An official website of the United States government
United States Department of Labor
United States Department of Labor
The .gov means it's official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information,
make sure you're on a federal government site.
The site is secure.
The
https:// ensures that you are connecting to the official website and that any
information you provide is encrypted and transmitted securely.
Consumer Expenditure Surveys
SHARE ON:
Consumer Expenditure Surveys: Database Getting Started Guide
This page provides information on the structure, uses, and considerations when using the Consumer Expenditure Surveys (CE) database.
Table of Contents
Section 3. The CE Database search tools
Section 4. The CE Database files
Section 1. The CE Surveys
The Consumer Expenditure Surveys (CE) are nationwide household surveys conducted by the Census Bureau on behalf of the U.S. Bureau of Labor Statistics (BLS) to study how U.S. consumers spend their money. The surveys are the only federal government data collection effort to obtain information on the complete range of consumers’ expenditures, income, and demographic characteristics, directly from consumers. The Interview Survey collects information on major or recurring items with the Diary Survey collecting data for more minor or frequently purchased items. CE data are primarily used to revise the relative importance of goods and services in the market basket of the Consumer Price Index (CPI). While the CE data are most notably associated with the CPI, the CE data have a long list of stakeholders in both federal and private-sector organizations. For more information on the program, see the overview section in the BLS Handbook of Methods.
BLS publishes 12-month estimates of consumer expenditures annually, with the estimates summarized by various income levels and demographic characteristics. BLS provides these data in tables, public use microdata files (PUMD), publications, and the CE database.
Section 2. The CE Database
The database enables users to obtain information on consumer spending quickly and efficiently. The database includes estimates on annual means from 1984 forward. The database also includes estimates on percent reporting and relative standard error from 2010, and a more detailed breakout of aggregate and aggregate share expenditures from 2011 for 18 demographic characteristics and for all consumer units.
2025 CE database changes
Military status tables. Military status within consumer unit estimates, which were previously available in tabular version as a research product, are available in the CE database starting with the 2025 data release. They include three main categories: all consumer units; military consumer units; and non-military consumer units. Military consumer units comprise three groups: all veteran and/or active duty; veteran only; and active duty only. Military consumer units with both types of members (at least one veteran and at least one active duty member) are included in the first group; however, they are not listed as a separate group due to small sample size.
See previous years' changes in the Improvements and Protocol Changes webpage.
Data available in the CE LABSTAT database
Data are grouped into one of four categories: ADDENDA Assets and liabilities, and other financial info; CUCHARS Consumer Characteristics; EXPEND Expenditures; and INCOME Income and Taxes. Most queries and data searches will focus on Expenditures.
Expenditures are organized into 14 expenditure subcategories. Each of these categories is broken out further into items. See Table 1 for Categories and Subcategories. For more information, see the CE Glossary for Categories, Subcategories, and Items.
Table 1: Categories and subcategories
|
Category |
Subcategory |
|
ADDENDA |
Net change in total assets and liabilities |
|
Other financial information |
|
|
CUCHARS |
Number of consumer units (in thousands) |
|
Consumer unit characteristics |
|
|
Percent distribution |
|
|
EXPEND |
Alcoholic beverages |
|
Apparel and services |
|
|
Cash contributions |
|
|
Education |
|
|
Entertainment |
|
|
Food |
|
|
Healthcare |
|
|
Housing |
|
|
Personal insurance and pensions |
|
|
Miscellaneous expenditures |
|
|
Personal care products and services |
|
|
Reading |
|
|
Tobacco products and smoking supplies |
|
|
Total average annual expenditures |
|
|
Transportation |
|
|
INCOME |
Income after taxes |
|
Income before taxes |
|
|
Personal taxes |
Statistics available in the CE database
- Estimates are the mean dollar amount spent per item. Means are calculated with all consumer units (CUs) with a certain characteristic within a demographic group or all CUs across the country rather than only those that purchased the item. These values are reported in dollars and are rounded to the nearest dollar. Annual means estimates are available from 1984 onward.
-
Share of Total Expenditures is a portion (percentage) of an estimate of the average annual total expenditures for the average CU with a certain characteristic in a given demographic group. For example, in 2024, the housing expenditure share of total expenditures for CUs with a reference person under age 25, was 35.6 percent. This indicates that, of the entire amount of expenditures by an average CU with a reference person under age 25, housing expenditures made up 35.6 percent of their entire budget, (the entire amount of spending for this group of CUs).
