1. Introduction

The Longitudinal Employer-Household Dynamics (LEHD) data at the U.S. Census Bureau is a quarterly database of linked employer-employee data covering over 95% of employment in the United States. [1] The LEHD data are generated by merging previously collected survey and administrative data on jobs, businesses, and workers. By integrating administrative data with existing census and surveys, a national longitudinal jobs database for the U.S. is generated at very low cost and with no additional respondent burden. The LEHD microdata is available to approved researchers in the Federal Statistical Research Data Center (FSRDC) network. LEHD data is also available to the public in tabular form in the Census Bureau’s Quarterly Workforce Indicators (QWI), LEHD Origin-Destination Employment Statistics (LODES), Job-to-Job Flows (J2J), and the Post-Secondary Employment Outcomes (PSEO) data.

In order to facilitate researcher use of the microdata, LEHD infrastructure files are periodically collected into a standardized form known as the LEHD Snapshot, which is then released into the FSRDC network. The purpose of this documentation and codebook is to provide a limited overview of the components of the LEHD Snapshot files, as well as detailed codebooks for researcher use. For an overview of the LEHD infrastructure, see Abowd et al. [2009]. This documentation is not intended to provide details of specific LEHD data products, or of the disclosure limitation measures used to produce those products. For more information on specific LEHD data products and their disclosure limitation techniques, see Abowd et al. [2006], Haney et al. [2017], Machanavajjhala et al. [2008] .

The LEHD Snapshot consists almost exclusively of restricted-use administrative microdata. Therefore, as with all projects using restricted data, any projects using the LEHD Snapshot must undergo standard review and approval processes. Researchers are required to demonstrate that they cannot conduct their research using public-use data sources, and they must identify the ways in which their project will benefit both the objectives of the U.S. Census Bureau and the broader research community. Before requesting access to the LEHD Snapshot, researchers are encouraged to examine the public data offerings of the LEHD program at https://lehd.ces.census.gov/data/, as well as the broader public offerings at https://data.census.gov/. For more details on the FSRDC program and the requirements for project approval, see the FSRDC about page as well as the approval guide below.

This documentation is current as of LEHD Snapshot release S2024_R2025Q4, which contains available data through 2024. Features may change across Snapshot releases. Updated versions of the Snapshot documentation are published in HTML format, and the most recent Snapshot is publicly available at https://lehd.ces.census.gov/data/lehd-snapshot-doc/latest/. References to documentation for earlier Snapshot releases can be found below. For details on how to cite the data and this documentation, see below.

1.1. Earnings Coverage

Covered private-industry employment in the LEHD data includes most corporate officials, all executives, all supervisory personnel, all professionals, all clerical workers, many farmworkers, all wage earners, all pieceworkers, and all part-time workers. Workers on paid sick leave, paid holiday, paid vacation, and the like are also covered. Workers on the payroll of more than one firm during the period are counted by each employer that is subject to UI, as long as those workers satisfy the preceding definition of employment. Workers have UI wages filed in every quarter they are covered, even though their wages may not be subject to UI tax in the latter months of the year.

Notable exclusions from UI coverage among private sector employers are independent contractors, the unincorporated self-employed, railroad workers covered by the railroad unemployment insurance system, some family employees of family-owned businesses, certain farm workers, students working for universities under certain cooperative programs, salespersons primarily paid on commission, and workers of some non-profits. States have some leeway in designating coverage, for a complete list, see the coverage section of the most recent Comparison of State UI Laws.

Covered employment in the LEHD data includes most employees of state and local governments with the exception of elected officials, members of a legislative body or members of the judiciary, and some emergency employees.

Federal government workers are not covered by state UI and do not appear in UI wage record data.

1.2. LEHD Confidential Microdata: Job, Employer, and Person-Level Files

The LEHD Snapshot is organized into sets of tables providing job-level, employer-level, and person-level data. The scope of each table is either national (all states) or a single state. National tables will have a US in the table name, while state tables will have the two-character state postal code. For purposes of this documentation, the postal code will be replaced with ZZ. Tables that are national in scope will not have the complete panel of states in all quarters because of limited availability of state job data both at the beginning and end of the time series. For more details, see Availability of Data.

