Imperfect Commitment, Social Constraints and Household Time Allocation Cristina Fernandez Almudena Sevilla-Sanz IESE Business School University of Essex, ISER December, 2005 Abstract Economic theories of the household predict that increases in rel- ative female human capital lead to increases in female labor force participation and, symmetrically, to decreases in the female time de- voted to household production. However, both at the longitudinal and cross-sectional level we observe that, despite the decline in the wage gender gap, specialization in home production continues to be high, with women providing most of household produced goods and services. We develop a simple model that recognizes the imperfect commitment associated to the contractual processes over household time allocation. In the light of the model, imperfect commitment is characterized as a constraint on the household division of labor 1
41
Embed
Imperfect Commitment, Social Constraints and …dse.univr.it/espe/documents/Papers/E/4/E4_2.pdfImperfect Commitment, Social Constraints and Household Time Allocation Cristina Fernandez
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Imperfect Commitment, Social Constraints
and Household Time Allocation
Cristina Fernandez Almudena Sevilla-Sanz
IESE Business School University of Essex, ISER
December, 2005
Abstract
Economic theories of the household predict that increases in rel-
ative female human capital lead to increases in female labor force
participation and, symmetrically, to decreases in the female time de-
voted to household production. However, both at the longitudinal
and cross-sectional level we observe that, despite the decline in the
wage gender gap, specialization in home production continues to be
high, with women providing most of household produced goods and
services. We develop a simple model that recognizes the imperfect
commitment associated to the contractual processes over household
time allocation. In the light of the model, imperfect commitment
is characterized as a constraint on the household division of labor
1
beyond what is considered to be "socially acceptable". The model
predicts that imperfect commitment problems are stronger (and thus
the social constraint more likely to bind) (1) the higher the woman’s
relative wage and (2) the less credible threats available. We test the
model using the 2002-2003 Spanish Time Use Survey, a time diary
survey with information on the time devoted to household production
activities by both partners. Empirical findings support the proposed
model of imperfect commitment in the allocation of household time.
Although a woman’s home time decreases as her wage goes up, this
effect is less pronounced as her wage is higher. Furthermore, the time
devoted to those household activities where no credible threats exist
(such as those involving care) are less elastic to an increase in the
relative female wage.
JEL classification: D13, J0, J1, J2, Z13
1 Introduction
Contrary to the predictions of comparative advantage or bargaining theories
of the household, higher female human capital has not led to a more egal-
itarian allocation of time within the household.1 The determinants of how
spouses allocate time to household production are based on the predictions
1Time spent producing goods and services within the household has been recognized as
important since Becker (Becker, 1965). However, the time devoted to household production
continues to be overlooked in most household economic models (Juster and Stafford, 1991).
2
of either the unitary household production models (Becker, 1991) or bargain-
ing models (McElroy and Horney, 1981). In the unitary framework family
members cooperate to produce utility for all, either through the purchase
of market goods and services with earnings from market work or through
household production. Specialization is thus efficient and the spouse with the
lowest opportunity cost (i.e. the lowest human capital or the highest home
productivity) contributes the most to household production and the least to
market work. Bargaining theories reach the same conclusion but are based
on the concept of threat points determined by either the cost of falling out
of marriage ((McElroy and Horney, 1981)) or of a non-cooperative marriage
((Lundberg and Pollak, 1993)).2 However, empirical findings using time-use
data contradict the prediction of both theories. For example, when a wife
works more hours than her husband outside the home, she still undertakes
a larger share of housework (Ackerlof and Kranton, 2000). Similarly, men’s
unpaid work increases with his wife’s wage but only up to the point where
the wife contributes as much as the husband to the household income. Be-
yond that point, the wife’s contribution to housework starts increasing again
2Chiappori (Chiappori, 1992) and (Browning and Chiappori, 1997) unified both set of
theories into a "collective" approach to the household, where efficiency in the household
maximization problem is secured due to spousal transfers of private consumption. Incor-
porating household production into the collective framework requires further restrictions
on preferences and technology in order to identify the sharing rule and raises questions on
the dichotomization of time into leisure and market work in household economic models
(Apps and Rees, 1996) and (Apps and Rees, 1997).