-
Aggregate Expenditures (in millions of dollars) are the entire amount of dollars spent on a particular expenditure. These values are expressed in millions of dollars. These estimates are available from 2011 onward.
-
Share of Aggregate Expenditures are the portions of aggregate expenditures (as percentages) allotted to distinct demographic groups. This figure allows users to compare the spending of different demographic groups. For example, the share of aggregate expenditures for eggs (CXU080110LB0102M) for the lowest 20 percent income quintile in 2024 is 14.5 percent. This can be interpreted as the lowest income quintile contributing 14.5% of the entire spending on eggs in 2024.
-
Consumer unit characteristics provide context for interpreting the associated dollar estimates. These characteristics are presented as percent distributions and include percentages of reference persons by sex, housing tenure, race, Hispanic or Latino origin, education, and ownership of vehicles and are stored with no decimal places. Consumer unit characteristics are reported averages and are stored with one decimal place.
-
Standard errors (SE) are a measure of the uncertainty in a survey's estimates caused by the use of data from a representative sample of households rather than a complete universe of households. Standard errors are the most common measure of sampling variability of a survey's estimates. Defined as the square root of the survey estimate's variance, they measure how much the estimates would vary if the survey could be repeated using a different sample of consumer units every time. SEs provide a general measure of an estimate's precision and are used to determine whether differences between various expenditure estimates are statistically significant. For more information, see the Standard Errors in the Consumer Expenditure Surveys.
-
Relative Standard Errors (RSE) are the ratio of the standard error to the mean of the expenditure. BLS considers mean estimates with RSEs of 25 percent or more to be unreliable. See the FAQ page for more information.
-
Number of consumer units is the estimated number of consumer units in the survey population that a given demographic represents. The estimates are weighted to reflect the population and are reported in thousands, rounded to the nearest person.
-
Percent reporting is the percentage of consumer units who recorded making a specific expenditure within a given data collection period. For example, the percent reporting for Vehicle purchases are recorded in the Interview survey which collects data on a quarterly basis. Not all CUs will have purchased a vehicle during that time frame. Percent reporting data will only capture those CUs who engaged in the transaction of purchasing a vehicle. This figure allows users to assess how prevalent an item is purchased and to estimate weekly or quarterly expenditure means for CUs that reported an expenditure. Expenditures for aggregate items that have data from both the Interview and Diary surveys do not have a percent reporting calculation because expenditure amounts from each survey reflect different periods (quarterly vs. weekly). Additionally, percent reporting values are not provided for Total average annual expenditures estimates.
Tabular data available in the CE database
Not all estimates are in the CE database. While some estimates are available on an annual basis since 1984, the CE database contains detailed annual means estimates from 2010 (see table 2). Data users can request pre-2010 all consumer unit detailed means tables by contacting CE. Also, the CE database does not offer data by metropolitan statistical area or cross-tabulations as these are 2 year estimates and the CE database is restricted to annual. Data users can find these estimates on the CE tables page.
Table 2: Types of tables available in the CE database
| Tabular Data Availability | Tables | CE Database |
|---|---|---|
| ✔ | ✔ | |
| ✔ | ✔ | |
| ✔ | ✔ | |
|
Calendar year aggregate shares by demographic characteristics |
✔ | ✔ |
| ✔ | ||
| ✔ |
Section 3. The CE Database search tools
This section provides another view of the CE database search tools. These search tools allow users to dynamically retrieve CE estimates and their series IDs. Series IDs are unique identifiers used to categorize and track data series and their data elements. Selected estimates are populated on a webpage, which can then be downloaded to Excel to develop further analysis and visualizations beyond the tool platform.
- Top Picks allows you to quickly retrieve BLS time series data from a list of those commonly requested.
- Data Finder allows you to key word search for any combination of search terms. Results are sorted by measures (including attributes), characteristics (including demographics and occupation), and survey allowing you to create charts and tables and to export the data into Excel or CSV.
- One-Screen Data Search allows you to produce customized time series by CUs and their characteristics. Data can be exported into Excel through a series of steps using a single-screen form.
