Rural Labor-Intensive Public Works: Impacts of Participation on

Rural Labor-Intensive Public Works: Impacts of Participation on Preschooler Nutrition: Evidence
from Niger
Author(s): Lynn R. Brown, Yisehac Yohannes and Patrick Webb
Source: American Journal of Agricultural Economics, Vol. 76, No. 5, Proceedings Issue (Dec.,
1994), pp. 1213-1218
Published by: Oxford University Press on behalf of the Agricultural & Applied Economics
Stable URL:
Accessed: 13-05-2015 10:46 UTC
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content
in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship.
For more information about JSTOR, please contact [email protected]
Agricultural & Applied Economics Association and Oxford University Press are collaborating with JSTOR to digitize,
preserve and extend access to American Journal of Agricultural Economics.
This content downloaded from on Wed, 13 May 2015 10:46:05 UTC
All use subject to JSTOR Terms and Conditions
Evidence from
Lynn R. Brown, Yisehac Yohannes,and Patrick Webb
The high incidence of preschooler malnutrition Time Use, Income Generation, and Child
continues unabated in many areas of the world. Nutrition
Between 1980 and 1990 the number of underweight preschoolers rose from 164 million to Time is a direct input to child nutrition in terms
184 million. The percentage of preschool chil- of child care as well as a complementary input
dren below two standard deviations of NCHS to both food and nonfood nutrition inputs. Genreference weight for age in 1990 ranged from eration of income, particularly for the poor, in58% in South Asia and 31% in South East Asia, volves allocation of time to labor activities.
to 30% in Sub-SaharanAfrica. A concern is the Low household income requires more housemarginal increase in Sub-SaharanAfrica during hold members, particularly women, to be enthe 1980s [Administrative Committee on Coor- gaged in income-generating activities. This exdination/Sub Committee on Nutrition (ACC/ acerbates already tight female time constraints.
SCN) 1992]. While the incidence of poverty Lipton and Ravallion note that "The burden of
and malnutrition are highly correlated, income the 'double day'-market labor and domestic
growth alone is not sufficient to reverse the labor-is more severe for women. Female agetrends in malnutrition. In Pakistan, despite a specific participation rates increase sharply as
6.3% annual growth rate in GDP per capita be- income falls toward severe poverty; yet so do
tween 1977 and 1990 (World Bank), and a the ratios of children to adult women."
In the current economic climate of structural
marked reduction in poverty (Malik), the improvement in the prevalence rate of under- adjustment, the use of social services such as
weight preschool children was less than 1% per health and education often entails both user
fees and a complementary time input for seryear (ACC/SCN 1993).
Weak linkages between income and improved vice use. Thus, the time demands of female
nutritional outcomes have led to a growing fo- child caretakers may conflict with good child
cus on nonfood inputs to nutrition. These in- nutritional outcomes. Time in income generaclude mothers education and social service
tion activities (translated into higher food exinfrastructure in terms of health care, safe penditures) and time in utilizing health and
education facilities improve child nutrition outwater, and adequate sanitation (Thomas,
Strauss, and Henriques; Strauss; Alderman
comes, but the loss of direct time spent in child
and Garcia). However an input that has re- care may worsen nutritional outcomes. The net
ceived relatively little attention is time in- effects, therefore, of female employment outside the home are complex, involving a realloputs to child care.
cation of time and changing expenditure patterns.
A further potential consequence of female
Lynn R. Brown and Yisehac Yohannes are research analysts, and
is a change in internal household
Patrick Webb is a research fellow, all at the International Food
decision-making processes. The resultant
Policy Research Institute.
The authors would also like to acknowledge the unfailing suphigher share of female income relative to male
port of Lawrence Haddad, and contributions from Harold Aldercan tilt the power base inside the
man, Saroj Bhattarai, Jane Hopkins, Shubh Kumar, Carol Levin
household toward women. Evidence indicates
and Tesfaye Teklu.
