DISCUSSION PAPER
"Z71?V
Report No.: 62
DETERMINANTS OF WOMEN'S TIME ALLOCATION
IN RURAL BANGLADESH
by.
Shahidur R. Khandlker
February 1987
Research Unit
Agriculture and Rural Development Department
Operational Policy Staff
World Bank
The views presented here are those of the author(s), and they should not
be interpreted as reflecting those of the World Bank.


The author is a consultant of the World Bank.
However,the World Bank does not accept responsibility for the views
expressed herein which are those of the authors and should not be
attributed to the World Bank or to its affiliated organizations.     The
findings, interpretations, and conclusions are the results of research
supported in part by the Bank; they do not-necessarily represent offi-
cial policy of the Bank.    The designations employed and the presenta-
tion of material in this document are solely for the convenience of
the reader and do not imply the expression of any opinion whatsoever
on the part of the World Bank or its affiliates concerning the legal
status of any country, territory, area or of its authorities, or con-
cerning the delimitation of its boundaries, or national affiliation.


DETERMINANTS OF WOMEN'S TIME ALLOCATION
IN RURAL BANGLADESH*
Shahidur R. Khandker
I. INTRODUCTION
Attempts are increasingly being made to adapt the economic theory
of the household, pioneered by Becker and Gronau, to the examination of the
time allocation behavior of rural households in developing countries.1 The
time allocation patterns of women in these countries, however, has received
particular attention.. This is perhaps because of two factors. The first
is whether women's time allocation is governed by economic constraints as
dictated by the economic theory of the household, or is determined
exclusively by local customs such as "patriarchy" as proposed by an
alternative theory.2 The second factor    -- since many women in developing
countries generate cash income from family enterprises -- focuses on
whether the women in the developing countries have a three-way or two-way
choice structure for work decision. The three-way choice structure is
whether "to work inside" or "to work outside" the family enterprise for
cash income, or "not to work" for cash income. This choice structure is
significantly different from the two-way,choice structure -- whether "to
work" or "not to work" for cash income outside the home -- which is typical
of women in more developed countries. The need to identify.the choice
structure is of concern because the observed choice structure of women's
work in the developing countries may warrant a more general economic theory
than the one developed for the women working in the developed countries.


-2-
Time budget studies from rural Bangladesh show that women
work for (a) nonmarket production, (b) cash income production from
employment in a family enterprise as unpaid family labor and (c) cash
income production from paid employment outside the family.     An econometric
study on the three-way choice structure of women's work in rural Bangladesh
found that this structure is not statistically appropriate.5 Instead, as
this study shows, a two-way choice structure (i.e., whether "to.work" in
the labor force, consisting of self-employment in a family enterprise and
paid employment in nonfamilial market production, both producing cash
income, or "to work" exclusively for home production) is more appropriate
for describing women's work patterns in Bangladesh. Thus, the transactions
costs of switching from self-employment in a family enterprise to market
wage production are not significant enough that ywomen in rural BAngladesh
find it difficult to substitute self-employment for market employment.6    A
woman's decision regarding her participation in the labor force may,
however, be endogenously determined.' This endogenity may cause important
selectivity bias in estimates of women's time allocation decisions.
The objective of this paper is to estimate a time allocation
model for Bangladeshi rural women where both women's labor force
participation decision and hours of work are jointly determined. For
policy purposes it is more beneficial to examine the factors influencing
time allocation of women rather than to note simple presence of women in
one type of work category. Moreover, this study will identify whether
sample selection bias is important for estimating women's time allocation
in nonmarket production, an important category of work where both groups of
women spend considerable amounts of their productive time. The study hopes


-3-
to shed light on the determinants of the economic roles of rural women in
Bangladesh and in similar developing countries.
The paper is organized as follows:   Section two outlines an
economic framework that shows how labor force participation decision and
hours of work are jointly determined and how participation decisions
affect women's time allocation decisions. This framework is used to derive
an econometric model of women's time allocation in Bangladesh. Section
three presents the survey data used to empirically implement the model.
Section four reports the results of the econometric estimates of the
model. The final section summarizes the results and presents policy
implications.


-4-
II. WOMEN'S TIME ALLOCATION AND LABOR FORCE PARTICIPATION:
AN ECONOMIC MODEL
The benefits of using the household unit as an organizing concept
in studying the econoT#c behavior of economic agents have been
established. The household unit is viewed as making collective decisions
that determine the behavior of each individual woman.
Assume that a household consisting of only husband (H) and wife
(w) derives utility from home produced goods (N), market goods (Z), and
leisure (Li, i=H,W) of the husband and wife. The production of
home-produced goods (N) requires resources, however. Hence, as the
household attempts to maximize utility subject to the constraints it faces,
home-produced goods competes with other goods, including the husband and
wife's leisure. In addition.to a time constraint, the household faces
market and community constraints. These constraints affect the transaction
costs of the goods that the household produces as substitutes for market
goods <nd services.
More formally, the household is assumed to maximize the following
utility function (1):
(1)       U = U (Z, N, LH, LW)
Assume that the home-produced good (N), is produced by a
combination of market-purchased inputs (X), and the time input of only the
wife (TWN). Thus, the nonmarket production of household is given by the
relation:
(2)       N =N (X, TWN)


