TRANSPORTATEON AND AGRICULTURAL IDEWLOPMENT
IN WESTERN KENYA: AN APPROACH TO THE
PROBLEM OF INVESTMENT PRIORITIES

Thesis for the Degree of PhD.

MICHIGAN STATE UNIVERSITY
Stephen S. Birdsall

196.8

 

 

 

r’ LIQRAFE

THESIS .. . ~ . .
"chlgdi:
t)“ ”(I .

4., L... ..

This is to certify that the
thesis entitled
TRANSPORTATION AND AGRICULTURAL DEVELOPMENT

IN WESTERN KENYA: AN APPROACH TO THE
PROBLEM OF INVESTMENT PRIORITIES

presented by
Stephen S. Birdsall

has been accepted towards fulfillment
of the requirements for

PhoDo degree in Geogra Elly

 

Q I {1 44,../ /. 12/ fl/
W ‘

o‘essor

 

0-169

us.
9 s tr
I b
.. '-
v
P.
O
In‘ -r
TV“ *5.
A“: CA.

$56 fine wt:

9;» I. .
suds, t;&en 1n

ABSTRACT

TRANSPORTATION AND AGRICULTURAL DEVELOPMENT
IN WESTERN KENYA: AN APPROACH TO THE
PROBLEM OF INVESTMENT PRIORITIES

by Stephen S. Birdsall

The efficient expenditure Of limited investment funds is
of vital concern to development planners in low income countries.
If the project chosen for investment from a set of alternatives is
not the one which will yield the maximum return for that invest-
ment, then in a normative sense, some portion of these funds has
been wasted. This problem is approached as it applies to agricul-
tural road investments. A methodology is Offered illustrating by
example the procedure for determining the most remunerative road
improvement that could be made in a section of western Kenya.

ii- Agriculture and transportation are important elements in
a country's total development effort. The majority of the popu-
lation in each newly independent African state are engaged in
agriculture. Agriculture is also an indispensible source Of
revenue for most Of these countries. The agricultural production
process can be made more efficient by encouraging regional
speCialization. Improved movement facilities do provide such en-
Clearly, decisions concerning rural roads in under-

couragement -

developed countries should be based upon an empirically determined

l J .- w';v‘.
.;s: of .nxcs:...-...

' ' : ; .,..‘..
:e mamas t. - .!

Oats.
'F
,.I..
I
.

a - V
22:12 agrzcutts

..." .' ' '..
.,,.;cat.on In .\.

L": manages
tears: with tho.

Q" I . ..
.... IESuzt lnClCi

:ntural change 1

PR.

.. J16 method is

N .-.
vrx‘s
.g‘

{ability 5'6

23:63 '
“‘11 Profit

Stephen S. Birdsall
list of investment priorities. It is argued that such a list must
be obtained through analysis founded upon the geographic perspec-
tive.

t’" The procedure for determining the priorities for invest-

1'

p.,

ment in agricultural roads is described in detail during its
application in Nyanza Region and Kericho District, Kenya. The
method used calculates the improvement which. can reasonably be
made in the profitability of agricultural production. The costs
and advantages of improving one road rather than another are then

compared with the agricultural potential located along both roads.‘

.d-r’

/

/ The result indicates which roads will assist the greatest agri-
cultural change for the least investment in facilities. Crucial
to the method is treatment of each economic element as a spatially

var i ab 1 e func tionlx’}

iThe pattern Of existing profitability for each of eight
crops was calculated by subtracting the pattern of production
costs from the Pattern of crop value per acre. {Using environ-
mental characteristics in combination as ecological regions, and
projecCed crop prices and yields, patterns Of potential crop
profitability were derived. :1 "The difference between existing and
potential profitability is the improvement of which the land is
capable. After modification for the variable pressure of popu-

lation on land resources, a map is derived illustrating the

place-to-Place variation in the increased returns from the land

"- ,;
5""; Cast 5.

" “cs.
":‘¢§A.s‘ Cv_‘.
l'..... ‘I th‘
:.:-:...S ,n ..

:;.;v-- “a? ‘ “a
.-.§.l~.tb 5“‘..

b o
..v . 9-, 7,.
use“: 1“: ‘Va

Stephen S. Birdsall
Inhich are possible in the near future given appropriate stimulus
to development.

The justification for investment in one road rather than
another must be based in part upon the direct costs and returns
expected from these investments. Road construction costs, the
cost of future road maintenance, and the reduction in vehicle
operating costs following road improvement are seen as the primary
elements in the transport sector to be considered in the priority
determination{ The ratio between these incremental costs and re-
turns from road investment is, in turn, related to the adjusted
pattern of agricultural profitability improvement. The result is
a variable index of relative priorities for road upgrading for
each unit distance of road in the study region. The method is

judged flexible, inclusive, and well suited to dynamic application.

.l-v- e‘vflp
.5533.

m

‘A

Trip

TRANSPORTATION AND AGRICULTURAL DEVELOPMENT
IN WESTERN KENYA: AN APPROACH TO THE
PROBLEM OF INVESTMENT PRIORITIES

BY

Stephen Si Birdsall

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY
Department of Geography

1968

ACKNOWLEDGEMENTS

As this volume is completed, I am indebted to many
people and wish to thank them here for their assistance.

Dr. Harm J. de Blij has provided intellectual stimulation and
encouragement throughout my graduate studies and typically in-
cisive comments and suggestions on the substance of this book.

Dr. Milton H. Steinmueller has also offered constructive criticism
on this work as well as previous research upon which part of the
present volume is based. I also want to thank Dr. Lawrence M.
Sommers for his help as Chairman of the Department of Geography.

I especially want to thank Dr. Charles C. Hughes, Director of the
African Studies Center, for the financial assistance without which
this study would have been seriously weakened.

Many individuals contributed to the completion of this
work by their assistance in Africa and earlier in London.
Professor R. J. Harrison Church provided numerous introductions
in London. Dr. E. D. Tingle, Head of Tropical Section, Road Re-
search Laboratory, supplied me with many relevant unpublished

reports.

In Africa, Mr. S. Kealey and Mr. K. Shabani of the East

African Railways and Harbours, and Mr. J. D. G. F. Howe of the

iii

 

. . .. .
9vu~an col‘fls aw
.v r
nhfibI-L..Ol~> h
. .
ou;:01 «as and I
u.»...»-- my

I van:

.vl— .~ “4‘
5““‘95: St...“ .

\‘N-aa
...\,: :‘r ts.“
. .h s
V
. a
v
UV ,. "‘-
.., “A.
at a
- L -... ‘ \_

Road Research Laboratory, Kenya Branch, were of immeasurable as-
sistance. A great many Kenyans, in addition to those interviewed,
were more helpful than they realized, I am sure, because of their
friendliness and their consistent good humor in the face of my
questions and requests.

I want to thank Mrs. Beverly Nadele for her frequent
assistance (and her patience) throughout five long years of
graduate study.

Special thanks must also be given to Mrs. Lawrence H.
Snyder for typing the complete final draft of this manuscript.

Mr. In Kim and Mr. Dennis Enberg provided some cartographic as-
sistance.

I am also indebted to Mr. S. M. Sepeku in many_ways.

As a friend, a fellow graduate student of geography, and as a
teacher, he was a constant source of wisdom and insight.into the
importance of being African.

Above all, my wife Sally has been an inspiration as well
as directly helpful. In addition to shouldering much of my re-
sponsibility as a parent to our two young children, she undertook,
with no previous experience, a good deal of the cartographic work

during the final weeks of work and did an excellent job.

iv

4T\”"" :. 4":\.r:‘
" .\ . ‘,- ‘1
----~ ~ saga“-..

'v:- qr

Y
..-. s; .....-l_

 

I
eat a.
-
vin‘v §:01.
.

III’ A RPM.

PR;

C051
POte
AFCVE

TABLE OF CONTENTS

Page
ACKNOWLEDGEMENTS . . . . . . . . . . . . . . . . . . . . . . . ”iii
LIST OF TABLES . . . . . . . . . . . . . . . . . . . . . . . . viii
LIST OF ILLUSTRATIONS . . . . . . . . . . . . . . . . . . . . X
LIST OF APPENDICES I O O O o I l O l O O c l O I I I 0 O C O O Xi ii
INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . 1
PART I. OVERVIEW
Chapter .
I. DEVELOPMENT, AGRICULTURE, AND THE NEED FOR
MOVEMENT . . . . . . . . . . . . . . . . . . . . . 11
Problems of Financing Development . . . . . . . . . 14
The Agricultural Sector . . . . . . . . . . . . . . 22~/
The Transport Sector . . . . . . . . . . . . . . . 26V
II. GEOGRAPHY AND ECONOMIC DEVELOPMENT . . . . . . . . . 36
The Necessity of a Geographic Perspective . . . . . 36
The Case Study . . . . . . . . . . . . . . . . . . 41
Models of the Macro-Regio . . . . . . . . . . . . 47
PART II. METHODOLOGY
III. A REGIONAL APPROACH TO RECONCILING INVESTMENT
PRIORITIES IN AGRICULTURE AND TRANSPORTATION . . . 54
Costs and Income . . . . . . . . . . . . . . . . . 56
Potential . . . . . . . . . . . . . . . . . . . . . 61
Movement Facilities . . . . . . . . . . . . . . . . 64

C S nL .. C : S C C m! a s. .. 7 .. .- .- .1
a «as r. C an .3 .3. C .r at a u. n... H» .3: my: «.5 k..
w... an .L P a.“ _ . In vL .L r; C. Ki. P S C; 7. pr. uh A 1A.. T; A
.nu UK. .N.
T .-,. T
us wk. 7..
DI- Y1.

 

Chapter Page

PART III. APPLICATION

IV. THE COSTS OF MOVEMENT . . . . . . . . . . . . . . . . 72
Factors of Transport Costs . . . . . . . . . . . . 75
Absolute Transport Costs . . . . . . . . . . . . . 83

V. THE COSTS OF PRODUCTION . . . . . . . . . . . . . . . 93
Problems of Cost Determination . . . . . . . . . 94
Reconciling Production Cost Variability . . . . . . 99
The Precision of Cost Patterns . . . . . . . . . . 101
Assumptions . . . . . . . . . . . . . . . 102
Location of Crop Production . . . . . . . . . . . . 104
Location of Production Costs . . . . . . . . . . . 115

VI. THE PROFITABILITY OF PRODUCTION . . . . . . . . . . . 129
Spatial Variation of Farm Income . . . . . . . . . 130
Determination of Income Variability . . . . . . . . 131
Patterns of Existing Profitability . . . . . . . . 154
Summary . . . . . . . . . . . . . . . . . . . . . . 165

VII. THE DERIVATION OF AGRICULTURAL POTENTIAL . . . . . . 167
Investment Decision Criteria . . . . . . . . . . 167
Foundations of Agricultural Potential

Determination . . . . . . . . . . . . . . . . . 171
Agro- ecology in Zambia . . . . . . . . . . . . . . 173
Agro- ecology in Rhodesia . . . . . . . . . . . . . 179
Agro- ecology in Kenya . . . . . . . . . . . . . 181
The Pattern of Ecological Zones . . . . . . . . . . 183
A Measure of Climatic Influence . . . . . . . . . . 193
VIII. THE VALUE OF POTENTIAL PROFITS . . . . . . . . . . . 209
The Expanded Areas of Future Production . . . . . . 210
The Need for Field Research . . . . . . . . . . . . 214
Present Data and Spatial Generalization . . . . . . 215
The Pattern of Future Profit . . . . . . . . . . 217
Improvement as an Investment Criterion . . . . . . 234
The Impact of Population . . . . . . . . . . 242
The Pattern of Maximum Profit Increase . . . . . . 254
Summary . . . . . . . . . . . . . . . . . . . . . . 262

vi

o
_.-O“

.—
-...‘v s 5

a

.y— (1"

I“: I'\

Chapter Page

IX. THE RELATIVE INVESTMENT PRIORITY INDEX . . . . . . . 265
The Pattern of Investment Costs . . . . . . . . . . 265
The Variable Costs of Road Improvement . . . . . . 267
The Variable Costs of Road Maintenance . . . . . . 271
The Problem of Absolutes . . . . . . . . . . . . . 279
The Priority Index . . . . . . . . . . . . . . . . 282
The Application . . . . . . . . . . . . . . . . . . 285
The Regional Priorities . . . . . . . . . . . . . . 298
Summary . . . . . . . . . . . . . . . . . . . . . . 307

PART IV. CONCLUSIONS

X. SUMMARY AND CONCLUSIONS . . . . . . . . . . . . . . . 309

The Context and the Orientation . . . . . . . . . . 309

The Existing Pattern of Agriculture . . . . . . . . 313

The Potential Pattern of Agriculture . . . . . . . 316

The Pattern of Road Investment . . . . . . . . . . 319

The Method's Potential and its Limits . . . . . . . 321

APPENDICES . . . . . . . . . . . . . . . . . . . . . . . . . . 323
BIBLIOGRAPHY . . . . . . . . . . . . . . . . . . . . . . . . . 329

vii

(7‘
l
LII

8~1

Table

4-1

4-2

4-3

5-1

5-2

5-3

6-1

6-2

6-3

6-4

6-5

8-1

LIST OF TABLES

New Motor Vehicle Registrations, 1960-1965

Vehicle Operating Costs, Operated in the Geita
Peninsula, Tanzania, 1966 (East African Cents
per Mile) :

Vehicle Operating Costs, Operated in Kenya,
1963(?) (East African Cents per Mile)

Ratio of Vehicle Operating Costs, by Road Surface
Quality

Primary Environmental Growing Limits on Major Cash
Crops Grown within the Study Region

Average Cost of Annual Crop Production .

Rice and Rice Flour Shipments, Mohoru Bay to Kisumu,
by East African Railways and Harbours

Average District Yields for Selected Cach Crops,
1963-1965 (per acre)

Average District Value per Acre for Selected Cash
Crops, 1963-1965 average (shillings)

Environmental Variables for Indices of Crop
Production Suitability .

Value per Acre Equivalents for Numerical Indices of
Environmental Suitability for Crop Production,
in shillings

Estimated Average Yields from Smallholder Tea
Plots, per Acre

Concurrence of Agro-Ecological Zones and Areas of
Crop Production Potential, as described by Brown

and developed separately in the present work .

viii

Page

74

78

79

79

109

117

119

132

135

137

148

164

218

afl»

0"
an

,a..

.P.

AC?

‘ a

;‘v-"‘
\..5-

P

g
5

'.a

a A.
d.\.u

92 C

 

a .

.. n
\- L.
&.

YT]
‘ \

C:

Iszai
gas

'9

C

m

A

per .

w. .
.. a VA
wk .1»
r .1
.t We
05 P
.t

\J

v)

.

pH]

Table

8-2

8-3

8-4

9-1

9-2

9-3

9-4

9-5

Estimated Average Potential Value of Production per
Acre, for Selected Cash Crops

Potential Value per Acre Equivalents for Numerical
Indices of Environmental Suitability for Crop
Production, in shillings

Land Carrying Capacity, by Ecological Zone

Road Grade Classification

Calculated Average Improvement Costs per Mile, by
road grade .

Modifications to the Basic Annual Road Maintenance
Grant

Total Cost of Road Maintenance for Roads Earning the
Basic Grant, 1964/65 (pounds sterling per mile

per year)

Determination of the Road Factor for the Investment
Priority Index .

ix

Page

221

226

245

270

272

273

274

289

tn
0

y 4 ,_4 ,.__..

n. n (J.

N '4) he he)
CL O (I m

Rcais

.
rgztc

Praduc

AVEra

AVEra

Figure
I- 1
I- 2
4- 1
4- 2
5- 1a

1b
1c
1d
5- 2a
2b
2c
2d
5- 3
5- 4
5- 5
5- 6
6- 1
6- 2
6- 3
6- 4
6- 5
6- 6

LIST OF ILLUSTRATIONS

Towns and Physical Features
Relief

Cost-Distance to Kisumu or Lumbwa,
truck .

Roads and Tracks
Producing Area: Tea

Pyrethrum

Coffee .

Potatoes
Producing Area: Cotton .

Sugar Cane .

Rice .

Groundnuts
Total Production Costs: Groundnuts
Total Production Costs: Cotton and
Total Production Costs: Sugar Cane
Total Production Costs: Coffee and
Average Annual Temperature
Average Annual Precipitation
Soils .
Average Value per Acre: Groundnuts
Average Value per Acre: Cotton and

Average Value per Acre: Sugar Cane

X

for 3-5 ton

and Tea .
Pyrethrum .
and Potatoes

Rice

and Tea .
Pyrethrum .

and Potatoes

Page

87

88
107
107
107
107
114
114
114
114
120
121
122
123
139
140
142
144
145

146

 

0..

g.‘..’:':
7

.' .‘
h. l
V at
"
J A.
o
o l
A

Figure Page

6- 7 Average Value per Acre: Coffee and Rice . . . . . . 147
6- 8 Average Existing Profitability: Groundnuts and

Tea . . . . . . . . . . . . . . . . . . . . . . . . 155
6- 9 Average Existing Profitability: Cotton and

Pyrethrum . . . . . . . . . . . . . . . . . . . . . 156
6-10 Average Existing Profitability: Sugar Cane and

Potatoes . . . . . . . . . . . . . . . . . . . . . 157
6-11 Average Existing Profitability: Coffee and Rice . . 158
7- 1 Agro-Ecological Regions: (after Brown) . . . . . . . 188
7- 2 Agro-Ecological Regions: (modified by soil type) . . 190
7- 3 Potential Evapotranspiration . . . . . . . . . . . . 197
7- 4 Water Surplus . . . . . . . . . . . . . . . . . . . . 198
7- 5 Water Deficit . . . . . . . . . . . . . . . . . . . . 199
7- 6 Moisture Index . . . . . . . . . . . . . . . . . . . 204

7- 7 Agro-Ecological Regions: (modified by soil and

climate) . . . . . . . . . . . . . . . . . . . . . 206
7- 8 Agro-Ecological Regions: (after D.I.E. Report) . . . 207
8- 1 Potential Profitability: Groundnuts and Tea . . . . 230
8- 2 Potential Profitability: Cotton and Pyrethrum . . . 231

8- 3 Potential Profitability: Sugar Cane and Potatoes . . 232
8- 4 Potential Profitability: Coffee and Rice . . . . . . 233

8- 5 Potential Profitability Improvement: Groundnuts
and Tea . . . . . . . . . . . . . . . . . . . . . . 237

8- 6 Potential Profitability Improvement: Cotton and
Pyrethrum . . . . . . . . . . . . . . . . . . . . . 238

8- 7 Potential Profitability Improvement: Sugar Cane
and Potatoes . . . . . . . . . . . . . . . . . . . 239

xi

Figure

8- 8

8-10
8-11

8-12

8-13

Potential Profitability Improvement: Coffee and
Rice

Population Density, 1962
Projected Population Density, 1972
Land Use Intensity

Maximum Potential Profitability Improvement, all
cash crops

Adjusted Potential Profitability Improvement
First Priority Ranking

Second Priority Ranking .

xii

Page

240
247
248

253

256
259
299

305

“‘~‘v~4‘{
ao.:,."‘
t.

1- The Meth
Suitah

IT

ta- Faming ‘

LIST OF APPENDICES

Appendix Page
I. The Method of Deriving Indices of Environmental

Suitability for Crop Production . . . . . . . . . . . 324

II. Farming System for Ecological Zones . . . . . . . . . . 325

xiii

.w. an ;u-~"' ‘
_: . L e»... A
a
._- Y” ,l-‘b 5"
,.. :...a\~L LI
. e
--.r

. —. -.: 3: {1‘1
qu—a. u" i-....;s

~cly . \—~ 5
. :s::::t if. t
2‘ :':se riais
: {-1

'L‘~- _
GATE: "" "r:
rte A .
.. .
_-?“.H
‘~¢-.a
v .
.‘ P
5._‘ .v-~
~§~ar
.
pr . .
.. ,..
., r-
cu“

INTRODUCTION

This study deals with the justification for decisions con-
cerning economic development investments. It outlines in detail
the explicit grounds on which an objective decision based on

4"—

empirical data may be made.) It does not approach the problem of
ti

whether or not to invest, but rather, where might the most remu-
nerative investment be made?? The scope of the study is limited to
investment in transport route improvement, and more Specifically,
to those roads which are used to market agricultural produce.’ The
procedure presented here allows the determination of a set of in-
vestment priorities for road improvements within an agricultural
region.

Before proceeding to the main body of this work, it should
be understood by all readers that no attempt has been made here to
provide a source for policy decisions. Whether an academician or
a politician, an economic theorist or an agricultural economist,
an overworked civil servant or a student of economic development
problems, it should be accepted by all that the purpose for
writing this volume is threefold.

First and foremost, it is an attempt to formulate a
general method for determining a system of economic development

investment priorities in regions of incompletely utilized

1

‘ w
. '7
O

f3

0"

9
~G‘.-

h~'.A
‘-l.

a.

a
u‘
.o

u ' . I I
D - l O ’ . ‘
.

.-...O‘.~

cf

1

...-"'..-

..-y~

‘-
_..>‘

s
' .
.Ah " ""
n.o -v‘b‘ ‘91.

so.

be-..

-\ “NF

'2‘”
yt‘

--\...
.. . .r"
. -2.

v
\

war as t

“.-
~

~..

Third

2
agricultural potential and inadequately developed transport
facilities. The primary emphasis, again, is on development of an
objective method of determining these priorities for the transport
sector.i;The small portion of the earth's surface chosen to pro-
vide an example of this method is not unique in its suitability
for such study, and the procedure developed here should be of use
elsewhere.

Second, the method formulated has sufficient latitude to
enable development planners, given adequate data, to gain insight
into non-transport limitations on a region's development. The
scope is not unlimited, for the basis for maximum economic im-
provement estimates pivotscnithe calculated agricultural potential
in the area. The method only incidentally allows manufacturing
development potential to be seen more sharply. Such aspects of
the economic picture as lack of public services and overly popu-
lated areas may be made more apparent under this method, however,
insofar as they are related to agricultural development and
potential.

Third, by the very nature of the data used in this appli-
cation of the method, specific requirements for future research
are clearly outlined.‘ This is as it should be. Whenever a new
approach to a problem is first proposed, it is bound to raise more
problems than it solves. Data are rarely available in the form

desired, especially in newly independent and low per capita income

a

r.
. .L. ..
. _ 4. \
an e a“

”a r.

r” T.
.n .. . . .
\a II;
o
.u .u.
: .a. _.
.a a. a.
. .a. ..
. .

 

5

Ya

 

H
‘ o-
‘ ~

3
countries such as Kenya. Trained personnel are at a premium; too
many immediate tasks are to be accomplished for grand new theories
to be conceived; and researchers from abroad are too often either
unaware of the dearth of relevant data and attempt to force mis—
placed information into involved models, or they are not inter-
ested in undertaking the extremely difficult job of gathering such
data. Throughout this study, problems of this kind are outlined;
hopefully some of the findings will be of use in later under-
takings.

The region of concern in this volume was shown for several
reasons. First, it is an area of high agricultural potential.
Perhaps the chief element of the environment which limits agricul-
tural production in Kenya is precipitation. When high rainfall
sections of the country are further subdivided on the basis of
roughness of relief and elevation, the region chosen for study be-
comes increasingly exclusive. Large areas even marginally suita-
ble for cultivation are limited to a narrow band along the Indian
Ocean coast, the central highlands-~much of which was restric-
tively scheduled for European settlement--and a fairly large area
of land to the east and north of Lake Victoria. This last is the
area under scrutiny here.

Second, the focus of this work (the problems of invest-
ment in agricultural feeder roads) diminishes the suitability of

some areas of high agricultural potential and makes other such

.L .n

 
 

.Au . . —c v. t . . . .
C .- .\. “A r. ok a . ,.\ a. Q» C o. .3 A
. . m L T. t . . 3 I t A. F
. . .2 L: t. a 1 “I S s ._ :a .l/ T. on
. . a . . . s . .t a S S .t a r . ... S
a». .0. a a .C .5. o. nu. a: u. A. .st Ft. {A It my .‘a 2
.d A. a a A; I r. a . C . S DU 0
n n‘. fly. in 0L L» I Adz
u‘. a o A 0.. G. v. i a r .J 1
o» J I. .3 ha ”a .Fs a“ *u “VI 3; Au
.3 . . . .u .5 as. ... ... .l -t C L»
. A , . , . u . .. . a .. ..~ . a c
v . ._ .. . to. . L. 5‘. It 5‘. fi‘v V]; n u
. a . .4. «r». . . -.. .1. L. . .4 n . s .~» ‘ a
u. A .s 2‘ .. .. u. ... . .m. r.. $1.. C.

 

 

4

areas more attractive. The coast was considered a poor choice
for a study of this nature because (1) movements within that area
would be dominantly linear and thus rather restrictive in the
general applicability of any conclusions reached, and (2) the
variety of environmental characteristics is not great, thus
yielding the same conclusion as that above with regard to appli-
cability elsewhere. The central highland area was also decided
against as a region in which to carry out this research because
(1) the attention given this area by the British resulted in a
relatively well-developed road network, too well-developed for
the purposes of this study, and (2) the presence of urban centers
in the area, not only in peripheral locations but throughout the
region, results in a marketing pattern which is much too complex
for what is meant to be an introductory research effort.

The region finally chosen for study comprises part of the
high precipitation area in Kenya adjacent to Lake Victoria. (See
Figure I-l.) The study region is the combined area of Central
Nyanza, South Nyanza, and Kisii Districts in Nyanza Region, and
Kericho District, which has been part of the Rift Valley Region
since 1962. Although it would have been more rational to examine
the entire high rainfall area (after all, production potential
does not vary along administrative boundaries), such an under-
taking would have been prohibitively expensive at this stage of

research development, in terms of both money and time. With the

 

34'] I ' sla‘ ' sill I it

~ NYANZA REGION and Kericho District

TOWNS AND
[Nd T" PHYSICAL FEATURES
C

It

 

 

 

 

 

 

 

\ A. \ (1% (/
~ “p" . E A Ins} C) J 5.2"" f \
9 \, MMMMMM f i/\
" 7 nnnnnnnn V. \
3 Q . \ .4“. I, \,J m\»"') o \
A I ~k~ \ » ----- ,.
. O . \ 1
.' ,/ N.
“ o 1 / x/ \ J”
" 4f . ales-Id h./‘ \ N
B . 9 f,‘ Mine ) e I,
‘y’ ' Ca 1 \\\r/ m
o a
\. Suns \-
\ V / A
.\\ ,,‘\
.\~ L. 35°]: . ’E
i k , \.
\\4 ms Primary Urban Center
Secondary Urban Centers
. ...... Other Centers
N District Names
nu Physiagraphic Features
E] Land Formed by Non-African Farmers
or otherwise excluded from study
as

 

 

 

 

FIGURE I- 1

'1

Ken
cs.
- -
_.e. -
d..

v..~....

c-o

V4,;n-s

p
._.e~.¢.e

o . Q. . - . .C .. . . t it
... .. .\\. ,c L .6 Nu T. 1 no
. . .u w. r. .n... :4 r r CA
.. a t a. A a ”C L a .c mu .2
C a .: Pt I; r. be
a .9. v. we .6 T. a u.
use N.» e a La .5 A: 3» PM .4
at. .F. b . .6. r; a A- L»
.P» .C ”a .t .: Le C .: Vol
.3 r u . A \a .A Cs {a e» A...
. v '5 u s s e .u t r ‘e 5e
.3 n s. ....u u: L; . a .e a a v .. u
.s i. ..s 'nQ .-\ tau . s a.»

f5

 

6
exception of that portion of Central Nyanza to the west of Maseno
and Yala in which there is no sharp change in physical character-
istics as one travels north into Busia District, the study region
is rather well bounded by a distinct change in agricultural
potential. (See Figure I-2.) The Nyando Escarpment provides a
sharp boundary along much of the remainder of the northern edge
of the study region. The Mau Forest occupies the largely unin-
habited eastern edges of Kericho District. The southern boundary
is indicated by the Tanzanian border in the west, and a gradual
drop in altitude and precipitation toward the southeast away from
the study region. The western boundary is, of course, clearly
indicated by the coast of Lake Victoria.

_ The focus of the present work is on agricultural roads
which facilitate movement for the African smallhold farmer.: Be-
cause of this, a portion of Kericho District and a smaller area
in Central Nyanza District are excluded from examination. These
excluded areas contain land owned or operated by non-African
farmers during the period for which data were available. It was
believed that the economics of cash crop farming are sufficiently
different in non-African areas from the present conditions of
African farm operation that the resulting patterns would be dis-
torted and less applicable to regions where long established
European or Asian farmers are absent.

Before proceeding to the study itself, several final re-

 

 

T Js°i l .

NYANZA REGION and Kericho District .
RELIEF

    

 
   

 

C
/[-'-—"-"1f“$fi$éy
. , ”7‘3? \53

 

 

 

 

 

 

 

5 U
I I" , I,
Altitude- feet
34’“

under 4000 7000-8000
4000—5000 - 8000—9000
5000-6000 - over 9000
6000-7000

KENYA

 

 

 

 

FIGURE I- 2

V as

-nr‘r‘
w' I.
. .

o

:J
ch

8

marks are appropriate.; Because this volume is not meant to be a
general thesis on economic development, some of the discussion of
problems of development (Chapter I) may appear rather superficial
to those well-versed in development theory. As might be implied
by the fact that such discussion takes place in the first chapter,
it is provided primarily as background for readers less familiar
with the basic arguments in this field.

The discussions in this volume are not meant to denigrate
the efforts and achievements of the civil servants in Kenya or
elsewhere. Many of those with whom the author had personal con-
tact, whether expatriate or African, were obviously doing more
work for less pay and under less-than-favorable conditions than
those academic critics who so often sit in judgement on the in-
adequacies of their results. I do not count myself among the
latter.

The research upon which the study is based was conducted
in East Africa during 1966. Very few data for 1966 have been in-
cluded because of their fragmentary nature when the author was
in the field. The conclusions, therefore, are relevant to the
data of 1965 and earlier. Monetary equivalents, of course, are
also as of 1965. One shilling consists of 100 cents (East African)
and is equivalent to $.14 (U.S.). Thus, one pound sterling =

Shs. 20/- = $2.80 (U.S.) throughout the volume. It is hoped that

I'ael,’
us. to
l

a

ELSCYaCCt I

 

9
neither the use of pounds-shillings-cents nor the retention of
British English terminology (and occasionally spelling) will

prove a hindrance to American English readers.

PART I

OVERVIEW

.3
CIT.
«KI- "
,_
e

\F'N‘
-u-v' A
.I - uh.

.
«W‘ "
..‘--
OF‘
...-. C
.
0;.«9‘
List-Q
CC":—
5'

_ I
~v-.fl-¥.s

.-.

A5
E
v;

E.

e‘y'

EVI‘N
55" 1e.
~ k¥lgar

CHAPTER I
.DEVELOPMENT, AGRICULTURE, AND
THE NEED FOR MOVEMENT

Rarely has a concept evoked such varied reactions as has
economic development. Academicians have toyed with abstractions
about its processes and stimuli for many years, challenged by
the complexities and the elusiveness of the object of their at-
tention. To politicians in low-income countries, economic
development is a serious and immediate task. And as so often
occurs with new ideas, those with whom the concept of develop-
ment most closely deals, i.e., the small farmer living in
relative isolation, are least likely to have any awareness of
the process that occupies their political leaders' attentions.

As varied as the reactions to economic development are,
the definitions of this process are often very much alike. The
customary discussion concerning economic development first
stresses the difficulties associated with any attempt at defi-
nition of the subject, then may make some reference to possible
indices of development, and finally arrive at something such as:
that process which results in a significant increase in average

annual per capita income well-distributed among the country's

ll

xv-

. .

-o.

A
r.-¢.—.o‘--'s

A«...

 

a” I L .. rt .. .. .. 4L 4.

.C . . .T an C C . e. .3 .3 .C C e.»
a. .. .t i .c. .. .. eh A: 2 a. n C.
. . Ca V. .t C rt 5 .C t. .C ... ~ L .t
.: nu _. . l A. F. C an r

.N. c. s u a» Y. 7. _.. a.“ a: .: ‘5 p y. .
.C .a .5 a L. c .. ... .s .7 LC . .. a.
c I. 9» .FL . . . p. .A Jr. .A C. c . . «c
r. ... .u L. .3 .. a .a. n. 2a .. .C (A
— .sl .7.. ..\. “A 9¢ L. .: .4. at .4» u- .N.

. __ .sa . . ad .: . a at a .. .\ in . C . . ~ . l. . a
.e. .. u. .. . . .s- er... «C -- .ah n». me .a .a- .es

12

1

population, or which results in a more efficient use of re-

sources.

The problem of settling for such definitions should be
obvious. There are numerous other measures of development which
might be used equally well, e.g., a production per capita index
based on one or more sectors of the national economy, relative
levels of consumption, or relative returns to investment.
Similarly, to say that development encourages or results in
greater efficiency simply says development results in a devel—
oped condition. The basic difficulty with these ”definitions,"
as with rising per capita income, is that they merely describe
and do not define, constituting one measure of the process but
not its essence, indicating one of the goals of development but
no aspect of the means of arriving there. In other words, such
indices describe economic growth or the economic results of
development but do not really define economic development.

In spite of these definitional problems, little con-

fusion results from the use of ”underdeveloped,” "less devel-

 

1For an example of this common introductory definition,
see, John R. P. Friedmann, The Spatial Structure 9f Economic
Development 12 the Tennessee Valley (Department of Geography
Research Paper No. 39; Chicago: University of Chicago Press,
1955).

 

 

 

2John W. Mellor, The Economics 2; Agricultural Develop-
ment (Ithaca, N. Y.: Cornell University Press, 1966), p. 3;
and, Santikumar Ghosh, The Financing 2f Economic Development
(Calcutta: The World Press Private Ltd., 1962), p. 1.

 

 

4"
“I

a

it:

,-.<

Pia-v.4
~s...ld‘si‘ 21:3“

in.
~-

-=C;aliza:i:n

2“
.

1.

..C:a

31' 7.3"

l3

' and "low income" with reference to

oped,” "developing,'
national well-being. Kenya, Guatemala and Burma are underdevel-
oped and Great Britain, New Zealand, and Sweden are not, regard-
less of the degree to which each country's resources are utilized.
Elaborate techniques are used to classify countries in terms of
their level of development, and description remains dominant as
countries ”develop" in the degree to which they are seen to
Change.

It is argued here, however, that historical perspective
is useful only when it is used to provide insights into possible
future occurrences. Thus, a description of processes that have
already occurred is meaningful to the extent that it may be used
to indicate trendsi: A useful definition, therefore, should
distinguish between growth and development, the results of devel-
opment and its primary characteristics. To this end, economic
development will be considered here to mean the evolution of
production outlook from that of local subsistence to one of
Specialization for the sake of commercial gain, whether individu-
al or national. The revolution in attitudes and values necessary
to bring about a change of personal emphasis from production for
direct support of oneself and one's family to production for
sale is the most distinguishing characteristic of what is common-
ly called economic development. It is almost pointless to speak

of per capita income when an individual provides for his own

.,.J. 3“ ”a” 1‘ "
.::_3 ... n 3 A..-.

u..‘_., g 1
....:.C.a-. SIN'JC'
r

" V r»... tr
.0 ant “v-
' LA

14
needs and has little excess. On the other hand, without per-
ceiving utility in the selling of produce, incomes will not
rise, based as they are on improvements in quantity or quality

3{It could be argued that the values of

I

of a regular surplus.
the producer may change but without knowledge of the means of
commercial production, it is unlikely that a significant surplus
will result. While it will not be so easily assumed here that
small-farm producers in low~income countries are unable to
produce a surplus with traditional methods if they see the value
in it,4 it will be conceded that investments must be made in
social overhead capital and infrastructure which will assist the
producer to pursue his gradually changing values. That is,
there is the need for investment capital, for funds with which
the necessary facilities for encouraging this revolution in out-

look may be best pursued.

Problems of Financing Development

 

The problem of finding adequate funds for development

has two major facets. The first of these relates to the ability,

 

3It is ironic that a study explicitly discussing the
necessity of changing values in the manner described should use
in its title the term "growth" rather than development. See,
W. W. Rostow, The Stages of Economic Growth (Cambridge: The
University Press, 1961), eSpecially Chapters 1 and 3.

 

 

4See, for example, the discussion in William Allan,
The African Husbandman (London: Oliver and Boyd, 1965).

 

, vh-I
" 0" f‘ ‘ '
" 1..

‘. bu- uL‘A

15
or the inability, of the low-income countries to generate the
needed capital internally. The other aspect of this problem
arises from the trade structure of these countries and the
manner in which this structure and the foreign exchange earned
are affected by world market conditions.

On a purely economic basis, the paucity of internal
capital generation in countries is a function of their low in-
come. With a low level of per capita income, a relatively poor
capacity to build up any significant amount of savings exists.
The closer is income to subsistence expenditures, the less will
remain for saving--even assuming a strong desire to save.
Similarly, the lower the level of savings, the lower will be the
ability to invest in improvements, whether individual or nation-
al. Without investment, it may be difficult to maintain levels
of productivity, let alone increase them. And finally, since
income is an immediate function of productivity (excluding
taxation), low output per unit of input will result in a mainte-
nance of low levels of personal income. This sequence of argu-
ments is commonly called "the vicious circle of poverty." The
importance of this basic concept here is the self-reinforcing
nature of the argument. A low-income country does not have the
financial-institutional resources available to generate its own

economic development, and thus rarely enough for growth.

c:;::er;art. E'.

ax‘a:.a:.e ICI' l.“

yx
f

 

16

As Nurkse has pointed out,5 this argument has its demand
counterpart. Even assuming, for example, an amount of capital
available for investment, there will be little incentive to in-
vest private capital if the market is small. The size of
market, of course, is limited by the low income - low productiv-
ity relationship which is, as has been shown, self-reinforcing.
The most obvious path out of this two-sided dilemma, and that
taken by most low-income countries, has been to set aside
dependence on private capital and make "public” investments in
industries which might ordinarily be more efficiently operated
in private hands, given the incentive of a potentially large
market.

With the acceptance of investment responsibility
residing in governmental rather than individual hands, attention
is focussed on obtaining a source, or sources of capital with
which to invest. Bilateral international aid, whether grants or
loans, is attractive financially in the short-term view but
usually has long-run political or economic liabilities. Taxa-
tion of the productive portion of the economy, on the other hand,
may result in depressed incentives to invest because of the rela-
tively small marginal increase in returns. Nurkse has pointed

out that with care and under somewhat unusual circumstances, this

 

5Ragnar Nurkse, Problems of Capital Formation in Under-

developed Countries (New York: Oxford University Press, 1962),
p. 5.

 

at lni.
”'5’" a ..-
‘.... :ouvu.r

it left i

 

17
6

repressive effect need not occur. For example, in an unsuccess-
ful attempt at indirect taxation of exports in Argentina, primary
products were marketed through a government monopoly. Prices

on the world market for the export goods were much higher than
the prices paid by the government to its internal suppliers. It
was hoped to finance the country's industrialization in this man-
ner. The result, however, was a drying up of export supplies.

In other words, Argentine producers were not willing to maintain
production levels at a depressed price merely for an impersonal,
national goal Of future industrialization. In Japan, on the
other hand, the government did not tax eXports but the land. "In
this way it left intact the incentive to produce silk for export
at a time of favorable export markets, while yet, in effect, in-
tercepting the foreign exchange proceeds of these exports for use

7 This comparison should be suffi—

in industrial development."
cient to illustrate that taxation of production is a risky propo-
sition, even when there appears no alternative. This is made
even more difficult by the fact that the primary source of export
produce lies in the hands of a majority of the population, i.e.,

the farmers.

Taxation of exports will have a less deliterious effect

 

6Ragnar Nurkse, "Trade Fluctuations and Buffer Policies
of Low-Income Countries," Kyklos, XI (1958), 152—53.

71bid., p. 153.

q . .
CZ EUCJCClCI‘. 'n’f‘c.
‘ ‘P'll‘ fl Anxr1
:. :_;..:r pr... _

. . , ,-

vuv )- nv- ,. .fi

'H‘n- PLaCLS 1-:
.

:5

1
ulna mark—.t C .,
u ‘,\“
W; - ‘1 '
’r‘c“ ddCt‘catL.

I

"r

 

18

on production when such production is increasing and the tax is
a smaller proportion of the total return. Given steady or rising
world prices for the export product, and a desire on the pro-
ducers' parts for the benefits of increased income, production
will increase and allow taxation to drain off a portion of the
gain for investment projects. Export products from low-income
countries are essentially primary products, however, and the
world market for primary products is subject to considerable
price fluctuation. It is to this point that we must turn next.

It is an unfortunate characteristic of primary product
markets that they are an unstable source of foreign exchange,
for it is the primary producing countries that most need
stability in earnings. Although there is some dispute over
the validity of figures illustrating this instability in the
long run, there is little argument over the existence of short-
run fluctuations in prices. A recent study by the United
Nations measured the deviation of export proceeds from a three-
year moving average, 1953-1960, for a large number of countries
and found fluctuations to be four times as high for countries

exporting primary products as for the industrial countries.8

Another unfortunate aspect of this situation is that

this instability in market prices is largely inherent in the

 

8United Nations Conference on Trade and Development,
Stabilization of International Commodity Markets, (E/CONF.46/8)
(New York, 1964), p. 14.

 

19

process of production. Where the products are agricultural,
supply is irregular; attempts to regulate it are clumsy and
achieve moderate success only where there is considerable
capital available to counter the unpredictable forces of nature.
Where the products are more directly extractive, as in lumber-
ing and mining, supply requires considerable investment, and
reliability of supply depends upon the accuracy of estimates
concerning the size and quality of reserves. Contrary to in-
dustrial production, primary production in low-income countries
is liable to be controlled by a very large number of individual
producers, especially in agriculture. The stubborn independence
of these producers adds to the problems of organizing the pro-
duction process. The unpredictable character of natural oc-
currences is amplified by the necessity that investment and
planting decisions be made well in advance of the harvest. If
these complications are added to the fluctuations in demand
which occur in the purchasing countries, the foundations of
short-term instability in primary product prices may be seen.

The severity with which these price variations affect
the producing country is further modified by the problem of ad-
justing the supply to the demand. The production process is
such that an input-to-output time lag limits any such effective
action. A crop planted in April in light of existing market

prices may be harvested to a very different price situation in

""1". M-

.......“~31 U:

1..

‘ ~§~run tre
455‘ tlallv
L

20
September. Attempts to control the supply in the producing
countries by holding excess supply during years of low prices for
later release during low production-high price years has failed
for several reasons. Either the price trend is downward over the
long-run and the government after a decade finds itself holding
large stocks of unwanted or very low price commodities, or the
individual producers, realizing that prices are not going to
drop below a certain level regardless of production volume,
merely increase production to gain higher returns. Both situa-
tions have the same result, namely the government eventually
finds itself holding large quantities of low value goods. Under
these circumstances, there is little benefit in encouraging pro-
duction with the aim of gaining capital for investment through
taxation.

\leuffer stocks and marketing boards can hold prices ap-
proximately stable, then, only when there is no change in the
long-run trend of consumption and price fluctuations are
essentially random. If demand increases, stocks will eventually
become depleted, but if demand declines, the program will become
very costly, requiring the disposal of large quantities of pro-
duce without allowing it to reach the world market where prices
would only be further depressed.

What is equally important, fluctuations in demand are

often brought about by production variations in the industrial

nr"
dul-

ES

tri

F""
b»"‘

FAN—O
5"“.
Q.“ .
C.u::YlES

:r‘v .
“an pro:

IRCT

H
\

2W”

21

countries and are not under the direct influence of the exporting
country. There is no assurance that a high production year in
the exporting country will be coincidental with a low production-
high demand year in the purchasing country, for many primary
products exported from low income countries compete in part with
similar products originating in the higher income countries.
Fluctuations in price, then, may result in an even slower accom-
modating response from the producer than initially considered.

The impact of such a situation on the primary producing
countries is serious. Most underdeveloped nations are also
primary producers with a majority of their export (foreign ex-
change) earnings based on two or three commodities. In many
instances, these commodities are agricultural and can expect a
slower increase in demand than manufactured goods as incomes
rise. Throughout the less developed world, rapid and general
economic improvement through development has become a dominant
concern. If development investments were planned from year to
year, variation in foreign exchange earnings would mean annual
variation in the degree of achievement for these development
plans. Furthermore, investment plans are not made merely from
year to year, but usually in increments of from three to seven
years. An unforeseen drop in world prices for the major com-
modities of a primary producer, and a corresponding drop in

export earnings, would be an even more serious blow to such

:1 s :1;
S U‘:
lave:

e
a: two

- .
‘n’?

paa-

.- ._C

__.....a~¢
‘

c?

9‘ s

22
investment planning.

Not only are low income countries limited in their
sources of investment funds, then, but they are also subject to
wide fluctuations in the earnings based on their most sub-
stantial source of funds. It becomes crucial under such circum-
stances that expenditures be made only on those aspects of the
national economy which will result in the greatest benefits rela-
tive to the amount invested. This is only economic self-interest
in any case, but when there are few funds available for develop-
ment, it heightens the urgency of a "waste not, want not” policy.
It is in this context that we now turn our attention to the first

of two investment sectors important to economic development.

