Svers ------- -
.. ---. -- -;-........ ... -.._... ... _..... l.J.~' ~ *- --.-- 4 I -a
GENDER ISSUES IN FARMING SYSTEMS
RESEARCH AND EXTENSION
S_ Intra-Household Gender Issues in Farming Systems in
Tanzania. Zambia and Malawi
Jean M. Due*
The Gender Issues in FSR/E Conference provides an excellent opportunity
to examine various aspects of the issue and of the FSR/E mode of
research/extension. This paper undertakes to: (1) test the Feldstein
(1985) conceptual framework by using FSR research results from Tanzania,
Zambia, and Malawi, and (2) highlight the contributions of social scientists
The Hildebrand-Poey (1985) conceptual framework suggests two areas for
the contributions of social scientists. The first is unencumbered social
science, which in its own right contributes new knowledge and understanding.
Examples are research on the economies (national, regional, international)
and societies into which new technology will be introduced. The second is
social science research which has continuous interaction with research of
agricultural scientists, is restricted to specific variables being addressed
by both groups, and in which there is active communication among all team
members. Examples are the farming systems, culture and households into
which new varieties will be extended. My illustrations of social science
research fit nicely into these two categories. The results/illustrations
come from research into development of new high yielding, disease, insect
and drought resistant beans in Tanzania and Malawi as part of the
Bean/owpea CRSP and from FSR undertaken in Zambia.
Now, shifting to the Feldstein framework, let me begin with the design
phase as that provides an understanding of the research in Tanzania.
*Jean M. Due is Professor of Agricultural Economics, University of Illinois
at Urbana-Champaign. The Bean/Cowpea Collaborative Research Support System
(CRSP) is funded by USAID in 18 countries in Latin America and Africa.
In the original design of the Tanzanian bean/cowpea CRSP research it
was agreed that traditional bean cultivars would be tested for resistance to
disease, insects, and drought. Then the best of these varieties would be
crossed with other materials which tested highly from international
institutes, surrounding countries, or U.S. seed banks. The breeding program
focused on breeding HVYs for specific agro-climatic conditions in Tanzania
and then included other desirable traits such as cooking quality, seed coat,
seed color, storability, and yield stability. Plant breeding is a long-term
process and the breeding program began before the diagnostic data were in;
however, the breeders were nationals who were cognizant of local conditions.
As the breeding program progressed and new data were fed into the
program from the diagnostic phase and FSR, priorities in the program changed
to include the cooking quality, seed coat, and storability aspects as well
as narrowing the potential crosses to those with best performance, based on
desired characteristics. Each year the pool was further narrowed so that
most promising lines could be tested on farmers' fields in both high
altitude-rainfall and low altitude-rainfall areas. A shortage of seed
limited the number of farms on which the new varieties (M3 101 and
Kabanima) could be tested initially.
It would seem important to have information on the level, quality, and
distribution of resources (land, labor, capital and other) of the community
and households as one ascertains the constraints facing farmers in the area.
Agro-climatic, soils, and census data provide a base of information for the
research domain. Visits with farmers and agricultural and government
personnel in the area can determine if this is the appropriate locale of the
research domain. Visits of the research team to the area are important. In
Tanzania when the team was consulting with agricultural officials as to
major bean growing areas, one area was suggested which, when visited and
farmers consulted, was one in which only one-third of the farmers grew
After visits to prospective research areas and rapid rural appraisal, a
more detailed FS-baseline study was undertaken to learn about the total FS,
the manner in which beans fitted into the system, distribution of size of
farms, production for consumtion or sale, and so forth. The average farm
was small, averaging 8.4 acres. The major inputs in agricultural production
were land and labor; little fertilizer or chemicals were used in these FS.
A few farms hired labor and/or oxen or mechanized ploughing; a few farmers
owned oxen. Capital equipment owned per household was limited to hoes,
axes, machetes and simple tools; cash reserves were also small but knowledge
of the farming system was impressive, having been acquired over the
In most areas of Tanzania land is available for increased production.
Technically the land is owned by the state but is allocated to individuals
by the community or ethnic group. Since the Arusha Declaration of 1967 when
rural households were moved from scattered sites to ujamaa villages, village
officials control the allocation of land within the village but, in general,
continue traditional land allocations. Households can obtain more land
outside the village unless the village is surrounded by other villages. In
one area of this research land was available, in the other (in the high
altitude-rainfall area) acreage could not be expanded.
