Methods and Data Sources for Calculating the Number of Full-Time Equivalent Employees By Industry

BACKGROUND:   For purposes of surveillance, many states would like to calculate injury and illness incidence rates by industry and by employer, using their own administrative workers’ compensation data in the numerator.  In some cases, states would also like to calculate these rates for sub-state geographies.  Although the U.S. Bureau of Labor Statistics (BLS) publishes incidence rates by industry at the state level in the Survey of Occupational Illness and Injury (SOII) program, rates by employer and at lower levels of geography are not available. Moreover, concerns have been raised about a SOII undercount of injuries and illnesses.  The ideal denominator for illness and injury incidence rates is full-time equivalent (FTE) employee counts, which control for hours worked.  Most states, however,  do not collect hours worked data from employers. 

METHODS:   The paper compares FTE employee estimates calculated using hours worked data by industry reported by households to the Current Population Survey (CPS) and the American Community Survey (ACS) with FTE estimates using data reported by employers to sources such as Current Employment Statistics and the BLS Labor Productivity and Cost data.  Data from 2009 for two states, California and Washington, are used as well as data for the nation.  California’s large population helps minimize sample size issues and the State of Washington is one of the few states that collect their own hours worked data from individual employers.  Two methods of estimating FTE employee counts are compared: the first method directly calculates the FTE estimates using ACS and CPS survey data for hours worked by industry; a second method calculates an FTE adjustment factor from the survey data, and then adjusts employee counts from the Quarterly Census of Employment and Wages. 

RESULTS: Preliminary results show that FTE employee estimates are similar using ACS and CPS data by industry.  In general, results from household surveys show higher estimates of FTE employees than data reported by employers.  This pattern held across most industries and suggests that hours worked are being over-reported in surveys by households, under-reported by employers or a combination of both.       

CONCLUSIONS: States have varied criteria for choosing between data sets and methodologies to calculate FTE employee counts.  This paper describes these criteria and provides guidance as to the suitability of different choices, including the desired level of industry detail.