Citation
Women in development

Material Information

Title:
Women in development A.I.D.'s experience, 1973-1985
Series Title:
A.I.D. program evaluation report
Creator:
Goddard, Paula O
Carloni, Alice Stewart
United States -- Agency for International Development
Place of Publication:
Washington D.C
Publisher:
U.S. Agency for International Development
Publication Date:
Language:
English
Physical Description:
v. : ; 28 cm.

Subjects

Subjects / Keywords:
Agricultural development projects -- Women ( lcsh )
Economic development projects -- Women ( lcsh )
Women in development -- Developoing countries ( lcsh )
Women -- Economic conditions -- Developing countries ( lcsh )
Genre:
bibliography ( marcgt )
federal government publication ( marcgt )

Notes

Bibliography:
Includes bibliographical references.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
21026095 ( OCLC )

UFDC Membership

Aggregations:
Women in Development
University of Florida

Full Text

Document No.


E-5


ENTER FOREVE* INORATON&-VAUAI-
CENTER FOR DEVELOPMENT INFORMATION & EVALUATION


3 ALE .f


AGENCY FOR INTERNATIONAL DEVELOPMENT
Washington, D.C. 20523


-f/s

















WOMEN IN DEVELOPMENT: A.I.D.'S EXPERIENCEr 1973-1985


VOL. II: TEN FIELD STUDIES


A.I.D. WORKING PAPER NO. 3 1 3










Paula 0. Goddard, Editor
(Center for Development Information and Evaluation, A.I.D.)









U.S. Agency for International Development

December 1989



The views and interpretations expressed in this report are
those of the authors and should not be attributed to the
Agency for International Development.








TABLE OF CONTENTS


Page

Preface . . . .. .. ... v

Glossary . . . . . vii

Chapters

1. Botswana Agricultural College Expansion Project
Mary B. Anderson and Edna McBreen . . 1

2. The Accelerated Mahaweli Program in Sri Lanka
Janet E. Benson and Jan Paul Emmert . . 23

3. Northeast Rainfed Agricultural Development Project
in Thailand: A Women-in-Development Assessment
M. Cristina Blanc-Szanton, Anamaria Viveros-Long,
and Nongluck Suphanchainat . . . .. 45

4. Association for the development of Micro-Enterprises,
Inc. (ADEMI) in the Dominican Republic
Rae Lesser Blumberg . . .. . 65

5. Work, Wealth, and a Women-in-Development "Natural
Experiment" in Guatemala: The ALCOSA Agribusiness
Project in 1980 and 1985
Rae Lesser Blumberg, with Maria Regina Estrada
de Batres and Josefina Xuya Cuxil .. .. 85

6. A Socioeconomic Assessment of the Arid and Semi-Arid
Lands Project in Kenya
Alice Carloni and Nadine R. Horenstein . ... 107

7. Women-in-Development Issues in the Nepal Resource
Conservation and Utilization Project
Alice A. Davenport, Tom Nickell, and Bina Pradhan .. 129

8. Appropriate Technology for Rural Women Project,
Inter-American Commission of Women, Organization
of American States
Cornelia Butler Flora . . .. . 146

9. The Morocco Industrial and Commercial Job Training
for Women Project
Margaret Lycette . . . . 172

L0. The Caribbean Agricultural Extension Project
Marianne Schrink and Paula O. Goddard . .. 192


References







PREF..CE


In connection with the 1985 World Conference to Review and
Appraise the Achievements of the United Nations Decade for Women
(1975-1985), A.I.D.'s Center for Development Information and
Evaluation (CDIE) undertook a year-long study of the Agency's
experience in women in development. The final report is
contained in two volumes: Women in Development: A.I.D.'s
Experience 1973-1S85, Volume I, Synthesis Paper, A.I.D. Program
Evaluation Report No. 18, and this volume, which contains edited
case studies of 10 field reports. The case studies, together
with six sector studies, formed the basis for the Synthesis
Paper.

Interest in A.I.D. and the larger development community in
women in development continues and in fact is again on the rise
in A.I.D. in response to recent Congressional pressure and
legislative action. But for all the intensity of interest and
the growing awareness that gender plays a key role in development
outcomes, frustration and lack of knowledge about how to apply
women-in-development concepts in specific projects and programs
continue, The shortage of women-in-development expertise is
serious, particularly in the design and implementation of broad
development activities where women and girls play an important
role. Few today would dispute the need for gender analysis in
policy reform programs, natural resource management, food
production, and social programs to improve health and education.
Yet the details of determining what role gender plays in the
success or failure of these activities still present a
significant challenge to even the most dedicated professional.

CDIE presents the 10 project case studies in this volume to
benefit the scholar or practitioner who aspires to greater
understanding of the role of gender in development. While the
overall picture drawn from these projects was summarized in the
Synthesis Paper, a reading of the individual case studies offers
further and in-depth insight into the projects. The cases cover
all the sectors in which women and girls in developing countries
are particularly active: employment, agricultural production,
and the provision of water, fuel, health, and education to their
families. Of course, each case is only a snap shot of a project
at a particular time, necessarily presented as though the project
were in static context. However, these projects have continued
to evolve since the case studies were prepared and may now bear
little resemblance to the trends anticipated and the predictions
made by CDIE's analysts. Indeed true students of the development
process must, at regular intervals and long after initial project
activities have ceased, revisit and reanalyze projects and







project impacts in order to understand the mysteries involved in
what we call "development." In the Alcosa project in Guatemala,
the CDIE analyst visited and studied a project 5 years after a
previous researcher had completed a case study and was able to
present another chapter in the ongoing story of this project. In
the ADEMI project, the analyst was able to update some of the
original research data from 1985 with additional recent data on
that project.

But for the other eight cases, it remains the task of future
researchers to return to these projects and plot the continued
course and ultimate impact of each. CDIE offers this volume of
10 field studies to the interested student of the now extensive
women-in-development literature as part of a historical record
and as a basis for continued research and better understanding of
this important topic.


Paula 0. Goddard
Deputy Director
Center for Development
Information and Evaluation
Bureau for Program and Policy
Coordination
Agency for International
Development
December 1989








-vii-

GLOSSARY


ADEMI


A.I.D.

ALCOSA

ASAL

bachot level




binna

CAEP

chena

comuna

deqa

FCDERUMA

el qasto

GDP

ha

IBTA

kqotla

mm

mwethya

NERAD

OFPPT


panchavat


Association for the Development of Micro-
Enterprises, Inc.

U.S. Agency for International Development

Alimentos Congelados Monte Bello, S.A.

Arid and Semi-Arid Lands project

educational level of those who have completed
secondary school but have not passed the exam
required to receive the baccalaureate diploma
(Morocco)

uxorilocal or matrilocal marriage (Sri Lanka)

Caribbean Agricultural Extension Project

dry-land agriculture (Sri Lanka)

Ecuadorean cooperative

virilocal marriage (Sri Lanka)



household expenses (Guatemala)

gross domestic product

hectares

Bolivian Institute of Agricultural Technology

traditional community meeting (Botswana)

millimeters

self-help (Kenya)

Northeast Rainfed Agricultural Development project

Office of Technical Training and Job Development
(Morocco)

traditional village council (Nepal)








-viii-

PTAMC Appropriate Technology for Rural Women Project

RCUP Resource Conservation and Utilization Project

tambon political subdistrict (Thailand)

WAND Women-in-Development Unit, University of the West
Indies

WID women in development









1. BOTSWANA AGRICULTURAL COLLEGE EXPANSION PROJECT

by

Mary B. Anderson, Economist
President, Collaborative for Development Action, Inc.

and

Edna McBreen, Agricultural Education Specialist
State University of New York


1.1 Summary


The effectiveness of the Botswana Agricultural College Ex-
pansion project in integrating women was evaluated at two levels:
(1) the extent to which women were integrated into the institu-
tion as faculty members and students and the extent to which
gender awareness and gender-related patterns of production were
incorporated in curriculum design and (2) the impact that the
inclusion or exclusion of women at the institutional level has
had on the productivity and well-being of Botswana's female
farmers.

The findings include the following:

1. The college has increased both its male and female fac-
ulty as a result of the expansion project. The ratio of female
to male faculty at the professional level has remained at 1:3.
At the technical level, the faculty increased from 11 to 23 mem-
bers, but the ratio of female to male faculty decreased from 1:11
in 1979 to 1:23 in 1985.

2. In the participant training component of the project,
women were underrepresented: 27 men and 5 women received train-
ing. The majority of degrees received by participant trainees
were in animal health-related studies (14) rather than in crop-
related studies (8), although 10 were in general fields equally
applicable to animals or crops. These figures show that the
effort to strengthen the college's staff was biased toward the
animal-related fields, which are dominated by men in Botswana,
whereas women predominate in crop-related fields.

3. An increasing number and proportion of women have been
admitted to the college as students in all four training categor-
ies. The proportion of women increased from 20 percent of gradu-
ates in the Agricultural Certificate program in 1979 to 39 per-
cent in 1984, from 0 percent in the Animal Health Certificate
program in 1979 to 6 percent in 1984, from 0 percent in the







-2-


Animal Health Diploma program in 1981 to 11 percent in 1984, and
from 0 percent in the Agricultural Diploma program in 1982 (first
year) to 33 percent in 1984. However, the attrition rate for
women students over this period was 33 percent, compared with i9
percent for men students.

4. The concentration of women students in the agriculture-
related areas and men in the animal health-related areas reflects
the general division of labor among Botswana's agricultural
population.

5. An annual limit of 15 women has been established for
admissions to the Agricultural Certificate program. The reason
given for this limit is that placement of women as agricultural
demonstrators in remote areas is difficult; there is, however,
disagreement concerning the validity of this argument.

6. There is little concrete evidence that women and men
differ in effectiveness as agricultural demonstrators, although
the agricultural extension services in general are criticized for
being ineffective. However, several issues did arise that war-
rant further investigation, including the following:

-- Women agricultural demonstrators who are pregnant are
not expected to travel or work in the same way as male
and nonpregnant female agricultural demonstrators.
There are no figures on the number of female agricul-
tural demonstrators who became pregnant or on actual
time lost to the job as a result of pregnancies. How-
ever, our interviews indicated that women agricultural
demonstrators lose 9 to 12 months of active work time
during pregnancy and nursing and, therefore, they can-
not serve their client farmers as well.

-- Among male agricultural demonstrators, the incidence of
alcoholism and the resulting numbers of days of work
lost are high. Exact figures on the numbers of days
lost have not been compiled.


1.2 Introduction


1.2.1 Project Goal


The Botswana Agricultural College Expansion project of th3
Agency for International Development (A.I.D.) was designed to
address the specific needs of small-scale farmers and herders
within Botswana's agricultural sector. It was chosen for the









women-in-development study because it is an institution-building
project of which the design included an explicit consideration of
gender roles, and because it allowed a fuller analysis of the
ways in which such a project design may or may not affect women
in the ultimate project population.

During FY 1984, the USAID/Botswana program included numerous
projects aimed at strengthening country institutions related to
agriculture, off-farm production, and income in r--al areas;
improving primary and postprimary education systems; and provid-
ing advanced training for administrators and technical personnel.
Usually, the Mission views women-in-development issues as an
integral part of its general work, so the majority of a Mission's
projects incorporate women into the mainstream of the project.
However, smaller efforts focusing directly on women and their
needs in development are evident.


1.2.2 Prolect Setting


The Botswana Agricultural College is administratively re-
sponsible for providing graduates in agriculture to work for the
Ministry of Agriculture. Thus, the policies and decisions of the
Ministry have had a strong impact on the project to expand the
college. Having estimated the staff it will need to fulfill its
goals with respect to Botswana's poorest farmers and herders, the
Ministry foresees a personnel shortfall of about 600 people by
1988, unless indigenous training is greatly increased. The Bo-
tswana Agricultural College Expansion project is an important
aspect of the strategy to increase the numbers of trained person-
nel in agricultural extension.

Several aspects of Botswana's general economic and political
context have also affected the Botswana Agricultural College
prrect.

The agricultural sector provides a livelihood for Botswana's
rural population--83 percent of the country's 1 million people.
That sector provides only about 10 percent of gross domestic
product (GDP), but it accounts for nearly 50 percent of all em-
ployment in Botswana. Livestock (mostly cattle) production ac-
counts for 80 percent of the sector's share of GDP. Ownership of
livestock, however, is highly skewed, with only 6.7 percent of
rural households controlling 54 percent of the national herd
(estimated at 2.8 million head). Most households are dependent
primarily on raising crops, which is mainly the responsibility of
women. Cultivation is difficult and risky under prevailing cli-
matic and ecological conditions. In years of moderate rainfall,









annual food grain production averages only 30 percent of national
requirements.

Botswana's economy is strongly linked to and dominated by
the neighboring Republic of South Africa. Botswana's membership
in the Southern Africa Customs Union generates 32 to 40 percent
of the Government's annual revenues, and South Africa accounts
for 88 percent of Botswana's imports and 15 percent of its ex-
ports. Approximately 20,000 Botswana are employed in South Af-
rica, primarily in the mining industry. The SoutL African com-
panies of DeBeers and Anglo-American hold significant. equity and
have managerial influence in Botswana's diamond, coal, and
copper-nickel production.

Botswana's physical characteristics present a challenge
to any agricultural project. It has a continental and semiarid
climate, with highly erratic rainfall averaging only '50 milli-
meters (mm) annually, but varying from less than 250 rm in the
southwest to more than 650 mm in the northwest. The Kalahari
Desert constitutes 80 percent of the country's land area, and
only about 6 percent of the land is suitable for farming. The
Limpopo Valley region on the eastern side of the country has
nearly 80 percent of the population, relatively good soil, and
generally just enough rainfall for some dry-land crops and
cattle.

Agricultural production in Botswana has suffered greatly
from current drought conditions. Planting, harvesting, and pro-
duction levels have all declined drastically. Three consecutive
years of drought have reduced Botswana's output to 5 percent of
its requirements.

The influence of gender on agricultural labor and production
in Botswana varies extensively according to location, tribe, and
family. The USAID Agricultural Development Officer, Dr. Anita
Mackie, described the variation in Botswana as the greatest she
has seen in her work in about 40 African countries. Tradition-
ally, however, agricultural production in Botswana involves a
complex system of gender-assigned responsibilities. At the macro
level, men are responsible for livestock, land clearing, and
plowing with oxen, and women are responsible for crops.

The actual delineation of work based on gender has been
affected by variations in family configuration: approximately 40
percent of families are headed by women. In addition t- single
women and widows, this figure includes a sizable number of mar-
ried women whose husbands are absent. Thus, strict delineation
of tasks by gender appears to be breaking down.







-5-

Data on persons performing crop and livestock activities
show that women are the major participants in 47.7 percent of all
crop activities but do not take part in livestock activities,
excluding those involving poultry. However, the data on persons
responsible for crop and livestock operations show that women are
responsible for 57.7 percent of all.crop activities and 12.1
percent of all livestock activities.

The traditional division of labor severely limits the
female-headed households' access to resources. Most women rely
on men for plowing with draft animals, so female heads of house-
holds must often hire someone to do this job. The constraint on
access to draft animals, therefore, quickly becomes a resource
constraint affecting both labor and economic resources. Because
women tend to be among the smallest landholders with the most
marginal operations, they cannot easily overcome this constraint.


1.3 Project Design, Implementation and Results, and Gender
Issues


1.3.1 Project Design


A.I.D. provided a grant of $7,1i9,300 to the Government of
Botswana to support the expansion of the Botswana Agricultural
College between 1978 and 1983. The objective of the project was
to improve the welfare of Botswana's small-scale farmers and
herders. As defined in the Project Paper, the project purpose
was to establish within the college a largely localized training
institution capable of serving rural sector needs. The project
inputs were specified as follows:

Long-term technical assistance in vocational and educa-
tional administration, extension, animal health, range
management, agricultural communication, agronomy, and
general science

-- Short-term technical assistance in animal breeding,
horticulture, rural sociology, and extension

-Construction of dormitories, classrooms, administration
and faculty offices, and other buildings; and new and
expanded facilities for several departments, a library,
and dining hall

-- Long-term participant training in the United States for
college staff and administrators to earn bachelor's and
master's degrees in agricultural economics, range









management, .)I ,r:ticulture, agronomy, veterinary medicine,
and educational administration

-- Provision of eAqipment and commodities including ve-
hicles, laboratories, and furnishings for.classrocms,
dormitories, and other facilities

The project was also designed to address an existing and pro-
jected shortfall in agricultural personnel by doubling the col-
lege's annual admissions capacity from 60 to 120 certificate-
level students, and by introducing a diploma-level program to
train 30 student- each year.

The primary project beneficiaries were to be 14 college
.staff members who were Lo receive training and 150 college stu-
dents per year who were Lo receive certificate- and diploma-
level training. Secondary beneficiaries were to be rural Bo-
tswana, who would benefit from improved service from agricultural
extension demonstrators and veterinary assistants.

Graduates from Botswana Agricultural College are employed
primarily as agricultural extension demonstrators aid as veteri-
nary assistants. Agricultural demonstrators are assigned to the
Ministry of Agriculture's Department of Agricultural Field Ser-
vices, and veterinary assistants are assigned to the Ministry's
Department of Veterinary Health. Together, the department pro-
vide the key link between the Ministry and rural Botswana's agri-
cultural and livestock activities.

At the time the project was designed, five women were serv-
ing as agricultural demonstrators and six women were enrolled in
the agricultural progr~;. The project intended to improve ad-
mission procedures to increase the number of women students at
the college, but it set no specific targets for women. This
approach--noting that women should be included without specifying
numbers or procedures--is in keeping with the stated intention of
the Government of Botswana and USAID/Gaboro:ie of "integrating"
women into mainstream projects. Furthermore, it reflects the
awareness by the Botswana Government and the Ministry of Agricul-
ture of the importance of women's roles in agricultural develop-
ment as set forth in the 1976-1981 National Development Plan.
The project's design documents, therefore, included discussion of
gender roles in agriculture but did not rebate. them Ln .any way to
the design for the co3iege expansion project.








-7-


1.3.2 Proiect Implementation and Results


On the whole, the project proceeded smoothly from its imple-
mentation in 1979. The number of participant trainees was in-
creased from 14 to 20, and technical assistance was extended to
ensure continuity until all Botswana returned from training (1986
projected). In general, the project is considered very success-
ful--.t achieved all its intended outputs. In addition, the
number of women students admitted to the college has risen.


1.3.3 Proiect Gender Issues


Botswana's Focus on Gender. The Government has established
a Women's Affairs Unit within the Ministry of Rome Affairs. Its
goals are to coordinate women's activities in Botswana at local,
national, and international levels; to disseminate information;
to conduct research on-the situation of women in Botswana; and to
interact with different Government departments on issues related
to women. The unit works in conjunction with other ministries,
using a cooperative approach aimed at integrating women into the
ministries' planning and ongoing work. The unit acts as a re-
source to other ministries rather than as an advocate for women's
rights.

The Ministry of Agriculture's Director of Field Services
stressed that the approach of his Ministry has been to attempt to
integrate women into the larger system both as staff members and
as project beneficiaries. The Ministry of Agriculture has di-
rected some programs toward women and, in fact, had a Women's
Unit until 1984, when it was phased out.

Availability of Gender-Related Data. Extensive gender-
related data on agriculture and extension activities in Botswana
as well as information describing Botswana Agricultural College
are available. In addition to some gender-disaggregated census
data and Government records, several studies based on interviews,
field observations, and national statistics have been written
about women in agriculture, intrahousehold management, and the
extension service's outreach to women farmers. (See, especially,
Bond 1974; Bettles 1984; and Fortmann 1981.) There is, however,
at least one research area directly related to the college and
its inclusion of women as faculty and students that requires
investigation: no data are available for comparing the success
of female versus male agricultural demonstrators and their ulti-
mate impact as agricultural extension agents; data indicating
whether women farmers are reached by extension efforts are also
scant.








-8-


1.4 The Field Study


1.4.1 Analytical Framework


This study addressed the following key issue: whether the
well-being and productivity of small-scale farmers and herders in
Botswana were more or less effectively improved because the Bo-
tswana Agricultural College Expansion project did or did not
integrate women into its design and implementation. The linkage
between the project (to increase the numbers of trained agricul-
tural extension staff) and the Government's goal of improving the
welfare of small-scale farmers and herders involves reaching
herders and farmers to help improve their productivity. The task
was to discover whether and how consideration of gender in the
college expansion project made any difference to the effective-
ness with which rural Botswana's productivity needs were served.

The primary focus of the study was the integration of women
into the college, particularly in the following three areas:

1. Faculty: Hiring and training

2. Students: Applications/adrLssions; retention/
performance; and placement

3. Curriculum:

-Content, breadth, priorities; relation to the real
situation in agriculture in terms of appropriateness
of technologies, gender divisions of labor, and
cultural practices

-Relation to agricultural research efforts: the
extent to which research findings are incorporated
into curriculum and the extent to which student
experience helps shape research priorities

Adaptation of methods for r-aching farmers and herd-
ers according to nature of target group

Another aim of the study was to assess the importance of
gender to the broader objectives to which the project was linked.
The following questions were basic to this analysis:

1. Who are the farmers and herders? Classified by:

-- Crop types (which vary by household type, decision-
making patterns, tribe, and land types)








-9-


-- Stages in crop production (plowing, planting, weed-
ing, harvesting)

-- Type of animal raised

-- Frequency and usefulness of contact with agricul-
tural demonstrators and veterinary assistants

2. Who are the agricultural demonstrators and veterinary
assistants?

-- Their ideal characteristics for this work

-- Advantages or disadvantages for male or female agri-
cultural demonstrators or veterinary assistants

3. What is the impact of the Agricultural College on pro-
ductivity?

-- When information is transmitted to farmers through
college graduates

-- By farmer type, crop type, or animal type


1.4.2 Evaluation Methodoloqy


In examining the key issues addressed by this study, we
relied on available documents and on information gained through
interviews.

