• TABLE OF CONTENTS
HIDE
 Copyright
 Title Page
 Introduction
 The farm production process: Some...
 The linear programming models
 The study area and elicitation...
 Results
 Conclusion
 Reference






Group Title: International working paper series ;, IW92-12
Title: Effects of resource constraints on the production systems and earnings potential of small farm households in the West Province of Cameroon
CITATION PAGE IMAGE ZOOMABLE PAGE TEXT
Full Citation
STANDARD VIEW MARC VIEW
Permanent Link: http://ufdc.ufl.edu/UF00056200/00001
 Material Information
Title: Effects of resource constraints on the production systems and earnings potential of small farm households in the West Province of Cameroon
Series Title: International working paper series
Physical Description: 54 p. : ; 28 cm.
Language: English
Creator: Stall, Steven J
Davis, C. G ( Carlton George ), 1936-
University of Florida -- Food and Resource Economics Dept
Publisher: Food and Resource Economics Dept., Institute of Food and Agricultural Sciences, University of Florida
Place of Publication: Gainesville Fla
Publication Date: [1992]
 Subjects
Subject: Farms, Small -- Cameroon -- West Province   ( lcsh )
Agriculture -- Economic aspects -- Cameroon -- West Province   ( lcsh )
Genre: government publication (state, provincial, terriorial, dependent)   ( marcgt )
bibliography   ( marcgt )
non-fiction   ( marcgt )
 Notes
Bibliography: Includes bibliographical references (p. 53-54).
Statement of Responsibility: by Steven J. Stall and Carlton G. Davis.
General Note: Title from cover.
General Note: "July 1992."
Funding: Electronic resources created as part of a prototype UF Institutional Repository and Faculty Papers project by the University of Florida.
 Record Information
Bibliographic ID: UF00056200
Volume ID: VID00001
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: aleph - 001761704
oclc - 27157308
notis - AJH4824

Table of Contents
    Copyright
        Copyright
    Title Page
        Title Page
    Introduction
        Page 1
    The farm production process: Some conceptual dimensions
        Page 2
        Page 3
        A linear programming model: Rationalization and specification
            Page 4
            Page 5
            Page 6
            Page 7
    The linear programming models
        Page 8
        Model 1
            Page 9
            Page 10
            Page 11
            Page 12
            Page 13
        Model 2
            Page 14
            Page 15
    The study area and elicitation procedures
        Page 16
        Study Area
            Page 16
            Page 17
            Page 18
        Data elicitation procedures
            Page 19
            Page 20
            Page 21
            Page 22
            Page 23
            Page 24
            Page 25
            Page 26
    Results
        Page 27
        Land
            Page 27
            Page 28
            Page 29
            Page 30
            Page 31
            Page 32
            Page 33
            Page 34
        Labour
            Page 35
        Capital
            Page 36
            Page 37
            Page 38
        Farm earnings
            Page 39
            Page 40
            Page 41
        Crop land production mix and earnings potential
            Page 42
            Page 43
            Page 44
            Page 45
            Page 46
            Page 47
            Page 48
    Conclusion
        Page 49
        Page 50
        Page 51
        Page 52
    Reference
        Page 53
        Page 54
Full Text





HISTORIC NOTE


The publications in this collection do
not reflect current scientific knowledge
or recommendations. These texts
represent the historic publishing
record of the Institute for Food and
Agricultural Sciences and should be
used only to trace the historic work of
the Institute and its staff. Current IFAS
research may be found on the
Electronic Data Information Source
(EDIS)

site maintained by the Florida
Cooperative Extension Service.






Copyright 2005, Board of Trustees, University
of Florida





/0, 51


IW92-12


EFFECTS OF RESOURCE CONSTRAINTS ON
THE PRODUCTION SYSTEMS AND EARNINGS
POTENTIAL OF SMALL FARM HOUSEHOLDS
IN THE WEST PROVINCE OF CAMEROON


by

Steven J. Stall and Carlton G. Davis

IW92-12 July 1992


INTERNATIONAL WORKING PAPER SERIES


FOOD AND RESOURCE ECONOMICS DEPARTMENT
Institute of Food and Agricultural Sciences
University of Florida
Gainesville, Florida 32611








EFFECTS OF RESOURCE CONSTRAINTS ON THE PRODUCTION SYSTEMS

AND EARNINGS POTENTIAL OF SMALL FARM HOUSEHOLDS IN THE

WEST PROVINCE OF CAMEROON


Steven J. Stall and Carlton G. Davis*
(Department of Food and Resource Economics,
University of Florida, Gainesville, Florida 32611, USA)


INTRODUCTION


In the years following independence in 1961, Cameroon has

experienced an average annual economic growth rate of 3.6 per

cent, which is almost three times higher than the Sub-Saharan

African nations' average of 1.0 per cent over the same period

(World Bank, 1987). The country's agricultural sector grew at an

average annual rate of 4.2 per cent over the 1965-80 period. This

was considerably better than the average growth rate of 1.9 per

cent of Sub-Saharan African countries over the same period. The

growth rate of the agricultural sector has slowed substantially

over the 1980-85 period, to about 1.3 per cent per annum. However,

even at the reduced rate, the country's agricultural sector growth

in the latter period was in excess of the Sub-Saharan African

average of 0.9 per cent (World Bank, 1987). For the most part,

Cameroon is self-sufficient in supplying its food needs. In

1984-85 only 2.6 per cent of the food consumed locally originated

from net imports (the difference between exports and imports), and

these imports were mainly cereals such as rice and wheat flour

(Ministry of the Plan and Regional Development, 1986).

*The authors extend thanks without implications to Max
Langham, Francois Kamajou and Bernard Guietang for their assistance
with study.









The West Province of Cameroon is the smallest of the

country's ten provinces, but is the third largest in population.

The province has the highest population density of any of the

provinces. For example, the population densities were 84.2, 95.7

and 107.7 persons per square kilometer in 1981, 1986 and 1991,

respectively. In contrast, the population densities for the

country as a whole were 18.9, 22.0 and 25.8 persons per square

kilometer in 1981, 1986 and 1991, respectively (Ministry of the

Plan and Regional Development, 1986). The Province is also an

important producer of agricultural commodities. Agriculture

accounted for over 65 per cent of the province's GDP in 1977,

which was more than twice the contribution of agriculture to the

national GDP (Nji, 1982). The province is a primary producer of

food crops and the main region for the production of Arabica

coffee. In addition, Robusta coffee, cocoa and tobacco are

produced in sizeable quantities. Given the land and other

agricultural resource constraints associated with rural population

pressures in Cameroon's West Province (and other regions), the

purpose of this paper is to report on the results of a study to

assess the effects of land and other resource constraints on the

agricultural production systems and income generation potentials

of small farm households in a sub-region of the West Province.


THE FARM PRODUCTION PROCESS: SOME CONCEPTUAL DIMENSIONS


In assessing the impact of land and other resource

constraints on Cameroon small farm systems, it is inter alia,

necessary to develop: (1) an understanding of the

characteristicsof the productive process and (2) identify and

2








measure the key factors impacting the process. Economic theory and

empirical studies provide the contextual framework for approaching

the assessment. In other words, the behavioral assumptions

underlying the form and the function of the production systems of

small farmers are rooted in the tenets of traditional economic

theory and the associated body of empirical findings relating to

observable economic phenomena. Based on these tenets, we

explicitly assume the following objectives as being typical of the

small farmer population: (1) maximization of net returns (2)

attainment of a reasonable level of self-sufficiency in household

food consumption and (3) minimization of risk in production

activities.

The choice of an appropriate empirical model for assessing

the behavioral characteristics of an economic system is

subjective. As such, the choice process needs to be informed by a

set of criteria. In the case of this study there are two

overriding criteria for the selection of the model used. These

are: (1) adaptability to the compositional theoretical framework

which is providing guidance to the behavioral assumptions and

characteristics of the system and (2) "appropriateness" in meeting

the general objective of the study. In this case, the objective is

to evaluate the impact of land and other resource constraints on

small farmer agricultural production systems and their income

generation potential. A number of studies of this kind have used

a production function approach in which regression analysis is

used to determine the coefficients under which small farmers









operate (Wolgin, 1975; De Boer and Chandra, 1978; Dillon and

Anderson, 1971; Yotopoulos, 1967; Hopper, 1965). These types of

studies have been primarily directed towards considerations of

risk, allocative efficiency and the nature of constraints on farm

production systems. Barnett et al (1982) used a linear

programming (LP) model to test various objectives of Senegalese

farmers. One study that is most closely related to our study in

terms of objective is that by Heyer (1971) of small farmers in

Kenya. Heyer used a linear programming model to estimate shadow

prices for resources, thereby estimating the effective constraint

of resources on the production process.

