Incorporating a Birth/Death Adjustment into the CES National Benchmark Revision

• Activity Number: 515
• Type: Contributed
• Date/Time: Wednesday, August 6, 2014 : 10:30 AM to 12:20 PM
• Sponsor: Government Statistics Section
• Author(s): Victoria Battista*+
• Companies: Bureau of Labor Statistics
• Keywords: Birth/Death Modeling ; Current Employmnet Statistics ; Industry Employment ; Estimation ; Forecasting ; Benchmarking

Abstract

The Current Employment Statistics (CES) program at the U.S. Bureau of Labor Statistics (BLS) produces monthly employment for a wide range of industries 3 weeks after each reference period, revising these employment totals twice in the following 2 months and twice more in the subsequent 2 years as a result of the benchmark process. During this time, business openings (births) and closings (deaths) are occurring, but cannot be collected because they are either not yet part of the sample frame or are simply no longer reporting. CES uses a birth/death (BD) model based on historic counts of these openings and closings taken from the CES sample frame to forecast net BD for the CES reference period. The actual net BD amounts, only used currently as inputs to net BD forecasts, are never incorporated into the last round of CES edits, the final wedged CES series. In this research, old net BD forecasts are replaced with actual net BD values before wedging the remaining error to calculate a final, BD adjusted, wedged CES estimate. The hope is that by incorporating the actual net BD amounts before benchmarking, final CES estimates will more accurately account for non-linear estimation error.