Federal Data Are Disappearing as Statistical Agencies Face Budget Cuts and Political Pressure

Federal statistical agencies face increasingly overt politicization by the Trump Administration on top of chronic underinvestment, uncertainty, and recent and proposed budget cuts, all of which threaten the accuracy and availability of vital federal data and public trust in official statistics.

Most recently, the U.S. Department of Agriculture (USDA) announced it was defunding the annual survey on food security, beginning with 2025 data collection.^[1] The elimination of this survey would prevent policymakers and researchers from tracking changes to household food insecurity just as the largest-ever cuts to food assistance through SNAP hit families and as food prices continue to rise, in large part due to tariffs. This action is part of a larger pattern that will reduce the availability of data or discredit data that are available.

Without sufficient funding and freedom from political interference, the federal statistical system as we know it — and our ability to make economic and policy decisions based in reality — are in jeopardy.

This report lays out pressing concerns about our nation’s most important federal data, including issues surrounding:

• disappearing federal data and defunding of the annual food security survey;
• underfunding and politicization of the decennial census;
• proposals to make surveys voluntary and limit follow-up activities;
• loss of expertise and threats to the independence of the Bureau of Labor Statistics and Census Bureau; and
• major proposed cuts to the Survey on Income and Program Participation.

Federal Data Are Disappearing, Including Our Best Measure of Household Food Security

The Trump Administration has taken a variety of actions to stop collecting or restrict access to federal statistical data. Some of these changes are likely attributable to understaffing but others appear to be motivated by politics.

On September 19, USDA announced that it would terminate not only the annual report on food insecurity in the U.S. but also eliminate funding for the underlying data, which are collected every December through the Current Population Survey Food Security Supplement (CPS-FSS). The USDA announcement inaccurately described these data as “redundant, costly, politicized, and extraneous.”^[2] These data are the nation’s single longest-running and most consistent measurement of a key indicator of material hardship, or families’ ability to meet basic needs. No other federal survey representative at the national and state levels provides such detailed information about people’s experiences with food hardship, data policymakers of both parties have relied on for three decades to inform policy decisions. Unless Congress acts, the 2024 data scheduled for release in October 2025 will be the last update and the data scheduled to be collected in December will not be collected.

Other decisions to collect or publish less federal data appear to be driven by the current administration’s views on issues like LGBTQ+ identities and climate change. The National Oceanic and Atmospheric Administration announced in May that it will no longer be updating its Billion Dollar Weather and Climate Disasters product, which tracks major weather and climate disasters that have total damages or costs of $1 billion or more; these data will be published through 2024.^[3] Additionally, Executive Order 14168 led to the removal of questions related to sexual orientation and gender identity in Census products such as the National Crime Victimization Survey and the National Health Interview Survey.^[4]

Some cuts to data collection and availability are likely related to agencies being under-resourced or understaffed. In April and again in July, the Bureau of Labor Statistics (BLS) announced it had suspended collection or reduced the sample for portions of the data used in the Consumer Price Index (CPI) due to a federal hiring freeze (BLS was granted an exemption to begin hiring CPI data collectors again in September).^[5]The CPI informs cost-of-living adjustments for Social Security, food assistance programs, and housing subsidies. Any reduction in data quality could directly affect benefit levels for millions of people across the U.S. In June, BLS also announced that “due to limited resources,” it had suspended access to all but one of its restricted-use datasets through the Federal Statistical Research Data Center program, stalling academic and policy research.^[6]

Mass layoffs threaten data and data-driven guidelines across the federal government. In March, the Department of Health and Human Services (HHS) suspended data collection for the Pregnancy Risk Assessment Monitoring System (PRAMS), a gold-standard database of maternal mortality data, and in April the full PRAMS team was put on administrative leave as part of a sweeping reduction in the federal workforce.^[7] While not a change in the federal statistical system, the Trump Administration in April also laid off all of the analysts at HHS who were responsible for updating the federal poverty guidelines that are used to calculate eligibility for more than 40 programs such as the National School Lunch Program, Low-Income Home Energy Assistance, and parts of Medicaid and Medicare.^[8]

