Population Research and Policy Review 22: 251–266, 2003.
© 2003 Kluwer Academic Publishers. Printed in the Netherlands.
A cohort analysis of the timing of ﬁrst birth and fertility in Ghana
STEPHEN OBENG GYIMAH
Department of Sociology, Queen’s University, Kingston, Ontario, Canada
Abstract. This paper examines the nature of the inverse association between age at ﬁrst birth
and fertility across successive generations of Ghanaian women. Within the context of en-
hanced non-marital opportunities for contemporary women and declining fertility, we develop
a rationale for and test the hypothesis that in a medium fertility environment as currently found
in Ghana, the effect of age at ﬁrst birth on fertility becomes more important than ever before.
Five birth cohorts were identiﬁed (1938–1944; 1945–1949; 1950–1954; 1955–1959; 1960–
1964) from a merged ﬁle of the 1988, 1993 and 1998 Ghana Demographic and Health Surveys.
The analyses were restricted to women over 35 years old at the time of the surveys, which
allowed us to use current parity as a reasonable proxy for completed fertility. Preliminary
results suggest that women who had ﬁrst births early tend to have a higher number of births
than those whose ﬁrst births occur late, regardless of birth cohort. In multivariate analyses, the
effect of age at ﬁrst birth as a determinant of fertility was found to be more substantial among
later cohorts. The implications of the ﬁndings are discussed.
Keywords: Age at ﬁrst birth, Birth cohort, Fertility, Ghana, Sub-Saharan Africa
There is a growing body of literature that associates early life events with
later ones. In social scientiﬁc research, the timing of the ﬁrst birth is known
to affect a variety of demographic and non-demographic phenomena in the
life course of women (see, for example, Mirowsky 2002, Taniguchi 1999,
and Morgan & Rindfuss 1999). In the absence of effective contraception,
for example, the total number of children a woman bears is principally a
function of the age at which childbearing begins. Invariably, women who start
reproduction very early in life tend to have a large number of children than
those who start late. This negative association has been known for some time
now and constitutes an important empirical regularity in the fertility patterns
(Trussell & Menken 1978; Bumpass et al. 1978).
A number of theoretical reasons have been given as to why age at ﬁrst
birth is an important predictor of completed fertility. Besides the direct demo-
graphic risk of the long exposure during the high fecundity years of the
late teens and early twenties, there are indirect socio-economic links. Early
childbearing can interrupt a young woman’s education and also limit outside