A Conceptual Framework for the U.S. Consumer Expenditure Surveys

By Scott Curtin, Adam Safir, Thesia I. Garner, Brett Matsumoto, and Jake Schild

Introduction

BLS studies of family living conditions rank among its oldest data-collecting functions, going back to the first consumer expenditure survey in 1888-91. The survey objectives have always been to provide data on the buying habits of U.S. consumers for use in a wide variety of economic research and in support of periodic revisions of the Consumer Price Index (CPI). Consumer Expenditure Surveys (CE) data are used for a variety of purposes, e.g., producing weights for the CPI, evaluating the effects of tax policy changes, examining the buying habits of certain groups of consumers, and obtaining a measure of economic well-being. These various uses require alternative concepts of expenditures or consumption.

The measurement of consumption is crucial for any assessment of living standards (or living conditions) or the examination of consumer behavior. Consumption has been defined as “the use of good or services to directly satisfy a person’s needs or wants, while consumption expenditure is the value of consumption goods or services paid for by a household¹.” Researchers have argued that consumption provides the “best” measure of economic well-being and that consumption should be the basis of a cost-of-living measure. Determining the actual consumption of a particular household or family is difficult because it depends on the household’s particular circumstances, as well as their choices and uses of purchases. As the UNECE Guide on Poverty Measurement (2018) notes, low levels of consumption may reflect individual decisions or non-monetary constraints, such as physical limitations. As a result, many studies use expenditures or outlays as a proxy to measure consumption and construct price indexes (as in the CPI).

A conceptual framework for the CE is an attempt to highlight both the conceptual and empirical issues in measuring the “spending patterns” of consumers. This conceptual framework describes both the scope and content of the CE surveys, with the goal of providing a reference to guide possible changes in the instrument, alternative uses of the data, and definition of terms used in publication. In general, the scope of the CE is defined to be what is theoretically within the nature of a consumer's expenditure, along with the necessary demographic characteristics, such as income, and other items necessary to provide the ability to analyze those expenditures.

To allow for a variety of research uses, the survey should allow for various definitions of expenditures and income. This document provides a component-level approach to measuring expenditures and other related information.

Background

The initial CE Conceptual Framework Team began meeting in the summer of 1999. The team's mission was to prepare written guidelines and rules to be used in selecting and assigning priorities to changes to the survey.

The conceptual framework began with the CE Mission Statement: "The BLS mission for the CE program is to collect, process, and disseminate high quality data that present an unbiased statistical picture of consumer expenditures for the Consumer Price Index, government agencies, and other data users, in support of a better understanding of consumer economic behavior. BLS staff accomplish this mission through a commitment to methodologically sound surveys, efficient data processing, timely data dissemination and analysis, continuous data user education, and high standards for accuracy, relevance, timeliness, and respondent confidentiality."

To meet this mission, the Bureau of Labor Statistics (BLS) must try to provide quality data and balance the needs of customers with the burden on respondents and total cost.

The objectives of the original CE Conceptual Framework Team were two-fold:

• To draft a general statement outlining the content and scope of the CE.
• 
  
• Upon review and acceptance of this statement by management, the team will then recommend an approach to evaluate future proposed changes to the content of the CE. The approach may include such criteria as difficulty, usefulness, potential error of the question/response changes, and should be consistent with the mission and objectives of the CE and the BLS.

This revised paper consists of an Introduction; Section I defining content and scope; Section II describing consumer units, expenditures, and other necessary information; Section III examining the population coverage and other content issues; and Section IV giving an overview of the theoretical scope and possible contents of expenditures and income. The revisions also reflect the introduction of income imputation and the use of TAXSIM to imputed federal and state income taxes.

Section I. CE Content and Scope

BLS should provide a continuous and comprehensive flow of income and consumer expenditure data in order to analyze the expenditure patterns of individual consumer units in the U.S., determine the aggregate expenditures for all consumers in the U.S., support periodic revisions of the CPI, and be of use in a wide variety of economic research.