To assess the extent of inclusion of women in the college,
we used student admissions records for the years prior to and
during the project, faculty records, records from the USAID/Bo-
tswana training officer regarding participant training in the
United States, and curriculum descriptions. We supplemented
information gained from these documents by interviews with the
college principal and vice principal and with members of the
A.I.D. technical assistance team. We also visited the college
and observed facilities for men and women and for a number of the
subjects taught at the college.

Assessing the quality of the extension services provided by
Botswana Agricultural College graduates and the influence of
gender was much more difficult. First, we attempted to learn
whether there was any difference in the quality of the work of
female or male extension agents who had graduated from the col-
lege. Then we looked for indicators of the quality of extension
advice provided by both female and male extension workers to








-10-


female and male farmers as it had been affected by tLe training
received at the college.

Botswana's agricultural extension service has been studied
and evaluated on numerous occasions. We reviewed a number of
previous studies for information on the relative effectiveness of
female and male extension workers. We also interviewed some of
the people involved in the research and Government officials in
the Ministry of Agriculture and the Women's Division about their
assessment of this research and its accuracy in describing the
current situation.

We alsD relied heavily on existing research on the gender
division of labor in agriculture and on productivity levels in
crop production and herding. From these studies we derived ques-
tions used in interviews with Ministry of Agriculture and other
extension service officials, as well as with the extension agents
and expatriate technical assistance staffs in the field. From
these interviews, we attempted to discover what priorities are
placed on providing extension services to the different activi-
ties in agriculture (as related to the gender division of labor)
and what relationships there may be between extension agents' ap-
proaches to their work and the possibility of increasing their
effectiveness in reaching female farmers and increasing women's
crops. We also interviewed district-level workers and extension
agents in three districts.

The fact that much research on gender roles in agriculture
and several evaluations of Botswana Agricultural College had
already been done was in some respects an impediment to this
study. Almost without fail, the gender focus of this evaluation
elicited standard responses of three types: (1) hostility,
(2) recitations of numbers, or (3) stereotypical reporting of
male/female roles and attributes. Several people were reluctant
to meet with us or to set up additional meetings for us; they
believed that the college had already been adequately evaluated
and that this evaluation was unnecessary and our requests were
impositions. In most cases we had to spend some time just get-
ting people to consider -he question we were actually asking:
whether gender matters i. the areas under discussion rather than
whether women should be or had been trained at the .Allege.

Thus this study represents a beginning of a new approach to
exploring where, when, and whether gender issues matter in proj-
ect design and not a conclusive study of the degree to which they
did matter in the Botswana Agricultural College Expansion
project.







-11-


1.5 Project Analysis


Findings are reported in two parts, corresponding to the
analytical framework discussed above. This section treats the
project itself and the integration of women into specific compo-
nents of the college--personnel and curriculum.


1.5.1 Personnel


Table 1 shows the gender distribution of the College faculty
at the beginning of the project in 1979 and in March 1985.


Table 1. Number of Male and Female Faculty at
Botswana Agricultural College in 1979 and in March 1985


1979 1985
Category Male Female Male Female

Professional
Expatriate 7 2 8 1
Local 4 2 13 6
Subtotal 11 4 21 7

Technical
Expatriate 9 -
Local 1 1 22 1
Subtotal 10 1 22 1


While the number of local staff increased markedly in both the
professional and technical categories, the percentage of women
relative to men fell in the technical category and remained at
approximately one-third in the professional category.

The Botswana Agricultural College Expansion project had a
significant participant training component. Eight students (male
and female) were in crop-related areas, 14 in animal health, and
10 in general areas related to both crop and animal production.
Of the 32 people who received training, 5 (17 percent) .were
women--somewhat below current female faculty representation. Two
of the women concentrated on crop-related courses and three on
general courses. None received a degree in animal health. The
greater number of degrees in animal sciences indicates an







-12-


implicit, if not explicit, priority on developing the competence
of the college staff in the livestock and other animal health
courses. This emphasis is directly counter to the Ministry of
Agriculture's stated priority of increasing food production and
reinforces the existing bias toward male-dominated livestock
production.

Of the additional faculty who received short-term training,
all three were men; two received general training and one focused
on a crop-related area.

Between 1973 and 1980, Botswana Agricultural College gradu-
ated 390 students, of whom 16 were women (Bettles 1984). As the
college expanded its program, the numbers of both male and female
enrollments and graduates increased, and the percentage of women
increased as well.

There are relatively more women in the two agricultural
programs than in the animal health programs. This distribution
reflects the general division of labor by gender in the country-
side. However, the proportion of female to male students at the
college underrepresents women both in the rural population in
general and in the populations engaged in crop- or animal-related
activities in particular. During the life of the project, en-
rollments and graduations of women tended to increase. In the
discussion below on placement of graduates, however, we note the
present possibility of a declining trend.

The attrition rate for students in the Agriculture Certifi-
cate program was significantly higher than the 10 percent esti-
mated by the Project Paper. The dropout rate as of '.984 was 22
percent--19 percent for men and 33 percent for women. Male/
female comparisons were not meaningful in the Animal Health Cer-
tificate program because the numbers were too low; at the time
this report was prepared the diploma program was only 4 years
old. That the dropout rate was higher for women than for men is
partially explained by the fact that women who become pregnant
leave the school for 1 year. They are, however, readmitted with-
out penalty. (Records on reentry of such students were not
available, but pregnancy could account for the fact that one
woman graduated in 1980 although no women were admitted in 1979.)

Everyone we interviewed who was connected with the college
noted that, academically, female students perform as well as or
better than male students. They noted that the graduating
students with the second highest academic record in 1983 and the
highest record in 1984 were women.

In 1981, the Department of Agricultural Field Services as-
sured the Permanent Secretary of the Ministry of Agriculture that







-13-


all women trained at Botswana Agricultural College would be em-
ployed. However, interviews with the college administration
revealed that the Ministry has, in writing, limited to 15 the
number of women who can now be admitted to the Agriculture Cer-
tificate program. A similar limit, transmitted orally, was
placed on female admissions to the Animal Health Certificate
program, but applications from women did not exceed the limit.
The reason given for these limitations was the difficulty of
placing women in the field, particularly in remote areas.

Table 2 shows the placement of all women graduates from the
Agriculture Certificate program for whom information was avail-
able as of 1984. Women graduates of Botswana Agricultural Col-
lege are hired by the Ministry and placed in appropriate posi-
tions. Women generally are placed in positions equivalent to
those of their male counterparts, except that women agricultural
demonstrators receive favorable treatment by being placed in less
remote and somewhat easier assignments for "health reasons."


1.5.2 Curriculum


From the desk review of A.I.D.-supported education projects,
it became clear that one area in which gender analysis could
significantly affect project design is curriculum. If curriculum
content addresses women's productive roles, its relevance to
girls' and women's lives will provide a strong incentive for them
to become involved. Similarly, when curriculum is intended to
train people who will train others, as at Botswana Agricultural
College, the explicit inclusion of information on women's roles
will help these students make their instruction relevant to the
people they teach.

Since the final evaluation of the college expansion project,
the college has published a course catalogue for the certificate
and diploma levels. The course descriptions are thorough, and
the sequencing of courses reflects the broad range of agricul-
tural and animal health activities in Botswana. There is no
apparent bias in favor of either subsistence farming or livestock
production. Where appropriate, course descriptions refer to
regional differences, including variations in soil and climate
types and cultural and traditional patterns of farming. The
curricula for these programs and the courses on extension, com-
munication methods, and rural sociology could incorporate issues
related to gender differences in labor and resource bases. We
were unable to meet with the instructor in rural sociology and
extension to discover whether he actually does deal explicitly
with gender.








-14-


Table 2. 1984 Location of All Women Agriculture Certificate
Graduates From Botswana Agricultural College



Position Number


Department of Agricultural
Field Services
Agricultural Demonstrators 26
Dairy Agricultural Demonstrator 1
Land Utilization 2
Agricultural Information 20
Subtotal 31

Botswana Agriculture College
Lecturer 10
Diploma Students 7
Subtotal 8

Department of Agricultural
Research
Research Staff 7
Left the Ministry of Agriculture 1
Subtotal 8

Total 47

OThese women now hold diplomas.
Source: Bettles 1984, 27.


In interviews, as we explored the ways in which Botswana
Agricultural College courses prepare graduates to work with women
farmers, we were told that "if they [graduates] go to the fields,
they will work with the farmers they find there--in many cases,
or most cases, women." In other words, students are instructed
in fieldwork methodologies but with little or no explicit consid-
eration of women.

We concluded that the comprehensiveness of the curriculum
might well constitute a good example of a "gender-blind" approach
that educates about agricultural practices in an inclusive and
integrated way. However, including gender awareness in course
descriptions could add an additional assurance that gender issues
will be addressed. It might also increase the sensitivity of
Botswana Agricultural College graduates to wa-s in which gender-
related patterns of production might affect their work priorities
as agricultural demonstrators.








-15-


1.5.3 Findings on the Relation Between Integrating Women
and AchievinQ Project Objectives


Findings en the importance of the relationship between the
integration of women into the project and achievement of greater
productivity and well-being for small farmers and herders are
discussed in the context of agricultural production and the agri-
cultural extension system.

It is impossible to determine the impact of an agricultural
education institution on production without a complex, long-term
study. Undoubtedly, the drought and the resultant declining
production have offset any positive impact that tht Botswana
Agricultural College may have had. However, there is a need for
greater understanding of the relative success of graduates in
their various roles as agricultural demonstrators, researchers,
ministry personnel, and the like. A follow-up study of college
graduates, planned for the near future, should begin to fill this
need.

Having learned about gender-related roles in agriculture, we
examined the agricultural extension system to see whether gender
was a factor in either staffing or effectiveness of delivery
(i.e., in affecting farmer productivity).

A 1983 study (Ramolemana and Hobbs) surveyed 213 people con-
cerned with extension to determine what factors were considered
to have the greatest influence on extension effectiveness and on
crop-production improvement. The study identified five major
problems affecting extension worker effectiveness:

-- adequate housing for agricultural demonstrators

-- Inadequate performance review and promotion procedures
for agricultural demonstrators, district agricultural
officers, and regional agricultural officers

Failures in the supply of needed seed varieties

-- Transportation difficulties in the field

Inadequate opportunities for in-service training.

Our interviews revealed the same issues; however, none, with
the possible exception of transportation, is specifically a
gender-related problem. Transportation, particularly riding
bicycles on sandy roads, was sometimes cited as a greater problem
for female than for male agricultural demonstrators. Using
motorbikes instead of bicycles to ease the problem worked to the







-16-


disadvantage of female agricultural demonstrators because women
traditionally have not ridden motorbikes. In our interviews,
however, we found widespread agreement that transportation diffi-
culties affected both male and female agricultural demonstrators
and that the use of motorbikes, given adequate supply, would
prove a good option for both.

Another problem faced by agricultural demonstrators was
identified by the principal of the college: the relatively poor
image of agricultural extension workers compared with tne image
enjoyed by community development, literacy, and health-extension
agents. He blamed Government policies toward agriculture for
this image. He considered the question of whether women could be
good agricultural demonstrators to be a diversion from the more
important issue of how Government promotion and support of all
agricultural demonstrators could enable them to be more effec-
tive. He pointed out that programs in social and community
development, health, and education/literacy all use women effec-
tively as extension agents and send them to remote regions with-
out unusual problems. .

There have been several studies and much discussion con-
cerning whether women are or can be as effective agricultural
demonstrators as men. Bond (1974) and Bettles (1984) both found
little or no prejudice against female agricultural demonstrators,
although Bond found a small number of crop farmers to be reluc-
tant to trust the knowledge of women in cattle matters. Bettles'
findings are more mixed but indicate that most groups questioned,
with the exception of male agricultural demonstrators, believe
that women can be effective agricultural demonstrators. The male
agricultural demonstrators stated that women could not do exten-
sion work because of transportation problems, lack of strength,
fear of traveling in the forest, and difficulties working with
livestock, among others.

To avoid repeating the work done in these studies, we took a
slightly different approach in our interviews. First, we asked
people to describe the characteristics they considered to be most
important in an effective agricultural demonstrator; then we
asked whether women or men had any special advantages or dis-
advantages in light of these characteristics.

There was considerable agreement on the qualities that
interviewees deemed important for agricultural demonstrators.
These included willingness to go into the field, work with
people, and work long hours; patience; good communication tech-
niques; and appropriate knowledge. Most interviewees believed
that these qualities had little or nothing to do with gender.
However, they pointed to three gender-related issues--pregnancy,
alcoholism, and location--that they believed did influence the








-17-


effectiveness of agricultural demonstrators. Table 3 summarizes
these responses.

Pregnancv. A high (unknown) proportion of female agricul-
tural demonstrators become pregnant, and both they and their
supervisors and colleagues agree that they lose about 9 to 12
months of active extension work during pregnancy and nursing.
They are not expected to travel to the fields or to do strenuous
work for much of this time. One woman explained to us that among
educated women (as female agricultural demonstrators are), there
is general consensus that they do not want to live as women farm-
ers do "who must work even when they are pregnant and nursing."


Table 3.


Survey of District Agricultural Officers and
Regional Agricultural Officers


Question .


What are the most common problems
you experience with your female
agricultural demonstrators?


What are the most common problems
you experience with your male
agricultural demonstrators?


Response


Transportation
Pregnancy
Health
Weakness

Drinking
Laziness
Absenteeism
Incompetence
Poor discipline


Source: Bettles 1984



Alcoholism. A high (unknown) proportion of male agricul-
tural demonstrators are frequently drunk and, as a result, do not
visit farmers regularly. This was repeatedly cited as a male
problem, and some people speculated that more job time is lost
because of alcoholism than pregnancy. One person noted that
alcoholism seems to be higher among male agricultural demonstra-
tors than among male extension agents in other ministries, and he
suggested that this was because of the unfavorable working condi-
tions for agricultural demonstrators .n general. Bettles' (1984)
survey of district agricultural officers and regional agricul-
tural officers also notes these two problems.


_








-18-


Location. When kqotlas (traditional community-based
decision-making meetings) are used to transmit information to
farmers, it is less acceptable for female agricultural demonstra-
tors than for male agricultural demonstrators to make presenta-
tions. When agricultural demonstrators must travel to work at
cattle posts, which often requires overnight journeys away from
villages, it is more difficult and less acceptable for women to
make the trip.

In our interviews, people frequently commented that female
agricultural demonstrators had easier access to female farmers,
whereas male agricultural demonstrators could work more easily
with male farmers. While this is widely assumed to be true, the
evidence that female farmers in Botswana are better served by
female than by male agricultural demonstrators (or that male
farmers are better served by male agricultural demonstrators) is
less clear. Bettles (1984) shows that both female and male agri-
cultural demonstrators spent more time with male farmers than
with female farmers. Male agricultural demonstrators' estimates
of time spent with female farmers were higher than those of fe-
male agricultural demonstrators, while female agricultural demon-
strators' estimates of time spent with male farmers were higher
than those of male agricultural demonstrators (see Table 4).

Our interviews, especially those with agricultural demon-
strators and regional and district agricultural officers, could
not substantiate greater access of women to women or of men to
men. People frequently did comment that (1) youth was a liabil-
ity that all agricultural demonstrators had to overcome by prov-
ing themselves to the farmers, and (2) female farmers in general
were more open to working with and learning from agricultural
demonstrators than were male farmers.

The problem mentioned concerning female agricultural demon-
strators' access to male farmers at kqotla meetings and at cattle
posts does affect extension delivery. However, the timing and
location of agricultural demonstrators' contacts with farmers are
possibly more important determinants of the number of farmers
reached and the effectiveness of the contacts. Female farmers
have far more demands on their time than male farmers (see
Section 1.1) and are less free to attend meetings. Thus, to
reach women, agricultural demonstrators must go to locations
where women perform their domestic tasks in close proximity to
each other. Agricultural demonstrators could also more effec-
tively reach female farmers by living in village areas with dem-
onstration homestead plots, water catchments, chickens, and the
like, or they can reach women by training villagers as demon-
stration farmers. Because much of women's work is done in and
near villages, the impact on women's activities of such efforts
by agricultural demonstrators is likely to be great.








-19-


0-


50









Ct2
W -
0 +04








0)

40

o a,






0




-4
0 -s
41 U







CO
















E'
c 4
0 0
0




54
0-.






















0
0
4C -







Q)
0


0


tri cO


r-4 v r-I (n


0 .-1

OC


0 :3
0 a0
o c
S3



0E
1> x
.C -- 0
4-i4J


c
C 6



0 0
o2 2


ZEZW ~


4J
ro-





C4
-,-
0





4-5
-,4









0003
Q 3: 4


(1)
) 0
0 z


4-


ci


-4
V-4 0








0H
-4


a 4 0



0 4


Qa


(0








-20-


Finally, the extension "package" or technology promoted by
agricultural demonstrators influences the decision about whom to
work with. If early plowing, for example, is chosen as a strat-
egy for increasing productivity, then the agricultural demonstra-
tor should know who decides when to plow and who does the actual
plowing. There are areas in Botswana where the chief or headman
still decides when to plow, and farmers will only begin when he
gives the word. There are other areas where men decide when to
plow and do the plowing, areas where men decide but women plow,
and areas where women decide when to plow and do the plowing.
Advice to men to plow early has been ineffective in areas where
they follow the headman's lead and in areas where women are ex-
pected to plow but have other major demands on their time.


1.6 Findings and Recommendations


1.5.1 Summary of Findings


The Botswana Agricultural College Expansion project made a
serious and, for the most part, effective effort to integrate
women into the faculty and student body of the College. However,
the female professional faculty has remained at approximately
one-third of the professional staff, even as the College ex-
panded, while the proportion of female technical faculty has
dropped significantly. Recruitment of female students received
the greatest attention and represents the area of greatest proj-
ect success in integrating women. The project was less success-
ful in promoting women's involvement in traditionally male-
dominated animal sciences; most women students continue to
concentrate on the agricultural sciences.

The curriculum redesign is broad, inclusive, and, appar-
ently, gender-blind. However, because "gender-blindness" is not
synonymous with explicit attention to the roles women play in
Botswana's agriculture and agricultural development, we suggest
that gender issues be explicitly incorporated into Botswana Agri-
cultural College courses.

Project design and implementation gave little attention to
the relationship between the college expansion and the impact on
agriculture in general and on female farmers in particular. No
data exist on the impact of Botswana Agricultural College gradu-
ates, on agricultural productivity, or on the relative effective-
ness of male and female agricultural demonstrators. There are
indications of some differences in male and female agricultural
demonstrators' access to farmers and number of productive days
spent on the job.








-21-


1.6.2 Recommendations


1. A program for the prevention of unwanted pregnancies should
be initiated at the college and maintained for agricultural
demonstrators. The impact of pregnancy on time spent on the
job was the only clearly gender-specific issue limiting the
effectiveness of female agricultural demonstrators. It is
our impression from our interviews that a large number of
these pregnancies are unplanned. A comprehensive family
planning program could be easily and inexpensively imple-
mented. Such a program should include the following:

-- Provision of family planning information, a medical
examination, and appropriate supplies on entry to the
college and as requested throughout the course of study.
Introducing such a program at the beginning of the
course of study will also lower the number of women who
drop out because of pregnancy.

-- Routine and continued provision of free information,
supplies, and medical services to female agricultural
demonstrators by the Department of Agricultural Field
Services.

2. The comprehensive and gender-blind approach to the agricul-
tural education curriculum at the Botswana Agricultural
College suggests two changes to ensure explicit attention to
the poorest subsistence farmers, many of whom are women.

-- Explicit mention of gender-based divisions of labor and
related gender issues (e.g., time, location, and labor
constraints) in descriptions of courses in rural sociol-
ogy and extension. Although current course descriptions
lend themselves to the treatment of gender issues, they
do not explicitly deal with them; explicit mention in
the course descriptions would not only ensure that
gender-related issues are addressed but it would also
increase students' awareness of gender issues and their
importance to agricultural demonstrator fieldwork.

-An increased number of courses related to homestead/
compound vegetable gardening. Such an emphasis could
have two advantages: (1) it would place farming within
the time, labor, and location constraints of many female
farmers in Botswana; and (2) the increased vegetable
consumption that might result would lower vitamin defi-
ciencies caused by lack of vegetables in the diet.







-22-


3. To assess the effectiveness of male and female agricultural
demonstrators in reaching male and female farmers, the pend-
ing follow-up study on Botswana Agricultural College gradu-
ates should collect the following data:

Specification of agricultural demonstrators' visits to
farmers (or farmer contacts) by gender of demonstrator,
gender of farmer, type of household (long-term absent
husband, single woman, widow, divorced woman, man/woman
farming together, single man)

Location of agricultural demonstrator/farmer contact,
with analysis of gender access to location

Timirn- of agricultural demonstrator/farmer contact, with
analysis of gender implications

-- Extension package contfer: and priorities and implica-
tions for who is cont zed where and when, and gender
implications ,

-- Number of days lost -o job because of pregnancy

-- Number of days los- to job because of other factors,
including drinking/alcoholism, by gender








2. THE ACCELERATED MAHAWELI PROGRAM IN SRI LANKA


by

Janet E. Benson, Anthropologist
(Kansas State University)

and

Jan Paul Emmert, Rural Sociologist
Center for Development Information and Evaluation, A.I.D.


