Berkeley Federal Statistical Research Data Center (FSRDC)
The Berkeley FSRDC is a joint project between the US Census Bureau and the University of California at Berkeley, hosted at the D-Lab. As one of the 30 sites in the Census FSRDC Program(link is external)(link is external),our facility provides a secure environment where researchers with appropriate security clearance on approved projects can access restricted-use datasets.
The Berkeley FSRDC provides researchers access to non-public federal datasets that can support research in economics, sociology, public health, and related fields. The Berkeley FSRDC is an access point for data from the US Census Bureau and from other partnering governmental agencies.
Broadly speaking, there are three types of Census data available through the Berkeley FSRDC:
Demographic Data:(link is external)(link is external) Censuses and surveys of households and individuals. Many of these datasets have a public-use version, but those available through the Berkeley FSRDC have more detailed geography (in some cases down to the block) and items such as income are often not top coded.
Economic Data:(link is external)(link is external) Economic data are data where the unit of observation is the establishment and/or firm. If a public-use version for these files exist, they are limited to data presented in aggregate form.
Linked Business and Household Data:(link is external)(link is external) The Longitudinal Employer–Household Dynamics (LEHD) program uses unemployment insurance records to match employers with employees to provide high-quality local labor market information. Firm and worker information are combined to create job level quarterly earnings history data, data on where workers live and work, and data on firm characteristics, such as industry.
Research projects wanting to use any of the above data must follow Census guidelines for data access. See the information on these guidelines below.
Partner Agency Data
The Berkeley FSRDC is also an access point for approved projects that use restricted-use data from the following partnering agencies:
Agency for Healthcare Research and Quality (AHRQ(link is external)(link is external))
Bureau of Labor Statistics (BLS(link is external)(link is external))
Bureau of Economic Analysis (BEA(link is external)(link is external))
National Center for Health Statistics (NCHS(link is external)(link is external))
Projects wanting to use partnering agency data will need to follow the proposal guidelines of the respective agency. Follow the agency links for more information.
Guidelines and Protocols
Census Proposal Guidelines
Researchers whose research agenda relies on restricted-use Census data work with the Berkeley FSRDC administrator, Angela Andrus, to develop a proposal before submission to the Center for Enterprise Dissemination (CED) at the US Census Bureau. The researcher should discuss the proposed project with the administrator to determine if the proposal demonstrates:
- Potential to benefit Census Bureau data programs
- Scientific merit
- Evidence of feasibility
- Clear need for non-public data
- Minimal risk of disclosure
Should the Berkeley FSRDC administrator and the researcher agree to proceed, the administrator will work with the researcher to develop a full proposal following these guidelines.
If you are interested in developing a Census data proposal, please contact:
Berkeley FSRDC Administrator
Special Sworn Status (SSS)
Researchers are required to obtain Special Sworn Status (SSS) before they can access the Berkeley FSRDC. To be eligible to apply for SSS, researchers must (a) be a US citizen or foreign national with 3 of the past 5 years of residency in the US, and (b) currently live in the US and have a US institutional affiliation. To apply for SSS, researchers must be fingerprinted, make a sworn statement in the presence of a notary public, and provide additional information for a background investigation. Costs for the preparation of these materials are the responsibility of the researcher. Application materials will be provided after the proposal has been approved.
Census Bureau Manual
Access and Project Fees
UC Berkeley Affiliates
Access and Census project fees are covered by the D-Lab for current faculty, postdocs, students, and staff of the University of California, Berkeley.
Non-UC Berkeley organizations involved in multiple projects or with multiple researchers can subscribe to a consortium membership that offers substantial discounts over individual projects and researcher fees.
Project hosting fees range from $15,000 - $20,000 per year, depending on the nature of the project, while annual researcher access fees on existing projects are lower.
For more information about our fees or for details of the consortium membership program, please contact Renee Starowicz (email@example.com(link sends e-mail), Co-Director of the Berkeley FSRDC.