1.2.1. Job-Level Files

Job-level data provide earnings for jobs, which is a link between a worker and a firm. The employment coverage in the LEHD job-level microdata files described below is UI-covered employment only (see Earnings Coverage for details on covered employment and wages). Researchers interested in federal jobs should request the (OPM) microdata separately as part of their research request.

Job-Level Files in the LEHD Snapshot

Table	Description	Scope	Key
Employment History File (EHF_ZZ)	Quarterly earnings for each job in a state as reported by the employer to the state’s UI system.	State	pik sein seinunit year
Job History File (JHF_ZZ)	Wide version of earnings history that contains imputed establishment (SEINUNITs) for multi-unit firms, and a job-level ID useful for tracking job spells across employer identifier changes.	State	pik sein spell_u2w
Earnings Indicators (EHF_US_INDICATORS)	Indicates whether or not a PIK had any earnings reported in any state in the LEHD system.	National	pik year
Earnings Availability (EHF_ALL_AVAILABILITY)	Indicates data availability ranges for all states.	National	state

1.2.2. Employer-Level Files

Employer-level data contain characteristics of firms, as well as summary measures calculated from job data.

Employer-Level Files in the LEHD Snapshot

Table	Description	Scope	Key
Employer Characteristics File, Establishment (ECF_ZZ_SEINUNIT)	An establishment level file with industry, size, and location.	State	sein year quarter seinunit
Employer Characteristics File, Firm (ECF_ZZ_SEIN)	Sourced from the SEINUNIT file, this file rolls up employer characteristics to the SEIN level.	State	sein year quarter
Employer Characteristics File, National Firm (ECF_ZZ_SEIN_T26)	Contains the national firm age and size data for employers in the ECF. This file requires IRS approval.	State	sein year quarter
Quarterly Workforce Indicators (QWI_ZZ_SEINUNIT)	The microdata version of the public use QWI data, containing establishment-level hires, separations, net job growth, and average earnings.	State	sein year quarter seinunit
Successor-Predecessor File (SPF_ZZ)	Flows of separations and hires between employers (SEINs) used to identify successor-predecessor relationships.	State	sein sein_succ qtime

1.2.3. Person-Level Files

Person-level tables provide demographic characteristics of individuals and observed residence geography.

Person-Level Files in the LEHD Snapshot

Table	Description	Scope	Key
Individual Characteristics File (ICF_US)	Demographics, place of birth, and education of workers.	National	pik
Date of Birth/Sex/Place of Birth Implicates (ICF_US_IMPLICATES_AGE_SEX_POB)	Multiply imputed date of birth, sex, and place of birth variables to complete missing information.	National	pik
Race/Ethnicity Implicates (ICF_US_IMPLICATES_RACE_ETHNICITY)	Multiply imputed race and ethnicity variables to complete missing information.	National	pik
Education Implicates (ICF_US_IMPLICATES_EDUCATION)	Multiply imputed education variables to complete missing information.	National	pik
Residence Geography, 1999-2011 (ICF_US_RESIDENCE_CPR)	Residence geography prior to 2012 for individuals found in the wage data.	National	pik address_year
Residence Geography, 2012 forward (ICF_US_RESIDENCE_RCF)	Residence geography from 2012 forward for individuals found in the wage data.	National	pik address_year

1.3. Linking the job, employer, and person data

• To link worker demographics to the jobs data, link the EHF or JHF to the ICF US via the PIK.

• To attach the characteristics of the employer (SEIN) to the job, link the EHF or JHF to the ECF SEIN using the SEIN.

• To attach the characteristics of the establishment to the job, use the first implicate of the worker-to-establishment imputation (SEINUNIT1) on the JHF to link to the ECF SEINUNIT table.

  • All implicates (SEINUNIT1-SEINUNIT10) can be used to produce standard errors that account for imputation variability [McKinney et al., 2021].

  • Because of longitudinal restrictions on establishment imputation, jobs may be active at establishments in quarters that the establishment does not appear on the ECF. See Adding Establishment Characteristics for more information.