- Multi-Screen Data Search allows you to produce customized time series by CUs and their characteristics. Data can be exported into Excel through a series of steps using a multi-screen form. Tables with multiple demographic characteristics can be created. Inputs are Category, Subcategory, Item, Demographic, and Characteristic.
- Series Report provides experienced users of CE data with one of the quickest ways to access the CE estimates on our website. Users that already know the Series ID for a given estimate can use this application by providing the Series ID as input. For more information on Series IDs, please see section 4 below.
Section 4. The CE Database files
This section provides the necessary documentation for working with the underlying text files. The text files house the data and parameters used to populate the estimates in the tools described above. Data files are in American Standard Code for Information Interchange (ASCII) text format. The data files cx.data.1.All.Data and cx.aspect contain all the estimates in the database. These files, used in conjunction with the series and mapping files, enable users to create custom table output that may better suit their research and analysis needs, beyond the capability of the tools. Working with these files does require users to have a firm understanding of the underlying data available. Prior to working with these files, users should first peruse the available data using either one of the search tools listed above or the mapping files detailed below, in an effort to identify the data elements needed for their individual research purposes.
File types and descriptions
This section describes the various files and file types available to generate time series of most CE estimates found on the CE tables page. Table 3 below identifies the files that are available.
Table 3: File types
| File Name | File Type | Description |
|---|---|---|
| Data | Contains means for expenditures, income and taxes, consumer characteristics, and assets and liabilities for 17 demographic groups. | |
| Data | Contains standard errors, expenditure aggregates, and expenditure shares for associated means found on the cx.data.1.AllData. | |
| Series | Contains a set of codes which, together, comprise a series identification code that serves to uniquely identify a single time series, such as "Women, 16 and over by Income Deciles: Ninth 10 percent." | |
| Mapping | Contains text descriptions that link to the category_code variable within the cx.series file. | |
| Mapping | Contains text descriptions that link to the subcategory_code variable within the cx.series. | |
| Mapping | Contains text descriptions that link to the item_code variable within the cx.series. | |
| Mapping | Contains text descriptions that link to the demographics_code variable within the cx.series. | |
| Mapping | Contains text descriptions that link to the characteristics_code variable within the cx.series. | |
| Mapping | Contains text descriptions that link to the footnote_code variable within the cx.series. | |
| Mapping | Contains text descriptions that link to the process_code variable within the cx.series. | |
| Contact | Contains CE program contact information. |
Data and mapping files and descriptions
Data files are tab delimited text files that contain all the estimates that are available in the database. The first record of each file contains the column headers for the data elements stored in each field. The cx.data.1.AllData data file contains average annual means as well as percentages that describe the consumer units within each demographic breakout. The cx.aspect data file contains associated estimates including the mean standard error, the expenditure aggregate or expenditure aggregate share, and the share of total expenditures. Associated estimates are identified by the aspect_type variable stored on file.
Below are the variables, their length, and possible values for the data file cx.data.1.AllData.
Table 4: cx.data.1.AllData file
| Variable | Length | Example value | Definition |
|---|---|---|---|
|
series_id |
30 | CXUMENBOYSLB0104M | Identifies specific series. |
|
year |
4 | 1984 | Identifies year of observation. |
|
period |
3 | A01 | Identifies period for which data are observed. |
|
value |
12 | 93 | Lists the observation for the series. |
|
footnote_code |
10 | It varies | Identifies footnote for the data series. |
Below are the variables, their lengths, and their possible values for the data file cx.aspect.
Table 5: cx.aspect file
| Variable | Length | Example value | Definition |
|---|---|---|---|
|
series_id |
30 | CXUMENBOYSLB0104M | Identifies the specific series. For a full list of possible series IDs, please see cx.series. |
|
year |
4 | 1984 | Identifies year of the estimate. |
|
period |
3 | A01 | Identifies the time period for which the estimate represents. The CE only develops annual estimates and therefore only has one begin and end period, A01. |
|
aspect_type |
2 | ES | Identifies the estimate type (see below of available estimates). |
|
value |
12 | 93 | Identifies the estimate. |
|
footnote_code |
15 |
1 | Identifies the footnote code for the data series. For a full list of possible footnote codes, please see cx.footnote. |
The cx.aspect data file can be merged with cx.data.1.Alldata, which allows the user to associate supporting estimates found on cx.aspect with the means found on cx.data.1.Alldata. Associated estimates include the mean standard error, the expenditure aggregate or expenditure aggregate share, and the share of total expenditures. Aspect_types are described in Section 5 below.