3
(Bittman et al., 2001). Furthermore, the unequal allocation of household
time persists after observable characteristics are taken into account (Alvarez
and Miles, 2003).
We present a simple household model that reconciles the theory and em-
pirical findings by recognizing that decisions over time allocation migtht be
subject to imperfect commitment mechanisms. These imperfect commit-
ment processes can be traced back to a couple’s inability to reach binding,
legally-enforceable agreements about future behavior because of the non-
observability by third parties (see (Basu, 2001) and (Rasul, 2002)) or the
inability to fullfil informal contracts because the lack of credible threats (Fol-
bre and Bittman, 2004). Imperfect commitment is usually characterized in a
dynamic bargaining setting as the inability of one spouse to make transfers
of private consumption to compensate the other partner for utility losses.
Thus inefficiencies may arise such as in Lundberg 2001 (Lundberg and Pol-
lak, 2001). The theory presented in this paper focuses on the inability of
potential partners to credibly commit to make transfers of time, rather than
private consumption. We characterize the imperfect commitment as a con-
straint on the household division of labor beyond what is considered to be
"socially acceptable". The model has two predictions. First, whereas the
constraint on household division of labor is not likely to bind for low wage
women where specializsation is high, it becomes binding for women with a
higher relative wage. Second, the model predicts that the constrain is more
likely to bind for those household actitivities where commitment failures are
4
more important (such as caring activities).
Empirical studies that test the theories of the household have focused
the identification of unitary from bargaining models of the household, which
assume efficiency inhousehold decisions. Studies that attempt to empirically
identify inefficient outcomes within the household are rare (An exception is
that of Mazzocco (Mazzocco, 2003)). We test the predictions of the model
using the 2002-2003 Spanish Time Use Survey, a time diary survey with in-
formation on the time devoted to household production activities by both
partners. Empirical findings support the proposed model of inefficient al-
location of household time. First, consistent with the presence of social
constraints in the division of household labor we find that a woman’s relative
time allocation to household production services decreases with her relative
wage up to a certain level, and remains constant afterwards. Second, we find
that for those household tasks subject to greater commitment problems (for
example childcare) the constraint binds for lower female relative wages.
The paper is organized as follows. Section 2 presents a simple model of
time allocation and household production. Section 3 specifies the empirical
methodology. Section ?? describes the data set used in the analysis. Section
?? presents the empirical results and section ?? concludes.
5
2 Theoretical Model
To begin a baseline model of the household is presented that focuses on two
specific aspects of the gains associated to a union: efficiency gains from spe-
cialization in household production and the consumption of market public
goods.3 We then present a model where imperfect committment is as a so-
cial constraint on the household division of household labor beyond what is
"socially acceptable".4 The rationale behind imperfect committment are the
non-observability by third parties of spouse’s time devoted to household pro-
duction and the absence of credible threats for certain household production
activities (especially those related to caring activities (Folbre and Bittman,
2004)). Once the surplus from the union is defined, prospective mates need
to form some notion as to whether families realize the potential gains and
how those gains are divided. However, while the division of household sur-
plus is done efficiently under the baseline model, inefficient allocations arise
in the presence of imperfect commitment and social constraints.
3Other dimensions to marriage such as risk pooling or consumption smoothing are left
out of the analysis for exposition purpuses.
4Similarly to Lundberg and Pollak (Lundberg and Pollak, 1993) and Ackelof et al.
(Ackerlof and Kranton, 2000).
6
2.1 Baseline Model: Efficient Allocation of Household
Time
A household is denoted by the superscript U (union) and is assumed to
be formed by two individuals a man m and a woman f . The joint house-
hold utility depends on the consumption of two public goods: household
maintenance (which is privately produced by the household members) and
a composite consumption good (which is purchased in the market). The
composite consumption good includes market consumption goods that are
jointly consumed by the household C (such as groceries, housing, child care,
etc.) and can be acquired in the market at a normalized price p = 1. For
the remainder of the paper we will refer to C as the market public good. The
household joint utility also depends on the production (and consumption) of
a particular public good, household chores or household public good, Z (these
are the "commodities" in Becker’s language (Becker, 1965) such as a cleaned
house or home-made meals). Consider Z as a lower bound for the amount
of household production that needs to be done in the household which differs
from C in that it cannot be purchased in the market and is produced using
both partners’ time in household production Hi for i = m, f such that
Z = afHf + amHm
with ai being the man’s and woman’s productivity in household produced
7
goods.5 Thus, whereas the output Z is consumed jointly by both partners,
each partner privately contributes to its production. I assume that each part-
ner derives disutility f(Hi) from the time devoted to household production
Hi for i = m,w, where f(.) is an increasing and convex cost function. I
normalize 0 ≤ Hi ≤ 1.