Amer. J. Agr. Econ. 76 (December 1994): 1213-1218
Copyright 1994 American Agricultural Economics Association
This content downloaded from on Wed, 13 May 2015 10:46:05 UTC
All use subject to JSTOR Terms and Conditions
December 1994
that income accruing to women, at a constant
household income, is more likely to be spent on
food and child-oriented goods (Hoddinott and
There is little evidence to date on the net impacts of female employment outside the home
on child welfare levels. A recent study using
Ghanaian data found that household calorie
availability was adversly affected by external
female employment. The net effect of women
working, accounting for extra income, was -4.4
calories per-capita per day at the mean
(Haddad). This confirmed the results of an earlier Canadian study (Campbell and Horton).
Haddad found, however, that the impact was
dependent on per-capita expenditure levels. For
households below median per-capita expenditure levels where women were engaged in market activities, per-capita calorie availability increased, while for those above median percapita expenditure levels calorie availability
decreased. This suggests that, for poorer groups
of households, the provision of market-based
employment-generating activities for women
can increase household calorie availability and
potentially improve child nutrition outcomes.
This suggestion is not without its caveats.
Higgins and Alderman showed that demanding
physical labor performed by Ghanaian women
had significant negative effects on their own
nutritional status. A review of earlier evidence
by Leslie did not provide conclusive evidence
that maternal employment adversly affected
child nutrition.
A common policy instrument used to alleviate poverty throughout much of Sub-Saharan
Africa is labor-intensive Public Works programs. These programs offer low wages so as
to target the poor. The work is physical and energy-intensive, often entailing building infrastructure such as roads, and generally employs
both men and women. While work has been
done to evaluate the success of these programs
in terms of alleviating poverty, little has been
done to disentangle differential impacts dependent on the gender of the program participant.
Almost nothing is known about the impact of
participation on the nutritional characteristics
of either the participants or other household
members. For example, male participation may
raise household income but females may, in
turn, have less power in the household decision-making process, resulting in poorer translation of income into household calorie adequacy. The energy-intensive nature of the
work may lead to higher energy demands by the
male participant, which, given their higher rela-
Amer. J. Agr. Econ.
tive income share and consequent power position, they are able to realize. Participation by a
female may, however, result in a greater share
of the household income increase being translated into calories. Thus, while female participation may result in less direct female child
care time, it may have a greater impact on
household calorie availability and the share being allocated to preschoolers, and consequently
yield a net positive impact on preschooler nutrition status. This study will examine the impact
of public works participation on the nutritional
status of children using household data collected in villages with access to public works
programs in Niger.
The model of choice is the standard Beckerian
model (Becker). Optimal household utility is
generated from both market-purchased and
home-produced goods subject to budget and
household technology constraints. Home-produced goods result from the combination of
market-purchased goods and household members' time, subject to a technology constraint.
One such home-produced good is child health
as measured by nutritional outcomes. Let the
utility of household h consisting of i members
where Mb represents conbe Uh = U(Mb,
Oc, Ni),
market-purchased goods by
household h,
consumption of c
Oc represents
home-produced goods by household h, and N,
relates to the health of individual i as measured
by a nutritional status indicator. The relationship of most concern to us is the production of
preschooler health as measured by anthropometric outcomes Ni = h(Ei, Xi, i, •Eh),where
N, is the health outcome of individual i who is a
preschooler; E is a vector of inputs chosen by
the household for individual i, such as mother's
time input and nutrients; Xi is a vector of exogenous inputs to health, such as age, sex of i;
is an error term specific to child i which represents unobservable health endowments of individual i, and sh represents unobservable household and community characteristics.
The inputs specified in the E vector represent
choices made by the household regarding time
spent in various activities, including food
preparation, consumption of foods, and other
goods and services. As such, the level of input
use is likely to be correlated with the unobservable error terms in the production function. An
instrumental variables technique would permit
estimation of a reduced-form child nutrition
This content downloaded from on Wed, 13 May 2015 10:46:05 UTC
All use subject to JSTOR Terms and Conditions
Brown, Yohannes, and Webb
Rural Public Works: Impacts on Preschooler Nutrition
production function.
