.26.: .3. gkruaxgi k.
V. V. 3......» $2 V
. .Wwv .. . .A.) .2 n) 3
. v2 . c 2
. . z . V .
82m»

‘
.73

.r 2..)
v

5‘
§§§f ' ?
435,152?
2:722:22:-
3 '2 ~2
2",“ a
tit!!! ‘ >1
.ggn t 2V
222
2:22;»...
Wagzfid -
6.32:2 2‘
lair

.. {v
’
£753.
» {332.
£2237
221:3?
3“ 1
gig}
: 22.:
47:22
2

. ~ .
7. VV
VV 1317..»Mm . L. V .
V. Irylpf er I .
4H§Fa§¥wmm42wfin mes: . _ V V V V V V
."t Qfiv. trip! :7 .73.“? 5?. .. a V . .r V V V .. V . .
. V.. 2; $2.". QR... . 3.2.... V. V V
8.15.».95 '9 7:! . .1 .f . § 3. \ 3 . . V VV “my .
Mn»? 32. . r If ”‘8'“. Q . .. . V . ~ 3.7
I . : V .
fifik“ ? VPYVL. l!” I .3. . V T 2 .2. . V . . V
25...; t 1' V: I; . F. vs. 2 . .
Sufi... 2!. v 2.2. .382 a... V... .Va V V . .. V .
.frbbra 9t? .\.~§izn§fl. V .\ .3. \. V V .
.9 9. title. v V V x w L .. ‘ .. V . .
. fig .2; t. .. gas-2i it 2. V .V 1. V k.—
. . I 2...... KV¢Q . 3W V31 . . 2m“
V, . $119.? . 2 .V V. VV.. 5’fqlfig .. 2M...
1 x

«I.

.-

xa .. V _. $22.1 fl \.V..u...h)!). _

-5? fiffihfifl. V. V. 2am... V? V :5 A. 2:. a... V . V

O‘ .- bl}.kd. " .l ‘ 'gtL ' ‘ A V
I"? 2

.2. 2 B .

aw“: .

V V .
V “.9 w
. . .1 V3
VV 51”.... 3% my .
$22.2... :%
.. 2.. \z , V .. V V V
. n V v.‘ m .0”
V influx ‘Nut. V V V 3V V I7.
. IV V12 LR. bk»; “223$ .V .
1 .V Shh...
V32. 32.2.1
. (29!: 531.21....
a.‘H¢v—.t043nfir.ufiruthfivza .. “I? V?

‘3‘! 1‘“
1).. VV
V V 2&9.de _
. V.V.. V . V i flaw fink)? V ..V a. .
”gnu-pf. V . . V g V: . ..
V V .k . . . . . .. V .5.
.. V . . . Va V“. . 2 . .2nv£fhr?v2flm.n
v V .. . . .... . V V .V. V . VV .3} In}...
a. .t v \taVini. . V V . I V . .V. . V V. V .. V." m V, . VV .. V .V... . . . . .
3.... A . . V . . V . V V I .
1». 31 14.... £350.05 V V . V V n . 1 . .V . .
7!? . V fl 1‘ 3.... ”19%.. V a V
. {£229.}. EM. V. 1 fix? .n Vuflzueu...\. . V
1:? cavtkfkrttvzfitvlc 2v V . 33332.1(. \iavh. . .
. : 2:\2...l£f2 .. . . V ..
. 4. V V
t. V V .
mu!

1
V‘

2.

.3
7':

. I?
‘. if;

.Y
.3

- .1
e

.. I! I?! zit-Ii. 1V V...

. V .2 3.252.232 2.6?! .gk 5....
2.2:“ . . .V V v
22.1.... V

.. a.
L V.
2.3.... .
V!
we. «mm V V.
1
.V .miu.‘ 5%....“ x V
V I.’ 1“?)1.) .12 .2 a». 13
1 .V
2. 3372522113.... ”.71..
*k1gux. 3a 31. 13!. 2i
grits. $1.2 w“
1121‘ I—Avk ‘Vo‘gV ‘9‘.

.3202.
'8

.
V V. .L2 .2

.t V.
.Ll... ;
s $302.?
an. .23....
h...

r
V v
V tun”; NM

5
V V .u
. . 2 . V

. «ii i5) V . V . :. . .. V .V .V... . L‘hLAR. Hf..-
.23...V!u_l.a...<.2mu.fi:§§§2 2%.. . . V .V 5....
‘51 321333101... 7793 1. V V . . V. A. V «.1.
ibfltjflaxslt ‘hhfifi?khuu: L. . _

. j V.$§§§...§Vll.uumnfih .
V . . . V V . V 712.5351 is
yr 32. V. 2. 313......1n25 .331» ‘21.. v VT...)
.\ 30x1$ a3. :61 .V 32.1.. 3‘0». )2. 5.1.2.... 29.3.2“... ‘
. 4nd uh V V \n. i5t1C2§513335172 2....

\(m . Cara—Van... . . . . 2 . L.
5‘»... RV V V a . . . . V
. 1. .3. $2..” .1606“. 2.4 .r 9. V . .
$3.3 .meu: _ .32... V... V . V. V
. najjuflhztfiafl . . v
i 4 ¢

V138} \ fig V
. . . sou. Ni 3
V .V...

. 2 32.25.... .9 .3322 V?3Vfl.23.m2&22u. .w: V we
.‘SV‘ ‘ z vV Jfi‘i .‘V‘gg k. I."
«EV

.
. . . V . V . V . b
.1132.‘ 3. V . 3V V L.

.Vév

. 221 . L. 32
V . 591 v .211. h... . . .
+5.22»; lfivfifiagfifiq 123%.». V. V V. V. V . . . V V V . V
v! 2%5233551 V .SVV. «9W2... in. 2. .3... V V V. V V : V .V . . . V
ivkfigfrb§Virw 32254221212323. n... «233222.225... .3. mm... . sVVV V .V . . V. . 1 . . V V. V. V M.» V . it?
Iauflfil.§5: xi 3175‘? hunk”, rn 13.1% . . .V . . 2 .. . 245.5 . Vnaflflffl.h.flplclh.1. I
V .25. V, 4? VE$§.§2 V V “snug.“fiztgméurf
3:55.: 2.8.3» 29.3.5»... V
. V V 13m”. x :2
2. V V

5.1:. . LsVetAfM. V...
252.22.. szVVVEmV... w...
a:

.3
5

2:22.23;

§

V J52.
5.5%]... L

V 72... u. n V V. .
V. V.. . S . V : . . . .
.. V . 2 3.. V V.V ..V V x .. V V . . . x1» . 2
.V. . .V V V : V. .3. V. 2. . V . .. V . .V V. V “$3329.33.
. . 1 5.2.7! . . V .. . .3. .VV L .V V.. I ..V i
V . V 2 3:22.21 2. V, . V .. . V V. V
:2 32222.21... . 2mm.» 2§§4J. 2.134.. . .V V. V VV V V . .V .V V .V
$2,233.... V $92. V2.1 L! V V . .. V . V V
ob V VA . V . V. V
.devs}. . V V .V
. . £2 1 .

3? iii;

27
is

“8.1.2.... .
.3
N.

. rnrshuhmunuh}
g 7 i2 an.

. .9 . 2.. . a

t .r}$s.§§fiffi 2 .V

if V.. . .&.§ME {Lméxa .Lmunnm¢&mvfimflm

9.1:)... {A 7!: . .V .... .1...
22.2. g V Ptaufimzififxiou Wang. 2

V .9..ng n. H V. .V £29 .28. .2 2.

. 5
.V. :
Y. 02.3 L
Lu... . .~$rL.tL..Hrr3 .g
. . fr...” 5...... V V. . .1 . 2.... influf..1.
.V. V .V . 2‘ 7 1 V . .. . w .
. . V 5.2.... gag. . V . 2; V V.
sufiflhufib. 2 2V. V 2mg 2.22%.. V .V V .V V .V . V. ”ma. V. .. L. 2.!
2122.... (”3.2.5 .2 . ..7V .3...“ Yr! .2... 23.. NW . :V V...” . as. ”VI/«AV . VV .VVs .lfifii .i
V V 22.22%} .f'litfit 22 .IfV. {£5.37 2. . 5b 4 V i V :V .V
FIVE...)2.mum.uub&tr¢2.2§.m.s2namk.§9 .2..- 5222.... V .V.
V.V.I.E§ffr.§.£ 21.1" V
2.0.

. f

H
23??
’53.??? '

£3 ‘

’7 fl!
7::

=8
:57

'1
'1'.’

e if... 2 .2 0
2. V .63.... . .tl 2.1.qunnmmflww
. .2 V 3. V1 .5. . . .fi . V VhVV..‘H.£vV4L....~l t1: V.~..r.{.!.?29??7 .13.}? .
V. IannWrE. .. . .. idfnuflm ..V. .nwermvzfa muffin... 2.954. LHwVTuzzm.
.. .vr..t~.u§£_.€§_ 2.1.2 9.5.5.9. ARA. Vg... {A 9.13.2.3... ...rfhna2ur.....a§..r
Virzhnyii 2...: t V. 2...... «H.432... @223: V . . $0.33”, .fisfiujlfifivwhztxv.«HR:
£9511}: .vanmmLflgi 2 2...... ..s&£§£ufibnfazgéV 2 2.92.»..mmufiv .Janhlwufifircs} 3V.
V... 2.2%.. 2.2. m 2.3.12.3.3222 V . V V 2
.ftz.2.l:gg Tr; 7a.? .34.}.an I! . z .V .
a , 2.. «dwarf?! 5.23.6.2 . . .
fitting...” $271.. p421..2.7..vusr..fihmaxrrfl.u: r . . . V: .2
VVEHuzammmmhvzrfgh flit. ”V as. gram... . V V V V V . _ x. :2...
V 2 v u. . -. .
.Vstt... .u%.2§¢3.u§nflifi%tfir. V. V . fihkmfiuh $212433..us twirl...
I... 3 V 2V fnfpVVhVuPb.truur VV .- .V _ . . :39? V. 52:2... in . .h.2...._.2..fiVVVI.V
flatly..- Vz.1t.k...h. konvkcs. , . .V V. . 3.32.32 5:52 I. Ln??? V V“ :.Pt..
36.2%:th . : L . .1)?! d.‘$~nlflfli+l gins"... flukfiaaéub
V 5:25.“; V V . total. I. Var: TE... .2 .22: a...
a. .1 v3 12; V . . .V 3437 3....I. T. .4) 5 “an”: .V» 732A 3.1....
. . L2. 2.. .VV1._.V2.. V V .ngfl2w. 2.49.9.2»... 1.252%..2. 3.23
V. “\fiabtxs V .23.}: . Jinn“. LJNXMVI.“7~... .T.«.m~uwtu2h& F.37k312x1Mu. .Frhfstfiaonur’
I. ‘ . § '7- .. 2 V . . F \. I
i flgnzfitrm 22.... 2.252%...22“ . 2. 23......9. .2243... .nhnohthxifflw
. . 2. .V: ._V n Laiatrfimuflhw; .211... . VIE. . fiva. +55 2.3.»? . 15122.91...
. V n. r... L. . a $22.“ .rfiinfyr..nr§.kvmvsrsflrlq 2%.. V. V .2; w.
2. . V. . VV . . . I. .95 . . .
VV V . .V 2.... rake»... 2 .V 4.2.... figufifiiei .. V. . . V
. V 2 . u . V: .. .22.. V: .. . . .1 V. V1.3. ‘VlfléPmMHn gun“. .2..V:.V. e .7 . ; .
Wheryvurgfifluw. .V . . . . _ V VV V V. . .VVV. .
V . .1.....:17.VV .. . V.. 2 .V. . ... V. . .

m
V:

 

 

. V V

:1. 2. . In»:
2 n y . \. {.32 .
.wV......\ VERIPMmLWQLBFHI

 

NW

HTMTW”iii“ IL,

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

THESIS

This is to certify that the

dissertation entitled

FARMING SYSTEMS FOR SMALL FARMERS
IN GEITA DISTRICT OF TANZANIA

presented by

HAIDARI K.R. AMANI

has been accepted towards fulfillment
of the requirements for

Ph.D. mgmemAg Economics

mm;

Major pron or

DMe February 2, 1982

Msu;.,...4mm,..;...1 - I", -n

 

 

 

 

MSU

LIBRARIES

 

 

RETURNING MATERIALS:
PIace in book drop to
remove this checkout from
your record. FINES wiII
be charged if book is
returned after the date
stamped beIow.

 

 

 

 

 

AN ANALYSIS OF MAIZE—COTTON
FARMING SYSTEM FOR SMALL
FARMERS IN GEITA DISTRICT

OF TANZANIA

By
Haidari Kanji Ramahani Amani
A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements

for the degree of

DOCTOR OF PHILOSOPHY

Department of Agricultural Economics

1981

 

 

ABSTRACT
AN ANALYSIS OF MAIZE-COTTON
FARMING SYSTEM FOR SMALL

FARMERS IN GEITA DISTRICT
OF TANZANIA

By

Haidari Kanji Ramadhani Amani

The problems of increasing agricultural output and productivity
in Tanzania vary from area to area. Some areas are faced with land
scarcity. In such areas more must be produced per hectare, and land—
saving technologies must be employed. Other areas face drought
conditions. Thus, drought resistant crops must be given top priority.

In areas where food crops compete for resources with annual export
crops, government policies may misguide farmers to allocate their re—
sources inefficiently. This is because the obejctives of the farmer
and the government are not the same. The farmer wants to utilize his
resources efficiently in order to maximize net farm income after meet—
ing family food consumption requirements. The government is concerned
with increased agricultural output that farmers can sell through the
controlled market.

The purpose of this study is to narrow the gap between the two by
analyzing the impact of the government's measures to increase agricul—
tural output on resource allocation, cropping patterns, and output.
The study was done in Geita district, MWanza region, Tanzania. Speci-
fically, the objectives of the study were to:

l. evaluate the impact of the constraints imposed by farm re-

sources on the production of cotton and maize;

Haidari K.R. Amani

2. assess the implications of oxen and tractor technological
choices with respect to their impact on net return per hectare, on
employment, and on labor productivity;

3. analyze the impact of rescheduling agricultural production
on resource allocation, agricultural output, and farm earnings; and

4. test the sensitivity of alternative input and output prices
on resource allocation, enterprise combinations and net farm income.

The data used in the study was collected from the farmers through
questionnaires. The methods of analysis included static and parametric
linear programming analysis.

The results of the analysis indicated that under optimum allocation
of existing resources, net farm income could be increased substantially.
However, further increases in farm income were hindered by seasonal
labor bottlenecks.

The net farm income was influenced by product and input prices.
The results showed that the official price of maize is significantly
below the free—market price so that farmers sell outside official
markets. Further, if input subsidies are removed, the improved farm-
ing practices experience larger declines in gross margins than the un—
improved practices. This suggests the important role played by input
subsidies in the system.

Depending on the realism of the assumptions made in the analysis,
the results obtained from the study could provide relevant insights and

guidelines for policy makers and researchers.

ACKNOWLEDGEMENTS

My study program and dissertation preparation benefited greatly from

 

the insights, suggestions and constructive criticisms of my Guidance
Committee which was comprised of Drs. Stephen Harsh, Carl Eicher, John
Ferris, Gerald Schwab and Anthony Koo. Special thanks, of course, are
due to Dr. Harsh, my Major Professor and thesis supervisor for his
humanity, help, farness, and flexibility. Special thanks are also due
to Dr. Eicher; his concern, humanity and encouragement helped when I
reached the walls throughout my graduate studies.

The Rockefeller Foundation fellowship award which enabled me to
start this program in fall 1976 as well as to finally complete it with
the Ih.D. thesis is gratefully acknowledged.. I also wish to place on
record the genuine interest and consideration that Mr. E. Mlay, J. Toya,
and A. Mhogolo afforded me whenever I was up in Geita district on field
work. Their active support in the organization and conduct of the field
research, and most of all their kind and considerate nature, went a long
way in enabling me to complete my field work on time and with good feeling.

Last, but by no means least, special appreciation is extended to
Joanne Gram who patiently and persistently typed this manuscript as well
as countless tables. My poor organization and penmanship never changed
her friendly way, and I am grateful for her.

Any errors or omissions found in this manuscript are solely the
responsibility of the author.

ii

Chapter

I

II

III

TABLE OF CONTENTS

INTRODUCTION . . . . . . . . . . .
Statement of the Problem . . . . . . .
Objectives of the Study . . . . . . .

The Organization of the Study . . .

BACKGROUND TO THE STUDY . . . . . . .

Production Patterns of Cotton and Maize in

Agricultural Pricing Policy in Tanzania
Production Patterns of Cotton

Production Patterns of Maize . . .

Review of Relevant Studies . .

Farming Operations . . . . . . . . . .
Land Preparation . . . . . . . .
Time of Planting . . . . . . . .
Weeding . . . . . . . . . . . . .

Harvesting .

Agricultural Mechanization in Tanzania:
Past Experience . .
Block Farming and Mechanization . .
Past Experiences in Mechanized Block

Farming of Cotton

o

Tanzania.

Tillage Tools Used in the Production of Cotton,

Maize, and Other Crops . . . .
Primary Tillage Tools .
Hand hoes . . .
Mouldboard plows
Disc plows . . . .
Heavy— duty disc harrows . .
Chisel plows . .
Rotavator . . . . . . . .

Ridgers . . . . . . .
Secondary Tillage Tools . . . .
Disc harrows . . . . .
Tyne harrows . . . . . .
Rotavators . . . . . . .
Tools for Weeding . . . . . . . .
Hand hoes . . . . . . . .
Tyne Implements . . . . .
Ridgers . . . . . . . .

THE AREA SELECTED FOR THE STUDY

Description of the Geography and Climate

the Study Area . . . .
Location, Size, and Topography .
Climate . . . . . . . . . . .
Soils . . . . . . . . .
Population . . . . . . . . . . . .

ll

22
25
25
27
27
27

32
38

39

 

Chapter

IV

Agriculture . . . . . . . . . . . . . .
Main Crops . . . . . . .
Cultivation Practices . . . . . . .
Agricultural Calendar . . . . . . .
Pests and Diseases . . .

Recommended Practices for Cotton .
Recommendations related to

cultivation practices . . .
Recommendations related to new '

Planting . . . . .

Thinning and Weeding .

Harvesting .

Uprooting and burning

Crop rotation . . . . . .

Fertilizers .

Insecticides

Recommended Practices for Maize .
Recommendations related to
cultivation practices

Recommendations related to new inputs

THE ANALYTICAL FRAMEWORK OF THE STUDY AND

RESEARCH METHODOLOGY . . .

The Analytical Framework . . . .

The Use of the Linear Programming Technique

Analyzing Small Farm Agriculture

The Linear Programming Model . . . . .
Research Methodology . . . . . . . .
Choosing the Research Site . . . .

Choosing the Sample . . . . . .
Procedures for Selecting Respondents

Sample Size . . .
Replacement . . . . . . . . . . . .
Data Collection . . . . . .

Construction of the Representative Farms
Characteristics of Farms in the Sample

Land Use . . . . . . .
Farm—Labor Force . . . . . . . . .
Farm Capital . . . . . . . . . . .
Cropping Patterns . . . . . . . .

for

THE STRUCTURE OF THE LINEAR PROGRAMMING MODELS

FOR THE STUDY ARE A . .
The Objective Function . . . . . . .
Activities in the Model . . . . . . .
Crop—Production Activities . . . .
Labor—Hiring Activities . . . . . .
iv

Page

59
59

64
69
7O

71
71
72

74
74

76
77
78

78
79

81

106

106
107
107
110

 

 

Chapter

VI

Working—Party Activities . . . . . . . .
Animal-Power Hiring Activities . . . . .
Animal— Power Owning Activities . . . . .

Tractor-Hiring Activities

Fertilizer— and Insecticide-Buying Activities

Consumption Activities . . . . . . . .
Crop—Selling Activities . . . . . . . .
Animal—Power Selling Activities . . . .
Transfer Activities . . . . . . . . . .
Restrictions in the Model . . . . . . . .
Agricultural— Land Restrictions . . . .
Labor Constraints . . . . . . . . .
Operating-Capital Restrictions . . . . .
Food Consumption Constraint . . . . .
Nonnegative Restriction . . . . . . . .
Some Limitations of the Model . . . . . . .

ANALYSES OF THE RESULT OF THE APPLICATION OF
LINEAR PROGRAMMING MODELS . . . .

Optimal Organizations with Existing Resources

and Prices for Participants . . .
Model A: Mechanization and Manual Technologi
Maize and Early Cotton . . . . . . .

Marginal— —Va1ue Products of Resources
Under Model A . . ...

Optimal Organizations of the Participants'
Representative Farms with Variable Produc
Fertilizer, and Insecticide Prices, and
Operating Capital Levels for Early Cotton
and Maize . . . . . . . . . . . . .

Alternative 1: Changes in the Relative Price
of Early Cotton and Maize

Marginal— Value Products (MVPs in Tshs) of
Resources Under Alternative I . . .

Alternative II: Increases in the Prices
of Fertilizers and Insecticides . . .

Marginal—Value Products (MVPs in Tshs) of
Resources Under Alternative II . . . .

Alternative III: Increase in the Levels
of Operating Capital . . . .

Marginal— —Va1ue Products (MVPs in Tshs) of
Resources under Alternative III . . . .

Model B: Mechanization and Manual Technologi
Maize, Early Cotton and Late Cotton

Marginal—Value of Resources under Model B

Optimal Organizations with Existing Resources
and Prices for Nonparticipating Farmers

 

THE

ES:

t,

S

ES:

0 o

Page

113
115
115
118
118
120
121
121
122
122
122
124
124
125
126
126

128

130
130

141

145
145
157
163
167
170
173

176
180

182

Chapter

Model C: Mechanization and Manual Technologies:

Maize and Early Cotton for Nonparticipants . .
Margina1——Va1ue Products of Resources

Under Model C . . . ...
Optimal Organizations of the Nonparticipants'

Representative Farms with Variable PrOduct,

Fertilizer, and Insecticide Ikices, and

Operating Capital Levels for Models

to C . . . . . . . . . . . . .

Alternative 1: Changes in the Relative

Prices of Early Cotton and Maize . . . . . .
Margina1——Va1ue Products of Resources

Under Alternative I . . . . . .
Alternative II: Increases in the Prices of

Fertilizers and Insecticides . . . . . . . . .
Marginal—Value Products of Resources

Under Alternative II . . . . . . . . . . . . .
Model D: Mechanization and Manual

Technologies: Early—Planted Cotton,

Maize, and Late— Planted Cotton for

Nonparticipants . . . . . . . . . .
MarginaL —Value Products of Resources
Under Models D to D . . . . . . . .

Comparison of Optimal Organizations

with Existing Resources and Prices

for Participants and Nonparticipants . . . .
Summary . . . . . .

VII SUMMARY, POLICY IMPLICATIONS, LIMITATIONS
OF THE STUDY AND SUGGESTIONS FOR
FURTHER RESEARCH . . . . . . . . . .
Summary . . . . . . . . .
Iblicy Implications . . . . . .
Limitations of the Study and
Suggestions for Further Research . . . . .
APPENDICES
A THE STRUCTURE OF GDP (FACTOR COST IN 1966 PRICES).
B NUMBER OF TRACTORS AND OX—ILOWS IN GEITA DISTRICT.

BIBLIOGRAPHY . . . . . . . . . . . . . . . . . . . . . . . .

vi

Page

183

188

190
191
198
200

205

207

209

211
215

222

222
232

236

240
241

242

 

 

 

LIST OF TABLES

'Table Page
1.1 Imports and Exports of Major Food Grains,

1965-66 - 1975-76 . . . . . . . . . . . . . . . . . . 4
2.1 Production and Exports of the Main Export Crops . . . 12
2.2 Sales of Cotton to local Mills . . . . . . . . . . . 13
2.3 Tanzanian Oil-Seed Crops: Value of

Production 1970-1977 . . . . . . . . . . . . . . . . 14

2.4 Estimated Cotton Production 1969—70 . . . . . . . . . 19
2.5 Recorded Commercial Transactions

of Main Food Crops . . . . . . . . . . . . . . . . . 20
3.1 Monthly Rainfall at Sengerema, Ukiriguru,

and Geita Towns . . . . . . . . . . . . . . . . . . . 51
3.2 Mean Annual Rainfall 1973-78 in Millimeters

and Days of Rainfall . . . . . . . . . . . . . . . . 52

3.3 Seasonal Distribution of Rainfall Since

1975 in Millimeters . . . . . . . . . . . . . . . . . 53
3.4 Estimated Land and Water Areas . . . . . . . . . . . 55
3.5 Land Distribution by District . . . . . . . . . . . . 56
3.6 The Population of Geita Since 1934 . . . . . . . . . 57
3.7 Sex and Age Distribution of Population

by District . . . . . . . . . . . . . . . . . . . . . 58

3.8a Allocation of Land to Major Crops . . . . . . . . . . 61
3.8b Allocation of Land to Major Crops . . . . . . . . . . 62
3.9 Calendar of Crop Operations for MWanza and

Shinyanga Regions . . . . . . . . . . . . . . . . . . 65
3.10 A.Typica1 Growth Calendar for Cotton in

Tanzania's western Cotton-Growing Areas . . . . . . . 67
3.11 Responses to Nitrogenous-Fertilizer Under

Varying Insecticide Applications . . . . . . . . . . 77
4.1 Composition of the Average Family in the

Study Area . . . . . . . . . . . . . . . . . . . . . 99

vii

Table

4.2

4.3

5.1

5.2

5.3

5.4

5.5

6.1

6.2

6.3

6.4

6.5

6.6

Total Labor Inputs by Month, wage Rates, and
Payments in Kind on the Representative Farms .

Operating capital of the Representative Farms
by Months . . . . . . . . . . . . . . . . . .

Crop Production Activities . . . . . . . . . .
Labor—Hiring and WOrking Party Activity . . .

Ox Plow Owning, Ox Plow Hiring, and Tractor
Hiring Activities . . . . . . . . . . . . . .

Fertilizer— and Insecticide—Buying and
Selling Activities, and Food Cbnsumption
Activity . . . . . . . . . . . . . . . . . .

Transfer Activities . . . . . . . . . . . . .

Comparison of Bench—Mark and Optimal

Organizations Under Existing Resources and the
Official Prices for Maize and Early Cotton for
Participants . . . . . . . . . . . . . . . . .

Comparison of Bench—Mark and Optimal
Organizations Under Existing Resources, the
Official Price for Early Cotton and the Free-
Market Price for Maize for Participants .

Margina1-Value Products (MVPs in Tshs) of
Resources for Models A1 to A for

Participants . . . . . . é . . . . . . .

Efficiency Measures for the Base Plans and
Changes in Relative Product Prices for Early
Cotton and Maize for Participants:

Alternative Ia . . . . . . . . . . . . . . . .

Efficiency Measures for the Base Plans and
Changes in Relative Product Prices for Early
Cotton and Maize for Participants:

Alternative lb

Efficiency Measures for the Base Plans and
Changes in Relative Product Prices for

Early Cbtton and Maize for Participants:
Alternative IC . . . . . . . . . . . . . . . .

viii

Page

101

104
108

111

116

119

123

131

135

140

147

150

153

Table Page

6.7 Marginal-Value Products (MVPs in Tshs)
of Resources for Mbdels to A for

Participants: Alternative Ia .4. . . . . . . . . . 158

6.8 Marginal-Value Products (MVPs in Tshs) of
Resources for Models A1 to A for Participants:

Alternative lb . . . . . . 6 . . . . . . . . . . . 160

6.9 Marginal-Value Products (MVPs in Tshs) of
Resources for Models A to A.4 for Participants:

Alternative IC . . . l . . . . . . . . . . . . . . 162

6.10 Efficiency Measures for the Base Plans and
Increases in Fertilizer and Insecticide Prices
for Early Cotton and Maize for Participants . . . . 165

6.11 Marginal-Value Products (MVPs in Tshs) of
Resources for Early Cotton and Maize for
Participants: Alternative 11 . . . . . . . . . . . 168

6.12 Efficiency Measures for the Base Plans and
Increases in the Levels of Operating Capital
for Early Cotton and Maize for Participants . . . . 171

6.13 Marginal—Value Products (MVPs in Tshs) of
Resources for Early Cotton and Maize, for
Participants: Alternative III . . . . . . . . . . 174

6.14 Comparison of Bench-Mark and Optimal
Organizations Under Existing Resources and
the Official Price for Early Cotton and Late
COtton, and Free—Market Price for Maize for
Participants . . . . . . . . . . . . . . . . . . . 177

6.15 Marginal-Value Products (MVPs in Tshs) of
Resources for Models B to B Under Early
Cotton, Late Cotton, and Maize, for
Participants . . . . . . . . . . . . . . . . . . . 181

6.16 Comparison of Bench—Mark and Optimal
Organizations Under Existing Resources
and the Official Prices for Maize and
Early Cotton for Nonparticipants . . . . . . . . . 184

6.17 Comparison of Bench-Mark and Optimal
Organizations Under Existing Resources
and the Official Price for Early Cotton
and Free—Market Price for Maize for
Nonparticipants . . . . . . . . . . . . . . . . . . 187

ix

Table

6.18

6.19

6.20

6.21

6.22

6.23

6.24

6.25

6.26

6.27

6.28

Marginal-Value Products (MVPs in Tshs) of
Resources for Medels C1 to C4 for

Nonparticipants . . . . . . . . . .

Efficiency Measures for the Base Plans and
Changes in Relative Product Prices for Maize
and Early Cotton for Nonparticipants:
Alternative Ia . . . . . . . . . . .

Efficiency Measures for the Base Plans and
Changes in Relative Product Prices for Maize
and Early Cotton for Nonparticipants:
Alternative Ib . . . . . . . .

Efficiency Measures for the Base Plans

and Changes in Relative Product Prices for
Maize and Early Cotton for Nonparticipants:
Alternative IC . . . . . . . . . . . . . .

Marginal-Value Products (MVPs in Tshs) of
Resources for Medels C to C for
Nonparticipants: Alternative Ia
Marginal-Value Products (MVPs in Tshs) of
Resources for Medels C to C for
Nonparticipants: Alternative Ib

Marginal-Value Products (MVPs in Tshs) of
Resources for Medels C to C for
Nonparticipants: Alternative IC . . .
EFficiency Measures for the Base Plans
and Changes in Fertilizer and Insecticide
Prices for Early Cotton and Maize for
Nonparticipants . . . . . . . . . .

Marginal—Value Products (MVPs in Tshs) of
Resources for Models C to C for
Nonparticipants: Alternative II .

Comparison of Bench-Mark and Optimal
Organizations Under Existing Resource
and Official Prices for Early Cotton
and Late Cotton for Nonparticipants .

Marginal-Value Products (MVPs in Tshs) of

Resources for Medels D1 to D4 for
Nonparticipants . . . . . . . .

Page

189

192

194

196

199

201

202

203

206

208

210

Table

6.29

Page

Comparison of the Optimal Organizations

of the Representative Farms for

Participants and Nonparticipants with

Existing Resources and the Official Price

for Early Cotton and Free-Market Price

for Maize . . . . . . . . . . . . . . . . . . . . . 212

Comparison of Marginal-Value Products

(MVPs in Tshs) of Resources between

Participants and Nonparticipants with

Existing Resources and the Official

Price for Early Cotton and Free-Market

Price for Maize . . . . . . . . . . . . . . . . . . 214

xi

LIST OF FIGURES

Cotton—Growing Areas . . . . . . . . .
Production in Tanzania . . . . . . . .
Primary Maize—Growing Areas . . . . .

Calendar of Crop Operations for
Cotton and Maize . . . . .

The Relationship Between Cotton
Yield and Planting Date .

The Impact of Relative Product Prices
on Maize Production for Participants .

The Impact of Relative Product Prices
on Maize Production for Nonparticipants

Page
15
18

21

26

28

219

220

 

CHAPTER I

INTRODUCTION

Slightly over 90 percent of Tanzania's population lives in the
rural sector and is primarily engaged in agricultural production.

Most farms are small and are owned and operated by families. These
farmers market approximately 40 percent of their food crops and

almost 100 percent of their other crops such as cotton, tea, and

coffee. They purchase about 10 percent of their labor inputs. Modern
inputs, such as fertilizers, improved seeds, insecticides, and machinery,
are used only minimally (Carr, 1976).

The small farms contribute over 40 percent of the Gross Domestic
Product (GDP) and 80 percent of the total export earnings (Malecela,
1980). It is the agriculural sector that, to a large measure, determines
the pace of economic and social development of Tanzania.1 This sector
is expected to supply the country's population with its nutrient require-
ments, to earn foreign exchange, which is badly needed for the develop—
ment of this sector and other sectors of the economy, to provide raw
materials for domestic industries, to provide a large and effective
market for domestic industrial goods, and to supply labor to other
sectors of the economy. The role of agriculture in the Tanzanian
economy was summarized by President Nyerere (1968):

The mistake we are making is to think that development

begins with industries. It is a mistake because we do not have

the means to establish many modern industries in our country.

Agricultural progress is the basis of Tanzania's development.

A great part of Tanzania's land is fertile and gets sufficient
rain. Our country can produce various crops for home

1
See Appendix A

Consumption and for export. We can produce food crops (which
can be exported if we can produce in large quantities) such as
maise, rice, wheat, beans, groundnuts, etc. And we can produce
such cash crops as sisal, cotton, coffee, tobacco, pyrethrum,
tea, etc. And because the main aim of development is to get
more food and more money for our other needs, our purpose must
be to increase production of these agricultural crops. This is
in fact the only road through which we can develop our country.

The Tanzanian government's preferred strategy for agricultural
development is based on the socialist strategy of organizing production.
In this strategy, the agricultural sector can be divided into four sub-
sectors: (1) ujamaa (communal) farms, in which a group of individuals
or families voluntarily own land, lives, and works together for the
benefit of all; (2) state farms, which are owned and run by government
agencies; (3) block farms, in which each family owns its own land,
makes its own production decisions, but can easily and voluntarily
share extension services, capital assets, and so forth; and (4) small
holder household or family farms.

The focus of this study is on the last two subsectors. These are
believed to bear the largest share of responsibility in Tanzania's
agricultural development. Although the government's agricultural
policy still emphasizes communal and state farms as the key of these
subsectors, their apparent failure has recently led to a quiet recogni—

tion of the importance of block farming and independent small holder

farming.

Statement of the Problem

 

The performance of the agricultural sector has not kept pace with
the growing population of Tanzania. Between 1967 and 1978, the popula-

tion increased by about 3.3 percent a year, whereas food production

increased by only about 2.5 percent per year over the same period
(Ellis, 1980). The government had to spend its foreign—exchange earn—
ings in order to import food, and these food imports have increased in
recent years (Table 1.1).

The production of export crops such as coffee, tabacco and cotton
did not improve either. Total output declined from 611,654 metric tons
in 1973-74 to 469,057 metric tons in 1978-79, a decrease of 23.3 percent.

The problems of increasing agricultural out put and productivity
vary from one area of the country to another. Some areas are faced with
rapid population growth that causes land scarcity. Such areas include
the Kilimanjaro region, parts of the Lake Victoria basin, the Usambara
highlands, parts of the Mbeya and Iringa regions, and around Mount Meru
in the Arusha region. In such areas of land scarcity, more must be
produced per hectare of farm land. Technologies that are land saving
or are substitutes for land need to be employed. Perhaps the most im—
portant means of increasing agricultural productivity in land—scarce
areas is increased use of high—quality farm inputs such as fertilizers,
higher-yielding seeds, and pesticides. In addition, investments in
such factors as rural feeder roads, marketing and storage facilities,
agriculatural research, and extension services are necessary. Other
areas such as Dodoma and Singida face drought conditions. For such
areas, there needs to be research in drought—resistant crops such as
sorghum, millet, cassava, and potatoes.

In areas where food crops compete for resources with annual export
crops, there is the problem that government policies (especially pricing

policy) may misguide farmers to allocate their resources inefficiently.

 

.omo_ .cmamq umLmAFQDQCD .mc:p_:oecm< co acumwcrz ”moczom

 

- - - - - w._s - _.m_ - - - mptoaxL
_.ea w.swe s.msm m.sa e.um_ - s.Nw - m.m_ m.se w.is muroael

:l‘i

Hm: quoh

 

- - - N.ol N.e - - - - - - mutoaxm
©.om o.m© o.mm - - - - - m.m m.© s.__ motoaar

- - - - - - - - u - - mucoaxm
m._m ©.mo_ m.mm N.w m.me o.__ m.mm n.0m m.m— P.Nm ©._m mucoaeH
cnoam a pawn;

 

- - - - - e.mm - m.2m m.o - - mutoaxm
m.me m.s~m o.mm_ m.mn m.mm . o.oe . - m.m m.w mucoaeH
mema

 

omimnmfi manemmfi emumxm_ maummmP NmuFNmF Pnnomm— ORIOQQF moumom_ wwuwomP nonwomF moumom_

 

mzoe z comm

 

om\mmm~imm\moo_ mzH<mo DOOA mow<z mo memoaxm oz< memoazH

_.a m_aee

 

Such areas include Tabora and Iringa where maize competes with tobacco,
and MWanza, where maize competes with cotton. Strategies for increas-
ing agricultrual output and productivity in these areas should be
based on a study of the impact of government policies on resource
allocation. Such studies should concentrate on the impact of input-
output price relationships on land and labor allocation, capital
budgeting, and utilization of modern inputs such as fertilizers and
insecticides. Where labor becomes a scarce resource, studies should
center on labor-saving techniques, such as mechanization, or on a
farming system that would change the demand for labor by utilizing
different farming operations. For example, perhaps one of the com-
peting crops could be grown a little earlier or later than the re—
commended time.

The government, however, has taken different measures in response
to the decreasing agricultural output. Since 1974, a single pan—
territorial (uniform) producer price for each crop has been fixed
annually by the Economic Committee of the Cabinet. This uniform
price applies to all major food crops, an increasing number of minor
food crops such as cassava, domestic oil seeds, and to all export crops
except coffee and sisal. Differentials in the uniform price are fixed
for different quality grades where applicable, but no distinction is
made with respect to location or transport costs.

In addition, the Revolutionary party (the only political party in
Tanzania) directs regions to meet crop targets set by the party; the
regions in turn set crop targets for each of their districts. This

process continues up to the farm level. In most cases, these targets

.are.quite arbitrary and are not accompanied by an incentive when they
are met or by punishment when they are not met.

Furthermore, in 1975 the government signed an agreement with the
'World Bank by which funds were to be made available for improving
agricultural practices in major crop-producing areas such as Geita
(in the MVanza region) where maize and cotton are the main crops or
Tabora where tobacco and maize are widely grown. It was decided to
introduce new management practices under a planned sequence of events.
The new management practices involved the use of improved seeds, farm—
land manure (FYM) and pesticides in conjunction with inorganic fertilizers.
This agreement did not cover every family farm primarily because the cost
‘would have been very high.

As for the family farms not included in this project, the government
required these farmers pay the same price for farm inputs such as
fertilizer as that paid by participating farmers. Although there were
no differences in input or output prices, nonparticipating farmers
usually could not get enough fertilizers, or if they could, the fer—
tilizers arrived too late because of the poor transportation system.

There is a wide gap between the knowledge of the small farmers
and that of the government policy makers. The farmer is concerned
‘with efficient utilization of his resources to maximize net farm
income after meeting consumption requirements of the family. The
policy makers are concerned with increased agricultural output that
farmers can sell through the goverment-controlled market.

The purpose of this study is to narrow the gap between the

objectives of family farms and policy makers, based on analysis of the

janact of the government's measures to increase agricultural output on
gulch factors as resource allocation, cropping patterns, and output of
tflne family farms. Analyses of labor-saving techniques, new management
Irractices, and changing price structures are also included in this

study.

Objectives of the Study

 

The objectives of the study concern both the participants and non—
puarticipants of the Geita Cotton Project. The objectives are as
follows:

1. to evaluate the impact of the constraints imposed by farm

resources on the production of cotton and maize;

2. To assess the implications of oxen and tractor technologi-
cal choices with respect to their impact on net return per
hectare, on employment, and on labor productivity;

3. to analyze the impact of rescheduling agricultural produc—
tion (by including early- and late—planted cotton) on
resource allocation, agricultural output, and farm earnings;
and

4. to test the sensitivity of alternative input and output
prices on resource allocation, enterprise combinations and

net farm income.

The Organization of the Study

 

In chapter 2, a background for the study is presented. It includes
a discussion of the production patterns of cotton and maize in
Tanzania, a review of past studies of these crops and of accounts of

past experiences with agricultural mechanization in Tanzania.

The methodology and analytical techniques used in the study are
discussed in Chapter 3. The analytical techniques consist of static
and parametric linear programming analyses. The application of linear
programming techniques in African agriculture is reviewed. The
sources of various data sets are described.

In Chapter 4, there is a description of the characteristics of
farming operations in the study area and methods for designating
representative farms are discussed. The structure of the linear
programming models used to represent the planning environment of
the representative farms in the study area are presented in Chapter 5.
Included are discussions of the model activities, technical coefficients,
prices, and resource restrictions used in the study area. In Chapter 6,
there is an analysis of optimum farm plans in terms of net farm income,
cropping patterns, and resource use. The impact of varying product and
input prices, capital and rescheduling cropping calendars on cropping
patterns and resource use is also examined. A summary of this study,
its implications for policy making, and suggestions for future re—

search are presented in Chapter 7.

 

9
CHAPTER II
BACKGROUND TO THE STUDY
Agricultural Pricing Policy and Production Patterns of Cotton and Maize

in Tanzania

In Chapter I, the general performance of Tanzania's agricultural
sector from 1967 to 1978 was briefly discussed. In Chapter II, a dis—
cnnssion in greater detail is presented concerning government pricing
Inalicy and production patterns of cotton and maize, the two crops
leOD.WhiCh this study is focused. Included in the discussion are
past production trends and problems, and brief review of relevant
literature pertaining to these crops.
yégricultural Pricing Policy in Tanzania

The stated objectives of Tanzanian development strategy, as
contained in the Arusha Declaration, are three in number. The first
objective is to achieve domestic food self-sufficiency. The second
is to expand the foreign exchange earning role of the agricultural
sector in order to increase the importation of capital and inter-
Inediate goods for the growth of the industrial sector. The last objec—
tive is to improve the standard of living of the rural population
through increasing rural incomes. These objectives are by no means
compatible. There is the problem of reconciling domestic food self—
sufficiency with the foreign exchange earning role of the agricultural
sector. What is consumed cannot be exported and vise-versa. There
is also the problem of reconciling the objective of increasing rural
incomes with the tendency for farmers to receive a progressively

smaller proportion of the market value of their crop sales (Ellis).

10

The main characteristic feature of the marketing system in Tanzania
is the high degree of centralized control dating back to 1967. The
principal feature is the movement toward greater stability and uni—
formity of output prices for each individual crop across the country,
coupled with the incorporation of an increasing number of crops into
official marketing channels. Most of the important agricultural com—
modities are under the control of crop authorities and other statutory
authorities. The Tanzania Cotton Authority (TCA) and the National
Milling Corporation (NMC) are the regulatory authorities for cotton
and maize respectively.

The Tanzanian government fixes the prices of food and export
crops just as it does for inputs like fertilizer and insecticides,
and consumer goods. These prices hold for all parts of the country.
There is no distinction made with respect to location or transport
cost. Differentials in the uniform price are fixed for different
quality grades for many crops including coffee, cotton and tobacco.

Administered prices are often not in harmony with the stated
development objectives of the government. Relative prices influence
the competitive positions of crop enterprises and, therefore, deter-
mine the output mix. There is evidence that the government's control
over the marketing system is one of the main causes that stagnated the
agricultural sector (Ellis, International Labor Organization, and
World Bank). It has been argued that the practice of fixing input and
output prices restricts economic forces and hampers the efficiency

allocation of resources.

11

Production Patterns of Cotton

 

The importance of cotton in the Tanzanian economy is summarized
in Table 2.1. Cotton accounts for more than 17 percent of the country's
total exports which in turn make up for about 30 percent of the country's
GDP. The crop is an important source of foreign exchange that is badly
needed to pay for imports. Because the government increasingly empha—
sizes rapid economic development as the long-term objective, means of
acquiring the foreign exchange to pay to import capital and intermediate
goods must occupy an increasingly important position. Consequently,
cotton exports have an even more important role in the economy.

In addition, during the last decade, Tanzania expanded her textile
industry and the production of cooking oil for the domestic market.
This expansion increased the country's demand for cotton lint and
cotton seed. As shown in Table 2.2, the domestic demand for cotton
lint increased by about 20 percent between 1973—74 and 1979-80. Among
the oilseed crops, cotton seed is ranked second to groundnuts (see
Table 2.3). Cotton seed is used in the manufacture of cooking oil,
margarine, and soap. Cotton seed cake, which is a valuable residue of
crushing the seed, has a high protein content and is a valuable live—
stock feed. Any future increase in the production of cooking oil,
margarine, soap, and animal feed may require asubstantial increase in
the production of oilseeds of which cotton seed is very important.

Over 90 percent of Tanzania's cotton is produced in the western
cotton-growing area (WCGA), which is composed of the regions of MWanza,
Shinyanga, Mara, Tabora, Kagera, Kigoma, and Singida (see Figure 2.1).

Shapiro, using data from the then Ministry of Agriculture, Food, and

—— _..._.._..__.-.—_ .-

12

Table 2.1

Production and Exports of the Main Export Crops

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

(310? 1968-69 1969—70 1970-71 1971-72 1972-73 1973-74 1974-75 1975-76 1976-77 1977-78 1978-79
Productiona 51,545 46,140 49,169 45,834 51,595 54,795 50,283 53,359 48,682 43,074 49,228
[.14
EExport 48,383 48,717 44,140 34,919 53,856 59,313 40,382 53,312 56,970 46,065 50,076
0
U
Valueb 265.1 255.1 312.2 227.4 383.0 495.3 375.1 483.0 1,282.7 1,857.2 1,302.1
Production 209,303 202,000 181,458 157,026 155,407 143,615 128,239 118,413 105,018 89,962 105,140
.4
<1
3 Export 169,227 217,236 160,813 154,917 112,601 93,594 101,866 90,294 67,325 78,729 76,932
(.0
Value 158.7 159.6 178.8 133.8 144.8 221.6 483.0 302.0 239.0 228.2 221.3
Production 285,472 289,134 421,332 360,116 428,033 359,139 359,139 326,264 370,441 358,637 285,706
6
[2 Export NA 281,750 338,850 297,118 354,750 329,890 269,715 209,143 316,695 232,405 256,314
8
Value 282.9 234.7 247.2 244.8 336.4 331.1 472.6 296.7 613.5 540.7 419.1
Production 17,156 113,500 111,270 121,500 122,517 145,080 117,153 83,521 96,807 68,488 NA
13
i Export 78,415 82,185 77,418 95,925 112,925 109,915 113,891 97,628 66,380 74,757 57,826
0
Value 11.4 136.4 138.7 148.1 172.8 173.9 242.8 221.0 207.4 272.0 160.9
0 Production NA NA 11,066 11,949 14,481 13,025 18,150 14,193 19,126 18,265 17,087
U
3 Export NA NA 6,947 4,783 5,396 6,116 8,831 6,372 11,184 11,561 11,294
0
[...
Value NA NA 44.8 43.1 49.2 55.5 88.1 82.3 NA 226.4 221.4
Product1on NA NA 8,492 10,457 12,706 12,658 12,974 13,732 14,074 10,415 10,946
<
EEXport 6,700 7,638 6,900 8,324 9,707 9,505 9,652 10,456 12,612 12,984 14,942
Value 45.0 48.0 42.0 49.0 53.8 45.2 69.1 81.2 134.5 177.8 168.0
Source: National Policy on Productivity, Income, and Prices, 1980.

 

a I I l O O
Production figures are given in metric tons.

Value of exports given in millions of Tanzanian shillings.

13

Table 2.2

Sales of Cotton to Local Mills

 

 

 

YEAR BALESa VALUEb
1973-74 56,186 96.241
1974—75 61,579 108.505
1975—76 59,841 105.443
1976-77 56,253 100.232
1977—78 54,913 124.911
1978-79 66,669 174.326
1979-80 70,267 232.024
Source: Tanzania Cotton Authority,.l980.

 

aOne bale is equivalent to 18 kilograms.

bValue of sales given in millions of Tanzanian shillings.

