Public-Use Microdata (PUMD)
The Consumer Expenditure Survey (CE) collects data on expenditures, income, and demographics in the United States. The public-use microdata (PUMD) files provide this information for individual respondents, without any information that could identify respondents. PUMD files include adjustments for information that is missing because respondents were unwilling or unable to provide it. The files also have been adjusted to reduce the likelihood of identifying respondents, either directly or through inference.
This format allows researchers to analyze expenditures, income, and demographic trends beyond what the published tables show. However, users of the PUMD files should have strong skills with statistical software. To learn more, you may want to explore the PUMD methodology with the CE PUMD Getting Started Guide and check the availability of data with the Dictionary for Interview and Diary Surveys (XLSX).
The 2017 PUMD were released on September 11, 2018.
Note: To users of the CE 2017 Public-use Microdata: some corrected files have been reissued on June 27, 2019. See Errata notice for details.
Experimental State Weights for 2017, supporting state-level analysis for selected states, are also available with their documentation.
To be notified when new datasets become available, please sign up for the CEX e-mail updates. If you have comments, suggestions, and questions about this page and its contents, please contact us.
Introduction to public-use microdata (PUMD)
The Consumer Expenditure Survey (CE) provides continuous and comprehensive information on the buying habits of U.S. consumers. The spending shares for items in the survey are used as weights for the Consumer Price Index. CE provides this data to the public in four formats, shown below:
What kind of data do you need?
- Major pre-defined expenditures, income, and demographic tables by age, race, region, housing, education, occupation, income categories, number of earners, Hispanic or Latino origin, and other aspects. In addition, we provide tables that cross tabulate two characteristics, for example age and income. To use these tables, users do not need statistical knowledge or software. These tables are available in three publications: News releases, published tables, and CE LABSTAT database.
- Detailed pre-defined expenditure, income, and demographic tables provide the same data as the tabulations described above with additional detail and the standard deviation measures. To use these tables, users need to take into account that some data points are based on a small number of responses, which may result in a standard error over 25 percent. These tables are available as experimental research tables or upon request from the methods contact.
Household-level survey responses with adjustments. These files contain the actual responses to the Interview and Diary Surveys. The responses can be used to create and correlate variables. CE adjusts this dataset in two ways. First, CE estimates components, like income, that respondents were unwilling to share and second, anonymizes the data to protect the respondents' identities.
To use this dataset, users need to be familiar with statistical concepts, such as standard error, and be proficient with a statistical software package, such as SAS, SPSS, Stata, or R. This dataset is available on the PUMD data page.
Household-level survey responses without adjustments. These data files are the same as the files described above except that they have not been altered to protect the respondents' identity, which provides users with slightly more information, for example on region and income.
To use this dataset, users need to be familiar with statistical concepts, such as standard error, and be proficient with a statistical software programming.
In addition, accessing this dataset is more involved than accessing the above alternatives. Users need to fulfill the following requirements: They need to be affiliated with a government agency, academic institution, or nonprofit organization, be a U.S. citizen or legal resident, have two years-time before they can access the data, be able to fund to travel to Washington D.C., show BLS that their research is in the public's interest and will be publically available, and receive clearance from both BLS and the U.S. Census Bureau to access the unadjusted data (Title 13). For more information on this avenue, see on-site visiting researcher program.
Last Modified Date: June 27, 2019