Local Governments Could Be Flying Blind as Federal Data Disappears

State and local governments have long understood they’d have to adjust to shifting federal priorities under the Trump administration. But many did not recognize they might have to adapt to a world where it’s harder to access crucial data.

Recent changes in federal data collection and distribution will make it harder for state and local governments to plan for things like disease outbreaks, disasters, crime prevention, transportation and housing policy, says Brian Castrucci, the president and CEO of the de Beaumont Foundation, a public health nonprofit.

“If you walked on an airplane and they had all the instruments blacked out, you would walk off that plane,” Castrucci says.

The changes have formed part of broader cuts during the Trump administration. As the Department of Government Efficiency went from federal department to department, it tended to view any data collection not clearly mandated by Congress as something that could be cut, says Amy O’Hara, executive director of the Georgetown Federal Statistical Research Data Center.

It will take time to know how much has gone missing from the more than 2,000 top-level federal government domains and the sub-sites and microsites under them. A February survey by the New York Times looked at just a dozen of these and found 8,000 pages had already been taken down.

Funding decisions to reduce support for staff and programs that collect and analyze data are another concern — the president has already proposed cutting all funding for research by the National Oceanic and Atmospheric Administration that could help governments understand changes in weather and atmospheric conditions.

“My assessment is that all federal data are now at risk,” says Denice Ross, U.S. chief data scientist during the Biden administration.

But, Ross says, the loss of experienced staff to collect and analyze that data will be even more destructive in the long run.

A Spectrum of Data
Even a routine government process like community planning draws on data, including land use patterns, demographics, environmental indicators and socioeconomic factors. Public health officials need a similarly wide range of data, in addition to health and illness trends, to stay ahead of emerging risks.

“It could be any number of a thousand things,” Castrucci says. “The thing you have to remember is that a number is somewhat meaningless without the context of other numbers.”

The current administration has taken particular aim at data it views as connected to diversity, equity and inclusion, or to environmental issues. A number of federal health agencies removed data sets from government websites earlier this year in attempts to comply with the president’s directives.

Those included resources like the Social Vulnerability Index — a database and mapping tool created to help planners and public health officials identify those at greater risk during emergencies. Other removed data sets include guidance on HIV treatment and the Youth Risk Behavior Surveillance System about student health. Experts warn that eliminating data about LGBTQ people is crucial for the work of clinicians, researchers and policymakers — a report from the University of California, Los Angeles School of Law suggests that eliminating data on this population could set their work back by decades.

The federal government also deleted a mapping and screening tool called EJSCREEN, which allowed local governments to assess the relationship between demographics and environmental risk factors. Without that tool, it’s harder for practitioners to know exactly where they might need to direct resources for disaster preparedness, says Gretchen Gehirke, co-founder of the Environmental Data and Governance Initiative. Local governments don’t have the resources to create their own version of this tool.

Congressionally mandated national climate assessments, which have guided policymakers for decades, are no longer available and have been discontinued. In 2023, the U.S. Department of Transportation described the newly published fifth report in the series as “the most comprehensive analysis of the state of climate change in the United States, providing authoritative, decision-relevant information.”

There are hundreds of millions of federal government webpages, Gehirke has learned. Identifying everything that has changed is a huge undertaking.

“There are definitely other databases that have been removed,” she says. “In the environmental space, what has been more common than removing the primary data sources has been removing the tools that the public uses to access that data.”

Weakening Trust
The response rate to household surveys from the Census Bureau has been declining over time, O’Hara says. These include resources that states and cities rely upon, such as the American Community Survey, the Consumer Expenditure Survey and the National Crime Victimization Survey.

Faltering participation is correlated with reduced trust. “If people lose trust in government, they have even less incentive to voluntarily provide information whenever they get a survey request,” O’Hara says.

To the extent that federal data collection becomes spotty or inconsistent, trust could decline further. The Trump administration has sown doubt itself, firing the commissioner of the Bureau of Labor Statistics over updates to monthly job reports that showed weaker job growth this summer. Trump accused the commissioner of rigging the data with no evidence for his claims. The president has said he wants to stop issuing monthly jobs reports, which his pick to replace the BLS commissioner has publicly endorsed.

Washington is also signaling distrust of outside voices. A June Federal Register notice announced the termination of 14 committees that had provided expertise and insights to federal agencies, including data strategies. Bodies addressing economic, census, energy, manufacturing and marine fishery issues were among the casualties.

No federal statistician was allowed to attend the recent annual convention of the American Statistical Association, says O’Hara, a first.

Preservation
Data generated by federal government programs and funded by taxpayers is in the public domain. A number of groups have taken on the task of archiving data sets they consider to be important.

Gehirke’s group is monitoring federal government web pages containing environmental information for changes, sharing what it is tracking and reporting on what it is seeing. It is partnering with the Internet Archive, a nonprofit library whose Wayback Machine lets users search the histories of nearly a trillion web pages, including federal websites and pdfs posted on them. (The archive includes government data of all sorts.)

Public Environmental Data Partners, a network with dozens of organizational and individual members, is archiving federal environmental data, including data sets that are still available on government sites. They also offer screening and mapping tools. (Additional examples below.)

A law Trump signed during his first term requires agencies to develop and share data to support policy decisions. “There’s a section of that that requires that data leaders across federal agencies engage with public users of the data, including their colleagues in state and local government,” Ross says. This opens the door for government users to give feedback regarding how they use data and why it’s important to their planning.

Ross is soliciting use case examples and posting them online. She notes that the Department of Homeland Security recently asked federal and state users of a geospatial data tool designed to support public safety and emergency response to provide use cases to help improve it.

As things shift upstream, the downstream need remains the same. Demographics, health indicators, economic conditions, weather patterns and environmental conditions are in constant flux. Local leaders need data that accurately reflects present realities as well as what might lie ahead when they make big decisions, says Beam Furr, the mayor of Broward County, Fla.

“It’s not possible to affect people to the most optimum levels unless I have good information,” he says. “It’s really as simple as that.”