The Impact of Parental Homeownership on Children's Outcomes during Early Adulthood

Abstract

Whether children benefit from being raised in a home owned by their parents has important policy implications and has been the topic of much scholarly debate. We match Panel Study of Income Dynamics data with census tract data to examine the impact of childhood experiences on adult outcomes for children followed over three decades. This allows us to document a wide range of characteristics.For children born between 1968 and 1974, we analyze data on their first 18 years and also various outcomes when they are between 25 and 31 in 1999. We control for a comprehensive set of observable parental characteristics and develop a method to control for unobservable child characteristics together with an instrumental variable for the remaining selection problems. Parental homeownership status and children's college education and home‐ownership status are closely related, although the former is generated partially by the greater residential stability associated with homeownership.

Full text

785
HOUSING POLICY DEBATE VOLUME 18 ISSUE 4
© 2007 METROPOLITAN INSTITUTE AT VIRGINIA TECH. ALL RIGHTS RESERVED.
The Impact of Parental
Homeownership on Children’s
Outcomes during Early Adulthood
George Galster
Wayne State University
Dave E. Marcotte and Marvin B. Mandell
University of Maryland at Baltimore County
Hal Wolman and Nancy Augustine
The George Washington University
Abstract
Whether children benef it from being raised in a home owned by their
parents has important policy implications and has been the topic of much
scholarly debate. We match Panel Study of Income Dynamics data with cen-
sus tract data to examine the impact of childhood experiences on adult out-
comes for children followed over three decades. This allows us to document
a wide range of characteristics.
For children born between 1968 and 1974, we analyze data on their f irst
18 years and also various outcomes when they are between 25 and 31 in
1999. We control for a comprehensive set of observable parental character-
istics and develop a method to control for unobservable child characteristics
together with an instrumental variable for the remaining selection problems.
Parental homeownership status and children’s college education and home-
ownership status are closely related, although the former is generated par-
tially by the greater residential stability associated with homeownership.
Keywords: Education; Families and children; Homeownership
Introduction and context
Much recent literature—both popular and academic—has focused on
conditions under which children are raised and the potential consequences
of these contextual factors for a variety of outcomes in later life. Indeed, the
body of scholarly statistical work seeking to identify the multiple predictors
of various social, economic, and psychological outcomes for children and

HOUSING POLICY DEBATE
786 George Galster, Dave E. Marcotte, Marvin B. Mandell, Hal Wolman, and Nancy Augustine
adults is voluminous and growing rapidly. It has also been the subject of sev-
eral recent comprehensive reviews (Earls and Carlson 2001; Ellen and Turner
2003; Galster 2005; Leventhal and Brooks-Gunn 2000; Robert 1999; Samp-
son, Morenoff, and Gannon-Rowley 2002). The bulk of this literature (e.g.,
Brooks-Gunn, Duncan, and Aber 1997; Furstenberg et al. 1999) examines
factors affecting outcomes at various stages of childhood. However impor-
tant such outcomes are, we believe that it is also crucial to examine factors
that account for young adult outcomes. In this regard, there is an established
literature examining such adult outcomes as welfare usage (Gottschalk 1996;
Gottschalk, McLanahan, and Sandefur 1994; Moff itt 1992; Pepper 2000;
Vartanian 1999); school dropouts (Clark 1992; Gleason and Vartanian
1999; Mayer 1997, Sawhill and Chadwick 1999); crime (Freeman 1991;
Grogger 1997; Peeples and Loeber 1994; Sullivan 1989); teen childbearing
(Barber 2001; Furstenberg, Levine, and Brooks-Gunn 1990; Haurin 1992;
McLanahan and Bumpass 1988; Sawhill and Chadwick 1999); economic
idleness (Haveman and Wolfe 1994; Mayer 1997; Payne 1987; Sawhill and
Chadwick 1999); and earnings (Corcoran et al. 1992; Haveman and Wolfe
1994; Vartanian 1999).
Of particular note for this article is emerging research that examines the
effects of the homeownership status of a family during child-rearing stages.
Although there is a considerable literature on the private and social benef its
of homeownership for such things as community participation, life satisfac-
tion, home maintenance, and accumulation of wealth (McCarthy, Van Zandt,
and Rohe 2001; Rohe, McCarthy, and Van Zandt 2000; Rossi and Weber
1996), only a few studies have attempted to link any of these effects to later
outcomes for children. The work of Green and White (1997); Boehm and
Schlottman (1999); Aaronson (2000); Boyle (2002); Harkness and Newman
(2002, 2003); Haurin, Parcel, and Haurin (2002a, 2002b); Haurin, Dietz,
and Weinberg (2003); and Kauppinen (2004) suggests that homeownership
status matters for children, although it is usually not clear whether the effect
is an independent one or is commingled with residential stability or neigh-
borhood conditions, or both.
As we shall amplify in this article, previous studies attempting to estimate
the relative importance of family background, neighborhood characteristics,
residential stability, and homeownership status on children’s later outcomes
as adults typically
1. Treat homeownership as independent of neighborhood characteristics
and residential stability

HOUSING POLICY DEBATE
Parental Homeownership and Young Adult Outcomes 787
2. Fail to measure homeownership over the whole childhood
3. Provide inadequate controls for children’s characteristics
4. Are limited in their ability to control for omitted parental variables (selec-
tion effects)
Our study offers what we hope will be advancements in these areas.
F irst, we treat neighborhood, stability, and homeownership as endogenously
determined (sometimes simultaneously so). Thus, for example, neighborhood
conditions during childhood may influence later adult outcomes directly
and also indirectly through their effect on household mobility and parental
choice of homeownership status. Second, we measure the cumulative impacts
of parental homeownership status during the entirety of childhood on sub-
sequent teen fertility, education, tenure status, and labor market earnings.
Third, we measure a host of observable parental characteristics and develop
a proxy for the unmeasured characteristics of children that likely affect
their outcomes as young adults but are spuriously related to their parents’
homeownership status. Fourth, we develop an instrumental variable (IV) for
childhood homeownership status as a further way of dealing with potential
endogeneity and selection bias problems.
We analyze data from the Panel Study of Income Dynamics (PSID) geo-
coded to census tract data. Using this panel data set, we follow children born
between 1968 and 1974 and observe their adult outcomes as of 1999 when
they were between 25 and 31 years of age. We can thus document the house-
hold environment of all of our sample children annually for all 18 years of
their childhood. In this fashion we can test the cumulative impact of parental
tenure status on children.
Our article is organized as follows. We f irst offer a holistic conceptual
framework for understanding how the homeownership status of parents
might influence outcomes for their children when they become young adults.
Second, we describe the preeminent omitted variables, selection, and endo-
geneity biases that must be overcome to gain accurate measurements of these
relationships. Third, we use our holistic framework as a vehicle for evaluating
earlier work and establishing a foundation for our modeling efforts. Fourth,
we describe our data set and the various procedures we use to attempt to
meet these challenges. F ifth, we present our empirical estimates of the key
relationships between parents’ homeownership status and children’s subse-
quent outcomes, exploring the sensitivity of results to alternative specif ica-
tions and controls. F inally, we discuss conclusions, policy implications, and
suggestions for further research.

HOUSING POLICY DEBATE
788 George Galster, Dave E. Marcotte, Marvin B. Mandell, Hal Wolman, and Nancy Augustine
The influence of homeownership status on young adult
outcomes: Theory, evidence, and empirical challenges
A holistic framework
To provide a framework for both assessing the limitations of previous
studies and guiding our own efforts, we present our structural model in f ig-
ure 1. We posit that young adult outcomes of interest (shown on the right
side of f igure 1) are determined by four sets of exogenous or predetermined
variables (followed by the appropriate path and examples in parentheses):
observed characteristics of individual children (path A: gender, race), unob-
served characteristics of individual children (path H: intelligence), observed
parental characteristics (path G: education, age), and unobserved parental
characteristics (path B: ambition, present orientation). These unobserved
parental factors (shown as dotted lines in f igure 1) are the source of the
omitted variables bias associated with selection, which we shall discuss later.
F inally, we see young adult outcomes as influenced by a set of intervening
endogenous variables: neighborhood characteristics (path C), parental hom-
eownership status (path D), and parental mobility expectations mediated by
actual mobility behavior (paths E and F). These variables we see as simulta-
neously determined, as we will explain later.
Our conceptual model is distinguished by the specif ication of homeown-
ership status/neighborhood location/mobility expectations as mutually causal
phenomena. Put differently, we argue that accurately measuring the relation-
ship of any one of these phenomena with young adult outcomes requires
that its relationship with all of the others be taken into account, a key point
to which we will return.
1
Earlier research has often modeled pairs of these
phenomena as jointly determined; examples are tenure and mobility (Boehm
1981; Ioannides 1987; Ioannides and Kan 1996; Kan 2000) and tenure and
neighborhood (Deng, Ross, and Wachter 2003; Gyourko, Linneman, and
Wachter 1997; Painter, Gabriel, and Myers 2001).
We posit the following structural equation system that will guide our
empirical work: Italicized acronyms indicate a simultaneous relationship,
and each variable is measured for a given household-year (not subscripted
for simplicity):
HO = f(N, ME, [X
1
]) (1)
1
We experimented in prototypes with modeling parental income as endogenous with
mobility expectations, neighborhood, and tenure choice and with developing IVs for them. It
proved challenging to identify instruments distinct from those used for tenure and neighbor-
hood, so these experiments are not reported here.

HOUSING POLICY DEBATE
Parental Homeownership and Young Adult Outcomes 789
F igure 1. A Structural Model of Childhood Determinants of Young Adult Outcomes
1
B
C
D
F
G
A
E
H
Educational
Attainment
Earnings
Home-
owner
Status
Teen Fertility
Young Adult
Outcomes
Observed
Individual
Characteristics
Unobserved
Parental
Characteristics
during Childhood
Observed
Parental
Characteristics
during Childhood
Unobserved
Individual
Characteristics
Neighborhood
Characteristics
during Childhood
Parental Home-
owner Status
during Childhood
Parental
Mobility
Expectations
Childhood
Parental
Residential Mobility
Behavior during
Childhood

HOUSING POLICY DEBATE
790 George Galster, Dave E. Marcotte, Marvin B. Mandell, Hal Wolman, and Nancy Augustine
N = f(HO, ME, [X
2
]) (2)
ME = f(HO, N, M-1, [X
3
]) (3)
M = f(ME-1, [X
4
]) (4)
where
HO = homeownership status (own or rent)
N = neighborhood poverty rate
ME = expectations regarding a potential move during the next year
M = actual mobility observed during the year
[X
i
] = vector of exogenous or predetermined predictors appropriate to
equation i
-1 = one-year lagged value of the variable
Rationale
We offer the following rationale for positing these simultaneous rela-
tionships among HO, N, and ME. For equation (1), if economic status (low
income and wealth) constrains a household to neighborhoods with numer-
ous social problems and concomitant expectations of depreciating property
values, there will be little motivation to buy a home. The same is true if the
household expects to move soon, because people will be unwilling to bear
the high transaction costs of buying. However, equation (2) suggests that
if a household would like to buy and expects to reside indef initely, certain
neighborhoods may not be selected because little appreciation in property
values or a poor quality of life is expected. For equation (3), someone who
can purchase a home in a good neighborhood will probably expect to move
less in the future. Conversely, current neighborhood conditions (especially if
they are declining) may trigger dissatisfaction and expectations of moving
out soon, but homeownership tenure may constrain out-migration if a home
must be sold to f inance an alternative dwelling in a different neighborhood
(equation [3]). Recent moves may reduce expectations of moving again in
the next year, given the substantial psychological, time, and out-of-pocket
costs involved. Equation (4) suggests that an actual move will typically be
preceded by the expectation of one.

