This piece originally appeared in The American Prospect. Read the original here.
On January 8, 2026, Donald Trump broke a foundational norm of American politics and barely anyone noticed. The political press corps, quite reasonably preoccupied with the administration’s numerous other entanglements, mostly turned a blind eye when Trump posted unreleased economic data to his Truth Social account the night before it was scheduled to be released.
Beyond the fire hose of news in the Trump era, the story also flew under the radar because it was a graph of cumulative jobs numbers over the course of 2025, featuring both the December figure and revisions for some previous months. But it was still a severe violation of the principles that have hitherto underpinned federal data, above all that it must be published in a consistent and neutral fashion so that nobody will be able to commit insider trading with data the taxpayer paid to collect. Federal data, in other words, should be a public good, not fodder for the president’s boasts or his cronies’ portfolios.
A former Bureau of Labor Statistics employee with knowledge of the incident, who was granted anonymity to speak candidly, described the statistics agency’s handling of the imbroglio to me. By their account, after the improper release, the acting commissioner of BLS called the acting chair of the Council of Economic Advisers, who shares the prerelease data along with analysis to the president, and threatened to revoke the White House’s early access if Trump flouted data release procedure again. The two eventually came to an understanding that the issue arose because only the first page of the prerelease materials bore an embargo label, and so the president assumed that all subsequent pages were available to publicize. Going forward, the acting CEA chair and acting BLS commissioner agreed to include an embargo note on every page to prevent confusion. (BLS did not respond to a request for comment.)
Erica Groshen, a senior economic adviser at Cornell University’s School of Industrial and Labor Relations, who also led BLS from 2013 to 2017, said this sort of slipup is not uncommon for new appointees with no experience handling sensitive federal data. Usually, it results in a “slap on the wrist” and it doesn’t happen again.
But since Nixon, offenders have not included presidents, and this was not Trump’s first offense either. In 2018, during his first term, Trump posted on Twitter that he was excited to see employment data about an hour before release.
Neither instance wound up being significant in terms of moving financial markets. But both spooked professional economists, financiers, and other statistics nerds. Over the course of the past two months for The American Prospect and the Revolving Door Project, I talked to a number of experts, including economists, former staff at the Council of Economic Advisers and the National Economic Council, consumer protection advocates, and others, about the implications of Trump’s improper release of economic data.
January’s employment data dispersal also wasn’t the last time Trump used not-yet-public federal data for his own crass purposes; on February 20, Trump leveraged disappointing GDP numbers to blast Democrats over the 2025 shutdown about half an hour before those numbers were released. OMB regulation requires that the White House and other political appointees not weigh in on the data until half an hour after release.
As Corey Frayer, director of investor protection at the Consumer Federation of America and a former Securities and Exchange Commission regulator, put it, “quantitative data is one of the last bastions of objective truth” that serves to “ground us in a shared reality.” That bastion is looking shaky; according to data from the American Statistical Association, collected by NORC at the University of Chicago, public trust in federal statistical data has fallen markedly since last summer. In June of 2025, an estimated 57 percent of Americans trusted federal statistics, falling to 50 percent as of February 2026, the latest month currently available. While the data fluctuates month to month and the dataset only goes back nine months, there is a clear downward trajectory; every demographic category except Hispanic Americans (which was at the same level in June 2025 and February 2026) is lower than it was last summer.
President Trump is infamous for helping to usher in a “post-truth” politics, through inveighing against “fake news” and bumbling through so many different lies, mistruths, deceptions, and piles of sheer nonsense, so quickly, that determining which (if any) of his statements are true is a full-time job. But even as public confidence in the media has declined and polarization over the very nature of reality proliferated, there have been some holdouts of truth that have weathered the frontal assault on objective reality. Federal data was one of those. Now, statistical agencies’ hard-won legitimacy and trust is being eroded, leaving us less able to make informed business and policy decisions and undermining one of our final shared realities.
