Sample Design in the Job Openings and Labor Turnover Survey

The Job Openings and Labor Turnover Survey (JOLTS) is a new program being fielded by the Bureau of Labor Statistics to compile measures of job openings and employee turnover for nonfarm businesses using a sample of 16,000 establishments. The job opening, employee turnover, and employment data are collected on a monthly basis. Employee turnover includes hires, and three categories of separations—quits, layoffs and discharges, and other separations. Most sample members will rotate out of the sample after participating for 18 months. Estimates will be produced for broad industry groups and Census regions. In this paper we will discuss JOLTS sample design, JOLTS estimators, sample coordination with other BLS surveys using permanent random numbers, and issues arising from sample selection using the Standard Industrial Classification (SIC) system as a strata identifier and estimating for North American Industry Classification System (NAICS) sectors. JOLTS is an establishment survey designed to assess rates of job openings and labor turnover relative to employment for key industrial sectors of the United States. Abraham (1983) discusses the job-openings rate as a labor demand companion statistic to the widely known labor supply measure, the unemployment rate. Clark, Cohen and Hyson (2000) discuss the concepts behind the measures produced by JOLTS in relationship to recent literature. Mueller and Phillips (2000) discuss the origins of JOLTS, and Mueller and Wohlford (2000) discuss JOLTS operations. Goldenberg and Phillips (2000) describe the extensive effort that went into defining concepts and developing the survey instruments for JOLTS. The scope of JOLTS is the same as that of the BLS monthly Current Employment Statistics (CES) survey--payroll employment in nonagricultural industries in the private sector, Federal and State and local government, across the 50 states and the District of Columbia. JOLTS will ratio adjust or “benchmark” estimates to employment estimates from the CES. (Butani, et al., 1997)