-5-
The household faces two separate time constraints, (3a) and (3b),
which state that the amount of time spent by each member on market work,
home production, and leisure cannot exceed his or her total amount of time
(-si, i=H,W) available:
(3a)       H = THM + LH
(3b)      Rq = TWM   TWN + LW
The household also faces a budget constraint (4) which states
that expenditure on market goods, used as inputs in household nonmarket
production, cannot exceed the family income (I):
(4)       PZZ +PxX WHTHM +WWTWM + V = I
where Wi (i=H,W) is the market wage rate of the husband and the wife,
the amount of time he or she spends producing the market good, V the
non-earned household income, and Pz and Px, respectively, the prices of-
market-purchased inputs Z and X.
Furthermore, for simplicity, assume that the production of home
good (N) is of a fixed-proportion type. This implies that inputs in
nonmarket production vary proportionately with the amount of home
goods produced:
(5a)      XN = yN,
(5b)     TNW = SN,
where y and 6 are, respectively, the marginal productivity of
market-purchased input (X) and the wife's time (TWN) in the production of
the home goods. Now, using the relations in (2), (3), and (5) into (4), we
rewrite the budget constraint as:


-6-
(6)    PZZ + TNN + WHLH + Wy4LW = WHQR +       + V
where I* is the household's "full income", defined as the income the
household earns if it decides to devote all of its time to market wbrk, and
1rN = YPx + SWw is the shadow price for home-produced goods (N).
Since the shadow price is a function of market prices, it follows Lhat a
change in market prices, such as the wife's market wage rate will have both
direct effects, and indirect effects through the shadow price on the level
of demand for the goods.
However, the wife's participation in market work is endogenous in
that the household unit may decide on whether the woman participates in the
labor market. If the wife does not participate, then the household faces
an additional constraint:
(7)       TWM = 0.
The maximization of utility function (1) subject to (6) and (7)
yields the following optimum conditions:
(8a)      UZ = XPz
(8b)      UN = X(N -6),
(8c)      ULH    XWH,
(8d) ULW = X(WW - 1)
(8e)      I* - PZZ - 7NN - WHLH - WWLW = 0,
(8f)      (QW - LW - 6N) = 0,
where X is the marginal utility of income (positive), and p is the ratio of
marginal disutility of market work to the marginal utility of income
(negative). However, since the constraint (7) is not binding, p is zero if
the wife decides to participate in the labor force.


-7-
The optimum conditions for the home-produced good (N) and the
wife's leisure (LW) vary according to whether she participates in the
labor force. A woman who participates in the labor force equates her
marginal utility of leisure with the "imputed" market price she gets from
market employment. in contrast, a woman who does not participate in the
labor force equates her marginal utility of leisure with the "imputed"
"virtual" or shadow wage, WW* = (WW - ) of her labor in
self-employment. This is defined as the market wage plus the value
assigned for her not supplying labor to market work.
For the woman who does not participate as well as for the woman
who does participate in the wage market, the marginal utility of a
home-produced good, such as a child, is equated with the "imputed" value of
its shadow price. However, a nonparticipant woman evaluates the shadow
price differently than does a participant woman. A woman who participates
in the wage market evaluates the shadow price of home-produced good in
terms of the market wage, while a woman who does not participate evaluates
it in terms of the "virtual" price ,f her labor. Thus, a woman's time
allocation is very much influenced by whether she participates in the labor
force.
As outlined, the model yields a system of reduced-form equations
for endogenous variables as functions of individual, household, market, and
community constraints that set prices for goods and services that the
household produces and consumes. In most cases, the effects of these
constraints have indeterminant signs for most of the home-produced goods.7
Of particular interest in this paper are the reduced-form equations for a


-8-
woman's time allocation in home and market productions, which are the
following.
(9a)      TWN~ = Tj (NH, WW, -y, 6, V, PX, Pz, QH,
(9b)      T1JM = TWM (WH, WW, y, 6, V, PX, Pz, Ri    )
Note, however, that for a woman who does not participate in the wage
market:
(9b')     TWM = 0
Explanatory Variables
Different reasons can justify the inclusion of market prices,
income earning (individual and household) assets, and individual- and
household-level constraints as explanatory variables in women's time
allocation.
Market-determined prices are the wages of two categories of
labor: adult male and adult female. An increase in the male wages will
induce an increase in the supply of the husband's -labor to market work.
This will induce a corresponding decrease in the wife's time allocation and
participation in market work, if husband and wife are substitutes in
production. Conversely, an increase in female wages, other things being
equal, is expected to increase the wife's labor market participation and
her time allocation to market production. However, if income effects from
an exogenous wage increase dominate the substitution effects, a backward
bending supply curve of labor to market work may result.
If schooling measures some form of child quality, household
distance to the nearest school will stand as a proxy variation in the price


-9-
of child quality.   Moreover, if a trade-off between child quality and
child quantity exists, an increase in the school distance represents an
exogenous increase in the price of schooling and decreases the quality per
.child. This induces a corresponding decrease in the shadow price of child
quantity, thereby increasing the number of children in the family.
Furthermore, if the children and the mother are substitutes in production,
this implies that the children can substitute the mother's time in work at
home; thus releasing the mother for market work.8
Household distance to a town, or market center, may serve as
proxy variation in:   (i) the transactions cost in purchasing goods for
which the household can produce substitutes and (ii) the costs of
market-oriented employment. Thus, an increase in the household distance to
market center means a higher transactions cost for market-purchased goods
and also higher employment costs for market-oriented employment. Together
these may imply a decrease in the probability of a woman's participation in
the labor force.
An individual characteristic -- education level -- can be treated
as an explanatory variable that may indicatq productivity potential, both
at home and in market productions. Holding wages constant, An increase in
the level of a woman's education can increase the probability of a woman's
participation and time allocation in market production, if it increases her
opportunity costs for staying at home. This may also imply an increase in
a woman's time allocation in home production if education increases her
home productivity. Thus, the ultimate effect of an increase in a woman's
education level is indeterminant. In contrast, a man's education level and


