Search BTN:

• HOME
• ARCHIVE
  • Archive By Date
  • Archive By Program
  • Archive By Publication
  • Archive By Theme
• ABOUT
• SUBSCRIBE

Changing how BLS defines "geographic areas" in the Consumer Expenditure Surveys

By Scott Curtin

The Consumer Expenditure Surveys (CE) provide the most detailed data on expenditures, income, and demographic characteristics collected directly from U.S. consumers by the U.S. Bureau of Labor Statistics (BLS). These data are not just nationally representative, they also provide varying levels of geographical information—this includes selected Metropolitan Statistical Areas (MSAs), regions, divisions, and different types of areas (e.g., urban or rural). Population growth and expansion among smaller cities and towns amplify the need to analyze and understand expenditure patterns for growing areas that may affect overall economic activity in the United States.

Starting with the 2021 CE data products, BLS updated its area definition to use both the Office of Management and Budget (OMB) principal city designation and the Census Bureau’s designations for urban and rural areas. This change is intended to update an outdated city definition and align urban and rural designations with the more broadly accepted definition maintained by the Census Bureau. This Beyond the Numbers article aims to provide an understanding of the changes, present findings for how these new definitions compare with the old definitions used to develop estimates prior to 2021, and discuss how these changes will impact data users.

How are area types redefined

BLS has historically used MSAs as the primary way for distinguishing urban consumers for the CE. An MSA is defined as an area that contains at least one urban area (UA) with a population of 50,000 or more.¹ Starting in 2003, BLS started to use the OMB designation of “central city,” to define and distinguish areas in their products too.² Originally defined as the largest city in a MSA, the central city designation continued to evolve with new MSA designations driven by population growth and urban development. Figure 1 highlights how these two concepts have been organized in CE products through 2020.

In December 2000, OMB announced standards for defining Metropolitan and Micropolitan Statistical Areas (MiSA), which included a new definition for designating area types called principal city, replacing the central city area definition.³ While the standards for defining MiSAs were once again updated in 2010, the principal city designation was preserved.⁴ As shown in figure 2, the Census Bureau’s urban-rural delineation does not utilize MSAs directly, but rather identifies an area as urban if the area is considered a UA or an urban cluster (UC). A UC is an area of at least 2,500 and less than 50,000 people. Rural areas are defined as all other areas that do not meet the urban classification requirements.⁵

The key difference between the Census Bureau’s definition and the definition used by BLS for the CE up until 2020 is that the Census Bureau’s definition of urban is not restricted to those consumer units (CUs) that reside inside of an MSA, thereby allowing urban and rural areas to exist in and outside of MSAs.⁶ The new definition shows a clear overlap in concepts, which allow for principal cities, MSAs, and MiSAs to be associated with rural areas. MiSAs are defined as an area that contains at least one UC with a population greater than 10,000, but less than 50,000. Together, these concepts form a core-based statistical area (CBSA) and have been the defining characteristics for determining urban and rural areas in CE. It is also important to note that the Census Bureau’s definition for urban and rural does not include a distinction for suburban, and instead uses the UA or UC designations to distinguish the level of urbanicity for a given area.

City redefined

In December 2010, OMB approved the designation of principal cities to create a less restrictive city definition that allows for the inclusion of smaller cities.⁷ Much like the central city definition, a principal city is predominantly defined as the largest place within a CBSA. Additional principal cities can and are established within a CBSA based on conditions related to population, job access, and employment.⁸ According to the Census Bureau’s March 2020 delineation of principal cities of metropolitan and micropolitan statistical areas, 1,269 principal cities were designated across 384 MSAs and 542 MiSAs.⁹ Chart 1 shows the number of principal cities within each state, with Texas (95) and California (91) having the largest number of principal cities among all states.

Looking at 2021 CE data, characteristics of CUs located inside of principal cities were very similar to those found in central cities, with population counts for principal cities only accounting for 1 percent less than the central city. As shown in table 1, demographic estimates among both groups are nearly identical.

Table 1. Selected demographics by central and principal city, Consumer Expenditure Surveys, 2021

Demographic	Central city	Principal city
Number of consumer units	48,486,782	47,997,974
Consumer unit characteristics
Age of reference person	49.45	49.32
Average number in consumer unit
People	2.3	2.3
Children under 18	0.5	0.5
Adults 65 and older	0.4	0.4
Earners	1.2	1.3
Vehicles	1.5	1.5
Reference person
Men	47	47
Women	53	53
Housing tenure
Homeowner	51	51
Renter	49	49
Race of reference person
Black or African-American	19	19
White, Asian, and all other races, not including Black or African-American	81	81
Hispanic or Latino	19	19
Not Hispanic or Latino	81	81
Education of reference person
College	69	70
Source: U.S. Bureau of Labor Statistics.