1.4. Approvals Needed for FSRDC Research

In addition to Census data, the administrative records that combine to form the LEHD infrastructure derive from several different federal and state agencies. Approval of a proposed FSRDC project by other agencies may be required, as follows:

• Requests for state-level tables will follow the procedure specified by the Memorandum of Understanding (MOU) with that state. Most commonly, the state agency is given the opportunity to approve proposed projects.

• Person-level demographic characteristics (ICF) require the approval of the Social Security Administration (SSA).

• Employer-level Title 26 (T26) data requires the additional approval of the Internal Revenue Service (IRS). This includes the Federal EIN, as well as age and size of the national firm.

The codebook for each table will indicate which approvals are required.

The job-level tables largely consist of state-provided data and so go through the LEHD state data owner approval process. The indicator and availability files do not require state approval. There is no federal tax information or Social Security administrative data in any of the job-level tables, so no additional IRS or SSA approvals are required.

1.5. Snapshot Packages in the FSRDC

LEHD Snapshot tables are delivered in packages according to the level of data in the tables. Tables that require additional approvals are provided in separate packages.

Snapshot Packages in the FSRDC

Package	Tables
Jobs Data	EHF_ZZ JHF_ZZ EHF_US_INDICATORS EHF_ALL_AVAILABILITY
Employer Data (non-T26)	ECF_ZZ_SEIN ECF_ZZ_SEINUNIT SPF_ZZ QWI_ZZ_SEINUNIT
Employer Data (T26)	ECF_ZZ_SEIN_T26
Person Demographics (Workers Only)	ICF_US ICF_US_IMPLICATES_AGE_SEX_POB ICF_US_IMPLICATES_RACE_ETHNICITY ICF_US_IMPLICATES_EDUCATION
Person Residence (Workers Only)	ICF_US_RESIDENCE_CPR ICF_US_RESIDENCE_RCF

1.6. Snapshot Data Formats

The LEHD Snapshot is provided in a set of SAS data tables. Beginning with the S2024 release, the Snapshot is also provided in the open source Parquet data format, with select metadata tables provided as comma-delimited text files rather than Parquet. Researchers who intend to conduct their analyses in tools other than SAS are encouraged to use the Parquet versions of snapshot files to limit the need for file conversion processing and to reduce data storage requirements for intermediate files. All Parquet files can be found in a parquet subdirectory alongside the SAS tables.

Parquet and SAS versions of files are functionally identical; minor differences in data storage types are indicated in this documentation. However, to simplify within-memory processing of the largest Snapshot datasets, worker- and job-level tables are partitioned into separate Parquet files based on the first two characters of the PIK, and these two characters have been added as a suffix to the corresponding file names (e.g., ehf_dc_00.parquet for DC, PIK group 00). PIKs where the first two characters are not numeric are placed into a file with an AA suffix. An AA suffix file may not exist if there are no PIKs with non-numeric characters.

A detailed discussion of working with the Parquet data format is beyond the scope of this documentation. However, popular libraries in R and Python are able to read and write the Parquet format directly, and cross-platform tools such as DuckDB may make Parquet data readily accessible to other programming languages as well. Several libraries also accept wildcard characters for filename partitions, so that PIK group partitioning can be used or ignored as needed depending on your use case. For example, the below example uses the Python Polars library to count the number of distinct PIKs found in the ICF_US. This can be performed all at once using filename wildcards, or it can be performed by adding up the number of distinct PIKs in each PIK group file.

import polars as pl

# Read all the ICF_US files into memory at once to get the total number of
# PIKs. Note that the ICF_US is already unique by PIK.
n_piks = pl.read_parquet('/path/to/icf/parquet/icf_us_??.parquet').height

# Create a list of all the PIK groups for looping, with leading zeros for the
# integer values.
pik_groups = [f'{x:02d}' for x in range(0,100)] + ['AA']

# Loop through PIK groups to count the number of PIKs in each group and add
# it to the total
n_piks_by_group = 0
for group in pik_groups:
   n_piks_by_group += (
      pl.read_parquet(f'/path/to/icf/parquet/icf_us_{group}.parquet').height
   )

# Confirm that summing the individual PIK groups yields the same number of
# total PIKs. This should return True
n_piks_by_group == n_piks

1.7. Quarter Indexing - qtime

On several LEHD tables, the year and quarter is replaced by or supplemented with an index for ease of counting quarters in longitudinal references. This is initialized with 1985Q1 being set to 1. The following SAS formulae can be used to switch back and forth:

• qtime = (year-1985) * 4 + quarter

• year = int((qtime-1)/4) + 1985

• quarter = mod(qtime-1, 4) + 1

For a table of all possible qtime values with year/quarter references see codebook.