- 10 -
the amount of resources he brought to the marriage affect a woman's time
allocation by allowing her more time for home production and less time for
market production. This may occur due to the positive wealth effects of
potential earnings influenced by physical and human capital endowments.
The household's assets variable, such as landholdings, may act as
proxy for productive household assets which exert both a price effect and
an income effect on a woman's time allocation. The price effect would
raise the marginal product or "shadow wage" of the woman's labor, while the
income effect would encourage the household to consume more of the woman's
leisure even at its given opportunity cost.
A single cross-section of household observations is unlikely to
contain direct price variation of a substantial degree. However,
geographical variations of different community-context variables (e.g.,
household proximity to market center or school) can provide variation in
the "implicit" price for many goods and services that households produce
for their own consumption. Thus, we can assess the effectiveness of
government intervention programs from a data analysis by inferring
information about the responsiveness of households to changes in prices
which are relevant to household nonmarket production and consumption
decision and which have been exogenously varied through government
intervention programs.


- 11 -
III. THE DATA AND ITS CHARACTERISTICS
The data for.this paper is drawn from a more expansive sample of
500 households in rural Bangladesh. The sample was randomly drawn from
eight Upazilas selected from regions to the north, east, and west of Dhaka
as well as from a central part of Bangladesh. Because of the geographical
variations, we expect price variations in this dataset.
Data on time allocation of women is classified under two
categories: home production and market production. Market production
includes work both inside and outside the home for cash income rroduction.
Our data analysis focuses on married women between the ages of 15 and 49.
This limited our sample from 500 to 444 housewives. Among these 444 women,
195 women participate in the market-oriented production, while the
remaining 249 women worked only in home production.10
Table 1 present the means and standard deviations of the
dependent and independent variables for the two groups of women.    As the
Table suggests, time use in home production is lower for women
participating in market production than for women who work only in home
production. This is, however, not unexpected. More interesting are the
characteristics that distinguish the two groups of women. For instance,
women who participate in market-oriented production are younger than women
who do not participate. Women who participate in market production also
live closer to a major market center than do women who do not participate.
Participating women (and their spouses) also have much more schooling than
do women who do not participate in market production. Since schooling
levels are much higher for the participant women and their spouses, the


- 12 -
Table 1:
Mean and Standard Deviation (S.D.) of Variables
(444 observations)
Nonparticipant        Participant
Variable name                           Mean       S.D.       Mean     S.D.
Endogenous variables:
Woman's time allocati-n
in home production                   278.2      46.66      134.5    64.16
Woman's time allocation
in market work                         -         -        213.8     63.17
Exogenous variables
Woman's age (in years)                   33.75      8.36       29.95    6.06
Woman's education (years)                 2.32       3.15       6.72    5.01
Husband's education (years)               5.15      4.62        8.43    5.01
Husband's premarriage asset
(in '000 current Taka)                48.84     75.60       17.98   33.26
Land holding (hectare)                    1.23       1.12       0.92    1.70
School distance (km)                      1.03       1.13       1.94    3.56
Distance to town (km)                     6.94      3.52        4.82    4.55
Predicted male wage (Taka)               33.21      11.33      40.98   12.57
Predicted female wage (Taka)             19.48       7.63      30.36   12.68
Sample size                             249                   195
Note: Nonparticipant group includes those women who do not participate in
the labor force. In contrast, participant group includes those
women who participate in the labor force. Time allocation data is
expressed in yearly mandays. Distance variables are calculated on
the basis of individual household's responses of its distance to the
nearest school and nearest major market center. Market center in
most cases refers to the Upazila headquarter, the source of all
institutional services. Predicted wages of husband and wife are
based on the estimates of separate wage functions for the subsample
of men and women who participate in the labor force for paid job.


- 13 -
predicted wage offers for both are much higher for the participant group
than for the nonparticipant group.11    However, the husband's premarital
assets are almost three times higher for women who do not participAte in
market production than for women who do. Thus, substantial differences
exist across these two groups of women, both in terms of time allocation
and specific characteristics. A woman's decision to participate in the
labor market influences her time allocation in home production differently
than from a woman who decides not to participate. This is, however, an
empirical issue which we now address.


-14-
IV. EMPIRICAL RESULTS
Table 2 presents the probit estimates of labor force
participation for women, while Table 3 presents the results of time-use (in
man-days) for women in home production. All women in the sample allocate
their time in home production so that time-use data for all women in home
production is used to first estimate this time allocation equation.
However, since whether women participate in nonfamilial market work is an
endogenously determ'ined decision, the estimates of home time allocation
equation may suffer from possible sample selection bias.12 If the
unmeasured characteristics influencing sample selection and home time
allocation are jointly normally distributed, then the influence of the
selectivity associated with participation in nonfamilial production can be,
"taken out" of the home time allocation estimates. This is done by first
estimating the participation equation for the whole sample using probit,
and then including the Mills-ratio estimate associated with the
self-selected subsample in the home time allocation equation. Thus, the
inclusion of the Mills-ratio variable in the home time allocation equation
not only eliminates the selectivity effects of participation in nonfamilial
production, but its coefficient also provides a consistent estimate of the
covariance between the unmeasured characteristics in the home time
allocation and participation equations.
Although the results in Table 2 suggest that women's labor force
participation is a decision endogenously determined, this endogeneity does


- 15 -
Table 2
Probit Maximum Likelihood Estimates of
Women's Labor Force Participation
Exogenous                        Parameter
Variables                        estimate               t-Statistic
Intercept                            -3.502                   -0.673
Woman's age                          -0.073                   -1.066
Woman's education                    -2.142                   -1.567
Husband's education                   0.780                     4.023*
Landholding                          -0.661                   -4.103*
Husband's premarriage assets         -0.014                   -3.591*
Distance to schooling                 0.175                    4.266*
Distance to town                     -0.032                   -0.854
Predicted male wage                  -0.313                   -4.345*
Predicted female wage                 0.906                     1.789*
(-2.0)x Log-Likelihood ratio        211.121
Sample size                         444
Note:   Dummy dcpendent variable = 1, if a woman participates in the labor
force (195 observations), 0 (249 observations) otherwise. Asterisk
refers to significance level of 10 percent or better.


