Income and Childbearing Decisions: evidence from Italy
ABSTRACT During the early 1990s, Italy has been one of the first countries to reach lowest-low fertility, i.e. below 1.3 children per woman. In this paper we focus on the period during which such fertility levels arose in order to assess the impact of income on fertility decisions. So far, analyses have suffered from the lack of appropriate data; we create a new data set making use of two different surveys from Bank of Italy (SHIW) and ISTAT (Labor Force Survey) and we apply discrete-time duration models. For first births, we find evidence of non-proportional hazards and of some 'recuperation' effect: women with high predicted wages tend to delay the first birth, subsequently recuperating. For second and third births, instead, the availability of a good child-care system seems to play a key role and income exhibits small intensity. In a final section, we explore the possible effect on fertility of an increase in financial support for poorer families that took place in 1999.
Income and Childbearing Decisions:
Evidence from Italy.
(Institute for Quantitative Methods, Bocconi University, Milan, Italy and Institute
for Social and Economic Research, University of Essex)
(Institute for Social and Economic Research, University of Essex)
Francesco C. Billari
(Institute for Quantitative Methods and Innocenzo Gasparini Institute for Economic
Research, Bocconi University, Milan, Italy)
ISER Working Paper
Thanks to John Erm isch, Stephen P. Jenkins, Heather Joshi, Cheti Nicoletti and Birgitta Rabe for their helpful
com m ents and suggestions. Concetta Rondinelli worked on this paper during her research visit at the Institute
for Social and Econom ic Research, ISER , University of Essex, whose hospitality she acknowledges.
Readers wishing to cite this document are asked to use the following form of words:
Rondinelli, Concetta; Aassve, Arnstein; Billari, Francesco C. (April 2006) ‘Income and
Childbearing Decisions: Evidence from Italy.’, ISER Working Paper 2006-06. Colchester:
University of Essex.
The on-line version of this working paper can be found at http://www.iser.essex.ac.uk/pubs/workpaps/
The Institute for Social and Econom ic Research (ISER) specialises in the production and analysis of longitudinal data.
MISOC (the ESRC Research Centre on Micro-social Change), an international centre for research into the
ULSC (the ESRC UK Longitudinal Studies Centre), a national resource centre to prom ote longitudinal surveys
and longitudinal research.
The support of both the Econom ic and Social Research Council (ESRC) and the University of Essex is gratefully
acknowledged. The work reported in this paper is part of the scientific program m e of the Institute for Social and Econom ic
Institute for Social and Econom ic Research, University of Essex, Wivenhoe Park,
Colchester. Essex CO4 3SQ UK
Telephone: +44 (0) 1206 872957 Fax: +44 (0) 1206 873151 E-m ail: firstname.lastname@example.org
All rights reserved. No part of this publication m ay be reproduced, stored in a retrieval system or transm itted, in
any form , or by any m eans, m echanical, photocopying, recording or otherwise, without the prior perm ission of
the Com m unications Manager, Institute for Social and Econom ic Research.
During the early 1990s, Italy has been one of the first countries to reach lowest-low fertility, i.e. below
1.3 children per wom an. In this paper we focus on the period during which such fertility levels arose in
order to assess the im pact of incom e on fertility decisions. So far, analyses have suffered from the lack
of appropriate data; we create a new data set m aking use of two different surveys from Bank of Italy
(SHIW) and ISTAT (Labor Force Survey) and we apply discrete-tim e duration m odels. For first births,
we find evidence of non-proportional hazards and of som e 'recuperation' effect: wom en with high
predicted wages tend to delay the first birth, subsequently recuperating. For second and third births,
instead, the availability of a good child-care system seem s to play a key role and incom e exhibits sm all
intensity. In a final section, we explore the possible effect on fertility of an increase in financial support
for poorer fam ilies that took place in 1999.
Keywords: lowest-low fertility, incom e and childbearing, tim ing of births.