The household’s utility is defined as the sum of individual utilities such
that:
V = U(C) + U(Z)− f(Hm)− f(Hf)
and the household’s maximization problem is:
maxCi,Hi
U(C) + U(Z)− f(Hm)− f(Hf)
st.
Z ≥ afHf + amHm
C =X(1−Hi)wi
at the optimum the household consumes all the joint disposable income
and produces the needed amount of household production. The amount of
time that each partner devotes to household production Hi is given by the
first order conditions:
Hi : −U 0(C)wi + aiU0(Z)− f 0(Hi) = 0
5For the reminder of this paper lets assume that af = am. This assumption is made for
exposition purposes only. The results are robust to general specifications of the production
function, which include market goods as inputs and other forms of substitutability.
8
The standard prediction follows thatHUf > HU
m as long as wm > wf . Further-
more the relative amount of housework Hf
Hf+Hmis decreasing on the female
wage wf .6
It is important to note that in this household there is a unique distri-
bution of individual utilities. This derives from the assumption that the
only private goods are essentially the disutility of time devoted to household
production. Thus, unlike collective (Browning and Chiappori, 1997) or non-
cooperative models of the household (Lundberg and Pollak, 1993), transfers
of private consumption between partners cannot compensate for time devoted
to household production. This distinction, which bares the substitutability
assumption between time spent at household production and money, is im-
portant on theoretical and empirical grounds. Empirically, time-use survey
data shows low levels of household services outsourcing, suggesting low sub-
stitution between time spent in household production and money spent in
market goods.7 On the theoretical front, Apps and Rees (Apps and Rees,
6Under the assumption of interior solution, the second order conditions are satisfied
such that:
Hi : 2U”(CU )w2i + 2aiU
”(ZU )− f”(HUi ) ≤ 0
for i = m, f .
7Using Australian time-use data, Bittman shows that during the period 1984-94 real
expenditure in outsourcing for cleaning, for instance, did not increase (Folbre and Bittman,
2004) (p.229-230). Moreover, only 4 percent of the households bought any cleaning services
during the two-week period of the survey. Similar evidence exists for the case of Britain
9
1996) and (Apps and Rees, 1997) extensively discuss the theoretical limi-
tations of the perfect substitutability assumption necessary to identify the
sharing rule in collective models of the household (Browning and Chiappori,
1997).8
2.2 Imperfect commitment, social constraints and in-
efficient allocation of household time
This section builds on the baseline model of section 2.1 by analyzing the role
of imperfect commitment processes associated to household production. Im-
perfect commitment is characterized as a constraint that effectively prevents
potential partners to perfectly contract upon the efficient division of house-
hold labor. Thus, the amount of time each partner devotes to household
production Hm and Hf is the value dictated by what is socially acceptable.
and the United States, where despite increases in income inequality the demand for paid
domestic services has not increased. Such low levels of outsourcing are also found in the
1994 and 2002 ISSP in this paper.
8Apps and Rees (Apps and Rees, 1997) point out that most of the goods consumed
within the household, except for arguably leisure, are public goods. They also show
that the range of goods where consumption differs significantly across individuals in the
household is relatively small. Similarly, Fella et al. (Fella et al., 2004) use the non-
divisibility of public consumption goods within marriage to explain the relevance of divorce
laws on the probability of divorce. In this setting, the allocation of private allocation of
time, rather than private consumption, becomes the variable of interest.
10
The household maximization problem becomes:
maxCi,Hi
U(C) + U(Z)− f(Hm)− f(Hf)
st.
Z ≥ Hm +Hf
C =X(1−Hi)wi
Hm ≤ Hm
where the constraint is characterized by the fact that the male partner
cannot increase his time to household production above what is imperative
by gender roles Hm (or symmetrically, that a woman cannot decrease the
amount of time devoted to household production below what is prescribed
by the existing social norms Hf). If the constraint is not binding we are
in the baseline model presented above, however when the constraint binds
the solution is straigth forward given that ZU is fixed and ZU = HU
m +HUf ,
HUf = H
U
f and CU =X(1−H
U)wi.
