Early and late motherhood: socioeconomic antecedents and consequences

1
Early and late motherhood socio-economic
antecedents and consequences
Sub-brand to go here

• Heather Joshi
• Conference on Economics and Demography
• NIESR November 5 2008

CLS is an ESRC Resource Centre based at the
Institute of Education
2
Outline

• Economic and social significance of the timing of
  first birth
• Aggregate trends in routine statistics
• Disaggregate Trends
• Determinants of Timing
• Consequence of Timing for children

3
1.The timing of childbearing

• The age at which childbearing starts is key
  driver of contemporary demographic change. No
  second births can occur before the first and in
  UK they usually do follow.
• While progression to the first and second birth
  have been relatively stable, the change in the
  age of initiation has brought about fluctuations
  in the period fertility rate tempo being more
  volatile that quantum and an important
  determinant in fluctuations in the age structure
  of children.
• Delay also tends to reduce eventual family size
  and the rate of population growth.
• Delayed motherhood is not socially neutral, but
  affects different social groups.
• The fact that it had not involved everyone is
  reflected in UKs still relatively high teenage
  fertility, which in turn has kept British
  fertility above the European average.

4
2. Trends in the First Birth

• Motherhood becoming later and (somewhat) less
  frequent
• women born 1900
  lt70
• born 1940
  ca 90
• born 1960
  ca 80
• But since the 1970 generation, motherhood is
  starting later.
• Increasing childlessness probably not all
  intended

5
A period perspective on postponement
Average age at first birth within marriage and
all outside marriage, EW 1958-2004
6
Births outside marriage, mostly to younger women
7
Entry to motherhood in post-war Britain, by
cohort year of womans own birth true birth
order
8
Age by which successive quintiles of the whole
cohort had entered motherhood
9
3 Data disaggregated by socio-economic
characteristics

• Bringing in information not recorded in the
  routine statistics
• Rendall et al, Population Studies 2005 -
  General Household Survey
• Rendall et al 2008 ONS LS
• Jenkins et al 2008 NCSD and BCS70
• Joshi, Hawkes and Ward 2004 mothers of
  Millennium Cohort
• Also recommended
• Ermisch and Pevalin ( 2005) 1970 cohort

10
Cohort differences in first birth probability,
EW, low vs medium education (Rendall et al 2005)
11
Trends by pre-birth occupation

• Rendall et al, 2008 comparing England and
  France in the two census longitudinal studies
  with occupational data hypothesize and find
  increasing polarization in age at first birth by
  pre-childbearing occupation between the 1980s and
  1990s in the U.K. but not in France.
• Early first childbearing persisted in the U.K.
  only among women in low-skill occupations, while
  shifts towards increasingly late first births
  occurred in clerical/secretarial occupations and
  above.
• Increases in age at first birth occurred across
  all occupations in France, but this was still
  much earlier on average than for all but
  low-skill British mothers.

12
4. Determinants of Delay

• Is education the driver of delayed motherhood?or
  just a proxy for something else?
• Those with human capital to accumulate, and
  facing higher opportunity costs of earnings
  interruption have more motive than others to
  delay. What else might motivate the
  differentiated behaviour?
• Andrew Jenkins, Heather Joshi Mark
  Killingsworth
• Educational Attainment, Labour Market
  Conditions and the Timing of First and
  Higher-Order Births in Britain
• CLS working paper imminent
• Uses data women from 1958 and 1970 Birth
  Cohorts
• Event history analysis of birth histories from
  age 16 and evidence on childhood circumstances

13
Hazard modelling with unobservered heterogeneity
hj, conditional hazard, for the j th birth (
1st up to 4th)
hj, conditional hazard, for the j th birth (
1st up to 4th) tj is the length of the jth
spell Z is a vector of covariates ? is a
person-specific unobserved heterogeneity
component ?, ß and f are transition-specific
parameters to be estimated
Heckman, J. J., Singer, B. (1984). A Method for
Minimizing the Impact of Distributional
Assumptions in Econometric Models for Duration
Data. Econometrica, 52, 271-320.
14
Hazard Regression for First Births in Two Cohorts
15
Why more early childbearing by the unskilled in
England vs France?

