Few economic ideas seem more intuitive than this: if new housing costs $400,000 to build, then building more expensive housing cannot help families who need affordable housing. Markets build for the wealthy and leave everyone else behind.
The economics tell a different story — one that is well-documented, empirically validated, and almost universally misunderstood by policymakers. The mechanism is called filtering. It may be the most powerful natural force for housing affordability in existence. And governments across the United States are systematically destroying it.
What Filtering Is
Filtering is the process by which housing units age out of the high-income market and gradually become accessible to lower-income households. The pre-war apartment building that once housed professionals is today a working-class rental. The 1970s ranch house marketed to middle managers is today entry-level for first-time buyers. The 1990s luxury condo now attracts households who could never have afforded its original price.
The mechanism operates through what economists call a vacancy chain. New units attract high-income occupants — the only households that can absorb new construction costs. Those households vacate their previous homes, which are occupied by slightly less affluent households, who in turn vacate still-older units, and so on down the income ladder. Every new unit at the top sends a ripple of available housing downward through the entire stock.
The empirical evidence is robust. Economist Stuart Rosenthal, in a landmark 2014 study in the American Economic Review, estimated annual filtering rates using panel data from 1985 to 2011. Rental housing filters downward at roughly 2.5 percent per year in real terms. A unit serving a median-income household tends to become affordable to a 40th-percentile household within 20 years — not by design, but because the market adjusted.
"While filtering has long been considered the primary mechanism by which markets supply low-income housing, direct estimates of that process have been absent. This has contributed to doubts about the viability of markets and to misplaced policy." — Stuart S. Rosenthal, American Economic Review, 2014
Rosenthal's note about "misplaced policy" is not incidental. The absence of direct empirical estimates allowed policymakers to dismiss filtering as theory. His data demonstrated it as fact.
The Supply Constraint That Kills Filtering
Filtering works only when there is enough new supply to set the chain in motion. The vacancy cascade requires a continuous injection of new units at the top of the market to free up older units for everyone below. That injection has not been happening at the scale the market requires.
Zillow estimated in June 2024 that the U.S. housing shortage had grown to 4.5 million homes — up from 4.3 million the year before. In 2022, the number of U.S. families increased by 1.8 million while only 1.4 million housing units were completed nationwide. Supply has chronically trailed demand for over a decade.
Census Bureau New Residential Construction data tells the story in stark terms. The U.S. completed over 2 million housing units annually in the mid-2000s. After the financial crisis, construction collapsed — and while it has partially recovered, the pace has never returned to levels sufficient to satisfy demand from a growing population. The Census Bureau's Characteristics of New Housing report shows approximately 1,627,000 total units completed in 2024 — 1,019,000 single-family and 608,000 multifamily. That pace barely keeps up with household formation, let alone closes the existing 4.5 million-unit deficit.
When the supply chain is throttled, the filtering mechanism stalls. There are no new units at the top to free up units at the middle. Units at every income level remain more expensive than they would be in a functioning market. The housing ladder remains structurally intact but has had its rungs removed.
Who Throttles the Supply
Government regulation imposes enormous costs at every stage of development. A 2021 analysis by the National Association of Home Builders found that regulatory requirements account for $93,870 of the average new home's price — roughly 24 percent of total cost. On the multifamily side, the National Multifamily Housing Council reported in 2022 that regulatory costs account for 40.6 percent of total apartment development costs. For every dollar spent building a new apartment, forty cents pays for government requirements with no market counterpart.
These costs compress new construction from both sides. Developers face higher per-unit costs, reducing projects that pencil out financially. Those that proceed must target the high end, where revenues absorb the regulatory premium. The moderate-income tier — exactly the middle rungs that filtering would otherwise supply — goes underbuilt. The NAHB has documented how "common zoning barriers — including minimum lot sizes, height restrictions and parking requirements — often minimize supply and hinder development."
Evidence From Markets Where Supply Was Freed
The theory is most legible when cities actually allow supply to grow. Minneapolis offers the clearest recent evidence in the United States. When the city adopted its 2040 Plan — eliminating single-family zoning citywide, reducing parking requirements, and streamlining housing approval — construction responded. The NAHB documented a 45 percent increase in permits for 2- to 4-unit buildings from 2020 to 2022, driven in significant part by the removal of parking mandates. The Minneapolis Federal Reserve examined the results in 2025, finding that Minneapolis rent growth has been subdued relative to peer cities — though the Fed noted difficulty disentangling the supply effects of the 2040 Plan from simultaneous demand-side factors. The supply response itself, however, is unambiguous.
The inverse is equally instructive. San Francisco, with some of the most restrictive housing regulations in the country, has seen rents rise despite persistently low vacancy rates. When the top of the market is locked by regulatory barriers, the vacancy chain never starts. Older stock that would filter downward becomes the only vehicle for demand at all income levels — prices for older housing rise rather than fall.
The Policy Implication Government Keeps Ignoring
The policy implication of filtering theory is uncomfortable: the most cost-effective affordable housing program is abundant market-rate construction. Not subsidized construction. Not voucher programs. Not inclusionary zoning mandates. Unencumbered market-rate supply at scale.
Government housing programs — Section 8 vouchers, Low Income Housing Tax Credits, public housing — each attempt to replicate administratively what filtering accomplishes automatically. They are expensive, administratively burdensome, and reach a fraction of households affected by supply constraints. The LIHTC program produces roughly 100,000 subsidized units per year at taxpayer expense. The housing deficit sits at 4.5 million units.
Thomas Sowell's insight applies directly: policy shaped by the "vision of the anointed" dismisses solutions that require no visible intervention. Allowing markets to build abundantly generates no ribbon-cutting ceremony, no political constituency, no funded program. It requires only that government stop preventing the natural mechanism from operating.
Friedrich Hayek argued that the price system carries information no central planner can replicate. In housing, that information flows through rents, vacancies, and filtering. When zoning codes, compliance costs, and administrative barriers suppress new construction, they jam the signal that would otherwise guide investment down through the income ladder. The market's housing ladder produces affordable homes naturally. Policymakers don't have to build anything. They only have to stop preventing the market from doing it for them.