Imperfect commitment has two effects: First, it diminishes the household
utility and second, it alters the distribution of household surplus. First,
household utility is lower because partners are constrained from reaching
optimal time allocations. Second, the distribution of the surplus is altered.
The decrease in total surplus occurs despite increases in the consumption
11
of the market public good.9
CU
Imp erf ect commitment = (1−HU
m)wm+(1−HU
f )wf > CUBaseline = (1−HU
m)wm+(1−HUf )wf
Under imperfect commitment in the way characterized above a man has
no incentive to deviate from what is constrained once the union is formed.
He not only enjoys a higher level of market consumption good but also a
lower disutility from household labor.
VU
m = U(CU) + U(ZU)− f(H
U
m) < V Um = U(CU) + U(ZU)− f(HU
m)
A woman is however worse off as a higher consumption of the market
public good is not enough to compensate her for a higher amount of time
devoted to household production.10
VU
f = U(CU) + U(ZU)− f(H
U
f ) < V Uf = U(CU) + U(ZU)− f(HU
f )
3 Empirical Specification
First we test for the existence of rigidities associated to the division of house-
hold labor. The model predicts that the constraint is more likely to bind the9Under social constraints the man devotes less time to household production and more
to market labor. Given that the man’s wage is higher, the consumption of market public
good increases.
10Given that the household’s utility decreases in the presence of imperfect committment,
and that man’s utility increases, woman’s utility must necessarily decrease.
12
higher the female relative wage. Intiuitively this is so because the degree of
specialization within a household i is decreasing with the female wage, so that
as the female wage increases Hi,m increases and Hi,f decreases. Let’s denote
w the binding wage defined as the female wage associated with an optimal
value H∗i,m, such that H
∗i,m = Hm. Then the constraint becomes binding
for any wi,f > w. Thus, for female wages less than w the ratio Hi,f
Hi,f+Hi,mis
decreasing in the female wage but becomes constant for any wi,f > w.
Denote the optimal share of female housework in a given household i as
hi,f =Hi,f
Hi,f+Hi,m. We showed that hi,f is a function of both spouses wages
wi,f and wi,m, spouses productivities ai,f and ai,m, the minimum amount
of housework Zi and the socially constrained housework H so that hi,f =
h(wi,ai, Zi,H) where the minimum amount of housework Zi, and individual
productivities ai,f and ai,m are randomly distributed across households and
are a function of the household’s and spouses’ observable characteristics and
tastes such that Zi = Z(Vi;ωi) and aj = a(Xi;ui). We can use the following
linear regression equation to test this first implication of the model
hi,f = wlowi,f β
lowf + whigh
i,f βhighf + wi,mβm +Xiγ + εi (1)
where i denotes the household, wf is the woman’s wage, wm is the husband’s
wage, andXi controls for other household’s and spouses’ characteristics. The
coefficients of interest are βlowf and βhighf . The theory predicts both coeffi-
cients to be negative, i.e., as female relative wage increase hf should decrease.
However, if our theory is right hf has to level off once wf = w is reached.
13
This would imply |βlowf | ≈ |βhighf |. The variables in Xi control for both the
technology of household produced goods ai and for the tastes for the level
of household production in the household Zi. Among these variables are the
existence of outsourced household production and the presence of microwave
and other devices used for household production that might affect spouses’
productivity in household production. Other variables included in Xi are
spouses ages, education and household composition, as well as a dummy to
control for regional residence.
