Dictatorships & Data Standards, Revisited

Corrupting government data is the most basic form of epistemic assault on rational and democratic governance — and that's clearly Trump's plan

As I wrote some seven ago in The American Interest, the integrity of government statistics is one of the essential foundations of democratic modernity. Systematically collected and disseminated government data — on inflation, employment, trade, demographics, weather, crime, pollution, traffic, poverty, disease, vaccination rates, etc. — provides the epistemic foundation on which citizens can scale their local experiences into a shared sense of national reality. Without this statistical scaffolding, politics collapses into pure narrative contestation, where truth is whatever serves the factional purpose of the moment. In liberal democracies, this trust in neutral data is not an incidental feature; it is a core political virtue, one that separates deliberation from demagoguery. Conversely, from the point of view of a would-be dictator, this trust represents a grave threat to their ability to dictate “reality” by fiat.

This is why Trump’s firing yesterday of Erika McEntarfer, Commissioner of the U.S. Bureau of Labor Statistics (BLS), strikes me as such a grave if sadly predictable development. The official justification, such as it is, followed the release of a weak July jobs report, which recorded just 73,000 new positions and downward revisions to previous months — just as most reality-based economists had anticipated would be the result of Trump’s chaotic, uncertainty-inducing economic policies. President Trump, seizing upon these figures, accused the BLS of intentionally “rigging the numbers” to make him and his administration “look bad.” Within hours, McEntarfer was out — fired not for any wrongdoing, but for overseeing the production of politically inconvenient arithmetic.

Some may be tempted to dismiss the firing of McEntarfer as just another Trumpian tantrum, but to do so would be to miss the structural significance of the act. This was not simply about one jobs report. It was a direct assault on the credibility of the institutional machinery that sustains shared reality in American public life. Public statistics are the foundation of public integrity. The BLS has long been viewed, by left and right alike, as an island of technocratic impartiality, one that month in and month out surveyed households to determine unemployment rates, businesses to determine job growth, and prices to determine inflation — in other words, to provide the most basic facts around which policy debate in a rational-legal system of governance is supposed to revolve. To punish it for statistical honesty sends the message that data itself is now fair game, at least for MAGA, in the partisan war over perception.

Insofar as there is not a hue and cry about the firing of McEntarfer, it represents the normalization of what might be called “epistemic authoritarianism” — that is, a shift away from the norm of disinterested statistical collection and toward a regime where numbers are evaluated not by their accuracy, but by their political utility (to MAGA). In autocracies, the manipulation of statistics is a well-honed craft: growth rates inflated, poverty figures massaged, dissent erased through numerical omission. The Chinese government, for example, was mocked for stopping the publication of data about youth unemployment when that information became embarrassing for the regime — and they have since stopped publishing hundreds of other economic statistics. The goal for epistemic authoritarians is not to understand reality but to fabricate one that consolidates power. That this logic is now migrating into the American context should be a matter of alarm for anyone who values either democratic accountability or rational policy-making.

The politicization of the BLS is emblematic of a broader epistemic fracture: personal experiences are no longer triangulated with trustworthy public data and mediated analysis. Instead, we now inhabit a landscape where each faction has its own “truth,” its own feeds, its own experts. Trump’s firing of McEntarfer is both a symptom and an accelerant of this fragmentation. He does not seek (or at any rate fails) to produce better results that would be revealed by data. He seeks to control the authority of facts themselves.

The same authoritarian impulse is evident in Trump’s recent slashing of the National Oceanic and Atmospheric Administration’s budget. NOAA’s mandate is to collect and publicly release climate data — and its data has formed the analytical bedrock of anthropogenic climate change that MAGA hates because it supports the case for phasing out fossil fuels. As with BLS and economic data, the “targeted attack on climate data” represents not just an environmental policy setback, but a deeper epistemological rupture. Like labor statistics, climate data is not merely information; it creates a form of public accountability for the results of global industrial pollution. By suppressing such data, Trump is not challenging scientific conclusions through argument or alternative evidence; he is seeking to preempt the possibility of argument altogether by removing the facts from circulation.

Or to take another example: health data. HHS Secretary Robert Kennedy Jr. has been undermining public health data, especially amid the 2025 measles outbreak and rising vaccine skepticism. HHS shelved a CDC forecast indicating that measles risk was particularly elevated in areas with low vaccination rates — data that could have informed and motivated public protection — on the ridiculous grounds that it did “not say anything the public doesn’t already know” (if so, why suppress it?). Meanwhile, the department has slashed funding and support for local health workers, canceled measles vaccination clinics, and laid off staff at key agencies including the CDC, FDA, and NIH — crippling the capacity to gather, analyze, and distribute essential health data. Also Kennedy removed all 17 members of the CDC’s Advisory Committee on Immunization Practices and replaced them with vaccine skeptics, diminishing the credibility of vaccine guidance and injecting partisan bias into once-fact-based epidemiological deliberation.

This is classic epistemic authoritarianism: if the data contradicts the narrative, eliminate the data as well as the stewards of that data. The goal is not to win debates in the public sphere, but to render them moot by controlling the informational terrain itself. Destroying public health data, halting climate data publication, firing the inconvenient BLS commissioner — it is all part of the epistemic authoritarian pattern. The aim in each case is to eliminate the instrumental infrastructure of shared public knowledge, that is, to disable society’s capacity to reason collectively about shared challenges — be it economic hardship, the value of vaccinations, or the planetary polycrisis — by seizing control over the means of statistical production that renders those challenges visible in a shared way.

As this trend continues, we risk arriving at a place where statistical outputs — on jobs, inflation, the weather, or anything else — carry no more civic weight than campaign speeches. The BLS, like the Census Bureau, the Congressional Budget Office, or the National Weather Service, is not just a technical body. It is a pillar of civic trust. To undermine it for partisan political advantage is to sap one of democracy’s most essential prerequisites: the belief that facts, though sometimes uncomfortable, are real, and that reality matters. When such acts are repeated, rewarded, and normalized, they create feedback loops that degrade the informational commons upon which rational governance depends.

What remains is the question of resistance. In the short term, our data-generating governmental institutions retain significant inertia, and many civil servants still take their role as neutral stewards of the facts seriously. But citizens and journalists need to recognize that defending the integrity of data is not merely a dry bureaucratic concern — it is a foundational democratic act. The facts do not defend themselves. We must. And we must do so not simply by rebutting lies, but by insisting on the institutional independence and statistical neutrality that make shared reality possible. Because if that vanishes, politics will becomes nothing more than performance — which in turn will mean that American democracy has become epistemically indistinguishable from the tyrannies we once defined ourself against.