A Linear Programming Model: Rationalization and Specification

Based on the model selection criteria discussed above, a

linear programming model was selected as being the most suitable

empirical procedure for evaluation of the system. One advantage of

a linear programming model within the context of traditional

agricultural production system such as Cameroon's is that it

avoids the question of allocative efficiency (Heyer, 1971).

Although an LP model arrives at a solution by maximizing the model

objectives, it avoids the assumption that technical maximization

has already occurred (Langham, 1968). Instead, the existing

(although possibly sub-optimal) allocative efficiency rates are

determined and used in the model. The model solution then points

to activities and levels of allocation that should be pursued if

the objectives of the model are desirable.








Further, given the objectives of the LP model, the shadow

prices of the resources are a measure of the constraining effect

of those resources. The degree to which the objectives and

constraints of an LP model match those of producers in the farm

production system being represented, is a function of the

specification of the actual model. Although an LP model must of

necessity maximize or minimize some objective function, the model

readily lends itself to attain other objectives through the use of

the constraints (Langham, 1968). Empirical studies by Barnett et

al (1982) indicate that the inclusion of other objectives besides

profit maximization may not improve the performance of the model.

In their study of Senegalese small farmers, a multiple objective

model did not exhibit superiority over a profit maximization model

with similarly-structured constraints. Also, a deterministic model

requires fewer observations to structure the model, which in the

context of the data limitation problems of developing countries,

could be an important consideration. Finally, within a resource

allocation LP model, the dual provides shadow prices or relative

measures of the resource constraints of their associated

resources. In this way such a model can show which constraints

are the most seriously binding and which merit the most research.

The general algebraic notion, the primal form of the LP model

specification,explicitly assumes the objectives of profit

maximization, food self-sufficiency and risk minimization. The

model is:








Maximize:


5 15
E PZ, + E Py (1)
i1 k-1




Subject to:

Land constraints.


x,j s d, (s=1,..4) (2)
J.1


Labor constraints.

4 8 5
_E B z, : b. (n=1,.4) (3)
-=1 j=1 1-4


Purchased input identities.

4 8
Zi E AOE = 0 (i=1,..3) (4)
S-1 J-1


Consumption requirement constraints.


E E c ,,X Yk a hk (k=1,..15) (5)
0=1 j-1


Xsj, Yk, Zi, Vk 2 0 for all s,k,i,j.

Indices:

i) = 1,..3 indexes purchased inputs, while i = 4,5 are
male and female purchased labour

k) = 1,..15 indexes the crops produced by the household.
k = 1,..13 are crops which are considered annually.
k = 14,15 are crops which are considered monthly.

s) = 1,..4 indexes the land types available to household.
















Variables

Pi

Zi

Pk

Yk

Xsj
ds

Bsjn


Zin
bn

Asji


Csjk

hk


S 1,..8 indexes the activities, or crop mixes, for
which land can be used. j = 1,..5 are activities in
the first season, while j = 6,..8 are activities in
the second planting season.

= 1,..4 indexes the types of labour required.


and coefficients:

= market price of the ith input

= quantity of the ith input purchased

= market price per bucket of the kth crop

= quantity sold of the kth crop, in buckets

= hectares of land of type s devoted to activity j

= hectares of land of type s available

= hours of nth labour type required for one hectare of
land of type s devoted to activity j

= hours of purchased labour of type n

= hours of nth labour type available

= quantity of ith input required for one hectare of
land of type s devoted to activity j

= quantity of kth crop produced by one hectare of land
of type s devoted to activity j in buckets

S minimum quantity of kth crop required for household
consumption, in buckets.


In the model, equation (1) represents the maximization of net

returns to resources by the household. Net return is defined as the

gross returns from farm products sold less the cost of inputs

purchased. The amount of product sold is the amount produced less

that reserved for home consumption. Equation (2) is the constraint

allocating the household land resources to the various production

alternatives or crop mixes. Equation (3) is the constraint
7









allocating household labour, borrowed labour and hired labour, to

the production alternatives. Equation (4) is condition guiding the

allocation of purchased inputs to the production alternatives and

equation (5) represents transfer constraint equations allocating

the production of each output to either home consumption or sale.


THE LINEAR PROGRAMMING MODELS


The linear programming model was separately specified for

each of the four household groups. Hyperlindo mathematical

programming computer software was used to solve the model. Two

detailed forms of the LP model was specified and these are

referred to as Model 1 and Model 2. Model 1 has as its operational

objective the determination of land and other resource constraints

on income potential of farmers. Model 2 objective is to assess the

relative value of alternative cropping activities of each group of

farmers and predict their response to factor and product pricing

policies. Each model was solved for a number of variation via a

series of runs. Model 1 had 4 runs: run 1 = measured land values,

without a binding capital constraint; run 2 = varying land

availability constraint, without a binding capital constraint; run

3 = holding land fixed, and varying the capital constraint values;

run 4 = varying both land availability constraint and capital

constraint values. Model 2 had 3 runs: run 1 = measured

coefficients, without a binding capital constraint; run 2 =

varying price received for coffee, without a binding capital

constraint; run 3 = varying price paid for fertilizer, without a








binding capital constraint. The detailed forms of models 1 and 2

are presented below.


Model 1

Model 1 holds the proportion of land allocated to coffee and

foodcrops fixed, based on the assumption that in the short run,

land cannot be converted from one enterprise to the other.

Objective function:

Maximize:

3 7 5 13 7 15
PZ, -i E PZ, + E PY + E E ^+ (IN FC) (6)
t.1 1.1 i4 kt1 t-1 k-14


pit = price of one hour of ith labour during period t

Zit = hours of ith labour purchased during period t

Pkt = price of crop k during period t

Ykt = quantity of crop k sold during period t
IN = average annual amount of income received from other
agricultural sales, such as fruit, chickens, etc.

FC = average fixed cost of annual general agricultural
purchases, including tools and equipment.

The objective function maximizes the revenue from crop sales less

the out of pocket cost of crop production. The time periods

represents the months during the period of the second phase survey

(November 1988 to April 1989) plus a single period representing

the rest of the year. There are therefore 7 time periods.

Constraints:

Land allocation constraints, in hectares:








8(7)
S X, c d, (s= 1,..4)
I=-


There are four land types: 1 = land available for coffee

production; 2 = land available for coffee food crop associated

cropping; 3 = land available for food crops in seasons 1 and 2; 4

= land available for food crop in season 2. Constraint (7) limit

the amount of land of a specific type to the amount of that land

available to the household. The eight cropping activities are: 1

= coffee production with land used throughout the year; 2 = food

crops intercropped with coffee (first season); 3 = beans

associated with potatoes (first season); 4 = corn associated with

food crops (first season); 5 = cabbage (first season); 6 = beans

with potatoes (second season); 7 = cabbage (second season); 8 =

red cabbage (second season).

Labour allocation constraints, in hours:

Type 1 labour constraints (female and large child):

4 a 1
E E b sXj (F7)l, (C17)i, 0 (t=1..7) (8)
=.1 j=2


Type 2 labour constraints (female and children)


E E b2j,, (F7) 1 (C172, -( (2 0 (t=..7 (9)
s1 =1 2








Type 3 labour constraints (all household members)


E E ~a,, ( (C27), (MC3, 0 (t=1,.7) (10)
1 (C27X3, W 31,, 1lO)
=1 j-1 2


Type 4 labour constraints (males only)


E bE b4sX (M)4, 0 (t=1..7) (11)
.-1 j-1


(FT)nt = hours of female labour available to type n labour
activities in period t.

(C1T)nt = hours of large child labour available to type n
labour activities in period t

(C2T)nt = hours of small child labour available to type n
labour activities in period t

(MT)nt = hours of male labour available to type n labour
activities in period t

There are four labour types, differentiated by gender, age

and task: 1 = tasks assigned to female or large child; 2 = tasks

assigned to female, large child or small child; 3 = tasks assigned

to general labour, including male, female or child; 4 = tasks

assigned to males. A large child was considered to be older than

10 years, while a small child was one of 10 years or less. Based

on discussions with local farmers, equivalencies were developed

for adult and child labour inputs. One hour of large child labour

was equivalent to one-half hour of adult labour, and one hour of

small child labour was equivalent to one-third hour of adult

labour. Constraints (8) though (11) allocate available labour to

the 4 labour types.









Female labour constraints.


(12)


E (F),, Z4, f, +fb, (t=1,..7)
-i1


Large child labour constraints.

3
E (C1T),,



Small child labour constraints.

3
E (C2n,,
nn2


Male labour constraints.


4
E (M), Z5, 2 m, + rb,
=.3


= hours

= hours

= hours

= hours

= hours

= hours

= hours

= hours


(13)


(14)


(15)


(t=1,..7)


of household female labour available in period t

of borrowed female labour available in period t

of large child labour available in period t

of small child labour available in period t

of household male labour available in period t

of borrowed male labour available in period t

of purchased female labour in period t

of purchased male labour in period t.


