The Consumer Expenditure Surveys (CE) collect information from the Nation's households and families on their buying habits (expenditures), income, and household characteristics. The strength of the surveys is that it allows data users to relate the expenditures and income of consumers to the characteristics of those consumers. The surveys consist of two components, a quarterly Interview Survey and a weekly Diary Survey, each with its own questionnaire and sample.
Data from the Consumer Expenditure Surveys are used in a number of different ways by a variety of users. One important use of the surveys is for the periodic revision of the Bureau of Labor Statistics Consumer Price Index. The Bureau uses survey results to select new market baskets of goods and services for the Consumer Price Index, to determine the relative importance of Consumer Price Index components, and to derive new cost weights for the market baskets. Market researchers find the data useful in analyzing the demand for groups of goods and services. The data allow them to track spending trends of different types of consumer units (see the response to question 27 for the definition of a consumer unit). Government and private agencies use the data to study the welfare of particular segments of the population, such as those consumer units with a reference person aged 65 and older or under age 25, or for low-income consumer units (see the response to question 29 for the definition of a reference person). Economic policymakers use the data to study the impact of policy changes on the welfare of different socioeconomic groups. Researchers use the data in a variety of studies, including those that focus on the spending behavior of different family types, trends in expenditures on various expenditure components including new types of goods and services, gift-giving behavior, consumption studies, and historical spending trends.
Consumer Expenditure Surveys contact information is provided here.
The information that respondents provide is used solely for statistical purposes. All Census Bureau data collectors and BLS CE staff take an oath of confidentiality and are subject to fines or imprisonment for improperly disclosing information provided by respondents. Names and addresses are removed from all forms, and are not included in any statistical release. As a further precaution, the Bureau of Labor Statistics applies certain restrictions to the public-use microdata. These include geographical and value restrictions that prevent the identification of respondents.
The CE COVID-19 webpage provides more information regarding the effects of the COVID-19 pandemic on the Consumer Expenditure Surveys.
The latest Annual Report is posted here.
Yes. Prior to 1980, the Consumer Expenditure surveys were conducted about every 10 years. Since that time, it has been an ongoing survey. Online data tables are available for 1961, 1972–73, and later surveys. For information about the availability of any CE data, including historical data, contact the Division of Consumer Expenditure Survey.
Caution should be used in comparing data from the current surveys with those gathered before the 1972–73 surveys, or even during the first few years of the current survey, due to changes in concepts and definitions. For example, integrated data from the Diary and Interview Surveys have been published for 1972–73 and from 1984 onward; prior to 1972–73, data from each survey were published separately. The Consumer Expenditure Surveys have electronic versions of integrated tables for 1972–73 and annually from 1984 onward. Also prior to 1972–73, published data covered only the urban portion of the population. Beginning in 1972–73 and from 1984 onward, the published data are for the total population, urban and rural.
CE data are used to estimate experimental poverty thresholds based on recommendations by the National Academy of Sciences and observations of a 2010 Interagency Technical Working Group. The most recent measure, referred to as the Supplemental Poverty Measure, is not designed to replace the current, official poverty estimates published each year by the U.S. Census Bureau, but rather is intended to provide additional information on poverty in the U.S. Details on the experimental poverty measures can be obtained from the BLS Division of Price and Index Number Research.
Detailed expenditure data for some items, such as food items, are unique to the Diary Survey. Data for other items, such as third-party reimbursements for medical care expenses or the cost of auto repairs, are collected only in the Interview Survey. However, there is considerable overlap in coverage between the surveys. Because of this overlap, integrating the data presents the problem of determining the appropriate survey component from which to select the expenditure items. When data are available from both survey sources, the more reliable of the two, as determined by statistical methods, is selected. As a result, some estimates are selected from the Interview Survey and others, from the Diary Survey. A 2011 Consumer Expenditure Surveys Anthology article describing the process is available (PDF) as well as a list showing which survey was used as the source for items in each year's published estimates (XLSX).
Yes. Starting with the 1972–73 tables, the Bureau of Labor Statistics has published data integrated from the Interview and Diary components of the survey. Because the two components are designed to capture different types of expenditures, integrating data from them combines the important features of both. The integrated data provide a complete accounting of consumer expenditures and income, which neither survey component alone is designed to do.
Information on assets and liabilities is collected from respondents to the surveys; however, like the income data, the assets and liabilities data are not as reliable as the expenditure data. Respondents may be unable or unwilling to provide accurate information on their assets and liabilities. Net changes in assets and liabilities are published in the CE tables. The public-use microdata also include information on assets and liabilities. An alternative source of data on assets, liabilities, and other financial information of consumers is the Survey of Consumer Finances, conducted by the Federal Reserve Board.
Information on spending on cellular phone services is found here.