Tanzanian Oil—Seed Crops:

14

Table 2.3
Value of Productiona

1970 to 1977

 

 

 

CROP 1970 1971 1972 1973 1974 1975 1976 1977)
Castor Seed 8.7 7.9 8.9 10.3 4.7 5.6 3.0 2.7
CopraC 5.8 7.7 7.5 4.9 5.8 6.1 4.4 4.9
Cotton 52.1 43.0 66.5 72.3 68.2 32.6 86.7 75.6
Groundnuts 24.9 19.5 56.0 51.9 67.4 113.3 78.4 83.7
Sesame 11.4 10.0 9.8 12.9 19.7 22.1 18.7 20.0
Sunflower 2.1 2.3 6.2 8.7 11.5 16.5 20.2 21.2
Source: 1970-75: United Republic of Tanzania, Economic Survey,

1976-77:

 

1975—76, 1977, p. 60.

United Republic of Tanzania, Economic Survey,
1977-78, 1979, p. 59.

8Value of Production given in millions of Tanzanian shillings.

bProvisional figures as noted in the source.

CValue of marketed quantities.

15

 

 

 

‘0 __“.__lntorn Beundouus
i ...—...... Outtict BaunConus
0 lllonultox 240
L— J

)0 12

4 A

 

Figure 2.1 Cotton-Growing

  
 
 
 
  

Areas

1’ V

KENYA

O \ O
é;\\'\\\.

J’l \ wfi
Atushu “V?“ o.

    
    
   
  
 
   
 

. (7105 h:
"°‘°' \\\- lusholo ;
Put. ,1.
Iorog-o

ondoo

  
 

londenu
., 0
1
.g' °
6.

'\\0 . .
s . quogoro

   

 

 

.Most important cotton-growing area

Other cotton—growing areas

l6

Cooperatives, estimated that Mwanza and Shinyanga regions accounted
for over 75 percent of the Cotton grown in the seven regions, with
MWanza producing slightly more than Shinyanga. Within MWanza region,
Geita is the most important cotton-producing district (see Table 2.4).
Kwimba district is second to Geita.

In Geita, cotton is a very important cash crop for small farmers.
In a sample of 89 farms, Collinson (1964) found out that average net
income (cash and imputed value of household consumption) was Tanzanian
shillings (Tshs) 1355.00 of which cotton sales accounted for Tshs
684.00 or 50.4 percent. Data collected by Larsen showed that cotton
was the single most important cash crop for farmers in Geita district
and in all Sukumaland, which includes both the MWanza and Shinyanga
regions. In 1968-69, Larsen estimated an average net farm income of
Tshs 1974.00 for a sample of 219 farmers in Geita, MWanza, Shinyanga,
and Kwimba districts. Of this total income which includes the im-
puted value of household consumption, cotton sales contributed Tshs
497.00 or 33.7 percent. There is enough evidence, therefore, to show
that cotton has a very important role to play in the Tanzanian economy
and that Geita district is the most important producing area.

Production Patterns of Maize

 

An assessment of progress achieved in the food-crop sector is less
easy to undertake than that of the export—crop sector because much of
the output of many of these crops does not enter commercial-marketing
channels. Sugar and wheat are more or less entirely commercialized,

but rice and, to an even greater extent, maize are consumed on the farm.

17

0f the food crops produced and consumed in Tanzania, maize is by
far the most important. The importance is illustrated in Figure 2.2
and Table 2.5. As shown in Figure 2.2, maize is the most important
grain produced in the country. And, as shown in Table 2.5, of the
four main food crops, maize is the most purchased and consumed. Maize
is the main staple food consumed in urban areas. Most of what is pur-
chased from the rural areas is sold to consumers in urban areas. Thus,
when maize production falls below domestic demand, the government has
to import it. Unfortunately, domestic demand for maize has often ex—
ceeded domestic supply. As shown in Table 1.1, the government spends
some of its scarce foreign exchange on importing food in general and
maize in particular.

The main maize—producing regions in Tanzania are Arusha, Dodoma,
Iringa, Kilimanjaro, Lindi, Mara, Mbeya, Morogoro, Mtwara, Rukwa,
Ruvuma, Tabora and Tanga (see Figure 2.3). Mwanza region as a whole
is not a primary producing area. However, Geita district, which is in
MWanza region, is one of the most important maize-producing areas in
Tanzania.

Although maize has been grown widely in Geita district for many
years, it has been quite commercialized more recently. At the moment,
there is a competition for resources between cotton and maize. In
1963-64 Collinson (1964) found that 90 percent of the households in
1113 Geita sample grew maize and some cassava with legumes and/or sweet
1x5tatoes. Nkonoki estimated that about 95 percent of all farmers in

the district grow maize.

1700 a
1600

1500

1400 .

1300 ‘
1200 -
1100 .

1.1
O
O
O
1

900 _
800 4

700 -
600 4

THOUSAND METRIC TONNES

500 _
400

300 .

 

 

‘v

 

200 4
' SORGHUM
100 -. V¢‘,/’ ‘ RICE
:g A—TILLET

é—-CEREALS

41 MAIZE

o ‘2” MB: WHEAT

_"

‘1961 62 63 64 65 66 67 68 69 70 71 72 73

YEAR

FIGURE 2.2 Grain Production in Tanzania

 

19

Table 2.4

Estimated Cotton Production 1969-70a

 

 

AREA AMOUNT
Ifiestern Cotton-Growing Area 420,000
MWanza Region 175,000
Geita Districtb 64,000
Kwimba District 62,000
Mwanza District 45,000
Ukerewe District 4,000
Shinyanga Region 160,000
Mara Region 45,000
Tabora Region 27,000
West Lake Regionc 10,000
Singida Region 2,500
Kigoma Region 500

 

Source: Ministry of Agriculture, "Cotton Prospects" and "Preliminary

Cotton Targets," (Mwanza: Tanzania, 1970), mimeographed.

 

a
Cotton production expressed in lB—kilogram bales.
b
Includes the new district of Sengerema.

c
Now known as Kagera Region.

20

Table 2.5

\ ' o 1 1 a 1 . | 3
Recorded Commerc1al Transactions of Ma1n Food Crops

 

 

 

 

 

 

 

 

 

 

1960-62 1970-72
FOOD .8 $3 .8 3::
U 4-J U 1.1
a o. m u -H a. m u
6%? as a as a: a as
g e ”1 Eu: 8’ a $3 8‘33
o 5‘ 8 s o o a o s 0
Q (I) O U) Q Q U) U U) Q
Wheat 12 49 -37 50 64 -14
Maize 59b NA — 94 133 -39
. c d
Rice 21 33 -12 48 37_ +11
Sugar 33e 6O -27 91 123 -32
Source: H.C. Kriesel, C.K. Laurent, C. Halpern, and H.E. Lazerele,

"Agricultural Marketing in Tanzania, Background Research and
Policy Proposals" (Michigan State University, East Lansing,
June, 1970).

Ministry of Agriculture, Market Statistical Report, July 1973;
Ministry of Economic Affairs and Development Planning, Second
Five-Year Plan.

 

a O O O O .
Transactions figures given in thousands of metric tons.

b

This is recorded marketed quantities.

c
An extraction rate of 65 percent was used to convert paddy into rice.

dData for 1961 and 1962 only.

eEx-mill, excluding jaggery.

21

    

3° ‘ "010 .3- .
‘_‘— .J
lot-9n

 

__n___lntnrn Boundonos
110

.«

 

>10 District Beuncunoi
'0 3‘ g:
0 lilonultox 240 o 2
I .5
.15
30 32 P E

 

 

 

 

Figure 2.3 Primary Maize—Growing Areasa

 

8Represented by dots

 

 

[‘ \____ /~\_/

Review of Relevant Studies

 

Although attempting to provide information for government policy
makers, in most studies on agricultural production in small—farm
economies little effort is made to analyze more than one crop-—
especially crops competitive in production. In addition, at times the
whole problem is seen from the national point of view. Only secondary
and aggregated data are used to analyze production problems of indivi-
dual crops. A few such studies are reviewed below.

Malima discussed the impact on cotton production of agricultural
research and extension, agriculatural infrastructure, agricultural
pricing, and specialization and exchange. The role of extension
services is to transmit research knowledge from research centers to
the farmers. In practice, however, the coordination between research
activities and extension services has not been effective. Malima
cited two main problems with the various extension services. First
there are not enough extension workers to visit the farmers regularly
and advise them on better methods of cotton farming. Second, cotton
farmers are illiterate and cannot utilize written extension materials.
As Lewis put it, for most less—developed countries, a breakthrough in
agriculture is "especially difficult because it is necessary to influ—
ence the decisions of hundreds of thousands of uneducated peasants"
(p.45).

However, there is another problem that affects production of
cotton. Quite a number of studies on cotton production pointed out
that food production has a very important role in determining farmers'

decisions regarding adoption of the "cotton package." In his study on

 

23

the effectiveness and impact of improvement methods on small—farm
cotton production in Mara region, Keregero concentrated on cotton
production and the profit motives of farmers. He concluded that
food production has a substantial impact on cotton yield. However,
his conclusion was not supported empirically.

In another study on cotton production in Mara (Keregero, De Vries,
and Bartlett) it was concluded that the real cost to the farmer of
adopting the "cotton package" can only be determined by detailed
micro—level study covering all major crops, including food crops.
Unfortunately, this microstudy was not done.

In their assessment of recommendations for cotton production as
applied by Niegerian farmers, Norman, Hayward, and Hallam found that
"because a security strategy still plays an important role in farming
decisions, the labor peak on a normal farm occurs in June and July
when food crops are being tended" (p. 270). The authors also found,
however, that where activity on unsprayed cotton began in July and
was at a peak in August, there was virtually no demand on labor during
June. Only when farmers adopted recommended practices and sowed their
cotton in June was there considerable competition for labor. The
competition would have been much greater had oxen not been employed.
These results also showed that labor inputs (excluding labor involved
in spraying) were on the average 94 percent higher for fields receiv—
ing recommended practices than for sole—cropped unsprayed cotton grown
under indigenous conditions. Sixty—four percent of this increase came
fron1 the extra labor required to harvest the heavier yields resulting

fron1 the adoption of improved techniques.

24

Collinson (1972) pointed out that among small farmers "...an
assured food supply and personal security needed to allow the produc—
tive activity to generate survival are extremely important" (p. 21).
It follows that the quantity of supply, quality, preferred taste,
and reliability of food supply are four specific motives dominating
decision making and resource allocation in small—farm agriculture.
On the other hand, Collinson argued that in any money economy it
would be totally wrong to assume that small farmers would not be moti—
vated by profits; this desire is necessary for further achievement.

Growth of agricultural output and productivity can be achieved in
many ways. The government may invest in rural feeder roads, marketing
and storage facilities, agricultural research, extension services, and
increased water supply and may also provide cost-price incentives to
farmers. These are necessary but not sufficient in themselves.

Based on numerous studies, some of which have been referred to
above, it can be concluded that any policy that is intended to
achieve national targets in agricultural output and productivity
must address itself to understanding the behavioral characteristics
of farmers. Farm management studies should be developed that deal
with alternative farming systems that are technically feasible and
economically profitable and that have the institutional, credit,
tenure and cost-price incentives required to induce farmers to adopt
new farming systems. Such studies would furnish indispensable informa-
tion and data for economic-development planning and for implementing

policy measures on the regional and national levels.

25

Cotton growing in Tanzania, and particularly in Geita district
'is management and labor intensive. Steps in growing cotton include
land preparation; selection of seed variety and seeding rates, time
of planting and planting operations; fertilizer application; spray-
ing; thinning; weeding three to five times; picking; elaborate
sorting of dirt from seed cotton before it is marketed; and uprooting
and burning of residue. In Figure 2.4, these steps are summarized and
compared with those for maize. Thus, since much labor is required for
cotton growing, labor-saving techniques may be required.

Maize on the other hand, requires only two to three sessions of
weeding and one or two applications of an insecticide. Labor require—
ments are comparatively fewer. The problem, however, is that cotton
and maize compete for farm resources, particularly for labor. This
competition for labor of cotton against food crops is especially

apparent during the following farming operations.

Farming_Operations

 

Land Preparation

 

Well performed land preparation insures a suitable bed for seed
germination, provides young seedlings with a weed—free fertile environ—
ment, guards against erosion, and conserves moisture. Keregero found
that land preparation requires large amounts of labor and that the
required amount of labor is often not available because farmers put a
hiifli'priority on the production of food crops especially maize. Percy
identified the increased reliance on maize as the major factor account—
iru; for late planting of cotton, since maize must be planted early to

ensure food for the family.

26

 

 

 

 

 

COTTON MAIZE
Date Operation Technology Date Operation Technology
Uprooting and
Aug.- Burning of Land
Sept. Crop Residue Manual Oct. Clearing Manual
Manual or Manual or
Oct.— Land Oxen or Land Oxen or
Nov. Tilling Tractor Oct. Tilling Tractor
Primarily Primarily
Nov. Ridging Manual Nov Ridging Manual
Nov.—
Dec Planting Manual Nov Planting Manual
First First
Jan. Weeding Manual Jan. Weeding Manual
Second Second
Feb. Weeding Manual Feb. Weeding Manual
Early Third Apr.-
March Weeding Manual May Harvesting Manual
Early Fourth
April Weeding Manual
May—
June Harvesting Manual
Figure 2.4

Calendar of Crop Operations for Cotton and Maize

 

27

Time of Planting

Research at Ukiriguru (Peat and Brown) showed that timely planting
of cotton is probably the most important single factor in increasing
yields. Early planting on the average increased yields by 169 kg of
seed cotton per hectare. In Figure 2.5, the relationship between
cotton yield and planting date is shown. Early planting forms the
base for many other practices in the recommended package. Returns
from fertilizer and insecticide in late-sown cotton are far less than
from early—sown cotton (Le Mare, 1969). However, early planting re—
quires early land preparation and thus requires a large labor input,
including fertilizer application, thinning and early weeding at a time
when labor is scarce. At this time, most food crops have to be weeded,
so farmers must either hire labor if it is available and they can

afford it, or delay planting.

Weeding

Hulls found a close relationship between the time farmers spend on
weeding and yields obtained. The recommended practice is to suppress
weeds as soon as they appear. However, Keregero noted that this is the
most labor-intensive activity in cotton growing and that it competes for

labor with food crops.

Harvesting

The timing of cotton harvesting is very important because fibres
in open bolls may be rained on and, hence, turn gray if left on the
plant too long. On the other hand, the open bolls should be left on

the plant for a few days to allow the fresh fibres to dry. Since all

 

 

28

 

Yield as a 100'
percent of
earliest
sowing
804
60.
40.
20-
0 r .

 

Nov. Dec. Jan. Feb. ' Mar.

FIGURE 2.5 The Relationship Between Cotton Yield and Planting Datea

Source: W. Reed, "Problems Posed by Late Sowing of Cotton in Lake
Region of Tanganyika."- 1964, p. 256.

a 2
Polynomial fitted to results from seven trials over seasons. R =O.74

 

29

bolls do not open simultaneously, more than one picking is necessary if
all the cotton is to be picked at the proper time. Labor bottlenecks
often occur during this period.

Shapiro outlined the timing of the various tasks involved in cotton
production. The first labor peak is in the second half of November
when ridging is done. Mest land preparation is completed by mid—to—late
December. After this until the end of April, weeding is necessary.
Harvesting begins in mid-May and reaches a peak in June.

Labor allocation becomes clearer when one considers all the main
crops. Collinson (1964) showed that peak periods of labor are November
land-preparation peak, January weeding peak, and June harvest peak.

The first two peaks coincide with the time when a great deal of labor
is also devoted to food crops, especially maize. Shapiro noted two
important consequences of this labor pattern. First, the coincidence
of labor peaks for cotton and food crops strains the supply of family
labor. Almost all hired labor in the area is for land-preparation
and weeding. Second, if cotton were planted one half month earlier
(a common recommendation) the peaks of cotton and food harvesting
would coincide and thereby force many farmers to hire labor in late—
April and late-May. This would be a third period of labor hiring,
because the land preparation and weeding peaks probably would still
occur (p.25).

This means that ways must be found to increase cotton and maize
production, given the scarcity of labor (family, hired, and labor from
working parties). Technologies that are labor saving or are substi-

tutes for labor must be found. Mechanical technologies are the only

30

substitutesfor labor. Mechanical changes in agricultural technologies
generally relate to the way in which agricultural tasks are performed
by a variety of power—implement combinations. The major sources of
farm power (human labor, animals, and mechanical sources such as the
combustion engine for tractors and diesel engines or electricity for
motors) and the implements with which these sources are used must
also be considered. Thus mechanical innovations involve new farm—
power sources and new implements and as such definitely change the
ratios of inputs (man, animal, machine) being used in farm operations.

Mechanical innovations are not strictly independent of biological
innovations i.e., the introduction of new inputs such as high-yielding
seeds, chemical nutrients, pesticides, or new cultural practices such
as crop rotations, new crops, crop calendars, and optimum timing of
agricultural operations. However, in many practical instances,
biological and mechanical innovations and their impacts can be treated
separately. This isespeciallythe case when it can be shown that the
choice of a particular power—implement combination to perform any
agricultural operation does not substantially affect the crop prac—
tices or yields. It is because of this that biological innovations
have been labeled as land—saving and mechanical innovations as labor—
saving, even though this is not strictly true.

In Geita, both biological and mechanical innovations are available.
Field evidence however, indicates that mechanical innovations are un-
likely to have any yield effects; that is, crop yields can be assumed

to be invariant with respect to whether or not agricultural operations

 

 

 

31

are performed manually, with oxen, or with tractors.l They are,
therefore, treated in this study as separate technologies.

Another possible way of dealing with labor bottlenecks is by
changing the cropping calendar. This would require much research and
economic analysis to determine which crop should be planted early or
late and what would be the impact of such a change on yields, farmers'
incomes, and resource allocations. Such a farming system should then
be compared to the existing one as well as to the application of
mechanization. The comparison should be based on cropping patterns,

total outputs, net farm incomes, and resource use.

1 . .

Plowing or seeding with oxen or tractors does not ralse yields. On
the contrary, there is some evidence, though small, that deep plowing
by oxen or tractors overexposes the soil, leading to moisture and

 

fertility losses. The use of ridges preserves the moisture and is the

most widely accepted innovation in Geita.

32

Agricultural Mechanization in Tanzania:

Past Experience

There have been various attempts in the past to introduce
mechanization in the form of tractor and oxen cultivation in Tanzania.
These attempts were made with a view to relieving seasonal labor
bottlenecks and raising labor productivity.

The early experiences of mechanization with the use of tractors
indicated that, although it offered some benefits through more effi—
cient use of labor—~particularly where land was abundant and available
by just clearing the bush, the costs were too heavy and could only be
justified if mechanized farming led to sufficiently higher production,
which often was not the case. The experiences of overcapitalized farm
operations in Konjera, Urambo, and Nachingwea run by the Overseas Food
Corporation and its successor, the Tanganyika ‘Agriculatural Corporation,
in the 19503 attest to these facts. Also, the lessons learned from the
famous colonial groundnut scheme in Nachingwea (1952-56) in which 21
farms of between 200-600 hectares were mechanized are quite instructive.

The main problem with early mechanized farming centered on the high
costs of European administration and management, machinery, declining
soil fertility, and the variability in yields experienced in spite of
intensive cash inputs such as fertilizers. The main advantages were
two: the ability to perform timely operations and presumably the
ability to bring more land under cultivation. Although it is possible
to bring more land under cultivation by the use of tractors, it is not

necessarily the cheapest way. In comparing the costs of bringingimore

 

33

land under cultivation (felling and clearing), Lord found manual methods
to be clearly more economical than the use of bulldozers and tractors
(pp. 125—129).

In spite of these lessons, interest in agricultural mechanization
did not end. Several attempts were made later on. First there were
the Settlement Schemes and the block—farm mechanization schemes of
the early 19603, which are well described by Taneja, and by the Bureau
of Research and Land Use Planning. At the same time, there were attempts
by the government to introduce tractor hire services through cooperative
unions. With the policy to establish Ujamaa (Socialist) farming in 1967,
there has been increasing free use of tractors in Ujamaa Villages through
the Regional Development Fund (RDF) which provides a sum to each region
each year for the purpose of effecting rural development. Taneja
reported that in 1971 alone, the Tanzania Rural Development Bank (TRDB)
loaned over 3.7 million Tshs for the purchase of tractors and other farm
equipment to Ujamaa villages in various parts of the country. The
experience of block-cultivation schemes is of special interest because
of their apparent potential for socioeconomic development in line with
the principles of Ujamaa.

By the late 19603, the failure of tractor technologies was already
apparent. A government study carried out in 1968 reported that “the
experience with these tractors had been so bad that it must be recog-
nized that the assumptions which were made at the time that they were
introduced were wrong and no degree of tinkering with the organizational
set-up can possibly make it work (Second Five-Year Plan, Appendix IV,

p.12).

34

The main benefit of the failure of the tractor technologies was that
many farmers turned to oxen cultivation. It was later concluded that
"there was no case for the introduction of tractors in areas where
ox—plowing is established, these can perform the same functions more
cheaply" (Collinson, 1974) (p. 8).

The failure of tractor technologies led the government to look
into the possibilities of the use of ox—powered equipment in the con—
text of an agricultural sector dominated by small farmers and in a
country where a large livestock population provided the potential for
draft animals.2

Yet relatively little is known about the role of oxen technologies,
particularly their potential within the context of the prevalent farming
systems in Tanzania. Although the Tanzania Agricultural Machinery Test—
ing Unit (TAMTU) was set up in Arusha in the early 19605 to look into
these problems, its main function has been the testing and modification
through field trials of a variety of farm machinery (including ox equip—
ment) and the processing of equipment submitted to it by designers and

manufacturers. Little attention has been paid to the development and

2"We are using hoes. If two million farmers in Tanzania could jump from
the hoe to the oxen plough, it would be a revolution. It would double
our living standard, triple our product. This is the kind of thing

China is doing."- President Nyerere, quoted in W.E. Smith, We Must Run
While They Walk, 1971. Also consider the more recent statements by
President Nyerere: "...the truth is that the agricultural results have

been very disappointing. Modern methods have not spread very widely;
the majority of our traditional crops are still being grown by the same

methods as our forefathers used.... People still think in terms of
getting a tractor for their farms——even when they are small——rather
than learning to use ox-ploughs." The Arusha Declaration: Ten Years

 

After.

35

adaptation of these technologies in order to successfully deliver them
to farmers. Certainly no detailed studies of the economic implication
of alternative technologies at the farm level have been carried out.

The only way of ascertaining that oxen, as opposed to manual
cultivation, were found economic in many cases was by the number of
oxen plows purchased by small farmers in the year following this change
in policy. In 1967, some 8,250 plows were bought; this increased to
9,100 in 1968 and to an estimated 15,000 by 1974. In some areas,
particularly Sukumaland which ineludes Shinyanga and MWanza regions,
it also led to a significant increase in cotton hactarages and to an
increase in farm incomes. It may also have generated additional em—
ployment although it was not possible to substantiate this.3

The government's cautious views on mechanization expressed after
the failure of tractor technologices in the 19608 did not last long.
They were overtaken by the events following the Iringa Declaration in
May, 1972 and the general speeding of planned villages as well as
Ujamaa villages, which took place from mid—1974 onward. The Iringa
declaration put emphasis on the modernization of farming and farming
techniques through gradual collectivization. This was interpreted in
practice to mean development of villages through a promise to make
tractors and other modern inputs available. Distribution of tractors
to selected Ujamaa villages was, therefore, continued.

In 1974, a FAQ mission reviewed the mechanization situation in

Tanzania and found that the TRDB gave villages loans to purchase

3See Clayton. However, the increase in hectarages on block farms
cannot be attributed to use of oxen. See Collinson, "Cotton
Development in Tanzania: A Review of the Cotton Program in
Sukumaland" 1974.

36

tractors and that villages were allocated tractors purchased with
rural development funds (RDF). Regions or district tractor—hire
services operated by the regional authorities or by such parastatal
organizations as the Tanzanian Cotton Authority were expanded. This
was done even when villages were not capable of supporting tractor
mechanization by subsidizing it in a variety of ways. These included
the nonrecovery of loans made to cooperative villages, the semiper-
manent loan of tractors to villages that had neither the capacity nor
the desire to meet any costs other than for fuel, the offering of
uneconomic and subsidized sales of hire services and through the
waiver of dues for hire services (FAO, P. 10).

On the other hand, the mission "found it extremely difficult to
find any concrete evidence in the regions of a concerted effort to
promote a wider and more intensive use of animal power. The program
of establishing ox—training centers, mentioned in the Second Five-

Year Plan (1969—1974) appeared to be virtually non-operative and little
technical progress had been made at the village level beyond the use of
the single harrow ox—plow (FAO, p. 9). The only exception was the work
done by TAMTU in developing an improved and wider range of animal—
drawn equipment and in fostering self—help at the village level by
training artisans to manufacture simple equipment. However, as men—
tioned earlier, the impact of TAMTU was limited not only by other
responsibilities given to it but also by the inability to either meet
the demand through its own manufacturing or to distribute and provide
services for the equipment it had tested or developed. Thus, in

spite of the importance of oxen as sources of farm power in certain

37

districts of Iringa, Shinyanga, Arusha, Mbeya, Mara, and MWanza regions,
there was "no area in which mechanization with oxen had progressed be—
yond the use of the simple harrow plows" (FAO, p. 13). Thus, in the
Mission's View, ”ox—equipment has yet to be given a real chance to
prove itself, even in.those districts where there is a tradition and
aptitude for its employment" (FAO, p. 13).

More recently, the United Nations Development Program (UNDP)
carried out a study that carefully compared manual—, oxen—, tractor—
cultivation methods by using simple partial—budgeting techniques. This
is one of the few available studies in which an attempt has been made
to compare the economics of alternative forms of farm power, either
manual, animal, or tractor, on the basis of carefully obtained cost
data. It was concluded that both at the family—farm and Ujamaa—village
levels, the promoting of comprehensive oxen techniques was most likely
to offer the greatest advantages. It was also recommended that on
larger, more successful Ujamaa villages, "well operated and maintained
power equipment" used for minimal cultivation techniques offered the
only feasible form of mechanization and that too in many cases in con—
junction with oxen—cultivation" (Beeney, p. 5).

In short, past studies suggested that under current conditions,
Tanzania should discourage tractor mechanization except in very special
circumstances after careful study of its economic efficiency and viability.
On the other hand, oxen cultivation should be further explored for its
potential to relieve labor bottlenecks. However, there is some justi-

fication for additional studies on the subject.

38

First, the little evidence there is from past studies on
agricultural technologies is based on partial budgeting methods
by which the impact of various technologies outside and apart from
the economics of the existing farming system (which include sub-
sistence and cash crops as well as livestock activities that compete
for farm resources and provide alternative sources of income and
employment) are analyzed. It is important to know how different
technologies will change the economics of such a farming system and
not just whether or not a particular task can be performed or a
particular crop grown at a lower cost by different methods.
Partial budgeting methods fail to analyze the total farming system.

Second, past studies put emphasis on analyzing only the economics
of various forms of mechanization. However, it is equally important
to emphasize yield-increasing technologies based on the application of
improved seeds, nutrients, pesticides, and cropping practices. Such
an integrated approach to biological and mechanical technolOgies and
their implication for the present farming system is needed.

Third, and particularly relevant to this study, is the need to
make a comperative study of this subject in the context of both the
household farm and block farming, which is now emphasized in Geita

and other cotton-growing anthobaccoegrowing areas.

Block Farming and Mechanization

 

In the future, most cotton in Tanzania is expected to be grown on
block farms. In block farming, each household owns farms that are

arranged close to farms owned by other households. The household farms

39

are separated by foot paths, and each block may hold from 20 to 40 hec—
tares. The economic considerations of block farming are as follows:

1. When farms are close together, they can be more effectively
attended by a small number of extension personnel. Improved husbandry
techniques, such as early planting, fertilizing, and spraying, can be
demonstrated on the participants' own plots.

2. It is more economical and easier to mechanize such farms.
Capital is a scarce production factor in Tanzania. Consequently, there
is little economic advantage in substituting capital for labor. The
main justification for the use of tractors is that they can supple-
ment already fully engaged family labor in the most labor—intensive
operations. In other words, rather than have 50 families cultivate
four hectares each, it would be desirable to have more participants
who would own one, two, or even more hectares on the block.

3. With regard to vermin, one farmer can look after 20 or so
hectares of cotton or maize per day on a ratational basis, instead of
each farmer looking after his own farm throughout the season.

4. Block farm schemes are also expected to insure economic and
proper land use and provide a much needed check against soil erosion.

Politicians View block farming as an approach that takes the rural
sector one step nearer to communal farming——an ultimate objective of

Tanzania's socialist policies.

Past Experiences in Mechanized Block Farming of Cotton

 

In 1964-65, the government introduced block cultivation schemes

in Mwanza and Shinyanga regions. In that year, the government turned

40

over to the then Victoria Federation of Cooperative Unions 159
tractors. The main objective of establishing block-cultivation
schemes was to help farmers use tractors economically by consoli-
dating cotton cultivation into large blocks. The main assumption
in which farm blocks were established was that labor was scarce in
these regions and that additional land could only be cultivated if
tractors were used to complement the labor: dAnother assumption
was that close supervision would improve farming practices. For
example, the use of fertilizer could be readily demonstrated by the
extension service with resultant improved yields. Klien, Green,
Donahue, and Stout reported that increases in cotton yields from
around 450 to 1100 kg per hectare were anticipated through the super-
vised use of fertilizers, insecticides, and greater cotton varieties.
A third assumption was that economies of scale could be effectively
realized if the land were farmed in large blocks. Between 1963 and
1965, the government imported about 673 tractors and distributed them
to cooperative unions throughout the country to use on a hire—service
basis.4

The outcome of the scheme was disastrous. A study by Amarshi
revealed that out of the 41 cooperative unions involved, only one
achieved any measure of success and about 65 percent of the Tshs 18

million spent on the scheme plus Tshs 2 million of arrears in interest

4Although the regional distribution was fairly widespread, nearly 41 per—
cent of the tractors (277) were given to Mwanza and Shinyanga. See
Report of Working Party No. 12 on Agricultural Mechanization, Second
Five-Year Plan, June 1968. Appendix 1.

 

41

payments were never recovered from the unions. Even "the one
successful cooperative union...foresaw the difficultires of running a
hire—service and, therefore, sold the tractors to its richer members
instead, on a hire—purchase basis" (Amarshi, p. 69).

A.number of studies cited the several factors that led to the
failures as follows: poororganization,inadequate skills for repair
and maintenance, unproductive and inefficient use of machines, high
costs of clearing land and inadequate increase in production. Klein,
Green, Donahue, and Stout noted that in Mwanza and Shinyanga regions,
supervision proved to be extremely difficult.7 This was so even though
40—65 percent of all extension workers in the country were employed on
the scheme and only a handful of the total farmers in the two regions
were affected. In addition, in spite of basing standard depreciation
estimates on 1,200 operating hours per tractor per annum, constant
losses were experienced on tractor operations due to inefficient or
inadequate use.5

Pupius argues that the form of agriculture under the block
schemes involved a shift from "low cost production, profitable with
low returns, to high cost production requiring high returns" (p. 4).
Under the block—farm schemes, a substantial improvement in crop yields
was required. Yet, as Heijnen found out, the yields were far lower

than anticipated and varied enormously between various blocks and

5Collinson (1964, p. 6) estimated a range of between 575—825 tractor
operating hours per annum, but evidence suggested that tractors did
not reach even this level of operation. De Wilde (1967, p. 439)
noted that of the 16.5 thousand hectares on 50 block farms, only 33
percent was cleared and 19 percent planted, amounting to an average
of 27 hectares per tractor of planted area.

42

individual plots within blocks. He came to a conclusion that in

order to meet the higher level of operational costs, cotton yields

of between 925-1200 kg per hectare were required. But cotton yields
varied between 300—465 kg per hectare in MWanza and Shinyanga regions.
In many cases, these could not even cover the cultivation costs.
Heijnen recommended that in order for higher yields to be achieved,
mechanization of agriculture should go hand—in—hand with the following:

1. Establishment of good soil fertility on block farms.

2. Areas selected should be suitable for cotton—growing and
mechanical cultavation.

3. Good crop husbandry would be necessary; i.e., late planting,
poor and late weeding and/or thinning, inadequate fertilizing,
and expensive spraying must be avoided.

In other words, yeild—increasing technologies must be emphasized when
a shift is made from low—cost production, profitable with low returns,

to high—cost production requiring high returns.

Tillage Tools Used in the Production of

Cotton, Maize, and Other Crops

There are three distinct groups of farming tools in Tanzania:
(1) primary tillage tools; (2) secondary tillage tools; and (3) weed—

ing tools. Each of these is discussed below.

Primary Tillage Tools

These tools are used to break and open up the land for subsequent

oPerations. These tools include:

43

1. Hand hoes. Hand hoes are used by many farmers in the country.
Where ridge cultivation is common, such as in MWanza and Shinyanga
regions, the general practice is to push all of the trash in the
furrow previously made. Then ridges are reformed on these furrows
when the trash is completely covered. Where plain cultivation is
practiced, the trash is first collected and burned before the land
is tilled. This practice, beside leavingthe soil unprotected from
erosion and sunshine, also increases organic matter decomposition.
With careful tilling and avoidance of burning, it is possible to
have a somewhat rough and trashy field. The tool, however, limits
extensive crop production because of low capacity and high human—
power requirements.

2. Mbuldboard plows. Presently, this is the only animal—pulled tool

 

used for primary tillage operations. The plow cuts, shatters, and
inverts the soild. Thus, it can pulverize the soil, but the extent

of pulverization depends on amongvother factors, the type of the
mouldboard. The tool requires very little trash, if any, in the field
for efficient operation. What trash exists is turned under and covered
by the soil. Hence soils from a field worked with this tool are un—
protected from rain and sunshine, leading to erosion and loss of
moisture by evaporation. Also such a condition is ideal for quick
decomposition of the organic matter. Moreover, the operation of this
tool is limited in areas with heavy soils and moisture because of that
stickiness or hardness. In relatively light soils, this plow can
Operate when the soil is dry. The result can be big clods that are

useful for protecting the soil from wind erosion.

44

Although there have been attempts by the Ubungo Farm Implement
Manufacturing Company (UFI), set up in Dar es—Salaam in 1970 with
Chinese aid to manufacture agricultural tools, to produce a mould—
board plow that can work successfully under all conditions, varia—
tions in local conditions have made it impossible.

3. Disc plows. This plow requires tractor power. Like the mould—
board plow it cuts, shatters, and inverts the soil. But unlike the
mouldboard plow, it leaves rough seedbeds and it can work under Stumpy,
stony and trashy conditions. The disc plow covers the trash in the
field but not as completely as the mouldboard plow.

4. Heavy-duty disc harrows. Harrows are sometimes used to open up

 

land. A harrow consists of two or more gangs of discs usually pulled
behind a big tractor. The tool pulVerizes the soil more than the disc
plow does. It requires large tractors with high horse—power output.
The whole system——tractor and harrow——is very heavy and can compact

the soil, rendering it impermeable.

5. Chisel plows. In some parts of Tanzania, chisel plows are common
tools for primary tillage. This kind of plow consists of a framework
on which either spring tynes of rigid shanks with soil—working points
are attached. The soil-working points are the pointed—teeth duck feed
and sweeps. The chisel plow cuts and shatters the soil but does not
invert it. It can work in fields with trash without covering the trash,
so it is a useful tool for protecting the soil from erosion and
sunshine.

6. Rotavator. The rotavator is not a common tool among small farmers,
but it can be found on some of the state farms. Where fields have pre—

viously been cultivated, it is used for primary tillage; otherwise it

 

45

is used in secondary tillage operations. It cuts and pulverizes the
soil, and it also incorporates the trash. Susceptibility of the soil
to wind and water erosion is increased when this tool is used. In
addition, decomposition of organic matter becomes rapid. Moreover,
high power is needed for operation of this tool.

7. Ridgers. The ridger is mounted on a tractor or pulled by animals.
It is used to make ridges either on cultivated or uncultivated land.
The ridger cuts and inverts the soil also but not to the same extent
that the mouldboard plow does. If properly used, it can leave some

amount of trash on the surface of the field.

Secondary Tillagg Tools

 

These tools are used after primary tillage to pulverize the soil
more, kill weeds, level the surface, and pack the soil into a firm
seedbed. The goal is to have a smooth trash—free field ready for
planting. Mechanical planters, in large—scale farms particularly,
are designed to work on such well-prepared fields. The tools commonly
used for secondary tillage include the following:

1. Disc Harrows. These harrows are available in various sizes but

are generally divided into trailing and mounted disc harrows. They are
used to pulverize the soil more, leaving it in afinetilth for the re—
ception of the seed. So they cover most of the trash, compact the soil,
and destroy the soil's physical properties. This leaves the soil very
susceptible to erosion, loss of water by evaporation, and high organic—

matter decomposition.

46

2. Tyne harrows. These harrows consist of a frame on which teeth are
attached. In spike—tooth harrows, the teeth are pointed and rigid,
whereas in spring-tooth harrows, teeth are curved and have a spring
mechanism. The spring—tooth harrows work efficiently even in rough
and trashy conditions. But harrows break the soil clods, creating a
fine seedbed. These harrows, and especially the spike-tooth harrows,
have a tendency to rake the trash.

3. Rotavators. These tools thoroughly cut and mix the trash while
pulverizing the soil more. They create a condition for rapid organic—
matter decomposition and also destroy the soil's physical properties,

leading to more erosion by water and wind.

Tools for Weeding

For weeding purposes, there are tools called cultivators.
Unfortunately, these tools are inadequately developed in Tanzania not
only on small—scale farms but also on large—scale farms. The few that
are available often lack steering and easy adjustment. Because of this,
destruction of crops by ripping is common. Weeding tools extensively
used in Tanzania are as follows:
1. Hand hoes. The most common hand hoes are light and short.
Depending on how they are used, they can produce a rough trash field
suitable for controlling soil erosion. But because of their high
human—power requirement, their capacity is limited to a few hactares.
2. Tyne implements. Like the chisel plow, these tools consist of a
framework on which shanks are attached to suit the crops to be weeded.
The soil—working points are fixed. For this kind of implement, there

must be completely trash—free conditions to avoid clogging and ripping

 

47

off the crop. Their ground clearance is often not enough, causing
easy clogging. Thus, they are efficient for weeding only in early
stages of the growth of the crop.

With enough ground clearance, the sweeps fitted to the shanks
have proved to be successful in working in trashy fields. They
penetrate the soil and cut the roots of weeds with very little, if
any, inversion and coverage of the trash. Unfortunately, these
implemetns are seldom used in Tanzania.

3. Ridgers. Ridgers are found only in a few areas of the country.
Although their main function is to make ridges, they can also be used
as weeding tools. They form small ridges on the crop rows covering
the weeds within the rows. But in a young crop where there is no
proper vegetative cover, the soil is exposed to sunshine and rain—
drop impact.

There are three main conclusions that can be drawn from the dis—
cussion of farm tools in Tanzania. First, most of the tools are
primarily used in preparing the land for planting. Other farm opera—
tions, such as weeding, still depend on manual technologies, with the
hand hoe playing an important role. Second, many of the tools require
literate farmers who can use the tools properly so as to avoid negative
impacts on their soils. Lastly, some of the tools, especially those
requiring large tractors with high horse—power output, are unlikely to
be profitable on small farms. This then calls for further consideration
of policy issues related to mechanical technology for small farms that

experience labor bottlenecks in some farming operations.

48
CHAPTER III

THE AREA SELECTED FOR THE STUDY

The data for this thesis was collected from Geita districtl during
the 1979—80 farming season. A sample of 80 farm householdswasselected
for this study. In this district, cotton is the main export crop,

whereas maize is the main food crop.

Description of the Geography and Climate of the Study Area

location, Sizgj and Typography

Geita district lies between 20 and 30 south latitude, and 320 and
330 east longitude. It is bounded on the west by Biharamulo district,
on the east by Smith Sound (a part of Lake Victoria), on the north by
Lake Victoria, and on the south by Kahama district. It is estimated
that Geita has a total land area of 9067 square kilometers. Much of the
district is characterized by gently rolling hills or ridges, which lie

approximately 1220 to 1370 meters above sea level.

Climate

The district receives an annual rainfall of between 940 and 1190 mm.
This amount of rainfall is generally more than that received by areas to
the east, where the Ukiriguru Research Center has recorded a 37—year
average annual rainfall of 847 mm. The rainfall pattern is marked by
definite dry and wet seasons, the latter showing a slight bimodal
distribution. Early rains begin in late August and September. The
first peak is reached in November or December, followed by a slightly

K

1

In this study, Geita district includes Sengerema district, which was
recently created from a larger Geita area. Both districts are equally
1Important for cotton and maize.

49

less wet period in January and February. The second peak comes in March
or April followed by a decrease in rainfall and then the dry season in
June and July.2

This rainfall districution (see Table 3.3) is well suited to cotton
farming, which needs about 760 mm of rain during its six—month growing
season. Cotton planted in late November or early December most likely
receives the proper rainfall distribution and thus avoids damage that
rain can cause to exposed cotton fibres.

The temperature in Geita is good for agriculture. Anthony and
Uchendu noted that during the cropping season, the mean maximum
temperature at nearby Ukiriguru is about 83oF and the mean minimum is
about 63oF. During July, the coldest month, the mean mimimum is 58°F.

Spilg. Most soil in the district is derived from granite and is
of a sandy nature. Generally, the soils are of average fertility.
Malcolm listed the following soil types in the granitic catena in order
from the top: isanga,kikungu,luseni, itogoro, mbuga (local names).
Samki gave the same list but labelled the highest layer skeletal soil
rather than isanga.

Isanga is described as "a coarse—grained sandy to gravelly soil of
a light reddish colour derived from granite with sporadic laterite,
which is a red ferruginous rock forming a surface or subsurface cover—
ing in some areas" (Malcolm, p. 176). The soil is said to be the most
favorable for Bambarra nuts and is also suited to sweet potatoes,
millet, cassava, and groundnuts.

E

2

The rainfall pattern in Sengerema, Geita Town, and Ukiriguru weather
Stations is shown in Table 3.1. The mean annual rainfall for Geita
and Sengerema districts are shown in Table 3.2.

50

Samki described Kikungu soils as being ”dark brown to dark reddish
brown, loamy sand to sandy loam...(with) low holding capacity? (p. 1).
These soils Show an advanced stage of weathering. Kikungu and luseni
are the most important for cotton growing. Kikungu is also suited for
cassava, millet, sorghum, groundnuts, and bambara nuts.

Luseni is a grayish brown fine sand that is low in organic matter
and is strongly leached. Apart from being suitable for cotton, Luseni
is also suited for cassava and sweet potatoes. -Itogoro is a sandy
clay—loam that is very soft when wet and forms a thin hard crust when
dry. It is the most favorable soil for cowpeas and is also suitable
for sorghum, cotton and sesame. However, itogoro is hard to work and
is now used mainly for grazing.

Samki described mbuga as being a clay or a clay—loam; the water
table is close to the surface and impedes drainage. Mbuga is high in
organic material and receives nutrients from the soils higher on the
slope. This soil is usually wet and hard to work in the rainy season.
Where it is found, mbuga soil is used for rice growing and grazing.

In 1966, the Ministry of Economic Affairs and Development Planning
claimed that of the 1,018,000 hectares of land in the district, 909,000
hectares were cultivable, 100,000 were in forests, and 9,000 were pri-
marily very rocky or swampy areas. Fifty percent of the 909,000 culti—
vable hectares was cited as having good soils (such as luseni, itogoro,
and kikungu) that would need little work prior to normal land preparation.
The other 50 percent was listed as having exhausted soils, heavy clays
that present problems to current technologies, waterlogged soils, or

soils overrun with weeds. The recent estimates of land and water areas,

51

Table 3.1

Monthly Rainfall at Sengerema, Ukiriguru, and Geita Townsa

 

MONTH SENGEREMA; SENGEREMA; UKIRIGURU2 GEITA TOWN3
1969 1970 (avg. for 1931— (period not
October 1967) specified)

 

 

January 173.44 101.53 95.71 101.28
February 184.32 79.76 89.63 84.32
March 193.44 111.66 128.62 120.77
April 86.85 188.38 145.59 155.72
May 114.69 84.57 68.87 87.35
June 0.51 0 8.86 11.14
July 0 O 2.03 1.01
August 0 72.67 11.90 12.41
September 76.72 33.93 23.04 31.90
October 18.99 66.84 46.84 85.33
November 150.40 109.64 114.69 136.47
December 109.64 111.66 111.66 122.04
Annual

Total 1,110.03 989.76 847.46 949.50

 

Sources: 1Unpublished daily rainfall records at the Sengerema sub
district station of the Tanzania Ministry of Agriculture
(K. Shapiro).

2Cotton Research Corporation, "Progress Reports from Experiment
Stations, Season 1967—68, Tanzania Western Cotton—Growing Area,"
1969, p.4.

3M.P. Collinson, "Farm Management Survey Report No. 4," p.3.

aRainfall figures given in millimeters. One millimeter is equal to 0.0394
inches.

52

.Hammcfimu mo m2mw woweoommw

 

 

 

 

 

 

 

 

 

 

 

 

.m.a :.owl@nma paw mmlwmma How uyommm Gowumnam>m uommoum: .uumnoum Gouuoo wufiwo "mousom

No mmo.H o2 mow.H mm 22H.H oq mwm mm Noo.H 02 www 20 me Hammcflmm

cam:

Hmncn<

ms mmH.H mm oom.~ 92 com am use mm moH.H so Nmm 02 «Ho mamumwcmm

No mow 02 0mm ow Nwm.a mm 202 no com mm 220 mm 0mm mufimw

m2ma .22 mama .22 mama .22 wsmo .22 mama .22 meme .22 mama .22 aonemHa

mo 2 mo 2 mo 2 mo 2 mo 2 «02 mo 2
AA<mzH<m
z<mz wnma nnma omaa mnma whoa whoa

m<mwlem

 

 

 

 

 

 

 

 

Hammcamm no «meme can muwumafifiafi2 ca wk-mso2 Hamuaamm Hmsaq< cmmz

N.m wanes

53

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

a a .
.N..Q ...Omlwhwfi mufim. ®lefi®fi POM UHOQQM COMUQDHNNVM UUQWOHWZ ome UUQWOHP COUUOU QUflwU . OUHDOW
1 _ _ 1 1
1 1 1 1 _ 1 K. ..v.
! 1 _ fl 11 1 17. n
_ 1. 1 _ 7w I.
1 _ 1 1 1 _io
z: 1 1 a: 1 m: 2: EH 2: 1 ...
1 1 v w
, 1 A
1 . 1
_ 1 Nu.—
0.002 0.23 wan.” _ 002 m0: 1 NT»; m 0.02 _ 0.0: Noné _ 0.02 4.3 000 0.03 0.2: an; 0.03 0.00 000 ....
U 1 a 1 1 G r.
1 , A 1 mm
n4: “.3 02 H 0.: ' odm :N 1 ~.0 ~.~n as m m néu 3 0.3 0.3 EN N6 1 0.2 cc d4.
. 1 _ 1 nu
, _ 1 _ x u.
1 _ _ 1 1 w.
M 1 u _ 1 u m
1 1 _ _ 1
_ , _ m 1 1 mm
2.4m m.n: ,1 03 1 Tmn w «.02 $2 1. 9mm N60" 03 0.0M _ «.02 man TR 002 m~n 0.3 m 2.00“ omm WW
1 1 1 1 1 "m1
1 1 1 _S
1 1 1 1 1 1m
1 1 1 1 1 1mr
1 1 1 1 1 1 1 1 1mm
0.: 1 ~.~0 1 J: 0.2 q a: mom 1 4.: W 900 N 00~ u 53 m ”.00 03 .1 0.0 0.3. 02 T: 1 0.02 «am Emu
1 1 1 1 1 1 1 1 1 1 1 1 11w
2 1 1 _ _ 1 _ _ . . 1N .
_ 1 . 1 1 1 1 1 1 w
_ h 1 m u ,1 1 H 1% o
_ . 1 1 1 _ 1 1 1 1 103
. 1 . 1 _ _ n 1 1 1 1H n...
2.2” _ mémfi 0U 1.2mm _ 2.0.: ~04 11 0.02 1 m0: 1 non _ a on ,1 2.02 _ men 1 4.0.” n62 . 03 2.2“ 1 2.02 v men .WW
1 1 2 1 1 1 1 _ 1
1 1 1 1 1 , 1 _ 1 1 1 1 1m...
_ 1 1 1 _ 1 i _ 2 . . .S
1 a u H + , 1 _ 1
1 1 1 1 . 1 1 , .
A3 _ 75: 1 2: 1 Airy 11 1 AC _ 75.: .1 33 . 7:5 1 u 2: 7:5 1 A3 a 2.5: 2 _
.«o x 322.321 ...).12 .«o 2 ”3.3921 7sz 1 20 N “$831,: 75: .«o N “owns—oz. Xi“: “ 20 N emuuwi. 2.2—.3 .3 N .wm=0>< H 2.5: _
u: 23 725505 Hauow mm. 2:55:05 “much 11: C: 13:23:. 230—. as C: “35.5: 1:03. _n- 30 2225:. 133. an 20 155:0: 14.302. 2 h
A: _1 30 , TC ,3 _ 2.: m 15 H :0 _ C: _, AC :0 _ 2.: _ 30 2 CV 30 11 23 CV 1 30 11 A3 " w
1
m3N>< infilq 1 a“: m :2” onoa _ mnoH ”mus—P1: x<fl>|oN “
1 1 1
WHUUQEWHHHE CH mNOH OUCHm HHNWGHMM MO COHUDDflHUWflQ HNGOmmwm

m.m mapme

54

and land distribution within the Geita Cotton Project are shown in

Tables 3.4 and 3.5 respectively.

