Modeling Adoption Potential of Improved Fallows in Southern Mali Using
Linear Programming


B. Kaya' and P.E. Hildebrand2

A Poster presented at the American Biennial Meeting of the Association for Farming Systems
Research-Extension (AFSRE), Guelph, Ontario, Canada October 20-23 1999


1School of Forest Resources and Conservation, University of Florida, P.O. Box 110410, Gainesville, FL
32611-0410, U.S.A.

2Food and Resource Economics, Institute of Food and Agricultural Sciences, University of Florida, P.O.
Box 110240, Gainesville, FL 32611-0240, U.S.A.









Abstract
Agricultural production in the Koutiala region, southern Mali, is based on cash sources
(cotton and groundnut), cereal sources (maize, sorghum, millet), and a store of wealth (livestock).
In these low-input farming systems, crop production is seriously constrained by soil fertility
decline. Research is being conducted in the region to investigate the potential of improved fallows
planted to leguminous agroforestry tree species to improve soil fertility and crop production. This
study examines the potential for adoption of this technology for a "Type A" household using
linear programming-based modeling. The model revealed that an improved fallow would be an
interesting venture only if fodder is valued and if maize yields of at least equal to or higher than
the regional average yield of 2500 kg ha' can be achieved. Improved fallows are not financially
attractive to farmers if they do not produce benefits other than improved grain yield. Any subsidy
program which would prevent farmers from cutting the fodder, as secondary output before the
end of the planned fallow length, would not have adoption potentials. A special fallow installation
loan program, similar to the one that cotton enjoys, would make the venture viable.








Introduction
Land degradation and soil fertility decline are the main constraints to food production in
low-input tropical farming systems such as in Mali. The use of fertilizers is far less than the
quantity of nutrients exported through erosion, leaching, crop residues and grains, and other
natural mechanisms. Lack of cash and adequate infrastructure, i.e. roads, transportation and
markets, limits further the use of chemical sources of nutrients. Traditionally, farmers relied on
long fallow periods to replenish the depleted soil fertility. However, due to increases of human
and animal populations and land use pressure, fallows have been reduced both in length and in
area or even abandoned in many farming systems.
The introduction of improved fallows planted to leguminous tree species to achieve the
function of natural fallows in a shorter period (3 to 4 years) has been initiated. Biophysical
research results (Batiano and Mokwunye, 1991; Palm and Sanchez, 1991; Mittal et al., 1992;
Kwesiga and Coe, 1994; Mafongoya and Nair, 1997; Mugendi and Nair, 1997) ascertain the
potentials of such managed fallows to soil fertility replenishment in the region. The present
research was conducted in the southern Mali Koutiala region in two supporting (on-farm and on-
station) experiments from 1995 to 1998. The results of these experiments and secondary data
collected by the Sikasso farming systems research team (ESPGRN) and the Malian Cotton
Company (CMDT) are used in this paper to model the potential adoption of improved fallows in
the region using linear programming for a "Type A" household.

Materials and methods
Description of the study area
The climate in the Koutiala region is sudano-sahelian with two main seasons: one rainy
season from May to September followed by a dry one from October to May. The rainy season
starts early in May with annual rainfall of 600 to 800 mm at 90% probability. Rainfall is
characterized by large intra-annual and inter-annual variability, further exacerbating the problems
of inadequate soil moisture (Coulibaly et al., 1998). Agricultural production is based on a rain-fed
cereal (maize, millet, sorghum) and cash (cotton, peanuts) crops. The rotation is mainly cotton-
maize/maize+millet-sorghum. Cotton is the main cash crop and the market is quite secure though
the farmers have little or no control over the price. It is the only crop which receives fertilizer and
other chemicals (pesticides and herbicides). The introduction of agricultural equipment has made
land preparation and weeding operations more rapid; however, most farm activities remain
manual. The drastic reduction of fallow area and length, or its total elimination from the rotation,
has resulted in a dramatic soil fertility decline and severe water erosion problems.
There is a large variation among farmers according to household composition, land
holding, wealth, farm equipment, and their risk bearing capacities. Thus, the Sikasso Farming
System Research team (DRSPR, now ESPGRN), in collaboration with Compagnie Malienne pour
le Developpement de Textiles (CMDT, the Malian cotton company) had identified in the early
eighties four major groups (or recommendation domains) of farmers (Kleene et al., 1989). This
study was conducted for a "Type A" household which has the following characteristics: 10 to 15
ha of land, all the required equipment (plow (s), drill(s), cultivator(s), and two pairs of draft
oxen), a herd of at least ten cattle, is well trained to properly use the equipment, and is self
sufficient in staple food (cereal). In short, the kind of household thought most likely to adopt
promising new technology.