The series file (cx.series) enables users to associate coded values with text definitions by merging them with the mapping files discussed in section 6 below. The series file can also be merged with the data files to associate estimates with estimate titles. The series file is in ASCII text format, which is a common format for text files. Data elements are separated by tabs and the first record of each file contains the variable names for the data elements stored in each field.
Series_IDs can be broken down into various segments described below. Below are the types of fields, their length, and example values.
Table 6: cx.series file
| Variable | Length | Example value | Definition |
|---|---|---|---|
|
series_id |
30 | CXUFRSHFRUTLB0201M | Series_id. For a full list of possible series IDs, please see cx.series. |
|
seasonal |
1 | U | Identifies whether the data are seasonally adjusted. All values are identified with a “U,” as all estimates are unadjusted/not seasonally adjusted. |
|
category_code |
10 | EXPEND | Identifies the category code for the data series. For a full list of possible category codes, please see cx.category. |
|
subcategory_code |
9 | FOOD | Identifies the subcategory code for the data series. For a full list of possible subcategory codes, please see cx.subcategory. |
|
item_code |
10 | FRSHFRUT | Identifies the item code for the data series. For a full list of possible item codes, please see cx.item. |
|
demographics_code |
4 | LB02 | Identifies the demographics code for the data series. For a full list of possible demographics codes, please see cx.demographics. |
|
characteristics_code |
2 | 01 | Identifies the characteristics code for the data series. For a full list of possible characteristics codes, please see cx.characteristics. |
|
process_code |
1 | M | Identifies the process code for the estimates found on cx.data.1.Alldata. (CE only provides Means on cx.data.1.Alldata so this will value will only be “M.”) |
|
series_title |
256 | Fresh fruits by Income Range: All Consumer Units | Title defining the unique series ID. |
|
footnote _code |
10 | It varies | Identifies the footnote code for the data series. For a full list of possible footnote codes, please see cx.footnote. |
|
begin_year |
4 | It varies | Identifies the first year the series is available. |
|
begin_period |
3 | AO1 | Identifies the starting period for which the estimate represents. The CE only develops annual estimates and therefore only has one begin and end period, A01. |
|
end_year |
4 | It varies | Identifies the last year the series is available. |
|
end_period |
3 | AO1 | Identifies the end period for which the estimate represents. The CE only develops annual estimates and therefore only has one begin and end period, A01. |
For example, the series_id CXUFRSHFRUTLB0201M can be broken out into these segments:
Table 7: Series IDs elements for series_ID CXUFRSHFRUTLB0201M
| Series ID elements | Example value |
|---|---|
|
survey specific abbreviation |
CX |
|
item_code |
FRSHFRUT |
|
demographics_code |
LB02 |
|
characteristics _code |
01 |
|
process_code |
M |
The cx.category file defines the available major category codes available in the CE database. Mapping files are in ASCII text format. Data elements are separated by tabs. The first record of each file contains the column headers for the data elements stored in each field. For CE, categories are limited to the following:
- Assets and liabilities, and other financial info
- Consumer Characteristics
- Expenditures
- Income and Taxes
Table 8: cx.category
| Variable | Length | Example value | Definition |
|---|---|---|---|
|
category_code |
10 | EXPEND | Identifies the category code for the data series. For a full list of possible category codes, please see cx.category. |
|
category_text |
60 | Expenditures | Title defining the unique category code. |
|
display_level |
2 | 0 | Identifies how columns should be grouped for display. |
|
selectable |
1 | T | Not applicable to CE. All series will be marked with “T.” |
|
sort_sequence |
5 | 100 | Sort sequence for how columns should be ordered. |
cx.subcategory: defines the available subcategory codes available in the CE database. cx.subcategory can be used to explore greater detail beyond what is available at the category level. For CE, subcategories include various expenditure types like food and transportation, as well as breakouts for income, like income before taxes.
