You are on page 1of 11

ECOS3027: Economics of the Family - Lectures Notes 8

Fertility II

Marian Vidal-Fernandez

Semester 2 2017

Key background readings **; required reading *


• **Hotz, V. J., J. Klerman, and R. Willis (1997): The Economics of Fertility in Developed
Countries: A Survey, Handbook of Population and Family Economics, Section 4, http:
//www.sciencedirect.com/science/article/pii/S1574003X97800244

• **BFW7 pp.209-211, OR BFW5 pp.194-196

• *Jane Waldfogel (1998): “Understanding the ‘Family Gap’ in Pay for Women with Chil-
dren”, The Journal of Economic Perspectives, Vol. 12, No. 1 (Winter, 1998), pp. 137-156

• **Korenman, Sanders, and David Neumark. ”Marriage, Motherhood, and Wages.” Jour-
nal of Human Resources (1992): 233-255.

• Lundborg P., Plug, E. and Rasmussen A.W., (2014) Fertility Effects on Female Labor
Supply: Evidence from IVF Treatments. IZA Discussion Paper No. 8609.

• Correll, Shelley J., Stephen Benard, and In Paik. “Getting a Job: Is There a Motherhood
Penalty? 1.” American Journal of Sociology 112, no. 5 (2007): 1297-1339.

• **Moschion, 2013, The Impact of Fertility of Mothers Labour Supply In Australia: Evi-
dence from Exogenous Variation in Family Size, The Economic Record, 89(286), pp.319-
338

• Livermore, Tanya, Joan Rodgers, and Peter Siminski. “The Effect of Motherhood on
Wages and Wage Growth: Evidence for Australia.” Economic Record 87, no. s1 (2011):
80-91.

• *Miller, Amalia R. “The effects of motherhood timing on career path.” Journal of Pop-
ulation Economics 24, no. 3 (2011): 1071-1100.

• *Goldin, Claudia and Larry Katz (1999), ”The Power of the Pill: Oral Contraceptives
and Women’s Career and Marriage Decisions”

• Bailey, Martha J. Fifty years of family planning: New evidence on the long-run effects of
increasing access to contraception. No. w19493. National Bureau of Economic Research,
2013.

• Ashcraft, A. and Lang, K. The Consequences of Teenage Childbearing. NBER Working


Papers 12485, 2006.

1
• *Kearney, Melissa S., and Phillip B. Levine. “Why is the Teen Birth Rate in the United
States So High and Why Does It Matter?.” The Journal of Economic Perspectives 26,
no. 2 (2012): 141-166.
• Kearney, Melissa S., and Phillip B. Levine. Media Influences on Social Outcomes: The
Impact of MTVs 16 and Pregnant on Teen Childbearing. No. w19795. National Bureau
of Economic Research, 2014.
• Buckles, K. and Munnich, E. Birth Spacing and Sibling Outcomes. The Journal of Human
Resources. 47(3). Summer 2012.

1 Fertility and labour supply choices


• Factors that influence the decision to have children and time allocation decisions: beyond
a static approach:

2 Features of dynamic lifecycle models of fertility


• The quality-quantity model of fertility discussed last week had a number of limitations,
not least a lack of empirical support
• One important feature is that it takes a static approach to the fertility decision
• If we want to understand changes in the Total Fertility Rate, we need to consider the
timing and spacing of births, not just completed fertility
• The interaction of fertility decisions with marriage/partnering and labour supply decisions
also requires a dynamic approach
• Dynamic lifecycle models of fertility choice are usually complicated with no closed form
solution: we will outline the approach and discuss some implications, but solving these
models is outside the scope of this course

2
• A typical preference structure for parents:
t
X
U= β t u(ct , lt , st )
t=0

where ct is parental consumption, lt is the woman’s leisure consumption and st is the flow
of utility derived from children (determined by the number of children, their ages, the
woman’s time investment in children in that period, and resources devoted to consumption
goods for children

– In most models, only the woman’s leisure/labour supply choice is considered: men
are assumed to supply labour inelastically
– There are many different ways in which the utility function is framed: some disregard
lt , others assume consumption of children only occurs at the end of the lifecycle and
is based only on the number of children...

