There is a troubling pattern emerging here in Washington. You have probably noticed it, but part of our job is calling this stuff out.

So here it is. When information is inconvenient, make it harder to find. When data produce an unwelcome result, question the data. When the people responsible for producing the numbers deliver bad news, attack the people. And when the rules are supposed to protect information from political interference, weaken the rules.

The latest example comes from the Commerce Department, home to the Census Bureau, Bureau of Economic Analysis, and National Oceanic and Atmospheric Administration. In August, Commerce quietly removed language from its scientific integrity policy that explicitly prohibited political interference with government research and data.

What could possibly be the reason for that? The previous policy said scientific findings should not be “suppressed, delayed, or altered for political purposes” or subjected to inappropriate influence. It identified protection from inappropriate influence as a hallmark of scientific integrity. This was a problem?

The timing seems, well, suspicious. The change was dated August 19, one day after the Census Bureau published an unusual report making claims about noncitizen voting. The report was not written by Census Bureau career experts and relied on research from a team with ties to a think tank founded by officials from President Trump’s first administration. Former Census officials expressed surprise that work produced outside the bureau had been published under the name of one of the nation’s most trusted statistical agencies.

Last year, President Trump fired Bureau of Labor Statistics Commissioner Erika McEntarfer hours after the agency released a disappointing jobs report with large downward revisions to previous months. Revisions, both downward and upward, are a normal part of producing economic estimates from incomplete information and updating them as better information becomes available. The president accused the agency of manipulating the numbers for political purposes but offered no evidence.

The administration also ended NOAA’s Billion-Dollar Weather and Climate Disasters database, which had tracked the frequency and cost of major disasters since 1980. We wrote about that decision at the time because the database was an important tool for understanding the growing federal cost of disasters. The administration stopped updating the database. The costs, of course, keep coming.

The same thing happened with information about government spending. In March 2025, the Office of Management and Budget took down a public database showing how it was directing agencies to spend money Congress had approved. The Government Accountability Office warned that this apportionment information was important for oversight of federal spending. TCS and others opposed this maneuver. A federal court eventually ruled that removing it violated the law and ordered OMB to put it back.

EPA has begun the process of dismantling most of its Greenhouse Gas Reporting Program, which requires more than 8,000 facilities, suppliers, and other reporting sites to submit data. The agency’s proposal would eliminate reporting requirements for 46 of the program’s 47 source categories and suspend most remaining oil and gas reporting until 2034.

These aren’t just a handful of examples we cherry-picked. A recent review by the Partnership for Public Service found that only about half of the major federal datasets it examined were still fully available and being kept up to date in 2025. Others missed updates, stopped updating altogether, or had features that no longer worked. Separate efforts by data experts have documented the termination or removal of information involving food insecurity, greenhouse gas emissions, future disaster risk, environmental justice, federal employment, public health, education, and other areas.

Federal datasets do not need to exist forever just because they exist today. Programs become outdated. Statistical methods improve. Some information may no longer be worth what taxpayers spend collecting it. But if that’s the case, then show us. That is what we call transparency. Give the public enough information to decide whether the government’s explanation holds up. That is what makes accountability possible.

And in case you’re thinking these are just feel-good ideas, consider how much depends on this information. The Federal Reserve relies on employment, inflation, and economic growth data to set interest rates. Investors use the same information to decide what they are willing to pay for Treasury securities. Businesses use it to decide whether to hire, build, borrow, or invest.

We’ve written recently about what happens when investors become less confident in the United States and demand higher interest rates to lend Washington money. With debt held by the public roughly the size of the entire economy and annual interest costs around $1 trillion, even small changes in borrowing costs can become very expensive.

There is nothing wrong with challenging government statistics. We do it all the time. Question assumptions. Check the methods. Compare estimates with what actually happened. Demand better data when the existing data aren’t good enough.

But that isn’t what we’re talking about here. If you don’t like what the data show, the answer isn’t to attack the people who produced them, weaken the rules protecting them from political interference, or stop collecting the information altogether. We’re talking about a pattern of removing or interfering with public data and offering explanations that seem designed more to brush off questions than answer them.

Getting rid of the data doesn’t get rid of the problem. It just makes it easier to pretend the problem isn’t there. But sooner or later, taxpayers get the bill anyway.