Table 9: cx.subcategory
| Variable | Length | Example value | Definition |
|---|---|---|---|
|
category_code |
10 | EXPEND | Identifies the category code for the data series. For a full list of possible category codes, please see cx.category. |
|
subcategory_code |
9 | PERSCARE | Identifies the subcategory code for the data series. For a full list of possible subcategory codes, please see cx.subcategory. |
|
subcategory_text |
50 | Personal care products and services | Title defining the unique subcategory code. |
|
display_level |
2 | 0 | Identifies how columns should be grouped for display. |
|
selectable |
1 | T | Not applicable to CE. All series will be marked with “T.” |
|
sort_sequence |
5 | 900 | Sort sequence for how columns should be ordered for display. |
cx.item: defines the available detailed items available in the CE database. For CE, items include various expenditure types like pets and major appliances, as well as breakouts for income, like self-employment.
Table 10: cx.item
| Variable | Length | Example value | Definition |
|---|---|---|---|
|
subcategory_code |
9 | ENTRTAIN | Identifies the subcategory code for the data series. For a full list of possible category codes, please see cx.subcategory. |
|
item_code |
10 | PETSPLAY | Identifies the item code for the data series. For a full list of possible item codes, please see cx.item. |
|
item_text |
50 | Pets, toys, and playground equipment | Title defining the unique item code. |
|
display_level |
2 | 1 | Identifies how columns should be grouped for display. |
|
selectable |
1 | T | Not applicable to CE. All series will be marked with “T.” |
|
sort_sequence |
5 | 11200 | Sort sequence for how columns should be ordered for display. |
cx.demographics: defines the available demographics available in the CE database. For CE, demographics include various breakouts like income quintiles, age, or race.
Table 11: cx.demographics
| Variable | Length | Example value | Definition |
|---|---|---|---|
|
demographics_code |
4 | LB01 | Identifies the demographics code for the data series. For a full list of possible demographics codes, please see cx.demographics. |
|
demographics_text |
150 | Quintiles before taxes | Title defining the unique item code. |
|
display_level |
2 | 0 | Identifies how columns are grouped for display. |
|
selectable |
1 | T | Not applicable to CE. All series will be marked with “T.” |
|
sort_sequence |
5 | 100 | Sort sequence for how columns should be ordered for display. |
cx.characteristics: defines the available detailed demographic characteristics available in the CE database. For CE, characteristics include various breakouts like “Lowest 20 percent income quintile,” or “Reference person under age 25.”
Table 12. cx.characteristics
| Variable | Length | Example value | Definition |
|---|---|---|---|
|
demographics_code |
4 | LB01 | Identifies the demographics code for the data series. For a full list of possible demographics codes, please see cx.demographics. |
|
characteristics_code |
2 | 01 | Identifies the characteristics code for the data series. For a full list of possible characteristics codes, please see cx.characteristics. |
|
characteristics_text |
120 | All consumer units | Title defining the unique characteristics code. |
|
display_level |
2 | 1 | Identifies how columns should be grouped for display. |
|
selectable |
1 | T | Not applicable to CE. All series will be marked with “T.” |
|
sort_sequence |
5 | 1010 | Sort sequence for how columns should be ordered for display. |
cx.footnote: defines the available footnote codes utilized in the CE database. For CE, footnotes identify suppression and other key details users should be aware of when interpreting the estimate.
Table 13: cx.footnote
| Variable | Length | Example value | Definition |
|---|---|---|---|
|
footnote_code |
2 | 1 | Identifies the footnote code for the data series. For a full list of possible footnote codes, please see cx.footnote. |
|
footnote_text |
250 | No data provided | Title defining the unique footnote code. |
cx.process: is a place holder used on cx.data.1.AllData. All estimates stored on cx.data.1.AllData are marked with a process code of “M."
Table 14: cx.process
| Variable | Length | Example value | Definition |
|---|---|---|---|
|
process_code |
1 | M | Identifies the process code for the estimates found on cx.data.1.AllData. (CE only provides Means on cx.data.1.AllData so this will value will only be “M.”) |
|
process_text |
30 | Means | Title defining the unique process code. |
|
display_level |
2 | 0 | Identifies how columns should be grouped for display. |
|
selectable |
1 | T | Not applicable to CE. All series will be marked with “T.” |
|
sort_sequence |
5 | 100 | Sort sequence for how columns should be ordered for display. |
Aspect file codes
Located in the cx.aspect file, aspect_type options are defined in Table 4 below. Note that expenditure means are located on the cx.data.1.Alldata file and do not have an aspect code.