All of the labour constraints (8) through (15) limit the labour

allocation to the respective labour types to that available either

from household members, borrowed, or hired.


ft

fbt

clt

c2t


mbt
mbt

Z4t

Z5t


s cl, (t=1,..7)


s c2, (t=1,..7)








Purchased input identities.
4 8 7
E E ajtj Xj 0o (i=1,..3) (16)
-1i J=1i -i


Purchased input identity (16) allocate purchased inputs to

cropping activities in the production process. The three types of

purchased inputs are: 1 = fertilizer purchased from coffee

cooperatives, assumed to be allocated to coffee production; 2 =

fertilizer purchased on the open market, assumed to be allocated

to foodcrops; 3 = other variable inputs, particularly seed and

gasoline or irrigation pumps.

Household consumption constraints for crops considered annually.

4 8 7
SEE Cj ,Xa Y hk (k=1,..13) (17)
1- j-1 t=1


Household consumption constraints for crops considered monthly.

4 8
E CjkXJ, Yr, hk, (k=14,15) (t=1,..7) (18)
*-1 j-1


Cjkt = buckets1 of crop k produced per hectare of activity
j on land type s in period t

Yk = buckets sold of crop k

Ykt = amount sold of crop k during time period t

hk = household consumption requirement for crop k

hkt = household consumption requirement for crop k during
time period t.




'Units of 15 liter buckets were used to measure production and
sales except in the cases of bananas and plantain, which were
measured in bunches, and corn, which was measured in sacks.

13









Constraints (17) are balance equations which move the

production of crops to sales and home consumption. Constraints

(18) balance the production of crops for which prices vary monthly

to sales and home consumption.

Capital constraint.

3 7 6
E P, Ei E PE pe + FC s1 (19)
-1 s.1 1-4


A capital constraint was introduced to test the effect of the

availability of capital on the system. The terms of the left-hand

side are the expenditure terms from the objective function,

and represent the total out-of-pocket costs, in CFA,2 of

production. The coefficient e is the amount, in CFA, of capital

available to the household.


Model 2

Model 2 allows land to be allocated freely to either coffee

or foodcrop production. Further, households are able to purchase

foodcrops on the local market to satisfy household consumption

needs. This is assumed to free land and labour inputs from

necessarily producing for subsistence, so that the highest level

of income generation can be attained. Thus, the only difference

between Model 1 and Model 2 are the constraints pertaining to land

and foodcrop allocation.



2The CFA is the Central African Franc, a currency unit shared
by many central and west-African nations, particularly those who
were formerly French colonies. The CFA is tied to the French Franc
(FF) at a rate of 50CFA = 1FF. In March 1989 US$1 = 310CFA.

14







Land constraints.


E X, sd (20)
I-1

Xj is hectares of land devoted to activity j.

8 4
S siE X (21)
J-6 I-1

Constraint (20) allows the total household land to be

available to planting by first-season cropmixes (j=l,..5). Coffee

is considered as one of these. Constraint (21) requires that

foodcrop land that is replanted during the second season (j=6,..8)

cannot exceed a certain proportion (g:(0
during the first season.

Household consumption constraints for crops considered annually.


EE cX -Y Wt k hk (k=I...13) (22)
j.1 J.1 t-1


Wk is quantity purchased of crop k.

Household consumption constraints for crops considered monthly.


4E c, XI, Y, + Wk, h (k=14,15), (=1,..7) (23)
s=1 J.1

Wkt is quantity purchased of crop k in period t.

The household consumption constraints are similar to Model 1

(constraints (17) and (18)) except that in this case the purchase

of food commodities to meet household consumption requirements is

permitted. A number of coefficients had to be calculated to









operationalize the linear programming models. These include:(1)

hourly labour wages, (2) input prices, (3) crop prices, (4) other

agricultural income, (5) land availability, (6) labour

availability, (7) input application rates, (8) yield rates, (9)

labour application rates, (10) household consumption requirements,

(11) household expenses. Details of these calculations are

available from the authors.


THE STUDY AREA AND ELICITATION PROCEDURES


Study Area

The data used in the analysis were obtained from the chiefdom

of Bafou in Cameroons' West Province. The chiefdom of Bafou is one

of the largest and most important of the Bamileke chiefdoms.

Bafou is located in the Department of Menoua,3 on the western edge

of the West Province, and is at its closest point, about 10

kilometers (km) from the city of Dschang, the departmental center.

Dschang is also the location of the University Center of Dschang,

the national agricultural university; patterned after the Land

Grant University System of the United States. The Bafou chiefdom

is roughly rectangular shaped, with dimensions 28 kms. long and 8

kms wide, with a surface area of 162 km2. A range of escarpments

divides the terrain into roughly highland and lowland areas. The



3A "Department" is an administrative unit. The West Province
has five administrative units (also called districts) Haut-Nkam,
Menoua, Mifi, Nde and Noun. The provincial capital of Menoua
district is Baffousam, which is the sixth largest city in Cameroon
(86,000 inhabitants). The district of Menoua has about 201,409 of
the West Province's 968,000 population.

16









operationalize the linear programming models. These include:(1)

hourly labour wages, (2) input prices, (3) crop prices, (4) other

agricultural income, (5) land availability, (6) labour

availability, (7) input application rates, (8) yield rates, (9)

labour application rates, (10) household consumption requirements,

(11) household expenses. Details of these calculations are

available from the authors.


THE STUDY AREA AND ELICITATION PROCEDURES


Study Area

The data used in the analysis were obtained from the chiefdom

of Bafou in Cameroons' West Province. The chiefdom of Bafou is one

of the largest and most important of the Bamileke chiefdoms.

Bafou is located in the Department of Menoua,3 on the western edge

of the West Province, and is at its closest point, about 10

kilometers (km) from the city of Dschang, the departmental center.

Dschang is also the location of the University Center of Dschang,

the national agricultural university; patterned after the Land

Grant University System of the United States. The Bafou chiefdom

is roughly rectangular shaped, with dimensions 28 kms. long and 8

kms wide, with a surface area of 162 km2. A range of escarpments

divides the terrain into roughly highland and lowland areas. The



3A "Department" is an administrative unit. The West Province
has five administrative units (also called districts) Haut-Nkam,
Menoua, Mifi, Nde and Noun. The provincial capital of Menoua
district is Baffousam, which is the sixth largest city in Cameroon
(86,000 inhabitants). The district of Menoua has about 201,409 of
the West Province's 968,000 population.

16








highland area covers the northern 10 kms. length of the chiefdom

with elevations ranging from 1600 to 2740 meters. The lowland area

occupy the southern 18 kms. length of the chiefdom, with

elevations ranging from 1400 to just under 1600 meters (Bergeret

et al, 1988). The climate is humid yet moderate, tempered by the

relatively high elevation. Annual rainfall averages 1.9 meters,

which is spread over an 8.5 month period, beginning in early March

and ending around mid-November. The annual variation in rainfall

pattern is considered the least varied in Cameroon. In a study of

the rainfall pattern over a 30 year period, enough rain had fallen

(25mm) to plant by March 31 in 9 years out of 10 (Bergeret et al,

1988).

The Bafou chiefdom is divided into more than 80 quartersr"

neighborhoodsds, each traditionally governed by a sub-chief,

whose authority is delegated by the chief. Favourable climate and

fertile volcanic soils contributed to the development of a

traditional agricultural system of associated cropping of corn,

beans, plantain, bananas and rootcrops. During the colonial

period, the plantation system of agriculture was introduced, with

Arabica and Robusta coffee varieties as the major plantation

crops. Coffee was rapidly adopted by small farmers between 1940

and 1960, such that at the present time coffee in association with

food crops has become the dominant agricultural practice

(Bergeret, et al, 1988). The major food crops grown are corn,

beans, cocoyams, yams, potatoes, cassava (manioc), bananas and

plantain. Also grown, but in smaller quantities are peanuts,








melons, onion, hot peppers and a variety of greens such as

amaranth. There are two cropping seasons, the first beginning with

the rains in March, and the second starting in August or

September. The first rainy season is the most important in the

production process, since the bulk of the annual plantings occur

during this period. Plantings during the second rainy season is

approximately one-quarter of the land planted during the first.

Although there is wide variation in crop mixes among plots, the

following three general cropping mixes are typical of the farming

system.


1) Coffee/Foodcrop Mix:- Coffee is planted at the usual density

of 2500 plants/hectare, or spaced 2 meters apart. Between

these coffee trees are planted any and/or all of the major

crops. However, very little is planted in this type of system

during the second season. Responsibility for cropping

decisions may be divided in cases where the cultivators of

the coffee and foodcrops are different individuals. Chemical

fertilizers are applied to coffee trees and are thus

available to the other crops. Fungicides are also applied to

coffee.

