Marital Trends Affecting Inequality over Time

1
Marital Trends Affecting Inequality over Time

• Christy Spivey (SIUE)
• cspivey_at_siue.edu
• and
• Ahsanuzzaman (Virginia Tech)

2
Newsweek, June 2006 Rethinking the Marriage
Crunch
3
Newsweek, June 2006 Rethinking the Marriage
Crunch

• Twenty Years Ago
• A single 30 year-old woman had a 20 chance of
  marrying, and that dropped to 5 by age 35 and
  2.6 by age 40 (Bennett, Bloom, Craig)
• Now
• 90 of baby boomers will marry at some time
  (Steven Martin, UMd)
• A single 40 year-old woman has greater than a 41
  chance of marrying

4
Newsweek, June 2006 Rethinking the Marriage
Crunch

• Women werent remaining unmarried because
  marriage was less appealing, but because it was
  becoming more appealing to wait. Steven
  Martin, UMd
• Breakdown of 50s cookie cutter life course,
  birth control, technology advancements allowed
  less household specialization
• Mens attitudes about marriage have changed?

5
Three Trends Affecting Inequality

• Are the more educated now more likely to get
  married?
• Are the more educated now more likely to marry
  someone like themselves (assortative mating)?
• Do the more educated have better marital
  quality and thus stay married longer?

6
of Adults who are College-Educated and Married
7
Education and the Likelihood of Marriage

• Marriage can confer economic, child-rearing
  advantages, as well as the opportunity to take
  advantage of economies of scale, production of
  household public goods, specialization, risk
  sharing, joint consumption
• Just another reason the less-educated might lose
  in a globalized economy.

8
Education and the Likelihood of Marriage

• We use U.S. Census (1960-2000) and American
  Community Survey (2006) data
• We consider all household heads and any spouse
  who are not currently enrolled in school
• We estimate probit equations for ever being
  married and multinomial logit equations for the
  various marital status categories

9
Education and the Likelihood of Marriage Probit
Results

• In 1960, everyone was less likely to have ever
  been married, compared to those with less than a
  high school degree
• High school grads and those with some college are
  now more likely to have ever been married, but
  this is not the case for college graduates.
  Similar results hold by gender.
• However, while college graduates are still less
  likely to have ever been married, the magnitude
  of the effect has shrunk over time

10
Education and the Likelihood of Marriage
11
Education and the Likelihood of Marriage
12
Education and the Likelihood of Marriage
Multinomial Logit Results

• Everyone is now more likely to be currently
  married, compared to those with less than a high
  school degree.
• Only college grads are still more likely to be
  single, compared to those with less than a high
  school degree. However, they are much less
  likely to be separated, divorced, or widowed.

13
Education and the Likelihood of Marriage
14
Education and the Likelihood of Marriage
15
Education and the Likelihood of Marriage
16
Education and the Likelihood of Marriage
17
Assortative Mating

• Whether assortative mating is positive or
  negative is an empirical question and depends
  upon the source of the gains from marriage,
  preferences, and supply and demand
• Negative assortative mating might occur when
  gains to marriage arise through specialization,
  but positive might occur if gains occur through
  production of household public goods (Lam, 1998).
  As the role of specialization has declined,
  might expect positive assortative mating over
  time.

18
Assortative Mating Over Time

• Increased supply of educated women harder to
  find an equally educated husband, or will
  financial freedom expand options (and result in a
  shifting balance of power within the household)?
• Have preferences/incentives for men to choose an
  educated spouse increased?
• Quite a few studies, but not all agree that
  educational homogamy has increased since the
  1960s
• Others (Mare, 1991) have argued that as the time
  between leaving school and marriage has
  increased, educational homogamy would decrease.

19
Assortative Mating Over Time

• Hou and Myles, 2007, Statistics Canada working
  paper
• The tendency of like to marry like has
  unambiguously risen since 1970s in both US and
  Canada
• The result of declining intermarriage at both
  ends of the educational spectrum
• In the US, the rise in homogamy was partially
  offset by an increased tendency over time for
  women to marry down the educational hierarchy

20
Assortative Mating

• We calculate absolute rates of homogamy for
  existing marriages for all adults
• This stock measure reflects the combined effects
  of assortative mating into first marriage,
  divorce, entry into additional marriages, and any
  tendency for spouses to grow alike in educational
  attainment after marriage

21
Assortative Mating

• Since 1970, the percentage of households
  resembling each other in terms of educational
  status has increased
• 1960 49.09
• 1970 47.08
• 1980 48
• 1990 48.54
• 2000 50.59
• 2006 52.05
• Not an immediately obvious trend since there is
  more variation in educational attainment now

22
Assortative Mating

• This overall trend obscures more subtle trends by
  educational category
• The of households in which both spouses have
  less than a high school degree decreases from
  1960 to 2006.
• The of households in which both spouses have a
  high school degree increases and then decreases
  again.
• The of households in which both spouses have
  more than a high school degree increases over
  time.

23
Assortative Mating 1960
24
Assortative Mating 2006
25
Assortative Mating 1970
26
Assortative Mating 1980
27
Assortative Mating 1990
28
Assortative Mating 2000
29
Education and Length of Marriage

• If more educated have better quality match and
  stay married longer, implications go beyond
  current inequality
• If loving, two-parent households are more common
  among the more educated, this confers additional
  benefits to their children, enhancing future
  inequality.

30
Education and Length of Marriage

• We use the NLSY79 panel data set that has some
  marital quality measures for women starting in
  1992
• Frequency of Arguments about chores and
  responsibilities, etc.
• Often (1), Sometimes (2), Hardly ever (3), or
  Never (4)?
• Frequency tell each other about day, etc.
• Almost Every Day (1), Once or Twice a Week (2),
  Once or Twice a Month (3), Less Than Once a Month
  (4)
• DEGREE OF HAPPINESS WITH MARRIAGE/RELATIONSHIP
  Would you say that your (relationship/marriage)
  is...
• Very happy (1), Fairly happy (2), Not too happy
  (3)

31
Education and Length of Marriage
32
Education and Length of Marriage
33
Education and Length of Marriage
34
Education and Length of Marriage
35
Inequality

• Wed like to know how these three trends affect
  inequality over time.
• We construct a data set using the Census, along
  with the Current Population Survey to fill in the
  missing years from 1960-2006.

36
Inequality
37
Inequality
38
Inequality

• Problems
• Reverse causation, inequality affects who and
  when we marry?
• Need to control for changes in the distribution
  of education over time
• Need a separate measure for the likelihood of
  ever being married, and the likelihood of divorce

39
Conclusions

• The most educated are not more likely to have
  ever been married
• However, they may be more likely to be married,
  which reflects other trends such as longer
  marriages
• Absolute assortative mating with respect to
  education is also on the rise and is positively
  associated with inequality