This is not an exhaustive list of data collection that has been reduced or suspended, and emerging stories suggest that these actions are only the earliest of many planned reductions of federal data.^[9]

Proposed Budget Cuts and Politicization Threaten the Decennial Census

The best-known Census Bureau product is the decennial census, which is required by the U.S. Constitution and forms the backbone of the federal data system. It determines congressional apportionment (the allocation of seats in the U.S. House of Representatives among the states) and helped guide the distribution of an estimated $2.8 trillion in federal funding in 2021 (the most recent year for which Census reported these estimates).^[10]

Preparation for the 2030 census is now at risk.^[11] While the President’s fiscal year 2026 budget proposes a $334 million increase from fiscal year 2025 in the category that includes the 2026 Census Test and other preparation for the 2030 census, the Senate Appropriations Committee reported out a Commerce, Justice, Science, and Related Agencies (CJS) bill in July with an increase of only $138 million, falling far short of the need.^[12] Although 2030 is five years away, decennial census research, testing, planning, contracting, and hiring are already on a tight timeline. Without sufficient funding, key preparations could be delayed or eliminated altogether, putting the count’s accuracy at risk.

In August, the President called on the Department of Commerce, which houses the Census Bureau, to “immediately begin work on a new and highly accurate” census that would exclude people without a documented immigration status from the count.^[13] In September, the House Appropriations Committee reported out a CJS bill that includes a provision excluding individuals not lawfully present in the United States when “rendering apportionment determinations in subsequent decennial censuses.^[14] This change would be an unprecedented departure from the way the census has been conducted throughout history and would be met with legal challenges under the 14th Amendment. Further, excluding people without a documented status would require adding a citizenship question to the decennial census, which is likely to suppress response rates and, as a result, severely compromise the quality of this constitutionally mandated count.^[15]

Restricting Follow-Up Actions Would Exacerbate Falling Response Rates

Investment in the statistical infrastructure is increasingly important as surveys worldwide face declining response rates.^[16] Policymakers are proposing measures such as cutting budget and staff, adding a citizenship question, and reducing follow-up actions for non-response that would accelerate the troubling decline in survey response rates, rather than seeking to reverse them by funding thoughtful modernization efforts.

The CJS bill reported out of the House Appropriations Committee in September would make the decennial census and the Census Bureau’s American Community Survey (ACS) voluntary and dramatically limit the Bureau’s ability to conduct follow-up with non-respondents for all surveys.^[17] (While mandatory participation is not enforced, Census tests found that response rates fell when surveys were designated as voluntary rather than mandatory, increasing survey costs and decreasing reliability of estimates).^[18] This would seriously exacerbate risks to data quality from non-response bias. Expert stakeholders and industry leaders have raised concerns to Congress that making Census surveys voluntary will dramatically reduce response rates and reliability of estimates that the private and public sectors depend on.^[19]

A separate proposal in a pair of companion bills, H.R. 4820 and S. 265, would make participation in the ACS voluntary. The ACS provides detailed social, economic, and housing data down to the local level. Like other national surveys, the ACS has faced flat funding in recent years, which has limited sample sizes, made it more difficult to update the questionnaire, and slowed efforts to reverse declining response rates. Making participation in the ACS voluntary would further weaken the reliability of estimates from these data, especially for rural areas and smaller demographic groups.

Bureau Of Labor Statistics and Census Bureau Are Losing Expertise and Face Threats To Independence

Data collected by the BLS allow us to understand the state of our economy by providing information about inflation, employment, wages, and more. BLS data — such as the monthly employment report — are relied upon by the public and private sectors to inform their decision-making.