In 2021, there were minimal differences across the two city definitions for each of the 14 expenditure categories. (See chart 2.) When comparing expenditure estimates from principal cities to those of central cities using the 2021 data, little change was observed. This lack of change can be expected given that most CUs associated with principal cities also reside in urban areas.

As seen in table 2, expenditures on transportation did decline under the principal city definition because of slightly less reliance on personal transportation for CUs within principal cities. Reclassification to principal cities also lowered expenditures on vehicle purchases, fuels, and other vehicle expenses.

Table 2. Income and expenditures by principal and central city, Consumer Expenditure Surveys, 2021

Item	Principal City	Central City
Income
Income before taxes	$80,740	$80,888
Expenditures
Food	8,072	8,034
Food at home	5,052	5,018
Food away from home	3,020	3,016
Alcoholic beverages	586	592
Housing	21,990	22,021
Shelter	13,297	13,369
Utilities, fuels, and public services	3,791	3,799
Household operations	1,487	1,449
Housekeeping supplies	721	722
Household furnishings and equipment	2,694	2,682
Apparel and services	1,804	1,786
Transportation	9,243	9,395
Vehicle purchases (net outlay)	3,822	3,955
Gasoline, other fuels, and motor oil	1,749	1,758
Other vehicle expenses	3,135	3,154
Public and other transportation	537	528
Healthcare	4,518	4,526
Entertainment	2,956	2,986
Reading	119	120
Education	1,339	1,350
Tobacco products and smoking supplies	295	298
Miscellaneous	866	849
Cash contributions	2,085	2,045
Personal insurance and pensions	7,281	7,267
Personal care products and services	748	737
Source: U.S. Bureau of Labor Statistics.

Urban and rural areas redefined

A change in definition is certainly expected to have implications on the underlying sample and how respondents are categorized. Chart 3 shows nearly a 12 percent shift in population from urban to rural in 2021 from the current CE urban-rural definition to the new Census Bureau definition.

Impacts on public use microdata

This definitional change is expected to have implications for CE public use microdata (PUMD) and tabular data users alike. For example, in the microdata, the variable “BLS_URBN,” which previously housed the indicator designating urban and rural CUs will be phased out and replaced with a new variable “URBAN.” This new variable will be coded in the same fashion as BLS_URBN with “1” for urban consumer units and “2” for rural consumer units, but the percentage of “1” and “2” observations in any given quarter of microdata is expected to change with the new definition. Similarly, on tables, one expects that there will be at least some change in demographic and expenditure patterns, based on the regrouping caused by the definitional change. For this reason, it is useful to compare characteristics computed using 2021 data under the current definition and under the new definition to assess the magnitude of the change.

Impact on demographics

While demographic characteristics of urban CUs showed little change under the two definitions, comparisons within the rural CU definition did yield some interesting results. As shown in table 3, both the average age of the reference person and percent of renters declined under the Census Bureau’s definition, while the percent of earners, homeowners, and CUs with at least one college graduate, increased under the new definition.

Table 3. Selected demographics by urban and rural, Consumer Expenditure Surveys, 2021

Demographic	Urban		Rural
	Old definition	New definition	Old definition	New definition
Number of consumer units	124,118,026	109,370,402	9,477,021	24,224,645
Consumer unit characteristics
Age of reference person	51.5	51.1	56.5	55.2
Average number in consumer unit
People	2.4	2.4	2.5	2.6
Children under 18	0.6	0.5	0.6	0.6
Adults 65 and older	0.4	0.4	0.6	0.5
Earners	1.3	1.3	1.1	1.2
Vehicles	1.8	1.7	2.4	2.5
Reference person
Men	48	47	46	48
Women	52	53	54	52
Housing tenure
Homeowner	64	61	79	84
Renter	36	39	21	16
Race of reference person
Black or African-American	14	15	5	6
White, Asian, and all other races, not including Black or African-American	86	85	95	94
Hispanic or Latino	16	17	4	5
Not Hispanic or Latino	84	83	96	95
Education of reference person
College	70	71	51	60
Source: U.S. Bureau of Labor Statistics.

