Last Friday morning, just as market participants were digesting a rather shockingly weak U.S. jobs report, an even weightier headline hit the screens. The disruptor-in-chief, Donald Trump, had fired the Commissioner of the Bureau of Labor Statistics (BLS), Erika McEntarfer. No advance notice. No press conference. Just a clean, clinical cut. The rationale put forward was, of course, ridiculous. Trump stated that the jobs report was rigged, not only in the below-consensus estimate of a 73,000 gain in July, but the combined 258,000 downward revision to May and June’s original numbers. He also ranted that the BLS had purposely altered data prior to the election and then after he came to the office.

While it may seem easy to disregard the Truth Social posts, the firing of arguably one of the world’s premier data collection agencies dents one of the most important elements of a properly functioning capital market – trust. Or, in this case, a lack thereof.  This week, I want to dive deep into the fragility of the numbers that fuel our models, portfolios, and policy expectations. Because when the referee is dismissed mid-game, it’s fair to wonder if the score still matters.

From the moment a major economics report hits the wire, analysts, traders and algorithms spring into motion. As the U.S. bond, currency and equity markets drive global capital movements, so too does the world move in response to the headlines and details of these reports. As such, there is an implied belief among market participants that the data released is representative of that segment of the economy. Prior to last week, many investors felt that even though the U.S. economy was slowing, it still retained a hue of exceptionalism. Job gains were close to the pace associated with an economy growing at or above potential. That in turn supported the rosy outlooks being espoused by companies reporting over the last number of weeks, thus ‘validating’ everything from lofty stock market valuations, below-average volatility and tight credit spreads.

The July payrolls report was definitely a jolt, especially with the revisions applied. The chart above shows the monthly changes in U.S. non-farm payrolls based on the June report and then from the July report. Trump’s claim that this report was rigged was buried by his own BLS Commissioner – Bill Beach – who he appointed back in 2017. Beach commented on the weekend that not only was it impossible for the Commissioner to alter the numbers, but his accusation betrayed a misunderstanding of how the data is put together.  The Commissioner does not see the report until Wednesday, after it has already been compiled for release Friday morning.

First, an initial monthly estimate of payrolls is based on phone and internet interviews with more than 100,000 business and government agencies (called the Current Employment Statistics, or CES, survey) and BLS models. In the case where responses were not received or where the firm or agency included an incorrect response, the BLS goes back and re-surveys and does this for two months. For example, when the May data is released at the start of June, those numbers can be revised with the June report and then again with the July report. After that, the number is embossed in stone, or sort of.  Since all of this data is based on survey, the BLS performs an annual benchmark adjustment that is based on comparing the year’s month’s estimates with data from the Quarterly Census of Employment and Wages (QCEW). Despite Trump’s view on things, revisions are normal and serve to make the final figures more accurate.

The firing of the BLS Commissioner isn’t merely symbolic. It signals a potentially seismic shift in how data may be handled under the Trump administration. While the official explanation hinted at 'performance misalignment,' insiders suggest it’s part of a broader effort to control the narrative ahead of a contentious trade season and a looming election. We’ve seen historical analogs—from Nixon’s interference in inflation stats to more recent tensions between the Fed and executive branch. But this move feels different. It opens the door to further politicization of agencies once deemed sacred and apolitical, and therefore essential to sound business and investment decisions, in addition to households.
 

Keep in mind that the BLS also compiles the CPI reports and based on the fiasco from last Friday, next Tuesday’s July CPI report is going to be even more scrutinized by the markets.  Economists are already forecasting a larger monthly increase of 0.3% and a one-tenth uptick in the year-over-year inflation rate to 3%. Anything higher and this could play against the increased expectations for a September Fed rate cut that came out of the payrolls data. Regardless of whether the CPI print comes in higher or lower than expected, investors are still more likely to ask - can we still believe this?

The concern over politicization would be troubling enough on its own. But it’s compounded by a more insidious, creeping issue: the declining quality and participation in the very data we rely on.

Let’s take the Nonfarm Payrolls report. As of July 2025, the initial response rate for the establishment survey had dropped to just 58%, compared to a historical average of around 72%. That's a substantial erosion in the sample size, which can partially explain larger revisions to initial estimates. Keep in mind, that outside of a few states, firms are not legally obligated to participate in the surveys.

Then there’s the Consumer Price Index. Before the pandemic, only about 2.5% of prices were imputed—estimated rather than measured. That figure jumped to 5.1% by late 2022, fell slightly in 2023, and is now back up to a staggering 15–19% due to suspended data collection in several cities. In simpler terms, nearly one in five prices you see in the official CPI may not actually be observed.
 

Statistical analysis has shown that even with the decline in survey participation, the average revisions to payrolls data today are not substantially different than back during the Biden administration or even in the fun days of Trump 1.0.  What does show, however, is that the revisions have been almost uniformly on the downward side and this suggests that the weakening in the U.S. economic fabric might be more severe than what have seen on the surface. If so, this would reinforce our contention that stocks are not fully pricing in economic risk.

Markets are incredibly good at adapting—but they’re terrible with ambiguity. If participants no longer trust government data, they could move toward private alternatives, like ADP employment and real-time credit card spending indices to name a couple. This could introduce more fragmentation into market narratives and increase volatility around official data releases. We might also see investor behavior shift toward real-time inputs and nowcasting models, where proprietary data sources become more valuable than government reports. The bigger question is if we can’t trust the scoreboard, do we still trust the game? 

The credibility of U.S. economic data has long been its silent superpower. Global central banks, pension funds, and risk managers build models—and allocate trillions—based on the perceived objectivity and transparency of American statistics. Coming back to the accuracy of CPI estimates, one market that would be extremely vulnerable would be U.S. Treasury Inflation-Protected Securities (TIPS). These are bonds where the principal is linked to the CPI whereby if inflation rises, the principal goes up and vice versa. When they mature, the U.S. treasury redeems either the original principal amount or the inflation-adjusted value, whichever is higher. Since retail and institutional investors buy these as a hedge against inflation, if there is doubt in the accuracy of the monthly CPI statistics, this could negatively affect demand for these bonds.

If the trust in official government statistics begins to erode, the consequences won’t stay contained. We could see international institutions begin shifting weight to alternative data providers, or even foreign economic indicators. Confidence in the U.S. dollar could be indirectly affected if the foundation of policy credibility starts to crumble. It may sound extreme, but it wouldn’t be the first time a crisis of confidence began with something seemingly technical. Again, this is something we have been talking to clients about for months, that if Trump could not achieve a weaker greenback the old-fashioned way, through lower rates, then injecting uncertainty into U.S. dollar assets could produce the same result. This of course is a very obtuse policy objective, but we have been at this long enough to know that it doesn’t matter how bad a policy seems, it's the effect on markets that has to be reflected on in deciding what to invest in and where.

On behalf of the Pyle Wealth Advisory team, have a wonderful weekend.   

Andrew Pyle