Population

The 1978 TAnzanian Census reported a total population of about
17.5 million. Of this population, 558,500 live in Geita district
(including Sengerema). The census reported an overall population
density in Geita district of 61.5 people per square kilometer.
However, the density is generally actually higher since the popula—
tion is concentrated in villages that were established during the
1974-76 period. Most of these villages are found along roads. The
average household size in the country is 4.7 people, whereas that in
Geita is 7.1. It is estimated that there are 86,000 farms in the
area. In Table 3.6 is summarized the population of Geita since
1934. In Table 3.7 is summarized the 1978 population distribution by
age and sex. It is clear from Table 3.6 that, since the population in
1978 was tentimes more than that of 1934, the average area of land per
person decreased by 90 percent. The growth in population was due to
migration of people from the dry areas of Shinyanga and Mwanza regions.

There are three important occupations that influence the settlement
pattern; these are agriculture, by far the most important; trading; and
mining. Geita and Sengerema twons, with populations of a little more
than 7000 and 13000 respectively, are the main urban settlements.

They are also the administrative headquarters of the two districts.

 

 

 

55

Table 3.4

Estimated land and Water Areasa

 

 

 

 

 

GEITA SENGEREMA TOTAL AREA
AREAS
2 Z of 2 Z of Z of
km Total km Total km Total
Area Area Area
Total Land
Area 6,752 69.6 2,315 32.5 9,067 53.9
Total Water
Area 2,953 30.4 4,817 67.5 7,770 46.1
Total Area 9,705 100.0 7,132 100.0 16,837 100.0
Cultivable Land 3,078 31.7 1,328 18.6 4,406 26.2
Grazing Land 3,674 37.9 987 13.8 4,661 27.7
Total Land 6,752 69.6 2,315 32.5 9,067 53.9

 

 

 

 

 

 

 

Source: Geita Cotton Project, "Project Evaluation Report for 1978—79
and 1979—80," p.2.

a . .
Areas given in square kilometers.

 

 

56

Table 3.5

Land Distribution by District

 

 

ITEM UNIT GEITA SENGEREMA TOTAL
Total Land 2
Area km 6,752 2,315 9,067
Cultivable 2
Land km 3,078 1,328 4,406
Families No. 48,183 41,056 89,239
Population No. 307,233 251,283 558,516
Family No. Per
Concentration km2 7.1 17.7 9.8
Population No. Per
Density ka 45.5 108.5 61.5
Cultivable Land
Per Family Hectares 6.39 3.24 4.93
Total Available
Land Per Person Hectares 2.19 0.93 1.62

 

 

 

 

 

Source: Geita Cotton Project, "Project Evaluation Report for
1978-79 and 1979—80," p.3.

57

Table 3.6

The Population of Geita Since 1934a

 

 

YEAR POPULATION POPULATION DENSITY AVERAGE SIZE OF LAND
('000) (per square km) Per Person (in hactares)
1934 55.8 6.2 16.23
1944 87.2 9.6 10.40
1947b 108.2 11.9 8.38
1957C 270.0 29.8 3.36
1960 285.0 31.4 3.19
1966 320.0 35.3 2.83
1967d 330.0 36.4 2.75
1970 371.4 40.9 2.42
1978e 585.5 61.5 1.62

 

Source: Bureau of Statistics, Statistical Abstract, 1980.

a . .
PrOJected estimates

b
’ C’ d’ eBased on respective year population census.

58

.msmcmo coaumfisaoa wmaa do wmmmmm

 

 

 

 

 

 

 

 

 

 

 

 

 

 

o n
wuuoawu vwcmwansmc: mawumwcwm 0cm mufiwo .wofiwwo unmeaoam>om uuwuumfio "mousom
66 6.662 626.666 622.622 262.262 6.662 662.262 666.622 666.262 6.662 662.266 662.662 662.662 26262
22 2.6 622.26 662.62 622.22 6.6 622.22 666.2 622.6 6.6 226.62 626.6 262.22 2666 6:6 66
22 2.62 666.26 266.66 662.26 2.62 222.22 266.62 662.62 6.22 626.62 266.62 662.22 26-66
662 2.62 666.22 626.26 622.26 2.22 262.26 666.62 262.62 6.22 622.22 626.22 622.22 26-62
662 6.62 626.262 262.26 666.62 6.22 266.66 662.66 226.66 6.62 626.66 226.62 662.62 22-62
662 2.66 226.662 626.66 226.66 6.66 666.66 266.62 626.22 6.66 262.222 666.66 626.26 6-6
use: .0062 6:520 260226266... 2:623. .0062 0:95 263an mg; .0022 6225.20 .6335..— code:
662 .66 26266 666 66 26266 666 66 2.666 666 66
62622650...— 60 N 236.2. 60 N 238 me x 2.30.2.
3.30.2 5v
666666 66<
2622626266 26266 <2m666266 <6266

 

 

 

 

muowuumfla 2n c06umasaom mo cowusnfiuum6a wm< 0cm xwm

n.m ma

662.

59

Agriculture

 

The physical environment--rainfall, temperature, topography, and
soils——of Geita permits a variety of crops to be grown. However, the
most important crops are cotton, maize, and cassava. Cattle and other
livestock such as sheep and goats are kept by many households, but
not much use is made of their manure nor their draft power. The long-
handled hoe is the universal cultivation tool. Shapiro in his research
sample of 76 noted that no farmer used an ox plow. There is a small
number of privately owned tractors available to private farmers on a
hire basis, but their use is very rare; in Shapiro's research sample,
only one farmer hired a tractor. Because of foreign—exchange problems
and the government's deliberate policy of controlling tractor—hire ser-
vices, even the small number of privately owned tractors has sharply
decreased. Most tractor-hire services are now provided by the Geita

Cotton Project (see Appendix A).

Main Crops

As previously mentioned, cotton is by far the most important cash
crop in the district. Maize and cassava are the two most important food
crops, followed by sweet potatoes and rice. Legumes, such as groundnuts,
beans, and cowpeas, as well as garden crops, such as cabbages, tomatoes,
bananas, and citrus fruits, supplement the above-mentioned starchy food.
All food crops also serve as cash crops although in a minor way for most
farmers. From 1975-79, however, maize was highly commercialized. Food
shortages in the country coupled with the government's emphasis on in—
creased maize production in the area led to a competition between cotton

and maize. Allocation of land to major crops has changed somewhat in the

60

last decade or so (see Table 3.8a). Currently, the crop mixture is
slightly different (see Table 3.8b). Since the move to planned
villages in 1974, followed by the government's direct involvement in
land allocation and emphasis on maize production, many farmers now

grow maize as a single crop. Millet and sorghum have become less
important among the starchy staples. Collinson reported that only

5 percent of his Geita sample grew either millet or sorghum. Shapiro
found that less than 10 percent of the farmers in his ample grew either.

In the sample for this study, no farmer grew millet or sorghum.

Cultivation Practices

 

Most, if not all, farmers plant their crops (except rice, vegetables,
and fruits) on top of ridges that are about 5 feet apart and about 18 to
24 inches high. The exact dimensions differ from crop to crop, those
for cassava being greater than those for cotton. Although ridge culti—
vation is very labor intensive, it has many benefits.

After harvesting, previously cultivated land is left untouched
until the rainy season is about to start. Then the farmers clear off
the weeds and other plant materials and put them into the furrows to
dry. The old ridges are then split in the middle and hoed into the
furrows on either side. The weeds and other plant materials that had
been hoed into the furrows are buried and the field is leveled. As
the rainy season begins, new ridges are built up over the dead weeds
in the furrows of the former ridges.

All farmers practice ridge cultivation. It is a traditional prac—

tice in Shinyanga and MVanza regions. The government has not attempted

61

Table 3.8a

Allocation of Land to Major Cropsa

 

PERCENTAGE OF MEAN OVER MEAN OVER
CROP OR MIXTURE OF CROPS FARMS GROWING SAMPLE GROWERS
THE CROP(S) (hectares) (hactares)

 

Cotton 93 1.61 1.73

Maize and Cassava with Legumes
and/or Sweet Potatoes 90 1.11 1.23

Maize with Legumes and/or

Sweet Potatoes 46 0.44 0.95
Rice 42 0.11

Old Cassava 40 0.43 0.26
Sweet Potatoes 24 0.04 1.08

Cassava with Legumes and/or
Sweet Potatoes 13 0.07 0.55

Sorghum or Millett with Legumes
and/or Sweet Potatoes 5 0.05 1.01

Total 3.86 6.96

 

Source: M.P. Collinson, "Farm Management Survey Report No. 4," p.8.

aAverage hectarage per farm in Lwenge Primary Society Area, Geita
district, 1963—64, sample of 89 farms.

62

 

 

 

 

Table 3.8b
Allocation of Land to Major Crops!
CROP OR MIXTURE PERCENTAGE OF MEAN OVER MEAN OVER
OF CROPS FARMS GROWING SAMPLE GROWERS
THE CROP(S) (hectares) (hactares)
P NP P NP P NP
Cotton 100 100 1.29 1.25 l 29 1 25
Maize 100 100 1.23 1.38 1.23 1.38
Cassava with Legumes 87 89 0.24 0.26 0.27 0.39
Cassava 61 69 0.19 0.20 0.48 0.61
Sweet Potatoes with 44 36 0.14 0.10 0 63 0 55
Legumes
Vegetables 28 31 0.03 0.04 0.10 0.18
Total 3.128 3.23 4.00a 4.36
Source: Computer from survey data.

aThese areas exclude 0.73 hectares of uncultivated land that was part
of 3.25 hectares allocated for growing cotton and maize.

63

to change this practice; for it is beneficial to crop growth and erosion
control. Shapiro listed some of the benefits of ridge cultivation as
follows:

1. Imporved weed control from burial of the previous year's weeds.

2. Use of weeds and other plant matter as compost.

3. Thorough, deep working of the soil.

4. Easier weeding.

5. Increased water—absorbing capacity due to increased surface

area.

6. 1ncreased water—holding capacity from thorough working of

the soil.

7. Improved drainage for plants.

8. Slower water runoff and hence greater water absorption

because ridges are parallel to the contour.

9. Decreased erosion because ridges are parallel to the contour

(p. 20).

Farmers in Geita seem to be particularly concerned about building
good ridges. Shapiro noted that some farmers "would hire imigrants from
Burundi to (sesa) a field but not to ridge it, because the imigrants do
not make ridges in their own country and hence have little skill in the
task" (p. 20).

The laborious task of land preparation has its complement in the
high level of field care throughout the farming season. Anthony and
Uchendu observed that "the standard of traditional cultivation in our
area was good and the practice of weeding and splitting ridges provides
good weed control in the early stage of crop growth (when it is most cri—

tical). Crops were remarkably weed free" (p. 59).

64

IXgricultural Calendar

In Table 3.9 are shown the timing of the various crop operations
for various crops grown in Mwanza and Shinyanga regions. However,
different survey sources give different calendars of operations de-
pending upon the particular district surveyed, the type of cropping
pattern prevalent at the time of the survey, and the variety of crops
being grown. For Geita district, Shapiro showed the timing of the
various tasks involved in cotton cultivation. The first labor peak
is in the second half of November when ridging occupies most of the
farmers' time. Most land preparation is completed by mid to late
December, after which weeding becomes most important, remaining
dominant until the end of April. Harvesting begins in mid—May and
reaches a peak in June.

Collinson showed the following peak periods for labor when he
considered all the main crops: the November land—preparation peak,
the January weeding peak, and the June harvest peak. The first two
peaks coincide with the time when a great deal of labor is also de—
voted to food crops, especially maize. Shapiro noted two important
consequences of this labor pattern. "First, the coincidence of labor
peaks for cotton and food crops strains the supply of family labor.
Almost all outside labor hired in the area is for land preparation
and weeding. Second, if cotton were planted one—half month earlier
(a common recommendation) the peaks of cotton and food harvesting
would coicide and thereby force many farmers to hire outside labor in
late April and early May. This would be a third period of labor hiring,
because the land preparation and weeding peaks probably would still

occur" (p. 25).

65

 

 

 

 

 

 

 

 

 

 

Table 3.9
. , a
Calendar of Crop Operations for Mwanza and Shinyanga Regions
___.__—_._ ——.—.—— __ ———-—-—-‘——-—-——-—~ ————— --- -~- 1 —— — " ““ '
l WEED WEED lST 2ND 3RD
OPfllAIION PLOH RIDGE PLANT l 2 “ARV. “ARV. HABV.
FOOD cnor ”— '
Pure Maize
or
Sorghun— Jan.- Feb.- Feb.~ Mar.-
Hlllet Jan Feb. Har. Mar. Apr. June July —————
Early
Leguneo
Caaaava- Nov.- Dec.~
Malta Oct. Oct Oct Nov Dec. Jan. ----------
Early Nov.- Dec.-
Sorghun Oct. Oct Oct Nov Dec. Jan. ----~ -----
Jan.- Feb.-
legunea Nov Dec Dec. Feb. Mar. Apr. May -----
Pure Jan.- Taken an
Caaaava Dec Dec Doc. Feb. ————— laqulrad -----
Pure
Sweet Taken an
Potatoea Nov Nov Nov Dec Mar. Required -----
.l______
Pure 1 1& Jan.- Mar.- Apr.-
lice Nov Jan. Feblii Apr. May May June -----
Pure Dec.- Dec.- Dec.— Jan.- Hay—
Groundnuta Jan. Jan. Jan. Feb. ---- June July -----
CASH CROP
Pab.- June- July-
Cotton Nov Dec. Dec. Jan. Mar. May July Aug.

 

 

 

 

 

 

 

 

 

Source: 1. Calendars for early maize, sorghum, millet, legumes, cas—
sava, sweet potatoes and cotton taken from field notes of
World Bank Project team in MWanza and Shinyanga regions.
2. Calendars for rice and maize mixtures from M.P. Collinson's
"Farm Management Survey Report No. 4," p.9.
3. Calendar for groundnuts from C.K. Klein, D.A.B. Green, R.L.

Donahue, and B.A. Stout, Agricultural Mechanization in
Equatorial Africa, p.2.

3 . . .
Thls IS an acceptable model calendar, With 25 percent of the crops
grown one month earlier and 25 percent one month later.

i, ii, and iii refer to seedbed preparation, cultivation and transplant-
ing respectively.

66

Because of the importance of the sowing date for cotton (often
stressed by extension agencies), it is important to further explore
the calendar of operations. Reference is made to Table 3.1 above, in
which is shown the distribution of rainfall in the district, and to
Table 3.10 below, in which is given a theoretical typical growing
calendar for cotton in the area.

Without reference to any specific area, Purseglove discussed the
ideal rainfall pattern for cotton:

Adequate, but not excessive moisture is required for early
vegetative growth; the first flowering period requires rela—
tive dryness, otherwise excessive boll shedding ensues; an
increase in moisture is required for boll swelling and re—
served growth, followed by dry weather for ripening and
harvest. Up to 15 percent more bolls are shed on days when
rain falls during flowering (p. 348).

Spence.and Spence and Littledyke did not mention the need for rela—
tive dryness during the first flowering period. But they did note that
available moisture should increase from germination to peak flowering
time when the water requirement is at its highest. This happens 100 to
130 days after sowing.

The dangers of too much, rather than too little, rainfall are
often emphasized. A possible reason, as Purseglove noted, is that
"wild (cotton) species are xerophytic and the ability to withstand
drought has persisted in modern cultivated cottons so that they can
recover from a dry spell and resume growth and fruiting" (p. 348).
Since Geita is a relatively wet area, this emphasis on excess rain—
fall is probably warranted. There is some evidence that shows the

effects of too much rainfall. For example, during the 1961—62 farm—

ing season, there was excessive rainfall with the following results as

67

Table 3.10

A Typical Growth Calendar For Cotton

In Tanzania's Western Cotton-Growing Area

 

 

STAGE DATE
Planting December 1
Germination Completed December 7
Flowering Starts February 7
Flowering Peaks March 12
Harvest Starts April 16
Harvest Peaks May 11
Harvest Ends July 20

 

Source: J.R. Spence, "The Importance of Sowing

Date for Cotton,” p.l.

68

reported by the empire Cotton Growing Corporation:

...the high rainfall reduced yields in many areas...
whereas at Ukiriguru the optimum rainfall from December
to March for cotton production is about 455.76 mm,
rainfall during this season was 633 mm. Excessive leach—
ing of the soil nitrates may have occurred, and also many
reports of rotting of seed and young plants have been re—
ceived from the district (p. 9).

During the 1967-68 growing season, the Cotton Research Corporation also

noted the following:

The crop harvested in 1968 was the smallest since 1964...
largely because of unusually heavy rainfall, particularly
at the beginning of the season in November and again in
February, March and April, making the year's total 1480
mm, the greatest record in Ukiriguru....

The heavy early rainfall created difficulties in
land preparation and sowing; later in the season there
were severe leaching in the light—textured soils and water—
logging in the heavy low—lying soils; farmers' weeding pro—
blems were increased and the season was favorable for
American bollworm (p. l).

Shapiro summarized the potential effects of excess rainfall as

follows:

1.

2.

Increased difficulty in land preparation and sowing.
Rotting of seeds and young plants.

Severe leaching on light soils.

Excessive boll shedding in waterlogged heavy soils.
Increased weeding problems.

Improved environment for pests during the growing season.
Damage to exposed cotton fibres.

Shortened dry season and hence better survival rates for

pests (p. 50).

69

As mentioned previously, most soils in Geita are light. Thus,
leaching is more likely to be a problem than is waterlogging. The ill
effects of leaching can be reduced, if not removed, by the application
of nitrogenous fertilizer, just as insecticides can combat the higher
pest populations resulting from heavy rains. The only problem is that
very little fertilizer and insecticide are used in the area as in
other areas of the country.

To minimize the effects of excess rainfall, research at Ukiriguru
recommended a planting period between November 15 and December 15.
This recommended period coincides with the time when a great deal of
labor and other resources are devoted to food crops. This relates to

the objectives of the study presented in Chapter I.

Pests and Diseases

 

The most damaging pest in Tanzania is the American bollworm. In
the western cotton-growing area (WCGA), spring bollworms and blue bugs
are singled out as being among the primary pests. Because Geita
District is relatively wet and cool, it has problems with two pests
that are not so important in most of the rest of the WCGA. These are
Lygus Spp. and Helopeltis Schoutendi. The former pierces the young
cotton buds after which they drop off the plant. Thus there is little
fruit and less output. The latter injects its toxic saliva into the
stem, leaves, and fruit of the plant when it pierces them to feed.

The saliva causes browning, scarring, and withering in the affected
Parts. Both pests can be controlled by the use of DDT or Thoidan.
American bollworms appear in great numbers soon after the plants

flower, about ten weeks after sowing. This pest can ruin most of the

70

early cotton crop. The spiny bollworm continues the job started by
the American bollworm. It begins by boring into and killing the
growing points of the plant. Late in the season it feeds on buds,
flowers, and bolls--the attack reaching a peak in June. Although
the latter can be greatly reduced by uprooting and burning all cotton
plants after harvest (before September 15 in Geita), it is difficult
to deal with the former by the same method. Reed (n.d.) argued that
farmers can prevent an excessive buildup of American bollworm and
might even reduce the population in the cotton fields if maize were
planted strategically:
American bollworm normally builds up on early maize where
it feeds on the cob tips and then migrates to the cotton.
The growing of early maize next to cotton...(almost en-
sures) heavy bollworm attacks. Wherever possible, maize
should be sown after cotton, for the bollworms may then
be attracted away from the cotton to maize, where they do
little damage (p.1).

DDT can kill American bollworms only if it is sprayed 6 to 8
times, as recommended. If it is sprayed only once or twice, as is
'now done by many farmers, the American bollworm population may build
'up to devastating numbers. The effectiveness of DDT or any other
insecticide depends on farmers knowing how to use it; however, the
price of DDT or Thiodan relative to output prices is an extremely

important factor, of great interest to this study as mentioned in

Chapter I.

Recommended Practices
The Ukiriguru research station in conjunction with the extension

service in the area have developed a set of general cotton—growing

71

recommendations for the WCGA. The recommendations have been developed

with the realization that the WCGA has many different microenviron—

ments, each perhaps requiring a unique set of detailed recommendations.

But because of manpower limitations, nothing more than general recommen—

dations could be developed. The following is a simplified list of

general recommendations for the WCGA:

Recommendations related to cultivation practices.

 

l.

2.

Plant all cotton between November 15 and December 31.
Plant six to ten seeds per hole, 25 millimeters deep and
cover firmly.

On ridges that are 1.52 meters apart, plant two rows 0.46
meters apart and space the seed holes 0.46 meters apart in
in each row.

Thin to two plants per stand after three weeks.

Weed for the first time while thinning.

Weed three to five times during the season to keep the
field weed free at all times.

Harvest as bolls open and cotton dries; do not allow open
bolls to remain on the plant very long.

Uproot and burn all cotton plants by September 15.

Rotate three years of cotton with three to five years of

cassava.

Recommendations related to new inputs.

 

1.

At sesa (land—preparation) time, put double superphosphate in
the former furrow of the split old ridge where the new ridge

is to be built. Use 125 killograms per hectare.

72

2. At six weeks apply ammonia sulphate in a shallow ditch
between the two rows of plants on top of each ridge.
Use 125 kg per hectare.

3. Start to spray Thiodan or DDT when the cotton plants begin
to flower, usually at ten weeks. Use half a kilogram of
active ingredient (75 percent DDT powder) mixed with 60
litres of water for each acre.

4. Spray Thiodan or DDT six times at two-week intervals.

5. If and when stainers or blue bugs build up, spray with

carbaryl.

These recommendations are discussed below under the following
headings: planting; thinning and weeding; harvesting; uprooting and
burning; rotation; fertilizers; and insecticides.

Planting. The method of land preparation in Geita is quite
satisfactory and acceptable to the Ministry of Agriculture and the
Tanzania Cotton Authority (TCA). However, the timing of land prepara—
tion is crucial if cotton is to be planted between November 15 and
iDecember 31. Spence (n.d.) noted that "the most important factor in
land preparation is correct timing: (p. l).

Planting during the recommended period increases the probability
that the moisture requirements of cotton will coincide with the rainfall
pattern. Since rains usually start in late September, by mid November
the land will probably contain enough moisture to insure germination.
beimum water requirement for cotton occurs during peak flowering,

which takes place 100 to 130 days after sowing. Since April is the

 

73

month of greatest rainfall, cotton planted in December probably will
be flowering under ideal rain conditions. Since rainfall diminishes
rapidly in May, for cotton planted during the recommended time,
there is not too great a risk of rain falling on open bolls.

In Geita, unlike the rest of the WCGA, the rainy season extends
late; thus cotton should be planted during the later part of the
recommended period. If cotton is planted too early, the bolls may
open during the rains.

The sowing recommendation of 6 to 10 seeds per hole, when only
two plants ultimately are to be left, is based on three considerations.
First, 100 percent germination is not expected. Second, seedling
mortability is high during the first three weeks of growth. Third, the
soils of the WCGA form a crust after rains, and the force of several
seedlings is required to penetrate it.

The reommendation that stands of cotton be 0.46 meters apart is
less widely adopted. Most farmers believe that this distance is too
close. There is no agreement bythe farmers as to why this is so.
Some say that such close planting leads to much vegetative growth but
few bolls. Others say that it encourages greater insect damage. In
the high temperatures of unshaded areas, insects may not breed well.
Under denser plantings, however, insects may breed faster because of
the shade. Spencer (n.d.) noted that farmers seem to choose wider
spacing "due to the superior appearance of the plants at wider
spacings...(they are) impressed by the high yield per plant given at

wider spacings" (p. 6). This can be interpreted as risk avoidance on

 

74

the part of farmers, since healthy plants may be better able to
withstand insect attacks and possible short periods of drought.

Thinning and weeding. Most farmers follow these recommendations.
They leave two cotton plants per stand and try to keep their fields
weed free. If thinning and first weeding are done simultaneously, the
plants will get off to a good start. Although thinning is easily done
when the ground is wet, it would not be delayed because of a dry spell
since in dry soil the competition of excess plants for water becomes
more damaging. Although it is easier and more effective to weed
during a dry spell, the timing for weeding like that of thinning, is
very important. Weeding should not be postponed until the weather is
just right.

Harvesting. The timing of cotton harvesting is very important
because fibres in open bolls may be rained on and, hence, turn gray
if left on the plant too long. On the other hand, the open bolls should
not be picked too soon; a few days are needed for the fresh fibres to
dry. Because bolls do not open or dry simultaneously, more than one
picking is necessary if all cotton is to be picked at the proper time.

After harvesting, farmers must sort cotton into clean (grade AR)
or dirty (grade BR) piles. Grade AR may constitute about 90 percent of
the total harvest. This sorting is important because the two grades
have different prices.

Uprooting and burning. It is recommended that all cotton plants

 

be uprooted and burned by September 15. Because of the failure of many
farmers to follow this recommendation, it has been made a bylaw. Farmers

are subject to fines if they have cotton on their fields after the

 

recommended date. The reason for passing a bylaw on this recommendation
is that cotton plants left on one farmers' field may serve as hosts for
insects that attack a neighbor's field. If all insects are to be eli—
minated, uprooting and burning must be done by every farmer, and done
most carefully.

Crop rotation. Spence (n.d.) noted that an ideal rotation
"system has not yet been achieved with cotton in the WCGA and no recom—
mendations can be made on cottonrotationsyet" (p. 10). Anthony and
Uchendu reported differently for Geita." In Geita it is recommended
that unfertilized cotton should be moved every fourth year and alter—
nated with cassava" (p. 35). They based this recommendation on the
following considerations:

Cassava, especially if left unweeded, has been shown to
have a beneficial effect on the following cotton crop.

In an experiment at Ukiriguru, cotton sown after three
years of cassava, both crops without applied fertilizer,
gave better yield over a three year period than continuous
cotton receiving 45.4 kg of sulphate of ammonia every year
and 45.4 kg of double superphosphate every third year

(p. 35).

Most farmers in Geita are aware that the cassava—cotton rotation
is a good practice. However, intercropping and crop rotation prac—
tices have been less and less applied during the last ten years. In
an attempt to provide close supervision of cotton and maize production,
the government allocates and regulates the use of land. Block farms
have been established in which only cotton and maize can be grown.

Other crops such as cassava, legumes, and vegetables can be grown on

land allocated around homesteads.

 

76

Fertilizers. The recommendation on fertilizer application is
quite general and does not take into account local conditions of
different areas within the WCGA that may lead to extreme variability
in responses to fertilizer. Shapiro reported that Geita district
differs in two important ways from most of the rest of the WCGA.
The district receives more rainfall than the rest of the WCGA.
Second, its soil was not tapped to any great extent, until after
World War II, whereas most of the rest of the WCGA was under cotton
cultivation since the early part of the century.

Fertilizer responsiveness in Geita, therefore, has been quite $4;
different from that in other areas. In general, the response to
superphosphate in Geita has been low relative to the response else—
where, whereas the response to ammonium sulphate may be relatively
high. LeMare (1967) in summarizing the Ukiriguru field trials, had
this to offer:

Yield increases due to superphosphate have been confined
to the area of Sukumaland which has been under cultiva—
tion for a long time. This comprises the districts of
Mwanza, Kwimba, Maswa and Shinyanga. Elsewhere (for
example, in Geita) responses to phosphate have been
small....
The pattern of nitrogen (for example ammonium sulphate)
response is more difficult to summarize because, unlike
phosphate, its effect is more variable between seasons,
and depends more upon the amount of water available....
Where ample water is available, and where other soil
factors are not limiting (conditions more common in
Geita than elsewhere in the WCGA), nitrogen can have a
very large effect on yield and can be profitable up to
very large dressings (p. 3).

It should be remembered also that the benefits of fertilizer

will vary, in part, according to the extent of insect damage. As

LeMare (1967, p. 3) put it: "the effect of fertilizer may be almost

77

completely lost if the crop is subject to severe insect attack."
Insecticides. Shapiro reported that many farmers and agricultural
officials believe that insecticides alone produce greater benefits than

do fertilizers alone. This observation is substantiated in Table 3.11.

Table 3.11

Responses to Nitrogenous-Fertilizer Under Varying

Insecticide Applicationsg

 

 

Insecticide Spraying Nitrochalk Applications (kg/hectare)
Regime 0 200
4'
None 419 424 “
Standard 1076 1173
Best 1515 1880

 

Source: P.H. LeMare, "The Importance of Insect Control for Fertilizer
Responses in Cotton," p.2.

a . .
The experimental plots were in the Mwanhala area.

The insecticide recommendation is probably more important in Geita
for two reasons. Recent settlement implies more remaining natural soil
fertility, whereas greater rainfall means that many pests may be more
damaging in unsprayed fields. However, the conditions attached to this
recommendation have somewhat impeded its adoption. The recommendation
calls for heavy labor inputs. To spray one hectare, the farmer would
need sixty litres of water to be mixed with DDT powder. Not only that,
but the process of spraying has to be done every two weeks for a total

of six times during the season. Not many farmers can readily get

water near their farms. And even if they catch and store rain water,
they would be unwilling to use this valuable clean water for spraying.
Fetching water from streams is too laborious for them.

In the last five years or so, farmers have been introduced to
two other insecticides, namely Thiodanand Cidial. Pumps are used for
spraying. But the pumps require batteries, which are sometimes un—
available in the market or are very expensive when they are available.
Thiodan and Cidial are extremely poisonous to human beings. Incidents
of people who have lost their lives after ingesting either of these
insecticides are known all over the district. This has impeded the

adoption of the recommendation.

Recommended Practices for Maize

 

In contrast to cotton, little research has been done on maize
production. The research station at Ukiriguru and many others in the
country have been researching export crops such as cotton, coffee,
tobacco, and tea. Only since the mid 1970's has there been research
interest in food crops, particularly in maize. The reasons for this
interest were presented in Chapter II. Thus, there is no detailed
and comprehensive list of recommended practices for maize. However,
research at Ukiriguru has provided some general recommendations that
are subject to further research.

Recommendations related to cultivation practices.

1. Plant maize between November and December.

2. Plant two to three seeds per hole and cover firmly.

3. On ridges that are 0.9 meters apart, plant one row and

space the seed holes 0.3 meters apart.

 

4. Thin to one or two plants per stand after three weeks.

5. Weed for first time while thinning.

6. Weed two to three times during the season to keep the
field weed free at most times.

7. Harvest when the grains are dry enough to be milled into
flour.

8. Uproot and burn all plants before next season.

Recommendations related to new inputs.

 

1. During planting, add about 50 kg of TSP per hectare.
2. Apply sulphate of ammonia three times during the season, the
first application should take place when the plant is about
21 centimeters high, the second when the plant has reached the
level of the knee, and the last during tassling.
3. Start to spray with DDT when the plant is about six weeks
along; spray three to four times at two-week intervals.
In addition, it is recommended that farmers use better seeds (mostly
hybrid seeds) instead of composite seeds, which are commonly used in
the area.
The absence of any field evidence has led to a conclusion
(based on observations) that unlike cotton, maize is not as critically
affected by insects or weather. Farmers, however, consider weather,
especially rainfall, as being crucial to maize production. In particular,
they are afraid of ashorterrainy season that occurs in the area every
few years. To avoid any risk, farmers plant maize as soon as the rains

start. That is the time when cotton should also be planted. And, as

80

discussed in Chapter II, this creates labor—allocation problems for
farmers, as do the recommendations for weeding and harvesting.

The recommendations for fertilizer and insecticide application
also create a problem for making capital allocation between cotton
and maize. This and the labor—allocation problems mentioned above
are aggravated by the fact that cotton and maize can be grown on the
same soils. Because cotton requires more resources and higher manage—
ment skills than maize, more variability in yields have been experienced

by farmers in the study area.3

 

Ninety—seven percent of the farmers in the sample experienced more
variability in cotton yields than in maize yields during the last five
years. Through informal discussions with extension agents in the area
the researcher found the farmers' response to the question was true.

 

81

CHAPTER IV

THE ANALYTICAL FRAMEWORK OF THE STUDY AND RESEARCH METHODOLOGY

The Analytical Framework

The choice of an analytical technique in constructing a model de—
pends upon the availability of data, the purpose for which the model is
intended, and the nature of the structural coefficients being sought to
elucidate a particular problem. In this study, linear programming was
used in the analysis of the data.

The use of LP as the computational tool in the farm—planning
exercise was based on the premise that small—holder farmers tend to
behave in ways that optimize their objectives given the constraints
within which they operate. Production and consumption considerations
are both important for small—holder farmers. Consequently, an attempt
was made to integrate the two deciSions into a single methodological
framework. Endogenous determination of consumption activities in LP
permitted the premise that staple food is grown for home consumption
and/or for sale in the market.

Risk factors are an important consideration in small—holder decision
making. Therefore, some method of incorporating risk factors into the LP
framework was needed. A number of approaches have been developed to take
into account risk factors in LP models (Kennedy and Francisco; Andrews and
John; and McCarl), but there is yet no clear guidance for choosing the
Inost appropriate method. Not only that, but time—series data on yields,
prices and production costs that are needed to measure income variability
may not be available. Data are required for the application of quadratic—

programming techniques to small—holder farm behavior under uncertainty.

82

It is more likely that small—holder farmers are concerned about achiev—
ing a minimum level of production with certainty rather than minimizing
income variance. In this study, risk factors were incorporated in the
analysis only as consumption constraints for the major food crop.

The Use of the Linear-Programming Technique for Analyzing
Small Farm Agriculture

 

The linear-programming technique has been applied increasingly in
recent years to solving small—holder problems in Africa and other third—
world countries. In his study of resource allocation among subsistence
farmers in Ghana to evaluate various policy recommendations designed to
increase agricultural production, Atta Kouadu employed the LP technique.
Thamrin Nurdin examined factors affecting farm decision making of small
farmers in west Sumatra by lexico—graphic programming in order to cope
with multi—objectives behavior. Ogunfowora, using Nigerian data, under-
took an analysis of the constraint posed by periodic specific capital
shortages and by quality of management as well as by labor. Subjective
limitations reflecting management differences and risk—aversion behavior
distinguish two farm models that represent different levels of commer—
cialization. Shadow prices for labor and capital suggest the types of
government policies that most efficiently increase income potential in
these respective farm types.

Ogunfowora also used a poly-period dynamic—programming model to
plan operations for a farm settlement scheme that would assure both an
adequate income and short—period repayment capability. Norman used the
LP techniques to evaluate the profitability of agricultural production

and labor utilization among the Hausa of northern Nigeria.

 

83

Specifically, he used the technique' to assess the profitability of
several adjustments in farm models based on data from the area. These
adjustments included reallocation of existing resources, increasing the
input of labor on a year—round basis, introducing currently available
new technologies for groundnuts, sorghum, and cotton, and increasing
prices of crops purchased by the marketing board. The adjustments
tended to increase farm income.

Heyer (1971) discussed several broader macro uses to which linear
programming micro analysis can be put, including the shadow pricing of
agricultural resources, the assessment of employment and mechanization
programs, and the evaluation of new variety profitability and research
priorities. Using data collected in Kenya, Heyer described as valid
the changing pattern of constraints that limit output under such alter—
native mixtures as the land—labor ratio. Nonfanm allocations of labor
time were not incorporated in the model. The analysis was extended,
however, to include uncertainty restrictions.

Norman and Ogunfowora used 1P techniques to assess profitability of
adjustment as well as to estimate specifically farm—firm fertilizer
demand and its elasticities with respect to own price, product price,
and capital, making it useful for policy prescription. The study also
showed that the linear programming technique can be used to estimate

resource demand in an environment lacking time—series data.

The Linear—Programming Model

 

Linear programming is a technique for maximizing (or minimizing) a

linear—objactive—function subject to some linear constraints. The model

84

has three components: (1) the objective function; (2) resource con—
straints; and (3) activities. According to Heady and Candler, the
mathematical formulation in matrix form is given as follows:

Maximize Z = c'x
subject to:

AX5.B

X 20
where:

Z = the objective function to be maximized (or minimized)

C = n x 1 vector of prices

X = n x 1 vector of activity levels

A = m x n matrix of input—output coefficients

B = m x 1 vector of resource restrictions

In order to obtain a determinate solution, several assumptions are
made: (1) additivity and linearity of activities; (2) divisibility of
activities and resources; (3) finiteness of alternative activities and
resource restrictions; and (4) single—value expectations; i.e., resource
supplies, input coefficients, and prices are known with certainty.

The main advantage of this LP model is that it "...allows for
several farm commodities as farm activities, seasonal labor and land
constraints, more than one production technique, land—labor—capital
substitution and a choice among several farm activities which are sub—
ject to different economic resource and behavioral constraints"
(Mudahar, p.2). Thus, unlike other commonly used calculation techniques
Of farm planning, linear programming can be used to provide a more

adequate analytical description of whole—farm situations.

 

85

An equally important advantage is that LP allows the determination
of certain important economic measures of the optimal plan. One can say,
for example, how stable the optimal plan is, measured in terms of the
change in the net revenue of each enterprise needed to bring about a
change in the levels of the activities in the optimal solution.

Similarly, the productivity of the farm resources can be assessed and
the importance of the various planning constraints evaluated.

The technique, however, has a number of limitations. First, the
standard LP model does not include any allowance for risk, which is
central to decision making among small—holder farmers. (New methods
such as MOTAD allow for handling risk.) The importance small farmers
attach to securing an adequate food supply as a primary objective is
well documented (Collinson, 1964; Heyer, 1969). Because of unreliable
marketing organizations, the wide gap between buying and selling
for identical or readily substitutable foods and the year—to—year
variation in prices and crop yields that accentuate the risk aversion,
even farms with well-developed markets continue to produce all or most
of their subsistence requirements. Thus, the objective function for
small farmers may indeed be security maximization rather than cash-
income maximization (Norman).

Another limitation of the LP technique results from the assumption
that the farmer will adopt any enterprise combination as long as it pro—
mises the highest income. This may not be the case. Often certain crops
and livestock weigh more heavily than others in the preference of farmers.
Heyer (1971) contended that the objective function is difficult to deter—

mine under small—holder farming because it is ambiguous, and risk

86

considerations tend to dominate production decisions. Further,

cultural and institutional factors constrain the production environment.
However, the standard LP model outlined above can be modified to include
allowance for risk and to estimate product supply as well as resource
demand. The modified standard LP model is called the parametric—
programming model. It enables the researcher to study the effects of

a wide range of costs or prices on the optimum solution to the standard
simplex method. Such a linear—programming problem with parametric ob—

jective functions has been conceptualized as follows (Ogunfowora;

Mwangi):
n
Maximize Z = 2 C. X,
a '-l J J
J—
subject to:
m
2 ainmib
. k
1=l
and
X.Z.O
J
where:
Z = z(Xl, X2, X3....Xj,....Xn)

C'fzc égc"
J’J’J’

CH _ cl
j j =kor C'.‘-C'=k
X J j

Z = the ath objective function to be maximized for a given price
level within the acceptable price range and for a given
technology.

bi = the level of the ith resource available.

87

C3 and 03 = the lower and upper limits of the jth activity.

ll

k the number of optimum solutions within the price.

X = constant increments in the price of the jth activity.

The farms are assumed to have achieved an optimum organization before
price and technology change occurs.

Even the modified LP model has not gone unchallenged. In a
comment on planned versus actual farmer performance under uncertainty
in under—developed agriculture, Palmer—Jones criticized the method em—
ployed by Heyer (1972) to analyze small-farmer behavior. Palmer—Jones
questioned the legitimacy of using average input—output coefficients in
the LP models used for the analyses because, first, farmers may in fact
alter their strategies or technical inputs under different environmental
conditions, and second, there is no theoretical reason to believe that
average inputs will give rise to average outputs. But Palmer—Jones did
not stop there. He went on to question the use and validity of the LP
techniques for studying small—farm situations in general. The main
argument Palmer—Jones gave for this criticism dealt with data problems.

Although Low (1978) agreed to the criticism directed at the Heyer
analysis of peasant—farmer behavior under uncertainty, he disagreed
with the criticism that the LP technique cannot contribute anything
useful in the study of peasant—farming situations. He supported his
disagreement by quoting two examples from his own work. First, an
unexpected relationship observed in southeast Ghana betweentfluaintroduc—
tion of tractors that enabled the expansion of cassava to the savanna
lands and the expansion of forest—cultivated maize for which tractors

were not used, were explained in terms of an LP specification (Low, 1974).

 

88

The explanation involving subsistence requirements, a third activity,
resource allocations between the three activities, and a maximum un-
certainty specification could not have easily been worked out in the
absence of the LP technique.

A related paper (Low, 1975) examined the implications of the LP
model for extension strategy. The conclusion reached in respect to
an improved maize recommendation, for example, seemed to explain,
contrary to what was expected on the basis of a partial—budget model,
why the innovation had not been readily adopted by certain farmers.
Because the model accounted for such factors as uncertainty, product—
product relationship, the allocation of clearing labor to a succession
of crops, and the relationship between production and consumption, in—
cluding the effects of storage losses and seasonal price fluctuations,
it probably constituted a more realistic representation of the
peasant—farmer decision—making environment than the partial budget
model.

Linear programming was used in the present study for the following
reasons:

1. In LP, a large number of interrelated variables can be handled,
and thus family-farm systems can be studied that are characterized by a
high degree of interdependence between production and consumption,
consumption and investment, investment and resource availability and
social and cultural constraints as mentioned at the beginning of this
chapter and in Chapter V below.

2. The maximum possible profit for a farm—planning problem is

guaranteed. This is difficult, if not impossible, to obtain with

 

89

ordinary budgeting for any complex problem.

3. With LP it is easy to vary available prices and resources as
well as input coefficients in order to simulate various management
levels. LP makes it possible to look almost instantaneously at a
range of possibilities, ones so laborious to determine with ordinary
budgeting that such possibilities could not be examined in practice.

In short, LP seemed to be the best technique for attaining the

objectives of this study as set forth in Chapter I.

Research Methodology
Choosing the Research Site

The choice of the research site was determined by two primary
reasons. First, it had to be an important area for cotton and maize
crops. The purpose for doing a micro—survey on cotton and maize was
broadly justified in Chapter I. That justification can be summarized
as follows:

1. Cotton accounts for more than 17 percent of the country's
total exports, which in turn make up for about 30 percent of the
country's GDP. Thus, it is an important source of foreign exchange,
which is badly needed to pay for the country's import requirements.
And, as was shown in Table 1.4, Geita is the most important cotton~
producing area.

2. Maize, on the other hand, is the main food crop in the
country, especially in the cotton—growing areas and more so in Geita.
For the past few years, Tanzania has not been able to meet her domestic
needs for this crop and has had to spend much of her foreign reserves
t0 import maize. Increased production of maize would thus reduce food

imports.

 

9O

3. The improvements in the production of both crops in Geita
are financed by the World Bank and involve many small farmers.

4. Both crops are grown primarily by small farmers under the
following conditions.

a. Neither crop is irrigated; hence they are subject to
particular erratic weather conditions;

b. Both crops use farming methods that are extremely time
consuming; and

c. Both crops are grown during the same season and are
thus competitive for labor and other inputs.

The second reason influencing the choice of the research site
was that the area had to have been the subject of enough prior re—
search so that adequate background information could be acquired
before the researcher went to the field. Malcolm, Anthony and
Uchendu, and Collinson (1964) among others provided basic agricul—

tural information about Geita.

Choosing the Sample

The population of the study consisted of all family farms in
the main cotton— and maize—growing villages in Geita district. For
the purpose of this research, the population of the study was not

confined to those villages in the World Bank's cotton project.

Procedures for Selecting Respondents

 

In order to increase the representativeness and precision of the
sample, the population was divided into two strata. One stratum con-

sisted of participants and the other of nonparticipants in the World

91

Bank's cotton project. This stratification was justified by the
objectives of this study. The two strata were believed to be different
in terms of the source, type, and extent of resource constraints im-
posed on their decision—making process. This belief was justified by
the assumption that the participants used the full package (fertilizer,
better seeds, insecticides, and so on) whereas the nonparticipants did
not; this assumption was not empirically tested).

The list of villages under each stratum was provided by the
evaluation unit of the Geita cotton project. During the same farming
season, the evaluation unit conducted farm—managment research.

The sample in the study consisted of 286 farmers involved in the
cotton project and 215 farmers not in the project. The unit had well—
trained enumerators with two to three years of experience in data
collection. It also had an easy access to the project's transport
system. The unit was short of supervisors, so the researcher was con—
sidered very useful. In exchange for the researcher's supervision,
the unit allowed the researcher to use its enumerators. In order to
minimize the work of the enumerators, the researcher's sample was
selected from that of the unit as long as the objectives of the re-
search were not compromised. The happy coincidence of the unit's and
the researcher's interests facilitated a smooth cooperation between
the two.

As a result of its previous research experience in the project,
the unit was able to provide adequate information about which villages
used plows; it also had a list of farmers who had already applied to
the project for tractor-hire services. This information was important

and necessary for the selection of the sample.

92

Sample Size

Ideally the sample size should be determined by the degree of
precision required (Yang). Statistical theory is most useful in help—
ing determine the size of the sample only when one variable is handled
and its variance is known in advance. However, more than one variable
was dealt with in this study; the variables used here included such
factors as family labor, hired labor, animal power owned and hired,
tractor services, prices received for crops sold, and wage rates. It
was thus impossible to apply a formal statistical procedure that would
achieve a statistical representativeness of the sample.

Another problem that hindered the use of a statistical procedure
for selecting the sample size was the cost that would be involved
especially in hiring enumerators and processing the data. In select—
ing the sample for this study, therefore, much consideration was given
to cost, time, availability and experience of enumerators, and last,
but not least, some degree of precision. As a trade—off between cost
minimization and precision, a sample of 80 households (40 from each
stratum) was considered to be sufficient.

The primary sampling unit was the family—farm household. The
family household represented both the production unit and the consump—
tion unit from each village. Enumerators, with the help of village
leaders, prepared four lists: the first list consisted of those
households that had decided to use labor (of any kind) for every
farming operation; the second consisted of those households that owned
plows and intended to use them in some farming operations; the third

consisted of those households that intended to hire plows for some

93

operations; and the last list contained those households that intended
to hire tractor services. The last list had already been prepared by
the evaluation unit of the Geita cotton project.

The main purpose of preparing these lists was to insure that the
sample included these different farming systems. Not many farm house—
holds owned plows; therefore, all households that owned plows were in—
cluded in the sample. From the list of households that had decided to
hire tractor services, a random sample of 10 households was drawn. This
was done for both participating and nonparticipating farm households.
Yang suggested at least 20 farm households as being necessary for re—
liable estimates for each stratum. Friedrich concluded that roughly 20
to 25 observations should be included in each stratum in order to make
a reliable comparison. However, the reliability of estimates very much
depended on the actual variability of the population. From each stratum,
4 substrata were constructed. The first was a substratum of farm house-
holds using only labor for farming operations; the second was for those
family farms using owned oxen plows; the third was for those using
hired ox plows; and the fourth was a substratum of family farms that
hired tractor services. Each substratum was expected to provide 10
observations. Lack of observations (especially for ox plows), however,
resulted in different numbers of observations. From the list of farm
households using only labor for farming operations, random samples of
21 and 22 farm households were selected for participants and nonparti—
cipants respectively. The main difference between the two samples re—

sulted from the difference in the application of ox plows.

