home sells near that figure, zoning is not doing much.

Then look at the coasts. In a small set of markets, greater New York and coastal California among them, the same house sells for more than double what it costs to build.¹ That gap has a name. Glaeser and Gyourko call it a regulatory tax: the wedge between price and production cost that no natural scarcity explains, because land, as they put it, "is worth far more when it sits under a new home than when it extends the lot of an existing home."¹ The constraint is not dirt. It is permission.

A rule that forbids anything denser than one house per lot restricts supply, and restricted supply raises price. Glaeser has run the arithmetic on the extreme case. The cost of adding a square foot at the top of a New York tower is "typically less than $400," and a decent 1,200-square-foot apartment costs "not more than $500,000" to build.² Such apartments sell for multiples of that, because the city stopped letting them go up. Manhattan permitted more than 11,000 new units a year from 1955 to 1964, and 3,100 a year from 1980 to 1999.² In the single year of 1960 it permitted 13,000 units, nearly two-thirds of everything it would permit across the whole of the 1990s.¹ Nothing about the island changed except the rules. "Opposing new building," Glaeser writes, "is the surest way to make a popular area unaffordable."²

The bill is not paid only by the family that overpays for one house. Glaeser and Gyourko put a floor on the national cost of restrictive land-use rules at "at least 2 percent of national output."¹

Who collects the difference

A tax collected by no government is still collected. The difference between price and cost does not vanish; it accrues to whoever already owns the scarce, permitted homes. Glaeser and Gyourko are blunt about who that is: "The big winners from the reduction in housing supply are a small number of older Americans who bought when prices were much lower."¹ The windfall concentrates in the expensive coastal markets and among the richest of the older cohorts, a redistribution, in their words, "from buyers" to "sellers."¹

This is no rounding error on the national balance sheet. Rising housing values account for 40 percent of the increase in the American capital-to-output ratio between 1970 and 2010, and housing capital explains 83 percent of the rise in private capital relative to income over the same span.¹ Much of the wealth the country banked over two generations is the same wedge, capitalized. And the rule that produces it is enforced in a specific room: the zoning board and the public hearing, where the people who already hold the asset outlast everyone who would like to buy in.

The workers who cannot move to the wages

Price is not the wedge's only cost. When the most productive cities are the ones that refuse to build, workers who would earn more there cannot afford to move, and the country produces less than it could. Chang-Tai Hsieh and Enrico Moretti built the canonical estimate of that loss: constrain housing in New York, San Francisco, and San Jose, and you misallocate labor across the entire economy.³

Here the honest advocate slows down. The magnitude is disputed among the very people best equipped to compute it. Hsieh and Moretti's published figure is a 3.7 percent gain in national output from deregulating those three metros.³ Brian Greaney, re-running their model, found coding errors and a sensitivity to units that in his account make the true effect "two orders of magnitude smaller."⁴ Bryan Caplan, pressing the other way, argues that a failure to compound growth makes the gain larger, and reports that the authors "graciously confirmed the mistakes."⁵ What survives the crossfire is the direction, not the decimal: supply constraints in high-wage cities misallocate labor. How much they cost is unsettled, and a careful case rests on the sign.

What building looks like when it is legal

The argument would be academic if no one had tried it. Three governments have, and their records are the best evidence available on what ending single-family-only zoning actually does.

Minneapolis went first and furthest. Its 2040 comprehensive plan, effective January 2020, ended single-family-only zoning on roughly 70 percent of the city's residential land and let duplexes and triplexes rise where only detached houses had stood.⁶ The city's rents then rose more slowly than in comparison cities,⁷ and the story hardened into a slogan: build more, and rents stop rising. The slogan runs ahead of the evidence. The Federal Reserve Bank of Minneapolis, studying the outcome in 2025, found that the city permitted only about 2 percent more multifamily housing than comparison cities and carried higher vacancies, and attributed the slow rent growth "more from a negative housing demand shock specific to the Minneapolis market in 2020 than the expected impact of changes in the housing supply."⁷ Minneapolis removed a legal barrier. It has not yet proven that removing the barrier is what held rents down.

Oregon made the same move statewide. House Bill 2001, passed in 2019, required cities to allow duplexes on nearly every residential lot, and in its larger cities to allow triplexes, fourplexes, cottage clusters, and townhouses: the missing middle by right, meaning approved against fixed standards without a discretionary hearing.⁸ Anticipating the capacity objection, the bill set aside $3.5 million for technical help and a formal process for areas where water and sewer genuinely cannot yet carry more.⁸

Portland shows what the Oregon mandate produced. The city permitted more than 1,400 middle-housing and backyard units in its single-family zones between August 2021 and mid-2024, and middle housing became the most common kind of home built in those zones, which supplied 43 percent of the city's permits in the first half of 2024.⁹ Those new units sold for $250,000 to $300,000 less than a new detached house, mostly because they are smaller.⁹ Cheaper homes, built by right, at scale.