The Agricultural Sector

Agriculture is no longer viewed by most development
economists as an unpromising sector for investment. During much
of the 1950's, agriculture was viewed as a sector which provided
manpower for industrialization but was a decidedly secondary
element in development stimulation. This attitude was largely
based on the assumption of the zero marginal productivity of

labor in agriculture in most underdeveloped countries.9 That is,

 

9For one of the earlier arguments for such an assumption,
see, W. Arthur Lewis, "Economic Development with Unlimited
Supplies of Labour," Manchester School of Economic and Social
Studies (May, 1954), pp. 139-92.

P‘s

“dd
1 b"

l

and i
I tum-cut a
.~r wi“

‘u

:‘ny
.u‘ae
L.
«fun.
.~

23

the productivity of agricultural labor was such that a relatively
large fraction of the rural population could be removed from pro-
ductive effort in agriculture without causing a drop in total out-
put. The validity of this assumption has been opposed recently
by Schultz10 and others, and while this argument has not yet been
settled to the satisfaction of all its participants, there is
less justification in ignoring agriculture's possible role in the
development process.11

The most widely supported position among development
economists argues that neither industrialization nor agriculture
alone is sufficient to provide the impetus for national economic
development. This "balanced" approach states that both agricul-
ture and industrial production are necessary in some proportion,
for without adequate support for production in one sector, the
other will be stalled by a "bottleneck" of low demand, low source

of investment capital, lack of adequate supplies, and so forth.12

 

10Theodore W. Schultz, Transforming Traditional_Agricu1-
ture (New Haven, Conn.: Yale University Press, 1964).

11For a review article of the methodological developments
on this topic, see, Charles H. C. Kao, Kurt R. Anschel, and Carl
K. Eicher, "Disguised Unemployment in Agriculture: A Survey,"
Agriculture in Economic Development, Carl Eicher and Lawrence
Witt, eds. (New York: McGraw4Hill Book Company, 1964), pp. 129-
44.

12For a review and discussion of the major theoretical
arguments current among economists concerning the "balanced" ap-
proach to development, see, Lawrence W. Witt, "Role of Agricul-
ture in Economic Development -- A Review," Journal of Farm
Economics, XLVII (February, 1965), 120-31.

nLITQUI rema-

.

.55C1“?‘.rrv
b~s ~“

‘F‘ mwa

~s ..... .55 [ha

24

Without reviewing the argument in more detail, it will suffice
here to point out that the importance of agriculture to under-
developed economies is apparent from the proportion of the popu-
lation so engaged and the degree to which many such countries'
export earnings are based on crop production. The precise
importance of agriculture is difficult to determine, and in fact
is so intertwined with other factors within the various national
economies that it may never be possible, or worthwhile, to
abstract its importance in detail. It may be useful at this
point, however, to describe the manner in which agriculture can
contribute to development, and some of the problems that may be
seen associated with these contributions.

One of the most basic contributions of agriculture
lies in the increased food available for home consumption. As
national population grows, there will naturally be an increased
requirement of food. Secondly, if there is a rise in per
capita income brought about by development, there will be ad-
ditional need for greater food supplies. The level of supplies
needed will be described by the income (rather than price)
elasticity of demand for food; in low-income countries this
tends to be higher than in high-income countries. Thus, demand
for food will rise relatively sharply during the early stages
of rising incomes in those countries which are less able to in-
vest in agriculture to meet this rising demand. Finally, be-

cause of the rural-urban migration, relatively more food will be

v-~

. a...
. .
t5.

25

consumed in urban areas; more food will have to be moved to the
cities. The importance of movement facilities in this is often
overlooked. If there is a considerable population shift, the
entire marketing system will require transformation.

(’In addition to providing more food to meet increasing
domestic use, there are several other contributions which
agriculture can make to the development process. As indicated
earlier, increased foreign exchange can be earned through greater
crop export if prices on the world market do not drop too sharply.
Increased productivity (per man-hour) in agriculture will theo-
retically release manpower for expanded use in non-farm sectors
of the economy. An expanding agricultural sector may also be a
growing market for industrial goods. These goods may be useful
inputs for agriculture, e.g., fertilizer, or they may be con-
sumer goods, more appealing to those with some surplus income.
And as also indicated earlier, agriculture may be considered a
source of capital funds to be invested in industry or social
overhead. Whether these funds are withdrawn from agriculture as
taxes on land, on agricultural inputs, or through controls on
prices of agricultural products is subject to individual vari-

bility.13

 

13For discussions on these topics, see, Bruce F. Johnston
and John W. Mellor, "The Nature of Agriculture's Contributions
to Economic Development," Food Research Institute Studies,
Stanford University, I (November, 1960), 335-56; also by Johnston

"natural a."

.
2'.-

era

eta

ch

:u
.:

1r cantr

s... K '
~..:

. aS<

fitlv

.1,
4C

26

The Transport Sector

 

‘- Closely associated with development in the more ap-
parently productive aspects of the economy is development of the
structural and institutional framework within which the pro-
ductive sectors operate and upon which they depend. In low
income countries, such elements of social overhead capital as
education and health facilities, and portions of national infra—
structure such as electrical power grids, large-scale irrigation
works, and transport and communication networks all compete with
each other and the agricultural and industrial sectors for
scarce investment funds. The intuition that these "non-produc-
tive" elements of a modern economy are vital to the growth and
development of the economy is widely held.

The extreme difficulty in determining the degree of
their contribution, precise or relative, has led to investment
procedures which do not do justice to the scarcity of the funds
available. As Adler has remonstrated, for example, "it is fre-
quently assumed that all transport improvements stimulate
economic development. The sad truth is that some do, some do

not, and that even some of those that do may not be economically

 

and Mellor, "The Role of Agriculture in Economic Development,"
American Economic Review, LI (September, 1961), 566-93; and
Mellor, gp..gi£., pp. 3-129. For another approach to the same
general problem, see, Simon Kuznets, "Economic Growth and the
Contribution of Agriculture: Notes on Measurements," Inter-
national Journal 2; Agrarian Affairs, III (April, 1961), 59-75.

 

 

 

27
justified in the sense that there may be better investment op-
portunities."14 As Hirschman has pointed out, this may be one
reason for the high interest in the transport sector by devel-
opment planners. The obscurity of the data before investment
insures acceptance of the plan, and the similarly vague data
which issue after the investment insure that no Specific portion
of the plan may be blamed if it fails.15 Although this is a
trifle unfair, for transportation is essentially a means in
development, not an end in itself, planners do seem to forget
this when particularly enamored with the concept of tranSporta—
tion's contribution to development.

When attending to the necessity and potential of trans-
portation for economic development, it is not difficult to lose
perSpective for the possible contributions of transport improve-
ment are many and varied. On a very elementary level, an
improvement in transport means that goods may be moved more
cheaply, more safely, and more dependably through space. In

another dimension, better transport also means faster movement

 

14Hans A. Adler, "Economic Evaluation of Transport
Projects," Transport Investment and Economic Development, Gary
Fromm, ed. (Washington: The Brookings Institution, 1965), p. 189;
also see, George W. Wilson,_gt.al., The Impact gf Highway Invest-
.ment on Development (Washington: The Brookings Institution,
1966), p. 174.

 

15Albert 0 Hirschman, The Strategy of Economic Develop-
ment (New Haven, Conn.: Yale University Press, 1958), p. 84.

b—‘

.1..'

t

 

. Y. 2. a H. L L.
S fix at .L 0 C s c :9 t ”V. 0
9c 11. T. .1 7.. r1 ”L. C b t 1 .
PA. a... «a. w. T .t .s . my n t
.t ... .1 .G : . 5c .1 .d‘ 1 .1.
: e C C C s c S .2 .Q C
{L a .Q {L .2 a a
i a .C 1 Nb al.. .
z a b a .: n.» .,1 a» D
.Z a P; {L r» lg \c\

28
16

of goods. To some economists, expensive transportation means

low revenue to those areas distant from markets, and therefore

17

an underuse of resources. Hawkins, on the other hand, has

laid emphasis on rate of movement, in addition to that of substi-
tution for less efficiently used resources.

The saving of capital brought about by the increased
speed of transport enables more and different types of
investment to be carried out with the same quantity of
savings. In under-developed economies this is important,
since capital tends to be the scarcest factor. It is also
a fertilising factor because its wider use will provide em-
ployment opportunities for the use of surplus land and labour.
There are many grounds for thinking that this is the major
connection between economic development and transport improve-
ments. Economic growth depends upon the growth of capital
employed per capita; transport improvements release working
capital which can then be used more productively as fixed
capital elsewhere in the economy.18

One of the more immediate results of increased transport
efficiency appears to be a rise in agricultural production in the
area through which the route passes. Although such cause-effect
relationships are impossible to determine with certainty, i.e.,
it is not possible to distinguish between rises in production
caused by the better transport and those which are only coinci-

dental to road improvements but would have occurred even without

 

16For a short discussion of these, see Wilson, g£_al.,
op. cit., p. 7.

17Mellor, 22. cit., p. 340.

18E. K. Hawkins, Roads and Road Tran5port 12_§g Under-
developed Country: .A Case Study of Uganda (Colonial Research
Studies No. 32; London: H. M. Stationery Office, 1962), p. 25.

 

x

‘\ '-
.:J.:f1'1::&:t C \

‘525 t... F.“

-»..tS t3

‘-
.‘\. 1

1’3 t

7 D ~COI§3 {I
”3‘32; . ..‘
“\M ‘3‘"
luszr \..
'h 1&1 hes
.3‘5 ‘

.. (1'1“ wk

('2 s
NL‘Po,‘ '

‘ b —. .
a \€C1

29
the better means of movement, empirical encouragement for theo-
retical considerations has come from several studies. In Uganda,
the research of Smith19 and Hawkins20 has reached conclusions

similar to those illustrated by March21 and outlined by Wilson.22

Hunter also had observed this phenomenon in an area of Ghana.23
In this area, a zone of coastal scrub with no transport facili-
ties, a sparse population maintained an isolated existence; the
nearest road was too far to headload produce. After the Ghanaian
Government constructed a road into this area, there was a fairly
rapid influx of food crop farmers, apparently because of the new

access to markets. In any case, real benefits are gained by in-

vestments to improve transport facilities, as shown by the rise

 

19N. D. S. Smith, "A Pilot Study in Uganda of the Effects
upon Economic Development of the Construction of Feeder Roads,"
Research Note No. RN/3408, Department of Scientific and In-
dustrial Research, British Road Research Laboratory, February,
1959 (unpublished).

20Hawkins, op. cit.

21Jose J. March, "Impacto Economico y Social de los
Caminos Vecinales," Mas Caminos, Mexico D. F., (November, 1956),
as referred to in: Organization of American States, General
Secretariat, Department of Economic Affairs of the Pan American
Union, Farm to Market Roads in_Latin America; their administra-
tion and financing (Washington: Pan American Union, 1964),
pp. 8-9.

 

 

 

22Wilson, gt al., op. cit., pp. 127-61 passim.

23Communication from Professor John M. Hunter, Michigan
State University, July 25, 1967.

1: traffic UhiCh

On a mar.
strage faciliti:
efficiency with
grds. largely I:
scale agricultur-
5211 units and
large number of
:a;ita';, emplsy '
Cfipltal and mucr.
is the loss to :
Mme, and M45
ture of his far."

_ Transpor

"iv-«l '
LCHJlUtlon in O.

 

 

30

in traffic which new highways, especially, seem to generate.24

On a more detailed level, the absence of suitable ’
storage facilities and cash reserves has had an impact on the
efficiency with which local transportation is used for marketing
goods. Largely as a result of these two factors, the small-
scale agricultural producer must market his harvest quickly, in
small units and at fairly short intervals. "This postulates a
large number of intermediaries, who, because of the high cost of
capital, employ methods of transportation using relatively little
capital and much labor."25 The obvious result, at least in part,
is the loss to the farmer of a sizable portion of his available
income, and this in turn maintains the marginal subsistence na-
ture of his farming.

(Tfansportation is also an important factor assisting the
revolution in outlook necessary for widespread and successful

development within a country. The extreme importance of the

proper marketing facilities when dealing with l'traditionally

 

24See the conclusions on this in Wilson, 35 _l., 22. c1t.,
pp. 177-78.

25P. T. Bauer and B. S. Yamey, "Economic Progress and
Occupational Distribution," The Economic Journal, LXI (December,
1951), 744. For another approach to this problem, see, United
Nations Economic and Social Council, Economic Commission for
Africa, East African Transport Problems in Relation to the
Promotion of Economic Development (Progress Report), (E/CN.
14/148) (January 24,1962).

":e underestittazc
the rise in last
Eruzner argues,
. 1“,
c: naverent .""
fully tested, '1:

the extension

..
lated distric:s H
2'; feeling to b:
triad range of I
tion of the pc; I
in this than an

V

and ECQaniC CC"

 

31

26 The easier move-

minded producers" is pointed out by Walker.
ment of ideas as well as the people who transmit them should not
be underestimated as a stimulus for change. As illustrated by
the rise in land values coincidental with better accessibility,
Brunner argues, "people are willing to pay for the convenience

of moving easily far more than the difference in the actual cost
of movement.”27 Although such a statement would have to be care-
fully tested, it does have a cynical ring of truth. More plainly,
the extension of good transport facilities into previously iso-
lated districts tends to weaken local loyalties and allow nation—
al feeling to become stronger. New movement facilities offer a
broad range of opportunities directly and to a greater propor-
tion of the population; road transport has greater potential

28 Internal political

in this than any other mode of movement.
v’

and economic cohesion is a vital element in generating this
development, for

. the dominating characteristic of the new African
national economies is their disjointedness, their lack
of internal links and connections. The first step to in-
crease the production of agriculture and to bring the

 

26David Walker, "Problems of Economic Development of East
Africa," Economic Development for Africa South 2; the Sahara,
E. A. G. Robinson, ed. (London: Macmillan, 1965), p. 110.

 

27C. T. Brunner, "Developing Countries: Roads Economic
Justification," British Road Federation, Bulletin, no. 310
(December, 1962), 239.

 

28See the discussion in Wilson, 33 al.,.gp. cit., pp.
198-202.

cial nee-2' for 17

it I.:1 do he “13
1"\\
. 4.. .
a. tier. F
//

:resent demand .
Fresert facilit.
ft: transport i-
tite may waste :
Beer. required t
railways, such .
Tanganyika Grog
E‘Gi‘n

‘flJ. 165 Of th"

transpartat inn.

 

32

subsistence sector more fully into the exchange economy

should be expansion of the internal transportation system,

followed by improvement of the marketing system generally.

While it may be accepted that, in general, there is cru-

cial need for improved transportation in the new African states,
it will do no harm to re-emphasize the misgivings of Hirschman
and Adler.3O Furthermore, transport investment zgfsatisfy
present demand often creates excess capacity for future needs.
Present facilities will affect future costs. If future demand
for transport is underestimated, the higher costs of that future
time may waste resources greater than those which would have
been required to upgrade the original investment. Abandoned
railways, such as the one built to support the abortive
Tanganyika Groundnut Scheme in southern Tanganyika, are extreme
examples of the waste possible from overestimation of demand for
transportation. It is well to be committed to worthwhile goals,
but if in practice, decisions based on this commitment are made,
without empirically testing the assumptions upon which the com-
mitment, in turn, is based, serious waste of investment resources
can result.

Agriculture itself differs from other economic sectors

with respect to the form in which information is useful for

 

29William 0. Jones, "Increasing Agricultural Productivity
in Tropical Africa," Economic Development in Africa, E. F. Jackson,
ed. (Oxford: Basil Blackwell, 1965), p. 47.

30See footnotes l4 and 15.

n ; M~A4..

Fry-"H

.hv‘

'

 

r . . C o; l . ~ . r .
. . A Rd» a~ ‘ h A I a .
r. rm . . .C s .. : . . 3 3. 7 .3. M. d
6 a. r a r C a in c: C .C r. O O S t
r C. —c. an . x CC u“ .. x .C .1 C r CL .Q r: O
a .C l .a C U .1 C r .3 Q C r S
9» F: a» at. : 6.. S *5 t A: A: “U. H.“
E a l _ u I a w u T. S C 1 W d 3 C
a H a t L n P. .s. an. n _... .9. .1... ..t {a .l .1
_ . .f .. .. t r. J c .l .. . E S S :i S pk P;

33
estimating transport requirements. This has been pointed out
by Fromm:
Long-range agricultural planning has usually been limited
to indicating production goals by food stuff types and

means of increasing production through seed, fertilizer,
mechanization, water, and credit programs. These programs

are not readily translated, even in the short-run and

especially in the longer run, into tranSport needs;ji-/
The implications of such an observation should have great im-
portance to non-geographer development planners. Fromm's
immediate implication is that the temporal perspective gained
through indication of goals in essentially static terms is not
meaningful in the context of transport planning which must be
dynamic. What may be more important, or is at least more often
ignored, is the intrinsically spatial asPect of transportation,
in sharp contrast to the spatially indeterminate character of the
most common measures used to increase agricultural production.

Agriculture, then, may be referred to as an important

sector in almost all underdeveloped economies. It may not be the
sole foundation for a country's economic development, and in a
few cases it may not even be necessary. That it must be con—
sidered important, however, is borne out by the many contribu-

tions to a state's economy that it can make, and the high propor-

tion of low-income countries' population which remains employed

 

31Gary Fromm, "Design of the Transport Sector," Transport
Investment and Economic Development, Gary Fromm, ed., op. cit.,

p. 96.

 

i

4.“

Ta

ral p

l
5..

pl.

.3
W.

“V.

thE

. .;
Du

2s

VL

34
in, or associated with agricultural production. Transportation
is definitely a necessary element in any national economy,
although there is some doubt that it is a factor sufficient by
itself to generate development.

Both the role of agriculture and that of transportation
in the development process are, therefore, in need of more care-
ful empirical analysis. That these analyses should be integrated
at least in part has been indicated by the authors of a United
Nations report on transportation and development in West Africa.

Improved transport must, for its full effect be combined

with a corresponding expansion of production and trade.

The latter must have reached a certain level before ex-

penditure on infrastructure is contemplated; and to judge

when that minimum has been attained is one of the most dif-

ficult problems in framing development policies. ’
That is to say, rationally planned expansion in either agricul-
tural production or the tranSport facilities which will move
these products presupposes a certain degree of expansion in the
other. Before an increased output can be efficiently marketed
(assuming capacity use of movement facilities previous to the
increase), there must be some improvement in the means of trans-
porting the larger production from the farms to the market. Con-
versely, and more difficultly, it may be wasteful to use scarce

resources improving transport facilities unless there are suf-

ficient increases in productive traffic prepared to use these

 

32United Nations, Tran8port Problems in Relation to Eco-
nomic Development iE_West Africa, (E/CN.14/63) (1962), p. 64.

 

 

better faciliti

. .
2115 DIESEHC WC
.

 

35

better facilities. It is specifically to this problem that

x

the present volume is addressed.

r;

a. p.

.m |..

nous that

u“

D .

v

D“%‘
“Jnllc diL"
‘1,

ECO

CHAPTER II

GEOGRAPHY AND ECONOMIC DEVELOPMENT

The Necessity pf 3 Geographic Perspective

 

 

The present volume has been written by a geographer. Be-
cause of this, it will be seen to treat economic development in a
manner different from what is ”customary" in development studies.
Although the impact of a geographer's approach on some of the
problems of economic development should become more apparent as
this study proceeds, it is thought relevant here to examine more
specifically, if only briefly, this approach and contrast it with
the way in which other disciplines, and most other geographers,
have treated the subject. Economic development is, after all, a
difficult topic to force into one or two academic disciplines.

2 Research which deals with economic development must be, in
the end, interdiSciplinaryli Whether the study in questiOn deals
with only one aspect of the development process or attempts to
treat reality to some degree, the study must extend beyond tra-
ditional discipline boundaries. In the latter case, it should be
obvious that reality is not bound by academic tradition; in
economic development, administration, agriculture, economics
(business and "pure”), geology, politics, psychology (group and
individual), soil science, zoology, and many other "separate"

36

ug'ns 3f 3039:?”
ad ge3grapuc .
"‘ciallv separa:
ecazeric lines.

. . » p g '
.ll; r)~:tnL [nag

Almast .-
'.':l‘.'es trar.s:i.~
is transmitted '

ltsvledge of d"

 

37
realms of academic study are tightly interwoven within an historical
and geographic framework. Even when there is an attempt to arti-
ficially separate a portion of reality for more detailed study along
academic lines, the study is valueless, in a pragmatic sense, to
the extent that it succeeds in eliminating the "complicating” fac-

tors of other disciplines. The degree to which ceteris paribus is

 

maintained at the conclusion of the study, explicitly or by as-
sumption only, the less real that study is likely to be, regardless
of any theoretical value such a study may have.

Almost all human activity takes place in space, i.e., in-
volves transmission of something between two locations. That which
is transmitted may not be visible on the landscape (information,
knowledge of decisions or innovations, etc.) except when it has
been utilized at the goal of its movement. Goods transport, popu-
lation migration, and military troop movements are all examples of
man's activities which are obvious aspects of the visible land-
scape. All too often, studies of these manifestations of human
activity unrealistically treat these phenomena in an abstract man-
ner, totally outside their spatial framework.

Economic development should, by definition, include as
much of reality as possible without losing sight of the value of
generalization. Both "economic" and, as defined here, "develop-
ment" are concepts with anthropomorphic associations. As such,

studies in either should include within their framework of assump-

38
tions the spatial nature of human activity. The reluctance of most
"pure" economists toward this idea has been amply demonstrated by
the necessity of some of their more reality-conscious brethren to
set themselves aside as "regional scientists."

Given the importance of space in any study of human activ-
ity, it appears strange that almost every published study concern-
ing economic development and containing some recognition of the
heterogeneity of spatial phenomena is either descriptive of existing
or past patterns, or when dealing in the abstract, treats space as
something homogeneous in relatively large blocks or as a simple cost
function within a larger economic model. Thus, Myrdal and Hirschman

l

treat rich and poor regions to be compared as wholes. Among many

others, on the other hand, Isard and Vining include "transfer" or
distance costs in discussing economic areas and economic systems.2

With so few exceptions that they are made prominent by their scarci-

ty, approaches to problems in economic development generally do not

 

1Gunnar Myrdal, Economic Theory and Underdeveloped Regions
(London: Duckworth, 1957); Albert 0. Hirschman, The Strategy pf
Economic Development (New Haven: Yale University Press, 1958).

 

 

2Walter Isard, Location and Space Economy (New York: John
Wiley & Sons, Inc., 1956), and also Isard's "Distance Inputs and the
Space Economy," The Quarterly Journal pprconomics, LXV (1951), 181-
98 and 373-99; Rutledge Vining, "Delimitations of Economic Areas:
Statistical Conceptions of the Spatial Structure of an Economic
System," Journal pf the American Statistical Association, XLVIII
(March, 1953), 44-64.

 

.2 T a».
H‘ :a

“r. .C .C
a . ...
. . .w. .H.
p... .2 «d
v. r 7..
.~ . >.. a c
a On .‘w

w t

C
2. .2
Rs. .\.A
~A P.
e'nu > .
V. A‘v
A... .Q.
.._._ 1

.59.... ‘
~§§CEJIL.
-

.
."
N~

-
. .
at.:?“

an

.at

H‘
-p
u

.a

1315

'f pr:

."
h

39
accept and treat the spatial aSpect of reality in its full complex-
ity.

In any absolute sense, this is not an easy task. Even dis-
regarding the vertical dimension of movement, i.e., the fact that
airplanes fly over, not on, the earth's surface, earth space is
three-dimensional. The well-discussed problem of representing any
sizable portion of the earth's spheroidal surface on a flat map, to
say nothing of relief variations, are several of the difficulties of
working with reality. There is no reason, however, why the Spatial
distribution of phenomena, the agents acting on them, and their
effects cannot be accepted rather than be ignored because they are
difficult to fit into a model, or cause mental Strain when there is
an attempt to conceptualize the problem in this manner. Granted,
the diversity and spatial heterogeneity of reality is difficult to
manipulate mentally; this is a basic value of cartographic repre-
sentation. It is through the imaginative use of maps and all they
represent that geographers can contribute greatly to the solution
of problems of economic development.

Geographers Should have the most suitable training for
dealing with real problems such as those in economic development.
Geography is necessarily interdisciplinary. Like history, it is

distinguished from other disciplines by its approach rather than

C

. .
filr‘

S Li

‘.
.1

5“

ha:

0
u

.n
.u

. V‘- F I
g.a':..':
.

f.--
A
..

.5

a,»
. a
3:
C

it

'p.
..

'vv-
I..

u ~

8
5‘51!) a"...
um

11

.H\

40
the subject matter it studies.3 Geography also deals with reality
in that it is primarily concerned with spatial patterns and rela-
tionships as they exist, existed, or are expected to exist, on the
earth's surface. Basically, it is this concern that has led many
geographers to intensive regional descriptions of places, areas,
regions, portions of the earth's surface, other geographers to
abstract that which deals with the spatial characteristics of a
real problem or phenomenon and present it as geographic analysis,
and Still other geographers to formulate models (one factor of
which is a distance function) in an attempt to make the analytical
process more rigorous, i.e., less liable to subjective qualifica-
tion and human error. In each of these ways, geographers have made
numerous contributions to the professional literature.

As the reader may have realized, however, in none of the
categories listed is there necessarily an attack made on real-world
problems in the manner described at the beginning of this chapter.
Geographic Studies of aspects of economic development fall into two
major groups. Geographers approach economic development either at
the extreme macro—level or as essentially unique "case studies" of
portions of the earth's surface. Furthermore, geographers using

"areal differentiation" and "spatial interaction" when giving

 

3For a discussion of this concept, one which has had con-
siderable methodological impact on the field, see, Richard
Hartshorne, Perspective pp the Nature pf Geography (Chicago: Rand
McNally & Company, 1959), esp. pp. 12-21.

 

.T

,—.v

5.-

‘ s

.u.

.-

.\~

.uk

C»

“H .

 

41
attention to the economic development concept, use these approaches
in a descriptive rather than a prescriptive manner. Examples may

be appreciated at this point.

The Case Study

 

There are innumerable possible examples of the case Study
type. This has resulted in part from the methodological emphasis
within geography on regional studies, until recently taken to mean
in-depth study of some part of the world that had not been studied
before, this tendency being reinforced in turn by the pressure for
"original" research--and what could be more original than Study of
an area no one else knows anything about? The great number of case
studies has also resulted in part from the fact that although it is
difficult to do case study research in the field, for data must be
collected on virtually every aspect of the region, it is much more
difficult both before and after the research to think of a general
principle to be tested in the study area or to analyze the results
to provide new insights into areal characteristics of which the
Study area provides one example. This is not to say that such case
studies are not useful, for they do provide information about many
parts of the world for which there are no ready data. One can
argue, however, that such studies' usefulness might be enhanced by
a hypothesis or method explicitly tested in the "case" context and

conclusions drawn concerning its utility.

P
«n‘ ‘S’ h" "' '
.u: J; In: S 5 ““

' C

Q.
l
on

n.
._.
' 3
('1
ID
' ‘
n;
U)
(‘V
, .
I J
7'

. .‘ 7
~- ..v- ~ . ‘.
L: u....K'~.\

 

 

 

42
A good example may be made on this point by reference to
one of the studies in the University of Chicago Geography series.4
The interesting thing about this thorough case study is that while
it is unlikely that a non-geographer would have undertaken the
research necessary for this Study, or written in the Same way,
there is almost nothing in the approach which calls attention to
the author's presumed training in geography. As legitimate re-
search, there is nothing against a geographer ”doing" economics or
history, or economists and historians "doing'I geography, as long
as the studies are well done. But a geographer who does not bring
something of the unique approach to reality that is part of his
discipline to bear upon the research problem under study sacri-
fices a greal deal of his potential contribution to the problem's
solution.

AS pointed out earlier, however, this non-Spatial type of
contribution from geographers has broad foundations in the rather
recent evolution of the discipline. In 1953, at a meeting of
geographers explicitly concerned with economic development, one of
the participants argued "that the geographer's contribution lies
in his ability to evaluate the interrelationships of various proj-

ects in over-all balanced development and that administrators with

 

4Darrell Randall, Factors pf Economic Development and the
Okovango Delta (Department of Geography Research Paper No. 47;
Chicago: University of Chicago Press, 1957).

 

43
a technical background often posses [Epig_7 a limited perspective
of the broader problems with which they must deal."5 Four years
later, another geographer wrote that a geographer contributes to
national development by pointing out the regional importance of
conservation and by functioning as a resource person to banks and
commercial firms, providing information on regions of the world

"6 In other

as "background to investment and trading policies.
words, geographers are to assist development indirectly and in
part, passively, by fulfilling the role of walking, talking ency-
clopedias and reference manuals.7
More useful are case studies designed to examine in detail
an assumption, hypothesis, or methodological approach to an eco-
nomic development problem. The Study by Larimore and several by

8

Hunter provide excellent examples of case Studies of this nature.

 

5Attributed to Frank Keller at ”Program Presented at the
49th Annual Meeting by the AAG Committee on Economic Development,"
Lloyd D. Black, Chairman, Professional Geographer, V (July, 1953),
16-17.

 

6B. J. Garnier, "Geography and National Development," The
Nigerian Geographical Journal, I (April, 1957), 4-5.

 

7For a well-done example of this approach, see, William A.
Hance, African Economic Development (New York: Harper & Brothers,
1958).

 

8Ann Larimore, "A Measure of Economic Change: Sequent
Development of Occupance in Busoga District, Uganda," Essays pp
Geography and Economic Development, Norton Ginsburg, ed. (Depart-
ment of Geography Research Paper No. 62; Chicago: The University
of Chicago, 1960), pp. 111-23; John M. Hunter, "Cocoa Migration

 

44
The study by Hunter on migration of cocoa farmers, for example,
provides not only an extraordinary amount of information about
the small areas concerned, most of it gathered ”fresh” in the
field rather from published and unpublished documents, but it also
gives an example of centrifugal and centripetal ("separatist” and
"unifying") factors bearing upon the new community made up of these
migrants. It is meaningful, then, to those concerned with develop-
ment in other parts of Africa in its indication of precisely where
non-European cultural traits may affect national unity. Because
of its completeness, the study also provides Specific data con-
cerning the manner in which these separating and unifying forces
have arisen for their particular Study area.

One unfortunate aspect of such studies, however, and one
which is understood by those aware of the dangers of induction,
results from unwarranted extension by those not aware of these
dangers. The implication may be taken that what has or has not
worked in area "A" will probably have the same results in area
"B." Researchers making this induction may not then carry out a

similar Study in area "B," as should be done, but use this assump-

 

and Patterns of Land Ownership in the Densu Valley near Suhum,
Ghana," Transactions and Papers, The Institute pf British
Geographers, no. 33 (December, 1963), 61-87; John M. Hunter,
"Population Pressure in a Part of the West African Savanna: A
Study of Nangodi, Northeast Ghana," Annals pf the Association pf
American Geographers, LVII (March, 1967), 101-14.

 

 

55p€Cia11..

V

f

Jr a more

45

tion, base conclusions on its veracity, and eventually, political
and economic decisions may be made on this basis. The failure
here is, of course, not on the part of the original researcher
but on those who make unfounded extensions of the initial work.
This is not a call for fewer such useful case studies, therefore,
but for more of equal quality aimed at Similar problems in other
parts of Africa.

That variety of case Study in which geographers use their
disciplinary training directly in developing a methodological ap-
proach of use to development planners in many countries is
especially valuable. Although as early as 1953, Vinge had called
for a more intelligent use of maps, beyond their role as "mere
illustrations,"9 the studies by Gosling10 and Moss11 are unfortu-
nately rare examples of this kind of presentation. Gosling at-
tempted to determine indices (e.g., rice production per agricul-
tural worker, Stored rice reserves per capita, and others) by

which areas relatively less suitable for development investment

 

9Clarence L. Vinge, "The Language of Geography,” Pro-
fessional Geographer, V (May, 1953), 8-9.

 

10L. A. Peter Gosling, "The Location of 'Problem' Areas in
Rural Malaya," Essays op Geography . . . , pp. cit., pp. 124-42.

11R. P. Moss, ”Soils, Slopes and Land Use in a Part of
South-Western Nigeria: Some Implications for the Planning of
Agricultural Development in Inter-Tropical Africa," Transactions
and Papers, The Institute of British Geographers, no. 32 (June,
1963), 143-68.

 

46
than other areas might be determined. This study, however, ap-
pears to assume administrative districts suitably uniform with
respect to these indices, for data are presented in this manner.
This is undoubtedly caused in part by the way in which the data
were gathered, but it is an unfortunate expediency. Furthermore,
in the case of this article, an unstated assumption could limit
its usefulness in more sparsely populated areas. That is, that
existing utilization of agricultural resources reflects the
absolute potential of these areas is an assumption that can be
made only with a serious danger of superficial assessment in non-
densely populated areas. In spite of this, the article is a use-
ful attempt at determining the areal variability of development
potential.

The article by Moss could be of even greater methodo-
logical utility. By combining the mapped distributions of soil,
soil parent material, slope, and existing land use, Moss was able
to form conclusions regarding a basis for regional comparison of
reaction to changing agricultural techniques. Not only is the
study useful for the detailed data presented and the substantive
conclusions made concerning the local ecosystem and planning for
agricultural development, but it is also important in demon-
strating the utility, even the necessity of dealing with the

Spatial variability of reality. Thus,

. r,,___

'k-neut.

1‘5 cf the
Betcrt

C7“;

EiCns AS Vl
there is "1

“All C h
search on as“
t.

D

{Tel n

 

 

47

In order adequately to plan agricultural development some
assessment of the impact of new crops and techniques upon

the ecosystem must be made, an end which may be achieved by
study of the relations of the ecosystems at present existing,
and by detailed investigation of mechanisms within them. In
either case detailed knowledge of the distributions of these
systems is an indispensable preliminary, and valid applica-
tion of the results also depends upon informative mapping.

It is this use of "informative mapping” that most closely identi-
fies Moss's article with the approach to be used in the present

volume.

Models of the Macro-Region

 

Before proceeding to the main body of this work, a few
comments should be made concerning the other major category of
geographic study of economic development, i.e., those at the macro-
1evel. By the term "macro-level," reference is not being made to
the in-depth case studies of large regions, but to discussions of
the variability of characteristics between relatively large re-
gions. As with discussion of the case study type of research,
there is much that is commendable in macro-level geographic re-
search on aspects of economic development. The Aplgp 2£.§EQEQE£E

Development by Ginsburg, for example, provides a tremendous amount

 

of information, containing the country-by-country variation of no

less than forty-eight indices of development.13 Furthermore, the

 

121bid., p. 165.

13Norton Ginsburg, Atlas of Economic Development (Depart-
ment of Geography Research Paper No. 68; Chicago: University of
Chicago Press, 1961).

48
statistical analysis of these differences provides insights to
the patterns not easily apparent.14 In sum, however, these pat-
terns and the analyses of them are descriptions of average dif-
ferences and can be of only marginal use for decision-makers in
the underdeveloped world. They may be encouraged, for example,
to discover that ". . . equatorial location apparently has little

"15

to do with stage of economic development and enlightened

that ". . . political colonialism, seems not to explain poverty
but even in some instances to have countered it . . . ,"16 but
it will not help planners decide where and how to use scarce
investment resources.17 Academically of interest, these studies
survey economic development problems with much too broad a brush

to contribute more than marginally to the increased well-being of

 

1['Brian J. L. Berry, "Basic Patterns of Economic Develop-
ment," Ibid., pp. 110-19.

151bid., p. 119.
16Ibld.

17For further examples of descriptively interesting studies
on the macro-level, see: Brian J. L. Berry, ”City Size Distribu-
tions and Economic Development," Economic Development and Cultural
Change, IX (July, 1961), 573-88; the articles by Wagner, Guyol,
and Berry in Essays op Geography . . . , pp. Ei£,, pp. 49-62,
pp. 65-77, pp. 78-107, reSpectively; and in a somewhat different
manner, Norton Ginsburg, "Natural Resources and Economic Develop-
ment," Annals of the Association of American Geographers, XLVII
(September, 1957), 197-212.

 

C. . . . 4 .t .
ad . a . x Cc E .l ‘ . 1
.5 8 "Nu N T. C Pr .t ., -. 71 FL T» a . l a v . n- l a. u. . ~ e 2
S . . an n. .1. .t C a C .3 U .3. a 3. V. a , 3 .t at D.
.t a C 5.. v» ha 4. T. t .1 C r. 7 T. n. .L C n.»
1 x n n .3 .V a r r v . x x .C .d a (L f; Cc x5 Y.“ l A 3
a at a C T n1. 9. .1 .C . .. .1 .: Pu an a.” A: W»
.n. f v. .... x r e a. . x w. M “a a r . t
. I L . - . .. . L. ..s . pl“ .HH v . .3. . . .
A r. t a .3 a . V. V. .al. h C la; Q. ?s t C ~l¢ at am .m A...“

a P4

49
any less developed country's population.18
Studies which use statistical and mathematical techniques
of analysis for model building often have the generally descrip-
tive character of macro-level studies whether examining large

19, 20

areas or small. For example, in an article written

specifically toward the problem of decision criteria for devel-
opment planners,21 Reiner becomes "hoist with his own petard”
when he attempts to regionalize variable phenomena and ends up
restricting their variability. Thus, he complains that "tra-
ditional texts and studies of economic development give scant

heed to subnational analysis, decisions, and aspirations.”22

 

18An exception to this generalization is the study by
Gould of one aspect of an entire country's development. Perhaps
by maintaining this focus throughout, Gould's work has taken on
some characteristics of a case study while remaining committed to
the analysis of the changing pattern of spatial interaction and
the impact of some of these changes on the development of the
country as a whole. See, Peter R. Gould, The Development pf Egg
Transportation Pattern ip Ghana (Studies in Geography No. 5;
Evanston, 111.: Northwestern University, 1960).

 

 

191 have the suSpicion that topological models would be
an exception, although I must confess ignorance of the principles
involved in topology.

20Gould's study is again a partial exception; Gould, op.
cit , pp. 96-114, passim.
21Thomas A. Reiner, "Sub-national and National Planning:

Decision Criteria," Papers, Regional Science Association, XIV
(1965), 107-36.

22Ibid., footnote, p. 109.

caflSll’L‘é’d ill or
“'2 and anaé'f
Finally. 11
hetercgene‘
misfit be a.
meaninglcS
there is D
equality m
regionaliz

And so, becaus

is "harness-n17...

In 81;.“
ducted by gram
development C'
aspects of le
CPEERt plann:

13 be concert

of phenom-

l..t‘l‘.a

phers intere-

 

afld Moss, up

 

his, ”The D

50
Before long, however, he indicates his assumption of, or predilec-
tion for, space that is made up of a number of relatively homo-
geneous areas; the implication is made that Space must be so
construed in order to be usefully included in development plan-
ning and analysis. He writes:
Finally, limits to the acceptable range of homogeneity or
heterogeneity within a given region must be set. Thus, it
might be argued, interregional equality is relatively
meaningless where, within regions there remains (or possibly
there is built up) greater disparity. Such internal in-
equality may be expressed spatially, in which case a finer
regionalization might resolve the problem.23
And so, because of a commitment to model-building, spatial variety
is "homogenized" and therefore made unreal.24
In sum, studies of economic development problems con-
ducted by geographers have varied from descriptive typologies of
development characteristics to detailed examinations of individual
aspects of less-developed countries, a few with utility for devel-
opment planners but most without- Although geographers profess
to be concerned with "areal differentiation," or the variability
of phenomena from place to place, there are only a few geogra-

phers interested in economic development, such as Hunter, Gould,

and Moss, who perceive this variability in a useful, analytical

 

23Ibid., pp. 115—16.

2Z'An indication that this need not hold when quantitative
techniques are applied with care and in circumstances where data
are available in suitable detail is provided by Julian Wolpert in
his, "The Decision Process in Spatial Context," Annals pf_£pg Asso-
ciation pf American Geographers, LIV (December, 1964), 537-58.

 

. n

u-

"

h‘
s:

r-O

51
manner. Even Hartshorne, the single most important geographer
in spreading the acceptance of the concept of areal differentia-
tion among members of the profession, sees generalized descrip-
tion of development as "a primary task for the geographer." More
explicitly, "a primary task for the geographer . . . is to find
ways and means of providing a more realistic description of the
current, and changing situation in economic development over the
world than that provided by statistical tables which list coun-
tries as though all of one kind.”25 There can be no doubt that
Hartshorne is aware of the variability of reality. His emphasis
here, however, remains on description of the development process
and its effects rather than on use of geography's approach in
the solution of problems associated with this process.

There are several implications which arise from geogra-
phers' abdication of their discipline's full utility in testing
problems of economic development. The least important is that
there has been a gradual deterioration of the discipline's
academic "reputation," as other social scientists have often
come to identify geography with the trivial case study form of
research. This is important, not because of any diminution of
the validity of the geographic perspective, but in a reduction of

expectations and a tendency to ignore the contributions which are

 

25Richard Hartshorne, "Geography and Economic Growth,”
Essays pp Geography . . . , pp. cit., p. 21.

present in the

Hare c:
:f approaching
As c'iscassed e

:azj: research t

tiers overlap .

.“P,' . .
“rifting 1n :

pcpflatiOU cha

trattntse s:
The P
u

 

 

52
present in the geographic literature.

More crucial is the continuing disregard of the necessity
of approaching reality in both a temporal and Spatial framework.
As discussed earlier, this minimizes the total applicability of
many research contributions to economic development problems;
the heterogeneous character of the real world makes less secure
the extension of isolated research suggestions for solution of
the immediate problems of decision-makers in underdeveloped coun-
tries. After all, it is only by indicating where abstract solu-
tions overlap with related problems, of a different nature but
occurring in the same location, whether dealing with institutions,
population characteristics, agricultural economics, or whatever,
that these solutions may be easily utilized.

The present volume attempts to face each of these
objections, at least by example. Although dealing with a rela-
tively small portion of the earth's surface it is not a case
study of the superficial variety. A methodology is offered which
is applicable in detail throughout much of the underdeveloped
world. This methodology is presented with clear recognition of
points at which further research is required and where more
Specialized, aspatial research may be necessary to provide pre-
liminary answers to problems raised. Most importantly, the study
treats space not merely as something to be described but as a part
Of reality that must be accepted in all its variety when dealing

‘With real problems.

PART II

METHODOLOGY

A REGIONAL A?

As ha~'

eccnamically

vestment, or a

the nation's
Yield a retur-
Pelitical COT.-
33‘31 of deYi—l
VGSt. Ill Ills.
”"9“ and tr.
Vlfl1a 10W 3?
trua- Of EEOSt
POlic} Vlllch .
for political;

The i
facilities f.
also been P0:
v'COl’I‘ectn 0r

Whether by

it

SEctor, , I

13 u:

 

CHAPTER III
A REGIONAL APPROACH TO RECONCILING INVESTMENT PRIORITIES
IN AGRICULTURE AND TRANSPORTATION

As has already been discussed, there is a vital need in
economically underdeveloped countries to minimize "wasteful” in-
vestment, or at least to maximize investment in those areas of
the nation's economic life which will be the most certain to
yield a return. This need is pressing not only because of the
political considerations but also because given the accepted
goal of development, there is a severe shortage of funds to in-
vest. In those countries where population growth, rapid urbani-
zation and the rising tide of personal aspirations are combined
with a low average level of personal income (characteristics
true of most of Asia, Africa, and Latin America) a governmental
policy which wastes funds by poor investment decisions is begging
for political disruption, if not disintegration.

The importance of both agriculture and adequate transport
facilities for a general improvement in individual welfare has
also been pointed out. The argument among economists over the
"correct" or the best path to national economic development,
whether by industrialization or concentration on the agricultural
sector, is unimportant here. It is sufficient to point out that

54

attic:

'

.‘t'fiv‘
'V—J-‘gc

.

In
.M

.c
.u an a.
v.1» L. 4.. .. .. have
.x on. .nm -._ . .
b. 1

r.\
.8 \

55

agriculture is the major occupation of the great majority of the
population in these countries. If improvements in income are to
come in the foreseeable future to more than a handful of people
in such countries, they may be most easily and profitably
achieved in agriculture; agricultural products, methods, at-
titudes, and marketing conditions are all improvable aspects of
the means of increasing average incomes. It is primarily with
this last aspect of agriculture that this study is concerned.

In the present attempt at developing a procedure to
minimize investment decision errors, attention is focussed on
the relative costs of farming and possible income which could
be obtained from better use of resources and better facilities
for marketing farm produce. Estimates of profit are obtained
from a comparison of costs and returns. This is not to imply that
this study uses the methods of cost-benefit analysis, often de-
scribed in these terms, nor is a sophisticated input-output model
attempted. The general language may be Similar because the
larger concepts are the same; it is to a great extent definition-
al. That is, the concept of "profit” is defined as the differ-
ence between costs and total returns. As presently constituted,
however, the methods of cost-benefit analysis and input-output
analysis are incomplete at best, and grossly misleading at worst,

for application in less developed economies although attempts at

greater refina
tcllected for

incarrect beca .
share the nun: "

'
A

'33 ies into wt.

“at is lfipor 1

 

lTlEC-l’y 32d 31'.

they have bErr

aEYLCultural

56

greater refinement of such methods continue to be made.1 Data

collected for use in such models are often inadequate, or even
incorrect because of errors in the collection procedure. Even
where the numbers are correct, there is no guarantee that cate-
gories into which the data are placed do not mask or distort
what is important and uniquely "non-Western" about the economy.
It has been attempted here, therefore, to maintain simplicity in
theory and approach real problems as they exiSt rather than as

they have been supposed to exist by some economic theory.