2.1 Summary


The Accelerated Mahaweli Program of the Government of Sri
Lanka is an extremely large and complicated river basin develop-
rent project using over $1.0 billion in assistance from a number
of international donors. The Mahaweli Program has dammed the
Mahaweli Ganga, Sri Lanka's longest river, and is extending
irrigation to several settlement schemes or "systems" being
established in the sparsely settled dry zone of north-central Sri
Lanka. By dramatically increasing the amount of land under in-
tensive agriculture, the project is intended to create employ-
ment, increase national production of rice, and provide land to
landless people from the heavily populated wet zone. It will
also generate needed hydroelectric power for industry.

Sri Lankan women play a key role in agriculture. They are
involved in transplanting, weeding, harvesting, and threshing
paddy (rice); preparing fields; caring for homestead gar:.ens,
tree crops, and livestock; and managing farms. Agency for Inter-
national Development (A.I.D.) planners anticipated that women
would not only continue this involvement on the Mahaweli projects
but would also participate in the income-generation schemes nece-
ssary to promote regional development. Project Papers stated
that families in which women held special skills (not specified)
were to be given priority in settler selection.

Although Project Papers suggested guidelines regarding
women, they gave no specific implementation targets. In fact,
the Mahaweli projects funded by A.I.D. focus narrowly on the con-
struction of irrigation works; settlement activities are carried
out entirely by the Sri Lankan Government. Thus, i:omen have been
consulted by project officials or included as recipients of
training more as a response to problems arising during the
settlement process than as a deliberate implementation of A.I.D.
guidelines.







-24-


Seven major issues emerged from this study.

1. Settler selection. Despite statements in the Project
Papers, we found that officials who were selecting settlers from
the electoral rolls considered only the husband's background. A
common assumption seems to be that all household heads are men
and that men have primary responsibility for farming.

2. Agricultural extension. When husbands hold off-farm
jobs or lack farming experience, women oversee hired labor and
make day-to-day management decisions. Female heads of household
(approximately 5 to 10 percent of settler households at present)
also act as farm managers, either sharecropping land or working
it themselves with the aid of children and other relatives.
Despite women's roles as both workers and decision-makers, no
agricultural extension work is aimed at married womea,, and almost
all field staff (except for community development officers) are
men.

3. Land tenure. -The current system of land tenure has
several implications for project success. Among new settlers
selected from the electorate, men rather than women obtain allot-
ments in 90 percent of cases, regardless of who will take the
leading role in farm management. Among resettlers (previous
occupants of the area) or evacuees from dam sites, residence
requirements result in allotments of land to women in up to 35
percent of settler households. Land is significant not only as a
basic resource but also as collateral for bank loans. Although
women will continue to have some customary rights in the next
generation, in most cases legal title will go to male heirs.

4. Labor constraints. Planners' assumptions about the
availability of family labor were overly optimistic. Although
project success is predicated on the production of two crops of
rice per year using only family labor, settlers in irrigated
zones must hire workers during peak seasons. Planners apparently
lacked reliable information concerning what household members,
particularly women, could or were willing to do in peak seasons,
given their domestic work loads. Although day care centers had
been introduced to free mothers for agricultural work, the major-
ity of settlers were not using them.

5. Gender, agricultural diversity, and nutrition. The new
farming system emphasizes irrigated rice, a cash and staple crop
in which male labor predominates, rather than dryland agriculture
(now illegal), which was previously practiced in the north-cen-
tral areas. The relatively high level of nutrition in Sri LankL
has been associated with diversity of food sources, but intensi-
fication of rice production has reduced the availability of food
for family consumption on the Mahaweli Program.








-25-


6. Homestead gardens. Homestead gardens, which both women
and men may cultivate, could contribute more to family nutrition
and supplementary income than they do at present. Women could
keep milk animals (dairy cattle or water buffalo) if fodder could
be grown; however, settlers complain that the 0.5-acre homestead
plot size is insufficient for dairying and other activities
formerly carried out by women. Lack of water for homestead gar-
dens during the second cropping season is also a common problem.

7. Income-generation projects and employment. Viable
off-farm projects for women during slack seasons would supply
needed supplementary cash for the farm household. Lacking income
from minor crops or cottage industries, farmers on System H must
often take subsistence loans to maintain themselves between paddy
harvests. Training programs target only men and young unmarried
women, however.


2.2 Introduction


2.2.1 Project Setting


Sri Lanka is a small island republic the size of West Vir-
ginia. Its population of 15 million is 75 percent Sinhalese
(mostly Buddhist); minority groups include largely Hindu Tamils
of south Indian origin (20 percent), Christians, and Muslims.
Following independence from Great Britain in 1948, the Government
invested heavily in health, subsidized food, and educational
programs. As a result, socioeconomic indicators for Sri Lanka
are unusually positive for a South Asian country: high literacy
rate (86 percent,, hich life expectancy (69 years), low infant
mortality, -ad low birth rates. Sri Lanka has remained a poor
country, however, with a per capital income of only $320 in 1984.
Despite z slowing birth rate, the population has more than dou-
bled since 1946. Low productivity, restriction of the private
sector, and low prices for traditional export crops like tea led
to food deficits and widespread unemployment by the late 1970s.

Women's status in Sri Lanka is higher than in other South
Asian countries. The island's positive socioeconomic indicators
are undoubtedly related to women's relative freedom, at least
among the Sinhalese ethnic group. As of 1981, literacy among
woman (82 percent) almost equaled that among men (90 percent).
In India and Bangladesh, by way of contrast, literacy among women
is less than half that among men. Sinhalese kinship structure is
bilateral, meaning that relationships are traced through women as
well as men; both women and men customarily inherit land, and a








-26-


husband may reside in his wife's community after marriage (Leach
1961).

Sri Lanka is characterized by two major climatic and ecolog-
ical divisions. The wet zone, located in the island's southwest
quadrant, receives more than 2,000 millimeters (mm) of rain per
year and is the most densely populated area of the country. The
dry zone, which encompasses the rest of the island, receives less
than 2,000 mm per year and has an "effective dry period" of 3
consecutive months. The unreliability of rainfall rather than
its average amount limits rainfed agriculture, making irrigation
essential for the production of two crops per year (Johnson and
Scrivenor 1981). The Mahaweli Program therefore has enormous
potential for increasing productivity and promoting population
redistribution.


2.2.2 A.I.D.'s Women-in-Development Perspective


Current A.I.D. policy regarding the role of women in devel-
opment goes beyond an earlier concern with equity. A.I.D. as-
sumes that gender roles have important implications for project
success or failure.

It is difficult to reconstruct what the USAID Mission orien-
tation toward women-in-development issues was at the time the
Accelerated Mahaweli Program was being designed. One statement
in a Mahaweli Project Paper (USAID/Colombo 1980) suggests an
orientation similar to present agency policy:

Obviously Sri Lankan women have a key role in agricul-
ture (transplanting, weeding, harvesting, and threshing
of paddy; care of the homestead gardens and tree
crops), but it is expected that they will also be heav-
ily involved in the agricultural and nonagricultural
diversification on which the project depends for re-
gional development. To encourage and promote this,
families in which women have special skills are being
given priority in the selection of settlers. In addi-
tion, training courses are planned to provide new
skills which the women can utilize to supplement family
income. Also, there is a special effort underway to
recruit women for the provision of services which will
be of prime concern to the project (e.g., women agri-
cultural extension agents, health workers, and
irrigationists).

However, despite this statement, it appears that the Mahaweli
projects funded by A.T.D. focused narrowly on the construction of







-27-


irrigation works and were not viewed as part of the Mission's
women-in-development effort.


2.3 Project Design


The Accelerated Mahaweli Program of the Gcv-rnment of Sri
Lanka is large and complicated river basin development project
that is receiving over $1.0 billion in assistance from 11 inter-
national donors. It is also a massive resettlement project,
originally intended to relocate 450,000 people. The Accelerated
Mahaweli Program has dammed the Mahaweli Ganga, Sri Lanka's
longest river, and is extending irrigation to several settlement
schemes or "systems" that are being established in the sparsely
settled dry zone of north-central Sri Lanka. By dramatically
increasing the amount of land under intensive agriculture, the
project is intended to generate employment, increase national
rice production with the goal of achieving self-sufficiency in
foodgrains, provide land to landless people from the heavily
populated wet zone, generate hydroelectric power, and stimulate
regional development.

The program allots small farms (usually two acres of irri-
gated land, with half an acre for the homestead), as well as some
nonfarm plots, to voluntary settlers. Because paddy (rice) is a
dietary staple as well as a major cash crop, the Accelerated
Mahaweli Program heavily emphasized paddy production until
recently.

The Accelerated Mahaweli Program is planned and implemented
by the Mahaweli Authority of Sri Lanka, under the Ministry for
Mahaweli Development, which oversees the complex settlement pro-
cess and supplies farmers with credit, marketing, and extension
services. Program implementation requires clearing large areas
of jungle, creating dams and irrigation systems, preparing
fields, selecting farmers, resettling thousands of families, and
creating an extensive infrastructure, including roads, wells, and
training centers. Activities are supported by specialists in
land allocation, agriculture, engineering, marketing, community
development, and livestock production.

Support of the Accelerated Mahaweli Program has been central
to A.I.D. assistance to Sri Lanka for the last 7 years. This
analysis considers activities on the older system and three cur-
rent A.I.D. projects in support of the Accelerated Mahaweli Pro-
gram. Two of the projects focus on irrigation construction in
System B, and che third is a sector support loan, which the
Government of Sri Lanka may use for a wide range of development
activities, such as road and minor canal construction.








-28-


2.3.1 Field Study Methodology


Because we had a very limited time of 16 days in the coun-
try, we began our research by interviewing project staff and
settlers. In all, we conducted approximately 110 interviews with
scholars and officials and with members of farm families, many
from the same family. We traveled to System H, the earliest of
the project schemes, to gain insights into situations that may
appear later in System B.

Our choices, with all the limitations they imply for the
Women-in Development study, were as follows: First, we chose to
concentrate more of our time on System B, which offered more
promise of insight into gender issues than irrigation activities;
but we also spent a day in each of the three projects of the
older System H. Second, we focused on resettlement as well as
irrigation. We particularly tried to examine the allotment pro-
cess, which determines who has access to land. These two choices
left us without a single definitive project design document.
None of the A.I.D. Project Papers provided very detailed informa-
tion on resettlement plans, which were clearly not the primary
concern of the A.I.D. planners. We therefore drew inferences
about the design of System B from any sources we could find: the
A.I.D. Project Papers, Mahaweli Authority documents, interviews,
and comments in reports.


2.3.2 Gender Assumptions Underlying Project Design


System B Project Papers assumed that the major beneficiaries
of the project would be poor or landless people, farm families
and nonfarm settlers alike, who would come mainly from outside
the project area through a voluntary settler selection program
emphasizing agricultural skills. In fact, not all settlers are
poor; some are evacuees from headworks dam construction sites;
also, there are more residents of old villages within the project
boundaries than had originally been estimated. It was assumed
that benefits would also accrue to nonsettlers and the second
generation, for whom jobs would be created as part of the re-
gional development spurred by intensified agricultural produc-
tion. Planners also assumed that each household would contain
only husband, wife, and unmarried children. Because of concerns
about land fragmentation in the next generation, plot subdivision
is'illegal, and homestead allotments are too small to contain
more than one house and garden.

Aside from some general guidelines, gender-related assump-
tions underlying project strategy are not readily apparent from







-29-


A.I.D. Project Papers, and no implementation plan for settlement
activities is outlined. Project personnel also assumed that
married women were not appropriate targets for agricultural
training, income-generation activities, or volunteer health work.
All programs in these areas were aimed at men and young unmarried
women. This pattern may be due to a need for off-farm employment
for settlers' adult children or, at least partly, to the assump-
tion that married women occupied with their families and unable
or unwilling to participate in training programs. However, offi-
cials do recognize women's participation in agricultural
production.

According to project land officers, settlers and project
personnel assume that the single descendant allowed to inherit
the land allotment will probably be male. The result will be
that most legal titles will go to men in the next generation,
except in families that have only daughters. Although we en may
in fact have some access to land, they will not possess the
allotment certificates required to receive bank loans.


2.3.3 Accommodation of Project Design to the Target Group


Given the projects' scale and the scattered location of
potential settlers, beneficiary participation in project design
was minimal. We can, however, analyze whether implementers of
the present Mahaweli schemes, especially System B, take account
of settler needs identified in earlier projects. A particularly
important question concerning women is whether they have effec-
tive input into local policy and implementation decisions on
agriculture and resettlement in general. Because women have
important roles as cultivators and farm managers, decisions made
without their input may involve unrealistic innovations and ig-
nore constraints to production (e.g., lack of child care or agri-
cultural training). The greater the participation of settlers,
both men and women, in problem-solving, the greater the potential
for developing effective solutions to local problems. A more
participatory approach, in which settlers select their own lead-
ers and control their own organizations, could facilitate commu-
nication between farmers and settlers and help reduce some of the
built-in dependence of settlers on Mahaweli staff.


2.3.4 Accommodation of Project Design to Women


System B Project Papers mention that women's skills and
interests were not taken sufficiently into account during settler
selection on System H (the pilot project), and they state that on







-30-


System B, "families in which women have special skills are being
given priority in settler selection." The "special skills" are
not defined. When we interviewed Sri Lankan project officials
about settler selection, they indicated that only the background
of the husband (all settlers must be legally married) is con-
sidered. An implicit assumption at the electorate level, there-
fore, seems to be that all heads of household are men and that
men have primary responsibility for farming. The Project Papers
mention that training courses are planned for women so that they
can supplement family incomes, again without describing the pro-
posed courses. Project Papers also call for special efforts to
recruit women as agricultural extension agents, health workers,
and other public servants (USAID/Colombo 1981). However, no
gender-specific target figures are given in any of the project
documents.

It is not clear to us whether women's input affected settle-
ment design. A System B Project Paper states that "the decision
to increase the homestead size from less than 0.2 ha [hectares]
to nearly 0.3 ha was based on complaints made by women settlers
in System H that the smaller size homestead did not provide suf-
ficient space for gardening, tree crop production, and livestock
raising, activities in which women were actively involved." In
fact, the Mahaweli Authority did not implement such a change in
System B.

Planners have successfully maximized social solidarity on
System B by settling in a single hamlet, whenever possible,
people from the same parent community, ethnic group, and some-
times the same extended family. Apparently System H's greater
social heterogeniety offers considerably less sense of community
and social support for families. This decision has had impor-
tant practical consequences for women, particularly those with
young children. Women's mobility was apparently much more re-
stricted on System H because they had no relatives or trusted
neighbors to care for children in their absence.

Women may be particularly disadvantaged by the installation
of agency employees as leaders of settler organizations. Women's
clubs organized in rural India under the administration of male
government officers have proved ineffective in promoting economic
development because of poor communication and the male adminis-
trators' stereotyped urban middle-class view of women's roles.

SAlthough there is no rigid bar to female participation in
irrigation turnout and community development groups, members are
men, and leadership by women is rare. There seems to be no regu-
lar mechanism for the expression of women's needs except through
male-dominated groups. As a result, women's opinions on a number
of issues, such as the type of child care desired and special







-31-


constraints to agricultural production, are not clearly
expressed.

One accommodation of project design to women on System H has
been to require that at least half the community development
officers be women. System B has only male officers at present,
but female staff are to be recruited at a later stage of the
project. Finally, day care centers (supported by UNICEF) have
been introduced in both systems, apparently in response to
settler requests.


2.4 Gender Issues


2.4.1 Local Production and Consumption Systems


There are three categories of Mahaweli settlers: prior resi-
dents of areas now constituting settlement schemes, evacuees from
areas flooded by large dams upstream, and eligible new settlers
selected from election rolls across the country. We do not have
detailed household-level data on women's economic roles among
settlers from these groups prior to resettlement, and it is
likely that considerable variation existed. Some roles are also
affected by structural features of resettlement such as the land
allocation process and community layout.


2.4.2 Division of Labor by Gender


The subsistence economy of villagers in the North Central
Province just prior to colonial rule was based on two kinds of
agriculture, shifting dryland cultivation (chena) and tank-irrig-
ated paddy (rice). Women and men participated in both types of
production, but women played a greater role in chena and men in
rice cultivation. Men cleared the dryland fields and guarded the
crops by night; both sexes sowed, weeded, and guarded the crops
during the day. The chena produced finger millet of high nutri-
tional value as well as other grains, pulses, and vegetables.
Men played a more dominant role in paddy production, although
women participated in harvesting and threshing. Women were also
responsible for post-harvest grain processing, cooking, fetching
water and firewood, and caring for children (Schrijvers 1983).

According to the Mahaweli settlers and officials inter-
viewed, current activities of men and women on the Mahaweli
schemes are similar to the precolonial activities, except that
the Accelerated Mahaweli Program is focused on irrigated








-32-


agriculture, in which male labor predominates, rather than on
dryland farming. Although, under the right conditions, irrigated
rice can maintain a much larger population than dry-land farming,
focusing on one crop has had unfortunate results. The present
farming systems on the Mahaweli schemes are relatively undiver-
sified compared with earlier systems in the north-central region,
where tank-irrigated rice cultivation and chena cultivation ex-
isted side by side, and fewer staple fruits and vegetables are
now available for family consumption. Lack of diversification
has had serious economic and nutritional consequences.

Homestead gardens are potentially important component of the
Mahaweli farming system. Such gardens may be planted and tended
by women, their husbands, or other family members. Who actually
c-11tivates the garden probably depends on a number of factors:
whether a woman has young children to care for without assis-
tance, whether a male settler is occupied with paddy cultivation,
what the settlers' farming backgrounds are, and especially what
the incentives for homestead gardening are.


2.4.3 Land Tenure


In the Sinhalese areas of Sri Lanka prior to British rule,
ownership and control of land by women was much more common than
in some other South Asian countries. Among all classes, both
sons and unmarried daughters inherited property. Marriages were
classified as either deqa (virilocal) or binna (uxorilocal or
matrilocal). Binna marriages, in which a husband came to live
with a land-owning wife or her family, were common. In such
cases, the married daughter, and not her husband, would inherit
land after her parents' death (Yalman 1967). The jungle outside
the village could be cultivated by anyone, male or female, but
rights were held only as long as the land was actually used.
Women heads of households (widows or those separated from their
husbands) could support their families by dryland cultivation and
by allowing their paddy fields to be sharecropped by others
(Schrijvers 1983).

According to some accounts, women's land rights began to
erode during the British colonial period. The Land Development
Ordinance of 1935 required that crown lands (land other than
house sites, paddy fields, and gardens in use at the time of
British survey) be given only on permanent lease; the tenant was
required to nominate a single heir for the holding. (This ordi-
nance contravened the cultural tradition of equal inheritance by
all children.) Our interviews with land officers on the Mahaweli
Program confirmed that this law is still in force and that in 90








-33-


percent of the cases people nominate a successor, usually the
eldest son.


2.4.4 Ownership and Control of Assets


Poor settlers in the Accelerated Mahaweli Program areas
possess little property other than houses and paddy plots.
Wealthier settlers may own tractors, draft animals, bicycles,
sewing machines, and other goods. We do not have detailed data
on ownership of property other than land, but some authors sug-
gest that machinery and animals handled by men, such as draft
animals, are considered male property.

In our household interviews, we tried to learn whether men
and women have separate income and expenditure streams and how
families handle finances. Men and women sometimes earn income
independently, although opportunities are fewer for women than
for men. Men are more-likely to market the main field crops,
paddy or chilies; women sometimes sell vegetables, fruits, eggs,
or small quantities of grain to itinerant traders or local mar-
kets. Particularly among poorer families, both women and men
work as agricultural laborers. We also met a few unusually
entrepreneurial women engaged in trading or cottage industries.

Earnings of both men and women usually go into a common
household fund. Most people interviewed stated that this fund
may be managed by either husband or wife but that both partici-
pate in major financial decisions. Some husbands purchase cigar-
ettes or alcohol without consulting their wives, and in a few
households men completely control the household budget. In other
cases, the husband admitted turning o-er all or most of his in-
come to his wife, who takes the leading role in expenditure
decisions.


2.4.5 Male Off-Farm Employment and Female Farm Manaaers


In several households we interviewed, husbands were engaged
in off-farm employment (usually Goverraent service) while wives
supervised hired laborers and cultivated on a day-to-day basis.
Our interviews also indicated that these women often undertook
much of the farm management as well. One bank official estimated
that approximately 30 percent of the agricultural loans made to
Pimburettewa (part of System B) settlers were in the wife's name,
because she held title. However, there were no firm gender-
disaggregated data on this point.








-34-


Female heads of household, usually widows, also manage
farms, sometimes with the help of children and other relatives.
Women may allow sharecropping of their land if their children are
small and they have no adult male assistance, but they try to
cultivate the land by themselves if they have sufficient family
labor. Perhaps 5 to 10 percent of settler households on System B
are headed by women who have responsibility for their families.
This percentage is likely to increase in the program area as the
population ages.

Because of serious labor shortages at peak periods in the
paddy cultivation cycle, there is little outmigration of men
seeking manual labor. On the contrary, laborers (mostly men, but
some young women as well) migrate seasonally from the wet zone to
Mahaweli Program areas.


2.5 Project Analysis


2.5.1 General Effects


The Accelerated Mahaweli Program seeks to generate employ-
ment and hydroelectric power, increase production, spur regional
development, and improve settlers' quality of life. What have
the accomplishments been to date?

Constraints related to irrigation canal construction, de-
layed settlement in certain areas, and the withholding of credit
to some settlers caused considerable disruption of the develop-
ment plans within the two systems (H and B) examined here.

Of the elements that affect quality of life, nutrition is
particularly affected by gender-related factors. The present
farming systems, which are based on monocrop rice cultivation, do
not provide sufficient diversity of production for good family
nutrition. The possibility of women's dryland chena cultivation
has been effectively eliminated by the loss of crown land to
irrigation schemes and homesteads, and sufficient water is not
usually available for homestead gardens during the dry season.
Farmers focus almost exclusively on cash crops such as rice and
chilies, despite recent efforts of project staff to encourage
diversification.







