The LEHD program maintains an active research program oriented on the use of longitudinally linked employer-employee data. Use of administrative data in the national statistical system is fairly new, and a core mission of LEHD research is to enhance the infrastructure for use of administrative data in the production of federal statistics. This includes activities such as developing statistical matching and imputation methods for data linkage, research comparing administrative and survey data to understand sources of error in each, and developing new prototypes for data products. LEHD economists share their research at academic conferences and publish in peer-reviewed journals and other scholarly outlets.
Research using LEHD microdata is also carried out by qualified academic researchers under approved projects using a secure network of Research Data Centers (RDCs). The RDC system is administered by the U.S. Census Bureau's Center for Economics Studies (CES). Please use the following link to view more information on accessing LEHD data through the RDCs.
While the importance of young small businesses to job creation and productivity growth is increasingly well understood, relatively little is known about the characteristics of the jobs generated by young firms. This project links microdata from the Census Bureau's Business Dynamics Statistics (BDS) to the LEHD linked employer-employee data. This enables us to do two things. First, we will add firm age and size to the set of employer characteristics available in LEHD-based public use data products QWI and LODES/OnTheMap. Second, we will add this firm data to LEHD microdata available to qualified researchers working on approved projects in the RDCs. This enhancement will allow for the first time a detailed look at jobs, earnings, and employment turnover generated by firms of varying ages and sizes, as well as the demographic characteristics of their workforce.
Beta QWI with firm age and size information began release in spring of 2012, and firm age and size are expected to be included in the 2013 release of LODES/OnTheMap and in the 2013 update to the LEHD microdata in the RDCs.
Employee movements between jobs, or job-to-job flows, constitute one of the major opportunities for the development of new statistical data products. An ongoing initiative of the LEHD program has been to categorize and tabulate the movements between employers present in the LEHD linked employer-employee data. This enhancement will allow a comprehensive picture of employee movements between employers, as well as movements between industries and across geographies, and across spells of nonemployment. Planned releases are for output data to be available by demographic and firm characteristics.
Prototype databases have been produced and analyzed, both to understand how to categorize job-to-job flows as well as to gage their economic significance. Major work on this product is planned for 2013 and 2014, with an initial release during 2014.
The most current QWI data cover 49 states and the District of Columbia, and about 98% of the private workforce in each of those states. However, due to incomplete data (Massachusetts has not submitted data to the Census Bureau as of this date) and inconsistent time-series availability across states, there is no national series available for the entire time series.
LEHD researchers have used the existing public-use data to construct the first national estimates for QWI. The national estimates are an important enhancement to existing series because they are the first series of both job and worker flows created from a single data source that include demographic and industry detail, compiled from data that have been integrated at the micro-level. Ongoing development focusses on creating national QWI by sex and education, as well as by race and ethnicity, using the same methodology.
Prototype national data can be downloaded here: http://www.vrdc.cornell.edu/qwipu.national
Until quite recently, federal workers and self-employed jobs were not available in the LEHD jobs data, due to their exclusion from state unemployment insurance programs. LEHD researchers have been working to fill this gap by integrating two additional data sources into the LEHD microdata: Office of Personnel Management (OPM) records on the federal workforce and tax data on the self-employed. The OPM data has been successfully integrated into the LEHD microdata, and were included in the spring 2012 release for LODES. Initial beta QWI time-series releases for federal workers began on a state-by-state basis starting in 2012.
Inclusion of the self-employed workforce in LEHD data reveals the high degree of interconnection between the wage and salary workforce and those who start their own businesses. The expanded frame permits a separate characterization of workforce dynamics for this group and makes it possible to produce statistics that capture the rich flows between wage and salary work and business ownership. Work on new data products to capture these dynamics is in early stages of development.
The potential benefits to Census Bureau program areas from linked survey administrative data are substantial, and yet largely untapped. In a time of declining survey response rates and budgetary cutbacks, integrated data has enormous potential to improve data quality and lower costs. A goal of the LEHD researcher program is to build the necessary knowledge and data infrastructure to allow administrative data to enhance Census Bureau survey program areas.
In addition to developing methodologies for linking survey and administrative data, LEHD researchers study discrepancies between survey responses and administrative data to better understand the sources of error in each. A recent project uses LEHD data linked to the Current Population Survey (CPS) to explore discrepancies between the two data sources. A newer project matches jobs in the LEHD data to the American Community Survey (ACS), with particular focus on improving data quality on place of work in both data sources. Construction of the linked job file will lay the groundwork for data quality improvements to LODES, QWI, and the Census Transportation Planning Package.