If you want to relate the expenditures of consumers to their income and characteristics, the Consumer Expenditure Surveys are the primary source of data. However, for users interested only in income information, data published by the Census Bureau of the U.S. Department of Commerce may be a better source of information. Data from the Current Population Survey are based on a much larger sample size. The Census Bureau also can provide income information from its Survey of Income and Program Participation, which focuses on low-income households. For information on this survey, visit the Survey of Income and Program Participation page.
No. The CE data in published tables show average expenditures and incomes of consumer units. The expenditure levels may vary across areas for a number of reasons. These include demographic and economic differences in age levels, income levels, size of consumer units, tastes, and personal preferences. A commonly used method of comparing the cost of living among areas involves developing an estimate of the cost of a similar bundle of goods and services for each area. The CE program makes no attempt to measure the cost of a standard bundle of goods and services, but instead provides actual expenditure levels of consumer units.
The CE data tables in Text format can be converted into XLS format (which can be used in Microsoft Excel or other spreadsheet applications) by taking the following steps:
The Public-Use Microdata (PUMD) files include the actual responses to the Interview and Diary surveys along with derived values, allowing researchers to analyze expenditures, income, and demographic trends beyond what the published tables show. If you are new to the CE PUMD, 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 CE PUMD files are located on the PUMD data files page. PUMD are available in four formats: SAS, SPSS, STATA, and Comma Delimited (ASCII). Each format presents the data for each year in two zipped files. Data collected with the Interview Survey are in the "Interview" file and those with the Diary Survey are in the "Diary" file.
You can link data between different PUMD files with the variable NEWID. NEWID is a concatenation of the Consumer Unit (CU) identifier and the interview/diary wave in which the data are collected. This variable is the only PUMD variable that is consistent across every PUMD file. Note some files require additional variables to link between files. See the CE PUMD Getting Started Guide for more detail on the PUMD file structure.
You can link data for a given CU across PUMD files by removing the last digit from the variable NEWID. Doing so identifies the CU. Starting in 2002, the variable CUID was added to the FMLI and FMLD files to link a CU across different interviews or diaries for a given CU.
Note: CUID follows the address not the CU, which may hurt the efficacy of quarterly longitudinal analyses that follow CUs, if the residents change.
You can identify the Interview or Diary Survey wave for a given data point with the last digit of the variable NEWID. For the Interview Survey, the digits range from 1 to 4, depending on the interview quarter. For example, a "1" corresponds to the first quarter after the CE program dropped the bounding interview in 2014. For the Diary survey, the last digit is 1 or 2 for the first or second diary week.
Prior to April 2015, the Interview Survey included a preliminary bounding interview, and each CU could be contacted up to five times over five quarters. This preliminary bounding interview collected household roster data and inventory information, which was meant to minimize errors in reporting information out of scope in future interviews. As a result, expenditures collected in this first interview were not released as part of the PUMD files. The bounding interview was dropped beginning in the second quarter of calendar year 2015.
No, you cannot link CUs across both surveys because the surveys use different samples.
The CE program protects respondents' confidentiality by adjusting responses that may reveal the respondents' identity. For more information, see our PUMD disclosure page.
The number of quarters you use to calculate annual estimates depends on the kind of annual estimate you chose. For example, a 2016 collection year estimate, that is, an estimate from all interviews conducted in 2016, requires files from Q161 through Q164. However, a 2016 calendar year estimate requires all five collection-quarters, Q161 through Q171, because CUs are asked about their previous 3 months of spending.
The estimates published by BLS are based on calendar periods that require the subsequent year's first quarter data. For more information on how to calculate collection and calendar year data, see the Interview documentation files on the PUMD documentation page.
CE adds an "X" to the names of quarterly Interview Survey files that appear twice, once as the fifth and final quarter of the previous year and once as the first quarter of the new year. The "X" indicates that this file differs from the same quarterly file of the previous calendar year release, because it uses the methodology for the new year.
The data in the published tables may differ from results calculated with PUMD for four major reasons:
Here is the glossary of survey terms and item descriptions page.
A consumer unit consists of any of the following:
The terms consumer unit, family, and household are often used interchangeably for convenience. However, the proper technical term for purposes of the CE data is consumer unit.
The reference person of the consumer unit is the first member mentioned by the respondent when asked "What are the names of all the persons living or staying here? Start with the name of the person or one of the persons who owns or rents the home." It is with respect to this person that the relationship of the other consumer unit members is determined.
The CE program includes a few lower levels of geographic data. For additional information, see the CE geography page.
The data shown in the published tables are averages for all consumer units, or for all the consumer units in a particular demographic group. For example, the expenditures, income, and characteristics for the group with a reference person under age 25 are averaged across all consumer units with that characteristic. Because not all consumer units purchase each item during the survey period, the average expenditure for an item is generally considerably lower than the expenditure by those consumer units that purchased that item. The less frequently an item is purchased, the greater the difference between the average for all consumer units and the average for those purchasing the item.