-35-


2.5.2 Effects of Project on Gender-Related Goals


Basic resources and services in the Mahaweli Program include
land, water, capital, credit, employment, and training. Although
not all of these resources and services were funded by A.I.D.,
they are all a part of the basic settlement package of the
Accelerated Mahaweli Program. No targets for women (e.g., number
of trainees, loan beneficiaries) appear in A.I.D. Project Papers,
perhaps because funding concentrated more on the physical infra-
structure than on resettlement. Gender-disaggregated data were
scarce. We were able, however, to locate some information on
allottees and project staff. Officials and bank managers also
gave us rough estimates of women given credit, managing farms, or
heading households.

Land. Some accounts of System H state that women have lost
rights to land under the project; the actual situation is con-
siderably more complicated and varies according to category of
settler. Evacuees and-'prior residents of an area being settled
are entitled to receive land, regardless of gender, if they are
married. Officials on System B estimated that approximately 35
percent of allottees there are women. In villages with many
evacuees or prior residents, the percentage is higher; for ex-
ample in one hamlet there were 75 female and 102 male land-
holders. The allotment process is more restrictive for new set-
tlers chosen from the electorate, with the result that usually
only men are considered for allotments and female heads of house-
hold tend to be excluded. Because most later settlers will be
chosen by this method, the percentage of women allottees in Sys-
tem B will gradually drop.

Several features of the Land Development Ordinance of 1935
appear to have negatively affected women's access to land in the
settlement schemes: restriction of land titi and land inheri-
tance to one person, and the assumption that the person inherit-
ing will probably be male. As long as land remains for allotment
on the settlement scheme, landless children of either gender can
apply for plots in their own right; once those plots are gone,
however, the eldest son will in most cases inherit the land.

Although legally land cannot be subdivided, in practice it
is. In fact, this is a common occurrence in settlement schemes
worldwide: land may officially remain in the name of one person
while actually being shared with others according to customary
practices. Women will continue to have some customary rights to
land in the next generation, but they may face difficulties ob-
taining bank loans and purchasing inputs because their land own-
ership is not legally recognized. This development in turn will
decrease the effectiveness of the project.








-36-


Water. Availability and management of water affect all
settlers, male and female. Officials assume that irrigation is a
concern predominantly of men, although women apparently
participate in turnout (irrigation self-help) groups if they are
heads of household or if their husbands are unable to attend.
On-farm water management would be most effective if the actual
farm manager, regardless of gender, received training and
information.

Credit. Our information on credit availability, at least
from banks, indicates that female allottees have no more dif-
ficulty than do male allottees in obtaining crop loans. However,
because almost all loans to settlers must be secured by a land
allotment, women whose spouses are allottees can seldom obtain
loans in their own names. According to personnel at the Bank of
Ceylon in System B, 25 to 30 percent of seasonal loan recipients
are women. Although most of these women have husbands, they not
only take loans, but they also actively manage the farms using
hired labor.

Crop loans are available (almost entirely for rice), but
both male and female entrepreneurs have difficulty obtaining bank
loans for anything other than farming. Almost everyone already
has a crop loan, and banks will not exceed certain credit limits,
a fact that inhibits the development of small businesses. Be-
cause bank officials believe that local demand is insufficient,
they are dubious about the profitability of most rural
enterprises.

However, not all settlers enter these schemes without re-
sources. We met a number of entrepreneurs, including several
women. The Mahaweli Authority has set aside commercial allot-
ments of up to 1.5 acres for various commercial activities rang-
ing from boutiques (small general stores) to rice mills. In the
Kalawewa project area of System H, 217 people have received al-
lotments for a little over 42 acres. No official gender-
disaggregated figures were available, but an administrator es-
timated that about 2 percent of these commercial allotments are
leased to women, mostly for boutiques and dressmaking shops.
Women from well-to-do families, who make up a minority of
settlers, obviously have more investment opportunities than do
poor women.

Off-Farm Employment. Employment opportunities for women, a-
for most men, are limited primarily to agricultural labor aind
construction work, when it is available. Women are reported to
constitute about 20 percent of the agricultural laborers on Sys-
tem B's experimental farm. Both men and women receive about 30
rupees (a little over $1.00) per day for agricultural and








-37-


construction work. The women laborers we interviewed, who were
usually married and had children, needed the extra income to help
support their families.

An issue related to employment is the extent to which the
Mahaweli Authority attempts to recruit female staff members. In
System B, where settlement began only in 1982, the only female
staff members were health volunteers and clerks and secretaries
in the project office in Aralaganwila. There were no female
field staff. The view of the project leadership was that under
the frontier conditions prevailing in much of System B, assigning
women to the field staff was not realistic; however, this situa-
tion was expected to change within a couple of years.

The health volunteers are young women between the ages of 18
and 25, selected from the settler population, educated up to the
10th standard, and trained by the Social Services Department in
such subjects as health, nutrition, child care, and family plan-
ning. Much of the work of health volunteers on both systems in-
volves treating malaria. There eventually should be one health
volunteer for every 50 families. Volunteers receive no salary
but value their training and the chance to contribute to the
community. In a society that holds education in high esteem,
this training also gives young women prestige.

In addition to the health volunteer program, System H has a
few female officers. There were far more men than women in all
positions except those of secretary and community development
officer. Officials make special efforts to select community
development officers from within the project. According to
Mahaweli Authority rules, each block (2,000-2,500 households)
should have one male and one female community development offi-
cer, with female officers concentrating on women's and children's
welfare. In the Kalawewa project area of System H, for example,
10 of the 13 community development officers were women. Because
men are apt to be promoted more rapidly than women, the tendency
has been for women to remain community development officers while
men move up, a fact that has resulted in a preponderance of women
in this post. At the unit (200-250 households) level, three of
the 88 positions were filled by women, all field assistants.

Where the Mahaweli Authority has required local recruitment
of women, it has been quite successful in hiring qualified female
staff. However, women have been assigned at the higher project
levels only in positions considered to be of interest to women--
that is, in community development positions dealing with health,
nutrition, and child care, and not in agriculture. Although the
use of women in key field positions at the unit level seems to be
successful even where women are providing agricultural extension
services primarily to men, there are still only a few women at








-38-


that level even in System H. As in many developing countries, a
middle-class urban model of female roles influences policy deci-
sions about appropriate employment for women.

Training Opportunities. The Mahaweli Authority provides
women with information on agricultural topics at several Home
Development Centers it created in System H. The Home Development
Centers incorporate several innovative approaches, and they
respond positively to settler requests for income-generation
projects. So far, only a few centers are operating (Kalawewa has
one for approximately 55,000 settlers), and their impact appears
to be minimal. However, some areas still lack development cen-
ters, and there are no programs for married women at any of the
centers. At one of the centers, UNICEF has provided equipment
and bicycles so that the women can commute daily from home, an
innovative solution to the lack of mobility that has limited
women's training opportunities in most South Asian countries.

Students attend the Home Development Center program after
completing the 10th standard. The Kalawewa project center's
4-month curriculum includes 4 days on national traditions and
customs; 2 weeks on home science; 2 weeks on health and sanita-
tion; 2 weeks on cattle and agriculture; and 2 weeks on needle-
work and handicrafts. The remaining 2 months are spent in vil-
lage projects such as home gardens. Doctors and health educators
give guest lectures about common problems such as worms, dysen-
tery, childhood diseases, and contaminated water. Agricultural
topics include dairy cows, seed beds, fertilizer, and agrochemi-
cals (not advised for home gardens).

During our visit, the Kalawewa project center was using two
borrowed sewing machines but expected to receive 20 more from
UNICEF. The students were producing jams and chutney, sewing
children's clothes, and making handicrafts to be sold at an up-
coming national rural mobile market. The center had initiated
the jam and chutney making in response to local demand; after
farmers had grown too many vegetables and glutted the market,
they asked officials for some way of preserving them. The
Mahaweli Authority gives the center loans to buy vegetables.
UNICEF supplied the canning equipment. Students who have fin-
ished the course may return to the center later to use its equip-
ment and make products for sale.

The Mahaweli Authority had provided the center with a shop
at Nochchiyagama, and another will be opened at Kalawewa. Be-
cause the sales outlet was in its first month of operation, it is
hard to say how profitable these endeavors will be. An apparent
demand for school uniforms and children's clothes suggests that
tailoring would be a good occupation to encourage; it is not
clear, however, whether the jams and chutneys will be as much in








-39-


demand. Other crafts such as wall hangings, wood carvings, and
handmade lace require an urban market. Neither staff nor
students could tell us what the expenses were or, in some cases,
how items would be priced.

In addition to income-generation projects at the Home Devel-
opment Centers, there were a few programs elsewhere on System H.
For example, the Mahaweli Authority started a vocational training
center a year before our visit to the Ipalogama Nochchiyagama
project area. This center teaches young women to make baskets
and straw hats, which are sold locally; 30 students per course
attend the 6-month program. In the Eppewela project, schemes
included a rexine (plastic products) center at Theriyama; a sew-
ing center at Pahalagama financed by a 10,000 rupee interest-free
loan for materials from the Mahaweli Authority; and a handicraft
training center (plaster of paris religious statues) at Tambut-
tegama, with mostly women participants.

Most of the income-generation schemes that were examined are
commonly started without any clear market orientation, analysis
of profitability, or training in elementary business skills. The
Mahaweli Authority in effect subsidizes some activities by provi-
ding interest-free loans, transportation, shops, or a ready mar-
ket for goods.


2.6 Lessons Learned


Providing new technology without adequately considering its
social impact may not only threaten project success in the narrow
sense of production goals but, more broadly, may also prove det-
rimental to farmers' well-being. For example, a more thorough
initial consideration of the farming system being introduced on
the Accelerated Mahaweli Program would have addressed issues of
crop diversification, gender roles in agriculture, potential
labor demands on men and women, and the necessity for training
all farmers in new agricultural methods, particularly water
management.

Several problems that Mahaweli Authority officials face
today might possibly have been avoided it gender analysis had
preceded project design. Larger homestead plots might have been
created, more attention might have been paid to gardening and
tree crops at an earlier stage, and greater emphasis might have
been placed on supplementary income projects for both married and
single women. More information on women's daily and seasonal
work schedules might have improved project design. Given a real-
istic appraisal of likely labor inputs from a farm family during
peak seasons, more emphasis might have been placed on crop







-40-


diversification at an earlier stage of the project. However, the
size and complexity of the Accelerated Mahaweli Program and,
particularly, the speed with which it proceeded probably pre-
cluded such considerations.


2.6.1 Gender Roles, Agricultural Diversity, and Nutrition


In most agricultural societies, including Sri Lanka, the
existence of complementary male and female farm activities helps
to maintain diversified farming systems. Women often specialize
in certain agricultural operations or crops compatible with
household duties while men concentrate on others. In Sri Lanka,
rice production has traditionally been considered the responsi-
bility of men (although women assist during peak seasons and
carry out certain field and postharvest tasks), and dryland agri-
culture receives more attention from women. Sri Lanka is unusual
among developing countries because of the importance of fruits,
vegetables, and fish in the diet.

In lowland areas of the dry zone, women contribute to the
diversity necessary for good family nutrition by tending fruit
trees in homestead gardens and cultivating pulses, millet, and
cassava on nonirrigated land. In the Kandyan highlands, from
which many of the new settlers came, home gardens are very impor-
tant for the nutritional status of the family and also provide
income throughout the year. Home gardens may contain 16 or more
different species of fruit and vegetables and provide a wide
range of food needed by a household, including all spices (which
are also cash crops) and coconut (a major dietary staple).

The virtual elimination of women's subsistence agriculture
in the Mahaweli Program in favor of irrigated rice has com-
promised family nutrition. Fewer foods are available for house-
hold consumption, particularly in the first years of settlement
when fruit trees are just being established, and farmers have
little money for food purchases between harvests. Officials on
System B now recognize the problem inherent in focusing on only
one crop and encourage homestead gardens and crop diversifica-
tion. The Mahaweli Authority suggests a model garden plan, cre-
ates demonstration plots, and provides settlers with seeds and
saplings.

If women's economic activities were focused on crops other
than paddy, the farm family could benefit from both improved
nutrition and an alternate source of income. A dairy animal
could provide milk for the household and for sale.







-41-


Viable off-farm projects for women during slack seasons
could also supply needed supplementary cash. Without income from
minor crops or cottage industries, farmers on System H must often
borrow money to maintain their subsistence needs between paddy
harvests.


2.6.2 Gender and Rice Cultivation


Women's labor is also central to project success as narrowly
defined in terms of rice production. According to an evaluation
carried out at the same time as this study (Nott et al. 1985),
the Mahaweli Project can be viable with paddy as the main crop
only if y elds exceed 100 bushels per acre for two crops a year
and if the family supplies all labor. Irrigated rice is very
labor-intensive compared to most other crops. Because women have
daily domestic and child care duties for which they receive
little help when they are in a nuclear family setting, the addi-
tional labor demands oA women during peak planting and harvesting
seasons are particularly stressful.

Although planners assumed that family members would carry
out all agricultural operations, in fact husband and wife alone
(most children are young or in school) are not able to handle the
work load during the highest labor-demand periods. Most families
hire half their labor requirements, reducing net income from the
12,000 rupees estimated as necessary to support the family to an
average of 8,400 rupees. Labor costs and availability are there-
fore major constraints to production. Given the right markets,
other crops can be more profitable than paddy alone and can les-
sen the heavy seasonal demands for labor.


2.6.3 Settler Selection and Training


One implication of women's active participation in agricul-
ture is that the selection process for settlers chosen from the
electorate would be improved if the qualifications of wife and
husband were considered jointly. Presumably, allotments would be
better managed if both husband and wife were experienced and
interested in agriculture. Women's other skills and entre-
preneurial experience should also be considered. Both male and
female farmers require training in new agricultural techniques,
particularly water management. Agricultural training should
include options for the most intense, integrated, and profitable
household garden cultivation possible under a diversity of local
conditions. Given women's child care responsibilities and







-42-


relative lack of mobility, special efforts may be necessary to
reach married women.


2.6.4 Impact on the Status of Women


Although we have no information on the settlers' status
prior to resettlement or on comparable groups of nonsettler
women, we do know that the land law will reduce women's chances
to legally inherit land in the next generation. This means that
female farm managers and entrepreneurs will not be eligible for
bank loans.

Settlers' situations also vary at different stages of the
settlement process. The first years are particularly stressful.
Women's work loads increase, settlers must cope with malaria and
diarrhea, and nutrition is apt to be poor. Daughters are more
likely than sons to be removed from school when the family moves.
After settlement is complete, the Mahaweli Authority and other
Government agencies try to involve women in the program by re-
cruiting female volunteers and some female staff locally and by
providing training programs. However, employment opportunities
outside agriculture are still very limited.


2.7 Recommendations


1. The Homestead Plot. Mahaweli Authority officials are
currently emphasizing agricultural diversification and the home-
stead garden more than before. Considering the plot's potential
economic importance for the family, increasing its size for new
settlers may be a worthwhile strategy. Settlers frequently com-
plained that plots were too small for dairying and other activ-
ities. Planting fruit- and nut-bearing trees and other crops
adapted to a dry climate could improve family nutrition and pro-
duce subsidiary income.

2. Settler Participation in Decision-Making. Women's reac-
tions tc various issues, such as day care centers, are difficult
to ascertain from the existing literature. By encouraging feed-
back from settlers, particularly women, project officials could
gain useful information and ensure better cooperation. Community
development groups should be headed by settlers (perhaps jointly
by'one man and one woman), not by project officials, and other
types of settler organization should be encouraged. Women's
involvement in settler associations should be encouraged so that
women members would constitute more than a token minority.








-43-


3. Selection. The settler selection process should con-
sider the qualifications and interests o' both wives and
husbands.

4. Agricultural Training. Agricultural training (including
water management) should be aimed at female farmers, married or
single, as well as at male farmers. Hiring and training female
officers and agents might not only facilitate communication with
female farmers (rapport is likely to be greater with members of
the same gender) but would also create models for young women to
emulate and would emphasize that farming is indeed "women's
business."

5. Income-Generation Proiects. More effort should be devo-
ted to designing viable supplementary income projects for women.
Market research should be undertaken before activities are star-
ted on any scale, and potential entrepreneurs should be encour-
aged to develop their own credit and marketing networks rather
than to depend heavily on the project. Local people frequently
mentioned dairying and-'tailoring as activities with local market
potential; one woman entrepreneur wanted to grow herbs and
another wanted to start a hand-loom industry. Sri Lankan busi-
ness people from outside the settlements might be recruited as
consultants.

The Accelerated Mahaweli Program has so far increased na-
tional rice production and provided land to many families that
were formerly landless. However, optimum production has not yet
been achieved, and numerous problems affect settlers' well-being
at this stage. A more thorough initial analysis of the farming
systems being introduced would have addressed issues of crop
diversification and productivity, household income, the role of
men and women in agriculture, potential labor demands on men and
women, and the necessity for training all farmers in new agricul-
tural methods, particularly water management.

Consideration of gender variables at an early stage of the
project might have resulted in improved settler selection, better
agricultural extension training, larger household plots, and
greater attention to gardening and tree crops. Household econ-
omic data and feedback from both male and female farmers might
have assisted project planners in making some of the early proj-
ect design decisions. But the size, complexity, and speed of
project implementation probably precluded such detailed advance
considerations. The fact that most beneficiaries were not yet
living in the project area and the long time between the initial
design and the settlement of people also complicated the planning
process.





Page
44
Missing
From
Original









3. THE NORTHEAST RAINFED AGRICULTURAL DEVELOPMENT PROJECT
IN THAILAND: A WOMEN-IN-DEVELOPMENT ASSESSMENT

by

M. Cristina Blanc-Szanton, Anthropologist
(Columbia University)

Anamaria Viveros-Long, Sociologist
(Agency for International Development)

Nongluck Suphanchainat, Economist
(Khonkaen University)


3.1 Summary


The evaluation of the Northeast Rainfed Agricultural Devel-
opment (NERAD) project demonstrates that in a Southeast Asian
rural setting where male and female roles in agriculture are not
clearly differentiated' gender constraints may play a significant
role in either ensuring or hindering the ultimate success of a
project.

The lack of systematic gender analysis in the NERAD project
stemmed from (1) faulty assumptions among project planners and
managers about intrahousehold patterns and behavior, and (2) the
lack of assessment of the project participant selection process.
These factors translated into project shortcomings in the areas
of labor output, technology transfer, trial success rates, and
overall improvement in the quality of life.

The main objective of NERAD was "to address the needs of the
rural poor in Northeast Thailand by establishing in eight sub-
districts a replicable, area-based agricultural technology devel-
opment program for increasing productivity and farm incomes in
rainfed agricultural zones." The major components of the project
included: (1) the identification and assessment of improved farm-
ing practices such as new subsistence or cash crops, improved
water use, and better animal husbandry; (2) improvements in basic
land and water resources; (3) a more effective extension system;
and (4) the establishment of interactive means of matching Gov-
ernment of Thailand technology development, programs, and resour-
ces with farmers' needs.

The primary implementers and ultimate beneficiaries of the
NEAAD project were to be the "poor Northeastern farm families."
These undifferentiated households were to be given tasks and
would presumably perform them in order to acquire the presumably
equally shared benefits. The intrahousehold structure--its mode




Pr^-3su Pags
~-.b Ed ~cii
f~r ~J i ~0l







-46-


of operation, its constraints, and its internally differentiated
incentive structure--were never adequately analyzed.

Although the NERAD Project Paper did mention that women were
"expected to benefit at least as much as men in most project
activities and [possibly]...more than men in a few of them such
as silk production and animal husbandry," the project did not
specifically target women as project participants or
beneficiaries.

The analysis of gender-related issues in the project design
only briefly acknowledged the flexibility in the division of
labor for agricultural activities in Northeast Thai households,
and the fact that women often controlled the land and usually
managed the household budget. Little significance was given the
fact that women and young children frequently were left on their
own to perform essential agricultural activities because the men
and older children had migrated to urban areas or to other rural
areas for cash income. These factors were never incorporated
into the project design or implementation documents. Further-
more, other background studies, including a socioeconomic survey
of households in 66 villages in the Northeast region, did not
analyze gender issues or consider the potential for farm house-
hold labor constraints.

Throughout the implementation phase the project had con-
siderable impact on women that directly affected project out-
comes. Women actively participated in the various components of
the project by providing at least half the required labor and by
selectively receiving some of the project inputs such as training
in silk production. Ultimately, the project has increased the
amount of labor required from women and children because: (1) the
new crop alternations are generally more labor intensive, and
crop diversification implies that agricultural labor is more
evenly distributed throughout the year; and (2) there is an in-
creasing incidence of seasonal and longer migration of men for
off-farm income. Women and their children have thus had to as-
sume the portion of farming activities usually performed by the
male family members. An increased female and child labor re-
quirement also has implications for the overall quality of life,
since women and older siblings are less available for household
and child care duties.

In addition, women have been excluded from benefits in a
number of domains immediately affecting the progress of the
project:

-- Female farmers were not consulted in the research phase
about farming activities that are their responsibility
and that are often required to increase rural







-47-


productivity (crop planting and processing, harvesting,
poultry raising, and vegetable growing)

-Female-headed or female-managed households, often a
significant number of the middle-income households in a
village, are systematically bypassed as specialist farm-
ers for crop trials or other farming activities

-- Rural women are consistently and systematically excluded
from training in the use of new technologies and other
inputs.

The development of appropriate agricultural technologies may be
jeopardized because the project solicits information only from
male farmers and does not acknowledge women's considerable labor
input to farming tasks, their ownership role, and their manage-
ment of the household budget.

