Explanatory Note


	The earnings series presented in this release 
are derived from the Bureau of Labor Statistics’ 
Current Employment Statistics (CES) survey, a 
monthly establishment survey of employment, payroll, 
and hours.  The deflator used for constant-dollar 
earnings series presented in this release is derived from 
the Consumer Price Index for Urban Wage Earners and 
Clerical Workers (CPI-W).
	For the purpose of the Real Earnings series, 
the CPI-W is converted from the base of 1982-84 that 
is used in the official, published series to a base of 
1982.  Thus, the constant dollar average hourly and 
weekly earnings series are in 1982 dollars.  To avoid 
confusion for users, the CPI data presented in Table A 
are the official, published CPI-W series.  These data 
may differ slightly from those used in the real earnings 
calculations.
	Seasonally adjusted data are used for 
estimates of percent change from the same month a 
year ago for current and constant average hourly and 
weekly earnings that are presented in Table B of this 
release.  Special techniques are applied to the CES 
hours and earnings data in the seasonal adjustment 
process to mitigate the effect of certain calendar-
related fluctuations.  Thus, over-the-year changes of 
these hours and earnings are best measured using 
seasonally adjusted series.  A discussion of the 
calendar-related fluctuations in the hours and earnings 
data and the special techniques to remove them is 
available in the February 2004 issue of Employment and 
Earnings or on the Internet under 'Technical Notes'
(http://www.bls.gov/ces/). 
	Earnings series from the monthly 
establishment series are estimated arithmetic averages 
(means) of the hourly and weekly earnings of all 
production and nonsupervisory jobs in the private 
nonfarm sector of the economy.  Average hourly 
earnings estimates are derived by dividing the 
estimated industry payroll--for all production and 
nonsupervisory jobs--by the corresponding paid hours.  
Average weekly hours estimates are similarly derived 
by dividing estimated aggregate hours by the 
corresponding number of production and nonsupervisory 
jobs.  Average weekly earnings estimates are derived 
by multiplying the average hourly earnings and the 
average weekly hours estimates.  This is equivalent to 
dividing the estimated payroll by the number of 
production and nonsupervisory jobs.  The weekly and 
hourly earnings estimates for aggregate industries, 
such as the major industry division and the total private 
sector averages printed in this release, are derived by 
summing the corresponding payroll, hours, and 
employment estimates of the component industries.  
As a result, each industry receives a "weight" in the 
published averages that corresponds to its current level 
of activity (employment or total hours).  This further 
implies that fluctuations and varying trends in 
employment in high-wage versus low-wage industries 
as well as wage rate changes influence the earnings 
averages.
	There are several characteristics of the series 
presented in this release that limit their suitability for 
some types of economic analyses. (1) The denominator 
for the weekly earnings series is the number of private 
nonfarm production and nonsupervisory worker jobs.  
This number includes full-time and part-time jobs as 
well as the jobs held by multiple jobholders in the 
private nonfarm sector.  These factors tend to result in 
weekly earnings averages significantly lower than the 
corresponding numbers for full-time jobs.  (2) Annual 
earnings averages can differ significantly from the 
result obtained by multiplying average weekly 
earnings times 52 weeks.  The difference may be due 
to factors such as turnovers and layoffs.  (3) The series 
are the average earnings of all production and 
nonsupervisory jobs, not the earnings average of 
"typical" jobs or jobs held by "typical" workers.  
Specifically, there are no adjustments for occupational, 
age, or schooling variations or for household type or 
location.  Many studies have established the 
significance of these factors and that their impact 
varies over time.
	Seasonally adjusted data (table 2) are 
preferred by some users for analyzing general earnings 
trends in the economy since they eliminate the effect of 
changes that normally occur at the same time and in 
about the same magnitude each year and, therefore, 
reveal the underlying trends and cyclical movements.  
Changes in average earnings may be due to seasonal 
changes in the proportion of workers in high-wage and 
low-wage industries or occupations or to seasonal 
changes in the amount of overtime work, and so on.
	For more information, see Thomas Gavett, 
"Measures of Change in Real Wages and Earnings," 
Monthly Labor Review, February 1972.
	Information in this release will be made 
available to sensory impaired individuals upon request.  
Voice phone:  202-691-5200; TDD Message Referral 
Phone Number: 1-800-877-8339.