- 16 -
not, however, yield any significant difference in home time-use behavior
between women who do and women who do not participate in the labor market.
This is evident in Table 3; the Mills-ratio coefficient is not significant
in the equation for either participant or nonparticipant group.    Both
groups behave similarly when using their time in home production. This
implies that in rural Bangladesh a woman's time-use in home production
bears a fixed-proportional relationship to the home-produced goods such
that sample selection bias in home time allocation is not important.13 it
follows, therefore, that the OLS estimates of home time allocation equation
that includes samples of both participant and nonparticipant women are the
appropriate estimates of women's time allocation in home production.
In contrast, time allocation to market production is clearly
subject to sample selection bias. Since time-use in market production is
only positive for women who participate, while zero for those who do not
participate, time allocation in a market-oriented production is a censored
variable. Ia this case, a censored regression technique (i.e., a Tobit
method) is used to estimate women's time allocation in market production.1
The censored regression results, presented in Table 4,suggest that women's
time allocation in market production is well explained by individual,
household, market, and community constraints as dictated by the economic
theory of the household.
The predicted male and female wages included in the regression
sets are based on separate regressions (not shown in the paper) of wage
offers for each group of participant men and women in terms of age,
schooling, and household distance to the market center.15 Apart from its


- 17 -
Table 3: Estimates of Women's Time Allocation in Home Production
Exogenous              OLS estimates       OLS estimates         OLS Estimates
Variables              of all samples    of nonparticipants      of participants
Intercept                 80.248  (0.336)   -943.892  (-3.816)     351.363  (1.124)
Woman's age               -2.192 (-0.704)    -17.283  (-5.065)*      1.599  (0.391)
Woman's education        -14.233 (-0.244)   -355.186  (-5.292)*     10.539  (0.129)
Husband's education      -19.524 (-2-133)*    18.093   (1.521)      26.178  (1.712)*
Husband's
premarriage assets       0.122  (2.111)*     0.022   (0.390)      -0.255 (-1.509)
Landholding               .20.552  (2.761)*    0.976    (0.103)    -21.642 (-1.693)*
Distance to town           5.677  (3.299)*     8.078   (4.656)*      1.57-2 (0.728)
Predicted male wage        8.330  (2.463)*    -6.833  (-1.492)      -9.632 (-1.625)*
Predicted female wage      1.667  (0.073)    138.589   (5.236)*     -5.357 (-0.163)
Mills' Ratio                                   9.714   (0.393)      38.192  (1.195)
R2                         0.3642              0.1523                0.2438
Sample size              444                 249                   195
Note: t-Statistics are in parentheses. Asterisk refers to significance level of
10 percent or better.


- 18 -
Table 4: Censored Regression (Tobit) of Women's Time Allocation in Market
Production
Exogenous variables           Parameter estimate        t-Statistic
Intercept                            -238.741                 -0.393
Woman's age                            -6.922                 -0.846
Woman's education                    -215.386                 -1.381
Husband's education                    90.435                  4.113*
Husband's premarriage assets           -0.899                 -4.568*
Landholdings                          -79.344                 -4.568*
Distance to schooling                  17.163                  3.108*
Distance to town                       -8.745                 -1.791*
Predicted male wage                   -36.733                 -4.453*
Predicted female wage                  93.709                   1.632*
Log-likelihood                      -1373
Estimation method                         maximum likelihood method
Sample size                           444
Note: Asterisk refers to significance level of 10 percent or better.


- 19 -
significant effects on market productivity, women's education seems to have
negligible effects in both the participation and the time allocation
equations. In contrast, men's education appears to have significant
positive effects both for women's labor force participation and time
allocation in market production.16
Women's wage and the household distance to school that increase
the probability of women's participation in the labor force have
significant positive effects on women's time allocation in market
production, and negative effects on time allocation in home production.
Thus, the factors such as female market wage that increase a woman's market
productivity can increase a woman's participation as well as time
allocation in labor market production. An increase in the household
distance to school means less school enrollment of children, perhaps
indicating higher participation and higher time allocation of women in
market production, if the children can substitute for the mother's time in
home produc;ion.17
On the other hand, variables such as a husband's premarital
assets, landholding, male market wage, and the household distance to market
center decrease women's participation in market-oriented production. These
have significant negative effects on a woman's time-use in market
production, and positive effects on her time-use in home production. These
results are consistent with a priori expectations. Thus, the factors such
as a man's premarital assets, landholding and male wage that reflect the
husband's earnings or wealth effects increase the wife's time allocation in


- 20 -
home production. In contrast, an increase in distance to market center
increases the transactions cost of market-oriented production, thereby
reducing the probability of a woman's participation and hence, time
allocation in market production.