HOUSING POLICY DEBATE
Parental Homeownership and Young Adult Outcomes 791
Challenges in accurately measuring determinants of young adult
outcomes
The holistic framework portrayed in f igure 1 makes it clear that there
are three preeminent challenges in obtaining accurate measurements of the
relationship between young adult outcomes and key childhood predictors of
interest, such as neighborhood, homeownership status, mobility, and certain
parental characteristics. They are selection bias and two sorts of omitted
variables bias.
Selection bias is now a well-known factor in determining the relation-
ship between environmental context and children’s outcomes. The basic issue
is that some parents who have certain (unmeasured) motivations and skills
related to their children’s upbringing would move to select neighborhoods,
choose a certain form of tenure, and manifest distinctive mobility patterns.
Any observed relationship between these contextual conditions and child or
young adult outcomes may therefore be biased because of this systematic
selection process, even if all of the observable characteristics of parents are
controlled for (Dietz 2002; Duncan, Connell, and Klebanov 1997; Dun-
can and Raudenbush 1999; Manski 1995, 2000). The problem can also be
formulated as omitted variables bias stemming from unobserved parental
characteristics. Is the observed statistical relationship between outcomes and
context indicative of an independent effect or merely unmeasured parental
characteristics that truly affected children’s outcomes but led to the observed
context as well?
2
We portray the implicit omitted parental variables’ relation-
ships in this selection problem as dashed lines in f igure 1.
When analyzing a sample of households that have chosen their neigh-
borhoods, tenure, and mobility through private market processes, this selec-
tion bias is likely severe indeed (Manski 1995; Tienda 1991). A variety of
econometric techniques, including sibling studies and IVs, has been used in
an attempt to overcome selection bias, but with incomplete success and/or
limited general applicability (see the reviews in Galster 2003 and 2005).
In addition, a few studies have attempted to explicitly model the selection
process into owner and rental tenures as part of a larger analysis of chil-
dren’s outcomes (Green and White 1997; Haurin, Parcel, and Haurin 2002a,
2002b).
3
2
The direction of the bias has been the subject of debate, with Jencks and Mayer (1990)
and Tienda (1991) arguing that neighborhood impacts are biased upward and Brooks-Gunn,
Duncan, and Aber (1997) arguing the opposite.
3
Green and White (1997) and Haurin, Parcel, and Haurin (2002b) come to opposite
conclusions about whether there are important selection effects between tenure choice and
outcomes.

HOUSING POLICY DEBATE
792 George Galster, Dave E. Marcotte, Marvin B. Mandell, Hal Wolman, and Nancy Augustine
There is yet another possible form of omitted variable bias (also shown
as a dashed line in f igure 1), stemming from the unobserved characteristics
of children whose young adult outcomes are ultimately measured. Such char-
acteristics are, in turn, likely a product of observed and unobserved parental
characteristics. If some of these parental characteristics also predict home-
ownership status, there will be a correlation (albeit spurious) between the
unobserved children’s characteristics and parental homeownership status,
thereby biasing the estimated coeff icient of the latter.
However, the challenge is even more complex than measuring all of the
relevant characteristics of parents and children. If f igure 1 is adopted as a
working premise, the selection process becomes much more complicated than
merely parents’ independent selection of a neighborhood, or a tenure, or a
pattern of mobility. In our view, the holistic challenge embodies the simul-
taneous selections of all three. Previous statistical studies have taken only a
partial view of the causal patterns in f igure 1 and equations (1) through (4);
virtually all have omitted one or more of the intervening variables. To the
extent that these variables are mutually causal, they will be correlated with
the neighborhood variable. Under these circumstances, the coeff icient will
be a biased estimate of the effect of neighborhood on outcomes because the
neighborhood variable is correlated with the disturbance term in the regres-
sion. As in the case of selection, there is an omitted variables bias problem,
but here it is due to the causal relationship between the neighborhood vari-
able and other uncontrolled variables that also affect outcomes. However,
the solution to this problem may not be as straightforward as including all of
the intervening variables in the outcome equation. If the causal relationships
are as strong as we have posited, these intervening variables may be so highly
correlated that multicollinearity may arise as a new econometric challenge.
Earlier work on homeownership status, residential stability, and
child outcomes
The literature consistently f inds that parental homeownership status has
a positive impact on children, ranging from early childhood cognitive, emo-
tional, behavioral, and social development (Boyle 2002; Haurin, Parcel, and
Haurin 2002a, 2002b); educational attainment (Aaronson 2000; Boehm and
Schlottman 1999; Green and White 1997; Harkness and Newman 2002,
2003; Kauppinen 2004); teen childbearing (Green and White 1997; Harkness
and Newman 2002, 2003); and earnings and welfare usage (Harkness and
Newman 2002, 2003) to the purchase of a home as a young adult (Boehm
and Schlottman 1999), although Harkness and Newman (2003) f ind stron-

HOUSING POLICY DEBATE
Parental Homeownership and Young Adult Outcomes 793
ger effects for children from lower-income households. Two primary causal
mechanisms through which parental homeownership status could produce
these felicitous outcomes for children have been advanced—a direct mecha-
nism (path D in f igure 1) and an indirect one (via mobility, paths E and F).
4
The direct effects that have been posited include the following:
1. Homeowners maintain their dwellings to higher standards than other-
wise identical rental households (Galster 1983, 1987; Mayer 1981); this
may differentially affect the health and cognitive and social development
of resident children (Parcel and Menaghan 1994a, 1994b).
2. Homeowners may acquire a distinctive set of skills, such as those related
to performing do-it-yourself home repairs, negotiating with contractors
or plumbers, and seeking ref inancing. Insofar as these may be transfer-
able to children, the latter will benef it (Boehm and Schlottman 1999;
Green and White 1997).
3. Homeowners may have more of a f inancial stake in the residence and
thus more motivation to monitor and control the activities of children
(both their own and their neighbors’) that might threaten local property
values (Haurin, Parcel, and Haurin 2002a, 2002b; Hoff and Sen 2005).
4. Homeowners may invest more in developing social capital and partici-
pating actively in the neighborhood; their children may benef it in a vari-
ety of ways (Austin and Baba 1990; Coleman 1988, 1990; Cox 1982;
DiPasquale and Glaeser 1999; Hunter 1975; Jeffers and Dobos 1984;
Rohe and Stegman 1994; Rossi and Weber 1996; Verba, Schlozman, and
Brady 1995).
5. Buying a home may yield gains in satisfaction and self-esteem, which in
turn translate into a more supportive, positive sociopsychological envi-
ronment for children (Balfour and Smith 1996; Rossi and Weber 1996).
6. To the extent that home appreciation outperforms other f inancial instru-
ments, homeowners may achieve a better equity position than renters
(for a review, see Herbert and Belsky 2006) and thus would be able to
invest more in the educational and nurturing aspects of the children’s
environment.
5
4
See Dietz and Haurin (2003) for a comprehensive review of the consequences of
homeownership.
5
After controlling for homeownership status and other parental characteristics, Haurin,
Parcel, and Haurin (2002a, 2002b) f ind that wealth was unrelated to either cognitive or emo-
tional dimensions of the home environment, children’s math and reading test scores, or an
index of children’s behavioral problems.

HOUSING POLICY DEBATE
794 George Galster, Dave E. Marcotte, Marvin B. Mandell, Hal Wolman, and Nancy Augustine
7. Homeowners may experience lower levels of stress because of greater
security of tenure, which produces more positive behavioral and cogni-
tive outcomes in the household (Cairney and Boyle 2004).
The indirect consequences putatively transpire through the effect of
homeownership status on residential stability. The argument proceeds as fol-
lows. Because of the high transaction costs involved in buying and selling a
home (Haurin, Hendershott, and Ling 1988), owners typically reside in any
given unit longer than renters do (Lee, Oropesa, and Kanan 1994; Rohe and
Stewart 1996). In turn, this enhanced residential stability can have numer-
ous positive impacts on children’s educational achievement and credential
attainment, substance use, social functioning, mental health, and sexual and
deviant behaviors (Buerkle 1997; DeWit 1998; Huff ines 2003; Potter et al.
2001; Stack 1994; Temple and Reynolds 1999). At least one potential reason
for this relationship is that as children remain longer in a neighborhood, they
are likely to become better known to adult neighbors, thus rendering them
more subject to behavioral modif ications through the exercise of neighbors’
“collective eff icacy” (Sampson, Morenoff, and Earls 1999).
Unfortunately, it is not possible to distinguish def initively among these
hypotheses from extant empirical work. The earliest study in this area, by
Green and White (1997), found that current parental homeownership was
associated with a reduced probability that a resident 17-year-old would drop
out of high school or give birth. Past residential mobility and homeowner-
ship status were not controlled for, however, and subsequent analyses by
Aaronson (2000) and Harkness and Newman (2002) showed that most of
the relationships were explained by the greater residential stability associated
with homeownership. After controlling for a variety of family and neighbor-
hood characteristics in his 2002 study of longitudinal samples of children in
Ontario, Canada, Boyle revealed a signif icantly lower rating for emotional-
behavioral problems for children of homeowners, but no attempts to control
for selection were provided.
Controlling for residential stability and wealth, Haurin, Parcel, and
Haurin (2002a, 2002b) offer the strongest support for a direct relationship
between homeownership and child outcomes. In a well-controlled treatment
effects model, they f ind that homeownership is positively related to both
indexes of the cognitive/stimulative and emotional/supportive dimensions
of the home environment. These two indexes, in turn, prove strongly pre-
dictive of children’s math and reading test scores and behavioral problems.
Moreover, homeownership still proves signif icant in predicting test scores

HOUSING POLICY DEBATE
Parental Homeownership and Young Adult Outcomes 795
(although not behaviors) when these home environment indexes (and other
parental and neighborhood characteristics) are controlled for.
Though provocative, the literature on how homeownership status may
affect children remains limited in many respects. Several studies omit key
control variables for parents and children and overlook selection effects that
may bias the apparent impacts of homeownership upward. Indeed, Barker
and Miller (2005) f ind that many of the reported effects of homeowner-
ship disappear when more controls and alternative estimation techniques
are introduced. No earlier studies have collected information on multiple
dimensions of context over the entirety of childhood and thus have not tested
for duration effects. But more important here, this work has taken only a
partial view of the causal patterns in f igure 1; none has modeled tenure sta-
tus, mobility, and neighborhood choices as mutually causal.
6
This potential
source of bias provides an additional reason for questioning the accuracy of
the relationships they measure between outcomes and key predictors of inter-
est. Our study attempts to improve on all the aforementioned shortcomings
of the literature.
Data to be analyzed and key measures
The PSID
A brief overview of the PSID data we analyze is a prerequisite for under-
standing our approach. Beginning in 1967, the PSID began interviewing 5,000
U.S. families. Each year since then, those families have been interviewed, as
have all families subsequently formed by individuals in those families and
by the spouses and children of those individuals. So, by 1999, the PSID was
following nearly 10,000 families. While at f irst it oversampled to obtain rela-
tively large sample sizes for poor households, the oversample was dropped
in the 1990s. Consequently, our analysis is limited to a sample designed to
be nationally representative of the U.S. population in 1967. We account for
differential attrition over the course of the panel by adjusting individuals’
PSID sampling weights by the inverse of the reciprocal of the attrition rate of
sample members with the same race, gender, and poverty status at birth. We
use a PSID geomatched f ile, which appends the child’s census tract identif ier
6
While other studies (Manski 1995) have discussed the issue of “simultaneity bias,” they
used the term to refer to the reflection problem of people tautologically causing the aggregate
neighborhood characteristics to be what they are and the neighborhood causing constituent
resident behaviors.

HOUSING POLICY DEBATE
796 George Galster, Dave E. Marcotte, Marvin B. Mandell, Hal Wolman, and Nancy Augustine
to each observation and interpolates the values of census tract variables for
observations between census years. We are thus able to observe annually
the household and (approximate) neighborhood environments in which our
sample individuals spend their childhood.
We focus our analysis on the PSID cohort of children born between 1968
and 1974 because it provides us with data on their f irst 18 years, as well as a
variety of outcomes measured in 1999 when they were between 25 and 31—
young adults who most likely had completed their education and had had
ample opportunity to enter the labor force.
7
Here, as throughout, we pres-
ent PSID weighted statistics, adjusted for group-specif ic attrition. Descriptive
statistics for the sample of children we analyzed—themselves, their parents/
households, and their neighborhoods as they were growing up—are pro-
vided in table 1.
Measures of key explanatory variables
Of particular interest to our inquiry are PSID-based statistics for paren-
tal homeownership status. On average, our sample spent 72 percent of their
childhood in households where the head owned the dwelling unit. Only 6
percent of the children in the sample grew up in a household in which the
head never owned the home in which they lived. Almost one in f ive (18
percent) lived up to 9 years in an owner-occupied home, almost one in three
(31 percent) lived between 9 and 17 years in an owner-occupied home, and
almost half (45 percent) lived all 18 years in a home that was owned by
the head. Although comparable tenure information is not available from the
decennial censuses, it is instructive to note that the national homeownership
rate during our cohort’s childhood was 59 percent in 1970 and 62 percent in
both 1980 and 1990.
8
To operationalize the neighborhood component of the childhood con-
text, we use information from the census tract, a homogeneous area of
roughly 4,000 inhabitants, tabulated in the decennial Census of Population
and Housing, with values interpolated for intercensus years.
9
In particular,
we report here on results when the neighborhood variable is the percentage
7
Such a longitudinal analysis has been strongly recommended as the vehicle for overcom-
ing the reflection problem (Manski 1995).
8
All national census f igures quoted here were obtained from the U.S. Bureau of the Census
(1972a, 1972b, 1983, 1984, 1992, 1993).
9
We use the GeoLytics Neighborhood Change Database, which adjusts data in 1970,
1980, and 1990 tracts that have changed their boundary def initions over the years to values
that would apply if boundaries had remained at their 1990 specif ications (GeoLytics 2007).

