Why ADP And BLS Job Reports Often Diverge

By Bob Iaccino

Each month, the financial community anticipates two key employment reports: the Automatic Data Processing (ADP) private payrolls data and the official Bureau of Labor Statistics (BLS) non-farm payrolls report. The ADP report is released two days before the BLS report. Both are important for understanding the health of the U.S. labor market, but the ADP report’s ability to predict the broader trends in the official BLS report is often questioned. How reliable is it as a predictor?

ADP and BLS data have a strong numerical correlation – around 94% over the past decade – but they frequently can deliver contradictory month-to-month signals. For instance, in June 2025, ADP reported a loss of 33,000 jobs, significantly below the consensus expectation of a 100,000 increase, while BLS reported a gain of 147,000 jobs in total non-farm payroll employment.

Another example of this occurred in October 2024. Days before the BLS reported a loss of 28,000 private sector jobs (the headline figure showed a total gain of 12,000 non-farm jobs), ADP data showed the private sector had gained 233,000 jobs. Such significant differences can create market volatility and be a challenge for policymakers making economic assessments.

The Methodology Gap

The correlation between ADP and BLS employment figures breaks down due to differing measurement methodologies. ADP counts workers on payrolls even if unpaid, for example during strikes or natural disasters, surveying 26 million workers. The BLS, conversely, surveys 122,000 employers and only counts workers who received pay during its reference period, leading to diverging data during economic disruptions.

In a 2019 Federal Reserve study, economists investigated the relationship between the CES (Current Employment Statistics) and ADP-FRB (ADP-Federal Reserve Board) series. Their research concluded that “the optimal predictor of the underlying gains puts approximately equal weight” on both series, demonstrating they are complementary, not competing measures. Combining their insights can reduce measurement error by approximately 20%. For detailed historical charts comparing the series, see the Federal Reserve’s 2019 FEDS Notes analysis “Tracking the Labor Market with Big Data”.

A growing challenge for the BLS has been declining survey response rates. Responses to the BLS establishment survey have been falling and stood at a low 43% in June 2025. This deterioration in data quality gives additional weight to alternative measures like ADP’s payroll data.

The Path Forward

Despite hopes among active investors that ADP might serve as a reliable preview of the BLS report, ADP explicitly states that “The ADP National Employment Report is an independent measure of private-sector employment. It is not intended to forecast the Bureau of Labor Statistics monthly jobs report.” This disclaimer became official policy after ADP retooled its methodology in 2022.

Economists increasingly see these divergences as beneficial rather than problematic. Former BLS commissioner Erica Groshen noted in an interview with CNN that ADP and BLS data “become checks on each other.”

Understanding that these series measure subtly different aspects of the same labor market – rather than expecting perfect correlation – provides a more nuanced view of employment trends. In an era of declining survey response rates and increasing economic complexity, having multiple independent measures of job growth is a crucial feature that makes our understanding of the labor market more robust, not less.

Original Post

Editor’s Note: The summary bullets for this article were chosen by Seeking Alpha editors.