- 21 -
V. Policy Implications and Conclusions
Determinants of women's work patterns in rural households need to
be identified for public policy programs that seek to improve the
well-being of the rural population. Since women, who constitute more than
half of the total population, are directly involved in the production of
both market and nonmarket goods, their participation in development
activities through their participation in market-oriented production may be
a precondition for economic and social development in a country such as
Bangladesh. Understanding the individual, household, market and community
factors that cause variations in women's time allocation may lead to more
successful government policies aimed at involving rural women in
development activities in a rural economy.
This paper investigates one important aspect of women's time-use
patterns in rural households, namely, whether women's time allocation in
Bangladesh is governed by economic constraints or is determined exclusively
by local customs such as "patriarchy". An economic model developed by
Becker and Gronau is utilized to test this alternative hypothesis.
However, to test the hypothesis, one needs to identify the number of
independent categories of work that a rural woman can participate in. This
is necessary for two reasons. First is the argument for possible
self-selection bias in estimates of a time allocation model to the extent
that a woman's presence in any distinct category of work is endogenously
determined at the household-level. The second is a possibility that a
three-way choice structure, if-identified, may oppose a two-way choice


- 22 -
structure, and consequently, may warrant a general economic theory for
women working in a country such as Bangladesh than the one developed for
the women working in the developed countries.
Based on the empirical evidence that suggests that a two-way
choice structure is more appropriate for Bangladeshi rural women, this
paper utilizes the economic theory developed within a two-way choice
structure to analyze the women's time allocation behavior in rural
Bangladesh. The results indicate that a woman's time-use pattern is not
fixed exclusively by the society but is partially influenced by individual-
and household-level economic constraints. This suggests that the
alternative hypothesis that the women's time allocation in rural Bangladesh
is inflexibly fixed by Local customs can be rejected. This also implies
that women in rural Bangladesh respond to market opportunities when such
opportunities are available.   In other words, it is not the "patriarchy"
but the limited market opportunities that restrict women from working
either inside or outside a family enterprise for producing cash income.
Therefore, if economic development proceeds in rural Bangladesh, it is
Likely that more women will increasingly participate in the market-oriented
production even if "patriarchy" exists in rural Bangladesh.18
Sample selection bias due to endogeneity of female participation
in the Labor force does not affect the estimates of women's time allocation
in home nonmarket production. This contradicts a priori expectation that
a Labor market participant woman and a nonparticipant woman may behave
differently in using their time in nonmarket production.    This finding


- 23 -
perhaps implies that a woman's time allocation in home prodtcetion may bear
a fixed-proportional relationship such that a woman's time allocation in
market production does not affect her time allocation in home production.
Among the important determinants of women's time allocation in
Bangladesh, it is the woman's spouse education that affects quite
significantly both a woman's labor market participation decision and her
time allocation in either market or nonmarket productions. In contrast, a
woman's education, apart from its significant effects on market
productivity, is not an important determinant for her time allocation. The
finding that a woman's time allocation is determined not by her own
endowment, but partially by her spouse's endowment, does not necessarily
reflect that a husband works as a "patriarch" who determines exclusively
his wife's time allocation in Bangladesh.   To the extent that an educated
woman is married to an educated man, this result may simply reflect a
spouse selection bias due to the endogeneity of the wife's and husband's
education levels determined partially by the parental characteristics of
the wife and the husband.19
An increase in women's market wages increases women's
participation in labor market production, a finding consistent with
expectations. Similarly, modernization programs such as urbanization that
increase market opportunities for women to participate in the
market-oriented production also.increase women's time allocation to labor
market production. In contrast, a husband's physical (as opposed to human)
endowments such as premarital assets or current landholdings can increase


-24-
a woman's home productivity and, hence, keep women away from
market-oriented production, perhaps due to their induced wealth effects on
women's time-use.
Government provision of schooling in the form of reducing
schooling costs for children may increase the quality and reduce the
quantity of children. However, since women and children are substitutes in
current production, this reduces women's participation in market-oriented
production. This need not be the case, however. As children, particularly
female ones, receive more education following a decrease in schooling
costs, they may choose more market work as adults. As this occurs, it is
likely that at some point children will stop being substitutes for women in
home-production, and instead become pure consumption goods (requiring more
home production by the mother). This would mean that more education of
today's female children will mean more education for the mother which in
turn will mean fewer children and .also more participation of mothers in
labor market production.20 This is currently observed in more developed
countries, and evidence suggests that this "pure consumption good" effect
of children is also present in Bangladesh, although at present it is
dominated by the children's labor substitutability effect.
It is indeed important for policy purposes to distinguish the
factors that influence women's labor market participation from the factors
that keep women away from market-oriented activity. This exercise clearly
indicates that it is not the "patriarchy" per se that restrict women from
working in labor market production. The other factors, such as limited


- 25 -
market opportunities, that may partially keep rural women in Bangladesh
away from market-oriented production. From this perspective, the relevant
policy variables of particular interest are clearly the market
interventions concerning women' s wages, public delivery system of
schooling, particularly for the female population, and above all,
modernization programs that increase market opportunities for women.
However, modernization programs and market interventions may yield more
immediate and direct results than the female schooling delivery system.
But, aevertheless,. it is the human capital investment rather than the
market interventions that can increase more sharply women's labor market
productivity and hence, labor market participation, and to this effect,
investment in female schooling is a necessary long-run policy objective
that deserves particular attention from the policymakers.