Chart 4 shows that the lack of change within the demographics among urban CUs caused little difference in urban expenditure patterns as observed under either definition. On the other hand, rural estimates did show deviation between the two definitions. Looking at the 14 major spending categories, all but 1 category—tobacco and smoking supplies—showed an increase in average annual expenditures, with housing increasing $5,253 (37 percent) to $19,583, and personal insurance and pensions increasing $2,400 (47 percent) to $7,533. Transportation and food also showed notable increases in average annual expenditures of $1,504 (12 percent) and $1,486 (24 percent), respectively.

The increase in rural expenditures under the Census Bureau’s definition is attributed to the reclassification of a large portion of CUs from urban to rural. The rural population more than doubled under the new definition from 9,477,021 to 24,224,645, while introducing many higher income CUs. As seen in table 4, income before taxes rose 38.5 percent and ultimately attributing to the widespread increases in expenditures. While urban expenditures did, for the most part, stay stable across all spending categories, a noticeable decline in transportation can be observed, which is primarily the result of personal vehicle reliant CUs being reclassified as rural under the new definition. The reclassification of urban CUs lowered expenditures on vehicle purchases, fuels, and other vehicle expenses in the urban category. Public transportation did increase slightly under the new definition for urban, further supporting the notion that consumer units under the new urban definition are less reliant on private means of transportation and more on public.

Table 4. Income and expenditures by urban and rural, Consumer Expenditure Surveys, 2021

Demographic	Urban		Rural
	Old definition	New definition	Old definition	New definition
Income
Income before taxes	$89,341	$87,642	$62,440	$86,487
Expenditures
Food	8,457	8,409	6,285	7,771
Food at home	5,340	5,311	4,300	5,036
Food away from home	3,117	3,098	1,985	2,735
Alcoholic beverages	579	578	253	449
Housing	23,266	23,304	14,330	19,583
Shelter	13,781	14,007	6,412	9,877
Utilities, fuels, and public services	4,225	4,130	4,207	4,644
Household operations	1,684	1,672	1,043	1,488
Housekeeping supplies	817	790	634	859
Household furnishings and equipment	2,759	2,705	2,033	2,715
Apparel and services	1,817	1,786	996	1,623
Transportation	10,870	10,362	12,161	13,665
Vehicle purchases (net outlay)	4,721	4,352	6,236	6,976
Gasoline, other fuels, and motor oil	2,108	2,008	2,664	2,777
Other vehicle expenses	3,571	3,512	3,054	3,633
Public and other transportation	470	490	208	280
Healthcare	5,412	5,169	5,973	6,727
Entertainment	3,577	3,389	3,450	4,366
Reading	118	123	59	77
Education	1,288	1,276	411	1,003
Tobacco products and smoking supplies	319	312	638	476
Miscellaneous	982	912	1,217	1,389
Cash contributions	2,416	2,299	2,391	2,937
Personal insurance and pensions	8,088	7,954	5,133	7,533
Personal care products and services	782	795	615	662
Source: U.S. Bureau of Labor Statistics.

Conclusion

Understanding precise definitions of demographic groups is key to interpreting related estimates, such as expenditures by group. With continued population growth and expansion among smaller cities and towns, the landscape for how urbanicity is defined is ever-changing. As this analysis shows, how boundaries of an area are defined directly impacts not only the socioeconomic characteristics but also the spending behavior. BLS plans to continue to monitor and evaluate further geographic definitional changes in the effort to ensure that CE data users continue to receive high-quality data that present an unbiased statistical picture of consumer spending.

Starting with the 2021 data release, detailed estimates utilizing the new definitions of urban and rural (table 1721) are available and can be found on the CE calendar year tables webpage. Table 1720, which utilized the old definition of urban and rural, was not published as part of the 2021 data release and will be discontinued starting with the 2022 data release. For comparison purposes, table 1720 populated with 2021 estimates is available upon request, allowing users a one-time (1-year) comparison table.

This Beyond the Numbers article was prepared by Scott Curtin, economist in the Office of Prices and Living Conditions, U.S. Bureau of Labor Statistics. Telephone: 202-691-6900. Contact the Consumer Expenditure Surveys program directly by email: CEXInfo@bls.gov; or by completing the form at https://data.bls.gov/cgi-bin/forms/cex?/cex/csxcont.htm.

If you are deaf, hard of hearing, or have a speech disability, please dial 7-1-1 to access telecommunications relay services. This article is in the public domain and may be reproduced without permission.