Opinion: The Fed is leaning toward another rate hike. The data behind it has rarely been weaker.

The Bureau of Labor Statistics said on October 14 that consumer prices rose 3.7 percent in the twelve months through September, up from 3.4 percent in August, with core inflation — stripping out food and energy — climbing to 2.6 percent. Gasoline alone, up 9.1 percent for the month and more than a third over the year, accounted for over half the increase. Within hours, traders had sharply raised their bets that the Federal Reserve will raise interest rates again at its October 27-28 meeting, rather than pause. Richmond Fed President Thomas Barkin, who has been among the more cautious voices on the Federal Open Market Committee, called the number higher than he had hoped.

That reaction was rational. What received almost no attention was the condition of the instrument that produced the number. The Consumer Price Index report that moved markets this week came from a statistical agency that spent fourteen months without a Senate-confirmed leader, that lost its first nominee to a social-media scandal, and whose own commissioner for the second half of that stretch would not take the job until he had publicly promised Congress, under oath, that the numbers were not rigged. The agency's survey response rates have been falling for a decade, and funding has not kept pace with the cost of chasing down the Americans who no longer answer government phone calls. None of that invalidates Wednesday's inflation reading. All of it should make everyone — the Fed most of all — treat the number with more humility than markets showed.

A rate increase built on thinner ground

On September 16, the Federal Open Market Committee raised its benchmark rate a quarter point, to a range of 3.75 to 4 percent, in a unanimous 12-0 vote — the first increase since 2023 and a reversal of two years of market expectations that the next Fed move would be downward. The committee's stated reasoning was spare: inflation remains elevated, and the increase would support a timelier return to the 2 percent target. When the minutes of that meeting were released on October 7, they showed most participants judging that a further increase would likely be appropriate before year's end, though without committing to October specifically.

That is the backdrop against which the September CPI reading landed like a confirmation. A central bank that has just reversed course on rates, and signaled more tightening to come, received a report showing inflation accelerating rather than cooling, with a hot producer-price number arriving the next day to reinforce it. It is exactly the kind of data point that can tip a committee from "probably one more hike" to "this meeting." The trouble is that the committee's confidence in that data point should be lower than it was two years ago, not the same — and lower than markets, which treated 3.7 percent as a clean fact, appeared to assume.

The commissioner who was fired over a number

The proximate cause of that lowered confidence is not abstract. In August 2025, President Trump fired BLS Commissioner Erika McEntarfer within hours of a weak jobs report that included substantial downward revisions to prior months, accusing her on social media of manipulating the figures for political reasons. He offered no evidence. William Beach, who had led the BLS during Trump's first term and had no partisan reason to defend his successor, called the firing groundless, noting that a commissioner has no hand in collecting the raw numbers and does not see a jobs report until the Wednesday before it is published.

What followed was not a quick replacement but a vacancy that stretched across fourteen months and two nominees. The first, Heritage Foundation economist E.J. Antoni, was withdrawn after old social-media posts surfaced. The second, career BLS economist Brett Matsumoto, told the Senate at his June confirmation hearing that he did not believe the agency's data had been fabricated or rigged — a statement that would have been unremarkable at any other BLS confirmation in the agency's history, and was necessary at his. He was confirmed in August, which means the inflation report that just moved the Fed was among the first major releases to carry his signature. That is a reasonable basis for confidence in the agency's integrity. It is not, by itself, a basis for confidence in the agency's capacity.

Confirming a commissioner answered the loudest political question hanging over the Bureau of Labor Statistics. It did nothing to answer the quieter, more consequential one: whether the surveys underneath the numbers still work.

The quieter crisis: who answers the phone

That capacity question is where the real erosion is happening, and it predates the McEntarfer firing by years. A Government Accountability Office report published in June found that response rates to the surveys underlying the monthly jobs report have declined over time, calling it a risk to data quality, and noted that the BLS has not even completed its own assessment of how much that decline biases the employer survey, in part because of data-linkage limitations. The GAO's recommendations — that BLS document its methodology more clearly and publish a bias assessment by the end of this year — were accepted by the agency, which is itself a tacit admission of the gap.

A separate account, laid out in an August letter from the House Education and Workforce Committee's Democratic staff, put the agency's response rate for one of its core voluntary surveys at a record low of 63.9 percent as of March 2026 and pointed to a staffing decline of more than 20 percent. A congressional minority letter is, by nature, an advocacy document, and its framing should be read accordingly. But its underlying numbers track a trend that nonpartisan reporting and the agency's own budget history corroborate: fewer employees processing fewer completed surveys, with less money to chase down the respondents who no longer pick up.

This is not a fringe worry confined to one party. A Reuters poll of 100 economists last summer found 89 concerned about the quality of official U.S. economic data, with most saying authorities were not treating the problem with sufficient urgency. Nearly two-thirds said deteriorating statistics could impair the Fed's own ability to set policy correctly — a prediction that looks less theoretical with each data-dependent FOMC meeting.

Why a good number from a weakened agency is still a problem

None of this is an argument that the September CPI figure is wrong. It is an argument that the margin of error around it — both the formal statistical kind the BLS publishes and the harder-to-quantify kind introduced by falling response rates and a thinned-out staff — is wider than the market's one-day, decisive reaction implied. The Fed is not supposed to make a tightening decision on a single data point in any case, and officials will have another jobs report and another inflation reading before they vote. But the psychological weight that this particular report carried, arriving as it did atop a committee already primed to look for confirmation that it had been right to raise rates in September, is exactly the scenario in which thinner data does the most damage: not by being dramatically wrong, but by being just uncertain enough, at just the wrong moment, to tip a close call.

Matsumoto's confirmation, after Antoni's withdrawal and more than a year of acting leadership under deputy commissioner William Wiatrowski, removed the most visible symbol of the crisis. It did not restore the surveys, the staffing, or the funding that the agency lost over the preceding several years, and the confirmation vote itself — 51-47, bundled into an en bloc package of unrelated nominations — suggests Congress treated the moment as a housekeeping matter rather than the structural repair it actually required. The GAO's bias assessment is not due until the end of this year. The online response option meant to arrest declining participation is not scheduled until 2027, and the agency has already warned that funding constraints could delay even that.

Markets and the Federal Reserve will keep having to act on BLS numbers in the meantime, because there is no credible substitute and because policy cannot simply wait for perfect data. That is precisely why the agency's capacity matters as much as its leadership. A thermometer that has been tampered with gets fixed by replacing the person who tampered with it. A thermometer that is simply wearing out from years of underinvestment needs new parts, not just a new reader. Congress funded the commissioner's confirmation hearing. It has yet to fund the instrument.