- 26 -
NOTES
*-- Views.,expressed-in this paper are those of theAuthor, and do not
necessarily reflect views of the World Bank.   T
comments  and Paul.McGuire for computer- assistancei *.Tim was  r-'sta
1. Gary S. Becker, "A Theory of the Allocation of Time."     Economic Journal
(September, 1965): 493-517; Reuben Gronau, "The Intrafamily Allocation
of Time: The Value of the Housewives' Time." American Economic Review
(September, 1973): 634-651; Robert Evenson, "Time Allocation in Rural
Philippine Households." American Journal of Agricultural Economics
(1978): 322-330.
2.  An alternative theory as opposed to the economic theory of the
household is given in Mead Cain, S.R. Khanam, and Shamsun Nahar,
"Class, Patriarchy, and Women's Work in Bangladesh." Population and
Development Review (1979): 405-438. According to this theory,
powerful local norms of female seclusion due to "patriarchy" extend to
the labor market and this severely restrict women from working outside
the family.
3.  The concern essentially is what makes it important about the
distinction between women's self-employment and market employment for
cash income and how would the independent variables be expected to
affect one differently than the other. Before addressing this concern,
however, one needs to identify whether the three-way choice is at all a
valid categorization of women's work patterns in a given country
context, since it is likely that validity of such a choice structure
may vary from country to country.
4.  S.R. Khandker, "Women's Role in household Productive Activities and
Fertility in Bangladesh." Center Discussion Paper No. 488 (New Haven:
Yale University, Economic Growth Center, 1985), Journal of Economic
Developement (forthcoming).
5.  S.R. Khandker, "Labor Market Participation of Married Women in
Bangladesh." Center Discussion Paper No. 514 (New Haven: Yale
University, Economic Growth Center, 1986).
6. The high substitutability between self-employment and market wage
employment may imply that women's market-oriented production in rural
Bangladesh remains at its pre-modern level such that a woman does not
find it difficult to switch from self-employment to market employment.


- 27 -
This contrasts the findings of a study which shows that a woman in
Japan finds it difficulc to switch from self-employment to either home
production or market wage production. See M. Anne Hill, "Female Labor
Force Participation in Developed and Developing Countries --
Consideration of the Informal Sector." Review of Economics and
Statistics (1983): 459-468.
7.  See, for example, Gronau for comparative static results of some
exogenous variables.
8. Another partial effect of an increase in school distance is to increase
the mother's time in home production for child bearing, if not for
child rearing. Therefore, whether a woman will participate in labor
market production due to an increase in school distance depends'on two
opposing partial effects.
9. Upazilas are "upgraded" Thana administrative units, covering roughly an
area of 100 square miles, and consisting of several Unions. Unions in
turn are collection of a number of villages. The Upazilas covered
under this study are respectively, Baidyerbazar (Dhaka district),
Ghatail (Tangail district), Kownia and Kotwali (Rangpur district),
Sherpur (Jamalpur district), Laksam (Comilla district), Ishardi (Pabna
district), and Gabtali (Bogra district).    For more detail of the data
collection, see Khandker (1985).
10. The original three-way choice categories of work are, respectively,
producing no cash income (i.e., working for non-market home
production), producing cash income-in family enterprise, and producing
cash income from employment outside the family. Producing no cash
income involves house chores such as cooking, cleaning, fetching water
and firewood, caring for the children and the elderly, washing and
serving food. In the rural areas of Bangladesh, it also includes
post-harvest operations of crops, such as threshing, drying, wiftnowing,
and sifting for paddy, kitchen gardening, and tending to domestic
animals. Producing cash income in family enterprise includes
activities such as family business, sewing, handicrafts, and rice
husking where a woman works either as a partner or an unpaid family
worker. Producing cash income froi employment outside the family
includes those activities that require women to work outside the home,
such as teaching, non-agricultural wage work, and other salaried
services. Although these three categories appear distinct, econometric
study shows that producing cash income in a family enterprise and
market cash employment are highly substitutable in that they are not
statistically two distinct categories of work.    See Khandker (1986).
11. The predicted wage offers are based on the estimated wage offers for
women and men working outside the home for cash using the approach
suggested by Heckman. See, James Heckman, "Sample Selection Bias as a
Specification Error." Econometrica (1979): 153-161.
12. See James Heckman.


- 28 -
13. In fact, following equation 8(b), home time allocation behavior of a
participant woman cannot be different from that of a nonparticipant
woman if p = 0 for a nonparticipant woman. Since marginal disutility
of market work is negative by assumption, U can be zero only if the
marginal utility of income of a nonparticipant woman is infinity.
14. See G.S. Maddala, Limited-dependent and Qualitative Variables in
Econometrics. (London: Cambridge University Press, 1983).
15. In addition to age, schooling and distance to market center, the mills'
ratio coefficient based on the participation equation is included in
the wage equation.   The inclusion of Mills' ratio then purges out the
selectivity effects in wage equation due to participation in cash
income production. These results suggest that education level of an
individual significantly affect his or her market productivity.
16. This finding seems to contradict a priori expectation that if the
husband's education reflects the effect of the husband's earning on
women's time-use, an increase in husband's education    acureases a
woman's labor market participation and, hence, time-use in
market-oriented production. See, Gronau, "Leisure, Home Production,
and Work -- the Theory of the Allocation of Time-Revisited." Journal
- of Political Economy (1977):1099-1123.
17. Results clearly indicate that a mother and children are substitutes in
current production in Bangladesh.
18. This is simply because even a "patriarch" cannot forego a high
opportunity cost for keeping women away from market-oriented activities
as a result of women's increasing market-opportunities with the pace of
economic development.
19. If information on parental characteristics were available, one could
have used predicted instead of actual schooling-levels of husband and
wife in the regressions that w6uld have removed the influence of spouse
selection bias from the estimates.
20. One study shows that as a woman's education increases, it reduces
fertility and also increases the woman's participation in
market-oriented production. See Khandker (1985).


