Payroll Survey
The Payroll Survey is the larger part of the Current Employment Statistics (CES) program, along with the household survey. It is a continual survey operated by the BLS.
Usage
Design
Sampling
The survey utilizes a probability sample of worksites drawn from unemployment insurance accounts. About 1/3 of the sample is rotated out each year. A supplemental sample of new businesses is selected midway through the year. It is designed to reflect non-farm payrolls.
The private sector sample is stratified by state, business size (8 categories), and industry (13 categories).
The government sample uses a quota system and is stratified by region.
Mode
For new sample members, the survey is administered by CATI. After a five month period, most sample is moved to a less costly mode per input from the sample member.
Frequency
Estimates are published monthly.
A major component of uncertainty are new businesses ('births') and business closures ('deaths'). Because there are lags between these events and administrative records being updated, and also lags between these frame updates and the next frame recreation, these dynamics have to be modeled.
A model of seasonal trends is also used to account for normal events such as school breaks.
Based on the continual data collection, monthly estimates are revised twice before being held constant.
On an annual basis, benchmark revisions are produced against the Quarterly Census of Employment and Wages (QCEW).