Analysis of Trends in Household RÙt RÙt RÙt BOOKMOBIÿÿÿ| % v )v 9v Iv Yv iv yv ‰v ™v ©v ¹v Év Ùv év ùv v v )v 9v Iv Yv iv yv ‰v ™v ©v ¹v Év Ùv év ùv v v )v 9v Iv Yv Mx $ MOBI ýém…+*
Analysis of Trends in Household Composition Using American Housing Survey Data
Analysis of Trends in Household Composition Using American Housing Survey Data
Disclaimer
The contents of this report are the views of the contractor and do not necessarily reflect the views or policies of the U.S. Department of Housing and Urban Development or the U.S. Government.
11
I am pleased to release this report commissioned by the U.S. Department of Housing and Urban Development, Analysis of Trends in Household Composition Using American Housing Survey Data. This report uses the longitudinal features of the American Housing Survey (AHS) to examine trends in household formation and the phenomenon of “doubled-up” households in the aftermath of the housing bubble and recession of 2007 through 2009. The stresses of that period, arising as they did from imbalances in the housing sector, reduced the rate of household formation among young individuals and families who would normally have moved out of their parents’ and shared units to live on their own. Moreover, the double stresses of the mortgage crisis and job loss induced many who had been living on their own to move in with relatives and other people to economize on housing expense.
The AHS is uniquely qualified to examine this phenomenon, because it is a nationally representative longitudinal survey that returns to the same housing units every 2 years. Thus, analysts using AHS data not only can produce aggregate estimates of household formation, they can examine changes in the composition of individual households. This study shows a steady decline in new household formations from 2003 through 2009. The primary contributors to new doubled-up households were adult children moving in with their parents. A notable increase occurred in the number of households consisting of more than one family living in the same housing unit, however.
The Department is interested in doubled-up households because they represent a disguised form of stress in the nation’s housing markets. Although doubled-up households may not exhibit the high housing cost burdens that are normally associated with “worst case housing needs,” and although their members are not homeless, their housing choices suggest that they are at best precariously housed and in danger of falling into the more easily observed categories of housing distress. This study provides information to policymakers and scholars to help them understand these undercurrents before they become manifest.
Kurt T. Usowski
Deputy Assistant Secretary
for Policy Development
iii
Executive Summary ................................................................................................................. vi
1. Overview .................................................................................................................................. 1
1.1. Background ...................... ................................................................................................... 1
1.2. Definitions........................................................................................................................... 1
1.3. Organization of the Report.................................................................................................. 2
2. Trends in Household Composition From AHS National Summary Data ................ 5
3. Comparison of AHS and CPS Analysis of Doubled-Up Households .................... 10
3.1. Methodological Differences Between the AHS and CPS................................................. 10
3.2. Comparison of the Counts of Doubled-Up Households, AHS Versus CPS ..................... 11
3.3. Comparison of the Characteristics of Doubled-Up Households, AHS Versus CPS......... 11
4. Characteristics of Doubled-Up Households and Their Housing Units ................. 15
4.1. Housing Characteristics of Units With Doubled-Up Households .................................... 15
4.2. Multivariate Analysis of Characteristics of Doubled-Up Households and Their
Housing Units ................................................................................................................... 16
5. Adult Children As the Primary Contributors to Doubled-Up Households ........... 22
5.1. Types of Other Household Members: 2003 to 2009......................................................... 22
5.2. Characteristics of Adult Children Living at Home ........................................................... 23
6. More-Than-One-Family Households As a Secondary Contributor to
Doubled-Up Households ................................................................................................... 25
6.1. Housing Units With Subfamilies, 2003 to 2009 ............................................................... 25
6.2. Characteristics of Householders in More-Than-One-Family Households........................ 27
7. Changes in Doubled-Up Status at the Household Level, 2005 to 2009 ................. 33
8. Doubled-Up Households As Reported by the 2011 AHS .......................................... 37
9. Concluding Observations ................................................................................................. 40
9.1. The Recession and Doubled-Up Households ................................................................... 40
9.2. Paths for Future Research ................................................................................................. 40
10. References .......................................................................................................................... 41
iv
Table 1A. AHS National Summary Data on Household Composition, All Households:
2003 to 2009 ................................................................................................................. 5
Table 1B. AHS National Summary Data on Household Composition, All Owner Households:
2003 to 2009 ................................................................................................................. 6
