Local Area Unemployment Statistics Program

U.S. Department of Labor Bureau of Labor Statistics

2 Massachusetts Ave., N.E.

Washington, D.C. 20212

DATE : January 27, 2020

LAUS Technical Memorandum No. S-20-06

MEMORANDUM FOR: STATE WORKFORCE AGENCY ADMINISTRATORS

AND REGIONAL COMMISSIONERS

FROM : JAY A. MOUSA

Associate Commissioner

Office of Field Operations

SUBJECT : Instructions for Monthly Local Area Unemployment Statistics

(LAUS) Estimation in 2020

Purpose: This memorandum contains instructions and the outline of the schedule for 2020 monthly production, finalization, and release of LAUS estimates.

Background: BLS is required by Office of Management and Budget Statistical Policy Directive No. 4 to publish monthly estimates on a timely basis. In order to facilitate that goal, it is necessary to establish and follow a schedule of important monthly activities and delivery dates. This memorandum describes the necessary activities and schedule; a comprehensive schedule of dates for these activities is provided in LAUS Technical Memorandum No. S-20-01-rev, Schedules for Monthly and Annual Processing in 2020, dated December 4, 2019 and revised December 30, 2019.

Overview of 2020: Most aspects of production and publication will parallel those of 2019:

• The current version of the State Time Series Analysis and Review System (STARS) and current LAUS models will be used for monthly production in 2020.

• BLS will continue to load Current Employment Statistics (CES) estimates into LAUSToo, in accordance with LAUS Technical Memorandum No. S-19-12.

• Geography changes will be discussed in a forthcoming LAUS technical memorandum.

• The Residency Assignment System (RAS) will begin using the new 2019 Census geographic boundaries and codes once updates are completed late in the day on January 27.

• An updated version of the Program for Measuring the Insured Unemployed Statistics (PROMIS) software, reflecting all necessary geographic changes, will be issued in January and should be used for production of 2020 estimates.

• The reference week for all months—January through December—will be the standard week including the 12^th day of the month.

• Emphasis continues to be placed on states providing comments on data that pass the Questionable Data Edits (QDEs) overall, but fail both over-the-month tests or both over-the-year tests by large margins.

Instructions for monthly estimation:

1. Model-based estimation:

Estimates for all states and the District of Columbia, the Los Angeles-Long Beach-Glendale Metropolitan Division, New York City, the Balance of California, and the Balance of New York State will be produced using the latest LAUS models with real-time benchmarking to national Current Population Survey (CPS) totals via census division models. BLS will update the STARS databases shortly after 8:30 a.m. Eastern Time (ET) on the day of the Employment Situation news release, enabling states to view or develop simulated monthly model-based estimates.

The official model-based estimates will be developed using the CPS data, along with the following state-entered inputs: Current Employment Statistics (CES) total nonfarm wage and salary employment estimates and Unemployment Insurance (UI) claims counts. States should continue to enter counts of Unemployment Compensation for Federal Employees (UCFE) claims and labor disputants, even though they are not used for developing model-based estimates. States should transmit preliminary CES, labor disputant, UI, and UCFE data for January 2020 estimation by February 28. In subsequent months, preliminary input data should be transmitted for the “current” month and revised input data for the prior month. Because of the interdependence involved, a state’s official estimates will not be available until all states in its census division have submitted their inputs and the national office has created the estimates.

Models will be used for five additional substate areas and their respective state balances:

• Chicago-Naperville-Arlington Heights, Illinois Metropolitan Division;

• Cleveland-Elyria, Ohio Metropolitan Statistical Area;

• Detroit-Warren-Dearborn, Michigan Metropolitan Statistical Area;

• Miami-Miami Beach-Kendall, Florida Metropolitan Division; and

• Seattle-Bellevue-Everett, Washington Metropolitan Division.

For California and New York, where there are no statewide models, the statewide estimates, both smoothed seasonally adjusted and not seasonally adjusted, are derived by summing the substate area and balance-of-state model estimates. In contrast, not seasonally adjusted data for the other five areas and their respective balances of state will be forced to sum to the separately calculated model-based estimates (reflecting real-time benchmarking) for the state.