A search of legislation involving the CE conducted by the BLS Division of Management Systems for this team did not turn up direct legal mandates spelling out what the BLS should measure and publish. No further efforts to identify direct legal mandates have been undertaken since. In 1999 Congress considered draft bills that would use certain CE data in bankruptcy proceedings. If these were to become law, or if legislation were enacted requiring the IRS or other federal agencies to use specified BLS consumer expenditure data products, then the BLS would work within that framework.

Although not legislation, two interagency technical working groups have recommended that the CE data be used as the basis for two poverty measures: the Supplemental Poverty Measure (SPM), and a consumption measure of poverty. In 2010, an Interagency Technical Working Group (ITWG) recommended that the BLS use the CE Interview to produce SPM thresholds². This measure built upon the recommendations in a 1995 report by the National Academy of Sciences Panel on Poverty and Family Assistance³. Another ITWG released a report in 2021 recommending that the BLS produce not only a consumption measure, but also a consumption measure of poverty⁴. The BLS has been producing SPM thresholds each year since 2011 and providing these to the Census Bureau, which then in turns produces SPM poverty reports⁵.

BLS must collect a time series set of current expenditures by the U.S. population for a variety of purposes. In addition to providing a market basket for the CPI by supplying weights and relative importances, the information is used by other Federal agencies including the Federal Reserve, the Bureau of Economic Analysis, and the Health Care Finance Administration.

The scope of the CE is defined to be what is theoretically within the nature of a personal consumer's expenditure, along with the items necessary to provide the ability to analyze those expenditures. In-scope expenditures include goods and services but exclude business expenses and home production. Other items considered necessary are income, supplemental demographic information, and other financial information.

The content of the CE may be different from the entire theoretical scope because of limits imposed by practicality, costs, and the difficulty of defining, measuring, or collecting the data, as well as the need for additional information to meet the needs of customers or to aid in recall by the respondents, such as asking for the reason for a trip.

Since the overall purpose of the CE is to obtain consumer expenditures, the next part of this document discusses the definition of the unit of analysis – consumer -- their expenditures and finally, the other necessary items.

The focus on consumer expenditures means collecting information about the value of the goods and services acquired during a given time period, including goods and services for the consumer unit’s (CU’s) own use and purchased as gifts to be given to people who live outside the consumer unit. This is different from consumption, which is the amount of goods and services that are used up in a given time period by the consumer unit for its own use. This also leaves out quantity, price per unit, and point of purchase information. Also not included in consumer expenditures are consumer’s business purchases. Finally, within the CE scope, consumer expenditures do not include the value of home production by the consumer, the value of goods and services received from people who do live within the consumer unit, and the transfer of goods and services from business government agencies.

Section II. Supporting Topics and Concepts

Consumer Unit

A consumer unit is a single financially independent person or a group of persons who share expenses. The CU is defined as: (1) all members of a particular household who are related by blood, marriage, adoption, or other legal arrangements; or (2) a person living alone or sharing a household with others or living as a roomer in a private home or lodging house or in permanent living quarters in a hotel or motel, but who is financially independent; or (3) two or more persons living together who are assumed to pool their income to make joint expenditure decisions. Financial independence is determined by the three major expense categories: Housing, food, and other living expenses. To be considered financially independent, at least two of the three major expense categories must be provided entirely, or in part, by the respondent. The purpose of measuring the CU is to obtain accurate information on the expenditures of groups of people who engage in joint expenditures, and who implicitly have access to the goods and services purchased.

This unit of measure is very similar to, but may provide more accurate expenditure data than, a "household" measure. A household, by definition, includes all people living in a housing unit with a share entrance to the living quarters. This can include people paying rent for a room but who do not share living expenses. The rest of the household is not likely to be able to accurately know what is purchased or how much is spent by roomers or borders, so the aggregate totals are not collectible.

This choice of measuring and publishing data at the CU level could change if we collect complete individual member data or a new unit of measure becomes theoretically more appropriate. The BLS investigated the use of web-based individual diaries in 2015, but determined that overall, providing CUs with such diaries did not represent an improvement over having CUs report by a single proxy diarist using a paper diary⁶. With improvements in technology and collection methods, the BLS could again consider the use of individuals diaries to improve expenditure estimates.