2) Corn Mix:- This is planted exclusively during the first

season, and consists of corn and primarily beans, but often

with some combination of cocoyams, potatoes, cassava and

yams. Chemical fertilizers are sometimes applied.








3) Bean/Potato Mix:- This can be planted during both seasons,

and occasionally includes small amounts of crops such as

cocoyams. Chemical fertilizers are usually applied.


Agricultural roles within the traditional farm households are

generally differentiated along gender lines. The tasks for males

are those associated with coffee, including planting, fertilizing,

spraying, pruning, harvesting, depulping and drying. In the

traditional system, among the food crop enterprises, only

plantains are cultivated and harvested by males. There are

indications, however, that the traditional gender designated roles

are in transition. Some males, particularly younger farmers, do

plant and harvest their own food crops, particularly those

produced for market. Also, women own and tend their own coffee

plots. Women perform the bulk of agricultural labour, including

participating in all of the coffee tasks, except pruning which is

designated exclusively as a male task. For food crops, land

preparation, planting, cultivating and harvesting are all

performed by women and women process and market the food crops.

Children in the farm household generally participate in most of

these tasks. Young boys will assist women in their tasks, but with

age both boys and girls assume their traditional roles.


Data Elicitation Procedures

Data relating to the production processes of typical small

farm households were elicited in a structured format over an eight

month period from September 1988 to April 1989 in the Bafou









chiefdom. To determine the characteristics of a typical farm,

knowledge of the essential characteristics of each household type

is essential. Sample survey procedures were used to acquire this

type of information. Such survey procedures need not be random.

A selection choice based on judgement will be at least as good and

perhaps better than random choices in determining the typical

characteristics. However, Mellor (1969) agree that a beginning

knowledge of the population to be studied is essential. In keeping

with these considerations a two-phase data elicitation procedure

was utilized to generate information required in the linear

programming model. The first phase consisted of a rapid informal

survey designed to provide: (1) descriptive analytical parameters

of the agricultural system in the study area:(2) an overview of

the general production and resource allocation patterns and (3)

guidance to the nature of the constraints to be faced in

conducting the formal survey. The second phase consisted of a

formal sample survey of small farmer households. This phase was

designed to generate data sets to be utilized in the linear

programming model specified in the empirical procedures.

Phase I Survey

During the first phase informal survey, 24 farm households

were identified through consultation with the area's Agricultural

Extension Posts, as representative of the farming system of the

Bafou chiefdom. At each location, both small and large farm

households were visited. Each household head was visited in plots

for approximately 90 minutes, and attempts were made to talk to at
20








least one female member of the household. The survey was conducted

in an informal manner, and all questions and responses were

recorded on voice tape. Questions were asked regarding number of

women and children in the household, land available to the

household, land sources, amount of land allocated to coffee and

other crops. Coffee production and land allocation data were

recorded from each farmer's coffee cooperative book. General

information was gathered on the production process, the

agricultural cycle, labour role in agriculture, availability and

use of inputs, markets for and sale of agricultural production,

and land acquisition patterns.

Analysis of the informal survey data clarified a number of

important points. It was made clear that much of the agricultural

production, particularly rootcrop and banana/plantain, was

harvested and sold on a day-to-day basis. Also, labour use was

seasonal, and with the exception of coffee sales records, no

records existed of either labour use or crop yields. Farmers,

however, were able to recall the amounts harvested of crops such

as corn and beans, which generally were harvested at one point in

time, with the exception of those plots that were close to farm

residences. Furthermore, the land holding size data are

approximate, since few farmers know the actual size of their

holdings. Farmers, generally would have a rough idea of how much

land was allocated to coffee, because a coffee cooperative

extension worker would have made such a size determination at some

point. These farmers generally had little idea, however, of the









size of their food plots. All in all, one important conclusion

derived from the first phase informal survey, was that important

elements of the data concerning the production process (and the LP

model) could not be elicited through single interviews of farmers

or even over several visits.


Phase 2 Survey

The approach to the second phase formal survey was informed

by the results of the first phase informal survey. Thus, the

decision was made to select a small number of representative farm

households and conduct repeated visits in order to obtain the

daily data necessary for the estimation of labour allocation and

certain crop production levels. Further, a small sample would

allow all of the households' land holdings to be visited and

measured in order to ascertain an accurate measurement of land and

other factor allocations. It was recognized that a small sample

survey would carry a much smaller probability of capturing

representative households, compared to a statistically large

sample. Budget and other constraints precluded the use of a large

sample survey. It was therefore decided to use the large sample

data base available for the study area, generated by the project

known as "Operation Bafou",4 to select households most closely

matching the typical households described by that data base. The

use of this procedure facilitated the selection of farm

4"Operation Bafou" is a data gathering project that has been
in operation for several years. The project is administered and
conducted by the Interdisciplinary Farming System Research Team
from the University of Dschang.








households, which although not randomly generated, nonetheless

permitted a selection of households that were representative, in

that they matched known parameters of household size, holding

size, and typical agricultural activities.

The second phase or primary survey was implemented via two

separate questionnaires. One questionnaire was developed to record

the daily activities of the household. The other questionnaire was

developed to record data relating to a particular plot of land

cultivated by the household. The two types of questionnaires were

administered by trained interviewers who were proficient in the

local dialect of the Bamileke language. A representative sample of

12 farm households were selected based on the "Operation Bafou"

sample frame. This number of households was determined to be the

largest that could be operationally accommodated, given the

comprehensiveness of the data requirement and other constraints.

However, because of certain data problems, only 10 farm households

are included in the empirical analysis. Selected characteristics

of these 10 farm households are presented in Table 1.

As shown in Table 1, three quarters or neighborhoods

represents the farm production variation found in Bafou, the study

area. The three neighborhoods are Tsinfou, Lefe and Loung.

Tsinfou is located in lowland terrain and is considered an average

quarter for the study area as a whole, with moderate population

density and holding sizes. A total of 3 representative households

were selected from Tsinfou neighbourhood, 1 large size household

and 2 small size households. Lefe is also located in lowland

terrain, where holding sizes are typically small, and population









Table 1.


Selected Characteristics of Representative Sample of
Small Farm Households


Large Size* Lowland Household Group
No. of
No. of No. of Children Total Ha. Av. Ha.
Area Quartier House- Adults in in House- of Coffee of Coffee
(Neighbourhood) holds Households holds land land

Tsinfou 1 4 11 2.62 2.62
Lefe 1 4 11 2.28 2.28
Subtotal 2 8 22 4.90 2.45

- Small Size* Lowland Household Group -

Tsinfou 2 4 5 1.47 0.73
Lefe 2 5 8 1.37 0.69
Subtotal 4 9 13 2.84 0.71

- Large Size Highland Household Group -

Loung 2 8 29 4.49 2.46
Subtotal 2 8 29 4.49 2.46

- Small Size Highland Household Group -
Loung 2 3 7 0.80 0.40
Subtotal 2 3 7 0.80 0.40
GRAND TOTAL 10 28 71 13.03 1.30

Source: Second Phase Formal Survey.

aHousehold size is based on the number of adults in the household
and the amount of land allocated to coffee production. A small
size household is one with 1 or 2 adults and less than 1.00 hectare
of coffee land, while a large size household is one with 4 or more
adults and more than 1.5 hectares of coffee land.



density is very high. A total of 3 households were selected from

Lefe quarter, 1 large size and 2 small size. Loung is located in

highland terrain, and is characterized by low population density,


large size holdings and a high incident of


market garden


(non-traditional cash crop) cultivation. A total of 4 households

were selected from the Loung neighbourhood, 2 large size and 2









small size. Summary characteristics of the 10 sample households

reveal the following aggregates: 28 adults, 71 children, 13.03

hectares of coffee land and 1.30 average hectare of coffee land

(Table 1). Summary characteristics of average size of crop land

allocations and yields are presented in Table 2. These figures are

the basis of the L.P. results.

Enumerators visited the sample households in each quarter

twice each week on designated days of the week. Each of the adult

members of the households were interviewed on each visit, at

which time both types of survey questionnaires were administered.

It was established earlier from the first phase survey that the

standard unit of measurement for marketing most foodcrops in the

area was a 15 litre bucket. Each of the sample household was

provided with a similar bucket and instructed that all information

gathered regarding yields, seed and even fertilizer use, would be

expressed in bucket units. Other common units of measurement used

were the coffee sack and fertilizer sack. Equivalencies in

measurement units were established as 1 coffee sack = 8 buckets,

and 1 fertilizer sack = 4 buckets. In administering the plot

questionnaire, plot measurements were made with a 15 meter length

of rope marked at 1 meter intervals.










