Industry Employment Projections
Preparing industry projections involves four steps.
First, historical time series of industry employment are developed using data from the Current Employment Statistics (CES) and
Quarterly Census of Employment and Wages (QCEW) programs based on three- and four-digit North American Industry Classification System (NAICS) codes.
For employment not covered by the CES or QCEW, the Current Population Survey (CPS) and the decennial census are the major data sources.
Several statistical methods and econometric models are then used to develop a set of preliminary projections for each industry.
Principal methods and approaches when projecting long-term industry employment are fully-specified econometric models; extrapolation or allocation models,
such as employment share or shift-share models; and single-equation regression models.
The short-term models include trend, ordinary least-squares, autoregressive-moving average, vector autoregressive, and Bayesian vector autoregressive models.