• The outcome of policy regime?
• Rendall et al ( 2008) analysis of births in 70s-
  90s suggest
• The targeted ( means tested) benefit system and
  lack of support for dual earner families tended
  to sustain, if not encourage early (often
  un-partnered) births in England and Wales
• The universal provisions in France, and support
  for dual earner families reduced the need for
  others to delay into their thirties.
• Higher rates of staying on in education/ training
  in France.
• This explanation may no longer apply after the
  Regime Change since late 1990s

16
Correlates of age at first birth among mothers of
Millennium Cohort Study (MCS)

• Mothers of MCS, themselves born between ca 1960
  and 1986, mostly also around 1970, around half
  had first child in the survey, born 2000-1,
  others had their first child at various points
  over the 1980s and 1990s.
• 29 of mothers entered motherhood 21 and 21 at
  31
• Early entry to motherhood associated with
• leaving school at minimum age
• low/ no qualifications
• family disruption in childhood
• Asian ethnicity, esp Bangladeshi
  and Pakistani
• unpartnered at time of 9 month
  survey
• not employed before or after birth
• living in disadvantaged area, on
  benefits, low income, etc
• partner low education and unemployment
• Converse applies to those who started in their 30s

17
Employment of Mothers with a child aged 3
18
5. Consequences for children?

• The low income of the families of young mothers
  clearly has some origins in the disadvantages
  which preceded parenthood.
• If relatively poor economic prospects meant there
  was little point in delaying, it is these
  prospects which help account for the predicament
  of poor young families.
• MCS enables us to assess how these disadvantages
  in the adult economy may have impinged on
  children
• At age 3 and 5 the children of the youngest
  mothers had on average worse scores on cognitive
  and behavioural assessments
• This is partly accounted for by the lower
  education of these mothers see analysis of age 3
  data below, but not entirely.
• There may be additional independent disadvantages
  to being born to a young mother

19
(No Transcript)
20
Conclusions on Early motherhood

• Catching up on entry to motherhood is less
  differentiated than avoiding an early start
  advantages of postponement tail off
• Both antecedent and current variables help
  explain differential outcomes by age at
  motherhood disadvantage facing families of young
  mothers not just the result of birth timing, but
  is compounded by it
• Early motherhood plays a part in an
  intergenerational transmission of disadvantage
• The bulge in the fertility schedule which helps
  keep UK TFR high comes with some problems.

21
Policy Implications

• Results suggest that Teenage Pregnancy Strategy
  (halve pregnancy rate under 18) should be aimed
  at economic and employment opportunities as well
  as sex education
• Flexible employment and education and child-care
  opportunities are needed for the younger mothers
  who have not been taking them.
• Young age at first motherhood say up to 25
  should be regarded as a marker for other problems
  and vulnerabilities, but it is important to guard
  against stigmatization compounding the challenges
  of early motherhood.
• Winwin for mother, and child generation and the
  public at large, if strategy succeeds in
  preventing the compounding of disadvantage in
  young families.

22
Any questions?

• Data and Documentation for NCDS, BCS70 and MCS
  now available from UK Data Archive
• www.data-archive.ac.uk
• Documentation available from CLS website
• http//www.cls.ioe.ac.uk/
• Guides to Initial findings (MCS1/MCS2/MCS3),
  working papers, searchable bibliography of
  published research based on the studies is all
  available from CLS website
• http//www.cls.ioe.ac.uk/

Institute of Education University of London 20
Bedford Way London WC1H 0AL Tel [phone]
6000 Fax [phone] Email
info_at_ioe.ac.uk Web www.ioe.ac.uk