The second test of the theory attempts to identify the imperfect commit-
ment processes associated to household production activities. If imperfect
commitment is the underlying cause for social constraints in the division of
household labor, and different household activities are subject to different
imperfect commitment problems, we would expect the constraint to be dif-
ferent for different types of household production activities. Consider two
types of household production activities, activity k and activity j. If ac-
tivity k is more prone to suffer from imperfect commitment processes due,
for example, to the inexistence of credible threats (such is the case for car-
ing activities) then we would expect that the constrained level of housework
would be lower for activity k than for activity j, such that Hk < Hj and
thus wk < wj for any household i. That is, the constraint binds for lower
levels of female wages for activity k than what it does for activity j. We test
this hypothesis by running equation 1 for different household activities:
14
hik,f = wlowikf β
lowkf + whigh
ikf βhighkf + wikmβkm +Xikγk + εik (2a)
hij,f = wlowij,fβ
lowf + whigh
ijf βhighf + wijmβm +Xijγj + εjk (2b)
where k and j stand for two different household activities and wlowif are low
female wages and whighif are high female wages. If imperfect commitment is
associated to activity k and not to activity j then we expect that whereas
|βlowkf | ≈ |βhighkf | for activity k, for activity j |βlowj,f | < |βhighj,f |. This is, theamount of woman’s time devoted to childcare (task k) is more likely to be
less responsive to a woman’s wage w as the female wage increases than the
amount of woman’s time devoted to dish washing (task j). Furthermore, we
would expect that |βlowkf | ≈ |βlowj,f |, i.e. when the constraint is not bindingthe effect of the female wage on the division of housework is similar for both
activities.
4 Data: 2002-2003 Spanish Time Use Survey
The data used for the empirical analysis is drawn from the 2002-03 Spanish
Time Use Survey. The Spanish Time Use Survey is part of the Harmonized
European in Time Use Surveys (HETUS) launched by the EU Statistics Of-
fice (Eurostat). This is a representative data set directed at a sample of
20,603 households, which obtains information on daily activities by means of
the completion of a personal diary and household and individual question-
naires. The sample is evenly distributed over the year in order to represent
15
all days of the week and is potentiated on the weekend in order to capture
a greater variety in the population’s behavior.For this purpose, the sample
is subdivided in two sub-samples of equal size, one which must complete the
diary from Monday to Thursday and one which completes it from Friday to
Sunday.
The activities diary constitutes the most characteristic instrument of the
Survey. All members of the household 10 years old and over must complete
it on a selected day (the same day for all members of the household). The
diaries time frame occupies 24 consecutive hours (from 6:00 in the morning
until 6:00 the following day) and is divided into 10 minute intervals. In each
of the intervals, the informant must note the main activity, the secondary
activity carried out at the same time (given the case) and if at that time
they are in the presence of other known persons. These activities are coded
according to a harmonized list of activities from Eurostat, which considers 10
large groups: personal care, work, studies, household and family, volunteer
work and meetings, social life and recreation, sports and open air activities,
hobbies and games, means of communication, and non-specified travel and
use of time.
Diary-based time use data avoids the biases associated to other time-use
surveys based on stylized questions where the total amount of time devoted
to a particular activity is recorded (Juster and Stafford, 1991), (Alvarez and
Miles, 2003) and (Kan, M. 2006). The Spanish Time Use survey proves
particularly useful for our study since, unlike other recent diary-based time
16
use surveys like the American Time Use Survey (ATUS), the Spanish Time
Use Survey contains information on time devoted to household production
by both spouses (as well as other members of the household).
Given the novelty of this data set, table ?? presents a comparison of
main variables to the Spanish Labor Force Survey (EPA), a well-known rep-
resentative panel data set of the Spanish labor market. We observe that the
main demographic and economic variables in both data sets resemble each
other. The education distribution is somewhat different between the two sur-
veys. This is likely to be due to a different classification method rather than
inherent differences in educational achievement. Labor force participation
confirms this hypothesis, as both data sets coincide in the percentage of men
and women in the labor force and unemployed.