Expenditures

The level of expenditures can depend upon the timing of the expenditure – does it occur at the time of acquisition, when it is paid for, or when it is consumed? The BLS definition of expenditures is based on what is spent or financially obligated at the time of acquisition in dollar terms for any type of payment method (e.g., a person used a credit card for a purchase with the expectation that what was charged will be paid in the future). This is true for all goods and services that are acquired by the consumer unit⁷. Depending upon the good or service, the reference period for “the time of acquisition” could be a day, a week, a month, in the past three months, or in the last 12 months. For the CE, goods and services are restricted to those intended to increase the current economic well-being (or consumption) of the consumer unit as well as those purchased to be given to people who live outside the consumer unit (e.g., a gift of clothing given to a friend). For all but purchased housing, the full cost of purchases is recorded even though full payment may not have been made at the date of purchase. As such, excluded from consumer expenditures are periodic credit or installment payments on goods or services already acquired. In the case of purchased housing, mortgage interest payments are included as well as those for home equity loans and lines of credit; mortgage and loan principal payments are not included in the official CE definition for owned housing shelter as these payments are considered savings. The Section IV provides alternate definitions of expenditures that could be obtained using CE data.

Income

The collection of income and financial resource information in the CE is important for the analysis of spending patterns and the verification of actual spending. Since one purpose of the CE survey is to measure “spending patterns”, it is expected that a measure of income would be associated with the expenditures at the CU level. It is important to note, however, that income is not a measure of achieved living standards, but rather only of potential living standards.

There are several alternate definitions of income used for different types of analysis. In order to support these different definitions, BLS should provide component-based collection of income. This will allow the CE Division, BLS researchers, and other data users to build the definition of income that is appropriate for their uses. For example, for CE publication purposes, the cash value of Supplement Nutrition and Assistance Program (SNAP) benefits is included income since it is assumed that SNAP benefits are used before other financial instruments to purchase food. However, to produce an income measure that is comparable to that published by the Census Bureau, the value of these SNAP benefits would need to be subtracted from the CE definition of income, which is defined as before tax and in-kind transfers. It is important to note that the CE publication table definition of income also includes the value of Supplemental Nutrition Assistance Program (SNAP) benefits, rent as pay, and any recovery rebates⁸; however, these are not included in the microdata definition of before or after-tax income. See Appendix A for a discussion of the definitions of income and a classification system of income components.

BLS staff are also collaborating with Census Bureau staff on several research projects related to the use of IRS income data to improve CE estimates. This includes a project using matched IRS data for potential use in income imputation, and a nonresponse bias study using title 26 household-level IRS income variables instead of publicly available zipcode-level IRS data for the survey’s nonresponse adjustment process.

Demographics

In addition to income, the CE should continue to collect useful and obtainable demographic information if it can be gathered in an accurate and cost-effective manner. There should be a collected core set of agree-upon variables as reliability, interview length and cost will permit.

The demographic data that the CE collects should be based on the following requirements:

1. be necessary for the CPI (such as location⁹, age of the reference person and spouse¹⁰, earner status or occupation)¹¹, or
2. be necessary for weighting (region, household size, contact attempts, zip-code level IRS adjusted gross income, age, race, rental/owner tenure, division, Hispanic/non-Hispanic, and urban/rural), or
3. be a key explanatory variable in economic analysis (education and age),
4. be necessary for survey research into improving survey quality, or
5. be necessary for comparing CE data to those collected by other major surveys.

The determination of what variables are "key explanatory" variables is subject to the particular use. Other interesting demographic variables may be collected as resources permit (for example, disability status), but should not be considered integral to the core scope of the CE. Additional variable collection will need to compete with other items for inclusion in the survey based on available resources.

Assets and Liabilities

A measurement of assets and liabilities is within the scope of the CE, because it can be useful to explain expenditures, and to support research into the expenditure life-cycle theory – the idea that CU expenditures are based on a combination of current and expected available financial resources. The net change in assets and liabilities could also be included in the content of the CE because it can be useful to examine the quality of the expenditure and income data, via the use of account balancing¹².

Section III. Coverage Scope and Content

Population

The CE should provide for the U.S. as a whole an accurate estimate of aggregate consumer expenditures by the civilian non-institutionalized population. The CPI requires accurate weighted relative shares at the urban core-based statistical area (CBSA) level. The CPI is published at the following levels: U.S. city average, Census region, and for select metro areas¹³.