Table 2. Summary Data for Representative Small Farm Households
Household Group

Small Large Small Large
Crop Land Type Lowland Lowland Highland Highland

- Average Hectare - -

Coffee land 0.498 1.796 0.342 2.246
Coffee food crop
land 0.586 1.580 0.351 1.873
Food crop land
(Season 1) 0.193 0.831 0.401 3.945
Food crop land
(Season 2) 0.073 0.296 0.114 0.917

Average Yield


- Kg./Hectare -- -

Coffee production 296.7 294.0 491.1 726.2

- -Buckets/Hectare -- -
Food crops -
Coffee mix:

Potato 0.2 1.2 0.0 1.1
Beans 15.2 19.0 1.9 5.9
Corna 186.1 31.5 0.0 38.0

Corn mix -
(1st Season):

Potato 51.9 166.3 133.6 294.0
Beans 20.9 21.8 20.7 5.8
Corn 320.1 71.0 53.5 76.4

Beans-potato mix -
(2nd Season):

Potato 3.6 29.6 126.1 129.5
Beans 16.7 33.4 19.3 19.3


Source: Second Phase Formal Survey

aCorn yield measured in sacks/hectare.









RESULTS


Effects of Resource Constraints on Production Systems and Earnings
Potential

Land

Results from run 1 of Model 1 indicate that the predicted

optimum allocation of land to alternative crop mix generally

matches the actual household land allocation as observed during

the survey. In all cases, all of the available land was allocated

by the model to cultivation. This result, however, does not

especially underscore the constraints imposed by land

availability, since in the model specification, land is the only

effective constraint, since labour, and all other inputs can be

purchased and capital was not binding. Table 3 compares the run 1

allocation of land to actual allocation. The primary difference

between the model allocation and observed values is the amount of

land left fallow. For example, in the case of the first season

corn mix planting, large highland farmers only planted 0.78

hectares in actuality, while the model allocates 3.9 hectares to

the mix. The model naturally allocated all land to production as

long as net returns to labour and other inputs were positive. This

difference could be a reflection of capital constraint which was

assumed to be non-binding in the model. It could also, reflect a

recognition on the part of the farmers to the limitations of land

to sustained annual croppings. Those farmers who still choose to

plant all of their land may simply be experiencing greater land

constraint with respect to their household consumption needs.









RESULTS


Effects of Resource Constraints on Production Systems and Earnings
Potential

Land

Results from run 1 of Model 1 indicate that the predicted

optimum allocation of land to alternative crop mix generally

matches the actual household land allocation as observed during

the survey. In all cases, all of the available land was allocated

by the model to cultivation. This result, however, does not

especially underscore the constraints imposed by land

availability, since in the model specification, land is the only

effective constraint, since labour, and all other inputs can be

purchased and capital was not binding. Table 3 compares the run 1

allocation of land to actual allocation. The primary difference

between the model allocation and observed values is the amount of

land left fallow. For example, in the case of the first season

corn mix planting, large highland farmers only planted 0.78

hectares in actuality, while the model allocates 3.9 hectares to

the mix. The model naturally allocated all land to production as

long as net returns to labour and other inputs were positive. This

difference could be a reflection of capital constraint which was

assumed to be non-binding in the model. It could also, reflect a

recognition on the part of the farmers to the limitations of land

to sustained annual croppings. Those farmers who still choose to

plant all of their land may simply be experiencing greater land

constraint with respect to their household consumption needs.










Recall that Model 1 does not allow for the purchase of foodcrops

and that household consumption requirements were calculated based

on observed production from existing crop mix allocation.


Table 3. LP Model Land Allocation and Actual Land Allocation, All
Household Groups
Household Group

Small Large Small Large
Lowland Lowland Highland Highland

Land Type Model Actual Model Actual Model Actual Model Actual
--------- Hectare-------


Coffee Production Land
Coffee 0.498 0.498 1.796 1.796 0.342 0.342 2.246 2.246
Coffee/Foodcrop Land
Foodcrop intercropped
w/coffee 0.586 0.586 1.580 1.580 0.351 0.337 1.873 1.873
Foodcrop Land, Season 1
Corn mix 0.193 0.188 0.831 0.780 0.391 0.340 3.904 0.778
Bean & potato mix NA* NA 0.000 0.048 0.000 0.047 0.000 0.154
Cabbage NA NA NA NA 0.010 0.014 0.041 0.339
Fallow 0.000 0.005 0.000 0.003 0.000 0.000 0.000 2.674
Foodcrop Land, Season 2
Bean & potato mix 0.070 0.070 0.296 0.287 0.104 0.092 0.597 0.523
Cabbage NA NA NA NA 0.010 0.022 0.253 0.264
Red cabbage NA NA NA NA NA NA 0.067 0.069
Fallow 0.000 0.061

*Not applicable.


The allocation of all available land in the model does

indicate clearly that there are positive returns to agricultural

inputs in general, and that in the context of the existing

constraints, larger land holdings are desirable for all groups.

The model allocation could also reflect a response to yield

uncertainty, whereby a large area is planted to ensure consumption

needs. This is true even at relatively high levels of land

availability (as shown in run 2), where the proportion of hired to








household labour is high, indicating positive returns to hired

labour. The model indicate that the cabbage crop is not a

preferred crop mix in terms of land allocation to each particular

crop mix, given the model constraints. As indicated in Table 3

only the two household groups in the highlands grow cabbage. In

each of these household groups the model allocated only enough

land to cabbage in the first season to meet household consumption

requirement for that crop. Only in the large highland group does

the model's land allocation solution include cabbage (red)

produced for sale in the second season. The foodcrop mix most

favoured by all of the groups was the first season corn mix. In

all household groups, more of the corn mix was in the model

solution than the actual plantings. The bean and potato crop mix

for the first season was not selected by the model for any of the

household groups, yet 3 of the 4 household groups actually planted

this crop mix. In the cases of coffee production land and

coffee/foodcrop land (the same land), only one mix is possible for

each of these land categories. In all household groups, the model

elected to plant all of the land available to these crop mixes,

indicating positive net returns to coffee. Observation of actual

foodcrop land plantings for the first and second seasons indicate

significant reduction in plantings during the second seasons. On

the average, the two highland groups planted about 26 percent of

the first season foodcrop land, while the two lowland groups

planted about 36 percent. This would suggest inter alia, that








highland farmers are less pressured by land and are able to leave

more land in fallow during the second season.

An interesting and informative dimension of the land

constraint question has to do with the relative scarcity of land

among household groups. As indicated earlier, the dual of the

linear programming model produces the dual of the resource

constraints and reflect the rate of change in the objective

function, with respect to each associated constraints. These dual

values are referred to as "shadow prices" or "implied values" for

the limited resources that are causing the constraint. A resource

whose availability poses no constraint to the model would carry an

implied value of zero. The implied values of land can be taken to

reflect the scarcity of land (and the value of the productive

capacity), given the available resources and technology. Thus, the

generation of implied land values for the four household groups

would provide some indication of relative land scarcity among the

groups. In general, it would be expected that households with

smaller land holdings would have higher implied land values, under

similar technology and farming practices, since they would have

the available labour to make use of more land. This condition,

however, would depend on the relative access to other resources,

particularly labour. Survey data indicate that in the highland

area, where land is generally considered to be less scarce, that

the labour/land ratios are smaller than in the lowland area,








reflecting a greater availability of land to labour.5 It would

be expected then that lowland households would reflect higher

implied values for land.

Table 4 shows the calculated implied land values for each

household type under run 1 of Model 1. Note that two types of

implied land values are computed in the table. One set of implied

land values in Table 4 are for coffee land, coffee/foodcrop land,

first season foodcrop land and second season foodcrop land. A

comparison of these values provide non-conclusive results, since

there is no correlation of land value to household size. The

relative values of coffee land, however, match closely the

relative yield rates (Table 2) that were computed for coffee for

each household groups. The coffee yield for both lowland groups is

approximately 295 kg/hectare, while small highland farms produce

491 kg/hectare, and large highland farms 726 kg/hectare. The

implied values of the land used for that purpose rise with

increased yield. As a general rule, implied land values were

higher for first season foodcrop land than for coffee land. This

finding is consistent with the model results reported in Table 3,

where the corn mix was favoured and the relatively high revenues

available from the sale of corn, potatoes and beans. The second set

of implied land values in Table 4 are estimates of total weighted

average implied land values for coffee land and food crop land.



5The ratios of total annual household labour to total land
holding for all households are: small lowland = 8,482 hours/
hectare; large lowland = 5,011 hours/hectare; small highland =
4,213 hours/hectare; large lowland = 3,195 hours/hectare.

