Trusted data is the cornerstone of markets

Following the release of a disappointing employment report in August that included large downward revisions to the prior two months, President Trump fired Erika McEntarfer, the head of the Bureau of Labor Statistics (BLS), the government agency responsible for compiling the monthly employment report. The president alleged in a Truth Social post that the Biden appointee had "faked" the job numbers, later adding that the report was “rigged in order to make the Republicans, and ME, look bad.”

The BLS under Commissioner McEntarfer’s watch certainly doesn’t have a spotless record. In February of 2024, a handful of Wall Street firms received an email from a BLS economist that contained non-public insights into how certain housing inflation statistics are calculated. In May of last year, the BLS inadvertently posted inflation data on its website earlier than the scheduled public release time. Three months later, several Wall Street firms accessed labor market data about 30 minutes before the official release. In each of those cases, the BLS said the information was released accidentally and called for investigations.

While the BLS has made mistakes, there is no evidence that it has been intentionally distorting employment reports to paint a weak picture of the jobs market. For the past several months, there have been multiple signs of slow deterioration in the labor market. For example, every payroll report this year has subsequently been revised lower, which is typically a sign of an accelerating decline in hiring. In addition, recent job gains have been very narrow, concentrated in sectors like health care that aren’t impacted by the state of the economy. Plus, the Federal Reserve’s most recent Beige Book reported declining labor demand in every single Federal Reserve district in the country. Additionally, consumer confidence surveys show people are saying that jobs are getting harder to find, which is why continuing claims for unemployment insurance have been grinding steadily higher. We’re also seeing fewer job postings on Indeed.com, perhaps because more and more companies are using artificial intelligence to replace entry level desk jobs. So, cracks in the labor market had been evident for several months prior to the most recent BLS jobs report.

While the downward revisions to prior payroll reports were unusually large, the reasons for the revisions are likely related to declining response rates from businesses. The BLS samples the payroll records of approximately 121,000 companies and government agencies each month across a full range of states, industries, and business sizes. The BLS then applies a birth-death model to account for new businesses opened and old businesses closed, then seasonally adjusts the data to account for factors such as holiday hiring patterns, weather disruptions and so on.

Since the beginning of Covid, the initial response rate to the survey has fallen markedly from 60 to 43%. The BLS contacts non-respondents after the initial survey results are announced, eventually incorporating their data into a revised payroll number. Because fewer firms are responding to the initial request, the subsequent revisions are necessarily larger. So, the big revisions we are seeing in the labor data are likely a function of companies not responding promptly to the BLS. An additional issue is the impact of the Covid recession on the economy, which has potentially distorted seasonal adjustment factors used by the BLS.

The real problem with the Trump administration’s attack on the government’s data reporting is that it creates the impression that the data have become politicized and can no longer be trusted by investors to provide an accurate snapshot of the economy. Combined with the administration’s continued attacks on the independence of the Federal Reserve, the markets are facing a collapse in trust.

While so far the market impact of the firing has been minimal, in the long run trust is the cornerstone of the American financial system. A loss of trust can lead to greater market volatility and a higher risk premium for all financial assets, even stocks which have been flying high on the promises of widespread AI adoption. Loss of trust can also lead to a steeper yield curve (due to an increase in the term premium) and erosion of the dollar’s status as the world’s reserve currency.

Once lost, trust is very difficult to reestablish.