There’s something wrong with the Federal Register

Photo Credit: Getty

The Trump-era Federal Register website has been glitching recently. Nearly two weeks ago, I noted on X/Twitter (tagging both @USNatArchives and @FedRegister) that the site’s “advanced search” function suddenly stopped returning precise rule counts. No announcement, just an abrupt change.

For instance, a query for total rules issued in 2025 as of this morning (September 2) displays  “1700+” instead of the precise 1,726.

The search tool historically supported detailed filtering by date, agency, rule type, small business or significant economic effects, and other attributes such as docket ID or Regulation Identifier Number. Even executive orders and other decree tallies were supported. Researchers, policymakers and stakeholders rely on precise figures; but now no matter the pursuit, that “+” approximation appears instead. Queries with fewer than 10 results just spit out “Limited results.”

One hopes there’s underlying reprogramming underway (say, something fancy like the DOGE engineering a Federal Register 3.0) that will restore precise figures as well as make the database more dazzling than ever. Imagine, for example, incorporating a new “Deregulatory” category to capture Trump’s “Unrules”—actions that relax certain enforcements or delay or revoke rules altogether.  

Ignored by the media as far as I can tell, these new imprecise tallies are a serious setback for transparency and disclosure. This lapse comes on top of the still-absent Spring edition of the “Unified Agenda of Federal Regulatory and Deregulatory Actions.” That agenda, which outlines agency priorities, is now later than any since 2012, the Obama administration failed to publish one altogether. Its absence is conspicuous, especially since it would show progress and plans on Trump’s one-in, ten-out directive.

In my most recent roundup of figures on the Ten Thousand Commandments page, I was able to compile numbers manually, counting search results page by page, intern-style, and adding them to the prior tallies. By that method the “actual” count for proposed rules so far in 2025 is 1,091, while the database serves out a vague “1,000+ documents.” Similarly executive orders, which total 207 (196 of them Trump’s) report as “200+ documents;” Significant rules today yield “80+ documents” (there are 81).

One must be vigilant even there, since search results are capped at 20 per page. One can search through pages 1 to 10, but once you hit “10” and try “Next,” you get a circle with a red slash, with no way to reach the end of the set, whether it’s sorted by “Relevant,” “Newest” or “Oldest.” This means rule trackers need to perform compilations weekly or lose track which is bound to happen regardless by so cumbersome a method.

But even the numbers obtainable by deeper hand-counts are suspect; for example, there are 18 significant final rules affecting small business so far in 2025. The query results, however, say “20+ documents.”

For as long as the Federal Register has been searchable online, earlier years’ numbers have sometimes inexplicably changed a hair or two percent, something visible over the years in Ten Thousand Commandments. That’s not a surprise in a database containing millions of components. But imprecision should be a bug, not a feature.

The administration should restore the discrete rule counts in the Federal Register. Better still, it should work with Congress to strengthen rule reporting by including deregulatory actions and creating annual regulatory report cards to facilitate ongoing regulatory reforms and streamlining, rather than adding to the pile.