Updates to the LAUS and MLS Programs

1
Updates to the LAUS and MLS Programs

• Sharon Brown
• Division of Local Area Unemployment Statistics
• Bureau of Labor Statistics
• Presented by Patrick Carey
• LMI Forum
• Sacramento, CA
• Nov 6-8, 2008

2
Local Area Unemployment Statistics Program Update

• Improvements to methodology---smoothed seasonally
  adjusted estimates
• Expanded releases of data
• Possible changes to annual benchmarking

3
Improvements to model based estimates

• LAUS has seen a progression in model-based
  estimation
• Historically, the trade-off has been between
  monthly smoothness and end-of-year revision
• First generation Implemented in 1989 in
  non-direct use States
• Small annual revision, significant month-to-month
  volatility
• Second generation Implemented in 1994 in
  non-direct use States, 1996 in direct-use States
• Smoother month-to-month movement, larger annual
  revision
• Issues with the endpoint of the benchmark and sum
  of States
• Third generation Implemented in 2005 in all
  States
• Smaller annual revision, no sum-of-State issue,
  but increased monthly volatility in current
  estimates

4
Current model estimation

• LAUS State employment and unemployment estimates
  are developed by bivariate models that use CES
  and UI inputs, respectively, in addition to CPS
  values.
• These estimates are controlled to division level
  model estimates of employment and unemployment
  which are in turn controlled to the National
  estimates of employment and unemployment.
• The process of controlling the State estimates to
  national estimates on a monthly basis is called
  real-time benchmarking.

5
Issues in current estimation

• Although problems with previous estimation were
  addressed, since 2005 States have noted
  occasional increases in volatility in current
  monthly estimates
• Volatility that masks real movement in the series
  may be related to
• Sampling error in the CPS
• Uncertainty at the end-point of estimation
• Real-time benchmarking
• Benchmark adjustment factors fluctuate and the
  relative relationship of the State within the
  division may be affected.
• Month-to-month direction of change may be
  affected, especially for seasonally adjusted
  series.

6
Smoothed seasonally adjusted estimates

• BLS has identified an approach that effectively
  reduces the variability in the monthly
  benchmarked seasonally adjusted series
• Smoothing eliminates fluctuations due to
  volatility but not those due to the business
  cycle
• Four options were developed related to the length
  of time under consideration, i.e., the business
  cycle
• 36, 24, 18, and 12 months

7
(No Transcript)
8
(No Transcript)
9
Status of the project

• A set of smoothed LAUS seasonally adjusted
  estimateshistorical and currentwere sent to
  each State in LAUS Technical Memorandum S-07-23,
  followed by monthly updates.
• States were requested to review the four levels
  of smoothed seasonally adjusted estimates and
  provide comments to BLS by September 14.
• Are the smoothed estimates less volatile?
• Do they more accurately reflect economic
  conditions?
• What degree of smoothness works best?

10
Status of the project--continued

• State responses to the smoothed LAUS estimates
• 29 States support the implementation of these
  estimates
• 12 States do not support their implementation
• 8 States were neutral in varying degrees
• 2 States were in support but for current
  estimates only
• The State comments will be discussed within the
  LAUS-MLS Policy Council and BLS.
• Decision will be made later in October, and will
  reflect State concerns.

11
Increasing the availability of LAUS data products

• Preliminary expanded Geographic Profiles
  employment status tables for States for 2005 and
  2006 were uploaded to LABSTAT on August 20.
• Statewide CPS data from 2004 on for full- and
  part-time employment and unemployment, employment
  by industry, employment by occupation, and hours
  at work were provided to regions and States on
  request in May.
• Statewide veterans employment status tables for
  2006 were placed on the LAUS Intranet site on
  April 9.
• Statewide 2005 data on unemployment duration was
  provided to ETA and to States

12
Increasing the availability of data--continued

• Tables of 2000-2006 annual average county labor
  force estimates were placed on LABSTAT to
  facilitate creation of rankings and listings.
• Model-based monthly and annual average
  employment-population ratios and labor force
  participation rates, along with error measures,
  was placed on the LAUS website September 26
• 2003 Geographic Profiles report will be issued
  shortly
• 2004 and 2005 Geographic Profiles reports will
  follow
• BLS looking at faster ways of issuing these data

13
Possible modification to LAUS revision and annual
benchmarking

• Monthly revisions in LAUS were eliminated as part
  of the FY 2006 permanent cut in State funding.
• Single estimate per month complicated corrections
  at the beginning of the year while benchmarking
  was in process.
• The 2007 benchmarking involved seven years of
  datathe most ever collectedfor about half of
  the States and five years of data for the
  remainder.
• Many States were required to submit or resubmit
  because of problems in other States.