Labor constraints at crucial times during the agricultural
cycle and the absence of essential technological knowledge appear
to have significantly limited cropping trial results. The proj-
ect strategy of crop diversification and improved inputs is very
labor intensive, but because of male migration, these increased
labor demands are being absorbed primarily by Thai female farm-
ers. Even though women are performing key agricultural activ-
ities, training and agricultural extension support continue to be
directed toward male heads of households. Serious crop losses
have occurred because of poor planting practices and misapplica-
tion of pesticides and fertilizers--often under the care of
females left to perform these tasks.

There were negative responses to training provided through
the project that could be directly attributed to lack of gender
consideration. Male farmers would enter a training program but
would drop out at the implementation stage because females in the
family refused to provide the added labor required of them, or
they voiced a preference for another, less labor-intensive
farming activity (fish raising instead of poultry raising). The
training that was targeted to women, silk production, was poorly
planned and implemented.


3.2. Introduction


The Northeast Rainfed Agricultural Development (NERAD) proj-
ect was selected for field study because it offered an oppor-
tunity to examine how gender-related issues affect the achieve-
ment of project objectives in a Southeast Asian rural setting.
There is still a strong assumption that because the roles of men







-48-


and women in rural areas of Southeast Asia are not as clearly
differentiated as in other cultures, development projects there
do not need to pay specific attention to gender-related factors.

The objective of the NERAD project is to increase the eco-
nomic well-being of poor farm families by establishing a repli-
cable agricultural development program in eight subdistricts in
Northeast Thailand. The project's short-range goals are to de-
velop and test appropriate technical packages for the intensifi-
cation and diversification of rainfed agriculture and to streng-
then the agricultural extension service's ability to disseminate
information and technical packages to farmers. The NERAD project
operates both at the institutional and at the farm level, but
this study focuses primarily on project interventions that are
directed to farmers. From a gender perspective, this is an
"integrated project"--that is, its funds have not deliberately
been targeted to men or to women farmers.

To elicit appropriate gender-related information, the fol-
lowing questions were asked about this project:

-- Were gender issues recognized in the assumptions and
design of the NERAD project?

-- Did project components reflect the gender dimension of
labor and other resources in the system of production
and consumption of Northeast Thai farm families?

-- What are the outcomes of the project to date for men and
women?

-- How do gender issues relate to the achievement of proj-
ect purposes and goals?


3.2.1 Project Setting


Aqroecological Characteristics. Because 95 percent of the
Northeast region has not been irrigated, crop production depends
solely on rainfall, which varies widely from year to year.
Erratic weather patterns can lead to flooding and drought within
the same field during the same year. Precipitation patterns also
vary considerably by province. Soils are generally high in acid-
ity, lack major natural nutrients, and are characterized by low
water-holding capacity.

Socioeconomic Characteristics. During the past 20 years,
agriculture in the Northeast region has evolved from almost to-
tally subsistence to a mixed cash-subsistence type of farming.







-49-


Farmers have responded to population pressure and to difficult
conditions by converting most forests into agricultural land.
While there has been a shift to cash cropping in upland fields,
subsistence crops (mostly rice) continue to be the major products
and the major concern in farmers' decision-making. Low crop
yields and low farm incomes have led farmers to minimize cash
inputs (such as fertilizers and pesticides) into agricultural
activities and to look for cash-generation opportunities in off-
farm employment.

The Northeast region is the poorest area in the country,
with the lowest per capital income. The NERAD project is directed
toward approximately 10,000 households distributed in the 101
villages of nine representative tambons (political subdistricts)
within four different provinces. Data collected during the de-
sign phase show considerable variation in poverty levels in the
project area. This variation has important implications for the
development of technical packages suitable for both the wealthier
and the poorer farmers.

The field study team identified major differences in house-
hold composition in the villages studied. The major types of
households are as follows:

-- Extended households
-- Nuclear households
-- Household headed by women
-- Marginal households
-- Households temporarily managed by women

These household patterns are determined by the Northeastern fam-
ily system, which emphasizes matrilocality (residence of daugh-
ters with or near their mothers) and customary inheritance of
land by females, and by outmigration for off-farm income. These
patterns of residence and inheritance, not uncommon elsewhere in
Southeast Asia or Thailand, have remained particularly strong in
the Northeast.

Outmigration from the Northeast region to other countries or
to Thai cities or other rural areas for off-farm employment has
increased greatly since the middle of this century. Nonfarm cash
income was already proportionately higher in the Northeast than
in any of the other regions of the country in 1975-1976.

Agricultural Research and Extension. Agricultural research
has not yet generated appropriate solutions for the difficult
conditions in the Northeast region. The variability of agroeco-
logical and environmental conditions makes the development of
improved technologies and their transfer to farmers very dif-
ficult. Moreover, agricultural research has been carried out in







-50-


experimental stations, and research results have not been inte-
grated with and adapted to the local farming systems in the area.

Agricultural extension services lack well-trained personnel
and local facilities. The subdistrict agricultural extension
agents, the main contact agents between the Department of Agri-
cultural Extension and the farmers, generally do not live in the
tambon, nor do they have local offices, storage space, or con-
venient places for meeting with farmers. Because of the strongly
centralized bureaucratic structure of the Royal Thai Government,
the extension agents tend to follow top-down agricultural poli-
cies formulated in the Ministry of Agriculture and Cooperatives
in Bangkok. They tend also to interact mostly with more success-
ful male farmers, thus limiting the dissemination of improved
agricultural practices to a small sector of the village
households.

Gender Roles in Production and Consumption. Land in the
project area is held in the woman's name and is inherited through
the mother. The division of labor in agriculture is relatively
flexible. Men prepare the land for rice and other field crops.
Women do most of the planting and transplanting. Both men and
women weed crops, although women do more than men. All household
members participate in harvesting. Women do all the time-consum-
ing work of processing kenaf. In addition, women perform most
domestic tasks.

Income from rice and field crops is pooled. Tobacco and
sugarcane are considered to be exclusively men's crops, and vege-
tables and poultry raising are women's activities. Women admin-
ister the household cash. Because rainfed agriculture is not
sufficiently productive to generate enough cash to support a
family, men and women engage in a variety of off-farm and nonfarm
income-earning activities. Among the poor, both men and women
hire out as casual wage laborers during the peak agricultural
season, and men seek construction work during the dry season.
Silk production is traditionally a female dry-season activity,
but the cloth is mainly for home use.

About 10 percent of the households in the project area are
headed by women, because male migration is widespread. Among the
poor or middle-income households, men leave the district during
the slack season in search of wage employment. Among all socio-
economic groups, women work longer hours than men. The poorer
the household, the greater the responsibility women have as farm-
ers and income earners.

Political Participation by Gender. The principal mechanism
for channeling project activities at the grassroots level is the
tambon council, which receives special central Government funds








-51-


and has directed local planning since 1975. Tambon councils are
dominated by prominent local families: the subdistrict headman,
the village headmen, and one other qualified or experienced per-
son from each village in the subdistrict. Tambon councils do not
have women members. But, since the 1982 Administration Act, 120
women have been elected village head and thus participate in
tambon councils as ex officio members. This number represents
only a beginning, however, since there are over 50,000 village
headmen in the country.

Tambon councils interact at the local level with village
development committees, whose female membership is only 3 per-
cent. The female members rarely attend meetings, and when they
do, they remain silent. To remedy the de facto exclusion of
women by the very centralized and hierarchical Thai bureaucracy,
the 1982-1986 Thailand Development Plan initiated a parallel
administrative structure of women's development committees to
supervise the development of women around the country. This
parallel structure has had limited funds an" impact; wumen con-
tinue to be generally excluded from key rural development deci-
sions (Vuthimedhi 1982).


3.3. Project Design and Gender Issues


3.3.1 Project Design


The goal of the Northeast Rainfed Agricultural Development
(NERAD) project in Thailand is to increase the economic well-
being of poor farm families by establishing a replicable agricul-
tural development program in eight subdistricts. Its immediate
objectives are to develop and test appropriate technical packages
for the intensification and diversification of rainfed agricul-
ture and to strengthen the agricultural extension services.
Project components include cropping systems research and trials,
village water resource development, village common-land manage-
ment, management training and extension, economic studies, and
other farming systems modifications. The project is implemented
by the Ministry of Agriculture.

The purpose of intensifying and diversifying rainfed farming
systems is to generate a cash surplus over and above household
consumption need. The main thrust is crop development. The
cropping system research and trials component is helping to de-
velop improved technical packages for five different agroeco-
logical zones.







-52-

The project strategy relies heavily on farmer-managed trials
as a source of feedback about proposed innovations. Participants
are selected by the village headman in consultation with the
subdistrict development council. As an incentive to cooperate,
participants are given free seed, fertilizer, pesticides, and
construction materials. At the grassroots level, project activ-
ities are channeled through the village council.

The project's impact on production depends on how many farm-
ers adopt the improved cropping practices. This, in turn, de-
pends on the extension service's ability to reach the average
farmer. The Ministry of Agriculture intends to strengthen exten-
sion services by doubling the number of outreach workers.

Gender Analysis and Inclusion in the Project Desiqn. The
Social Analysis Appendix in the NERAD Project Paper contains a
one-page statement acknowledging the fact that women own the
land, manage household cash, and play a major role in agricul-
ture, but the text of the Project Paper makes no reference to
these facts. It merely states that "women are expected to bene-
fit at least as much as men in most project activities and may
benefit relatively more than men in a few activities such as silk
production and animal husbandry."

No attention was paid to a case study entitled "Northeast
Rainfed Agricultural Development in Thailand: A Baseline Survey
of Women's Roles in Household Resource Allocation for a Farming
Systems Approach," prepared by Ingrid Palmer for the Population
Council in 1983. Although the study provided important informa-
tion on interhousehold and intrahousehold behavior within the
NERAD project region, its findings were neither incorporated in
the project nor translated into Thai.

The suitability of institutions, delivery systems, technical
packages, and feedback mechanisms for reaching and benefiting
women has not been considered. The implicit assumption is that
communication within the farm household is adequate. Unfortu-
nately, evidence from project households confirms that the oppo-
site is actually the case, a fact that argues for more personal-
ized targeting to make technology transfer more effective.


3.3.2 Project Implementation from a Gender Perspective


I The following subsections will examine major project com-
ponents from a gender perspective. To the extent that the data
permit, we will examine the participation of men and women in
each component, the benefits that resulted from the intervention,







-53-


and implementation problems generated when greater attention is
provided gender issues.


Cropping Systems


The primary aim of the cropping systems component was to
test at least 13 crop modifications that were designed to help
stabilize subsistence rice production and to increase cash crop
production. The cropping component of the project included the
following activities:

-- Selecting villages and villagers to conduct NERAD crop-
ping trials

-- Gathering information on current local farming practices
and problems through rapid assessment techniques

-- Monitoring yearly cropping trials, with accompanying
inputs of equipment, training, and information

-- Expanding dissemination of trial results and other use-
ful farming information

All farmers in the project area actively participate in
project implementation by providing information about problems
and practices. Although rural women contribute at least half of
the farming labor, have major responsibility for designated agri-
cultural tasks, and do all the farm marketing, they are seldom
consulted for information because men feel that they themselves
know more about farming. In fact, we found that women could
articulate their farming problems more clearly than the men in
their household. Despite women's acknowledged contributions to
the agricultural sector, the NERAD.project staff did not consult
them during the implementation phase.

Participating farmers, selected by extension agents, vil-
lage headmen, NERAD field managers, and Department of Agricul-
tural Extension representatives for various cropping trials, also
provide land, labor, and occasional cash for fertilizers and
pesticides. Specialist farmers are targeted by extension agents
and village headmen for selected training and dissemination of
technological information. Participating farmers and specialist
farmers also receive direct inputs including equipment and free
advice.

The selection of participating and specialist farmers was
distorted socioeconomically and by gender. Participants were
usually contact farmers who had been working with the extension







-54-

agents in the principal villages of the tambons and who had been
selected in the past during village meetings called by the head-
man. These contact farmers were frequently friends of the head-
men and usually were the larger, wealthier farmers in the area.
This selection process systematically discriminated against small
landholders with less than 30 rai (4.8 hectares), who were the
majority of farmers in many tambons. It also discriminated
against female-headed or temporarily female-managed households.

NERAD and extension agents are usually more concerned with
producing successful cropping trials than with assembling a re-
presentative selection of farmers for those trials. The selec-
tion process was also affected by the implicit assumption that
the selected "leading" farmers would be the most effective in
disseminating the results of the trials, and by the belief that
women were not technically farm operators, despite the overwhelm-
ing evidence to the contrary.

Women's participation in formal training sessions providing
initial information about crops and equipment was also sharply
limited. Women farmers were very rarely invited and, because
they seldom were selected as specialist farmers, they rarely
participated.

Child care constraints and inconvenient meeting times fur-
ther reduced women's participation. If the training session was
located some distance from home, women were less willing than men
to leave the family for several days at a time. Such constraints
are very real and create special problems for women, even though
extended family support permits women to work in the fields daily
or to seek wage labor.

Despite women's increased participation in plowing and other
phases of farming and marketing, formal technology transfer has
clearly focused on the male heads of household. Participating
farmers continue to receive advice and information directly
through field and home visits by NERAD projects or by extension
staff trained by the project. The exclusively male extension
agents and NERAD field managers easily established personal con-
tacts with the wealthy male farmers whom they visited most often
and who provided them with local information. Only rarely did
extension or project staff meet with women in the fields to dis-
cuss crop technology with them.

Gender constraints are related to the way Government offi-
cials have consistently interacted at the local level and to the
way they are perceived by villagers. In a less politically ori-
ented context, we were told, women would feel quite free to dis-
cuss farming issues. But women's participation in the project's







-55-


cropping systems component remains formally peripheral, sporadic,
and indirect.


Marketing


A key component of local economic success is the marketing
of crops. In mid-1984, a Marketing Working Group was created
within NERAD after a vegetable marketing report suggested the use
of trials, rapid assessment techniques, and farmer-merchant meet-
ings and encouraged group buying and selling, procurement, and
periodic market activities at the local level. Nowhere in the
report was it acknowledged that the growing, processing, and
marketing of vegetables, as well as group procurement and sell-
ing, are predominantly women's activities. Nor was it mentioned
that women play a key role in deciding which vegetables to plant
and when to plant.


Village Water Resources Management


Water resources are a very serious constraint on productiv-
ity in the Northeast, both in terms of managing runoff from heavy
rainfall to keep paddies continuously flooded but not overflow-
ing, and to prevent them from drying out after rice is trans-
planted, as well as increasing water availability for rice and
other seasonal crops. Suggested solutions examined by a Water
Resources Working Group within NERAD included, among other
things, diversion weirs and modified shallow wells.

Women regularly participate in the volunteer labor supply
for maintaining weirs and wells. Gender-disaggregated data were
not kept on water resource management activities, but women de-
rive direct benefits from their participation. Weirs and wells
support rice paddies owned by women and help the survivability of
other crops as well. In addition, women learn how to construct a
weir or well when they volunteer their time.

Women farmers were usually not selected to have a village
well located on their land--a decision made by the village head-
man. Although a well could be used by neighboring farmers, the
greatest benefit was naturally derived by the farmer on whose
land it was located. Poorer or female-headed or -managed house-
holds had little chance of obtaining wells since they were not
likely candidates in the eyes of the village headmen. Male heads
of household usually decided upon the actual location of the
well, although the land often belonged to the wife; and the needs
of cash crops--for example, tobacco with its assured market--







-56-

often overrode considerations of irrigation for vegetables or
water for domestic use.


Village Common-Lands Management


Common lands are important in Northeastern farming systems
for rotating livestock foraging and fuelwood collection. During
the last 30 to 40 years, these common lands have increasingly
been converted to private use and farmed by squatters.

The NERAD Common Lands Working Group has planned seed and
forage improvements and fast-growing eucalyptus intercropped with
cassava and hemata grass for forage. The Working Group used
rapid assessment techniques to study the role of common lands in
village farming in 1984, and by 1985 had started tree seedling
nurseries in many villages to satisfy both the need for common
woodlots and private farmers' transplanting needs. These tree
nurseries were operated by wage workers, often young women,
according to our observations.

The Common Lands project activities have so far provided
only limited inputs. Nevertheless, implementation plans do not
reflect the work women do in these areas nor the need to inves-
tigate further how to include females as well as males in
implementation.


Other Farming Systems Modifications


Fish Production Improvement. Fish is an important element
of the Northeastern rural diet and a source of supplemental in-
come. Activities in both newly constructed and existing fish
ponds included training, stocking, fertilization, and planned
harvesting. Selected villagers offered their labor or ponds and
received in exchange fingerlings, fertilizers and pest control
chemicals, food, and technical assistance. The major goal, an
increase in fish production, was set at harvesting over 200 tons
per year, well above the current 80 tons per year in the project
area.

Since 1982, the NERAD project and the Department of Fisher-
ies have encouraged public and private fish production by stock-
ing existing village ponds with improved fingerlings and provid-
ing technical advice to neliy created village fish committees.
Thirty-eight fish farm-er specialists were trained in 1984; all
were men.







-57-


Although women actively participate in all phases of fish
production, males are usually considered the fishermen. The
project has provided no breakdown of fish production activities
by gender. However, the provision of supplies (fingerlings) and
technical assistance follows the usual village political leader-
ship patterns: they are provided to male farmers who become
specialist farmers and sit on village fish committees.

We learned during our interviews that wives often influence
the selection of the supplementary activities, such as fish pro-
duction. As household financial manager, a wife's greater knowl-
edge of relative inputs, labor commitments, and estimated profits
from each activity greatly influences her husband's decision-
making. In some cases, women favored fish production because it
is less labor intensive than poultry raising, for example.

Poultry and Livestock Production. Native chickens and ducKs
have long contributed to the household diet and provided an
occasional cash supplement for women. Livestock--cattle and
water buffalo for draft power--belong to the men but are cared
for primarily by women and children. The project targeted an
increase from the current 100,000 to 275,000 birds and an in-
crease to more than 55,000 dozen eggs and 500 head of large ani-
mals per year.

The specialist chicken farmers formed village poultry com-
mittees and received free hatched native chickens, medicines,
vaccinations, use of egg hatcheries, and general information and
training. Care and supervision of chickens were provided by both
men and women. Women shared heavily in the daily cleaning and
feeding tasks, while men concentrated on technical procedures,
such as vaccinations. Women occasionally assumed this task if
male members of the household were not available, but they had
not been encouraged to do so.

Women are active participants in chicken farming, but they
are losing their traditional control over that activity as back-
yard production is replaced by larger scale commercial chicken
farming. Men's and women's labor inputs have increased because
of the increasing number of birds, yet women are not included in
the intensive technical training and assistance provided by the
project.

Sericulture. Rural Northeastern women produce silk to weave
cloth for family use and occasionally for sale. With appropriate
production improvements and marketing supports, silk production
could provide significant income, given the current international
price for woven silk cloth. However, the quality and quantity of
silk cloth produced in this region are low, mainly because of
poor environmental conditions and poor management techniques.







-58-


Consequently, silk production is considered a marginal activity
for older women and not an important home-based, cash-generating
endeavor.

In 1984, the project sponsored its first silk training ses-
sions for 72 women specialist farmers, and village training was
offered to interested farmers in 19 villages. In additions, 23
tambon extension agents were trained in improved silk production
practices at the Korat silk station. Although demonstration
rooms have been built and improved mulberry varieties and hybrid
silkworm varieties from the experiment stations have been intro-
duced in several villages, these technological improvements have
not been effectively transferred to women farmers.

Women are the main participants in both silk production and
training activities. However, this component of the project has
started very slowly, and training efforts have not been very
successful because of poor scheduling and inappropriate training
techniques. The training was conducted at an inconvenient time
for women--during the busy rice-growing season--and required a
25-day stay far from home. Training coincided with flooding,
which caused women to worry about their crops, and per diem rates
were not considered sufficient to cover the time away from the
farm. Thus, one-fourth of the women attending left before the
end of the training. Farmers and extension agents agreed that
the training was not appropriate for local conditions.

Thus, the one activity specifically aimed at women farmers
has not yet developed successful inputs and has had very limited
outputs. More testing and better planning are needed if seri-
culture is to become a viable income producer.


Management, Training, and Extension


Participation of Women. The personnel involved in the
management of the project and in its very important extension
activities consist of both NERAD and Ministry of Agriculture and
Cooperatives staff. In addition, many of the NERAD project com-
ponents are implemented and field-monitored by agriculture offi-
cials. All Ministry personnel receive training and staff sup-
port--an important component of the project.

The NERAD managerial staff consists mostly of well-trained
Western and Thai technicians, all of whom art male, except for
two regional government agriculture officials. The clerical
staff, on the other hand, is heavily female. The differential
participation by men and women that is common in most Royal Thai
Government offices is mirrored in the NERAD staff.







-59-


Social Science Training. The Ministry staff for the NERAD
project are professionals in key positions in the Ministry of
Agriculture in Bangkok and in provincial managerial positions,
and agricultural officers and subject-matter specialists at pro-
vincial and district levels. In 1984, approximately one-third of
the field-level extension agents were female. These female ex-
tension agents were assigned to supervise home economics and
youth programs rather than agricultural extension activities.

Would it help the overall extension process if there were
more social scientists in the planning and training departments
of Department of Agricultural Extension? Would it help if more
field-level extension agents were female?

Informal discussions with various agricultural officers and
NERAD project personnel suggest that more social science training
would help field-level extension agents to clarify the mediating
role they are supposed to play between farmers, technology, and
the Government, to better understand household organization, and
thus to become more successful change agents in rural commun-
ities. USAID/Bangkok personnel also suggested informally that
gender distortions and handicaps in the project might be addres-
sed by increasing the proportion of female extension agents. A
general impression emerging from a number of recently completed
rural projects in the Northeast region is that women extension
agents are more effective than men and associate with a wider
range of villagers when they are assigned to agricultural exten-
sion work.