1.8. Availability of Data

The longitudinal availability of data in the LEHD Snapshot will vary by type of data and the data source, as described in this section.

1.8.1. Job and Employer Data (State Provided)

The following table indicates the availability of LEHD data by state for the EHF, ECF, and QWI tables. This includes all states that are or have been members of the LED partnership. In general, LEHD snapshot release S2024_R2025Q4 contains all available data through 2024.

• The EHF range represents all of the quarters of UI wage data that have been received.

• The QWI range may begin later than the EHF range if certain data quarters do not meet quality standards for publication in LED data products.

  • Earnings quarters on the JHF are restricted to the QWI range.

• The ECF range may start before the EHF range if employer-level data submissions (QCEW) are available prior to the beginning of the UI wage data series.

• End dates will usually be the same for all tables within a state. States that are not active partners may not be available for the latest quarters.

Available Quarters of LEHD State Provided Data

State	ECF Start	EHF Start	QWI/JHF Start	SPF Start	ECF End	EHF End	QWI/JHF End	SPF End
Alabama	2001Q1	2001Q1	2001Q1	2000Q3	2025Q1	2025Q1	2025Q1	2025Q3
Alaska	1990Q1	1990Q1	2000Q1	1989Q3	2016Q2	2016Q2	2016Q2	2016Q4
Arizona	1992Q1	1992Q1	2004Q1	1991Q3	2025Q1	2025Q1	2025Q1	2025Q3
Arkansas	2002Q3	2002Q3	2002Q3	2002Q1	2025Q1	2025Q1	2025Q1	2025Q3
California	1991Q1	1991Q3	1991Q3	1990Q3	2025Q1	2025Q1	2025Q1	2025Q3
Colorado	1990Q1	1990Q1	1993Q2	1989Q3	2025Q1	2025Q1	2025Q1	2025Q3
Connecticut	1996Q1	1996Q1	1996Q1	1995Q3	2025Q1	2025Q1	2025Q1	2025Q3
Delaware	1997Q1	1998Q3	1998Q3	1996Q3	2025Q1	2025Q1	2025Q1	2025Q3
District of Columbia	2000Q4	2002Q2	2005Q2	2000Q2	2025Q1	2025Q1	2025Q1	2025Q3
Florida	1989Q1	1992Q4	1997Q4	1988Q3	2025Q1	2025Q1	2025Q1	2025Q3
Georgia	1994Q1	1994Q1	1998Q1	1993Q3	2025Q1	2025Q1	2025Q1	2025Q3
Hawaii	1995Q4	1995Q4	1995Q4	1995Q2	2025Q1	2025Q1	2025Q1	2025Q3
Idaho	1990Q1	1990Q1	1991Q1	1989Q3	2025Q1	2025Q1	2025Q1	2025Q3
Illinois	1990Q1	1990Q1	1993Q2	1989Q3	2025Q1	2025Q1	2025Q1	2025Q3
Indiana	1990Q1	1990Q1	1998Q1	1989Q3	2025Q1	2025Q1	2025Q1	2025Q3
Iowa	1990Q1	1998Q4	1998Q4	1989Q3	2025Q1	2025Q1	2025Q1	2025Q3
Kansas	1990Q1	1990Q1	1993Q1	1989Q3	2025Q1	2025Q1	2025Q1	2025Q3
Kentucky	1996Q4	1996Q4	2001Q1	1996Q2	2025Q1	2025Q1	2025Q1	2025Q3
Louisiana	1990Q1	1990Q1	1995Q1	1989Q3	2025Q1	2025Q1	2025Q1	2025Q3