Time Period

CE data are published as annual calendar year expenditures, usually in September for the previous year. Microdata, or CU-level records, are also released at the same time as the published estimates.

Time Series Continuity

As the CE evolves and improves, the ability to provide long-term signposts becomes important. Changes to definitions of key variables, and the addition or deletion of key variables should still allow for the production of a continuous set of survey results. Users of the published and microdata should still be able to compare earlier years' data with current data as much as possible. Changes should be documented clearly so that future BLS staff and customers can understand what occurred, when it changed, and why the change was made.

Section IV. Theoretical Scope of Expenditures

Table 1 in this section describes the theoretical scope of expenditures by major categories. The following are the definitions of the columns used in the table, with additional details on select terms.

Consumption Expenditures

The Consumption Expenditures column represents the expenditures incurred by consumer units on consumption goods and services that are acquired by purchase or through transfers received from government or other private institutions. These goods and services are restricted to those intended to increase the current economic well-being (or consumption) of the consumer unit. This measure includes the acquisition costs as measured by actual financial obligation at the time of acquisition in dollar terms, for any type of payment method, of all goods and services that are acquired by the consumer unit from outside the consumer unit. This measure is similar to final consumption expenditure used in the System of National Accounts (SNA) if the “expenditure” on owned homes is measured by the value of the flow of housing services.

Total Expenditures

The Total Expenditures column is the current definition of expenditures used in the CE published estimates. This measure includes the expenditures on goods and services for current consumption plus expenditures for goods and services expected to be used for the future consumption of the consumer unit or to be transferred to other consumer units. Total expenditures consist of the transaction costs, including excise and sales taxes, of goods and services acquired during the interview or recordkeeping period. Expenditure estimates include expenditures for gifts but exclude purchases or portions of purchases directly assignable to business purposes. Also excluded are periodic credit or installment payments on goods or services already acquired. The full cost of each purchase is recorded even though full payment may not have been made at the date of purchase.

CE Total Outlays

The Total Outlays column represents the out-of-pocket outlays of consumers. This measure is similar to Total Expenditures, but with the modifications that the net purchase price of financed vehicles is excluded, payments on principal on all financed vehicles (identified in both transportation and entertainment in the CE) are included, and payments of principal on home mortgages are included.

CE Current Consumption Expenditures

The Current Consumption Expenditures column was used in earlier versions of the CE publications¹⁴. In earlier publications, these represented expenditures that a family (as opposed to a consumer unit) incurred for current consumption--that is, what the family spent for itself. This same definition can be applied for CUs. Current consumption expenditures consist of the transaction costs, including excise and sales taxes, of goods and services acquired during the interview or recordkeeping period for the CU’s own consumption. Current consumption expenditures exclude those for cash contributions; endowments, annuities, life insurance, and other personal insurance; retirement, pensions, Social Security, and principal payments on mortgages and other loans. When data are available, expenditures for gifts of goods, services, and money (or other financial instruments) for people outside the CU are also excluded. Depending upon the one’s interpretation of expenditures for education and health – as investments or for current consumption – these too could be excluded from the definition of Current Consumption Expenditures.

Consumption (theoretical)

Consumption represents the value of purchased services and non-durable goods, and the value of the service flows of durable goods and owner housing. This consumption reflects both purchases made by the consumer and for the consumer in the case of in-kind transfers. This measure includes the items in current consumption expenditures but excludes the purchase price of durables and includes an estimate of the service flows of these items. This measure also includes the value of consumption goods that are received from outside the consumer unit through barter or as in-kind transfers, and the value of home-produced goods and services for the CU’s own consumption.

Barter

If the receipt and giving of bartered goods and services from others is considered to entail acquisition costs as measured by actual financial obligation at the time of acquisition in dollar terms, for any type of payment method, then they are included in the theoretical scope of consumer expenditures. However, the barter of goods and services received from others is excluded from the content of the CE because of the extra burden involved in collecting the information, and the potential difficulty in determining accurate dollar values for in-kind reciprocal transactions.