HOUSING POLICY DEBATE
Parental Homeownership and Young Adult Outcomes 797
Table 1. Characteristics of Sample Individuals and Their Mean Circumstances from
Ages 0 to 18
Mean
Characteristics of individuals observed in 1999
Black female [blackfem] 0.041
Black male [blackmale] 0.057
White female [whitefem] 0 . 3 31
Order of birth (1 = f irst, 2 = 2nd, etc.) [birthorder] 2.233
Age in years [age99] 28.74
Married [married] 0.481
Exogenous characteristics of parents and household (calculated over ages 0 to 18)
Proportion of years lived in poverty [pro_live_in_poverty0to18] 0.069
Proportion of years lived with two parents [pro_livew_2_parents0to18] 0.842
Proportion of years lived in a metropolitan area [ave_smsa0to18] 0.731
Average number of neighbors the head knew by name [ave_num_neigh_known] 12.33
Proportion of years lived with family within walking distance [ave_relatives] 0.392
Education of the household head [ave_education_head0to18] 13.24
Occupational prestige of the household head [ave_hdocc_pre0to18] 43.92
Proportion of years the head was self-employed [ave_self_employed0to18] 0.141
Proportion of years the wife of the head was employed [ave_employed_wife0to18] 0.491
Annual hours the head worked [ave_annu_hrs_wkd0to18] 2123
Head self-identif ied as Protestant, Catholic, or Jewish [religion] 0.901
Proportion of years the head read a newspaper every day [ave_readnewspaper] 0.802
Proportion of years the head belonged to a union [ave_union] 0.285
Proportion of years the head did not attend a religious service weekly [ave_nochurch] 0.226
Proportion of years the head never participated in social clubs [ave_no_socialclubs] 0.538
Proportion of years the head “planned his/her life ahead” [ave_plan_ahead] 0.581
Proportion of years the head “trusted most people” [ave_trust] 0.604
Head is a veteran [veteran] 0.392
Mother f irst gave birth as a teen [momteen] 0.045
Head raised in a large city (not a suburb) [largecity] 0.41
Head raised in a rural or small town (not a suburb) [farm] 0.174
Endogenous characteristics of the household (calculated over ages 0 to 18)
Proportion of years when the head owned the home [pro_parents_own0to18] 0 . 7 2 2
Proportion of years when the residence was not changed [pro_stability_year0to18] 0. 8 0 9
Average percentage of the population below poverty, census tract [ave_neigh_in_pov0to18] 10.25
Source: Authors’ analysis of PSID data for a select sample; N = 755 (weighted). The source for Average
percentage of the population below poverty, census tract, was GeoLytics 2007.
Note: The variable abbreviations used in the appendixes are shown in brackets.

HOUSING POLICY DEBATE
798 George Galster, Dave E. Marcotte, Marvin B. Mandell, Hal Wolman, and Nancy Augustine
of persons living in households with incomes below the federal poverty line.
10
This variable has been widely used in the literature as a powerful proxy for
neighborhood disadvantage. On average, children in our sample experienced
a census tract having a poverty rate of 10.25 percent during their childhood.
This compares with a national poverty rate during our cohort’s childhood of
12.6 percent in 1970 and 11.7 percent in both 1980 and 1990.
Residential stability and mobility expectations were measured from PSID
data. Each year, the respondent is asked whether the household plans on mov-
ing within the next 12 months; we code aff irmative responses as a dummy
variable for that year. Because we also know actual changes in respondent
addresses, we can calculate the number of times the household moved dur-
ing each child’s f irst 18 years. We def ine a stability variable that denotes the
proportion of these 18 years during which the child did not move; the mean
for our analysis sample was 81 percent. By contrast, the national percentage
of households that did not move was 77 percent in both 1970 and 1980 and
78 percent in 1990.
A necessary condition for the precise measurement of contextual effects
is that the widest possible array of the characteristics of children and their
household while growing up should be included as controls in the model
(Ginther, Haveman, and Wolfe 2000). We believe that our work has met this
condition in a way that is superior to what is seen in earlier studies. We can
control for a wide range of objective characteristics of the household, such
as marital, poverty, and occupational status; education; religion; teen fertil-
ity; and location. Moreover, unlike earlier studies, we can control for several
attitudinal characteristics (such as trust in others, present orientation) and
behavioral characteristics of the head, such as how often newspapers are
read, social clubs and religious services attended, and so on).
F inally, we develop an innovative proxy measure for unobserved chil-
dren’s characteristics based on the ability to be a homeowner as a young
adult. We begin by positing that our cohort’s probability of homeownership
status by 1999 (H) can be modeled as function of
1. The individual’s observable characteristics [X
6
] that are both time-invari-
ant (race, gender) and time-varying (marital status, income, education)
2. The individual’s unobservable characteristics [X
u
] (future orientation,
credit rating, etc.) and exogenous observable characteristics of the parent/
10
We also experimented with homeownership rates in the neighborhood (Ding and Knaap
2003), but were unable to f ind an IV that was not highly correlated with the IV for parent’s
tenure status.

HOUSING POLICY DEBATE
Parental Homeownership and Young Adult Outcomes 799
household during the individual’s childhood [X
7c
] (marital status, pov-
erty, etc.)
3. Endogenous observable characteristics of the parent/household during
the individual’s childhood [X
8c
] (homeownership, neighborhood pov-
erty, mobility)
4. 1999 characteristics of the individual’s metropolitan area [A] that affect
the home purchase decision (home prices, rental costs, etc.)
prob (H) = f([X
6
], [X
u
], [X
7c
], [X
8c
], [A]) + ε (5)
Next, we estimate a preliminary logistic regression model (5) (obviously,
excluding [X
u
]) for our cohort in 1999. F inally, we save the residuals of
this regression and use them as a proxy for [X
u
] in our outcome equations
for young adult fertility, education, and income.
11
We cannot be certain, of
course, what unobserved attributes [X
u
] measures. Nevertheless, we think it
reasonable to assume that many of the same unobservables that will inf lu-
ence becoming a homeowner will also be predictors of these other outcomes
and thus should ideally be controlled for.
Measures of outcomes for children as young adults
Our goal here is to relate the childhood average experience of paren-
tal homeownership status, controlling for all the other characteristics of the
child’s environment noted earlier (and listed in table 1), to the following out-
comes: tenure status, teen fertility, school credential attainment, and labor
earnings as of 1999. Descriptive statistics for these outcomes for our analysis
sample are presented in table 2.
By 1999, 44 percent of the children born between 1968 and 1974 had
formed their own household and were residing in a home that they had pur-
chased. A total of 88 percent had graduated from high school or obtained a
general equivalency diploma, and 14 percent had graduated from a four-year
college. Some 94 percent had reached age 18 without having had a child. The
PSID collects income information only from respondents who have formed
11
An analogous procedure for estimating proxies for time-invariant individual character-
istics has recently been used by Musterd et al. (forthcoming).

HOUSING POLICY DEBATE
800 George Galster, Dave E. Marcotte, Marvin B. Mandell, Hal Wolman, and Nancy Augustine
their own household, so the income statistics we report refer only to these
members of our cohort. This group earned $18,509 on average in 1998 and
had a total family income of $34,752.
12
Estimation procedures and methodological issues
Model overview
Our models for fertility, education, and earnings outcomes of young
adults are as follows:
FER = f(HO
c
, N
c
, M
c
, [X
6
], [X
7c
], [X
u
]) +
ε
(6)
HS = f(HO
c
, N
c
, M
c
, [X
6
], [X
7c
], [X
u
], FER) +
ε
(7)
COL = f(HO
c
, N
c
, M
c
, [X
6
], [X
7c
], [X
u
], FER) +
ε
(8)
INC = f(HO
c
, N
c
, M
c
, [X
6
], [X
7c
], [X
u
], FER, HS, COL, HRS) +
ε
(9)
where
FER = 1 if reached age 18 without having a child, 0 otherwise
HS = 1 if received a high school diploma or equivalency degree by 1999, 0
otherwise
COL = 1 if received a bachelor’s (4-year) degree by 1999, 0 otherwise
Table 2. Descriptive Statistics for Outcomes of Young Adult Sample
Mean Standard Deviation
Had No Children Before Age 18 0.944 0.107
Completed High School or More, 1999 0.881 0.352
College Degree or More, 1999 0.144 0.351
Owned Home, 1999 0.442 0.162
Ln Annual Earnings, 1998 9.826 1.102
12
In 1998, 85 percent had enough family income to keep them out of poverty. In prelimi-
nary runs, we experimented with several other measures of labor market outcomes: annual
hours worked, employed during the previous year, and not in poverty during that year. None of
the endogenous variables in our model ever proved related to these outcomes in a statistically
signif icant way, so for parsimony we omit them from the discussion.

HOUSING POLICY DEBATE
Parental Homeownership and Young Adult Outcomes 801
INC = the natural logarithm of 1998 income from earnings (only for those
who had formed a household and were employed at some time during
1998)
HO
c
= proportion of childhood years that parents owned the dwelling they
occupied
N
c
= average poverty rate in the census tract in which the child resided
between birth and age 18
M
c
= proportion of childhood years that the household moved between
dwellings
[X
6
] = exogenous (observed) characteristics of the individual in 1999; see
table 1
[X
7c
] = exogenous characteristics of parent/household during childhood; see
table 1
[X
u
] = proxy for unobserved individual characteristics (see the earlier
discussion)
HRS = hours worked during 1998
c subscripts = variables computed for the entire childhood
We estimate the coeff icients of variables in this model using ordinary
least squares (OLS) when the outcome is continuous (equation [9]) and logit
when the outcome is dichotomous (equations [5] through [8]). The sample
for estimating these coeff icients includes all children in our initial cohort who
have “survived” in the sample to the point at which the outcome in question
is observed: 1999. Equations for fertility, education, and earnings outcomes
have virtually identical right-hand sides measuring (exogenous or predeter-
mined) characteristics of the individual and the individual’s household and
(endogenous) aforementioned childhood conditions. Descriptive statistics of
these last variables are presented in table 1. For all of these variables, we use
proportional f igures calculated over the f irst 18 years of the child’s life (or for
however many years we have data).
We model our set of outcomes as causally interrelated, as shown on the
right side of f igure 1. Educational attainment is a function of fertility before
age 18. Earnings are a function of fertility and education. Homeownership is
a function of income, fertility, and education.

HOUSING POLICY DEBATE
802 George Galster, Dave E. Marcotte, Marvin B. Mandell, Hal Wolman, and Nancy Augustine
Instrumentation procedure
We suspect that HO
c
will be correlated with the disturbance terms in
equations (5) through (9) because of the issues we have discussed. There-
fore, we will experiment with IVs. Our approach for estimating IVs proceeds
through the following three steps.
F irst, we estimate an OLS regression based on observations of individual
child-years. In this regression, the left-hand side is the observed value of the
census tract poverty rate in a given child’s neighborhood in a particular PSID
year, and the right-hand side contains observed values of every exogenous
variable [X] in the system of equations (1) through (4). These exogenous
variables include contemporaneous values of countywide characteristics cor-
responding to the HO
c
, N
c
, and M
c
variables, additional exogenous pre-
dictors for each outcome, and dummy variables for the calendar year. The
complete listing is shown in appendix A. In this f irst step, the regression is
estimated based on all annual observations from age 1 to age 18 for each
child in our sample.
13
We included all observations of children having data
for at least 10 years of their childhood. What is of prime importance here
is how well the f irst-stage linear probability model regressions predict the
values of HO
c
, not their estimated parameters in and of themselves, because
this will determine the power of the instrument (Murray 2006). As a result,
for this f irst stage we use OLS, not needlessly complicated panel estimation
procedures.
In the second step, the regression is used to generate predicted values of
household homeownership probability for each of the f irst 18 years of each
child’s life, based on values of all exogenous variables appropriate for the
given year. Now we must switch from a child-year unit of observation to a
child-childhood average unit, which necessitates a step not normally required
in two-stage least squares. In the third step, we compute the average of these
predicted values over all observed years of childhood. These averages for
each sampled individual become our IV measures for the duration of home-
ownership experienced during childhood.
Identifying and evaluating instruments for homeownership status
during childhood
To satisfy the rank condition in performing two-stage least squares, there
must be at least as many exogenous variables excluded from equations (1)
through (4) as there are endogenous variables included in each one. We meet
13
We begin with age 1, not birth, because some explanatory variables are lagged.

