What is the establishment vs household survey?
The BLS produces two monthly employment measures: the establishment survey (payrolls from 119K businesses) and the household survey (60K household interviews). They can diverge substantially and tell different stories.
Current Value
Updated 8 hours agoWhy It Matters
The Bureau of Labor Statistics releases two employment surveys each month in the Employment Situation report. The establishment survey (also called the Current Employment Statistics or payroll survey) contacts approximately 119,000 businesses and government agencies covering roughly 629,000 worksites. The household survey (also called the Current Population Survey) interviews approximately 60,000 households. These two surveys measure different things, use different methods, and frequently tell different stories.
The establishment survey produces the nonfarm payrolls number, which counts the total number of jobs at surveyed worksites. A person holding two jobs is counted twice. The survey captures employment, hours worked, and earnings with high reliability due to its large sample size and the fact that businesses have exact payroll records. However, it misses agricultural workers, self-employed individuals, unpaid family workers, and workers in private households.
The household survey produces the unemployment rate, labor force participation rate, and other demographic breakdowns. It counts people, not jobs, so a person with two jobs is counted once. It captures self-employment, gig work, and agricultural employment that the establishment survey misses. However, its smaller sample size makes it more volatile, and responses depend on subjective self-classification of employment status.
Persistent divergences between the two surveys often signal structural shifts in the labor market. During 2022-2023, the establishment survey showed robust job growth while the household survey was notably weaker, a gap that raised questions about whether payroll data was overstating employment strength. Possible explanations included the growth of multiple jobholding (inflating establishment counts), population estimation errors in the household survey, and an increase in self-employment that the establishment survey misses. When the two surveys disagree, the truth is usually somewhere between them.
Related Pages
More Employment Questions
Related Analysis
Continue Across Convex
Get daily macro analysis with context on employment, regime signals, and what the data is telling us.
Educational content for informational purposes only, not financial advice. Data sourced from official statistical releases and market feeds. Updated periodically.