JEL Classification: J13, J18, C41
During the last two decades of the Twentieth Century, Italy has been, together with Spain, the first
country to reach the so called lowest-low fertility, i.e. below 1.3 children per wom an. During this period,
total fertility has sharply decreased, while both the percentage of wom en having com pleted at least
upper secondary school and fem ale labor force participation have significantly increased. Given this
setting, it is of prim ary interest to establish the extent to which wom en's educational attainm ent and
potential incom e, and therefore their opportunity cost, m ay affect their childbearing decisions. As Billari
and Kohler (2004) point out, the em ergence of the lowest-low fertility in Southern Europe is not
connected to the steep increase in childlessness but it is due to the sudden decrease of the progression
to the second, third and subsequent children.
Com bining two different data sets (from the ISTAT Labor Force Survey and the Survey of
Households' Wage and Wealth led by the Bank of Italy in 2002), we estim ate the effect of wage on the
postponem ent of m otherhood. The wage effect is negatively correlated with having children, the
m agnitude, however, varies according to parity. Consistent with the opportunity cost theories, wage has
a strong and negative effect in the tim ing of first birth and wom en with higher wages tend to delay
m otherhood. We find a non-proportional hazard and a recuperation effect (even if not com plete),
suggesting that wom en with higher wage start to have children later, but recuperate after som e tim e.
Furtherm ore, there is no evidence that institutional effects are responsible for the postponem ent of
m aternity. Different is the pattern for second and third birth: the wage effect is always negative but it
has sm aller intensity when com pared to the first one. Nevertheless, in line with Erm isch (1989), we find
evidence of institutional effects affecting the decision of having m ore than one child: the risk of
experiencing a second and third birth is higher for a Northern wom an because she is m ore confident in
the availability of childcare.
It is reasonable to suppose that increasing financial support to households with children m ay
have an im pact on the probability of having the third and fourth child, m ainly for poorer households. To
this extent, the Welfare Minister Livia Turco (law num ber 448 of the Year 1998) introduced two policy
m easures with the explicit purpose of supporting poor households with children. These two m easures
could cause a significant increase in incom e for low-incom e fam ilies, covering a non-negligible
proportion of the cost of an additional child. Our estim ates give som e support for the law having an
im pact. But this effect is not clearly identifiable and quantifiable given our m ethodological approach,
which in turn is driven by the data available. Despite this lim itation our results provide interesting
evidence of the key role wage plays in fertility decisions.
In this paper we assess the effect of income on fertility in Italy during the period
1983-2003. During this period, fertility levels have significantly declined and
in the middle of the period (around 1992-1993) passed the critical thresholds of
1.3 children per woman defining so-called lowest-low fertility. Moreover, during
the same period Italian women have increased their educational attainment and
labor force participation. Given this setting, it is of primary interest to establish
the extent to which women’s educational attainment and potential income, and
therefore their opportunity cost, may affect their childbearing decisions.
So far, research has been hampered by the availability of data combining fer-
tility and income. We overcome this problem by combining two different data
sets. Data from the Labor Force Survey (ISTAT, 2003) are used in order to re-
construct the basic demographics and the fertility history of a woman. In order
to link these features with potential income1, we use income and earnings data
from Bank of Italy (Households Income and Wealth, 2002). Data from Bank of
Italy’s survey are thus used to predict women’s potential income, which is then
introduced as explanatory variable in discrete time hazard regression models for
first, second, and third births. In order to obtain a measure of the income we
use a Tobit model, censoring the augmented log-wage at the smallest value of the
a-incremented log-income distribution.
Women aged between 15 and 40 years in 2003, linked with their co-residing
children at the moment of the interview, are our unit of analysis. Our measure of
birth decisions. We are mainly interested in showing if income plays a key role in
the postponement of motherhood in Italy and its different impact in the transition
to first, second and third birth. More specifically, we assess if socioeconomic
features are the only determinants in such a transition or if other socio-cultural
situations are responsible for delaying motherhood and deciding whether to have
the second and third birth.
The remainder of this paper is organized as follows. In Section 1 we review
the existing literature and describe the main features of the Italian setting as far as
childbearing decisions are concerned. Section 2 provides a description of the data
and methods we use. In section 3 we present our main results. Section 4 includes
concluding remarks and some policy considerations.
1In this paper we use the word income which stands for wage (i.e. income per hour).