California is the cautionary tale. Senate Bill 9, in 2021, let owners split a single-family lot and put up to four units where one had stood, with ministerial approval, meaning a sign-off by staff against objective standards and no hearing.¹⁰ In its first year almost nothing happened. Los Angeles fielded 211 unit applications and approved 38; San Diego, which permitted more than 5,000 homes in 2021 by other means, saw 7 applications and approved none.¹¹ The trickle reflects the statute's design. Owner-occupancy strings, sensitivity to lot size, and competition from California's more permissive backyard-unit law throttled it.¹¹ Oregon wrote a broad by-right mandate and got housing; California wrote a narrow one hedged with conditions and got paperwork.

Where the reform is challenged hardest

The strongest objection is constitutional, not aesthetic. In 2024 five California charter cities, Redondo Beach, Carson, Torrance, Whittier, and Del Mar, persuaded a Los Angeles judge that SB 9 intruded on their authority over "municipal affairs," and won.¹² The court held the state's lot-split mandate "is not reasonably related to the production of affordable housing, interferes with local government, and is therefore unconstitutional," and struck it down for the five.¹³ This is local control at its most serious. Not a neighbor grumbling about a duplex, but a court agreeing that a state overrode a city's right to govern its own land. The reliance interest behind it is real too: people bought into a neighborhood zoned a certain way and expected it to stay that way.

The answer lies in which local public the hearing represents. Jerusalem Demsas calls the American development process one where "whoever yells the loudest and longest wins," and argues that deference to community input is "a big part of why the U.S. is suffering from a nearly 3.8-million-home shortage."¹⁴ The people who fill a zoning hearing to defend the existing rules are the incumbents the wedge already pays. Democratic self-governance describes a town. It does not describe a Tuesday-night hearing run by the handful of owners with the most to protect. And the ruling binds five cities on one statute; no court has found that legalizing housing exceeds a state's power, which is why the Attorney General has appealed.¹²

The second objection is capacity. Roads, sewers, and habitat cannot absorb unlimited density, and the process built to check that, environmental review, is a genuine constraint rather than a pretext. In Minneapolis it stopped the plan cold. A coalition led by Smart Growth Minneapolis argued the 2040 plan would cause "materially adverse environmental effects," and in 2023 a judge enjoined it under the Minnesota Environmental Rights Act, reverting the city to its older plan.¹⁵ The injunction dissolved only because the legislature then amended the law to exempt comprehensive plans from that review. The reform survived by statute, not by winning the environmental argument on its merits.

The answer comes from the same mechanism that built the case. Oregon did not wave capacity away; it funded it, with $3.5 million and a formal infrastructure-deficiency process for places where service genuinely cannot stretch.⁸ The honest form of the capacity worry is an argument for phasing and investment where the pipes are full, not for a permanent ban everywhere the pipes are fine.

The third objection comes from the left, and in the short run it is the likeliest to be right. Yonah Freemark studied Chicago's transit-area upzonings and found something inconvenient. Parcel values rose significantly, yet there were "no impacts of the reforms on the number of newly permitted dwellings over five years."¹⁶ Upzoning capitalized into land prices ahead of any new building, which is the exact channel by which legalization without protection can raise rents and push out tenants before it houses anyone. The Urban Displacement Project, the most careful academic home for this worry, finds the evidence "limited and conflicting" and warns that new market-rate housing "may result in rent increases in lower-priced residential buildings nearby" in the short term.¹⁷

Freemark himself limits his finding to the short-term, parcel-by-parcel case, which is precisely not the reform on the table.¹⁶ Chicago upzoned narrow strips, lot by lot, an invitation to speculate on the lucky parcel. Broad, by-right, citywide legalization works the other way: when every lot can add units, no single lot wins a scarcity lottery. Portland is the tell, where a by-right mandate produced cheaper homes at volume, not merely higher land values.⁹ And the Urban Displacement Project's own conclusion is not "stop building." It is build and protect, pairing new supply with tenant protection and preservation.¹⁷ That is a friendly amendment to the reform, not a refutation of it.

None of these objections reaches the mechanism. Restrict supply and price rises above cost; the wedge flows to incumbents; the room that guards it is not the public. What would move this paper off the position is not another thin year of SB 9 numbers, which the statute's own design explains. It is a city that legalized the missing middle broadly and by right, waited out the land-price adjustment Freemark measured, funded the infrastructure, bypassed the hearing, and still built nothing. That result would mean the constraint was never really permission. The record so far, Portland most of all, points the other way.