Durable Goods

How to best measure expenditures for durable goods is evolving as research continues in BLS and elsewhere. For owner housing, presently the CE collects data on both the out-of-pocket expenditures and flow-of-services rental equivalence. From 1953 through 1983 the CPI required a modified data set of cost information for owned housing but switched to requiring flow of service estimates from then to the present. The CE will need to continue to provide both kinds of data to its chief customers.

For CE publications, and use for the production of the CPI, the purchases of vehicles are measured as a net price (e.g., the cost of a new car minus any trade-in value). The CE does not collect a flow-of-service estimate for the amount of the vehicle consumed or depreciated annually. However, research within the BLS Division of Price and Index Number Research (DPINR) is being conducted to estimate the value the flow of services from cars and trucks using CE Interview data, building on the work of other researchers¹⁵. Research developments by CPI staff has recently focused on rental leasing arrangements which may lead to an approximation of the flow of services of cars and trucks.

Goods or Services

For CE publications and Total Expenditures, these categories also include purchases intended as gifts for others outside of the CU as well as charitable contributions. Included in total food are expenditures for food as pay and SNAP benefits; included in total renter expenses is rent as pay. Excluded from the Goods or Services categories are the net value of bartered goods and services, those from home production. Similar to the SNA, bater is excluded because of the extra burden in asking questions to collect the information and expected unreliability of determining dollar value.

Home Production

The CE publication of total expenditures excludes home production. In contrast, unlike for the CE, food produced on farms for own consumption is included in the SNA as part of PCE. In addition, the BEA produces a Household Production Satellite Account¹⁶.

1. OECD (2013), OECD Framework for Statistics on the Distribution of Household Income, Consumption and Wealth, OECD Publishing, Paris, Chapter 5. Household Consumption, p.104
2. Flow of services from major appliances (e.g., stoves, refrigerators, dishwashers, washers, dryers) are assumed to be included in reported rental equivalence of owner-occupied housing since in the U.S. such appliances are included in rental units when they are rented and thus are included in reported rents paid by renters.
3. Cost above interest.
4. Healthcare and Education contain aspects of both consumption and investment.
5. "Rent as pay" and "Meals as pay" are included; implicitly SNAP benefits are included as they are considered like monetary instruments and assumed to be a source to pay for food.

Appendix

The measurement of income can depend upon the particular use, e.g., tax analysis, savings, or general well-being. One concept of income is the potential to spend for desired goods and services. Our component-based collection of income will depend on the predictability (or regularity), liquidity (i.e., the potential to convert the component into cash or other financial instrument) and flexibility (the potential to increase spending on desired goods and services) of the component. The more liquid and flexible the component, the closer the component is to actual income. The component-based measurement of income depends on the possibility of determining a dollar value for the component, the total value of the component, and the number of people receiving the component. For example, wages and salaries are reliably measured, large and received by many people. They are also liquid and flexible. While SNAP benefits as electronic benefits satisfy these criteria, these benefits are as not as liquid or flexible. In addition, one could question the reliability of SNAP benefits recalled over the past 12 months.

In comparison to the original 2000 BLS internal Framework document, variable names for income have been added to reflect the use of income imputation and the use of TAXSIM to estimate federal and state income taxes.

Income Definitions

The following definitions of income should be supported by the CE collection instruments

1. The net increase in the purchasing power to consume. (Haig-Simons definition)
2. The total amount of regular cash income plus “near-cash” government transfers defined as before tax income¹⁷ and after-tax and transfer income¹⁸. Included in these are the cash value of Supplemental Nutrition Assistance Program (SNAP) benefits¹⁹. Income imputation was introduced in 2004 quarter one, and the imputation of federal and state income taxes in 2013 quarter two with the use of the NBER software TAXSIM.
3. The amount a household can afford to consume during the reference period without having to finance its consumption by reducing its cash, by disposing of other financial or non-financial resources, or by increasing its liabilities.
4. Census Bureau’s definition for published reports: Money income excluding capital gains before taxes and in-kind benefit transfers.

Classification of Income Components

Given all of the various definitions, what follows is a possible classification of income components. The Concepts Team did not recommend that all of these be included in the theory of income or be collected by the surveys. Instead, they noted that some of these components are more income-like than others, based on liquidity, flexibility, measurability, and because they occur on a regular or predictable basis. The dividing line on the ambiguous elements between what should be in the CE and what should be out needs to be determined by a further research effort.