Ingrid Palmer (1983) also received strongly favorable an-
swers from both male and female farmers in the NERAD tambons to
the question: "Would you prefer extension agents to be women?"
Male farmers we interviewed said that women are more willing to
deal with detailed issues of economic management and that working
with a woman avoids the status-conscious political relationships
of the male culture, often accompanied by social drinking, and
places the interaction within a business- or service-oriented
context. Women preferred female extension agents because the
interaction was more constructive and practical. However, the
assumption that female extension agents would interact more fre-
quently with female farmers about issues related to farming and
farming systems was not confirmed by our interviews in the field.
This is an issue that needs further study.

More important than the sex of the extension agent is the
content and context of the message delivered. Future research
must explore these areas to determine more effective ways to
communicate with farmers.







-60-


3.4 Project Outcomes


3.4.1 Project Effectiveness


By March 1985, the implementation of project components,
especially at the farming level, was considerably behind sche-
dule. This fact is in part attributable to institutional dif-
ficulties and the resulting slow delivery of agricultural inputs.
However, observations from this field study and supporting evi-
dence from other studies suggest that the lack of attention and
adaptation of the project to gender issues and the simplistic
assumptions about the household as an undifferentiated unit may
also have contributed significantly to project inefficiencies.

Womens' Participation. Of the more than 3,000 NERAD par-
ticipating farmers selected during the first 4 years of the proj-
ect (1982 to April 1985) for cropping trials, rapid assessment
techniques, fishery, poultry, sericulture, machinery, livestock
and water control, and fertilizers and pesticide production and
adoption, only 7 percent were women; and most of them were i.n-
volved in sericulture. Less than 2 percent participated in for-
mal fishery, poultry, livestock, or crop production activities.
Among the more than 1,000 specialist farmers during the first 4
years of the project in poultry, fertilizer, crops, fisheries,
fruit trees, and sericulture, only 70 were women, and they were
trained only in sericulture.

Women's cooperative participation is required throughout the
farming cycle. Even production credit, supplied by government
and commercial banks, requires the signature and approval of the
women, who generally own the land.

Data Collection Limitations. The development of appropriate
agricultural technologies may have been jeopardized because the
project solicited information only from male farmers during rapid
assessment technique reports and did not acknowledge the fact
that women perform at least half the farming tasks, own a major-
ity of the land, and manage household finances. Thus, women
farmers, with their own perspective of farming tasks and consid-
erable knowledge of farm economics and labor needs, did not con-
tribute to the project decision-making process.

SLabor Constraints Affecting Productivity and Adoption. The
project strategy of crop diversification and improved inputs is
very labor intensive. Off-farm work and male migration have
increased labor demands on Thai female farmers. Labor con-
straints affect the timely care of rice, corn, and other crops
during the peak labor season and lead to lowered productivity








-61-


levels. Seasonal or yearly migration for off-farm cash also
affects the kinds and number of crops planted and their subse-
quent care.

Our interviews disclosed cases in which male farmers had
migrated overseas for work and the wives had taken over the De-
partment of Agriculture crop trials, often with poor results.
Such labor constraints were exemplified by the case of a woman in
Nong Pan village, Roi-Et tambon. Her husband had been away for
18 months, and, even with the help of her four teenage children,
she had had a very difficult time in growing her usual quantity
of rice. She had planted only half a rai of tobacco instead of
the usual two rai, and she had not dug the temporary shallow well
needed to irrigate tobacco and vegetables. She had tried to keep
up with fertilizer and pesticide applications on tobacco, but the
harvest was poor and she received a low market price as a result.

Technology Transfer Constraints Affecting Productivity and
Adoption. Because technology transfer and crop information are
addressed to male farmers while many of the key farming opera-
tions are performed by women, decision-making and action are
delayed, information and technology transfer is not always com-
plete, and cropping trials yield poor results.

A frequently reported reason for poor yields is farmers'
lack of expertise in key farming operations, especially with new
crops. Woman farmers, lacking regular access to extension ser-
vices, are at a particular disadvantage. For example, a female
farmer in the Roi-Et district was planting her root crop stocks
sideways instead of straight up until her husband, who happened
to be bringing her the roots, told her how to plant them cor-
rectly. This single error resulted in the loss of half her crop
that year.

Our observations also indicated that wcmen frequently do not
use new equipment, which is introduced into farming households
through male heads. Because men are frequently absent from the
farm, it is likely the machinery will not be used properly. Women
continue to work with old technologies unless they are included
more directly in the technology transfer process.

The field interviews further demonstrated the negative con-
sequences of aiming technology training to men only. Women who
performed key agricultural activities in the absence of the
trained men often misplanted seeds and misapplied pesticides and
fertilizers. Serious yield losses resulted.

Negative responses to training provided by the project in
other farming systems activities such as poultry, fish raising,
livestock improvement, and fruit trees could also be directly








-62-


attributed to lack of gender considerations. Because of the
particular selection procedures of trainees among specialist
farmers, this training was provided exclusively to male farmers.
Male farmers would enter a training program but then drop out at
the implementation stage because their wives and daughters re-
fused to provide the labor or preferred another, less labor-
intensive, farming system activity (fish raising rather than
poultry raising, for example). As has been mentioned, in silk
production, the only activity expressly targeted to women in the
project, the location, timing, and duration of the training did
not respond to the needs and constraints of the female trainees,
so they often dropped out.

In marketing new and old crops and other outputs from the
Northeastern rainfed farming system, the project experienced
serious problems that it is just beginning to address. Since
some of the pre- and post-rice crops are marketed exclusively by
woren, NERAD's failure to include women in the implementation
planned by the Marketing Workgroup is likely to cause more prob-
lems in the future.

In general, the NERAD project placed increased labor re-
quirements on female farmers, who were already overburdened. The
field study indicated that the fact may have some potentially
negative long-term effects on household nutrition and well-being.

The evaluation of the NERAD project from a gender perspec-
tive demonstrates that in social contexts in which women have
strong decision-making and managerial positions in rural house-
holds, neglecting to encourage their active cooperation in proj-
ect design and implementation may lead to significant project
inefficiencies and delays and may jeopardize the ultimate goals
of the project. In the case of NERAD, the lack of systematic
gender analysis and adaptation seems to originate from (1) faulty
assumptions among project planners and managers about household
internal behavior and (2) the lack of careful assessments of
existing socioeconomic and gender distortions in the farmer spe-
cialist selection procedures currently used by the Department of
Agricultural Extension. These factors translated into project
shortcomings in the areas of labor constraints, technology trans-
fer, trial success rates, and overall improvement in the quality
of life.


3.5 Lessons Learned


1. The activities that men and women perform in local farm-
ing systems should not be overlooked in the development, adop-
tion, and dissemination of new agricultural technologies. This







-63-


consideration is particularly important in a Southeast Asian
context, where women own land and men and women actively parti-
cipate in most stages of agricultural production, processing,
marketing, and division of resources.

Because the development of appropriate farming systems re-
quires thorough knowledge both of pre-existing agricultural prac-
tices, problems, and constraints and of the needs of individual
farmers, related information should be gathered from the men and
women directly engaged in farming.

The adoption of new farming practices depends mainly on (1)
the availability of both male and female labor resources in the
household at specific points in the calendar year and (2) an
awareness of the alternative costs and incentives of the new
practices to both male and female farmers throughout the calendar
year. The dissemination of farming practices depends on the
techniques used to reach everyone involved in different aspects
of the farming system. If the information is targeted only to
male farmers, the adoption of new practices often suffers.

2. Gender analysis is an essential component in understand-
ing how project goals can be achieved, and it must be integrated
into the project design and implementation. Gender analysis and
the adaptation of project design and implementation need to be
performed by Agency for International Development (A.I.D.) staff
and consultants and host country personnel responsible for the
project. An understanding of the political, social, cultural,
ethnic, and religious context at both the country and local
levels is crucial to raising awareness about gender issues in
development projects.

3. In complex projects such as NERAD, gender analysis and
incorporation of women's resources into project design and imple-
mentation should occur in a context that acknowledges the impor-
tance of other factors that affect project success. The promo-
tion of agricultural technologies that increase demand on women's
limited time may have negative consequences for child care, child
nutrition, and overall household well-being. Agricultural proj-
ects that demand inc-eased farm labor should provide households
with alternative options such as appropriate child care
arrangements.


3.6 Recorrm~ndations


1. A consistent effort should be made to incorporate both fe-
male and male farmers into project activities, including gather-
ing information, selecting farmers, and providing agricultural







-64-


inputs, equipment, and training. This goal could best be
achieved in the following ways:

-- By holding interviews with both women and men as part of
the periodic project information gathering process.

-- By using these data to determine the target populations
of each project component.

-- By holding training sessions for both women and men.
Project training activities should be timed and organ-
ized for maximum convenience: meetings should be sched-
uled for evenings rather than mornings- in convenient
village locations, training sessions should be shorter
and should be held in the village, and seasonal labor
peaks should be avoided. The training sessions, to be
most effective, should be arranged for the convenience
of the farmers rather than that of the government
agencies.

-- By including a sample of female-headed or female-managed
households among participating farmers and specialists
to ascertain the range of problems faced by different
kinds of households and to broaden the impact of the
project among poor farm families.

2. The project should help Thai agricultural ministries improve
training of extension agents in communication techniques so that
messages are delivered to all farmers engaged in project
activities.

3. The possibility should be explored of encouraging women's
cooperative child care at the village level. This could relieve
labor pressures during peak seasons as well as facilitate women's
participation in training sessions.









4. ENTREPRENEURSHIP, CREDIT, AND GENDER
IN THE INFORMAL SECTOR OF THE DOMINICAN REPUBLIC:
THE ADEMI STORY

by

Rae Lesser Blumberg
University of California, San Diego


4.1 Summary


One of the major purposes of this research was to explore
the women-in-development (WID) implications of a mainstream de-
velopment project that had not been designed specifically to
"take women into account," but did, in fact, include them. The
project chosen was the Dominican Republic's ADEMI (Association
for the Development of Microenterprise, Inc.), which gives short-
term working capital loans at market rates of interest to infor-
mal sector microentrepreneurs.

In this successful, innovative project, the involvement of
women was inadvertent. ADEMI eliminated almost all the lengthy
delays, bureaucratic red tape and collateral requirements that
generally exclude poor people from qualifying for most Third
World credit projects. Since ADEMI, a private sector organiza-
tion, aims its small loans at poor and very poor microentrepre-
neurs of the urban informal sector, and this group includes a
high proportion of women, the removal of the barriers meant that
significant numbers of women were able to become credit
recipients.

As the data of this study show, women clients proved better
than their male counterparts at job creation--a primary goal of
the project. This was particularly true of women in the cloth-
ing/textile sector, where nearly half the numerous loan clients
were female. Moreover, women's payback record proved to be at
least as good as that of men.

Overall, not only has ADEMI been successful in promoting
microentrepreneurship and incorporating women, its program also
has helped to create a sizable number of new jobs at low cost
while stabilizing even larger numbers of existing jobs in a coun-
try where unemployment and underemployment are on the rise.




The author gratefully acknowledges the editing assistance of
Kathleen Moran of A.I.D. and Suzanne L. Willis of the University
of California, San Diego.







-66-


ADEMI was founded in 1983 with two program components: One
provided loans to individual microentrepreneurs (17 percent of
whom were women) who had from roughly one to six employees. The
second made financing available to the very lowest level of the
microentrepreneurial hierarchy--street sellers, recyclers, home-
based mini-workshop ventures, and so forth. These loans were
provided through Solidarity Groups, composed of four to eight
people who knew and trusted each other and who would band to-
gether to guarantee each other's loans. Fully 43 percent of
Solidarity Group members were women.

Unfortunately, the Solidarity Group component was suspended
in late 1984 because of growing top management concerns over a
worsening economic situation, fears of loan abandonment, a low
job-creation rate, and other subjective reasons. In effect, this
single policy change eliminated about 77 percent of all women
beneficiaries. Yet the findings of this study indicate that the
fears about the Solidarity Groups were not borne out by empirical
data. Therefore, to resurrect the Solidarity Group component
would be the single most important WID action ADEMI could take.

Still, the data in this study show that the ADEMI project
provides a model of great potential. In less than 2 years of
operation, it began approaching administrative self-sufficiency,
and it clearly had developed a methodology of getting credit at
market rates of interest to poor microentrepreneurs who never
before had access to credit below usurious moneylender rates.

ADEMI represents an usually successful story in a new type
of development project (private sector organizations offering
microenterprise credit for the informal sector) that is both
growing rapidly and enjoying a high level of success. The notion
of private sector organizations may eventually compete to give
credit to large numbers of poor male and female informal sector
entrepreneurs may be stretching reality too far. But, even on a
greatly reduced scale, if such organizations as the one examined
in this study prove replicable and capable of achieving self-
sufficient growth within several years, enormous progress could
be made toward two achievable development goals: incorporating
women into mainstream development efforts, and promoting "growth
with equity" for a substantial stratum of the Third World poor.

Lessons learned from this examination include:

-The ADEMI concept could serve as the model on which to
build the "McDonald's of Third World credit projects."
It is fast, profitable, efficient, low in cost, and
capable of reaching large numbers of people.







-67-


-- ADEMI has adopted procedures that eliminate the barriers
that prevent poor people and women from benefitting from
most credit programs: it has eliminated multiple visits
for loan applications, abolished collateral requirements
for most loans, substituted "streetwise" college student
"promoters" (asesores) for traditional bureaucrats in
selecting loan clients, and guaranteed subsequent loans
if previous ones are paid on time.

-The economic benefits derived from making Icans avail-
able to women, who as a group have been responsible and
successful, can bring considerable social benefits in
their wake, ranging from women clients' increased self-
confidence to increased spending on their children's
education and welfare.


4.2 Introduction


The overall purpose of my 1985 field research on women in
development in the Dominican Republic and Guatemala, when I in-
vestigated two "mainstream" development assistance projects of
the United States Agency for International Development (A.I.D.),
was to analyze the extent to which they were incorporating women
as well as men--and with what effect on both sexes and on the
project in question. I used a Rapid Rural Appraisal methodology
to explore the WID implications of two projects that involved
substantial sums of money and had not been designed specifically
to "take women into account."

The ADEMI project proved to be of particular interest for
both WID and "mainstream" development reasons. Microenterprise
credit projects aimed at the burgeoning informal sector of the
Third World are coming increasingly into vogue as a development
tool. As the recent, growing literature on such projects shows,
such projects tend to be very successful in providing direct
benefits to quite poor people, who use the small loans in ways
that create appreciable numbers of jobs for people even poorer
than themselves (Boomgard 1989, Blumberg 1989). Moreover, as
Boomgard's (1989) synthesis report, based on a sample of 32
microenterprise credit projects, concludes, such projects "are
well suited to the goal of integrating women into the development
process. The proportion of women beneficiaries is high in many of
the programs reviewed" (p.xv).









4.3 Project Design and Related Gender Issues


4.3.1 A Successful Start


The Association for the Development of Microenterprises,
Inc. (ADEMI) proved to be a strong example of both of Boomgard's
generalizations. It is an extremely successful, innovative,
well-run and soon-to-be-profitable credit project that gives
small loans to poor and very poor urban entrepreneurs. And,
without deliberately attempting to target women, its methodology
is such that it lowered the barriers that usually exclude both
women and the poor from credit programs.

Indeed, the fundamental fact about ADEMI is its remarkable
success. At the time of the 1985 field research, ADEMI was 2
years old and had grown spectacularly in that time. It was clear
that this innovative private sector organization had found a way
to dispense with most of the red tape and constraints that almost
everywhere limit lending to the small microentrepreneurs of the
informal sector. At ADEMI, procedures are simple and fast:
short-term loans for working capital are processed and ready only
a few days after application has been made. Even more impressive
than the streamlined organizational procedures is the association
staff. From the executive director to the messenger/photocop-
yist, they are typically dedicated, hard-working, competent,
efficient, honest, and empathetic young people. More adjectives
would strain credulity, but from the top to bottom, I found ADEMI
people working with enthusiasm and elan.

Background research in preparation for my field work had
prepared me to expect a project with an exceptional track record.
If anything, the figures I found proved even better: after only
2 years, ADEMI is on the verge of making a profit, creating a
self-sustaining and growing mechanism for generating both income
and jobs for the urban poor. In April 1985, Mirtha Olivares,
ADEMI's executive director, summarized the organization's
activity and impact as follows:

-- It creates one job for every US$818 it lends (paying
US$67, RD$209, per month per employee).

-- It creates an average additional monthly profit of per
microentrepreneur.


-- It rotates its portfolio 3.5 times a year.








-69-


-Therefore, with a US$100,000 loan portfolio ADEMI can
lend US$350,000 and can generate US$1,184,000 in addi-
tional income per year, as follows:
US$343,000 for new workers
US$841,000 in additional microentrepreneurs' income

-- Thus, each dollar of the original US$100,000 generates
almost US$12 of new income a year.

ADEMI'S program also has been successful in creating new
jobs. The average microentrepreneur loan client has added nearly
1.5 new workers. The Dominican Republic is suffering a serious
crisis of unemployment and underemployment, which is especially
worrisome in the urban sector. Thus, the fact that loans to
microentrepreneurs have created sizable numbers of new jobs at
low cost while stabilizing even larger numbers of existing jobs
will prove to be quite significant to the discussion below.

ADEMI is an indigenous private voluntary organization crea-
ted in February 1983 by a group of influential Dominican business
leaders. Before ADEMI existed, informal sector microentrepre-
neurs could get credit only at usurious rates and by pledging
capital assets. In contrast, ADEMI provides its short-term work-
ing capital loans at market rates of interest. A.I.D. has been
supportive during ADEMI's formative period, providing DR$150,000
from local currencies into the credit account and DR$500,000 over
2 years for overhead expenses until ADEMI reaches organizational
self-sufficiency.

One additional background detail is needed. ADEMI was crea-
ted with two program components. The first involved loans to
individual microentrepreneurs who had some means of production
and from roughly one to six employees. The second was even more
innovative and provided loans to the very lowest level of the
microentrepreneurial hierarchy: street sellers, recyclers, home-
based mini-workshop ventures, and so forth. The vehicle for the
second type of loan -as the Solidarity Group, in which some four
to eight people who knew and trusted each other would guarantee
each other's loans. They would be loaned a sum to be divided
among the group members. Common to both components was the idea
of the gradually increasing loan size. With each successful, on-
time payment, borrowers were eligible to receive a larger sum, up
to the then-existing ceilings on loans tc individual microentre-
preneurs and to Solidarity Groups.

In reviewing project documents in preparation for my field
work, I learned that the Solidarity Group component had been
suspended. Why this occurred will be discussed below. Nonethe-
less, ADEMI statistics continue to list Solidarity Group data.
(During 1983, ADEMI's first year of operation, 214 of the 215








-70-


Solidarity Groups were formed; the last was formed in February
1984, although the program was not formally suspended until
September-October, 1984. But by the spring of 1984, loans to
Solidarity Groups began to be delayed, frozen, or cut--even to
groups that had never been late with a single payment.) Table 1
shows important data, compiled by Mirtha Olivares, about ADEMI's
operations.


Table 1.


ADEMI Activity to March 31, 1985


Individual Solidarity
Micro- Groups Totals
Measure enterprises (n=215)


People Benefited

Business financed 847 1,150 1,197

Persons benefited
Directly 5,521 1,150 6,671

Indirectlya 27,605 5,750 33,355

New jobs created 1,150 1,150

Loan Activity

Loans granted 4,408 1,805 6,213

Total amount
loaned (US$) 1,544,882. 392,375 1,937,257

Average amount/
loan (US$) 350.47 237.38 311.81

Loan portfolio
(current) RD $606,563b
US $189,551

OAssuming 5 people per household of each direct beneficiary
bIn'this table, US$1.00 = RD$3.20; elsewhere, US$1.00= RD$3.125








-71-


4.3.2 Related Gender Issues: The Policy Change That Eliminated
Most Female Clients

Two statistics about ADEMI must be highlighted. First, as
Table 1 shows, no new jobs were attributed to the Solidarity
Groups, versus 1,150 for the individual microentrepreneurs.
Second, in a study entitled, "Women's Participation in ADEMI,"
Rebecca Reichmann (1984) found that as of February 1984, only
about 14 percent of the individual microentrepreneurs were fe-
male, whereas fully 43 percent of the 1,150 members of the Soli-
darity Groups were women--almost 500 individuals. The figures I
found (May 1985) show that 17 percent of the microenterprises
receiving individual loans are run by women--150 out of a total
of 874 ventures. In short, by suspending the Solidarity Gr-up
component, ADEMI had eliminated about 77 percent of all women
beneficiaries.

Why this dramatic policy shift? Project documents, Reich-
mann's study (conducted for Accion International/AITEC, which
originated the innovative ADEMI methodology and continues to
provide technical assistance and evaluation for ADEMI), and my
own interviews in Boston with Accion International/AITEC offi-
cials (Steve Gross and Jeffrey Ashe) provided no more than inter-
esting clues. None of the ADEMI documents mentioned the dispro-
portionate effect on women of suspending the Solidarity Group
component. What were mentioned (very briefly) were vague and
unsubstantiated rationalizations for the suspension: the economy
was deteriorating and those on the bottom would be most hurt,
making future loans risky; some recipients may have encouraged
the migration of "country cousins" to join their Solidarity
Groups and share the ADEMI bonanza; other recipients might aban-
don the city to return to the country if the economic situation
grew more precarious; Solidarity Groups did not create jobs;
delinquencies seemed to be increasing.