Yet in August, hours after the release by BLS of large downward revisions for May and June employment data, President Trump instructed staff to terminate BLS Commissioner Erika McEntarfer and made unfounded accusations of data manipulation against BLS. (Revisions are not manipulations, but rather changes that improve accuracy by accounting for delayed responses to surveys.) At the same time, the White House released a blog post attacking the agency’s credibility. As William Beach, a former BLS commissioner appointed by President Trump, observed, “the totally groundless firing [of Commissioner McEntarfer] sets a dangerous precedent and undermines the statistical mission of the Bureau.”^[20]

Statistical agencies need highly qualified and nonpartisan leaders. The next BLS commissioner will be responsible for safeguarding the Bureau’s long-standing commitment to credibility and neutrality, providing deep expertise on economic statistics, and quickly filling BLS senior staff positions, one-third of which are currently vacant.^[21] Economists across the political spectrum have raised serious concerns that E.J. Antoni, the President’s nominee to serve in the now-vacant position of BLS commissioner, lacks these qualifications and instead would play the role of a highly partisan ally.^[22]

The Census Bureau has similarly experienced a “brain drain” from recent reductions in force and high-level departures, including the early departure of previous bureau director Robert Santos.^[23] On September 19, career Census Bureau economist and current Deputy Director Ron Jarmin was replaced as acting Director of the Census Bureau by George Cook, a Trump political appointee. Cook also currently serves as acting Undersecretary for Economic Affairs at the Department of Commerce (a position that oversees the Census Bureau Director) and as chief of staff to the same role. Per his official biography, Cook has no prior government experience and no advanced training in statistical methods.^[24]

The Trump Administration has also disbanded volunteer advisory panels of outside experts, including the 2030 Census Advisory Committee; the Bureau of the Census Scientific Advisory Committee; the Census Bureau National Advisory Committee on Racial, Ethnic, and Other Populations; and the Technical Advisory Committee and the Data Users Advisory Committee at BLS.^[25] Terminating advisory groups isolates these agencies from expertise at a critical juncture for both decennial census preparation and broader survey modernization efforts.

Policymakers Will Lose Critical Information if the Survey of Income and Program Participation (SIPP) Is Cut

SIPP is a Census survey that provides information about a range of economic well-being indicators — including the impact of public benefit programs — and enables policymakers to understand how proposed laws will change eligibility and costs. The President’s fiscal year 2026 budget and the CJS bill reported out by the House Appropriations Committee both propose cutting $40 million from the Census budget category that includes SIPP.^[26] The Administration appears to envision a sharply reduced version of SIPP, relying on links to unspecified administrative records from various other government agencies.

This reduced version of SIPP is unlikely to provide the uniquely rich content (such as month-by-month income data) and structure (such as following children as they move between different caregivers’ homes) that allow the current SIPP to answer policymakers’ questions about families, their needs, and the programs that serve them. For example, policymakers may be interested in how a longer recertification period or different asset limit would affect the cost or the number of people eligible for Medicaid, Supplemental Security Income (SSI), WIC (the Special Supplemental Nutrition Program for Women, Infants and Children), SNAP, or school breakfast and lunch, along with their poverty status. The current SIPP collects the required longitudinal income and other monthly data needed to answer these questions; the reduced version envisioned by the Trump Administration likely would not.

SIPP provides detailed longitudinal data on individuals’ eligibility for and participation in programs like SNAP, SSI, and Temporary Assistance for Needy Families.^[27] It also collects data on select dimensions of material hardship including food hardship, but with far less detail than the CPS-FSS, and the sample is not large enough to generate reliable state-level estimates for all states. SIPP is the only nationally representative survey of its size that tracks how long people receive various government benefits. This information is required for annual reports to Congress under federal law.^[28] Cutting SIPP funding would likely derail recent plans to modernize and improve the survey and weaken its overall usefulness for understanding economic well-being.

Federal Statistical Agencies Provide Critical Infrastructure for Our Economy and Democracy

Each of the proposals detailed above would take federal statistical agencies in the wrong direction. Instead, to continue producing timely, reliable estimates that policymakers and the public can trust while meeting modern data collection challenges, statistical agencies need highly qualified and independent leaders, freedom from political interference, and the resources necessary to fulfill their mission.