17
EPA Time Use
Both Men Women Both Men Women from 16 to 19 5.43 5.73 5.14 5.55 5.68 5.42 from 20 to 24 8.43 8.86 8.02 8.46 8.91 8.03 from 25 to 29 10.08 10.58 9.62 10.58 11.12 10.08 from 30 to 34 10.13 10.63 9.66 9.75 10.26 9.27 from 35 to 39 9.75 10.15 9.37 9.94 10.23 9.66 from 40 to 44 8.94 9.22 8.68 8.95 9.18 8.74 from 45 to 49 7.85 8.04 7.67 7.95 8.23 7.7 from 50 to 54 7.09 7.22 6.97 7.13 7.2 7.07 from 55 to 59 6.65 6.68 6.61 6.54 6.62 6.46 from 60 to 64 5.56 5.49 5.63 5.59 5.43 5.74 from 65 to 69 5.86 5.44 6.25 6.43 6.09 6.76 more than 70 14.22 11.95 16.37 13.12 11.05 15.07Total 100 48.56 51.44 100 48.66 51.34
Both Men Women Both Men WomenSingle 31.05 35.19 27.15 30.27 33.76 26.96Married 58.61 60.32 57.00 59.55 61.39 57.8Widow 7.58 2.54 12.33 7.28 2.54 11.77Divorced 2.76 1.96 3.52 2.91 2.32 3.47Total 100 100 100 100 100 100
Observations 1892.00 1892.00 1892.00 1882.00 1882.00 1882.00 1877.00R-squared 0.04 0.05 0.07 0.07 0.08 0.09 0.1Absolute value of significant t statistics in parentheses* significant at 10%; ** significant at 5%; *** significant at 1%
Table 5.1.a: Binding Constraints on the Division of Housework
26
5.2 Heterogeneity in Household Production and Im-
perfect Commitment
The second test attempts to identify the imperfect commitment processes as-
sociated to different household production activities by exploiting the infor-
mation in the data regarding the different commitment problems associated
to different household activities. For this purpose we exploit three pieces of
information: The frequency of household production activities (whereas it is
a daily activity or a less frequent activity), when this household production
is undertaken (if during the working days or in the weekend) and the type of
household production activity (whether it is caring activities or housework
activities). In the first case we expect imperfect commitment problems to be
more important for those activities that are performed daily. In the second
case we expect imperfect commitment problems to be more important for
those activities performed during workdays. In the third case we expect im-
perfect commitment problems to be more important for caring rather than
other activities.
Frequent and infrequent household activities One would expect that
the type of housework done daily is that type of housework that "needs to
be done" and cannot be postponed. However, other type of housework, the
one that is done less often, is considered to be not such a necessity so people
can live without it. This would imply that whereas there might be some
credible threats availiable for non-daily housework, credible threats might be
27
more difficult to find for daily housework. If activity k is performed daily
and activity j non-daily, then Hk < Hj and thus wk < wj for any household
i. That is, the constraint binds for activity k (daily) for lower female wages
than what it does for activity j (non-daily).
We test this hypothesis by running equation 1 separately for daily and
non-daily activities. Daily activities are those that are done for more than
half of the sample and non-daily activities are those not done for more than
half of the sample in the day when they are interviewed. Daily activities
are thus cooking, cleaning, clothes and shopping. Non-daily activities are
household repairs and maintenance, household management and gardening.
The results of are presented in table
We observe that most of the variation observed in table ?? comes from
daily activities. The coefficients for daily activities are very similar to those
of total housework. As it was the case in ??, we cannot reject the hypothe-
sis that the female wage coefficients are not different from each other, which
would be consistent with the existence of constraints in the division of house-
hold labor.
28
Wife's share of housework time - Non-daily [1] [2] [3] [4] [5] [6] [7]wife's net monthly earnings 500-999 -0.11 -0.09 -0.09 -0.11 -0.11 -0.13 -0.12
Observations 428.00 428.00 428.00 425.00 425.00 425.00 425.00R-squared 0.07 0.07 0.07 0.09 0.10 0.13 0.21Absolute value of significant t statistics in parentheses* significant at 10%; ** significant at 5%; *** significant at 1%
Table 5.2.b: Social Constraints in the Division of Housework. Daily vs. Non-daily Activities
29
However, for non-daily activities we observe that the coefficients on the
female wage are very different from those in table ??. A significant coeffi-
cient of -.11 denotes a big drop in the share when women move up from less
than 500 to between 500-1000 euros per month. The coefficients on higher
earnings are not significant, except for a significant coefficient of .3 for earn-
ings between 2000-2500. This would not suggest the existence of a social
constraint in housework for these activities.
Weekdays vs. weekend One would expect that individuals prefer to
postpone housework activities when possible to those days when there is
more time availiable.16 Thus, one would expect that during the weekend
individuals perform those types of housework that are more easily postponed
and not so important. In other words, one would expect that for housework
performed during the weekend there are credible threats availiable, and thus
commitment problems are less important. If activity k is performed during
the weekend and activity j during the week, thenHk > Hj and thus wk > wj
for any household i. That is, the constraint binds for higher levels of female
wages for activity k (weekend) than what it does for activity j (week). We
test this hypothesis by running equation 1 separately for week and weekend
housework shares.
