Lacking data policy is more than a research problem, it's a government performance problem

Effective democracies require modern data policies that direct the country’s ability to govern and hold legitimacy. Data policies offer information justice by providing formal guidelines and rules for the collection, management, protection, and disposal of public information — quantitative and qualitative. They can also bolster trust by setting oversight of information quality — ensuring accuracy and accountability. This policy domain of “data” is understudied, undervalued and underemphasized, despite the accelerated use of data against a backdrop of growing public mistrust.

The federal statistical system has long operated as background infrastructure, as something agency leaders and program managers relied on without much thought, the way they rely on relevant domains like electricity or broadband. Data from the Census Bureau, the Bureau of Labor Statistics, the Centers for Disease Control, the National Center for Education Statistics, and their counterparts across the government showed up when needed most in democracy: to justify budget requests, measure program outcomes, allocate resources and demonstrate results to oversight bodies and the public.

That infrastructure is now under serious strain with pain being felt inside and outside government, at federal, state and local levels. Earlier this year, the social science research company SSRS conducted a survey of more than 500 users of federal statistical data across academia, nonprofit organizations, state and local governments and the private sector. The core question was "have changes to the federal statistical system impacted your ability to do your work?" The findings were resounding, with 93% of respondents reporting that changes to the federal statistical system since the start of 2025 have damaged their ability to do their work. It is quite likely staff inside the government are experiencing similar difficulties.

These respondents describe specific challenges. They report datasets have been taken offline without notice. Expected data publications have been delayed or canceled. Restricted-use data approvals that once took days are now taking months, if they happen at all. Staffing cuts at federal statistical agencies have hollowed out the technical assistance functions that helped users work with complex federal datasets. These stories are not just about citizens, they also represent real breakdowns in the informational infrastructure that federal agencies themselves also depend on to manage programs, respond to Congress and serve the public. These are not failings of politics, they are failings of policy.

For federal leaders, the implications of not having earnest data policy cut in several directions.

The data crisis is a workforce crisis

Respondents to the survey were consistent on this point: the most immediate source of their difficulties is not missing datasets but missing people. When experienced staff depart, they take with them decades of applied knowledge about how data collections work, where the edge cases are and how to help users get reliable answers from complex products. One respondent described the situation plainly: institutional knowledge has been forever lost, and the data user community will suffer in ways that may not be fully known or even detectable. Agency leaders managing through the current period of workforce reduction should treat the departure of statistical expertise as a long-term operational risk, not just a position that can be filled in the next appropriation cycle.

Broken time series are not self-correcting

One of the most consequential findings from the survey involves longitudinal data, the long-running collections that allow agencies, researchers and policymakers to track trends over time. Respondents were clear about what interrupted collections mean in practice. A gap in a time series is not simply a missing data point, it undermines the comparability of all subsequent data, potentially for years. The government shutdown in fall 2025 produced the first interruption to the Current Population Survey in more than 70 years. The October 2025 Consumer Price Index was lost entirely. These are not recoverable losses. They are permanent holes in the historical record, with downstream consequences for economic forecasting, policy evaluation and program management that will persist long after the immediate disruptions are resolved.

There are no adequate substitutes

Public data users in this survey are doing their best to adapt. They are delaying programs until data is available, turning to older archived datasets or private sector sources, leveraging state and locally collected data or building statistical models as workarounds. But they are nearly unanimous that none of these options adequately replace timely, high-quality federal data.

Trust, once lost, is hard to rebuild

More than 70% of these public data users express concern about the long-term erosion of public trust in federal statistics. This is not a soft concern. Trust in official statistics is the foundation on which their practical utility rests. When state and local government planners, nonprofit service providers or academic researchers begin to doubt the integrity or continuity of federal data, they stop building systems that depend on it. The downstream effects on evidence-based management inside agencies are real and compounding.

What would it take to stabilize the situation? Beyond appropriations decisions, effective data policy should direct agencies to preserve statistical expertise as national security and workforce retention priorities, not afterthoughts. Where staff departures are unavoidable, structured efforts to document institutional knowledge, how data collections work, where errors arise and what common user needs exist, could help preserve at least some of what would otherwise be lost.

These public data users are instructive for data policy, they tell us that the federal statistical system is not only a research amenity, it is operational infrastructure. This is equally true for the federal workforce in general. When the data system degrades, government's capacity to know what is working, who needs help and whether resources are reaching their intended targets degrades with it — affecting how ordinary people experience democracy and their government. Rebuilding that capacity will require sustained attention from federal executives who recognize the need for serious data policy that modernizes America’s data infrastructure and take steps to safeguard it.

David C. Wilson is dean and professor at the Goldman School of Public Policy, UC Berkeley, and Chris Jackson is senior vice president at SSRS.