Estimation of a reduced-form nutrition production function, however, will not indicate
how changes in the determinants of the choice
of these inputs affect child nutritional outcomes. For example, it will not reveal how
changes in labor market activity by household
members affects child nutrition outcomes. This
effect would be transmitted through changes in
household members' time allocation patterns,
which may potentially affect the level of other
inputs to the child nutrition production function. Given that the objective of this study is to
examine the relationship of labor market activity in public works programs to child nutrition
outcomes, the methodology will be to use a
quasi-reduced form, a hybrid production-demand function.1
Our final child nutrition estimation equation
will contain both exogenous and endogenous
variables. The latter relate to household public
works labor supply, total household labor supply and household per capita calorie availability. All labor variables are at the household
level, disaggregated by gender and instrumented. Measures of other nonfood nutrition
inputs, such as access to and quality of health
care, quality of water, and sanitation, are not
available at the household level. Village-level
dummy variables, therefore, will be used to
control for community-level fixed effects.
The data is cross-sectional and was collected
during 1991-92 from public works programs
operating in three rural sites-Banimate,
Galmi, and Kalfou Rafi-in Niger (Webb). All
the sites have had publics works activities since
at least 1987. Banimate is located in a poor
northern zone of Niger, close to the Malian border, with very low rainfall. Infrastructure is
poor although there is safe drinking water in
1 The ideal methodology would link changes in labor market activity directly to demands for individual inputs which enter into,
the nutrition production function. Econometrically, however, this is
close to impossible. Many of the input demands will relate to
mother's time in a health-related household activity. For example
breastfeeding is conditional on mother's time and is also a direct
but endogenous input to child nutrition. Estimation of input demand for time in breastfeeding conditional on mother's labor market activity, however, would present a problem as the same price,
the mother's wage in the labor market, is relevant for both. It
would be difficult to come up with a set of exclusion restrictions
that permit estimation of time-related nutrition input demand equations, such as breastfeeding, which include endogenous explanatory variables such as labor supply.
both the wet and dry seasons. The nearest
health clinic is a three-hour walk away, as are
the nearest market and grain mill.
Galmi is a relatively wealthy rural area close
to the Nigerian border (ensuring a thriving contraband trade) on a main Niamey to Zinder
tarmac road, with a dam providing irrigation
for some land. Galmi benefits from schools, a
local daily market, and a grain mill. The nearest
health clinic is a fifteen-minute walk away, and
safe water, a five-minute walk away.
Kalfou Rafi, with a rainfall of just 350 mm
per year, is very dependent on farming and livestock. Infrastructure and services are limited.
The nearest tarmac road is a ninety-minute
walk, the health clinic and daily market a
thirty-minute walk. There is year-round access
to safe water.
The public works programs, which are both
labor and energy intensive, were all aimed at
natural resource protection including soil terracing, reforestation, erosion control, stone terracing, and windbreak planting.
All projects use food as the payment medium,
with a daily time wage equivalent to 2.25 kg of
cereals (typically sorghum or millet), plus additional noncereal commodities such as dried
milk, cooking oil, canned meat, and sugar when
The total number of households surveyed was
275 (74 in Banimate, 126 in Galmi, and 75 in
Kalfou Rafi). The survey included modules on
demographics, production, assets, employment/
income, time use, food consumption, expenditures, nutritional status measures, and public
works project experience.
Stratifying the sample into high- and lowpublic-works-participation households showed
that high-participation groups were poorer and
had an average of three workers per household,
compared to one for low-participation households. The average number of days worked per
participant was higher for high-participation
groups. Women comprised 59% of project participants, providing 61.5% of all public works
labor. Most were between the ages of sixteen
and thirty-five and were married. For the most
part, men were in the same age category.
