Beyond the Numbers
January 2014 | Vol. 3 / No. 1
PRICES & SPENDING

New education classification better reflects income and spending patterns in the Consumer Expenditure Survey

By Ann C. Foster

An individual’s level of education and associated earnings profoundly influence spending patterns.1 Published Consumer Expenditure Survey (CE) data tables have shown average expenditures, income, and other consumer unit (CU) characteristics classified by education of the reference person. With the release of calendar year 2012 CE data on September 10, 2013, the education of reference person classification was replaced by the highest education level of any member in the consumer unit.2

The major reason for this change is that the highest level of education attained by any household member more accurately reflects income and spending patterns than does the education level of the reference person only.3 For example, data from the Census Bureau show that the proportion of married couples where the wife is the more educated spouse increased during the 1996—2010 period.4 This means that the education level in families where the husband is designated as the reference person could be understated.

Table 1 shows selected characteristics, mean annual expenditures, and expenditure shares for consumer units classified by the highest level of education of any CU member, which is the new breakdown. Table 2 presents the same data classified by education of the reference person, which is the old breakdown. (Both tables show data for 2012.)

Table 1. Highest education level of any member: Annual expenditure means and shares, 2012
ItemAll consumer unitsLess than college graduateCollege graduate
TotalLess than high school graduateHigh school graduateHigh school graduate with some collegeAssociate's degreeTotalBachelor's degreeMaster's, professional, doctoral degree

Number of consumer units (in thousands)

124,41676,78910,57126,60125,79313,82547,62628,06919,557

Percentage of consumer units

100.061.78.521.420.711.138.322.615.7

Consumer unit characteristic:

 

Pretax income

$65,596$44,603$25,159$39,357$48,224$62,809$99,444$85,802$119,023

Age (reference person)

50.050.756.253.047.548.348.947.451.1

Average number in consumer unit:

 

Persons

2.52.42.22.42.42.72.62.52.7

Children under 18

0.60.60.70.60.60.70.60.60.6

Persons 65 and older

0.30.40.50.40.30.30.30.30.4

Earners

1.31.20.71.01.21.51.51.51.5

Vehicles

1.91.71.11.71.82.22.12.12.2

Percent distribution:

 

Sex (reference person):

 

Male

474443454541515052

Female

535657555559495048

Race (reference person):

 

Black or African-American

131516151614898

White, Asian, and all other races

878584858486929192

Hispanic or Latino origin (reference person):

 

Hispanic or Latino

131635141212785

Not Hispanic or Latino

878465868888939295

Housing tenure:

 

Homeowner

645847595668747179

With mortgage

393116283345514954

Without mortgage

262732322322232225

Renter

364253414432262921

At least one vehicle owned or leased

888470848693939394

Total annual expenditures:

$51,442$39,107$24,582$34,786$43,041$50,836$71,151$63,135$82,606

Food

 

Mean

6,5995,4093,9134,9445,7496,6588,4357,9289,143

Share

12.813.815.914.213.413.111.912.611.1

Food at home

 

Mean

3,9213,4422,8623,2633,5423,9804,6544,3805,032

Share

7.68.811.69.48.27.86.56.96.1

Food away from home

 

Mean

2,6781,9671,0511,6822,2072,6783,7823,5494,111

Share

5.25.04.34.85.15.35.35.65.0

Housing

 

Mean

16,88713,1979,38812,14314,24116,14622,81520,23026,512

Share

32.833.738.234.933.131.832.132.032.1

Apparel and services

 

Mean

1,7361,3291,0421,0831,5401,5882,3662,2262,565

Share

3.43.44.23.13.63.13.33.53.1

Transportation

 

Mean

8,9987,3813,8356,2258,42210,34211,59410,81712,710

Share

17.518.915.617.919.620.316.317.115.4

Healthcare

 

Mean

3,5562,8852,0042,8232,9493,5604,6354,1965,263

Share

6.97.48.28.16.97.06.56.66.4

Entertainment

 

Mean

2,6051,9011,1341,6302,1182,5713,7253,2204,443

Share

5.14.94.64.74.95.15.25.15.4

Pensions and Social Security

 

Mean

5,2383,2201,3742,6453,5985,0328,4937,13910,436

Share

10.28.25.67.68.49.911.911.312.6

Other(1)

 

Mean

5822378418943293442449379088737911533

Share

11.39.77.79.510.39.712.811.714.0

Footnotes:
(1) Includes cash contributions, alcohol, tobacco, personal care products and services, reading, education, life and personal insurance, and miscellaneous expenses.
 