Costs and Income

 

In the development of a useful method of determining
agricultural and transportation investment priorities, it is
important to begin with a thorough understanding of the economics
of the existing situation This means that existing levels of
costs and returns must be established preliminary to a determina-
tion of existing levels of profit, or net income.

Direct costs to the farmer may be viewed in two parts:
the costs of production and the costs of marketing the produce.
Costs of production will vary in many ways. The type of crop

grown, the suitability of the land for production of each crop,

 

1For example, see the dissertation, Henry Malcolm Steiner,
Criteria for Plannipg Rural Roads ip p Developing Country: .333
Case pf Mexico (Project on Engineering - Economic Planning, Report
EEP-l7; Stanford: Stanford University Institute in Engineering -
Economic Systems, 1965).

 

 

 

'1

pr!

57

variations in taxation, crop labor requirements and many aspects
of the individual farmer's make-up, e.g , commitment to farming,
receptivity to new methods (which often involve more capital cost
but do not necessarily result in a proportional increase in out-
put), etc., are all aspects of production costs. As such, it is
important that their variation from place to place be determined.

Production costs are admittedly one of the most crucial
elements in a program of improved farming. Without knowledge of
such costs, there can be no clear understanding of the motivation
of a farmer to grow one crop rather than another, or indeed, not
to grow any crops at all. It is apparently only a recent revela-
tion to Western economists that perhaps the subsistence farmer is
not acting "irrationally" in refusing to grow certain crops or
in maintaining subsistence food production along with minor in-
vestments in cash crop production. And yet, in Spite of the im-
portance of knowledge on production costs, very little research
has been conducted to gather such information. Although blame
for such a lack may be accepted because of the difficulty of
gathering correct data--a generally illiterate, often suspicious
farm population do not make data collection easy even assuming
adequately educated and motivated data collectors--surely the
present situation has been made worse by the biases of research-
ers who avoid collecting precise data on costs because "it is too

difficult to gather and it is probably rather like costing in .

'F. P mar.
Lu 5 .‘n';3\

or item

“L Ci:

(1"
'1

'-I
(I)

58

(Yorkshire, Illinois, the Rhone Valley, etc.) anyway " The fact
that most African farming is obviously not similar to European
or American agriculture has not persuaded more than a few research-
ers to undertake studies to determine the elements of costing in
small-farm African agriculture.2

In comparison to costs of production, costs of marketing
are deceptively more easy to determine. Essentially, they are
comprised of the cost of moving the produce from the fields to
the market, and any costs which may arise from charges levied by
a middle-man or marketing board. Since movement costs are costs
per unit distance, it is theoretically possible to determine
actual movement costs on the basis of the mode of transport used,
the type of goods shipped, and the distance to the market. This,
however, provides another example of the difficulties which may
arise if theoretical considerations are not examined in the light
of circumstances prevalent in the region with which the research-
er is concerned.

In order to be of practical value, a theoretical approach
to a problem must at least imply the exceptions or variations
which will be encountered in the problem's Specific examination.

That is to say, theories should not be so generally stated that

 

2For a recent condemnation of the results of ethno-
centricity in economic development research, see, Polly Hill,
"A Plea for Indigenous Economics: the West African Example,"
Economic Development and Cultural Change, XV (October, 1966),
10-20.

«urn

s...»

.y»
in

i. .

59
they lose all meaning when attempts are made to test them
empirically. In the present example, unless it is known whether
farmers carry their goods to market on their backs, by bicycle,
or by wagon, a real estimate of transport costs cannot be made.
Perhaps more importantly, more difficult to determine, and there-
fore less likely to be measured, the inclination of the farmers
to use a more highly mechanized form of transport to market should
be known. If private trucks are not used to pick up produce close
to the field in which it was grown, it is very important to know
whether this is because of high deterioration costs on trucks re-
sulting from the low quality of roads in the area, greedy middle-
men who demand high rates for movement of the goods, the farmer's
perception of the value of their own time, the distances involved,
or incidental benefits derived at the market place, e.g., social
contacts or relief from relative isolation, Unless such attitudes
and behavior patterns are well understood, decisions should not
be made concerning improvement of transport facilities, primarily
roads, which may have no effect on relative costs to the farmer,
thus limiting the expected response of greater production.

The average farmer's gross income may be more easily
determined. For present purposes, agricultural output not con-
sumed on the farm is considered to have cash value, regardless of
its cultural context. The price and the quantity of produce sold

provide the relative level of gross farm income. Such minor or

60
basically non-farm sources of income as sale of mulch or occas-
ional pieces of timber, share of bridewealth, and so forth, can
usually be excluded from any calculation of total farm income;
they are too irregular and most often a very small proportion of
the total.

With costs and total farm income determined, it is
obviously possible to calculate the spatial variability of net
farm income, or profit. On a regional as on an individual basis,
the profitability of farming varies inversely with total costs;
in areas of high total cost, whether caused by production or

marketing, profits will be lower, ceteris paribus. The difficulty

 

is that rarely are "other things equal." Better prices may be
obtained by farmers frequenting some markets rather than others
on the basis of quality, reputation, or favoritism. Without
adequate supervision, farmers in one area may be given less than
the standard market price by an unscrupulous trader, a practice
made worse by the lack of accurate information on such matters
available to the farmer. Nevertheless, a map of total cost of
production, given all the variability of the elements of cost,
subtracted from a map showing total value of existing production
(i.e., volume multiplied by price) will result in a map illus-
trating the variability of farm profits, or spatial differences

in "profitability of farming" at the present time.

61

Potential

The next element in the method of determining transport
investment priorities is that of total production potential.
This pattern may be best derived and mapped by the use of eco-
logical regions and the suitability of each such region for
agricultural production.

Although there are numerous definitions of the concept
of ecology, the basis of almost all involves the interrelation—
ships between, even the interdependence among the various aspects
of the environment. This general juxtaposition is dominated by
the study of the interdependence of living things with respect
to each other and also to the impersonal forces or actions of
nature. Thus, the way in which man affects the "natural balance"
and conversely, the way in which man is affected by nature are
both seen as legitimate fields of research open to the ecologist
For the present purposes, the manner in which plants of use to
man are limited by environmental characteristics is of primary
concern.

At a given level of technology, the upper limits placed
on the productivity of a given crop are set by the suitability
of environmental characteristics in the area in which it is being
grown. Other elements of agricultural production act upon these
limits so that they are rarely, if ever, achieved. But as a

determinant of the absolute yield, it is the environment which

62
is crucial. Thus, incomplete knowledge or a value system which
does not contain maximizing output as a relevant issue will pre-
vent the environmental bounds from being approached. Consistent
with this argument is the position which holds that the limits
which man places on the production of crops by his own behavior
are not as immutable as those levied by nature.

The errors contained within this argument, however, pivot
on the assumption of a given level of technology. As with Malthus,
the fact that a change in agricultural technology, in effect,
changes the environment cannot be ignored. Irrigation, fertili-
zers, and the steel-bladed plow are clear examples of techno-
logical innovations which brought about changes in the effective
environment. They have permitted crops to be grown in areas
which could not support production at the same level before the
new tool or technique was developed.

It is important, however, to remember the use to which
the concept of ecological regions is to be put. The concern
here is with total agricultural production potential. Changes
in environmental suitability for agriculture based on future
technological developments are unpredictable, and therefore ir-
relevant for the present purposes. They are important in theory,
but not pertinent in practice until they occur. Furthermore, if
a study area is divided into a number of distinct ecological

regions, i.e., "homogeneous" regions defined on the basis of

63
soils-vegetation-relief-climatic association, a technological
improvement affecting the agro—suitability of one or more of
these regions will merely increase that region's production
potential by some factor. In turn, this will affect the present
attempt in substance but not in style or validity.

These, then, are the next steps in the procedure under
consideration, the method by which a priority for regional in-
vestment in agriculture and transport facilities may be objective-
ly determined. Ecological regions, which by their internal simi-
larity of environmental characteristics indicate a basically
similar support for plant growth, must be delimited. Each eco-
logical association indicates a potential level of agricultural
production per unit area. Once the estimates of these levels
have been made, relative cropping capacities may be determined
and total value of these production capacities may be areally in-
dicated for the entire study region. By then subtracting the

Spatial distribution of existing cost of farming, as determined

 

above, from this map illustrating total potential income from
farming, a map of the total potential profitability of farming is
obtained. By then subtracting the spatial distribution of the
existing profitability of farming from this derived estimate of
total potential profitability, a final map will result illus-

trating the variations in profit imppovements possible in farming

 

within the study area.

64

The general concept of the use of ecological regions to
define agricultural potential is not new. Mr. L. H. Brown, agri-
cultural economist, ecologist, natural historian, and ex-civil
servant in Kenya, outlined in a number of unpublished reports
the ecological regions of two of the major agricultural areas in
Kenya.3 In these reports, Brown describes the major ecological
divisions within these areas and then offers combinations of farm
production which would take best advantage of the environmental
characteristics of each such sub-region. The immediate purpose
of these reports was to indicate reasonable production goals for
African farmers and guidelines for approaching these goals. The
purpose of the present paper is to make a similar use of these
ecological regions, and then extend the results to achieve much

broader ends.

Movement Facilities

 

The final task preliminary to completing a methodology
for determining investment priorities for agricultural transport
facilities deals with the costs of goods movement. Assuming for
the moment that no human variables are involved, the cost of

moving goods is proportional to the type of good, the mode of

 

3For a more available publication of the statistical sec-
tions of the report not used in the present study, although
presented with no reference to Mr. Brown's authorship, see, Eric
Clayton, Agrarian Development 12 Peasant Economies (New York:
The Macmillan Company, 1964), pp. 35-39.

 

 

6S
transport, the quality of the road surface, the rate of movement,
and the distance traveled. Bulk goods are less expensive to
transport than freight which requires special or individual
handling. The smaller the volume moved and the shorter the
distance it is to be carried, the more costly it is, per unit
volume or per unit distance, to transport the freight by railway
relative to road carrier. Maintenance costs and rate of vehicle
deterioration rise sharply with a decline in quality of running
surface. Unless the freight has high utility, and therefore high
value, the increase in speed of carriage beyond a certain point
becomes less economic, i.e., more costly in proportion to the
returns gained by faster service. Finally, in absolute terms,
it always costs more to move goods a greater rather than a lesser
distance; more out-of-pocket funds, more time, more effort is
involved. In a precise accounting of transport costs, all of
these factors must be considered. The present task will treat
each of these aspects of transport costs, but will do so to
varying degrees of thoroughness.

As mentioned earlier, man and his idiosyncrasies cannot
be so easily ignored when theory is applied to reality. In terms
of transport costs in a facility-poor and an agricultural region,
this can mean many things. To a farmer who pushes his produce to
market piled high on the rear of his bicycle, the "effective" cost

may be no greater on unimproved earth roads than on gravel roads.

66

If his farming is such that he has extended periods of non-
working time, a greater distance to market may not be perceived
as involving greater costs, whether in time, effort, or capital.
An improvement of facilities, therefore, would not result in
greater individual profits in such instances, because these
aspects of transport costs were not seen to exist as costs.

Further such complications arise if a middle-man is in-
volved in produce transport. As often occurs in parts of the
world where small-scale cash cropping is beginning in isolated
rural areas, an entrepeneLLr-trader may grasp the opportunity to
gain profit from the wide separation between farmer and market,
separate in physical distance and in knowledge of market prices.
The "marketing agent" travels through the rural area in a truck,
purchasing produce at a price enough below the market price to
obtain a profit for himself after paying his own transport costs.
If the farmers do not know the price their produce would obtain
at the market, there may be little reduction in their total
costs of production even after road improvements; the agent
would simply receive a larger profit. On the other hand, if the
farmer does know the market price, or becomes aware of it after
improvements in road facilities have been made, there could be a
dramatic shift in methods of marketing, perceived costs of
marketing, and thereby a change in total output. It is important,

then, before attempting to apply the method being developed here,

67

to determine the farmers' degree of awareness of true market
prices, their reaction to these prices in terms of the decisions
made to gain full advantage of such prices, the importance, if
any, of perceived transport costs to their production efforts,
and if possible, the differences between the true cost of
transport and the rates charged by goods carriers, whether con-
tracted or itinerant.

As will be seen in the application of the present
methodology, many of these complications of reality have not
been taken into account. This is due to expediencies involving
time and resources, for most such research, designed to make
theoretical derivations more realistic for a specific region,
requires considerable time and manpower. In the application,
then, the following assumptions are made: costs of movement
over land distances are those of road motor hauliers in all cases,
unless specifically designated otherwise; a reduction in transport
costs will result in a comparable reduction in the total average
production costs of the farmers; a reduction of total average
production costs will lead the farmers to attempt to approach
more closely the maximum potential production of which their land
is capable. With these costs determined and the assumptions
understood, it is possible to outline the final phases of the

investment priority determination technique being developed here.4

 

4Some readers might ask: What of the assumption that a
decrease in costs and a proportionate increase in profit will

68

At any given level of production costs, quality of road
surface, and volume of goods produced, the priority for invest-
ment in transportation facilities in agricultural areas may be
easily determined. By determining the changes in movement costs
which would follow from an equivalent investment in each road in
the study region, depending in part upon the amount of traffic
to be carried and variable with the existing quality of the roads, 0
a map of relative road investments may be constructed. This map
may then be compared with that map derived earlier illustrating
the variations from place to place in the potential improvement
of farming profitability. Obviously, then, those areas with the
greatest improvement potential which can be reached by the least
investment in upgrading transport facilities should be the first
areas to benefit from reduced movement costs. That is, those
roads connecting areas of greateSt potential profit improvement
while themselves requiring the least relative investment should
have the highest priority in an investment schedule for road

facilities.

 

necessarily result in a greater desire to produce? Can it
legitimately be assumed that more money for the same effort will
induce more work in order to gain even more money? It can be
assumed, but it should be tested. The present study, however,
is not concerned with this problem. The focus, again, is con-
centrated upon the reasons for making one decision rather than
another, i.e., on investment decisions. The question is pppz

To invest, or not to invest? but whether objective justifica-
tion may be made about a decision concerning investment in an
A-type or a B-type project.

69

This procedure is one which may be made dynamic. That is,
with suitable adjustments in several of the factors, a series of
differing investment priority schedules may be determined. Changes
in volume, and therefore total value of crops produced, in any of
the agro—ecological regions will result in an alteration of the
difference between total existing and total potential profitabil-
ity. Thus, as regions are improved and more produce grown, they
are less likely to justify further transport investments unless
they have considerably more potential relative to the other parts
of the study area. Furthermore, by gradually upgrading in the
procedure the road quality of those roads which have been improved,
the ratio between costs of investment in movement facilities and
the level of profitability improvement possible in the areas to
be connected with market outlets will also change. That is, as
higher investment costs are required and existing profits are
increased, the overall differences between improvement costs and
improved profits will be diminished, and a change in investment
priorities will result.

A final note should be added to this outline regarding
non-economic considerations. Ideally, there should be no con-
flict between economic and non-economic pressures on decision-
makers. In reality, however, an efficient economic decision may
be a despotic decision. Political or religious expediencies often

force economically sound undertakings to be set aside, modified,

70

or in some way diminished in effectiveness. The present author

has no quarrel with such considerations.

They are made or not at

the option of the national leadership, or the people as repre-

sented by this leadership. It is simply

this volume, or indeed, the jurisdiction

to take such considerations into account.

a method for making objective decisions,
here, could be implemented without undue
it is recognized that such a hope may be

recognized and it is admitted, but it is

not within the scope of
of its author to attempt
While it is hoped that
such as the one presented
personal dislocations,
somewhat fatuous. It is

not excused.

PART III

APPLICATION

CHAPTER IV

THE COSTS OF MOVEMENT

A basic concern in determining total costs of production
to small-hold farmers in rural areas of Africa is that portion of
costs required to overcome location separation of the producer
and the consumer or his agent. Transport charges may be a crucial
element in the productivity of an area. Transport costs are
"crucial" not necessarily in the sense of raising output per unit
input or per unit area, but more in the distinction between
limited, subsistence production with an occasional surplus and
commercial production with considerable commitment by the farmer
to production for sale. The distance between production and de-
mand centers is obviously of little relevance to the producer if
his primary goal is subsistence.

There are numerous components of transport costs which,
in different combinations, result in varying levels of movement-
vehicle operating cost. These costs may be classified as fixed
or variable. For highway carriers, the former include such
factors as vehicle purchase cost and licenses, terminal construc-
tion and maintenance, administrative wages, and the like. Fixed
costs are those which must be borne regardless of the volume of
traffic moved. Variable costs, then, include costs of fuel and

72

73
oil, tires, vehicle maintenance, and in some proportion, vehicle
depreciation and drivers' wages. In the following discussion,
and in construction of the map of tranSport costs, it is the
strictly variable costs of fuel, oil, tires, and maintenance
which are used as a measure of these costs. In the present case,
these variable, or "out-of-pocket" costs will be taken to ap-
proximate the transport cost to the small-scale agricultural
producer. This may be justified not only on the basis of data
availability, but also in the argument that since few such pro-
ducers will own their transportation, there will be a tendency
for trucking charges to fluctuate with the amount of freight car-
ried. Freight charges, reflected by operating costs, will also
fluctuate with several other factors. It is to these factors
that the discussion would now be turned but for an objection that
may have arisen in the mind of the reader.

It may be objected that there need be no relation between
variations in cost, borne immediately by the transport operator,
and the rates charged the farmer for shipment of his goods. The
former figure, after all, does not include operator profit while
the latter does. Since it is cost to the farmer which is of
concern in this volume, considerable error is possible by substi-
tuting operating cost for transport charges. In reply to this
argument, it must be remembered that entry into the trucking

industry is relatively easy. Without restrictive licensing, there

74
is some evidence that road transport is viewed by African entre-
peneurs as an area fertile with development possibilities.l Under
such conditions of competition, discriminatory pricing between
commodity classes will be minimized and charges will be closely
related to costs.

Data, unfortunately, are not available to substantiate an
assumption of large volume entry into the trucking industry by
numerous individual Kenyans. There is some question over the pre-
cision of road license statistics, but as Table 4-1 indicates,

the trend in commercial vehicle licensing is unclear. It may be

TABLE 4-1

NEW MOTOR VEHICLE REGISTRATIONS, 1960-1965

 

 

1960 1961 1962 1963 1964 1965
Motor cars 5,007 3,653 4,305 4,466 4,058 5,031
Utilities 2,914 2,118 2,422 2,542 2,974 3,402
Lorries 857 714 618 799 658 785
Buses 243 186 206 289 281 252
Motorcycles 1,151 634 536 602 650 729
Other 740 452 557 753 980 850
Total 10,912 7,757 8,644 9,451 9,601 11,049

 

Source: Kenya, Economic Survey, 1966, p. 56.

 

too early after true independence, however, for indications of

 

1See, E. K. Hawkins, Road Transport_ip Nigeria (London:

Oxford University Press, 1958).

 

75
entrepeneurial activity of this nature to be reflected in owner-
ship or registration statistics. It should also be pointed out
that many African smallholders engaged in cash crop productions
operate through co-operative marketing societies. Under such an
arrangement, profits on transport of produce are negligible since
most societies are large enough to operate their own "lorries,"

or trucks, and maintain administrative operations by more direct

levy.

Factors pf Transport Costs

 

Vehicle operating costs will vary under a number of dif-
ferent operating circumstances. For example, the quality of the
road surface will affect out-of-pocket costs. As road quality
declines, maintenance costs increase, as does tire wear. Average
speed is reduced, resulting in fewer trips per week and possibly,
higher fuel consumption. Similar influences on truck operating
costs result from variations in road alignment, gradient, Speed
of operation, volume of traffic or congestion, and traffic
"composition," or the types of vehicles operating in what propor-
tion. Of all these factors of haulage costs, those relating to
road conditions (surface, alignment, gradient) are likely to have
the greatest effect on costs in Nyanza Region and Kericho
District. Maximum vehicle speeds are limited in Kenya by weight

class, traffic volume is rarely high enough to warrant a label of

76

"congestion," and the factors making traffic composition important
are a function of gradient and deterioration of road surface.

Of the three elements of route condition, the effects of
alignment, or horizontal road curvature, are the most difficult
to assess. This difficulty arises in the individualized nature
of the source of a high proportion of cost variation based on
this factor. As Winch has pointed out with regard to road align-

ment, the exact correction factors needed in a particular

case will depend on observed driver behavior, and may well differ

2 Therefore,

for different types of roads in different areas."
while road alignment may be a factor of significant weight,
seriously affecting vehicle operating costs, there is no clear
indication that it is such a factor. For present purposes, ef-
fects of alignment will be considered proportional to the effects
of the other factors of road quality, at least within the study
region.

There is an inverse relationship between the quantitative
factor of vehicle operating costs and the qualitative one of road
surface condition. Costs increase as road surface quality de-
clines. Unfortunately, it is rather problematic to determine this

relationship with any precision. In an attempt to discover the

manner in which these costs varied within three general categories

2David M. Winch, The Economics pf Highway Planning
(Toronto: University of Toronto Press, 1963), p. 70.

77
of road quality for five countries in East and Central Africa,
Bonney and Stevens were forced to conclude:
No direct comparison can be made on a cost basis between
the various countries owing to local variations in unit
costs but the results show a consistent pattern although
data on improved routes are restricted. . . . These
variations can result from local standards of road con-

struction and road maintenance, terrain, vehicle age,
traffic volumes and climate.3

In spite of these reservations, it is necessary to discover the
approximate cost-proportions for the several qualities of road
surface.

‘In'addition to the research conducted by Bonney and
Stevens, data on cost variability with road surfaces is provided
by a study in Kenya by Klein4 and one near Lake Victoria in
Tanzania by Halcrow and Partners.5 All three studies list average
operating costs per mile for several vehicle sizes over a number

of road surfaces. Although it would be difficult to defend

 

3R. S. P. Bonney and N. F. Stevens, "Report on the Investi-
gation into Vehicle Operating Costs in East and Central Africa,
Part III. Kenya and Uganda," Laboratogy Note No. LN/606/RSPB.NFS,
Road Research Laboratory, June, 1964 (unpublished), p. 6.

 

4Martin S. Klein, Economic Feasibility pf Improvipg Two
International Highway Links lg Kenya (Cambridge, Mass.: United
Research Incorporated, 1965).

 

5Sir William Halcrow & Partners, Report pp the Feasibil-
ity pf Constructing Roads £9 Serve the pragu Cotton Zone and
Preliminagy Route Location for a Road from Busisi via Buyagp.£p
Kharumwa and for a Link £2 Kasamwa (Ministry of Communications
and Works, Government of Tanzania, August, 1966).

 

 

 

78
comparability of road surface criteria, it is useful to group the
more numerous classes used by Halcrow and Partners, and by Klein,
into the three general categories of paved, improved, and unim-
proved roads after Bonney and Stevens. These operating costs,

and the grouped averages, are shown in Tables 4-2 and 4—3.

TABLE 4-2

VEHICLE OPERATING CCSTS,* OPERATED IN THE
GEITA PENINSULA, TANZANIA, 1966
(EAST AFRICAN CENTS PER MILE)

 

Road type cars, pickups, 3-5 ton lorries 6-8 ton lorries
vans
grouped grouped grouped
Bitumen -/42 -/42 -/62 -/62 -/83 -/83
Gravel "spine" -/53 -/77 1/03
Graded earth: -/59 -/83 1/16
"spur" -/64 -/89 l/28
existing -/76 l/l7 1/59
Track, -/86 1/27 1/77
cross-country -/95 1/37 1/94

 

*Cost components included are only those attributed to fuel,
oil, tires, and maintenance and repair.

Source: Sir William Halcrow & Partners, Report pp the
Feasibility . . . , 1966.

 

Even quick calculation of the ratios represented by the three cate-
gories of road surface demonstrates a surprising similarity among
these proportions regardless of the weight of the vehicle. These

ratios, shown in Table 4-4, are also similar to that determined by

79
TABLE 4-3

VEHICLE OPERATING COSTS,* OPERATED IN KENYA, 1963 (2)
(EAST AFRICAN CENTS PER MILE)

 

Road type Light vehicles Heavy vehicles
grouped grouped

Bitumen -/7O -/70 1/40 l/40
Improved gravel -/84 1/68

slightly impr. gravel -/98 -/98 1/96 1/96
improved track 1/11 2/23

Slightly impr. track 1/26 1/33 2/51 2/65
unimproved track 1/40 2/79

 

*
Cost components are those used by Klein and may not be
directly comparable to the figures in Table 4-2.

Source: Martin S. Klein, Economic Feasibility . . . ,
1965, p. II-15.

 

TABLE 4-4

RATIO OF VEHICLE OPERATING COSTS,
BY ROAD SURFACE QUALITY

 

unpaved
Study vehicle weights paved improved unimproved
Halcrow cars, pickups, vans 1.00 1.40 2.05
Halcrow 3-5 ton lorries 1.00 1.34 2.05
Halcrow 6-8 ton lorries 1.00 1.40 2.13
Klein "light" 1.00 1.35 1.90
Klein "heavy" 1.00 1.40 1.90

 

80
Bonney and Stevens for Zambia, the only country for which an
"improved" category of road was available. The variation in
costs on "unimproved" roads is not unexpected; it is this low-
quality road category which would contain roads most susceptible
to variations in relief, precipitation, maintenance frequency,
and so forth. The single departure from a rather close grouping
of operating costs on unpaved but "improved" roads may be at-
tributable to a great many variables such as those described by
Bonney and Stevens, above. Indeed, one of the most startling
aspects of this comparison of cost ratios on the three cate—
gories of road surface is the general propinquity of the figures
within each road class. On the basis of these figures, the ratio
used in the present volume for vehicle operating costs over roads
which are paved : improved : unimproved has been accepted as
1.0 : 1.4 : 2.0, respectively.

This ratio of out-of—pocket operating costs designates
the manner in which transport costs vary with differences in
road surface quality. Although there are numerous factors af-
fecting operating cost other than that used to determine this
ratio, for simplicity it is assumed that all but one of the ad-
ditional cost factors vary directly and in the same fashion as
costs due to variable road surface quality.. In other words,
the ratio 1.0 : 1.4 : 2.0 for paved, "improved" (gravel), and

"unimproved" (earth) road surfaces, remains constant with the

81
inclusion of the other cost variables. The factor which can-
not justifiably be considered to vary in this manner is cost
variability due to topographic relief, or road gradient.

The effects of gradient, primarily the rate of grade,6
on transport operating costs are numerous. The vehicle's power
unit will have to work harder per unit distance on sloped vis-a-
vis flat conditions. There is an increase in fuel consumption and
an increased time on the road when the route consists of steep
slopes. Furthermore, by reducing the maximum load which can be
carried, it becomes necessary to use additional carrying units,
thus increasing average costs per ton-mile. "There are many fac-
tors governing the extent of these effects, the most important of
which are: rate of grade, length of grade, combination of grades,
over-all rise and fall, approach conditions, terminal conditions,

power of vehicles, and the necessity to change gear or brake."7

A fairly large proportion of the region under examination
in the present work contains considerable topographic variation.
Virtually all of Kisii District and almost all of Kericho District
are very hilly while much of South Nyanza District has rolling to

moderately hilly relief. A truck operating exclusively within

 

6In the United States, anix percent grade is a vertical
rise ofux feet per 100 feet horizontal distance.

7Winch, The Economics . . . , pp. 215., p. 69. For further
discussion on most of these factors, see, William W. Hay, Ag Intro-
duction 52 Transportation Engineering (New York: John Wiley &
Sons, Inc., 1961), pp. 217-22.

 

 

82

these regions would be continuously confronted with relatively
high road grades with an associated increase in operating costs.
At a four percent average grade, fuel consumption is approxi—
mately 15% above that on level roads; at 5 percent grade, fuel
consumption increases to better than 40% beyond that for level
roads; and with a 6 percent grade, fuel consumption is nearly
double the amount averaged on 0-1 percent grade roads.8

The basic concern in the present volume, however, is
with movement from production areas to market outlets, and re-
turn. It is on these costs, under the proper conditions of
competition, that cost to the producer will be based. The two
major outlets for the study region are Kisumu and Lumbwa, both
of these stations being located at altitudes lower than the
region which uses them as railheads. Most of the gain in fuel
costs made by lorries (trucks) transporting goods to these
markets--more downhill than uphill grade--will be lost on the
return trip. Between the fact that grades on most main roads
are less than four percent, and the fact that cost increases or
decreases on inbound and outbound transport will tend to balance
each other, the impact of gradient variation on a map of average

transport costs will be small. In the present volume, therefore,

 

86. J. Roth, "The Boonomic Benefits to be Obtained by
Road Improvements, with special reference to Vehicle Operating
Costs," Research Note No. RN/3426/GJR, March 1959, Road Research
Laboratory, Crowthorne, U.K. (unpublished), see Figure 4.

 

83
this element of cost variation will be excluded. If more detailed
data justify, or are required for, more precise application of
cost factors in some future instance, this judgment may be false.

At the present scale, it is not.

Absolute Transport Costs

 

In order to construct a map of transport cost between
any point in the study region and the major market outlets along
the railway, some estimate must be made of absolute levels of
cost rather than the relative costs described by the ratio
1.0 : 1.4 : 2.0. On the other hand, the problems associated
with determining a single level of cost for each class of road
are considerable, as has been discussed before. By referring to
Tables 4-2 and 4-3, the reader can see that the impact of vehicle
weight on operating cost is as great as the impact of differing
road surface quality. Klein shows that a light weight vehicle,
operating on unimproved track, costs no more to operate per mile
than a heavy vehicle traveling on bitumen. Similar difficulties
are raised because of the inclusion, or exclusion, of different
cost components by different authors. Klein, again, does not
designate the components of his figures of "average operating
costs" per mile; they undoubtedly include such factors as
driver's wages, depreciation, and possibly profit.

Hints toward the solution of these problems come from

84
personal Observations in Nyanza Region and Kericho District during
1966. Very rarely were vehicles above 6-ton tare weight observed
in this region. Almost all were three to five ton lorries, or
the smaller pick-ups and vans. The latter were operated chiefly
by small merchants from towns such as Kericho, Kisii, Sotik, and
the like, carrying non-agricultural goods. It Should be accepta-
ble to assume, then, that under present conditions of vehicle
availability and operation, the dominant trucks transporting
surplus agricultural produce are within the three to five ton
category. The only direct data on lorries of this weight group
is in the Halcrow report, the cost per mile on bitumen roads
being -/62 cents. Confirmation of this general level of costs
does come from the studies conducted by Bonney and Stevens. Also
dealing with vehicles of a wide tonnage range, their figures are
averages of costs for each vehicle and are therefore liable to be
closer to the middle category of vehicle sizes than to either
the very light or the very heavy trucks. By converting their
figures9 from "pence per gross ton mile" to "East African cents
per ton mile," taking into account the relative distances
traveled by each vehicle and the variation in gross weights of

the trucks, the average operating costs for vehicles moving on

 

9The detailed data for Zambia is given in the authors'
"Report on the Investigation into Vehicle Operating Costs in East
and Central Africa, Part I. ZHmbia," Laboratory Note No. LN/344/
RSPB.NFS (revised), Road Research Laboratory, February, 1965
(unpublished).

85
bituminous roads was calculated to be -/61 cents per ton mile.

With these figures, the foundation for accepting an
absolute level of transport operating costs has been laid.
Although the degree of similarity between the data for both the
Bonney and Stevens report and the study made by Halcrow and
Partners may prove coincidental, the fact that they are very
close provides support for the use of a similar cost figure here.
For purposes of simplicity in calculation, the cost of movement
per mile over surfaces of paved, improved unpaved, and unim-
proved unpaved will be taken to equal Shs. -/60, -/84, and 1/20,
respectively. These figures are, of course, based on the ratio
of vehicle operating costs over the three road quality cate-
gories as determined earlier.

Further support for these cost levels is received from
two agricultural officers dealing with movement to and from the
study region. Although actual cost data were not available from
trucking operators, ostensibly on the grounds of inadequate in-
formation, both men provided independent educated estimates which
in their similarity lend credence to the cost estimates accepted
for the present work. The District Agricultural Officer Of
Kericho District pointed out that maize grown in the district may
be sold to the Produce Marketing Board. There is a seven cent

(-/07) cess per 200 1b. bag per mile for transport of this maize

86
from market stall in the rural areas to the Board store.10 The
cost to the farmer for movement of his corn surplus to the Board
store therefore is -/70 per ton-mile. The present author was also
informed by a staff member of the Pyrethrum Marketing Board in
Nakuru that pyrethrum is carried from Kisii District to the
Nakuru factory at about five shillings (5/-) per mile for a one-
way load.11 Since the lorry carries about 130 bags of pyrethrum
flowers at 65 pounds per bag, or approximately 8450 lbs., this is
a cost to the producer of about 1/20 per ton-mile, or -/60 per
ton-mile £§EE.EEX on the predominantly bitumen road. With these
additional indications of transport cost levels, the absolute
costs are well established, albeit in an indirect manner.

Using estimated movement Costs for each category of road
surface, a map was drawn showing trip costs for three to five ton
trucks from any location in the study region to either of the
major produce outlets for the region, i.e., Kisumu or Lumbwa
(Figure 4-1). All road distances were first measured regardless
of road surface quality. Such measurements were made on virtually
every road or track "suitable for wheeled vehicle" in the study

region (see Figure 4-2). In converting these distances to cost-

 

10Interview with Mr. J. Birir, D. A. O. Kericho District,
November 12, 1966.

11Interview with Mr. R. J. Munro, November 11, 1966.

 

14": .31

Ads

 

d

NYANZA REGION and Kericho District

COST-DISTANCE TO
KISUMU OR LUMBWA

FOR 3-5 TON TRUCK

  
   

l:-

 

Total Trip Distance In Shillings

   

 

m. o to |O/- - 60/— to 70/-
IO/- to 2m- 70/- to 80/-
20/- to 30/- 80/— to 90/-
30/40 40/- n 90l-to IOOI-
E 40/- to 50/- - om IOO/-
50/. 10 em-
KENYA

 

 

    

 

 

 

 

 

 

 

FIGURE 4- 1

 

 

 

34°I r ' .',, ' as‘I I

NYANZA REGION and Kericho District

/ j ROADS AND TRACKS
I --

 

 

 

 

— Unpaved Roads

 

 

KEN;:\\\K\\\j/dé>

 

 

FIGURE 4- 2

 

 

89
distances--based on the various qualities of road surface--and
constructing Figure 4-1, a series of assumptions were made. While
these assumptions were necessary for clarity and do not undermine
the theoretical development of the present illustration, they
should be made explicit.

One of the most basic assumptions involved the consistency
of road quality within each category. It was assumed that trans-
port operating costs per mile would be the same over any road
within each category. While necessary, this assumption is not
valid in detail. The unpaved-improved road between Ahero and
Oyugis, for example, was being upgraded during the latter half of
1966 preliminary to bitumenizing. Until funds are obtained to
pave this road, it will continue to appear a more costly route
between Kisii and Kisumu than it actually is. The question may
also be raised whether all tracks rated "suitable for wheeled
vehicles" are actually passable in any season other than the
very driest. The so-called "all-weather” roads are also often
found mis-named during the heavy rains. Without detailed data on
all the roads in question, and their quality variation from season
to season, each category of road surface has been considered of
even quality and possible seasonal variations ignored.

The isopleths of transport operating cost, shown on
Figure 4-1, have been drawn on the assumption that minimum cost-

distance to Lumbwa or to Kisumu was the distance which would be

90
traveled by a lorry from any given point in the region. Thus, a
shorter cost-distance route via an unimproved track may be Shown
rather than a more costly route which may be preferred and
actually covered by an individual driver because it is an improved
road. Similarly, the railway rate differential between Kisumu-
Nairobi shipments and Lumbwa-Nairobi shipments was not taken into
account when Figure 4-1 was drawn. The fact that these railway
charges vary with the product being shipped, and in any case are
relatively low, argued for this exclusion. Finally, isopleths of
movement cost-distance in areas containing no vehicular roads,
only foot- and bicycle-tracks, were drawn on the basis of one to
two miles per ten shillings movement cost. While these costs
would undoubtedly be much larger if the labor involved were as-
signed even moderate wages, it would be unrealistic to do this at
the present time.

Examination of Figure 4-1 shows nothing startling in its
pattern of transport cost levels. The lower cost-distances from
Lumbwa and Kisumu extend outward from these railheads along major
roads toward each other and the interior of the study region.
Total movement costs to these foci merge between twenty and thirty
shillings, with the major trend of these costs increasing rather
regularly beyond thirty shillings to the southwest through Kisii
and South Nyanza Districts, and to the west in that portion of

Central Nyanza. Major variations from this pattern are those

91
areas containing no vehicular roads at all.

Two objections may be raised in regard to the cost-
distance pattern of Figure 4-1. One of these is that the iso-
pleths should be drawn with reference to the nearest railhead,
not only to Kisumu or to Lumbwa. Examination of freight con-
signments in each of the stations between Kisumu and Lumbwa, how-
ever, showed that traffic moving through these railheads was
almost entirely local in character, most of the outbound freight
originating in the large commercial farms north of the railway.
While more of the freight shipped from the stations north of
Kisumu does originate within the study region, much is trans-
ported directly to Kisumu by road. For the sake of consistency,
Kisumu was kept as the sole focus of produce movement in the
western part of the study region.

The second possible objection to the pattern derived
for Figure 4-1 is that since the several ports operated by the
East African Railways and Harbours were not part of the cost-
distance calculations, the map does not Show actual movement
costs to Kisumu. While this may have been a valid criticism a
decade ago, there are indications that the cost difference be-
tween moving freight via the road-lake route and completely by
land is rapidly disappearing. Improved roads, inadequate "go-
down" or warehouse facilities at the ports, speed, flexibility,

and an increasing demand for consumer goods in the interior

92

portions of the study region have combined to encourage movement
to Kisumu from Kisii District and much of South Nyanza solely over
land routes. Most of the parchment coffee, for example, shipped
by the Kisii Farmers Co-operative Union, now moves directly to
Kisumu by lorry. There are cases where lake movement is likely,
but for present purposes these are exceptions and have been
Specifically pointed out later in the volume.

With the spatial variation in transport costs determined
it is necessary to turn to the somewhat more problematic deter-
mination of the spatial variability of agricultural production

COStS .

CHAPTER V

THE COSTS OF PRODUCTION

When the question of crop production costs is raised to
agricultural economists, the responses will be widely varying;
only rarely will they have ready replies. There have simply
been too few studies made in detail to provide definite answers
to this question even in the dollars-and-cents (or pounds-
shillings-pence) framework of Western economics. That this
cultural context may not be pertinent in Africa has been ques-
tioned before.1 Whether because of the wide range of variables
involved, the "irrational" or at least "unclear" motivation of
the African smallholder from Western points of view, or colonial
predilection for focus on total output rather than relative in-
come, formal attempts at objective determination of farm pro-
duction costs in Kenya--and other part of Africa-~have been
quite recent.

It is not difficult to accept the importance of crop
production cost knowledge in newly independent Africa. A great

majority of the population of Kenya are farmers. The well-being

 

1See the discussion of this point in Chapter III.

93

94
of the nation as a whole cannot be improved solely by industriali-
zation in the few large urban centers, although political pres-
sures may be stronger and more immediate in the cities. The per-
sonal income of smallhold African farmers can be increased by
improving the volume or quality of output, thereby increasing
gross income, or income can be made larger by focussing attention
on production costs in an attempt to reduce those costs which are
unnecessarily large. To do this latter, of course, a clear
picture of the various components of cost must be appreciated.
On a practical level, production costs are an element to be in-
cluded in the proper consideration of agricultural loan applica-
tions from individual farmers. In other words, once the objec-
tive of "improved personal well-being for every citizen, urban
and rural" is accepted as a national goal, it becomes imperative
to be capable of assessing input-output relationships with some
detail both for purposes of planning resource allocation and for

more immediate economic operations.

Problems pf Cost Determination

The various expenses to be enumerated and analysed in a
detailed accounting of crop production costs are numerous and
difficult to determine. Many of those costs directly attributa-
ble to a particular crop or stock unit are of a type which varies

with the productive unit. Coffee, for example, requires cost

95
components of fertilizer, mulch, manure, Sprays, and replacement
seedlings in addition to labor costs associated with each of
these plus labor for cultivation, planting, weeding, and harvest-
ing. Such components for cotton, tea or sugar cane are at a
completely different volume, if they are even part of the pro-
duction process. Some crops have high seasonal labor demands
while others may be grown and harvested without any need for
hired labor. And on this last point, on what basis should "un-
paid" family labor be evaluated--if it is to be included at all?
How are joint capital costs (e.g., communal grazing land) to be
subdivided among the several farmers? How can those elements of
cost which apply to the general productivity of an enterprise
without contributing directly to any specific process (e.g.,
materials for structural improvements, fuel for farm machinery
in general use, and other "overheads") be included in a cost ac-
counting of any of the individual crop accounts?2 It should be
evident that it is not an easy matter to correctly estimate the
many directly accountable cost variations from crop to crop and
the other cost variables more indirectly contributing to farm

production.

 

2For one attempt to allocate the various farm expenses,
see, Government of Kenya, The System pf Recording and Analysis
used 13 Production Cost Studies pf Large Scale Farming 35 Western
Kenya (Report No. 17; Nairobi: Farm Economics Survey Unit, Eco-
nomics and Statistics Division, Ministry of Finance and Economic
Planning, 1963).

 

 

96

In addition to the methodological difficulties posed by
these complexities, there are a number of practical problems
raised by the characteristics of African agriculture. Most of
the population with whom cost researchers must deal are illiter-
ate. Information concerning farm costs must either be obtained
by frequent (and expensive) observation of farm operations by
the research worker, or it must depend upon the memory of the
farmer. If the farmer associates farm cost research with the
taxes he must pay, or for some other reason doubts the motives
of the questions being asked him, the accuracy of the data col-
lected will be impaired. Many population censuses taken in
Africa during the latter portion of the colonial period are of
doubtful validity for this very reason. Fortunately, there are
indications that data collected in this manner is available and
fairly reliable, regardless of the degree of farmer literacy, if
sufficient care has been taken preliminary to data gathering. A
study of farm economics in western Tanzania specifically under-
taken to answer such questions resulted in the general conclusion
that "there is no question that the African farmer knows most of
the details which we need to fully analyse his farm. His accumu-
lated experience over a number of years must have left its impres-

sion on his mind, both quantitative and qualitative."3 As the

 

3"Report on Bukumbi Survey, Completed February 1962,"
Western Region Research Centre, Ministry of Agriculture,
Tanganyika (unpublished), p. 23.

97
report proceeds to point out, however, there are two basic
barriers which must be overcome in gaining access to the farmers'
information: the cultural and linguistic framework in which the
questions and answers will be given, and the confidence and
interest of the farmers which must be maintained throughout the
research period. The accuracy of the results are largely a
function of the efficiency with which these barriers are over-
come.

Further problems of determining cost-of-production in
agriculture are raised by variations in environmental conditions
and in quality and quantity differences in yields. It is inex-
pensive to grow a given quantity of rice in well-watered, fertile,
flat terrain which receives warmth all year long relative to the
cost of growing the same amount of rice in a hilly, sub-arctic
environment with a growing season of 80 days and effective
precipitation under 15 inches annually. In the latter case,
obviously, considerable capital in necessary to modify the en-
vironment so that the rice may be produced in yields even ap-
proaching the former environment's output, produced with only a
relative modicum of labor or capital input. For most forms of
agricultural produce, the relative suitability of the environ-
ment is a primary factor to be considered when studying the
variability of production costs.

In reality, it is impossible to set, for a given product,

98
an arbitrary quantity or quality level as a base for comparison
of production costs between environments. The range of yield
and produce quality is likely to be as wide within similar en-
vironments as between different ones. A poor or disinterested
farmer will usually have low quality produce in spite of natural
advantages. On a cost basis, this low level of output would
ideally be represented by a low labor or investment input. How-
ever, such qualitative measures as the attention given by the
farmer to weeds, pests, or market conditions, the desire of a
farmer to provide for himself and his family more than a sub-
sistence level of living, and the felt importance of extra—
agricultural activities which may distract the farmer from his
farm work are all extremely difficult to quantify.

From the foregoing discussion, it is not difficult to
understand the less-than-positive reaction that an agricultural
economist may be expected to make when faced with questions con-
cerning average crop production costs. In spite of these prob—
lems, however, the importance of gaining a detailed understand-
ing of the farm economics of small-farm African agriculture has
been made clear. A number of studies of large-scale and small-
scale farming in Kenya have been completed by the Farm Economics
Survey Unit in response to the need for this knowledge. Although
the cost figures obtained from these studies may be applied,

strictly speaking, only to those areas in which the research was

99
undertaken and for those years during which the data were gathered,
they do provide guidelines for similar research needed in the area
under consideration in this volume as well as tentative levels of
average costs in environmental conditions Similar to those in

which the studies were made.

Reconciling Production Cost Variability

 

It is certainly possible to surmount the difficulties in-
herent in research of agricultural production costs, given suffi-
cient time and money. Cost-of-production is first determined by
crop and stock unit; these total costs are simply the sum of the
costs of those operations directly contributing to each crop's
output. Those over-all farm costs which cannot be attributed
directly to individual crop output are totaled under a general
category such as "overheads."4 In several of the Kenya FESU
Studies, the difficulty of whether or not to include unpaid

family labor is avoided by showing total costs for various crops

both without any value attached to family labor and with family

 

4There is no conceptual reason why several levels of such
general farm costs could not be determined on the basis of grouped
individual farm characteristics, say, by means of a modified form
of factor analysis. This would provide others interested in farm
production economics the means by which total farm costs could
be estimated with a reduced number of factors describing the
relative levels of these costs. To this author's knowledge, such
factors have not been determined for large-scale or small-scale
African agriculture.