94

Among the participants, only 5 owned plows and only 4 were ready to
hire ox plows. Among the nonparticipants, only 2 owned plows; only

6 were ready to hire them.

Replacement

Some investigators mentioned the problem of uncooperative farmers.
This problem was expected in this study because of the political sensi—
tivity of the cotton—maize projects. To guard against such an eventu-
ality, the possibility of the nonresponse rate was assumed to be 10
percent for each stratum, thus increasing the target sample for each
sample by 10 farm households using only labor. To guard against non—
cooperation from plow owners and from those who intended to hire either
tractor or plow services, much depended on the cooperation of village
leaders and enumerators. Although enumerators were expected to be
cooperative and friendly to all selected farm households, they were
instructed to work more closely with those farm households owning ox
plows and hiring plows or tractor services. This measure was taken to
avoid losing any observation from these groups whose number was limited.
The outcome was extremely favorable.

From those farm households selected for their use of labor in
every farm operation, only four dropped out (one because of death in
the family, two because of family quarrels, and one because of misunder—
standing with village leadership). None dropped out from those who
owned plows or from those who intended to hire tractor services, one
household changed its mind about hiring the services. Since there

was no other farm household in this category in the selected villages

 

95

to replace this farm, the household was left in the sample as one
that used only labor for all farm operations.

The good cooperation that the study received from the selected
farm households and village leaders can be explained. First, the
enumerators were well known by villages; they had been in the area
for the previous three years. Second, most of the enumerators knew
the local language; thus they were not suspected of anything covert.

They were trusted.

Data Collection

The researcher designed and developed questionnaires for the pur—
pose of collecting the needed data. This included data on the produc—
tion, prices of inputs and outputs, and the resources that were available
or could be available on the farm. Data for only one cropping season,
1979—80,were collected.

The data were collected by six enumerators who were well trained
by the evaluation unit of the Geita cotton project. Enumerators
visited each farm household twice a week for a full year. For each
household member, the enumerators recorded all the activities done and
for how long prior to the interview. More emphasis was given to the
collection of data on labor use, and family—labor allocation during
the peak demand for labor (especially during weeding). This format of
twice—a-week visits allowed for relatively short recall on the part of
respondents. The enumerators were supervised by the researcher.

Input prices as well as output prices are determined by the govern—
ment prior to the farming season. So these were known by each farm

household, and all farmers had given the same response to questions

96

related to prices. However, the researcher observed that most farmers
did not sell maize through the established government agency, the
National Milling Corporation. Maize is sold on the black market

where prices may rise as high as 250 percent above the government
price. In fact, most farmers would wait a little longer after harvest
to sell their maize output. When asked what would be a fair price for
a kilogram of maize, most farmers responded Tshs 2.50, which in fact

was a black market price two to three months following harvest.

Construction of the Representative Farms

One method of analyzing the data collected would have been to
program every farm household using a case—study approach. However,
such an approach would not only have been too costly and, therefore,
prohibitive, but also might not have led to meaningful results.
Consequently, in carrying out the linear-programming analysis, it was
essential to set up a representative farm.

In areas in which there is a reasonable homogeneity with respect
to major resources (particularly natural resources such as soil type,
topography, and climate) and cultural practices, LP can be used to
obtain a representative farm in order to guide planning for individual
farms. The manner by which the representative farm is constructed
limits its usefulness. Collinson (1974) (p. 125) discussed three alter—
native techniques for deriving representative farms. They are

l. The identification of a particular farm as the typical farm.

2. The use of an average farm (derived from average resource,
input-output, and net price coefficients of a sample farm) as a repre—

sentative farm.

 

97

3. A hypothetical or synthesis—of—composit farm from different
components of the population.

It is not easy to find a typical farm. It not only requires the
consideration of a wide range of criteria but also the selection and
the construction of the criteria are difficult tasks; data for this
purpose may not be available nor easy to collect.

The average—farm approach brings with it the aggregation bias.
Buckwell and Hazel, Carter, and Miller discussed the agregation bias
inherent in the average-farm approach. Aggregation bias exists when
the sum of the solution from the individual farms in the set does not
equal the estimate obtained by the optimum solution to the entire set
directly.

Although the hypothetical farm approach reduces the aggregation
bias, it has a practical weakness in that it is difficult to identify
several institutional variables and human factors and their distribu-
tion within the population. These nontypical variables involve such
factors as institutional constraints, motivations, preferences, and
managerial ability that have an important impact on farm organization,
production efficiency, and earning (Plaxico and Tweeten).

The choice of the method for construction of the representative
farm depends on the purpose for which the result of the study is to be
used. In this study, the objective was to identify resource constraints
and farm adjustments (both for participants and nonparticipants in the
Geita cotton project) and to estimate the degree of farmers' responses
to input and output price changes. The use of the average farm as a

representative farm was justified and was preferred for this study.

 

98

The farms in the sample were classified into two groups: the
participants and the nonparticipants in the Geita cotton project. The
farms in each group were assumed to be sufficiently homogeneous with
respect to the key variables that affect farm adjustment. For the
analysis, only one average farm was used for each group; this offered

an opportunity for more detailed analysis using parametric techniques.

Characteristics of Farms in the Sample

 

Land Use

The average size of holdings for cotton and maize in the study area
was 3.25 hectares for the participants and 2.63 hectares for the nonpar—
ticipants. The cultivated areas were 2.52 and 2.63 hectares respectively.
Land is allocated by local government officials. About 85 percent of the
land allocated is under block farming. Since renting of land is for—
hidden by law, farmers can only expand their holdings through official
land allocation. Land allocated to cotton and maize are located 1 to 2
kilometers from homesteads. This distance has discouraged farmers from
using animal manure which has to be carried manually from homesteads
where it is kept. Farmers also grow other minor food crops such as
cassava, sweet potatoes, and legumes; these are grown around or near
homesteads. The average size of holdings for such crops was 0.6
hectares in each group.
Farm—Labor Force

The major source of farm labor in these farms was the family; this
was expected because family farms predominate in the study area consisted
of 7.7 persons in the participating families and 6.5 persons in the non—
participating families. The composition of the average farm family in

each group is shown in Table 4.1.

99

Table 4.1

Composition of the Average Family in the Study Area

 

 

 

 

 

 

FAMILY CONVERSION FACTOR NO. IN FAMILY ACTIVE & DEPENDENT
MEMBERS T0 MAN . MEMBERS AS Z OF TOT.
EQUIVALENT Pa NPb P NP
Head of the “N ‘W
Family (Male) 1.00 1 1
Number of
Wives 0.75 1.3 1.2
Children:
Male
(15 yrs. or
more) 1.00 1.6 1.2 5>88.29 >89.23
Female
(15 yrs. or
more) 0.75 2.1 1.9
Male &
Female
(7—14 yrs.) 0.50 0.8 0.5 _‘ _}
Children
(younger than
7 yrs.) 0.00 0.5 0.4
11.71 10.77
Dependent
Adults
(over 60 yrs.) 0.00 0.4 0.3
Total 7.7 6.5 100.00 100.0

 

 

 

 

 

 

Source: Compiled from the survey data.

aParticipating families

b
Nonparticipating families

100

Family labor is usually supplemented by hired labor or labor
acquired through working parties. This is especially true during
peak labor demands. Hired labor is paid in cash, whereas labor from
working parties is paid in kind (food and beer). The allocation of
monthly labor inputs on the representative farm is shown in Table 4.2.

The total labor input on the representative farms was 3948.70
man hours for the participants and 4374.86 man hours for the nonparti—
cipants. 72.02 percent of the participants' total labor input came
from the family, 7.55 percent from hired labor, and 20.43 percent from
working parties. The sources of labor input for the nonparticipants
was remarkably different. 95.82 percent came from the family, 1.55
percent from hired labor and 2.63 percent from working parties.

As is shown in Table 4.2, participating farmers used more hired
labor and working parties than did nonparticipating farmers. The main
reason for this was that the former group followed crop-husbandry
recommendations including three to four weeding operations.

Farm Capital

Farm capital refers to manmade goods or assets that are produced
for the purpose of being used in the process of agricultural production.
It includes items such as machines, tools, buildings, livestock, seeds,
fertilizers, and insecticides. Such assets are usually classified,
according to the length of their productive lives, into fixed— or long-
term capital and operating— or short—term capital. The former consists
of items such as machines, tools, land improvements, and buildings with

productive lives that extend beyond one production cyclel, whereas the

1Production cycle in this study refers to (l) clearing and cultivation of
land; (2) sowing and fertilizing; (3) weeding; and (4) harvesting.

101

 

.mwflaflamm wcfiumawofiuumacoz

n

.moflawfimm mafiummfiofiuummm

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

00.00H 00.00H W mo.~ mq.0~ mn.H mm.n um.mo N0.N~ W N
i
ew.q~m.q 0n.w¢0.m 0n.qHH 00.500 00.00 oo.mo~ 0H.NOH.q 0n.mqm.~ W HmUOH

00.0 I I I N~.mn~ um.oo~ 00.0w 00.0c I I NN.MHN em.mau meow
I I I mm.H qm.aam Nu.ooq I mm.0m I mn.nq qm.aam qo.~mm an:

-.H ~0.a I I H0.Hwa mm.qmm ~H.<H 0¢.mm I I ow.on< ma.HmN .uac
I I I I na.mqa on.mo~ I I I I na.m<~ on.oo~ .uw:

q0.H ~a.a mo.q wo.< n~.0nn m~.00n mH.mN w 0H.0NH m0.am mn.oo 00.m~n om.mmm .nom

on.o 05.0 nH.q ma.q 00.0mm mn.~om mm.~H w on.mNH Nn.0~ 00.Nm 00.00n ow.Hmm .cmn

mm.0 n0.H I mm.m ~0.q0n mo.an an.qa m mw.q~H I 00.mm am.owc mn.nnm .oun

no.0 «0.0 m~.m mN.m mm.mam m<.mom m~.0H m Nm.00H nq.0a 0m.c~ w~.nw¢ Ho.wm~ .>oz

.

mm.0 M 00.0 I 0H.m No.q~m am.mmq qm.oa mm.oma I 00.m~ m0.m0n m0.H~m .uuo
I N I I I 0m.mo~ N~.~0H I I I mm.mo~ NN.~0H .unom
I I I I 00.no 00.<n I I I 00.n0 00.qn .w:<
m2 m A2 a m2 m L2 m m2 m 9&2 mm

1.562
A.u=\mnmav A.u=\mzmHv Amuaoccwxv _ Amunozcmzv Ampsoncmzv Awuaoccmxv
asz zH .»<m mh<m m0<3 m0m<A A<HOB _ >Hm<m quxmoz x0m<A 0mmH= mom<q qux<m

 

 

 

 

 

mapmm m>au¢ucommumom

ecu so wcHM :H mucoezmm new .mwumm owmz .nuuoz mo musacH Momma Hmuow

N.q waan

 

102

latter is made up of assets such as fertilizer and seed that are used
up in a single production cycle (Upton, p. 149; Herbst, p. 8; Barnard
and Nix, p. 50).

The level of fixed capital in the study area was quite low compared
to labor. Use of capital equipment to substitute for labor was very
small. Each family farm had an average of four hand hoes, two axes,
and three chopping knives. Ox plows and tractors were used minimally.
Among the participating farm households in the selected villages, only
four owned ox plow teams and only four hired animal power. As for the
nonparticipants, only two owned ox plow teams and only six hired ox
plows. There was no farmer who owned a tractor in either group.
However, 25 percent of the farmers in each group hired tractor services
for land cultivation.

Livestock is quite important in the study area. Livestock includes
cattle, goats, and sheep. Although the Sukuma (the main tribe in Geita)
are traditionally cattle-owning people, they are essentially agricultur-
alists. Cattle ownership fills two important roles: one is economic,
and the other is social. Cattle provide a means of storing wealth, and
they are sold to provide cash for hiring labor, for fertilizers, insec—
ticides or for any emergency need. Farmers regard investment in live—
stock as an important contribution to the security of family members.
The average number of livestock per family in the study area consisted
of 8 cattle, 3 goats, and 1 sheep for the participants and 11 cattle, 5

goats, and 2 sheep for the nonparticipants.

 

103

Operating capital is required to purchase farm inputs, such as
fertilizer and insecticides, and to pay for hired labor and working
parties.2 The main source of capital in the area is personal savings,
which are generally low due to low incomes. Money lenders as a source
of credit are nonexistent because the government has banned them; they
are considered to be exploiters since they make profit for doing little
or no work. Individual family farms cannot take advantage of the
little institutional credit (offered by the TRDB or National Bank of
Commerce) since suchcreditcan only be given to communal or Ujamaa
farms. However, the governmenn in collaboration with the World Bank,
provides input price subsidies, especially for fertilizers and
insecticides.

The average value of operating capital for each group in the study
area is shown in Table 4.3. It is clear from the table that partici—
pating farmers had more operating capital than the nonparticipating
group.

Cropp ing Pat terns

The crops grown in the study area can be divided into two groups:
those crops grown in block farms and those grown near or around home—
steads. Cotton and maize fall into the first group, whereas crops
such as cassava, sweet potatoes, groundnuts, beans, and peas as well as
garden crops such as cabbages and tomatoes fall into the second group.

Most crops in the second group are intercropped. The most common
mixtures are cassava with legumes and/or sweet potatoes. Physical and

socioeconomic considerations interact to determine the types of crops

 

2Payment for working parties does not involve cash expenditures. Labor
of this kind is paid in kind, a payment that includes food and/or beer.

104

Table 4.3

Operating Capital of the Representative Farms by Months

 

 

 

 

OPERATING CAPITALa
MONTH

Pb NP
August 0 0
September 0 0
October 420.30 275.00
November 317.75 120. 85 5
December 223.00 80.20 “E
January 159.05 100.65
February 172.40 95.00
March 0 0
April 61.45 60.50
May 82.20 0
June 116.00 124.15
Total 1,552.15 856.35

 

 

 

Operating capital given in Tanzanian shillings;
it includes cash expenditures and payments in kind;
the latter was converted into monetary terms.

b
Participating Families.

CNonparticipating Families.

105

and mixtures to be grown. Among the physical factors are rainfall, soil,
and temperature. The socioeconomic factors include the need to maximize
returns to the limiting factors——especially land and labor; the need to
obtain higher output; and the need for security. The last factor indi—
cates that mixed cropping, used as a means of increasing returns to land,
is also used as a form of crop diversification, which is a strategy
against risk. Such crops, however, are not the direct concern of the
government. Most investments by the government and/or the World Bank

are geared to improving cotton and maize production in block farms. It
is also important to note here that there is no apparent serious compe-
tition for resources between cotton and maize on the one hand and
cassava, beans, sweet potatoes, groundnuts, vegetables, and so forth

on the other.

106
CHAPTER V

THE STRUCTURE OF THE LINEARrPROGRAMMING MODELS FOR THE STUDY AREA

In Chapter IV, the mathematical framework for the linear—program—
ming models used in this study was presented. In this chapter, the
linear—programming models are described, each having the following
elements: (1) an objective function; (2) an activity set; and (3) a

constraint structure.

The Objective Function

Small farmers often entertain a number of objectives (including
income maximization, output maximization, security and cost minimiza—
tion) that are not necessarily mutually exclusive. A.number of studies
(Schultz, Wolf and DeWilde) showed that a variety of objectives exist
among small farmers.

In this study, the objective function maximized the net farm in—
come on fixed factors subject to the satisfaction of household food
consumption. Every farm household studied gave first priority to the
provision of food to its members. This objective function has been
referred to as security and profit maximization (Norman). Net farm
income was defined as the total value of production less variable
costs of production.

There are two alternative approaches to incorporating more than
one objective in a single linear-programming model (Upton). One
approach is to combine the various objectives into a single—decision
criterion such as expected utility maximization. The other approach

(popularly known as the lexicographic approach or multigoal programming)

Ex) .11

107

is to employ a hierarchy of objectives, all but one being treated as
constraints. The second approach has been used in quite a number of
studies on African farmers such as those by Ogunfowora, Low (1974),
and Mwangi. It is this approach that was used in this study.

The security objective of producing staple food for the family was
specified in the matrix as a constraint to force the production of
necessary amounts of maize for meeting the minimum family-consumption
level.

Activities in the Model

 

There were eleven groups of activities included in the model.
They were as follows:

1. Crop—production activities

2. Labor—hiring activities

3. Working—party activities

4. Animal—power (ox) hiring activities

5. Animal-power (ox) owning activities

6. Tractor—hiring activities

7. Fertilizer and insecticide-buying activities

8. Crop—consumption activities

9. Crop-selling activities

10. Animal—power selling activities

11. Transfer activities

Crop—Production Activities

 

In columns A A , and A3 of Table 5.1 are outlined in the crop-

l’ 2
production activities for the representative farms in the model. There

are two main crops: cotton, which is grown solely for the market and

108
Table 5.1

Crop Production Activities

 

 

 

OIJICTIVI
h.(Cj) A1 A, A]
(lady) (bu)
cm mu on
ma usw-cns um I) o 0 "an LIL:
1 mm la 1 1 1 5. 3.25
2 IL Aug. In. 1L0!) 1 366
3 FL Sept. Ira. 76.67 101.63 n.00 : 371
4 IL Oct. Ira. 1A5.“ 176.61 10.67 i “9
5 FL Nov. In. 113.55 105.06 145.“ _<' ‘26
6 n Dec. Ilra. 151.09 126.69 113.55 1 us
1 FL Jan Mr. 121 31 160.“ 153.09 5 M6
6 In. Feb. Ilrl. 220.75 166.61 221.31 E 676
9 n Har. Nu. 169.76 - 120.75 1 :62
10 n. ”I. Im. 126.96 136.21 169.16 1 so:
11 n Hay urn. “a.” In.“ 124.96 1 us
12 n. Jun. lira. 216 37 - us.“ i £31
1] 7L July lira. - - 210.37 1 All
16‘ Hired a; Plow 0: Mn. 16.00 16.95 1 o
29 Med <- Plow on. 0. Nu. 10.00 25.” 1 us
30 " awed (I. Plow Nov. 0- un. ll 02 u so 1 96
)1 Med 0: “on Dec. cu Im. 5.1.1 ~ 73
12c Hired Tractor Tr: In. 5.50 5.20 i 0
n TSP n. 72 oo «5.00 71.00 i o
34 SA 1.. 50.00 16.00 Ao.oo i o
35 mm Taha. 13.34 - 11.11. 1 0
)6 our Taha. - s 06 - g o
u on cm x“. -675.50 - ~su.6o : o
1.6 on In: K5. - 4381.00 - i o

 

 

 

 

 

 

 

 

 

Abbreviations: CTN = Cotton; MZE = Maize; FL = Family Labor;

TSP = Triple Superphosphate; SA = Sulphate of Ammonia;
THDN = Thiodan; OPT = Output; and A = Activity.
aApplied only to farmers who hired ox plows.

bApplied to farmers who owned ox plows.

c
Applied to farmers who hired tractors.

109

and maize, which is grown for the market and for home consumption. In
columns A1Iand A2, two alternative ways of growing cotton are shown-—
early cotton growing and late cotton growing. Because it results in
lower yields per hectare, late planting of cotton is strongly dis—
couraged by the government through political campaigns and close super—
vision by extension agents. As a result, 95 percent of farm households
in the sample planted early cotton. In this study the economic—
engineering method was used to include late—planted cotton as an
activity. This method was used for analyzing the impact of an alter—

native farming system on cropping patterns, resource allocation, and

net farm income. Because maize is the main food crop, farmers would
not take any risk in planting it late. So there is only one activity
for maize. Cotton and maize are significant in terms of their contri—
bution to family—food requirements and farm income. Consequently,
they were identified as the enterprises that most adequately depict
the important production opportunities available to the family farm

in the study area.

Other crops such as cassava, sweet potatoes, groundnuts, beans, cow—
peas, and vegetables such as tomatoes and cabbages were not included in
the crop—production activities for the following reasons. These crops
do not compete for land with cotton and maize. Land allocated by the
government for cotton and maize cannot be used to grow any other crops.
All these crops together constitute about 2 percent of the total area
cultivated by each farm household. Third, in growing these crops, there
is no apparent competition for labor. For example, cassava and sweet
potatoes can be planted much earlier than maize because they are drought—

resistant crops. It was observed that most farm households began land

110

preparation for these crops in late August although planting took place
in September or early October. Beans, cowpeas, groundnuts, and vege-
tables also require light rains. They are planted anytime between
October and December. With such an extended period of planting, their
demand for labor is not as critical as that of cotton and maize. All
of these crops are weeded once and this operation can always be left
until the labor demand for cotton and maize is low. Cassavas, groundnuts,
and sweet potatoes can be harvested anytime after May without affecting
yields per hectare, whereas vegetables, beans, and cowpeas are harvested
in mid—April when labor demand for cotton and maize is not critical.
Finally, the cost of collecting data that includes all these crops was
beyond the resources of this study.

The input—output coefficients of the crop—production activities
were derived from the survey data. The activity unit (i.e., the amount
of crop production that each unit of activity represented) was one
hectare. The objective—function coefficients (Cj) for the crop—production
activities represented the costs of fertilizers, insecticides, hired labor,
working parties, animal owning and hiring, and tractor-hiring services.

Negative signs were assigned because costs reduce the income of farmers.

Labor—Hiring Activities

 

In the study area, family labor available for work on the family
farm was augmented by hired labor. Labor—hiring activities are out—
lined in columns A4 to A9 of Table 5.2. The prices used are the wage
rates per man-hour prevailing in the study area at the time labor was
hired. This was important because wage rates for hired labor varied

from one activity to another, depending on the demand for hired labor,

availability of labor from working parties and so on.

111

 

 

momma >Hwamm u Ah “mowuumm mcfixpoz u m3 muoemq wewufim am ”meowumfi>mupp<
:3 w T T .3: 25.. .E 2
2: w. T T .85 .3: .E S
New Iv. HI T .9; 33¢ .E 0H
2.. w T T .3: Sun .5 m
.3... w T T .fim .Sn .5 a
MS w T T .3: .08 .E o
o? w T T e .3: .>oz .E m
a: w. T T ...5 .So .5 a
2.? £6- 8.T NTT 2.? 3.T $6. $6- $.T 8:7 2:7 36. 2.? 3.? $958.5
m . F. B 05:. .3: un< . new . can. . own .>oz . uoo hm: . pom . can. . own .>oz . uuo m 30m
x m 5 5 5 5 5 5 5 5 E 5 5 .5 .5 .E n
S
28 .57
F. or 2.. z. 24 2.. :4 S... a. p. 2.. a. m. a Essen

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

moflufl>fiuu< xuemm I wcflxuoz 0cm wcHuH: I uonmq

N.m waan

112

-Lst

uflEfiA u BEA ”Hmuwemo mafiumuweo n 00 “mofiuumm wcflxeoz n m3 ”gonna mewuwm u an umc0flumfl>muna<

... ... 2. ... ... ... ... 2. z

 

mwflufl>wuo< zuumm I wcfixuoz 0cm wcflpwm I Momma

emscaucoo N.m magma

113

These activities have negative coefficients in the family—labor
rows, indicating that an increase in one unit of hired labor relaxes
the family—labor constraints by one unit. The wage rate of hired
labor is positive in the operating—capital rows, meaning that an
additional unit of hired labor decreases operating capital by its
wage rate. Thus, the amount of hired labor is determined not only
by the family—labor constraint but also by the availability of operating
capital.

Labor—hiring activities have negative values on the Cj of the
objective function because each unit of hired labor reduces the value
of the objective function by its wage rate. It should be noted that in
the study area, the average farm household was a net buyer of labor;
hence the selling of family labor in the form of off—farm work is not

provided for in the model.

Working—Party Activities

 

Unlike hired labor or labor obtained through contract systems
(exchange labor), labor from working parties is based on the trust the
community has in an individual or a farm household. A farm household
invites members of the community to come and work on its farm. All the
host farm household provides is food and beer. The community members
(mostly farmers) who turn up for the work do not demand any particular
type of food or beer. They simply expect to be well fed. The system
is different from exchange labor, in which farm households agree to
Provide labor for each other on arranged days, in the sense that the

Parties do not discuss any payment either prior to or after the work.

114

Working parties provide the cheapest labor apart from family
labor. However, this is not unlimited labor. The availability of
this type of labor depends on how busy other farm households are.
During certain farm operations, such as first weeding, most farm
households cannot leave their farms and work on another household's
farm. In this case, hiring of labor becomes prevalent. Working
parties are also constrained by the amount of operating capital
available to the farm household. If farmers engage in an alterna—
tive farming system (one including late—planted cotton) there are
shifts in the demand for working parties, but one cannot tell in ad—
vance how such shifts will take place.

Working—party activities are presented in columns A10 to A17 in
Table 5.2. For analysis purposes, the in—kind payments (food and beer)
were converted into money terms bynnfltiplyingthe quantity received by
its existing price or (where this was impossible) by having the farmers
evaluate the food and beer provided.

For all farming operations, family labor, hired labor, and labor
from working parties were assumed to be perfect substitutes. As with
labor-hiring activities, working-party activities have negative coeffi—
cients in the family-labor rows, indicating that an increase in one unit
of labor from working parties relaxes the family labor constraints by
one unit. Similarly, the payment to this type of labor is positive in
the operating-capital rows, meaning that an additional unit of this
labor decreases operating capital by the amount paid.

These activities have negative Cj values in the objective function

Since each unit of working—party labor reduces the value of the

115

objective function by the amount paid to it. The average farm house—
hold in the study area is a net user of labor from working parties.
Hence the use of family labor in the form of working parties is not

provided for in the model.

Animal—Power Hiring Activities

 

These activities are presented in columns A20 to A26 of Table 5.3.
The prices used are the rents paid per hectare. Animals hired for farm
work usually include an ox team; one or two men come to drive the team.
For land cultivation and ridging, family labor, hired labor, community
labor and animals hired (after appropriate conversions) are assumed to
be nearly perfect substitutes. The animal—hiring activities have nega—
tive coefficients of five in the family labor rows, indicating that
for land cultivation and ridging an additional unit of animals hired
relaxes five units of family labor constraints.

The hiring of animals decreases the operating capital by cost of
rent per hectare. Thus, the availability of operating capital deter—
mines the extent to which animal power can be substituted for family
labor.

Animal—hiring activities have negative Cj values in the objective
function since each unit of hired animal power reduces the value of the

objective function by the amount paid per hectare.

Animal—Power Owning Activities

 

A number of farm households in the study area own ox plows that are
used to augment the stock of family labor available for cultivating and

ridging the land. These activities are outlined in columns A21 to A23

116

I...

.ufleflq n
”henna %HHEmm u an ”Momma I

HZQ new ”Houflamo wcwumuoao n 00
I m<q ”cowum>fiuado

.58 33% u we: Eofioo n 2.5

”mCOflumH>munn¢

.uoBom scuomuu pone: cues Honda pokoaafio Aqv new ”scan xo pope: cues poema pwmoaefiw AmV ”soad xo pwcso
sufi3 uonma pwzoaaEm ANV muonma pmxoaan Aav 0:3 muwaumm pow pmusmfioo %Hucopcoewpcfl mums woflua>auom mmwcH
. m

 

 

in s..-
. 3 '
an.“ I lung
5 a E
I ... .: ... ... ... ... ... .... i ... ... ... fig...

mwmflufi>auo< wCHnH: HOuumuH can .wcwuHmIonm xo .wafl:30IBoam x0

m.m oHan

117

and A27 to A 8 of Table 5.3. An average farm household in the study

2

area has four oxen and ox—drawn equipment that can work for an average

of 250 ox-hours a year. Hourly operating costs were based on depre—
ciation, interest, housing, veterinary costs, and so on.1 Based on these
factors, the hourly operating cost is Tshs 3.33.

Animal—power owning activities reduce the value of the objective
function; they are, therefore, assigned negative coefficients in the
objective function. Since ox plows can substitute labor for land culti—
vation and ridging, there are cotton cultivation and ridging as well as
maize cultivation and ridging by owned—animal—power activities. These
are assigned zero coefficients. However, animal—power owning activity
requires labor to run it. So there are positive family—labor coeffici—
ents in the family—labor rows, meaning that any use of animal power de—

pletes family labor by the number of hours indicated.

1(a) Depreciation: Purchase price for 4 oxen @ Tshs 450, 1 plow at Tshs
350; and 1 cart at Tshs 1700. Use-life for ox assumed to be 4 years
each; for plow, 5 years; and for cart, 6 years. Total cost for six
years is Tsh 4825; less resale value of Tshs 525 for each ox at
slaughter: (4825—2100) = 2725. Assuming straight—line depreciation,
annual cost = Tshs 454.

Interest: Charges at 8% percent on Tshs 2725 compounded over 6 years
= 287/year.

Maintenance: Usually consists of feed requirements, but since oxen
in the area are grazed on communal land, the opportunity cost of
maintenance is taken to be zero. However, the oxen are lightweight
(200-250 kg) and their work effort is very low. A day's work is not
more than 4 hours.

Housing, veterinary, etc. These are charged at 20 percent of depre—
ciation on 91 Tshs per year.

Cost of oxen and equipment based on data from Work Bank, Appraisal of the
National Maize Project, Report No. 89a/TA, Washington, D.C. Dec. 8, 1975,
and field notes. Costs updated for 1980. The project was conducted for
the United Republic of Tanzania.

(b

v

(c

v

(d

v

118

Tractor-Hiring Activities

 

In columns A29 and A30 of Table 5.3, tractor hiring activities
are outlined for the representative farm in the model. The prices
used are the rents paid per hectare. For land cultivation, the
following are assumed to be nearly perfect substitutes; family labor,
hired labor, working parties, and hired animal power (after appro—
priate conversions). They have negative coefficients of 15 in the
November family—labor row, indicating that for land cultivation an
additional hour of hired tractor use releases 15 units of family—
labor constraint.

The hiring of tractors decreases the operating capital by cost of
rent per hectare. Thus, the availability of operating capital deter—
mines the extent to which tractors can be substituted for family labor.
Tractor—hiring activities have negative Cj values in the objective
function since each unit of hired—tractor power reduces the value of

the objective function by the amount paid per hectare.

Fertilizer— and Insecticide-Buying Activities

 

In Table 5.4, columns A31 to A fertilizer- and insecticide—buying

34
activities are listed. Farmers buy two types of fertilizers (triple
super—phosphate (TSP) and sulphate of ammonia (SA) and two types of in—
secticides, Thiodan and DDT. Although the two types of fertilizers are
used for both cotton and maize, Thiodan is sprayed only on cotton and
DDT only on maize.

Data on different levels of fertilizers and insecticides needed to
test their economic impact on farmers net farm income and resource allo—

cation was not available. Consequently, only one level for each of those

inputs was used in the analysis.

Vanilla-r - and lnaacllcule - Buying and Selling Acttvltlea.

119

Table S.‘

and Food Con-Innov- Actlvuy

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

I
\ 'I
\E'fm" ‘31 ‘32 I ‘11 ‘34 ‘35 "36 ‘37 ‘35 ‘39
\ n. (CL) .
BUY IUY IUV BUY SEJ. SEA. cms. KENT KENT
IOU \ UNITS TSP SA TNDN DDT CTN. MZE. HIE. 0X 0X SIGN LILS.
\ PW PW
\ 00:. NOV.
IESDUICES ~ I-l M —l.0J -l].y~ -5.0b 3.10 2.50 0 180.00 180.00
4 Fl Oct. lira. 55.15 a Ah?
5 1 n Nov. llra 40.10 no
29 I 06m ()1 Plan (I
I an. N". l 1 115
10 um 0a flaw On
I Nov. lira. l : 9!
33 I 15? In. —I i 1 o
I
34 SA 1;. l —l 1 i 0
35 mm 'l’lhs I l — I <_ u
.
36 nor 'r-m I -l L 0
17 0: on. Ly. | 460.00 ; 1.20.30
35 0C Nuv. Taha. ' LO] 480.00 5 ”7.73
)9 06 Dec. Tuba. 1 13. ll. l 223.00
45 ' (12 June I 5.00 -3.20 -2.50 1 “0.00
I40 0(' July Tahl ’ -3.20 i 0
L7 DPT cm. In. 1 l 1 0
AB on n22. lg. I I l 0
49 cans. nzz. u. I _<_ .346
66 Own 0! How Ila 55.15 1 0
our.
Capacity
67 can 0- "av u. L0.10 _<_ 0
luau
Capacity ‘
bbreviattuna: TSP - Triple Superphoaphate; SA - Sulphate of Dnla' 1ND" - Thioda . Cl'N - Cotton: HZ! - Nata ;
CONS. - Contention: l1. - Pauly Labor: (X: - 07:“qu capital; on - Output

 

120

The prices of these activities reflect the cost associated with
the purchase of these inputs. The prices used are those prevailing
in the area in the 1980—81 season. The Cj values of these activities
are negative since they reduce the value of the program. The activities
have negative coefficients in the row columns indicating that an increase
of one unit of fertilizer or insecticide in the basis increases the
stock (assumed initially at zero levels) of these inputs. The positive
coefficients in the operating—capital rows show that the purchase of
fertilizer and insecticide requires an expenditure of operating capital

equal to the price of the fertilizer or insecticide.

v.5

The prices of these inputs are fixed by the government prior to
the farming season. Farmers are subsidized, so the prices used in the
model are subsidized prices. TSP is subsidized by about 30 percent,
SA by 40 percent, Thiodan by 50 percent, and DDT by about 80 percent.

These prices are indicated by Cj values.

Consumption Activities
When this model was formulated, it was assumed that household food
consumption is satisfied before any sale activities are undertaken. In
this study there is only one consumption activity, which is shown in
Table 5.4, column A37. The minimum household—consumption requirements
were determined from the farm-survey data. The positive coefficients
. . . . . . .th
attached to the consumption activ1ty indicate that one unit of the j

commodity to be consumed depletes the corresponding output in the output

row.

121

Crop-Selling Activities
Crop—selling activities are shown in Table 5.4 columns A35 and

A The model is contracted in such a way that the selling of food

36'
crops takes place only after consumption needs have been satisfied.
It is also assumed that selling prices reflect the cost of storage;
therefore, no storage activities are included in the model. The
price of maize (Cj value) is that prevailing in the local (black)
market in 1980-81. Local markets are assumed to be competitive, and
prices usually differ from the government-controlled price. There
is no black market for this crop.

The objective—function coefficients are positive because selling
adds to the value of the objective function. The row coefficients of
the output of the crops are also positive since selling activities re—

duce the stock of that crop.

Animal-Power Selling Activities

 

These activities are shown in Table 5.4, columns A38 and A39.
When a farmer owns an ox plow, he is likely to rent it to other farmers
after employing it on his own farm. This is very common in the area.
It is a source of income that farmers need to meet their cash expenses.
These activities have positive coefficients in the family-labor
rows, indicating that a unit increase in the level of these activities
reduces the amount of family labor available by the value of the
coefficient. The same can be said for the positive coefficients in the
animal—power owning rows. The negative coefficients in the operating—
capital rows indicate that a unit of animal power sold increases operating
Capital by the amount shown. The objective function coefficients are posi—

tive because selling adds to the value of the program.

122

Transfer Activities
In columns A40 to A47

that are used to pass surplus capital from one month to the next

of Table 5.5, activities are represented

month during the farming season, even if this transfer is not re-
quired directly to finance operations (Barnard and Nix, p. 443).

These activities have zero value in the objective function, a posi-
tive sign for the last month in the operating—capital row, and a nega—
tive sign on the amount—of—transfer row, which means that the capital

accumulation at the end of the crop year is increased.

Restrictions in the Model

 

In the study area, farmers are faced with certain restrictions or
contraints in their production activities. These restrictions are out-
lined in Tables 5.1 to 5.5 in the columns labeled R.H.S. The restric—

tions are all defined below.

Agricultural—Land Restrictions

The land available to the representative farms influences both the
acreage allocated to various crops and the cropping patterns undertaken
by the farm firm. In the study area, land for cotton and maize is allo—
cated by local government officials. Farmers are not expected to grow
crops other than those on land allocated. No farmer is allowed either
to sell or buy land. There was no evidence of land renting. Consequently,

in the linear programming model, these options are not provided for.

123

. ._..._ w—-

Hmo«umu acHHmHmao I 00 "HmHHamo mcaumuoac co Hmemcmuw a doe

HmcoHumH>munL<

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

I H 1 1
~ . n _
um.oNq _ H 1 meme “ HmuHamo 1 an
_ . mCHucw “
_ n _
o i _ H H. _ meme >.HsToo m as
. _ .
.I. . _
oo.oHH m M H H. . meme mcseIoo ms
.II.
OH.Hm m H HI meme Hszoo as
I._
m..He .. i H HI . meme .Ha<Ioo ms
.
0 H4 H 1 HI meow .meIoo m.
oa.NHH 4. a H H. meme .H»H.oo H4
..Il . . H _
mo.omH M . w H H. meme 1 .cmeIoc Os
“II. . _ _
oo.mmu n . _ m n H H. meme _ .umoIoo _ mm
_ a .
_-I_ _ i w .
ma.uHm m t i M M , H HI meme 1 .>oquo . mm
. . H _ H
_ * H _ . _
_ .
mm.c~q . I u w u H H . H meme 1 .Huquo 1 am
.... . _ u _ _ m
- h v H m H _ . H I . II
A M Hooo.o H o H o I a 1 o . o 1 o n o H o o a . .
m 1 m, mHue H mass 1 ,mx _ .Haq .me H .Hmu M .cme H .uwo _ .po.a _ mmumaummm _
.m.x.m w arsHOU . -Icas m Isa: m I.ta< _ ..mr _ -.HmH , -.cme 1 I.uoo 1 I.>oz a I.Huoi mHHzH t u
.U. . . _ . . . H a II
.NW 1 M M . _ u _ H 1 . rm
. «Hnmzeap _ Hoe _ oce _ no» N use use ooh . UOH 1 nos . UOH 1 .
_ . H H i m H u a H _
. 4 4 I H a H ...I . .
. . . 1 _ - _ .A ”UV cw //_
as H. I a. 1 o. . .4 1 In mg. m. _ Hg 0. _ H I .
< .4 H < u < m < « ¢ e _ < w < a a WHHHImmmo
.IH H .. 4 .HHmceHw

124

Labor Constraints

Family labor restrictions are specified on a monthly basis in the
model.2 The row unit is man hours. The amount of family labor avail—
able to each representative farm for each month was estimated from the
farm-survey data. The family—labor stock could be supplemented by labor
hiring or by labor obtained through working parties. The amount of the
last two types of labor, however, depends on the total labor requirement
relative to the amount of family labor available, the amount of hired
labor available in relation to its wage rate, the amount of labor from

working parties, and the amount of operating capital, both cash and food.

gyerating—Capital Restrictions

In this study, cash expenses were used as an indication of the
amount of operating capital. Part of these cash expenses, however, were
derived from payments in kind (food and beer) for working parties. Food
and beer offered for working parties were valued in Tanzanian shillings.
The amount of funds available for cash expenses on the representative
farm was set equal to the amount estimated to have been spent on hired

labor, working parties, fertilizers, insecticides, hiring of ox teams,

2Beneke and Winterboer (p. 64) indicated that including a single labor
restraint implies that labor can be freely substituted among seasons
of the year. Labor is likely to have different opportunity costs in
different seasons. Realistic planning requires taking account of the
seasonality of labor requirements and restraints. Restraints should
be formed to focus on those periods of the year in which labor allo—
cation is critical. The remaining noncritical periods can also be
included to provide a complete accounting of labor within the system.

125

and hiring of tractors for the crop-production activities during the

1979-80 cropping season. These were estimated from the data obtained

in the survey conducted for this study.
The operating—capital constraints were also specified on a

monthly basis. Barnard and Nix (p. 439) offered two primary ways of

incorporating operating capital into a linear programming matrix. One

is to use transfer activities to pass surplus capital from one period

to another during the year. The other is to accumulate capital balances

in successive periods. In this study, transfer activities were used to

pass surplus funds from one month to another during the cropping year. ?
There was no data on short-term credit availability or savings. -”

Thus, no borrowing activities was included in the linear programming

models to supplement operating capital. Hence the operating—capital

restriction used was the minimal estimate of capital availability.

However, this restriction was relaxed in the analysis by the assump-

tion that increases in farmers' incomes due to increased product

prices would lead to an increase of 10 percent in operating capital.

Food-Consumption Constraint

 

It was assumed that the representative farmer is motivated by the

"first security rule,"

that is, the first priority is to produce the
family's food—consumption requirements. Therefore, the farmer needs

to produce the minimum amount of maize for the family consumption. Thus,
a minimum amount. The amount of food consumed was estimated from the
survey data. These data were aggregated to obtain the average consump—

tion of the household per year. The average was then used as a con—

straint for maize produced by the farm firm.

126

Nonnegative Restriction

One of the requirements for the use of linear programming is the
nonnegativity of the activities in the model. None of the activities

discussed above could be operated at negative levels.

Some Limitations of the Model

What has just been presented does not exhaust the list of activities
and restrictions that could be included in a linear programming model of
small-farm agriculture. Examples of activities not included in the
model would be the following: nonfarm activities of family members;
various types of the activities in the farming process, including pro-
duction of cassava, sweet potatoes, groundnuts, beans and cowpeas.
Additional levels of fertilizer or insecticide use could have been in—
cluded as separate activities. The number of activities in the model
depends on the availability of data and the objectives of the study.

Large and complex models such as systems simulations, which are
valuable for analyzing domains of problems, both of economic and non—
economic disciplines (Dent and Anderson; Manetsch) are costly in terms
of time, money, and other resources. It is not always certain that the
benefits to be derived from additional activities in terms of precision
for planning purposes are sufficient to justify the additional costs of
employing a systems approach for analyzing the defined problem.
Further, results from such complex models may be difficult to interpret
in terms of tracing the logical causal relationships between a change
in this study was kept as simple as possible but complete enough to

reflect the farm situation in the area of the study.

127

In this chapter, a detailed description of the structure of the
linear—programming models employed in this study and how the support—
ing data was generated were presented. The application of the models

as described here and their variants are presented in Chapter VI.

128
CHAPTER.VI
ANALYSES OF THE RESULTS OF THE APPLICATION OF THE LINEAR-PROGRAMMING

MODELS

In this chapter, the analysis is presented of the results of the
application of the linear—programming models for the participant and
nonparticipant representative farms. The analysis is focused on:

(1) the possibilities of increasing farm income through improved allo-
cation of existing resources; (2) the determination of optimal cropping
patterns under existing resource constraints, prices, and technology;
and (3) the extent of resource use and productivity. At a later stage,
the impact of changing relative product prices, input prices, and
operating-capital level on farm income, cropping patterns, and resource
use is explored.

First, the product price was varied while other resources, input
prices, and operating capital remain unchanged. There were three
relative product-price changes. These were 0 percent and 80 percent;
25 percent and 80 percent; and 25 percent and 150 percent increases
for cotton and maize, respectively. These price increases were in
keeping with the government—price changes that ranged from zero per-
cent between the 1975—76 and 1976-77 farm seasons to 25 percent between
the 1979-80 and 1980-81 season for cotton; and market-price increases
for maize (within MWanza region) that ranged from 80 percent to 150
percent above the government price between July and October 1980.

‘Many farmers interviewed by the researcher confirmed that such unofficial
price increases for maize have been common since the late 19603 and have

since been used in decision making.

129

Second, input prices, especially those of fertilizer and insecticide
were increased by up to 30 percent for TSP, 40 percent for SA, 80 per-
cent for DDT, and 50 percent for Thiodan, with existing product prices,
other resources, and operating capital remaining unchanged. These in-
creases caused the farmers to pay the actual input prices, which were
subsidized by the respective percentages. The goal was to lower these
subsidies as farmers' incomes increased.

Finally, the operating-capital levels were increased by 10 percent
(if the resource was limiting) with existing resources; product and
input prices remained unchanged. The purpose of this increase was
that, given favorable product prices, farmers' incomes would increase,
and, in turn, farmers would increase their operating—capital levels.

In the first section of this chapter, a discussion is presented of
the comparison of the optimal plan of the participants' representative
farm enterprises under (1) the maize and early-cotton farming system;
and (2) under the maize, early-cotton, and late-cotton farming system.
The comparisons are based on the changes on farm income, resource use
and productivity, and return to resources. In the second section of
the chapter, a review is included of nonparticipants' representative
farm enterprises in order to curve a basis for comparing the optimal
plan between the participants and nonparticipants in the cotton project
in Geita District financed by the WOrld Bank.

The objective of the third section is to show the different effects
of improved agricultural practices (such as use of better seeds,
recommended amounts of fertilizers and insecticides, and recommended

number of weedings) on enterprise combinations, resource use and

130

productivity, and average returns to limiting resources. A.summary of
the results of the various linear-programming analyses is given in the

final section of the chapter.

Optimal Organization of Existing Resources and Prices for Participants,

 

Model A

Mechanization and Manual Technologies: Maize and Early Cotton

 

In Table 6.1 a comparison is given of bench—mark and optimal organi-
zations of the representative farm in the Geita cotton project. Four
models are shown in the table: one for only labor user (A1); one for
farmers employing labor and owned ox—drawn equipment (A2); one for those
hiring ox—drawn equipment in addition to using labor (A3); and one for
farmers using labor and hiring tractors (A4). As a basis for comparison,
the following economic measures were employed: net farm income per hec—
tare, net farm income per man hour, and net farm income per unit of
operating capital.

The optimum net farm income for Model A1 totaled Tanzanian shillings
(Tshs) 4450.96 compared to Tshs 3884.47 from the actual average for the
representative farm in the sample. This represented a 14.58 percent
increase. The optimum plan included 2.05 hectares for cotton compared
to 1.29 hectares from the actual average, and 0.39 hectares for maize
compared to 1.23 hectares. These represented an increase of 58.91 per-
cent and a decrease of 68.29 percent for cotton and maize respectively.
The big decrease in hectarage planted in maize requires some explanation.
The price of maize used in deriving the optimal organization was the
official price of Tshs 1.00. Based on this price, farmers would plan
to grow maize for household consumption only and very little if any for

market

l3l

.A§¢v Houomyu coma: mam momma so Am<v Emma x0 vwhflfi mum HOQMH .AN<V Emwu
xo wwcao man noan go AH< Hovozv macaw Momma uocuflm hoamfiw mquhmm .stuwwou wwzoamam uoc mum owwSHo.w.o
.wapmw w>wumumwmmumou

Scum woumuocww mucoaofiwwwoo Eonm vmcfiwmw spam Hmduom wnu ms Mumelsocmn oau .moanmu ucmsvmmnaw wow wasp :Ha

.mcomwumano quufififlumm Ou moanmu ocmnwwmnsm cam wasp CH powaaocfl ms coaumNHcmmuo xmeISUGop mnhm

 

 

 

 

 

 

 

IIII _ 1
.6? Ho; 0.5. 1 3; 8.124 ,, 8x 1 is? 1 $6 ,1 24 :1: 1 H335 mat-33:2
no.“ I 2; is I ,1 2H 1 a.“ I , NTH ,1 am.“ I 1 SH 1 an; 1 E: , :51. 339., 1 .51:
3.2. 2:5: 3.2.. 1 3.132 1 2.2+ 1 2 is: .H 2+ 1, 2ng _ 9.321 23 371%
3.0? 1 ”figs HTS- 1 2.23 11 3.97 1 $.51 2.3. 1, 2.2,. 1,. 3.23, 2: “5:5 21:08
8HI 1 o I 1 I 1 I I I 1 I 1 mn.m 19:. Q...“ ,. €0.55 noun-b.»
I I on o I 1 «0 mm 1 I , I I 1 1 omAq 11 mum H5 11 Avmuumv Adah 5v
I I I 1 I 1 no H + ,1 mo 5 . 1 I 8.2 1 E: .7 2.25: it sou
. SJ I 8.th «TH I 1 2.2.2 1 3.0 I 1, 3.x: , 3;. I ,_ 2.th 1 2.31:1 ...: 1 Se: .38
9:? Six 1nd? 3.03 1 3.3. 1 8.2; 1 84? 1 $.ch _ 8.8m 1 9:1 ,, It: €23:
3 so- ,1 a n 12 so. 1 a H 1 2.2.- ,1 2H 1, 2.3- 1 Ra 1 8&2 1 :H1 1 :51 1.3:.
8.2. 3.3: 2.2+ ,1 2.3.: 1, 8.3+ 1 3.22 . 2H? 1 3.8: 2.331 3:. _, :31. Si:
no. ...H 2H- 11.; 109T. 12H no. ...; 1 2N 1 .111 1 21:38
, ,
S a? an o S m? 1 on o 1 3.3- 1 an o 1 2 3- ,1 and 1, 2H 1 :1 1 135: Had
Ho 2. S N 2 H? 1 2: 1 2:? _ 24 1, Has? 1 8H _ zH 1 :1 1 26:08 123
3.3+ 3.2: 3.3+ 1 8.39 1 2.8.. 1 2J2? 2.2+ 1 3.3311 3.;31 :1: 1 CE: I35 I: ooz
1 >
H , ,
$56 8 H08: nun-6 1 n< H23: _ $36 1 3 Hana. $56 1, H< Honor, 1 1,
u u 1 u u , 1 5!: SE. 1 EH
.25.:5 1 $52: 1 1

 

mucmmfiowuumm pom couuoo xaumm mam wNfimz pow ooflum Hmwuwwwo mfiu cam
mwopsomwm wafiumflxm sows: wCOHumwwcmwuo Hmfifiuao cam mxumzlfiocmm mo comflummsoo

H.o mHLmH

132

Since less land was used, most resource use also decreased.
Whereas family labor increased by 21.92 percent, hired labor and labor
from working parties decreased by 96.10 percent and 62.66 percent
respectively. Total labor decreased by 4.47 percent.