Table 1C. AHS National Summary Data on Household Composition, All Renter Households:
2003 to 2009 ................................................................................................................. 7
Table 2. Changes in Measures of Household Crowding................................................................ 9
Table 3. Comparison of Estimates of Doubled-Up Households Based on AHS and
CPS Data ................................ ........................................................................................ 11
Table 4. Comparison of Characteristics of Doubled-Up Households Based on AHS and
CPS Data ........................................................................................................................ 12
Table 5. Housing Characteristics of Doubled-Up Households in 2009 ....................................... 15
Table 6. Multivariate Analysis of the Characteristics of Doubled-Up Households..................... 17
Table 7. Units With Doubled-Up Households by Type of Other Household Member,
2003 to 2009 .................................................................................................................. 22
Table 8. Characteristics of Adult Children Living at Home ........................................................ 24
Table 9. Housing Units With Subfamilies by Type of Subfamily, 2003 to 2009 ........................ 26
Table 10. More-Than-One-Family Households, 2003 to 2009 .................................................... 27
Table 11A. Characteristics of Reference Persons in 2003 ........................................................... 29
Table 11B. Characteristics of Reference Persons in 2009 ........................................................... 30
Table 12. Type of Other Household Member in Units That Changed Doubled-Up Status
From 2005 to 2009 ....................................................................................................... 34
Table 13. Type of Other Household Member in Units That Were Doubled Up in Both 2005
and 2009 ....................................................................................................................... 35
Table 14. AHS National Summary Data on Household Composition for Occupied Housing
Units, 2003 to 2011 ...................................................................................................... 37
v
This report uses data from the American Housing Survey (AHS) to analyze changes in household composition from the 2003 survey through the 2009 survey (hereafter, from 2003 to 2009), particularly those changes that reflect an increase in doubled-up households. The term “other household member” is applied to people who are not the householder nor the householder’s spouse, partner, or minor children. Other household members include adult children, parents, grandchildren, siblings, and nonrelatives. Households containing other household members are labeled “doubled-up households.” Census Bureau studies suggesting a link between the recession and the increase in doubled-up households motivated this research. The data reported here
provide several indications that the recession and the preceding financial crisis had an effect on
doubled-up households.
Trends in Household Composition From AHS National Summary Data
Several findings stand out in the AHS national summary data (see Tables 1A through 1C).
• The growth rate of households declined during the study period; 3.0 million households were added from 2003 to 2005, 1.8 million were added from 2005 to 2007, and only 1.1 million were added from 2007 to 2009.
• Households with single, adult offspring ages 18 through 29 years increased substantially from 2003 to 2009, registering a 12.9-percent change overall and a 24.6-percent change among renter households.
• The number of households containing one or more unrelated subfamilies more than tripled during the period, growing by more than 400,000 households.
Table 14 uses data from the 2011 AHS to update these trends. The latest AHS data do not provide reliabl e comparisons with previous AHS data because the Census Bureau changed the benchmark used to estimate counts in 2011 (from the 2000 census to the 2010 census). With this caveat in mind, the 2011 AHS shows that (1) 3.0 million new households were added from 2009 to 2011, and (2) the count of households with unrelated subfamilies declined by 80,000 from
2009 to 2011.
The AHS national summary data indicate that the increase in doubled-up households has not caused more crowding in housing units. During the 6-year period from 2003 to 2009, only minimal changes occurred in measures of crowding, and, in general, these changes show less overcrowding.
Comparison of AHS and Current Population Survey Analyses of Doubled-Up Households
AHS data indicated trends similar to those reported in Census Bureau studies using data from the Current Population Survey (CPS). Even without making adjustments for certain definitional and methodological differences, the AHS counts of doubled-up households are close to those in the Census reports. If children 18 years old and older are considered to be adults living at home, the AHS counts 26.0 million doubled-up households in 2007 and 26.5 million in 2009, compared
vi
with Census Bureau counts of 26.1 million in 2008 and 28.4 million in 2010. If children 21 years old and older are considered to be adults living at home, the AHS counts 20.5 million in 2007
and 21.0 million in 2009, compared with 19.7 million in 2007 and 21.8 million in 2011, as reported in the Census Bureau’s income and poverty report for 2010 (see Table 3). Further
analysis of the characteristics of doubled-up households shows that the AHS data describe these households as having the same demographic characteristics as reported in a Census Bureau study
based on CPS data. The similarity in counts and in the descriptions of doubled-up households demonstrates that the AHS can be safely used to study changes in household composition. A
planned supplement on doubled-up households will enhance the information available in the
AHS.