31









Table 4. Implied and Weighted Average Implied Land Values, all
Households

Household Group

Lowland Highland

Land Type Small Large Small Large

- Implied Land Values -
(CFA/Hectare)

Coffee land 105,200 128,400 213,100 333,200

Coffee/
Foodcrop land 262,200 77,400 162,400 67,400

Foodcrop land
1st season 370,100 305,400 200,800 379,800

Foodcrop land
2nd season 27,200 72,700 420,100 207,400

- Total Weighted Average -
Implied Land Values
(CFA/Hectare)

Coffee land 367,400 205,790 376,510 400,570

Foodcrop land 379,990 331,270 320,140 427,990

Weighted average 370,520 245,490 345,990 418,050



These weighted average implied land values were calculated using

a modified version of the formula used to compute the implied land

values for the upper portion of the table. The total weighted

average figures are the average value of all land to the farmer

based on coffee and foodcrop holdings. The weighted average

implied land values indicate as expected, that land values are

higher for small farmers, and lower for large land holders, with

the noted exception of large highland farmers.

32








Closer examination of the pattern of implied land values was

conducted through the relaxation of the land constraints, allowing

increasing land availability to each household group, and

calculating implied land values (Model 1, run 2). The land

constraints for each land type were relaxed by equal percentages.

In each household group implied land values were assessed under

conditions of an increase in land availability to a maximum of

300 per cent of existing land holdings. Land holdings beyond that

point were considered unrealistic in light of the requirement for

high proportions of hired labour. Table 5 shows total weighted

average implied land values for all household groups under these

conditions. Note that in Table 5 only land values for which the

basis of the model is changed are presented, so that the

incremental increases shown are not all equal. The implied land

values for the small and large lowland farmers and the small

highland farmers follow roughly the same trend. The results for

the large highland farmers are significantly different. The

patterns exhibited in the total weighted average implied land

values are also reflected in the calculated implied land values

for coffee land and foodcrop land. These results are interpreted

as reflecting a fundamental difference in productivity of inputs,

including land, labour and other factors, and could also reflect

some economies of size. The large highland farmers are probably

operating at a level of technology which is superior to the other










Table 5. Total Weighted
Households


Average Implied Land Values, all


Household Group Land

Total Large Small Large Small
Land Lowland Lowland Highland Highland


Hectares


0.8
0.9
1.0
1.1
1.2
1.3
1.4
1.5
1.6
1.7
1.8
1.9
2.0
2.1
2.4
2.5
2.6
2.9
3.2
3.4
4.0
5.0
5.8
6.1
6.8
7.1
7.4
8.0
8.6
9.0
10.0
11.0


370.5
363.7
363.7
363.7
325.9
325.9
284.4
284.4
259.3
255.7
255.7
233.0
220.0
220.0
NAb
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA


1000 of CFA/Hectare


INFa
INF
INF
INF
INF
INF
INF
INF
INF
INF
INF
INF
INF
INF
INF
INF
245.5
245.5
220.7
195.1
195.1
195.1
149.4
142.0
140.8
138.0
138.0
NA
NA
NA
NA
NA


346.0
346.0
327.0
327.0
327.0
327.0
293.9
263.7
263.7
263.4
260.8
260.8
260.4
259.5
259.5
259.5
199.6
199.6
NAb
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA


aIndicates infeasible solution due to constraints of household
consumption requirements.
bIndicates solutions are no longer applicable because of
unrealistic land availability.


INF
INF
INF
INF
INF
INF
INF
INF
INF
INF
INF
INF
INF
INF
INF
INF
INF
INF
INF
INF
INF
INF
INF
418.1
418.1
418.1
415.3
403.7
384.9
384.9
384.9
362.0








groups. The lowland as well as the small highland farmers face

generally the same land constraints, since their resource base,

factor efficiency, activities productivity and available

alternatives are not markedly different.

Although not presented here, it was observed that implied

land value for coffee the traditional "cash crop", is not markedly

different from that of foodcrop land. Indeed, there is strong

evidence suggesting that the implied value of the coffee land,

particularly for the two small farm household groups (lowland and

highland) comes from the value of foodcrops planted

in association with coffee. Examination of data collating

increasing land availability with implied land value, shows that

in 3 of the 4 household groups, foodcrop land exhibits a higher

implied value. In all four household groups implied land value

declines more slowly when available land is increased.


Labour

Results from run 1 of Model 1 indicate that labour is not

generally a constraint at existing levels of land availability.

Seasonally, however, some labour constraints occur in two of the

household groups. Although the model allows for the hiring of

labour, labour was only hired in two cases: (1) by the small

lowland group during the month of March (94.3 female hours) and

(2) by the large highland group during March (499.9 female hours).

These household hired labour quantities represents 3.6 per cent

and 4.3 per cent of total annual labour for small lowland and

large highland groups, respectively. According to the results, as

35









land is made available to households (run 2, Model 1), labour

becomes increasingly more constraining. Table 6 shows the

relationship of hired labour to total labour under increasing land

availability. In the case of both large highland and small lowland

farmers, the proportion of hired labour increases to approximately

11 percent with a 50 percent increase in available land. All of

these increases come from female labour.6 In the case of large

highland households, although only a small percentage of labour is

hired at existing levels of land availability, that percentage

represents a large quantity of labour demanded from the local

labour market. Local labour supply might not be available to meet

that demand, or capital constraint might not permit the hiring of

that quantity of labour. In either case, the net effect could be

less than optimal allocation of land. This finding adds further

insights into the earlier discussion regarding the relatively

lower proportion of cropland planted to foodstuffs by the highland

household group (26 percent) during the second season, compared to

lowland household groups (36 percent).


Capital

Capital constraints may have significant impacts on the

production systems. In all of the model results discussed so far,

a solution was feasible as long as revenues exceed the out-of-

pocket costs of production. In order to assess the impact of



6Model solutions include hired male labour after a 90 percent
increase in land availability in the case of large highland farm
households, and after a 120 percent increase for small lowland
households.









Table 6. Hired Labour as Percentage of Total Labour with
Percentage Increase in Land Available, all Households

Household Group
Increase in
Total Land Small Large Small Large
Available Lowland Lowland Highland Highland

- Percent - -

00 3.6 0.0 0.0 4.3
10 5.2 0.0 0.0 5.8
20 6.5 0.7 0.0 6.6
30 7.9 2.0 0.5 6.9
40 9.1 4.1 1.4 8.5
50 10.9 5.9 2.2 11.0
60 13.0 7.6 2.3 13.3
70 14.9 9.1 3.5 15.5
80 16.6 10.5 4.3 17.8
90 19.0 11.7 5.6 21.1
100 21.8 12.9 7.7 24.3
150 36.1 22.9 21.3 38.0
200 51.5 34.0 32.4 46.5



capital limitations the model was run over a range of capital

constraint values (run 3, Model 1), and the calculated implied

value of capital plotted against capital availability. Two data

sets are generated, one reflecting actual land constraints and the

other reflecting available land equal to 150 per cent of actual

land holdings. The results begin at the minimum amount of capital

sufficient to arrive at a feasible solution, and continue until

the implied value of capital goes to zero, in which case capital

is no longer a binding constraint. In all cases, the implied value

of capital is very large at the point of minimum feasibility. The

largest implied value occurs for large highland households, where

at a minimum feasible amount of capital (227,000 CFA per

household), the implied value is more than 14 times the actual
37









capital. This reflect the relatively high productive potential of

these farmers. In all four household groups, however, the implied

value of capital is initially large and remain positive for some

time, as capital availability increases. With increased capital

availability, at 150 percent of actual land holdings, the implied

value is initially generally greater, and remains positive over an

even wider range. Capital is therefore highly valued by the

production system at levels close to minimum amount feasible. As

such, a lack of capital availability would severely constrain the

income generating potential of the production system.

Capital constraint effects at different levels of land

availability are examined for a range of land holding levels (run

4, Model 1). Results indicate that the minimum capital requirement

is the same for any level of land holding for each household

group, reflecting the minimum requirement of meeting household

consumption demands. With the exception of the large highland

household group, the results show that at the smallest level of

land holding (actual), the capital requirement range is quite

small. We further computed the ratio of capital requirement to

land used at the point where the implied value of capital goes to

zero. Results indicate no markedly different ratios, with the

least amount of capital required per hectare found among the large

highland farmers. Similarities of these capital/hectare ratios,

combined with the proportion of land left fallow by the large

highland farmers, suggest differences between these farmers and








the other household groups in terms of adequacy of land holdings

size, with respect to household consumption requirements.


Farm Earnings

Since off-farm employment activity was negligible in all four

household groups, crop sales were the only means of income

generation specified in the model. Coffee sales for each household

group generated in the solution (run 1, Model 1) matches observed

sales. In the case of foodcrop sales, the model solution had much

higher sales than were actually observed, particularly with

respect to corn, beans and potato. This was due to the allocation

of land to those crops in the solution. Table 7 shows a comparison

of the model's prediction of foodcrop sales to actual sales in

terms of percentage of total production. As indicated in the

table, not only are sales of corn, beans and potato higher in the

model solution, but sale of cabbage is generally lower.