Maine	1996Q1	1996Q1	1996Q2	1995Q3	2025Q1	2025Q1	2025Q1	2025Q3
Maryland	1985Q4	1985Q4	1990Q1	1985Q2	2025Q1	2025Q1	2025Q1	2025Q3
Massachusetts	2002Q1	2002Q1	2010Q1	2001Q3	2024Q3	2024Q3	2024Q3	2025Q1
Michigan	1998Q1	1998Q1	2000Q3	1997Q3	2021Q4	2021Q4	2021Q4	2022Q2
Minnesota	1994Q3	1994Q3	1994Q3	1994Q1	2025Q1	2025Q1	2025Q1	2025Q3
Mississippi	2003Q3	2003Q3	2003Q3	2003Q1	2025Q1	2025Q1	2025Q1	2025Q3
Missouri	1990Q1	1990Q1	1995Q1	1989Q3	2024Q2	2024Q2	2024Q2	2024Q4
Montana	1993Q1	1993Q1	1993Q1	1992Q3	2025Q1	2025Q1	2025Q1	2025Q3
Nebraska	1999Q1	1999Q1	1999Q1	1998Q3	2025Q1	2025Q1	2025Q1	2025Q3
Nevada	1998Q1	1998Q1	1998Q1	1997Q3	2025Q1	2025Q1	2025Q1	2025Q3
New Hampshire	2003Q1	2003Q1	2003Q1	2002Q3	2025Q1	2025Q1	2025Q1	2025Q3
New Jersey	1995Q1	1996Q1	1996Q1	1994Q3	2025Q1	2025Q1	2025Q1	2025Q3
New Mexico	1990Q1	1995Q3	1995Q3	1989Q3	2025Q1	2025Q1	2025Q1	2025Q3
New York	1990Q1	1995Q1	2000Q1	1989Q3	2025Q1	2025Q1	2025Q1	2025Q3
North Carolina	1990Q1	1991Q1	1992Q4	1989Q3	2025Q1	2025Q1	2025Q1	2025Q3
North Dakota	1998Q1	1998Q1	1998Q1	1997Q3	2025Q1	2025Q1	2025Q1	2025Q3
Ohio	1994Q1	1994Q1	2000Q1	1993Q3	2025Q1	2025Q1	2025Q1	2025Q3
Oklahoma	1999Q1	2000Q1	2000Q1	1998Q3	2025Q1	2025Q1	2025Q1	2025Q3
Oregon	1990Q1	1991Q1	1991Q1	1989Q3	2025Q1	2025Q1	2025Q1	2025Q3
Pennsylvania	1991Q1	1991Q1	1997Q1	1990Q3	2025Q1	2025Q1	2025Q1	2025Q3
Puerto Rico	2001Q1	2001Q1	2010Q1	2000Q3	2025Q1	2025Q1	2025Q1	2025Q3
Rhode Island	1990Q1	1995Q1	1995Q1	1989Q3	2025Q1	2025Q1	2025Q1	2025Q3
South Carolina	1998Q1	1998Q1	1998Q1	1997Q3	2025Q1	2025Q1	2025Q1	2025Q3
South Dakota	1994Q1	1994Q1	1998Q1	1993Q3	2025Q1	2025Q1	2025Q1	2025Q3
Tennessee	1998Q1	1998Q1	1998Q1	1997Q3	2025Q1	2025Q1	2025Q1	2025Q3
Texas	1990Q1	1995Q1	1995Q1	1989Q3	2025Q1	2025Q1	2025Q1	2025Q3
Utah	1990Q1	1999Q1	1999Q3	1989Q3	2025Q1	2025Q1	2025Q1	2025Q3
Vermont	2000Q1	2000Q1	2000Q1	1999Q3	2025Q1	2025Q1	2025Q1	2025Q3
Virginia	1995Q3	1998Q1	1998Q3	1995Q1	2025Q1	2025Q1	2025Q1	2025Q3
Washington	1990Q1	1990Q1	1990Q1	1989Q3	2025Q1	2025Q1	2025Q1	2025Q3
West Virginia	1990Q1	1997Q1	1997Q1	1989Q3	2025Q1	2025Q1	2025Q1	2025Q3
Wisconsin	1990Q1	1990Q1	1990Q1	1989Q3	2025Q1	2025Q1	2025Q1	2025Q3
Wyoming	1992Q1	1992Q1	2001Q1	1991Q3	2025Q1	2025Q1	2025Q1	2025Q3

1.8.2. Residence Data

Annual residence data is provided for workers who appear in the Job data. Residences are sourced from the Composite Person Record (CPR) tables or the Residential Candidates File (RCF). The year ranges for which each data source is available follow.