Characteristics of Doubled-Up Households and Their Housing Units
The richness of information in the AHS microdata provides additional insights into the characteristics of doubled-up households (see Tables 4 through 6).
• Households with older householders are more likely to be doubled up.
• Households with householders who are Black, non-Hispanic; other race, non-Hispanic; or
Hispanic are more likely than households with White householders to be doubled up.
• Households with foreign-born householders are more likely to be doubled up than those with householders born in the United States or associated areas, such as Puerto Rico.
• The percentage of doubled-up households that live in single-family homes is substantially higher than the percentage of nondoubled households that live in single-family homes.
• Doubled-up households are more concentrated in central cities and in the urban parts of suburban areas than their nondoubled counterparts. The quality of housing units, as measured by the AHS adequacy standard, does not differ between doubled-up and nondoubled households.
Adult Children As the Primary Contributors to Doubled-Up Households
Children ages 21 and older are the most common type of other household member;
approximately one-half of the doubled-up households in all four surveys (2003, 2005, 2007, and
2009) included a child ages 21 and older. The percentage of doubled-up households with a child or children ages 21 and older increased by slightly more than 3 percentage points during this period, which includes the financial crisis and subsequent recession (see Table 7).
• In all four surveys, approximately 50 percent of a ll adult children in doubled-up households were ages 21 through 25. The percentages of those ages 26 through 30 and those ages 50 and older were highest in 2009. The substantial increase in doubling up among those ages 26 through 30 in 2009 is consistent with young adults having difficulties finding employment. The percentage of adult children ages 50 and older in doubled-up households increased steadily across the four surveys and may reflect both a worsening of economic conditions and an increase in the number of adult children moving in with aged parents to take care of them (see Table 8).
vii
• Worsening economic conditions are clearly evident in the limited information in the AHS
on the work experience of adult children living with a parent or parents. The percentage
of adult children working in the past week fell from about 60 percent in the first three surveys to approximately 57 percent in the 2009 survey. The percentages of wage, salary, and self-employment income steadily declined across the surveys.
More-Than-One-Family Households As a Secondary Contributor to Doubled-Up
Households
The most dramatic demographic change in household composition from 2003 to 2009 was the more than tripling of the number of households containing unrelated subfamilies.
• The number of households with related subfamilies grew from 2,656,000 to 2,846,000, and the number of households with unrelated subfamilies grew from 199,000 to 622,000. The increase in the number of unrelated subfamilies and the shift in the composition of this group lend credence to the proposition that economic hard times have led to doubled- up households; in this case, doubled-up households among unrelated households.
• The key demographic changes among households with unrelated subfamilies were an increased percentage of non-Hispanic subfamilies; a higher percentage of younger, unrelated subfamilies, particularly among those ages 26 through 35; an increased percentage of American citizens among the reference persons (that is, that person within each family unit chosen by the Census Bureau to distinguish households on the basis of race, age, employment history, and so on) in unrelated subfamilies; and a better educational profile for those reference persons.
Changes in Doubled-Up Status at the Household Level, 2005 to 2009
By focusing on households that occupied the same housing unit throughout the 2005-to-2009 period, we were able to observe how households became doubled up, stayed doubled up, or became undoubled (see Tables 8 and 9). A shift in the character of doubled-up units occurred during the period that included the financial crisis and recession.
• Children remaining at home dominated the evolution of doubled-up households during the recession period. Of the households that were doubled up in 2005 and became undoubled by 2009, 58.6 percent had a child or children ages 21 and older move out, whereas among the households that became doubled up between 2005 and 2009, 79.1 percent contained a child age 21 and older in 2009.