It’s easy to lose sight of just how much data collection is done by the federal government: public-health statistics, labor market measures, agricultural production levels, macroeconomic indicators, education enrollment figures, and much more. Few people know that there is a chief statistician of the United States (CSOTUS?) who is responsible for coordinating the sprawling, byzantine federal statistical system—comprising 16 agencies whose primary mission is the sampling for and distribution of statistical data and over 100 total agencies whose ambits include statistical work.
The Office of Management and Budget maintains a list of roughly three dozen (at the end of the Obama administration, there were 38; in the current year’s release schedule, there are 35) of the most important, market-moving datasets. Designated as “principal federal economic indicators,” these are the data that are most foundational to running the economy, including agricultural indicators, construction starts, GDP, trade figures, job creations, and inflation rates.
When it comes to Trump’s data impropriety, “some of it is impulse control, [but] a lot of it is malicious,” according to Groundwork Collaborative Chief of Policy and Advocacy Alex Jacquez, who previously was an economic adviser on President Biden’s NEC. This administration’s attack on independent data is a rollback of half a century’s worth of work to ensure a professional statistical system that is both accurate and apolitical. It is, to borrow a Biden-era buzzword, a “whole-of-government” approach to eviscerating the foundation of our information ecosystem.
The modern information ecosystem surrounding federal economic data traces back to 1971, when then-Secretary of Labor James Hodgson and the Nixon White House tried to put their thumb on the scale in how the employment data was characterized in Bureau of Labor Statistics releases. First in January 1971 and again in March, Hodgson gave rosier readouts of jobs numbers than BLS itself. Hodgson then canceled the independent BLS press briefings in March. The Nixon administration continued to insert itself in BLS data dissemination well into that summer, with recordings on the Nixon tapes of the president calling BLS staff “little bastards” and saying “we just have to get rid of him [then-BLS commissioner Geoffrey Moore] and put a political person in.” This episode, detailed by the late former commissioner Janet Norwood, set off a congressional investigation that ultimately led to the implementation of OMB Statistical Policy Directive No. 3, which centralizes and depoliticizes the compilation and distribution of federal statistical data. The directive was further strengthened in 1985 with additional restrictions on providing access to the data prior to release to the White House via the Council of Economic Advisers. (Nixon’s meddling in BLS also included tasking his aide and private equity mogul Fred Malek to create a list of the “Jewish cabal” in the agency.)
BLS is far from the only important statistical agency in the federal government—the Department of Agriculture and the Department of Commerce both have multiple statistics units that are responsible for some of the principal federal economic indicators—but the employment and inflation data it manages is probably the most well known and resonant in our day-to-day lives. It also has among the most storied legacies, dating back to 1884 and operating as its own government department for more than a decade before it was rolled into the new Department of Commerce and Labor in 1903. What’s more, the agency embodies the Nixon-era inflection point that began the modern era of our information ecosystems.
The agency is also widely known for its professionalism and accuracy. Even as the Trump administration has gutted many statistical measures and demonstrated a pattern of manipulating or quashing data it does not like, everyone I spoke to remained confident in BLS. One reason for this, according to Mike Konczal, senior director of policy and research at the Economic Security Project and an alum of the Biden NEC, is that “a lot of financial markets run on data from BLS. If it were disrupted or politicized, hedge funds would call the White House” to complain, whereas, when it comes to a lot of the data that has been attacked more aggressively, like “a lot of the research at NIH or elsewhere that is trying to help vulnerable people, those communities have a lot fewer phone lines into the White House.”
Wall Street’s influence, in other words, is protecting the integrity of economic data critical to their business model. The same cannot be said for those fighting addiction now having to go without the Drug Abuse Warning Network data on emergency drug abuse–related medical treatment, people fighting hunger without the Department of Agriculture’s Household Food Security Report, pregnant mothers and their advocates who can no longer rely on the CDC’s Pregnancy Risk Assessment Monitoring System, farmworkers who cannot look to the Agricultural Labor Survey, or educational professionals unable to turn to the Institute of Education Sciences.