Available Years of LEHD Residence Data

Source	Start Year	End Year
CPR	1999	2011
RCF	2012	2023

1.9. Previous Versions of Snapshot Documentation

Links to previous versions of documentation can be accessed in the table below. Please note that documentation for Snapshot versions prior to S2021 are in working paper format.

Previous Versions of Snapshot Documentation

Version	Link
S2024_R2025Q4	2024 Snapshot
S2023_R2024Q4	2023 Snapshot
S2022_R2023Q4	2022 Snapshot
S2021_R2022Q4	2021 Snapshot
S2014	Vilhuber [2018]
S2011	Vilhuber and McKinney [2014]
S2008	McKinney and Vilhuber [2011]
S2004	McKinney and Vilhuber [2011]

1.10. Citing the Data, Sponsors, and Documentation

1.10.1. Acknowledging the Sponsors

The LEHD Snapshot draws on a data infrastructure that received substantial funding from a number of funding agencies and foundations. We strongly encourage researchers to acknowledge that funding in their paper’s “Acknowledgements” or data appendix. The following statement may be used:

This research uses data from the Census Bureau’s Longitudinal Employer-Household Dynamics program, which was partially supported by the following National Science Foundation grants SAS-9978093, SES-0339191, and ITR-0427889; National Institute on Aging grant AG018854; and grants from the Alfred P. Sloan Foundation.

1.10.2. Citing the Data

The data may be cited as follows:

Suggested citation

• U.S. Census Bureau. LEHD Snapshot Release S2024. [Computer file], U.S. Census Bureau, Center for Economic Studies, Research Data Centers [distributor], Washington, DC, 2025.

1.10.3. Data Access Statement

Many journals have adopted stringent data availability requirements. Researchers will need to work with the Census Bureau in order to ensure availability of their programs and research extracts. A statement similar to the following has been successfully used for accepted papers:

The data used for this paper were prepared in the U.S. Census Bureau’s secure computing facilities under an authorized project using the Federal Statistical Research Data Center network. The exact analysis files have been fully archived so that the programming sequence submitted in compliance with the [JOURNAL]’s editorial policy can be run in its entirety, except for the component that extracts the analysis sample from the underlying confidential databases. I grant any researchers with appropriate Census-approved project permission to use my exact research files provided that those files were among the ones that they requested when the approval was obtained (a Census Bureau requirement). In compliance with the [JOURNAL]’s editorial policy, I am submitting the list of those files, and the last known location of the archive on the Census Bureau’s RDC network as of [date]. I authorize the editorial staff of the [JOURNAL] to release this list and my statement of cooperation to any researcher who requests it, as well as to the U.S. Census Bureau or any agency cooperating with the Census Bureau in supervising research that uses the restricted-access data that I have used.

1.10.4. Citing this Documentation

The latest version of this documentation is published in HTML format, and is publicly available at https://lehd.ces.census.gov/data/lehd-snapshot-doc/latest/. However, we suggest that you cite this documentation from its latest working paper publication as follows:

Suggested citation

• Matthew Graham, Erika McEntarfer, Kevin McKinney, Stephen Tibbets, and Lee Tucker. LEHD Snapshot Documentation. Working Papers 22-51, Center for Economic Studies, U.S. Census Bureau, November 2022. URL: https://ideas.repec.org/p/cen/wpaper/22-51.html.

[1]

This research describes data from the Census Bureau’s Longitudinal Employer Household Dynamics Program, the original creation of which was partially supported by the following National Science Foundation (NSF) Grants SES-9978093, SES-0339191 and ITR0427889; National Institute on Aging Grant AG018854; and grants from the Alfred P. Sloan Foundation. Additionally, the current authors acknowledge the extensive contribution over the years by many, many individuals to the cumulative knowledge reflected in this document, too many to adequately enumerate here. This document has been reviewed to ensure that no confidential information is disclosed.