• Usually have period-by-period budget constraints for the mother’s time, eg. lt +ht +tct = 1

• Financial budget constraint may allow borrowing and saving; may be credit constrained

• There may or may not be uncertainty over wages and fertility processes

• Maternal human capital investment may also be modelled: wages responding to education
and experience

• Assume certainty:

– Transitory changes in income or wages may affect the timing of births but are unlikely
to affect the total number of births, as their impact on lifetime resources is small
– Overall, the opportunity cost of children determines the best time to have children:
expect fertility to occur at low cost times
– Where the woman’s human capital depreciation rate is higher and she has positive
initial earnings potential, the costs of children are higher and so childbearing is likely
to be postponed
– Poor initial opportunities in the labour market encourage earlier childbearing
– Since having children reduces parental consumption goods and the couple cannot
borrow against future income, the optimal childbearing time is when the father’s
income is at its highest point (lowest marginal utility of parental consumption)
– If children bring immediate utility, this incentivises earlier childbearing: traded off
against the opportunity cost of human capital investment or leisure
– There is also a precautionary incentive to postpone the first birth if contraception
is not perfectly effective or costless

• What about birth spacing? Why wait between children?

– A rising income trajectory and imperfect capital markets may explain this
– Variations in the relative prices of children over time, generated by variations in the
mother’s wage (avoiding excessive depreciation) could also affect this
– Maybe spacing affects outcomes of children and that is relevant for their quality
(Buckles, Munnich, 2012)

3
• We can also incorporate contraceptive choice into this framework: trading off the costs
and benefits of various methods

Key point: these models are not straightforward, and the assumptions made drive
the conclusions drawn regarding optimal timing and spacing of childbearing.

3 Fertility, Labour Supply and Wages: Empirical evidence


• The relationship between fertility, labour supply and wages is not clearly predicted from
lifecycle models of fertility
– An increase in a woman’s wage/labour market opportunities will have income and
substitution effects
– Different assumptions on wage depreciation will affect predictions from these models
– A low wage may cause higher fertility due to a lower opportunity cost of home
production
– But at the same time, a lower wage may reflect depreciation resulting from time out
of the labour market for home production
– The direction of any causality, if any correlation is found, is not clear
– Moreover, a lack of correlation may mask opposing causal effects
• The Family Wage Gap – ie. the wage gap between women with children and without
children, was bigger in the US in 1991 than the gender wage gap
• In Australia, a first child is associated with a 5% reduction in wages, and an additional
child with a further 4.4% decrease that is unexplained by education, experience and
occupation (Livermore et al 2011)
• Do a woman’s wages fall when she has children? Potential mechanisms:
– Less labour market experience (time out of labour force; time in part time work)
– Tenure interruptions reduce the gains from firm-specific training
– Larger responsibility for home production/childcare
– Discrimination – statistical? Audit studies in the US suggest this is the case (see
Correll et al 2007)
– Selection
– Mitigated by institutional arrangements? (we will come back to this)
• Korenman and Neumark 1992 – Marriage, Motherhood and Wages
– Data from the United States for women aged 28-38 – observations in 1980 and 1982
– Attempt to estimate the direct effect of marriage and fertility on wages, not the in-
direct effect through reduced human capital accumulation, so controls for education,
experience and tenure are included
– Raw averages: never-married women earn most, married with spouse present least.
Wages decline with number of children. Wage growth is not that different across
categories
– Control for experience and tenure: very little association between children, marriage
and wages

4
– Sources of bias in OLS regressions:
∗ Endogeneity: marriage, fertility, experience and tenure treated as endogenous
variables; instrumented with family background measures
· Experience and tenure found to be endogenous; instrumenting these suggests
that having two or more children is associated with lower wages
· Valid instruments?
∗ Fixed effects to eliminated unobserved time-invariant characteristics
· Children not significant determinants of wages: “women with wage-enhancing
characteristics (net of observables) appear less likely to have (two or more)
children
∗ Employment selectivity bias: only observe wages for those who are working in
both time periods: does this matter?
∗ Estimates broadly unchanged
– Overall: having one child doesn’t appear to lower wages, but having two or more
does, even when controlling for experience and tenure (but experience and tenure
appear endogenous so including them as controls may not be appropriate)

• Lundborg et al. 2014: Use Denmark’s in-vitro fertilisation (IVF) registry to compare
women who are succesful in their first attempt to conceive versus counterparts conditional
on education and experience. They find long-lasting impact on wages of first child. Maybe
Korenman and Neumark 241
differences in generosity of parental leave policies between U.S. and Danish results.
Table 2
Wage Equation Estimates for White Working Women, 1982 Ordinary
Least Squares (dependent variable: natural logarithm of hourly
earnings)a

(1) (2) (3) (4)