• Units that entered the period doubled up with adult children or grandchildren were more likely than all other doubled-up households to stay doubled up.
viii
Concluding Observations
The Census Bureau studies suggesting a link between the recession and the increase in doubled- up households were an important motivation for this research. The data cited in this report provide several indications that the recession and the preceding financial crisis had an effect on doubled-up households. The most notable indications were the following:
• The steady decline in new household formations from 2003 to 2009.
• The increased number of adult children living at home.
• The fact that, by 2009, adult children living at hom e were more likely to be recently out of school and were less likely to be employed.
• The increased number of unrelated families living in one household.
• The characteristics of these unrelated families changing in such a way as to suggest that economic hardship became a more important cause of unrelated families living together than immigration.
The report also contains suggestions for further research on changes in household composition.
ix
1.1. Background
In October 2009, Econometrica, Inc., entered into a contract with the U.S. Department of
Housing and Urban Development (HUD) to support the American Housing Survey (AHS). Task
13 of that contract required the Econometrica team to analyze the information collected in the national AHS on changes in household composition during the most recent recession. HUD asked Econometrica to track the proportion of new household formations as the economy has
moved through the current business cycle and to track the incidence of subfamilies and crowding
among recent movers in different survey years.
In September 2011, the Census Bureau published its annual P-60 report on income and poverty in the United States (DeNavas, Proctor, and Smith, 2011).1 This report called attention to the “doubled-up household” effect that the recent recession has had on household composition.
Doubled-up households are defined as households that include at least one “additional” adult: a person 18 or older who is not enrolled in school and is not the householder, spouse or cohabitating partner of the householder. In spring 2007, prior to the recession, doubled-up households totaled 19.7 million. By spring 2011, the number of doubled-up households had increased by 2.0 million to 21.8 million and the percent rose by 1.3 percentage points from 17.0 percent to 18.3 percent (Census Bureau, 2011c).
The analysis in the published report followed a more detailed Census Bureau study of the changes in household composition from 2008 to 2010 by Laryssa Mykyta and Suzanne Macartney (2011).
The Census Bureau report and previous paper focused mainly on how the recession slowed household formation by discouraging young adults from moving out and forming their own households and by making it necessary for single adults and families to move back to live with parents or to join together to form more-than-one-family households. Mykyta and Macartney pointed out that doubled-up households can include parents moving in with children. An extensive article in the weekend edition of the Wall Street Journal (Browning, 2012) highlighted this latter phenomenon.
This report, prepared under Econometrica’s 2009 contract with HUD, builds on the Census
Bureau’s previous work, using information from the AHS.
1.2. Definitions
This report uses the following terms, for which definitions are important.
• Household. A household consists of all people who occupy a particular housing unit as their usual residence. Each occupied housing unit contains one and only one household.
1 See page 21. The P60 report for 2011 did not discuss doubling up.
2 American Housing Survey for the United States: 2009 contains the following statement on page A-21 following the
1
By definition, the number of households is equal to the number of occupied housing units.
• Householder. Householder is a term used only in survey work. A household may contain several people with different demographic or economic characteristics. In analyzing survey data, it is useful to have one person whose characteristics can be referenced to distinguish households on the basis of race, age, employment history, and so on. In gene ral, the choice of householder is arbitrary. The householder characteristics are used as an imperfect proxy for the characteristics of the other members of the household. In the AHS, the householder is the first household member listed on the questionnaire who is an owner or renter of the sample unit and is 18 years old or older.
• Other household member. “Other household member” includes any person who is not the householder, the householder’s spouse or partner, or a child of the householder younger than age 21. Parents, grandchildren, and unrelated children are included in “other household members.” Sometimes the report includes children ages 18 and older (instead of ages 21 and older) in “other household members.” When this alternative definition is used, the report clearly indicates the difference.
• Doubled-up household. A household with an “other household member” is a doubled-up household. In most contexts, a child age 21 or older will count as an “other household member,” but sometimes the analysis will count a child age 18 or older as an “other household member.” When this alternative definition is used, the report clearly indicates the difference.
• Family. In this report, a family consists of: (1) a married couple with or without children; (2) an unmarried, widowed, divorced, or separated woman with a child or children; or (3) an unmarried, widowed, divorced, or separated man with a child or children.
• Primary family. If the householder is a member of the family, then the family is the primary family.