HOUSING POLICY DEBATE
Parental Homeownership and Young Adult Outcomes 803
this condition; indeed our equation system (1) through (4) is overidentif ied.
Moreover, each equation must have one or more clearly exogenous variables
that appear only in the given equation as strong predictors. In the case of
household homeownership status during childhood, we use the following as
unique identifying instruments:
1. Index of owner-occupied housing prices in the metropolitan area (lag 1
year)
2. Index of owner-occupied housing prices in the metropolitan area (lead 1
year)
3. Index of gross rents paid by renter occupants in the metropolitan area
(lag 1 year)
4. Item 3 times renter status in the preceding year
5. Home mortgage interest rate for a 30-year f ixed-rate loan
6. Item 5 times renter status in the preceding year
7. Ratio of the costs of renting to the costs of owning in the metropolitan
area (lag 1 year)
8. Item 7 times renter status in the preceding year
9. Cases where the household head received a lump-sum monetary pay-
ment since the child’s birth (e.g., inheritance: 1 = yes; 0 = no)
10. Item 9 times renter status in the preceding year
11. Difference in the household’s real income from the preceding year to the
current one (if greater than 0; 0 otherwise)
12. Item 11 times renter status in the preceding year
13. Homeownership rates in the county (lag 1 year)
The rationales for these variables are straightforward. We would expect
that households would be more willing and able to own their dwelling in
a metropolitan area with lower absolute prices of owner-occupied dwell-
ings, lower relative prices of owner-occupied versus renter-occupied dwell-
ings, higher expected dwelling appreciation (for which we use leading values
as a proxy), and lower mortgage interest rates. They should also be more
able to own if they received a lump-sum payment during the child’s lifetime
(our only proxy for down payment constraints) and experienced a growth in
income. Homeownership rates in the county serve as a proxy for unobserved
characteristics of the area’s housing stock and the household’s preferences for

HOUSING POLICY DEBATE
804 George Galster, Dave E. Marcotte, Marvin B. Mandell, Hal Wolman, and Nancy Augustine
homeownership.
14
We allow coeff icients of most of those variables to differ
between households that rent and those that own.
Additional exogenous variables coming from equations (2) through (4)
and used in the f irst stage are presented in appendix A. Overall, our f irst-
stage linear probability model for household homeownership status during
a given year performed only moderately well (the R
2
was 0.29), although we
had 12,500 child-year observations in this f irst-stage regression and only 31
regressors.
Complicating issues
F ive issues require further discussion. The f irst is the operational def ini-
tion of neighborhood. Although census tracts are imperfect, we use them as
our preferred approximation to neighborhood, as is common in U.S. studies.
However, until 1990, rural areas were not divided into census tracts. To avoid
the potential problems of missing data and a mixture of urban and rural
scales of “neighborhood,” we conf ine our analysis to children who spent at
least 12 of their f irst 18 years in tracted, metropolitan-area neighborhoods.
Second, the attitudes and behaviors of the household head that we use as
controls (see table 1) are not measured annually in the PSID. Indeed, for most
variables, the questions were asked only from 1968 through 1972.
15
Each
attitude and behavior we used as a control proved stable over time. Pair-wise
correlations between responses to the question “carry out plans” over the six
points in time at which this question was asked ranged from 0.17 to 0.40.
Cronbach’s alpha, a measure of internal consistency, for a scale consisting
of the sum of the responses to this question over the six years, was 0.70.
Pair-wise correlations between responses to the question “plan ahead” over
the six points at which this question was asked ranged from 0.20 to 0.46;
Cronbach’s alpha was 0.77. Pair-wise correlations between responses to the
question “trust” over the f ive points in time at which this question was asked
ranged from 0.40 to 0.54; Cronbach’s alpha was 0.81.
The third issue requiring some discussion is the handling of homeowner-
ship status in a given year—a dichotomous variable. As noted in Wooldridge
(2002), it is not appropriate to apply probit or logit models to such variables
and then use the predicted probability on the right-hand side in second-stage
estimation. Hence, in our f irst stage, we applied a linear probability model to
14
The instrumentation strategy follows from the seminal works by Evans, Oates, and
Schwab (1992) and Foster and McLanahan (1996).
15
However, some were asked again in 1975, and a question about union membership was
collected from 1968 through 1981.

HOUSING POLICY DEBATE
Parental Homeownership and Young Adult Outcomes 805
this dichotomous endogenous variable and used the predicted value obtained
from that estimation in our second stage.
Fourth, given our conceptual model in f igure 1, it would have been desir-
able to simultaneously instrument not only for HO
c
but also for all of the
endogenous variables in equations (5) through (9), including N
c
, and M
c
. This
would have permitted the estimation of less biased parameters for these vari-
ables as well. Unfortunately, in our preliminary experiments, it proved chal-
lenging to identify instruments that uniquely identif ied all of these variables.
The result was that the resulting instruments for HO
c
, N
c
, and M
c
proved
too intercorrelated to be meaningfully employed in the same regression.
F inally, as noted, our instrumentation procedure involves estimates for
HO
c
that are multiyear averages of predicted values. Given that the distri-
bution of this new, “average” instrument is not known, the standard errors
yielded by conventional OLS or logit procedures cannot be interpreted in a
straightforward fashion. Thus, as is standard practice under these circum-
stances, we will report “bootstrapped” parameter values as estimated by
Stata when examining our IV estimates.
Estimates of the relationship between parental homeownership
status and outcomes for young adults, with and without controls
for residential stability
Overview and discussion of control variables
We estimated logit models for equations (5) through (8) and an OLS
model for equation (9). Before turning to the results for the homeownership
variable, we will briefly highlight some of the more interesting relationships
involving other variables; the details are presented in table A.1. As an over-
arching assessment, all outcome equations had decent explanatory power
according to the criteria appropriate for logit and OLS estimations (see table
A.1). Moreover, there is strong support for our specif ication of recursive
relationships between teen fertility, educational attainments, subsequent
earnings, and homeownership status. Not surprisingly, having a child before
age 18 clearly appears to reduce the chances of graduating from high school.
Educational attainments, especially a college degree, are strongly related to
earnings and homeownership in turn. F inally, there is strong support for our
claim that the intervening variables (which we argue are mutually causal
with homeownership) are important predictors of young adult outcomes.
This reinforces our contention that models of homeownership effects that
do not control for these contexts likely suffer from severe omitted variables
bias.

HOUSING POLICY DEBATE
806 George Galster, Dave E. Marcotte, Marvin B. Mandell, Hal Wolman, and Nancy Augustine
Owning a home in 1999 was associated with living more childhood years
with both parents, especially if the mother did not work outside the home
(see table A.1). Ownership was also more likely if the mother did not give
birth while a teen, the young adult was already married and college educated,
and the parent or parents were more trusting of others.
As for predictors of not having a child before age 18, growing up with
a mother who did not give birth as a teen and with a household head who
ascribed to “planning ahead” proved eff icacious. The former result may be
due to unmeasured characteristics of the mother related to norms and values
surrounding teen childbearing. Future-oriented parents may be more prone
to instill these same attitudes in their children, thereby encouraging them to
avoid future prospect-stunting actions like teen child-rearing.
Consider next the educational attainment equations. Not surprisingly,
having a well-educated parent or parents was strongly correlated with greater
chances of later graduating from high school and college. The same pattern
held for children raised by parents who knew more neighbors by name. We
cannot be sure why this greater degree of parental neighborhood social inte-
gration (controlling for mobility) seemingly translates into greater educa-
tional achievements, although it may be due to the implied intensif ication
of neighbors’ “bonding social capital” and monitoring of children’s behav-
iors, both pro- and anti-educational (Brisson and Usher 2007). This explana-
tion would also be consistent for the f inding that children of parents who
never belonged to social organizations were less likely to graduate from high
school. Older members of our cohorts evinced higher achievements, prob-
ably because they had more time to obtain graduate equivalency exams and
complete college coursework.
Other statistically signif icant relationships do not have obvious expla-
nations. Children from homes where the head was more trusting of other
people, was a veteran, or was not a union member were less likely to gradu-
ate from high school. Children who were raised in a large city (instead of a
suburb) and came from homes where parents were not members of social
clubs were more likely to get a college degree.
F inally, in the wage earnings equation, we observe that all else being
equal, children raised by a head who was more future-oriented earned more,
suggesting that these children learned attitudes and behaviors that are related
to delayed gratif ication and longer-term strategizing and have substantial
labor market payoffs. Children from households experiencing longer spells
of poverty and/or single parenting earn less, consistent with the hypothesis
that the material and psychological deprivation associated with these cir-
cumstances creates developmental disadvantages with a lasting impact on

HOUSING POLICY DEBATE
Parental Homeownership and Young Adult Outcomes 807
earnings. Older cohorts earn more, as would be expected from their typically
longer tenure in the workforce. Also as expected, employees earn more if
they are better-educated white males and work more weeks during the year.
It is less clear precisely why children raised in dual-parent homes with moth-
ers who worked more earned more as young adults, although role modeling,
familial resources, and/or intrafamily social dynamics of unspecif ied content
may be at work.
Baseline estimates of parental homeownership and effects of
successively adding controls
Earlier we explained how the potential impacts of parental homeown-
ership on children can be either direct or indirect through the impact on
increasing residential stability. To get our initial estimate of the full effect of
homeownership through both means, we estimated models where the resi-
dential stability variable was omitted. Our approach involved entering blocks
of control variables to ascertain the sensitivity of the estimated coeff icient of
parental homeownership. These parameter estimates are presented in the top
panel of table 3.
Our baseline estimates control for all of the observed characteristics of
individual children as listed in table 1, plus the basic (exogenous) socioeco-
nomic characteristics of their parents and their neighborhood.
When only these rather rudimentary controls are used, our results indi-
cate that parental homeownership is related in a statistically signif icant way
only to the education and homeownership outcomes. The logit model results
(summarized in the f irst row of the top panel of table 3) indicate that, all else
being equal, the proportion of years during childhood that parents owned
their home was positively associated with the probability that children would
earn a high school diploma, obtain a college degree, and own a home as
young adults. When controlling for education, we found no statistically sig-
nif icant association between parental homeownership status and teen fertil-
ity or children’s subsequent wage earnings. This is consistent with f indings
from Boehm and Schlottman (1999) and Harkness and Newman (2002). It
thus appears that whatever influence parental homeownership status may
have on young adults’ economic success occurs indirectly through an effect
on educational attainments.
Do the two statistically signif icant relationships persist when we add
more controls? The second row of results in the top panel of table 3 shows
the effect of adding all remaining observed parental control variables (atti-
tudes and behaviors) from table 1. Several of the behaviors and attitudes of

HOUSING POLICY DEBATE
808 George Galster, Dave E. Marcotte, Marvin B. Mandell, Hal Wolman, and Nancy Augustine
the head that were measured early in the child’s life (which have been omit-
ted from previous studies) prove to hold statistically signif icant explanatory
power, as noted earlier. The addition of these controls attenuates the magni-
tude of the parental homeownership coeff icient only slightly in the case of
college degree and high school attainment, but has no effect on homeowner-
ship parameters.
The third row in the top panel of table 3 shows the effect of adding our
proxy for the unobserved characteristics of the young adults along with all
previous controls. This proxy performs powerfully as a predictor of young
Table 3. Estimated Parameters for Childhood Average Parental Homeownership as a
Predictor of Young Adult Outcomes in 1999, by Model Specif ication
High School College No Child
Diploma or Degree before Owner of Ln (Wage
Controls More or More Age 18 Home Earnings)
Stability not controlled
Individual characteristics 0.943 1.223 0.474 1.045 0.096
and parental SES and (0.404)** (0.473)*** (0.453) (0.281)*** (0.147)
neighborhood
Above plus parental 0.833 1.15 0.350 1.065 0.02
attitudes and behaviors (0.421)** (0.524)*** (0.501) (0.360)*** (0.156)
Above plus proxy for 0.834 1.115 0.343 NA 0.051
unobserved individual (0.387)** (0.543)** (0.482) (0.136)
characteristics
Above, using IV 1.278 0.851 1.116 2.226 0.342
(1.194) (1.258) (1.409) (0.937)** (0.498)
Stability controlled
Individual characteristic 0.765 0.927 –0.056 1.195 0.087
and parental SES and (0.467) (0.513)* (0.509) (0.337)*** (0.160)
neighborhood
Above plus parental 0.767 0.983 –0.001 1.158 0.61
attitudes and behaviors (0.468) (0.524)* (0.508) (0.366)*** (0.186)
Above plus proxy for 0.756 0.842 –0.283 NA 0.041
unobserved individual (0.538) (0.544) (0.640) (0.159)
characteristics
Above, using IV 0.993 0.285 0.360 2.157 0.330
(1.217) (1.228) (1.446) (1.014)** (0.541)
Note: Robust standard errors estimated by bootstrapping are shown in parentheses. Ln(wage) parameters are
estimated by OLS; all others are estimated by logit.
NA = not applicable; SES = socioeconomic status.
Full results are reported in tables A.1 and A.2.
*p < 0.10. **p < 0.05. *** p < 0.01 (two-tailed tests).

