Allocation within the household: direct survey evidence

Abstract

We examine the physical and mental health effects of providing care to an elderly mother on the adult child caregiver. We address the endogeneity of the selection in and out of caregiving using an instrumental variable approach, and carefully control for baseline health and work status of the adult child using fixed effects and Arellano-Bond estimation techniques. Continued caregiving over time increases depressive symptoms for married women and married men. In addition, the increase in depressive symptoms is persistent for married men. Depressive symptoms for single men and women are not affected by continued caregiving. There is a small protective effect on the likelihood (10%) of having any heart conditions among married women who continue caregiving. Robustness checks confirm that the increase in depressive symptoms and decrease in likelihood of heart conditions can be directly attributable to caregiving behavior, and not due to a direct effect of the death of the mother. The initial onset of caregiving, by contrast, has no immediate effects on physical or mental health for any subgroup of caregivers.

Full text

Page 1

ISSN 1471-0498











DEPARTMENT OF ECONOMICS
DISCUSSION PAPER SERIES




ALLOCATION WITHIN THE HOUSEHOLD: DIRECT SURVEY
EVIDENCE



Martin Browning and Jens Bonke





Number 429
April 2009
Manor Road Building, Oxford OX1 3UQ

Page 2

Allocation within the household: direct
survey evidence.?
Martin Browning
Department of Economics,
Oxford Universityy
Jens Bonke
SFI, Copenhagen
September 2008
JEL Code: D13, C81,
Keywords: sharing rule, intrahousehold allocation,
expenditures.
Abstract
?The collection of the survey information used in this paper was funded by the Danish
SSF project "Intra-family Allocation of Resources in Denmark", 24-00-0212. We are very
grateful for this funding. We also thank the Danish National Research Council for support
through its grant to CAM.
yCorresponding co-author: Martin.Browning@economics.ox.ac.uk
1

Page 3

We report on a new diary based expenditure survey that for the
…rst time collects direct information on the allocation of all expen-
ditures to di¤erent members of the household. The most important
…nding from the survey is that the mean share that wives have of all
expenditures that are assignable to husband or wife is well determined
and very close to one half. Despite this equality at the mean, there is
considerable dispersion across the population and in half of households
one partner receives twice as much (or more) as the other. Moreover,
these expenditures comprise a sizable part of the household budget.
For example, the mean joint expenditure by husbands and wives on
their own private, assignable goods accounts for 11% of disposable
income.
The main observable determinants of the variation in sharing ex-
penditures across couples are a mixture of variables found in previous
studies and variables that have not been considered before. As regards
the former, we …nd that wives in higher educated and/or higher income
households have a higher expenditure share. The other familiar e¤ect
is that the wife’s share of assignable expenditures is increasing in her
share of gross income. Turning to variables that have not previously
been considered in the literature, we …nd that some of these are highly
2

Page 4

signi…cant and also have a strong impact on sharing. If the husband
had a mother who was in full-time employment when he was aged 14
then he receives over two percentage points more of assignable expen-
diture than if his mother was not employed full-time. The strongest
e¤ect is also the most puzzling. If the wife has a child who is not the
natural child of her partner then she receives six percentage points of
assignable expenditure less than an otherwise comparable woman. On
the other hand, men do better if they have had a previous child.
3

Page 5

1Introduction.
Within economics and sociology, the household was for a long time conceived
of as a unity with goals that are agreed by all members (the so-called ‘uni-
tary’ model). In the last twenty years, however, economists have questioned
that assumption and have elaborated models allowing for di¤erent prefer-
ences concerning the allocation of time and money within the family (see
Lundberg and Pollak (2003) and Browning, Chiappori and Lechene (2006)
for references and discussion). Simultaneously, economic psychologists and
sociologists have studied the decision-making processes and power-relations
within the family (see, for example, Vogler (1998) and England and Fol-
bre (2006) for surveys). Although there is widespread agreement that the
unitary assumption is inappropriate, there is no consensus on what should
replace it. We believe that the lack of appropriate data is the most im-
portant impediment to progress in this area. Acting on this belief we have
developed and conducted a new survey that is designed to throw more light
on intra-household allocation.
Our data are drawn from the Danish Household Expenditure Surveys
conducted in the years 1999 ? 2003. This is a conventional diary based ex-
penditure survey for which we have collected three novel sets of additional
4