The nutritional indicator chosen for this study
is weight for age. Fifty-seven percent of children under six years of age in the sample were
more than two standard deviations below the
NCHS reference standards. The figure rises to
68% in Banimate and falls to 44% in Kalfou
Rafi. Labor days supplied to public works programs are lowest in Banimate, with a yearly
mean of 74 days and 60 days for males and fe-
This content downloaded from on Wed, 13 May 2015 10:46:05 UTC
All use subject to JSTOR Terms and Conditions
December 1994
Amer. J. Agr. Econ.
Table 1. Variables used in the Analysis: Means And Standard Deviations
Weightfor age z score
Predictedhouseholdpercap. calories
Totalpredictedhouseholdmale labordays
Male sharepublicworks*
Age of child in months
Sex of child (1 = male)
Numberof childrenin householdage < 6yrs
Numberof childrenin householdage 6-14 yrs
Numberof femalesin householdage 15-65 yrs
Numberof malesin householdage 15-65 yrs
Numberof adultsover age 65 yrs
Sex of householdhead(1 = male)
Dependencyratio(persons<15yrs/persons>15 yrs)
Household Size
Per capitahouseholdassets (valueCFA)
Source: IFPRI/INRAN survey 1991-92.2
* Share public works is total household predicted
public works days in a year (male/female) as a proportion of total household predicted
labor days including public works in a year (male/female).
males respectively. In Kalfou Rafi labor days
rise, at the mean, to 190 and 215 days per year
for males and females respectively. This is the
only district where female labor supply to public works is higher than males. Mean daily percapita calorie consumption ranges from 2,148
in Banimate to 3,247 in Kalfou Rafi.
The means and standard deviations of variables used in the analysis are reportedin table 1.
The gender disaggregated instrumenting equations, using predetermined variables to predict
public works participation, labor supply to public works, and total household labor supply, are
not reported for space reasons. Instruments included asset holdings, household demographic
structure, and individual- and community-level
characteristics. Coefficients in these regressions had the anticipated signs. For example,
the presence of children less than six years of
age strongly deterred both female participation
and the magnitude of female labor supply to
public works programs, but had no influence
for men.
The results of the per-capita calorie instrumenting equation and preschooler nutritional
weight for age regression are shown in table 2.
The predicted share of total household public
works labor in total labor supply, disaggregated
by gender, is included in the per-capita calorie
instrumenting equation to measure potential direct impacts of public works employment.
While for women the coefficient is positive, for
men the reverse is true. This result may be
driven by many factors, including but not limited to changes in resource allocation due to
women's increased relative income contribution; men paid in food may reduce other
intrahousehold transfers resulting in lower food
Within the child nutrition relationship, the
coefficient of the share of public works labor
variable represents the net effect of several
components. It captures income effects outside
of calorie availability (hypothesized to be positive for men and women), changes in resource
control/allocation (hypothesized positive for
women and negative for men), changes in child
care time/practices (hypothesized negative for
women, zero for men), and the increased competition for calories between participant and
child (negative for men and positive or zero for
women). In our results the net impact of a rising share of household public works employment for women is positive resulting in an increase in weight for age of children. There is no
impact with an increasing share of male public
works labor supply. An increase in total household female labor supply also improves child
nutrition, whereas for men there is a negative
direct impact of increasing total labor supply.
This content downloaded from on Wed, 13 May 2015 10:46:05 UTC
All use subject to JSTOR Terms and Conditions
Brown, Yohannes,and Webb
Rural Public Works:Impacts on Preschooler Nutrition
Table2. Results of Calorie InstrumentingEquation and Determinantsof PreschoolerNutrition
Log PredCals
t stats
Male sharepublicworks
Sex of householdhead(1 = male)
Numberof childrenin householdage < 6yrs
Numberof childrenin householdage 6-14 yrs
Numberof femalesin householdage 15-65 yrs
Numberof malesin householdage 15-65 yrs
Numberof adultsover age 65 yrs
Dependencyratio(persons<15yrs/persons215 yrs)
Age of child in months
Age child squared
Sex of child (1 = male)
Predictedhouseholdper cap. calories
Adj. R Sq.