HOUSING POLICY DEBATE
Parental Homeownership and Young Adult Outcomes 809
adult earnings (see table A.1) and employment status (not shown), thereby
giving us conf idence that it is indeed serving as a measure of unobserved
characteristics that deserve to be controlled in the outcome equations.
16
As
before, however, its inclusion in the model does not substantially alter the
results for parental homeownership and educational attainments.
Thus, even when an extensive set of controls is added, a statistically sig-
nif icant relationship persists between parental homeownership status and
children’s later educational and homeownership attainments. As for the eco-
nomic signif icance of these relationships, the college parameter indicates that
every additional year that children lived in a home owned by their parents is
associated with a 0.008 (or 5 percent from the mean) increase in the prob-
ability of obtaining a college degree. The relationship is much weaker in
the case of a high school diploma, however, with the added year associated
with only a 0.005 increase (0.005 percent of the mean) in the probability of
graduating. The ownership parameter is the largest, indicating that an added
year in a home owned by parents is associated with a 0.013 (2 percent of the
mean) increase in the probability that the person will become a homeowner
in early adulthood.
Effect of controlling for residential stability
What happens to the relationship between parental homeownership and
children’s outcomes when residential mobility is taken into account? We re-
ran all the foregoing models but included the proportion of childhood years
the young adult did not change residence as an additional explanatory vari-
able. Results for the parental homeownership parameters are presented in
the bottom panel of table 3; the full model results are shown in table A.1.
Residential stability during childhood is positively associated with the
probability of not having a child before age 18. The estimated coeff icient
indicates that every additional year during which a child did not move is
associated with a 0.008 (or 0.8 percent from the mean) increase in the prob-
ability of not having a child before adulthood. Once fertility is controlled
for, however, stability never proves statistically signif icantly related to edu-
cational attainments, tenure status, or earnings; see table A.2. This suggests
that the common observation of a direct relationship between residential
stability and a child’s educational performance (Hanushek, Kain, and Rivkin
2004; Pribesh and Downey 1999) may be largely due to the intervening teen
fertility relationship.
16
Results are available from the f irst author.

HOUSING POLICY DEBATE
810 George Galster, Dave E. Marcotte, Marvin B. Mandell, Hal Wolman, and Nancy Augustine
The inclusion of stability in the model dramatically alters the basic con-
clusions about the apparent effects of parental homeownership. By including
residential stability in the most completely controlled model, the coeff icient
of the parental homeownership variable is reduced by 9 percent in the high
school attainment model and 24 percent in the college degree model, and in
both models the coeff icients are no longer statistically signif icant (see the bot-
tom panel of table 3). These f indings suggest that the relationships between
parental homeownership status and educational outcomes for children are
mediated in a important way by residential mobility (paths E and F in f igure
1). This result is consistent with the conclusions reached by Aaronson (2000)
and Harkness and Newman (2002), although they measured parental home-
ownership only during one short period late in childhood, not cumulatively
as we have done. However, the relationship between parental and children’s
homeownership status is virtually unaffected by the addition of childhood
stability; in fact, the coeff icient grows slightly higher.
Housing tenure or housing wealth?
We investigated the extent to which the strong correlations between
parental homeownership and children’s later college attainments and tenure
status might have been due to the superior wealth of the parents. For no out-
comes did we f ind proxies that parental housing wealth was statistically sig-
nif icant (results not shown). This suggests that it is owner-occupancy tenure
itself, not the potential wealth associated with the dwelling, that is more pre-
dictive of children’s educational and tenure outcomes. Although Boehm and
Schlottman (1999) found home value positively related to college completion
rates, they did not control for residential stability, neighborhood characteris-
tics, or many other parental characteristics, as we do.
IV estimates of the relationship between parental homeownership status
and outcomes for young adults
F inally, we report a model wherein our key variable of interest—propor-
tion of childhood years spent in a home owned by parents—is replaced by
an IV produced by the aforementioned procedure. Results are displayed in
the last rows of the two panels of table 3, with runs where stability is both
controlled for and not controlled for; all other controls are entered.
Examination of these results indicates that IV point estimates are of
roughly comparable magnitudes to those obtained in the rows above, with
the exception of the homeownership outcome, where IV estimates are twice
as large. Unfortunately, they are estimated much less precisely; the standard

HOUSING POLICY DEBATE
Parental Homeownership and Young Adult Outcomes 811
errors are on the order of three times larger for the IV results. This is to be
expected, because IV estimation focuses only on the change in the potentially
endogenous variable that can be explained by exogenous factors employed
in the f irst stage. In our case, the relatively weak power of the instruments
leaves us with little variation to exploit. As a result, we cannot detect even
the sizable effects. Thus, our IV results should be treated as suggestive only.
Nonetheless, the relationship between parental homeownership and housing
tenure remains substantial and statistically signif icant. Indeed, the IV point
estimates of the association between parental homeownership and education,
fertility, and housing tenure are larger than the noninstrumented estimates.
In sum, we view our attempts to instrument by use of exogenous variables
in our structural model as appropriate but, ultimately, only partially success-
ful. We are not conf ident that we have removed all the selection effects. We
believe that progress in estimating more robust and precise IV estimates in
the current application will require additional information at the individual
level beyond what is available from the PSID.
Conclusions, caveats, and implications
An emerging empirical consensus
A nascent consensus on the relationship between parental homeowner-
ship and children’s educational outcomes is emerging from several studies
using different samples and analytical techniques. Green and White (1997)
conducted several analyses of different data sets and concluded that for the
household with an average income, owning rather than renting reduced the
chances that a 17-year-old in the household would drop out of school by 3 to
4 percentage points. Using a PSID sample of children who left their parents’
home from 1975 to 1982, Boehm and Schlottman (1999) found that spend-
ing the last seven years in a home the parents owned instead of rented was
associated with a 15 percentage point increase in completing high school, a
14.5 percentage point increase in completing some postsecondary education,
and a 27 percentage point increase in graduating from college.
Using an unusually wide array of controls for parental socioeconomic
characteristics, attitudes and behaviors, unobserved individual characteris-
tics, and neighborhood conditions, we f ind that parental homeownership
status cumulated over childhood remains a statistically signif icant predictor
of children’s educational attainments. Further, the magnitude of these rela-
tionships is not systematically eroded when we use IV estimation techniques.
However, the magnitude of the implied impact on high school completion is
very small, whereas the effect on college completion is large. Our smallest

HOUSING POLICY DEBATE
812 George Galster, Dave E. Marcotte, Marvin B. Mandell, Hal Wolman, and Nancy Augustine
parameter estimate over all of our model specif ications indicates that those
spending all of their childhood in a home owned by their parents would be
predicted to have a 0.19 higher probability of obtaining a college degree
than children with otherwise identical backgrounds whose parents always
rented. Given that the mean college completion rate in our sample was 0.14,
this predicted difference is large indeed. Nevertheless, some doubt remains
over whether these apparent impacts follow from homeownership itself or
from associated residential stability (Aaronson 2000; Newman and Hark-
ness 2002).
When we control for the proportion of years during childhood when the
child did not move, the magnitude of the college effect drops by roughly one-
fourth and barely falls below the minimal criterion for statistical signif icance.
Of course, to the extent that the choice of owning a home is simultaneously
a choice of longer tenure, to add residential stability is to overcontrol the
model. We therefore conclude that parental homeownership is related to the
probability that a child will eventually graduate from college to a statistically
and economically signif icant extent.
These conclusions comport nicely with the earlier work of Haurin, Par-
cel, and Haurin (2002a, 2002b), who found that parental homeownership
for the six-year period during which the children matured from 5 to 8 to
11 to 14 years old was associated with test scores that were 7 to 10 percent
higher, both directly and indirectly through their relationship with various
indexes of cognitive and emotional functioning and with lack of behavioral
problems. Boyle (2002) also identif ied an association between parental home-
ownership and improved emotional and behavioral outcomes for children.
Logically, it is a short step to suggest that such improvements in achievement
test scores, emotional states, and behaviors are related to staying in school
longer and eventually graduating from college.
In addition, we have explored the relationship between growing up in
a home owned by parents and the likelihood that children will own their
own homes by young adulthood. Unlike the case with education, this strong
relationship was not mediated by childhood residential stability. Again, using
the smallest of the point estimates our various models obtained, we calculate
that a person who spent an entire childhood in a home owned by parents
had a 0.09 (or 16 percent from the mean) higher probability of owning a
home in 1999 than an otherwise identical person who never lived in a home
owned by parents. This result can be interpreted in several ways that are not
mutually exclusive.
On the one hand, it may indicate that children raised by homeowners
come to uncritically view this tenure as normative (Helderman and Mul-

HOUSING POLICY DEBATE
Parental Homeownership and Young Adult Outcomes 813
der 2007) or perhaps more consciously perceive the value of control over
environment and mobility and the potential for generating wealth associ-
ated with homeownership. On the other hand, many of the same positive
cognitive and behavioral outcomes associated with homeownership noted
here may lead children to save and plan for the future in a distinctive way
that promotes this tenure for them. Moreover, they may acquire from their
parents certain skills and experiences that are associated with buying and
maintaining a home and reduce the informational and psychological costs of
homeownership (Green and White 1997).
Of course, our study has identif ied statistical associations, not proven
causal links. However, we have tried to purge the measured association of
the common confounding biases—from omitted variables, selection, and
endogeneity—in a fashion that we believe offers important methodological
advances. Further, we have noted several hypotheses that are not mutually
exclusive and that offer plausible causal mechanisms to explain how owning
one’s home might provide an independent enhancement to the environment
in which children are raised, thereby producing several favorable outcomes
during young adulthood.
Caveats and directions for future research
Although we believe that our study offers many advances in the measure-
ment of homeownership effects, there are several areas in which our work
falls short and implicitly points to where future research would be especially
productive. F irst, our measure of homeownership using the childhood aver-
age experience, although a reasonable measure of cumulative impacts, is not
the only measure that might be used. An alternative could be that home-
ownership has different effects depending on the developmental stage of the
child. Unfortunately, when we tried such a specif ication, our results were not
robust enough to report. Second, we treated only one aspect of the neigh-
borhood—poverty rate—as endogenous, and clearly the neighborhood is a
more complex, multidimensional bundle of attributes than that. Future work
could productively explore the extent to which other aspects such as racial-
ethnic composition or homeownership rates might contribute to children’s
well-being in ways other than neighborhood poverty. Third, while we believe
that an IV strategy is called for here, further effort in developing more pow-
erful instruments is warranted, with a goal of estimating the full structural
equation system of endogenous neighborhood, mobility, and tenure choices.
Fourth, our analysis investigated parental homeownership effects from 1968
through 1992, an era predating the numerous reforms and innovations in the

HOUSING POLICY DEBATE
814 George Galster, Dave E. Marcotte, Marvin B. Mandell, Hal Wolman, and Nancy Augustine
mortgage market that we have since witnessed. There is no assurance that
results from this earlier era apply equally today, especially with regard to
lower-income households that have been able to attain homeownership only
because of these changes. F inally, our approach does not permit us to look
inside the “black box” of homeownership to discern the specif ic causal pro-
cesses at work. Qualitative work that probes more deeply into uncovering
the differential environments, attitudes, and behaviors between homeowner
and renter households is thus called for.
Should policy makers care about expanding homeownership
opportunities?
For many decades, federal housing policy has encouraged owner-occu-
pancy over rental tenure (Shlay 2006). There have been two, non–mutually
exclusive, categories of rationales for this encouragement. The f irst is that
homeownership is essentially a “merit good,” one that has certain intrinsic,
private benef its that all members of society deserve a chance to consume
on the grounds of distributive justice. This rationale considers such reputed
benef its as increased wealth, social status, security of tenure, control over
dwelling, pride, and life satisfaction (McCarthy, Van Zandt, and Rohe 2001;
Rohe, McCarthy, and Van Zandt 2000;
The second category is that homeownership conveys external benef its—
or positive externalities—on the greater society over and above the benef its
accrued by the homeowners themselves. The externalities typically cited
include enhanced home maintenance, social and political participation, and
attachment to community (McCarthy, Van Zandt, and Rohe 2001; Rohe,
McCarthy, and Van Zandt 2000). On the basis of our study and the earlier
literature we have cited, the benef its to children—future workers and citi-
zens—who live in homes owned by their parents can be added to this list.
Thus, we believe that there is a substantial and growing body of research
that provides justif ication for some public sector intervention to enhance
opportunities for homeownership. Indeed, federal, state, and local govern-
ments and nonprof it and philanthropic organizations have been engaged
for many years in a dizzying array of initiatives designed to lower barri-
ers to attaining and sustaining homeownership, especially for lower-income
households. (See Collins 2004, Cortes et al. 2006, Herbert and Belsky
2006, Herbert et al. 2005, Lubell 2005, and Retsinas and Belsky 2002 for
thorough reviews.) Our position should not be interpreted, however, as a
blanket endorsement of the present mélange of instruments aimed at this
goal or of the position that homeownership is best for every household in
every circumstance.