- 29 -
DISCUSSION PAPERS
AGR/Research Unit
.Report No.: ARU 1
Agricultural Mechanization: A Comparative Historical Perspective
by Hans P. Binswanger, October 30, 1982.
Report No.: ARU 2
The Acquisition of Information and the Adoption of New Technology
by Gershon Feder and Roger Slade, September 1982.
Report No.: ARU 3
Selecting Contact Farmers for Agricultural Extension: The Training and
Visit System in Haryana, India
By Gershon Feder and Roger Slade, August 1982.
Report No.: ARU 4
The Impact of Attitudes Toward Risk on Agricultural Decisions in Rural
India
by Hans P. Binswanger, Dayanatha Jha, T. Balaramaiah and Donald A.
Sillers, May 1982.
Report No.: ARU 5
Behavior and Material Determinants of Production Relations.in Agriculture
by Hans P. Binswanger and Mark R. Rosenzweig, June 1982
(Revised July 22, 1985).
Report No.: ARU 6
The Demand for Food and Foodgrain Quality in India
by Hans P. Binswanger, Jaime B. Quizon and Gurushri Swamy, November
1982.
Report No.: ARU 7
Policy Implications of Research on Energy Intake and Activity Levels with
Reference to the Debate of the Energy Adequacy of Existing Diets in
Developing Countries
by Shlomo Reutlinger, May 1983.
Report No.: ARU 8
More Effective Aid to the World's Poor and Hungry:   A Fresh Look at
United States Public Law 480, Title II Food Aid
by Shlomo Routlinger, June 1983.
Report No.: ARU 9
Factor Gains and Losses in the Indian Semi-Arid Tropics: A Didactic
Approach to Modeling the Agricultural Sector
by Jaime B. Quizon and Hans P. Binswanger, September 1983
(Revised May 1984).
Report No.: ARU 10
The Distribution of Income in India's Northern Wheat Region
By Jaime B. Quizon, Hans P. Binswanger and Devendra Gupta, August 1983
(Revised June 1984).


- 30 -
Discussion Papers
(continued)
Report No.: ARU 11
Population Density, Farming Intensity, Patterns of Labor-Use and
-    Mechanization
by Prabhu L. Pingali and Hans P. Binswanger, September 1983.
Report No.: ARU 12
The Nutritional Impact of Food Aid: Critria for the Selection of Cost-
Effective Foods
by Shlomo Reutlinger and Judith Katona-Apte, September 1983.
Report No.: ARU 13
Project Food Aid and Equitable Growth: Income-Transfer Efficiency First!
by Shlomo Reutlinger, August 1983.
Report No.: ARU 14
* Nutritional Impact of Agricultural Projects: A Conceptual Framework for
Modifying the Design and Implementation of Projects
by Shlomo Reutlinger, August 2, 1983.
Report No.: ARU 15
Patterns of Agricultural Projection
by Hans P. Binswanger and Pasquale L. Scandizzo, November 15, 1983.
Report No.: ARU 16
Factor Costs, Income and Supply Shares in Indian Agriculture
by Ranjan Pal and Jaime Quizon, December 1983.
Report No.: ARU 17 -
Behavioral and Material Determinants of Production Relations in Land
Abundant Tropical Agriculture
by Hans P. Binswanger and John McIntire, January 1984.
Report No.: ARU 18
The Relation Between Farm Size and Farm Productivity: The Role of Family
by Gershon Feder, December.1983.
Report No.: ARU 19
A Comparative Analysis of Some Aspects of the Training and Visit System
of Agricultural Extension in India
by Gershon Feder and Roger Slade, February 1984.
Report No.: ARU 20
Distributional Consequences of Alternative Food Policies in India
by Hans P. Binswanger and Jaime B. Quizon, August 31, 1984.
Report No.: ARU 21
Income Distribution in India: The Impact of Policies and Growth in the
Agricultural Sector
by Jaime B. Quizon and Hans P. .Binswanger, November 1984
(Revised October 1985).


- 31 -
Discussion Papers
(continued)
Report No.: ARU 22
Population Density and Agricultural Intensification: A Study of the
Evolution of Technologies in Tropical Agriculture
by Prabhu L. Pingali and Hans P. Binswanger, October 17, 1984.
Report No.: ARU 23
The Evolution of Farming Systems in Agricultural Technology in
Sub-Saharan Africa
by Hans P. Binswanger and Prabhu L. Pingali, October 1984.
Report No.: ARU 24
Population Density and Farming Systems - The Changing Locus of
Innovations and Techni al Change
by Prahbu L. Pingali and Hans P. Binswanger, October 1984.
Report No.: ARU 25
The Training and Visit Extension System: An Analysis of Operations and
Effects
by G. Feder, R. H. Slade and A. K. Sundaram, November 1984.
Report No.: ARU 26
The Role of Public Policy in the Diffusion of New Agricultural Technology
by Gershon Feder and Roger Slade, October 1984.
Report No.: ARU 27
Fertiltzer Subsidies: A Review of Policy Issues with Special Emphasis
on Western Africa
by Haim Shalit and Hans P. Binswanger, November 1984
(Revised November 1985).
Report No.: ARU 28
From Land-Abundance to Land Scarcity:   The Effects of Population Growth
on Production Relations in Agrarian Economies
by Mark R. Rosenzweig, Hans P. Binswanger and John McIntire,
November 1984.
Report No.: ARU 29
The Impact of Rural Electrification and Infrastructure on Agricultural
Changes in India, 1966-1980
by Douglas F. Barnes and Hans P. Binswanger, December 1984.
Report No.: ARU 30
Public Tractor Hire and Equipment Hire Schemes in Developing Countries
(with special emphasis on Africa). A study prepared by the Overseas
Division, National Institute of Agricultural Engineering (OD/NIAE)
by P. J. Seager and R. S. Fieldson, November 1984.
Report No.: ARU 31
Evaluating Research System Performance and Targeting Research in Land
Abundant Areas of Sub-Saharan Africa
by Hans P. Binswanger,-January 1985.