Operating capital also decreased from Tshs 1552.15 to Tshs 673.28,
or a 56.62 percent decrease. The decrease in cash and payment in kind
resulted from the decrease in the use of hired labor, working parties,
and inputs such as fertilizers and insecticides. The average per hec-
tare income was Tshs 1824.16, an 18.35 percent increase. The return
per hour of family laber was Tshs 1.28 compared to Tshs 1.36, a 5.88
percent decrease, whereas that of a unit of operating capital was
Tshs 6.61 compared to Tshs 2.50, an 84.4 percent increase.

When the use of owned animal—drawn equipment was introduced
(Model A2), the area planted in cotton increased from 1.29 hectares
to 2.23 hectares, an increase of 72.87 percent. The area planted in
maize remained unchanged. Consequently, the optimum net farm income
came to Tshs 4675.26 compared to 3884. 47 from the actual average.
This was a 20.35 percent increase.

As as result of more land being used in the optimal organization,
it would be expected that total labor use would increase. However,
farmers devoted fewer hours during land preparation because of the em—
ployment of owned ox—drawn equipment. The analysis showed that land
preparation accounted for only 102 man hours per hectare (compared to
264 man hours per hectare in Model Al). Since ox—drawn equipment was
used on land preparation alone, time saved in the early part of the

farming season was almost compensated for by the increased labor

133

required for later weedings. Thus, despite the increased cropping
area, total labor use declined by 0.61 percent. Most of this decline
came from hired labor (96.10 percent) and from working parties (51.29
percent). Family labor increased from 2843.70 man hours to 3527.85
man hours, or a 24.06 percent increase. Use of owned ox—drawn equip-
ment increased by 3.98 percent from the actual average in the sample.

As a result of changes in the use of hired labor, working parties,
owned ox—drawn equipment, and operating capital decreased by 46.22
percent. The net farm income per hectare was Tshs 1784.45 compared to
Tshs 1541.45, a 15.76 percent increase. The net farm income per man
hour of family labor decreased from Tshs 1.36 to Tshs 1.32, a 2.94
percent decrease whereas that of a unit of operating capital was Tshs
5.60 compared to Tshs 2.50, an increase of 124 percent.

In Model A3 (use of labor and hired ox—drawn equipment) the cotton
area increased from 1.29 hectares to 2.08 hectares, an increase of 61.24
percent. There was no change in area planted in maize. As a result of
these changes, use of family labor increased by 22.77 percent, whereas
hired labor and labor from working parties decreased by 96.10 percent
and 50.32 percent respectively. Total labor decreased by 1.34 percent;
this decline was explained by the decrease in total area cropped as well
as by employment of hired ox—drawn equipment. Operating capital de-
creased from Tshs 1552.15 to Tshs 982.30, a 36.71 percent decrease.

Although the net farm income per hectare increased from Tshs 1541.45
to Tshs 1828.38, an increase of 18.61 percent, the net farm income per
man hour of family labor declined by 5.14 percent, and the income per

unit of operating capital increased by 87.6 percent.

134

The optimal organization for Model A4 was similar to that of A1.
Hired tractor services were not included in the optimal solution.

This left labor activities to determine the optimal organization.
Since labor coefficients and the other coefficients in Model A4 were
the same as those in Al, the optimal results of the two models were
the same.

In deciding how to allocate farm resources between the two crops,
however, farmers did not use the official price for maize. In contrast
to cotton, it is easy to find buyers for maize in the black market;
these buyers pay a much higher price. This price is determined com—
petitively in the food market. It was observed during the research
that the price for a kilogram of maize was about Tshs 1.50 to 1.80
during the period immediately following harvest, but increased quite
sharply to Tshs 2.50 within two to three months. Through observation
and informal discussions with farmers in the area, it was concluded
that many farmers made their resource—allocation decisions on an ex—
pected maize price of Tshs 2.50 per kilogram. It was learned that
this practice has been going on for the last six years and is likely
to continue in the future.

In Table 6.2, a comparison is given of bench—mark and optimal
organizations of representative farms when the official price for
maize was replaced by a free—market price of Tshs 2.50. The results
of the analysis showed that in Model Al; the area planted in maize in—
creased from 1.23 hectares to 1.30 hectares, or an increase of 5.69
percent from the bench—mark average. The area planted in cotton increased

from 1.29 hectares in the bench—mark average to 1.65 hectares in the

135

 

 

 

 

 

 

 

 

 

 

 

monH<NHz<omo A<ZHHmO

 

 

 

Mmo.ms+ am.m so.~q+ om.q mm.~m+ No.q mo.wm+ am.n mo.m acme amuaamo wcauauuao\Hez
.qo.o . om.a mm.n . Hm.H mq.m - ~m.a qe.a - om.a oo.H mama Henna wanna: do pm\ch
om.o + q¢.shma -.a + nn.ooma mm.a + H~.amwa ow.o + sq.nawa ow.soma onus mm\Haz
mm.mm- «s.o~oH om.ma+ oq.soma No.0N- nm.aq~a mm.mm- ca.o~oa ma.~mma anus Hmuaamo maaaaumao
ooH- o u - - - - a mm.m mum .u»s ouuam honours
- - ~m.~fi+ ~m.oq a - - . om.aq mp: xo Awopamv sums xo
- - u . mm.mH+ ma.om - . ao.ms up; No Anmnzov anus xo
Nm.NH+ sm.~qss mo.ma+ oo.nmqq oa.oa+ mm.qmms mm.NH+ q~.~qqq o~.mqom mum gonna Hauos
mm.¢H- N~.Nmo mm.qal Hm.nmo No.HH- oo.mah mm.oa- Na.~mo oo.aom my: sauna moaxuoz
qa.sou mm.qu qs.4o- mm.qoa «n.40- mm.soa «n.40- mm.qu oo.ma~ mum roams can“:
mm.N~+ mm.qmom qm.o~+ 84.noom m~.~m+ oa.aosm Nm.~N+ oa.smom oa.mqm~ mum roams saaama
_oo.aa+ ma.~ nw.aa+ sm.~ qw.aa+ No.m so.sa+ mo.~ ~m.~ am new; Hmuoe
mo.m + om.a Nm.s + mm.a mm.HH+ nm.a as.m + om.H m~.H mm Amuamxo can;
so.sN+ mo.a Ho.aN+ mo.H Ha.a~+ mo.a Ho.s~+ mo.H a~.H mm Accouoov can;
so.na+ oq.mmmm aN.mH+ mo.qaom mn.H~+ oo.omam so.aa+ oq.wmmn om.oo~q arms fichv wagons spud “62
“ego M s... ammo: 8736 m< EB: 8546 N< 58: 835 a... .58: g: .22: 5.:
N a N _ N N .623

 

muamafiuauumm Mew mafia: no“ moaum omxumzlmouu on“ ncm acuuoo mapmm new oofium HmfiUfiwmo och

.wmopzommm wcaumfixm Hubs: mCOwumNficmmpo Hmafiuao vcm xumznnucmm wo comfipmaaoo

N.o mfipmh

optimal organization. This represented an increase of 27.91 percent.
Consequently, the optimum net farm income came to Tshs 5538.46 com-
pared to 4706.86 from the bench-mark average. This was a 17.67 per—
ent increase.

As a result of more land being used in the optimal organization,
total labor use increased from 3948.70 man hours in the bench—mark
average to 4442.24 man hours in the optimal organization, a 12.53
percent increase. All the increase, however, came from family labor,
which increased by 22.82 percent. The use of hired labor and labor
from working parties decreased by 64.74 percent and 19.53 percent
respectively. Changes in the use of hired labor and labor from work—
ing parties led to a decrease in the use of operating capital from
Tshs 1552.15 in the bench—mark average to Tshs 1026.74 in the optimal
solution. The change represented a 33.85 percent fall.

Although there was a 17.67 percent increase in the optimum net
farm income, the optimum net farm return per hectare increased by 0.50
percent, and the return per hour of unpaid labor actually fell by 9.64
percent. In the optimal organization, more land and more labor were
used in the production of cotton and maize. This explains why the
returns to land increased by a small percentage and the return to
family labor fell. The net farm income per unit of operating capital,
however, increased remarkably from Tshs 3.03 in the bench-mark average
to Tshs 5.39, an increase of 78.03 percent. This remarkable increase
was explained by the increase in the optimum net farm income on the one
hand, and on the other, by a substantial decrease in the optimal amount

of operating capital used.

137

In Model A2, the inclusion of owned ox—drawn equipment led to some
notable changes in the use of resources. Although the area planted in
cotton remained the same as that in Model A1 (i.e., increased by 27.91
percent from that in the bench-mark average), the area planted in maize
increased by 11.38 percent from the bench-mark average (and by about
5.38 percent above that in Model A1). Total land use increased from
2.52 hectares in the bench—mark average to 3.02 hectares in the optimal
organization, an increase of 19.84 percent. The net farm income ins
creased from Tshs 4706.86 to 5729.60, representing a 21.73 percent
increase from the bench-mark average.

The increase in land use led to an increase in total labor use of
16.10 percent from the bench-mark average. This increase was wholly
accounted for by the increase in family labor used, which increased
from 2843.70 man hours to 3761.99 man hours, an increase of 32.29
percent. Most of the increase in family-labor use came during weeding,
although some increases were also noted during ridging. The results
indicated that during land preparation, family-labor use was only 126
man hours per hectare (compared to 267 man hours in Model A1). During
weeding, family-labor use increased from 394 man hours per hectare in
Model A1 to about 450 man hours per hectare. The reason (as explained
earlier in this chapter) was that weeding was still done by labor.

Use of hired labor and labor from working parties declined by 64.74 per-
cent and 11.02 percent respectively. Further, the use of owned ox—
drawn equipment increased from 75.09 ox hours in the actual average

to 89.98 ox hours; this represented an increase of 19.82 percent.

138

The changes in the use of farm resources discussed above led to
a decline in the amount of operating capital used; it declined from
Tshs 1552.15 to Tshs 1241.35, representing a 20.02 percent fall. The
net farm income per hectare, and per unit of operating capital in-
creased by 1.58 percent, and 52.33 percent respectively, whereas the
return to unpaid labor decreased by 8.43 percent.

In Model A3, the change in the area planted in cotton was the
same as in MOdels A1 and A2. The area planted in maize increased
from 1.23 hectares in the actual average to 1.32 hectares, an increase
of 7.32 percent. Compared to Models A1 and A2, this change was
better than that in the former model and worse than in the latter.
This indicated that the use of owned ox—drawn equipment might be more
profitable than that of hired ox-drawn equipment. (See also the
difference in the net farm incomes for Models A2 and A3). The net
farm income increased from Tshs 4706.86 to Tshs 5614.93, and increase
of 19.29 percent.

The use of family labor increased by 29.94 percent whereas that
of hired labor and labor from working parties decreased by 64.74 per—
cent and 14.83 percent respectively. Labor use as a whole increased
by 13.63 percent with the whole increase accounted for by the increase
in the use of family labor. The decline in the use of family labor
during land preparation was not as substantial in this model as it was
in Model A2 (126 man hours per hectare in Model A2 compared to 241 man
hours in A3). The reason was that part of land cultivation was done
by labor and part by hired ox—drawn equipment; the use of hired ox—

drawn equipment was limited by the amount of available operating

139

capital in October and November (see Table 6.3). Although in the
actual average, 41.50 ox hours were hired, in the optimal organiza-
tion 46.82 ox hours were hired. This was an increase of 12.82
percent. Operating capital decreased from Tshs 1552.15 to Tshs
1304.40, or a 15.96 percent fall.

There were increases in net farm income per hectare, and per unit
of operating capital. Net farm income per hectare increased from Tshs
1867.80 in the actual average to Tshs 1890.55 in the optimal organiza—
tion. This represented an increase of 1.12 percent. Although net
farm income per unit of operating capital increased from Tshs 1867.80 to
1890.55, a 1.22 percent increase, that per unit of unpaid labor de—
creased from Tshs 1.66 to Tshs 1.51, a decrease of 7.83 percent.

Again, the results for Model A.4 were the same as those for Model
A2 since the hiring of tractor services was not included in the opti—
mal organization. The exclusion of tractor-hiring activity in the
optimal organization was explained primarily by the cost of hiring
it. It cost almost Tshs 300 per hectare to hire a tractor compared to
about Tshs 180 per hectare to hire an ox team.

The results in Table 6.2 clearly show that risk on food production
was removed when the black market price for maize was used in the
analysis. Compared to the analysis under the official price for both
cotton and maize, the black market price for cotton increased the area
in maize by about seven times. Even under poor weather and other
adverse conditions, this expansion in maize hectarage could produce
enough food for household consumption. Further, year to year higher

variability in cotton yields (as expressed by many farmers during

140

.mamvoE_HHw sH muosvoua Hmcfiwume oumu o>m£ wowsHUCH uoc moousommu .mmapmu osmovmmnsm mam wasp ch

 

 

 

 

 

 

 

 

 

 

 

 

 

 

oo.a Ho.o mm.a mo.o ma.a am.o oo.a Ho.o pm .>oz sza m3
am.a mN.o o ma.o o o am.a mN.o Hm .uuo 9x4 m3
aq.o 0 No.0 0 o~.o o oq.o 0 meme .>02 00
mq.o o mm.o o Hm.o o mq.o o mama .uuo oo
oo.m so.m oo.m oo.m so.m so.m so.m oo.m meme Boo
am.ma sm.ma am.ma am.ma qm.ma am.ma sm.ma qm.ma mama znms
mo.H mo.H mo.H mo.H mo.a mo.H mo.H mo.H mm «m
am.H am.a am.H am.a qm.a am.a am.a am.a we awe
mm.o o oa.a o oH.H o ma.o o as am: as
NH.H o HN.N o mo.m o NH.H 0 pm .nmm as
mm.o mo.o mH.H cm.m wN.H mo.~ ma.o mm.o Hm .cme as
wo.H o oq.H o oo.a sq.o wo.a 0 pm .umo as
ma.m Hq.m mm.~ mm.m sm.m a©.m ma.N Hq.N Hm .>oz as
aa.~ om.a o o o o am.~ oa.a um .uuo as

om.NnmN2m oo.Humsz om.NumN2m oo.aummzm om.mumwza oo.Humsz om.mnmmzm oo.anmN2m

o~.muzsom oN.muzsom oa.muzeua o~.muzsom o~.muzsom Om.muzsom o~.mnzsom om.muzeom
9H2: momDOmme

q< amooz m< ammo: Na ammo: H< gage:
q

mucmawofiuuwm now

< oo Ha mauve: pod

m

m.© wanme

mmousommm mo Amsmh aw mm>ZV muosvoum msam>lfimcwwpmz

141

informal and formal interviews) means that any solution that leads to
more production of maize stabilizes farmers"net farm income

The results, shown in Tables 6.1 and 6.2, deviated substantially
from the farmers' current practices. This deviation is a reminder
that linear programming is an exercise in normative economics. Based
on the assumptions used and the constraints given, it indicates how
the farmers' incomes could be maximized. It should be noted, however,
that the omission of still other factors from the models may prevent
the models from representing in the study all aspects of farmers'

behavior.

Marginal—Value Products of Resources Under Model A

 

The marginal—value product (MVP) of disposable activites is de—
fined as the increase in the value of total output that is obtained
from the use of additional units of the resource with all other inputs
held constant. This condition may not be met in a linear-programming
framework because technological coefficients for activites are assumed
to be at fixed ratios to one another. Thus, the increase in one unit
of only one input requires the increase in other inputs in order to
keep the ratio of coefficients fixed. Despite this deficiency, MVPs
can provide information on the most likely resources to be expanded in
order to increase the value of the objective function. The MVP of a
resource is constant over the specific range, and the solution holds
until other resources become limiting. At that point, another enter—

prise organization becomes optimal, and the MVP of the resource changes.

142

The MVPs indicate the productivity of resources on the farm.
They show the amount of total farm income that can be increased by the
utilization of the additional unit of the resource. Thus, they give
information about the possible gains or losses in farm income that are
possible through the acquisition of the scarce resource. The MVPs of
slack resources are zero and positive for the limiting factors or con—
straint resources. The more limiting the resource, the higher its MVP.
In order to be meaningful, therefore, the MVP of the resource should be
compared to its cost of acquisition of its marginal-factor cost (MFC).
It is profitable to acquire an additional unit of a resource if its
MVP is greater than its acquisition cost. Maximum farm income can be
obtained only when all MVPs of all resources are equal to their MFCs.

The MVPs of resources used in the production of cotton and maize
in Models A1 to A4 are given in Table 6.3. For each model, there are
two columns of MVPs. One column contains MVPs of resources used in the
production of cotton and maize when the price for maize was Tsh 1.00 and
the other when the price for maize was Tshs 2.50.

At the maize price of Tshs 1.00 or Tshs 2.50, land was in excess

supply in Mbdels A to A as shown by its zero MVP. However, farmers

1 4’
could not increase their incomes by using more land because they were
constrained by other resources, particularly labor. The use of ox—
drawn equipment or tractors did not bring enough land into use to make
it a constraint on production.

Irrespective of which of the two maize prices was used, not one of
Models A1 to A showed family labor constraints in August, September,

4

March, and April. However, family labor in November and January was a

 

143

limiting factor in all models. In November, much labor was required
to cultivate the land and/or to make ridges. The factor was also
limiting in January, when weeding for cotton and maize required a
great deal of labor to fight weeds that grow rapidly as a result of
heavy rains.

The effect of partial mechanization on resource constraints was

clearly shown when MVPs of family labor for Models A_ and A2 were

1
compared, irrespective of which maize price was used. First, the use

of ox—drawn equipment for land cultivation in Models A and A3 reduced

2
labor demand in October, thus removing the conStraint that was observed
in Mbdel Al. Second, ox technology, by expanding the area under culti-
vation, increased the demand for labor during ridging, weeding, and
harvesting. The analyses showed that the MVPs of family labor in
November, December, January, February and May were higher in Model A2
than in A1. At the maize price of Tshs 2.50, for example, the MVPs of

family labor in Model A were Tshs 2.75, 1.08, 0.93, 1.12, and 0.95

1
respectively. The corresponding MVPs in Model A2 were Tshs 2.87, 1.60,
1.28, 2.03, and 1.16. Apparently, ox plow cultivation aggravated the
seasonal labor bottleneck.

Although no allowance was made in the model for selling farming
labor, the wage rates for hired labor can be taken to reflect the
family-labor opportunity cost. The wage rates in October, November,
December, January, February, and May were Tshs 3.10, 3.25, 3.87, 4.13,
4.98, and Tshs 1.85. Thus, in any model, farmers could not increase

their income by hiring more labor because MVPs were less than the MFCs

for the months mentioned above.

144

The farmers could increase their income either by working extra
hours or by organizing working parties. The latter alternative might
not be possible. Labor from working parties was a constraint in
October and November for Models A1 and A.4 and in November for A2 and
A3 (see Table 6.3). Since these months were the peak labor periods for
farmers (the main source of this type of labor), working parties became
very limited.

The MVP of an owned ox team was zero. No more of this resource
could be used until more labor became available for weeding and har—
vesting, or until farmers were taught how to use ox—drawn equipment
for farming operations other than land cultivation.

In October and November, operating captial was a constraint in all
models when the price for maize was Tshs 2.50. The resource was more
constraining in October for Models A2 and A3. With the expansion of
land under cultivation, more cash expenses were made to pay for hired
oxen technology (in the case of Model A3), operating costs for the in-
creased use of owned ox technology (in Model A2), and expenditures on
increased use of fertilizers and insecticides.

The MVPs of fertilizers and insecticides were equal to their prices.
The MVP of TSP = MFC of TSP = Tshs 1.34; the MVP of SA = MFC of SA = Tshs
1.03; the MVP of Thiodan - its MFC = Tshs 13.34; and the MVP of DDT = its
MFC - Tshs 5.06. It was not known, however, with existing fertilizer
and insecticide prices and under existing input-output relationships,
whether optimum use of these resources had been achieved. Data on pro—

duction responses to these inputs were not available.

145

Optimal Organizations of the Participants' Representative Farms with
Variable Product, Fertilizer, and Insecticide Prices, and Operating—
Capital Levels for Early Cotton and Maize

 

 

 

The extent to which the optimal allocation of existing farm resources
under the present state of technology and prices would increase the net
farm income, change existing cropping patterns, and improve resource use
was shown in the analyses of Models A1 to A4' It is necessary to explore
the impact on Models A1 to A4 of variable—product prices, fertilizer and
insecticide prices, and operating—capital level on: (1) net farm income;
(2) cropping patterns; and (3) resource use. In Tables 6.4 to 6.15 and
in the accompanying discussion, these changes are presented as Alternatives
1, II and 111. Alternative I represents increases in the relative prices
of cotton and maize; Alternative 11 represents the increase in the prices

of fertilizers and insecticides; Alternative 111 represents the increase

in the level of operating capital.

Alternative 1:

 

Changes in the Relative-Prices of Early Cotton and Maize

 

As stated earlier in this chapter, the relative prices of cotton
and maize were increased by zero percent and 80 percent; 25 percent and
80 percent; and 25 percent and 150 percent for cotton and maize respec—
tively. These increases were in keeping with government price increases
that ranged from zero percent between 1975-76 and 1976-77 to 25 percent
between 1979—80 and 1980-81 for cotton and with market—price increases
for maize (within MWanza Region) that ranged from 80 percent to 150

percent above the government price between May and October 1980.

146

Although relative product prices were changed, input prices and level
of operating capital remained unchanged.

Since there are three product—price changes, Alternative I was
divided into three subalternatives. In Alternative Ia, produce prices
of Tshs 3.20 for cotton and 1.80 for maize were used. In Alternative
Ib, the prices of Tsh 4.00 for cotton and 1.80 for maize were used.

In Alternative Ic, product prices of Tshs 4.00 for cotton and 2.50 for
maize were used.

The changes in net farm income, cropping patterns, and resource
use for those product-price changes are given in Tables 6.4 to 6.6.

As shown in Table 6.4, Alternative Ia led to a remarkable fall
in hectarage planted in maize and a remarkable increase in cotton hec—
tarage in Models A1 to A4' In Model A1 the area planted in cotton
increased from 1.65 to 2.01 hectares, whereas maize hectarage decreased
from 1.30 in the base plan to 0.48 in the alternative plan. These re—
present an increase of 21.82 percent and a decrease of 63.07 percent
in the areas planted in cotton and maize respectively. In Model A2, the
cotton area increased by 28.48 percent, Whereas the maize area decreased
by 55.47 percent. There was also an increase of 23.03 percent and a de—

crease of 56.06 percent in the areas of cotton and maize, respectively,

for Model A3. As in the base plan, the changes in Model A. were similar

4
to those in A4' The total land planted in the crops, however, declined
in Models A1 to A4'

As a result of the changes in land use, the use of other resources
also changed. In Model Al, the decrease in cropped area led to a de-

crease in the use of labor; family labor declined by 10.07 percent, and

147

.comaquEoo momofiafioaw

cu mwflnmo ocmscwmnsm vcm manu CH nonsaocfi who woman ammo «chm

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

«.0 wHAMH

A _ _
ma. NH+Wmo. o mm.m ~a.om-.mq.m cm.q mm.aH+ mg. m _ No.q “mo.NH+ ao.o _ om.m w memew .amo wcHumumao\an
mm. m -.om. H om.H om.HH- am.H Hn.H $0.0 - an. H . mm.H mm.m - om.H 0m.H 1 mgms_ roam; .aacs rm\Haz
am.m -moo.momH aq.aama. am.a - No.qomH mm.oamH mm.m - mo. ammH_ H~.ammH _om.m . oa.aowH_ as.aamHm mems_ m:\Haz
so.ams mm.mma qa.omoH ma.mm- ao.som cm.mOMH mo.o~- oq..Ha . mm.quH am.am-_~n.ama ~ sa.oNon mews. HmuHamo mcHumrmao

o o o - u - m - - - - Tm opal AumuHxv hangars
- - - ca.oH- «H.om mm.oq l a - - - sum go. HamuHxv anus xo
- - - - n - .mm.a “mm. .Hm _ mm.am - - - as: xo_ Hawazoo amma xo
on.oH- no.5aom «N.quq AA.NH- ma.MHmm oo.amaq ms. HH ”NT ocoqm mm.qmma 05.0H- no.5mom qm.~qqqm anew roams Hmoos
ma.Hq-_mH.omm Na.Nmo mo.mq- mq.Ham Hm.awo “so. .Hq-Mmm.mNa eo.mHa _ma.Hq- mH.omm Na.mmo W anew sauna wceroa
mm.oo- am.m mm.qu wn.oau am.m mm.aoH mm.oo-,am.m mm.q0H mm.oa- am.m mm.qu “ mum“ roams smer
so.oH- ma.mHmm ao.qmom on.q - mo.mHmm 0H.mmom Na.m . sm.mmom as.Honm so.oH- ma.mHmm_ mo.qmom mu:_ roams AHHamm
am.mH- oa.~ ma.~ ~H.NH- Ho.~ am.~ o.a . MN.~ No.m am.mH- ma.~ mm.~ mm_ sums Hayes
no.mo- mq.o om.H so.om- mm.o mm.H as.mm-_Ho.o am.H no.mo- mq.o om.H new HmNHmzo scma
“Nw.H~+ Ho.~ mo.H mo.m~+ mo.~ mo.H ma.m~+ NH N m mo.H mm.HN+ Ho.N mo.H «mascuuoo aHummv sang
mo.mH- mm.oqu os.mmmm «H.0Hu Hm.mo~q mm.QHon so.mH- mm. Hma _ oo.amam mo.mH- mm.oomq oq.mmmm mnmstmzv maoucH arms umz
- .H ..
s< How ya Hana: m< you m< Have: . £H0um m4 Hana: H< new H< Have: M
.mwcmzu swam new mwamzo swam now mmcmcu cmHm new wwcwsu swam new HHzD _ :mHH
. N 3oz swam mmmm N 3oz swam wmmm _ N 3oz .cmam mmmm N 3oz swam mmmm m
wH m>fiumChwuH<
”manmaHUHuumm mnu you mufim: new acuuoo xaumm new mmoaum oozvoum m>Humem a“ muwcmfiu Una mmcmHm mmmm mLu wow mousmmmz mocmfiofiwwm

148

hired labor and labor from working parties declined by 96.58 percent
and 41.75 percent respectively. Total labor use declined by 16.76 percent.

In Model A2, the use of family labor decreased by 3.42 percent, that
of hired labor by 96.58 percent, and that of labor from working parties
by 41.04 percent. In all, labor use declined by 11.43 percent. Total
labor use in Model A3 declined by 12.77 percent. Of this, family labor
decreased by 4.76 percent, hired labor by 96.58 percent, and working
parties by 43.05 percent.

The use of ox—drawn equipment in Models A2 and A also declined.

3

Although in Model A2 the use of owned ox—drawn equipment declined by
9.59 percent, that of hired ox-drawn equipment declined by 16.40 per-
cent. The decline in the use of farm resources that require cash ex—
penses, such as hired labor, working parties, ox technology, and so on,
led to a fall in the use of operating capital. It fell by 27.97 percent
in Model A1, by 26.09 in Model A2, and by 33.75 percent in Model A3.
The changes in net farm income were also remarkable. The esti—
mated net farm income in Model A1 decreased from Tshs 5538.46 in the
base plan to Tshs 4506.83 in the alternative plan; this represented a
fall of 18.63 percent. Consequently, the net farm income per hectare
declined by 3.59 percent. This small decline in return to land, de—
spite a big fall in net farm income, was also due to a decline in
cropped land. The return to unpaid labor (family labor) decreased
from Tshs 1.50 to Tshs 1.36, a 9.33 percent fall. Finally, net farm
income per a unit of operating capital increased by 12.98 percent.

In Model A2, the net farm income was estimated to be Tshs 4981.28,

compared to Tshs 5729.60 in the base plan. This represented a decline

149

of 13.06 percent. The return to land was Tshs 1824.65. This was a
decline of 3.82 percent from that in the base plan. Although net
farm income per unit of unpaid labor declined from Tsh 1.52 to 1.37,
a 9.06 percent fall, that per unit of operating capital increased by
17.53 percent.

The estimated net farm income for Model A3 decreased from Tshs

5614.93 in the base plan to Tshs 4708.51 in the alternative plan.
This represented a 16.14 percent fall. As a result, net farm income
per hectare declined by 4.57 percent, the return to unpaid labor de—
clined by 11.26 percent, and the return to a unit of operating capi—
tal increased by 26.72 percent.

In Alternative lb, the official product prices increased by 25
percent for cotton and 80 percent for maize. As in the previous
case, there were substantial changes in areas planted with each
crop, use of labor, and other resources, as well as in net farm in—
comes for Models A1, A2, A3, and A4' The results are presented in
Table 6.5.

In Model Al’ the cotton area increased by 25.45 percent, whereas
the area planted in maize decreased by 65.38 percent. The same changes
were observed in Model A4° Even higher increases in the area planted
in cotton were achieved in Models A2 and A3. In the former, the
cotton area increased by 33.94 percent and in the latter by 27.27
percent. Although the area planted in maize declined by 59.12 percent
in Model A2, that in Model A3 declined by 59.84 percent. Furthermore,
in each model, total land cropped was less than that in their respec—

tive base plans.

150

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

cm.mm+ o~.~ mm.m mq.nm+ 00.0 om.q an.nm+ om.o No.q qm.mn+ mm.n mm.n mnmh Hmufiamo wcaumuwao\Hmz
Hm.HH+ mo.H om.H n~.o + mo.H Hm.H wH.HH+ oe.a mm.H Hm.HH+ no.H om.H man uonmq was: we um\Hmz
mn.o~+ :a.nomm q<.nwma mo.oH+ me.HmNN mm.omma om.aa+ qo.~omm H~.~owa mn.oN+ ~H.~omm qe.mnma moms m:\sz
-.m~I mH.mmn qw.o~oa q~.HmI mm.omm oq.<oma qo.o~I mo.mmm mm.aq~a N~.mNI nH.mwn am.o~oa mcmH HmOfiamo mcfiumuoao
o o o I I I I I I I I I mum ore Anmufixv nouomph
I I I HH.MHI mo.oq Nw.oq I I I I I I mu: x0 Avmufimv Emmy x0
I I I II I I Ed I .33 no.8 I I I fix 5 Sande 98“. xo
ma.mHI nm.mmmm qm.mqqa mm.oaI oq.mmom 00.nmqe q<.m I oo.amaq mm.qmw¢ mm.mHI nm.mmmm q~.~qqq mum gonna Hmuob
mo.mMI T~.wmm mn.mmo um.OQI oq.moq Hm.nwo om.mMI ma.a¢¢ oo.mHn om.QMI ofl.wam Nn.Nmo mu: muumm wcaxuo3
wm.omImmm.m mm.qu mm.omI mm.m mm.moH mm.omI mm.m mm.qu mn.omt mm.m mm.<oa my: “coma woufim
mo.o I Wo.-qm oo.qwom mo.m I @q.awmmm oa.moom oq.H I oo.oo~m mm.Homm mm.o I Ho.mmqm ma.qmom mu: uonmq maaamm
mm.qHIWNm.N mm.~ mq.HH+ mo.~ U mm.m mm.w I mn.~ w No.m nm.an mm.N mm.~ mm ucmg Hooch
mm.moI Vq.o om.H qw.omI mm.o . mm.H NH.mmI om.o m em.H mm.mol mq.o 0m.H mm Amuamxv ccmA
mq.mm+_mo.~ mo.H m~.n~+ 0H.N mo.~ qm.mm+ HN.N fl mo.a mq.mm+ mo.m mo.H m: Accuooo szmmv ecQH
. .
mH.m + WH.mH~L oq.wmmm oa.m + Hq.HNom mm.qaom mm.m +mHm.moNoM 00.0mmm mH.m + NH.Mfimn oq.mmmm mLmH Aszv mEoocH some uwz
a T a ..w H
3< no“ q< Hove: m< pow _m< Hone: m mm< u0w mm< Honor _H< uOml fi¢ Hone:
owcm£u_ cmam Mow owcmLU swam new ”mwcmnol swam . u0m mwcmfiu cmam_ pew HHZb XMHH
N _ 3oz Wmam ommm N 302 :mam ommm _ N _ 3oz Wmam wmmm n 3oz Tmam mmmm
9H o>Homcuwods "mucmoaofiupmm new
mnemx new scoooo hapmm pow mwofipm uoanoum m>Humamm ca mmwcmco new woman mmmm mcu qu mousmmo: zucofiofimmm

n.o

manmh

151

The reduction in cropped areas resulted in less use of other

resources. In Models A1 and A. family labor, hired labor, and labor

4:

from working parties decreased by 6.98 percent, 96.58 percent, and

38.99 percent respectively; in Model A2 these changes in labor use

decreased by 1.46 percent, 96.58 percent, and 38.56 percent; in Model
A3 family labor decreased by 3.08 percent, hired labor by 96.58 percent
and working parties by 40.57 percent. The use of ox—drawn equipment
declined by 8.27 percent in Model A2 and by 13.11 percent in Model A3.
Since the use of farm resources that require operating capital to

acquire them declined, the use of operating capital fell by 23.72 per~

cent in Models A and A., by 20.64 percent in A

l , and by 31.24 percent

2
in A3.

Unlike Alternative la, in which net farm incomes in Models A1 to
A.4 declined remarkably from those in their respective base plans, net
farm incomes in Alternative lb increased. The estimated net farm in—

come in Models A1 and A increased from Tshs 5538.46 to Tshs 5713.12,

4
or an increase of 3.15 percent. As a result, the retUrn to land in—
creased by 20.75 percent, the return to unpaid labor increased by

11.31 percent, and the return to a unit of operating capital increased
by 35.34 percent. In Model A2, the net farm income was estimated to be
Tshs 6267.51 compared to Tshs 5729.60 in the base plan; this was an in—
crease 9.38 percent. As a result of this increase and a decrease in the
use of farm resources, net farm income per hectare increased by 19.26
percent, return to unpaid labor increased by 11.18 percent, and return

to a unit of operating capital by 37.71 percent. The estimated net

farm income for Model A3 came to Tshs 5921.41 compared to Tshs 5614.93

152

in the base plan. This represented an increase of 5.46 percent. In

turn, the return per hectare of land, the return to unpaid labor, and
the return per unit of operating capital increased by 19.09 percent,

9.27 percent, and 53.48 percent.

The results for Alternative Ic, as shown in Table 6.6, were also
substantial. In Models A1 and A4’ the cotton area increased by 2.42
percent whereas that planted in maize fell by 4.62 percent. In Model
A2, the area planted in cotton increased by 3.64 and that planted in
maize decreased by 6.57 percent. Adthough the cotton area in Model A3
increase by only 1.82 percent, the area planted in maize declined by
4.54 percent. The results also showed that total cultivated land de—
clined in each model.

The above changes in land allocation resulted also in notable
changes in the use of other resources. In Model A1, family labor and
working parties increased by 0.56 percent and 7.25 percent respectively,
whereas hired labor declined by 10.87 percent. Total labor used,
however, increased by 1.26 percent. The use of family labor and work—
ing parties in Model A2 increased by 1.64 percent and 0.73 percent
respectively. There was no change in the employment of hired labor.
Total labor use increased by 1.46 percent. Despite a decline in
total cultivated land, labor use in these two models increased. The
reason lay in the increase in the cotton area. Cotton demands more
labor per hectare than does maize. Thus, the declines in the maize
area of 4.62 percent for Model A1 and 6.57 percent for Model A2 did
not release enough labor to meet the demands of an-increased cotton
area. In Model A3,

maize area was more than sufficient to meet the demand for labor by

however, the decline of 0.06 hectares in the

 

153

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

00.0 +.Ha.n am.m No.H~+ m~.m om.q wH.HH+ aH.m No.4 oo.m + am.m om.n mews HmuHaao mcHomrmao\Huz
oo.oH+.va.H on.H om.oH+ ea.H Hm.H N®.MH+ ma.H mm.H oo.oH+ as.H om.H mama henna sac: co r:\Haz
Ha.aH+.vn.qo~u qe.aamH Hm.oH+ Ho.~o~m mm.ommH om.eH+ aa.aowm Hm.aoma Hq.AH+ om.qo~m qq.aamH mews mx\Hux
mm.o +.mo.NaoH «n.0NoH no.m + -.mmNH oq.qomH om.m + OH.mm~d mm.quH Nn.o + mo.~mod sa.o~oH meme HmUHamo acHumroac
o o o I I I I I I I I I my: one Avopfimv neuumph
I I I o mm.oq mm.oq I I I I I I mum x0 Avmuwzv snow 5
I I I I I I no.0 I oaém wo.ow I OI I mum KO :5ch Bomb KO
0N.H ..mm.maQL q~.~eqq mo.H I mm.amqa oo.amaq ca.H + mm.qmma mm.smmq 0N.H + mm.maqa 4N.que mum roams Hmuos
m~.a.+rm.moo Na.~mo am.~ I mm.0ao Hm.amo ma.o + oo.wHa so.mHa mN.a + mm.mmo Na.~mo my: spams aspires
.
am.oHI H.ma mm.aoH mo.NHI om.Hm mm.qu o mm.qu mn.qu am.oHI 0H.ma mm.aoH a»: roam; nmrH:
om.o.+qm.moaa aa.smom maq.o I so.ansm 0H.maom so.H +lom.mmmm aa.Hoam om.o + Hm.moam aa.smsm up: ronma aHHami
mo.o_Ina.~ ma.~ Ho.H Imqo.~ mo.~ mo.o +.am.~ No.m mo.o I ma.~ mm.~ mm mama Hades
. . _
u "
No.s.Ir~.H om.H lam.q Imo~.H NM.H am.o Ilm~.H " am.H No.q I s~.H om.H m: HmNmee acme
_ . N _ m . n l
Na.~ Iamo.H _ mo.H “mm.H +wmo.H mo.H so.m +_HA.H m mo.H ” NQ.N + mo.H mo.H _ mm M Hecaooo aHrmmv acmu
. . . l .
_ l _ . . l _ .
No.od+ma.wmqs es.mmmm Mam.MH+_ao.maqs ma.qHsm .HN.mH+_oH.Hooa oo.omam No.oH+ wa.mmas os.wmmm% mews ”Hchv 0985 gram on
_ M I“ .r l _ ”I _ l I
r H q H H H T l N 1 l n I.
U.< rod“ e Hose: “ Mm< soc me Hanoz y _.< not” < Home: H< row H< Hmeoxl "
owcmcuu swam “ no“ ”mmcmnufl swam hem _mwcmnod cmflm l new mwcmnu swam new _ HHZB M XMFH
n W 302 ”swam mmwm _ u _ 3oz ”cmfla owmm l N U 3oz "swam vmmm N 302 spam mmmml

 

 

l

UH w>HomcpmuH< “mocmofiofiuumm new
mNHmz ncm couuoo mfipmm uOM moofipm ooznoum o>aomHmm cw mmwcmno cam mcmfim mwmm mnu pew mosammmz zocofiofiwmm

0.0 manmh

154

cotton for which the area increased by 0.03 hectares. Thus, family
labor declined by 0.49 percent, hired labor by 12.03 percent, and
labor from working parties by 2.39 percent. Labor use declined by 1.05

percent. The use of owned ox-drawn equipment in Model A declined by

2

0.97 percent, whereas that of hired ox—drawn equipment in Model A3 did
not change. Operating capital increased in all models; these increases
were 5.62 percent, 3.50 percent, 5.07 percent and 5.62 percent for
Models A1, A2, A3, and A4.

Increases in net farm income were much higher ianlternative Ic
than in Ib; better product prices and more area cropped led to the
differences in net farm income. In Models A1 and A4’ net farm income
came to Tshs 6458.78 compared to Tshs 5538.46 in the base plan. This
was a 16.62 percent increase. As a result of this change and a de—
cline in the area cultivated, net farm income per hectare increased
from Tshs 1877.44 in the base plan to Tshs 2204.36 in the alternative
plan. This represented an increase of 17.41 percent. Return to unpaid
labor increased from Tsh 1.50 to 1.74 or a 16 percent increase. Despite
an increase in operating capital, the return per unit of operating
capital increased from Tshs 5.39 to Tshs 5.97, an increase of 9.66
percent.

The estimated net farm income in Model A2 was Tshs 6601.16. This
was an increase of 15.21 percent from Tshs 5729.60 in the base plan.
Net farm income increased from Tshs 1897.21 in the base plan to Tshs
2207.74 in the alternative plan. This represented an increase of 16.36
‘percent. Whereas return to unpaid labor increased by 13.82 percent,

from Tsh 1.52 to Tsh 1.73, the return to a unit of operating capital

increased from Tshs 4.62 to Tshs 5.14, an 11.18 percent increase.

155

In Model A3, net farm income increased from Tshs 5614.93 in the
base plan to Tshs 6475.69 in the new plan. As a result of this
change and changes in the use of farm resources, there were remark-
able changes in return to land, unpaid labor, and operating capital.

The return to 3 hectare of land increased from Tshs 1890.55 to 2202.61,
an increaserof 16.51 percent. Although the return to unpaid labor in—
creased by 16.56 percent, from Tshslaffl.to Tshs 1.76, the return to a
unit of operating capital increased from Tshs 4.30 to Tshs 5.23, a 21.62
percent increase.

These results provided important insights into farmers' responses
to changes in product prices. It was quite clear from the results in
Alternative Ia that maize was less competitive than cotton. Although
the area planted in maize declined remarkably, that planted in cotton
increased substantially. Further, net farm income was lower than that
in the base plan, and so were the returns to land, to unpaid labor, and
to a unit of operating capital. The employment of owned ox—drawn equip—
ment (Model A2) declined by a smaller percentage than that of hired ox-
drawn equipment (Model A3).

In Alternative Ib, maize became even less competitive in all models.
The area planted in maize was smaller than that in Alternative la. The
amount produced was just enough to meet household consumption needs. On
the other hand, the cotton area increased by a larger percentage. And,
as in Alternative Ia, the decline in the employment of hired ox—drawn

equipment was larger than that of owned ox—drawn equipment.

156

Alternative Ic showed better results. Maize was more competitive
than it was in the other two alternatives. Food produced in this alter—
native was more than double that produced in the other two alternatives.
In addition, net farm income in each model was higher than the respec-
tive incomes in the base plans.

Apart from the insights related to each alternative, there were
general observations to be drawn from the results. First, the current
official price for maize was too low to make maize competitive with
cotton. Even an increase of 80 percent (from Tsh 1.00 to Tsh 1.80),
although keeping the price for cotton at Tshs 3.20, did not lead to a
substantial change in the area planted in maize or, consequently, to
maize output. The black market of Tshs 2.50 seemed to be more influ-
ential in farmers' decision-making processes. Second, the results
showed that farmers were sensitive to relative produce prices. Given
the importance of both crops to the national economy, it was necessary
to carefully consider the impact of changes in product prices on
farmers' resource—allocation decisions and their eventual impact on
output. Third, the use of ox-drawn equipment seemed to be profitable,
despite its limited use. The results showed that in each alternative,
Model A2 produced the highest net farm income followed by Model A3, and
A1 (or A4). Tractor hiring (Model A4) was not profitable in any alter—
native. Finally, land cultivated was less than the amount available.
The main reason was lack of enough resources—-especially labor to farm
additional land. The availability of ox—drawn equipment in Models A2
and A3 did not utilize all land because the technology was limited to
land cultivation; this left ridging, weeding, and harvesting operations

to be done by labor.

157

Marginal-Value Products (MVPS in Tshs) of Resources
Under Alternative I

 

The marginal—value products of resources for the three alternative
relative product prices are presented in Tables 6.7 to 6.9. Under
Alternative Ia (see Table 6.7), the MVP of land was zero in each model.
This implied that land was in excess supply. The same conclusion was
reached with respect to family labor in August and September. As a
result of a decline in land cultivated in each model, monthly labor
requirements for the two crops declined. Thus, in Models Al and A4’
MVPs for family labor in October to February were lower than those in
the base plans. The MVP of family labor in May dropped from 0.95 Tshs
to zero; the reason for this was the substantial decline in the area
planted in maize, the crop that needs more labor in May since this is
the harvesting period. The MVP of family labor in June was Tshs 2.13
compared to zero in the base plan. This was a result of the increase
in the area planted in cotton, which demands more labor for harvesting
during June.

Other unstable changes in Models A1 and A.4 were the MVPs of working
parties and operating capital. In October and Nbvember, the MVP of
working-party labor declined from those in the base plan; the changes
were from MVPs of Tshs 1.37 to 0.98 in October, and from Tshs 1.66 to
1.03 in November. The MVPs of operating capital became zero in the new
plan. Again, these changes resulted from the decline in land cultivated,

whiCh in turn led to a decline in the use of labor, fertilizer, and

insecticides.

158

 

 

 

 

 

 

 

mo.H oo.H mm.H mm.H HH.H ma.H mo.H mo.H as .>02 exp m3
mm.o sm.H o o o o mm.o sm.H a: .ouo Baa as
o me.o 0 No.0 0 om.o o ma.o mama .>oz oo
o ma.o o mm.o o Hw.o o mq.o mama .ooo 00
amqu oo.m am.MH oo.m em.MH so.m am.mH oo.m meme Ham
am.mH em.mH am.mH am.MH qm.mH am.mH am.mH em.mH mews zams
mo.H mo.H mo.H mo.H mo.H mo.H mo.H mo.H me am
am.H am.H em.H em.H am.H em.H am.H em.H we ems
o ma.o om.o oa.H Hs.o eH.H o mm.o am am: as
om.o NH.H oo.H HN.N Hm.H mo.N om.o NH.H Hm .nmm as
Ha.o mm.o as.o oH.H mm.o mN.H He.o no.0 um .cmm as
mo.o (mo.H No.H oa.H «N.H oo.H mo.o mo.H om .omm as
wH.H ma.~ ae.H mm.m om.N aw.m mH.H ma.~ Hm .>oz as
ms.H mm.N o o o o ms.H mm.N um .uuo an

Egg 2.3m 72.: 233 23m 23m 23m 23m

3mz mm: zmz mm<m 25H mam zmz mmfiH
9H2: amazemmm

ea ammo: m< page: N< ammo: H< ammo:

 

 

 

m
H m>flumcuoua< "mucmmflofiusmm How

q

 

 

 

 

a.o mHnme

< co H< mHopoz How moouaowwm mo AmsmH cw mm>zv muoswoum maam>lawswmumz

 

159

The changes in MVPs of resources for Models A2 and A3 were not

very different from those for Models A1 and A4° The MVPs of family
labor in November to February and May were lower than their corres—
ponding MVPs in the base plans. However, the MVPs of family labor in
June were Tshs 2.98 and Tshs 2.35 compared to the MVPs of zero under
the base plans in Models A2 and A.3 respectively. In addition, operating
capital, which, under the base plans, was a constraint in October and
November, was in excess in the new plans.

In summary, the reduction in demand for labor and other inputs as
a result of the decline in the area planted in maize more than offset
the increased demand for these resources as a result of an increased
cotton area. This explained the marginal decline in MVPs of most farm
resources. The only exception was the MVP of family labor in June,
which increased substantially because of the increased demand for labor
to harvest the expanded cotton area.

