Bad housing data? UC Irvine is on it

Nicholas J. Marantz, associate professor of urban planning and public policy at UC Irvine, is investigating how effectively current data sources track changes in residential housing stock. His aim is to understand how policy changes, such as new zoning laws and broader housing market forces influence the availability and creation of homes, particularly affordable homes.

The research, supported by a grant from Arnold Ventures, aims to evaluate the quality and reliability of existing information tools used to monitor housing units — from single-family homes to larger residential developments — at the neighborhood level. By identifying the strengths and limitations of these sources, Marantz hopes to give planners, policymakers and researchers better tools to understand how housing supply shifts over time, even within small, specific communities.

Marantz recently answered questions about his research.

Why is it so difficult to accurately track changes in housing stock, especially at the neighborhood level? What gaps or failures in current data sources inspired this research?

“Comprehensive, national data on new construction simply isn’t collected at all. The comprehensive national data we do have covers building permits (rather than housing starts or construction), and it is aggregated to the municipal-level. So, it doesn't tell us how many units have actually been built, and it doesn't tell us anything about what's going on in specific neighborhoods. County property assessors collect data that helps us to estimate new development, but those datasets are often incomplete in important ways. For example, they frequently don't include the number of units in apartment buildings. In addition, there are over 3,000 counties in the U.S., and not all of them make their data readily available. Even among the ones that do, the quality and format of the data varies substantially. The Census Bureau provides annual counts of housing units at the neighborhood level, but those counts don't enable us to distinguish between housing types (single-family vs apartments) or housing tenures (rental vs. ownership). In addition, they don't enable us to determine what portion of a given change is due to demolitions as distinguished from new construction, and this distinction matters a lot for policy.”

How could unreliable housing data lead to bad policy decisions — and what are the real-world consequences for communities?

“If you can’t measure what’s actually happening, you can't tell whether your policies are working. Consider California’s reforms to make accessory dwelling units easier to build. Federal building permit data, which excludes units added inside of existing structures, makes those reforms look far less productive than state-level data that captures a broader range of construction activity. If policymakers or researchers rely on the wrong source, they risk abandoning effective reforms or doubling down on ineffective ones.”

Why does it matter that we can track housing changes at the local, neighborhood level rather than just regionally or statewide? Can you give an example of how a zoning change might affect housing supply in ways that current data tools simply can’t capture?

“California’s housing strategy involves lots of different reforms that we'd expect to have different outcomes. To cite just a few examples, there are laws opening commercially zoned land to residential development, ADU policies that create additional units within existing structures, and laws intended to facilitate development in areas where more housing would mitigate greenhouse gas emissions. Without granular data identifying outcomes on specific parcels, there's no way to know whether a given law is driving production on the sites it targets. Policymakers trying to evaluate these reforms with the available data risk drawing the wrong conclusions about what's working and where. In California, the state has tried to remedy this limitation by asking municipalities to submit parcel-level data to the state's Department of Housing & Community Development, but we've identified a number of reporting inconsistencies among jurisdictions, and in many cases the data are incomplete.”

How could better data tools change the way policymakers respond?

“California has been at the forefront of state-level housing reform. It requires local governments to plan for and report on housing production, and it has passed sweeping legislation on ADUs, density, and streamlined approvals. But the data collected by the state are incomplete and sometimes inaccurate. Better measurement tools would facilitate evaluation of the myriad housing laws it has adopted over the past decade. It would also help the state to move from a compliance-based framework, based on whether cities have zoned for enough units, to an outcomes-based one, assessing whether (and where) housing actually was built.”

What would a truly reliable housing tracking system look like, and how far are we from having one?

“The ideal system would integrate information on new construction, demolitions, conversions, and changes to existing structures into a single, regularly updated dataset that's available at fine geographic scales. It would be timely enough to provide feedback on policy changes within a year or two, not a decade. And it would be consistent across jurisdictions so you could make valid comparisons. We're further from that than most people realize, but the pieces are starting to come together. The housing unit counts from the Census Bureau's Master Address File are a step in the right direction. California's reporting requirements, despite their flaws, represent a model other states could build on. The real challenge is institutional. We'd need sustained investment in data infrastructure at the federal level and better quality control at the state and local level. We've historically treated housing data as a byproduct of other statistical programs rather than as a priority in its own right, and that needs to change.”

Who do you hope will ultimately use the findings from this research — city planners, state legislators, housing advocates?

“All of the above. City planners need to know whether their zoning reforms are producing results. State legislators need reliable benchmarks to evaluate whether their policies are having the intended effects or being circumvented. Housing advocates need evidence to hold state and local governments accountable. I'd also include researchers in that list. There's been an explosion of state-level housing policy experimentation over the past several years, with over half of all states adopting some form of housing production policy between 2017 and 2025. The research community has an opportunity to learn from that variation in ways that could inform policy, but only if the data are up to the task.”

Marantz hopes his research will set off a chain reaction of improvements: federal data programs that capture the full range of housing stock changes at a finer geographic scale; state-level investment in data infrastructure with genuine quality controls rather than mandates that simply hope for local compliance; and a cultural shift in how the country values housing information.

“We track jobs, inflation and disease outbreaks with relative precision and speed,” he says. “Housing, which shapes nearly every dimension of people's lives, deserves the same rigor.”
— Mimi Ko Cruz