HOUSING POLICY DEBATE
Parental Homeownership and Young Adult Outcomes 815
On the contrary, our reading of the literature (for details, see Galster
and Santiago 2007) leads us to the conclusion that the advantages of home-
ownership are highly contingent on changes in the local housing, mortgage,
and labor markets and on the unforeseen familial challenges that may befall
homeowners after their purchase. To the extent that these changes force
more vulnerable households to leave homeownership prematurely (especially
through default), substantial economic penalties ensue. Thus, the rationale
for public support of expanding homeownership must be weighed soberly
against the associated risks of putting vulnerable households into untenable
homeownership situations.
Appendix A
Exogenous and predetermined variables [x] from equations (1) through
(5) used in the f irst stage of instrumentation procedure
1. Index of owner-occupied housing prices in the metropolitan area (lag 1
year)
2. Index of owner-occupied housing prices in the metropolitan area (lead 1
year)
3. Index of gross rents paid by renter occupants in the metropolitan area
(lag 1 year)
4. Item 3 times renter status in the preceding year
5. Home mortgage interest rate for a 30-year f ixed-rate loan
6. Item 5 times renter status in the preceding year
7. Ratio of costs of renting to owning in the metropolitan area (lag 1 year)
8. Item 7 times renter status in the preceding year
9. Whether the family’s oldest child reached age 5 in the preceding year (1
= yes; 0 = no)
10. Whether the family’s oldest child reached age 13 in the preceding year (1
= yes; 0 = no)
11. Whether any other child in the family reached age 5 in the preceding year
(1 = yes; 0 = no)
12. Whether any other child in the family reached age 13 in the preceding
year (1 = yes; 0 = no)

HOUSING POLICY DEBATE
816 George Galster, Dave E. Marcotte, Marvin B. Mandell, Hal Wolman, and Nancy Augustine
13. Age of the household head
14. Whether the household head received a lump-sum monetary payment
since the child’s birth; (e.g., inheritance; 1 = yes; 0 = no)
15. Item 14 times renter status in the preceding year
16. Difference in the household’s real income from the preceding to the cur-
rent year (if greater than 0; 0 otherwise)
17. Item 16 times renter status in the preceding year
18. Poverty rate of the county (lag 1 year)
19. Homeownership rate of the county (lag 1 year)
20. Whether the household expects to move next year (lag 1 year)
21. Whether the household owns the home it occupies (lag 1 year)
22. Logarithm of def lated household income (lag 1 year)
23. Year (denoted by a set of dummy variables, 1968 = excluded year)
Table A.1. Estimated Parameters for the Full Model (Excluding Stability)
Young Adult Outcomes
High School College No Child ln (wage)
Explanatory Variables Own Home Diploma or Higher Graduate before Age 18 1999
blackfem –0.06 0.823 –0.13 –1.577 0.199
(0.351) (0.661) (0.879) (0.716)** (0.353)
blackmale 0.232 1.703 –0.957 0.492 0.519
(0.386) (0.797)** (0.801) (0.973) (0.342)
whitefem –0.28 0.218 –0.439 –1.727 –0.431
(0.208) (0.383) (0.283) (0.498)*** (0.098)***
birthorder –0.102 0.059 0.009 0.041 –0.084
(0.059)* (0.098) (0.106) (0.108) (0.043)**
age99 0.044 0.145 0.41 –0.193 0.081
(0.042) (0.078)* (0.065)*** (0.094)** (0.025)***
pro_live_in_poverty0to18 0.534 –1.804 1.216 –1.033 –1.375
(0.606) (1.118) (1.668) (1.189) (0.474)***
pro_livew_2_parents0to18 1.126 0.163 –1.753 0.692 1.066
(0.409)*** (0.858) (0.810)** (0.958) (0.286)***
pro_parent_own0to18 1.071 0.56 1.396 –0.286 –0.151
(0.316)*** (0.622) (0.616)** (0.809) (0.183)
(pro_live_in_poverty0to18) –0.003 –0.064 –0.04 –0.085 –0.023
(0.016) (0.029)** (0.032) (0.036)** (0.013)*
religion 0.441 –0.713 –0.545 0.104 –0.038
(0.340) (0.532) (0.584) (0.745) (0.167)
largecity 0.029 0.46 0.989 0.719 0.01
(0.202) (0.463) (0.308)*** (0.464) (0.122)
farm –0.069 –0.164 0.423 0.424 –0.028
(0.237) (0.462) (0.421) (0.514) (0.138)
veteran 0.204 –0.905 0.238 –0.7 0.023
(0.189) (0.466)* (0.295) (0.411)* (0.113)

HOUSING POLICY DEBATE
Parental Homeownership and Young Adult Outcomes 817
momteen –0.709 0.83 –0.618 –0.48 –0.548
(0.347)** (0.509) (0.873) (0.682) (0.332)*
married 1.644 N/A N/A N/A 0.104
(0.179)*** N/A N/A N/A (0.091)
at_least_hs 0.468 N/A N/A N/A 0.179
(0.293) N/A N/A N/A (0.155)
collgrad 0.456 N/A N/A N/A 0.435
(0.249)* N/A N/A N/A (0.117)***
no_child_before_18 0.218 1.235 –0.056 N/A 0.062
(0.304) (0.461)*** (0.560) N/A (0.240)
ave_education_head0to18 –0.102 0.648 0.343 0.19 –0.029
(0.081) (0.286)** (0.105)*** (0.194) (0.048)
ave_hdocc_pre0to18 0.007 0.031 0.005 0.038 –0.009
(0.011) (0.034) (0.017) (0.023) (0.007)
ave_self_employed0to18 –0.488 –0.356 0.547 –0.038 0.189
(0.428) (1.353) (0.550) (0.999) (0.266)
ave_employed_wife0to18 –0.79 –0.622 0.683 0.146 –0.462
(0.360)** (0.721) (0.526) (0.804) (0.200)**
ave_smsa0to18 0.231 –0.336 0.961 –0.089 –0.023
(0.273) (0.510) (0.483)** (0.555) (0.147)
ave_annu_hrs_wkd0to18 0 0 0 0 0
(0.000) (0.000) (0.000) (0.001) (0.000)
ave_readnewspaper –0.794 0.735 0.13 –0.446 0.21
(0.305)*** (0.570) (0.541) (0.597) (0.191)
ave_union –0.521 1.025 0.017 –0.717 0.075
(0.292)* (0.513)** (0.464) (0.582) (0.149)
ave_nochurch 0.467 0.391 –0.674 0.625 0.136
(0.325) (0.660) (0.579) (0.711) (0.201)
ave_no_socialclubs –0.256 –1.234 0.891 0.063 0.051
(0.288) (0.664)* (0.405)** (0.652) (0.134)
ave_relatives 0.273 –0.008 0.099 –0.11 –0.07
(0.255) (0.540) (0.374) (0.527) (0.135)
ave_num_neigh_known 0.009 0.055 0.043 0.024 0.012
(0.013) (0.032)* (0.019)** (0.032) (0.007)*
ave_plan_ahead 0.137 –0.273 0.639 1.312 0.306
(0.268) (0.491) (0.461) (0.542)** (0.155)**
ave_trust 0.63 –1.107 –0.008 –0.591 –0.133
(0.285)** (0.525)** (0.438) (0.699) (0.123)
wage1000s 0 N/A N/A N/A N/A
(0.001)
proxy_ unobs. child charact –0.594 –0.396 0.064 0.407 N/A
N/A (0.363) (0.348) (0.420) (0.103)***
annu_hrs_wkd99 N/A N/A N/A N/A 0
N/A N/A N/A V (0.000)**
Constant –3.179 –10.39 –19.68 5.021 9.328
(1.797)* (4.131)** (2.670)*** (4.135) (1.151)***
Observations 775 755 755 755 541
Wald Chi–squared (d.f.) 217.1 (33) 194.8 (30) 96.4 (30) 104.7 (29) NA
Pseudo R
2
/R
2
0.21 0.26 0.19 0.29 0.34
Note: All parameters areestimated by logit except for the natural log (wage) model, which was estimated by OLS.
Robust standard errors are in parentheses.
*p < 0.10. **p < 0.05. *** p < 0.01 (two–tailed tests).
Table A.1. Estimated Parameters for the Full Model (Excluding Stability) Continued
Young Adult Outcomes
High School College No Child ln(wage)
Explanatory Variables Own Home Diploma or Higher Graduate before Age 18 1999

HOUSING POLICY DEBATE
818 George Galster, Dave E. Marcotte, Marvin B. Mandell, Hal Wolman, and Nancy Augustine
blackfem –0.041 0.861 –0.132 –1.516 0.196
(0.352) (0.674) (0.877) (0.699)** (0.349)
blackmale 0.245 1.713 –0.954 0.784 0.516
(0.386) (0.816)** (0.803) (0.903) (0.336)
whitefem –0.29 0.194 –0.437 –1.724 –0.424
(0.208) (0.377) (0.282) (0.514)*** (0.097)***
birthorder –0.091 0.075 0.006 –0.015 –0.091
(0.060) (0.101) (0.102) (0.101) (0.044)**
age99 0.045 0.148 0.41 –0.21 0.081
(0.042) (0.078)* (0.066)*** (0.095)** (0.025)***
pro_live_in_poverty0to18 0.527 –1.769 1.222 –1.395 –1.397
(0.609) (1.098) (1.659) (1.233) (0.473)***
pro_livew_2_parents0to18 1.138 0.155 –1.756 0.664 1.059
(0.410)*** (0.842) (0.809)** (0.939) (0.282)***
pro_parents_own0to18 1.227 0.832 1.354 –1.293 –0.26
(0.367)*** (0.675) (0.741)* (0.912) (0.236)
pro_stability0to18 –0.528 –0.804 0.149 2.802 0.33
(0.621) (1.040) (1.254) (1.332)** (0.375)
ave_pert_inc_below_pov0to18 –0.001 –0.062 –0.041 –0.101 –0.023
(0.016) (0.029)** (0.032) (0.035)*** (0.013)*
religion 0.478 –0.701 –0.558 0.067 –0.057
(0.343) (0.534) (0.562) (0.715) (0.168)
largecity 0.027 0.471 0.989 0.655 0.011
(0.202) (0.464) (0.308)*** (0.452) (0.122)
farm –0.073 –0.174 0.423 0.426 –0.023
(0.238) (0.461) (0.421) (0.518) (0.138)
veteran 0.2 –0.95 0.241 –0.599 0.031
(0.189) (0.466)** (0.294) (0.439) (0.115)
momteen –0.715 0.872 –0.617 –0.63 –0.559
(0.347)** (0.520)* (0.873) (0.636) (0.331)*
married 1.645 N/A N/A N/A 0.103
(0.179)*** N/A N/A N/A (0.092)
at_least_hs 0.464 N/A N/A N/A 0.179
(0.293) N/A N/A N/A (0.155)
collgrad 0.464 N/A N/A N/A 0.436
(0.249)* N/A N/A N/A (0.116)***
no_child_before_18 0.246 1.269 –0.063 N/A 0.053
(0.305) (0.471)*** (0.571) N/A (0.234)
Eave_education_head0to18 –0.099 0.659 0.343 0.199 –0.031
(0.081) (0.285)** (0.106)*** (0.212) (0.048)
ave_hdocc_pre0to18 0.008 0.03 0.005 0.032 –0.01
(0.011) (0.034) (0.017) (0.026) (0.007)
ave_self_employed0to18 –0.481 –0.289 0.54 –0.042 0.185
(0.428) (1.367) (0.562) (0.977) (0.267)
ave_employed_wife0to18 –0.796 –0.639 0.685 0.275 –0.451
(0.360)** (0.717) (0.526) (0.812) (0.203)**
ave_smsa0to18 0.247 –0.334 0.95 –0.115 –0.037
(0.274) (0.507) (0.496)* (0.583) (0.145)
ave_annu_hrs_wkd0to18 0 0 0 0 0
(0.000) (0.000) (0.000) (0.001) (0.000)
ave_readnewspaper –0.771 0.801 0.125 –0.662 0.197
(0.307)** (0.587) (0.540) (0.615) (0.190)
ave_union –0.497 1.108 0.011 –0.99 0.051
(0.293)* (0.517)** (0.461) (0.618) (0.150)
Table A.2. Estimated Parameters for the Full Model (Including Stability)
Young Adult Outcomes
High School College No Child ln (wage)
Explanatory Variables Own Home Diploma or Higher Graduate before Age 18 1999