The MVPs of resources for Alternative Ib are presented in Table
6.8. As with Alternative la, the MVP of land was zero in each model,
as were the MVPs of family labor in August and September. The changes
in the MVPs of other resources followed the same pattern as those in
Alternative Ia; MVPs of family and working-party labor were slightly
lower than their corresponding MVPs in the base plans for each model.
There were two exceptions to this general conclusion. One was that the
'MVP of family labor in May was zero in each model because there was

much less land cultivated for maize, causing much less demand for

labor during harvesting. The other exception was that the MVPs of

160

 

 

 

 

 

 

 

 

 

 

 

 

 

 

mw.o o 4N.H o am.~ o mw.o c He muss sea as
am.H so.H mo.o Nm.H am.H as.H am.H as.H Hm .>oz sea as
0H.H am.H o o o o 0H.H sm.H Hm .uoo sea as
o as.o 0 No.0 0 ON.o o ma.o mama .>oz oo
o mq.o o wa.o o Hw.o o ma.o mews .uuo so
so.m oo.m oo.m oo.m oo.m oo.m oo.m oo.m mean sea
em.mH am.MH am.MH am.mH am.MH am.MH am.MH am.mH mama zany
mo.H mo.H mo.H mo.H. mo.H mo.H mo.H mo.H we am
em.H am.H em.H em.H em.H em.H am.H em.H we awe
Ha.a o ww.a o NN.w o Ha.a 0 pm mass as
o ma.o o ON.H o 0H.H o ma.o on am: as
mo.H NH.H aN.H HN.N mm.H mo.N mo.H NH.H rm .ama as
No.0 ma.o ma.o mH.H ww.o mN.H as.o ma.o um .sme as
oa.o mo.H ma.o os.H ao.H os.H oa.o wo.H rm .Uwo as
mm.H ma.~ NN.H am.N «H.N am.N mm.H ma.m rm .>oz as
aw.H aa.~ o o o o ew.H aa.m rm .uuo as
Z<AQ Z<Am Z<Am Z<Am z¢gm Z<Hm Z<Am Z<Am
Zmz Mm<m Bmz mm<m zmz mmmm Bmz Mm¢m
HHZD momDOmmm
q< AMQOZ m< AmQOZ N¢ QMQOZ H< Ammo:
a a

H m>wumfiuoua<

"mudmaflowuumm sow

w.© maan

¢ ou H< Howoz How moosdomom mo AmSmH aw mm>2V muoswoum ode>IHmGHmumz

161

family labor in June increased tremendously in the model. From
zero MVPs in the base plans, the new MVPs were Tshs 7.71 for Models

A1 and A4’ Tshs 8.22 for A2, and Tshs 7.88 for A The reason was

3.
that the area planted in cotton had increased to even more than it

was in Ia resulting in more demand for labor during harvesting. The
changes in the MVPs of family labor in June were accompanied by changes
in the MVPs of working-party labor, for which the value in the base
plans was zero. The new MVPs of this resource were Tshs 0.85 for
Models A1 and A4,

The MVPs of resources in Alternative Ic, as shown in Table 6.9,

Tshs 2.39 for Model A2, and Tshs 1.24 for Model A3.

were generally lower but closer to their respective MVPs in the base
plans than were those in Alternative Ia or 1b. The reason was that
the areas planted in maize and cotton were too different from those
in the base plans. The most notable changes were in the MVPs of
November and December for family labor in Model A2 and the MVPs of
October and November for operating capital in all models. In Model
A2, the MVPs of family labor in October (Tshs 2.90) and November
(Tsh 1.68) were slightly higher than their respective MVPs in the
base plans. Because of the expanded cotton area, more labor for
ridging and planting was required than the labor released by the re-
duction in the maize area. In all models, the MVPs of operating
capital in October and November were higher than their corresponding
MVPs in the base plans. The increases in MVP may have resulted from

an increase in demand for resources such as fertilizer and insecticides

by the expanded area planted in cotton. This demand could not be met

162

 

 

 

 

 

 

 

 

 

 

o
H o>HumaHouH<

 

 

 

21H 84 SH SH SH EH 24 SH a: :62 ES a:
84 SH 0 o o o .84 NTH .5 .80 .53 .Hz
mm.o oq.o NH.o no.0 om.o om.o mm.o aq.o moms .>oz 00
No.0 oq.o sw.o wo.o Nw.o Hw.o No.0 me.o mama .ooo oo
oo.m oo.m oo.m oo.m oo.m oo.m oo.m oo.m mama Boo
qm.ma qm.mH em.MH em.ma qm.mH qm.mH em.mH em.mH moms zome
mo.H mo.H mo.H mo.H mo.H mo.H mo.H mo.H we <m
em.H em.H em.H em.H em.H em.H em.H em.H we owe
ow.o mm.o No.a on.H «n.o oH.H ow.o no.0 pm mmz 4m
qo.H NH.H ma.m HN.N oo.N mo.m qo.a NH.H um .omm 4m
ow.o mm.o oo.H mH.H 0H.H mm.a om.o mm.o um .cmm Am
Ho.H mo.H mm.H oq.H wo.H oo.H Ho.H wo.H Hm .oma gm
qo.m mm.~ Hm.N mm.m om.m mw.m ¢©.N mm.m um .>oz Am
..Nm.m mm.~ o o o o No.m oa.m pm .uuo as

Z<Am z<qm Z<Am z<Am z<gm z<qm z<qm z<dm

Bmz mm<m Bmz mm<m zmz mm<m Bmz mm<m
HHZD mUMDOmmm

q< Ammo: m< Ammo: N< ammo: H< Ammo:

 

Housmafiowuumm Mom g< ou H< maowoz wow moousowmm mo AmfimH ma mm>2v muoooosm oDHm>Ichwwumz

m.© oHQmH

163

by the release of operating capital from the reduced maize hectarage,
since cotton required more of these inputs per hectare than maize did.
The reduction in working-party labor, as indicated by the fall of MVPs
for this resource, released operating capital that was used to help
meet the demand for other resources.

In summary, as shown by the results, the main limiting factor for
increasing production was labor, especially during the peak seasons
discussed in Chapter II. It is possible that if farmers were paid
higher prices than those discussed, they would work extra hours and
thus release family labor constraints in the optimal plan. It is most
unlikely, however, that the government would increase prices of cotton
and maize beyond the ranges discussed in this study. ~Price increases
in the past ten years or so have not shown any indication of this
happening. Nor was there any indication that the black market price
for maize would increase by a big margin from the current price of
Tshs 2.50 per kilogram; if this happened due to forces of supply and
demand, consumers would shift their demand to other food items such
as rice, sorghum, and cassava and eventually push down the price of

maize.

Alternative 11:

 

Increases in the-Prices of Fertilizers and Insecticides

 

The existing fertilizer and insecticide prices were subsidized by
‘Ehe World Bank and/or by the government. Eventually, however, the
farmers would bear the full costs of these inputs. In the following

discussion, unsubsidized prices are used in order to show their impact

164

on net farm income, cropping patterns, and resource use. In Tables
6.10 and 6.11, the results of the analysis are shown.

Increased fertilizer and insecticide prices reduced the area of
maize and cotton in all four models. In Models A1 and A4’ the cotton
area decreased from 1.65 hectares to 1.53 hectares, a 7.27 percent de—
cline, whereas the area planted in maize decreased from 1.30 hectares
to 1.21 hectares, a decline of 6.92 percent. As a result of these
changes, net farm income came to Tshs 5121.97 compared to Tshs 5538.46
in the base plan.

Reduction in the total area cultivated led to a decline in the
employment of labor. Family labor decreased by 4.79 percent, hired
labor by 96.58 percent, and labor from working parties by 31.09
percent. In all, use of labor decreased from 4442.24 man hours in the
base plans to 3961.43 man hours in the new plans. This represented a
10.82 percent decrease. In addition, the use of operating capital
decreased from Tshs 1026.74 to Tshs 993.56, a reduction of 3.23 percent.
This resulted from the reduced use of fertilizers, insecticides, hired
labor, and labor from working parties, all of which fell because of a
decrease in cotton and maize areas.

The decline in net farm income and in levels of the use of resources
led to the fall in returns to land, unpaid labor, and a unit of operating
capital. The return per hectare of land was Tshs 1869.32 computed to
Tshs 1877.44 in the base plan, a decrease of 0.43 percent» And
although the net average return per man hour of unpaid labor decreased
from Tsh 1.50 to Tsh 1.46, a decline of 2.67 percent, the net return to a
unit of operating capital decreased from Tshs 5.39 to Tshs 5.16, repre-

senting a fall of 4.26 percent.

165

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

. _ _ r _
o~.e I H.m am.n wa.o~+.oH.m " om.q oa.HH+_AH.m No.q o~.q I SH.“ am.m mans Hmuaamo wcHomomao\Haz
so.~.I a.H om.H am.m I oq.H Hm.H 0N.n I mm.H am.H no.~ I oe.H om.H mama _ roam; sacs so pm\Haz
. _ .
mq.o.Iwm.aowa «c.5amH ~H.H I am.aoma mm.oamH ma.H ImHN.ssma Hm.aamH mq.o I ~m.aomfi as.aamH menu 1 mm\Hux
. . p . .
m~.m ..wm.mam an.o~oH om.m~I on.mmo oq.qoMH sH.NNI mq.ooa mm.Ha~H m~.m I on.mom sn.a~oH meme * HmuHamo acaamrmo:
o o o I I I I I I I I I _mum uuH Aooufimv uOuomuw
I I I ooHI o mm.oq I I _ I I I I _ mum xo Aamerv anus xc
_
I I I I I I om.~HImom.mn wo.om I I I mum Ko Humcsov some x0
mm.OHI_wa.Hoan QN.~sqs Ha.HHI mq.Hoam oo.am¢q om.mHI Nm.mamm mm.qmms .Nm.oHI mq.Hoan 4N.quq mom roam; Hmuoa
so.Hm..sa.asq ma.~mo sm.qMI an.aaq Hm.amo MA.N4I 0N.HHQ so.mHa ao.HmI «n.0qq ~a.~no mum Aroma waaxrou
_

mm.omI~m.m mm.qu mm.omI am.m mm.qu mm.oaI.am.m mm.soH mm.omI am.m mm.soH mum roams emer

oa.a.IWH.moma aa.qwom oo.m I NH.momm SH.maom so.m I ao.mmq ma.Hsam as.q I NH.womm am.smomm mum roama MHHaml
l _ l . n
. + _ .

HH.a..¢a.~ _ mo.~ rah.“ I as.~ no.N o~.HHI_ms.N w No.m HH.~ I as.~ na.~ 1 mm _ yams HmooH
“ l h l ,

Na.o_IH~.H _ om.H mam.m I.H~.H _ Nm.H an.QHIlaH.H w AH.H ~m.o I H~.H w om.H ‘ mm HmNmee mama
._ . . . . l
_ . . . . l l l -

R; ImTH _ SH m2; Immmé _ $.H Tim IMHTH H 3H 2.5 ImeH 8H W 3: 2538 .353 23.
_ _ .. l _ _ ._ .
. . . _ . _ .

mm.“ ..mm.HNHn sq.mmmm “ma.m Ilsa.HNHm mo.sHom _aa.NHI_mq.eaoe oo.a~am mm.“ Ilsa.HNHn oa.mmmm_ mews ”AHazv maoucH Erma .Ix
l u “I l l. a a l H . _
m H J n *1 H .fi .44 _ J u M ..I
Hq< row” .< Hana: _ nm< coc_ m< Have: 1 uc< Howl Aa. Hmnox HH< pow H< Hmeozw w

oucmzon cmam q sew Momcmcu_ cmam " new ”omcmno~ cmHm m p0w owamno. swam Ho“ m HHZD ” xmsH

u m 302 .cmHm ommm u u * 3oz ”swam ommm w N _ 3oz “cmam ommm N ~ 3oz ”swam mmmm_ _

_

 

mucmafiofiuumm new m~amz vcm aeouoo xfiumm new
moofium opaofiuommcH new uoNfiHHuumm ca mommopocH can madam mmmm mcu u0w mousmmmz wocmwofiwmm

OH.o wanes

166

The results in Model A2 were the worst. Although cotton area de—
creased from 1.64 to 1.51 hectares, the area planted in maize decreased
from 1.37 to 1.17 hectares. Consequently, net farm income came to Tshs
4996.43, compared to Tshs 5729.60; this represented a 12.79 percent
reduction. Employment of family labor, hired labor, and labor from
working parties declined by 8.07 percent, 96.58 percent, and 42.73
percent respectively. The use of owned ox-drawn equipment decreased
from 89.98 to 78.86 ox hours, a 12.36 percent fall. The reduction in
the employment of hired labor, labor from working parties, as well as
the decrease in the amount of fertilizers and insecticides used de—
creased operating capital from Tshs 1241.35 in the base plan to Tshs
966.46 in the new plan; this was a 22.14 percent decline.

With the exception of the return to a unit of operating capital,
which increased from Tshs 4.62 to 5.17, an 11.90 percent increase, the
changes in the return to land and to unpaid family labor followed the
same pattern as they did in Models A1 and A4. The net farm income per
hectare declined from Tshs 1897.21 in the base plan to Tshs 1864.34 in
the new plan. This represented a decline of 1.73 percent. The return
to unpaid labor came to Tshs 1.44 compared to Tshs 1.52 in the base plan,
a decrease of 5.26 percent.

In Model A3, the increase in fertilizer and insecticide prices
resulted in more or less the same changes as in Model A1. The area
planted in cotton was 1.53 hectares, a decrease of 7.27 percent from
1.65 hectares in the base plan. The area planted in maize decreased
from 1.32 hectares in the base plan to 1.21 hectares in the new plan,

a decrease of 8.33 percent. The reduction in cultivated land led to

167

a reduction in the use of other resources. Family labor declined
from 3695.16 to 3508.12 man hours, a 5.06 percent decrease. The em—
ployment of hired labor and labor from working parties declined by
96.58 percent and 42.73 percent respectively. In addition, no ox-
drawn equipment was hired compared to 46.82 ox hours hired in the
base plan. The reduction in land cultivated meant that the demand
for labor decreased and that, given no changes in the supply of labor,
there was enough labor for all farm operations. Hiring of ox-drawn
equipment become unprofitable. Operating capital fell substantially;
whereas it was Tshs 1304.40 in the base plan, in the new plan it came
to Tshs 993.56, a decline of 23.80 percent.

As a result of these changes, net farm income decreased from
Tshs 5614.93 in the base plan to Tshs 5121.97 in the new plan.
Further, net farm income per hectare decreased from Tshs 1890.55 to
Tshs 1869.32, a reduction of 1.12 percent; the return to unpaid labor
decreased from Tshs 1.51 to Tshs 1.46, a 3.97 percent decrease. The
only increase was realized in the return to a unit of operating capi-

tal; it increased from Tshs 4.30 to 5.16, an increase of 26.98 percent.

Marginal-Value Products (MVPs in Tshs) of Resources Under Alternative 11

The MVPs of resources under Alternative 11 are shown in Table 6.11.
The MVPs of most resources in all four models were either equal to or
lower than their corresponding MVPs in the base plans. This was especi—
ally true with land, family labor, hired labor, labor from working

Parties, ox—drawn equipment, and operating capital from December to June.

168

 

 

 

 

 

 

 

3H 8H o SH 0 SH SH SH E .52 as .Hz
36 SH 0 o o 0 $6 HTH um .30 H5 H3
mm.H oe.o OH.H No.0 NH.H om.o mm.H mq.o mzma .>02 00
34 3.0 $4 mod 3H 3.0 84 $6 moms .30 oo
om.mm oo.m om.m~ oo.m om.m~ oo.m om.m~ oo.m moms Hon
3.3.. 3.2 3.3 HTmH moém «TS woém «TS moms zone
2H. 8H TH 8H 2H 8H 2H 8H we am
SH «TH NmH HTH 3H +HTH NmH HTH mg mg
mm.o mm.o om.o om.H HH.H oH.H mN.o mm.o Hm so: Hm
om.o NH.H ow.o HN.N om.H mo.m ow.o NH.H pm .omm Hm
qw.o mm.o aw.o mH.H mo.H wN.H ew.o mm.o um .cmw Hm
mo.o wo.a no.o oq.H Hw.o oo.H no.0 wo.H Hm .omm Hm
Ho.N ms.m Ho.m am.m mo.o mw.m Ho.m mm.N Hm .>oz Am
ma.m mm.m mH.N o o o mH.N mm.m Hm .uoo Hm

23a 2.3m 2...: 23m 722m 24.5 23H 23m

52 mm: 52 3% 52 mm: RE 3%
$2: Empomma

Na ammo: < .582 a. £82 H< ammo:

 

 

 

 

 

 

 

HH 0>Huwfihwu2

ill.

HH.© maan

”mucwafiowusmm pom omwmzmzmwcouuoo kauwm pom mmuusomom mo AmzmH cw mm>zv muoswosm woam>lamcwwsmz

169

The MVP of land remained zero in all models, and so were the MVPs of
family labor in August, September, March, April and June. In the other
months, the MVPs of family labor were lower than those in the base plans.
The MVPs of working party labor in Octdber and November also declined.
In Models A2 and A3, the MVPs declined from Tshs 1.37 to 0.85 in October
and from Tshs 1.66 to 1.02 in November. The MVPs of TSP, SA, Thiodan,
and DDT were higher than their corresponding MVPs in the base plans.
However, the difference in the two plans lay in the different prices
used for those inputs. In fact, the MVPs of these inputs were equal
to their prices. As in the base plans, it was not known, with these in—
put prices and under existing input-output relationships, whether the
optimum use of these resources was achieved. Data on production re—
sponses to these inputs were not available.

The most notable changes in the MVPs of resources were those of
operating capital. In each model, the MVPs of October and November
operating capital were substantially higher than those in the base

plans; in Models A1 and A the MVPs of operating capital increased

4,
from Tshs 0.49 to Tshs 1.26 in October and from Tshs 0.49 to 1.53 in
NOvember to Tshs 1.75 and 1.17 respectively. In Model A3, the MVPs in—
<2reased from Tshs 0.98 and 0.07 to Tshs 1.64 and 1.10 in October and
‘NOvember respectively. Undoubtedly, the increase in fertilizer and
insecticide prices led to a higher demand for operating capital in
those months during which these inputs were purchased. The reduction

in area cultivated could not remove the constraints imposed by higher

input prices.

170

Alternative III:

Increase in the levels of Operating Capital

 

In the base plans, operating capital was a limiting factor in‘
October and November. An assumption was made that farmers could
increase the monthly levels of operating capital by 10 percent if
given favorable prices and if their net farm incomes increased. The
results of this change in operating capital are presented in Tables
6.12 and 6.13.

It is clear from Table 6.12 that the change in area cropped was
remarkable in all models. Maize became more profitable than cotton.

In Models A1 and A. the area for cotton was reduced from 1.65 no 1.52

4,
hectares. Theareaplanted in maize increased from 1.30 to 1.47 hectares.
The total area of land cultivated increased by 1.34 percent. The employ-
ment of other resources also increased; family labor increased by 0.56
percent; there was no change in the amount of labor hired; and although
labor from working parties increased by 6.19 percent, operating capital
increased 2.70 percent.

Consequently, net farm income increased from Tshs 5538.46 in the
base plan to Tshs 5672.89, an increase of 2.43 percent. Because this
percentage of increase was higher than those of area cultivated or
family labor, the returns to land and to unpaid family labor increased
by 1.06 percent and 2.00 percent respectively. The return to a unit
of operating capital, however, declined from Tshs 5.39 to Tshs 5.38;
the percentage of increase in the amount of operating capital was

higher than the percentage of increase in net farm income.

‘31... .. _

 

 

171

 

 

 

 

 

 

 

 

2.0 13.0 00.0 0.0.0 I 3.0 omé 8.0 + ~05 $3 3.0 I 0T0 00.0 2:3. H338 53308:;
oo.~ + 004 004 00H + 00H RH moé + .04 ~04 oo.~ + 004 00H 22. H003 was: 00 .5202
8H 13.8.; .3.th and + 3.32 352: 3:0 + 3.82 8.33 8H + 8.83 .3.th was 25%..
SH + $303“ $.33 .3H + 2.33 3:52 .20 + 00.002 2.33 oTN + meémod $.33 0:2. H388 33308
0 0 0 I I I I I I I I I mu: ova Akufimv Heuomph
I I I SH + 00.3 3.00 I I I I I I mum x0 332.: same U.5
I I I I I I 00.0 + 00.00 00.00 I I I mum x0 300.300 800.0 x0
mmé + 00.0000 $.30.» 30 + 0H.mm0g 00.503 H04 + 00.0001 00:50.0 004 + 00.0001 3.3.3 0.5 0003 233.
3.0 + 3.000 2&00 nod + 0535 Hm§00 004 + 00.02 00.0: 2.0 + 01000 «5.30 mu: .3000 wcfitoa
0 00.02 00.00H 0 00.43 00.02 0 .0033 00.02 0 00.00." 00.00..” mp: .303 005x
00.0 + 2.00? 00.0000 2.0 + $.00? 010000 00.0 +Mqo.;\.m 340mm 00.0 + 0N.00nm 00.0000 0.5 .303 .3ng
qm.H +.m0.~ 00.N H0.H + 00.0 mm.m 00.0 +mqo.m No.0 qm.a + 00.N 00.N 0: 0:00 HmHOH
00.md+“a.fi 00.H n0.n + ~0.H Nm.H NN.0H+MHO.H mm.H 00.ma+ 50.0 00.H mm A0~H0z0 0cm;
mm.m I N0.H 00.H qm.q I m0.H 00.0 nm.n I_m0.a 00.H mw.m I N0.H 00.H mx m Acouooo xflummv vcwH
. _
med + 00.200 04.0000 004 + .0.0000 00.300 .No.0 1 Kimmnnm 00.030 mim + 00.?00 00.0000 0:00. :CE 08005 Emu 0.07.
l l .1
q< HONM q< H000: m< 000 m< H000: mm< H00 m< kuozm a< H00 H< H0002“ M
wwcmoo swam M 000 0wam£U cmHm 000 mwcmnu. swam “ 000 1 000000 swam 000 w Hsz m XMHH
N 302 Woman 0000 N 302 .cme 0000 u M 302 Woman mmmmm N 302 wcmHm 0000“ M

 

 

 

 

 

 

 

 

 

mucmafiuwopmm H00 oNHm: 000 000000 sauna ~00
kuHamo mcfiumpoao 00 0H0>QH 0:0 ca mowmouocH 0:0 woman 0000 0:0 ~00 mwpammmz wocoaofiwmm

NH.0 manmh

172

Although the area planted in cotton decreased from 1.65 to 1.53
hectares in Model A2, the area under maize increased from 1.37 to
1.51 hectares; these changes represented a decrease of 7.27 percent
and an increase of 10.22 percent for cotton and maize respectively.
The total area cultivated, however, increased from 3.02 to 3.04 hec-
tares, an increase of 0.66 percent. As the total area cultivated ins
creased, the employment of other resources also increased. Family
labor increased by 0.93 percent; labor from working parties increased
by 1.54 percent; use of owned ox—drawn equipment increased slightly
by 0.68 percent; and operating capital increased by 0.74 percent.

The changes in the use of resources led to an increase in net
farm income from Tshs 5729.60 in the base plan to Tshs 5782.71 in
the new plan; this was an increase of 0.92 percent. In turn, the net
farm income per hectare increased by 0.26 percent; the return to un-
paid labor increased by 2.63 percent; and the return to a unit of
operating capital increased by 0.21 percent.

In MOdel A the cotton area decreased from 1.65 hectares in the

3,
base plan to 1.58 hectares in the new plan. The area planted in
maize, on the other hand, increased from 1.32 to 1.42 hectares. These
changes led to an increase of 0.03 hectares in the total area culti—
vated and to increases in the employment of other resources. The
employment of family labor, labor from working parties, hired ox-

drawn equipment, and operating capital increased by 0.15 percent, 2.97

percent, 1.77 percent, and 1.94 percent respectively.

173

The impact of the increased level of operating capital on net
farm income was positive. From a net farm income of Tshs 5614.93,
in the base plan, net farm income increased to Tshs 5693.04 in the
new plan, an increase of 1.39 percent. This change, together with the
changes in the employment of unpaid labor, area cultivated, and opera-
ting capital, led to an increase in return to unpaid labor and to
land of 1.98 percent and 0.38 percent respectively, and to a decrease

in return to a unit of operating capital by 0.46 percent.

Marginal Value Products (MVPs in Tshs) of Resources Under Alternative III

 

 

In Table 6.13, the MVPs of resources under Alternative III are
shown. As in Alternative II, the MVPs of land; family labor in August,
September, October (for M0dels A2 and A3), March, April, and June;
owned ox-drawn equipment; operating capital in December to June; hired
labor in October to May; and working parties in October (for Mbdels A2
and A3), December, January, February, April and June were zero in all
models and equal to their corresponding MVPs in their respective base
plans. The MVPs of family labor in October (for MOdels A1 and A4) and
in November, December, January, and February were lower in all four
models than in the base plans. This reflected the fall in demand for
labor mainly because of the decline in the cotton area; a 1 percent
decrease in the cotton area released more labor than required by a 1
percent increase in the maize area. In May, however, the MVPs of
family labor were higher than the corresponding MVPs in the base plans.
As a result of the increase in the maize area, demand for labor for

harvesting, which takes place in May, increased. This higher demand

 

 

 

 

 

 

 

 

 

00.0 0 0H.0 0 00.0 0 55.0 0 02 502 Hz; 03
0H.H 00.H 00.0 mm.H N0.H 05.H 0H.H 00.0 02 .>02 020 03
H0.H 5m.H 0 0 0 0 .0.H 5m.H 02 .000 Hz; 03
0 00.0 00.0 50.0 ma.0 00.0 0 00.0 0:00 .>02 00
0N.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 mama .000 00
00.0 00.m 00.0 00.0 00.0 00.m 00.0 00.m mama 000
sm.mH qm.ma 0m.mH qm.ma qm.mH 00.0H 00.0H sm.ma 0000 2020
m0.a 00.0 mo.H m0.H 00.H m0.H mo.a m0.H 02 <0
sm.a 0m.H qm.a 00.0 am.H qm.H sm.a 00.0 02 000
0H.H 00.0 H5.H 05.0 0H.N 0H.H 0H.H 00.0 00 202 00
mm mm.H 50.0 00.5 05.5 mm.m m0.m mm.H 50.0 02 .000 02
05.0 no.0 00.0 0H.H 00.0 0~.H 05.0 00.0 00 .000 02
50.0 00.H NH.H 00.0 0N.H 00.0 50.0 00.H pm .000 02
mm.N m5.N 00.N mm.m 00.0 50.0 mm.~ m5.N um .>oz 02
m0.m 00.N 0 0 0 0 00.N 00.N pm .000 Ah

2<00 2<00 2<00 2000 2<05 2400 2<00 2<00

302 0000 302 mm<0 202 mm<0 302 mm<m
0H2: 00200002

0< 0000: m< 0000: N< 0000: H0 00002

 

 

 

 

 

 

 

....I.

HHH 0>Huwcumua< "mucmmfiowunmm Mom 0302.000 couuou 0.20m How mmousommm mo 3an 5.” 002.3 muoswoum mDHmSIHmGkumZ

0H .0 maan

175

for labor was met by working parties in May, which in turn became a
constraint. Operating capital, which was a limiting factor in October
and November, was still a limiting factor in October for all models and
in November for Models A2 and A3. These MVPS of operating capital,
however, were lower than the MVFé in the base plans.

The results under Alternative III, like those in the other alterna-
tives, indicated that, with more operating capital, farmers could in—
crease their net farm incomes and reduce the resource constraints in
production. However, this would clearly be against the objective of
the government, i.e., to increase the output of cotton and maize, not
to increase one at the expense of the other. The results in the three
alternatives indicated that seasonal labor bottlenecks were still the
main constraints to expanding the area under cultivation and consequently
the cotton and maize yields. The employment of owned or hired ox-
drawn equipment was limited to land preparation. Farmers did not know
how to use ox teams for planting, weeding, or harvesting.

There are two possible ways of alleviating seasonal labor bottle—
necks. One solution is to space out farm operations between maize and
cotton by having one of these crops planted late. The other is through
mechanization (whether by animal draft power or tractor) of farm operae
tions that are currently limited by labor bottlenecks. The latter
option is discussed in the summary of this chapter and in the discussion
of policy issues in Chapter VII. The first cropping option is examined

in the discussion of Model B below.

176

Model B:
Mechanization and Manual Technologies:

Maize, Early Cotton, and Late Cotton

 

If farmers were to plant one of these crops late, maize would not
be the likely choice. Maize is their main food crop, and farmers would
not take any risk by growing it late. Cotton, therefore, is a better
choice.1 Late—planted cotton has lower yields. In the analysis,
January was taken as the planting month for late cotton. If cotton
were planted in January, yields per hectare would fall by about 20
percent.

By including late—planted cotton, crop production activities in-
crease to three. The input coefficients for late-planted cotton were
similar to those for early cotton except that they were shifted
downward. As an example, the family labor input coefficient for early
cotton in August was shifted to September, that in September was shifted
to October, and so on. The results for the analysis are presented in
Tables 6.14 and 6.15.

By introducing late—planted cotton as an optional production
activity, early-planted cotton was not included in the optimal solution

except in Model B As shown in the results for Models B and B4, the

2' 1
area planted in maize expanded from 1.23 hectares in the actual average

to 1.34 hectares in the optimal solution, an increase of 8.94 percent.

The area under late—planted cotton came to 1.75 hectares.

 

1About 93.5 percent of the farmers interviewed indicated that if they

were to choose which crop to grow late, they would choose cotton.

177

 

 

 

 

 

“0000000 0000000005002

 

 

 

mZOHH<NHz<omO A<rwhac

00.0m+ _ 00.0 ” 00.00+ 0 00.0 05.0~+ 0 00.0 i 00.00+ W 00.0 00.0 _ 0000
00.00- a 00.0 w 00.00- M 00.0 50.00- m 00.0 W 00.00- w 00.0 00.0 0000 _ 00000 000000 005000
00.00- m 00.00000 00.00- _ 50.0000 50.0 ” 00.0050 M 00.00- . 00.0000 00.5000 0000 w 000000
00.00- W 00.0000m 00.0 + W 05.0000 00.00- w 00.0000 W 00.00- fl 00.0000 00.0000 0000 N 0000000 0000000-c
ooHI M o m I w I I m I M I _ I . mm.m 00$ 00% m Anwufizv 0000000
0 n . m i .
- m - W - m - “ 00.0~+ M 00.00 M - m - 00.05 000 00 M 0000300 0000 00
00.00+ w 00.0050” 00.0~+ W 00. 0050" 00.0N+ 0 00.0050 W 00.0N+ m 00.0050 05.0000 000 00000 00005
. _ . . . . 0

00.0 - H 00.000 W 00.0 - w 50. 005 _ 00.0 - m 00.005 W 00.0 - m 00.000 00.500 000 00000 0000000
ooHI W O W ooHI m o OOHI . o m ooHI w o oo.w¢~ mum ponma nmufix
00.0q+ w 00.0000“ 00.00+ % 00.0000 00.0q+ M 00.0000 W 00.00+ M 00.0000 05.0000 000 _ 000000
00.00+ _ 00.0 W 00.0~+ m 00.0 00.00+ W 00.0 0 00.~N+ m 00.0 00.0 . 00 W 0000 00005
000+ 05.0 w 000+ m 05.0 000+ W 00.0 W 000+ W 05.0 w 0 m 00 N 0000000 00000 0000
00.0 + 00.0 W 00.00+ “ 50.0 H 50.00+ m 00.0 0 00.0 + w 00.0 W 00.0 W 00 m 0000000 0000
000- 0 M 000- W 0 50.05- W 50.0 000- W 0 m 00.0 W 00 m 0000000 000000 0000
00.0 + _ 00.0000“ 50.0 + m 50.5000“ 05.00+ " 00.0000 m 00.0 + m 00.0000w 00.0050 . 0000 “50000 000000 0000 0.0

u m 0 u a _. I. _

022H00 M 092a0" A weamsm N . 092E. M0 w 02: m

N u 0 00000. 0 m 0 00000m u w 0 00000 W N M 00000: -00000 g 0020 m

_ _ i

---—.1»- -

 

 

mucmawuwuhmm hOw mNmeI MOM mofihm umxHNZI mmhm Ucm .SOquU mumd UCQ COuuQU AHHMM
000 00000 00000000 000 000 0000:0000 00000000 00000 000000.000000 0000000 000 0000- -00000 00 0000000000

0a.o manmw

178

Changes in the use of other resources were also significant.
Whereas the employment of family labor increased by 40.38 percent,
that of hired labor and working parties declined by 100 percent and
0.61 percent respectively from those in the actual situation. The
use of operating capital decreased from Tshs 1552.15 to Tshs 1296.60,

a decline of 16.46 percent.

After these changes, the net farm income came to Tshs 5129.03,
an increase of 8.96 percent from Tshs 4706.86 in the actual average.

The net farm income per hectare and the return to unpaid labor de—
creased by 11.13 percent and 22.89 percent respectively. Only the
return to a unit of operating capital increased-—it increased from
Tshs 3.03 in the actual average to Tshs 3.95 in the optimal plan, an
increase of 30.36 percent.

The optimal solution in Model B2, which included the use of owned
ox—drawn equipment included 0.27 hectares of early cotton, 1.49 hectares
of late cotton, and 1.36 hectares of maize. These changes represented a
79.07 percent decline in the area planted in early cotton, a 100 percent
and a 10.57 percent increase in the area under late-planted cotton and
maize respectively. Total land cultivated increased by 23.41 percent.
Whereas the employment of family labor increased by 40.12 percent, hired
labor and working parties declined by 100 percent and 6.48 percent respec-
tively. Whereas employment of owned ox-drawn equipment increased by
23.24 percent, use of operating capital fell from Tshs 1552.15 to Tshs
1391.44, a 10.35 percent decrease.

The net farm income increased from Tshs 4706.86 in the actual

average to Tshs 5304.68 in the optimal solution. This represented an

 

179

increase of 12.70 percent. This increase together with the changes
in the use of family labor, area cultivated, and operating capital
led to a decrease of 19.87 percent and 8.97 percent in the return to
unpaid labor and land respectively, and an increase of 25.74 percent
in return to a unit of operating capital.

In the results for Mbdel B3, which included the hiring of ox-
drawn equipment, the highest increase was shown in the area planted
in maize; it increased from 1.23 to 1.37 hectares, an increase of
11.38 percent. Although the optimal solution did not include early—
planted cotton, the area under late—planted cotton came to 1.73
hectares. Compared to the actual average employment of hired labor
and working parties declined by 100 percent and 3.32 percent respec—
tively. Meanwhile, employment of family labor, hired ox—drawn equip-
ment, and operating capital increased by 40.00 percent, 9.95 percent,
and 8.59 percent.

As a result of the above changes in the use of resources, the
estimated net farm income came to Tshs 5157.35, an increase of 9.57
percent from Tshs 4706.86 in the actual average. The net average
resturn per hectare was Tshs 1663.67 compared to Tshs 1867.80 in the
actual average, a decrease of 10.93 percent. Net farm income per man
hour of unpaid labor was Tshs 1.29 compared to Tshs 1.66, a 22.29 per-
cent decline. Finallygthe net return per unit of operating capital was
Tshs 3.63 compared to Tshs 3.03 for the actual average situation, an

increase of 19.80 percent.

180

Marginal—Value Products (MVPs in Tshs) of Resources Under Model B

In Table 6.15, the MVPS of resources under Model B are given.

1
Land was still in excess supply as shown by its zero MVPs. The in—
corporation of late-planted cotton did not bring enough land into

use to make it a constraint on production. However, much more land
was employed in this model than in Model A (see Tables 6.4 to 616).

The MVPs for labor in November to February and in May were lower

in Mbdels B to B than their corresponding MVPs in the base plans for

l 4
Models A1 to A4. In March and April, however, the MVPs for labor were
positive as compared with zero for the same months in Models A1 to A4.

These changes in the MVPs for labor were quite understandable. In
Model A, land cultivation, ridging, planting, first weeding, and second
weeding for early-planted cotton and maize took place during the same
months, which required a great deal of labor per farm operation. In
Model B, however, these farm operations were spaced out by introducing
late-planted cotton. Thus, the coincidence of the same operation taking
place for both crops at the same time was avoided. By doing this,
seasonal labor demands were reduced. The MVPs of family labor in March
and April were exceptions to the above argument. The high MVP of

to A. was caused by a

l 4

high demand for labor that resulted from an expanded area under late

family labor in March compared zero in Models A

cotton. It should be noted that for late-planted cotton, March would
be the period for second weeding, an operation that requires more labor
than either the first, third, or fourth weedings. As for April, there
vuould be a coincidence of demand for family labor between the third

“needing of cotton and harvesting of maize. Such a high demand for labor

181

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Hq.o oo.H o Nm.H o oN.H Hq.o oo.H um .>oz 924 m3
0 mm.H o o o o O wm.a Hm .uuo BEA m3
NH.o o mN.o o oo.o o NH.o o mama .omm oo
mm.o mq.o Nq.o no.0 oH.o om.o mN.o mq.o mLmH .>02 00
o mq.o o no.0 o Hw.o o mq.o mLmH .uoo oo
oo.m oo.m oo.m oo.m oo.m oo.m oo.m oo.m man Ham
qm.ma qm.mH qm.ma qm.mH qm.ma qm.ma qm.ma qm.ma mnme zamH
mo.H mo.H mo.H mo.H mo.H mo.H mo.H mo.H wM <m
«m.H qm.H em.H qm.H qm.H qm.H qm.H cm.H mm mmH
Nm.o mm.o 00.0 om.H 05.0 oa.a mm.o mm.o Hm >mz Ah
no.H o Ho.H o mm.H o no.H o u: .ua< Ah
ww.m o qm.m o mH.q o ww.m o um .um: Ah
No.0 NH.H Nw.o HN.N qo.H mo.N No.0 NH.H um .nmm up
00.0 mm.o mN.o wH.H Nm.o wN.H 00.0 mo.o um .cmm Ah
qfi.o wo.H wm.a oq.H no.0 oo.H «H.o wo.H um .owo gm
om.H mm.m oo.H mm.m Om.a mw.m om.H mm.N um .>oz an
o mm.N o o o o o ma.m um .ooo Ah
swam cmfim cmfim swam swam swam swam swam
Bmz wwmm 3oz wmmm sz ommm 3oz ommm HHZD momDOmmm
cm Homo: mm Hoooz mm ammo: Hm Home:
{ll\

.couuoo mflpmm umwcj ed ou mfl

mH.o mapme

cmmaufiupmm new mafia: mam coauoo mum;
m mameoz bOm mwupsommm so Amnme as mm>zv mueseOHm msfim> Hmckumz

182

would also be caused by an expanded cultivated area for crops in
Model B. Although the MVPs for November working parties declined

from 1.66 Tshs in Models A1 and A to Tshs 0.41 in Models B and B

4 1 4’
the MVPs of working parties in October for Models B1 and B4 and
November for Models B2 and B3 declined from positive MVPs to zero.

Unlike Models A1 to A4, in which the MVPs of operating capital
were positive in October and November, the MVPs of operating capital
were positive for November and December in Models B1 to B4. The in—
corporation of late—planted cotton, therefore, did not totally remove
the operating-capital constraints; it only shifted the constraints.

Although the results in Model B were not better in terms of net
farm income, than those in Model A, they indicated a substantial poten—
tial for minimizing seasonal labor bottlenecks and increasing farm
income if biological technologies that might produce high-yielding
late—planted cotton seeds were introduced. The empirical findings
indicated that land is not a limiting factor. Such information is of
value to policy makers in determining the extent to which emphasis
Should be put in land—saving techniques or labor—saving techniques.

gptimal Organizations with Existing Resources
and Prices for Nonparticipating Farmers

In this section, a discussion is presented of the nonparticipating
farmers. This discussion is parallel to that given for participating
farmers. First, Model C for maize and early—planted cotton is dis—
cussed. Then Model D for late—planted cotton is discussed. Alternative
ways of applying the models are also presented as they were for Models

A and B.

 

183

Model C:

Mechanization and Manual Technologies: Maize and Early Cotton
for Nonparticipants

 

 

In Table 6.16, a comparison between the bench-mark and optimal
organizations of the representative farm for the nonparticipants is

given. Four submodels are represented in the table: Model Cl’

representing farmers who employ only labor for farm operations;

Model C2, representing farmers who employ labor and owned ox—drawn

equipment; Model C3, representing farmers who, in addition to labor,

 

employ hired ox—drawn equipment; Model C4, representing farmers who
employ labor and hired tractor power.
In the analysis, it was shown that the results for Models C1,

C3, and C were the same. Since there was no use of hired ox—drawn

4

equipment in Model C or of tractor power in Model C4, all farm oper-

3
ations were done by manual technology. As a result, the areas planted
in the crops were to be determined by the seasonal availability of
labor and other resources.

Compared to the bench-mark, the areas planted in maize and cotton
changed substantially in all four models. In Models Cl’ C3, and C4,
the area planted in maize decreased from 1.38 hectares under the bench—
mark to 0.43 hectares in the optimal organizations. This represented a
68.84 percent decline. 0n the other hand, the area planted in cotton
increased from 1.25 to 2.02 hectares, an increase of 61.60 percent. The
total area cultivated declined by 6.46 percent.

The decline in land cultivated led to a decline in the employment

of other resources. If no labor was hired, family labor and working

184

 

 

 

 

 

mm.mq+ o~.m mm.sq+ o~.n mm.o +
Ho.o~+ am.o Ho.om+ mm.o mm.am+
wo.Hm+ mo.emsH mo.Hm+ mm.omsfl mo.Hm+
Ns.aHI mm.~oa No.5HI mm.~os ~5.¢HI
ooH- o I I I
I I ooHI o I
I I I I so.~ I
mo.m I N~.NmHs wo.m I -.~mHs oa.w I
am.a~- mH.mm Hm.a~- ma.mm ~m.mm-
ooH- o OOH- o cod-
so.~ I so.msoq sa.~ I so.ooos mo.o I
os.o I _ os.~ os.o I os.~ mm.~ I
- m - m - - . -
w .
sm.ms- m ms.o m sw.wou ms.o _ sm.mo-
oo.a¢+ m No.~ m oo.Ho+ No.N CN.HNI
Ho.-+ m as.amom w Ho.~N+ H¢.amom_ mo.wmi
.r _
$726 w _. 8250 SEW-
N M «U amaozm N no ammo: N

 

 

 

mN.m

no.0

mm.owcH

qw.omn

wm.oh

mo.ommm

wq.on

o

mm.mamm

nn.m

mq.o

QH.N

<m.ammm

 

 

 

 

 

 

 

 

mZOHH<NHz<UmO A<£HHmO

-..—db—-i_-—q ..

 

mm.¢¢+ om.m mq.m mnma Hmufiamu wcfiumuooo\Huz
Ho.oN+ sm.o as.o mama roams sacaco t:\Huz
wo.Hm+ ms.omqa ~m.smHH mama mxaHmz
Nm.aau mw.~os mm.omm mama Hmuaamo assumtmcc
. I I ow.q mu: one Avouwmv pouumps
I I mn.o~ mum so Anmufixv amoH x0
I I om.wn mum x0 Auocsov ammw xo
wo.m I NN.~qu om.samq mum gonna Hades
Hm.a~- ma.mm oa.sHH up: sauna wcaxtoz
ooHI o oo.mo mum roams amt“:
sm.~ I so.ooos oa.~aas mum roams sflaamu
04.0 - o¢.~ mo.~ m: can; Hmuos
I I w I m: Accuuoo mumAV mam;
w qw.mcI mq.o w mm.a mm AoNfimzv wand
W oo.a¢+ No.N m m~.H mm . Accuses harass was;
M Ho.NN+ Hm.amomm a~.mwm~ mzms MAHmzv «acucH arms “02
_
AII
w usage H IVE:
_ a 0 sun muzmm HHz: zmHH

-' ~ --——~-—-—-—— -4».._-

 

qu moofium Hmaofimuo ozu can mmopsommm wcaumaxm nova: mGOHumNHcmwuo Hmaauoo cam xusznocom

mucmofiowuumacoz yew GOuuoo haumm mam unwmz

0H.o maan

WC

comfiumoaoo

185

parties declined by 2.94 percent and 27.51 percent respectively.
Total labor employed declined by 5.08 percent, which was lower than
the percentage decline in total land cultivated. This was simply
because more labor was required to farm the expanded cotton area
because cotton requires more labor per hectare than does maize.
Operating capital declined from Tshs 856 in the bench-mark to Tshs
702.85, a 17.92 percent fall.

The optimum net farm income came to Tshs 3657.91 compared to Tshs
2983.27, an increase of 22.61 percent. The changes in net farm income
and in the employment of resources led to increases in net farm income
per hectare, per unpaid labor, and per unit of operating capital; these
increased by 31.08 percent, 26.61 percent, and 49.55 percent respectively.

The results for Model C2 were noticeably different. Whereas the
change in the area planted in maize was the same as those in Models C1’
C3, and C4,

from 1.25 hectares in the bench—mark to 2.14 hectares in the optimal

the area planted in cotton was the largest. It increased

solution. The decline in total cultivated land was only 2.83 percent.
As in the other models, no labor was hired in the optimal solution.
Employment of family labor, labor from working parties, and owned ox—
drawn equipment declined by 6.65 percent, 33.32 percent, and 2.94 per-
cent respectively. In addition, use of operating capital declined from
'Ishs 856.35 in the bench—mark to Tshs 730.24, a fall of 14.72 percent.
The estimated net farm income increased from Tshs 2983.27 to Tshs
3821.54, a 28.09 percent increase. Consequently, the return to land,
to unpaid labor, and to a unit of operating capital increased by 31.09

percent, 37.53 percent, and 0.58 percent respectively.

 

 

186

In Table 6.17, bench-mark and optimal organizations are compared
with the official price for maize replaced by the free market price of

Tshs 2.50 per kilogram. Again, the results for Models C C3, and C

1’

were the same. The area planted in maize increased by 9.42 percent,

4

whereas that planted in cotton declined by 10.40 percent. Total land
cultivated did not change. Although the decline of the cotton area

by 0.13 hectares was offset by an increase of the same size for maize,
employment of labor and operating capital declined. Family labor, labor
from working parties, and operating capital declined by 9.87 percent,
76.02 percent and 33.46 percent respectively.

The change in net farm income was not substantial; it increased
from Tshs 4027.18 in the bench-mark to Tshs 4260.83, an increase of
5.80 percent. As there was no change in total land cultivated, the
return to land also increased by 5.80 percent. However, substantial
declines in the employment of family labor and operating capital led
to substantial increases in return to unpaid labor and to a unit of
operating capital by 32.29 percent and 92.96 percent respectively.

In the optimal solution for MOdel C2, a further decline in the
cotton area was observed. From 1.25 hectares in the bench-mark, the
cotton area declined to 1.09 hectares. The area lost by cotton was
gained by maize, for which the area increased from 1.38 to 1.54 hec—
tares. Consequently, there was no change in total land cultivated.
Because of the use of owned ox—drawn equipment, employment of labor
was much less in this model. Use of family labor declined by 16.35

percent, whereas that of working parties declined by 88.95 percent.

187

 

 

 

 

 

 

 

 

 

.