Additional fees may be charged to projects that use partner agency data or that impose other special costs on CED, the Census Bureau, or the Berkeley FSRDC. These costs are the responsibility of the researcher and are not covered by the Berkeley FSRDC access and hosting fees.
Current Research Projects
Current Berkeley FSRDC Projects Using Data from the U.S. Census Bureau
"Quantifying the Eligibility, Enrollment, and Retention of Low- and Moderate-Income Populations in Means-Tested Programs"
This research will quantify short-run earnings volatility among low- and moderate-income families in California to estimate the incidence of family income changes and gauge implications for means-tested public insurance program eligibility. Moreover, this project will examine uptake of insurance programs and assess causes of disenrollment. To identify low- and moderate-income families in California, quantify earnings volatility, and assess public insurance eligibility, we will use quarterly earnings records contained in the Employment History File from the Longitudinal Employer-Household Dynamics (LEHD) program linked to household records in the American Community Survey and Current Population Survey, and person-level characteristics in the LEHD Individual Characteristics File. To this, we will link Medicaid Statistics Information Systems records, in order to capture spells of enrollment to Medicaid and SCHIP. With this, we will generate measures of inappropriate disenrollment (program drop-out despite continuing eligibility) and legitimate disenrollment where return to these programs occurs in a short period of time (churning). In 2014, millions of Americans became eligible for government-subsidized health insurance programs where eligibility is determined by income falling within precisely defined ranges. We find it important to know whether eligible individuals have taken up benefits and, for families with more variable incomes, whether the costs of maintaining enrollment result in an increased likelihood of churning—moving off and then back on a program within a short period of time—which is costly both for households and programs.
"Energy and the Environment - Plant-Level Production and Market-Level Incidence"
The extent to which environmental, energy, and other types of regulations influence firm-level production decisions has important implications for aggregate welfare and distributional incidence. Empirical research in this area is limited though growing. Much economic research on environmental regulation and environmental goods still compares how "clean" versus "dirty" industries respond to different regulatory or economic forces. However, the availability of firm and plant-level data makes it possible to recognize that firms within an industry differ very widely. Even within a narrowly defined industry firms differ enormously in the quantity and mix of pollutants that they emit, in the stringency of regulations they face, in productivity, trade exposure, market power, product quality, input mix, and product mix. Some of these differences may reflect measurement error or idiosyncratic productivity shocks, but others reflect fundamental economic forces. The lack of high-quality, firm or plant level microdata has substantially limited research in this area.
"Technology Lock-In: Evidence from U.S. Manufacturing"
This project aims to quantify the extent to which initial market conditions determine manufacturing establishments' subsequent input mix. We focus on the role of initial energy prices for short and long-run energy usage as an example of this technology lock-in. We propose to conduct this analysis using data from the Annual Survey of Manufactures, the Census of Manufactures, the Longitudinal Business Database, the Manufacturing Energy Consumption Survey, the Annual Capital Expenditure Survey, and the Survey of Pollution Abatement Costs & Expenditures from the Census Bureau, in combination with several external data sources on U.S. energy prices. First, we will use these data sets to provide new descriptive evidence on the importance of energy prices in the year of an establishment's entry for subsequent energy efficiency, summarizing the effects by entry cohort and by initial energy price. Second, we will estimate a dynamic structural model of entry, exit, and capital investment to simulate how the energy intensity of production would change for entrant and incumbent establishments under several counterfactual market conditions. Our goal is to use the structural model to study how raising energy prices and reducing capital adjustment costs affect the investment decisions and the input mix of U.S. manufacturing establishments.
- Angela Andrus, Administrator, firstname.lastname@example.org(link sends e-mail)
- Renee Starowicz, Co-Director, email@example.com(link sends e-mail)
- David Harding, Co-Director, firstname.lastname@example.org(link sends e-mail)
350 Social Sciences Building
University of California
Berkeley, CA 94720