100
labor costs included at the rates that would have been required
if all work had been done by hired labor.5 For a given farm,
then, production cost estimates could be based on the quantities
of each crop grown or stock unit carried and the amount of un-
paid family labor available. The subjective element, dealing
with farmer attitudes and the like, would still remain un-
measured.

Much more difficult is the resolution of the problem
posed by variations in environmental suitability for a given
farm enterprise and the accompanying variation in support for
the optimum output of that enterprise per unit input. It is im-
portant to remember here that the Specifics of optimizing crop
yields continue to be rather elusive, even to agricultural re-
search specialists. Environmental optima for any crop must be
given as ranges rather than as singular criteria; it may never
be otherwise, considering the interdependence of the many environ-
mental characteristics even now recognized as crucial to crop
growth. The effect on crop yields of one degree-day change in
temperature may be countered by, say, a 0.3 rise in soil pH, or

two inches greater precipitation during the week of highest

 

SSee, for example, Government of Kenya, "Some Aspects of
Agricultural Development in Nyeri District, 1962," Report N9. 21
(August, 1964), and ". . . 1963" Report N9. 24 (August, 1966),
Farm Economics Survey Unit, Economics and Statistics Division,
Ministry of Finance and Economic Planning.

lOl
temperature, or slightly lower soil porosity, or lower average
wind velocity, or . . . . When the variations in human commitment
to obtaining the highest possible yield, and the continually
changing possibilities for technological modifications of the en-
vironment are considered in all their variegated combinations, it
is not difficult to understand the uncertainty of an "optimum" en-
vironment. That some environments will give a higher yield per
unit cost--or will have a lower cost of operation per unit output--
is certain; that the concept of an optimum environment has any

reality is less clear.

The Precision pf Cost Patterns

 

 

It may be helpful here to review the use to which crop
production costs will be put in this work and the precision re-
quired by this application. The production costs for agricultural
output within the study region are needed to complete the picture
of total production costs which must be borne by the individual
farmers in this region. It is the variation of current produc-
tion costs from place to place, i.e., the spatial variation of
these costs, which is of primary interest. The precision required
in the present work is not such that detailed policy decisions may
justifiably be made from its conclusions. This was explicitly
stated earlier. When the present methodology is used preliminary

to investment decision-making, detailed cost studies should be

102
made on the basis of, perhaps, a one percent stratified sample,
randomly chosen from the African farms located within each of
the ecological regions comprising the study region. In this
volume, however, the pattern of farm cost variation will be
drawn on bases of less practical rigor but of equivalent con-

ceptual soundness.

Assumptions

 

In order that a map may be drawn illustrating the
spatial variation of the major cash crops in the study region,
a number of simplifying assumptions have been made. Again, in
a real application of the present method, these necessary as-
sumptions may be much fewer in number, depending upon the
financial resources available for research. At the scale of the
present undertaking, the only crops included in the study are
those which (1) have been marketed in large quantities during
the first half of this decade, (2) show promise of continuing
to be significant sources of cash income to the African small-
holders in the study region, or (3) are cash crops which are not
yet well established in the study region but could be grown there
profitably if cultural-psychological factors do not prevent their
success. By placing these limitations on the products to be in-
cluded here, marginal crops with a very low year-to-year surplus

for sale to external markets (e.g., finger millet or wimbi, sesame

103

or simsim, grams and pulses, sorghum, castor, most vegetables,
etc.) do not cloud the more individually remunerative cash crop
pattern based on high value, often high output crops. In ad-
dition, possibly transitory crops (e.g., passion fruit6) and agri-
cultural products originating almost entirely on non-African
operated farms (e.g., wheat and wattle) are also excluded. In
spite of these restrictions, the crops which remain are suffi-
cient in variety and number to test fairly the proposed approach
being presented.

It should also be pointed out that each crop comprising
a portion of the pattern of crop production costs should be part
of a system of crop rotation. This rotation may be three years
pasture alternated with three years arable, as is common on the

7

lower land in Kisii District, or some other series. Some form

 

6Passion fruit experienced a rapid rise in popularity
among farmers in Kisii District as the following production fig-
ures indicate:

 

 

Quantity marketed Value recorded
1963 nil nil
1964 316,498 lbs. £2,374.00
1965 2,228,794 lbs. Ll6,715.75

(Source: Department of Agriculture Annual 32‘
port, Kisii District, 1965, p. 14.)

An even more rapid decline in 1966 output was feared; the 1966

crop was being severely damaged by a fungal disease and the future
of passion fruit as a source of income for small-scale farmers in
the district was therefore difficult to predict. - Interview

with Mr. Ogoda, Kisii District Cash Crops Officer, November 14,19ai

7Interview with Mr. Ogoda, Kisii District Cash Crops Offi-
cer, November 14, 1966.

104

of rotation is necessary in almost every case, and small-scale
farmers are discouraged from complete dependence upon cash crops.
On the other hand, it is these crops--and regularly marketable
surpluses of high-value food crops--that are of primary concern
here. The factor of land use rotation and the important, but
non-marketed, food crops that are part of this rotation will not
be included, therefore, in considerations until much later in the

work.

Location pf Crpp Production

 

Two important sets of information must be obtained before
a map of the spatial variation of crop production costs can be
made. The cartographer must know EEEEE the given crop is grown,
and he must know Egg 2222 it costs to grow that crop in any
location within the crop's growing area. Some of the major theo-
retical difficulties in determining the latter have been discussed.
For the reader unfamiliar with the very real problem of data
availability (and reliability) associated with field research in
low income countries, it may be said here that even the presumably
simple task of determining the area in which a given crop is grown

becomes an exercise in extrapolation, interpolation, and logic.

Thus, a number of assumptions often have to be made and acted upon.

A discussion of the various aspects of this exercise, and the as-

sumptions associated with developing a map of the variations in

105
production costs for the crops to be considered, provides a clear
example of the problems with such data. More importantly, the
discussion illustrates the manner in which several important
steps were made in the application of the method for developing
a series of priorities for development investment presented in
Chapter III. Readers directly concerned with such investment
decisions, therefore, may be better able to test some of the
assumptions made here, or to obtain more specific data not avail-
able to the author.

Knowledge of the crop production areas' locations in the
four Kenya districts under consideration in this volume (Central
Nyanza, South Nyanza, Kisii, and Kericho) varies in degree of
reliability with each crop and each district. Tea production can
be pinpointed with the greatest accuracy. All smallhold tea pro-
duction in Kenya is under the tight administrative control of
the Kenya Tea Development Authority (KTDA). The KTDA was founded
to assist smallhold African farmers in establishing tea in a
profitable manner through control of loans, sale of tea stumps,
and marketing arrangements. Tea is an expensive crop to establish,
about Shs. 3000/- per acre in Kericho and Kisii Districts,8 re-

quires particular care during the maturing period, and must be

 

8Interview with Mr. R. Moore, Acting Senior Tea Officer,
West of Rift, Kericho, November 12, 1966.

106
marketed quickly and in large quantities because of rapid leaf
deterioration. The green leaf must reach the factory for
processing within six to eight hours after it is picked.9 In
order to maintain high prices, there is also a high quality
standard demanded--and maintained. For all these reasons, the
KTDA carefully limits tea production to those areas environ-
mentally suitable for successful production. One result of this
control has been the explicit definition of the areas in Kericho
and Kisii which are available for smallhold tea development.10
These areas are shown on Figure 5-1a.

The areas in which pyrethrum and coffee are grown can be
deduced from a combination of each crop's environmental limits
on successful growth and the distributions of pyrethrum col-
lecting points or coffee factories. By outlining the distribu-

tion of pyrethrum buying posts in Kisii District,11 for example,

 

91bid.

10Boundaries were taken from: Edwards and Burrow, Con-
sulting Civil Engineers, "Preliminary Report to the Permanent
Secretary, Ministry of Works Communications and Power, Government
of Kenya, on Tea Roads," Nairobi, l964(?) (unpublished). Note:
The Kericho limits may not be precise; a map of Kericho's tea area
posted on the wall of the Senior Tea Officer's office in Kericho
town showed another boundary enclosing a slightly larger area.

11Obtained from the office of the Cash Crops Officer,
Kisii District, on November 14, 1966.

 

 

 

 

 

in ' Y ' i“

- NYANZA REGION and Kericho District

 

 
 
 
  

 

 

“I f I V H. j

- NYANZA REGION and Kericho District

I fl”? PRODUCING AREA

/)~v '\
1C5“?—

  
  

PYRETHRUM

 

 

1.?

 

 

 

 

 

 

 

 

 

 

“I J-

- NYANZA REGION and Kericho District
(«r-"j PRODUCING AREA

j

 

 

 

COFFEE

\\\:

7/

 

IIIV.

 

 

 

 

 

h‘n l v ‘41

~ NYANZA REGION and Kericho District

/.~‘“j PRODUCING AREA

 

I

4

 

 

 

 

 

 

 

 

 

 

 

 

108

and using as a guide the minimum altitude of 6200 feet, above
whicH all pyrethrum is grown in Kisii,12 an area enclosing
pyrethrum production in Kisii can be mapped (Figure 5-lb).
Similarly, the study region's coffee production areas have been
drawn on the basis of the distribution of coffee factories in
each district with some modifications made to approximate
temperature, soil, or precipitation limitations (Figure 5-1c).

Beyond these three crops (tea, coffee, and pyrethrum),
it is increasingly difficult to determine the areas in which one
or another of the remaining major cash crops is being grown. The
1965 Annual Report for the Department of Agriculture in Rift
Valley Province reported that 30,000 bags of European (i.e.,
white) potatoes were sold as surplus in 1965 from Kericho

13 This considerable surplus could have been grown in

District.
a great many areas within the district according to the general
growing limits available for this crop (see Table 5-1). By ex-
amining the East African Railways and Harbours shipment billings
at Lumbwa Station, it was discovered that all shipments of

potatoes out-bound from this station originated in either Sotik

or Bomet. Partial confirmation of this was made by reference to

 

2Interview with Mr. Ogoda, pp. cit.

13Loc. cit., p. 16.

109

 

emu
-eeHuemae-HHea

 

 

 

meom umoE .GH ca .GHE .uHm .um ooomuoq
wcaeaon
spoons wan
unmwam >uwmouoa woow Hawk Hon umuumn ow umahua
>uo> Ou ESHCOB .cH omuma .cwa .O>m homo asfiwcaam cho Human
.uHa .um ooomuoq
ucmucoo uHHm
cam mafia swan
mm.e me
m83H>5Ham pmumwwuufi homo .GHE
cam .mHOmoumH no you» you w whom
umaw .mHomoaauw :oumsvowm: omuwmp ooo¢amm moan
monuauau
.um oououoowau
ouuumpoa owed “Hm Ou scans mafia .O>m moms asawxda ,AuuwnaH4v
Hmuusmc .anOH moses“ omima moonuoo asaaunoH mewou
Onon Hfiom aofiumuwafiumum Opsumumaame homo

mmOMU mm<U m0h<z zo mHHSHA OZHSOmu A¢HZMZZOMH>ZN MM<ZHMM

to eeeeaeaa

zOHomm wQDHm may szHHS ZBOMU

Him made

110

.awmmmm

 

 

 

 

 

 

 

.omth .da .AmomH .mmmum .>ch vuowxo “coccoqv mdwaouw ommmoo .umummm .m .4 .H umwuusom
.uHm .uw oooo
Oumumpoa Eowpwa O>onw ummn asp Aammmousmv
cu uswfiam ou uawwa .cw mm .cfia .um oomq .uHm .cwaq mOOumuoa
mHHom Huwucomwo
powwoauouua no ma common Hwaucommm ma
wannabes m>awn co uoc Bums .%uv cowaom Sana .muv AmuscmOdv
on unease “Sago eeemaae m.ea mu .eAe m.uae .ec oocm-oe eeeeeeeouw
coauom wowsouw moon uo>o .O>u
wcfiusv .:H w a .aamu no“? mzucoa
Haom zoo .ua\.a« om .cwa O>wu=oomaoo o>wmm
Hwom ham .cw mm .cwa .uHm um ocomuoe acuuou
.CH maioq .um come
manquomoua o>onm umon
.ea mm .eSa .em come .uae .eaam sensuouse
QHDQOEQQ Q .QHHUOHUIHHO3
ouauovoa m.oto ma .xman was .cw om .cwa .O>u mono usonmo do»
macaw afiom cofiumuaawomum ouaumuoaaoe memo

emacaeeoo--a-m

no ee:OHOH<

mgmdfi

lll

.aa-ema .ee .Ameaa ..e:H
.Hammioowucoum ".h .z .mwmwao @003Ochmv mammuwmoo OHanoom .umpcwxma< .3 snow

.uwo .mm..BOH>umu:« .mpowo .Hz
.NH .m can OH .m .AwmmH ..oo can coouu .mamawcoq "novcogv OOH .awvm .H
.uHu .mm..3OH>uOucH .Ouooz .m .u:

.w~-e~ .ee .Ameaa .eeeeeeem Heaeom mo eeeuaueeH
cmowum< ummm one "anouamzv mowum< ummm.mm.huum=ch powdm may .xawum .m mwfiuoso

 

 

.l .aammmm
.mema .HH was H muumm .MMdOM now howaom odouo ammo Hmaowumz < .asoum .m .A

 

 

 

 

.awmmmm .NHIHH .aa .AmoaH .mamewooq "cocoon m.uo nuqv moan .umwuo .m .n

 

:.m: vaGOaa< .NomH .uowuumwn “Hafiz .uuommm Hmaac¢ usuammo Hmuauaaowuw< uowuumwn

vmaawucooiianm mAm¢H

112

14 and an early beginning of

increased yields in Bomet Division
production, also in Bomet.15 With attention thus re-focussed in
southeastern Kericho District, the environmental limits on white
potato production could be applied to a specific portion of the
district and a map of the study region's potato production area
was drawn (Figure 5-ld).

It may be necessary to make clear here that the present
author does not propose that the boundaries of these and the
following crop growing areas are precisely those which exist. An
attempt has been made to construct the boundaries of these mapped
areas in an inclusive, rather than a specific manner. Since the
data are not available for the latter, it was deemed desirable
to include areas in which the given crop is not currently being
grown rather than not provide later in the work such cost-return
information as will be derived for these areas. If the purist
balks at these admitted errors, he should be reminded that a major
purpose of the present work is to offer a first application of a
method of investment priority determination. If this method can

be accepted as one valid in form and function after the example is

completed, then details of this method and the specific data that

 

14District Agricultural Officer Annual Report, Kericho
District, 1965, p. 2.

15Government of Kenya, Kenya African Agricultural Sample
Census, 1960/61, Part 11, Economics and Statistics Division, Of-
fice of the Minister of State for Constitutional Affairs and
Economic Planning, 1962, p. 100.

 

113
are required for policy decisions can be obtained directly.

The growing areas of the other cash crops to be studied--
cotton, rice, sugar cane, and groundnuts (peanuts)--were deter-
mined in much the same manner as those already covered. Cotton
buying posts were located in the study region by names supplied
by the Kenya Cotton Lint and Seed Marketing Board. Using the
environmental growth limits listed in Table 5-1 and the distribu-
tion of these buying posts, with aggregate production indications

16 the area in which most

from published and unpublished reports,
cotton is grown was delimited (Figure 5-2a). The sugar cane

growing area was based on personal observations, published and un-
published reports,17 as well as the plant growth limits listed in
Table 5-1. This area is shown on Figure 5-2b. The areas in which

rice and groundnuts are presumed grown were delimited on the basis

of environmental suitability and references made to divisional

 

lélhig-3 Nyanza Province, report for 1966-1970 Develop-

ment Plan (mimeographed), p. 8.

17In addition to the reports used for cotton (footnote
16), also: District Agricultural Officer Annual Report, South
Nyanza, 1965, p. 6; Colony and Protectorate of Kenya, Report of
the Committee to Carry_ Out an Economic Survey of South Nyanza:
and Kericho District - with, a view to advising whether the
economic potential would justify Rail Development, (C. J. Martin,
Chairman) (Nairobi: The Government Printer, 1957), passim; and,
Republic of Kenya, Report pp Agricultural Roads Miwani:Chemelil:
Muhoroni Areas (Nairobi: Sir Alexander Gibb & Partners (East
Africa), 1965).

 

 

 

 

 

 

I.

H. r A In ' b

NYANZA REGION and Kericho District

 

iHHI II'
V“
\*

II
N.

 

Id

 

 

 

 

COTTON

E

 

 

 

 

 

W T *7

NYANZA REGION and Kericho District

 

 

SUGAR CANE

 

 

 

 

 

 

 

 

 

 

 

 

p

 

is “I

NYANZA REGION and Kericho District

«1‘: PRODUCING AREA
,6 c...

b.

   

 

 

 

 

ERIC E

 

 

 

NYANZA REGION and Kericho District

M

” l1 PRODUCING AREA

1

 

.......

GROUNDNUTS

 

 

 

 

 

 

 

 

 

 

 

 

 

 

FIGURE

5- 2

 

 

115
(sub-district) agricultural units18 (Figures 5-2c and 5-2d). It
is within these areas that the spatial variation in crop produc-
tion costs, total Production costs, and profitability of farming

have been determined.

Location pf Production Costs

The figures representing the annual cost of producing
each of the crops concerned were taken from several sources.
None of these sources deals specifically with agricultural con-
ditions within the study region. In a number of cases, however,
the ecological conditions under which the cost studies were
performed (in Nyeri District, Kenya) are the same as those found
in the Kisii highlands and throughout much of Kericho District.
The ecological similarity between the region in which these cost
figures have been determined and the region of interest in this
volume can only be seen as fortuitous, remembering the considera-
ble difficulties of cost-of-crop-production studies, as discussed
earlier, and the scarcity of such Studies. For those crops not
covered in the Nyeri studies, other sources of production cost

data have been found. Although there is less support for their

 

18Kenya, Sample Census, pp. cit.; Nyanza Province, report

 

for 1966-1970 . . . , pp. cit.; Colony and Protectorate of Kenya,
Report pf the Committgg . . . , 2p. cit.; and, District Agri-

 

cultural Officer Annual Reports for the districts, 1960-1965.

116
application to conditions in the study districts, these data do
provide some basis for costing on which the method for deter-
mining investment priorities may be demonstrated. Of course,
before policy decisions are made, cost studies of some sort would
have to be conducted within the region of concern.

The annual per acre costs of producing the crops to be
examined in this volume are listed in Table 5-2. These costs
should be viewed as those necessary regardless of the yield level
or the environmental suitability of the regions in which the
crops are grown. That is, they will be presumed constant. Al-
though this presumption cannot in fact be made, the manner in
which available data are reported makes such an assumption a
practical one. This will be further explained later during the
discussion dealing with the value of production.

The spatial variation in total agricultural costs, as
defined earlier in the work, is obtained by summing the variation
in transport cost from the locations of crop production to the
major regional outlets and the costs of producing these crops.
There are costs of marketing and production other than those con-
sidered, of course, but these additional costs have not been in-
cluded for reasons of simplicity and clarity. In most cases, the
pattern of total agricultural costs for each crop has been drawn
by adding the production cost for that crop (Table 5-2) to the

movement cost-distance pattern (Figure 4-1) for the area in which

117

TABLE 5-2

AVERAGE COST OF ANNUAL CROP PRODUCTION

 

Crop Cost/Acre/Year Source
coffee1 Shs. 290/- FESU
rice2 Shs. 180/- Brown:CC
sugar cane3 Shs. 440/- Brown:CC
tea4 Shs. 240/— FESU
pyrethrum Shs. 80/- FESU (1962-63 ave.)
cottonS Shs. 100/- Smith & Brown
groundnuts Shs. 200/- Brown:CC
potatoes Shs. 500/- Brown:CC

 

1Average costs, third through fifth years after planting,
1963 report (rounded).

2Two-thirds of Brown's figure, as there are no irrigation
costs here; all rice is grown on seasonal swampland.

3Five year average: planting, first ratoon, and second
ratoon crops with 20-month growing season for each mature crop.

4Does not include cost of establishing tea plots; labor
costs are 1962-63 average, materials costs are from the 1962
report.

5Does not include labor costs; Shs. 50/- per acre each
for spray and for pump use.

Sources: FESU - Kenya, "Some Aspects of Agricultural De-
velopment in Nyeri District," 1962
(Report N9. 21 - August, 1964) and 1963
(Report N9. 24 - August, 1966), Farm
Economics Survey Unit.

118

TABLE 5-2--Continued

Brown:CC - L. H. Brown, A National Cash
Crops . . . , pp. cit. (Part II).

 

Smith & Brown - R. Smith and R. A. Brown,
"Improved Methods of Cotton
Growing in Eastern Region,
Tanganyika," The Empire Cotton
Growing Review, XL (1963),
26847.17

 

 

 

19Smith and Brown refer to a Shs. l40/- cost per acre of

cotton for pump and insecticide (p. 273). Another article

refers to costs per acre, probably including labor costs, of
about Shs. 200/- for a farmer producing 720 lbs. of seed cotton
per acre ("Revenue and Cost of Production of African Grown Crops
in Nyasaland," Nypsaland Farmer and Forester, VI (1962), 8,
abstract in Empire Cotton Growipg Review, XL (1963), 145-46.).
Because of the low cotton yields per acre in Central and South
Nyanza (about 300 lbs. per acre per year), the cost level for this
region was set at Shs. 100/-.

 

 

119
the crop was grown (Figures 5-la to 5-1d, Figures 5-2a to 5-2d).
The areal distribution of different levels of total production
costs for groundnuts, for example, were constructed in this
rather straightforward manner (Figure 5-3). The variation of
these costs from place to place for pyrethrum (Figure 5-4),
potatoes and sugar cane (Figure 5-5), and coffee (Figure 5-6)
were also drawn in this way.

The marketing arrangements for cotton, tea, and in
part for rice, are so different from those of the five crops
referred to above that each total cost pattern required in-
dividual treatment. Rice grown in the southwestern corner of
South Nyanza, for example, is marketed by moving it to Mohoru
Bay pier by land and then shipment to Kisumu by lake services.
Although quite a bit of this rice has moved via East African

Railways and Harbours lake carriers (see Table 5-3), an unde-

TABLE 5-3

RICE AND RICE FLOUR SHIPMENTS, MOHORU BAY TO KISUMU,
BY EAST AFRICAN RAILWAYS AND HARBOURS

 

Year weight, in lbs.
1965 3,264,468.

1964 1,712,500. (estimated)
1963 556,894.

1962 491,289.

 

Source: E. A. R. & H. monthly records

 

 

fh

NYANZA REGION and Kericho District

   
  
  

 

 

      
  

    
  
 
 
 

 

 

 

        
    
 
 
 
 
 

 

 

 

 

 

 

 

   

 

N‘ (.17 AGRICULTURAL COSTS
GROUNDNUTS and TEA
shIIIings/acre to market
. '\ “
I it“ ii": /'\_
, I . II '
III
ii'I'" . \‘\
' J
| / z"
"1. \§séa \~\ IIJ-‘V' 11
\ I. V
\,\ . 1‘
‘\.\ KL ”a” ,.
s o .- m ., n ,, \\\\ \
‘\.'j Shillings
L 2IO/- to 220/— Hlfl] 270/- to 280/-
: 220A m 230% ‘ - w 290%
- m 240% S ZSOL w 300A
- Io 250/— 300/- to‘3l0/—
- m 260% iiii EOA m 320A
- Io 270/— I 320/- In 330/-
KENYA
L . ..

 

 

 

 

FIGURE 5- 3

 

 

 

 

i.

ails

   

{4

 

 

tum

 

afl!

NYANZA REGION and Kericho District
, TOTAL
AGRICULTURAL COSTS-

COTTON and PYRETHRUM

shiIIings/acre to market

 

 

Shillings

IOO/- Io IIOI-
IlO/- Io |20/-

  

- I30/- to MOI-

J's-.11.. MOI- to ISO/-

 

E I50/- Io ISO/-
7 -

//// l60/ Io I7OI-
I I70/- to ISO/-
- ISOI- Io I90/-
- I90/- to 2001-

 

 

FIGURE 5- 4

 

 

 

 

 

u’Iu

   

 

 

L...

 

afl-

~ NYANZA REGION and Kericho District

TOTAL

AGRICULTURAL COSTS

   

fur'flw

”A Ix,»

v /
3,3,. I " '

 

SUGAR CANE and POTATOES

5'3 shiIIIings/acre to market

 

ShNHngs

440/- Io 450/.

 

450/- to 460/-

  

460/- to 470/-

. 470/- Io 480/—

480/- to 490/-

 

490/— to SOD/-

500/- to 5IOI-

.-__:=?$
E1536

   

SIOI- to 520/-

 

. 520/- to 530/-
§ 530/- Io 540/-
§¥ 540l- Io 550/-

E 550/- to 560/-

560/- to 570/-

. 570/- to 530/-
. 580/- to 590/-

 

 

 

FIGURE 5- 5

 

 

 

 

 

afl-

 

 

 

\~
\.
\\.
\
5 ° 5 IO U 10 1’
m:
I: E N YA '

 

:1

       

Ii,

:51:

~ NYANZA REGION and Kericho District

TOTAL

fl

COFFEE and RICE

shillings/acre to market

If”? AGRICULTURAL COSTS-

 

 

Coffee

[INN 3001. on are/-

E sue/- Io 320/-

. no; In 330/-
% 330/- Io 340/-
% 340/— Io 350/-
350/- Io see/-
a sea/- Io 370/-
m 370/— to sea/-
- 3eo/- Io 390/-
. 390/- Io 400/-

ShHans

Rice

mo;- Io ISO/-

1, : I901- to 200/.
2001- Io ZIOI-
- 2I0/- M 220/.
- 220/- to 230!-
2301- Io 240/.

its“ 2401- Io 2501-

 

 

 

FIGURE 5- 6

 

 

124
termined portion of it must move by cheaper--and probably less
reliable--private dhows. In order to make the present applica-
tion of the method proposed in Chapter III as realistic as
possible, the transport cost isopleths have been revised for
this somewhat isolated rice-growing region. On Map 5-6, which
shows the spatial variation of total production costs for rice
as well as for coffee, the pattern of costs for rice in south-
western South Nyanza is based on land-water movement costs plus
the production cost per acre estimated for rice. The cost of
lake freight transport has been taken to be -/20 cents per ton
mile;20 the distance from Mohoru Bay to Kisumu is 98 miles.21
This explains the relatively low total cost of rice output in

this area considering the distant Shs. llO/- "cost-miles" it is

located from Kisumu by land routes.

 

20One E. A. R. G H. office memorandum refers to a 1958
figure of 16 cents per ton mile for "working expenditure." -
(E. A. R. & H. Memorandum, from Chief Operating Superintendent
to the General Manager, 22nd February 1960, Appendix A, p. 2.)
Another memo refers to the "average for Lake Victoria traffic of

17.9 cents." - (E. A. R. & H. Memorandum CD/97, 2nd March
1962, Statement 'B,' Note (b).) And more recently, the operating
cost for the newly instituted wagon ferries at 1964 tonnage levels
was set at 22 cents per ton mile, although this cost could be much
lower with closer-to—capacity loads. - (Office Communication from
Chief Operating Superintendent to the General Manager, E. A. R. &
H., 19th October 1965, p. 5). As a probably high compromise, the
figure of 20 cents (Sh. -/20) per ton mile will be used as the
cost of goods transport on Lake Victoria.

21East African Railways and Harbours, Tgmporary Distance
Tables (being sections 371 to 377 of Tariff Book No. 3, January
1965).

125

The marketing systems for cotton and tea have the same
general impact on the pattern of total agricultural costs as
occurs in the "anomalous" area of rice production, although the
reasons for the distinct market patterns are different. For
cotton, the transport costs paid by the individual smallholder,
whether paid directly through a cess or indirectly by means of
lower prices, is based on the cost of moving the crop from the
farm to the ginnery via one of the numerous cotton buying posts.
There are four such ginneries within Nyanza Region: two in South
Nyanza (Homa Bay and Kendu Bay) and two in Central Nyanza (Kibos
and Ndere, the latter just over 17 miles west of Yala). The
isopleths of cost-distance, to which production costs are added
to obtain the spatial variation of total agricultural costs, have
been drawn to focus upon one of these cotton ginneries. All
cotton grown in South Nyanza, for example, south and west of a
line extending roughly Six miles eastward from the small bay
northeast of Homa Bay, and then running southeast to Kisii, is
taken to Homa Bay ginnery. This "line" results from adherence to
a rule of drawing the isopleths on the basis of least-cost to
either of the adjacent ginneries. That portion of cotton produc-
tion arising in the isolated southwestern corner of South Nyanza
is presumed land carried to Mohoru Bay before being shipped by
water carrier the 86 miles to Homa Bay. The remainder of the cot-

ton grown in the district is taken to Kendu Bay for ginning.

 

126
It is interesting to note that the locations of cotton
buying posts associated with each ginning factory are within the
minimum transport-cost region for that ginnery relative to the
others. That is, each buying post located by the author sends

its raw seed cotton to the closest cost-distant ginnery, as

 

determined here. Although there is no formal policy reason for
this, as far as could be determined, this fact does give ad-
ditional support to the practical validity of the relative levels
of movement costs.

An unusual pattern of movement costs for tea marketing
results from the nature of the crop itself. Because of rapid
deterioration, it is necessary to transport newly picked green
leaves to the factory for fermenting within six to eight hours
after it has been picked. With requirements this stringent, the
patterns of movement costs are obviously focussed on each dis-
trict's reSpective tea factory. Total movement costs are the
cost-distance of shipping green leaf to the factory plus the cost
per ton mile for made tea from the factory to Lumbwa.

A brief comment might be made here concerning sugar cane
cost patterns. As with tea leaf, there is rapid deterioration
in the cane after cutting. The sugar content must be extracted
from the cane without delay, at least in some crude form. One

estimate places 10 miles as "the maximum economic distance sugar

127

."22 Thus, factories must be frequent

cane can be hauled .
in large areas of sugar cane production. These factories may

refine sugar to its clean, white form or they may stop at an

earlier stage with jaggery.23 Since the only factory in Nyanza

Region refining white sugar is adjacent to the large Asian
estates at Miwani--although another is under construction at
Muhoroni--much of the cane grown in smallholdings must be first
converted to jaggery and then shipped to Miwani for further re-
finement. As the nature of the crop demands that small jaggery
factories be widespread and numerous in areas of production,24
the pattern of movement costs pertinent to sugar marketing has
been taken to be equivalent to the pattern drawn directly from
the fields where the crops are produced to either the Kisumu or
Lumbwa market outlets. While this is not strictly accurate, the
distortion is not great.

As a summary, a set of maps has been constructed illus-
trating the place-to-place variation in total agricultural costs

to produce and market the major cash crops in Nyanza Region and

 

22District Agricultural Officer Annual Report, Kericho
District, 1964, p. 4.

23Jaggery, or "jaggre," is a coarse, brown, lumpy sugar
only partially refined.

24There are some 65 factories converting cane to jaggery
located in the small but important region of sugar production
immediately east of Kisumu. - Republic of Kenya, Report pp
,Aggicultural Roads, Miwani: . . . , pp, 235., p. 6.

 

 

128
Kericho District. These several patterns of cost outlay for the
African smallholder are based on the costs of producing each
crop and the cost of transporting the produce from the fields
where they are grown to the major market centers. The trans-
port costs are, in turn, based on operating costs for a three to
five ton vehicle over road surfaces of differing quality. The
patterns of total cost may also differ somewhat among the crops
considered because of variations in marketing requirements. A
striking feature of these maps is their complexity. Although
customarily it is a principle of map construction that the re-
sult must be clear and readable, the areal heterogeneity of
reality cannot be presented in anything other than a very in-
volved pattern. It is with one portion of reality that the

present work is directly concerned.

CHAPTER VI

THE PROFITABILITY OF PRODUCTION

A necessary step in the determination of investment
priorities in agricultural regions is some calculation of
existing and potential profitability of the land within the
region under study. It is accepted that profitability of land
use, or more specifically, that profitability of endeavors using
the land as a primary resource, will be higher in some places
than others. Land is not, after all, everywhere equally suitable
for any given agricultural use. The patterns of both present and
potential profitability can be expected to be as complex as those
already drawn. These maps, illustrating the spatial variation of
African smallhold farmers' annual outlay, have been drawn as
necessary preliminaries to determination of equivalent variation
in income and profit. It is necessary in this chapter to, first,
calculate the place-to-place variation in gross income, and
second, obtain maps showing the real variation in profitability
of present agricultural efforts by subtracting the patterns of
cost from those of income. Later, it will be necessary to de-
termine the spatial differences in potential profitability of
agricultural undertakings in order that the locations of areas of

129

130

greatest profitability improvement may be made clear.

Spatial Variation pf Farm Income

One of the most difficult problems to be overcome in
mapping locational differences in agricultural profitability is
concerned with the income derived from crop sales. If a given
crop yields the same amount per unit area wherever it is grown
and is sold at the same price wherever it is marketed, there is
little difficulty in calculation of the Spatial variation of
profit, for only costs would vary with location. This is rarely,
if ever, the case. As pointed out earlier during discussion of
production costs, the environmental suitability for a given crop
is clearly the dominant factor in its productivity within regions
having few of the "environment compensating" aids developed
rather recently by more technologically advanced countries. This
suitability varies from place to place in an extremely involved
manner, for slight deficiencies in one environmental requirement
may be compensated for in crop growth by a super-sufficiency in
one of the other factors. Although there may be an "optimum"
association of climate-soil-relief-angle of solar radiation-etc.
for any crop, no such grouping of environmental characteristics
has ever been determined. It is unlikely that such an optimum
will be settled upon as long as it is technically more efficient

to develop a new crop strain to suit the environment than to modify

131
the environment to the needs of existing crop varieties. The
important points here are that there is considerable variation
in environmental support for crop production and that this
variation is reflected in an individual smallholder's farm income

and farm profit.

Determination pf Income Variability

 

It is clear that there are gross income differences in
the region under study in this volume. For every cash crop being
examined here, there is a district-by-district variation in
average annual yield per acre, average sale price, or both. The
yield differences, for example, are illustrated in Table 6-1.
Similarly wide differences exist between the prices paid for these
crops in the several districts. By ruling out such possible
causes of price differences as consistent error in reporting the
figures1 and purposeful prejudice against those receiving low
prices by the agencies setting the prices, the dominant sources
of price variation from place to place for a given crop must be
roughly equivalent to variation in yield quality. The spatial
pattern of value per acre for each of the cash crops being ex-

amined must be attributable in large part to differences in en-

 

1A number of apparent errors of this nature have been
discovered in the DA0 Annual Reports' yield figures; many of these
are referred to below in association with Table 6-2. There is no
assurance that all such errors have been discovered.

 

.moma .qoma .mema .muaomom amsaa< aouamwo amaauanoaumd uuaauman "moousom

 

 

 

as me on - - - I- - - I- - I- nmouuuoa
aemeee
an an m om ea ea e e e m m e nee:.m
anmnee
-- - -- -- I- -- com com see CON com com ameae noeeoe
-- - -- cos cos see -I - - I- I- - suggeeass
........... aneae
sooa coca coca some some some - I- I- -- - - aanmae nee
aneae
- - - -- - - om om ON ON mm mm same names
Amcouv
- - I- - - - e e e om em em Aewsee mean
. n: an N - e - e m m e asses
m n on w m: e 0 ma ma 0 I- - II kahuna
”ommwoo
330v
mesa eesa mesa mesa sesa mesa mesa eesa mesa mesa eesa mesa soao
A.oam0v ocoaumx aamaM mucmhz nusom mncmmz Hmuuamo
ammu< muse

mesa-mesa .eaomo mesa omeomamm mom mnamaw seamsmas mo¢¢m><

Hue mAm¢H

133

vironmental suitability for production of these crops in those
areas where they are presently being grown. That is to say, dif—
ferences in price per unit output must be caused primarily by
differences in the quality of the good sold, and differences in
yield per unit area must be caused by differences in the quantity
and quality of the inputs--which may be either environmental or
human. While it is recognized that the quality of human inputs
are often equally as important in determining output quality as
are non-human inputs, an accurate judgement of the relative
quality of the human inputs in this, or any other, area requires
much more detailed research than was possible at this stage of
model development. The primary source of place-to-place dif-
ferences in a crop's average value per acre (price per unit weight
multiplied by weight yield per acre) is here considered to be
spatial variation in environmental suitability for that crop's
production in the locations where it is grown.

0n the basis of this conclusion, a major problem is pre-

sented: How can the differences in average district value per

 

acre be translated into place-to-place variations in crop value
per acre as reflected by the appropriate level of environmental
support for that crop? And equally as important, how can this be
done without loss of areal specificity, without resorting to gross
areal generalizations which would undermine the rationale of the

present undertaking? That the differences in average crop value

134

per acre between the districts are too great and varied to be at-
tributed entirely to measurement or recording error may be seen
by examining Table 6-2.

These difficulties have been met by accepting a small
amount of generalization, well suited to the present scale of
the problem, and making several assumptions concerning the eco-
logical requirements of the crops being examined. First, a series
of suitability levels was outlined for each important element of
the environment for each crop. Each level was assigned a number.
The more "optimum" the temperature level, the higher the number
assigned (see Table 6-3). The major bases for the level rating
assigned each crop were the environmental growing limits, or re-
quirements, as outlined in Table 5-1. With this suitability
rating established for each crop, a square grid with units of one
square mile was placed in turn over each of the maps illustrating
a pattern of some environmental element pertinent to each of the
crops. When determining the pattern of suitability for cotton,
for example, the grid was placed alternatively over the map of
average annual temperature (Figure 6-1), the map of average annual
precipitation (Figure 6-2), and the map showing the pattern of
soils (Figure 6-3). In each square mile within the area in which
cotton is grown (see Figure 5-2a) the appropriate number was noted
for that square mile's suitability level depending upon the pat-

tern of temperature, rainfall, or soil type. By adding the

135

TABLE 6-2

AVERAGE DISTRICT VALUE PER ACRE FOR SELECTED CASH CROPS,
1963-1965 AVERAGE (SHILLINGS)

 

Crop Central Nyanza South Nyanza Kisii Kericho
coffeel 1270/- 1840/- 2100/- 1480/-
rice2 780/- 260/- -- —-
sugar cane 1080/- 530/- -- --
tea3 -- -- 1550/- 1550/-
pyrethrum -- -- 600/- --
cotton 140/- 180/- -- --
groundnuts4 350/- 430/- .375/- -_
potatoesS -- -- -- 675/-

 

1Since the quality, and therefore price, difference be-
tween clean and mbuni coffee is based almost entirely on the
quality of human inputs in the production process, value per acre
was adjusted to the proportional level that would have existed if
all coffee produced had been clean; only 1965 figures were used
for coffee; corrections in multiplication have been made.

2A correction in multiplication (price x yield) for the re-
ported 1964 value has been included.

3Because of possible error in price calculation for re-
ported Kericho District data, the value per acre of tea production
in Kericho has been assumed to be the same as in Kisii District;
also, corrections in multiplication of Kisii figures have been made.

4The figure for Kisii District is based on the author's
calculations. Recorded yields rose sharply from 5 bags in 1962
to 20 bags in 1963 and 15 bags in 1964 and in 1965 (three to five
times the yield recorded elsewhere). At the same time, the re-
corded prices for groundnuts in the district fell even more sharply

136

TABLE 6-2--Continued

to Shs. ll/- per bag in 1963, one-seventh the price of 1962 and
one-seventh to one-ninth that in other districts at the same time.
Therefore, based on total Sales recorded elsewhere in the re-
ports for Kisii District, the average Shs. 375/- figure was cal-
culated. Data for Kericho District groundnut production was
considered by the author to be unreliable after 1962.

5This figure is based on a yield of 45 bags per acre and
a price of Shs. 15/— per bag.

Sources: District Agricultural Officer Annual Reports,
1963, 1964, 1965.

137

 

 

 

ass aaon a noN.Ne-oo.ee
no axons a .Ea oe-se N noo.Ne-oN.Ne
aase N .ea oe ne>o N soN.Ne uses: see
a oNe .sas
NNN .oNs
N as .Ne .eNe a .eN oN-os a soo.oN-oN.Ne
N aNs .ee .Nae N .eN oe-oe N EON.NN-oo.oN
e NNN N .Ea oe ne>o N noo.NN-oN.NN ease names
A mac:
H mumsuo N H
N aNe .ONN N N a noo.oN-oe.Ne
N eNe e N N noe.NN-oo.oN
s NNe N ence no s N moo.NN-oN.NN wean
suma mzuaoa
m>nu=ommcoo
a NNe .sae
N was a .Ea os-os a mooN en>o
N es N .Ea oe-oN N soo.oN-oN.Ne
s Nae .Ne N .ea oe ne>o N see.Ne-oN.Ne menace
.OG .06 .0G
Roman canmanm> xmpsn manuanm> woman Oanmnnm>
acaumunanomnm mvsuaua< no .mnSnm
make aaom Hmscn¢ Owanm>< InmaEOH Hmscn< mmmnm>¢ mono

NHHAHm<HHDm ZOHHUDDOMM momu mo mmoHQZH mam mmam¢Hm<> A¢HZMZZOMH>ZW

muo NAM¢H

138

.sucoa umsu CH mmsoan oau wanpmmoxm no: coNumunaNumnm menm>m mm pmanmop onus :hnv:

“canoe umsu ca monocw pawns mcwvmmoxm conumnnawomnm owmno>m mm COGNMOC onus :nma

 

 

 

 

2*
A NN* a .:N o: nova: H .um coon nova:
N ma§ N .:n om-oq N .um coco-coon
m m* m .:n on nm>o m .um oooo nO>o nOOuduoa
a mnonuo .NN* a woo:
N aa* N H H .uw ooom nm>o
N was .Nae N N N .ua sees-ooos
s Ne .ee e N N .ae sees seen: neeeeeeonw
sac unacce-
O>au=ummaoo
H maa
N ONN .aNs a .eN es noes:
N es .Nae .eNe N .ea oe-oe a soo.oN-oN.Ne
s es .NNN .NNe N .Ea ee-oe N see.NN-oo.oN
N Nae .Ne s .ea oe ne>o N soo.NN-oe.NN scenes
anoaeeaeeeneaaae
mnom non 03Hm> now Ommn ononoom an conumnnm> unmNonmsmch Ednsumnha
.o: .o: .0:
Roman manmnnm> xmpaw manmanm> woman manmnnm>
SONnunnaNomnm upsunua< no .OnSum
make Hwom Hmsaa< wwmnw>< Inumame Hmsaa< mwmnm>< aonu

emanaucoo-IN-e maa<n

 

pd

oil:

351!

T

NYANZA REGION and Kericho District
a... AVERAGE ANNUAL

 

 

 

A“ -’ TEMPERATURE

["0 °L
\1,

 

 

Eh

 

 

KENYA

 

 

 

,— 1110— Degree 3 Forenheit

Weather Recording

Station

 

 

FIGURE 6- l

 

5.

Still r so’ ' 351s ' I:

NYANZA REGION and Kericho District I
a... AVERAGE ANNUAL

41"; (”I

H. we, PRECIPITATION

2,
o
I .
a/ 5 ‘5 K Lvr- . RN", ,,
m ‘9 - . . I, ,~ "3'77 . “I; .
;-. Q

   

 

 

 

 

 

 

o ____/ Inches

III.

9 Weather Recording Station
K! NYA
1" I J

 

 

 

 

FIGURE 6- 2

141

Key to Soil Type Numbers (Figure 6-3)

Well-Drained Soils

Yellow-red sandy clay loams (latosolic)

Dark red friable clays (latosolic)

Red friable clays (latosolic)

Red to yellow-red, gritty, sandy clay loams
(latosolic)

comm-b

Soils with Slight Seasonally Impeded Drainage

15. Brown clay loams to clays

18. Red to dark red friable clays with laterite
horizon

19. Brown to yellow-red sandy clay loams with laterite
horizon

20. Strong brown to pale yellow loamy sands with
laterite horizon

Seasonally Waterlogged Soils

21. Light grey to white mottled loamy sands with
laterite horizon (ground-water laterite)

Soils with Impeded Drainage
22. Black to dark grey clays (grumosolic)
Poorly Drained Soils
26. Greyish brown, mottled, clays (Gley or "Vlei")
27. Dark grey to greyish brown compacted loamy sand
(solodized solonetzic)
28. Very dark brown clays with light textured top-
soils (planosolic)

Others - Lithosols and Regosols

30. Alluvium, recent lacustrine deposits and peaty
swamps
35. Shallow stony soils with rock outcrops

 

 

Z
-<
>
Z
N
>
:1
m
m
D
Z
a:
=
a.
X
co
2
n
=-
a
G
en
1-0-
2
n
I—I-

. SOILS

 

 

 

 

I .“— ‘ ‘ ‘ i :; /;jfv/
. I" ./
ter1te Wm\
I .- V V I I” raucuo .' \‘\
. I i" /,/'/
’ " I" fill ,
.. . . . ”II-I'll,” Il“""" "'" II II' \
I . " :iifli l'IIlllllllll " H (A
.\ \\\\\ -II'!1I'I'I"'II| ” '
' ” \ “ III' 'I "
, _ I. ., \ I “lmflli .
'.\\<§§ llml 'l ‘
’- . _ . \ -
. >. I'll”...- “I
. \ ‘
' \ \ \II-i-I III IIIIIII I"-
)d 15']. . w
' ' s .- g m u m u SOIL TYPES (see key for soil names)
W
'*"- Well-Drained Seasonally Waterlogged
E 4 2'
3n- IIIIIIIII 5 impeded Drainage
as s I 22
Poorly Drained
C/,
I 4% a .I 26

Slight Seasonally
Impeded Drainage

E '5 - 28
Other Soils
I so
- 35

 

KENYA

 

 

 

 

 

 

 

 

 

FIGURE 6- 3

143
individual levels in each square mile, the result is a numerical
pattern of relative environmental suitability for production of
cotton (or tea or . . .) within the area accepted as producing
that crop. For an example of this process, see Appendix I.
Based on the principle that an average value per acre figure for
any district masks production values above and below itself, it
was possible to transform the relative numerical values, or in-
dices, into relative shilling values per acre. The resulting
patterns of average value per acre of production are shown on
Figures 6-4, 6-5, 6-6, and 6-7.