Married, spouse present -.02 .01 .02 .05


(.05) (.04) (.04) (.04)
Divorced or separated -.00 .05 .05 .10
(.05) (.05) (.05) (.04)
One child -.13 -.05 -.05 -.04
(.04) (.04) (.04) (.04)
Two+ children -.30 -.18 -.18 -.07
(.03) (.03) (.03) (.03)
Education - .06 .06 .07
(.01) (.01) (.01)
South -- -.05 - .05
(.02) (.02)
Urban - .15 .15
(.03) (.02)
Experience - -- .02
(.004)
Tenure - - - .03
(.003)
F-testb .00 .00 .00 .03
R2 .10 .19 .23 .33

a. There are 1,207 observations.Standarderrorsare reportedin parentheses.Sample


weights were not used in computingestimates. Observationsare includedonly if the
wage reportedis for a job at which the respondentis currentlyworking.Never married
and no childrenare the referencecategories. Single-yearage dummyvariablesare in-
cluded in all specifications.
b. P-value for joint test of significanceof maritalstatus and fertilityvariables.

Source: with
results are consistent Korenman andof Neumark
the findings many of the1992
studies reviewed
in Section II: after controlling for experience and tenure, marriage and
children have relatively little association with wages.9'10

9. We explored the sensitivity of the coefficientsreportedin Column(4) of Table 2 to the


inclusionof controlsfor, in turn,years marriedandyears divorcedor separated;the number
of preschool-agechildren;and age of the mother at first birth. The results do not differ
qualitativelyfrom those in Table 2: addingyears marriedand years divorcedor separated

5
Source: Lundborg et al, 2014

• Later age at first birth correlated with better career outcomes – why? Miller
(2011)

– Could be due to anticipated earnings determining fertility, outside influences (eg.


ambition) driving both outcomes, or early fertility impacting on earnings and career
– Use miscarriage, conceptions when contraception in use and delays in birth after
stated contraception cessation as plausibly exogenous variation in age at first birth
– A year of delay associated with 9% increase in career earnings, 6% increase in work
experience and 3% increase in wage rates
– Larger returns to women with more education
– Impact of early childbearing on level and growth of wages – suggestion of a ‘mommy
track’ (this may be voluntary or externally imposed)

6
The effects of motherhood timing on career path 1083

Table 2 Effects of motherhood timing on career outcomes

(1) (2) (3) (4) (5) (6) (7)


OLS IV IV OLS IV OLS IV
Dependent variable Ln(earnings) Ln(wage rate) Ln(hours)
A1B 0.0963a 0.0878a 0.0610a 0.0319a 0.0306a 0.0644a 0.0572a
[0.0098] [0.020] [0.022] [0.0052] [0.010] [0.0070] [0.014]
Birth cohort −0.0658a −0.0623a −0.0533a −0.0259a −0.0254a −0.0399a −0.0370a
[0.015] [0.016] [0.016] [0.0078] [0.0086] [0.011] [0.012]
HS diploma 1.123a 1.127a 1.156a 0.237a 0.238a 0.886a 0.889a
[0.16] [0.16] [0.15] [0.083] [0.082] [0.11] [0.11]
College or higher 1.258a 1.276a 1.308a 0.379a 0.382a 0.879a 0.894a
[0.17] [0.17] [0.17] [0.089] [0.091] [0.12] [0.12]
Black 0.180b 0.174b 0.154c 0.0735 0.0726 0.106c 0.102
[0.087] [0.087] [0.086] [0.046] [0.046] [0.062] [0.062]
Hispanic 0.301a 0.297a 0.300a 0.138a 0.138a 0.163b 0.159b
[0.089] [0.089] [0.087] [0.047] [0.047] [0.063] [0.063]
AFQT percentile 0.827a 0.836a 0.935a 0.578a 0.579a 0.249b 0.257b
[0.15] [0.15] [0.15] [0.079] [0.079] [0.11] [0.11]
Report contraception −0.105c −0.0933 −0.0666 −0.0293 −0.0274 −0.0756c −0.0659
[0.060] [0.064] [0.064] [0.032] [0.034] [0.043] [0.046]
Smoking −0.144b −0.146b −0.172b −0.0712c −0.0715c −0.0731 −0.0747
[0.073] [0.073] [0.072] [0.039] [0.038] [0.052] [0.052]
Alcohol 0.135b 0.135b 0.127b 0.0805b 0.0805b 0.0549 0.0546
[0.063] [0.062] [0.061] [0.033] [0.033] [0.045] [0.044]
Marijuana −0.131 −0.13 −0.196 −0.179c −0.179c 0.0479 0.0488
[0.20] [0.20] [0.20] [0.11] [0.10] [0.14] [0.14]
Cocaine −0.584c −0.589c −0.517 −0.163 −0.164 −0.420c −0.425c
[0.35] [0.35] [0.35] [0.19] [0.19] [0.25] [0.25]
# children by age 34 −0.210a
[0.040]
Observations 1,030 1,030 1,030 1,030 1,030 1,030 1,030
R-squared 0.27 0.27 0.29 0.24 0.24 0.18 0.18
Over-identification test of instrumental variables
Sargan statistic 1.767 0.511 1.277 1.053
P value 0.413 0.775 0.528 0.591