The LP model
Source and type of data
The data for this research originated primarily from an existing long term data base of the
Sikasso farming systems research team (ESPGRN) and from farm surveys conducted in the
Koutiala region in 1996 (ESPGRN and Projet Jachere) and at the N'Goukan village during 1997
and 1998 cropping seasons. Data related to crop yield and dry matter production in the research
area come mainly from the experiments conducted on farmers fields at N'Goukan (1996 to 1998)
and at the N'Tarla research station in 1997 and 1998. Secondary sources such as other IER
programs, CMDT annual reports, and review of literature were used to complement and refine the
collected data.

Production activities
Though we are aware of the fact that farmers in this region undertake many activities in
their production systems, activities are limited here, for simplicity, to crop (cereals and cash
crops) and fodder production. Cotton is solely for sale while the cereals (maize and sorghum) can
be sold and/or used for consumption as staple food. The increasingly important integration of
livestock and agriculture in the region and the acute shortages of feed during the six-to-seven-
month dry season have made fodder production a necessity for households with cattle. Thus,
besides the soil fertility improvement aspects an improved fallow will directly benefit the system
through high quality fodder production.

Constraints/limitations
The Malian Cotton Company (CMDT) provides all the required inputs for cotton on loan
to the farmers at the beginning of the rainy season. The other crops do not benefit from this pre-
financing. Cereal consumption is estimated to be 350 kg person'" year' and it is assumed that
there is equal preference for maize and sorghum in the region. To reduce weight loss during the
dry season, cattle feed should be supplemented by 2 kg fodder head-" day'1 i.e., an estimate of
2000 kg of DM for three months.

Scenarios
Scenario 1.
The farmer pays all expenses for an improved fallow, except for protection against cattle
during the dry season. An estimated 825 gliricidia seedlings at 25 CFA/seedling and 10 kg of
stylosanthes seeds at 1000 CFA/kg are required for one ha of a gliricidia + stylosanthes-improved
fallow. The total amount is 30625 CFA/ha. After meeting the annual fodder requirement of the
farm draft animals (2000 kg DM), the opportunity to sell fodder is given at an estimated price of
10 CFA kg-' (based on ESPGRN Sikasso figures for a similar quality fodder made of dolique and
maize). Dry matter yield of the improved fallow is 4000, 5000, and 7000 kg ha" at ages 2, 3, and
4 years respectively.
Scenario 2.
This scenario introduces a 50% subsidy on the installation cost of the improved fallow but,
at the same time, reduces by half the monetary value of fodder.
Scenario 3.
Total subsidy of the installation cost and no monetary value for fodder.








Results


Scenarios 1 and 2.
In year 1, there was a 12.4% reduction of the cotton area to make room for the improved
fallow. Consequently, the farm has forgone 7.45% of the total annual income as compared with
the basic plan (simulation) to accommodate the new technology. Cash and labor for weeding
cereals and the improved fallow are constraining. The second year there was a reestablishment of
the cotton area to its original level of 5.79 ha and a 58% increase in the area allocated to fodder
banks. This has resulted in a 6% increase in the original farm income. The only scarce resource
was weeding labor for all crops. Year 3 was similar to year 2 in all aspects but the farm income
increased by 7%. At year 4 it was profitable to put the improved fallow into maize. At yields of
3000 kg ha7' (achieved in the region from our on-farm trial) this enterprise was more profitable
than fertilizer-applied maize which was undertaken only after all the land under 4-year old
improved fallow has been farmed. The farm income increased by 12% mainly due to fodder and
higher maize grain yield. Labor for ploughing (all crops) and weeding (cotton, maize, and fodder
bank) was constraining. There was a 1.38 ha of available land which may be planted to fallow if
the cycle was more than four years. The behavior in scenario 2 was similar to the one in the first
scenario except for the fact that the farm income at the end of the year was reduced further
because of the change in the price of fodder.