HOUSING POLICY DEBATE
Parental Homeownership and Young Adult Outcomes 819
ave_nochurch 0.476 0.377 –0.682 0.618 0.125
(0.325) (0.647) (0.578) (0.693) (0.201)
ave_no_socialclubs –0.269 –1.273 0.894 0.112 0.064
(0.288) (0.659)* (0.404)** (0.645) (0.134)
ave_relatives 0.292 0.001 0.098 –0.147 –0.071
(0.257) (0.545) (0.373) (0.542) (0.135)
ave_num_neigh_known 0.009 0.054 0.043 0.02 0.012
(0.013) (0.032)* (0.019)** (0.031) (0.007)*
ave_plan_ahead 0.138 –0.26 0.642 1.351 0.306
(0.269) (0.489) (0.460) (0.569)** (0.155)**
ave_trust 0.641 –1.122 –0.011 –0.542 –0.145
(0.286)** (0.531)** (0.439) (0.693) (0.122)
wage1000s 0 N/A N/A N/A N/A
(0.001) N/A N/A N/A N/A
residown N/A –0.602 –0.407 0.045 0.417
N/A (0.361)* (0.354) (0.435) (0.104)***
annu_hrs_wkd99 N/A N/A N/A N/A 0
N/A N/A N/A N/A (0.000)**
Constant –3.097 –10.308 –19.715 5.008 9.279
(1.801)* (4.050)** (2.697)*** (4.123) (1.161)***
Observations 775 755 755 755 541
Wald chi–squared (d.f.) 217.1 (33) 194.8 (30) 96.4 (30) 104.7 (29) NA
Pseudo R
2
/R
2
0.21 0.26 0.19 0.29 0.34
Note: All parameters are estimated by logit except in(wage) model, which was estimated by OLS. Robust
standard errors are in parentheses.
*p < 0.10. **p < 0.05. *** p < 0.01 (two-tailed tests).
Table A.2. Estimated Parameters for the Full Model (Including Stability) Continued
Young Adult Outcomes
High School College No Child ln (wage)
Explanatory Variables Own Home Diploma or Higher Graduate before Age 18 1999
Authors
George Galster is the Hilberry Professor of Urban Affairs in the Depart-
ment of Geography and Urban Planning at Wayne State University. Dave E.
Marcotte is an associate professor and Marvin B. Mandell is a professor in
the Department of Public Policy at the University of Maryland at Baltimore
County. Hal Wolman is a professor and Nancy Augustine is an assistant pro-
fessor in the Public Policy Program at The George Washington University.
This research was supported by a grant from the Ford Foundation. The
opinions expressed are those of the authors and do not necessarily ref lect
those of the boards of trustees of the Ford Foundation or of our respec-
tive universities. We acknowledge the helpful suggestions made by Katherine
Kiel, Stuart Rosenthal, and anonymous referees on an earlier draft. Richard
Ban and Caitlin Malloy provided excellent production assistance.

HOUSING POLICY DEBATE
820 George Galster, Dave E. Marcotte, Marvin B. Mandell, Hal Wolman, and Nancy Augustine
References
Aaronson, Daniel. 2000. A Note on the Benef its of Homeownership. Journal of Urban
Economics 47:356–69.
Austin, Mark, and Yoko Baba. 1990. Social Determinants of Neighborhood Attachments.
Sociological Spectrum 10:59–78.
Balfour, Danny, and Janet Smith. 1996. Transforming Lease-Purchase Programs for Low-
Income Families. Journal of Urban Affairs 18:173–88.
Barber, Jennifer. 2001. The Intergenerational Transmission of Age at F irst Birth among
Married and Unmarried Men and Women. Social Science Research 30:219–47.
Barker, David, and Eric Miller. 2005. Homeownership and Child Welfare. Paper pre-
sented at the midyear meeting of the American Real Estate and Urban Economics Associa-
tion, Washington, DC. May.
Boehm, Thomas. 1981. Tenure Choice and Expected Mobility: A Synthesis. Journal of
Urban Economics 10:375–89.
Boehm, Thomas, and Alan Schlottman. 1999. Does Home Ownership by Parents Have an
Economic Impact on Their Children? Journal of Housing Economics 8:217–32.
Boyle, Michael. 2002. Home Ownership and the Emotional and Behavioral Problems of
Children and Youth. Child Development 73:883–92.
Brisson, Daniel, and Charles Usher. 2007. The Effects of Informal Neighborhood Bond-
ing Social Capital and Neighborhood Context on Homeownership for Families Living in
Poverty. Journal of Urban Affairs 29:65–75.
Brooks-Gunn, Jeanne, Greg J. Duncan, and J. Lawrence Aber, eds. 1997. Neighbor-
hood Poverty. Vol. 1. Context and Consequences for Children. New York: Russell Sage
Foundation.
Buerkle, Karla. 1997. Mobile Children and Families: Qualitative and Quantitative Explo-
rations of the Meaning and Impact of Residential Mobility and School Changes. Disserta-
tion Abstracts International, Humanities and Social Sciences Section A 58(6A):2067.
Cairney, John, and Michael Boyle. 2004. Home Ownership, Mortgages, and Psychologi-
cal Distress. Housing Studies 19:161–74.
Clark, Rebecca. 1992. Neighborhood Effects of Dropping Out of School among Teenage
Boys. Working paper. Urban Institute.
Coleman, James. 1988. Social Capital and the Creation of Human Capital. American
Journal of Sociology 94:S95–S120.
Coleman, James. 1990. Foundations of Social Theory. Cambridge, MA: Harvard Univer-
sity Press.
Collins, J. Michael. 2004. An Assessment of Policies to Promote F inancial Savings and
Asset Building in the U.S. Washington, DC: PolicyLab Consulting Group, LLC.

HOUSING POLICY DEBATE
Parental Homeownership and Young Adult Outcomes 821
Corcoran, Mary, Roger Gordon, Deborah Laren, and Gary Solon. 1992. The Association
between Men’s Economic Status and Their Family and Community Origins. Journal of
Human Resources 27:575–601.
Cortes, Alvaro, Christopher E. Herbert, Erin Wilson, and Elizabeth Clay. 2006.
Improving Homeownership Opportunities for Hispanic Families: A Review of the Litera-
ture. Washington, DC: U.S. Department of Housing and Urban Development, Off ice of
Policy Development and Research.
Cox, Kevin. 1982. Housing Tenure and Neighborhood Activism. Urban Affairs Quar-
terly 18:107–29.
Deng, Yonheng, Steven Ross, and Susan Wachter. 2003. Racial Differences in Home-
ownership: The Effect of Residential Location. Regional Science and Urban Economics
33:517–56.
DeWit, David. 1998. Frequent Childhood Geographic Relocation: Its Impact on Drug
Use Initiation and the Development of Alcohol and other Drug-Related Problems among
Adolescents and Young Adults. Addictive Behaviors 23:623–34.
Dietz, Robert. 2002. The Estimation of Neighborhood Effects in the Social Sciences.
Social Science Research 31:539–75.
Dietz, Robert, and Donald Haurin. 2003. The Social and Private Micro-Level Conse-
quences of Homeownership. Journal of Urban Economics 54:401–50.
Ding, Chengri, and Gerrit-Jan Knaap. 2002. Property Values in Inner-City Neighbor-
hoods: The Effects of Homeownership, Housing Investment, and Economic Develop-
ment. Housing Policy Debate 13(4):701–27.
DiPasquale, Denise, and Edward Glaeser. 1999. Incentives and Social Capital: Are Hom-
eowners Better Citizens? Journal of Urban Economics 45:354–84.
Duncan, Greg J., James Connell, and Pamela Klebanov. 1997. Conceptual and Method-
ological Issues in Estimating Causal Effects of Neighborhoods and Family Conditions on
Individual Development. In Neighborhood Poverty. Vol. 1. Context and Consequences
for Children, ed. Jeanne Brooks-Gunn, Greg J. Duncan, and J. Lawrence Aber, 219–50.
New York: Russell Sage Foundation.
Duncan, Greg J., and Stephen W. Raudenbush. 1999. Assessing the Effect of Context in
Studies of Child and Youth Development. Educational Psychology 34:29–41.
Earls, Felton, and Mary Carlson. 2001. The Social Ecology of Child Health and Well-
Being. Annual Review of Public Health 22:143–66.
Ellen, Ingrid Gould, and Margery Austin Turner. 2003. Do Neighborhoods Matter and
Why? In Choosing a Better Life? Evaluating the Moving to Opportunity Experiment, ed.
John Goering and Judith D. Feins, 313–38. Washington, DC: Urban Institute Press.
Evans, William, Wallace Oates, and Robert Schwab. 1992. Measuring Peer Group Effects:
A Study of Teenage Behavior. Journal of Political Economy 100:966–91.

HOUSING POLICY DEBATE
822 George Galster, Dave E. Marcotte, Marvin B. Mandell, Hal Wolman, and Nancy Augustine
Foster, E. Michael, and Sara McLanahan. 1996. An Illustration of the Use of Instrumental
Variables: Do Neighborhood Conditions Affect a Young Person’s Chance of F inishing
High School? Psychological Methods 1:249–60.
Freeman, Richard. 1991. Crime and the Employment of Disadvantaged Youths. Working
Paper No. w3875. National Bureau of Economic Research.
Furstenberg, Frank Jr., Thomas Cook, Jacquelynne Eccles, Glen Elder Jr., and Arnold
Sameroff. 1999. Managing to Make It: Urban Families and Adolescent Success. Chicago:
University of Chicago Press.
Furstenberg, Frank Jr., Judith Levine, and Jeanne Brooks-Gunn. 1990. The Children of
Teenage Mothers: Patterns of Early Childbearing in Two Generations. Family Planning
Perspectives 22:54–61.
Galster, George. 1983. Empirical Evidence on Cross-Tenure Differences in Home Mainte-
nance and Conditions. Land Economics 59:107–13.
Galster, George. 1987. Homeowners and Neighborhood Reinvestment. Durham, NC:
Duke University Press.
Galster, George. 2003. Investigating Behavioral Impacts of Poor Neighborhoods: Toward
New Data and Analytical Strategies. Housing Studies 18:893–914.
Galster, George. 2005. Neighborhood Mix, Social Opportunities, and the Policy Chal-
lenges of an Increasingly Diverse Amsterdam. Amsterdam, the Netherlands: University of
Amsterdam, Department of Geography, Planning, and International Development Studies.
World Wide Web page <http://www.fmg.uva.nl/amidst/object.cfm/objectid=7C149E7C-
EC9F-4C2E-91DB7485C0839425> (accessed April 10, 2007).
Galster, George, and Anna Santiago. 2007. Low-Income Homeownership as an Asset-
Building Tool: What Can We Tell Policymakers? Paper presented at the Brookings Institu-
tion’s Conference on Urban and Regional Policy Effects, Washington, DC. March 29–30.
GeoLytics. 2007. Neighborhood Change Database. World Wide Web page <http://www.geo-
lytics.com/USCensus,Neighborhood-Change-Database-1970-2000,Data,Order,Products.
asp> (accessed September 24).
Ginther, David, Robert Haveman, and Barbara Wolfe. 2000. Neighborhood Attributes
as Determinants of Children’s Outcomes: How Robust Are the Relationships? Journal of
Human Resources 35:603–42.
Gleason, Philip, and Thomas Vartanian. 1999. Do Neighborhood Conditions Affect High
School Dropout and College Graduation Rates? Journal of Socio-Economics 28:21–41.
Gottschalk, Peter. 1996. Is the Correlation in Welfare Participation across Generations
Spurious? Journal of Public Economics 63:1–25.
Gottschalk, Peter, Sara McLanahan, and Gary Sandefur. 1994. The Dynamics and Inter-
generational Transmission of Poverty and Welfare Participation. In Confronting Poverty,
ed. Sheldon Danziger, Gary Sandefur, and Daniel Weinberg, 85–108. Cambridge, MA:
Harvard University Press.