Capital for off-farm inputs (fertilizer, chemicals) was theoretically
available through the Tanzania Rural Development Bank if the borrowing was
undertaken in the name of the village. In general the Bank does not
allocate credit to individual farmers. Individuals can borrow from friends
and relatives, usually at zero rates of interest. However, more important
than the official credit constraint to individual farmers was. the scarcity
of foreign exchange to import farm inputs. The economy of the country is
depressed and foreign exchange reserves depleted. The FS study showed that
few farmers borrowed through the bank during the study period and that
almost no fertilizer or chemicals were available in the area during the
1980s. Credit, when available, was only in kind; it was not available for
tools, hiring labor, oxen, or tractors for ploughing or for consumption
expenditures. Capital was available for livestock purchases only an state
farms and institutions. The rural development bank does make loans to
villages for mechanized machinery and petroleum when foreign exchange
permits; however, villages in this region had not qualified for mechanized
machinery loans during this period.
Probably the most significant information that was discovered regarding
inputs was that women choose the bean seeds for planting. This finding has
important consequences for extension of new varieties developed. If there
is a perception that male extension workers cannot visit female farmers,
adoption of new varieties will de delayed or distorted.
In the diagnostic phase visits with government and other officials did
not indicate that other than male labor was involved in agricultural
production. However, on the drive from the university to the potential
research sites it was obvious that females were significantly involved. How
involved? When labor data by gender were included in the baseline study of
the farming system, it was found that females contributed one-half the labor
on major agricultural crops and more than half on minor crops. In addition
females contributed more than one-half the labor on household tasks, child
rearing, and fetching water and fuel. Labor allocation on the major crops
by cropping practice by gender is shown in Table 1.
INSERT TABLE 1 HERE.
It is shown that females contributed more labor than males to planting,
weeding, and harvesting of all the major crops and 59 and 67% of total labor
for beans and rice, respectively, in this area of Tanzania.1 Even in bean
areas where crop acreage has been expanded with mechanized or ox ploughing,
the percentage of labor contributed by females did not decrease markedly
(Due et al., 1985).
Barnes-Mconnell (1985) collected daily data from farm households in
bean growing areas of northern Malawi; she found that over the year males
and females in the household each provided about half of the total labor
used (with child care omitted), with females contributing slightly more
(52.3%) than males (47.7%). Her data show that females contributed 42.5% of
the agricultural labor per year, 73% of the domestic requirements, and 42.7%
of the economic inputs (buying, selling, brewing beer, etc.). While the
total was a team effort and comparable across gender, there were large
differences in gender contributions both among and within categories.
All farming system studies in tropical Africa show tight labor
constraints at certain periods of the crop calendar. One of the most
interesting insights in Barnes-McConnell's data is the relationship between
1 Labor on marketing of beans is underreported. This is obvious when
visiting markets in the research area.
labor demands and health. Families spontaneously reported illnesses and
deaths, with the most illnesses toward the end of the rainy season (February
through April), and the most deaths within the March-June season. Dampness,
water problems inside the houses, the heavy work load and reduced food
availability explain this relationship.
Few farming systems studies interview women unless they are "heads" of
households. In a study in Zambia, Due and colleagues at the Rural
Development Studies Bureau of the University of Zambia (1985) interviewed
women in farm households to obtain the relative hours of labor allocated to
agricultural and other tasks during the farming and other seasons. Her
results astounded her male colleagues. Females contributed more total hours
per day (8.5) than males (7.4) .in agriculture as well as in non-agricultural
tasks (5.0 hours for females compared to 1.1 hours for males).
In some parts of tropical Africa men and women farm fields jointly, in
other parts women have some separate fields and farm other fields jointly
with men; men may work fields communally with other males; or work parties
may do some crop operations (e.g., weeding or harvesting) by gender. In
other situations labor is allocated by crop with males taking primary
responsibility for the export/nonconsumable crops (cotton, coffee) while
females take primary responsibility for food crops. Knowledge of these
labor patterns is important for the agricultural scientists as they plan
Hired labor for crop production may be either male or female with
female labor always being less well paid than male. Although hired labor is
not widely used on most smallholder farms, it is used as acreage is
expanded. In Tanzania work parties do weeding and harvesting on some farms
but most of the labor is family labor. Work parties may be paid in
food/refreshments or in cash.
