March 13, 1997

Announcement of the March 1996 Benchmark Revision
for State and Area Estimates from the Payroll 
Survey Once a year, the Bureau of Labor Statistics
adjusts the payroll survey’s sample-based employment
estimates to incorporate the previous year’s March 
universe employment counts in a process known as
benchmarking. These universe employment counts are 
derived principally from State unemployment insurance
tax reports that nearly all employers are required to
file. Revised payroll survey employment, hours, and 
earnings estimates incorporating these benchmark 
revisions for States and metropolitan areas are 
published with the release of January preliminary data
in March of each year. The March 1996 benchmark 
revision was released on March 13, 1997, and affected
all unadjusted data from April 1995 forward and the 
seasonally-adjusted State estimates back to 1992. More 
detailed information on the effect of the benchmark 
revisions to State and industry series will be available
within the next month. 

State data are derived independently from national data
and are intended for individual state analyses only. 
Because the state data are not designed for national 
aggregation, a summation could present a misleading 
picture of national trends. The same is true with the 
benchmark revision process. For information on the 
national March 1996 benchmark revision , please refer 
to the note on the National Current Employment Statistics
section of this website, or contact the Division of Monthly
Industry Employment Statistics on (202) 606-6555. 

February 10, 1997

CAUTION -- CES STATE and AREA LABSTAT USERS

Between February 7 and March 9 the CES state and area 
estimates are being revised to incorporate new 1994 
benchmarks. During this annual procedure it is possible that 
all data from April 1993 forward will be revised.  LABSTAT 
will be updated state by state on a flow basis as data 
become available.  Therefore, interstate comparisons are 
discouraged.  No portion of a state will be updated until the 
entire state's not seasonally adjusted estimates are ready 
so that all current data within a state will always be on the 
same benchmark.  Users are cautioned to check the 
benchmark years whenever extracting data.  A '94' 
benchmark year and 1994 annual average indicate that data 
have been revised.

New benchmark data are used to calculate new seasonal 
adjustment factors and to revise the last five years of 
historical seasonally adjusted All Employee statewide data 
at the major industry division level.  Because data must be 
benchmarked before seasonal adjustment can take place 
there will be a lag between the availability of the two data 
types.  Therefore it is possible that at any given time data 
for a state may be available not seasonally adjusted on the 
March 1994 benchmark while seasonally adjusted data for 
that same state, current and for the past five years, is 
available on the March 1993 benchmark.

Again, it is imperative that users check benchmark years 
before making data comparisons.  All revisions and 
seasonal adjustments should be final by March 9.