Source: U.S. Bureau of Labor Statistics, Consumer Expenditure Survey.
 

Table 2. Education of reference person: Annual expenditure means and shares, 2012
ItemAll consumer unitsLess than college graduateCollege graduate
TotalLess than high school graduateHigh school graduateHigh school graduate with some collegeAssociate's degreeTotalBachelor's degreeMaster's, professional, doctoral degree

Number of consumer units (in thousands)

124,41685,17816,24631,02225,62312,28739,23824,79814,440

Percentage of consumer units

100.068.513.124.920.69.931.519.911.6

Consumer unit characteristic:

 

Pretax income

$65,596$49,901$33,154$47,221$55,987$66,122$99,667$89,438$117,233

Age (reference person)

50.050.754.652.447.547.948.646.951.4

Average number in consumer unit:

 

Persons

2.52.52.72.52.42.62.42.42.5

Children under 18

0.60.60.80.60.60.70.60.60.6

Persons 65 and older

0.30.40.50.40.30.30.30.20.4

Earners

1.31.21.01.21.31.51.41.41.4

Vehicles

1.91.81.41.91.92.22.02.02.1

Percent distribution:

 

Sex (reference person):

 

Male

474444464540515152

Female

535656545560494948

Race (reference person):

 

Black or African-American

131515141415888

White, Asian, and all other races

878585868685929292

Hispanic or Latino origin (reference person):

 

Hispanic or Latino

131635121010674

Not Hispanic or Latino

878465889090949396

Housing tenure:

 

Homeowner

646051635968737078

With mortgage

393321333646504952

Without mortgage

262730312322232126

Renter

364049374132273022

At least one vehicle owned or leased

888575868793939293

Total annual expenditures:

$51,442$41,983$31,194$39,989$46,118$52,414$71,926$66,420$81,363

Food

 

Mean

6,5995,7994,8135,6046,0406,9708,3147,9448,943

Share

12.813.815.414.013.113.311.612.011.0

Food at home

 

Mean

3,9213,6413,4923,5853,5554,1504,5174,3514,799

Share

7.68.711.29.07.77.96.36.65.9

Food away from home

 

Mean

2,6782,1571,3202,0202,4852,8203,7973,5934,144

Share

5.25.14.25.15.45.45.35.45.1

Housing

 

Mean

16,88713,91311,19713,32714,99216,72123,33921,59826,322

Share

32.833.135.933.332.531.932.432.532.4

Apparel and services

 

Mean

1,7361,4411,3271,2721,6031,6532,3692,2942,498

Share

3.403.404.403.203.503.103.403.403.20

Transportation

 

Mean

8,9987,9025,5147,6958,7219,86311,37411,05011,933

Share

17.518.817.719.218.918.815.816.614.7

Healthcare

 

Mean

3,5563,0722,2183,1193,2893,6304,6084,3405,066.590

Share

6.97.37.17.87.16.96.46.56.2

Entertainment

 

Mean

2,6052,0591,3931,8122,3672,8833,7873,4344,392

Share

5.14.94.54.55.15.55.35.25.4

Pensions and Social Security

 

Mean

5,2383,6432,0853,4014,1375,2858,7017,58610,615

Share

10.28.76.78.59.010.112.111.413.0

Other(1)

 

Mean

5,8224,1552,6473,7584,9705,4109,4348,17511,593

Share

11.39.98.59.410.810.313.112.314.2

Footnotes:
(1) Includes cash contributions, alcohol, tobacco, personal care products and services, reading, education, life and personal insurance, and miscellaneous expenses. Source: U.S. Bureau of Labor Statistics, Consumer Expenditure Survey.
 