Data users may notice that average annual expenditures presented in the income tables sometimes exceed income before taxes for the lower income groups. Consumer units whose members experience a spell of unemployment may draw on their savings to maintain their expenditures. Self-employed consumers may experience business losses that result in low or even negative incomes, but are able to maintain their expenditures by borrowing or relying on savings. Students may get by on loans while they are in school, and retirees may rely on drawing down savings and investments.
The income taxes increased with the 2013 data because the CE program introduced new estimates of state and federal tax liabilities using the TAXSIM calculator produced by the National Bureau of Economic Research. Beginning with the second quarter of 2013, all state and federal tax amounts used in the tables are estimates based on the expenditures and income and family characteristics. Because changes were introduced part way into 2013 for calculating Federal and State taxes, estimates for 2013 are not strictly comparable to prior or subsequent years. Estimates beginning in 2014 are not strictly comparable to earlier years. The Consumer Expenditure Surveys introduced these estimates to improve the quality of the tax liabilities, which suffered from low response rates, and to improve the estimates of after-tax income. The Consumer Expenditure Surveys gratefully acknowledges the support of the National Bureau of Economic Research for improving the tax estimates.
Prior to the 2013 tables, the published income tax data were collected in the Consumer Expenditure Surveys from respondents and were subject to large non-response errors. The Consumer Expenditure Surveys did not impute values for federal and state income taxes when the amounts were unanswered. An alternative source of data on income taxes as well as other taxes is Tax Stats, produced by the Statistics of Income Division of the Internal Revenue Service.
Nonresponse is a common problem in household surveys, particularly for questions regarding income. Nonresponse means that the respondent either does not know, or refuses to provide, the information requested. Prior to publication of the 2004 tables, the Consumer Expenditure Surveys handled nonresponse to income questions by publishing income data for complete income reporters only. To be classified as a complete income reporter, the respondent had to provide a value for at least one major source of income for the consumer unit. However, not all "complete" reporters provided a full accounting of income for all sources for which receipt was reported.
Starting in 2004, the Consumer Expenditure Surveys introduced multiple imputation to fill in the blanks resulting from nonresponse to income questions. In this method several estimates are made each time the respondent reports the receipt of, but no value for, a particular source of income. The estimates are made based on characteristics of the member or consumer unit for which receipt is reported. The average of these estimates is used to provide the final figures shown in the tables. Because all consumer units now have actual or imputed values for income data available to produce means and other information, the old complete income reporter data are no longer published in tables.
The introduction of multiply imputed data allows for use of the full set of income data from all consumer units, and therefore more accurate comparisons of income and expenditures. For example, instead of producing estimates for complete income reporters, some of whom are still missing income information, income data are now provided in tables for all consumer units with these blanks filled in, resulting in smaller gaps by which expenditures exceed income for low income consumer units. In addition, when examining income by demographic - such as age or composition of consumer unit-income data now are presented for all consumer units within that category rather than for complete reporters only. Consider, for example, consumer units whose reference person is under 25 years old. Prior to 2004, the age tables show average annual expenditures for all consumer units in this age group, but income is shown only for complete reporters in this age group. Starting in 2004, both expenditure and income data are shown for all consumer units within this age group. Therefore, the data are more appropriately compared. For additional information, see Income Imputation Introduced With 2004 Data.
Data collection is carried out by the U.S. Census Bureau under contract with the Bureau of Labor Statistics. In the Interview Survey, each consumer unit is interviewed every 3 months over four calendar quarters. In the initial interview, information is collected on demographic and family characteristics and on the consumer unit inventory of major durable goods. Income and employment information is collected in the first and fourth interviews. In the fourth interview, a supplemental section is administered in order to account for changes in assets and liabilities over a one-year period.
In the Diary Survey, respondents are asked to keep track of all their purchases made each day for two consecutive 1-week periods. Participants receive each weekly diary during the initial visit by a Census Bureau interviewer.
The two survey components, the Interview Survey and the Diary Survey, are designed to collect different types of expenditures. The Interview Survey is designed to obtain data on the types of expenditures respondents can recall for a period of 3 months or longer. These include relatively large expenditures, such as those for property, automobiles, and major durable goods, and those that occur on a regular basis, such as rent or utilities. Each consumer unit is interviewed once per quarter for four consecutive quarters. The Diary Survey is designed to obtain data on frequently purchased smaller items, including food and beverages, both at home and in food establishments, housekeeping supplies, tobacco, nonprescription drugs, and personal care products and services. Each consumer unit records its expenditures in a diary for two consecutive 1-week periods. Respondents are less likely to recall such purchases over longer periods. Although the diary was designed to collect information on expenditures that could not be easily recalled over time, Diary respondents are asked to report all expenses (except overnight travel) that the consumer unit incurs during the survey week.