• Subfamily. If the householder is not a member of the family, then the family is a subfamily.
• Reference person. Reference person is another term used only in survey work. A family may contain several people with different demographic or economic characteristics. In analyzing survey data, it is useful to have one person whose characteristics can be referenced to distinguish families on the basis of race, age, employment history, and so on. The reference person’s characteristics are used as an imperfect proxy for the characteristics of the other members of the family or subfamily. For primary families, the reference person is always the householder. For subfamilies, the choice of reference person is arbitrary. Usually, the reference person of a subfamily is the first person from the subfamily age 21 or older whose information is recorded.
1.3. Organization of the Report
This report uses the AHS to investigate recent changes in household composition. In addition to this Overview section, the report consists of eight other sections.
2
• Section 2—Trends in Household Composition From AHS National Summary Data.
Consistent with HUD’s requirement of Econometrica to analyze household composition
changes, Section 2 presents household composition information from the national summary tables. Summary data are presented from four AHSs: 2003, 2005, 2007, and
2009. The AHS data track household composition from 2005 (before the recession and the financial crisis) to 2007 (when the financial crisis hit and the very beginning of the
recession) and to 2009 (when the recession came to a conclusion and the slow recovery began). After reading this section, the reader should be left with an understanding of how
household composition changed from the 2003 survey through the 2009 survey
(hereafter, from 2003 to 2009). Examination of changes in the number of persons per
household and the number of persons per room during this period indicates that the increase in doubled-up households did not result in more overcrowding.
• Section 3—Comparison of AHS and CPS Analysis of Doubled-Up Households.
Section 3 ev aluates the extent to which the AHS replicates trends reported by the Current
Population Survey (CPS). To accomplish this goal, AHS microdata are used to replicate,
to the extent possible, the trends reported by Mykyta and Macartney (2011) and in the Census Bureau’s income and poverty report for 2010. The third section also uses the microdata to replicate information on the characteristics of doubled-up households from Mykyta and Macartney’s analysis. On the basis of Table 3, the AHS reports similar trends in household composition to those based on CPS data.
• Section 4—Characteristics of Doubled-Up Households and Their Housing Units.
Section 4 presents a deeper analysis of the characteristics of doubled-up households and
their housing units. Tabular and multivariate analyses indicate that doubled-up households differ from other households in housing and unit. For example, households with foreign-born householders and those in single-family, detached units are more likely to be doubled up.
• Section 5—Adult Children As the Primary Contributors to Doubled-Up Households. Section 5 finds that adult children living at home represent most other household members in doubled-up households. Section 5 examines the characteristics of these adult children and finds that, although the characteristics changed only slightly from 2003 to 2009, the changes are consistent with the effect of the recession on the economic opportunities of young adults.
• Section 6—More-Than-One-Family Households As a Secondary Contributor to Doubled-Up Households. Section 6 examines the characteristics of a second important type of doubled-up households, those containing more than one family. The section describes these households in terms of the characteristics of the householders in the primary family and those of the reference person or people in the other family or families. The most important changes in more-than-one-family households were a tripling of the number of households with unrelated subfamilies and demographic shifts among the unrelated subfamilies consistent with economic hardship.
• Section 7– Changes in Doubled-Up Status at the Household Level, 2005 to 2009.
Section 7 analyzes the dynamics of household composition during the period of the
3
recession by studying an important subset of the AHS microdata, namely, those housing units with the same households in 2005 and 2009. Section 7 compares the composition of three groups of households: those that were doubled up in both 2005 and 2009, those that doubled up between 2005 and 2009, and those that undoubled between 2005 and 2009. Adult children living at home were the predominant cause of the differences among these groups.
• Section 8—Doubled-Up Households As Reported by the 2011 AHS. Section 8 looks at the national summary tables from the 2011 AHS, which was released after the analysis of
2003-to-2009 data in this report was complete, to see if any notable changes have occurred in household composition. Table 14 reports these numbers, with the important
caveat that the counts are not comparable because the Census Bureau changed the benchmark used to estimate counts.
• Section 9—Concluding Observations. Because this analysis was stimulated by the attention given to changes in household composition during the recession, Section 9 pulls together results reported in the previous sections that bear on the alleged effect of the recession on household composition. Section 9 also presents ideas for future research.