 

 

 

 

mZOHH<NHz<omO A<ZHHmO

 

 

oo.m¢+ l so.o m em.~o+ m so.o m os.om+ so.a om.~o+ w no.o ca.s mama Hmuwamu “caumtmao\zuz
om.~m+ m RN.H . ¢N.Nm+ s~.H HH.mN+ oN.H om.~m+ M sN.H o¢.o mama roams vacate tz\Haz
om.m + m so.o~oH ow.m + ao.o~sH mm.s + mm.HooH om.m + M ao.o~sH s~.HmmH mgms mx\Hux
cs.MMI W om.aom as.mmI ow.scm ms.omI mm.msm sq.mmI om.mom mm.omw mama Hmuaawo wcfiumtmcc
ooHI H o I I I I _ I M I om.q mu: ops Auouwmv p0oumuw
I W I r ooHI m o I I I m I mm.om mum x0 . Abouamv amoH x0
I w I I I m o om.ma ” I w I om.ma mum xo Acmcaov anus so
oo.mHI m sa.mowm oo.msI m sa.mommw om.aHI so.mmmm W oo.MHI W as.momm ow.q~ms mum W roams swoon
No.0aI m Hm.a~ No.o~I m Hm.sm W ma.wmI as.~fi W No.oaI W Hm.s~ m ON.QHH my: . auras mcaxtox
ooHI o ooHI W o W ooHI o W ooHI W o m oo.mo mum bonus oats:
aw.s I . mm.mham_ aw.s I M mm.maamw mm.oHI “ oq.00mm w am.o I W mm.mssm" 0H.Noaq mum rooms sasamt
o mo.~ m o m ms.~ W o w ms.N m o m mo.~ w mo.~ m: cams Hauoh
I . I m I m I W I I m I W I W l . m: m ACOquU 0..“qu USN]
ms.s + Hm.H m Ns.a + W Hm.H W mm.HH+ m sm.a m Nq.m + W Hm.H M mm.H I mm W AmNHmzc acn_
os.oHI NH.H W oq.oHI W NH.H m ow.NHI m mo.H w os.oHI w NH.H W m~.a mm M Acouuou ssnmmc scma
om.m + mm.oo~s_ ow.m + m mw.oomsm mm.s + m as.HHNs m ow.m + m mm.oo~qm wfi.a~oq mcms mAHmzc meoUCH arms “I,
a n _ l a ”I i l
I 1 _ I I I 4
isaxoh _ usafio“ . mezac_ " msaso “ 82: w
,h so doozm .... . mo deozm N _ No ammo: ” N M Ho ammo: Imozmm 32: m ES
1

 

muomoaowuumacoz “Om mufimz HOW mofium umxHQZImmum van coauoo maumm
pew oofium Hmofiwwo wSu van mmousommm wcfiumfixm nova: mCOfiumuacmwuo Hmaauoo can xMNZIcocmm mo comwumqeoo

NH.© mHan

188

The decline of 30.45 percent in the use of operating capital, however,
was a little lower than that in the other models. If a farmer employed
owned ox—drawn equipment in Model C2, he incurred operating costs (see
Chapter V). The decline in operating capital would have been much lower
in this model if the area planted in cotton declined by a smaller
percentage. This was because more inputs were used in cotton produc—
tion than in maize production.

From Tshs 4027.18 in the bench-mark, the estimated net farm in-
come increased to Tshs 4211.49, an increase of 4.58 percent. As a
result, net farm income increased by 4.58 percent, return to unpaid
labor increased by 25.11 percent, and return to a unit of operating

capital increased by 50.46 percent.

Marginal-Value Products of Resources Under Model C

The MVPs of resources used in the production of maize and cotton

'in Models C1 to C4 are shown in Table 6.18. The MVPs of resources in

1’ C3, and C4 were the same. At the official maize price of

Tshs 1.00, land was in excess supply in Models C1 to C4, as shown by

Models C

its zero MVP. At a free-market price for maize, the MVP of land be-

came Tshs 465.22 for Models C1, C3

both reflected the scarcity of land. Thus, if one hectare were

, and C4, and Tshs 514.67 for Model

C2;
Iarought into cultivation under Models C1, C3, and C4, income would in—
czrease by Tshs 465.22, whereas under Model C2, the increase in income
vvould be Tshs 514.67. Similarly, a reduction of land under cultivation

by one hectare would decrease income by the amounts of MVPs indicated

above. The MFC of land was assumed to be zero since no land was

189

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

, M
oo.m oo.m ©O.m co.m ©O.m m ©O.m ©O.m ©O.m ” mLmH HOD
_
MM.MM MM.MM Mm.MM 0m.MM MM.MM MM.MM MM.MM Mm.MM m Mama Mama
. .
mo.H mO.H mO.H mo.H mo.H H mo.a MO.H mo.H m mm <m
Mm.M 0m.M 0m.M 0m.M Mm.M m Mm.M. 0M.M Mm.M _ MM Mms
_ .
qo.H O «O.H o mm.H w o MO.H o ” um .me gm
_ M
o o o o 0o.M M o o o m Mm .cms Ms
MN.M MM.o MM.M MM.o wo.o . mm.o M0.M MM.o m 0: .>oz MM
_
q©.H mN.o q©.H AN.O O w O q©.H NN.O hm .uoc Am
mm.m©q O NN.m©q 0 IF. m©.qam _ O NN.m©q 0 mm Ucmq
M M
q .M .M .
d d i .d d d d d w d d _ d d d d d d .
. I I i Q . Q . _
MM ,Mm ,M ,IM “MM MM Mm M0. 1
MM lMM MM MM MM MM MM MM“ 1
7uwc _ W. To 70 To W. 90 M 70 Co ~ I. 90 “ 70 f0 T. Cc ” _
. MM . ..I.., _.. ...l... H
MM 1mm mm. MM m MM 1 MM w mm 1mm 1 :3 l, 853:
_ [MI I IIIIIIIIIITIIII n M .
M m _ I l _ H
o MMMoz 0 Moss: M 00 Mesa: fl M0 Mass: M fl
mucmoflowopmaccz Low 00 Ho mambo: p00 mmopsommm Mo Awnwe CH mm>zv muosuopd msflm>IHmcflmpmr

ma.o oflfimb

190

bought or rented. The MVPs of family labor in August, September,
October (for Model C2), December, and March to June, the MVPS of
owned ox-drawn equipment in October to December, operating capital,
hired labor, and labor from working parties were all zero irrespec-
tive of which of the two maize prices was used. Family labor con-
straints were observed in October for Models C1’ C3, and CZ; in
November for all models; in January for Model 02 (when the maize
price was Tshs 2.50), and in February for all models (when the price
for maize was Tshs 2.50). These were the peak labor periods for
these farmers. A great deal of labor for cultivating and ridging
the land was required in October and November; in January and parti-
cularly in February, much labor was required for weeding of cotton
and maize to fight weeds that grow rapidly due to heavy rains. The
MVPs of TSP, SA, Thoidan, and DDT were equal to their prices. In
the absence of data on production responses to these inputs, it was
not known if the optimum use of these resources had been achieved.
_thimal Organizations of the Nonparticipants' RepresentatiVe Farms

with Variable Product, Fertilizer, and Insedticide Irflces, and

_Qperating-Capital Levels for Models C1 to C4
The impact of input and output price changes on net farm income,

cropping patterns, and resource use under Models Cl and C4 is dis-
cussed in this section. No attempt was made to explore the impact of

Changing operating capital levels since the resource was not a con—

straint in the base plan.

 

191

Alternative I:

 

Changes in the Relative Prices of Early Cotton and Maize

 

The relative prices of cotton and maize were increased by the same
proportion as those discussed in section one of this chapter. These
price changes are discussed under Alternatives Ia, Ib, and Ic. Because
of relative price changes under Alternative Ia (Table 6.19), the area
planted in maize decreased by 63.57 percent; the area planted in cotton
1’ C3 and C4. There was no

change in total land cultivated. Compared to the base plan, employment

increased by 85.71 percent under Models C

of labor increased from 3805.74 to 3978.08 man hours. All of this in-
crease, however, came from family labor; its employment increased by
4.78 percent whereas that of working parties declined by 31004 percent.
Use of operating captial increased from Tshs 569.80 in the base plan to
Tshs 624.77 in the optimal organization, an increase of 9.64 percent.
Most of this increase resulted from the increase in cotton hectarage,
for much more purchased inputs per hectare were required by cotton than
by maize.

The estimated net farm income came to Tshs 3908.62 compared to
Tshs 4260.83 in the base plan. This represented a decrease of 8.26
percent. As a result, net farm income per hectare fell by the same
percentage, return to unpaid labor and to a unit of operating capital
declined by 22.83 percent and 31.09 percent respectively.

Under Model C the results followed more or less the same pattern

2,

as those discussed above. Whereas the cotton area increased by 88.07

percent, the are planted in maize declined by 62.34 from that for the

 

 

192

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

8qu T0 85 84? N0 84. 2.07 as B; 84TW0 SM 2: M338 chSM8:.E
SM? m5 2.4 mas? To 24 SAT we; 84 8.2? “.35 RA Ems 803 3895 to 5;:
2.0 I.M.00.; 8.80M 80 H.003 8.83 S; $0.33 2.80M 3:0 I 3.00.: 8.80M £2. £5:
0N.m +.a~.0~0 om.m0m 0N.m +.ws.0~0 om.m0m oh.¢ +.Hm.mm0 mm.mmm 00.m +.ss.0N0 ow.m0m mama Hmuflamu waumumao
o o o I I I I I I I I I mum ouH Anouwxv heuomus
I u I I o o o I I I I I I 8m S 38:: 5.3 Us
I --I- I I I _ I o 90.2 m 3.0a I I I 3m .5 9238 53H 5
2.0 + 3.38 Mimosa 2.0 + 8.33 simommmooé “8.003% 8.23 .2... + 00.2% 058%. mu: . 803 M88
3.2- 3.2 3.: was? E 3 m 3.8 was: 0 0.“ WEAVER M05 m 2x _ has Munro;
0 _ o . o m o n o w .o M o H o . o _ o o o 3m 803 ME;
wn.q IwH.mmo mm mm.mnnmmmn.qI Ir .omomm mm.mnnmmmm.0.+wo.000& 00.0ommwms.0.+fla.mmmn mm.mnnm mum w Monmg smwams
o W0M 1 SM _ o WOM M SM . o T: 1 M0.“ o .m0.~ M SM .5 _ 23 M30...
_ . a _ . _ _
Rafwfio m M04 “3. m0I Wm. o W 24 mqm.N0IT.o m 004 “Rafummd m 34 mm m 323: 23
mmflwos _ 24 m2. 3+3. N SM m.o.mw+.mo.m m 84 MIAMIWMOM . .24 mm . 9038 p.230 23
£018.83. 8.0038: 0 Ir0. .83 m0. 002:0; 23.095 3.:N30NM 40.00am“ 8.00.3. 23. . 3sz 2.85 ES 7.x
. l M M _. .. M .I .. _. l
- 06 80m 0 3001 _mu 8: mg :63 .mo 80m 0 30on Mo 80 Mo 309... m
Mummfi SE . SM .mmwmcoma Sm H 80 Mwmmfi. .53 M 80 “mmcm6_ cm: _ 80 u SE M ES
9 3oz “ swam ommmw n w z_ cmam mmmmm a m 362 m: as mmmmp N * 3oz ~ swam mmmmm w

 

 

 

mH m>fiumaumudq

”mucmafiowupmocoz you GOuuoo
haumm vcm mafimx you mmofium uospoum m>fiumamm CH mwmcmcu cam mamam mmmm msu “cw mouswmo: %ucwfiowwmm

mH.o mHQMB

193

base plan. Again, the total area cultivated remained unchanged.
Although there was no employment of hired labor and although labor
from working parties declined by 53.94 percent, use of labor increased
by 4.00 percent due to a 4.38 percent increase in the employment of
family labor. The change in the employment of owned ox-drawn equip-
ment was zero; operating capital increased by 9.70 percent.

The estimated net farm income decreased from Tshs 4211.49 in the
base plan to Tshs 3896.58 in the optimal organization, a fall of 7.47
percent. With no change in total area cultivated, return to land
changed by the same percentage. The average return to unpaid labor
declined by 11.67 percent, whereas that to a unit of operating capital
decreased by 15.70 percent.

As in Alternative Ia, there were substantial changes under
Alternative Ib (as shown in Table 6.20). The results in Models Cl,

C , and C

3 were again the same. The area planted in cotton increased

4
by 91.07 percent, that planted in maize declined by 67.55 percent, and
the total cultivated area remained the same. Despite the decline in
the area planted in maize, the increase in the cotton area led to in—
creases in the employment of labor and operating capital. The optimal
use of family labor, labor from working parties, and operating capital
increased by 6.29 percent, 140.60 percent, and 25.71 percent respectively
from those under the base plan.

The optimum net farm income came to Tshs 4718.48 compared to Tshs
4260.83, an increase of 10.74 percent. Consequently, the return to

land increased by 10.74 percent since there was no change in total

area cultivated. In addition, the average return to unpaid labor

 

 

194

 

 

 

 

 

 

 

 

 

 

 

 

 

 

3.3- 00.0 85 3.3- 3.0 8.8 2.0180 8g 00.3800 BM. ME 1338 83208:;
0:0 ISM RM 030 I :4 GA 8.0 +84 84 0N0 .24 RM 88 :33 33.5 0.. :5:
0531.88: 8.8.: £818.82 8.80M 8.2+ M040: 8.80M £818.82 8.82 28 25M:
2.8+ 8.8: 8.800 :0? 8.0: 8.80m 8.8+ 3.8a “0.8“ 2.21802 8.88 95 2338 8228.0
0 o o I I I I I I I I I mum onH Anuuwmv neuumuh
I I I o o o I I I I I I w»: x0 Aumuwxv amok x0
I I I I I I o 8;: 80s I I I 3x 5 8280 $3 5
0mg +8.33 05.88 0N; + 8.33 :88 3.0 + 3.25 8.28 0mg + 8.83 :88 E: :53 ~28
88:58 2.3 8.8128 2.3 2.32.8.3 H “0.2 .wo0.o.:+2.00 2.: 3: its .823:
o o o o o o o o . o _ o o 0 fix 3.3 00::
8.0 +8.23 8.2% 8.0 +8.23 2.22 8.0 00.20”. 8.08180 +8.20. 8.23 E: 303 :32
. . _ , .
o WSM SM 0 SM 3.0.. 0 ”SM m SM m o ~m0.~ SM mm 83 M38
. _ .
3.00.88 34 3.3.8.0 3.4 2.2- 00.0 M04 W00.M0I8.o RM mm $29.0 8.3
3.81:.“ 24 8.8+ 3M .2; M85828 U 84 $04818 24 m: 28:8 3:5 83
£813.25 8.003 «5.8+ 8.020 8.83 8.01.383. T. fifidfmimg $.82 28 2.sz .38.: ES 3....
.0 m 0 m m M N _ N H H
o to: o 38: o .80 u 38: o 30 o 302: A u 80 0 M80:
cmHm " pew mmCMLU :Mfiw no“ .wwcmzo swam yew _wmcmfiu“ swam new HHZD ZMHH
3oz m GMHA wmmm u 3oz :maz wmmmm N awz swam mmmmw N m 3oz swam ummm_

 

 

 

 

PH gin—C32

"wuamofiuwuumacoz Mew

couuoo hfiumm can oufim: pew mwuwum uuavoum m>fiumawx a“ mowcmnu vcm woman mmmm osu wow mmusmmmx huanUMuwu

o~.o manmw

195

declined by 8.26 percent and that to a unit of operating capital fell
by 27.45 percent.

In Model C the area planted in maize declined by 70.13 percent;

2,
this decline was offset by an increase of 99.08 percent in the cotton
hectarage. Although there was no change in total cultivated land or
in the employment of owned ox-drawn equipment, employment of family
labor increased by 4.60 percent, of labor from working parties by
255.24 percent (27.51 to 45.01 man hours), and of operating capital
by 29.00 percent. The estimated net farm income came to Tshs 4632.79,
an increase of 10.00 percent from Tshs 4211.49 in the base plan. The
net farm income per hectare, per man hour of family labor, and per
unit of operating capital increased by 10.00 percent, 5.00 percent,
and 14.71 percent respectively.

In Table 6.21 the results for Alternative 1c are given. As for

the first two alternatives; the results for Models C C3 and C were

1’ 4

the same. Both cotton and maize became fairly competitive. The area
planted in cotton increased from 1.12 hectares in the base plan to 1.24
Inectares in the optimal organization. This represented a 10.31 percent
increase. The area planted in maize, on the other hand, declined by
7.94 percent. These changes left the total cultivated area unchanged.
Since sotton required more inputs per hectare than maize, employment of
labor and operating capital increased. The use of family labor in-
creased by 1.28 percent, labor from working parties by 42.20 percent,
and operating capital by 7.34 percent.

The optimum net farm income came to Tshs 5514.13 compared to 4260.83

in the base plan; this was an increase of 29.42 percent. The changes in

 

196

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

nn.o +.wo.a no.o mm.o +. o.m No.0 5N.Hm+;w~.m no.5 mm.o +.wo.a N.06 mnwh Hmufiomu wCNumumao\Hmz
m
an.na+.v¢.a nm.a me.ma+. q.H “N.a oo.om+.vm.H 0N.H sq.ma+.v¢.a n~.H mnmfi NopmA vfimnaa mo N:\Huz
.Nq.m~+.wo.o¢0m mo.omoH N¢.m~+. c.0mom mo.o~oH n¢.mm+.wo.mmam mm.HooH Nq.mm+.mo.o¢o~ mo.o~oa man mm\Hmz
cm.n +.mo.HHo ow.mom Mm.“ +. o.HHo ow.mom mo.H +.wm.moo mn.mmm qm.m +.mo.aao ow.mon man . amusamu mcfiunumao
o o o I I I I I I I I I mum ouH W Auoufimv Neuomph
M
I I I o o o I I I I I I mum #0 m Anmufimv Bomb x0
I I I I I I o 8.2 8.3 I I I fix .8 _ 2.258 8% 5
A
mm.H +.*o.mowJ q~.momm mm.H +.Wo.momm qu.mowm qm.~ + wo.moom no.mHmm mm.a +.no.mowm an.momm mum w ponmq Hmuoh
ON.N?+WH.mm . Hm.n~ om.Ns+NH.mm Hm.mm .mH.0W+mn.mH no.ma ON.N?+NH.mm Hm.nm mum W muums mcfixuoa
” H
o o o o o o o o o o o o mum H Noam; not“:
_
mN.H + o.momm mm.mnnm mN.H +.mm.o~mm mm.mnnm 50. N +. m. Nmmm o¢.oomm mN.H +.mm.cmmm mw.mnnm mu: W conmq mafiamm
. _
o _0.N SN 0 W0.N SN SN 8 m0.N SN 8 m 83 M88
0m.m_IUm . Hm.H 0m.mITm.H Hm.H Mm. NHI 0%“ qm.a cm.m Imm.~ an.H m: Awuwmxv ucmg
. n .
2.3+ 0N4 “ NH; 2. 2+ .wN. M NN.N 3.21:. so; 2.3+ NM NM; mm M .8808 .285 8.2
. _
_ . . .
Nq.o~+.wm.qamm mm. oomqmmq. om+.wm. «awn mm.oomn m0.mm+.wm.0mom mq.HHNq Nq.m~+.wm.qamm mw.oo~q mnmw m Aszv meoocH sums “ox
.r b- ..I . r s “r
. u . * _ H q u
. . _m . . _
”so to: so 300: 6: mo 80 o 300: _ 8 No 382 o 30 o 38: i
mmcmfiu“ cmam ~ p00 .mmcm5 _ mjm Now wwcmgo swam so“ .mmcmzu swam HOw HHZD _ XMHH
N m 362 ~ swam ommm. N _ 3oz cogs mmmm N “ Roz swam mmmm N 362 swam mmmm M
. . _ _ M

 

GONNQU mfiumm

(H m>wumauoua<

”mundafioauumacoz NOW

new «mam: how moofihm uosnoum m>fiumamm a“ mmwamfiu can mamam wwmm wgu New mmNSmmmx moaowofiwwm

HN.o mHAMH

197

net farm income and employment of farm resources led to increases of
29.42 percent, 13.47 percent, and 0.55 percent in net farm income per
hectare, per man hour 0f family labor, and per unit of operating capi—
tal respectively.

Contrasted to the results with Alternatives Ia and Ib, the best
results for Alternative Ic were achieved under Model C2' Although the
area planted in maize declined by 12.34 percent, that planted in cotton
increased by 17.43 percent, leading to no change in the total cultivated
land. Employment of family labor increased by 2.47 percent, working
parties by 24.15 percent, and operating capital by 1.65 percent.

Whereas the net farm income in the base plan was Tshs 4211.49,
that in the optimal organization came to Tshs 5620.52. This income

was higher than those in Models C C , and C

1’ 3 4°

in net farm income per hectare, per man hour of family labor, and per

The resulting changes

unit of operating capital were 33.45 percent, 30 percent, and 31.27
percent respectively.

These results provided more or less the same conclusions as those
drawn in the first section of this chapter. However, there is one
major difference that needs some attention. In the analyses presented
so far, there was no situation in which the employment of hired ox—
drawn equipment (and tractor power) was profitable. Unlike the parti—
cipating farmers, nonparticipating farmers owned less average land and
did not weed their farms as many times; thus demand for labor was
reduced. Consequently, labor was relatively abundant for land prepara-

tion (cultivating and ridging) as well as for other farm operations.

 

198

Marginal-Value Products of Resources Under Alternative I

 

The MVPs of resources under Alternative Ia are shown in Table 6.22.
Compared to the base plans, the MVPs of land fell in each model. Whereas

in Models C C , and C the MVPs of land fell from Tshs 465.22 to Tshs

1’ 3 4’

371.60, that in Model C2 fell from Tshs 514.67 to Tshs 401.48. Other

notable changes were in the MVPs of family labor in November, January
(for Model C2), and February for all models. In November, the MVPs of

family labor for Models C C3, and C increased from Tshs 1.81 in the

1’ 4,
base plan to Tshs 1.96 in the new plan. For the same month, the MVP of

family labor in Model C2 increased from Tshs 0.98 to Tshs 1.15. For

Model C2, the MVP of family labor in January increased from Tshs 1.06

to Tshs 1.32. Finally, the MVPs of family labor increased from Tshs

1.04 to Tshs 1.23 in Models C1’ C3, and C4

in Model C2. The increase in cotton hectarage was the cause of these

, and from 1.55 to Tshs 1.89

changes. In November, ridging of land for both cotton and maize took
place. Cotton ridges were bigger (in terms of width and height) and
required more labor to make than maize ridges. This increased demand
for labor and pushed up the MVP of family labor. In January and
February, first and second weeding took place. An hectare of cotton
required more labor than an hectare of maize, because during the first
cotton weeding, farmers also applied greater quantities of sulphate of
ammonia in addition to thinning and spraying. During the second weed—
ing, spraying added to the demand for labor. Although SA application,
thinning, and spraying were also done for maize, they were not done

to such a great extent.

 

 

 

199

 

 

 

 

 

 

 

 

A "
oo.m oo.m mo.m oo.m oo.m oo.m oo.m oo.m M mcmw Foo
Mm.MH qm.mH i Mm.ma Mm.MH Mm.mN MM.MH 0m.ma Mm.mH w mcme zone
mo.H mo.H mo.H mo.H mo.H mo.H mo.H mo.H mx «m
qm.a Mm.H qm.H 0m.H 0m.H 0m.H qm.H Mm.H wx mmh
MN.H Mo.H mN.H 00.H ow.H mm.H mm.H qo.H u: .nmm gm
0 o o o Nm.H oo.H o 0 pm .cmn Am
om.H Hw.H om.a Hm.H mH.H wo.o ow.H Hw.H u: .>oz Am
M©.H qo.a Mo.H qo.a o o Mo.H Mo.H u: .uoo an
oo.Hmm NN.moq oo.Hnm NN.moq quwoq no.0am oo.Hnm mm.moq mm mama
swam swam swam swam swam cmHm swam swam
Boz wmmm 302 wmmm 3oz omwm 3oz wwmm HHZS mom20mmm
00 Home: mu vaoz mu vaoz Ho waoz

 

 

 

 

 

 

mH o>flumauwua< ”madmaflowuumacoz How «0 cu Ho maowoz New mwousommm mo Aman Cw wm>2v muosvoum msam>lamafimumz

NN.© canoe

200

In Alternative Ib (see Table 6.23) more or less the same results as
those for Ia were obtained. The MVPs of land declined in each model,
from Tshs 465.22 in the base plans to Tshs 316.97 in the new plans for

MOdels C 03, and C4, and from Tshs 514.67 to Tshs 382.34 for Mbdel C

l’ 2.

The MVPs of family labor in November, January (for Mbdel C2), February,
and June (for Model C2) were higher than their corresponding MVPs in the
base plans. The resons for these changes were the same as for those in
Alternative Ia.

The MVPs of resources under Alternative Ic are given in Table 6.24.
The results followed the same pattern as those for Alternative Ia and Ib.
The only exception was the increase in the MVPs of land in each model.

In Models C C3, and C the MVPs of land increased from 465.22 in the

l, 4,
base plans to Tshs 483.14 in the new plans. The MVP of land in Mbdel

C2 came to Tshs 559.36 compared to Tshs 514.67 in the base plan. These
changes implied that land became more restricting. The November MVPs

of family labor increased from Tshs 1.81 to Tshs 1.92 for Models C1’

C3, and 04’ and from Tshs 0.98 to Tshs 1.06 for MOdel C And, although

2.
the MVIS of family labor in January and February increased from Tshs 1.06

to Tshs 1.39 and from Tshs 1.55 to Tshs 1.74 respectively for MOdel C2,

the increase for MOdels C1’ C3, and C4 went from Tshs 1.04 to Tshs 1.22.
Again, the reasons for these changes were the same as those discussed

for Alternative Ia.

Alternative II:

 

Increases in the Prices of Fertilizers and Insecticides

 

In Table 6.25, the changes are given for cropped areas, net farm in—

come, and resource use that resulted from substituting subsidized with

 

 

201

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

_ _ . . _ 0
n _ g i . W
_ .
oo.m so.m m so.m oo.m so.m m so.m _ oo.m _ so.m m meme _ Hoe
I _ _
. . . _ .
qm.ma am.MH m sm.ma qm.ma qm.ma l qm.ma " am.MH _ sm.MH ” mews name
i u l ”
mo.H mo.H ” mo.H mo.H mo.H W mo.H mo.a mo.H m me am
. $ _ .
«m.a . qm.H qm.H qm.H qm.H M sm.a am.H qm.H we Ame
. l I
A¢.H so.a aq.fi so.a mo.~ w mm.H sq.H so.H pm ’ .pmm as
i
o . o o o mm.H . oo.H o m 0 pm _ .cmm gm
.
I .
OH.N Hm.H OH.N H®.H 0N.H _ mm.o oH.N Hm.a , u: w .>oz as
i _ _
l i .
co.H qw.a qw.a qo.H o n o q©.H qo.H W p: " .ouo an
. .
_ _ , _ _
no.0Hm m Nm.moq ha.ofim NN.m©q sm.~wm so.qfim sm.oam mm.moq_ mm mama
__
_ . l J“
swam . swam swam swam swam I swam swam swam ” “
3oz _ mmmm M Bmz mmmm 3oz mmwm 3oz _ mmmm m HHZD M momDOmmm
I .t l . w
a . H i
so Munoz l mu Hweoz as Hmuox Q Have: [R ”
pH m>wumCHmuH< "mucmaflowupmacoz MOw «o co Ho mawwoz now Amzme :fl mm>2v muoowopm m3Hm>|Hmcwwum2

mm.o magma

 

202

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

.
.
_
oo.m oo.m oo.m oo.m oo.m oo.m oo.m oo.m mzmH M Han
a
em.MH «m.ma qm.ma qm.mH qm.mH qm.ma _ qm.mH am.mH mcmH w zomh
. . _
_ l
mo.a mo.H mo.H mo.H mo.H mo.H mo.a mo.H _ mx a 4m
. _
qm.H . qm.H am.H qm.H qm.fi . sm.4 qm.a sm.a we . awe
.
_
NN.H _ qo.H NN.H qo.H qu.H mm.H NN.H qo.H p: .nmm am
.
o m o o o mm.H ©O-H ” o o _ p: .cmm an
t . .
N¢.H Hw.H No.H Hm.a oo.H . wo.o W Hm.H Hm.H _ pm .>oz gm
. .
. _
qw.a q©.H «©.H «w.a o M o m qw.a q©.H H or “ .uoo Am
. . _ . ~ .
. .
qa.qu mm.moq qa.qu mm.moq om.mmm m mo.qflm w «H.qu mm.moqfl mm w mama
_ i L. “r _ b
u N l. n
swam swam M cmam swam swam . swam w swam swam _ M
3oz mmmm M 3oz mmmm 3oz m mmmm H Bmz A mmmm W HHZD w mommbmmm
[If m m . H m
_ .
_ ... . . *
so Have: H mu Hmwo: so ”use: m Ho Honoz m “
UH m>wumcumufi< "mucmawofluumacoz hem «0 cu Ho mampoz pow szmH cw mm>zv muosvopm msam>tamcwwumz

«N.c wHLmH

 

203

 

 

 

 

 

 

 

«m.o~I WH.n no.m qw.omI_wH.n mo.m mn.m I o.e no.n ¢N.o~I.WH.m no.m mnmh flouaamu wcfiumumoo\sz
«N.OHI VH.H mN.H q~.0HI.qH.H n~.H mw.o I H.H om.H «N.OHI.VH.H nm.a man uonmA mamas: mo Hm\Hmz
mm.o.ImH.mood mo.omoH mm.o.IuH.mooH mo.o~oa no.0.Imo.oomH mm.HooH mm.o.IUH.mooH mo.omoH mzmH mm\sz
<~.m +.mN.mwm ow.oom om.m +.m~.wwm om.mom am.o +.om.mmo mm.mom 4N.m +.mN.wmm ow.mom mLmH Hmufiamu mafiumumdo
o o o I I I I I _ I I I I mum one Anmufimv “cuomuh
I I I o o o I I I I I I my: x0 Aowhflmv ammH x0
I I I I I I o om.mn om.mn I I I mu: x0 Avmc3ov ame x0
mo.m.umo.omnm qn.momm No.N Iwo.mmmm qw.momm Hm.o.IxH.Homfl no.mHmm No.~ Imo.mmnm qw.mowm mu: uonmq Hooch
” . .
mo.mmI mm.oa Hm.nm mo.mMI.um.oH Hm.nm ma.om..wa.oa _ no.NH .mo.mMI mm.oa Hm.nm mum >uumm wcflxpoz
. . _
_
o o o o o o o o u o o o o mu: HopQH nmnwm
.
_
qn.H.Iqq.NHmm mm.mnnm qu.a.Ifiq.NHnm mm.wnmm qq.c Imd.Hoqfl oq.oomm qu.a.ucq.~anm m~.wnnm mu: popmu wfiflamm
o wo.m mo.~ o wo.m mo.m o wo.m m mo.m o mo.~ mo.~ m: was; Mmuow
~ ._
HN.m +.mm.H Hm.H Hm.m +.wm.a Hm.a mm.H +.Tm.fi ” qm.a Hm.m +.cm.a Hm.H mm Amemxv mam»
.
oq.<.uno.fi NH.H oq.q Ixo.a NH.H mm.m Imo.a w mo.H oq.q Ixo.H NH.H mm Accuuoo zfiummv Ucma
Na.o‘Imo.Hmmn mw.oomq N¢.o.Imo.HNN1 mw.oomq.no.o.Imq.mmHL mq.HHNq mm.o_Imo.HNNg mm.oomq mamH AHmzv ascocw Esme umz
a .I _ _r
fl 4 . .u I u H H
“nu HOW U Hmooz mu How mu Home: No p0m_ cu Home: 0 MOM o Honoz
mmcmnu_ swam . pOM mwomnu cmdw “cm mwcmnu swam _ no“ wmcmnu cmam new HHZD szH
N 3oz " Mam wwmm N 302 swam mmmm N 3oz swam mmmm N 3oz swam mmmm
i

 

 

 

 

 

 

 

 

 

 

 

mucmoaofiuumocoz wow mNHmz mam

mN.o manna

acuuou %Humm Mew wuuaum moaowuommcw vcw umuwawuumm cw mmwcmgu mam woman mmmm onu pow

mounmomz mocoaofiwwm

204

unsubsidized fertilizers and insecticides. The results for Models C1,

C3, and C showed that, although cotton area declined from 1.12 hectares

4
in the base plans to 1.07 hectares in the new plans, the area planted in
maize increased from 1.51 to 1.56 hectares. The total cropped area re—
mained unchanged. Operating capital increased by 3.24 percent; the use
of other resources declined. The employment of family labor and labor
from working parties declined by 1.74 percent and 39.69 percent respec—
tively; as a result, the use of labor declined by 2.02 percent.

The estimated net farm income came to Tshs 4221.65 compared to Tshs
4260.83 in the base plan, a decrease of 0.92 percent. In turn, the re-
turn to land, to unpaid labor, and to a unit of operating capital de-
clined by 0.92 percent, 10.24 percent, and 20.84 percent respectively.

In Model C2, the area planted in cotton declined by 2.75 percent,
whereas that planted in maize increased by 1.95 percent. The employment
of family labor and labor from working parties declined by 0.44 percent
and 20.13 percent respectively leading to a decline of 0.51 percent in
the use of labor. The use of operating capital increased by 6.54 per-
cent from that in the base plan.

The optimum net farm income decreased from Tshs 4211.49 in the base
plan to Tshs 4183.47 in the new plan, a decline of 0.67 percent. Con—
sequently the net farm income per hectare, for unpaid labor, and per

unit of operating capital declined by 0.67 percent, 0.83 percent, and

5.79 percent respectively.

 

205

Marginal-Value Products of Resources Under Alternative II

The MVPs of resources under Alternative II are presented in Table
6.26. The MVP of land declined in each model, from Tshs 465.22 in the

base plans to Tshs 297.53 in the new plans for Models C C3, and C

1’ 4'

For Model C the MVPs declined from Tshs 514.67 in the base plan to

2,
Tshs 318.32 in the new plan.

Unlike Alternative I, in which the MVPs of family labor in November,
January (for Model C2), and February were higher than their corresponding
MVPs in the base plans, in Alternative II they were lower. The reason
lay in the decline in the cotton area, which released much more labor
than the increase in the area planted in maize could absorb. The in—
crease in the prices of fertilizers and insecticides produced a con—
straint in the December operating capital. Although this was not the
time for spraying Thiodan, many farmers bought it in December because
distribution is very poor when the rains become heavy in January and
February.

In summary, the increase in TSP, SA, Thiodan, and DDT made cotton
less competitive by reducing its area. The impact on net farm income
was, therefore, negative since in each model the farmers' incomes were
lower than they were under the base plans. In addition, the input—
price increases constrained the December operating capital so that farmers

could not expand their production even if land and labor were not con-

straining factors.

 

206

 

 

 

 

 

 

 

 

 

 

 

 

 

 

10.0 0 1 10.0 1 0 1 00.0 1 0 1 10.0 1 0 1 0000 1 .000 00
1 1 . 1 1 _ 1

00.0 00.0 00.0 1 00.0 00.0 1 00.0 1 00.0 1 00.0 1 0001 1 100
1 1 1 1 1

.1 1 1 1 1 0

00.01 00.01 1 00.01 00.01 00.01 1 00.01 1 00.01 1 00.01 1 0001 1 7001
1 1 1 1
1 1 1 1 1 1

00.1 1 00.1 00.1 00.1 . 00.1 1 00.1 1 00.1 1 00.1 1 00 1 00
1 1 1 _ 1 1 - 1

00.1 1 00.1 00.1 00.1 1 00.1 1 00.1 1 00.1 00.1 1 00 1 001
1 1 1 1

01.0 00.1 01.0 00.1 00.1 1 00.1 01.0 00.1 1 00 1 .000 10
1 1 1
1 1

0 0 0 0 10.0 1 00.1 0 0 _ 00 1 .000 10
. 1 1 1

00.1 10.1 1 00.1 10.1 00.0 1 00.0 00.1 10.1 1 00 1 .002 10
1 1 _

1 1 . 1 1 .

00.1 00.1 1 00.1 00.1 1 0 1 0 00.1 00 1 1 00 1 000 10
1 1 1 1
1 1 .1 I 14 . 1 . 1

00.100 1 00.0 1 00.100 00.000 1 00.010 1 10.01 00.100 1 00 0001 00 1 0001
1 LI? 1 1 t 0 1 W
1 1 1 1 1 1 1 1 H
0010 . 0010 1 0010 1 0010 1 0010 . 0010 1 0010 1 0010 1 1

300 1 0000 1 302 1 0000 1 302 1 0000 1 302 1 0000 1 1120 1 00000000
1 1r 1 1 1 01 . 1 1
00 10000 N 00 10002 1 00 10000 1 10 10002 1 1

HH m>1umcumua< "manmafiowupmmcoz pow 10U 00 Ho mawpoz 00m AmLmH C0 mm>zv mauscoum msfim>IHmckumz

om.© manmy

207

Model D:

Mechanization and Manual Technologies:

 

 

Early—Planted Cotton, Maize, and Late-Planted Cotton for Nonparticipants

Late—planted cotton was included in the same way it was included
for participating farmers. The results of the analysis are shown in
Tables 6.27 and 6.28.

In Mbdels D D , and D

1’ 3 4’

from 1.25 in the bench—mark to 0.61 hectares, that planted in maize in—

the area planted in early cotton declined

creased from 1.38 to 1.89 hectares, and that under 1ate+planted cotton
increased to 0.13 hectares. There was no change in total cultivated
land. However, the use of labor and operating capital declined.
Employment of family labor, hired labor, labor from working parties,
and operating capital declined by 7.68 percent, 100 percent, 50.76 per—
cent, and 16.89 percent respectively. Although the decline in the use
of labor resulted from a decline in hectarage planted in early cotton,
the decline in operating capital was caused by the decline in hired
labor and working parties in addition to the decline in inputs such as
fertilizers and insecticides that resulted from the reduction in
cotton hectarage. The estimated net farm income Came to Tshs 4135.89
compared to Tshs 4027.18, an increase of 2.69 percent. The returns to
land, to unpaid labor, and to a unit of operating capital increased by
2.69 percent. The returns to land, to unpaid labor, and to a unit of
Operating capital increased by 2.69 percent, 10.41 percent, and 23.62

percent respectively.

 

208

 

 

 

 

 

No.m~+ Hm.n No.mN+ Nm.m NN.oq+ No.o No.mm+ Hm.m oN.q mama NmuNamo mcfiumquONHmz
Nq.oa+ No.N Nq.0N+ No.N Nw.NN+ NN.N NQ.ON+ No.N om.o asap Noam; cacowo NmNHmz
No.N + mm.NNmN @0.N + mm.~NmN qm.a + mm.qmmN mo.N + mm.~NmN «N.NmmN mama mmNsz
mm.oNI No.NNN am.cNI qo.NNN mm.NNI qo.wNo ¢m.oNI co.NNN mm.omm msma NQUNamo wcfiumumao
O O I I I I I I OWIQ mum.— URH. Avmhfimv HOuUmHH
I I o o I I I I mN.o~ my: xo Anuufixv Emma No
I I I I o oo.mN I I ow.mN mum xo Aumcsov same xo
q~.oNI mm.o~¢m qm.oNI om.o~am oo.mNI ON.QNmm qN.ONI mm.o~mm om.quq mum Nona; Nmuoa
0N.omI Nq.om 0N.omI Nq.om mm.NmI oo.o~ 0N.omI Nq.om oN.¢NN mum Nupmm chxuoz
ooNI o ooNI o ooaI o ooNIH o oo.wo my: Henna woufim
mo.N I NN.onm mo.N I NN.onm oo.oNI qm.moqm mo.N I NN.onm 0N.NaNq mum Nona; NNNamm
o mo.N o mo.N o mo.N o mo.~ mo.~ m: vcmg Nmuoy
ooN+ MN.o ooa+ mN.o ooN+ mo.o ooa+ mN.o o a: Acouuoo mumgv new;
m¢.om+ mm.N ma.om+ mm.N mm.Nm+ mm.N mo.om+ aw.N mm.N mm AmuNmzv ncmN
ON.NmI No.0 oN.NmI No.0 _ oo.NqI MN.o om.NmI No.o mN.N mm Acouuoo NNNmmV ncmq
mo.~ + om.mmNc mo.N + ow.mMqu qm.N + NN.mmoq mo.~ + am.mqu mN.NNoq mnmy AHmzv «soocN apmm umz
8255 8256 m " mafia N SEE N NEE
N o Ammo: N a game: N a ammo: N a ammo: Imozmm HHz: zmHN

 

 

 

_

 

 

 

monH<NHz<omo A<ZHHAO

 

 

 

 

muchNUNuummcoz New mnwmz Now muaum umxumzumwum nan .couuoo mama can
zauwm New moaum HMUNMMO van mwounomwx wcfiumfixm quCD mcofiumuwcmwho Hmafiuno new Musznucmm mo comfiumaaoo

mm.o wfinme

209

In Model D the results were not much different from those in

2,
D3, and D

Models D The early-planted cotton hectarage decreased

4.

from 1.25 in the bench-mark to 0.73 hectares, the area under maize

1,

increased from 1.38 to 1.82 hectares, and the area under late-planted
cotton came to 0.08 hectares. As a result of these changes in areas
cultivated, the bench-makr employment family labor, hired labor, work—
ing parties, and operating capital declined by 16.66 percent, 100
percent, 81.98 percent, and 27.83 percent respectively.

The estimated net farm income rose from Tshs 4027.18 in the bench—
mark to Tshs 4088.21, a slight increase of 1.54 percent. Consequently,
the net farm income per hectare, the return to unpaid labor, and the
return per unit of operating capital increased by 1.54 percent, 21.87

percent, and 40.77 percent respectively.

Marginal—Value Products (MVPs) of Resources Under Models D to D4

 

l

The MVPs of resources are presented in Table 6.28. Land became

more constraining in all models. In MOdels D , D , and D the MVPS

1 3 4’
of land increased from Tshs 465.22 in the base plan to Tshs 591.64 in

the new plan. In Model D the MVP of land under the new plan came to

2,
Tshs 658.13 compared to Tshs 514.67 in the base plan. As labor became

less constraining, the MVP of land went up. The decline in the MVPs

of family labor was evident in October (for Models D D3, and D4),

1,

in November, in January (for Model D2), and in February. In October,
the MVP of family labor declined from Tshs 1.64 to Tshs 1.27. In

November it declined from Tshs 1.81 to 1.56 for Models D D , and D

1’ 3 4,

and from Tshs 0.98 to 0.60 for Model D2. In February, the MVP of

 

 

210

 

 

 

 

 

I [I I'llvlliil [Ill -

 

 

 

 

 

 

 

 

 

1 1 1 1
1 1 1 1 1 1
so.m 1 so.m 1 so.m 1 so.m so.m so.m 1 so.m so.m 1 meme 1 Non
1 1 1 1 1 1
1 1 1 1 1
qm.ma 1 qm.ma 1 qm.ma 1 qm.ma qm.mH qm.mH 1 qm.ma gm.MH 1 mLmH 1 zone
1 1 1 1 1
1 1 1
mo.H mo.H 1 mo.H mo.H mo.H mo.a 1 mo.H mo.a 1 wx <m
1
1
qm.N gm.N qm.N qm.N qm.1 am.N qm.N qm.N 1 NN ems
1
1
mo.o qo.H mc.o 1 qo.H No.H mm.H mo.o qo.a 1 gm .nmm gm
1
o o o o gm.o oo.a o o 1 pm 1 .Cmm 4p
1 1 _
1 1
omefi Hw.fi ©m.a Hm.H oo.o mm.o om.H Hw.H 1 pm 1 .>02 an
1 1
1 1
mm.fl qo.H RN.H 1 qo.a o o NN.H q©.H 1 pm 1 .uoo Am
1 1
«o.amm mm.mcq qc.flmm 1 mm.moq mfi.wm@ m©.qfim qo.amm mm.m©q mm 1 Ucmg
IIIIIIIII, 1 -I;I r - 1 1 I-
1 1 1
swag swap 1 cmflc 1 exam cmam swam swam swam 1 1
362 ommm 1 Bow 1 mmwm 3oz ommm 3oz mmwm 1 21 1 mompcmmm
1 1 - - 1 1
1o None: N: 16cc: Na Nave: Na 1 1
wucmamoflup11C3: H21 q: Cu HG w o 1. p5. baa _ IJ . I . I J
. ... . ; - I . , ,1=w1r rt I It; 0.x do NmH cw mc>TV vuoswoug msflm>lamcflwpmz

Lwome

211

family labor declined from 1.04 to 0.63 Tshs for Models D D3, and D

1’ 4’

and from 1.55 to Tshs 1.02 for Model D2.

The results for Models D1 and D4, as with those for Models B1 to
B4, indicated some potential for reducing seasonal labor bottlenecks
and increasing farm income. In contrast to Models B1 to B4, however,
labor was not the only bottleneck. As the analysis showed, land was
a major constraint among the nonparticipants. Labor- and land-saving
technologies were more or less equally important issues to these
farmers.

Comparison of Optimal Organizations with Existing Resources and
Prices for Participants and Nonparticipants

 

In this section, the analyses of the results from the linear—
programming models for the participants and nonparticipants are
compared. The apparent assumption in this comparison is that parti-
cipating farmers adopted, at least partially, the recommended crop
practices such as the application of plant nutrients (organic and
inorganic), spraying of insecticides, and three to four weedings,
whereas the nonparticipating farmers did not. The assumption is based
on the following: that participating farmers used more TSP, SA and
insecticides per hectare; that they weeded their fields at least
three times; that they were closely supervised by extension workers;
and that they had more access to the recommended inputs.

A comparison of the optimal organizations of the two groups of
farmers is presented in Table 6.29. Models Al to A referred to the

4

optimal organizations for the participants; Models C1 to C4 referred to

the optimal organizations for the nonparticipants.

 

212

 

mucmafiufiuumacoz wsu mo mauve:

 

 

 

mucmawuauumm 0;» mo mamcon

n~.wo+ no.m mm.m mm.oHa+ no.m om.q mo.mm+ no.5 No.4 I mm.wo+ no.a mm.m man Hmuaamo wcfiumuwno\th
mm.mHI nN.H 0m.H mw.mHI 5N.H an.H mo.H~I 0N.H Nn.H mm.mHI NN.H om.H msmH Henna wanna: «0 u:\Hmz
an.mHI no.0NOH qq.nnwa om.an ao.o~oH nm.oowH om.mHI Nm.HooH Hm.nmwa HN.MHI mo.ouoH eq.nnma mama mm\Hmz
cm.QQI om.mom an.o~oa mm.omI om.mom oq.<oma mo.NmI mm.mmm mm.aq~a om.q¢I ow.mom «n.0NoH man Hmuaamo wcwumuumo

o o o I I I I I I I I I mum ans Auuuamv Heuumup

I I I ooHI o Nw.oc I I I I I I mum x0 Akufimv ammH x0

I I I I I I Hm.NaI oa.wn mm.mm I I I I mum x0 Anwaaov Each x0
NM.<HI qu.mowm an.~qqq mH.nHI an.momm oo.mmq¢ qm.m~I mo.mHmm wn.qwmq Nm.qal qu.mowm qw.~¢aq mum nonmq amuse
mn.mml Hm.n~ Nn.~mo mo.mml Hm.n~ Hm.nmo qm.wmI1uo.~H oo.man mn.mmI Hm.nm ~n.~mo mum muumm wcwxuoz

OOHI o mm.qoa ooaI o mm.qu ooHIW o mm.q0a OOHI o mm.<ofi mum nonmA cone:
mm.~ + mN.wNNm mm.qwom mN.N + mm.mnnm oa.mmom mn.o I1oq.oOmm am.Hon mm.~ + mm.wnnm wm.qmom mum ponmg zfiflamm
«w.oHI mo.m mo.~ mq.HHI mo.~ no.m Hm.NHI mo.m No.m qw.oaI mo.~ mm.N mm wand Hmuow
nH.oH+ Hm.H Om.H om.qa+ am.H mm.H Hq.NH+ qm.H um.H mH.oH+ Hm.H om.H mm Amuwmzv vcmA
NH.~mI NH.H mo.H NH.~MI NH.H mo.a ,qm.MMI mo.H mo.a NN.NMI NH.H me.a mm Accuuoo zapmmv ucmg
oo.m~I mm.oomq ec.mmmm HH.¢NI mm.ooN¢ mo.qum mq.oNI oq.HHNq oo.o-w mo.MNI mw.oouq oq.mmmm mzmh AHsz wEOUCH sump uwz
mwcmzu qu q< mwcwgu mo m< wwqmflu No N< wwcm:U Ho H<

N Nmuox NW8: N Hwnoz NW8: N $18: $8: N vaoz m3: ES E:

 

 

 

 

 

 

 

 

 

 

 

 

monH<NHz<omo A<ZHBAO

 

 

 

unamz VON wuaum uwxumxIwmuh can
wucmafiuauumacoz vcm qumaHuwukmm yew

acuuou maumm Mow wuaum amauauwo mzu mum mmuuaoumm wcaumfixm nu“:
mahmm w>auMucwmwunmx wnu mo mcoaumNficmwuo Hmafiuao wnu mo comfiumnfioo

0N.o

UHDMH

213

The optimum organization for farmers using only manual technologies
showed that the area planted in early cotton was 32.12 percent greater
for the participants (Model A1) than for the nonparticipants (Model Cl).