Category and pretax income differences

Classification by highest level of education shifted almost 7 percent of the CUs in the sample from the less-than-college-graduate section to the college-graduate section, resulting in 61.7 percent of CUs categorized as less than college graduate and 38.3 percent categorized as college graduate. The greatest drop was in the less than high school graduate from 13.1 percent to 8.5 percent, a reduction of 5.7 million households, and the high school graduate categories from 24.9 percent to 21.4 percent, a reduction of 4.4 million households. The proportion of CUs falling into the “high school graduate with some college” and the “associate’s degree” categories was similar in both tables. Using the new breakdown increased the proportion of CUs falling into the bachelor’s degree category (22.6 percent, compared with 19.9 percent in the old breakdown) and master’s, professional, doctoral degree category (15.7 percent, compared with 11.6 percent in the old breakdown). 

Tables 1 and 2 show that average pretax income increased with the level of education. With one exception, the average dollar amount of pretax income was higher when using the old breakdown. In the new breakdown, pretax income for the master’s, professional, doctoral degree group was $119,023—compared with 117,233 for the same group classified by the reference person’s education. 

Annual expenditures

In tables 1 and 2, total annual expenditures also increased with the level of education and associated income. Except for the master’s, professional, doctoral degree level, mean annual expenditures at every level of education were less when classified by highest level of education of any member (the new breakdown), compared with those classified by the reference person’s education (the old breakdown).

Food, housing, and transportation

The dollar amount spent on each of the three categories increased with level of education and associated level of income. The relative share of these amounts differed by level of education. For example, food spending accounted for a greater share of the household budget at lower education levels. One exception is found in the old breakdown where households at the high-school-graduate-with-some-college level spent 13.1 percent of the budget on food, compared with 13.3 percent of the budget for households at the associate’s degree level. The reason is that food at home—a component of food spending—accounted for a greater share of the budget of households at the associate’s level 7.9 percent, compared with 7.7 percent for the high-school-graduate-with-some-college category). 

Housing expense patterns varied somewhat between the new and the old breakdowns, most likely because of the differences in housing tenure. When classified by highest education level, the homeownership rate ranged from 47 percent for the less-than-high-school-graduate category to 79 percent for the master’s, professional, doctoral degree category. When classified by reference person’s education, these proportions were 51 and 78 percent, respectively. These differences may account for the greater differential in dollar outlays between the less-than-high-school-graduate category and the master’s, professional, doctoral degree category in the new breakdown ($9,388 compared with $26,512). In the old breakdown this differential was $11,197 compared with $26,322). The budget shares accounted for by housing narrowly declined as education and associated income increased in both breakdowns. One factor could be the higher outlays for mortgage interest, property taxes, and repair and maintenance by homeowners. 

Transportation expense patterns also varied between the old and the new breakdowns. In both breakdowns, the dollar amount accounted for by transportation expenses increased with level of education and associated level of income. The share of the household budget accounted for by transportation varied by breakdown. In the new breakdown, transportation’s share of the household budget increased from 15.6 percent for those with less than a high school education to 20.3 percent for those with an associate’s degree and then declined from 17.1 percent for those with a bachelor’s degree to 15.4 percent for those at the master’s, professional, doctoral degree level. In the old breakdown, transportation expenses claimed 17.7 percent of the budget of households in the less-than-high-school-graduate category, increasing to 19 percent for each of the other three categories in the less-than-college-graduate section before declining to 16.6 percent for those in the bachelor’s degree category and 14.7 percent for those in the master’s, professional, and doctoral degree category. In both breakdowns, households in the less-than-high-school-graduate category were less apt to own a vehicle than households in the remaining categories; 70 percent in the new breakdown and 75 percent in the old breakdown did not own a car. As a comparison, almost all (94 percent in the new breakdown and 93 percent in the old breakdown) households in the master’s, professional, and doctoral degree categories owned at least one vehicle. In both breakdowns, lower vehicle ownership among households with less than a high school education would reduce the amount spent on vehicle purchases, insurance, and repairs, lowering transportation costs relative to households in the remaining categories.

Other spending

In both breakdowns, the dollar amount spent on healthcare increased with level of education and associated income. Households at higher levels of education and associated income spent a lower share of the budget on healthcare than those at lower levels, regardless of breakdown. 