In particular, the administration has shredded datasets that provide valuable information for disfavored minority communities, especially civil rights–related measures of racial equity and statistics covering queer people. Facially inveighing against “DEI” or “radical gender ideology,” the administration has used culture-war battle cries to cease collecting data on trans people being raped in prison or victimized by violent crime, remove racial demographic data from federal workforce data, and so on.
And even at BLS, while direct meddling with the data would likely provoke a rash of resignations and whistleblowing, Trump can still do a lot of damage.
To start, the staff at BLS has been cut by about 20 percent, and that’s at an agency where funding and capacity have stagnated for years. That’s almost certainly why former commissioner Erika McEntarfer announced last summer that the agency would be reducing the sample size of the monthly household survey by 5,000 due to survey costs “increasing faster than the budget.” Understaffing is such an issue that Ph.D. economists are being dispatched to carry out grunt work like monitoring phone lines, according to reporting from Bloomberg. Also last summer, BLS announced that it is collecting about 15 percent less data on consumer prices, used to calculate the Consumer Price Index, which in turn is used by the Federal Reserve to set interest rates. In some areas, BLS has stopped collecting consumer price data entirely, relying on statistical inference to fill in the gaps.
Then last August, Trump fired McEntarfer, alleging without evidence that she had manipulated jobs numbers to make him look bad. Trump then nominated E.J. Antoni, chief economist at the increasingly fascistic Heritage Foundation. Antoni, who thinks Social Security is a Ponzi scheme and called for DOGE to “take a chainsaw to the BLS,” was panned as an unqualified partisan even by conservative economists.
In January, with Antoni’s nomination dead in the water, Trump instead nominated Brett Matsumoto, a career statistician with a decade working at BLS, to lead the agency. For the last year, Matsumoto has been on assignment to the Council of Economic Advisers. The agency he’ll likely return to is not the same one he left in 2025.
All this takes its toll. According to Groshen, the former BLS commissioner, “trust is mission critical for statistical agencies like BLS.” That trust is frayed by the president spinning economic data before it is even released. Claudia Sahm, an economist who previously worked at the Federal Reserve and the Obama Council of Economic Advisers, said that “it’s a degrading of an institution, it’s a degrading of a norm” that runs the risk of “chipping away at the integrity of the data.”
“For people to trust the data, they have to believe that it was collected and processed without political influence,” added Groshen. That belief is shaken when the president fires the BLS commissioner for unfavorable revisions and starts disseminating a political narrative before independent experts and reporters even have a chance to read the latest release.
It can be hard to understand how dire the stakes around a functional, robust, apolitical BLS are. Fundamentally, the crucial decision-making information that the economy relies on is “a trust system,” as Jacquez described it. The wide-ranging attacks on BLS independence showcased by McEntarfer’s firing and the attempted installation of Antoni invite skepticism of federal data. That skepticism itself then threatens to further degrade the data. Both Sahm and Groshen relayed that it is becoming increasingly difficult to survey a large number of people for a variety of reasons, including the drop in response rates in traditional phone surveys and rising costs of research. But it also does not help if potential respondents don’t trust the survey they’d be participating in to be accurate and unbiased. Why bother responding if it will be warped to political ends anyway?
The erosion of legitimacy in federal data features cascading shocks to public trust, each feeding into the next. The president posting data before its release does not, in itself, undermine the statistics’ integrity. But it “blurs the line between the generation of objective statistics and the government’s political objectives,” according to Groshen. As that line blurs, people trust the statistical agencies less. When people trust the agencies less, they won’t participate in data collection, resulting in worse data. Then, if the quality of data declines, it prompts more criticism of the agencies. Thus has Trump created a doom loop within the agencies that produce some of the most critical decision-making inputs to the global economy.
Image credit: “BLS” by brownpau is licensed under CC BY 2.0.