In one high altitude-high rainfall area of Tanzania where beans are
grown on steep hillsides with 66% for the market and 34% for family
consumption, 22 of 60 sampled families hired some labor but the average
length of time per household was only 3 days. In another area where farms
were larger and more hired ploughing was done, only 26 days of male labor
and 9 days of female labor was hired by 85 sampled smallholders. Hired
machinery accounted for 26% of the total farm operating costs for these
families but hired labor accounted for only 22%; female labor on average was
paid 71% of male labor per day.
(i) Crop and livestock production
There are high incentives for these smallholder families to
enter into agricultural production because that production provides over 90%
of the family food supply. If that food supply fails, there may be little
food available in the market as surpluses are moved to the urban areas.
Eighty percent of the families reported that their first priority in farming
was the provision of the family food; the second was income for consumption
and other purchases. During this time of a depressed economy in Tanzania,
wage goods were extremely scarce and prices high. Thus the incentive to
produce to feed the family was very high but price incentives provided by
the government through the official agency, the National Milling Company,
were low. Most governments in tropical African countries have kept farm
prices low to. placate vocal urban consumers (World Bank, 1981, 1984, Eicher,
1982, Due, 1986). Most of the farm families interviewed in the Tanzanian
baseline/FS studies sold their food crops at local markets rather than
through official government agencies, as prices in the local markets were
three to five times higher. In addition farmers were paid promptly, as does
not happen through official marketing channels. There is also active trade
in "over the border" markets in most countries; this trade does not enter
official export statistics.
In the major FS studies in Tanzania, it has been found that in general
families consumed 50% of total crop production and sold 50%; percentages
consumed and sold varies by crop; all of the cotton and coffee was sold, and
in some areas (the high altitude bean growing areas where land is a
constraint) all of the maize was consumed. But in other areas about 50% of
beans, maize, and sorghum were consumed and 50% sold.
Most of these smallholder families in Tanzania have some livestock and
some of the families have considerable cattle, sheep and goats which are
herded away from the farm. Reported consumption of livestock (including
poultry) was low; livestock provided only 23 of 2,295 calories consumed per
person per day for the sampled families. But livestock sales were a
substantial addition to income from crop sales, providing between 3 and 39%
of average total crop sales on sampled farms.
(ii) Non-crop income or "off-farm" income
Most of the FS studies undertaken have not generated data on
earnings of non-crop and livestock income, that is, income obtained from
selling small quantities of the minor crops (vegetables, fruits, etc.) in
the markets, selling beverages along the road, brewing beer, and employment
off the farm. In tropical African rural areas these avenues of income were
thought to be too low to include. But our data have shown that important
small sums are generated in this manner, especially by females. Beer
brewing is the most important source of non-crop income for women; it is, of
course, another use of crops which are not usually sold through the market.
Females also engage in working an other farms and selling small quantities
of fruits and vegetables. In general females received little child support
(even in female-headed households) or gifts from friends/relatives. Males
worked in construction, butchering, tailoring, farming and road work to
supplement their farm income. This so-called "non-crop income" provided 84%
as much as crop and livestock sales in Arusha area where there were more
opportunities for off-farm income and 11% of crop and livestock sales in
other areas more remote from urban areas. Males earned more income from
employment off the farm while females earned most of their income from beer
brewing and selling fruits and vegetables.
The Zambia data (when farm females were interviewed) showed these off-
farm earnings to be even more important than in Tanzania. In Zambia these
income sources contributed 32% as much as crop and livestock sales; females
earned more income than males (K125 compared to K89 for males per
smallholder family); females earned 58% of the off-farm income and males
Data on income generation from FS studies in three different areas of
Tanzania are shown in Table 2.
INSERT TABLE 2 HERE.
(iii) Decisions on use of cash income
Little is really known of the intra-household decision-making
on allocation of cash income in smallholder families in tropical Africa.
This is not surprising as little is really known about these decisions in
North American families, and decision-making varies by family and by
consumer good. However some data were generated in the Zambian FS studies
when the farm women were being interviewed. These data are only indications
Table 2. Average Value of Crop Production and Cash Earnings per
Sampled Family in Tanzanian Shillings, by Area, Tanzania,
Sample size .........................