For the small highland household group the model produce

only enough cabbage for household consumption. These results are

supported by comparative data of model solution and actual gross

revenues from all crop sales, as a percentage of total revenue

(data not presented here). These data indicate that while coffee

sales in actuality make up the majority of gross revenue in all

household groups, in percent terms, gross revenue from coffee

sales were smaller than those from foodcrop sales in model

solutions.









Table 7. Percent of Crop Production Sold, Model Results and
Observed Values, all Households

Household Group

Small Lowland Large Lowland

Crop Actual Model Actual Model

- - Percent - - -

Banana 22.2 22.8 15.6 18.1
Beans 0.0 9.2 12.4 47.8
Cabbage NAa NA NA NA
Cocoyam 0.0 0.5 0.4 2.3
Corn 3.6 89.5 29.6 49.3
Manioc 0.0 0.0 62.9 62.2
Plantain 38.9 39.1 8.3 12.0
Potato 0.0 49.0 9.2 68.7
Yam 2.4 3.1 0.0 3.3


Household Group

Small Highland Large Highland

Actual Model Actual Model


- - Percent - - -
Banana 0.0 0.0 2.8 12.6
Beans 0.0 77.9 3.7 45.4
Cabbage 70.7 0.0 90.7 85.3
Cocoyam 0.0 0.3 5.1 23.2
Corn 0.0 28.2 31.7 83.2
Manioc 0.0 0.0 0.0 2.0
Plantain 0.0 0.0 3.0 21.2
Potato 27.4 66.6 27.5 86.3
Red cabbage NA NA 100.0 100.0
Yam 0.0 0.0 0.0 39.2


aNot applicable.










Table 8 shows a comparison between the objective function

value (earnings) result and estimated average annual family

expenses. Of the four household groups, only in the case of the

large highland group was the net returns from farming expressed

in the objective function, enough to meet the estimated living

expenses of the household. Although it is likely that some income

was unreported, the consistency and magnitudes of these outcomes

are consistent with those of Fouda (1988) in her study of farm

households in the Western Province. She found that the average

household's net revenue of 1,106,014 CFA did not cover family

living expenses (1,222,447 CFA). The implication being that the

available resource base, particularly land, is not large enough to

support the households. In the case of the large highland

households, it would appear that the level of land holding is

sufficiently large to provide for the household's need. However,

in light of the relatively large amount of land actually devoted

to fallow by this household group, it is expected that their

income will not match the income level in the objective function

of the model solution. In fact, as shown in Table 8 their reported

gross revenue of 1,133,201 CFA does not meet their estimated

annual living expenses (1,450,523 CFA).









Table 8. Comparison of Objective Function Values and Gross Annual
Revenues and Annual Living Expenses

Objective Average Annual Average Annual
Function Gross Revenues Living Expenses
Value
Household Group (Earnings)
- - CFA- -- - -

Small Lowland 220,299 73,182 414,915
Large Lowland 245,969 94,395 1,231,250
Small Highland 151,268 111,171 219,375
Large Highland 2,154,097 1,133,201 1,450,523



Crop Land Production Mix and Earnings Potential

Under specifications of Model 2, complete flexibility of land

use is permitted, as well as purchase of food for household

consumption. Run 1 of Model 2 was used to examine the long run

optimality in land allocation to alternative crop mix. Table 9

shows the crop land allocation results as predicted by the model.

Of some note is the finding that during the first season, the only

period when non-irrigated cultivation of cabbage is possible, no

cabbage crop was planted by either of the highland household

groups. This result contrasts strongly with the tendency of most

highland farmers to allocate land to cabbage during this season,

as indicated earlier in Table 7. During the survey, farmers

complained of lower prices received for cabbage. The model results

indicate that at least for the first season, prices are such that

cabbage production might be unattractive. The location of Douala

and Yaounde (the two largest and rapidly growing urban areas),

adjacent to the West Province, might be creating transitional

trends in the demand for cabbage. There is evidence that the









supply of cabbage has increased significantly as farmers seek

alternatives to the perceived low returns to coffee. Nevertheless,

the rate of growth in this demand might not be significant. All

the results indicate, however, that of the various food crops,

corn and to some extent potatoes, generate large positive returns.

Under Model 2, the solution for both of the large household groups

included exclusively, the first season corn cropping mix.


Table 9. Optimal Land Allocation, Model 2, all Households


Cropmix


Coffee

Season 1
Foodcrops intercropped
with w/coffee
Cornmix
Bean and potato mix
Cabbage

Season 2
Bean and potato mix
Cabbage
Red cabbage


Household Group

Highland Lowland

Large Small Large Small

- - Hectares - -

0.00 0.75 0.00 0.58


0.00
6.19
0.00
0.00


1.54
0.25
0.07


0.75
0.00
0.00
0.00


0.00
0.00
NA


0.00
2.63
0.00
NA


0.78
NA
NA


0.58
0.20
0.00
NAa


0.07
NA
NA


aNot applicable.


The optimality of land allocation to coffee production shows

mixed results from Model 2. Large household groups chose not to

plant coffee, even though there are indications that the rate of

yield is higher among the large highland farms. This tendency can

be understood in terms of the weighted average implied values for









coffee land and foodcrop land shown earlier in Table 4. Recall

that in that table, the weighted average value of foodcrop land is

greater than that of coffee land in both of the large household

groups. Also, the value of both cropland types are virtually the

same for the two lowland household groups, suggesting that there

might be a high degree of sensitivity to coffee prices. As such,

reduction in the price of coffee could quickly shift the basis of

the solution to reflect greater allocation of land to foodcrop

production.

Results of run 2 of Model 2 show the effects of changes in

coffee price on optimal land allocation. Table 10 shows the

results for the small and large highland household groups and

Table 11 shows the results for the small and large lowland groups.

Results are shown for only those prices at which a change in the

basis of the model occur. Since at the existing prices, neither of

the large household models chose to plant coffee, the results were

determined for rising coffee prices. In the large highland group,

a moderate 16 per cent increase in coffee price (550 CFA/kg)

resulted in the model allocating land to coffee. However, a 68 per

cent increase in coffee prices (800 CFA/kg) was required to induce

the model to abandon food entirely and produce only coffee. This

is further indication of the attractiveness of food crop in the

production system. It has been estimated that the implicit tax on

Arabica coffee production was about 71 percent at the time of the

study. The 68 percentage required increase in coffee prices is

thus less than the implicit tax, so even if farmers received the









Table 10. Optimal Land Allocation Under Changing Coffee Prices,
Highland Household Groups

Crop Mix Price of Coffee

Small Highland Household Group

CFA/Kg

325 375 400 475a


- Hectares - -

Coffee 0.00 0.34 0.61 0.75

First Season
Foodcrops
with coffee 0.00 0.34 0.61 0.75
Corn mix 0.75 0.41 0.15 0.00
Cabbage 0.00 0.00 0.00 0.00

Second Season
Beans & potatoes 0.23 0.12 0.05 0.00
Cabbage 0.00 0.00 0.00 0.00


Large Highland Household Group

CFA/Kg

475 550 600 625 650 800

- Hectares - -

Coffee 0.0 2.9 3.0 3.1 4.3 6.2

First Season
Foodcrops
with coffee 0.0 2.9 3.0 3.1 4.3 0.0
Corn mix 6.2 3.3 3.2 3.1 1.9 0.0
Cabbage 0.0 0.0 0.0 0.0 0.0 0.0

Second Season
Beans & potatoes 1.5 0.8 0.9 0.9 0.5 0.0
Cabbage 0.3 0.1 0.0 0.0 0.0 0.0
Red cabbage 0.1 0.1 0.1 0.1 0.1 0.0


aPrice of coffee at time of study.








world market price for coffee, foodcrop production would still be

an attractive alternative.

In the large lowland household group the optimal model

solution also did not include coffee planting at the existing

price level. It took an 89 per cent increase in the price of

coffee (900 CFA/kg) to induce the model solution to allocate land

to coffee production (Table 11). However, even at that price

level, the model solution included some corn cropping mix. Since

food purchasing is permitted in the solution, the corn cropping

mix is most likely for sale, rather than household consumption. In

the case of both small household groups the optimal model solution

included coffee production at the existing price level (475

CFA/kg). However, in the small highland household group, a 21 per

cent decrease in coffee price to 375 CFA/kg, leads to a decrease

of more than one-half of the land allocated to coffee production,

and 31 per cent decrease leads to a complete shift away from

coffee. In the case of the small lowland household group, a 16 per

cent decrease in coffee price to 400 CFA/kg leads to a decrease in

coffee land allocation by more than one-half. The general results

of the optimal model solution indicates that under even moderate

decreases in coffee prices, the tendency is to shift away from

coffee production to foodcrops. Thus, at prices existing at the

time of the study, coffee production appears to be at best

marginally attractive to some of the household groups.