Dependent variables are calculated over the age range from 21 to 34, adjusting dollars for inflation.
The sample includes all women with full experience profiles, who had their first child between the
ages of 21 and 33, and the years 1983–2000. Standard errors in brackets
a Significant at 1%
b Significant at 5%
c Significant at 10%

Source: Miller (2011)

bill 1-year
• Distinguishing risk-freeselection
between rates instead.
andForcausal
robustness, career
effects ofearnings areon
fertility alsotime use –
calculated over expanded windows: from age 21 to 35, 21 to 36, and on, until
Moschion 2013
21 to 42. The longer windows include smaller samples of women, but yield
substantially similar results for career effects. Nevertheless, one should remain
– Again,cautious
direction of causality between fertility events and labour supply/time use
about extrapolating the findings to motherhood delays past age 33.
changes is difficult
Hours to disentangle
worked each year are computed using actual hours reported in the work
history files (covering
– Use the exogenous variation all weeks from 1979
in number to 2000), from
of children and averaged over
the birth ofmissing
twins and gender
weeks within a year. Periods of non-participation are included as zeros; they
mix ofdo
first two children to estimate the impacts on hours, participation and domestic
not induce sample selectivity for career earnings. The average wage rate
work
– More children reduce labour market participation, reduce hours of work and increase
domestic work for women
– The impact is high relative to other countries, and occurs at all educational levels
– Also in contrast to other countries, men appear to reduce hours and participation
with a higher than expected number of children

7
2013 THE IMPACT OF FERTILITY ON MOTHERS’ LABOUR SUPPLY IN AUSTRALIA 327

T ABLE 4
Effect of Having More Than One (Two) Child(ren) on Mothers’ Outcomes (Labour Market Participation, Number of
Hours in Paid Work and Unpaid Domestic Work Per Week)

OLS IV IV
Controls 1 No controls Controls 1

LMP
More than one child
Twins 1 !0.155*** (0.005) !0.100** (0.050) !0.119*** (0.046)
N 59,573 59,573 59,573
More than two children
Same sex !0.183*** (0.005) !0.241*** (0.074) !0.195*** (0.071)
Twins 2 !0.183*** (0.005) !0.103*** (0.031) !0.106*** (0.030)
Twins 2 & Same sex !0.183*** (0.005) !0.123*** (0.029) !0.119*** (0.027)
(Hansen stat) – 2.9732 1.3379
(Hansen P-value) – 0.0847 0.2474
N 40,962 40,962 40,962
Hours paid work
More than one child
Twins 1 !6.293*** (0.165) !4.504*** (1.637) !4.393*** (1.504)
N 58,827 58,827 58,827
More than two children
Same sex !6.362*** (0.172) !9.249*** (2.537) !7.782*** (2.390)
Twins 2 !6.362*** (0.172) !2.287** (1.031) !2.253** (0.976)
Twins 2 & Same sex !6.362*** (0.172) !3.299*** (0.950) !3.037*** (0.901)
(Hansen stat) – 6.5415 4.5175
(Hansen P-value) – 0.0105 0.0335
N 40,439 40,439 40,439
Hours domestic work
More than one child
Twins 1 6.264*** (0.121) 5.283*** (1.239) 3.989*** (1.144)
N 58,356 58,356 58,356
More than two children
Same sex 4.356*** (0.133) 0.744 (1.887) 0.857 (1.837)
Twins 2 4.356*** (0.133) 4.091*** (0.741) 3.768*** (0.720)
Twins 2 & Same sex 4.356*** (0.133) 3.592*** (0.688) 3.344*** (0.668)
(Hansen stat) – 2.7157 2.1908
(Hansen P-value) – 0.0994 0.1388
N 40,072 40,072 40,072

Levels of significance: *10%; **5%; ***1%. LMP, Labour market participation.


Sample: Mothers with at least one child of their own (biological or adopted), aged 18 or less, and such that no child has left the
household (N = 59,573). In the second part of the table, the sample is restricted to mothers with at least two children in the
household (N = 40,962).
Note: Robust standard errors are reported in parentheses. Controls 1 include age, level of education (dummy for completing year
12), age at first birth, ancestry (seven dummy variables), sex of the first and, eventually, the second child.
Source: Census 2006 (5%), ABS.