Value of crop production ...........
Percent of crop cosumed (%).......
Value of crop sales ................
Value of livestock sales...........
Total crop and livestock sales.....
Farm operating costs...............
Net farm cash income ..............
Off-farm income & gifts ............
Total cash income...............
Family living expenses .............
Balance ................. .......
Source: Due et al, 1985.
or generalities; anthropologists would insist that one would have to observe
as well as ask to ascertain who within the family is the major decision-
maker. These farm women responded that husbands and wives jointly made
decisions on the use of husbands' cash earnings in 58% of the families with
husband and wife present; in 40% of the cases the decision was made by the
husband alone and in only 2% by the wife alone. When the same question was
asked regarding wives' cash income, the decision was made jointly in 50% of
the cases, by the wife alone in 40% and by the husband alone in 10%.
(d) So What? Results?
(i) The whole emphasis of FSR/E is one from the "bottom up" that is
from the farmer's point of view. However, it is clear that most FSR assumed
that the farmer was a male and it was sufficient to interview or discuss
constraints only with males. Often host country colleagues were shocked
when the extent of female labor participation was documented. These data
were obtained not from rapid rural appraisals but from more detailed FS
research; the rapid rural appraisals did not imply female labor was
important, but that women grew a few vegetables near the house.
Agricultural and university officials talked about "women's gardens" as
their contribution to the farming system.
(ii) When male farmers were asked about the major constraints to
increasing agricultural production, the responses never included a shortage
of labor. Drought, rains which were too early, too late, or too little,
insects, pests, etc. were the major constraints mentioned. Yet when the
farm wives were interviewed in Zambia a very different list was forthcoming.
In Zambia the question was phrased differently. Rather than being asked
their major constraints in increasing agricultural production, the women
were asked what types of development would most assist them in farming. The
overwhelming reply was improvements in farming, followed by availability of
credit, clinics and wells. Twenty of the 54 responses on farming
improvements were for labor-saving devices oxen and ploughs, and tractors
and farm equipment for hire. This is not surprising when one documents the
hours which females contribute to the FS! Smaller numbers requested higher
farm product prices, lower input prices, loans to purchase cattle and
grinding mills and assistance in drainage. Credit was listed separately as
credit is not available to females through official government channels in
Zambia unless they are widowed or divorced.
(iii) The FS mode emphasizes the complete FS rather than individual
crops. Thus, even when our-research focused on breeding higher yielding
bean cultivars, those new cultivars must fit within the present FS including
the tight labor constraints at certain seasons. The FS mode also emphasizes
the household-farm model of smallholder production with production for both
consumption and sale. Innovations also must fit within other input
constraints in the short run; if this is an area where fertilizer and
chemicals are not available most years, the high yielding varieties must be
higher yielding without these inputs. It is difficult to persuade
agricultural scientists to test under "Zero input" conditions!
It was also difficult to persuade the agricultural scientists to test
the HYVs under intercropping conditions, the manner in which most
smallholder beans are grown in Tanzania. It is much easier to obtain yield
data under monoculture. As a result of the FS studies, the focus of the
breeders has changed; data on family preferences of beans for consumption,
sale and storage have been incorporated into the breeding program. It is
now understood by all the researchers that at least one HYV is needed for
the cooler, high altitude areas and one for the hotter, lowland areas. On-
farm trials, therefore, were undertaken in each area on representative
(iv) When research on a major innovation is being undertaken jointly
with host country scientists it can be assumed that much of these FS results
are already known by the national researchers involved and the FS-baseline
data are generated primarily to increase the knowledge of the expatriate
colleagues. That assumption is not necessarily valid; FS data are necessary
for both groups. New innovations which fit into or relax the labor
constraints will be more rapidly adopted than others which accentuate labor
constraints. he CRSP researchers have decided that short maturing
varieties of beans/cowpeas will better fit the labor constraints and food
needs of smallholder families in many areas than longer maturing varieties.
III. ., TEI'TNG-EVATUATIONC
A shortage of seed (due to breakdown of irrigation equipment in the .dry
season) and shortages of petroleum and spare parts for the transport
vehicles severely constrained the number of farmers which could be included
in on-farm trials of IM101 and Yabanima cultivars approximately 50 miles
fro Sokoine University of Agriculture.