Prior to going to the field, I was able to find data about
only one of the vaguely stated justifications for freezing the
Solidarity Group component--delinquencies. During 1983, delin-
quencies were only 1 percent among Solidarity Group members, and
they rose among both microentrepreneurs and Solidarity Groups in
early 1984 as economic conditions deteriorated. Furthermore,
preliminary analyses from the only study done to date that com-
pared the performance of different subgroups of loan recipients
found no tendency for the smallest loans to be more often delin-
quent than larger ones. Clearly a field follow-up of the Soli-
darity Group puzzle was needed.








-72-


4.3.3 Methodology: Rapid Rural Appraisal Technicues


Follow-up Interviews. Following up on Rebecca Reichmann's
early 1984 investigations of women's participation in ADEMI, I
interviewed 16 microentrepreneurs (10 women and 6 men) and 20
members of Solidarity Groups (10 women and 10 men) representing
11 different groups. The microentrepreneurs were interviewed
individually, in their places of business or in their homes. The
Solidarity Group members were interviewed both in groups and in a
marathon group session including 10 men and 2 women, that went on
for hours.

I developed a questionnaire that incorporated much of Reich-
mann's 1984 instrument. She had stressed the factors, such as
transaction time and costs, collateral requirements, and so
forth, that frequently restrict women's access to credit--factors
that ADEMI has almost wholly overcome as part of its normal ope-
rating procedures, which apply to both men and women. The ques-
tionnaire also included topics relating to my overarching WID
research agenda (such as the internal economies of clients'
households and changes in workload, relative decision-making
power in the family, self-confidence, and economic plans). After
completing the questionnaire items, I asked open-ended questions
about institutional issues (ADEMI's policies, delivery systems,
and practices and how they have changed). With the Solidarity
Group members, I also asked about their group's payment record
and history and explored their perception of ADEMI's policy
change and the impact on their lives.

Records Comparison. A review of ADEMI records enabled me to
check the information about the various people and Solidarity
Groups interviewed. Loan amounts compared exactly, but inter-
viewees were often vague as to whether they were on their eighth
or eleventh ADEMI loan. They also tended to mention more part-
time or family workers than their ADEMI records showed.

Informant Interviews. I also interviewed key informants in
ADEMI: the dynamic president, Lic. Camilo Lluberes, the three
Peace Corps Volunteers who provided ADEMI with accounting and
computer expertise and design advice for clients, and five of the
asesores, who acted as counselors, promoters, and credit investi-
gators and who formed the foundation of ADEMI's successful
strategy.

Data Gathering. I took advantage of the new computer system
to request a number of sex-disaggregated statistics. As these
statistics were generated programs were written that will permit
such information to be routinely disaggregated by gender in the
future.








-73-


Data Analysis. To explore the rationale for suspending the
Solidarity Groups, I coded, tabulated, and analyzed the available
data on the 46 Solidarity Groups that had never been a day late
on a payment (21.4 percent of the 215)--even in the months when
ADEMI was winding down the Solidarity Group component and loans
were frequently delayed or cut.

In all of the steps described above, I attempted to use
"triangulation" techniques, in which the multiple approaches
provide a broader and deeper picture of what is being investi-
gated, despite the rapidity of the process. Triangulation, in-
volving getting more than one source of data for a strictly deli-
mited number of variables and topics, is the heart of Rapid Rural
Appraisal. With it, one can generate data that are reliable and
valid enough to support the real-world decisions affecting
people's lives that are often based on RRA research: projects
are developed, continued, or discontinued; groups of benefi-
ciaries are included or excluded.

In the case of ADEMI, since the microentrepreneur component
is continuing and thriving, a.-d since the female microentrepre-
neurs, although relatively few, were shown by every method of
used to be doing at least as well as the men, let us begin the
analysis with a gender-disaggregated examination of the data for
individual microentrepreneurs.


4.4 Project Analysis


4.4.1 Male and Female Microgentrepreneurs: Present Performance
and Future Prospects


The data from the computer runs show a picture of remarkable
success among male and female microentrepreneurs. Interviews
with 16 of the clients revealed people who, for the most part,
proved remarkable in person, while appearing mediocre on paper.
What would a bank have said about even the two most highly
educated?

-A 38-year-old man with a university degree in social
communication and with artistic inclinations. Unable to
find work in his field, he needed a friend's advice to
produce tourist handicrafts. For 8 years he has opera-
ted a modest workshop producing arber handicrafts and
jewelry.








-74-


-- The 32-year-old wife of a top executive (he administers
200 employees), trained--but not happy--as a Certified
Public Accountant. Some five years ago, she began to
move into the dressmaking business on a small and often
fluctuating scale.

What, indeed, would a bank have said about some of the less
well educated, and thus more typical, microentrepreneurs?

-- A 55-year-old women household head with six dependent
children and a fourth-grade education who, although
trained as a dressmaker, finds she earns more running a
hardware store. Six years ago, she opened a small-scale
backyard factory to make cement blocks--a highly non-
traditional business for a female.

-- A 52-year-old bachelor who has fathered 14 children, has
a sixth-grade education, and who worked in shoe repair
in New York for many years. He brought back some anti-
quated shoe repair equipment that he has installed in a
shop on the main street of one of the poorest and most
remote shacktowns of Santo Domingo.

These four cases are not the most successful. One couple
went from 5 to 50 employees in an amber and handicrafts business,
fueled by 12 ADEMI loans over an 18-month period. Another man, a
bakery owner, expanded from 8 employees to 20, spurred by 8 ADEMI
loans over a 23-month period. Nevertheless, the stories of the
people described above:

-The man with the social communication degree saw his
amber handicrafts and jewelry business grow from
2 employees to 8, and from DR$1,380 to DR$4,800 in
sales--using 9 ADEMI loans over 19 months. Now he plans
to expand and upgrade "as far as I can."

-The women exaccountant is already on her fourth ADEMI
loan in 7 months. Her loans have increased from DR$400
to DR$1,300. She has added one employee and stabilized
and increased the working hours of her other four
employees. She feels her business is now organized;
soon she hopes to make it grow by adding men's shirts
and retail sales to her women's wear line. She feels
more self-confident and secure, and thinks that because
she now feels better organized, "I have more control" in
economic decisions in her home.

-- In 18 months with ADEMI, the woman with the backyard
cement block factory has had seven loans rising from
DR$200 to DR$2,000. She has added two employees and, as








-75-


is common, especially among female microentrepreneurs,
most of her children also help. Her volume has in-
creased from one to three truckloads of sand delivered
per week. "Sometimes I never even close; sometimes I'm
up wetting the blocks at 6 a.m., but it doesn't bother
me." She is proud and pleased with her business's
growth and she now wants to buy more land to expand it.
Any small savings go for the children "so that they can
study" and for the house. And her modest house shows
it: it contains a set of encyclopedias, a new refrige-
rator, and the children's stereo.

-Finally, the man with the shoe repair business has al-
most more business than he can handle. Sales have in-
creased from DR$1,400 to DR$19,000/month during his
23 months with ADEMI. He has had nine loans, rising
from DR$300 to DR$5,000. He has gone from two to nine -
employees, and, during busy hours, he has customers
waiting in line. He even uses his machines to do addi-
tional work for other shoe repairers who don't have such
equipment. He says his income has gone up 500 percent,
but he reinvests it, anticipating a business with
20 employees. He also is more active in community
charity work "for sick children--one has to help the
others." He is more confident, tranquil, and secure, he
tells me. But during the whole interview, during the
afternoon rush hour, the jammed shop is closer to pande-
monium than tranquility.

Some Gender Statistics. From this small glimpse of the
people involved, let us move to some selected aggregate statis-
tics. Even though ADEMI's microentrepreneurs are engaged in a
wide array of businesses, over 80 percent of the 150 female
clients are found in only three economic sectors: food, cloth-
ing, and ceramics. The single largest economic category is
clothing: 150 clients engage in mostly small-scale tailoring,
dressmaking, and related production/sales activities, and almost
half are women.

Table 2 compares the statistics for men and women micro-
entrepreneurs in general, as well as those in the clothing sec-
tor. The results for the clothing sector are particularly strik-
ing: women's businesses grew faster than men's in five of the
six categories. In the sixth category, fixed assets, women's
businesses initially had more than the men's, which are now
catching up.

Even more significant, given the increasing concern in ADEMI
policy for job creation, the women microentrepreneurs have crea-
ted more jobs than have men. The comparison is especially drama-










Sr- 0 Nn~


N 01


to )


* -


c0, t- M' ('4
In C4 V-f CM4

04 co 0%
<0 CM


-0)


I

0:


0


14
a u
0



0






g o
-4












3e
0*












0

-I
0c








0.1
013





0
C



-r.
2:


40 u) 4 wo C
4= (r) m? r. w. O
I- W) CM4 01 q (4


en N





00


'U > m
*> -^ wto &to
0 (o0 D 14 V V
0 02 ) 0B 0
0 b 2 <0 b-
aca
0 W 0i H2 0 0 V
2 a (i CO
0 )0 > a
14 04 c c- C
U4 a A4 to X to)
1l 0 14 0 0 0e oi-
(Iv) 0V Ci v 0


0

2-
I
o



5


IN
II-
I I
IC





























4)u
Ic,












0o2
140

*)0
S00


e-
.-I
'4:


SCU)O
1-i c U)


0 N
t N

0 0


r-


Vs. m r- 10
fol CE> C- Ifp
0 to U)
C- C


CV ) v
N U
-41


01 C')
C, C
(I) (1
-4


*q. D i
U)
w q-4 -


Ul) U)


CV) %D
CO <
CO tO, .-


< r1 -
Un ('2


* a-4
in CM ,


CW) 0 0n
4D~


412 w 4
( = 0n
w 0



@2(1 ( 0 01 0
0 o> 0 0 -4 *C -

030
x 0 > 0 0 0
O) >w (/2,4:~
X t o > g 0 a
w a4 B Q3 a o
(14 Cl 'a M T 2


0


o w
00
OO
QMc


C

-4
Q





a
O
-4



41
0
E4



4C0
00







cl

41
}


C
01





L' Un .4 N 01


c %0 -4 O 4
Ut) C$ -W 30
% r V4


14
0
41
co
0
0





o 0
00
(0 -4








-76-


tic in the clothing sector, where men's businesses have increased
by .64 employees, while women's businesses have added an average
of 1.4 jobs each. For the total 874 businesses, women's enter-
prises added 1.5 jobs each versus 1.3 for the men's enterprises.

In short, despite the fact that the economy has suffered
greatly since early 1984, both men and women microentrepreneurs
have used their ADEMI loans to make their businesses grow.
That's a success story.

The fact that women microentrepreneurs perform so well in
their businesses is new. Nevertheless, those working for ADEMI
already had subjective impressions that women's payback record is
somewhat better than men's, and women tend to be more responsible
than men in their business ventures. This attitude came out in
each of the interviews with the asesores and their supervisor,
who had daily contact with both male and female clients, and it
was echoed in the comments of ADEMI's top management. Data to
test these impressions are available, at least in part. Male
versus female performance would have to be compared in smaller
versus larger businesses, and with smaller versus larger loans.

Small Versus Large: The Risk Factor. In an analysis funded
by the Inter-American Development Bank, Accion Interna-
tional/AITEC found that those with small businesses/small loans
were no more likely to be delinquent than their larger scale
counterparts. These findings had not yet been presented to ADEMI
in final form, however, and some, from the president to several
of the asesores, mentioned their feelings that, as economic con-
ditions have deteriorated, smaller microentrepreneurs may repre-
sent a worsening risk because of the greater precariousness of
their businesses.

Moreover, some ADEMI people seem to have acted (consciously
or not) on their impressions. In comparison with the heady days
of 1983, there seems to be a tendency for the loans to "trickle
up" a bit toward somewhat larger and more established micro-
entrepreneurs. A comparison of the distribution of ADEMI loans in
the last quarter of 1984 with the distribution of microentrepre-
neurs surveyed in a large-scale 1980 study of the microentrepre-
neurial/informal sector shows something of an upward skew. The
average microentrepreneur in the 1980 study had 1.6 employees;
ADEMI's recent loan recipients averaged 3.8. Table 3 gives fur-
ther details.








77-

Table 3. Distribution of hMcroentrepreneurs and of ADEMI Loans



1980 Study: Microentrepreneurs Clients of ADEMI as of July,
with Fixed Location, 1984, per Initial Loan
Santo Domingo Application Forms


No. of No. of
employees Percent ezplcyees Percent


0 33.8 0 2.2
1-2 47.9 1-2 26.6
3-4 12.0 3-4 35.0
5-6 2.1 5 11.3
7 and up 4.2 6-10 19.0
11 and up 5.9


Source: Study for the Dominican
Development Foundation by Stephen
H. Gross; personal communication
from Jeffrey Ashe


Source: Personal communi-
cation from J. Ashe; based
on analysis requested
from ADEMI


In other words, ADEMI has given 36.2 percent of its loans to
microentrepreneurs with five or more employees (6.3 percent of
the 1980 study's universe), and 71.2 percent to those with three
or more employees (18.3 percent of the 1980 study's universe).
Only 28.8 percent of ADEMI loans went to those with two employees
or fewer (81.7 percent of the 1980 universe).

Implications for Gender Consideration. Why is this upward
skew of ADEMI loans important? Quite simple: it is commonly
agreed that the higher up one looks in the microentrepreneurial
hierarchy, the fewer females one finds. To date, only 17 percent
of the microentrepreneurs to whom ADEMI has given loans are wo-
men, and the proportion of female beneficiaries would almost
surely drop sizably if the loans were to "trickle up" still fur-
ther to the larger, better established businesses of the microen-
trepreneurial sector. Moreover, because of strong loan demand
ADEMI does little promotion, an activity that might attract new
female clients if women-linked communications channels were used.
Furthermore, with the imminent plans to buy small motorcycles for
the asesores, there has been serious talk of not hiring any more
women for the job--and anecdotal evidence indicates that women
asesores have more women clients. In sum, it appears that future
prospects are for fewer rather than more women microentrepreneurs








-78-


as clients of ADEMI, despite their solid performance and general
reputation as nore reliable borrowers.


4.4.5 Male and Female Solidarity Group Members


It was not until the last few days of my stay in Santo
Domingo that I was able to make contact with the former head of
the Solidarity Group component and begin meeting with the Soli-
darity Group members. Meeting the people and hearing their
stories brought home the human impact of the policy decision to
suspend the Solidarity Group component. And it provided addi-
tional support for my tentative conclusion that, in its keyday,
the Solidarity Group program had been a success in economic,
social, and gender terms. In other words, ADEMI had had two
winning components.

What is clear is that, faced with a resource crunch (delayed
receipt of Inter-American Development Bank funds) and an eroding
economy, ADEMI top management in early 1984 began opting for the
job-creating, less poverty-stricken microentrepreneurs over the
survival-sustaining Solidarity Groups. What is not clear is that
any empirical evidence exists to buttress management's admittedly
subjective impression that the performance of the Solidarity
Groups was questionable and their prospects very risky. To the
contrary, I found indications of positive, rather than negative,
performance. Only a full-scale study can fully resolve the ques-
tion of whether any misconduct on the part of Solidarity Groups
contributed to their demise.

Analysis of the Solidarity Groups. Table 4 shows data on
the "super performers"--those out of the total 215 Solidarity
Groups that had never been even one day late in repaying their
loans.

Three things are clear from Table 4. there are no gender
differences in the composition of the "super performer" 46 Soli-
darity Groups versus the full 215: all-fem='e groups, all- male
groups, and mixed groups are represented in .e same proportions
among the 46 as in the 215. Second, the "super performer" groups
averaged from 7.4 to 9 loans each--thus, their performance held
up over time. Third, residence data do not support the "country
cousin" allegation contained in ADEMI's 1984 third-quarter report
that Solidarity Group members might be bringing their rural rela-
tives to join in their groups and share ADEMI's largesse.








-79-


Table 4. Data on 46 Top Solidarity Groups



46 Super Performers

215 Average
(Total) Mean No. Group % Members w/
Percent Percent Loans Size <2 Yrs. in City
All-female
groups 4 3 9.0 5.0 0%

All-male
groups 33 33 7.7 .5.6 9.5

Mixed groups: 64 63 7.4 5.7 3.0
>50% female 25 24
Female coord. 28 26



In addition, in data not shown, my analysis of the members
of the "super performer" groups found no apparent relationship
between years of education and income. Worse yet, there was a
clear trend for younger Solidarity Group members to have more
education, with no consistent effect on their fortunes. It
appears that unemployment and underemployment and the state of
the economy are now such that many young people with well over
primary school (sixth-grade) educations are unable to be incor-
porated into the mainstream economy; and in the informal sector,
as evidenced by the top 46 groups, the do not seem to be doing
any better than their less educated counterparts.

All this was brought home during the first day I interviewed
the Solidarity Group members. My last interview of the day was
with three women from two of the all-female groups. The house
was a humble wooden shack but it had a refrigerator and TV. Two
middle-aged women (48 and 45, with 3 years and no education,
respectively) represented a group composed of five female heads
of household street peddlers. The third woman was the 26-year-
old daughter (with two children and an eighth grade education) of
the 48-year-old in whose house the interview took place. Her
group of seven young women with children included four dress-
makers and three street sellers of cosmetics and cloth. The 45-
year-old said that before ADEMI she had been "enslaved by the
moneylender." Her group had had 12 ADEMI loans, but as 1984
progressed, they had had some delays and cuts in their loan
levels. This depleted their working capital to the point that
they sometimes had to go to the moneylenders when an ADEMI pay-
ment was due. After concluding with a discussion of their hard-








-80-


won economic independence ("Some women who earn money accept the
man's authority, but we don't."), I was asked by one of the group
if I could drop her off near her night school. She had changed
to stylish clothes and her make-up was stunning. I didn't recog-
nize her. She was studying "to improve her life." I didn't
mention my data analysis showing no relation between education
and income among Solidarity Group members.

Earlier that afternoon, I had been involved in a meeting
that took place in a barrio fronting the Ozama River that fought
a yearly battle with flooding and impassable dirt roads. The 10
men and 2 women included representatives of two of the "super 46"
Solidarity Groups. We aired the problems with ADEMI's lengthen-
ing delays in 1984 loans. We discussed the economic crisis; the
people's frequent response was to change business in order to
survive. For example, one man had sold plastic on the street.
When that became unprofitable, he sold vegetables on the other
side of the river. Now he is working as a driver. He had been
the coordinator of one of the two "46" groups represented, "but
you have to be flexible."

I also learned the story of the coordinator of the other
"46" group represented at the meeting. In order to maintain his
group's perfect on-time record in what turned out to be their
last loan, he pawned his TV. The group never got another loan,
and he lost his TV.

Nevertheless, most of the Solidarity Group members inter-
viewed said they would be willing to work with ADEMI again if the
Solidarity Group program were resurrected. Even though they
agreed that the loans never reached a sufficient level to "get
them over the threshold" and create a growing and sustainable
business, most had experienced some economic improvement. And
even though the Solidarity Group members I met are much poorer
than the microentrepreneurs, they spoke with the same entrepre-
neurial spirit. They explained to me in detail just how much of
a working capital loan a person in each of their varied economic
activities would need to get a viable business off the ground.
The amount varied from DR$200 to DR$500 (for an upholsterer),
probably double what was needed in 1983. ADEMI's Solidarity
Group loans had never gotten up to the needed breakthrough level.

In none of the meetings with Solidarity Group members did I
hear of evidence to corroborate another of the subjective impres-
sions advanced by management (in ADEMI documents) as a rationale
for suspending the Solidarity Group component: that Solidarity
Group members might well flee the city. To the contrary, most of
these people were long-time residents of the city. Most were old
enough that they had growing children for whom the city meant








-81-


schools and a possible better life. ("How could we abandon our
children's future, to go back to the country?")

In fact, the only ADi)MI contention that the Solidarity Group
members supported was that the groups should be smaller rather
than larger. Groups of six to eight were seen by some as too
large. Smaller group sizes of three to five were mentioned when
we talked about Lheir attitudes toward participating in a resus-
citated Solidarity Group program.

A Future for Solidarity Groups? By late summer of 1985
(after I returned for a second field visit), the president of
ADEMI had become willing to restart an experimental Solidarity
Group effort, if it were accompanied by a baseline study and
ongoing monitoring and evaluation. But he felt that because, in
his opinion, women Solidarity Group members proved more respon-
sible than some of the men in paying back loans, such a program
should be limited exclusively to women. In fact, he would prefer
that they be female heads of household who have the urgent neces-
sity for economic activity and display responsible behavior.

In that direction, however, lies the possibility of a small
"charity" program rather than a large, growing, and self-sustain-
ing economic program. What is so unique about the ADEMI Soli-
darity Group program is precisely its economic viability. Lend-
ing costs are low (presumably lower than in the individual micro-
entrepreneur component, although comparative data are not
available), and large numbers of people of both genders who al-
ready are experienced in microentrepreneurial activities would be
eager to join. If, indeed, women's performance is at least as
good as the men's, then, given their high representation in the
lower levels of microentrepreneurs--the Solidarity Group target
population--it behooves management to include them. But not
exclusively. The "lessons learned" of a decade of WID indicates
that "all women" means "small." And ADEMI's ultimate promise is
for something very big.


4.5 Lessons Learned


4.5.1 Credit Projects in "McDonald's" Wrappers?


Jeffrey Ashe summed up the promise of the original ADEMI
conception, a twin Solidarity Group-plus-microentrepreneur focus,
in a striking phrase: as the model on which to build the
"McDonald's of Third World credit projects to help the poor."
Fast, profitable, efficient, low-cost, capable of reaching large
numbers of people--it could as well be an achievable prognosis








-82-


for an ADEMI-type program as it proved to be for the U.S. fast
food industry that grew out of McDonald's breakthrough idea. If
the ADEMI model were refined and "franchised" worldwide, its
potential to provide substantial economic benefits to conceivably
scores of millions of Third World women and men informal sector
entrepreneurs--while paying its own way--is incredible. Further-
more, economic benefits can bring considerable social benefits in
their wake, ranging from the virtually universal increase in
self-confidence among ADEMI beneficiaries to increased spending
on their children's education and welfare. Additionally, if the
ADEMI model, including the Solidarity Group component, proves to
be a successful private sector venture, then this new type of
"McDonald's" would presumably generate its own competitors.