States are reminded that, for the production of both preliminary and revised model-based estimates, only the official CES employment data should be used, unless an atypical/exception has been granted. The national office compares the STARS inputs with the CES estimates to ensure conformity. Any differences beyond those attributable to rounding are likely to delay the production of estimates for all states in the census division.

2. Substate estimation:

A. As with the statewide estimates, preliminary substate estimates will be produced for all months, and revised estimates will be produced for January–November. BLS will load both preliminary and revised CES employment inputs for all CES areas (primarily metropolitan areas and metropolitan divisions). States will enter both preliminary and revised UI and UCFE counts, total nonfarm wage and salary employment (M01) estimates for non-CES areas, and labor disputant (M02) estimates for both CES and non-CES areas. Revised UI and UCFE counts for a given month will likely differ from preliminary counts for the same month because they are more complete. Revised non-CES area M01 inputs for a given month may or may not differ from the preliminary inputs for the same month, depending on the methodology used to produce them. For example, states that use the Composite Quota Estimator (CQE) in ACESweb to derive non-CES area M01 should incorporate the monthly revisions at all levels of geography each month. States that use the BLS-provided Small Area Employment Estimates (SAEE) should incorporate monthly revisions when a new quarter of estimates becomes available.

B. Monthly and quarterly BLS-supplied substate inputs will be loaded into LAUSToo by BLS according to the schedule in attachment 2 of S-20-01-rev. BLS-supplied inputs include (1) statewide unemployed new entrants and reentrants, (2) statewide agricultural employment and all-other employment, (3) Railroad Retirement Board (RRB) claims, and (4) survival rates and unemployment rate ranges for unemployed exhaustee estimation. Of those, only RRB claims counts will be revised each month. Inputs that are provided by BLS once per year for annual revisions and monthly production are discussed in separate memoranda.

C. Given the complexity of interstate areas and the use of Dynamic Residency Ratios (DRRs), the state-to-state exchange of data continues to be critical. LAUSToo is designed to facilitate this exchange. Attachment 3 of S-20-01-rev contains a schedule of dates for the exchange of various data items.

D. The reference periods for the continued claims counts and final payments data are provided in attachment 4 of S-20-01-rev. As noted above, reference periods for all months will be the standard week including the 12^th day of the month.

3. State finalization and due dates:

LAUS due dates for 2020 are similar to those of recent years. As is typically the case, due dates for January statewide inputs and substate estimates and February substate estimates (but not statewide inputs) will be relatively later than for other months, due to the time required to perform annual processing. The due dates for other months will be based on the “regular” schedule, typically the Friday after the Employment Situation release for statewide inputs and the Friday two weeks after the Employment Situation release for substate estimates. See attachment 1 of S-20-01-rev for specific dates.

The January due date will pertain to January preliminary data only. Revised data for the prior month, December 2019, will be estimated as part of annual processing. Due dates for February through December will pertain to preliminary data for the “current” month and revised data for the prior month.

Timely completion of LAUS data is a requirement in the Labor Market Information (LMI) Cooperative Agreement. Therefore, states are required to transmit their inputs in STARS and finalize substate area estimates in LAUSToo by the due dates—and are encouraged to do so earlier whenever possible. However, inputs and estimates should not be transmitted or finalized early unless they have been thoroughly reviewed.

1. Model-based estimates:
   
   States will use STARS to transmit monthly inputs for statewide models no later than 12:00 noon ET on the dates specified in attachment 1 of S-20-01-rev. The due dates for the District of Columbia, model-based substate and balance-of-state areas, and Puerto Rico will be the same as those for states. (Estimates for Puerto Rico are developed from a household survey and not through STARS, although the Commonwealth will transmit estimates by entering them into STARS.) States should alert their regional office if the availability of any input has been delayed. Given the interdependence of estimation within census divisions, timely entry of model inputs is critical.

Monthly DEMECON files containing demographic and economic characteristic data will be made available to each state via EUSWeb shortly after the Employment Situation release at 8:30 a.m. ET.  Twelve-month moving-average data for the same components will be made available to each state at the same time.  On the day of the Employment Situation, each state will receive data for that state only.  Shortly after the corresponding State Employment and Unemployment news release, all 12-month moving-average tables will be made available to all states.