Several months before planting in each farm testing domain, university
personnel visited the villages and the village chairperson (a government
official) and met with groups of farmers to inform them of the request to
test this potentially HYV on some of their fields. At that time farmers
were enthusiastic about this opportunity. However, at planting time when
the team again returned to the villages farmers, dissatisfied with
government agricultural policy, were increasingly less interested in
participating. Unfortunately the same day the university faculty members
went out to villages to make final arrangements for the on-farm trials, a
government official was meeting with farmers to urge them to plant at least
a minimum acreage of cotton, a crop with a very low price and yield and very
high labor requirements. Farmers were angry at the insistence. The
university personnel had to follow this official on the podium! Because of
the general hostile attitude toward the government official (that same
attitude was transferred to the university faculty, also paid by
government), the university personnel decided it was best to ask for
volunteers to try this new potentially HYV. Since there was seed sufficient
for only 12 farmers, 12 volunteers were selected and the planting
instructions and seed left with the village chairperson and the extension
agent (both government officials). Time did not permit university personnel
to stay and plant that day, nor could they return to visit the sites for a
When the researchers returned and the FSR data were collected, they.
found that the volunteers in one village included 4 "farmers" who earned
most of their income from non-farm sources (one village chair, one bar
parlor operator, one butcher) and in both villages and the distribution of
farms by size included more of the larger farmers than in the village as a
The new varieties outyielded the traditional varieties by about 70% but
that the variation in yield from farmer to farmer was very high. Since
yield stability is important to these smallholders, this aspect of the HYVs
will be investigated. In order to increase the seed pool for the next year
all available seed was purchased from the farmers; therefore cooking
comparisons could not be ascertained at that time. Other data relating to
farmers' preference of plant structure, disease resistance, insect
resistance, etc. were also obtained.
IV. REMMENDATION, ADOPTION & DISSEMINATION
Further testing and breeding will be done before final recommendations
will be made. However, serious questions are being raised about the ability
of the present government seed multiplication service to provide adequate
bean seeds and the present extension service to extend HYVs. These issues
will be discussed with government officials. A marketing study is being
planned to estimate the impact on the present marketing structure if
significantly increased production of beans is realized.
The FSR/E still has not addressed the needs of female-headed households
which are estimated to make up 15% or more of smallholder farm households in
many parts of Africa. Mollel (1986) is investigating the T & V Extension
System implemented in one region of Tanzania (Tanga) from 1980 to 1984 by an
international donor agency. He ascertained relative crop acreage, crops
grown, net income and visits of extension agents to contact, non-contact,
and female-headed farm households in th T & V area. His data are shown in
INSET TABLE 3 HERE.
It is evident that the female-headed families were unable to prepare as
much land and plant as much acreage as the joint families (with both husband
and wife present); thus their ability to sell surplus production for cash
was decreased. These women compensated partially by allocating their labor
to non-crop income generation brewing beer and selling small quantities
of fruits and vegetables in the local markets. The net cash income per
family of the female-headed households was only 54% as large as that of the
contact farmers and 55% of that of the non-contact farmers.
The female-headed households were visited much less frequently than
joint families by extension agents; 72% of the female-headed farmers
-14 (a) -
Table 1. Percentage of Labor Days Contributed by Females by Operation by
Crop, Sampled Families, Kilosa, Tanzania, 1980*
Maize Sorghum Rice Cotton Beans Sunflower TOTAL
% % % % % % %
Land preparation 44 37 61 39 55 34 46
Planting 52 41 77 48 60 40 56
Weeding/thinning 51 43 65 51 59 40 52
Spraying 0 0 0 0 0 0 0
Harvesting 54 46 71 51 69 43 58
Marketing 17 16 50 31 0 12 21
TOAL 48 40 67 39 59 39 48
Source: Due and Anandajayasekeram (1984), p. 589.
Table 3. Comparison of Crop Acreage, Income, and Extension Visits Between
Contact, Non-Contact, and Female-Headed Households,
Tanga Region, Tanzania, 1985.
