Coronavirus (Covid-19): Latest updates and information
Skip to main content Skip to navigation

Session 1: Rationale, problems and pitfalls

Understanding Industry Staffing Patterns: The Occupational Employment Statistics Survey and the National Employment Matrix

The U.S. Employment Projections methods culminate in the application of the National Employment Matrix to translate industry employment projections to occupational projections. The Matrix depicts industry staffing patterns – distribution of each industry’s employment by occupation -- and facilitates projecting change in these patterns to the target year.

The Matrix, in turn, is derived primarily from the Occupational Employment Statistics (OES) survey, a large employer survey yielding employment by occupation and wages for over 800 occupations, and distribution of employment by occupation for about 350 industries. OES survey data were introduced in the U.S. projections work in the late 1970’s, replacing data from the decennial Census of Population.

The OES survey has the advantage of being statistically robust, with a large sample and relatively small sampling errors on national data, and providing a great amount of occupation and industry detail. For use in sub-national projections, OES provides geographic detail that permits generation of projections using state and area industry staffing patterns. For industries and classes of workers not covered by the OES survey, primarily agriculture industries and self-employed workers, we use household survey data.

To project industry staffing patterns, BLS economists examine past changes in staffing patterns and analyze information on technology, business practices, and other factors that affect the occupational mix of each industry. Based on this analysis, they adjust the staffing pattern to reflect their judgment of likely changes over the projection period. For the 2008-2018 projections, using projected staffing patterns rather than keeping staffing patterns unchanged resulted in employment change equivalent to nearly 30 percent of the projected change in total employment.

Rationale and Uses: Lessons from French PMQ (Prospective de métiers et des qualifications) Exercises

In the mid 1990’s, the French Prime Minister launched a national exercise of forecasting of occupations and skills employment. This exercise is undertaken by a team of experts from various French administrative agencies, and is led by the Centre for Strategic Analysis (CAS, which replaced the Plan Commission (Commissariat général du Plan)). The goals of these forecasts are linked to the failures of French labour market (high level of structural unemployment, bad school to work transitions) and especially its segmentation. Some new issues enhanced the process since the beginning of 2000’s: aging, Lisbon strategy, greening economy. The four generations of the project have been discussed and used by social partners and Government in their informal discussions. So qualitative results and key messages are also important than quantitative results.

Tools and methods varied for 20 years, but there have always been employment and exits of labour market forecasts by occupation using a specific classification crossing statistical and administrative approach. The macroeconomic analysis varied in the time and was often separated from the detailed employment forecasts. In the past, the exercises were very close to the demographic projections made by the French Institute of Statistics (INSEE) and focused on trends. The current project is quite different with three scenarios and three employment forecasts base on a multisectoral macroeconomic model (Nemesis from Erasme Team).

Some difficulties can be identified. First of all, the quality of the data generates some problems, despite the big role of national institute of statistics. Second, the lack of studies on occupations and skills in France. Third, the lack of knowledge in economics in the debate on labour market.

Speakers

  • Dixie Sommers
  • Tristan Klein