Table 5.2 presents the amount of housework performed during the week-16This is true since the marginal utility of leisure is greater in those days when time
for leisure is scarce, such as working days. Thus, it is expected that individuals pospone
housework time during the week to the weekends in order to maximize their utility.
30
end by both wives and husbands, as well as the wife’s housework over total
housework ratio. We observe that when compared to 5.2 the amount of time
devoted to household activities is greater for almost all activities during the
weekend. However, the degree of specialization as measured by the ratio of
interest is slightly lower during the weekend than during the week.
Observations 804.00 804.00 804.00 798.00 798.00 798.00 798.00R-squared 0.06 0.07 0.12 0.16 0.16 0.18 0.21Absolute value of t statistics in parentheses* significant at 10%; ** significant at 5%; *** significant at 1%
Table 5.2.c: Social Constraints in the Division of Housework. Weekend vs. Weekday Activities
33
Caring activities Table 5.2 presents the results of running equation 1 for
caring activities. We observe that the earnings coefficients are all very close
to zero and insignificant for almost all the specifications, with magnitudes
ranging from half percentage to 1 percentage points. This would suggest that
the constraint is binding overall the wage distribution, which is consistent
with the model.
34
Wife's share of caring time- Care [1] [2] [3] [4] [5] [6] [7]wagewife999 0.00 0.03 0.02 0.01 0.00 -0.02 -0.01
Observations 1885.00 1885.00 1885.00 1875.00 1875.00 1875.00 1870.00R-squared 0.01 0.01 0.14 0.14 0.17 0.55 0.56Absolute value of t statistics in parentheses* significant at 10%; ** significant at 5%; *** significant at 1%
Table 5.2.d: Social Constraints in the Division of Housework, Caring vs. Non-Caring Activities
35
6 Conclusion
Economic theories of the household predict that increases in female human
capital lead to increases in female labor force participation and, symmet-
rically, to decreases in the female time devoted to household production.
However, both at the longitudinal and cross-sectional level we observe that,
despite the decline in the wage gender gap, specialization in home produc-
tion continues to be high with women providing most of household produced
goods and services. We develop a model that recognizes the inefficiencies
inherent to household production activities. We test the model using the
2001 Spanish Time Use Survey, a time diary survey with information on the
time devoted to household production activities by both partners. Empirical
findings support the proposed model inefficient allocation of household time.
36
References
Akerlof, G., and Kranton, R. (2000), “Economics and Identity”, Quar-
terly Journal of Economics, Vol. 115, Issue 3: 715-753.
Apps, P. (2003), "Gender, Time Use and Models of the Household", IZA
Discussion Paper No. 796.
Apps, P. and Rees, R. (1997), "Collective Labor Supply and Household
Production", The Journal of Political Economy, Vol. 105, No. 1, 178-
190.
Apps, P. and Rees, R. (1996), "Labour Supply, Household Production
and Intra-FamilyWelfare Distribution", Journal of Public Economics
(60), 199-219.
Alvarez and Miles (2003), ”Gender effects on Housework Allocation:
Evidence from Spanish two-earners couples”, Journal of Population
Economics 16: 277-242.
Basu, Kaushik (2004) "Gender and Say: AModel of Household Behavior
with Endogenously-determined Balance of Power", Harvard Institute
of Economic Research Discussion Paper No. 2054.
Becker, G. (1965), ”A Theory of the Allocation of Time”, Economics
Journal 75: 493-517.
Becker, G. (1973), ”A Theory of Marriage: Part I”, The Journal of
Political Economy, Vol. 81, No. 4, pp. 813-846
Becker, G. (1974), ”A Theory of Marriage: Part II”, The Journal of
37
Political Economy, Vol. 82, No. 2, Part 2: Marriage, Family Human
Capital, and Fertility, pp. S11-S26.
Becker, G. (1991), "ATreatise on the Family", Harvard University Press,
Cambridge, Massachusetts.
Becker and Murphy (2000), Social Economics: Market Behavior in a
Social Environment . Cambridge, MA: Belknap Press.