The CE program is exploring a new design to improve the quality of its estimates. Some key features of the redesign include combining the separate Interview and Diary samples into one, reducing the diary keeping period to one week, providing online personal diaries to allow each household member to enter their own expenses, encouraging respondents to refer to their records when reporting expenses, restructuring the interview to take place over two visits to the household, and having each household participate in the survey for only two waves. The CE redesign (also called the Gemini Project) is currently undergoing field testing. For more information, see the Gemini Project to Redesign the Consumer Expenditure Survey.
More consumer units reported expenditures for health insurance in 2014 than in 2013, and because of an improvement in interview collection methods, higher expenditures were reported. The percent of households reporting quarterly expenditures on health insurance increased from 65.5 percent in 2013 to 68.0 percent in 2014. The insurance questions were revised from 3-month recall questions to questions about the amount of the last payment and the payment period.
The new estimates are more accurate because the respondent does not have to calculate a quarterly estimate-instead the estimate is calculated by BLS, using the amount of the last payment which respondents are more likely to know. On the basis of cognitive testing of these questions, BLS concluded that these new questions produce better estimates. For those consumer units whose time in sample encompassed reporting health insurance expenditures using both the old questions and the new questions, the mean expenditure using the new questions increased by 26.2 percent compared to the old questions. In the 2014 tables, some of the over-the-year change in the healthcare expenditure data, especially in the health insurance subcomponent, is due to these improvements to the survey questionnaire.
The Interview and Diary Surveys are sample surveys and are subject to two types of errors, nonsampling and sampling. Nonsampling errors can be attributed to many sources, such as differences in the interpretation of questions, inability or unwillingness of the respondent to provide correct information, mistakes in recording or coding the data obtained, and other errors of collection, response, processing, coverage, and estimation for missing data. The full extent of nonsampling error is unknown. Sampling errors occur because the survey data are collected from a sample and not from the entire population.
Caution should be used in interpreting the expenditure data, especially when relating averages to individual circumstances. The data shown in the published tables are averages for demographic groups of consumer units. Expenditures by individual consumer units may differ from the average even if the characteristics of the group are similar to those of the individual consumer unit. Income, family size, age of family members, geographic location, and individual tastes and preferences all influence expenditures.
Average expenditures on items at finer levels of detail might not be as reliable as those published for more aggregate levels because there are sometimes few reports of expenditures on more detailed items. A small number of unusually large purchases of infrequently reported items or an increase in the number of consumers reporting such expenditures might cause a large change in the average expenditure from one period to the next. The tables published on the CE website show the expenditure component level at which the estimates are considered to be reliable. However, even in those tables, data in some cells are footnoted as being likely to have large sampling errors due to the scarcity of reports. More detailed tables are available upon request.
Coverage is the extent to which a survey represents its target population. Under-coverage is when some population members do not have a chance of being selected for the sample, and over-coverage is when they have more than one chance of being selected or their addresses are nonresidential. It is important to measure coverage because every segment of the target population needs to be properly represented in order to minimize bias in the survey estimates. The target population of the Consumer Expenditure Survey is the U.S. civilian noninstitutuional population.
The list of households from which the sample of the Consumer Expenditure Survey is drawn is based on the U.S. Census Bureau's Master Address File (MAF) plus a group quarters file.
Sampling error is the difference between the survey estimate and the true population value. The most common measure of the magnitude of sampling error is the standard error. The primary purpose of standard errors is to provide users with a measure of the variability associated with the mean estimates. This variability measures how close different estimates would be to each other if it were possible to repeat the Consumer Expenditure surveys over and over using different samples of consumer units. A small standard error indicates that multiple samples would produce values that are consistently very close to each other, whereas a large standard error would indicate that multiple samples would produce values that are not close to each other.
The coefficient of variation is the standard error divided by the mean expenditure, and it is expressed as a percent. It gives the relative amount of variability instead of the absolute amount of variability in the expenditure estimates.
Beginning with year 2000 data, the CE program has made available standard error tables using integrated data from both surveys. The separate standard error table format has been discontinued with the release of the 2012 data. These separate data tables have been replaced by the combined expenditure, share, and standard error tables.
Also available is an article that describes standard errors of the expenditures and income estimates in the Consumer Expenditure Surveys.
Variability limits your ability to use the data with precision and certainty. Measures of variability inform you about the range of possible values around a particular data point. Generally, the larger the range, the lower your confidence in the precise data point. For more information, see "How does the variability of Consumer Expenditure data impact your analysis?"
Last Modified Date: June 9, 2022