4
Tables 1A through 1C contain information on household composition drawn from Table 2-9 in the AHS national summary tables for 2003, 2005, 2007, and 2009. These data are used to study trends in household composition during this period.
Table 1A. AHS National Summary Data on Household Composition, All Households: 2003 to 2009
2003
2005
2007
2009
Households (thousands)
105,842
108,871
110,692
111,806
Single-person households
28,171
29,181
29,996
30,108
Two-or-more-person households
77,672
79,691
80,695
81,698
As a percent of all households
Single-person households
26.6
26.8
27.1
26.9
Two-or-more-person households
73.4
73.2
72.9
73.1
Persons other than spouse or children (thousands)
With other relatives
22,012
22,724
23,133
23,656
Single adult offspring 18 to 29
12,050
12,397
12,889
13,511
Single adult offspring 30 years of age or over
3,649
3,629
3,778
3,862
Households with three generations
3,049
3,335
3,202
3,301
Households with one subfamily
2,428
2,541
2,480
2,622
Subfamily householder age under 30
1,225
1,252
1,218
1,318
30 to 64
1,060
1,156
1,180
1,219
65 and over
143
133
82
85
Households with two or more subfamilies
113
100
138
111
Households with other types of relatives
7,244
7,474
7,174
7,165
With nonrelatives
9,564
9,979
10,265
10,717
Co-owners or co-renters
3,153
3,410
2,683
3,428
Lodgers
1,316
1,445
1,262
1,207
Unrelated children, under 18 years old
1,105
1,058
1,100
1,133
Other nonrelatives
4,808
4,914
6,179
6,049
One or more subfamilies
196
274
494
621
Two-person households,
none related to each other
4,956
5,204
5,396
5,403
Three-to-eight-person households, none related to each other
730
711
825
862
AHS = American Housing Survey.
5
Table 1B. AHS National Summary Data on Household Composition, All Owner Households: 2003 to 2009
2003
2005
2007
2009
Households (thousands)
72,238
74,931
75,647
76,428
Single-person households
15,455
16,205
16,686
16,777
Two-or-more-person households
56,784
58,725
58,961
59,651
As a percent of all households
Single-person households
21.4
21.6
22.1
22.0
Two-or-more-person households
78.6
78.4
77.9
78.0
Persons other than spouse or children (thousands)
With other relatives
16,501
17,172
17,170
17,454
Single adult offspring 18 to 29
9,619
9,911
10,179
10,481
Single adult offspring 30 years of age or over
2,957
2,960
3,046
3,148
Households with three generations
2,227
2,441
2,300
2,342
Households with one subfamily
1,699
1,813
1,697
1,837
Subfamily householder age under 30
762
843
778
858
30 to 64
804
846
847
906
65 and over
133
124
73
73
Households with two or more subfamilies
81
50
86
74
Households with other types of relatives
4,713
4,919
4,532
4,543
With nonrelatives
4,383
4,780
4,509
4,816
Co-owners or co-renters
898
1,024
806
1,000
Lodgers
545
725
488
533
Unrelated children, under 18 years old
683
663
627
647
Other nonrelatives
2,649
2,793
3,080
3,160
One or more subfamilies
111
146
265
352
Two-person households,
none related to each other
2,189
2,414
2,341
2,440
Three-to-eight-person households,
none related to each other
163
183
192
190
AHS = American Housing Survey.
6
Table 1C. AHS National Summary Data on Household Composition, All Renter Households: 2003 to 2009
2003
2005
2007
2009
Households (thousands)
33,604
33,940
35,045
35,378
Single-person households
12,716
12,976
13,310
13,331
Two-or-more-person households
20,888
20,966
21,735
22,047
As a percent of all households
Single-person households
37.8
38.2
38.0
37.7
Two-or-more-person households
62.2
61.8
62.0
62.3
Persons other than spouse or children (thousands)
With other relatives
5,510
5,553
5,963
6,202
Single adult offspring 18 to 29
2,432
2,486
2,710
3,030
Single adult offspring 30 years of age or over
692
669
732
714
Households with three generations
822
894
902
958
Households with one subfamily
729
728
783
784
Subfamily householder age under 30
463
409
441
459
30 to 64
256
309
333
314
65 and over<