What Explains the Decline in First Marriage in the United States? Evidence from the Panel Study of Income Dynamics, 1969 to 2013
| Author | Daniel Schneider,Matthew Stimpson,Kristen Harknett |
| DOI | http://doi.org/10.1111/jomf.12481 |
| Date | 01 August 2018 |
| Published date | 01 August 2018 |
D S University of California Berkeley
K H University of California San Francisco∗
M S University of California Berkeley∗∗
What Explains the Decline in First Marriage in the
United States? Evidence from the Panel Study of
Income Dynamics, 1969 to 2013
Rates of entry into rst marriage have declined
sharply in the United States during the past half
century, and there is evidence of broad gaps in
marriage entry by race and education. Although
a large literatureexplores the inuences on mar-
riage for single cohorts, there is little research
that tests explanations for this decline across
multiple cohorts. The authors use individual
and contextual measures of employment and
incarceration to predict transitions to rst mar-
riage in the Panel Study of Income Dynam-
ics (1969–2013). They test two prominent the-
ories of why marriage rates have declined: the
decreased availability of “marriageable” men
and the increased economic standing of women.
They nd that men’sreduced economic prospects
and increased risk of incarceration contributed
to the decline in rst marriage rates during the
past 45 years in the United States, although these
basic measures of economic and carceral condi-
tions cannot explain the entire decline.
The institution of marriage has changed
dramatically during the past half-century in
Department of Sociology, Universityof California
Berkeley, 480 BarrowsHall, Berkeley, CA 94720
(djschneider@berkeley.edu).
∗University of California San Francisco.
∗∗University of California Berkeley.
Key Words: employment, incarceration, marriage, social
demography.
the United States. Americans now marry later
and less often than at any time in the past
100 years (Ruggles, 2015). Furthermore, there
are pronounced divides in marriage entry such
that African American men and women are less
likely to marry than their White counterparts
(Cherlin, 1998; Tucker & Mitchell-Kernan,
1995), and those with less education are less
likely to marry than their more highly edu-
cated counterparts (Ellwood & Jencks, 2004;
McLanahan, 2004). Demographers, sociolo-
gists, and economists have offered a number of
well-known explanations for this demographic
shift. Notably, these explanations include the
decline in men’s labor market position, the
increasingly long reach of mass incarceration,
and the rise in women’slabor force participation.
However,there is a conspicuous divide within
this research on marital change. Demographic
descriptions of the “retreat from marriage”
use census or vital statistics data from several
decades to show an increase in the proportion
of unmarried adults and a rise in the age at
marriage (Lundberg, Pollak, & Stearns, 2016;
McLanahan, 2004; Pew Research Center, 2010).
However, studies that seek to test explanations
for this decline almost exclusively use data
from single cohorts to estimate the effects of
economic and social factors on marriage entry
(e.g., Apel, 2016; Clarkberg, 1999; McClendon,
Kuo, & Raley,2014). There is very little research
that puts these two pieces together to actually
Journal of Marriage and Family 80 (August 2018): 791–811 791
DOI:10.1111/jomf.12481
792 Journal of Marriage and Family
examine the extent to which broad changes in
marriage entry over time are explained by fac-
tors such as reductions in men’s and increases
in women’s labor force activity and increases in
incarceration.
We analyze data from the Panel Survey
of Income Dynamics (PSID) for the years
1969 to 2013. The PSID (https://psidonline.
isr.umich.edu/) is unique in capturing
individual-level longitudinal data on marriage
entry, labor force activity, and incarcera-
tion experiences for multiple cohorts of men
and women. We rst show time trends in
age-adjusted rst marriage rates for men and
women overall and by race and educational
attainment. Wethen examine the extent to which
labor force factors and incarceration explain
these time trends in rst marriage entry for men
and women by race and education group.
C F M A
Data from Monthly Vital Statistics and the
American Community Survey show that rates
of rst marriage declined steadily from 1970
to 2008. In 1970, the rate of rst marriages
was about 95 per 1,000 never-married women,
whereas in 2008 this rate had dropped to 42
per 1,000 (Payne, 2010). An analysis of marital
timing nds that this “retreat from marriage”
represents a delay in marrying more so than an
eschewing of marriage altogether (Goldstein &
Kenney, 2001). Nevertheless, these declines in
marriage rates and delays in marital entry add up
to a decline in marriage prevalence, especially
in early adulthood. For instance, census and
American Community Survey data show that
the percent of adults aged 18 and older who are
married declined from 72% in 1960 to 52% in
2008 (Pew Research Center, 2010).
Alongside the reports of overall declines in
marriage over time, the period between 1960
and the 2000s witnessed a wide divergence in
marriage prevalence across education groups
(Ellwood & Jencks, 2004; McLanahan, 2004)
and across race groups. In 1960, the percent of
those currently married among college-educated
Americans (76%) and those with a high school
degree or less education (69%) differed by seven
percentage points, but that gap had widened to
16 percentage points (64% vs. 48%) by 2008
(Pew Research Center, 2010). Analyses of mar-
ital timing also nd a similar pattern of class
divergence in the projected proportion who will
eventually marry (Goldstein & Kenney, 2001).
The racial gap in marriage also widened dur-
ing this same period, but not as dramatically. In
1960, the gap between Whites (74%) and Blacks
(61%) was 13 percentage points compared with
24 percentage points (56% vs. 32%) in 2008
(Pew Research Center, 2010).
T M C
What precipitated these sweeping changes in
marriage in America over the past half century?
American social and economic life changed
dramatically during this period, and increases
in school enrollment (Buchmann & DiPrete,
2006), life expectancy (Bulled & Sosis, 2010),
and contraceptive access (McLanahan, 2004;
Stevenson & Wolfers, 2007)—to name only
a few factors—may have inuenced the tim-
ing and extent of marriage entry. However,
scholarship has overwhelmingly focused on the
presence (or absence) of “marriageable” men
and the possibility of economically independent
women who lack the incentive to marry as
explanations for this change.
Marriageable Men
The idea that men are expected to possess some
threshold level of economic resources to be
normatively marriageable is longstanding in
demography (i.e., Davis & Blake, 1956). Wilson
and Neckerman (1987) famously built on that
accepted insight to argue that such men became
relatively scarcer with stagnant wage growth,
deindustrialization, and rising incarceration dur-
ing the latter half of the 20th century, trends that
disproportionately affected African American
men. By the time of Wilson and Neckerman’s
writing in the 1980s, these social and economic
changes were already apparent. Since that time,
there is only stronger evidence that men’s labor
force position has declined while their exposure
to incarceration has increased. This has been
especially true for African American men as
well as for men without a college degree. For
instance, while men at all points in the income
distribution experienced strong wage growth
up until 1970, wage growth has essentially
been at, or even negative, since then for men
in the bottom 40% of earners (Autor, Katz,
& Kearney, 2008), and the male labor force
participation rate has declined steadily during
this time period, especially for men without a
college degree (Aaronson et al., 2014).
To continue reading
Request your trial