- 32 -
Discussion Papers
(continued)
Report No.: ARU 32
On the Provision of Extension Services in Third World Agriculture
by Alastair J. Fischer (Consultant), January 1985.
Report No.: ARU 33
An Economic Appraisal of Withdrawing Fertilizer Subsidies in Indiia
by Jaime B. Quizon, April 1985 (Revised August 1985).
Report No.: ARU 34
The Impact of Agricultural Extension: A Case Study of the Training and
Visit Method (T&V) in Haryana
by Gershon Feder, Lawrence J. Lau and Roger H. Slade, March 1985.
Report No.: ARU 35
Managing Water Managers: Deterring Expropriation, or Equity as,a Control
Mechanism
by Robert Wade, April 1985
Report No.: ARU 36
Common Property Resource Management in South Indian Villages
by Robert Wade, April 1985.
Report No.: ARU 37
On the Sociology of Irrigation: How do we Know the Truth about Canal
Performance?
by Robert Wade, May 1985.
Report No.: ARU 38
-Some Organizations Concerned with Animal Traction Research and
Development in Sub-Saharan Africa
by Paul Starkey, April 1985.
Report No,: ARU 39
The Economic Consequences of an Open Trade Policy for Rice in India
by Jaime B. Quizon and James Barbieri, June 1985.
Report No.: ARU 40
Agricultural Mechanization and the Evolution of Farming Systems in
Sub-Saharan Africa
by Prabhu L;. Pingali, Yves Bigot and Hans P. Binswanger, May 1985.
Report No.: ARU 41
Eastiasian Financial Systems as a Challange to Economics:   The Adventures
of "Rigidity," with particular reference to Taiwan
by Rober Wade, June 1985.
Report No.: ARU 42
Education, Experience and Imperfect Processing of Information in the
Adoption of Innovations
by Alastair J. Fischer, June 1985.
Report No.: ARU 43
A Review of the Literature on Land Tenure Systems in Sub-Saharan Africa
by Raymond Noronha, July 1985.


- 33 -
Discussion Paoers
(continued)
ReDort No.:. ARU 44
Policy Options for Food Security.
by Shlomo Reutlinger, July 1985.
Report No.: ARU 45
Credit Markets in Rural South India: Theoretical Issues and Empirical
Analysis.
by H. Binswanger, T. Balaramaiah, V. Bashkar Rao, M.J. Bhende and
K.V. Kashirsagar, July 1985.
Report No.: ARU 46
The Impact of Agricultural Extension: The Training and Visit System in
India.
by Gershon Feder and Roger Slade, June 1985.
Report No.: ARU 47
Methodological Issues in the Evaluation of Extension Impact,
by Gershon Feder and Roger Slade, July 1985.
Report No.: ARU 48
Escimation of Aggregate Agricultural Supply Response.-
by Hans Binswanger, Yair Mundlak, Maw-Cheng Yang and Alan Bowers
August 1985 (Revised October 1985).
Reoort No.: ARU 49
Land Values and Land Title Security in Rural Thailand.
by Yongyuth Chalamwong and Gershon Feder, June 1985 (Revised October
1985).
ReDort No.: ARU 50
Land Ownership Security and Capital Formation in Rural Thailand.
by Gershon Feder and Tongroj Onchan, December 1985 (Revised February
1986).
Report No.: ARU 51
Land Ownership Security and Farm Productivity in Rural Thailand.
by Gershon Feder, April 1986.
ReDort No.: ARU 52
Social and Cultural Aspects of Land Inheritance and Transactions in
Rural Thailand.
by Charles B. Mell, June 1986.


- 34 -
Discussion Papers
(continued)
Report No.:. ARU 53
Land Ownership Security and Access to Credit in Rural Thailand.
by Gershon Feder, Tongroj Onchan and Tejaswi Raparla, April 1986.
Report No.: ARU 54
The Management of Common Property Resources: Collective Action as an
Alternative to Privatization or State Regulation.
by Robert Wade, May 1986.
Report No.: ARU 55
Land Policies and Farm Productivity in Thailand's Forest Reserve Areas
by Gershon Feder, Tongroj Onchan and Yongyuth Chalamwong, August 1986.
Report No.: ARU 56
On the Determinants of Cross-Country Aggregate Agricultural Supply
by Hans Binswanger, Yair Mundlak, Maw-Cheng Yang and Alan Bowers,
September 1986.
Report No.: ARU 57
What Can Agriculture Do for the Poorest Rural Groups? by Hans P. Binswanger,
and Jaime B. Quizon, September 1986.
Report No.: ARU 58
Population Density, Market Access and Farmer-Generated Technical Change
in Sub-Saharan Africa by Prabhu L. Pingali and Hans P. Binswanger,
September 1986.
Report No.: ARU 59
Credit Markets, Wealth and Endowments in Rural South India by Hans P.
Binswanger and Mark R. Rosenzweig, December 1-5, 1986.
Report No.: ARU 60
Resource Endowments, Farming Systems and Technology Priorities for
Sub-Saharan Africa by Hans P. Binswanger and Prabhu Pingali, November
1986.
Report No.: ARU 61
Rural Factor Markets in China After the HOusehold Responsibility System
Report by Justin Yifu Lin, November 1986.
Report No.: ARU 62
Determinants of Women's Time Allocation in Rural Bangladesh by Shahidur
R. Khandker, February 1987.