The distribution of the shilling values per acre was
constructed so that the mean numerical index, with regard to
frequency of occurrence, and almost always, the modal numerical
index would also be the mean value level after shilling quanti-
ties had been assigned each number. The resulting values for
each crop (Table 6-4) always included, at the bottom of the
suitability scale, at least one index of production value per
acre which was less than or equal to the level of total produc-
tion costs for that crop. This provides a check on the total
spread of values, for as was stated earlier, the areas to be
considered producing areas for each crop were drawn to accept
errors of excessive areal inclusion in order to minimize possible
errors of exclusion. Thus, there are undoubtedly areas in which

no profit can be gained, i.e., where cost exceeds value, without

 

 

u‘l :

NYANZA REGION and Kericho District
AVERAGE

fl”?

14AM

  
 
  
  
 

 

 

 

an:
KENYA I

 

{
31

F ..
‘\""’1 \

W
"x

VALUE PER ACRE

GROUNDNUTS and TEA

 

\
j

_ ;
32;/rm”?

 

 

ShHHngs
Groundnuts

 

Tea

550/-
E 5450/-

 

 

 

 

FIGURE 6- 4

 

 

 

NYANZA REGION and Kericho District ‘

AVERAGE
VALUE PER ACRE

COTTON and PYRETHRUM

  
   

 

 

 

 

 

 

\ '
\. ‘\
\ 33']! . :1
\~
i o s to I: 10 u \.
W \~ \
\*~\ Shillings
an:
75/- - 200/-
t 1 noox- - 225/-
mm
; I25/- I 250/-
150/-
..r‘ V
@ I751- E 600/-
KENYA
9' u

 

 

 

 

FIGURE 6- 5

 

 

 

d

  

34‘!- ' u'c ' 3‘1:

NYANZA REGION and Kericho District
AVERAGE

VALUE PER ACRE

 

 

 

 

 

5,
g 1,567
AIM -
K ......
\.
./\
ShHHngs
300/. 700/-
3 375A 750%
? 450/- 900/-
fig 500/_ IO50/—
- 525/- race/-
2:: I350»

 

- 600/-

 

 

 

 

FIGURE 6- 6

 

 

 

 

 

u'lu ' F ' W

NYANZA REGION and Kericho District

H AVERAGE
{J VALUE PER ACRE

 

 

s o a lo u no u ~\~ \'\ Shilllngs
h-dEEa-h=fl-had
' ‘ "' \~\~l Rice Coffee

200/- 550/-
kl' 225l- noes/-
250l- r375;-
z7sx~ E rue/-
300/- @ woo/-
w 725/. a raw/-
Q; 750/— m 2000/-
. 775/. - 2:501-
mflfl] soc/-

 

 

  

KENYA

 

 

 

 

 

FIGURE 6- 7

 

148
TABLE 6-4
VALUE PER ACRE EQUIVALENTS FOR NUMERICAL INDICES OF

ENVIRONMENTAL SUITABILITY FOR CROP PRODUCTION,
IN SHILLINGS

Crop District Index number Value per Acre

 

6 850/-
7 1025/-
8 1200/-
9 1375/-

OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO

coffee Central Nyanza

OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO

Kericho

149

TABLE 6-4--Continued

Crop District Index number Value per Acre

 

sugar cane Central Nyanza 4 600/-
5 750/-

6 900/-

7 1050/-

8 1200/-

9

1350/-

OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO

cotton Central Nyanza
and
South Nyanza

l-‘I-‘

groundnuts Central Nyanza, 3
South Nyanza, 4
Kisii, and 5

Kericho 6 325/-
7
8
9

150

TABLE 6-4--Continued

 

Crop District Index number Value per Acre
potatoes Kericho S SOD/-

6 550/-

7 600/-

8 650/-

9 700/-

 

w

151
elimination of some cost factor, for example, by considering
family labor costs as non-accountable.

This method for determining the spatial variation of
environmental support for a crop's production, and the impact
of this suitability variation on the pattern of production value
per acre, has a number of problems as well as advantages which
should be recognized before proceeding. The advantages are
several. Little detail of the place-to-place variation is lost
even when using the existing data. Considerable flexibility also
is inherent in the process. It may be used with possibly greater
validity on problems at the micro-level, for the achievement of
accuracy in problems at this scale is both more likely with the
suggested approach and more generally desirable in the solution
of such research. It is also flexible in the sense that an im-
provement in the information relevant to any aspect of "suitabil-
ity" determination may be accepted and the entire pattern modi-
fied, if necessary, by insertion of the appropriate data within
a relatively short time.

The most important difficulties associated with this
method of determining spatial variation in crop value per acre
result from the assumptions necessary to its operation. In the
first place, it might be argued, the method implies that the same
irelative quantities of each crop are grown in every square mile

considered for that crop. Without data indicating where crops

152

are produced and in what volume, however, this assumption must
be made in spite of the slight "homogenizing" of real Spatial
variation which results. The objection loses some of its force
also when it is remembered that by working through environmental
suitability to the value of production per unit area, the method
describes what the value of production would be if that crop were
to be produced there. With the proper data, labeling suitability
levels with their correct shilling value is much more possible,
including the proper spread of these values along the numerical
index range. A second objection could be that the method assumes
equal importance for each of the ecological factors considered
relevant to the crop's growth. It also assumes an equal division
of importance between the several numerical levels within each
factor. While it must be accepted as a valid objection, for ex-
ample, that in calculating the value per acre of cotton, the
impact of a difference of 2.5 degrees in average annual tempera-
ture is not likely to be the same as a difference of 10 inches
of precipitation, until more precise information is available
dealing with ecologic support for crop production in the region
concerned, or in one ecologically similar, there is no alterna-
tive to acceptance of this assumption.

Exception to the stated exclusion of human quality-
inputs and their impact on crop value per acre might also be

raised again. Careful examination of Table 6-4, however, will

153

show that some non-environmental, and presumably human, factor
has been included in making the monetary assignments to a
number of the numerical index scales. Sugar cane marketed in
Central Nyanza, for example, averaged slightly over twice the
price per unit volume of sugar cane marketed in South Nyanza.
This difference is not justified on the basis of the patterns of
environmental support for cane production in the districts. Al-
though the price difference may be related to the higher grade of
refinement possible in Central Nyanza, much of it must be caused
by the pool of cane-growing experience available to Central
Nyanza farmers from the Asian-owned cane plantation located im-
mediately north of the railroad between Kibos and Koru. A similar
difference in shilling - numerical index assignments was accepted
in the case of rice produced in South and Central Nyanza. In
this case, the average price received for rice in each district
was the same but yields were three times as great in Central
Nyanza. It was deduced from this that output quality was approxi-
mately equal while inputs were not; the scale for South Nyanza is
therefore one-third that of Central Nyanza. There were similar,
but smaller, adjustments made to the district scales for coffee
value per acre figures, based on differences in average value per
acre.

In sum, although possible objections to the procedure

used to establish value per acre variation must be given due con-

154
sideration in any operation of the present method resulting in
policy decisions on investment priorities, it is believed that
these objections could be answered with proper research into the
relevant aspects of agricultural production in the region being
examined. This belief is held regarding procedural as well as
factual elements of the model. The procedure is sufficiently
flexible to allow drastic revision, for example, in the as-
sumptions made concerning the relative weighting of the several
environmental characteristics when developing a pattern of eco-
logical support for a crop's production. Similarly, the method
has enough latitude to include measures of human influences on
crop output or on crop marketing procedures and costs, should
such measures ever be determined. And of course, improvements
in data reliability can be readily absorbed. Arguments raised at
this point are concerned primarily with the example being pre-

sented rather than with the model itself.

Patterns of Existing Profitability

The maps drawn by subtracting the patterns of cost per
acre from the patterns of average return per acre, i.e., the maps
showing the average profitability of the several cash crops
(Figures 6-8, 6-9, 6-10, and 6-11), are even more difficult to
read at a glance than their predecessors. There is a general

trend in most of these maps indicating relative cost levels;

 

 

l

NYANZA REGION and Kericho District ..

AVERAGE EXISTING
”Ex PROFITABILITY -

Nx

 

 

GROUNDNUTS AND TEA

Munr‘VJ” / , _ ,' , 91

fixrsumu
.. ,:.

l:-

 

 

 

 

 

 

\. 35"]: . W
\~\ . Profit per Acre in Shillings
s o a In is an as .
..... \_ \ Groundnuts
\‘i .
- no prom IOOO/— to l049l-
am 0 to 49/- |050/— 10 I099/-
50/- I0 99/— I|00/- '0 ”43/-
I00/— '0 l49l- |I50/- '0 ”99/-
l50/- ro 199/- I200;- to 1249/-
. 200/- to 249/. rzsox— to 1299/-
250/- to 299/- I300/— 10 349/-
KENYA
I 300/- to 349/- rum to I399!-
1 L n

 

 

 

 

 

FIGURE 6- 8

 

 

 

 

 

-i

 

u'I: ' lo- ' :s‘Ir ' .‘r

NYANZA REGION and Kericho District
AVERAGE EXISTING
PROFITABILITY

COTTON and PYRETHRU

1?”

  
   

[2."

 

 

 

   

.....
ShHHngs
341:
cotton Pyrethrum
400/— to 449/-
[flflflm 450/— to 499/-
KENYA

 

 

 

 

 

FIGURE 6- 9

 

 

3'1 1

 

*r

is?

NYANZA REGION and Kericho District

AVERAGE EXISTING
PROFITABILITY

SUGAR CANE and POTATOES

  
   
   

 

\I Profit per Acre in Shillings

- no profit 400/— to 449/-
- 0 i0 49/- mm 500/- to 549/-
‘\‘

22;»; 501-10 99/- 550/- in 599/-

r00/- to I49/- [mm 650/- to 699/-

-. 3:9: 700/- t 749/—
- :11; 250/ to 299/- - °

 

 

 

  

 

 

 

- 350/- to 399/-

“ 850/- to 899/-

 

FIGURE 6-10

 

 

 

 

 

.l‘

34°I l

m ' .1.

NYANZA REGION and Kericho District
AVERAGE EXISTING
PROFITABILITY

COFFEE and RICE

 

 

KENYA

 

 

1°.

 

   

a

N \. 351i . '-

\‘\ \ Profit per Acre in Shillings

‘\‘ \ Coffee Rice
\‘A 500/» to 599/. - no pyom
- 600/— m 699/— .:' :-: 0 re 49/-
~é§é 1000/- io |099/- 50/- to 99/-
@ rr00/- to n99/- 500/- to 549/-
W r200/— to r299/— 550/— fo 599/-
rsoox- io r399/— , soc/s in 649/-
V/A I400/— io mew.
E mom to 1599/.
MW 1600/. '0 I699/»
fit}: nooz— lo I799l<
m I800/— lo l899/—
‘f’

 

 

 

 

FIGURE 6-11

 

159
profits tend to be lower within a given crop pattern at greater
distances from the relevant market outlet.2 On the other hand,
the dominance of the pattern of crop production value per acre
is obvious, not only in the broad boundaries but also in the
occasional countering of the higher cost - lower profit trend.
Some areas are simply well-suited to production of a crop, and
high profits may be expected even though this area is among the
higher cost locations for that crop.3 The levels of production
profitability for each crop have been based on fifty shilling in-
tervals, except in the case of coffee where Shs. lOO/- intervals
were used. This has maintained comparability between the eight
crop profitability patterns in spite of the wide absolute range
of such profitability, as high as Shs. 1800/- to 1900/- per acre
for some places of coffee production and as low as no apparent
profit for occasional crop producing locations.

The fact that areas of no profitability do exist on the
maps for groundnuts (Figure 6-8), cotton (Figure 6-9), sugar cane
and potatoes (Figure 6-10), and rice (Figure 6-11) should not be
surprising. It may be objected that no area can be cropped at a

loss for very long. Eventually, such sub-marginal land will re-

 

2See, for example, the changing profitability of groundnut
production in South Nyanza, Figure 6-8.

3Again, see Figure 6-8; this dominance is also very clear
in the sugar cane pattern, Figure 6-lO.

160
vert to subsistence production, and should therefore not be shown
as part of the growing areas for the relevant crops. In response
to this argument, it must be remembered that the crop growing
areas were drawn as large as believed necessary in order to in-
clude all locations where each crop was currently grown. This
admittedly may have included areas that are not producing any of
the particular crop for the very reason put forth by the objec-
tion. More importantly, however, are the assumptions, incorrect-
ly made, that costs have been completely and accurately accounted
in the present example of investment priority determination.

Also important is the assumption that the African small-
holder cannot or will not remain in an area, producing a given
cash crop, if he must regularly "invest" more in that crop's out-
put than the return justifies. At the macro-level, this may ap-
pear correct. But for the individual farmer who can draw upon a
family for labor, upon less direct kinship ties for transport from
field to market at "bargain" rates, upon fresh soil every three
years in regions where shifting cultivation is possible, the land
may offer returns that are acceptable. Such returns may be viewed
as even more acceptable, in spite of their apparent position be-
low the no-profit margin, because of the likely concentration,
emphasis, and dependence upon subsistence activities for the
family's immediate needs with cash cropping being of relatively

peripheral concern. It is not accidental that the "no profit"

 

161
portions of the growing areas for the cash crops under study are
all located in either poor environmental conditions or outside
the portions of the study region cultivated by well-established
African smallholders. On this latter point, for example, the
only sub-profitable farming carried on in Kisii or Kericho
Districts, both containing aggressively profitable farmers, is
that area in the extreme southeastern corner of Kericho District;
this area comprises part of the expanding frontier of African
agricultural settlement in the district.4 In sum, there are
strong arguments for indicating crop areas where costs appear
to exceed returns.

The four maps of "Average Existing Profitability" contain
several strong patterns. Coffee, especially that grown in Kisii
District, is by far the most profitable source of African small-
hold agricultural income. As Ruthenberg has written: "Small-
holder development in Kenya would be bright and easy indeed if
more coffee could be planted. No other event has put such a
strain on future African agricultural development as the re-

striction on coffee."5 As can be easily seen on Figure 6-11,

 

4For an earlier reference to this, see, Kenya, Ministry
of Agriculture, Animal Husbandry, and Water Resources, African
Land Development in Kenya, 1946-1955 (Nairobi: The Government
Printer, 1956), pp. 145-60, passim.

 

 

5Hans Ruthenberg, African Agricultural Development Policy
in Kenya 1952-1965 (Berlin-Heidelberg, New York: Springer-Verlag,
1966), p. 114.

 

162

wherever coffee is grown, it brings a high return per acre.
Although the present author does not hold such unreserved
feelings concerning the implications of the current coffee re-
strictions, it was definitely in Kenya's long-term interest to
accede to the International Coffee Agreement. Kenya's first-
year quota under the agreement, in 1964, was set at 30,100 tons.
Because large coffee acreages had been planted during the last
years preceding Kenya's accession to the ICA, a ban on further
planting was imposed by the government at the end of 1963 so that
production would not too rapidly outrun the ICA sales quota.6
Nevertheless, total coffee output was expected to reach 60,000
to 70,000 tons per year, even without further planting.7
Emphasis, therefore, has recently been placed-~and will continue
to be so directed-~on quality improvement among the existing
plantings without additional acreages being put under cultivation.
The high output per acre, therefore, can be expected to increase
even though the area under coffee will not be enlarged.

It is also apparent from the four profitability maps that

tea is the only cash crop which approaches coffee's net return

per acre in Kericho and Kisii Districts. The primary cost dif-

 

 

6Kenya, Some Economic Aspects . . . , 1963, op, cit.,
p. 12.

7 . . .

Brown, A National Cash Crops Poligy . . . , o . c1t.,

 

p. 10.

163
ference between tea and coffee production by the smallholder
is not truly indicated by the derived pattern of profitability,
because this cost is short—run and the present interest is in
average existing long-run costs. There is a rather formidable
cost barrier for most Africans with regard to tea production.
The bottleneck in smallholder tea production lies in the
establishment expense which mainly consisted of the cost
of the planting material and labour. Although in fact the
price per piece of the tea stumps is not very great (30
cents each which is somewhat less than the price of coffee
seedlings) the cost of putting down an acre of this crop
is high because of the number of stumps required to fill
one acre (around 3,500 at normal Spacing). Establish-

ment also requires a lot of labour for the preparation of
the land before planting.8

First year costs for coffee establishment in ecologically similar
Nyeri District were estimated to be Shs. 584/89 per acre while
expenses per acre to establish tea were set at Shs. 1586/31.9
Furthermore, significant yields from tea plants are not achieved
until the fourth year after planting while coffee plants yield a
year more quickly (see Table 6-5). It is apparent that although
tea can be a relatively lucrative cash crop for the African small-
holder, the more rigorous environmental requirements and eco-
nomics of production make tea only a partial replacement for new

coffee in smallhold agriculture.

 

8Kenya, Some Economic Aspects . . . , 1963, op. cit., p. 27.

 

91bid., p. 15 and p. 28.

164
TABLE 6-5

ESTIMATED AVERAGE YIELDS FROM SMALLHOLDER TEA PLOTS,
PER ACRE

 

450 lb. green leaf (100 lb. made tea)
1350 lb. green leaf (300 1b. made tea)
2700 lb green leaf (600 lb. made tea)
4050 lb. green leaf (900 lb. made tea)

third year after planting
fourth year after planting
fifth year after planting
sixth year after planting

 

Source: The Operations and Development Plans of the Kenya
Tea Development Authority (Nairobi: K.T.D.A.,
3lst December, 1964), p. 16.

 

Most of the remaining patterns of crop profitability
follow closely from environmental or cost variation. Rice ap-
pears to be a profitable crop in Central Nyanza on the Kano Plain
but close to the no-profit margin in South Nyanza. This latter
pattern is in contrast to that which might be expected on the
basis of rice and rice flour shipments from this area (see Table
5-3). The explanation which will be offered for this apparently
low-return area producing an increasing quantity of rice is an
acceptance by the farmer of these low returns per acre as
adequate, or at least sufficient to encourage efforts the follow-
ing year. Groundnut patterns of profitability appear to vary
more closely with environmental support for production than cost.
The pattern of net returns from sugar cane production is very
mixed; it is marginal in South Nyanza, related to environmental

variation in northern Central Nyanza, and highest on the Kano

165
Plain in Central Nyanza adjacent to the well-established sugar
plantations immediately north of the railway.

Cotton is obviously a marginal crop wherever it is grown
in Nyanza Region. The pattern of profitability appears related
to both the variation in value per acre and the spatial dif-
ferences in total cost. The cost levels used are extremely low
considering possible additional sources of cost, such as labor,
implements, bags, and the like. The relatively low average
return over most of the cotton producing area is largely at-
tributable to the low average yields obtained. Yields of about
300 pounds per acre are common in the region adjacent to Lake
Victoria while yields may exceed 1000 pounds per acre in other
cotton growing areas in Kenya.10 It is difficult to ascribe

this difference to environmental variations alone.

Summary

The work thus far on the applied example of the method
proposed earlier (Chapter III) has dealt with existing costs,
existing production, and estimates of existing profit gained per
acre. Each of these factors of farming economics has been pre-
sented as far as possible in a manner approaching its actual

place-to-place variation within the chosen study area in Kenya.

 

0Interview with Mr. J. Adero, Kenya Cotton Lint and Seed
Marketing Board, Nairobi, November 25, 1966.

166
The data were presented in this fashion because it is felt
that attempts to formulate investment policy in a given portion
of a country cannot fully succeed unless the reality of spatial
heterogeneity is recognized, accepted, and the proper variations
in policy are made. Possibly of even greater importance is the
interaction within the many segments of this space such that a
change in one part alters the conditions in the other portions
of the total. Investment produces--or presumes--change. A
series of investments should therefore include some consideration
of changing conditions within and around the location of the
previous investment of that series. In the present volume, it
is necessary to prepare an estimate, or estimates, of potential
output before any priority schedule of systematic investments
can be developed. It is this pattern of potential net income

that must next be determined.

CHAPTER VII

THE DERIVATION OF AGRICULTURAL POTENTIAL

Investment Decision Criteria

 

Justification for proposed investments should be made
on the basis of expected returns. Those investments having the
greatest possibility of reaching fruition with profitable re-
sults are obviously investments which should receive first at-
tention, non-economic factors aside. Investments expected to
return low benefits, or moderate benefits with low probability
of occurrence, should under normal circumstances be deferred
until higher-return investments have been made. The so-called
"cost-benefit" analysis approaches the problems of investment
by making an accounting of expected investment project costs,
and comparing total costs with the sum of expected benefits.1
Although there are methodological difficulties with most

estimates of total benefits, this approach has been used fre-

quently in the United States to justify or oppose public in-

 

1The literature is much too extensive to list an
adequate cross-section of applications. For a good review
article, however, see, A. R. Prest and R. Turvey, "Cost-
Benefit Analysis: A Survey," The Economic Journal, LXXV
(December, 1965), 683-735.

 

167

168

vestment projects.

Most road improvement investments in Africa are made on
a largely subjective basis. A road benefits the area through
which it passes in a number of obvious ways: without the road
there could be little interchange of goods, people, information,
ideas, and so forth, with all the associated benefits of regional
and personal specialization. A road is a means of facilitating
movement, of minimizing the costs needed to overcome distance.
Unfortunately, although construction of a new road may have
easily accountable results (increased total expenditure on fuel
following a rise in traffic, and the like), many presumed
benefits are confounded with events that may or may not be a
result of the new road, and in any case are difficult to evalu-
ate (e.g., a rise in school attendance). With assurance of in-
definite results from objective attempts to analyse proposed
road investments, it is not surprising that subjective decisions
are made.3

As far as could be determined, choices between road

projects to receive investment funds are made, basically, on a

 

2For a procedural discussion on this method of analysis,
see either, Arthur Maas, 35 31,, Design of Water Resource Systems
(Cambridge, Mass.: Harvard University Press, 1962); or, Roland N.
McKean, Efficiency in Government through Systems Analysis (New
York: John Wiley and Sons, Inc., 1958).

 

3Additional problems along this line, presented when the
road improvements are considered within agricultural areas, were
discussed in Chapter I, above.

169
subjective basis in Kenya. To be sure, cost research is conducted
to determine on a crude basis the possibilities for agricultural
development. The final decision, from which limited funds will
be spent, is liable to be founded on reports from the Ministry
of Agriculture and the Ministry of Communications and Works,
and some over-all plan developed within the various units of the
Ministry of Economic Planning and Development. Within each
ministry, a series of subjective decisions have undoubtedly
already been made. In agriculture, for example, the ministry
headquarters must condense the reports received from each province
or region relating, among other things,the areas within the region
which "deserve" better roads because of their potential contribu-
tion to the regional and national economies. If some form of
priority listing was not made at this level, the final planner
would be overwhelmed with information on which to base his deci-
sion. Such is one advantage of bureaucracy. Each provincial re-
port, however, is largely based on the set of district reports
comprising each province's total. The district reports will also
have been condensed, in this case by the provincial office, on
order to facilitate ministerial decisions. The district reports,
in the rare instance when they do pertain specifically to road
improvements, usually present comments on road quality rather
than the various investment possibilities and their justification.

Such comments as are offered are, in turn, based on subjective

170
estimates made by the District Agricultural Officer or his as-
sistants. It is little wonder that limited investment funds may
be allocated for low-return road improvements--or what may be
worse, not allocated for high-return projects-~when the long
series of subjective priority "thinnings" are considered.4
Dependence upon subjectivity and indefinite foundations

for road investment decisions are by no means newly developed
means of avoiding responsibility for unfulfilled expectations.S
Nor is this method of determining investment priorities neces-
sarily irresponsible; the more objective means of determining
these priorities can be very expensive, time-consuming, and re-
sult in too vague a series of priorities to be acceptable. No
less an authority than the World Bank survey mission to Kenya
has written, with respect to agricultural feeder roads:

The mission is not in a position to recognize firm or de-

tailed priorities as between either areas or individual

roads. In this group, changing developments in different

parts of the country, will require a flexible plan rather

than the establishment of rigid priorities. The method
used by the authority of nominating a list of roads in

 

4This system also tends to be detrimental to inter-
regional road projects, whether Kenya-Tanzania or Kenya-Uganda,
because it focusses project information almost entirely on
internal needs and potentialities.

5See the reference to Hirschman, The Strategy . . . ,
op. cit., in Chapter I, above.

 

171
each area, with work to be performed on selected items,
or substitutes, as funds are found, satisfies this
criterion.
Thus, with easy assurance of "flexibility” the means of selecting
projects is left to subjective judgement. The present volume, of
course, is offering a means of transport investment priority de-

termination in rural areas which is objective and yet maintains

methodological flexibility.

Foundations 2; Agricultural Potential Determination

 

 

At the root of subjective attempts to formulate invest-
ment priorities is the difficulty of estimating probable returns
from a given investment in a given region. The technique most
commonly used, after subjective ones, is that of accounting the
monetary benefits expected to result from the given investment.
The expectations are based on population, traffic, and crop pro-
duction trends within the area. As such, they are relatively in-
flexible with respect to regions and individual projects as
argued by the World Bank mission, above. The sum total of the
expected benefits is also liable to error having its source in
judgements made by the research team concerning which aspects of
the area's society will benefit, and to what degree these benefits

will accrue in the short- and the long-run.

 

6International Bank for Reconstruction and Development,
The Economic Development 9: Kenya (Baltimore: The Johns Hopkins
Press, 1963), p. 184.

 

172

It is argued here that a more dependable approach to the
problem of establishing possible agricultural returns from trans-
port improvements can be achieved through examination of the eco-
logical potential of the region. Ecological regions, as defined
in Chapter III, are for present purposes contiguous areas on the
earth's surface which indicate a basically similar support for
plant growth by the similarity of their environmental character-
istics.

To argue that the non-human factors in plant growth are
more important than the human elements, or to argue that they are
the final control in maximizing production even if labor, atti-
tude, and other such human inputs are "perfected," is to ignore
the nature-changing abilities of human technology. Malthus was,
after all, wrong in the narrowest, short-run sense. On the other
hand, to imply that in a given set of environmental circumstances
there is no maximum output which can be expected--because the
maximum at one level of technology may be increased with some
higher technical knowledge--is to pass from reality to theory.
The concern here is with the levels of output within the reach of
African smallholders engaged in cash crop production. In this
sense, engineered alterations in the environment are less likely

than changes in human attitudes toward production.7 Environmental

 

7Retired District Officers, some theoretical economists,
and others may disagree with this. The author trusts that anthro-
pologists and Africans would not.

173

interaction, described in the form of ecological regions, are
therefore used as the basis for setting levels of crop production
potential in different portions of a given agricultural region.

The use of environmental characteristics as indicators
of possible crop production can scarcely be considered new. On
the other hand, the use of these elements in combination, as
ecological units, in order to indicate place-to-place variation
in crop productivity is not common in the literature. Three
outstanding examples of this approach in Africa are known, al-
though only two have been published in a complete form. It may
be helpful here to make brief, but explicit reference to each of

these examples before proceeding.

Agro-ecology $3 Zambia

 

William Allan, while Assistant Director of Agriculture in
Northern Rhodesia (now Zambia) during the 1940's, undertook with
Colin G. Trapnell and others to classify a great deal of the land
in the country then devoted to African agriculture. The primary
purpose of this task was apparently to classify these areas in
terms of their suitability for different crops.8 0n the basis of
information gathered during field surveys, Allan devised a method

of determining the total number of people who could be supported

 

8See "Foreword," by Max Gluckman, in William Allan, The
African Husbandman (London: Oliver and Boyd, 1965), pp. v-viii.

 

174
by a given unit area of land, i.e., the total "carrying capacity"
of the land.

It is obvious that Allan and Trapnell were well aware of
the spatial variability of land suitability for their method
deals directly with this variability. In contrast to the often
glib implication of a homogeneous land input in some production
equation, Allan used Trapnell's previous work as a guide in con-
ducting vegetation-soil association surveys. The vegetation-soil
data gained from these surveys were mapped, and the areas of each
distinct association were measured. Using the fact, often missed
or ignored in the averaging of data, that varying proportions of
the land are cultivable in the different association areas, Allan
could then derive a series of cultivability levels which applied
throughout each of the vegetation-soil areas.

In any given portion of the survey area, it was possible
to determine how much of the area was suitable for continuous
or partial cultivation and how much was truly useless for agricul-

ture.9 Each major vegetation-soil association was assigned to one

 

9Allan proposed six primary land categories:

Permanent Cultivation Land, where land is cultivated
and left fallow for roughly equivalent alternating
periods;

Semi-Permanent Land, land which may be cultivated con-
tinuously for fairly long periods but which requires
longer periods of "rest" than permanent land;

Recurrent Cultivation Land, requiring three to five

 

 

175
of a set of land quality categories and a land use index was, in
turn, assigned each category. The index for a given category
indicated roughly the number of fallow years required for each
year of cultivation. The better the vegetation-soil indications
for crop growth, the lower the index because fewer fallow years
are necessary to maintain soil fertility. Then, using the average
family size and the number of bags of corn required to feed this
"producing unit" from a farm of given size with given average
yields together with the land use factor and the cultivable per-
centage of each quantity of the various land qualities present in
the region, Allan was able to calculate the "Critical Population
Density" for that region. "For a single Vegetation-Soil type, the
area of land required per head of population may be expressed by
the formula 100 L.U. x C/P, when L.U. is the Land-Use Factor, C
the Cultivation Factor and P the Cultivable Percentage for the
type."10 The Cultivation Factor, in turn, is ". . . the average

acreage cultivated per 'producing unit' ."11

 

times the cultivation period for rest and recupera-
tion of fertility;

Shifting Cultivation Land, which must be left fallow
extended periods, say, twenty to thirty years, after
only two to three years cultivation;

Partial Cultivation Land, basically uncultivable land
but containing small pockets of usable soil;

Uncultivable 25 waste Land.

 

 

 

 

Allan, The African Husbandman, 22. cit., pp. 30-35.

 

10Ibid., p. 89.

111bid., p. 42.

176
The goal in this method is to permit determination of
the critical population density of places before this density is
reached. Gradual deterioration of the land may in this way be
circumvented, at least in theory. As Allan points out,

The term "Critical Population Density" was coined as a de—
scriptive label for this concept of the human-carrying
capacity of an area in relation to a given land-use system,
expressed in terms of population per square mile, and it
was defined as the maximum population density the system is
capable of supporting permanently in that environment with-
out damage 52 Egg land.12

 

 

Although designed to prevent land destruction by over-cultivation
in semi-subsistence economies, this approach could be modified to
apply to over-use and erosion by cash crop producers as well.
Since this volume deals with the spatial variability of cash crop
production, at existing and potential levels of production, Allan's
procedure must be examined from the vieWpoint of its possible use-
fulness in determining these levels of crop output.

There are several aspects of the method used by Allan
which are of direct applicability to the requirements of the pres-
ent volume's procedural outline. In order to delimit a region's
ecological vegetation-soil associations, field surveys were made
along traverse lines forming a grid having unit sides of one mile

and two miles.13 Using this fine a grid with supplementary guides

 

12Ibid., p. 89.

13See Allan's earlier, Studies 33 African Land Usage 33
Northern Rhodesia (Rhodes-Livingstone Papers Number 15; 1949),
pp. 3-7.

 

177

from aerial photographs, it was possible for considerable varia-
tion to be discerned and mapped. It was necessary, of course, to
have a thorough knowledge of the local region's expected joint
vegetation-soil occurrences. Two associations placed in the same
group by Allan, for example, from a survey taken in the Ndola
district of (then) Northern Rhodesia, are called

Red Earth Types: (a) Woodland of Brachystegia spiciformis

and B. longifolia with evergreens in the understory; on

generally deep, brownish-red to lighter red clay-loams.

(b) Woodland of B. floribunda, Julbernardia tomentosa, and
J. paniculata; on bright red to orange-red clay loams.14

 

 

Once such associations are made, however, a sizable area may be
rapidly mapped by a small survey team.
Allan recognized that ecological surveys taken as a means
of assessing land potential have limitations. An ecological survey
is not readily applicable and may be of little value in
regions where biotic and climatic factors, most commonly the
former, obscure the relationship between vegetation and soil.
By burning, cutting, cultivating, and grazing, man and his
animals can change vegetation radically.15
Dependence upon vegetation as the primary indicator of ecological
suitability for crop production may, therefore, be rather tenuous
in regions that have been long, or destructively, cultivated.

Fortunately, the association which can be made between land poten-

tial and vegetation remains relatively undisturbed across much of

 

1['Allan, The African . . . , op. cit., p. 17.

 

Ibid., p. 13.

 

178
Africa, and especially in eastern and south—central Africa. For
purposes of the present study, then, value from Allan's work can
be associated with recognition of the varying agricultural po-
tential of vegetation-soil areas and with the relatively rapid
field mapping which can be done by survey teams basing their
classification of areas on these vegetation-soil groups.

There is little, however, that can be directly obtained
from Allan's further use of these ecological areas for this
volume's immediate needs. In developing a procedure to calculate
the total or per acre population which may be supported by a
given region's agricultural capabilities, Allan first determines
the existing yields per acre for each of the six to eight eco-
logical associations. The total area of each vegetation-soil
grouping is then obtained by field surveys. This allows cal-
culation of total yields in the study region. Once the average
acreage required to feed one person is determined, it is a simple
matter to calculate the total population which may be supported
by the land resources without soil destruction or deterioration.
This procedure is not usable for present purposes because of its
dependence upon existing yields. Undoubtedly because of a major
concern with a land use situation which was even then becoming
serious for future production, Allan dealt basically with existing
land pressures and output rather than some future level of pro-

duction. Unfortunately, it is precisely this potential output,

179
and its relation to the locational variation in environmental
support, which is of primary interest here. Therefore, it may
be concluded that although there are points of support in Allan's
works for the ecological method of determining spatial patterns
of crop production, they offer no direct contributions to the
problem of relating ecological indicators to potential levels of

agricultural output.

Agro-ecology lg Rhodesia

 

A two-part survey which was much closer in its goals to
those of the present work was made in Southern Rhodesia during
the later 1950's.16 Although the authors of Part I, "Agro—
Ecological Survey," based the rainfall-soil-vegetation relation-
ships used in this study upon an earlier work by J. Smith dealing
with the Sudan,17 there was some apparent influence from the much

closer efforts of Colin Trapnell.18 This set of reports attempted

 

16Federation of Rhodesia and Nyasaland, Ag Agricultural
Survey of Southern Rhodesia, Parts I and II (Salisbury: The
Government Printer, 1959).

 

17Reference made, loc. cit., p. 32, to J. Smith, "Distribu-
tion of Tree Species in the Sudan in Relation to Rainfall and Soil
Texture," Agric. Pub. Comm., Khartoum (1950).

18In, Federation of Rhodesia and Nyasaland, Ag Agricul-

tural . . . , pp, cit., p. 32, footnote 22, reference is made to
an anonymous work, "The Soils, Vegetation and Agriculture of North
Eastern Rhodesia,": Report of the Ecological Survey: Govt.

Printer, Lusaka (1943). This work is referred to by Allan as

". . . Colin Trapnell's remarkable pioneer survey of Northern
Rhodesia . . . ," Allan, The African . . . , gp.‘gi£., p. 13, and
bibliography item no. 171, p. 482.

 

180
to classify Southern Rhodesia into ”Natural Regions" based on
associations of environmental characteristics bearing upon crop
production. The classification was made in order to define
". . . the field of liland47 uses toward which, with due regard
for economic influences, the planners will aim in any programme

."19 The aim of the

to achieve optimum physical outputs
study, in other words, was to delimit areas of varying total po-
tential output on the basis of their ecological characteristics.
In this regional determination, primary emphasis was

placed on soil-formation and soil-fertility in so far as it could
be related to crop yields. Climate, for example, was considered
important in crop production but was placed, by strong implica-
tion, in a position secondary to the nature of the parent material
in the soil-formation process.20 Vegetation was also seen to fol-
low soil character here rather than the reverse. Nevertheless,
the relationship between plant growth and rainfall-soil texture
combinations was given considerable emphasis, and as in the case
of the earlier Trapnell-Allan surveys, vegetation was used as the
basic indicator of soil-precipitation areas of relative agricul-

tural suitability.21 On the basis of these environmental indi-

 

19Fed. of Rhod. and Nyas., AA Agricultural . . . , 22. cit.,

 

201bid., p. 16.

21See, Ibid., pp. 32-39.

181
cators, the authors of this study divided their total area of
concern into Natural Regions, and then into Natural Areas within
each Natural Region. Each National Area contained smaller sub-
areas of specified agricultural potential. Unfortunately, for
present purposes, the means by which the total potential of each
of these ecological groupings was determined is not clearly and
explicitly outlined. Furthermore, the resulting areas of produc-
tion potential are not strictly transferable to the present vol-
ume's study region because of environmental differences. In
spite of these limitations of detail, IRE Agricultural Survey‘gf

Southern Rhodesia does provide greater conceptual justification

 

for using ecological associations as indicators of "optimum," or
at least equivalent levels of potential crop output. This report
also is an example of the research thoroughness necessary in

order that policy decisions may be made with reasonable confidence,
although again, a more complete discussion of the precise manner
in which the potentialities of each ecological region were calcu-

lated would have lent more credence to the policy proposals made.

Agro-ecology 12 Kenya

Kenya's agricultural policy has fortunately been guided
by several productive ecologists. After a period of service in
Northern Rhodesia, Colin Trapnell apparently spent a number of

years in Kenya initiating the outlines of present agricultural

182
policy insofar as it is based upon ecological criteria. Leslie
Brown carried on the work begun by Trapnell, expanding its cover-
age and proceeding with directness into applications of ecological
criteria to specific farming programs. Trapnell's influence, for
example, is apparent in Brown's observation that
natural vegetation is the best indicator available of
the potential of any area of land, resulting as it does from
the sum of the effects of rainfall, soil type and temperature.
From a knowledge and understanding of the grasses and trees--
especially grasses--it is possible to arrive at a fair es-
timate of the agricultural potential of almost any area. It
is the natural vegetation which is the key to the whole mat-
ter.2
With this approach toward estimating the agricultural potential
of African farming areas, at least two draft policies were out-
lined by Brown, one each for the two provinces containing the most
productive areas of African agriculture, Central Province and
Nyanza Province.

Fortunately for the purposes of the present work, Brown's
note on agricultural potential in Nyanza Province has remained
accessible, insofar as the files of the Kenya Ministry of
Agriculture are accessible, in spite of its being overshadowed by

23

the better-known Swynnerton Plan. In this unpublished, mimeo—

 

22L. H. Brown, "Development and Farm Planning in the African
Areas of Kenya," East African Agricultural Journal (October, 1957),
quoted in Allan, The African . . . , gp. cit., p. 14.

 

 

23R. J. M. Swynnerton, Plan for the Intensification 2;
African Agriculture lg Kenya (Nairobi: The Government of Kenya,
1954).

 

 

183
graphed, indeed, unsigned and undated report,24 Brown describes
the general locations of a series of ecological zones grouped on
the basis of their potential agricultural productivity. The
farming systems suitable for each of these zones are given in
detail, providing for the present work guidelines for profitabil-
ity estimates which also vary from zone to zone. That is to say,
suggested systems of operating a farm of given size within each
of the zones of farming potential are detailed, including the
proper series of annual rotation and allowance of land area for
non-cash crop efforts. Unfortunately, the only known map of these
ecological regions is restricted to the Central Nyanza District
alone, leaving South Nyanza, Kisii, and Kericho District patterns
in verbal form only. It has been necessary, therefore, to con-
struct such a map on the basis of the verbal description given in
the report, available soil-vegetation—relief information, and per-
sonal observations of a non—systematic nature. Once the general
map of the ecology and the potential of the several land areas is
drawn, it is possible to apply knowledge of crop prices to the

various levels of potential output.

The Pattern 2E Ecological Zones

 

The most direct source, almost the only source, of informa-

 

24L. H. Brown, "Draft Agricultural Policy: Nyanza Prov-
ince," for the Government of Kenya, Ministry of Agriculture,
about 1957 (unpublished).

184
tion concerning Brown's conception of the several agro-
ecological regions within Nyanza Province25 is his "Draft Agri-
cultural Policy" referred to above. Because of its importance
for the present work and for others of the same nature yet to be
undertaken, and also because of the difficulty of obtaining this
reference outside of Kenya, that portion of the note describing
the ecological "farming zones” is quoted at length. These
descriptions follow, in what Brown himself referred to as ". . . a
"26

rough classification of the main ecological zones.

A. Balanced Mixed Farming Zones.

 

(l) Kikuyu - Star Grass Zone. This Zone occurs on Mount
Elgon in Maragoli - Bunyore, Kisii Highlands, and Kericho,
and is characterised by the presence of Kikuyu grass or
Star grass dominant in pasture. The original vegetation
was either forest of the Acacia abyssinica - Vernonia -
Hyparrhenia cyarbaria associations common in Sotik. Although
there is considerable variation in altitude, temperatures
etc., this zone is all capable of intensive cultivation
with one or other high-priced cash crop and of supporting

a heavy stocking of productive animals. It is the most
important and valuable Agricultural area in Nyanza Province,
as is demonstrated by the very high populations carried in
parts of it, e.g. Maragoli - Bunyore and Kisii Highlands.

 

 

(2) High Rainfall Savannahs. This is the next most important
zone in Nyanza province. It consists of short tree-high grass
associations of various sorts, but the following grasses are

 

 

25At the time this report was written, Nyanza Province com-
prised the land now divided into Western Region (less properly,
Western Province although this earlier terminology persists),
Nyanza Region, and Kericho District, the last now included within
the Rift Valley Region.

26
pp. 1-3.

Brown, "Draft . . . ," 92. cit., p. l. Passage quoted,

185

important; Cymbopgon validus, Hyparrhenia rufa, Hyparrhenia
filipendula and Imperata cylindrica; wherever one of these
grasses, with Combretums, Erythrinas, and Bauhineas among
trees are dominant this is the zone concerned.

 

 

Most of this zone is of high potential agriculturally,
though less so than (1) above; a proportion of it is useless
for arable land owing to stones and murram. At present
much of the grassland is poor and coarse, and this zone
generally is capable of great development.

B. Impeded Drainage Sub-Zones.

Poor drainage is a condition affecting possibly one
thousand square miles in Nyanza province, and the areas
which result require different treatment to the balanced
mixed farming zones although they may have equally favour-
able rainfall. Generally, they contain small pockets of
good arable land surrounded by large areas of land which
are useless for arable cultivation or of very limited
value only; their proper usage thus seems to entail the
gathering of fertility from outside the arable and more
or less continuous cropping on the small pockets of good
soil. The typical appearance of such zones is the grouped
tree grassland of Butende, Trans Mara, and Sotik, but one
important area is simply a grass plain dominated by
Penaisetum.ca£abasis. The following are distinct types:-

(1) Sub Zone of Ay(ll, Sotik highlands and part of Sotik
division of Kipsigis. The hillside pockets in this case
originally supported Acacia Abyssinica - Hyparrhenia cymbaria
associations, still to be seen in Sotik, but heavily culti-
vated in Kipsigis. Hilltops are stony and covered with
Themeda triandra, Loudetia kggerensis grassland; on the

flats the grasslands between the hillocks are dominated

by Setarias and pennisetum catabasis.

 

According to our present knowledge this type of country
is only of limited usefulness agriculturally. The hilltops
are potentially very poor, and the very large area of flats
are only of limited value for stock keeping unless the drainage
problem can be solved. 1

(2) Sub Zone of A.(2). Butende country - 150-200 square
miles on Tanganyika border. This zone is similar to the
above except that it is generally flatter with no sharp
hills. The patches of potential arable land carry vegeta-

186

tion similar to Zone A.(2) and occur on ridges where the
soil is deeper; cropping range in these patches will be

similar to A.(2). In impeded drainage conditions grouped
tree grassland dominated by Themeda trianda and Loudetia

kagerensis occurs.

 

This zone is generally valuable stock country with
patches suitable for cultivation. The potential of the
land will depend upon the proportion of it which is suitable
for arable farming.

(3) Sub Zone of A.(2). Pennisetum catabasis seasonal
swamp. This occurs over about four hundred square miles
of South Nyanza district. It is simply a grass plain on
black clay soil with an almost uniform association of
Pennisetum catabasis with Star grass. As opposed to the
last two zones the impeded drainage is caused by the soil
type rather than by the presence of a hard pan just below
the surface. This zone is thought to be of high potential
for sugar production.

 

 

C. Less favourable Zones.

These zones are not referred to as Marginal, as the
rainfall is nowhere so low that cultivation under a
reasonable system would be really precarious, as in'drier
parts of Kenya.