Entertainment spending followed similar patterns in both the old and the new breakdowns, increasing in dollar amount with level of education and associated income. In both breakdowns, entertainment spending narrowly ranged from about 4.5 percent of the household budget for the less-than-high-school-graduate category to 5.4 percent of spending for households in the master's, professional, doctoral degree category.

Outlays on pensions and Social Security differed in the old and new breakdowns. In both breakdowns, the average annual outlay and share of the household budget increased with level of education and associated level of income. In the new breakdown, however, the average annual outlay and share of the household budget were below comparable categories in the old breakdown, particularly at the lowest levels of education and associated levels of income. For example, households classified as less than high school graduate averaged $1,374 (5.6 percent of total spending) for pensions and Social Security in the new breakdown, compared with $2,085 (6.7 percent) in the old breakdown. Households in the high school graduate category averaged $2,645 (7.6 percent of total spending) in the new breakdown, compared with $3,401 (8.5 percent of total spending) in the old breakdown.  One reason could be that there were fewer earners in the new breakdown compared with the old. For example, households classified as less than high school averaged 0.7 earners in the new breakdown compared with 1.0 earner in the old breakdown; for households classified as high school graduate these averages were 1.0 and 1.2 earners, respectively.  Another reason could be the lower pretax income, earlier mentioned, among household groups in the new breakdown in relation to comparable households in the old breakdown.  Because saving usually increases with income, households in the new breakdown might have reduced voluntary pension contributions to 401 K plans or IRAs, in order to free up additional funds for consumption.

Income sources

For households at the lowest levels of education, sources of pretax income differed somewhat between the old and the new breakdowns. These differences might explain some of the variations in spending patterns mentioned earlier. One difference is that in the new breakdown (table 3), wages and salaries accounted for a lower proportion of pretax income for those households with less than a college degree than in the old breakdown. (See table 4.) For example, in the new breakdown, wages and salaries averaged 56.8 percent of pretax income for households classified as less than a high school graduate and 68.7 percent of pretax income for households classified as high school graduate, compared with 67.4 percent and 72.8 percent, respectively, in the old breakdown. Another reason is that in the new breakdown, Social Security, private, and government retirement benefits accounted for a higher proportion of pretax income for households with less than a college education than in the old breakdown. For example, in the new breakdown, Social Security, private, and government retirement benefits accounted for 28.8 percent of pretax income for households classified as less than a high school graduate and 21.3 percent of pretax income for households classified as high school graduate, compared with 21.3 percent and 18.1 percent, respectively, in the old breakdown. It was mentioned earlier that in the new breakdown, the average annual outlay and share of the household budget for pensions and Social Security for households at the lowest levels of education were below those found in the old breakdown. The variation in income sources may explain some of the variation in outlays for pensions and Social Security.

Table 3. Highest education level of any member: Annual pretax income means and shares, 2012
ItemAll consumer unitsLess than college graduateCollege graduate
TotalLess than
high school graduate
High school graduateHigh school graduate
with some college
Associate's degreeTotalBachelor's degreeMaster's, professional,
doctoral degree

Number of consumer units (in thousands)

124,41676,78910,57126,60125,79313,82547,62628,06919,557

Percentage of consumer units

100.061.78.521.420.711.138.322.615.7

Pretax income:

$65,596$44,603$25,159$39,357$48,224$62,809$99,444$85,802$119,023

Wages and salaries

 

Mean

51,73032,84514,29427,02136,47751,46082,18071,13598,031

Share

78.973.656.868.775.681.982.682.982.4

Self-employment income

 

Mean

2,9171,8741,3271,7801,9502,3294,5984,2115,154

Share

4.44.25.34.54.03.74.64.94.3

Social Security, private and government retirement

 

Mean

8,0217,4207,2348,3886,8886,6928,9917,24211,501

Share

12.216.628.821.314.310.79.08.49.7

Interest, dividends, rental income, other property income

 

Mean

1,3586913046079307052,4321,8613,253

Share

2.11.61.21.51.91.12.42.22.7

Unemployment and workers' compensation, veterans' benefits

 

Mean

428464357353544611370408317

Share

0.71.01.40.91.11.00.40.50.3

Public assistance, supplemental security income, food stamps

 

Mean

5347191181765571552235279172

Share

0.81.64.71.91.20.90.20.30.1

Regular contributions for support

 

Mean

380366206277529356401471302

Share

0.60.80.80.71.10.60.40.50.3

Other income

 

Mean

229224257167334104235195293

Share

0.30.51.00.40.70.20.20.20.2

Source: U.S. Bureau of Labor Statistics, Consumer Expenditure Survey.
 