Source: Moschion (2013)


rate (Table 4). OLS estimates indicate that LMP. With controls (column 3), having more
mothers with at least two children have partic- than one child decreases LMP by 12 percentage
ipation rates that are 16 percentage points lower points, and having more than two children by
4 Contraception, fertility and labour market activity
than their peers with only one child (column 1). between 11 and 20 percentage points, depending
Mothers with at least three children participate on the instrument used. Comparing the different
less than mothers with two children by 18
• The ability to control (perhaps stochastically)2SLS
percentage points. With or without controls,
specifications, adding control variables
fertility is assumed in dynamic models
does not alter the results, which is consistent
of
fertility and labour
instrumental supply
variable estimates confirm that with the assumption that the instruments are
fertility has a negative impact on mothers’ exogenous.
• Expanded access to contraception, particulary the oral contraceptive pill (OCP), had
substantial impacts on women’s labour market attachment and success
© 2013 Economic Society of Australia

• Goldin and Katz (2002) show that women living in states that made the OCP available
outside of marriage and to minors were more likely to enter professional occupations and
had a higher age at marriage

• Further research (see, for example, Bailey 2013) has found impacts on later family incomes
(2% higher) through men’s earnings and hours worked. The children of affected women
had higher college completion rates

8
5 Teenage pregnancy and birth rate
• The lifecycle model can also be used to consider teenage fertility choices: the decision to
give birth as a teenager may reflect expectations about future opportunities

• Once again, future opportunities may also be shaped by early childbearing: so the direc-
tion of causality between poor outcomes and early childbearing is unclear

• Australia’s rate of teenage fertility (births per 1000 of those aged 15-19) is moderate
internationally, and has declined dramatically over the past fifty years

Figure 1: Teenage fertility rate (age 15-19), 2012, for selected countries and regions
35  

30  

25  

20  

15  

10  

5  

0  
United  States   United   New  Zealand   OECD   Canada   Australia   European   Sweden   France   Germany  
Kingdom   members   Union  

Source: World Bank

• Kearney and Levine (2012) summarise the evidence, mainly for the United States

– Dramatic declines are driven by lower teen pregnancy rates, not higher abortion
– Higher US teen birth rate appears to reflect lower sexual activity alongside lower
contraceptive use compared to other countries, and in explaining the time trend

Figure 2: Teenage fertility rate (age 15-19) in Australia


60  

50  

40  

30  

20  

10  

0  
1963   1968   1973   1978   1983   1988   1993   1998   2003   2008  

Source: World Bank

9
Source: Kearney and Levine (2012)

– Policies associated with reductions in teen childbearing: Medicaid access to family


planning and lower welfare payments – but the effects are small. Abortion availability
has no effect
– More important: background disadvantage. Growing up in poverty or in a single-
parent household are risk factors – but captured by spatial and time variation most
importantly
– Perception of long-term prospects matters: if few long-term benefits perceived, teen
childbearing is more likely
– For teens from low socioeconomic status backgrounds, living in a high-inequality
state is associated with a 5 percentage point increase in being a teenage mother,
even after controlling for many other characteristics

This content downloaded from 129.78.233.210 on Tue, 14 Oct 2014 01:10:56 AM


All use subject to JSTOR Terms and Conditions

10
Source: Kearney and Levine (2012)

– Does teen childbearing cause poorer later outcomes?


∗ Need to adjust for selection: teens with poorer prospects are more likely to have
children as teenagers
∗ Within-family approach, miscarriage approach: estimates reduced substantially,
often insignificant and reversed sign
∗ Overall, differences in outcomes reflect differences in selection into teenage moth-
erhood (Ashcraft and Lang, 2006).

• A final note: social norms and the media are almost certainly important

– Kearney and Levine (2014) find that the MTV show 16 and Pregnant had a causal
effect on teenage pregnancy – a 5.7 percentage point reduction in teen births, ac-
counting for one third of the decline over the time period in question. There was
also an increase in internet searches regardning birth control and abortion

Short answer
1. Give three reasons why mothers may earn lower wages than women without children.

2. Briefly outline whether empirical evidence suggests that delaying the age at first birth
improves a woman’s career outcomes.

3. Give two examples of factors


This contentaffecting fertility
downloaded from andonlabour
129.78.233.210 Tue, 14 Oct supply decisions
2014 01:10:56 AM in a dynamic
All use subject to JSTOR Terms and Conditions
lifecycle model of fertility that are not captured in a static model.

11

You might also like