(1) Lake shore savannahs. (Acacia-Balanites plus
Bothriochloa). These are somewhat similar to the Central
Province savannahs in potential. Typically the grassland
is Themeda - Bothriochloa with some Hyparrhenia, and
Loudetia kagerensis dominant on stony ground. Trees are
mainly Acacias with Balanites also very common. There

are patches where impeded drainage occurs in slight degree,
indicated by an association of Setarias with Acacia sayal
as on the Uyoma peninsular.

 

 

 

 

 

With care this zone can support a balanced agriculture
provided the importance of fodder feeding is recognised and
allowed for by growing a fodder crop. Probably four months
fodder feeding a year, or even more, may be necessary, -
possibly in the form of silage.

(2) Kano Plains (Acacia-Balanites plus Bothriochloa).
This zone appears in similar rainfall conditions to (1)
above but on a different soil, which is a silt loam with

 

187

some impeded drainage. The vegetation is largely Acacias
and Balanites with Star Grass, Bothriochloa, Themeda, and
Sporobolus. Parts are definitely swampy and subject to
flooding from rivers. The pristine state of this type of
country may possibly be seen in the Lambwe Valley. The
conditions occur not only on the Kano plains but in many
smaller areas round the Lake shore.

 

There is a high potential under irrigation in this
zone (for rice), and a farming system should be possible
for non-irrigible areas similar to, with slight variations
from, that suitable for (1) above. In both this and (l)
the importance of fishing must be recognised as a food
source.

,(3) Masai Grasslands. This name is given to a small zone,
marginal for agriculture, in Sotik division in the lower
part of the Amala basis. Vegetation is Acacia and shrubs
with Star grass, Eragrostis, and Digitaria. This zone
should ideally be left uncultivated and used as ranching
country. Cultivation would only destroy it.

 

A map drawn exclusively from this verbal description
would indeed be general. With slight reference, therefore, to
the map of soil types (Figure 6-3), a map was drawn approxi-
mating that described by Brown in the above passage (see Figure
7-1). Perhaps the most striking pattern of agricultural poten-
tial illustrated by this map is the dominance of high potential
areas (A.l. - Kikuyu-Star Grass) in Kisii and central Kericho
Districts. This ecological indicator is also found on the
periphery of northern Central Nyanza with the major area covered
by this second zone of high potential located around Maragoli and
Bunyore in what is now Western Region. Of equal important for
the study region is the sizable portion of South Nyanza District

classified A.2.--high rainfall savannah--for the presence of this

 

 

34°] K )C

NYANZA REGION and Kericho District ..

AGRO-ECOLOGICAL
REGIONS <

(after Brown's description)

  
  
   
    

2.2% a”.
, \.\
' . n.
I. 2 .2
. D
F7. ‘.,- \. rim-WA, ‘ TM?
‘7’T‘9‘I~"7 "12' '
.. , - /

    

' «urcuo

   

\\

 

 

 

 

 

 

 

— 35°12 . W
s a s m ., ,0 ,. WW Ecological Zones
M'“ Q E A.I Kikuyu-Star Grass
1‘," A.2 High Rainfall Savannah
B.I Impeded Drainage-sub-zone of AI
ii 8.2 Impeded Drainage-sub—zone of A.2
8.3 Seasonal Swamp—sub—zona of A.2
C.I Lake Shore Savannah
C.2 Kano Plains
KENYA C.3 Masai Grasslands
1" . .5

 

 

 

 

 

 

189
ecology is said to represent an area "generally . . . capable
of great development."27 A point should be made here of the
nearly total inclusion of Central Nyanza within one or the other
of the two major "less favourable zones," Lake shore savannah and
the Kano Plain. Emphasis is laid here on the extent of these
zones because of the demands made on this environment by the
rather large population living in Central Nyanza. The agricul-
tural aspect of these demands will be discussed later.

This map of ecological regions may be made more specific
by a careful comparison with the derived soil map (Figure 6-3).
The resulting modifications to the pattern of ecological varia-
tion are shown on Figure 7-2. The dominant patterns remain; the
generalized pattern of Figure 7-1 was, after all, guided by asso-
ciations between major soil groups and the verbal description by
Brown.

In addition to increased attention given to soil-type
boundaries, an additional category of land potential has been
added to Figure 7-2. The land in this new category is, in many
instances, land which is waterlogged throughout the year, i.e.,
swamp and marsh land. Some drainage of these areas may produce

good agricultural land; under present conditions, only the margins

are feasibly used. Another portion of this new category repre-

 

27
Ibid., p. 2.

 

 

- NYANZA REGION and Kericho District

AGRO-ECOLOGICAL
REGIONS

(modified by soil type)

     
 

I
r’ .'- - \7
'r :9
- ./ V”?
/’fi/v'

 

         
 
 
 
 
  
    
    

 

 

:\=.\\‘.\\\\\
Q - \ .A—ap“ —\\‘§§\\‘ 4,

'- _ ___.__-—-—_-—P.§J

 

 

 

 

 

 

coogrca ones

B.l

%/

r

swamps and
shallow soils

     

 

 

 

 

 

 

FIGURE 7- 2

 

191
sents land which is not usable for agriculture because of the
absence of soil, either on steep slopes or hilltops. Several
volcanic remnants may be identified by the presence of this land
category along the northern shore of South Nyanza.

The most drastic changes in Figure 7-1 were made on the
basis of soil associations. A large portion of southern Kericho
District was changed from B.l. to C.2. because the entire area
classified in the Kano Plains zone associated with the so-called
”black cotton" soil is also indicated as present in southern
Kericho. Similarly, a large section in southern South Nyanza was
re-classified 3.1. from A.2. because of the dominance of a soil
with impeded drainage there as was typical of the described B.l.
zone. On the whole, greater attention to boundaries between soil
types (which are rarely boundaries at this scale, in any case)
provides detail not present in the verbal report of these zones'
locations, but the over—all pattern is basically unchanged.

It may be relevant here to discuss briefly the basis upon
which the soil map was constructed. The importance of soil pat-
terns in providing a foundation for the agro-ecological zones is
obvious. And since the ecological differences are crucial to an
understanding of agricultural potential, a proper soil map is of
equivalent importance. Because there is no soil map of the study

region available, the data used here are derived from patterns of

192
soil-topography associations compiled by R. M. Scott.28 The
regular association which can be made between a succession of
soils down a slope is called a ”catena," and it is on this con-
cept that the combinational map drawn by Scott was based. As
Scott writes: ”It has been noted in the field that a given topog-
raphy having a similar climate and parent materials tends to give
similar soil sequences. As the topography or parent material

29 Since associated soil-topog-

changes so do the soil sequences."
raphy patterns may be constructed by indicating the location and
extent of dominant catenas, it is not illogical to presume that
the relief pattern may be "subtracted” leaving a map of soil types.
This, in fact, is what has been done. By matching soil—topography
associations with moderate scale relief maps of the study region

and defining soil differences on the basis of topographic changes,

a map of major soil patterns was drawn (Figure 6-3).30

 

28R. M. Scott, "The Soils of East Africa," The Natural
Resources 2f East Africa, E. W. Russell, ed. (Nairobi: East
African Literature Bureau, 1962), pp. 67-76, and map.

 

 

291bid., p. 71.

30The only available alternative source of soil information
was the soil map provided in the Atlas 2; Kenya (Nairobi: Survey
of Kenya, 1959). This map was judged unsatisfactory for present
purposes because of the peculiar manner in which relative soil
dominance is represented. The fractions of sub-primary soils in
an area are shown as proportionately wide linear patterns within
the dominant soil, but the specific locations of these secondary
soils within their general area of occurrence are not indicated
in any respect.

193

The agro-ecological regions delimited in Figure 7-2 were
based primarily on the derived soils map for several reasons.
The soil pattern has within it an element of the variations in
relief, an important element in land utilization. The soils map
was, after all, drawn from soils—relief associations. The
boundaries between soils could also be drawn with a fairly strong
degree of certainty as they were often accompanied by a sharp
change in relief. And finally, the description by Brown of the
agro-ecological zones he had based on his experience in this
part of Kenya seemed to follow most closely the pattern of soils

as drawn from the map by Scott.

 

A Measure 2; Climatic Influence

A complete definition of agro-ecological zones cannot be
obtained from soil characteristics alone. In an area where the
soil-vegetation associations can be accurately mapped, further
environmental factors may, in fact, be no more than refinements
of the derived pattern. In the present case, however, although
the vegetation indicators of agricultural potential are well
described by Brown, the precise areal extent of each set of in-
dicators is not. It is therefore necessary to include some
further elements of the environment which will be surrogates for
the unknown distribution of vegetation. In much of East and

Central Africa, climate may be used as this proxy for natural

194
vegetation when soil type is also included. "Vegetation and soil
are closely related, but it is the soil together with the climate
that determines the vegetation, not the vegetation the soil as in
some parts of the world."31
A desirable climatic measure would be one which relates
in some manner to vegetation requirements for temperature and

precipitation. Evapotranspiration is such a measure. Developed

32 33

independently in 1948 by Penman and Thornthwaite, the concept

of potential evapotranspiration relates plant moisture require-
ments to the availability of this moisture. Thus, potential
evapotranspiration

. . represents the water that would be lost by evaporation
from the ground surface and transpiration from vegetation in
an area where there is always an adequate moisture supply.
This water loss is unaffected by type of vegetation cover,
soil type, cultivation practice or changing soil moisture con-
tent. Potential evapotranspiration is primarily a function of
the available energy from the sun. Thus it is also an index
of thermal efficiency. It is not merely a growth index since
it expresses growth in terms of the water that is needed for
growth. It combines both the moisture and heat aspects of
climate.34

 

31Fed. of Rhod. and Nyas., AA_Agricultural Survey . . . ,
‘22. cit., p. 16.

32H. L. Penman, "Natural Evaporation from Open water, Bare
Soil and Grass," Proceedings 9; the Royal Sociegy, CXCIII (1948),
120-45.

 

330..W. Thornthwaite, ”An Approach toward a Rational Classi-
fication of Climate," Geographical ReView,‘XXXVIII (1948), 85-94.

 

34C. W. Thornthwaite, J R. Mather and D. B. Carter, "Three
Water Balance Maps of Southwest Asia," Publications 12 Climatology,
XI (1958), Laboratory of Climatology, Centerton, N. J., p. 23.

195
Potential Evapotranspiration (PE) therefore provides a means of
including the major climatic impact on crop growth within a
single measure and is suitable to be used here as a modifying
agent to the agro-ecological zones based largely on soil type.

In spite of the conceptual attractiveness of this measure,
there is still much controversy over the best method of determin-
ing a location's PE. Mechanical means of measuring PE have yielded
varying results. A number of formulas have been offered by re-
searchers following Penman and Thornthwaite as more accurate or
easier to manipulate. The precise relationship between individual
crop water requirements in terms of PE and crop yields has not yet
been firmly established. The variability of results found to
follow calculations of the various methods for determining PE in
different environmental extremes has brought each of these methods
into question.35 Regardless of these problems of detail, it is
not difficult to accept the conceptual attractiveness and poten-
tial of this approach to agro-climatology, and to agree with
Baier that ". . . eventually those methods which include soil
moisture estimates, either measured or calculated, are the most

promising as a basis for evaluating agro-climates, land use, agri-

 

35For a review article of these problems and much of the
literature on PE, see, W. Baier, "The Interrelationship of
Meteorological Factors, Soil Moisture and Plant Growth," Inter-
national Journal 9f Biometeorology, IX (1965), 5-20.

 

196

cultural production potential of a region and for crop fore-
casting."36

As a means of modifying the previously defined, soil-based
agro-ecological zones within the study region, the levels of poten-
tial evapotranspiration, water surplus and water deficit were cal-
culated for 85 weather recording stations in Nyanza Region and
Kericho District. The Thornthwaite method of PE determination

was used because of the simple data requirements.37 The map of

PE (Figure 7-3) can be considered an illustration of water need in

 

the sense that it represents the amount of water transpired from
plants which have sufficient moisture for complete growth-use, i.e.,
enough moisture to forestall any diminution of plant growth due to
less-than-optimum water availability. The patterns of water sur-
plus (Figure 7-4) and water deficit (Figure 7-5) represent the
variation in annual totals of monthly excesses and inadequacies in
the available moisture with respect to that amount needed. In

this way, high water surplus indicates an area wherein soil
moisture is almost always sufficient for successful crop growth,

while high water deficit is an indication of an area where a crop

 

36Ibid., p. 14.

37Mean monthly temperature, mean monthly precipitation,
average length of day (not pertinent here, because the study
region is located astride the equator), and the water-holding
capacity of the soils. Temperature and precipitation records are
available from the East African Meteorological Organization.

 

34‘1- ' J: ' as‘lt

NYANZA REGION and Kericho District ..
POTENTIAL
(j EVAPOTRANSPIRATION

 

 

 

 

 

  

0 Weather Recording Station

 

 

 

 

 

FIGURE 7- 3

 

a?“

‘1

all: ' a'r

NYANZA REGION and Kericho District

  
  
 

 

 

 

 

 

 

I
“I.“

KISUMU

 

‘I

 

  

 

,...—ao-—— Inches

0 Weather Recordlng Statlon

 

 

FIGURE 7- 4

 

 

 

 

 

34°Iu ' ,1, ' M! l :1

NYANZA REGION and Kericho District

 

14“:
l lnches

 

 

 

f—a—

. Weather Recording Statlon

KENYA

 

 

 

 

FIGURE 7- 5

 

 

 

200
will yield poorly unless specifically suited to a dry climate or
an adequate source of ground water is available.

The pattern of potential evapotranspiration in the study
region is much as might be expected 2 priori. If PE is viewed as
water need, it is not difficult to imagine the close relation-
ship between PE and temperature. When PE isopleths are drawn with
some reference to the modifying importance of altitude, the pat-
tern shown by Figure 7-3 is not unexpected. The highest altitudes
are the coolest parts of the region and, therefore, have the
lowest annual PE. Conversely, the low-lying lands adjacent to
Lake Victoria generally have a rather high PE. Although the
precise boundaries between the interval classes used to illustrate
the pattern may be inaccurate--there are, after all, several large
areas with little change in elevation and few meteorological
stations--the general pattern of water need is undoubtedly close
to that shown.

The map of water surplus (Figure 7-4) is also one whose
general form is easily predicted. Since water surplus ". . . re-
fers to the part of the precipitation at a place which cannot be
evaporated, transpired or held in the soil for later evapotran-
spiration . . . ,"38 those places having low PE and high precipi-

tation will have a large average annual water surplus. In Nyanza

 

38Thornthwaite, Mather, and Carter, "Three Water Balance
Maps . . . ," 22. cit., p. 28.

201
Region and Kericho District, the Kisii Highlands and the rising
land in eastern Kericho District show the highest mean annual
water surplus while the warmer lake margin, Kano Plain and south-
western Central Nyanza average less than a lO-inch annual water
surplus. Although water surplus is ordinarily an important ele-
ment of the water balance of an area because of the inferences
which may be made about stream flow and irrigation needs, it may
be seen here as an indicator of relative precipitation dependabil-
ity. In those parts of the study region showing a high average
water surplus, there is a lower probability of region-wide drought
having agriculturally serious consequences.

The map of water deficit (Figure 7-5) is generally an in-
verted water surplus pattern. This is not to imply that water
deficit is the reverse of water surplus, for each is related to
the other only through the concept of water need, or PE. Rela-
tive levels of water deficiency do, however, provide a more ac-
curate map of drought probability than lists of successive rain-
less days. Thornthwaite and Mather have pointed out

". . . that we cannot define drought as a shortage of rain-
fall alone because such a definition would fail to take into
account the amount of water needed . . . . Drought does not

begin when rain ceases but rather only when plant roots can
no longer obtain moisture in needed amounts."39

 

39C. W. Thornthwaite and J. R. Mather, "The water Balance,"

Publications $3 Climatology, VIII (1955), Laboratory of Climatology,
p. 36.

 

202
When water deficit is considered in this light, Figure 7-5 pro-
vides clear climatic justification for Brown's judgement: ”The
great majority of the Province is of high potential agricultural-
ly, and there are only comparatively small areas, chiefly around
the Lake shore, which can be regarded as marginal for agricul-

ture."40

The map clearly indicates that only along the shore
margin in western South Nyanza, and possibly in southwestern
Central Nyanza, is drought a sufficiently regular occurrence to
seriously handicap most agricultural undertakings.

As informative as Figures 7-3, 7-4, and 7-5 are, it is
more meaningful to combine the effects of water need, surplus and
deficit as support for, or inhibition of, successful crop pro-

duction. Thornthwaite and Mather provide such a measure in the

form of a moisture index, where

 

. water sur lus - water deficit
M01sture Index = p .41

water need
By relating surplus to deficit at a place and dividing by the
amount of precipitation required to maintain optimum crop growth
at that location, the relation between these concepts is stand-
ardized. Obviously, where mean annual water surplus exceeds the

average water deficit for a year, the index is positive, while the

 

40Brown, "Draft . . . ," 92. cit., p. l.

41
p. 71.

Thornthwaite and Mather, "The Water Balance," 22. cit.,

 

203
reverse sign results where average deficit is larger than average
surplus, regardless of the potential evapotranspiration. By
dividing the surplus-deficit difference by water need, the index
is standardized as areas of large surplus are related to areas
of large deficit through a common factor, i.e., how much of a
proportion of water need is extra, or lacking, as the case may be.
This has the effect, shown on Figure 7-6, of portraying more ac-
curately the rather small proportion of the study region which is
climatically marginal for agriculture. Perhaps more importantly,
the relatively mild water deficit in these coastal areas is cor-
rectly shown in contrast to the exaggerated impression which might
be gained from Figure 7-5. Since the map of standardized moisture
availability provides an illustration of the spatial variation of
an index combining the agro—climatic factors of water need (PE),
water surplus and water deficit, this map is a suitable pattern
by which the soil-based agro-ecological zones might be modified,
or strengthened, to include the obviously important elements of
precipitation and temperature.

The agro-ecological zoning is, in fact, little modified
by the pattern illustrated on Figure 7-6. At first thought, this
may seem a bit surprising, for one map is based on soil and,
partly, vegetation patterns while the other map is explicitly
founded on climatically derived indices. After somq considera-

tion, however, the two patterns can be seen to approach the same

 

 

34°“ ' I'd ' 35°I l

NYANZA REGION and Kericho District .

4%:7° MOISTURE INDEX

 

 

34°12 /|.0-/ water surplus—water deficit

 

 

water need

0 Weather Recording Station

KENYA

 

 

 

 

 

 

FIGURE 7— 6

 

 

205

problem, i.e., relative environmental support for potential
agricultural production, albeit from two different directions.
The patterns also have another factor in common--relief. Both
the soil determination and the temperature pattern (and in part
precipitation and vegetation) are closely related to relief.
Because of this, the overriding patterns are very close within
the clearly sectioned topographic divisions--the flat Kano Plain,
the Kisii Highlands and the central Kericho upland, the rolling
low plateau of central South Nyanza not too different from the
lower near-plains of western Central Nyanza, and the drier
southern margins of both Kericho and South Nyanza, the latter
emphasized by inland extensions of "coastal" conditions in the
Migori River Valley. The only alterations encouraged by the map
of moisture index variation on the soil-based agro-ecological
zone pattern are some indication of the drier coastal conditions
in South Nyanza, changing a small area from ”High rainfall
savannah," although this latter area is not as well endowed with
moisture as central South Nyanza (see Figure 7-7). 0n the whole,
then, the major zonal outline is strengthened rather than modi-
fied.

An interesting comparison can be made between Figure 7-7,
derived in the manner described, and Figure 7-8, from a report

recently published by the German Development Institute.42 Al-

 

42Deutsches Institut fur Entwicklungspolitik, "Vorgutachten
zur Raumplanung in der Region Kisumu (Kenia)," Berlin: D.I.E.,

 

 

i.

NYANZA REGION and Kericho District ..

AGRO-ECOLOGICAL
REGIONS ‘

modified by soil type and by climate

  

 

 

 

 

   

 

i
11
\ V§§§ m- n
w/j/l’1//"
I O S lilo.‘ I! 20 )9 //
Q Favorable zones Less favorable zones
m. III AJ
Impeded drainage _
8" Unsuitable
[/92 3.2 swampy and
shallow soils
E 3.3
KENYA
. y . I.

 

 

 

 

 

FIGURE 7- 7

 

 

 

 

i

14“!

ad! ' a'r

NYANZA REGION and Kericho District

 

 

 

KENYA

 

ECOLOGICAL REGIONS

(after the D.I.E. report)

 

Zone CMsflfwaHon

Star Grass (A.l)

 

High Rainfall Savannah (11.2)

Lakeshore Savannah (C.I)

Lakeshore Savannah, Kano Type (C.2I

 

 

 

 

FIGURE 7- 8

 

 

208
though the region of concern for this report did not include
Kericho District, an ecological zone map was drawn using the
primary categories outlined by Brown. Specific indication of
the criteria on which the map was drawn is lacking, but the in-
fluence of Brown's report is obvious. What is most important for
the present work, the over-all pattern of zones is very close to
that obtained by careful climatic and pedologic modification of
Brown's written description. Further modifications to Figure 7-7
are not justified, although additional support of the pattern
shown on this map is provided by the independently derived, but

similar generalized zones of Figure 7-8.

 

1967, map 9 (p. 59). The research for this report was done by a
team of German students from the Development Institute under

Dr. P. Waller during the present author's period of research in
Kenya. Although paths crossed frequently, it was not possible to
arrange a meeting until departure from the field was very near
for both parties. Therefore, unfortunately, there was no ex-
change of data.

CHAPTER VIII

THE VALUE OF POTENTIAL PROFITS

The most important task to be accomplished during this
chapter is that of translating the agro-ecological zones, defined
and delimited in the immediately preceding chapter, into zones of
cash crop output potential, in shillings per acre, and in a form

directly comparable to the series of maps in Chapter VI illus-

 

If!“-

trating existing value of production per acre. This will be
done both within the context of data available at this prelimin-
ary stage, and with reference to the manner in which the trans-
lation could be done with the more precise data possible from
detailed team surveys of the regional ecological pattern. Once
the pattern of relative potential has been altered to indicate
maximum production expectations on an absolute level (in shil-
lings) as well, further modifications based on the human factor
in production (traditions, production attitudes, and so forth),
population density, and rotation of land use within a suitable
farming system, could be introduced. In short, the spatial
variation in cash crop potential must be made explicit and at
the same time more realistic by attention to some of the im-
portant non-environmental influences on the average annual poten-
tial of the land.

209

210

The Expanded Areas 2: Future Production

 

Only a moderate amount of new land must be considered
when examining areas of growing potential for the cash crops
under study. On the basis of previously established environ-
mental growth limits (Table 5-1), there are several growing
areas which must be expanded. Previous areas delimited as en-
closing the major existing production of the eight cash crops in-
cluded in the present study were not concerned with possible crop

production but with probable existing areas of output.

 

I‘W 4" 5...“)

Not all crop areas can be considered expandable. Coffee,
for example, is now under strict planting control. The area de-
voted to coffee production may even diminish as less productive
areas are reverted to other crops when producers in these margin-
al areas find they cannot maintain quality output. For the
present, emphasis is placed on increased quality of output rather
than new plantings.l The area delimited as suitable for tea
production, based on KTDA maps, is not likely to be enlarged.

The schedule for new district plantings in Kisii and Kericho has
been outlined by the KTDA but will undoubtedly be within the care-
fully planned administrative, economic, and geographic framework
set by the Authority. The boundaries enclosing present groundnut

production, an area drawn large because of poor data on the

 

1See the discussion of this in Chapter VI, above.

211

location of this production, have also been maintained as the
most likely outer limits of feasible groundnut production in the
near future.2 As with groundnuts, the areas drawn to include
locations of existing rice production were made large, on the
basis of a broadly suitable environment, and remain the same
size for the present concern with potential rice production.
The only locational exception of any significance which might
have been made, but was not, about future rice growing areas is
the margin of the large Yala Swamp in western Nyanza, presently
being investigated as a possible area suitable for irrigated
agriculture.3 The lands usable for coffee, tea, groundnuts, and
rice production in the near future are eXpected to remain much as
shown when considering areas of existing output.

Each of the other four cash crops included in the present
study--sugar cane, white potatoes, pyrethrum, and cotton--have
larger areas of production under the criterion of possible source

of profitable output in the near future. Sugar cane potential was

 

2This reluctance to enlarge the area suitable for growing
groundnuts is based objectively on a comparison of soil, tempera-
ture, and precipitation records with the limits established
earlier. It is possible, however, that groundnuts have become
synonymous with agricultural investment bungling and is therefore
a crop extra sensitive to predictions of potential returns.

3Republic of Kenya, Development Plan, 1966-1970 (Nairobi:
Government Printer, 1966), p. 142.

212
enlarged in western Central Nyanza, in a small area immediately
south of the railway extending to Kibos on the Kano Plain, and
expansion of the South Nyanza growing area to include all of the
3.3. ecological category designated by Brown (see Figure 7-7).
Although this last area appears to lack adequate and dependable
precipitation for sugar cane, the impeded drainage caused by the
soil may be a sufficient source of moisture.4

The land deemed suitable for potato production is much

increased over that considered to contain most of the present

 

output. Using the previously listed environmental limits as
guides (Table 5-1), the potato growing area in Kericho District
was more than doubled. Much of the eastern half of Kisii
District was also considered an area with suitable soil, tempera-
ture, and moisture resources to be a potential source of this
crop. The only major reservation one might have with respect to
the Kisii area, the basis of which is not shown on the maps of
this district, is the very hilly nature of the countryside. The
slopes of this rather strongly dissected upland are possibly too
great for successful potato cultivation on a large scale.

A large area in Kericho District appears suitable for

pyrethrum, although part of this area is beyond the land available

 

4Whatever the reason, Brown wrote of this region, it is
". . . thought to be of high potential for sugar production." -
Brown, "Draft ," 2p. cit., p. 2.

213
to the African smallholder, i.e., land owned by the large tea
estates immediately east of the main Lumbwa-Sotik road. In
spite of the small amount of land devoted to pyrethrum production
in Kericho,S and the fact that the land there closely meets the
production suitability criteria, reservations must be expressed *
on the propriety of encouraging expanded production of the crop.
Because pyrethrum is refined to make an insecticide, the threat a
of an artificial competitor with similar qualities but lower cost

of production is always present. While the author has no firm

 

[Alli ‘13“.-

knowledge of the existence of such a competitive insecticide, the
added qualification on future levels of supply and demand should
be made explicit.

The only significant addition to the area considered
suitable for cotton production is that in the Lambwe Valley, in
South Nyanza to the southwest of Homa Bay. Although this area
is commonly referred to as having potential for future cattle
grazing,6 the ecological character is compared by Brown to that

of the Kano Plain.7 On the basis of this latter reference, and

 

5Approximately 300 acres in 1966. - Interview with Mr.
J. Birir, District Agricultural Officer, Kericho District,
November 12, 1966.

6Kenya, Develppment Plan . . . , pp. cit., pp. 144-45.

7Brown, "Draft . . . ," pp. c1t., p. 3.

214
the comparable environmental characteristics, it is supposed that
cotton could be grown in the valley with some success. The Lambwe
Valley is presently unattractive for either crop or animal agri-
culture because of the extensive areas dominated by the tsetse fly,
carrier of trypanosomiasis, or "sleeping sickness." Eradication
of the tsetse fly in this area (tentative plans have already been

made8) would lead to both cultivation and stock-keeping land use.

The Need for Field Research

 

It should be understood in this, and the following dis-
cussion, that crops and crop areas may often be appealing on only
a theoretical basis. Considerable research, field testing, and
the like, are absolutely necessary before any crop is offered to
individual farmers as a new source of possible income. Govern-
mental files are undoubtedly filled with reports relating the un-
fortunate tale of some cash crop that was "supposed" to be suit-
able for the environmental conditions in which it was planted.
The grand Tanganyika Groundnut Scheme is only one of the more
famous experiences of this nature.9 Low income countries cannot
afford the losses which accompany these enterprises. The Govern—

ment of Kenya, at least, appears well aware of this danger, for

 

8Kenya, Development Plan . . . , pp. cit., p. 145.

 

9For a full report of this venture, see, Alan Wood, The
Groundnut Affair (London: The Bodley Head, 1950).

 

215
it has written: "Any major development must be preceded by a
pilot project in which tenant performance is observed under con-
ditions identical to those which are projected for the major

scheme.”10

Present Data and Spatial Generalization

 

A guide to translating the derived agro-ecological zones
from relative to absolute (shilling) levels of potential is pro-
vided again by the report by Brown. In making this translation,
however, the cautioning that Brown himself made when formulating
his draft policy is a reticence as pertinent to the present work
as to Brown's preliminary one.

Knowledge is lacking on suitable cash crops to use in the
areas suited to them, with the exception of some Coffee areas.
A programme of experimentation needs to be worked out which
will give us this knowledge

It is possible to work out potential targets for cash
crops in the province, but this can only be done in a vague
way until ppp distribution pg ecological zones 1p known, and
I have not, thprefore, attempted it here.11 lfemphasis
mine - S.S.B3_/

 

 

As far as could be determined, the distribution of ecological
zones is not known beyond the description of them left by Brown.
The "vague" targets offered by Brown, supplemented by inferences
made on the basis of recent data, remain the primary guidelines

available for present needs. To pursue the objective of providing

 

10Kenya, Development Plan . . . , pp. cit., p. 139.

 

11Brown, "Draft . . . ," pp. cit., p. 6.

216
an applied example of the method outlined above (Chapter III)
with explicit identification of sections of the application
demanding more detailed research than has previously been con-
ducted, the guidelines and ecological variation will be assumed
more definite than in fact they are.

It may be necessary here to emphasize the degree and
manner in which the areas of potential crop production are spatial
generalizations. Some uneasiness may be present in the reader be-
cause of the stress laid earlier on a methodology which minimized
the generalization of spatial variation in order that the truly
complex variability of reality be adequately represented, i.e.,
represented in detail sufficient to maintain the character of in-
teracting elements of the real world rather than suffer great
loss of detail in homogeneous administrative or physical regions.
This possible unease may be settled in several ways. First, the
scale at which data are available does not permit more detailed
separation of the localized areas with true production potential
from the larger region so classified here. This separation 1p
possible, however, and would provide more justification for any
application upon which policy decisions will be based. Second,
there will always be a need to balance the desirable level of
specificity with the available time and money for research. This
balance will, of course, be weighed in light of the marginal in-

formation obtained from additional research expenditures. Third,

217
it is possible to make the adjustment necessary to approximate
the true proportion of any region's land having potential for
production, if the information is known, by multiplying the land
area by the percentage of each region which has that potential.
If, for example, the Kano Plain was treated as a regional whole,
suited in its entirety to rice production, while in fact, part
of this area averaged 90% of the land suitable for rice, and of
the remainder, only 70% could correctly be considered rice-growing
land, the true potential of the whole would be diminished by the
proportional amounts in the appropriate locations. In sum, the
generalizations accepted here are limited, explicit, and cor-
rectable within the over-all methodology if proper data are avail-

able and deemed useful.

The Pattern pg Future Profit

 

In comparing the derived ecological zones with the
boundaries of land areas considered as having suitable potential
for production of each of the cash crops, a strong resemblance be-
tween the two is seen. Table 8-1 has been constructed to compare
occurrences of each of the cash crop areas with the pattern of
ecological zones described as suitable for these crops by Brown.12

Potential rice production in the High Rainfall Savannah zone is

the only occurrence of a description of crop potential by Brown

 

12Ibid.,passim.

218

 

a

.aofiuuavoua Hawuaauom mo maoamou ages onu on Hananawuaa momma Hanan i
.Hawuaauoa vowmwaaav wasamaom wa«>a: was mu“

iHHnauauoua manaaowumaov mo panopwmaoo on umaa sags; aowuoavoua wawumaxa mo mamas iumm
.Mm.uo mm.uoa asp .xuoa uaamaua can a« vanwuomav
ma muHAAAauwam Hauaaaaouw>am can coauuapoua waaumwxa mo mauauuaa on» Scum i 3
.mmm..mm :. . . . umaun: .aaoum aw acauawuumov can Baum i m "3a:

.ou:«>oum wuaaaz aw :oHuosvoum uncuaaouo mo mawuwfiawuaauom map on woman uo: maov cabana

 

3 m mucuauom
3 muaavaaoum
m 3 m MW 3 m 3 aouuoo

3 m asunuau3m

3 m may
3 m 3 m 3 m aaau Human
mm mm. m auwu

mm. 3 m oommou

 

m.0 «.0 ~.o m.m N.m ~.m N.< H.< macho

 

macaw HaufimoHoomiouw<

Mmo3 Hzmmmmm mmH zH Muma<m<mmm Dmmodm>mn 924 230mm 3m nmmHmumwn m<
.A<HHZMHom ZCHHUDQOMm AOMU mo m<mm< 02¢ mMZON AdUHUOAOomiOMU< ho MUZMMMDUZOU

aim mnm<H

219

not indicated in some manner by the area of potential ouput ac-
cepted in the present work. This has not been included because
rice production in this zone is likely to be an extremely local
undertaking proximate to the small swamps scattered throughout
South and Central Nyanza. In addition, aside from areas of
groundnut potential which are not referred to in Brown's report,
only cotton grown in the Kikuyu-Star grass zone in western Central
Nyanza, and the potatoes grown near Bomet in southern Kericho
District are crops definitely produced with substantial profit in
zones not specified as appropriate by Brown or closely adjacent
to these suitable zones. There is considerable justification,
then, for using the monetary equivalents set forth by Brown,
modified and updated where necessary.

In his approximately ten-year-old "Draft Agricultural
Policy" for Nyanza Province, Brown does not always provide satis-
factory information for current predictions about potential levels
of shilling returns which can be expected from future agricultural
efforts. Groundnut information is not present in the report. In
other cases, such as that of the white potato, both yield and
average price had already exceeded his projections by 1965. The
potential levels of farming value per acre for each of the cash
crops of interest have been based on Brown's estimates of possible
yield per acre and price per unit output with frequent adjust-

ments to these figures based on other, more up-to-date data. These

220
potential values per acre, with their sources, are summarized in
Table 8-2.

A somewhat more detailed discussion of the derivation of
these levels of potential value per acre may be justified. Pre-
sumably, through appropriate efforts toward production control
and marketing procedures, the average prices of seven of the
eight cash crops could be expected to rise, tea being a possible
exception because of the excellent control now being exercised.
On the other hand, fluctuations in the world market price and
the general long-run decline in the prices of primary products
threaten even the existing price levels in all except a few of
these crops. In balance, the 1965 prices, as reported in the
respective District Agricultural Officers' annual reports for
that year, have been accepted as conservative estimates of future
price levels.

The estimated annual yields of these crops are also
rather moderate and well within the reach of African smallholders
in Nyanza Region and Kericho District. Both coffee and sugar
cane yield levels were taken from Brown's "Draft Agricultural
Policy," although coffee growers in Kisii achieved yields of only
6 cwt. per acre in 1965.13 Brown's estimate of annual sugar cane

yields was reduced for present purposes from an average of 26 tons

 

13District Agricultural Officer Annual Report, Kisii
District, 1965, Appendix F.

221

.E .aa< .mpaH .omm .a . mam tum

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

.upompa prca< oA\opam .euam
Q<n fiwmax "Scum .Ho>oa .ahcax "Bonn Aaaauwv
omxi .mnm .mnH ooam i\o~oH .msm aau
.>H canoe
.moma .uuomwm .H .aa<
Hfl—JGE 03 a. .HIIU .mm. a . o .
aucamz .uaoo "scum awaun .aaoum "Baum
i\om .mnm moan 3N i\oo- .msm aaau woman
.>H capes
.momH .auoamm
Haaaa¢ 04a
anaaxz .uaao
Qua .3
.OH .Q amo¢H ..IIIUHU .NW . . . .
.uuommw Hasca< mmouu some
o<n auaaNz .w. ”scum .uaz .aaoum "Baum
-\os .pnm Hams; o~ -\oow .pem was»
.m .aa< .moaa .H xfipcpaa<
.uuonom Haaaa< .mmw 2mm . . . .
mfimflwmwm "Boom a .535 "aoum Acaoauv
i\omm .mcm .uso m i\oow~ .mnm aommou
Haas Hon auwum vaafiw Haaca< auu<\aaaa> mono

whomo =m<u nmaomqmm mom
.mmu< Mam ZOHHUDDomm ho MDA<> A<Hazmkom mu<mm>4 QMH¢ZHHmm

Nim mam¢H

222

NNN.H vawauo>a muasoam aausuauha “amau .oomH .uanaauaam i wood .uanouuo Hash anu magnum
.vHaH3 ooimoofi any a>ona uaauaou cannuouma a“ uaaaa>oumaw uaau you NH a so vamamu

 

 

.pam .p ..mmm..mm_.oaaa-opafi

.caHm acaamoHa>an madaM asu aw uam auua you mwan ma mo Haow an» o>ona ma mwsafi

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

.m .a .moaa .m .a .moaa
.uuoaam fiaaca< ammmmwm.wmmmmfi
o<d osuwuaM "Beau dflm onowuax "Scum
-\o~ .pem mamas op -\oo~H .pem ppoppuoa
.mufi .a
.@H .nm «m0@# «.3 .NNW a . o .
.uuomwm,aaaaa< amouo ammo
mflm.auaaNz .w. "aouw awmm .asoum "Eoum
i\ooH .msm amen m -\oow .mnm muscvaaouw
.>H agape .moH .a .pm .a
.WQGH «g «.mwlm. .Nm. . s s .
Hunca< mfim amouu ammo
anca33 .uaau "aouw .mmm .aaoum “Eouw
om\a .msm .mnH coo i\oom .mnm couuoo
.m .aa< .momH
.uuomom Aaaaa<
mmm “Mafia "soup
om\a .pnm ~.pn3 owe -\meo .mem appnuppsa
owe: you aufium wfiaww Haaaa¢ ouu<\aaaa> mono

vosawucooiimim mam<e

223

.uaa3 uazu vaofih mmmmmmm.anu o>ona wuwcuiaao nwsonuaa .moma

wcwuap uuwuumfiw mflcu you wouuomou pfiawh Esawxma asu mo mvuwnuIOSu 3H:o ma wage

m

.oomH .HH nonEa>oz .auaxmz .puaom wcfiuoxumz Eaunuouxm .oucaz h .3 .Hz saws 3mw>uaucw

Eouw mumo i

.wom.H mo Haow a was wa.a mo awaua>a ahcmMiHHa aw msmuo> ucouaoo swunuauma

wascwucooiimiw mqm<e

224

per acre to 24 tons per acre. Of the three crops whose poten-
tial yields were taken from Brown's National Cash Crops Policy,
groundnuts, rice, and cotton, the last has the greatest like-
lihood of being grossly underestimated. Both Brown and a member
of the Kenya Cotton Lint and Seed Marketing Board place ppppp- F‘
_££pl production levels in Nyanza Region at 900-1000 lbs. per acre
per year.14 Research at Kibos has yielded seed cotton from ex-
perimental strains at 2260 lbs. per acre.15 Because of the per-

sistence of low yields in this region, however, only a moderate

 

increase to 600 lbs. per acre is used. The estimated levels of
these and the remaining crops' potential yields are therefore
reasonable in a context of African smallhold farmer abilities
which may be developed within a decade or so.

It is important to recognize that the crop production
values per acre listed in Table 8-2 are averages for the entire
study region or some major portion of it. In order to be con-
sistent in avoiding the screening effect of homogeneous criteria
within the large areal units, these averages must be broken into
their constituent classes in some manner meaningful to the present
problem. This difficulty was also faced with respect to existing

values of production. As in that case, it is possible to argue

 

14L. H. Brown, National Cash Crops . . . , pp. cit., p.
103; and, interview with Mr. J. Adero, November 25, 1966.

15M. S. Hastie, (abstract) in Empire Cotton Growipg
Review, XLII (1965), 318.

225
that by holding constant the human quality inputs to farming,
for they do not, after all, have spatial relevance except on a
broad scale, variation in product price and product yield are
caused primarily by environmental suitability for production of
that crop. By adding the distributions of environmental cor- i
rectness for each crop, the place-to-place variation in support

l

of each crop's production can be determined. i
I

Fortunately, much of this has already been accomplished. j

A

 

After determining the numerical indices of combined environ-
mental support based on the criteria shown in Table 6-3 for those
areas within the study region considered suitable for cash crop
production but not presently producing these crops, new shilling-
equivalents for these indices could be assigned using potential,
rather than existing, value per acre data. Table 8-3 summarizes
this breakdown, and has been framed in the same way as Table 6-4
to facilitate comparisons. As with the earlier data, the pat-
tern of agricultural support indicated by these indices was de-
termined by superimposing a one-square-mile square grid over each
pattern of a relevant aspect of the environment, covering the
land not yet developed but having potential for production, and
then adding the indices of each grid unit overlying a given square
mile. It is from the mapped pattern described in Table 8-3 that

levels of profitability improvement have been constructed.

226
TABLE 8-3
POTENTIAL VALUE PER ACRE EQUIVALENTS FOR NUMERICAL INDICES

OF ENVIRONMENTAL SUITABILITY FOR CROP PRODUCTION,
IN SHILLINGS

Crop District Index number Value per Acre

 

6 1100/-
7 1350/-
8 1600/-
9 1850/-

OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO

South Nyanza

coffee Central Nyanza

 

OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO

OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO

Kericho

OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO

227

TABLE 8-3--Continued

Crop District Index number Value per Acre

 

sugar cane Central Nyanza 4 600/-
5 800/-

6 1000/‘

7 1200/-

8 1400/-

9

1600/-

00000000000000000000000000000000000000000000000000000000

------------------—----------—---—------------—----------------------

cotton Central Nyanza 4 150/-
and
South Nyanza

groundnuts Central Nyanza, 3
South Nyanza, 4
Kisii, and 5

Kericho 6 650/-
7
8
9

potatoes Kericho and Kisii

 

 

228
It should be noted that where there is a difference be-
tween crop value per acre in several districts for the same crop
(e.g., coffee)--a feature of the dependence of data on adminis-

trative divisions--then the range of gross returns per acre is

narrower for estimates of potential vis-a-vis existing output. 9-
i.
i
Although there is no empirical justification at hand for this, 5

the decrease in difference between more progressive and less r
progressive agricultural areas may be theoretically attributed

to the implementation of better land use practices (e.g., en-

 

closure, rotation of crops, terracing, maintenance of proper stock
densities, and so forth). This implementation should affect the
areas of poorer agriculture relatively more than the better areas
because of the already established improved farming practices in
the more developed areas.

The problem of potential cost levels is not so easy to
solve. Undoubtedly, greater farming costs will be required to
gain the higher returns per acre listed in Table 8-3. The
spatial variation in these costs, higher outlays required in more
demanding environments and lower production costs in more suitable
areas, need not be of concern. Although this may seem contrary
to the basic methodological goal in this volume, some reflection
will indicate that if production costs are treated as Spatially
variable because of variability in environmental support for pro-

duction, then the same environmental factor is being counted

229
twice, in both cost App return accounting. The proportions by
which the increased value per acre of production would be
diminished because of the greater cost requirements necessary to
achieve these higher returns are presently unknown. It should be

recognized that approximations of these proportions can be

. [1

determined only by careful research into the economics of farm
production within the study region. For the present, to facili-

tate calculations, the reduction of increased profitability due

~m'r—

to higher costs will be considered zero. This is, after all,

11',“

 

consistent with the mapping of existing profitability (Figures
6-8 through 6-11), for costs were there assumed constant regard-
less of output. If cost research supplies answers for existing
levels of production, these answers may also be used for esti-
mates of potential production costs.

Using the data of Table 8-3 and the patterns of environ-
mental support for agriculture, maps illustrating the spatial
variation of potential profitability may be drawn (Figures 8-1
through 8-4). Comparison of each map with the analogous map of
average existing profitability of crop production (Figures 6-8
through 6-11) shows little change in the patterns, with the
exception, of course, of those areas added to the growing regions
of each crop. Ignoring this last factor, the patterns of poten-
tial profitability for cotton and white potatoes are those showing

the greatest change. Although the potential of these two crops

 

 

NYANZA REGION and Kericho District

  
   

 

 

 

 

 

 

 

ni ’1
GROUNDNUTS Ctnd TEA
it if.
3' l
11 ll
~\~ '1 \
\~ \
\,\~ \ ”.1, ‘ ,.
\x. ShHHngs

, o e - :0. u m 15 \,\~ '\ antitfln 650/- lo 699/-

\.j E 50‘ M 9% [HM] 700/- to 149/-

m: IOOI- '0 ”9‘ E 750/. to 199/-

250/— to 299/- 800/. '0 “9"

'0 39% E 350/. to 999/-

Tea

'° “9" IZOOI- to race/-

' 1° “95" % izeoI- io- I299!-

‘ '° 5‘9" m I300/- to uses/-

KENYA , ~. ' '° 599" @ ISSOI- lo r3991-
! - 5°04 '° 5‘9" E i400/- to new. ‘.

 

 

 

FIGURE 8- l

 

 

 

T

NYANZA REGION and Kericho District

POTENTIAL
PROFITABILITY

COTTON and PYRETHRUM

  
 
      
  
   
 

......

