The Occupational Employment Statistics (OES) program estimates employment and wages for most occupations by industry and sector at the national level, and by occupation at the state and metropolitan statistical area (MSA), and non-MSA levels in the 50 states and the District of Columbia. OES accounts for 1.2 million establishments and 62% of national employment, including railroad, but excluding military, agriculture, fishing, forestry, private households, self-employment, and others.
How Emsi Incorporates OES
OES is our primary source of occupation data, but we compensate for OES’s general weaknesses and lack of valid historical data by utilizing stronger, more accurate industry employment counts from QCEW, County Business Patterns (CBP), and American Community Survey (ACS), among others. We then apply regionalized, OES-based staffing patterns to the industry data to show the distribution of jobs by occupation.
Emsi gathers occupation earnings data from OES. We use unsuppression techniques to fill in missing values as appropriate, and also build a time series of OES data in order to present historical occupation earnings.
For a more detailed explanation of how Emsi incorporates OES data into occupational processes, see this article.
Strengths of OES
- OES has estimates for specific industries, including national industry-specific occupational employment and wage estimates.
- OES has estimates for individual states, including cross-industry occupational employment and wage estimates for individual states.
- OES has estimates for metropolitan and nonmetropolitan areas, which together cover the entire United States.
Weaknesses of OES
- OES is merely a survey and is not based on administrative records like Quarterly Census of Employment and Wages (QCEW) from the BLS; because of this, OES’s figures aren’t as comprehensive as most industry data.
- Not all metropolitan and nonmetropolitan areas have information for all occupations.
- Only 57% of employment is covered in the OES survey (compared to the 95% of wage-and-salary jobs captured by QCEW), which excludes all industries under NAICS 11 (agriculture, forestry, fishing, and hunting) except for logging, support activities for crop production, and support activities for animal production.
- The OES survey takes up to three years to complete, so the BLS states that it is less useful for measuring change in job counts or wages over time. An apparent increase in wages, for example, could just as likely be due to different businesses responding to the survey in one year, changes in the occupational, industrial, and geographical classification systems, changes to collection or estimation methods, or changes to other methodologies in the survey. Emsi’s occupation methodology (see article referenced above) is designed to counteract this weakness in OES data.