4.5.2 Replication to Extend Mainstreaming?


The notion that private sector organizations would compete
to give credit to large numbers of poor male and female informal
sector entrepreneurs may be stretching reality too far. But,
even on a greatly reduced scale, if honest and efficient ADEMI-
type organizations prove replicable and capable of achieving
self-sufficient growth within several years, enormous progress
could be made toward two achievable development goals: incor-
porating women into mainstream development efforts and promoting
"growth with equity" for a substantial stratum of the Third World
poor.

All in all, the computer data confirmed subjective impres-
sions of women microentrepreneurs as good loan clients. Their
businesses grew impressively on all major indicators and outper-
formed those of their male counterparts in the clothing sector,
the largest single category. And although the Solidarity Group
component remained suspended, the president of ADEMI proclaimed
his willingness to revive it on an experimental, research-backed
basis. Finally, there was the alluring promise that the ADEMI
model might be a replicable way to extend mainstream development
benefits to women along with men in the urban informal sector of
much of the Third World.

Accordingly, let us conclude this examination of the full
ADEMI concept (i.e., both the individual microentrepreneur and
Solidarity Group components) by putting it in the context of
other microenterprise credit projects and then underlining the
features that simultaneously allow it to achieve success as a
mainstream development project while facilitating the incorpora-
tion of women.








-83-


First, the fact that women prove as good as or better than
men as loan clients is not unique to ADEMI. Lycette's review of
the literature (1984) shows that women's loan performance is at
least as good as men's. This finding is supported by two more
recent studies of Guatemala's SIMME urban microenterprise pro-
ject. By August 1988, 1,623 individual loans averaging $1,300
had been granted; 287, or 17.68 percent, were given to women,
although there was no special emphasis on reaching women clients.
Of the 1,623 loans, 248 (15.3 percent) were in arrears. Yet only
26, or 10.48 percent of the 248 were women's loans (Blumberg and
Reibel 1988). In August 1989, the SIMME project was revisited
shortly after an upsurge in late payments had been stemmed by a
new emphasis on one-on-one contact between individual loan client
and their asesores. By this time, women were 22.8 percent of
SIMME's 4,824 loan clients. Once again, they were less likely to
be in arrears: 41 percent of women, versus 49.87 percent of men,
were a month or more behind on their payments. Moreover, women
were less likely to be seriously in arrears: 20.35 percent of
men, versus. 12.36 percent of women, were four or more payments
behind (Blumberg and Revere 1989).

Furthermore, microenterprise credit projects with many women
clients have been found to be the most successful. For example,
according to Timberg (1988), Bangladesh's renowned Grameen Devel-
opment Bank has a minimal default rate, under 3 percent. As of
1988, it had provided loans to over 400,000 of the "poorest of
the poor." At the end of 1986, women were 69 percent of loan
clients, accounting for 55 percent of the cumulative loan amount.
Since then, the bank has further expanded loans to women, based
on their better performance. Timberg (1988) also cites
Indonesia's enormous and successful BKK program, which extended
2.7 million loans amounting to over $55 million just between 1972
and 1982, and which involves 60 percent female clients, mostly
traders. In addition, Tendler's (1987) meta-evaluation of 100
"livelihood, employment and income generation" projects (1987)
found that microenterprise credit projects were among the most
successful. Of eight best performing projects, five dealt exclu-
sively with women and the other three gave serious attention to
women clients from the start (McKee 1988).

What specific lessons can be learned from ADEMI? Perhaps
the most general one is that ADEMI ',ved so successful in reach-
ing poor and women informal sector entrepreneurs because it eli-
minated the high "up front" costs and barriers that typically
exclude them. ADEMI's procedures included the following:

-- It reduced the number of trips that a person had to make
to a credit office in order to process a loan applica-
tion (generally, to a single visit).








-84-


-- It did not require collateral for most loans (indeed,
the Solidarity Groups functioned as a group guarantee
for the poorest clients).

-It used "streetwise" promoters (asesores were typically
college night students who soon learned and cared about
the microenterprise sector) to select loan clients and
aid them with the application process. This contrasts
with the bureaucrats administering most credit programs,
who tend to intimidate poor, uneducated potential
clients, and who often have a deep-seated negative view
of clients' loan repayment inclinations.

-- It rewarded on-time payment by offering an immediate,
new, slightly larger loan for prompt and full repayment.

These procedures proved sufficient to allow the entry of substan-
tial numbers of women, who make up an appreciable proportion of
those working on their own account in the informal sector of much
of the Third World (see, e.g., Berger 1985; Berger and Buvinic
1988).

One final benefit could flow from extending the ADEMI model:
projects that put income in the hands of women get an additional
bonus. It has been repeatedly documented that income under fe-
male (versus male) control is more likely to be spent on family
"basic human needs," especially children's nutrition (Blumberg
1988 summarizes the empirical evidence). And control of this
income tends to enhance these women's role in household decision-
making, giving them greater leverage in fertility, domestic, and
economic decisions (Blumberg 1989 documents this fact). Thus,
the increased resources of women microentrepreneurs who receive
credit not only create jobs in the informal sector but also con-
tribute to such other development goals as lower fertility, bet-
ter child nutrition, and increased emphasis on children's educa-
tion and health (two other areas to which women are more likely
to dedicate income under their control). In conclusion, the
ADEMI model may provide a "win-win" scenario for both mainstream
and WID goals.







-85-


5. WORK, WEALTH, AND A WOMEN-IN-DEVELOPMENT
"NATURAL EXPERIMENT" IN GUATEMALA:
THE ALCOSA AGRIBUSINESS PROJECT IN 1980 AND 1985

by

Rae Lesser Blumberg
University of California, San Diego

with the assistance of
Maria Regina Estrada de-Batres
and
Josefina Xuya Cuxil


5.1 Summary


Women, it has been documented, are increasingly important
producers of wealth in today's world economy. In much of the
development community, to have women creating wealth makes a
stronger argument for including them in development projects than
to have women creating welfare.

Fortuitously, the results of this study, which looked at the
socioeconomic effect a food processing plant had on local women,
provide preliminary support for the idea that when women work to
create wealth, and get to keep some of it, both the efficiency
and equity aims of development are enhanced. Project data indi-
cate that income flowing into female hands increases their sense
of efficacy, which, in turn, bodes well for their future produc-
tivity and contributing to family welfare.

Using the groundwork laid during an earlier research effort
in an agricultural area west of Guatemala City, Guatemala, this
follow-up study compared the original research findings to the
circumstances found 5 years later. How had women fared over time
in a project that was not targeted to them but nonetheless
greatly affected them? By reexamining the four original sites,
the findings showed how the gender division of labor and of re-
sources affected both the men and women involved and the overall
development process associated with the introduction of a food
processing plant in the area.


5. Introduction


This paper presents something unusual in the social sci-
ences: a follow-up on a "natural experiment." The research







-86-


being followed up was not originally focused on women in develop-
ment, nor had it been planned as a "natural experiment." Rather,
Ken Kusterer, a sociologist from American University, had spent 4
months in 1980 conducting a study published as The Social Impact
of Agribusiness: A Case Study of ALCOSA in Guatemala (Kusterer,
Estrada de Batres, and Xuya Cuxil 1981). The research covered
three villages where poor peasants--mainly Cakchiquel-speaking
Mayan Indians--grew cauliflower, broccoli, and snow peas on con-
tract for ALCOSA (Alimentos Congelados Monte Bello, S.A.), a
wholly-owned subsidiary of a U.S.-based multinational company,
Hanover Brands. In addition, Kusterer studied the Ladino women
who worked at ALCOSA's processing plant, where the vegetables
were frozen and packed for export to the U.S. market. Kusterer's
work represents development research at its best: an insightful
combination of qualitative and quantitative techniques that
sharply illuminates the human impact of large-scale development.
And because Kusterer's well-written and richly detailed monograph
paid attention to both men and women, the ground was laid for a
serendipitous discovery: by rearranging the data from the four
sites studied, one could construct a "natural experiment" that
shows that as women's involvement in the activities and benefits
of a planned development project increased, so, too, did the
positive impact on both the women themselves and the project.

Although the women-in-development (WID) field has accumu-
lated a large body of knowledge and increasing policy prominence
in the years since Ester Boserup's landmark 1970 book, Woman's
Role in Economic Development, funding for in-depth field research
is not abundant. By returning to Kusterer's research site, one
could, at low cost, do a "five years after" follow-up of what
unexpectedly turned out to be a striking array, from low to high,
of women's involvement in the process and rewards of development.

This paper examines the findings from the ALCOSA Agribusi-
ness Project: To what extent is the project a success, indepen-
dent of its attention to women? To what extent is it a success
from the WID perspective? This report also attempts to highlight
the wider implications of what happens to women, to their men,
families, and communities, and to the project itself, when
planned development deliberately or incidentally affects the
gender division of labor and resources.


5.2.1 Project Setting


The Socioeconomic Status of the Area. The three villages of
poor contract growers and the town where the processing plant are
located all are part of the lush, green areas of Guatemala. The
three villages are to the west of the capital, Guatemala City, in








-87-


the "Indian Highlands" area. Although Guatemala must be con-
sidered a racially divided society, with the various Mayan Indian
groups definitely subordinate to the dominant Ladinos, moving
from one group to another is possible. To do so, one must speak
Spanish, not an Indian language, and must abandon Indian customs
and clothes. Nevertheless, many Indians resolutely cling to
their traditional identity even in the face of so many obvious
disadvantages.

The women retain much more of their visible Indian personas
than do the men. They are likely to speak almost exclusively in
their Indian tongue. Indian females receive less schooling than
males, so often their Spanish is rudimentary, at best. Large
numbers of them continue to wear their magnificent native cos-
tume, a hand-woven marvel of brilliant hues and distinctive de-
signs. Wearing the costume is an act both of proud tradition and
of courage.

The Mayan Indians of the highlands generally adopt a "family
subsistence strategy" ih order to survive. This does not mean
that every cent earned by every family member is automatically
pooled into a "common pot." The prevailing pattern is that the
male household head and the grown children who are still partici-
pating members of the household (even though they may not live
there year-round) contribute some proportion of labor and/or
earnings; they keep remainder of their earnings for themselves.
Guatemalan wives also tend to engage in economic pursuits. More-
over, Guatemalan wives' traditional income-generating ventures
(e.g., weaving) and market trading are done partly on their own
account. But qualitative evidence indicates that Guatemalan
women, like women around the world, tend to devote virtually all
income they control to household subsistence expenses.

The main subsistence crops in the highlands are corn and
beans. In addition, a variety of traditional horticultural crops
are grown: radishes, carrots, cabbage, and so on. While the
extent of women's involvement in field work varies from village
to village, they have a long history as market vendors, marketing
the family's surplus production, their own weavings, or other
products in the town and city markets. The women also raise
chickens and other small animals, tend kitchen gardens in many
areas, and perform all the relatively invisible tasks that keep
the household economy going and reproduce its labor force. They
and their children fetch water, fodder, and fuel; process har-
vested crops; and perform the cooking, cleaning, child care, and
other domestic activities. An important activity for many is
weaving for both household use and sale. Seasonal migration
(especially by men) to the coastal plantations and other possible
job locations helps round out the household survival strategy.








-88-


The survival of many families was jeopardized during the
years of political and economic instability that peaked in 1981-
1982. The research described in this paper was done prior to the
November 1985 election that put a popular civilian government,
under President Cerezo, in power for the first time since 1954.

The Development Context: Agency for International Develop-
ment (A.I.D.) Strategy. In 1975, LAAD de Centroamerica, a sub-
sidiary of the Latin America Agribusiness Development Corpora-
tion, promised a low-cost loan to Hanover Brands' new wholly-
owned subsidiary, ALCOSA. TLe world economy was in deep
recession at this time as a result of the first energy crisis.
A.I.D. was looking for new programs and ways to reach the rural
poor. It also sought to promote both agribusiness and "agricul-
tural diversification" (away from overemphasis on a small number
of traditional export crops). The LAAD de Centroamerica loan
project seemed to touch all the bases. According to the 1975
Capital Assistance Paper (the principal project proposal docu-
ment), the overall long-term development goal was "to create an
environment or the conditions in which the rural poor will have
increased opportunities for participation in the benefits emanat-
ing from development." To accomplish this:

This project will contribute substantially to the pro-
motion, development, and expansion of agribusiness
systems and enterprises in Central America through
addressing constraints in agricultural production and
supply systems where capital, management, training,
technical, and financial assistance can be productively
applied. To this eLd, LAAD will provide its speci-
alized resources to agribusiness opportunities which
offer maximum potential for beneficial economic impact
on a priority target group composed of small farmers
and landless workers.

The more specific short-term project purpose was twofold:

1. To develop agribusiness activities in Central
America in nontraditional agriculture which
increase the participation of the rural poor;
and

2. To develop LAAD into a self-sustaining financial
operation based on commercially available capital

A subloan to ALCOSA fit in well with the project purposes.
ALCOSA, backed by Hanover Brands, would introduce contract grow-
ing of nontraditional horticultural crops (principally cauli-
flower, broccoli, and snow peas) to poor and very poor Guatemalan
farmers. As people in the United States changed their health,







-89-


exercise, and dietary habits in the 1970s, the market for many
comznodities boomed. COpportunities were created for enterprises
located in the Third World to enter those markets. And, given
the sometimes strange interdependencies of the world economy, the
lives of the poor peasants in the 17 contract-grower villages
where ALCOSA was operating by 1980 would be transformed. Indeed,
the lives of the several hundred female employees in its process-
ing Plant would be revolutionized.

In 1975, women in development was not yet a real A.I.D. con-
cerl. In fact, the word "women" does not even appear in the 1975
Capital Assistance Paper for the LAAD loan project that would
fund AICOSA. Nor was women in development a concern of the Gua-
temalan Government. Absolutely nothing in the 1975 LAAD project
or its ALCOSA subloan was geared to even the mere recognition of
the gender--let alone gender-disaggregated analysis or the adap-
tation of the project to incorporate women. In short, although
women would participate and benefit in varying degrees in this
project, it was not because of any deliberate program design that
they did so.

Accordingly, the main aim of Kusterer's research was to
clarify ALCOSA's impacts (favorable or otherwise) on the lives
and fortunes of its contract farmers and processing plant work-
ers, women as well as men. As it turned out, many of the effects
on women, both positive and negative, fell into the "unintended
consequences" category.


5.3 Project Desion, Methodologyc and Gender Issues


5.3.1 Project Design


Some Guidinc Hypotheses. To address the broader concerns of
the A.I.D. study, the Guatemala research was guided by a series
of hYPotheses. Hypotheses 1 to 7 deal with gender stratification
and women in development; Hypotheses 8 and 9 are about planned
development.

1. For women, control of economic resources (relative to
men) is the most important variable affecting gender
stratification and a variety of other consequences
affecting their lives.

2. Independently controlled income (versus control of land
or other property) is the most readily accessible form
of economic power open to women.







-90-


3. Therefore, women's independently controlled income is
associated with their greater leverage in fertility de-
cisions and household economic and domestic decisions.

4. Women's activities that do not result in independently
controlled income are not associated with women's
greater leverage in family fertility and economic
decisions.

5. Women's independently controlled income is associated
with their greater sense of efficacy and self-esteem.

6. Women are more likely to spend income under their con-
trol on family well-being, especially children's nutri-
tion and necessities.

7. When women's independently controlled income falls, they
lose household power more sharply and quickly than they
t when their income rises.

8. rce development resources are targeted to poor,
s groups, such as women, the landless, and so
-here is a tendency for those resources to be
c uted in terms of "power and privilege" (Lenski
19o-;, which is manifested as a "trickle up" of
benefits.

9. Vulnerable groups (women) may be shut or pushed out of
project benefits unless (1) barriers to the group's re-
ceipt of benefits are reduced by deliberate selection
procedures and appropriate delivery mechanisms, and (2)
there is continued oversight of project benefit
allocation.

Kusterer's research sites, arranged in the WID-related pro-
gression in which they will be presented here, were as follows:
three villages in the Cakchiquel-speaking western highlands
areas--Patzicia, Chimachoy, and Santiago Sacatepequez--and the
town of San Jose Pinula, where the ALCOSA plant is located, in
the Ladino area just south of Guatemala City.

Village of Patzicia. What is striking about Patzicia in WID
terms is that (1) neither Cakchiquel nor Ladino women worked in
the fields in pre-ALCOSA days; (2) the ALCOSA vegetables are ex-
tremely labor-intensive; (3) by 1980 such critical labor short-
ages had developed that large farmers had generally abandoned
contract growing entirely and some of the poorer farmers were
forced to spend scarce resources in expensive "labor-saving"
farming methods; and yet (4) as of 1980, women still did not work
in the fields, to their husbands' clear disadvantage. As








-91-


Kusterer reported (Kusterer, Estrada de Batres, and Xuya Cuxxl
1981, 17): "Only Patzicia has such sharp problems of labor sup-
ply because only Patzicia overlooks what has emerged in the other
two villages as the obvious solution: farm labor for women."
What had happened in the intervening 5 years? Would there still
be farmers engaged in contract growing for ALCOSA 1985? And if
so, would the women of Patzicia still be as male-dominated, sub-
missive, and timid as they were in 1980?

Village of Chimachoy. 1980 was a bad year for this Cakchi-
quel village of about 100 families, located on top of a mountain
reached by a barely passable dirt track. Kusterer found that, of
the three villages, Chimachoy was the most negatively affected by
ALCOSA's continual calls for more production. Chimachoy's farm-
ers cut back on their traditional food crops to expand cauli-
flower production. So when ALCOSA was hit with an unexpected
torrent of cauliflower from its 17 grower villages in July 1980
and abruptly suspended its purchase, some two-thirds of Chimachoy
farmers were left with no viable outlet for their crop (one-third
were not then harvesting). Hunger, anger, and drastic losses
followed. There was even an attempt to burn down ALCOSA's buying
shed. Many of the villagers were left with onerous levels of
outstanding debts. The better-off villagers had been building
concrete block houses to replace their cane and straw huts, many
with special loans; many poorer families were paying in install-
ments for small purchases.

With respect to the WID hypotheses presented above, what is
significant about the Chimachoy case is that the women, who pre-
viously had helped in the fields only during planting, now were
pulled into 2 to 3 days of horticultural labor each week on top
of their normally overburdened schedules. As a result, they had
to cut back on their marketing trips to town, thus reducing their
only independently controlled income. In consequence, Kusterer
found, they were increasingly dependent on their husbands and had
a possibly decreased voice in household affairs. Furthermore,
although the women were working 2 to 3 days per week on the
ALCOSA vegetables, the ALCOSA payment came in the form of a check
(that had to be cashed in the nearest big town) made out solely
to the husband. In short, in Chimachoy, Kusterer found the situ-
ation that has become a classic in the WID literature: a devel-
opment project that increases women's labor burden while failing
to give them a direct share of the new benefits, and eroding
their existing resource base. In the WID literature, such a sce-
nario often has been tied to lowered productivity and efficiency
ana even to the failure of the project itself. (For supporting
evidence, see Blumberg 1989.)

Village of Santiago Sacatepequez. In 1980, this was the
most successful of the three villages with respect to the







-92-


project. A Swiss-formed cooperative, Four Pines, dealt directly
with ALCOSA on behalf of its more than 400 members, and, in order
to sell to ALCOSA, farmers had to be a co-op members. Results
were exceptional (Kusterer, Estrada de Batres, and Xuya Cuxil
1981, 61):

The average farmer in Santiago has lower costs than his
counterparts in the other two towns, yields that are
more than twice as high, proportions of first-quality
product that are 15 percent higher than Patzicia and 20
percent higher than Chimachoy, and net income per unit
of land that [is] many times higher than elsewhere.

As a possible explanation, Kusterer cites the labor-inten-
sive farming methods used on the small plots (Santiago is only 20
kilometers. from Guatemala City and on the fringe of suburban
development, so holding are very tiny), better initial knowledge,
and better advice from the co-op's agronomists than from
ALCOSA's. Might there also be a connection with the project's
1980 "WID status"?

Specifically, in 1980, Kusterer found that the women worked
more as partners than as helpers to the men. Although the ALCOSA
vegetables were new, these women already had a horticultural
tradition and were skilled, experienced farmers. Moreover, in
1980, women seemed to be sharing in direct benefits. Women were
almost as likely as men to deliver produce to the co-op and were
often lined up with the men at the window where cash payments
were made. They had paid a price: in addition to an added
"double day" burden, the women had to cut back on their formerly
frequent marketing trips to Guatemala City's main terminal market
(their principal source of independently controlled income).
Kusterer did not address directly the extent to which women actu-
ally shared in direct benefits and the consequences of their
involvement in the project. Nevertheless, it was clear that
Santiago women were getting more direct benefits than women in
the other two villages. Questions for 1985 would include how
Santiago women were faring relative to their husbands, and what
the relative success of the project was for the co-op and for the
people affected, both collectively and disaggregated by sex,
class, co-op membership, and so forth.

San Jos6 Pinula's Processing Plant Workers. In 1980, some
85 percent of the processing plant's permanent workers and 100
percent of the seasonal ones were female. ALCOSA, unlike most
Guatemalan employers, paid the minimum wage and legally mandated
benefits. Moreover, during most of the 8- to 9-month "high sea-
son," ALCOSA worked very long shifts--12 to 16 hours. As a re-
sult, according to Kusterer's figures, the plant's female em-
ployees made from 150 to 300 percent as much as they could have