LEHD PRA Federal Register Response

On December 8, 2016 the Census Bureau published a Notice in the Federal Register (i.e., pages 88662-88663, Citation 81 FR 88662, document number: 2016-29395) inviting public comments on the Census Bureau’s proposed information collection, the Longitudinal Employer-Household Dynamics (LEHD) program. A single comment was received that questioned the underlying data quality, the importance of the LEHD statistical products, and the additional burden that participation in the LED Partnership might impose. The Census Bureau has reviewed this comment and provides the following response.

Underlying Data Quality

The LEHD program is fully aware of its responsibility to provide data that meets statistical quality standards. To achieve this goal the program has standard operating procedures that determine the format of the files to be submitted, verify file contents, and define the submission protocols of the data that is provided by its partners. Each secure data submission needs to pass multiple conformity checks and statistical consistency checks. When a state first enters the partnership, statistics and tabulations for that state are thoroughly reviewed and approved by both the Census Bureau and the State’s Labor Market Information (LMI) agency prior to its initial public release. Additionally the LEHD program, with support from its state partners, conducts the following activities to insure the quality of the data being released:

• Continual research focused on improving the quality of the statistics the Census Bureau provides. The resulting output of this research is submitted to academic and internal peer review, as well as discussed with state partners in monthly calls and annual conferences. Research is currently being conducted into ways to mitigate bias when workplaces are not broken out or are absent from the data for a variety of reasons, when individual wage records are not reported, and when establishment-level records contain potentially inconsistent data.

• LEHD supports research which may lead to new statistical products as well as improvements to the current products. Participation by states and other agencies is voluntary and recent projects addressed working with Unemployment Insurance benefit records to better identify what happens when a state's workers are not working; and, working with records from a state's educational tracking systems to assess cross-border work activity of graduates.

• At any point in time, one or two states may be temporarily missing from the latest release due to observed quality indicators that are constantly monitored by the LEHD program. Issues may be abnormal low or high record counts or changes in administrative systems that prevent historical linkages. These data quality issues are reviewed with a state’s staff that may require additional or replacement data submissions. The typical resolution time is 1-2 quarterly production cycles. The Census Bureau’s production includes processes that allow for the missing data for a specific state to be included in subsequent releases once the data quality issue has been resolved. There has been no case where a data quality issue has not been resolved to the mutual satisfaction of a state partner and the Census Bureau.

Importance of the LEHD Statistical Data Products

The data products developed by the LEHD program provide statistics on employment, earnings, and job flows at detailed levels of geography and industry and for different demographic groups. The potential uses of these data products and their supporting dissemination tools are far-reaching, both for unraveling many important questions in economic research and for the provision of new statistical products. For example:

• Census Bureau web server metrics show over 105,000 visits to its dissemination tools between January 1, 2017 and June 1, 2017.

• The New Jersey State Data Center used OnTheMap for Emergency Management to quickly learn the impact of hurricane Sandy with regards to identification of Federal Disaster Declaration Areas and its affects on communities (i.e., population and workforce).

• The state of Nevada has used the Job-to-Job Flows data product to understand the migration of its workforce that supports the hotel industry.

• The Department of Transportation encourages the use the LEHD data products and supporting dissemination tools for planning: https://www.fhwa.dot.gov/planning/census_issues/lehd/.

• A Bloomberg Business article provided an analysis using the LEHD data that showed the share of industry workers by age, based on a moving average of seasonally adjusted data.

• The Philadelphia Center City District used LEHD data to understand the details of the area’s workforce and economy in order to monitor the effectiveness of economic programs and policy initiatives.

Additional examples of how the LEHD data products and supporting dissemination tools have been used can be found at the LEHD website: https://lehd.ces.census.gov/led_in_action/.

Additional burden to states’ resources

An objective of the LEHD program is to reduce burden in the development of its unique statistical products. This reduced burden incudes the elimination of traditional surveys and working to minimize man hours required to support this program by its data providing partners. For participating states, this data sharing partnership is voluntary and is defined via a Memorandum of Understanding (MOU), which establishes how and what data is provided to the Census Bureau, how the Census Bureau can use this data, and products that can be developed. The Census Bureau has worked to minimize the time required for the processing of the MOU by creating a template that outlines the specific requirements on provision, use, and products. This template has helped in reducing time required of both Census and state legal staff to review, incorporate specific changes, and approve.

It should also be noted that the LED partnership renewal cycle was changed by the Census Bureau from 10 years to 5 in order to provide state partners a means to adjust to changes (e.g., legislative, available data, etc.). The partnership has begun the process for voluntary renewals to the 5 year MOU and as of April 2017, 44 states have renewed their membership and 4 states are in the process for renewal. There are 5 states under the older 10 year MOU that will be renewing in the next year. One state, Wyoming, has withdrawn from this data sharing partnership.

In summary, the LEHD program uses existing administrative data to create new and relevant statistical products that is cost effective with minimal respondent burden.