However, the area planted in maize in Model C was larger by 16.15

1

percent. The optimal use of family labor was 2.53 percent higher for

Model Cl than for Mbdel A1; in the latter model labor was not only hired

but some was obtained through working parties. Such a use of labor

reflected the difference in the optimal cropping pattern presented

above. In addition, much more operating capital was used in MOdel A1.
The net farm income in MOdel A1 was 23.06 percent higher than in

Model Cl' Consequently, the net farm income per hectare and return to

unpaid labor were also higher by 13.71 percent and 15.33 percent
respectively. Only the returns to a unit of operating capital was

higher for Model C The difference in net farm income was not only a

1°
result of differences in cropped area but also of differences in yields
per hectare. By comparing measures of efficienty, i.e., net farm in—
come per hectare, return to unpaid labor and to a unit of operating
capital, the optimum farm organization of the representative farm for
the participants was to a larger extent, more efficient. The same
conclusion was reached when a comparison was made between Models A2

and C A and C3, and A and C

2’ 3 4 4°

The MVPs of resources for the two grups of farmers are shown in

Table 6.30. The MVPs of land for the participants (Models A to A4)

1

was zero; those for the nonparticipants (Models C to C4) were positive.

1

This indicated that land was a limiting factor among the nonparticipating

farmers.

 

214

 

 

 

o mo.H o mm.H o mN.H o oo.H um .>oz HEM m3
0 nm.H o o o o o mm.H Hm .uuo BEA m3
0 mq.o 0 No.0 0 0N.o o mq.o mfimH .>02 00
o @q.o o wm.o o Hw.o o mq.o wfime .uoo oo
oo.m oo.m oo.m oo.m oo.m oo.m oo.m oo.m mSmH Han
qm.ma qm.ma qm.ma qm.ma qm.ma qm.ma qm.ma qm.mH wfimH zmmH
mo.a mo.H mo.H mo.H mo.H mo.H mo.H mo.H wM <w
qm.a qm.a qm.a qm.a qm.a em.H qm.H qm.H wM mmH
o mm.o o om.H o oH.H o mm.o Hm km: Ah
qo.H NH.H qo.H HN.N mm.H mo.N qo.H NH.H Hm .nmm Ah
0 mm.o o mH.H oo.H wN.H o mo.o Mm .cmh an
o wo.H o oq.H o oo.H o wo.H Hm .omn Ah
Hw.H mn.N Hw.H am.N wm.o mw.N Hw.H mm.N Hm .>oz Ah
qw.a mm.N qo.H o o o co.H am.~ Hm .uoo Ah
NN.m©q o NN.moq o mo.qam o Nm.moq 0 mm puma
1No q< mo m< No N< Ho H¢
AMQOZ amooz AWQOE AMQOZ Hmooz ammo: Ammo: Hmooz HHZD mUMDOmmm

 

 

 

 

 

 

 

 

 

 

mafia: How moflpm uwxpmz wmum cam coauoo kaumm How wofium Hmowwmo ms» bum mwounommm wdfiumflxm :uHB

mucmaHUfluumaaoz mam mquQHUHuHmm cwo3uwm mwuusommm mo AmsmH CH mm>zv muosvoum mDHm>IHmckumz wo nomfipmmaoo

om.o wapmH

215

In general, family labor was more restricting for the participants.
The MVPs of family labor were positive in the months of October (ex—
cept for Model A2), November, December, January, February, and May
compared to October (except for Model C2), November, January (for Model
C2) and February for the nonparticipants. There were two explanations
for this difference. One was that the participating farmers employed
more labor because they cultivated more land than the nonparticipants
did. Another was that participating farmers used more labor per hectare
during weeding because they followed the recommended practice of weeding.

Another main difference in the use of resources was shown by the
MVPS of operating capital. For the participants, operating capital was
a constraint in October and November; for the nonparticipants, opera—
ting capital was not a constraint in any month. This did not mean that
the latter group of farmers had more operating capital than the parti—
cipants. Rather participating farmers spent more cash on buying

fertilizer and insecticides and paying for labor.

sm

The results of the study indicated that both groups of family
farms had potential for achieving increases in farm incomes, resource
use, and productivity irrespective of which mechanical technologies
they used. Among the participating farmers, however, Model A2
(the employment of labor and owned ox teams) was slightly more pro—

fitable than Models A1, A3, or A Because of labor bottlenecks

4.
during land preparation, the owned ox team in Model A2 helped expand

the area planted in maize. Model A.3 (use of labor with hired ox

teams) was the next most profitable farm practice. But the shortage

 

216

of operating capital limited its application. Hired tractors were
considered to be unprofitable.

For the nonparticipants, manual technologies were the most
profitable. With an average of 2.63 hectares, neither owned ox
teams, hired ox teams, nor hired tractors were profitable. Thus,
those farmers who owned ox teams would have been better off if they
had employed labor for all farming operations. The same conclusion
applied to those who hired ox teams. For owners of ox teams, there
was an advantage of receiving income from custom work. This income,
however, was subtracted from the farmers' net incomes.

For both groups, the variations in relative product prices pro—
duced substantial changes in cropping patterns. Under Alternative“Ia,
the area planted in maize was drastically reduced for both groups; the
area planted in cotton, on the other hand, increased tremendously.
Under Alternative Ib, the increase in cotton price further reduced the
area planted in maize and increased that planted in cotton. Under
Alternative Ic, the area planted in each crop was not far from the
area in the bench—mark. The latter alternative also provided the best
results in terms of net farm income and return to land, labor, and to

a unit of operating capital.

 

The impact of changing relative product prices for cotton and maize

on maize production by participants and nonparticipants is illustrated
in figures 6.1 and 6.2 respectively. As the ratio of the price of
cotton to maize is increased, the percentage of land planted to maize

decreases. At point a, land planted to maize is 44.4 percent for

217

participants and 5718 percent for nonparticipants. If the price of
cotton is increased by 25 percent (point b), land planted to maize
declines to 42.7 percent and 52.5 percent for participants and non—
participants respectively. This is a slight change compared to the
change between Alternatives TC and la (points b and c) in which land
planted to maize drops from 42.7 percent and 52.5 percent to 20.9 perv
cent and 20.1 percent respectively. In this case, a 1 percent increase
in cotton price while keeping that for maize constant leads to a more
than 1 percent decline in land planted to maize. Apparently, the
production of maize is very sensitive to price changes. This extreme
sensitivity to price changes has two main effects. First, it creates
wide variability in the quantity of maize marketed because farmers will
satisfy family maize consumption needs before they sell. Undoubtedly,
the government's objective of food self-sufficienCy wil be undermined.
Second, farmers' net farm income would be destabilized as a result of
more land planted to cotton. As mentioned in Chapter III and in the
early part of this chapter, farmers in the area experience year-to-year
higher variability in cotton than in maize yields. Thus, any increase
in land planted to cotton at the expense of maize leads to high variability
in farmers' net earnings. As a result, farmers' ability to purchase farm
inputs such as labor, fertilizers and insecticides, and consumer goods
such as other food items not produced on the farm, is also destabilized.
From Alternatives Ia to Ib (point c to d) area planted to maize
drops from 20.9 percent to 19.2 percent and from 20.1 percent to 18.3

percent for nonparticipants. A similar slight change in land planted

218

to maize is seen between Alternative Ib (point d) and point e (when
official prices for cotton and maize are employed). Between point c
and c, area planted to maize is generally stable. The ratio of cotton
to maize prices is too high to induce farmers to produce maize for
both household consumption and for sale. Instead, they decide to pro-
duct only for household consumption. With more land devoted to

cotton production the wide variability of farmeré' earnings is also
experienced between point c and e.

It is quite evident from Figures 6.1 and 6.2 that in terms of meet—
ing the government's objectives of (1) food self—sufficiency; (2) in—
creasing cotton output for export and domestic use, and (3) improving
the standard of life for the rural population through increased and
stable incomes, the range between points a and b provide the best
guidance to price policy issues.

The impact of increased fertilizer and insecticide prices on net
farm income, cropping patterns, and resource use was also more or less
the same for both groups. Net farm incomes were reduced in both groups
and for all models. Although there was a reduction in the area planted
in cotton and maize for the participants, the area planted in maize was
increased at the expense of the cotton area for the nonparticipants.
Maize demanded less inputs (TSP, SA, DDT) per hectare than did cotton
for the latter group.

An increase in the October and November operating capital for par—
ticipants decreased cotton hectarage and increased the area planted in
maize in each model. This change, however, led to an increase in net

farm income.

219

 

 

 

Percentage 100 "
of land
planted to
maize.

80 .

60 q

a
40.1 b
20 -
c
e
0 A r .
1 2 3 Cotton Price

Maize Price

FIGURE 6.1 The Impact of Relative Product Prices on Maize Production
for Participants

3 . . . . .
Free market price for maize and official price for cotton.
b
Alternative Ic
C 0
Alternative Ia
d .
Alternative Ib

e . . . .
Off1c1al prices for cotton and maize.

220

 

 

Percentage lOOII
of land
planted to
maize
80%
60'
a
b
40‘
20‘ c d
0 v I c
l ' 2 3

Cotton Price
Maize Price

 

FIGURE 6.2 The Impact of Relative Product Prices on Maize Production
for Nonparticipants

 

a O O O O C

Free market price for maize and offiCial price for cotton.
b .

Alternative Ic

c .

Alternative Ia

d .

Alternative Ib

e . . . .
OffiCial prices for cotton and maize.

221

The impact of spacing out farm operations between maize and cotton
on net farm income, cropping patterns, resource use, and productivity
was also analyzed. This was done as a way of alleviating labor bottle-
necks that were so apparent in the optimal solutions. For participating
farmers, both maize and cotton output would increase as a result of an
increase in the area planted in both crops and despite a 20 percent de—
cline in cotton yield per hectare associated with late—planted cotton.
Higher yields per hectare for cotton could be increased if research at
Ukiliguru could produce a drought—resistant cottonseed.

The results for nonparticipants, however, were rather different.
Maize output would increase tremendously due to a substantial increase
in maize hectarage; cotton would decline because the area decreased and
an hectare of late-planted cotton would produce 20 percent less than an
hectare of early—planted cotton. Land shortage was also a factor here.
Under existing resources and prices, one crop could be expanded only at
the expense of the other.

Net farm income would increase for the participants as a result of
changes described above. Increased cropping area for the participants
would offset any possible minimization of labor bottlenecks; notable
changes would be in March and April, when family labor would be scarce.
Labor bottlenecks would occur during the same months as those under the
base plans for the nonparticipants. If late—planted cotton were included,

the scarcity of labor could be minimized.

 

222
CHAPTER VII

SUMMARY, POLICY ISSUES, LIMITATIONS OF THE STUDY, AND SUGGESTIONS
FOR FURTHER RESEARCH

Summary

The general objective of this study is to analyze empirically the
impact of the government's measures to increase agricultural output on
farm income, cropping patterns (including output), and resource pro-
ductivity on small-holder family farms in Geita district of Tanzania.
Specifically, the objectives of the study were (1) to evaluate the im—

pact of the constraints imposed by farm resources on the production of

 

 

cotton and maize; (2) to assess the implications of oxen and tractor
technological choices with respect to their impact on net return per
hectare, on employment, and on labor productivity; (3) to analyze the
impact of rescheduling agricultural production (by including early—
and late-planted cotton) on resource allocation, agricultural output,
and farm earning; and (4) to test the sensitivity of alternative input
and output prices on resource allocation, enterprise combinations and
net farm income.

Cotton and maize production is influenced by the overall national
and locational problems facing the Tanzanian agricultural sector. The
national problems range from poor marketing and pricing policies to
insufficient research studies, insufficient roads, lack of road main—
tenance, insufficient low-cost transportation facilities, unclear bureau—
cratic structure, poor training, and lack of trained planners. The
locational problems of increasing agricultural output and agricultural

productivity include land scarcity, drought conditions, and labor shortages.

223

Solutions to such problems require specific studies as a basis for
formulating specific policies. So far, little attention has been
paid to these areal problems and solutions.

Instead, different measures have been taken in response to the
falling agricultural output. The following few examples reflect Some
of the alternatives attempted. First, since 1974, a single pan—
territorial producer price for each crop has been fixed annually by
the economic committee of the cabinet. Differentials in the uniform
price are fixed for different quality grades where applicable, but
no distinction is made with respect to location or transport costs.
Second, each year the Revolutionary Party (the only political party
in the country) directs regions to meet crop targets set by the national
party leaders. These targets are often arbitrary and are not accom—
panied by any incentive for being met or punishment for not being met.
Third, in 1975 an agreement was made between the Tanzanian government
and the World Bank in which the latter would provide funds (on a loan
basis) for improving agricultural practices in the major cotton— and
tobacco-growing areas. The main purpose of this agreement was to intro-
duce biological innovations that involve the use of improved seeds,
pesticides, and fertilizers, both organic and inorganic.

The production of cotton and maize in Tanzania is a complex process.
These two crops play a very important role in the Tanzanian economy.
Cotton accounts for more than 17 percent of Tanzania's total exports
which in turn make up for about 30 percent of the country's GDP. The
crop is a source of foreign exchange that the country needs badly to

pay for imports. In addition, more cotton is needed each year to meet

 

224

its demand by expanding domestic textile industry. Maize, on the other
hand, is the Staple food in Tanzania. If the country cannot produce
enough maize to meet domestic demand, the government has to spend some
of its scarce foreign exchange to import it. Unfortunately, domestic
demand for maize has often exceeded domestic supply.

The measures taken to increase the production of such important
crops as cotton and maize have not been based on an understanding of
the complex interactions between resource allocation and risk minimi-
zation that frequently influence the adoption of new innovations among
small—holder family farms. Such complex interactions become more evi—
dent in areas in which main food crops compete with main export crops
for the resources available to the family-farm households. Information
on the utilization of land, labor, and implements in particular environ—
mental and cultural practices is needed by policy makers. Unfortunately,
such information is hardly available. In most studies (examples given in
Chapter II) on agricultural production in small-farm economies little
effort is made to analyze more than one crop. This is especially true
in situations where crops are competitive with each other. Often the
whole problem is seen from the macro point of View and is studied by
using secondary and aggregated data. It was urged that any policy that
is intended to achieve national targets in agricultural output and pro-
ductivity must address itself to understanding the behavioral character—
istics of farmers.

Geita district, in MWanza region of Tanzania was chosen as the area
for the study because it is the most important cotton producing area and

one of the main areas for maize production in Tanzania. In Geita district,

 

225

cotton and maize compete for farm resources, especially labor during
land preparation, planting, weeding, and harvesting.

The Geita district, is an area which receives enough rainfall for
cotton, maize, and other food crops such as cassava, legumes and
vegetables. Most of the soils are also suitable for these crops.

The district is estimated to have 909,000 hectares of cultivable land.
About half a million people live in the district, with an average house—
hold size of 7.1 persons. Cotton is the main cash crop while maize is
produced both for household consumption and for sale. Production of
maize for consumption is given first priority by farmers. All farmers
surveyed practice ridge cultivation which is regarded as beneficial to
crop growth and erosion control. However, it is a practice that re-
quires a lot of labor to operate.

Generally, the farming season begins in August with land clearance
and ends in June with harvesting and sorting of cotton into clean and
dirty piles. Most cotton and maize is grown in block farms. Farmers
included in the Geita Cotton World Bank project (participants) practice
improved farming by following recommended practices such as employing
a good weeding program, use of fertilizers, and spraying crops with in—
secticides. Nonparticipant farmers seldom follow such practices.

The research methodology used in this study for both the partici—
pants and nonparticipants in the Geita Cotton area was static linear
programming and parametric linear programming. These techniques were
used to determine the organization that would maximize net farm income
under existing resources, prices, and technology; varying relative

products, and input prices, as well as varying levels of operating

 

226

capital when it was necessary. Linear programming was used in the
study because (1) a large number of interrelated variables can be
handled, and thus family-farm systems that are characterized by a
high degree of interdependence between production and consumption,
consumption and investment, investment and resource availability,

and social and cultural constraints; (2) the maximum possible profit
for a farm-planning problem is guaranteed; and (3) it is easy to vary
available prices and resources as well as input coefficients.

The sample in the study consisted of 80 farmers, 40 from parti—
cipants and 40 from nonparticipants. This stratification was justified
by the objectives of this study. For each group, a two—stage sampling
method was employed. The first was to select villages in such a way
that the final sample included farmers who used either manual, oxen or
tractor technologies. The second stage dealt with the selection of
farm households; this was done randomly from four prepared lists:
households that had decided to employ labor only; households that owned
plows; those who intended to hire plows; and the last list contained
households that intended to hire tractor services. The needed data
on production, prices, and resources availability and uses were collected
through questionnaires. Data for only one cropping season, 1979—80 were
collected by six well trained enumerators. In order to estimate optimum
plans, data from the farms were averaged and assumed to form a repre—
sentative farm for that group.

The structure of the linear—programming model is tailored to fit
the unique aspects of the study. The objective function to be maxi-
mized was net farm income subject to meeting the minimum maize-

consumption requirements of the farm household. The construction of

 

227

linear programming models included crop—production activites, labor—
hiring activities, working party activities, animal-power owning
activities, animal—power hiring activities, tractor—hiring activities,
fertilizer- and insecticide-buying activities, crop—consumption acti—
vity, crop—selling activities, animal—power selling activities, and
transfer activities. The last group of activities was included to
transfer surplus operating capital from one month to another.

The results obtained from the linear programming analysis led to
the following conclusions. First, there is a need to remove labor
bottlenecks if production is to be increased. These bottlenecks can
be removed in four ways: (1) by increasing the intensity of the family—
labor inputs; (2) by hiring additional labor if it is available in the
area for the peak periods at the going wage rate or by organizing-
working parties; (3) by mechanizing through switching to animal draft
power or tractor cultivation; or (4) by changing the cropping calendar
so that seasonally required labor resources do not exceed seasonally
available labor resources.

Farmers will increase the intensity of the family labor input only
when it is economically beneficial to them. Input-output pricing
policies and their intended purpose must be quite beneficial before farm
households intensify the use of family labor. Hiring of additional labor
is constrained by two factors. First, it is politically discouraged by
the government's policy of building a socialist society in which no
individual would work for another individual for money. Second, as a
result of the above factor, there is no guarantee that such a type of
labor would be available in the future. If this type of labor remains

available, the difference between demand and supply for hired labor will

 

228

increase the wage rate. Thus, more operating capital will be demanded.
Working parties seem to be a better alternative as a source of labor;
however, since most of this labor comes from other farm households,
there is no assurance of getting it since other farm households are
often also experiencing labor bottlenecks. Exchange labor would have
been the best alternative and it is encouraged by the government.
However, the limiting factor is farmers' suspicion 6f this system since
it is always referred to as a step towards communal (socialist) farming.
The third possible way of removing labor bottlenecks is through
mechanization, either by switching to animal draft power or tractor
cultivation. The analysis has shown that under optimal organization,
tractor power is not profitable; thus, it should not be used. It
should be remembered that the use of tractors in the study area is
confined to land cultivation. Although this reduces the labor require—
ment per hectare for this farm operation, it does not remove labor
bottlenecks during other operations. This is not the only problem.
Tractors are not manufactured in the country. To import them, the
government has to spend its meager foreign exchange reserves. It is
also worthwhile to note here that, whereas a tractor would not only
require increased amounts of foreign exchange for purchasing fuel to
run it, the price of oil on the world market has increased tremendously
over the last eight years, and signs are that it will continue to
spiral. The cost of hiring tractors would then be too high for small
farmers to afford unless the government subsidizes them. For these
reasons the government has correctly discouraged small farmers from

using tractors. Instead, emphasis has been put on using ox plows.

 

229

There are four potential benefits attributed to using animal draft
power. First, it allows the expansion of hectarage by reducing labor
time required per hectare. Second, it leads to higher yields, which
result in the short run from better and more timely performance due to
the use of manure and crop residues. Third, labor time saved may be de—
voted to other activities of value to the farm household. Fourth, crop
removal and marketing can be facilitated by the use of animal drawn
carts and can thus provide a source of income from custom transport
where the demand for that service exists.

The results obtained from the analysis, however, indicate that
the economic benefits of animal draft power not only depend on the
extent of adoption but also on the intensity of its use. There are
critical labor bottlenecks during ridging, weeding, and harvesting for
most farmers. Given this situation, there is a need to adopt animal
draft power for those operations. This type of power is particularly
important for the participants, who face more serious labor bottlenecks
during those operations; it should also be introduced to the nonparti—
cipants as larger areas of land become available for cultivation.

In order to fully adopt the use of animal draft power, the govern—
ment, through extension agencies, must train household farmers about
how to use animal-drawn implements for the various farm operations, as
well as the intensification of land use and maintenance of soil fertility.
Not only that, but a credit system must be available so that farmers can
borrow enough funds to purchase the required animal—drawn implements.
The present institutions for inputs supply, repair and maintenance,
animal health services, extension services, and marketing must be im-

proved in order to meet the new demands for these inputs.

 

230

The early experiments with mechanization called for too many
changes too quickly. This led to high costs, poor management of
machinery, decline in soil fertility, and the variability in yields.
Based on these bad experiences,farmersshould be encouraged to adopt
new technology, including the introduction of animal power, through a
gradual process. For example, they should not be asked to undertake
animal weeding until they have the trained animals and have learned
how to weed without damaging crops. Step—by—step adoption of the tech—
nology may allow the farmer to move to higher technology farming
methods and also keep his debt-service obligations within his ability
to pay.

Equally important is the need to conduct research on the biologi-
cal and mechanical aspects of animal-drawn equipment so that it can be
adapted to diverse local conditions. Designs for plowing and weeding
equipment should be developed for the different agronomic conditions
found in different regions and districts.

labor bottlenecks can also be removed if the cropping calendar is
changed. Instead of maize and cotton competing for labor during the
plowing, weeding, and harvesting operations, cotton could be planted
later. The results indicated that this farming system is possible.

For the average participating farmer, maize output could increase by
about 9.94 percent while cotton could decline by 27 percent due to the
20 percent decrease in yield per hectare that resulted from late—planted
cotton. Net farm income, however, could increase by about 10.2 percent.
For an average nonparticipating farmer, cotton output could decrease by
as much as 37.2 percent due to a reduction in the area planted in the

crop as well as a 20 percent decline in yield per hectare. Maize output

 

231

could increase by 34.99 percent as a result of a 35 percent increase in
cropping area. The change in farm income could increase by 2.4 percent.
Unlike an average participating farmer, a representative nonparticipa-
ting farmer is already facing land constraints. Thus, it is impossible
to increase the area of one crop without decreasing that of the other.
By changing the cropping calendar, maize seems to be more profitable to
these farmers than cotton.

The adoption of this new farming system, however, is quite risky.
It depends on the weather, particularly rain. If the rainy season is
short, i.e. it ends in late February, late—planted cotton may not sur-
vive, and thus cotton output might fall drastically.

Another major finding of this study was that the official price
of maize is significantly below the free—market price. Therefore,
farmers are more likely to sell outside the officially approved market—
ing channels, in order to increase the profitability of the farm.

This research project also addressed the issue of input subsidies.
If these are removed, the improved farming practices currently being
followed by participants provide larger declines in gross margins
than the unimproved practices currently being followed by the non—
participants, whether output of maize is valued at the official (lower)
price or at the free-market (higher) price. This suggests very clearly
the role of input subsidies as presently used in the system. Improved
practices provide higher gross margins over unimproved practices and
hence an incentive to adopt them but only when modern inputs are
highly subsidized. Thus, following improved practices tends to increase

the level of physical output per hectare because the relative cost and

232

return relationship has been altered by the subsidy program. This is
a direct income transfer to the farm sector that is being paid by
other sectors in the economic system.

Thus, the present policies that subsidized modern inputs,
coupled withtflmacontrol of official market channels (output prices
are fixed below the free—market price) offset each other to a large
extent. The subsidized inputs are considered incentives to adopt new
practices. From the farmer's viewpoint, what the government gives in
input subsidies, it takes away in lower fixed prices. In addition,
there is the added burden of administering and policing both input and
output markets, a burden that has to be born by other sectors of the

economy.

Policy Issues

 

On the assumption that the data, the analytical framework, and the
unit of analysis all have a reasonable degree of validity, the quanti-
tative results obtained from the study could provide relevant insights
and guidelines that would aid policy makers and agricultural researchers.
Some policy implications of the results obtained are presented here.

In the linear programming analysis, it was shown that small—holder
family farms from both groups experience labor constraints during land
tilling and ridging, weeding, and harvesting. The present use of trac-
tors or ox plows does not solve seasonal labor bottlenecks because
their use is confined to land tilling. Labor bottlenecks were mentioned
by farmers in the area as one of the major limiting factors in increas—

ing agricultural output. It follows, therefore, that any policy aimed

 

233

at increasing“ agricultural output through expanding crop hectarage or
even through intensification must be supported by labor—saving techniques.
There may be a need to introduce new crops or crop mixtures that can in—
crease productivity through flexibility in the timing of farm operations.
There is also the possibility of complementing ox plows with a wide

scope of low—cost implements such as ox—drawnmulticultivatorsand seeders.
Demand for such relatively simple farm implements can be substantial,
especially when the introduction of such implements is accompanied by
other yield—increasing technologies.

The potential demand for such implements needs to be supported by
growth in the small-industry and service sectors of the rural or national
economy. The promotion of draft equipment is, however, limited by its
cost, which, though lower than motorized equipment on a per unit basis,
still represents a considerable investment for small farmers. This calls
for a formulation of credit policy based on the productivity of capital.
This policy will place additional demand upon the Tanzania Rural Devel—
opment Bank, and regional and district cooperatives as the source of
short—term credit to farmers. Farmer—owned cooperatives may be more
important in filling this role because they can more easily involve a
large number of small—holder farms. Also, they may be in a better posi-
tion to collect on the repayment of the loans. The success of such a
policy, however, will also depend on yield—increasing technologies and
higher ouput prices.

The results of the study also indicated that farm income, output,
resource use and productivity are influenced by product prices and

inPut prices to a much greater extent than they are by operating capital.

 

234

However, if the government removes the subsidies on inputs without
increasing product prices and also not provide a source of credit to
meet the high inputs costs, then operating capital is going to be a
limiting factor. This emphasizes the complementarity of credit ser-
vices and new techniques and suggests that credit should be made an
important component of any new technology package.

Another policy issue that needs a careful and immediate review
by the government is the producer price in particular and the market-
ing system in general. The present price policies have proved self—
defeating. Lower and controlled prices for food crops have not succeeded
in securing adequate supply of food stuffs. It also has increased
farmers' risks in marketing food surpluses outside the official market.
As a result, over time it has destabilized food supplies. Further, the
present single national producer price for each crop, which is intended
to insure that small farmers in remote areas are not disadvantaged on
the grounds of location, involves costs that have to be recognized and
taken into account when the trade—off between income and resource-
allocation objectives is examined. Often these producer prices give
wrong signs to farmers. Thus, in areas where maize cannot grow well
because of weather conditions, a higher price of maize, compared to
other food crops, tends to influence farmers in such areas in allocating
more resources to maize than to other food crops. In areas like Geita,
where maize and cotton are favored by weather, relative prices of these
crops play a very important role in influencing farmers' decision—

making processes.

 

235

Currently, the price for maize is too low compared to what farmers
can get in the black market. Given the importance of such areas in
producing both export crops and food crops, it may be necessary to
replace the single pan—territorial pricing system with a locational
pricing policy that could influence farmers indifferent areas of the
country to make use of their comparative advantages. This change
may have to be accompanied by a change in the present price-controlled
system: the more crops that are brought into the price-control system,
the more difficult it becomes to get their relative prices correct.

It would be much easier to arrive at correct relative prices if prices
of minor crops such as legumes, cassave, and potatoes could be deter-
mined by local markets.

The use of fertilizers and insecticides appears to have some posi-
tive impact on yields. The comparison in yields per hectare between the
participants and nonparticipants is an indication of that impact since
the participant farmers used more of the chemical nutrients than the
nonparticipants. However, nonavailability, limited availability, and
late arrival of these inputs caused many farmers to limit their use of
these resources. Often it is the poor road system and lack of trans—
port to bring the inputs to the farmers that cause those problems.
Investment in all weather feeder roads will enable fertilizer to be
delivered in the rural areas even during the rainy season when some of
these roads are now impassable. In addition, a credit system should be
available to local cooperatives to construct permanent storage facil-'

ities. The proximity of available inputs will reduce the already

 

 

 

236

substantial transport costs. Investments in such infrstructures will
definitely have high payoff in the long run.

It is important to stress here that the policy suggestions made
here are not mutually exclusive. They need to be discussed and re—
viewed as complementary since their simultaneous application will

result in greater impact.

Limitations of the Study and Suggestions for Further Research

The scope of the application of the results of the study is limited
by the reliance on one year's data and by the fact that the consequences
of variations in input—output coefficients were not analyzed. Thus, the
results of the study need to be complemented with the results of similar
studies and personal experiences from other areas and for different years
in order to obtain a comprehensive picture that may help policy makers.

Further, linear programming estimates are limited in that they are
generated in the context of assumptions and model specifications under—
lying the linear model. The closer these assumptions and model speci—
fications approach an accurate reflection of the decision environment
for small—family farms in the study area, the more valid the results
are likely to be. In the study, it was assumed that small-family
farmers in the study area had as their objective maximization of net
farm income subject to meeting minimum food requirements. However,
farmers are more dynamic, and this assumption may not hold. To the
extent that the assumption fails, farmers' actual decisions may differ
significantly from those indicated as optimum by the results of this

study.

 

. ...--

 

 

237

Apparently, the approach adopted in this study provides only
partial—equilibrium solutions. It is most unlikely that in the long
run the price of maize for instance, can go up if that for cotton re-
mains constant. Further the exclusion of other food crops such as
cassavas and legumes has obviously affected the results obtained in
this study; this is particularly so since the price for maize was
determined in a free (black) market for food crops.

In addition, the lack of risk and uncertainty considerations is
another characteristic of the static economic assumption under which
the linear programming models were constructed. However, small
family farmers will be faced by uncertainties regarding changing;
economic and political institutional elements, technologies and so on
in the decision—making context. The results might be quite different
from those obtained in this study.

A comprehensive nationa].policy on agricultural production
cannot be based on such isolated studies. There is a need for similar
research in other areas of the country where similar problems are
found. The most well—known example is Tabora region where tobacco
competes with maize for resources.

There is also a need for research on the impact of input subsi—
dies vis—a—vis higher product prices. Currently, there are no empir—
ical studies on these issues that can help policy makers who often have
to choose between input subsidies and product—price supports.

During the last decade, Tanzania has spent quite a substantial
amount of scarce resources on input subsidies without adequate knowledge

of the payoffs or the distributive effects. Given the government's

238

concern with equity considerations, research is required an the impact
of past expenditures on agricultural output and their income-distribution
effects.

Research is also needed to determine the impact of High levels of
fertilization on agricultural output, farm income, and the acidity of
the soils. This need arises because of the unknown impact of fertili—
zation. Similar research is needed to show the impact of insecticides
on output and farm income. These studies should be as locationally
specific as costs and manpower constraints can allow.

The absence of detailed recommendations on agricultural practices
for maize (see Chapter III) call for more research on cultivation
practices, input uses, and so on. Such research may provide valuable
information that can be used in resources allocation decisions.

Further, much research has to be done in an attempt to develop
cotton seeds that may endure a relatively dry or short rainy season.
This might turn out to be very costly, and it will take a long time
before farmers fully adopt it. Yet it is a possibility that should be
explored along with the others that are discussed above.

The long-term solution to.the labor problem that was discussed in
this study requires comprehensive research on mechanization, especially
on animal—drawn equipment. The research should cover the issues of the
adaptability of the technology to diverse local conditions; the required
supportive services, such as extension services, supply of inputs,
credit, repair, and maintenance; the impact on desired cropping
patterns, production, resource use, and their financial returns; and,

above all the nature of interaction between mechanical technology and

239

labor bottlenecks on the one hand, and between mechanical technology and

biological innovations on the other

APPENDICES

240

APPENDIX A

 

THE STRUCTURE OF GDP (FACTOR COST) IN 1966 PRICES

 

1965-67 1971-73 1973—75
Agriculture 44.5 39.7 38.1
Mining and Quarrying 2.8 1.2 0.7
Manufacturing 8.1 10.0 9.9
Trade 12.3 12.1 12.0
Transport 7.4 10.1 10.3
Finance 10.2 9.9 9.9
Public Administration 10.9 12.6 14.6
Other Services 3.8 4.4 4.5

 

Total 100.0 100.0 100.0

 

241

APPENDIX B

NUMBER OF TRACTORS AND OX PLOWS IN GEITA DISTRICT

 

 

 

OX
YEAR TRACTORSa PLOWS
1978 24 290
1979 1 333
TOTAL 25 623
Source: Geita District Annual Report, 1979

aAs owned by the Geita Cotton Project.

(Unpublished).

 

242
BIBLIOGRAPHY
Amarshi, A. K. "The Tractor Fiasco in Tanzania," Agricole, Vol. No. 2,
1969.
Andrews, M. S. and John, R. M. "An Integrated Production - Consumption

Model for Dominican Republic," Agricultural Experiment Station,
University of Maryland, College Park, Maryland, 1976.

Anthony, Keneth R. M., and Uchendu, Victor. "A Field Study of Agricul—
tural Change: Geita District, Tanzania," Stanford, California:
Food Research Institute, Stanford University, 1969 (mimeographed).

Atta Konadu, Y. K. "Economic Optima in Resource Allocation for Small—
holder Subsistance Farming in Ghana," unpubliShed Ph.D. Thesis,
Department of Agricultural Economics, Michigan State University,
1974.

Barnard C. S. and Nix, J. S. Farm Planning and Control, Cambridge
University Press, Iondon, 1976.

 

Beeney, J. M. "UNDP Report to the United Republic of Tanzania on
Agricultural Mechanization," FAQ, Rome, 1975.

Beneke, R. R. and Winterboer, R. Linear Programming Applications to
Farm Planning, Iowa State University, 1971.

 

Buckwell, A, E. and Hazel, P. B. R. "Implication of Aggregate Bias
for the Construction of Static and Dynamic Linear Programming
Supply Models," Journal of Agricultural Economics, Volume 23, 1972,
pp. 119—134.

"Bureau Calculates Growth Rate," Tanzania Daily News, 21 November 1979,
p. 1.

Bureau of Research and Land Use Planning. "The Mechanized Block
Schemes in Mwanza Region, 1964—69," University of Dar—es Salaam,
Tanzania, 1969.

Bureau of Statistics. "Results of the 1978 Census," Dar-es—Salaam,
Tanzania, 1980.

Carr, S. "The Use of Insecticides on Cotton in Mwanza and Shinyanga,"
World Bank Project Preparation Report, Annex V, Appendix 16 and 17,
Mwanza, Tanzania, July 1976, (mimeographed).

243

. "The Use of Inorganic Fertilizers on Maize in MWanza—Shinyanga
Regions," WOrld Bank Project Report, Annex V, Appendix 15, MWanza,
Tanzania, July 1976, (mimeographed).

Carter, Harold 0. "Representative Fareruide for Decision Making,"
Journal of Farm Economics, Vol. 45, 1963, pp. 1458-1468.

 

Clayton, E. S. "Mechanization and Employment in East African Agriculture,"
International Labor Review, 1972, p..27.

 

Collinson, M. P. "Farm Management Survey Report No. 4: Lwenge Primary
Society Area, Geita District, MWanza Region, Nov. 1963 to October
1964 Season," Ukiriguru, Tanzania: Kilimo Research and Training
Institute, 1964, (mimeographed).

. Farm Management in Peasant Agriculture: A Handbook of Rural
Development Planning_in Africa. Praeger, New York, 1972.

 

 

. "Some Guidelines for Farm Economics Data Collection in Rural
Areas of Africa," ADC, Beirut, 1974.

"Cotton Development in Tanzania: A Review of the Cotton
Program in Sukumaland," Employment and Rural Division, Study No. 13,
World Bank, 1974.

Cotton Research Corporation. Progress Reports from Experiment Stations,
Season 1967-68, Tanzania Western Cotton Growing Area, (London), 1969.

 

 

Dent, J. B. and Anderson, J. R. Systems Analysis in Agricultural
Management, John Wiley and Sons, Australasia Pty. Ltd., Sydney, 1971.

 

 

DeVries, J. and Fortman, L. "A Study of Ujamaa Villages in Iringa Region,
Tanzania," Department of Rural Economy and Extension, University of
Dar—es-Salaam, 1974.

DeWilde, J. C. Experiences with Agricultural Development in Tropical
Africa, Volume II, The Case Studies, Baltimore, The John Hopkins
Press, 1967.

 

 

Ellis, F. "Agricultural Pricing Policy in Tanzania 1970-79: Implications
for Agricultural Output, Rural Incomes and Crop Marketing Costs,"
Economic Research Bureau, University of Dar-es—Salaam, Tanzania,

July 1980.

Empire Cotton Growing Corporation. Progress Report from Experiment
Stations's Season 1961—62, Tanganyika Lake Region, (London), 1963.

 

 

FAO. "Assistance in Agricultural Mechanization, Tanzania," (Mission

Report, September-October), FAO, Rome, 1974.

Fiedrich, K. H. Farm Management Data Collection and Analysis. FAO,
Agricultural Services Bulletin No. 34, FAQ, Rome, 1977, pp. 70—81.

 

 

244

Geita Cotton Project. "ProjeCt Evaluation Report for 1978-79,":l979.
. "Project Evaluation Report for 1979-80," 1980.

Government of Tanzania. "Report of WOrking Party No. 12 on Agricultural
Mechanization: Second Five Year Plan, Appedix IV, 1968.

 

Heady, E. O. and Candler, W. Linear Programming Methods, Iowa State
University lress, Ames, 1958.

 

Heijen, J. D. The Mechanized Block Cultivation Schemes in MWanza Region
1964-69, Research Paper No. 9. Bureau of Resource Use and land Use
Planning, University of Dar-es—Salaam, Tanzania, 1969.

Herbst, J. H. Farm Management Principles, Budgets, Plans, Stipes Publish—
ing Company, 1974.

 

Heyer, J. "Linear Programming Analysis of Constraints on Peasant Farms
in Kenya," Food Research Institute Studies, Volume X, No. 1, Stanford
University, 1971.

 

. "An Analysis of Peasant Farm Production Under Conditions of
Uncertainty," JAE 23, 1972, pp. 135—146.

Hulls, R. H. "The Assessment of Agricultural Extension in Sukumaland,
Western Tanzania," Economic Research Bureau Paper 71—7, University
of Dar-es-Salaam, Tanzania, 1972.

International Labor Organization. "Jobs and Skills Programme for Africa,"
Towards Self—Reliance: Development, Employment, and Equity Issues in
Tanzania, Addis Ababa, ILO, 1978.

Kennedy, J. O. S. and Fransisco, E. M. "On the Formulation of Risk Con-
straints of Linear Programming." American Journal of Agricultural
Economics, Volume 25, No. 2, 1974.

Keregero, K. J. B. "Investigations into the Effectiveness and Impact of
Improved Methods on Smallholder Cotton Production in Mara Region,"
University of Dar—es—Salaam, Morogoro, Tanzania, 1975.

Keregero, K. J. B., DeVries, J., and Bartlett, C. D. S. "Farmer
Resistance to Extension Advice: Who is to blame? A Case Study of
Cotton Production in Mara Region, Tanzania," Rural Economic Review
Paper No. 5, University of Dar—es—Salaam, Morogoro, Tanzania, 1976.

Klien, C. K., Green, D. A. G.,Donahue, R. L., and Stout, B. A. Agricul-
tural Mechanization in Equatorial Africa. College of Agricultural
and Natural Resources, Research Report No. 6, Michigan State Univer-
sity, East Lansing, 1969.

245

Kriesel, H. C., Laurent, C. K., Halpern, C., and Larzelere, H. E.
'ngicultural Marketing in Tanzania, Background Research and
Policy Proposals,"MSU, East Lansing, 1970.

Larsen, Arne. “Variations in Income Among Farming Areas in Sukumaland
and Related-Policy Implications." Economic Research Bureau paper
No. 70.6, University of Dar—es—Salaam, Tanzania, 1970 (mimeographed).

LeMare, P. H. "Research on the Use of Fertilizer in Relation to Cotton
Growing," Ukiriguru Research Notes, Tanzania, Report No. 8, 1967.

"Experiments on the Use of Fertilizer for Cotton in Western
Tanzania," Ukiriguru Research and Training Institute, Progress
Report 14, 1969.

Lewis, W. A. Development Planning, Harper and Row, New York, 1966.

Lord, R. F. Economic Aspects of Mechanized Farming in Nachipgwea,
Tanzania, London, Her Majesty's Stationary office, 1963.

Low, A. R. C. "Decision Taking Under Uncertainty: A Linear Programming
Model of Peasant Farmer Behaviour,"JAE 25, 1974, pp. 311—321.

"Small Farmer Improvement Strategies — The Implications of a
Computer Simulation Study of Indigenous Farming in South East
Ghana," Oxford Agrarian Studies 4, 1975, pp. l-l7.

___ "Linear Programming and the Study of Peasant Farming Situations,
A Reply,"JAE 29, 1978, pp. 189-190.

Malcolm, D.W. Sukumaland, London: Oxford University' Bess, 1953.

Malecela, J. (former Minister of Agriculture). "Agricultural Policy
Towards Socialist Rural Development in Tanzania," A speech made at
the University of Dar—es—Salaam, 1980.

Malima, K. A. "The Economics of Cotton Production in Tanzania: An
Examination of the Factors that Influence Agricultural Development,"
Ph.D. Thesis, Princeton, New Jersey, 1971.

Manetsch, T. J. A Generalized Simulation Approadfi'ko Agricultural
Sector Analysis with Special Reference to Nigeria, produced by the
MSU Simulation team under USAID. Contract AID/csd—1557, East Lansing,
1971.

 

 

McCarl, B. A. "Indonesian Supply Prospects for Food Crops, A Farm Level
Linear Programming Analysis of Dryland and Westland Food Crop Pro—
duction," World Bank, Washington, 1978.

Miller, Thomas A. "Sufficient Conditions for Exact Aggregation in Linear
Programming Model,"égricultural Economic Research, Volume XVIII,
No. 2, 1966.

 

246

Ministry of Agriculture. "Cotton Prospects" and "Preliminary Cotton
Targets," MWanza, Tanzania, 1970 (mimeographed).

"Market Statistical Report," Tanzania, 1973.

"Producer Prices for the Main Scheduled Crops, 1972—73 to
1980—81," Tanzania, 1981.

 

Ministry of Economic Affairs and Development Planning. "Development
Programme for a Rural District: Geita District 1966—1969,"
Dar-es—Salaam, Tanzania, 1966.

Report of Working Party No. 12 on Agricultural Mechanization,
Dar-es-Salaam, Tanzania, 1968.

 

Second Five-Year Plan, Dar—es—Salaam, Tanzania, 1969.

 

Mudahar, M. S. "Dynamic Analysis of Direct and Indirect Implications of
Technological Change in Agriculture. The Case of Punja, India,"
Occasional paper 79. Department of Agricultural Economics,

Cornell University, 1974.

Mwangi, M. W. "Farm Level Derived Demand Responses for Fertilizer in
Kenya," Ph.D. Dissertation, Michigan State University, East Lansing,
1978.

Nkonoki, S. R. ”Land Holding, Politics and Agricultural MechaniZation
in Geita District," Institute of Development Studies, University
of Dar-es-Salaam, Tanzania, 1977.

Norman, D. W. "Economic Analysis of Agricultural Production and
Labor Utilization Among the Hawsa in the North of Nigeria,"
African Rural Employment Study, Rural Employment Paper No. 4,
‘MSU, East Lansing, 1973.

and Ogunfowora, A. ”Farm Firm Normative Fertilizer Demand
Response in the North Central State of Nigeria," JAE 24, 1973.

Hayward, J. A., and Hallam, H. R. "An Assessment of Cotton
Growing Recommendations as Applied by Nigerian Farmers," Cotton
Growing Review, 51, 1974, pp. 266-280.

 

 

Nyerere, J. K. Freedom and Socialism, Oxford University Press, London,
New York, 1968.

 

. The Arusha Declaration: Ten Years After, Dar-es-Salaam,
Tanzania, 1977.

 

Nurdin, T. "Factors Affecting Farm Decision Making: A Case Study in
west Sumatra, Indonesia," 1977.

247

Ogunfowora, A. "A.Linear Programming Analysis of Income Opportunities
and Optimal Farm Plans in Peasant Farming," Bulletin of Rural
Economics and Sociology, Iowa State University, 1972.

Palmer—Jones, R. W. "Field Research MethodsIUsed in Farm Management
Studies," I.A.R. Samaru, 1977 (mimeographed).

Peat, J. E., and Brown, K. J. "Effects of Management on Increasing
Crop Yields in the Lake Province of Tanganyika," East African
Agricultural and Forestry_J0drnal, Vol. 16, 1960.

 

 

Percy, H. C. "Factors that Affect Cotton Production in the Western
Growing Area," Cotton Growing Review, 52, 1975, pp. 253—277.

 

Plexico, J. S. and Tweeten, L. G. (1963). “Representative Farm for
Policy and Projection Research," JFE, Vol. 45, No. 5, pp. 1458—
1468.

Purseglove, J. W. Tropical Crops, Dicotyledons, 2, London: Longmans,
1968.

 

Pupius, J. A. "The Effects of Block Mechanization Schemes on Cbtton
'Production Within the Lake Region," Ukiriguru, Tanzania:
Western Region Research Center, 1964 (mimeographed)

Reed, W. "Cotton Pests," Ukiriguru, Tanzania, no date, (mimeographed).

"Problems posed by Early Sowing of Cotton in Lake Region,
Tanganyika," Empire Cotton Growing Review, 1964.

 

Samki, J. K. "Granitic Soils in the Western Cotton Growing Zone of
Tanzania," Ukiriguru Research Notes (Tanzania), 18, 1968.

 

Schultz, T. W. Transforming Traditional Agriculture, Yale University,
1964.

 

Shapiro, K. "Efficiency and Modernization in African Agriculture,"
Ph.D. Thesis, Stanford University, 1974.

Smith, M. G. "The Economy of Hausa Communities of Zaria," Colonial
Research Studies No. 11, Her Majesty's Stationary Office, 1955.

Smith, W. E. We Must Run While They Walk, Random House, New York,
1971.

 

Spence, J. R. "The Importance of Sowing Data for Cotton," Ukiriguru
Research Notes 32, Tanzania, 1969.

"Cotton Agronomy" Ukiriguru, Tanzania, no date (mimeographed).
Spence, J. R. and Littledyke, A. D. "Rainfall at Ukiriguru with Special

Reference to the Growing of Cotton," Ukiriguru Research Notes 41,
Tanzania, 1970.

248

Tanegja, M. L. "Role of Mechanization in Rural Development," University
of Dar-es—Salaam, Tanzania, 1972 (mimeographed).

Tanganyika African National Union, The Arusha Declaration and TANU's
Policy on Socialism and Self Reliance, Dar-es-Salaam, TANU, 1967.

 

 

Tanzania Cotton Authority, Production and Marketing Statistics for
Cotton, Dar—es-Salaam, 1980.

 

The Revolutionary Party of Tanzania, "National Policy on Productivity,
Income, and Prices," National Printing Company, Dar-es-Salaam,
1980.

United Republic of Tanzania. Economic Survey, Dar—es—Salaam, 1977.

 

Economic Survey, Dar—es-Salaam, 1979.

Upton, M. "Some of the Problems in Using Planning Techniques,"
Development Studies No. 14, Reading University, 1974.

Wolf, E. R. Peasants, Prentice Hall, Inc., 1966.

World Bank, Tanzania Basic Economic Report, Report No. l6l6-TA,
Washington D.C., 1977.

Yang, W. Y. "Methods of Farm Management Investigation for Improving
Farm Productivity," Agricultural Development Paper, No. 80, FAQ,
Rome, 1965.

 

 

 

   

MICHIGRN STRTE UNIV. LIBRRRIES
IIHIIIHIIIIIIIIIHIIIHIIIIHIIIIIIIHIIHHIIIIIIIIIIIIIHHI
31293008537965