Table 4. Education of reference person: Annual pretax income means and shares, 2012
ItemAll consumer unitsLess than college graduateCollege graduate
TotalLess than
high school graduate
High school graduateHigh school graduate
with some college
Associate's degreeTotalBachelor's degreeMaster's, professional,
doctoral degree

Number of consumer units (in thousands)

124,41685,17816,24631,02225,62312,28739,23824,79814,440

Percentage of consumer units

100.068.513.124.920.69.931.519.911.6

Pretax income:

$65,596$49,901$33,154$47,221$55,987$66,122$99,667$89,438$117,233

Wages and salaries

 

Mean

51,73037,70722,35934,36843,40854,54382,17174,55495,254

Share

78.975.667.472.877.582.582.483.481.3

Self-employment income

 

Mean

2,9172,0601,5122,0282,1982,5744,7784,7364,850

Share

4.44.14.64.33.93.94.85.34.1

Social Security, private and government retirement

 

Mean

8,0217,5927,0548,5707,1826,6918,9537,05312,215

Share

12.215.221.318.112.810.19.07.910.4

Interest, dividends, rental income, other property income

 

Mean

1,3588292897201,3617102,5061,8433,643

Share

2.11.70.91.52.41.12.52.13.1

Unemployment and workers' compensation, veterans' benefits

 

Mean

428462376406506626355356354

Share

0.70.91.10.90.90.90.40.40.3

Public assistance, supplemental security income, food stamps

 

Mean

5346881,165678498478198207184

Share

0.81.43.51.40.90.70.20.20.2

Regular contributions for support

 

Mean

380348189276516392447496362

Share

0.60.70.60.60.90.60.40.60.3

Other income

 

Mean

229215211174318108259193371

Share

0.30.40.60.40.60.20.30.20.3
 

Source: U.S. Bureau of Labor Statistics, Consumer Expenditure Survey.
 

Conclusions and implications

With the release of calendar year 2012 CE data, classification of household expenditures by the education of the reference person breakdown was replaced by classification by the highest level of education attained by any consumer unit (CU) member. The rationale for the change was that the highest level of education attained by any household member more accurately reflects income and spending patterns than does the education level of the reference person only.

A comparison of expenditures was made using the old and new breakdowns. The new breakdown reduced the proportion of households in the less-than-college-graduate section, with the greatest reduction in the less-than-high–school-and-high-school-graduate classifications. With the exception of the master’s, professional, doctoral degree classification in the college-graduate section, pretax income and annual expenditures were lower in the new breakdown.

In both breakdowns, pretax income and total annual expenditures increased with level of education. Spending on most items followed similar patterns in both breakdowns. One exception was outlay for pensions and Social Security in the less-than-high-school and high-school-graduate classification which was lower in dollar amount and share of the household budget in the new breakdown, compared with the old.

Wages and salaries, the largest component of most households’ pretax income, is a reflection of the value of the productivity the marketplace assigns to a worker. Michael’s neutrality model also assumes that education increases efficiency in all nonmarket activities and that its influence is neutral (has the same effect on each activity).5 This greater efficiency increases a household’s real income at any given level of money income, thus increasing a household’s ability to purchase goods and services. Michael hypothesized that the increase in real income associated with increased education would result in proportionately higher outlays on luxury goods—but proportionately lower outlays on necessities.