Bertrand, Luttmer and Mullainatham (2000), "Network Effects and
Welfare Cultures", Quarterly Journal of Economics 115:3.
Bittman, M., England, P., Folbre, N., Matherson, G. (2002), “When
Gender Trumps Money: Bargaining and Time in Household Work”,
MIMEO.
Blau Francine D., and Kahn, Lawrence M. (1992), "The Gender Earn-
ings Gap: Learning from International Comparisons", The American
Economic Review (May 1992) 553-538.
Bryant, W.Keith, The Economic Organization of the Household, Cam-
bridge University Press 1995.
Browning and Chiappori (1997), "Efficient Intra-Household Allocations:
A General Characterization and Empirical Tests", Econometrica,
Vol. 66, No. 6. (Nov., 1998), pp. 1241-1278.
Chiappori, P-A (1992), “Collective Labor Supply and Welfare”, The
Journal of Political Economy, Vol. 100, Iss. 3.437-467.
De Laat, J. and Sevilla-Sanz, A. (2004). "Working Women, Husband’s
38
Home Time, and Lowest-Low Fertility in Europe" (mimeo).
Drewianka, Scott (2003), "Estimating Social Effects in Matching Mar-
kets: Externalities in Spousal Seach", The Review of Economics and
Statistics, 85(2): 409-423.
"How Europeans spend their time Everyday life of women and men".
Luxembourg: Office for Official Publications of the European Com-
munities, 2004 (ISBN 92-894-7235-9). European Communities, 2004.
Fella, Giulio, Mariotti, Marco, Manzini, Paola (2004) "Does Divorce Law
Matter?", Journal of the European Economic Association, 2(4):607—
633
Fernández, Raquel, Fogli, Alessandra and Olivetti, Claudia (2004),
"Mothers and Sons: Preference Formation and Female Labor Force
Dynamics"Quarterly Journal of Economics, Volume 119, Number 4.
Folbre N. and Bittman, M (2004), "Family Time: The Social Organiza-
tion of Care", Routledge, New York, NY.
Gornick, Janet C., Meyers, Marcia K (2003), "Families That Work:
Policies for Reconciling Parenthood and Employment", New York :
Russell Sage Foundation.
Hersch, J. and Stratton, L.S. (2002), "Housework and Wages", Journal
of Human Resources, Vol. 37 Issue 1, p217
Hersch, J. and Stratton, L.S. (1994), "Housework, Wages, and the Di-
vision of Housework Time for Employed Spouses", The American
39
Economic Review, Volume 84, Issue 2, Papers and Proceedings of
the Hundred and Sixth Annual Meeting of the American Economic
Association, 120-125.
Juster, T. and Stafford, F. (1991), "The Allocation of Time: Empir-
ical Findings, Behavioral Models, and Problems of Measurement",
Journal of Economic Literature, Vol. 29, No. 2, pp. 471-522.
Lundberg, S. and R.A. Pollak, (2001), “Efficiency in Marriage,”NBER
Working Paper 8642.
Lundberg, S. and R.A. Pollak, (1993) Separate Spheres Bargaining and
the Marriage Market, Journal of Political Economy, 101, iss. 6, pp.
988-1010
Manski, Charles F. (1993), Identification of Endogenous Social Effects:
The Reflection Problem, The Review of Economic Studies, Vol. 60,
No. 3, 531-542.
Manski, Charles F., (2000), "Economic Analysis of Social Interactions",
Journal of Economic Perspectives, Volume 14, Number 3, pp. 115-
136.
Mazzocco, Maurizio (2003) "Individual Euler Equations rather than
Household Euler Equations", Working Paper , University of Wis-
consin.
McElroy, M. and M.J. Horney (1981), “Nash-Bargained Household De-
cisions: Toward a Generalization of the Theory of Demand”, Inter-
40
national Economics Review, Vol. 22, Issue 2 pp. 333-349.
Morgan, S.Philip (2003), ”Is Low Fertility a Twenty-First-Century De-
mographic Crisis?”, Demography, Vol. 40, Number 4 pp 589-604.
Rasul, Imran, Household Bargaining Over Fertility: Theory and Evi-
dence from Malaysia, working paper, London School of Economics,