B. Substate estimates:

Substate estimates generally are due on the Friday of the week after the statewide inputs are due, as shown in attachment 1 of S-20-01-rev. February estimates are the exception; they are due two weeks after the statewide inputs are due. Estimates should be finalized by 4:45 p.m. ET.

States will need to have effective communication with nearby states in order to ensure timely data finalization. The schedule in attachment 3 of S-20-01-rev should help in this regard, as should the data exchange functionality in LAUSToo. It should be noted that those dates are generally the latest dates that will allow states to meet the BLS due date. States may want to share data items with neighboring states earlier than those dates, especially if a state has an early publication schedule.

States will not be able to finalize estimates unless they have run all of the required edits and entered comments, where applicable. As always, clear and meaningful comments should be provided regarding any questionable data, not just those that fail the QDEs. In addition to overall failures, states should examine data values that fail both over-the-month tests or both over-the-year tests by large margins and provide comments. These results, which are shown in yellow on the screen, also may indicate erroneous data. Edits are to be run on preliminary data for January–December and revised data for January–November.

C. Late finalization or corrections:

In the case of statewide inputs, states must notify their regional office if they believe that a correction is necessary. Because of the interdependence of estimation, it is likely that a correction will affect the other states in the census division.

Regarding substate estimates, states should contact their regional office as early as possible if they know that they will be finalizing substate data after the due date. BLS can extend the processing period in LAUSToo for selected states, if needed. Data corrections more than one week after the due date will be submitted using the LAUSToo corrections module.

BLS publication of LAUS data: Publication dates for the first two reference months of the year will be relatively later than for other months, due to the time required to complete annual processing. Following publication of February substate data, the schedule becomes much more “regular” for March–December, though it is still affected by holidays and conflicts with other BLS releases.

Both monthly news releases—State Employment and Unemployment and Metropolitan Area Employment and Unemployment—will be available via the BLS website at 10:00 a.m. ET on the dates that they are issued. (See attachment 1 of S-20-01-rev.)

In the January and February Metropolitan Area news releases, the prior year’s data will be the annual processing data that have been only partially reviewed and are not yet in LABSTAT. Those data are the most comparable to the 2020 estimates. A box note in the releases will highlight this.

The national office will update the time series database on the BLS website with preliminary data for the current reference month, revised data for the prior month (except when December is the prior month), and any historical revisions or corrections on the dates listed in attachment 1 of S-20-01-rev. The issuance dates for all-area (also known as “small area”) data will coincide with those for the Metropolitan Area news release.

Regional offices and states should access LABSTAT or LAUSToo’s LAUS Estimates (BLS RELEASED) report to obtain published BLS data for areas outside their region/state. For security reasons, LABSTAT may not be available until approximately 10:30 a.m. ET. Once the time series database has been updated, the “At a Glance” and “Latest Numbers” features of the BLS website are updated. In LAUSToo, data for all states will become available by approximately 10:30 a.m. ET as well.

Smoothed seasonally adjusted metropolitan area estimates: BLS will continue to provide “unverified” smoothed seasonally adjusted metropolitan area estimates on the Monday following the substate due date. These estimates are based on not seasonally adjusted substate data that are still under review. “Verified” estimates generally will be provided on Thursday of that week, barring any outstanding issues with the not seasonally adjusted data. BLS will publish these data on the LAUS website in conjunction with the Metropolitan Area news release.

Alternative measures of labor underutilization: The national office will continue its practice of issuing four-quarter moving averages of the alternative measures of labor underutilization, also known as U-1 through U-6, at the state level each quarter. Issuance of those data will occur on the Friday after the State Employment and Unemployment news release is issued for the last month of each quarter. See S-20-01-rev for specific dates.

Actions required: States should transmit STARS inputs (including monthly revisions); enter LAUSToo inputs (including monthly revisions); and generate, edit, and finalize LAUS estimates according to the instructions and schedules contained in this memorandum and S-20-01-rev.

Workload impact: Neutral.

Inquiries: All inquiries should be directed to the appropriate regional office.

Effective date: These instructions are effective immediately.