Contact Non-contact Female-Headed
Sample size..................... 32 34 32
Acres in crops.................. 2.8 2.1 1.4
Crop sales less farm cash
operating expenses (Tsh.).... 2,773.60 1,364.80 318.75
Off-farm income (Tsh.).......... 962.50 2,294.40 1,685.50
Net cash income (Tsh.)....... 3,736.10 3,659.20 2,004.25
Visits by extension agent:
No visits ................... 5 19 23
1 & 2........................ 8 9 4
3 & 4........................ 8 4 1
More than 4.................. 10 2 4
Source: Mollel (1986).
received no extension visits compared to only 16% of the contact farmers and
56% of the non-contact farmers. Only 5 (16%) of the female-headed families
had 1 to 4 extension visits per year while 50% of the contact and 38% of the
non-contact farmers were visited 1 to 4 times. With less information,
female headed households were less able to try the recommended practices of
improved seed, early planting, spacing, weeding, and fertilizer application.
From this experience, FSR which does not have a person on the team who
is sensitized to gender issues will miss important information. In these
examples the importance of female labor in the FS, of females in decision-
making about crops planted, seeds chosen, and income allocated would have
been overlooked. Female-headed households continue to be invisible in FSR/E
in Africa. If the FSR/E is a tool to improve the total FS, then gender
issues must be considered. Mhe Feldstein model assists in focusing
attention an these gender issues.
1985 1985 Bean/Cowpea CRSP Malawi Project, 1985 Annual
Reports from the Social Science Cmponent, Michigan State
University, (mimeo). East lansing.
Due, Jean M.,
and P. Anandajayaskeream.
"Ccntrasting Farming Systems in Morogoro Region, Tanzania,"
Canadian Journal of African Studies, 18(3).
Due, Jean M., Marcia White and Timothy Rocke.
1985 Beans in the Farming Systems in Two Regions of Tanzania,
1980-82, Technical Report No. 4, Tanzania, Department of
Rural Economy, Sokoine University, Morogoro, and the
Department of Agricultural Economics, College of
Agriculture, University of Illinois at Urbana-ampaign,
Due, Jean M., and Timothy Mudenda.
1985 "Women's Contributions to Farming Systems and Household
Income in Zambia," MichiQan State University WID Series
No. 85, East Lansing.
Due, Jean M.
Eicher, Carl K.
"Agricultural Policy in Tropical Africa: Is a Turnaround
Possible?" Department of Agricultural Economics Staff
aer, No. 86 E-338. Urbana.
"Facing Up to Africa's Food Crisis," Foreign Affairs,
"Intra-Household Dynamics and Fanning Systems Research and
Extension Conceptual Framework," a paper presented at the
Farming Systems Conference, Kansas State University,
(mimeographed), Manhattan, IS.
Hildebrand, Peter E., and F. Poey
1985 On-Farm Agronomic Trials in Farming Systems Research and
Extension. Lynee Reinner Publishers, Boulder, CO.
"An Evaluation of the T & V Extension System in Tanga
Region, Tanzania," M.S. Thesis, University of Illinois
Accelerated Development in Sub-Saharan Africa: An Agenda
for Action. World Bank, Washington, D.C.
Toward Sustained Development in Sub-Saharan Africa. World
Bank, Washington, D.C.
Mgeta and Kilosa districts
b. 1200 to 1800 metres at Mgeta, 500-1,000 ms. in Kilosa
d. Precipitation 760 to 1,600 mm in Mgeta, less than 1,000 at
Slope severe in Mgeta, flat in Kilosa
Socio-Economic explained in text
Nature of cropping system explained in text
Trial details crop beans, intercropped and monocropped
previous traditional varieties offshoots of Canadian
'~D 101 and Kabanima
d. planting: Mgeta, Nov. and March, Kilosa, April harvest
4 mos. later
g. plot size 100 sq.m. laid out beside one of identical size
for the traditional variety commonly used
h. farmer managed; treated the two plots alike re weeding, etc.
6. Hypothesis: That higher yielding, insect, disease and drought resist-
ant varieites would increase smallholder family incomes and well-
being and nutrition.
b. In a country short of food and protein, beans could improve caloric
and protein availability; present price policy too low and infra-
structure and transport might not handle markedly increased pro-
duction at this time. The economy of the country is very
c. Farmer assessment limited due to limited seed available for the
trials in the initial year. Farmers who were involved liked the
7. Explained in text.