Table 15: aspect types
| Aspect type | Definition |
|---|---|
|
AG |
Aggregate Expenditure |
|
E |
Standard Error |
|
ES |
Expenditure Shares |
|
RP |
Percent Reporting |
|
R0 |
Relative Standard Error |
|
AS |
Share of Aggregate Expenditure |
Below are more detailed explanations of each aspect type:
AG (Aggregate Expenditure in millions of dollars):
Series defined with this aspect type represent the aggregate expenditure, or sum of all purchases of an expenditure or expenditure subcategory for all consumer units (characteristic_code = 01). For example, the Series Report for expenditures on eggs for all consumers (CXU080110LB0101M ) shows that the aggregate expenditure (AG) in 2024 was $14,257. This is the complete dollar amount presented in millions representing $14,257,000.
Table 16: series ID CXU080110LB0101M
| Year | Estimate | Standard Error | Relative Standard Error | Percent Reporting | Share of Total Expenditures | Aggregate Expenditure (in millions of dollars) | Share of Aggregate Expenditure |
|---|---|---|---|---|---|---|---|
|
2024 |
105 | 3.58 | 3.4 | 30.38 | 0.1 | 14257 | -(5) |
|
Footnotes: |
|||||||
AS (Share of Aggregate Expenditures):
Series defined with the aspect_type AS represent the share (or portion) of aggregate expenditures that a particular demographic group (characteristic_code > 01) makes up of the whole aggregate. For example, the share of aggregate expenditures for eggs (CXU080110LB0102M) for the lowest 20 percent income quintile in 2024 contains a percentage of 14.5. This can be interpreted as the lowest income quintile contributing 14.5% of the entire spending on eggs in 2024.
Table 17: series ID CXU080110LB0102M
| Year | Estimate | Standard Error | Relative Standard Error | Percent Reporting | Share of Total Expenditures | Aggregate Expenditure (in millions of dollars) | Share of Aggregate Expenditure |
|---|---|---|---|---|---|---|---|
|
2024 |
77 | 6.94 | 9.06 | 22.79 | 0.2 | 2068 | 14.5 |
ES (Share of Total Expenditures):
Series defined with aspect_type of “ES” will provide a share of total expenditure, which represents an expenditure item expressed as a percent of total expenditures for a given demographic group. For example, the share of total expenditures for series_ID CXUHOUSINGLB0402M, housing expenditures for CUs with a reference person 25 and younger, in 2024 was 35.6%. This estimate indicates that of their entire budget (the entire amount of discretionary spending for this group of CUs), housing made up 35.6% of that total.
Table 18: series ID CXUHOUSINGLB0402M
| Year | Estimate | Standard Error | Relative Standard Error | Percent Reporting | Share of Total Expenditures | Aggregate Expenditure (in millions of dollars) | Share of Aggregate Expenditure |
|---|---|---|---|---|---|---|---|
|
2024 |
16853 | 1044.99 | 6.2 | -(5) | 35.6 | 112261 | 3.1 |
|
Footnotes: |
|||||||
E (Standard Error):
Series defined with aspect_type of “E” will provide the standard error of the mean. For example, in 2024, the standard error for housing expenditures for CUs with a reference person 25 years old or younger (CXUHOUSINGLB0402M) was 1044.99 This indicates that if the CE surveys could have been repeated with different samples of consumer units, the mean expenditure could have varied by up to plus-or-minus $1044.99. For more information on standard errors please visit the following link: CE FAQs.
R0 (Relative Standard Error):
Series defined with aspect_type “R0” will provide the relative standard error (RSE) of the mean. For more information on relative standard errors, see the following link: CE FAQs
RP (Percent Reporting):
Series defined with aspect_code “RP” will provide percent reporting (PR), the percent of consumer units making the expenditure during the data collection period. For example, the percent reporting for Vehicle insurance is per quarter, and all CUs do not pay car insurance premiums every three months. The percent reporting will be weekly or quarterly, depending on the survey used as the source of that expenditure item. Aggregated rows that have data from both surveys do not have a percent reporting calculation because expenditure amounts from each survey represent different time periods (quarterly vs. weekly), which disallows the development of a percent reporting figure at the aggregate level.
Last modified date: November 2, 2026