Does Maternal Age Affect Children’s Test Scores?*
Andrew Leigh
Economics Program
Research School of Social Sciences
Australian National University
andrew.leigh@anu.edu.au
http://econrsss.anu.edu.au/~aleigh/
Xiaodong Gong
Australian Treasury
xiaodong.gong@treasury.gov.au
Abstract
We estimate the relationship between maternal age and child outcomes, using indices
aimed at measuring overall outcomes, learning outcomes, and social outcomes. In all
cases, we find evidence that children of older mothers have higher outcomes. Not
only do children born to mothers in their twenties do better than children born to teen
mothers, but children born to mothers in their thirties do better than children born to
mothers in their twenties. However, when we control for other socioeconomic
characteristics, such as family income, parental education, and single parenthood, the
coefficients on maternal age become small and statistically insignificant. The only
exception is an index of social outcomes, which is positively associated with maternal
age, even controlling for socioeconomic factors. For cognitive outcomes, young
motherhood appears to be a marker, not a cause, of poor child outcomes.
Keywords: cognitive development, non-cognitive skills, intergenerational human
capital transmission
JEL Codes: I2, J13
*
This paper uses unit record data from the Longitudinal Study of Australian Children
(LSAC) Survey. The LSAC Project was initiated and is funded by the Australian
Government Department of Families, Housing, Community Services and Indigenous
Affairs (FaHCSIA) and is managed by the Australian Institute of Family Studies
(AIFS). The findings and views reported in this paper, however, are those of the
authors and should not be attributed to either FaHCSIA or AIFS. We are grateful to
Mark Wooden, two anonymous referees, and FaHCSIA staff for valuable comments
and suggestions on earlier drafts. Susanne Schmidt and Elena Varganova provided
outstanding research assistance.
,1. Introduction
Most studies of early childbearing have focused on the consequences for the mother
in terms of employment, human capital accumulation, and earnings. However, it is
possible that a mother‘s age also has a causal effect on her child‘s outcomes. Further,
while most studies have concentrated on teenage motherhood, policymakers might
also be concerned with the relationship for older mothers. Put simply, we might want
to know not only whether children of 28 year-old mothers do better than children of
18 year-old mothers, but also whether children of 38 year-old mothers do better than
children of 28 year-old mothers.
While most government policies ignore maternal age in a statutory sense, they may
have a differential effect in an economic sense. For example, a Maternity Payment
(‗Baby Bonus‘) of $5,000 is likely to have a larger impact on the behaviour of an 18
year-old mother than of a 38 year-old mother, due to the different hourly wages and
wealth levels of the two individuals. Conversely, higher levels of government
assistance for in vitro fertilisation treatments (IVF) are likely to increase the share of
older mothers. Better understanding the development of children born to mothers of
different ages is therefore relevant to shaping government policy even in an
environment where maternal age is rarely an explicit condition of benefit receipt.
In theory, there are various channels through which maternal age might affect
children‘s outcomes. Theories suggesting that maternal age will be positively
2
,correlated with child outcomes include the possibility that younger mothers may be
emotionally unprepared for motherhood, that they may have less parenting skills than
if they had become parents at an older age, that they may be less sensitive in
responding to their infants, or that they may make more poorly informed choices
about daycare, preschool and kindergarten. Conversely, it might be the case that
maternal age will be negatively correlated with child outcomes – for example if older
mothers are less physically fit, have less support from grandparents, or spend less
time with their children because their opportunity cost (market wage) is higher.
In this paper, we estimate the relationship between the age of a child‘s mother and the
outcome indices for their children in early years (ages 4-5). We use data from
Australia, which has seen a steady shift towards women having children at an older
age. Each year, the age of the typical mother rises by approximately two months. Yet
there has been little evidence on whether this substantial social change might have a
positive or negative impact on children.
A major empirical challenge in estimating the causal effect of maternal age on
children‘s outcomes is that any observed correlation will not necessarily imply
causality. For example, women who have babies at a young age may also be poorer,
less educated, and less likely to be married than older mothers. Younger mothers may
also be different with respect to unobservable characteristics, such as intrinsic
motivation, academic aptitude and social networks. If the kinds of women who have
3
, children at a young age are systematically different from the kinds of women who
have children at older ages, then we cannot assume that differences in the outcomes of
children of younger mothers and older mothers tell us what would happen if the same
woman chose to delay childbearing. To a modest extent, we can address some of these
difficulties by carrying out a multiple regression in which we hold constant
observable factors such as family income, whether there are two parents present, and
the presence of siblings.
To preview our findings, we observe a positive and statistically significant
relationship between maternal age and children‘s outcomes. This relationship exists
not only when comparing teenage mothers with mothers in their twenties, but also
across other ages. However, the relationship is quite sensitive to the addition of other
controls, suggesting that mother‘s age may be a proxy for other indices of
disadvantage. For the overall index and learning index (though not for the social
index), mother‘s age has no significant association with child outcomes once
socioeconomic controls are added to the regression.
The remainder of this paper is structured as follows. In Section 2, we provide some
background to our study by presenting an overview of trends in maternal age over
recent decades, and reviewing the literature on the relationship between maternal age
and child outcomes. In Section 3, we discuss the data from which our results are
drawn: the Longitudinal Survey of Australian Children (LSAC). In Section 4, we
4