Scenario 3.
With 100% subsidy, the farmer did not have to reduce either the area allocated to each
crop nor his end of year income which remain identical to the ones in the initial/basic plan. The
subsidy favored the farming of more land than in the basic plan (14.6 ha vs 13.4 ha). This means
that the remaining land (1.20 ha) converted into improved fallow would have been left as a
traditional fallow or used otherwise with low or no return. In such a situation, the improved
fallow represents a much better alternative especially because it can improve crop yield and
increase the family income. However, because fodder has no monetary value, the fodder bank was
dropped out of the system after year 2 when the improved fallow started supplying all of the
required fodder, freeing more time and land for other activities.

Conclusions
The model revealed that improved fallow would be an interesting venture only if fodder is
valued and if maize yields of at least equal to or higher than the regional average yield of 2500 kg
ha1' can be achieved. From our on-farm trial, fallows planted to the association Gliricidia sepium
/Stylosanthes hamata achieved a 3000 kg ha"1 grain yield after three growing seasons of age,
maybe because P deficiency was prevented by a blanket application of 300 kg ha"' of the Tilimsi
rock phosphate (PNT) during seed bed preparation at year 1. Allowing for less than ideal weather
conditions and poor crop nutrition, a four year duration is suggested for improved fallows in the
Koutiala region.
Improved fallows are not financially attractive to farmers if they do not produce benefits
other than soil fertility improvement measured in terms of crop yield. Because of the strong
animal component of the region's farming systems, fodder is a real constraint, especially during
the dry season. Thus, any subsidy program which would prevent farmers from cutting the fodder,








as secondary output before the end of the planned fallow length, would not have adoption
potentials. Cash being a real constraint in such systems, a special fallow installation loan program,
similar to the one that cotton enjoys, would make the venture viable and attractive to farmers.

References
Batiano, A. and Mokwunye, A.U. 1991. Role of manures and crop residues in alleviating soil
fertility constraints to crop production. Fertilizer Research 29:117-125.

Coulibaly, O. Vitale, J.D., and Sanders, J.H. 1998. Expected effects of devaluation on cereal
production in the sudanian region of Mali. Agricultural System 54(4):489-503.

Kleene, P., Sanogo, B., and Viestra, G. 1989. A partir de Fonsebougou: presentation, objectifs et
methodologie du "Volet Fonsebougou" (1977-1987), Systemes de production rurale au Mali.
Volume 1, IER/Bamako/Mali, KIT/Amsterdam, Pays-Bas.

Kwesiga, F and Coe, R. 1994. The effect of short rotation Sesbania sesban planted fallows on
maize yield. For. Ecol. Man. 64:199-208.

Mafongoya, P.L. and Nair, P.K.R. 1997. Multipurpose tree prunings as a source of nitrogen to
maize under semiarid conditions in Zimbabwe. Part 1. Nitrogen-recovery rates as influenced by
pruning quality and methods of application. Agroforestry Systems 35:31-46.

Mugendi, D.N., Nair, P.K., 1997. Predicting the decomposition pattern of tree biomass in tropical
highland microregions of Kenya. Agroforestry Systems 35: 187-201.

Mittal, S.P., Grewal, S.S., Agnihotri, Y., Sud, A.D., 1992. Substitution of nitrogen requirement
of maize through leaf biomass ofLeucaena leucocephala: agronomic and economic
considerations. Agroforestry Systems 19: 207-216.

Palm, C.A. and Sanchez, P.A. 1991. Nitrogen release from the leaves of some tropical legumes
as affected by their lignin and polyphenolic contents. SoilBiol. andBiohem. 23:83-88.