Existing research using CE data provides partial support for Michael’s neutrality model, indicating that the relationship between education and household expenditures is significant, but complex. Michael, using data from the 1960—1961 CE found that, controlling for the effects of income and other factors, education was associated with a proportionately greater outlays on food away from home, household operations, leisure (recreation), and education, which were classified as luxuries. Of the items classified as necessities, education was associated with a proportionately lower outlay on food at home, tobacco, and personal care.6 Abdel-Ghany and Foster, using data from the 1972—1973 CE, also found that, controlling for the effects of income and other factors, education was associated with a proportionately greater outlay on household operations, recreation (leisure), and education and proportionately lower outlay on tobacco and utilities.7 Because of the nature of the data examined in this article, whether education has an influence on spending separate from its association with household income cannot be determined.

This Beyond the Numbers summary was prepared by Ann C. Foster, an economist in the Office of Prices and Living Conditions, Bureau of Labor Statistics. Email: foster.ann@bls.gov; telephone: (202) 691-5174.

Information in this article will be made available to sensory-impaired individuals upon request. Voice phone: (202) 691-5200. Federal Relay Service: 1 (800) 877-8339. This article is in the public domain and may be reproduced without permission.

Suggested citation:

Ann C. Foster, “New education classification better reflects income and spending patterns in the Consumer Expenditure Survey ,” Beyond the Numbers: Prices & Spending, vol. 3, no. 1 (U.S. Bureau of Labor Statistics, January 2014), http://www.bls.gov/opub/btn/volume-3/education-classification-and-income-and-spending-patterns.htm

1For example, in 2012, median weekly earnings for full-time workers age 25 and over were $471 for those without a high school diploma, $652 for high school graduates (no college), and $1,735 for those with a professional degree, such as medicine or law. For more information, see “Earnings and unemployment rates by educational attainment,” (Bureau of Labor Statistics, May 23, 2013), www.bls.gov/emp/ep_chart_001.htm.     

Data from the Census Bureau’s 2011 American Community Survey were used to estimate work-life earnings—the expected earnings over a 40-year time period for the population age 25-64 who maintain full-time, year-round employment for the entire period. Median earnings were $936,000 for workers with an eighth grade education or less, compared with $1,371,000 for those with a high school degree, but no college classes and $4,159,000 for those with a professional degree. For more information, see Tiffany Julian, “Work-Life Earnings by Field of Degree and Occupation for People with a Bachelor’s Degree: 2011” (U.S. Census Bureau, October 2012), www.census.gov/prod/2012pubs/acsbr11-04.pdf.

2The reference person is the first household member mentioned by the respondent when asked to “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. In two-parent families, the reference person can be male or female. In one-parent families, the gender of the reference person is usually that of the sole parent. 

3In the Consumer Expenditure Survey (CE), the consumer unit is the entity on which expenditure reports are collected. Consumer units include families, single persons living alone or sharing a household with others but who are financially independent, or two or more persons living together who share expenses. While “consumer unit” is the proper technical term for the purposes of the CE, “household” or “family” often is used interchangeably for convenience. This article will use “household” instead of consumer unit. For more information, see the BLS Handbook of Methods, chapter 16, “Consumer Expenditures and Income.”  

4The proportion of married couples where the husband has more education than the wife decreased, while the proportion where the husband and wife had the same level of education remained about the same. For more information, see Rebecca Chenevert, “Changing Levels of Spousal Education and Labor Force Supply,” Working Paper 2012-22 (U.S. Census Bureau, 2012) www.census.gov/people/laborforce/publications/Chenevert_MEA2012.pdf.

5For more information, see Robert T. Michael, “The Effect of Education on Efficiency in Consumption,” the National Bureau of Economic Research Occasional Paper, no. 116 (New York: Columbia University, 1972), at http://papers.nber.org/books/mich72-1

6For more information, see Robert T. Michael, “The Effect of Education on Efficiency in Consumption,” the National Bureau of Economic Research Occasional Paper, no. 116 (New York: Columbia University, 1972), at http://papers.nber.org/books/mich72-1.

7For more information, see Mohamed Abdel-Ghany and Ann C. Foster, “Impact of Income and Wife’s Education on Family Consumption Expenditures,”  Journal of Consumer Studies and Home Economics, vol. 6, no. 1, March 1982, pp. 21–28.

Publish Date: Monday, January 27, 2014