Regular (Or Predictable) Monetary Income

A. Earnings of regular monetary income from employment

1. Wages, salaries,
2. Self-employment

B. Regular monetary income from investment

1. Net rental income
2. Interest, dividends
3. Royalties or income from estates or trusts

C. Regular income from retirement, survivor, disability pension plans

D. Regular monetary income from government

1. Social security Benefits and Railroad Retirement
2. Supplemental Security Income (SSI)
3. Veteran’s Administration
4. Unemployment compensation
5. Worker’s compensation
6. Cash assistance (e.g., TANF, short term emergency, job training such as Job Corps)
7. Earned Income Tax Credit (EITC)

E. Other regular cash income from non-CU sources

1. Alimony
2. Child support
3. Other regular income transfers from non-CU members including from individuals, cash scholarship and fellowships, stipends not based on working, or from the care of foster children

Irregular Monetary Income from Government²⁰

F. Covid-related Economic Stimulus Payment (EIP)

G. Expanded Child Tax Credit (CTC)

H. Expanded unemployment compensation

Irregular Monetary Income

I. Irregular monetary income from changes in investments

1. Realized capital gains

J. Lump sum monetary or cash equivalent income from non-CU sources

1. Lottery winnings (games of chance)
2. Payments from insurance, estates, trusts, royalties, alimony, prizes
3. Income transfers from non-CU members

In-Kind Transfer of Goods and Services

K. In-kind transfers from employers (regular or irregular):

1. Meals as pay, rent as pay, etc.
2. Health insurance (full or partial employer contribution for the purchase)

L. Non-housing in-kind transfers from government

1. Supplemental Nutrition Assistance Program (SNAP).
2. Health insurance (Medicare, Medicaid, Tricare, CHIP, Affordable Care Act, etc.)
3. National School Lunch Program (NSLP) benefits
4. Special Supplemental Nutrition Program for Women, Infants, and Children (WIC) benefits
5. Low Income Home Energy Assistance Program (LIHEAP) benefits

M. Housing assistance from government

1. Rental vouchers or Section 8 housing
2. Other types of public housing

N. Income from home production of services from home production

1. Net value of housing services from owner-occupied housing
2. Net value of services from household consumer durables (not already accounted for in housing services from owner-occupied housing)
3. Net value of unpaid domestic services

O. In-kind benefits from non-CU members

1. Food banks
2. Charity handouts
3. Gifts
4. Implicit income from living rent-free

Change in Net Worth

P. Potential Income through changes in net worth

1. Unrealized capital gains (losses)
2. Transfers of non-monetary assets from non-CU members
3. Financial contributions by employer to employee pension or retirement plan

Q. Actual income in current period through changes in net worth

1. Realized capital gains
2. Withdrawals from savings, pension or retirement plans, liquidation of assets
3. Changes in debt

Definitions Using Components

1. Consumer Expenditure Survey definition of before tax income: FINCBTAX = Regular (or Predictable) Monetary Income (excluding tax credits such as EITC, EIP, CTC) + In-Kind Transfer of Goods and Services (employer provided food and rent as pay) + Non-housing In-Kind Transfer from Government (SNAP)
2. Census Bureau’s official definition of money income²¹ = Regular (or Predictable) Monetary Income (including expanded UI and excluding tax credits such as EITC, EIP, CTC).

Consumer Expenditure Conceptual Framework Team 1999-2000

• Jeffrey Blaha, SMD
• Sioux Groves, P&C
• David Johnson, DPINR
• David Swanson, SMD
• Richard Dietz, P&C
• Steve Henderson, BIA, Chair
• Bill Passero, BIA

Updated in 2022 by

• Scott Curtin, BIA Chief
• Adam Safir, CE Chief
• Thesia I. Garner, DPINR Chief
• Brett Matsumoto, DPINR
• Jake Schild, DPINR

If you are deaf, hard of hearing, or have a speech disability, please dial 7-1-1 to access telecommunications relay services or the information voice phone at: (202) 691-5200. This article is in the public domain and may be reproduced without permission