HOUSING POLICY DEBATE
Parental Homeownership and Young Adult Outcomes 823
Green, Richard, and Michelle White. 1997. Measuring the Benef its of Homeowning:
Effects on Children. Journal of Urban Economics 41(3):441–61.
Grogger, Jeffrey. 1997. Incarceration-Related Costs of Early Childbearing. In Kids Having
Kids: Economic Costs and Social Consequences of Teen Pregnancy, ed. Rebecca May-
nard, 95–143. Washington, DC: Urban Institute Press.
Gyourko, Joseph, Peter Linneman, and Susan Wachter. 1997. Tenure Choice and Intra-
Urban Location: The Roles of Race and Wealth. Working paper. University of Pennsylva-
nia, Wharton Real Estate Center.
Hanushek, Eric, John Kain, and Steven Rivkin. 2004. Disruption vs. Tiebout Improvement:
The Costs and Benef its of Switching Schools. Journal of Public Economics 88:1721–46.
Harkness, Joseph, and Sandra J. Newman. 2002. Homeownership for the Poor in Dis-
tressed Neighborhoods: Does This Make Sense? Housing Policy Debate 13(3):597-630.
Harkness, Joseph, and Sandra Newman. 2003. Differential Effects of Homeownership
on Children from Higher- and Lower-Income Families. Journal of Housing Research
14:1–19.
Haurin, Donald, Robert Dietz, and Bruce Weinberg. 2003. The Impact of Neighborhood
Homeownership Rates: A Review of the Theoretical and Empirical Literature. Journal of
Housing Research 13:119–51.
Haurin, Donald, Patric Hendershott, and David Ling. 1988. Homeownership Rates of
Married Couples: An Econometric Investigation. Housing F inance Review 7:85–108.
Haurin, Donald, Toby Parcel, and R. Jean Haurin. 2002a. Does Home Ownership Affect
Child Outcomes? Real Estate Economics 30:635–66.
Haurin, Donald, Toby Parcel, and R. Jean Haurin. 2002b. Impact of Home Ownership
on Child Outcome. In Low Income Homeownership: Examining the Unexamined Goal,
ed. Eric S. Belsky and Nicolas P. Retsinas, 427–46. Washington, DC: Brookings Institu-
tion Press.
Haurin, R. Jean. 1992. Patterns of Childhood Residence and the Relationship to Young
Adult Outcomes. Journal of Marriage and the Family 54:846–80.
Haveman, Robert, and Barbara Wolfe. 1994. Succeeding Generations: On the Effects of
Investments in Children. New York: Russell Sage Foundation.
Helderman, Amanda, and Clara Mulder. 2007. Intergenerational Transmission of Hom-
eownership. The Roles of Gifts and Continuities in Housing Market Characteristics.
Urban Studies 44:231–47.
Herbert, Christopher, and Eric Belsky. 2006. The Homeownership Experience of Low-
Income and Minority Families. Washington, DC: U.S. Department of Housing and Urban
Development.

HOUSING POLICY DEBATE
824 George Galster, Dave E. Marcotte, Marvin B. Mandell, Hal Wolman, and Nancy Augustine
Herbert, Christopher, Donald Haurin, Stuart Rosenthal, and Mark Duda. 2005. Hom-
eownership Gaps among Low-Income and Minority Borrowers and Neighborhoods.
Washington, DC: U.S. Department of Housing and Urban Development.
Hoff, Karla, and Arijit Sen. 2005. Homeownership, Community Interactions, and Segre-
gation. American Economic Review 95:1167–89.
Huff ines, Margaret. 2003. A Study of the Impact of Mobility on Student Social Adjust-
ment. Dissertation Abstracts International Section A: Humanities and Social Sciences
64(4–A):1184.
Hunter, Albert. 1975. The Loss of Community: An Empirical Test through Replication.
American Sociological Review 40:537–51.
Ioannides, Yannis. 1987. Residential Mobility and Housing Tenure Choice. Regional Sci-
ence and Urban Economies 17(2):265–87.
Ioannides, Yannis, and Kamhon Kan. 1996. Structural Estimation of Residential Mobility
and Housing Tenure Choice. Journal of Regional Science 36:335–63.
Jeffers, Leo, and Jean Dobos. 1984. Communication and Neighborhood Mobilization.
Urban Affairs Quarterly 20:97–112.
Jencks, Christopher, and Susan Mayer. 1990. The Social Consequences of Growing Up in
a Poor Neighborhood. In Inner-City Poverty in the United States, ed. Lawrence Lynn and
Michael McGeary, 111–86. Washington, DC: National Academy Press.
Kan, Kamhon. 2000. Dynamic Modeling of Housing Tenure Choice. Journal of Urban
Economics 48:46–69.
Kauppinen, Timo. 2004. Neighbourhood Effects in a European City: The Educational
Careers of Young People in Helsinki. Paper presented at the meeting of the European
Network for Housing Research, Cambridge, England. July.
Lee, Barrett, R. S. Oropesa, and James Kanan. 1994. Neighborhood Context and Resi-
dential Mobility. Demography 31:249–70.
Leventhal, Tama, and Jeanne Brooks-Gunn. 2000. The Neighborhoods They Live In.
Psychological Bulletin 126:309–37.
Lubell, Jeffrey. 2005. Strengthening the Ladder for Sustainable Homeownership. Wash-
ington, DC: National Housing Conference and Annie E. Casey Foundation.
Manski, Charles. 1995. Identif ication Problems in the Social Sciences. Cambridge, MA:
Harvard University Press.
Manski, Charles. 2000. Economic Analysis of Social Interactions. Journal of Economic
Perspectives 14:115–36.
Mayer, Neil. 1981. Rehabilitation Decisions in Rental Housing. Journal of Urban Eco-
nomics 10:76–94.

HOUSING POLICY DEBATE
Parental Homeownership and Young Adult Outcomes 825
Mayer, Susan. 1997. What Money Can’t Buy: Family Income and Children’s Life Chances.
Cambridge, MA: Harvard University Press.
McCarthy, George, Shannon Van Zandt, and William M. Rohe. 2001. The Economic
Benef its and Costs of Homeownership. Working Paper No. 01–02. Research Institute for
Housing America.
McLanahan, Sara, and Larry Bumpass. 1988. Intergenerational Consequences of Family
Disruption. American Journal of Sociology 94:130–52.
Moff itt, Robert. 1992. Incentive Effects of the U.S. Welfare System: A Review. Journal of
Economic Literature 30:1–61.
Murray, Michael. 2006. Avoiding Invalid Instruments and Coping with Weak Instru-
ments. Journal of Economic Perspectives 20:111–32.
Musterd, Sako, Roger Andersson, George Galster, and Timo Kauppinen. 2008. Are Co-
Ethnic Clusters Good or Bad? Exploring the Relationships between Individuals’ Incomes
and the Immigrant Character of their Neighborhood. Environment and Planning A,
forthcoming.
Painter, Gary, Stuart Gabriel, and Dowell Myers. 2001. Paths to Homeownership: An
Analysis of the Residential Location and Homeownership Choices of Black Households
in Los Angeles. Working Paper No. 01–03. Research Institute on Housing America.
Parcel, Toby, and Elizabeth Menaghan. 1994a. Early Parental Work, Family, Social Capi-
tal, and Early Childhood Outcomes. American Journal of Sociology 99:972–1009.
Parcel, Toby, and Elizabeth Menaghan. 1994b. Parents’ Jobs and Children’s Lives. New
York: Aldine de Gruyter.
Payne, Joan. 1987. Does Unemployment Run in Families? Some F indings from the Gen-
eral Household Survey. Sociology 21:199–214.
Peeples, Faith, and Rolf Loeber. 1994. Do Individual Factors and Neighborhood Context
Explain Ethnic Differences in Juvenile Delinquency? Journal of Quantitative Criminology
10:141–57.
Pepper, John. 2000. The Intergenerational Transmission of Welfare Receipt: A Nonpara-
metric Bounds Analysis. Review of Economics and Statistics 82:472–88.
Potter, Lloyd, Marcie-Jo Kresnow, Kenneth Powell, Thomas Simon, James Mercy, Roberta
Lee, Ralph Frankowski, Alan Swann, Timothy Bayer, and Patrick Carroll. 2001. The
Influence of Geographic Mobility on Nearly Lethal Suicide Attempts. Suicide and Life-
Threatening Behavior 32(suppl):42–48.
Pribesh, Shana, and Douglas Downey. 1999. Why Are Residential and School Moves
Associated with Poor School Performance? Demography 36:521–34.
Retsinas, Nicolas P., and Eric S. Belsky, eds. 2002. Low-Income Homeownership: Exam-
ining the Unexamined Goal. Washington, DC: Brookings Institution Press.

HOUSING POLICY DEBATE
826 George Galster, Dave E. Marcotte, Marvin B. Mandell, Hal Wolman, and Nancy Augustine
Robert, Stephanie. 1999. Socioeconomic Position and Health: The Independent Contribu-
tion of Community Socioeconomic Context. Annual Review of Sociology 25:489–516.
Rohe, William, George McCarthy, and Shannon Van Zandt. 2000. The Social Benef its
and Costs of Homeownership. Working Paper No. 00–01. Research Institute for Housing
America.
Rohe, William M., and Michael A. Stegman. 1994. The Impact of Home Ownership
on the Social and Political Involvement of Low-Income People. Urban Affairs Quarterly
30:152–72.
Rohe, William M., and Leslie S. Stewart. 1996. Homeownership and Neighborhood Sta-
bility. Housing Policy Debate 7(1):37–81.
Rossi, Peter H., and Eleanor Weber. 1996. The Social Benef its of Homeownership: Empir-
ical Evidence from National Surveys. Housing Policy Debate 7(1):1–35.
Sampson, Robert, Jeffrey Morenoff, and Felton Earls. 1999. Beyond Social Capital: Spatial
Dynamics of Collective Eff icacy for Children. American Sociological Review 64:633–60.
Sampson, Robert, Jeffrey Morenoff, and Thomas Gannon-Rowley. 2002. Assessing
“Neighborhood Effects”: Social Processes and New Directions in Research. Annual
Review of Sociology 28:443–78.
Sawhill, Isabel, and Laura Chadwick. 1999. Children in Cities: Uncertain Futures. Survey
Series Working Paper. Brookings Institution, Center on Urban and Metropolitan Policy.
Shlay, Anne. 2006. Low-Income Homeownership: American Dream or Delusion? Urban
Studies 43:511–31.
Stack, Steven. 1994. The Effect of Geographic Mobility on Premarital Sex. Journal of
Marriage and the Family 56:204–08.
Sullivan, Mercer. 1989. Getting Paid: Youth Crime and Work in the Inner City. Ithaca,
NY: Cornell University Press.
Temple, Judy, and Arthur Reynolds. 1999. School Mobility and Achievement: Longitudi-
nal F indings from an Urban Cohort. Journal of School Psychology 37:355–77.
Tienda, Marta. 1991. Poor People and Poor Places: Deciphering Neighborhood Effects of
Poverty Outcomes. In Macro-Micro Linkages in Sociology, ed. J. Haber, 244–62. New-
bury Park, CA: Sage.
U.S. Bureau of the Census. 1972a. 1970 Census of Housing. Metropolitan Housing Char-
acteristics United States and Regions HC(2)–1. Washington, DC: U.S. Government Print-
ing Off ice.
U.S. Bureau of the Census. 1972b. 1970 Census of Population. General Social and Eco-
nomic Characteristics United States Summary PC(1)–C1. Washington, DC: U.S. Govern-
ment Printing Off ice.

HOUSING POLICY DEBATE
Parental Homeownership and Young Adult Outcomes 827
U.S. Bureau of the Census. 1983. 1980 Census of Population. Vol. 1. Characteristics
of the Population, Chapter A: Number of Inhabitants, Part 1 United States Summary
PC80–1–A1. Washington, DC: U.S. Government Printing Off ice.
U.S. Bureau of the Census. 1984. 1980 Census of Housing. Vol. 2. Metropolitan Housing
Characteristics, United States Summary HC80–2–1. Washington, DC: U.S. Government
Printing Off ice.
U.S. Bureau of the Census. 1992. 1990 Census of Population and Housing. Summary
Social, Economic, and Housing Characteristics CPH–5–1, Washington, DC: U.S. Govern-
ment Printing Off ice.
U.S. Bureau of the Census. 1993. 1990 Census of Housing. Detailed Housing Character-
istics, Section 1 of 3 Metropolitan Areas CH–2–1B, Washington, DC: U.S. Government
Printing Off ice.
Vartanian, Thomas. 1999. Childhood Conditions and Adult Welfare Use: Examining
Neighborhood and Family Factors. Journal of Marriage and the Family 61:225–37.
Verba, Sidney, Lehman Schlozman, and Henry Brady. 1995. Voice and Equality: Civic
Voluntarism in American Politics. Cambridge, MA: Harvard University Press.
Wooldridge, Jeffrey. 2002. Econometric Analysis of Cross Section and Panel Data. Cam-
bridge, MA: MIT Press.

HOUSING POLICY DEBATE
828 George Galster, Dave E. Marcotte, Marvin B. Mandell, Hal Wolman, and Nancy Augustine