Z score
t stats
* significant at 10%
** significant at 1%
Fears have been expressed about the impact on
child nutrition of participation in public works
programs by women; this study allays those
fears. For poorer households, increasing the
share of public works employment for women
has positive impacts on child nutrition over and
above its impact of increasing calorie availability, despite potential reductions in direct time
inputs into child care. Male participation in
public works, however, has no direct impact on
child nutritional status, but leads to a reduction
in household per capita calorie availability
which has a negative impact on child nutrition
status. Similarly, increases in female labor supply have direct positive impacts on child nutritional status while for men the opposite is true.
Poverty alleviation schemes, such as public
works employment programs which have design characteristics that discourage participation by women, may have a lesser impact on reducing preschooler malnutrition. That does not
mean, however, that there are no disadvantages
to rising female labor supply. This study has
taken no account of public works participation
by women on their own nutritional status, nor
the impact of reduced child care time on other
aspects of child welfare.
Administrative Committee on Coordination/Sub
Committee on Nutrition (ACC/SCN). Second
Report on the WorldNutrition Situation. Volume
1. Global and Regional Results. Geneva: ACC,
Administrative Committee on Coordination/Sub
Committee on Nutrition (ACC/SCN). Second
Report on the WorldNutrition Situation. Volume
II. Country Trends, Methods and Statistics.
Geneva: ACC, 1993.
Alderman,H., and M. Garcia. "Food Security and
HealthSecurity:Explainingthe Levels of Nutritional Status in Pakistan."Econ. Develop. and
Cultur. Change 42(1994):485-507.
Becker, G. "A Model of Time Allocation." The
Econ. J. 75(1965):493-517.
Campbell,C., and S. Horton."Wife'sEmployment,
Food Expenditures,and ApparentNutrientIntake: Evidence from Canada." Amer. J. Agr.
Econ. 73(1991):784-94.
Haddad,L. "The Impactof Women'sEmployment
Statuson HouseholdFood Securityat Different
and Nutr. Bull.
Higgins, P.A., and H. Alderman. "Labor and
Women'sNutrition.A Study of EnergyExpenditure, Fertility, and Nutritional Status in
This content downloaded from on Wed, 13 May 2015 10:46:05 UTC
All use subject to JSTOR Terms and Conditions
Amer.J. Agr. Econ.
Ghana." Washington DC: World Bank, Policy
Research Working Paper WPS 1009, 1992.
Hoddinott, J., and L. Haddad. "Does Female Income
Share Influence Household Expenditure Patterns? Evidence from C6te D'Ivoire." Oxford
Bull. Econ. and Statist. in press, 1994.
Leslie, J. "Women's Work and Child Nutrition in the
Third World." World Develop. 16(1988):134162.
Lipton, M., and M. Ravallion. "Poverty and Policy."
Handbook of Development Economics, Vol. 3. J.
Behrman and T.N. Srinivasan, eds. Amsterdam:
North Holland, forthcoming.
Malik, S. "Poverty in Pakistan: 1984/85 to 1987/88."
Including the Poor. M. Lipton and J. van der
Gaag, eds. New York: World Bank, 1993.
Strauss, J. "Households, Communities, and Preschool Children's Nutrition Outcomes: Evidence from Rural C6te D'Ivoire." Econ. Develop. and Cultur. Change 38(1990):231-62.
Thomas, D., J. Strauss, and M.H. Henriques. "How
Does Mother's Education Affect Child Height?"
J. Human Resour. 26(1991):183-211.
Webb, P. "Food Security Through Employment in
the Sahel: Labor-Intensive Programs in Niger."
Unpublished, International Food Policy Research Institute.
World Bank. World Development Report 1992. Oxford: Oxford University Press, 1992.
This content downloaded from on Wed, 13 May 2015 10:46:05 UTC
All use subject to JSTOR Terms and Conditions