 

 

 

KENYA

 

‘ 'WLJV‘K J‘Vov‘.

v

W mm Jan

_/ :fw/l

KIIICNO \\

 

  

 

\ Cotton

~\"A D O to 439/-

50/- to 99/-

IOOI- to l49l-

ISO/- to I99/-

200/- to 249/-

 

250/- to 299/-

I 300/- to 349/-
- 350/- to 399/-

 

Shillings

Pyrethrum

[HIM 500/— to 549/-

550/- to 599/-

 

 

 

 

FIGURE 8- 2

 

 

 

NYANZA REGION and Kericho District .

POTENTIAL
PROFITABILITY

SUGAR CANE and POTATOES

I? .

 
   

 

 
 

 

. ShHHngs
\\ - no profit 450/- to 499/-
D o to 49/- [HIM] 500/- to 549/-

 

 

 

34" 50/- to 99/- E 550/- to 599/-
1001- to max- 6007‘ to 649/-
l50/- to |99/- [Inn 650/- to 599/.
2001- to 249/- E 700/. to 149/-
2501- to 299/— 750/- to 199/-
%% 300/- to 349/. m esox- to 099/-
;3; 350/— to 399/- fi 900/- lo 949/-

. 400/- to 449% m IIOOI- to INS/-
I‘

 

 
 

KENYA

 

 

 

 

 

FIGURE 8- 3

 

uh.

 

234
was underemphasized when values were being assigned, they are
crops for which a considerable profitability increase could be
accepted as feasible. In any case, these four maps of poten-
tial profitability are intermediate steps in the derivation of
maps showing the improvement possible in the profit levels of

African smallholders.

Improvement pp pp Investment Criterion

 

 

In the real world of near-subjective decision making,
the customary guidepost used to accept or reject road investments
is the area's total productive potential. If some agricultural
region is considered (believed-reported-"seems") to have high
potential for lucrative output of a crop or crops, improvement
in the infrastructure connecting this region to established
markets is felt justified. Where there is no existing develop-
ment, where there is no cultivation, no settlement, no road
system (or virtually none), then this.method of making decisions
may be satisfactory. There are, however, few places left in East
Africa where these conditions hold. Certainly, no sizable portion
of the study region is so bereft of development, except, possibly,
the tsetse-infested Lambwe Valley. The framework from which in-
vestment alternatives are to be chosen should not be given the
latitude of such subjectivity.

Rather than using EQEEl agricultural potential as the

measure by which road investments are judged, it would be more

235
correct, both in theory and in practice, to base this judgement

on the variability of the possible improvement in production.

 

There is a certain level of production possible, and often ap-
proached, at a given place without the advantages of improved

road connections to market foci. It follows that the benefits to
be derived from road upgrading must originate in, and be limited
by, the potential improvement in production rather than the total
productive capacity of the area. Some crops will reach the
markets even with poor roads unless the land is greatly overpopu-
lated. The amount of such production and the state of the exist-
ing road system provide the two-fold foundation upon which esti-
mates of improvement can be made. Those areas where production
can be improved the greatest amount, i.e., have the largest
absolute difference between potential and existing output, justify
the greatest upgrading in the road system. Adherence to this
principle also has practical utility in that it is difficult to
estimate subjectively several areas' improvement levels for
comparative purposes; planners are required to rely on more objec-
tive methods of determining investment priorities.

Theoretical models and methods derived in the context of
well-developed economies often circumvent this distinction between
total output and improvement but lose a measure of applicability
by doing so. The cost-benefit approach to transport needs, for

example, may attempt to make a value accounting of the costs of

236
road upgrading to compare with the sum of the direct and indirect
returns, or benefits, expected to follow the investment. The
emphasis in this approach is on whether or not the benefits
exceed the costs for that transport line or isolated network.
Little concern can be directed toward the variability of relative
success; this is usually not possible, given the difficulty of

16 By focussing

determining the number and volume of benefits.
attention on the road, instead of on the regions to be served more
efficiently by the road, the pppp of agricultural development
(land use optimization) are made secondary to the ppppp (in part,
the transport route improvement) and priorities for investment
cannot be established properly, i.e., on the basis of what is to
be developed. The method of analysis outlined in the present
volume examines road investments as means of encouraging and al-
lowing agricultural development in light of the potential improve-
ment of the land itself.

By subtracting the patterns of average existing profit-
ability of crop production (Figures 6-8 through 6-11) from the
respective maps of potential profitability (Figures 8-1 through
8-4), a set of patterns were derived illustrating for each of the
cash crops included in the study, the spatial variability of
average potential profitability improvement (Figures 8-5 through

8-8). For each crop, the average improvement possible in the

 

16McKean, Efficiency 1p Government . . . , pp. cit.

 

 

 

 

 

u'Ic

NYANZA REGION and Kericho District
POTENTIAL PROFITABILITY
IMPROVEMENT

F f :51!

 

Shillings

 

 

KENYA

S no improvement

 

 

 

 

250/-
300/-
E soc/-
550/-

 

 

FIGURE 8- 6

 

 

 

 

 

i.

NYANZA REGION and Kericho District
POTENTIAL PROFITABILITY
IMPROVEMENT

s SUGAR CANE and POTATOES

  
   

 

 

ShHHngs

 

 

In I I
W ‘
lilies \~ \
\‘A C] no improvement 400/-
m. - 450/—
3 sow-
% 600/-

700/-

E 750/-

900/-

 

MOO/-

   

 

 

KENYA '
'.'

 

 

 

FIGURE 8- 7

 

 

 

 

JI'I I

 

 

‘r
k

151! ' 1'!

~ NYANZA REGION and Kericho District

POTENTIAL PROFITABILITY
IMPROVEMENT

   

     
    
   

 

\

. .
/_. In
:1, s'» “II
A,
ll
.—
-—
-_ ;
.— ..
- I

 

 

 

ShHHngs

C] no improvement

“‘3‘ 400/.

 

500/-

mm 600/-

----- 700/-

 

 

FIGURE 8- 8

 

 

241

level of annual profit gained by the African smallholder, within
the limits of assumptions made earlier, varies in the manner and
degree shown on these last four maps. In those areas where there
was no apparent profit in current pp foreseeable cultivation, the
areas were not shown on the improvement patterns, even as areas
of "no improvement." The average levels of profitability in-
crease were calculated using the lower value for each range
(usually Shs. 50/-) of profitability. The land to the west of
Yala, in Central Nyanza, for example, was calculated to have an
average existing profitability for sugar cane production of
Shs. 400/- to 450/- per acre, while it was estimated that sugar
cane ppplp yield African farmers returns of Shs. 500/- to 550/-
per acre on the same land in the near future. The mean improve-
ment potential was considered Shs. lOO/- in such areas, for al-
though the range of improvement is Shs. 50/- to 150/-, the stand-
ard error is undoubtedly small.

One of the strongest patterns to appear is the relatively
high potential improvement indicated for central South Nyanza.
In this area, both groundnuts and sugar cane are shown to be
capable of increased returns at the level of about Shs. 400/-
for each crop. Other regions and crops capable of great improve-
ment are found in those portions of the study region where environ-
mental conditions are suitable for a crop's production but are not

presently supporting any measurable output of this crop. Potato

242

production in Kisii, for example, may increase per acre profit-
ability by a large amount because there is no significant exist-
ing potato production in Kisii. Most striking in a negative
sense is the very low increase in profitability of rice produc-
tion, especially on the apparently attractive Kano Plain. This
lack of profit improvement reflects the necessarily conservative
estimates of potential rice yields, a conservatism forced by the
lack of reliable data concerning the land's capabilities in this

regard.

The Impact pi Population

 

 

A simple composite of these patterns of potential profit
improvement is not an adequate basis for investment policy
decisions; it is both incomplete and misleading. Any study such
as the present one cannot limit its scope to generalized "average
totals." It must be concerned with the proportion of the total
area which is likely to be devoted to production of each crop.
The true potential of a region lies not only in its capability of
yielding high value-per-acre crops in localized areas, but in the
possibility of producing a large quantity of these crops at high
quality levels throughout the growing areas. Knowledge of total
over-all output from a given area is more important as justifica-
tion for road investments than some average value of output for

an indefinite quantity of land.

243

Also important in attempts to estimate relative potential
returns from a given region is the concept that some proportion
of each average square mile must be subtracted from the total
for "unusable" land. Land which cannot be used for cash crop
production includes non-agricultural land (trees, housing, and
wasteland, such as swamps), land which must be regularly devoted
to food crops and pasturage,17 and a factor which accounts for
reduced land use due to proper rotation of crops. In this way,
the returns expected from a specific area can be viewed in a more
realistic perspective. Without such adjustment, any region would
appear to hold much greater promise than was actually present.

The work by Brown is again helpful. In his "Draft Agri-
cultural Policy: Nyanza Province," Brown attempts "to assess the
population carrying capacity of each [Tecologicalp7 zone on a
basis of minimum economic holdings for each, based upon the farm-

ing systems devised as suitable for each type of country."18

 

17The government presently recognizes that individual
African smallhold farmers will probably cultivate at least sub-
sistence quantities of food crops before devoting land to cash
crops, regardless of possible economic arguments against food
crop production. Most farmers, that is, will provide themselves
and their families with at least minimum security in case of a
price drop for the cash crop, or a general crop failure in the
major Kenyan food producing regions, as occurred in 1965. See
Kenya, Development Plan, . . . , pp, p1p., p. 168.

 

18Loc. cit., p. 10.

244
Based on admittedly arbitrary but entirely reasonable farming
schemes for each major producing zone, with family returns be-
tween Shs. 2195/- and Shs. 2870/- per year for the several
schemes, the minimum size farm which will support a family in
each zone is outlined.19 Extending further, by using an average
family size of 7.5 people per family, Brown offers the relative
population carrying capacities of the several agro-ecological
zones (Table 8-4). Much higher population densities than those
shown could be supported by the agricultural resources of the
area, but the inhabitants' level of living would be substantially
lower than that outlined by Brown. Greater‘denSities, in other
words, could be supported at carefully managed subsistence
levels, but this is obviously an undesirable situation both for
the local population and for national economic development ef-
forts.

In order to correctly estimate the proportion of a unit
area which may be devoted to cash crop production, the variable
density of population to be supported by the land must be ap-
praised and included with the other factors that reduce cash crop
land use. Unfortunately, population density data are not avail-

able in detail more precise than the Location level, these

 

19For the reader's reference, these schemes are repro-
duced in Appendix II.

245

.OH .3 ..uuu 2mm :. . . . awake: .caoum "aunaom

.pamaouucw an hficwauuou umoaaa canoe shaman mgsas

 

 

aaoaxaa m.o
mum «em «H «.0
mum 33m «a H.o
oom oem wow sm.m
#506an N s m
w azoaxas m H.m
mmo oo¢ NH N.¢
«wo.H mac n H.<
uaaahoHnaa nasuo Aaaaa museum Many macaw

was mavauu mo an: annuasuwuwa a>wuuavoua a“ Amouuav

Hana saws 3uamcav manwamom mufimcav newuaasmoa manwwmom wawvaom abaflawz aGON

mzou gauHooaoom an .waHoamau oszmmau 9244

«um mAm<H

246
Locations averaging 70 to 100 square miles each in area.20 Even
at this scale, however, significant variation in density is ap-
parent (see Figure 8-9). The Kisii Highlands stand out as a
region of dense population ranging from an average 488.6 people
per square mile in Wanjare Location, immediately to the west of
Kisii town, to the high average population density of 876.9 per
square mile in the Kitutu Locations. A second concentration of
population can be seen around Kisumu and stretching to the north
and northwest toward the very high densities in Bunyore and
Maragoli (exceeding 1200 per square mile in 1962) outside the
study region. Equally striking are the low densities in the
southwestern third of South Nyanza District, a result, in part,
of poor soils, and the relatively empty land from the Lambwe
Valley westward resulting from both the large areas of steep re-
lief near the lakeshore and the presence of tsetse flies through-
out much of the Valley.

Potential land productivity must be based on future popu-
lation patterns rather than on densities which were counted in
1962. To this end, an average population increase of 2 per cent
per year for the ten years, 1962-1972, has been assumed and a map

of the resulting 1972 densities drawn (Figure 8-10). The patterns

 

20Data on this and other population statistics in this

section are from, W. T. W. Morgan and N. Manfred Shaffer, Popula-
tion p; Kepya; Densipy and Distribution (Nairobi: Oxford Univer-
sity Press, 1966).

 

 

 

:u'I I ' J: ' ss'I I

- NYANZA REGION and Kericho District
POPULATION DENSITY1

   

\\\\\\\ .

 

 

    

PopMafion per Square MHe

0—99 - soo-sss
- lOO- l99 a 600-699

- 200- 299 “in: 700-799

  

*1“, 300-399 - coo-ass

\‘\\‘\\ 400 499

 

 

 

 

 

 

FIGURE 8- 9

 

 

 

u'lr ' 1' ' :51 r

NYANZA REGION and Kericho District .

    

PROJECTED
POPULATION DENSITY —
' I972

llllllll

 

 

 
  

   

r Population per Square Mile

‘11—— '~:-:-:~:r:>:-:r: V‘ '- 3 .
' o—ss .:,:...,;;,.;, 600-699

 

lOO—I99 ”5% 700-799

- 200- 299 I 300-399

300-399 I 900-999
- I -
§§ 400 499 lOOO r099
- soc-599
KENYA
r

 

 

 

4L

 

 

 

 

 

FIGURE 8-10

249

formed only intensify the range in 1962 population density
variation as shown on Figure 8-9. Admittedly, the two per cent
annual rate of increase is somewhat arbitrary, and possibly a
little low; there are no reliable indications of real population
growth, however, as the 1948 Census of Population does not pro-
vide fully comparable data units. This rate of population growth
is set below the possibly higher natural increase in order to in-
clude intra-regional rural to urban migration and some migration
to Nairobi or other extra-regional centers. Planned migration,
such as possible large-scale settlement in the Lambwe Valley,
cannot be foreseen but could be accounted for in any real applica-
tion of the proposed method for determining investment priorities.

Although the level of productivity per unit area which
can be reached in the region's agriculture must reflect the
pressure of future population upon the land resources for minimum
food requirements, the spatial variability of these resources must
also be taken into account. Obviously, the land is more or less
able to produce foodstuffs as well as cash crops depending upon
the ecological characteristics of the place. A sufficiently small
population may well be able to live at a higher standard on large
quantities of poor land than an extremely large group of people
on a small amount of land regardless of the excellence of the agri-

cultural environment. A reflection of this is the variation in

250
size, from ecological zone to zone, of minimum holdings required
to return to a single farm family produce valued at three to
four hundred dollars, shown in Table 8-4. The place-to-place
variation of the combined effects of population pressure and
land suitability can be illustrated in a derived index, or
factor.

The cash crops Land Use Intensity Factor is intended to
measure the relative proportion of a given area of land which
may be devoted to production of cash crops given the area's
population density and the ability of the land to supply minimum
food needs for that population. Each farm system offered by
Brown, regardless of the ecological zone, included only one acre
of the total farm to be devoted to cash crops. Non-cash crop
acreage in the Kikuyu-Star Grass Zone (A.l) should average six
out of the total seven acres per farm. In the zone of High
Rainfall Savannah ecology, eleven out of the average farming
family's twelve acres must be devoted to non-cash crop efforts.
The index of Land Use Intensity, then, is the difference between
this subsistence land requirement and the average amount of land

available to each family, divided by average farm size.21 That is,

 

211m calculating the average number of farm acres per
family in each of the locations, some portion of each square mile
should be taken from "total agricultural land" as non-arable.
On the other hand, some of each location's population is not farm
population. It has been assumed here that these balance each
other. The population density totals were used, as was the con-
sideration that there are 640 acres per square mile. Also, in

251

farm size - subsistence land re uirements b zone
Land Use In- = q ’ y

tensity Factor farm size

 

As population pressure decreases, there is more available land per
family and the proportion of each farm that can be devoted to cash
crops increases, assuming, of course, that average farm size also
increases to occupy all the land available to each family. Although
this assumption could (and should) be eliminated by appropriate ad-
justments in practice, the assumption has been kept in the present
case. It is likely to apply only in the Lambwe Valley and to the
west, regions of very sparse population at present.

No adjustment in the final average farm size per family
was made in those locations containing sizable uninhabited areas.
For example, the low settlement in the Nyando River Swamp on the
Kano Plain undoubtedly raises the average population density--and
decreases the average family farm size--in the remainder of West
Kano Location above the average figure recorded. On the other
hand, there is sure to be an above average drain on the rural
population caused by the proximity to Kisumu and the employment
opportunities there, real or imagined. In any case, with the
more detailed field research encouraged throughout this volume as

a requisite to rational policy decisions, adequate information

 

lieu of reliable methods of estimating varying farm family size
from district to district, Brown's figure of 7.5 people per
family, implied in calculations on page 10, Brown, "Draft . . . ,
pp. p15., has been used.

252
would be available to eliminate this objection. The required data
should, in fact, be contained within that gathered specifically
for estimates of production and ecological zoning.

The Intensity Factor is attractive in that it can also
indicate realistically zones of overpopulation. When the availa-
ble land is less than the space required for a subsistence exist-
ence by the average family, the factor becomes negative. Although
an appalling proportion of the study region is expected to suffer
such a condition by 1972, according to the growth assumptions and
calculations made here, no portion is shown to have a negative
Land Use Intensity Factor (see Figure 8-11). Wherever popula-
tion pressure was such that little or no land could be expected to
be devoted to cash crops--all the land being required for bare
subsistence, and the average available land per family reduced
below the quantity required under Brown's farming systems-~then
the figures reverted to those which would hold if the minimum
amount of land per farm could be devoted to cash crop production.
Thus, one acre in twelve or one in fourteen (factors of .08 and
.07) were used as representative of conditions in zones A.2 and
C.1 or C.2, respectively, wherever population density was over
the carrying capacity of the land. A minimum land use factor of
.14 was similarly applied to portions of A.1 ecology which were
very densely populated. These minimums were used primarily to

maintain the complexity of the demonstration, but would not, of

 

 

NYANZA REGION and Kericho District

 

 

 

KENYA

 

LAND USE INTENSITY

A

 

 

 

 

9'4
./
/’fi/*/
d
a
Factors based on projected
I972 popMafion
o II|.m-so
[:1 OhJO §§§ shso
u-zo ' .m-Jo
2 .21-50 I .7|—.so
III 3|n4o III srnso.

 

 

 

 

 

FIGURE 8-11

 

254
course, be proper in a real application, for their use would mask
those areas unlikely (or unable) to produce significant quanti-
ties of cash crops from smallhold farms. Well over half of Central
Nyanza, perhaps one-third of South Nyanza, and fully 70 per cent
of Kisii District are expected to be subjected to this level of

population pressure by 1972.

The Pattern.2£ Maximum Profit Increase

 

With the pattern of Land Use Intensity Factors determined,
providing an estimate of the proportion of any unit area which
may be devoted to cash crop production considering both popula-
tion density and environmental conditions, the shilling values per
acre which can be expected from improved levels of farm profit-
ability may be calculated and mapped in a relatively straight-
forward manner. As a first step in this final mapping, the pat-
terns of potential profitability improvement for each of the crops
must be combined. Under conditions of adequate information and
motivation, the cash crop grown on any farm in the study region
should be that which will yield the farmer the greatest return
for his efforts. Although these conditions may not be met in a
real situation, the goal will invariably be maximization of in-
dividual well-being.

The optimum pattern of potential improvement in cash crop

profitability, therefore, is comprised of a combination of those

255
portions of each crop's potential improvement (Figures 8-5 through
8-8) which can be expected to yield the greatest profit per acre
(see Figure 8-12). The resources of Kisii and Kericho Districts
stand out again in this pattern. The very high potential improve-
ment in Kisii District, with only a small portion of the produc-
tive land capable of as little profit improvement as Shs. 500/-
per acre, reflects the high expected return gains from coffee
(western Kisii) and white potatoes (eastern Kisii). It is a bit
surprising that this portion of the study region, already pro-
ducing the highest value-yielding cash crops, also has the greatest
average potential improvement in crop profitability of any district
here. The generally high potential in Kericho District is brought
about primarily from the expected yield gains, and possible acreage
increases, for potatoes, although pyrethrum, groundnuts, and some
tea zones are also represented. The large area of high potential
in South Nyanza is due chiefly to possible groundnut yield improve-
ment with some peripheral gains from coffee. The remaining areas
in the district indicate expected maximum returns from sugar cane,
cotton, and in a very small area, rice. Central Nyanza, produc-
tive to a large extent only in cotton (which has a relatively low
improvement potential), appears capable of improving maximum crop
profitability at the level of the remainder of the study region
only in the area to the south and west of Yala and Maseno. This

zone of potentially higher gains arises from a complex combination

 

 

 

NYANZA REGION and Kericho District

MAXIMUM POTENTIAL
PROFITABILITY
IMPROVEMENT

ALL CASH CROPS

i

 
     
 

J.‘

 

l/I

 

   

ShHHngs

“\\3 [::] 50fl

5“* 450A

500A

 

550A
600A
' 700A
750A

900A

   

KENYA HOOL

 

 

 

 

 

 

FIGURE 8-12

 

257
of sugar cane, groundnut, and coffee profitability dominance, a
combination that is pleasing in its diversity but disappointing
in its areal extent.

Some areas in the study region contain a sufficiently low
average population density to have a sizable amount of land
capable of some cash crop production available per family. Al-
though it must be remembered that the fact of sparse population
probably indicates that farmers in these areas are not likely to
take full advantage of conditions of low pressure on the land,
the potential remains. The possibility of obtaining a more broad-
ly based family income, through diversifying the cash crops pro-
duced, is higher in these low density areas. A series of guide-
lines were developed to take this additional possibility into
account. Figure 8-12 was constructed on the basis of these rules.

1. In any cash crop production area where more than one
crop may be grown, the more lucrative crop, A, takes preced-
ence if there is land for only one cash crop.

2. If land is available over and above the minimum
holding, the additional land will be used according to #1,
above, except that a second crop, B, will be grown if there
is more than (an admittedly arbitrary) 10 acres available for
cash cropping; annual profit yields may be determined by

A (10) + B (X-lO)
Yield =

 

X

258
where X is the total acreage which may be devoted to cash
crops at that place, with the condition that lO<X520.
When more than 20 acres are available for this purpose, both
crops A and B are farmed on equal acreages.

3. Rule #2 will not hold if the second cash crop, B,

would yield a return improvement less than one—third that of
the most lucrative crop, A.
In spite of the fact that in the present work, the requirements
for either rules #2 or #3 were absent in almost every case where
there was an overlapping of cash cropping capability, some guide-
lines or rules such as these should be included in future applica-
tions of the present methodology.

The final pattern of shilling value per acre, resulting
from improved farm profitability, may be constructed directly from
Figure 8-12. Multiplying the patterns of Figures 8-12 and 8-11,
the maximum profit improvement which can be expected and the
proportion of land devotable to such profit improvement, re-
spectively, yields a third pattern which reflects more accurately
than any of the previous maps the real levels of cash crop im-
provement of which the region is capable (Figure 8-13).

Figure 8-13 does not indicate which areas are able to pro-
duce the greatest absolute cash crop surplus. The combined re-
turns from Kisii District, for example, may include income from

coffee, tea, pyrethrum, and potatoes, all high value-per-acre

 

 

« NYANZA REGIDN and Kericho District

ADJUSTED POTENTIAL
IMPROVEMENT
PROFITABILITY

   
   

n“

 

 

 

m“ ' \ \ Increase in Profit per Acre, Shillings
A

m, E] o to 49/- 300/. to 349/-

 

 

 

50/- to 99/— E 350/— to 399/-
. _ - \§ - -
q 1001 to I49/ \\\\\\ 400/ to 449/

I50/~ IO l99l-

    

450/- to 499/—

200/- to 249/- 500/— to 549/-

 

250/— to 299/—
K I N YA

 

 

 

 

FIGURE 8-13

 

 

260
crops. Smallholders in Kisii may take advantage of the low
variability in climatic conditions to reduce the proportion of
their land devoted to food crops below that possible elsewhere.
Because of the very high average population density in Kisii,
however, and because, under the assumptions of the farming
schemes proposed by Brown, income equity among smallholders is
maximized and only one of the several possible cash crops was
"counted” in mapping the patterns of optimum profitability gains,
the average value per acre increase which can be expected is
kept lower than in other regions where the population pressures
are much less severe relative to the capacity of the land to
produce.

The region of greatest potential improvement in profit-
ability is found in Kericho District around, and to the east and
south of Litein (see Figure 8-13). After adjustments are made
for the relative carrying capacities of the land and the popu-
lation pressure which must be "carried," the improvement in prdfit-
ability in this area reaches levels as high as Shs. SlO/- per acre
with much of the area capable of profit per acre increases above
Shs. 300/-. The fact that this same portion of Kericho District
presently contains high-quality smallholder tea indicates it is
presently the most flexible 22d promising of areas for individual
African small-farm development in the study region.

Other zones capable of high profit improvement in the

261

region are smaller in size, lower in value of improvement, or
both. The small area in southeastern South Nyanza showing fairly
high potential profit increase is a zone environmentally capable
of coffee production; because of present restrictions on new
coffee planting, however, this increase can be expected only if
the coffee is already established there. If the coffee factories
located there are, for some reason, spatially separated from
the areas of current coffee farming, then the estimated profit
improvement is much too high. In west-central South Nyanza, a
moderately large area of reasonable profit improvement (Shs. ZOO/-
to 300/-) might be expected if the land and climate are, in fact,
well suited to groundnut production. The adjacent land in the
Lambwe Valley showing slightly lower, but substantial potential
may be less accurately assessed. The level of improvement pos-
sible (Shs. 162/- to 202/- through most of the middle Valley) is
based almost entirely upon the very low projected (1972) popu-
lation density. Development of this land for agriculture is not
likely to proceed rapidly enough to place returns of this level
within "the near future"; population pressure will undoubtedly be
much greater than that projected for the Valley by the time such
land development has taken place.

The most surprising, and disappointing, results obtained
from this map pertain to Central Nyanza District. While Kisii

District profit gain potential could not be expected to be large

262
considering the tremendous population already supported by the
land resources, the real potential is even lower throughout
Central Nyanza. With only two moderate exceptions, each having
a small areal extent, the average small farm profits can be
expected to rise less than Shs. 77/- per acre anywhere in the
district. For most of this low increase area, the profit im-
provement is likely to be less than Shs. SO/— per acre. As this
general paucity of improved profit capabilities results from the
already high population density per carrying capacity in Central
Nyanza, the situation is liable to deteriorate further unless
some unforeseeable agricultural (or non-farm employment) possi-
bilities improve the bases upon which these low figures are cal-

culated.

Summary

Investments should be justified on the basis of the
returns which can be expected from the investment. If the
desired returns are primarily economic, the grounds for their
acceptance or rejection should be as objectively determined as
data permit. Decisions on the correctness of road investments in
agricultural regions are, however, rarely based on a thoroughly
objective study of the true potential of the various regions which
will benefit from the improved transport facilities. Unprejudiced

estimates of this varying potential, it was argued, could well be

263

based upon ecological regionalizing with the relative support
for different aspects of agriculture determined for each
ecology. Several examples of this procedure conducted in
Africa, with one fortunately in Kenya, were briefly discussed.

The true potential of the land, however it is measured,
must be treated as spatially variable, i.e., different from place-
to-place. This rather obvious observation is too often ignored
in preliminary and post-research gathering stages of investment
considerations. The described ecological divisions of the study
region were delimited cartographically and further subdivided on
pedologic and climatic criteria. Using these and other environ-
mental criteria, a series of complex patterns were derived showing
the spatial variety of potential agricultural profit for each of
the cash crops included in this study. When the concern is with
road investments, as it is here, priorities should be based on
the potential increment in profit levels rather than the absolute
maximum profit totals. After combining the patterns of potential
profitability improvement for the several cash crops, these
patterns were adjusted to account for the relative availability
of land for such production vis-a-vis land needed for foodstuffs.
The resultant distribution of profit potential indicated sizable
areas of high and moderately high improvement levels in portions
of Kericho and South Nyanza Districts, respectively, with rather

low profit per acre improvements likely in Kisii District and

 

.. .9- .

minima

course
improx
throug
resul‘

ROW 1:

264
minimal profit gain capabilities throughout Central Nyanza.

Road investment priorities cannot be determined, of
course, solely on the basis of possible agricultural production
improvements. The characteristics of the roads, the regions
through which they are to pass, and the network changes which will
result must all be examined. It is to this task that attention is

now turned.

CHAPTER IX

THE RELATIVE INVESTMENT PRIORITY INDEX

Where to invest in road improvements in agricultural areas
is a decision that must he based on a consideration of both the
support to be gained from the land's output and the costs of im-
proving the efficiency of movement to and from the land. Deriva-
tion of the former factor has been outlined in detail in the pre-
ceding chapters. SeVeral facets of the second consideration must
now be examined. The importance of improvement costs and annual
maintenance costs must be shown in light of the need to provide
access to areas of high potential farm profitability improvement.

The sources of variation in both construction and maintenance

(I)

cost should be discussed. The applied pattern of the sum of
these costs, taking into account the sources of variability

where they apply to the region concerned, must also be derived.
It is only after these steps are completed that a decision may be

made on the relative investment priorities for road projects in

the agricultural region under study.

The fattern of Investment Costs

 

Just as proper justification for road investments must be
based upon the true potential improvement in agricultural output,

265

266
some consideration must be given to costs which will have to be
borne in order to facilitate this rise in farm productivity.
Obviously, it may not be feasible to encourage output increases
by improving the means of transportation, even from land which
is extremely well-endowed with agricultural resources, if this
land is also very distant from the major markets. Similarly, a
road which requires frequent and expensive maintenance in order
to keep its running surface at a given quality level will be a
less attractive investment than one which does not, unless the
expected benefits are proportionately higher on the former road.
Investment in transportation projects should not be undertaken,
non-economic factors aside, if costs exceed benefits over a given
period. The decision to invest in one road improvement rather
than another should be based on the relative excess of potential
gain over probable cost for the two investments.

In applying this elementary economic principle to the
situation with which this volume is concerned, the variations
are numerous but the theme remains. As has already been shown,
it is reasonable to expect a considerable spatial variety of
farming profitability improvement in the near future. To say the
same thing in another way, there are wide differences from place
to place in the quantity of returns which can be expected from
improved produce movement facilities. By the same token, the

heterogeneity of nature and the different levels of existing road

267
surface, present in spatial as well as qualitative variety,
results in a very complex pattern of investment costs. As it is
necessary to examine a region as a whole in order to determine
investment priorities, for roads examined in isolation are un-
realistically removed from the spatial context in which they do
exist on the earth's surface, the combined pattern of costs and

that of returns are both required in this priority determination.

The Variable Costs 2: Road Improvement

 

 

The costs which must be borne for annual maintenance and
for the initial construction or quality improvement of the
road surface are extremely varied and in some cases, rather dif-
ficult to determine. This variation is in large part a result of
the spatial heterogeneity of the earth's surface features. Road
construction and improvement costs will be different for every
unit distance of a single road. The sum of these units will vary
so that average construction cost per mile will not be the same
for different routes and different roads. Spatial variations in
relief, climate, drainage pattern, and soils are usually sufficient
to cause these different costs. In addition, however, there are
numerous human decisions which inject cost variation into any road

1

improvement considerations. For example, problems of traffic

 

1For a discussion of these and related problems, see,
William W. Hay, An Introduction £2_Transportation . . . , gp. cit.,
pp. 377-393.

 

 

268
allocation must be decided. Where will traffic demand be greatest
in the near future and by what mode of transport should this
traffic move? An estimate must be made concerning the proper
level of demand to be accepted as having the most efficient cost
results. To what traffic level between mean and peak flows should
the road capacity be adjusted in order to avoid both the high
costs resulting from congestion and the waste of excessive invest-
ment? Furthermore, to what economic degree is time spent on the
road an important factor in operation costs? And in detailed
routing of the road, to what standard grade and alignment will the
construction be designed? Are there "works of man" to be avoided
or approached by the route? With this variety of cost differences,
it is easy to accept the individuality of a road's initial or im-
provement construction costs.

There are severe difficulties for the present work result-
ing from this characteristic of transport engineering economics.
In order to provide a balanced accounting of costs throughout the
study region, it would be necessary to survey each major length of
road between junctions to determine the initial cost of upgrading
the road in question from its existing quality to the next highest
road grade. Some estimate would also be required of construction
costs for roads into areas now sparsely settled but containing
land of high agricultural potential. These surveys need not be in

great detail, for the preliminary nature of the methodology out-

269

lined in this volume does not justify this. It would be necessary,
however, for a team of trained highway engineers, familiar with
the environment and costing situation of East Africa, to calculate
roughly the financial requirements for every significant road in
the study region.2

Although it is not possible at this stage of research,
for the reasons discussed, to present detailed estimates of im-
provement costs for all roads in the study region, general levels
of cost may be used here as preliminary guidelines. Detailed
surveys of several parts of the study region have been carried out
and the improvement costs were calculated for the major routes in
the area. Thus, figures are available for the cost of upgrading
roads in flat to rolling terrain,3 and in rolling to quite hilly

topography.4 The stages through which road improvements are

 

2While this author traveled extensively in the study
region over most of the main roads and many of the minor ones,
he had neither the training nor the time to provide realistic
cost estimates on the scale needed.

3Sir Alexander Gibb and Partners, Report on Agricultural
Roads . . . , 22. cit.

4Republic of Kenya, Ministry of Works, Loan Application
for Trunk Road Development 1967-1970, "Preliminary Project
Report on the Wanjare - Suna Road B4/3," (Item A2) and "Pre-
liminary Project Report on the Ahero - Nyakoe Road B4/l & B4/2,"
(Item A3), August, 1966; and the large table compiled by Eduard
Irgens which was enclosed in this application, DProposed I.D.A...
Trunk Road Programme 1967/1970 : Discounted Cash Flow Analysis,"
Ministry of Works, Roads Branch, drawing number MPR/lOS, August,
1966.

270
measured are more numerous than the three quality levels used in
Chapter IV when assigning vehicle operating costs. Although there
is a difference in terminology used to refer to these road grades
in the several reports, the actual levels of road quality are
similar in corresponding categories. A "Type II gravel road" in
one report,5 for example, is apparently the same as a "Grade 4
road" in another.6 The quality differences between roads are best
described by Irgens (see Table 9-1), and it is with these general
grade categories in mind that average improvement costs are

offered for use here.

TABLE 9-1

ROAD GRADE CLASSIFICATION

 

Grade 0 - Unimproved track
Grade 1 - Slightly improved track
Grade 2 - Improved track
Grade 3 - Gravel Road, slightly improved
Grade 4 - Improved Gravel Road,

or low standard tarmac road
Bitumen Road (paved)

 

In addition to the paucity of data concerning road im-
provement costs in the study region, the individualized character
of such costs makes the process of deriving average costs dif-

ficult. The Gibb and Partners report, for example, cautions

 

SKenya, Ministry of Works, Loan Application . . . , pp. cit.

 

6Irgens, "Proposed I.D.A. Trunk . . . ," pp. c t.

271

. phat the soil conditions prevailing 1 on the Kano
P1ain_/ together with the lack of suitable surfacing
materials within reasonable haul distance have contribu-
ted significantly to the high overall cost. The average
cost per mile of road cannot therefore be directly com-
pared with the cost of similar roads constructed else-
where in more favourable conditions.7

Further questions are raised when the same road is assigned two
widely divergent cost per mile estimates in different portions of

the same report! Thus, the bitumenizing of the road from Wanjare

 

to Kamagambo (south of Kisii town toward Suna) is estimated to
cost £19,000 per mile in the text of the loan application,8 while
the same improvement for the same road is set at £26,400 by
Irgens.9 The cost per mile figures used in the present volume,
then, are obviously no more than indications of the real costs of
improving any road in the study region. The calculated average
grade-to-grade road improvement costs per mile which will be used
in this volume, rounded to facilitate computations and to avoid
any mistaken impression of detailed accuracy, are shown in Table

9-2.

The Variable Costs pf Road Maintenance

 

 

The sources of variation in road maintenance costs have

fortunately been grouped into more clearly defined categories than

 

7

 

Report pp_Agricultural Roads . . . , pp. cit., p. 40.
8Kenya, Loan Application . . . , (Item A2), pp. cit.,

Appendix A.2 (IV).

9Irgens, "Proposed I.D.A. Trunk . . . , pp. cit., column
7, line A.2A.

272
TABLE 9-2

CALCULATED AVERAGE IMPROVEMENT COSTS PER MILE,
BY ROAD GRADE

 

rounded calculated
Road Grade: from to cost/mile cost/mile

 

gravel (4) bitumen £10,000. £9,990.
gravel (3) gravel (4) £ 9,000. £8,936.
earth (2) gravel (3) £ 4,000. £4,130
earth (1) gravel (3) £ 5,000. £4,754.
earth (1) earth (2) £ 3,000. £2,888.

 

have the sources of improvement costs. Because the administration
of regular annual maintenance grants has been under the jurisdiction
of official agencies--first the Road Authority, and since July 1,
1964, the Ministry of Works' Road Branch--justification for region-
al variation in demand for these grants has been carefully
scrutinized. In a recent examination of Kenya's road maintenance
costs, it is reported that since 1952, "the allocation of mainte-
nance grants was calculated as a certain rate/mile from a formula
based on traffic density with modifying factors for local climatic
conditions and the availability of materials."10 From the basic
rate, governed by the traffic to be borne, an addition or subtrac-

tion of funds may be made if the rainfall and the relative availa-

 

10British Road Research Laboratory, "A Study of Road Mainte-
nance Costs in Kenya," East African Transport Planning Research
Unit (Kenya), KN/Z (unpublished), p. 2.

273
bility of suitable surfacing material warrant this modification.

The schedule of reported incremental factors is shown in Table 9-3.

TABLE 9-3

MODIFICATIONS TO THE BASIC ANNUAL ROAD MAINTENANCE GRANT

 

Condition Increment

 

Average Annual Rainfall

0 - 20 inches - 5%
21 - 40 inches 0%
41 - 60 inches + 5%
above 60 inches +10%

Surfacing Materials

Location requiring haul over 5 miles + 5%
Hard (vis-a-vis soft) material + 5%

 

Source: British Road Research Laboratory, "A Study
H

, pp. cit., p. 2.

In addition to the basic road maintenance grant, variable
in the manner described, two further factors are applied toward
total maintenance costs. A small sum is paid to local authori-
ties throughout Kenya for the administration of local road mainte-
nance. This figure varies with the daily traffic density of the
roads and is one-quarter of the sum of the basic and the re-
gravelling grants. A much larger factor in total costs is this

re-gravelling grant. As is noted in the Road Research Laboratory

report:

274

Engineers have found from experience that the loss of gravel
from the surface of the road is of the order of l" per year
for every 100 vehicles using the road per day. The cost of
replacing gravel is not likely to be less than £600/mile

for a 4" compacted thickness, two lanes wide.11

The total cost of maintenance, then, is the total of these three
elements plus any modifications based on climatic or materials
characteristics. Table 9-4 summarizes these total costs by road

"class,' where the class is determined by the traffic density.

TABLE 9-4

TOTAL COST OF ROAD MAINTENANCE FOR ROADS EARNING
THE BASIC GRANT, 1964/65
(POUNDS STERLING PER MILE PER YEAR)

 

 

Class Vehicles Basic Re-gravelling Administrative Total
of per Grant Grant Grant Cost
Road Day
I 25-50 43 50 23 116
II 51—100 66 100 42 208
III 101-200 89 150 60 299
IV 201-300 150 200 88 438
V over 300 270 250 130 650
bitumen:
l—lane 200* - 50 250
2-lane 265* - 66 331

 

The rates quoted for bitumen roads include periodic re-
sealing and resurfacing at about 5-year intervals.

Source: British Road Research Laboratory, "A Study . . . ,
pp. cit., p. 9.

 

11
Ibid., pp. 6-7.

275
Using the data presented here in Table 9-4, the author
of the Road Research Laboratory note on maintenance costs derived
an equation describing the relationship between the total cost of
maintenance for gravel roads and the traffic flow in vehicles per
day. The linear equation is T.M.C. = 95 + 1.370 using the mid-
flow values, where T.M.C. is total maintenance cost and Q is the

12 The equation states that the total

average daily traffic flow.
maintenance cost, in pounds sterling, increases 1.37 times for
every unit increase in vehicle flow, over the initial £95. of
fixed costs (in this case representing maintenance of non-surface
portions of the road, e.g., ditches and culverts). By adding the
climatic and material availability increments, where suitable, to
the T.M.C. derived from this equation for each road in the study
region, it should be possible to cartographically represent the
pattern of road improvement costs.

Some of the most important, and typically some of the
least available, data needed in order to construct the map of
total road maintenance costs are the complete traffic flows for
all roads in the study region. Several surveys of traffic volume

have been made in selected portions of the study region or include

the Nyanza Region and Kericho District within a larger survey

 

 

276

area 13

The data resulting from these surveys are not wholly
satisfactory, however, for not only are the data reserved to the
few main roads in the study region, but they are also widely
divergent in the level of average daily traffic reported for the

same locations on several roads.14

Later applications of the
present methodology should include traffic counts on all roads
in the region, from Grade 2, which is generally a packed earth
road, through bitumen. Grade 1 roads, or motorable tracks, may
be checked in the areas containing top priority investment pos-

sibilities (as determined below) but will generally have fewer

than 25 vehicles per day.15 With suitable traffic data, total

 

 

 

13Kenya, Loan Application . . . , (Item A2) and (Item A3),
pp. cit.; Sir Alexander Gibb and Partners, Report on Agricultural
Roads . . . , pp. cit.; and, British Road Research Laboratory,

”Preliminary Note on the Origin and Destination Survey of Road
Traffic in Kenya 1964/65," East African Transport Planning Re-
search Unit (Kenya), KN/4 (unpublished), p. 35.

14 .
Although unusual circumstances might cause a conSidera-

ble variation in traffic counted during different surveys, the
variation reported is disturbingly large. The Road Research
Laboratory survey counted 175 vehicles per day just outside Kisii
town on the road to Oyugis, while the Ministry of Works loan ap-
plication reports 381 vehicles per day as the average. - British
Road Research Laboratory, "Preliminary Note . . . ," pp. pip.,

p. 35; and, Kenya, Loan Application . . . , "Preliminary Report

on the Ahero - Nyakoe Road . . . , pp. ci ., Appendix A.3 (II).

 

15Roads of similar quality used for tea collection have
traffic generally less than ten vehicles per day. See, John A.
King, Jr., Economic Develppment Projects and their Appraisal
(Baltimore: The Johns Hopkins Press, 1967), p. 302.

277

maintenance costs may be determined for each road in the region
being studied. Proper increments would, of course, be added to
or subtracted from the basic grants because of climatic or
material availability variations.

For present purposes, as a means of deriving a pattern
of road costs reflecting both construction and maintenance costs,
a series of assumptions will be made regarding road traffic

16 the study

patterns. Although not strictly an assumption,
region's roads are re-classified into five categories, rather

than the three shown on Figure 4-2. A distinction is made here
between one-lane and two-lane bitumen roads. The sole one-lane
bitumen road in the study region connects Sotik with a short two-
lane bitumen road extending south from Kericho town. The only
gravel roads realistically called "all weather" roads are (ex-
cluding short bitumen sections): Kisumu through Yala, Ahero to
Miwani, Ahero to Kisii via Oyugis, Sondu to Sotik, Sotik to Kisii
via Keroka, and Kisii to Kamagambo. These roads will be assumed
to carry traffic suitable to Classes IV and V (see Table 9-4).

The remaining "unpaved" roads shown on Figure 4-2 are generally of
mixed quality and seemed more suitable to a combined gravel/earth

category. These roads will be assumed to carry Class II and III

traffic, or an average 50 to 200 vehicles per day. Those routes

 

Some justification for the new categories is provided by
qualitative judgements based on the author's personal observations.

278
connecting major centers will be assumed to carry close to the
maximum traffic in this class while roads of less obvious im-
portance may be expected to carry lower traffic volumes. The
Oyugis - Kendu Bay road, for example, appeared to the author to
be of much higher general quality than, say, that passing through
Mirogi to Lake Victoria. The remaining roads, or "vehicular
tracks," will be treated as though they carry daily traffic at
the level of 25 to 50 vehicles, undoubtedly an excessive estimate
in many cases. These traffic volume assignments permit here, at
the least, a consistent assessment of relative maintenance costs
for each road within the study region.

It is not, however, costs based on existing traffic which
are of primary concern here. Because it is the total road costs
to be borne i£_road improvements are made that need to be cal-
culated, estimates of future traffic volumes are required. In
Kenya, it has been found that road improvement leads to a 40
percent increase in daily traffic during the first year after up-
grading and an 8 percent increase thereafter. An average 10
percent increase per year could be assumed when calculating mainte-
nance costs.17 The pattern of costs based on these maintenance

requirement estimates must be added to the varying upgrading costs,

 

17Republic of Kenya, Ministry of Works, Loan Application
for Trunk Road Development 1967-1970, "General Information on All
Road and Bridge Projects," August, 1966, p. 2.

 

279
the latter converted to annual costs by using the appropriate
loan interest rate and amortization period. In the present case,
a 20 year loan period and a 4 percent rate of interest will be

used in calculations.

The Problem pf Absolutes

A serious difficulty in the logical development of the
present method emerges when consideration is given to the problem
of obtaining a meaningful and applicable measure of the variable
contribution which can be made by upgrading the study region's
roads. It is desirable to make this complication explicit so
that it may be treated directly, and more importantly for the
future, so that the research necessary to resolve the matter may
be conducted.

