Over the past decade, the idea of a nationwide housing shortage has become conventional wisdom. As rents and home prices reached historic highs, policymakers increasingly focused on restrictive zoning and inadequate housing construction as key drivers of rising housing costs. The prescription seemed straightforward: Build more housing.
That diagnosis emerged in high-cost metropolitan areas under extraordinary housing pressure, especially coastal California and parts of the Northeast. As housing costs increased nationwide during and after the COVID-19 pandemic, it became a national issue.
But is every metropolitan area actually short on housing? The answer matters because housing policy should be rooted in local market conditions, not broad assumptions.
To explore the question, we examined housing vacancy levels, or whether homes are available for occupancy, across metropolitan areas and over time.
Our findings challenge the idea of a universal housing shortage. While some metropolitan areas face persistent availability constraints, others have relatively accessible housing markets—even when homebuilding has been limited.
Housing affordability is a nationwide issue, but housing shortages differ by metropolitan area
The United States unquestionably faces a housing affordability crisis. Millions of households struggle to afford housing, homelessness remains high, and some metropolitan areas haven’t built enough homes to meet demand.
But affordability problems don’t stem from the same market conditions everywhere. New York City faces severe supply constraints and has a vacancy rate of just 1.4 percent. Other areas face persistent vacancies, aging housing stock, or neighborhood disinvestment, combined with too-low incomes. In Detroit, local incomes aren’t high enough to afford available housing. Treating these fundamentally different markets as one national housing shortage risks prescribing the wrong remedies.
We examined housing supply and demand trends from 1990 to 2020 in the 107 US metropolitan areas with populations above 500,000 in 2020. We included only tracts in census-classified urban areas in 2020. We combined owner and rental vacancy rates—including dwellings that the US Census Bureau reported as vacant for rent, for sale, or sold or rented and not occupied—into a vacancy-balance index based on long-established estimates of healthy housing-market vacancies. Then, we classified metropolitan areas as relatively tight (needing more homes to meet demand), balanced, or in surplus (having more than enough homes).
Source: Authors’ analysis of US Census Bureau data, 1990–2020.
Notes: CBSA = core-based statistical area. To account for differences in the mix of rental and owner housing across metropolitan areas, we calculate a vacancy balance indicator. It compares actual vacancy with benchmarks of 7 percent for rentals and 1.5 percent for owner housing. We do not include dwellings classified as second homes, farmworker housing, or vacant for some other reason. Values above the benchmark indicate surplus housing; values below it indicate a tight market. Graph shows only the 107 largest metropolitan areas.
We find that, overall, housing markets were more often loose than tight over the past three decades. In every census year from 1990 through 2020, more metropolitan areas showed housing surpluses than shortages. Even in 2020, after a decade of rapid market tightening, 46 metropolitan areas exhibited housing surpluses, while only 36 qualified as tight markets. At the other extreme, by 2010, following years of high construction rates, the foreclosure crisis, and declining incomes, 92 metropolitan areas had housing surpluses, and only three had tight markets.
Source: Authors’ analysis of US Census Bureau data, 1990–2020.
Notes: To account for differences in the mix of rental and owner housing across metropolitan areas, we calculate a vacancy balance indicator. It compares actual vacancy with benchmarks of 7 percent for rentals and 1.5 percent for owner housing. We do not include dwellings classified as second homes, farmworker housing, or vacant for some other reason. Values above the benchmark indicate surplus housing; values below it indicate a tight market. Map shows only the 107 largest metropolitan areas.
Housing scarcity was also highly concentrated over time. California dominated the map of persistent shortages. The San José metropolitan area—which has long been constrained by low housing production despite high demand from the Silicon Valley economy—was the only market that remained tight throughout the study period. But by 2020, every California metropolitan area in our sample had become tight, including inland markets like Modesto, which had substantial surpluses a decade earlier.
Elsewhere, shortages appeared repeatedly in parts of the Northeast and Pacific Northwest but were less persistent, with markets moving between shortages and surpluses over time.
Persistent housing surplus followed a very different pattern. Florida stands out most clearly: Cape Coral, Miami, and Tampa, for instance, all remained in surplus throughout the 30-year study period. Many other fast-growing Sunbelt metropolitan areas—including Charlotte, Houston, and Phoenix—also consistently maintained surpluses. Only a handful of older industrial metropolitan areas, like Indianapolis, St. Louis, and Wichita, remained consistently loose.
Overall, the relationship between housing growth and vacancy balance has been inconsistent. Metropolitan areas like Charleston added housing units at a fast clip between 2010 and 2020, and kept relatively high vacancy rates by 2020. Provo-Orem, Utah, added homes even more quickly over that time, but experienced a very tight housing market by 2020. Both the Modesto and St. Louis metropolitan areas added few housing units but had dramatically different outcomes in terms of housing vacancy.
Source: Authors’ analysis of US Census Bureau data, 1990–2020.
Notes: To account for differences in the mix of rental and owner housing across metropolitan areas, we calculate a vacancy balance indicator. It compares actual vacancy with benchmarks of 7 percent for rentals and 1.5 percent for owner housing. We do not include dwellings classified as second homes, farmworker housing, or vacant for some other reason. Values above the benchmark indicate surplus housing; values below it indicate a tight market. Graph shows only the 107 largest metropolitan areas.
Housing markets can become constrained for many reasons. In some cases, adding supply can ease demand pressures and improve availability. In others, local incomes and family sizes may be just as influential.
Moreover, vacancy is just one measure of housing supply. Understanding a region’s housing needs requires evaluating housing costs, particularly for those with low incomes, across neighborhoods. It also requires considering whether housing costs enable people to move.
Stakeholders should tailor housing policy to local market conditions
These findings suggest that policymakers should distinguish between a national housing affordability crisis and a national housing shortage. Where markets are genuinely constrained, housing production and zoning reform are essential to ensuring adequate supply across locations and housing types.
But in metropolitan areas with persistent surpluses, challenges can stem from low incomes, deteriorating housing quality, neighborhood decline, discrimination, or poor access to jobs and opportunity. In those areas, policymakers should focus on housing availability and quality.
In all markets, zoning reforms add value, though not always as a mechanism to boost supply. Zoning reforms can expand housing choice, support fair housing, encourage walkable neighborhoods, facilitate adaptive reuse, and reduce unnecessary barriers to redevelopment. These are important goals regardless of metropolitan context.
The findings also suggest caution toward policies that assume every community primarily needs more housing production to lower costs. Production incentives may be effective where housing shortages are demonstrable, but they are less likely to address weak demand, obsolete housing, or neighborhood disinvestment.
The United States does not have a single housing market. This reality doesn’t diminish the severity of the housing crisis. Instead, it underscores its multiple causes and the need for multiple solutions.