14
Possible modification to monthly revision and
annual benchmarking

• In FY 2006, BLS eliminated regular monthly
  revisions of current LAUS estimates to address a
  budget rescission. For this activity, State
  funding was reduced by 600,000.
• Most States, including those on the Policy
  Council, have expressed interest in resuming
  monthly revisions of LAUS estimates for both
  operational and methodological reasons.
• BLS is considering the restoration of monthly
  revisions with the streamlining of substate
  annual processing identified as the offset.
• The State members of the Policy Council are
  surveying States.
• Thus far, 41 States have commented
• 35 support the monthly revision
  restoration/annual benchmarking reduction
• 3 do not
• 3 had no strong opinion

15
Mass Layoff Statistics Program update

• Status of MLS Redesign projects
• The use of a structured interview approach for
  collecting important information from employers
  on the nature of the layoff.
• The review, revision, and expansion of the
  economic reasons for the layoff, and the
  development of seven higher-level categories.
• A new data collection the Business Function
  involved in the mass layoff.
• A new use for initial claims using all initial
  claims entering the MLS system to define small
  layoffs.
• Operational enhancements and improvements

16
Use of a structured interview in MLS

• Structured interview approach was primarily
  introduced to support the collection of Business
  Function.
• Other important reasons to use a structured
  interview
• Good tool for new analysts to use (get up to
  speed faster).
• Good tool if an analyst doesnt do a lot of
  employer-contact interviews.
• Memory jogger if a respondent gives you all the
  information at once.
• Helps control the human urge to elaborate, to
  paraphrase, or to explain in more detail (and
  possibly change meaning).
• Improves the consistency of approach and the
  quality of the collected information.

17
Improvements to economic reasons for layoff

• New reasons
• Domestic competition
• Excess inventory/saturated market
• Cost control/cost cutting/increase profitability
• Government regulations/intervention
• Additional probing/Interest in secondary reasons
• Creation of higher level categories for economic
  reasons
• Business demand
• Organizational changes
• Financial issues
• Production specific
• Disaster/safety
• Seasonal
• Other/miscellaneous

18
(No Transcript)
19
Extended Mass Layoff Events by Reason
Categories1Second quarters, 2001-07
p
1 The chart excludes information on layoffs due
to other/miscellaneous reasons. p preliminary.
20
Business Function and Business Process

• In June 2006, a proposal was developed to collect
  information on business function involved in the
  layoff.
• Business Function is defined as an activity
  that a firm does in order to produce its product
  or achieve its objective.
• A feasibility study was conducted in September
  and October in which 11 States tested the
  collection of business function.
• In January 2007, the collection of Business
  Function involved in mass layoffs and plant
  closings was implemented nationwide.
• Categorization of Business Functions into
  higher-level Business Processes was introduced in
  March.
• Business Processes describe the full range of
  activities a firm engages in to conduct its
  business.

21
MLS experiences in collecting Business Functions
and identifying Business Processes

• During the first and second quarters, the
  employer interview was not adversely affected by
  the Business Function question
• Refusals were 4.4 percent of all events in the
  first quarter and 3.4 percent in the second
  quarter.
• These represent overall refusals to participate,
  not item nonresponse.
• Does not know reports were quite low1.2
  percent and 1.1 percent, respectivelyindicating
  that the correct person was being interviewed
• Unable to contact reports were relatively
  high6.7 percent and 3.6 percent,
  respectivelylikely reflecting the requirement to
  wait 31 days before attempting contact.

22
MLS experiences--continued

• Business Function guidance
• Addressing occupation and industry responses
• Reliance on follow-up questions
• Addressing variations in Business Functions
  reported to BLS
• 1st Quarter 489 unique terms used for 1,873
  Business Functions
• 28 terms with at least 10 responses accounted for
  1,113 Functions
• Other 461 terms, mostly used for one or two
  Functions, reflecting minor differences in
  wording or overly detailed Functions
• 2nd Quarter 442 unique terms used for 2,160
  Business Functions
• 40 terms with at least 10 responses accounted for
  1,434 Functions
• In an attempt to standardize terms, guidance was
  provided for a number of functions including
  clerical support, food services, lodging
  services, general management, construction
  activities, entertainment services.
• Addressing specific industry issues
• Temporary help firms, retail trade establishments
• Layoffs related to summer school closings

23
Total and Main Business Processes involved in
Mass Layoffs, January-June 2007
24
Next steps for Business Function and Process

• Further automation of the collection is very
  important.
• Further standardizing of Business Function terms
  is essential, but this must be done without
  compromising the collection.
• Additional qualitative analysis of the Business
  Function collection and the coding to Business
  Process continues, leading to publication.

25
Operational improvements and activities

• North Carolina Longitudinal Tracking System
• WinMLS operating system
• 2007.2.0
• Reduced number of employer interview pages
  printed.
• IC totals, establishment employment, important
  layoff dates will appear on the interview
  printouts.
• Check box added for hiring hall
• 2008.1.0
• Entry of Business Function and Business Process
  information will be automated.
• NAICS 2007 will be implemented.
• Extract function for LTS will be included.
• Archive function for information older than 5
  years (maybe).

26
The year ahead for LAUS and MLS

• Continued focus on program improvement
• Methodology, concepts, operations.
• Timeliness.
• Customer use.
• Communications with States.
• LAUS-MLS Policy Council meeting in New Orleans
  October 10-11.
• LAUS National Training Conference in San
  Francisco October 30-November 1.
• LMI Forum in States November 6-8.
• MLS National Training Conference in San Antonio
  November 28-29.