A special police member monitors a protest, while inside the Consumer Financial Protection Bureau building, the day after members of Elon Musk's Department of Government Efficiency (DOGE) moved into the CFPB, in Washington, on Feb. 8.Nathan Howard/Reuters
In recent years, economists have raised concerns about companies and consumers shunning surveys used to produce vital information about business conditions in the United States.
But in the early days of a new Trump administration, their worries have shifted to something more pressing: Trusted data is disappearing.
Soon after the inauguration of U.S. President Donald Trump last month, scores of government websites and datasets – including those of the U.S. Census Bureau and the Centers for Disease Control and Prevention (CDC) – went offline, ostensibly to ensure they complied with executive orders on gender, diversity and other matters.
The U.S. government is also slashing costs, which has brought an abrupt cancellation of some studies. And the nascent Department of Government Efficiency (DOGE), led by Elon Musk, has access to sensitive records in the Labor Department, which is responsible for a bounty of key economic releases that move financial markets.
To date, there are no reported signs of data manipulation, and many of the affected websites have returned. Major economic reports have also been published as scheduled.
But researchers say this period of volatility has undermined trust in U.S. statistical agencies and raised broader concerns about political interference.
“There’s a chilling effect on a lot of things,” said Beth Jarosz, senior program director at the Population Reference Bureau, a non-profit research organization based in Washington “People who do public health, school administrators, business owners and hedge fund managers all rely on this data.
“And even if it’s just temporary, not being able to get what they need creates uncertainty and reduces trust in the federal statistics and the information we all need.”
For example, when the government websites began to disappear, microdata for the American Community Survey – which offers detailed information about demographics and economic characteristics such as income and poverty status at the local level – were temporarily unavailable.
The U.S. government has said little about the outages, although the CDC website said it was “being modified to comply with President Trump’s executive orders.” The U.S. Office of Personnel Management, which manages the federal civil service, instructed government departments and agencies to remove communications that “inculcate or promote gender ideology,” among other directives.
The disappearance of data has wide-ranging effects. Governments produce reports that the private sector can’t replicate. And these statistics inform everything from academic research to corporate decision-making.
“I think of public data as information infrastructure,” said Erica Groshen, a former commissioner of the U.S. Bureau of Labor Statistics (BLS), which is part of the Labor Department and produces market-moving reports on inflation and employment. “The federal government can provide data that the private sector can’t and won’t.”
In order to produce their reports, U.S. statistical agencies frequently rely on the participation of households and businesses to fill out surveys with private information. Ms. Groshen said she’s concerned that response rates – which are already falling – could weaken further if people question the confidentiality of what they’re providing, particularly as DOGE gains broad access to the federal government.
Statistical agencies in the U.S. and elsewhere have struggled with weaker survey participation for many years. In one notable example, only about one-third of businesses approached to fill out the BLS’s Job Openings and Labor Turnover Survey do so – about half the proportion in the 2010s.
The BLS and other agencies contend that data quality remains high, although critics point to non-response bias – the idea that non-respondents may be inherently different than those who continue to fill out questionnaires, which would skew the numbers.
If response rates continue to fall, there is a greater likelihood that economic data will become less reliable. The danger is that reports “will stop telling us about who’s doing well and who’s not well by any degree of disaggregation,” said Armine Yalnizyan, a Canadian economist and Atkinson fellow on the future of workers.
Funding is another concern, particularly as the Trump administration makes sweeping cuts. These include the termination of roughly US$900-million in Education Department contracts, spelling an end to various research projects on academic performance.
When data disappear or become less reliable, it becomes tougher to challenge the policies of the Trump White House, Ms. Yalnizyan said. “You can’t see what is really happening, so you cannot dispute what they say.”
Ms. Jarosz said the public has paid for data produced by the government – and that information should remain in the public domain.
“I think part of what is so concerning about this is it sets a really dangerous precedent that any administration could delete data they don’t like for any reason,” she said.