Table 11. Optimal Land Allocation Under Changing Coffee Prices,
Lowland Household Groups

Crop Mix Price of Coffee

Small Lowland Household Group

CFA/Kg

100 250 400 450 500 550 800


- - Hectares - -

Coffee 0.0 0.07 0.23 0.53 0.58 0.68 0.78

First Season
Foodcrops
with coffee 0.07 0.07 0.23 0.53 0.58 0.68 0.78
Corn mix 0.71 0.71 0.55 0.25 0.20 0.09 0.00

Second Season
Beans &
potatoes 0.26 0.26 0.20 0.09 0.07 0.04 0.00



Large Lowland Household Group

CFA/Kg

475a 900 1000


- Hectares - -

Coffee 0.00 2.21 2.63

First Season
Foodcrops
with coffee 0.00 2.21 2.63
Corn mix 2.63 0.41 0.00
Cabbage 0.00 0.00 0.00

Second Season
Beans & potatoes 0.79 0.12 0.00


aPrice of coffee received by farmer at time of study.









Model 2 also generated an optimal solution for cropping mix,

over a range of fertilizer prices (run 3). The results indicate

some resistance in optimal land allocation to changes in the price

of fertilizer. In the small highland and large lowland groups, an

increase in fertilizer price by more than 100 per cent effected no

change in the model's basis. In the large highland household

group, the rise in fertilizer price led to a shift away from

cabbage in the second season, but this reallocation represented a

small portion of total land. In the small lowland group, a rise in

fertilizer price7 to 4500 CFA/sack for both coffee and foodcrops

led to a significant shift away from coffee production and towards

the corn cropping mix. The general implication is that, assuming

that the removal of subsidies increases the price of fertilizer to

the farmer, the implementation of such a policy would have minimal

differential effect on cropping patterns. The exception might be

cabbage production, which like other non-traditional cash crops,

heavy fertilizer use would put them at a disadvantage when

fertilizer prices rise. The small lowland household group model

gave results suggesting that an increase in the price of

fertilizer could have an additional negative impact on coffee

production. The model indicate that in the other three household

groups, the effect of capital is neutral with respect to

fertilizer price increases. Thus, in terms of the small lowland

households, fertilizer price increase in combination with capital

7Fertilizer prices at the time of the study were measured at
an average of 3500 CFA/sack when purchased on the open market, and
2500 CFA/sack when obtained from the coffee cooperative.

48








constraint, could make the outlook for coffee production even less

attractive.


CONCLUSIONS


The farmers of the Bafou study area can be categorized into

four representative farm household groups: (1) small lowland

households, (2) large lowland households, (3) small highland

households and (4) large highland households. The households can

be further categorized into two groups with respect to the effects

of land constraints. Group 1 consists of both lowland household

groups and the small highland household group. The implied values

of land to the households in Group 1, show that smaller land

holders generally face higher land values than do larger land

holders. That is, other things being equal, additional land is of

more value to those with less to begin with. As such, the effects

of land constraints on income generation are more severe for those

farmers with smaller holdings. However, all members of Group 1

share similar levels of productivity and efficiency, in terms of

rates of yield of labour and other factor use. Also, results

indicate that for all members of this group, returns from

agricultural production are not enough to meet household living

expenses. The conclusion is that for this group of householders,

their resource base is too small to support the household size.

Group 2, the large highland farmers, display greater resource

productivity and allocative efficiency than Group 1 farmers,

resulting in slightly higher implied land values. Thus, due to









simply technical differences, land scarcity has a greater impact

on the income generation of Group 2 households, inspite of the

relatively large size of their holdings. Further, Group 2 farmers

have enough land to be able to generate enough income to meet

household living expenses, and they do not allocate all of their

land to productive activities. All households have a strong need

for capital, and the lack of access to capital might be as large

factor in their behaviour as are land constraints. Under existing

levels of farm technology, most households, even most large

households, do not have enough land to generate sufficient income

to meet household needs. As population increase, so will

population densities. Unless off-farm employment opportunities

become available, holding sizes will become even less adequate to

sustain households.

In terms of optimal crop land allocation, the results

indicate that the future for the continued widespread cultivation

of Arabica coffee in the West Province is at risk. At existing

prices, two of the four household optimization models selected

coffee over foodcrop production. However, each of the four

household group models showed sensitivity to coffee prices. As

such, even moderate reduction in coffee prices caused the profit

maximization model solution to indicate a near complete shift from

coffee to foodcrop production. Any shift away from coffee

production may not necessarily take the form of clear-cutting

coffee plants. Farmers were observed to be increasing the spacing








between plants when replanting their coffee, and thereby

increasing the area available for alley-cropping with cabbage,

beans and other foodcrops in association with the coffee.

Cabbage (including red cabbage) is the most widely planted

non-traditional cash crop. Recent years have witnessed its

increasing cultivation, as farmers search for alternatives to

coffee production. Results indicate, however, that cabbage as a

non-traditional (or market garden) cash crop might not be as

financially attractive alternative as is believed. Of the two

household groups who cultivated cabbage, only one household group

selected the crop in the optimal allocation solution, and the

selection occurred only during the second season. During the

second (dry) season cabbage must be irrigated, and the resulting

small supply assures higher prices. In the aggregate, however, the

demand side of the market is small and prices have declined in

recent years as supply increased. The implied land value results

of the model indicate that in general, a higher land value exists

for land allocated to foodcrops than for coffee production. It was

found that even on land allocated to coffee, much of the land

value comes from foodcrops planted in association with coffee. In

the optimal land allocation model, two of the four household

groups chose to plant only the corn crop mix during the first

planting season, consisting of corn, beans and potatoes. In one of

these household groups (the large highland farmers), the corn crop

mix was selected inspite of the highest coffee yield rate. In this

household group,nearly all land is reallocated to the corn, bean









and potatoes crop mix when there is a moderate decline in coffee

prices.








REFERENCES


Barnett,Douglas, Brian Blake and Bruce McCarl, 1982. "Goal
Programming via Multidimensional Scaling Applied to Senegalese
Subsistence Farms." American Journal of Agricultural
Economics, 64(4): 720-727.

Bergeret, Pascal, Guy Ducret, Isabell Grangeret, M.N. Roux and J.L.
Schafer, 1988. "Le Systeme Agraire de Bafou (Ouest Cameroun)
Chefferie du Pays Bamilike." Dschang: University Center of
Dschang, Operation Bafou Report.

DeBoer, John A. and Satish Chandra, 1978. "A Model of Crop
Selection in Semi-subsistence Agriculture and an Application
to Mixed Agriculture in Fiji." American Journal of
Agricultural Economics, 60(3): 436-444.

Dillon, John L. and Jock R. Anderson, 1971. "Allocative Efficiency,
Traditional Agriculture, and Risk." American Journal of
Agricultural Economics, 53(1): 26-32.

Fouda, Therese M., 1988. "Assessing Small Farmers' Credit Needs: A
Case Study of the West Province of Cameroon." Gainesville:
M.Sc. thesis, Food and Resource Economics Department,
University of Florida.

Hayer, Judith, 1971. "A Linear Programming Analysis of Constraints
on Peasant Farms in Kenya." Food Research Institute Studies,
10(1): 55-67.

Hooper, W. David, 1965. "Allocative Efficiency in Traditional
Indian Agriculture." Journal of Farm Economics, 47(5): 221-
239.

Langham, Max R., 1968. "A Dynamic Linear Programming Model for
Development Planning." In: W.W. McPherson (Editor), Economic
Development in Tropical Agriculture. Gainesville: University
of Florida Press.

Mellor, John, 1969. "The Subsistence Farmer in Traditional
Economies." In: Clifton R. Wharton (Editor), Subsistence
Agriculture and Economic Development. Chicago: Aldine
Publishing Company.

Ministry of the Plan and Regional Development, 1986.Sixth Five-
Year Economic, Social and Cultural Development Plan. 1986-
1991. Yaounde: Republic of Cameroon.

Nji, Ajaga, 1982. "Socioeconomic Factors Related to the Adoption of
Agricultural Technology in the Western Province of Cameroon."
Dschang: University Center of Dschang, Mimeograph Report.









Wolgin, J.M., 1975. "Resource Allocation and Risk: A Case Study of
Smallholder Agriculture in Kenya." American Journal of
Agricultural Economics, 57(4): 622-630.

World Bank, 1987. World Development Report. New York: Oxford
University Press.

Yotopoulos, P.A., 1967. Allocative Efficiency in Economic
Development. Athens: Center of Planning and Economic Research.




University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - - mvs