The road quality to which all roads can be improved, in
the end, is the highest quality--two-lane bitumen. It would seem
theoretically correct to measure costs and relative well-being
derived from the road improvements to this level. It is only by
reference to the best movement facilities (based on the costs of
reaching these levels and the increased benefits to be derived

from the new facilities) that an absolute measure of road improve-

 

18The factors to use in calculating annual capital recovery
for these, and other, rates of payment are shown in tabular form
in Hay, Ap Introduction . . . , pp. cit., p. 390.

 

280

ment costs and returns may be obtained for comparison with the
level of potential agricultural improvements. Without such an
absolute measure, it might appear, there can be no consistent
reference point to provide a basis for investment decisions.

0n the other hand, it is not realistic to expect all
roads to be upgraded to bitumen quality. Recent research in
southwestern Tanzania under Robert Bonney has indicated that the
move to bitumen road quality may be justified, ". . . where the
proportion of medium/heavy traffic is especially high . . . at a
traffic volume as low as 60 vehicles/day - depending on the
standard of maintenance."19 It is not likely, however, that the
funds for such extensive road improvements will be available when
they are needed. Furthermore, it is more meaningful to devise a
method of investment priority determination which allows com-
parison between, say, improvement from track to earth/gravel
quality and improvement from gravel to bitumen, both in terms of
the facilities required to reap the possible returns from improved
agricultural production and the relative levels of these pos-

sibilities. What is needed, more specifically, is a justifiable

 

19R. S. P. Bonney, M. N. Daniel, and A. T. Noorani, "Traffic
and Road Characteristics in South Western Tanzania," E.A.T.P.R.U.
Research Note No. 5, British Road Research Laboratory, Ministry of
Transport, East African Transport Planning Research Unit, Tanzania
Branch, May, 1966, (unpublished), p. 35.

 

281
method for choosing the optimum location and degree of road up-

grading with regard to expectable returns. But if only the poten-

 

tial returns are used pp p basis for compprison, the differences

 

 

 

pp returns from transport improvement are ignored, for regardless

 

 

pf the amount pf transport route ppgrading, the potential returns

 

 

from pgriculture are the same. This is a formidable dilemma. To

 

resolve it, it is necessary to return from theoretical considera-
tions to the reality presented by meager research results.

There are no known guidelines for correctly calculating
the actual returns resulting from road investments. This has
been emphasized before (Chapters I and IV). Cost-benefit studies
do attempt to estimate these future returns, but with very few
exceptions, even the direct returns from increased productivity
in areas adjacent to the new or improved road have been "esti-
mated" with no more than a cursory regard for empirical justi-
fication. In less developed regions, these studies have been
hindered by a lack of long-term, detailed statistical records.
Although increased production has been observed in a number of
20

cases (see Chapter I), only in the study by Smith in Uganda,

and possibly Hawkins' later research in the same country,21 are

 

20N. D. S. Smith, "A Pilot Study in Uganda . . . ," pp.

0
n

21E. K. Hawkins, Roads and Road Development . . . , o . cit.

 

282
there cases in Africa which lend quantitative indications of pos-
sible production increases.22 Other secondary benefits are even
more difficult to measure.

The only immediate benefits which deal with individual
movement costs23 and can be used with reasonable assurance of
accuracy are the reduced operating costs that follow from improved
road surfaces. Considering this limited source of dependable
return data and the objectives of the present methodological of-
fering (i.e., that a system or priorities be established concern-
ing the best alternative among numerous investment possibilities),
it is more meaningful to measure comparative investment possibil-

ities with respect to the maximum returns that can be obtained

from the land.

The Priority Index

 

There are three primary variables involved in the present
problem-~agricultural potential, existing road quality, and the
priority for investments. It is only by fixing two of these

factors that the third may be eXplained. By dealing with maximum

 

22Robert Bonney, who had worked on this problem on Sabah in
the East Indies, was engaged in research during the latter part of
1966 in southwestern Tanzania on the relationship between agricul-
tural production and its proximity to movement facilities. His con-
clusions may provide a partial solution to the present problem.

23Other quantifiable benefits, such as license fee increases,
higher fuel tax income, and the like, are basically a function of
increased traffic volume.

283
agricultural potential, this factor is dealt with in a standardized
form even though the level of potential may vary from place to
place. In the same sense, although existing road quality is
variable from road to road, that aspect of this factor crucial
to the present undertaking may be made constant by standardizing
the improvement costs to be invested. Thus, by relating the same
level of road improvement costs to the known returns which result
from these investments (the reduced vehicle operating costs), re-
gardless of the initial road quality, and then by relating the
resulting factor, in turn, to a consistent measure of agricultural
potential (the maximum potential, as defined by explicit criteria),
the variable ratio which results indicates the routes' respective
priorities.

This relationship, resulting in the establishment of in-
vestment priorities, may be more clearly seen by stating it in
another form.

Relative Priority Index =

Potential Improvement in Farm Profitability

 

construction costs + maintenance costs
reduction in vehicle operating costs

It is vital that the resulting fraction be seen as a quantitative
statement of relative investment justification. These fractions
have meaning only with respect to one another and within the system

of roads and agricultural regions being examined. The index rela-

284
ting to a given route cannot be compared with some standard value,
for none exists, and a large or a small result must be analysed
within the context of the complete range of values for all similar
routes in the study region.

Careful consideration of the manner in which this rela—
tionship is expressed, above, reveals the simple consistency in
this statement of any road's relative investment priority level.
The denominator is an expression of the relation between road in-
vestment costs and the benefits to be derived from these invest-
ment expenditures. If returns are large relative to annual out-
put, the denominator is small; it is large when costs are rela-
tively large compared to returns due to the reduction in vehicle
operating costs. The larger the returns relative to the costs
(i.e., the smaller the denominator), the larger will be the
priority index for a given potential profitability improvement.
It should be noted that the denominator need not be less than 1.0
(i.e., costs may be greater than benefits) in order that a route
achieve a high investment priority rating. If the agricultural
potential is sufficiently large relative to the potential else-
where in the region, a route which might be rejected by cost-
benefit analysis may still be justified as the top priority in-
vestment project.

Although it has been frequently stated that the present

concern is with the setting of priorities, not with the establish-

285
ment of acceptable or unacceptable grounds for investment, the
author should make a related point clear. It is recognized that
there is a danger that this methodology may strongly indicate the
best of a dozen alternative investment projects, all of which are
wasteful, because the region of concern is too small. It is also
believed, however, that considering the present stage of agri-
cultural development in most African countries and the funds
available for road improvement, there is little danger of over-
investment in transport facilities. These nations contain
regions which, even on the basis of broad (but informed) sub-
jective judgement, are of high agricultural potential yet lack
good internal communications. If the proper statistical records
are kept during and after road improvements in these regions, suf-
ficient data should be available to calculate the proper mathe-
matical constant indicating the level at which high agricultural
potential is not adequate to overcome an excessively large de-

ficiency of returns relative to costs.

The Application

 

The major task which remains is to demonstrate the manner
in which the measure derived above may be applied to a real
situation resulting in the establishment of a series of road in-
vestment priorities. As with earlier portions of the present ap-

plication, a number of assumptions are necessary, but also as

286
before, they are made explicit, are eminently testable given
adequate resources and correctable if required, and do not
drastically affect the outline of the methodology.

Because of the limitations imposed on the analysis by
the requirement of a consistent increment in road improvement
cosrs, the value used should be suitably proportionate to the
average improvement costs between road grade categories. By
examining Table 9-2, several alternative factor combinations
are apparent in the calculated improvement costs per mile. The
proportion to be used in this volume is that applicable to the
costs required to upgrade roads from Grade (1), basically an earth
track, to Grade (3), a mixed earth/gravel road, from Grade (3) to
the top quality gravel road, Grade (4), and finally, from Grade
(4) to a bitumen surface. This alternative was chosen for several
reasons: the proportion of £5000 to £9000 to £10,000, or 1 to
1.8 to 2, is very manageable, and most vehicular tracks in the
study region are of a quality that is best described as unim-
proved, or Grade (1).

This ratio, 1:0 to 1.8 to 2.0,,represents the inverse of‘
the relative class improvement by investing similar amounts of
capital into each quality of road. That is to say, by spending
£5000 on every mile of road in the study region, all earth tracks
would be improved to the level of Grade (3), or earth/gravel. All

Grade (3) roads would be improved, but none would be upgraded suf-

287
ficiently to place the bettered road in Grade (4). The quantita-
tive improvement in costs of movement and traffic and maintenance

costs is only _l_ (or 0.55) times the Grade (3) to Grade (4) in-
1.8

crease in these factors. The £5000 invested in every Grade (4)
road results in an increase of half that which would have occurred
had the full improvement to two-lane bitumen been possible.
Clearly, several assumptions must be made at this point.
The change in traffic and maintenance costs and in operating
costs for those investment situations where the full grade in-
crease cannot be made with the investment increment chosen to ap-
ply to all roads (the costs of which are indicated in Table 9—2)
are assumed to be proportional to the reduction from the total
investment required. Although under ordinary circumstances, im-
provement costs are not reported on projects resulting in less
than a full grade increase in road quality, this does not pre-
clude its feasibility and in many cases because of scarce funds,
its desirability. That returns will be proportional to inputs is
also assumed, but this is consistent with the linear relationship
between these factors with reSpect to maintenance costs and traffic
volume.24

The second assumption which must be made is related to

bitumen roads. Much of the road between Kericho and Sotik is

 

24British Road Research Laboratory, "A Study of Road

, pp. cit., p. 9.

288
bitumen but only a single-lane bitumen road with gravel shoulders.
Because there are almost no data relating to one-lane paved roads,
the costs of improving to or from this level and the effects
generated by improvement will be assumed to be half that required
to upgrade a Grade (4) gravel road to two-lane bitumen. Although
it will be admitted that the empirical foundation for this as-
sumption is weak, firmer knowledge could easily be substituted for
the £5000 cost per mile "convenience figure” used in the present
work.

It is easiest if the relationship between the investment
inputs and outputs (i.e., the denominator in the relative priority
factor) is determined in detail before any attempt is made to
combine it with the spatially variable agricultural potential.
Although any application of the method, when full traffic data
are available, would require a more specific set of calculations
for each length of road, the limitations of these data on the
present, preliminary application allow generalized factors to be
determined for each level of road improvement. In this applica-
tion, the "before investment" levels of daily traffic were set
at 35 vehicles per day on regional "tracks," 125 vpd on the mixed
earth/gravel roads, an average of 300 vpd on the Grade (4) gravel
roads, and 300 vpd also on the only l-lane bitumen road. The
calculations following from these figures are summarized in

Table 9-5.

289.

.Hw4.+ .ommfl n .momd + .owm + .U.Z.H

aw mwcuso o i n u w
umo%\ammm umm%\ommm umm%\mmmm umo%\mNNM ucwsm>ouaaw
paou Houmm m
umo>\ommfl umo>\oomw uao>\oomm umom\mqam ucmEo>ouaEH
pmou muommn o
€20.24.
woumasoamo
ea> 0mm ea> amm ea> ass ea> Ne flavouuuauu
mauwsm> >>umn m
ea> mmm ea> amm ea> «mm sa> me ewemeo
oESHo> owmwdpu N i m u s
pm> amm pa> mmw pa> QO pa> mm Aavucm8o>oumEH
Hound uwmmmuu m
pa> oom pa> oom pm> mNH pm> mm ucmso>oum8w
ouomon oflmwuuu N
ooomm ooomm ooomm ooomm oHHE\umoo
sowuosuumcoo A

cu u Eoum oo . Eouw 0o u Eoum ou i Eouw
.uan mcmaim .uwn mcmHIH Amvao>muw Ho>muw\cuumm xomuu

 

NMQZH VHHmonm HszEmm>zH Mme mom mOHo<m Q¢Om mmH ho ZOHE<ZHZMmHmn

mum mdm¢H

290

.ucosumw>cfl
poems mumoh 0H owmmmuu woumEHumo mum :3onm mouswwm .uoummmuonu pooh some ommmuocw uaouuoa m cm cam
ucoEo>ouaEH Houmm new» umuflm onu cw ommmuocw ofimmmuu unmouoa oq m no woman mum macaumasuHmuAnv

.AmV one: mom .uoHHme maoumcowuuoaoua mH AHA 3ouv muwoo

wswumummo oaofinm> :H cowuospmn ego .ucmEumo>cH one mp pmuuoaasm on.mmw :owfiB uaoEo>onaEfi HmHuumo
umsu pom .mum: pew: osmaumo>cw wouflpumpGMum one amsu “museum ma Aa souv memo» Hm>muw Aqv opmuw
Hazy ou memo“ Ho>mum\nuumo map e>ouaEH ou wouwsvou umoo cowuosuomaoo Hasm can omsmoom

 

 

 

Amy
meadow“ .umo>sw ma
mN.H Hm.o oo.~ oo.m iflflmoo .umo>aw Mfl u ea
oH.mema oH.momu mm.mmmm em.omaa vamca>mm umoo
wcwumuomo .saw ma
em\HN sm\HN am\aa mo\oH Queues“ s>uoe
Ham .mmcw>mm
eases .m>m m x as u NH
oo\- oo\- mH\- e~\- Amv.we\umoo wage
umumao mHoH£w>
CH coHooswmu Ha
m.w¢¢@ o.HHHm m.¢mo$ m.mmqm .umm>cw Heuwm
.H8\umoo .ccm w + m H OH
m.moma a.momw a.~omfl. m.momd onmawa\umoo
.umcoo Humans m
ou i Eouw 0o . Eoum ou s Eouw cu . aonm
.uwa museum .an ocmHiH Amvfim>wuw Ho>muw\nuumo xomuu

 

wmdcwucoonnmim made

291

.mmmpcsm co umwxm cu peasmmu mw mEDHo> owmmmnu kawdp HmEuoc ecu mo mamnimcoawv

.oo\- " eo\- “ NN\- " em\- " om\u ma oases menu .mxoauu cos m-m pom mans use mwceauaem cH
.cwEDan «swarm ”Emanufln ecaaua “Hm>wuwufie>auw\nuumm"xoauu .paou mo wwwmuw one so mumoo wcwuuuwmo
somzuon :oHomHeu ecu pom o.~ " H.~ . ~.H “ o.H " o.~ owuou one so woman mum mouswwm ommneamv

.umeumucfi Hanson unwouoa d um powuoa “mom
om m um>o ponwuuoEm .zuwamsv econ some cu pwfiaamm endow ooomm use monomeummu msaa> mHnHon

.mnowuuwuu> muwawnaaww>d

Hafluouoa no oflumewHo HwooH mo mwmwn one :o powmwumsm mucmswuocfi owsaoaw uo: op mouswwm «mosH

.wfinmu mane cw soap ecu paw .qim wfinme oo wouaaou scammoomwp uofifiuam one mew .Bon owmmuuu mawmp
swans: m5 3 c was .034 + mm u .059 memes .38 0255338 :33 of 3 6.23.23

.oflmmmuu Hmuoo use mo mpuwnunozu we cu poaswmm mg owwwouu “masofinm> >>moonv

vmscfiucooiumim mqm<H

292

The calculations necessary to determine the transport
element of the relative investment priority factor are not dif-
ficult once the preliminary cost and return information has been
gathered. From a beginning traffic volume and a known rate of
traffic change, given road improvements, the actual change in
volume of traffic may be determined. The difference between
pre-improvement and post-improvement traffic yields the altered
maintenance costs which must be borne annually because the road
quality was upgraded. The sum of these generally larger mainte-
nance costs and the construction costs required to allow this
increased traffic (in the present case, a constant £5000 reduced
to a realistic annual cost) yields the total costs which must
be borne because of the improved infrastructure. By then calcu-
lating the annual benefits received because of the better road
surface, here limited to the reduction in vehicle operating
costs, a benefit level is derived comparable to the total cost
level. This, of course, allows calculation of the denominator
of the investment priority index for each road in the study
region.

Several observations may be made concerning the figures
shown in Table 9-5. It should be remembered that the ratios of
costs to returns cannot be taken as absolute cost-benefit ratios.
Aside from the restraining assumptions, the ratios derived in

the manner indicated tend to be very pessimistic because the only

293

benefit counted in the calculation is that resulting from lower
movement costs. Keeping in mind the many data qualifications,
it would appear that the top-grade gravel roads in the study
region should be upgraded to at least one-lane bitumen quality.
The low figure of 0.31 leading to this conclusion is, of course,
a function of the existing high traffic volume on these roads25
and the sharp decrease in annual maintenance requirements after
a road surface has been paved.26 These conclusions are supported
by the construction work already begun on the Ahero-Kisii road in
late 1966 preliminary to applying a bitumen surface. In any
case, it is these ratios between improvement costs and the change
in vehicle operating costs that are one half of the information
needed to calculate the roads' relative investment priority in-
dices.

The index which will provide an indication of a road's
relative position among a short priority sequence for investment

is calculated, as discussed above, by determining a quantitative

measure of the relation between potential improvement in farm

 

25Reported in, British Road Research Laboratory,
”Preliminary Note on the Origin . . . ," pp, cit., Figure 4.
26

See Table 9-4. Furthermore, Klein has reported that
Kenya's policy has been to bitumenize road with traffic exceeding
300 vehicles per day, judged very conservative in contrast
with Tanzania's general policy to bitumenize at 150 vehicles

per day. - Martin S. Klein, Economic Feasibility . . . , pp. cit.,

p. II-27.

294

profitability and the relative costs of road upgrading. In
other words, this measure describes the relation between the ex-
pense of improving a given length of road and the economic quality
of the land through which this length of road passes. There is
no assumption that the land will yield to its full potential be-
cause of easier and cheaper transportation, but it is believed
that this is highly probable.

As a guide in carrying out the calculations necessary to
complete the choice of top priority investment projects among the

”rules” were drawn up. These

roads in the study region, a set of
rules (together with some explanatory argument) follows:

1. The relative investment priority index is defined by

the relationship between the land's potential and the

road or roads passing across this land, as was stated

earlier. For a given length of road, the index equals

Potential Improvement in Farm Profitability

 

costs resulting from road improvement
returns received from road improvement

 

2. The land for which the agricultural potential is con-
sidered responsive to changes in road quality is
limited to one mile distant from the road adjacent
on each side. This distance is considered here (sub-
jectively and arbitrarily) the no-profit margin for
walking large quantities of produce to the road. Land

lying at a distance greater than one mile from the

295
nearest road, or track suitable for wheeled vehicle,
is taken to have a sufficiently reduced profitabil-
ity to be deemed non—productive for present purposes.
It is recognized that this margin is probably located
too close to the road to accurately represent con-
ditions in the study region. There is, however, no
known function, applicable in Africa, expressing
the decline in profitability, productivity, or at-
traction to the farmer with distance from the nearest
motorable route.
The important factor in priority determination is the
ratio between agricultural potential and road returns
for the costs incurred. The index resulting for pppp
pile of road (based on Table 9-5 and on Rules #1 and’
#2) indicates the relative priority which should be
given that length of existing road.
In order to deal with realistic road lengths, the
priority index will be determined on the basis of road
lengths longer than the one mile units of Rule #3.
The indices for each mile will be summed over the
length of the connected road. The maximum resulting
12EE§.2£ combined, one mile units will be taken to
represent the best investment possibility among the

various composite routes.

296
Each route whose index is to be compared with others
must have at least one direct connection (not through
another route) with a non-improvable (i.e., a two-
lane bitumen) road which, in turn, connects with one
or the other of the primary regional market outlets,
Kisumu or Lumbwa. Of course, this rule limits to some
degree the flexibility of the present application, for
there may be areas of high agricultural potential
proximate to non-bitumen truck routes, but having no
good connection with a major bitumen road. This may
be shown in future applications if desired; it is
within the scope of the method. Here, however, im-
provement priorities will be shown in light of the
development of the region as a whole. This procedure,
combining the economics of improving a (non-bitumen)
trunk road with those of the feeder roads, is justified
by reference to a more complete examination of the
relations involved. It is only by considering the
entire road length, as long as the priority index re-
mains large, that the true relationship may be shown
between the rural profitability potential and the means
of moving the products of these rural areas the entire
distance to the market.

Routes not presently in existence extending into re-

297
gions of high, but as yet undeveloped agricultural
potential will ppp be considered. (Suggested areas
requiring such a road net may be seen by implication
from the combined patterns of Figures 4-2 and 8-13.)
Using these broad guidelines, the calculations may be made which
yield the pattern of road investment priorities.

Before proceeding, a final objection might be made. The
resulting priority indices, it could be argued, do not accurately
represent the desired priorities because they do not include a
measure of p£}p_movement costs. Such indices would contain opera-
ting costs over bitumen sections of the total routes as well as
over those roads capable of improvement. This, however, is a
reversion to the old cost-benefit approach. If agricultural out-
put is increased sufficiently to result in income greater than the
costs of moving the goods to market, pppp the road construction is
justified. As has been argued before, this is not the issue. A
variable level of output will reach the market without better
routes; this is the existing level of product sales. The problem
solved by the present method is: Given the desire to improve
transport facilities, what agricultural pattern justifies, through
the potential improvement in cash crop profitability and the exist-
ing road connections with the market, the decision to upgrade the
roads connecting this area with the market? The costs of operating

vehicles on the paved roads are not directly relevant because there

298
is no change in these costs with a change in the condition of
other roads--assuming maintenance remains adequate-~until conges-
tion becomes a serious cost burden. The agricultural areas through
which the bitumen roads pass will not be assisted toward a closer
approximation of full potential output by better road conditions.
The road conditions, obviously, will not be better in this

instance.

The Regional Priorities

By applying the rules set down above to the data for the
present study region, a series of priority lines are established.
The successive sums of relative priority indices, drawn from the
points where unpaved roads contact paved roads, show very clearly
a stong dominance in encouragement for investment along a few routes.
Those routes having the greatest concentration of justification are
shown in the form of a flow map (Figure 9-1). Although the calcu-
lations resulted in numerical conclusions, the flow map offers
better visual support for the road investment priorities.

That road for which the relative investment prioripy index
remains the largest for the greatest distance 3p the road which
should take investment precedence over all others in the region.

In a situation where a decision must be made between a short road
with a high priority and a completely separate long road with only

a moderately high index, the shorter road should be upgraded first.

 

 

~ NYANZA REGIDN and Kericho District .

«“:7 FIRST
trim PRIORITY RANKING

 

 

 

 

 

ttoppflmhyimeflmem
second pfiorny inveshnent

. other roads

 

KENYA

 

 

 

 

 

 

FIGURE 9- 1

 

300

The larger index for the short road indicates that investment in
this route will make better movement facilities available to
land with greater potential gains in profitability than would an
improvement in the longer route. The leading statement in this
paragraph explicitly states the guide which should be used when
choosing the optimum route among the many which may be connected
to a single base.

Examination of the pattern of investment indices
(Figure 9-1) indicates that the route with the greatest justifi-
cation to be classified "top priority" for investment is the road
between Ahero and Kisii via Oyugis. Beyond Kisii, the extension
of top priority may not be clear from the map but the calcula-
tions show it to be the extension passing southwest from Kisii
through Kamagambo. The road toward Sotik is deceptively longer
with almost exactly the same index (382.4) just before reaching
the Kericho District border, as the southwestern road's index
(382.6) at Kamagambo. The choice between these two roads is clear
when it is pointed out that much of the apparently longer dis-
tance southeastward is over the bitumen section of this road from
Kisii. Adjacent to the town, a portion of each of these three
roads entering Kisii has been paved and cannot, therefore, be in-
cluded in calculations yielding the relative investment indices.

The explanation for this road's high index is not difficult

301
to make. The primary cause lies in the fact that this road has
a top quality gravel surface. The class into which the road would
be raised upon investment of the unit improvement decided upon
(i.e., £5000 per mile) is one-lane bitumen. Because of the sharply
diminished maintenance costs required by paved roads relative to
gravel surfaces, especially when the latter must bear a high
traffic volume, the change would cost less relative to that of
other roads. Thus, even the moderately high potential improve-
ment (Shs. 300/- to 400/-) of the small area through which the
road passes after Sondu is sufficient to raise the index sharply.
The low denominator and the occasionally moderate numerator values,
supplied by the land through which the route passes, are adequate
to maintain this large index as far as Kamagambo. Beyond this
point, the quality of the road changes with the result that the
index diminishes rather quickly. Nonetheless, the top priority
investment rating is justified ten miles byond Kamagambo along the
road toward Suna.

The length of the first investment priority route is
limited by the index size of the next best route for investment.
As the top priority route is followed from its base, it eventually
reaches a level equivalent to the investment index which is next
largest after that associated with this primary route. At that
point, the road classified as having the greatest relative justi-

fication for investment ends and a second, slightly lower priority

302
index road is followed to ipp termination, and so forth.

The setting of priorities, however, is not simply a
matter of choosing, in a step-by-step fashion, the next smaller
index among a map of these indices for the entire region. It is
important to remember that a change in one section of this net-
work of indices may result in an altered importance for other
portions of the net. The upgrading of the Ahero-Kisii-Kamagambo
road will alter the index of every road and track connected with
this route. The eventual road improvement may only be to a one-
lane bitumen quality or to a two-lane, 20 foot roadway; the
determination of first priority does not limit the improvement
to the standard level used in the calculation. This decision
should be made after the setting of priorities. In the present
case the high traffic volume already carried in 1964 and 196527
would seem to justify two-lane pavement unless funds simply can-
not be allocated. In either case, the cost of further improve-
ment, maintenance, and the reduction in operating costs will be
different for roads associated with this route, and the pattern
of priority indices will also be different.

Assume, for example, that the road from Ahero to Kisii,

and also the unpaved section of the road from Kisii to Kamagambo,

have been bitumenized to a width of two lanes. Furthermore, ten

 

27British Road Research Laboratory, "Preliminary Note
I!

. , pp. cit., Figure 4.

303

miles of the road beyond Kamagambo toward Suna have been up-
graded to a Grade (4) gravel surface. In addition, there exist
the desire and the funds to further facilitate the process of
moving goods to market outlets by improving the quality of other
road surfaces. It is appropriate to ask what new top priority
road investment may be seen.

Following these assumptions and demands, a new pattern
of indices was calculated. It was clear that the major priorities
remained with those roads of gravel quality which passed through
areas of moderate agricultural potential. The reduction in annual
maintenance costs which can be expected after reconstructing
these roads to bitumen surfaces was again a sufficient benefit in
its own respect to justify early upgrading. Thus, the investment
required to improve the 10 miles beyond Kamagambo raised to gravel
quality under the top priority investment and that required to
extend the bitumen portion of the road between Kisii and Sotik to
at least the District boundary are the next priority investments.
In order that undue importance not be put on the assumed invest-
ment desirability of upgrading 311 gravel roads, it should be
pointed out that the index calculated for the Kisii District
portion of the Sondu - Sotik gravel road does pp; justify an in-
vestment before several other roads' upgrading. In spite of the
sharp drop in annual costs after paving, the land in this corner

of Kisii District is apparently not capable of enough profitabil-

304

ity improvement to result in a high priority index for this
road.

The route which would justify a third priority rank
for upgrading is discovered in the same manner as previous
priorities. This route, as shown on Figure 9-2, is obviously in
Kericho District between the end of two-lane pavement just south
of Kericho town and a few miles beyond Bomet. In the case of
this route, it is the land's extremely high potential profitabil-
ity improvement that is the predominant source contributing to
the resulting large priority index.

It is interesting to note that improvements to the roads
included in the first two priority determinations are already
under way, or are at least to a stage where requests for the

d.28 In contrast to

necessary funds have been formally submitte
this, only the widening of the one-land bitumen section of the
third priority route has been included in the published program
of road development.29 A major portion of the route remaining

in this priority, between Litein and Bomet, is partially included
within the improvement plan for tea roads in this area, but this

30

inclusion is obviously coincidental. This reflects one advantage

 

28Kenya, Development Plan 1966—1970, pp. cit., pp. 281-82.

29Ibid., p. 283.

30See the maps accompanying, Edwards and Burrow, "Prelimin-
ary Report to the Permanent Secretary . . . ," pp. cit.

 

 

NYANZA REGIDN and Kericho District

NA, SECOND
ff ;\ PRIORITY RANKING

   
  

     
 

/ n
H 'muf~\,/H_YLI ’

 

 

 

 

 

m- EE% Priority Index Values

' top priority investment

  

" '3; other roads

KENYA

 

 

 

 

 

FIGURE 9- 2

 

 

306

of the objective determination of road investment priorities as
demonstrated in the present work; projects of potentially high
return are not by-passed in favor of programs of more obvious,
but shorter term remuneration. It should be remembered, how-
ever, that there are a great many data limitations in the
present application of this procedure. The agricultural poten-
tial might possibly have been exaggerated sufficiently to distort
the real priority ranking. Only more complete data collection
could support or refute this reservation.

It is also important to keep in mind the basis upon which
the final determination is made. If there are reasons, economic
or not, which justify road improvements through an area with
little or no potential profitability improvement, the present
methodology will not indicate this. The road connecting Kisii and
Homa Bay, for example, would obviously be very low on the priority
scale using the method outlined in this volume (see Figure 9-2),
and yet it is slated for upgrading as soon as funds are available.31
The purpose may be an attempt to encourage traffic now moving from
Kisii to Kisumu entirely by land transport to return to the cheaper
route using E.A.R. & H. lake services from Homa Bay to Kisumu.
Even here, however, the method is useful, for it does indicate the

relative level of support which can be expected from the agricul-

 

31Kenya, Development Plan 1966-1970, pp. cit., p. 283.

307

tural producers among whom the road passes.

Summary

In sum, the method by which an objective decision may be
reached concerning road improvements in agricultural areas is
based upon the costs of improving the efficiency of movement and
the potential returns from the land which might be reaped if
better transport existed. Annual road costs, variable from place
to place, may be calculated for each road. The spatial dif-
ferences of such costs will follow from a great many environ-
mental factors as well as the variable character of the roads'
surfaces at the present time. An application of the method was
carried through several stages of priority determination for the
roads within the study region. To proceed much further would be
an injustice to the assumptions made necessary by sparse data.
Nonetheless, it is clear that the method is workable and could
be of great assistance in reducing the waste of sparse funds

that follows from less-than-optimum investments.

PART IV

CONCLUSIONS

CHAPTER X

SUMMARY AND CONCLUSIONS

The Context and the Orientation

 

The method outlined in this volume offers an answer to
the question: How can the decision to invest in one road project
rather than another be made more rational, more realistic, and
with greater empirical justification? Granted that subjective in-
vestment decisions may result in less than optimal allocation of
funds, in what manner can the knowledge required for a decision
based on the relevant characteristics of the region be brought
together and translated into an outline for choosing the best among
many alternatives?” Theuproblémqisgnobtwhetherito invest or.not'
to invest. The concern here is discovering which of several in-
vestment possibilities promises the greatest returns. Decisions
concerning road investments, especially investments dealing with
an improvement in road quality, vis-a-vis the construction of a
new road, are usually founded on essentially subjective factors.
What is needed, it is believed, is an objective method for de-
termining the priorities of road investment projects in agricul-
tural areas.

This problem has been treated within the context offered

309

310
by an "underdeveloped" country. Even aside from the possible
sensitivities bruised by the term "underdeveloped," it is important
that the author's concept of this word be made explicit. Under-
development, as used throughout this volume, does not relate in
any way to the intelligence, capabilities, or character of the
people living in the countries commonly so designated. Nor does
the term refer to the intensity with which the land is used, in
spite of the recently fashionable view that Canada, Australia, and
the United States are truly underdeveloped. The concept is here
used to refer to the relative position of a country along a dif-
ficult-to-define continuum of economic nature, the position being
determined primarily by the attitudes of the dominant portion of
the population toward participation in a commercial economy and

the acceptance of the changed mode of economic behavior required

 

by such participation. This, of course, is reflected in such
trends as rising per capita income, a rising proportion of non-
agricultural employment, increasing urbanization, and often a re-
orientation of the political structure. It is important to the
development process, then, that all possible encouragement be
given to the population to break with subsistence activities and
contribute to a growing national market economy.

It is clear that agriculture and transportation are among

the most important facets of a country's development efforts. The

311
majority of the population in each newly independent state in
Africa are engaged in agriculture. With several notable ex-
ceptions, these countries will have to depend in large part upon
revenue gained from the agricultural sector. Agricultural products,
however, are an unstable source of foreign exchange because of the
wide fluctuations in world supplies, world demand, and world market
prices. This instability, plus the factor of an apparent long-term
decline in the competitive position of most such products, has led
development planners to place at least a balanced emphasis on in—
dustrial, as well as agricultural, development. Nonetheless, agri-
cultural development will remain crucial to national development
within the foreseeable future.

Because of the source of most internally generated develop-
ment funds and the fluctuations in the availability of these funds,
there is an added urgency to use such funds in the most efficient
manner possible. It would be important that development funds not
be wasted in any case, because of the tremendous need for project
financing throughout the national economy of a low income country.
The choice between one investment project and another, therefore,
should be made on the basis of a thorough investigation of the
alternatives so that the project offering the best expectation of
returns may be clearly seen.

Increased transport efficiency obviously contributes to

a nation's development. An operational economic specialization in

312

a regional framework, in addition to the customary occupational
sense, is allowed and encouraged only with low-cost movement
facilities. Agricultural production appears to increase sharply
in areas in which the means of transportation are improved. A
more controlled marketing procedure is possible for the small
farm producer when transportation of goods is frequent, rapid,
and proximate to his farm. Transportation is also important in
encouraging the broader outlook that is necessary for a revolution
in farming (and marketing) methods, to say nothing of the contri-
bution to national political integration. This last point, in
turn, is vital to the maintenance of stability and unified effort
required in countries which have few development resources.

Investment in transportation is necessary to development.
Investment projects should be undertaken only in light of the
returns to the total economy available from alternative projects.
Decisions concerning the importance of rural roads in under-
developed countries, therefore, should be based upon an empirically
determined list of investment priorities. Commitment to such a
procedure, if it is correctly conceived, will reduce the possibil-
ity of inefficient resource allocation caused by decisions founded
on subjectivity.

An argument was presented in this work that the formula-
tion of a procedure for investment priority determination is a

problem that might best be considered from a geographic perspective.

313
Reality is complex and does not lend itself to compartmentalized
analysis along the lines of academic disciplines. Geography ex-
amines reality from a Spatial perspective and is distinguished
from other disciplines by this approach rather than any of the
specific types of subject matter studied. That few geographers
have concerned themselves with economic development in this way
vis-a-vis the "case study" approach and the building of world
scale descriptive models does not invalidate the applicability of
a spatial perspective to detailed, practical problems in develop-

ment.

The Existipg Pattern pf Agpiculture

 

 

The procedure by which the priorities can be determined
for investment in agricultural roads has been discussed in detail
during its application to a region in an underdeveloped country.
The method described here is, in the abstract, to determine the
improvement which can reasonably be made in the profitability of
agricultural production and then compare the costs and advantages
of improving one road rather than another with the agricultural
potential of both. This is done in order to discover which roads
will assist the greatest agricultural change for the least in-
vestment in facilities. Crucial to the method is treatment of
these costs, returns, profitability, and so forth, as patterns

containing ppatially variable functions pf each economic element.

 

314

The pattern of existing crop profitability in the study
region, Nyanza Region and Kericho District in Kenya, is seen as
the difference in patterns of costs per acre of production and
those of value per acre. The costs to the farmer for production
of the cash crops included for analysis is also divided into the
cost of producing the crop on the farm and the cost of moving the
crop to market. Both of these costs are variable from place to
place. Transport costs are smaller for farmers located near the
market centers. Furthermore, these costs will be less for farmers
whose land is close to existing roads of high quality, for vehicle
operating costs are lower on paved roads than on gravel roads or
earth tracks. Although costs of production are also undoubtedly
spatially variable, it was believed that such qualitative factors
as farmer motivation and knowledge were strict deterrents to a
full understanding of this variability. In any case, the non-
human sources of cost variability are included in treatment of
differences in value per acre of production; inclusion of these
factors in both aspects of the method would multiply their true
effect on the resulting patterns. Total costs of production are
taken to be the sum of the uniform cost of production inputs for
each crop plus the cost of moving that crop from the land on
which it is grown to the major market centers.

With adequate data, the pattern of existing profitability

of cash crop production could be easily obtained. Because these

315
data are not available at the present time, it was necessary to
simulate the patterns of farm income. Due to the difficulties
associated with an assignment of quantitative indices to the
variable quality of human inputs, it was argued that place-to-
place variations in the environmental suitability for the pro-
duction of each of the relevant cash crops were the over-riding
factors affecting the pattern of value per acre of these crops.
Using derived indices of environmental suitability and the pat-
terns of each pertinent environmental characteristic, the spatial
variation in crop value per acre was determined. Objections to
the method used for this determination were answered. Because
of its flexibility, both in terms of easy revision and of ap-
plicability to more detailed studies of the same nature, this
method is an attractive substitute for the fully detailed data
which should be gathered prior to an actual application of the
total procedure.

The patterns of existing profitability, determined by
subtracting the patterns of total production cost from the respec-
tive patterns of crop value per acre, showed coffee and tea to be the
most profitable cash crops within the study region. Although there
were areas of highly profitable farming elsewhere, these two crops,
with support from pyrethrum and white potatoes, give Kisii and
Kericho Districts the zones of greatest profitability of produc-

tion.

316

The Potential Pattern pf Agriculture

 

One of the most serious problems faced by those concerned
with development when attempting to formulate a series of invest-
ment priorities is the difficulty of estimating the returns which
can be expected from an investment. Even in the present volume,
where immediate returns to investment are only of indirect concern,
the level of future agricultural profitability is very important.
Only by estimating in some consistent manner the potential return
per acre from cash crop production can the potential improvement
in agricultural profitability be calculated. It was argued here
that a dependable, and more meaningful, approach to this problem
of determining agricultural potential can be made through use of
ecological regions.

The use of environmental characteristics in combination,
as ecological areas, in estimating agricultural production po-
tentialities is not common. William Allan and Colin Trapnell
surveyed a good deal of Northern Rhodesia in order to classify the
land devoted to African agriculture into areas of varying suitabil-
ity for different crops. Allan was later able to utilize the pat-
terns of vegetation-soil associations in determining the variable
pressure of population on the land resources. A detailed survey
of Southern Rhodesia during the 1950's was made explicitly to
provide development planners there with the information required

by them concerning the optimum agricultural use to which the land

317
could be put. In Kenya, L. H. Brown formulated preliminary agri-
cultural policies for the two most productive farming regions in
the country. It is the report by Brown on the ecology and agri-
cultural potential of Nyanza Province that provides the beginnings
for a detailed determination of potential profitability in this
region.

Some modifications of the description of ecological zones
offered by Brown were made on the basis of the patterns of other
environmental elements. Climatic influences in the form of poten-
tial evapotranspiration and water surplus and deficit were also
used to modify this delimitation. The resulting ecological regions
were found to correspond to the patterns of production for those
crops which can be expected to produce well in each region.

Using the extent of each ecological region as a guide,
the portion of the study region capable of supporting each cash
crop was indicated. These areas of potential production were re-
fined using the previously stated environmental growing limits. It
was then possible to translate this pattern into one showing the
potential levels of farming value per acre in the not-too-distant
future for each cash crop. Future value per acre patterns were
based on reasonable estimates of the yields per acre and prices
per unit output, and on the crops' patterns of environmental suita-
bility. The patterns of improvement in profit per acre, i.e., the

difference between existing and potential profitability configura-

318
tions, were then drawn. These patterns illustrate the variable
gain of which each crop is capable given the proper inputs, one
of these inputs being better transportation.

As was pointed out, the pressure of the population upon
the land forces modification of potential production levels.

Under systems of land use designed to give each farm family a
moderate annual income regardless of the ecological region in
which the farm is located, it was apparent that population
pressure was a crucial variable in any estimate of crop produc-
tion per acre. If the population per square mile exceeds the
productive capacity of the land for even moderate subsistence
requirements, then little dependable cash crop surplus can be
expected. The greater the surplus of land per family above that
required for each family's subsistence, the larger is the pro-
portion of the total average farm acreage which may be devoted to
the appropriately lucrative cash crop. A severe pressure of popu—
lation on the agricultural resources was shown to be imminent
throughout a large portion of the study region, even under reasona-
ble estimates of population grOWth.

By combining those portions of each cash crop's potential
improvement which can be expected to yield the greatest profit in-
crease per acre for each separate square mile in the study region,
the optimum pattern of potential improvement in profitability was

determined. This pattern, however, does not account for popula-

319

tion pressure. The map of profit increase per acre, adjusted for
the variable demands placed on the land resources by population,
was drawn as a product of a land use intensity map (reflecting
population and land carrying capacity) and the absolute pattern
of likely profitability improvement (reflecting costs, yields,
and prices of the best crop for each unit of land). This product
illustrates the place-to-place variation in the increased returns
from the land which are possible in the near future given appro-

priate stimulus to the development.

The Pattern pf Road Investment

 

 

The justification for investment in one road rather than
another is based partly upon the costs of upgrading the roads'
quality and the immediate benefits which will be derived by trans-
porters from the better roads. Construction costs are variable
with environmental conditions in the area through which the road
passes, traffic demands, deviations from the least-cost route be-
cause of human settlement patterns, and the like. Annual mainte-
nance costs, both before and after upgrading, are variable with
the quality of road surface, the demands placed on the surface
by traffic, environmental conditions, and the availability of
materials required in road maintenance. Although there are many
benefits, both direct and indirect, which undoubtedly result from

road improvements, the difficulties of obtaining empirical

320
measures of these benefits applicable to the region of concern
here were seen as justification for restricting the direct re-
turns from investment to a measurable element of the transport
factor. This direct return was the reduction in vehicle opera-
ting costs resulting from the change in road surface.

The ratio between the incremental costs of road invest-
ment and the reduction in operating costs following this in-
vestment is the denominator of an index of relative priorities
for road upgrading, the numerator, of course, being the adjusted
pattern of agricultural profitability improvement. For a given
level of agro-potential, the smaller the denominator, because of
a better relationship between returns and costs from investment,
the larger is the priority index, and vice versa.

It was found that because of the existing high traffic
volume on major non-bitumen roads in Nyanza Region and Kericho
District and the decrease in annual maintenance costs when a road
is paved, those roads which would be improved to bitumen quality
with any reasonable investment were clearly the top priority road
projects in the region if even moderately high agricultural im-
provement is possible along their route. Of those existing roads
not this near to bitumen quality, those within the high potential
improvement areas of Kericho District had the next greatest justi-

fication for early investment.

321

The Method's Potential and its Limits

 

There are numerous assumptions made throughout the volume.
These assumptions are almost entirely limited to the application
of the proposed methodology rather than to the method itself.

The assumptions should be tested thoroughly in any application
which is to lead to real investment decisions. This cannot be
emphasized too often. An attempt has been made to consider the
implications of each assumption in as clear a manner as possible
to facilitate their later examination. The procedure itself,
however, stands without bolstering assumptions. Given the re-
sults of the necessary research, this model for investment
priority determination within a spatial framework should be a
valuable addition to development methodology.

The procedure is especially attractive because of its
flexibility and inclusiveness. Although designed explicitly for
road investment problems, numerous other questions are answered
in the process of its application. The locations for agricultural
investment which must accompany any road investment in rural areas
are clearly indicated by the pattern used in the index fraction's
numerator. The locations of extreme population pressure are also
made plain by the land use intensity factor. The implications
from this map of the real ability of the land to carry existing
and projected populations are many. The origins of rural emigra-

tion, and local, small town unemployment, may be implied from

322
this pattern as can the areas able to support much larger popu-
lations, and how much larger. The procedure may also take on a
dynamic character if it can be determined that a given invest-
ment in a certain level of agricultural production will result
in changes in both investment priorities after some time period.
Perhaps the most attractive feature to harried development
planners, in Kenya and elsewhere, should be the fact that once
the data has been gathered, the updating and revision necessary
to continue the model through several time periods do not re-
quire much additional data that is not already gathered under
existing procedures.

This volume relates a preliminary application of a method
for determining road investment priorities in agricultural areas
There are many aspects which could be modified with more special-
ized knowledge and more complete survey data. It should make
clear, however, to those interested in economic development that
the spatial perspective is an important element in many of the
real problems facing underdeveloped countries. It is hoped that
this methodology will be of assistance in the solution of some

of these problems.

APPENDICES

Appendix I The Method of Deriving Indices
of Environmental Suitability
for Crop Production

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

31744444 1111117'22
1,
$44444 1111 222
PTTTAA 11/1/{2222
333333\ $222222
}3\33333 22222223
2\33333 22222223
223333 22222/333
223333 222/33333
precipitation temperature
indices indices
113333 55668899
133333 55688999
222\\333 56677899
222933 77777788
2222444 667777910
2 2pr_—g,/2 4 4 6 6 6 7 7 9 9 10
2f|44444 66889101010
44I44444 688 910101010
soil indices composite environmental

suitability rating

324

Appendix II Farming System for Ecological Zones

from: L. H. Brown, "Draft Agricultural Policy: Nyanza
Province," Appendix I.

Farming System for Zone 1. Star - Kikuyu - Grass Zone

 

(1) 9

(2) 2%

(3) 2%

(4) 1

acre. House and vegetable and banana garden
Surplus eggs poultry etc. 100/-

acres. Arable Land. Production as follows -
LR. Maize 1 ac. Mtama or Wimbi 1 ac. Potatoes % ac.
15 bags 8 - 4 bags 20 bags

SR. Beans or other Legume 1% ac. Grain 1 ac.
6 bags 4 bags

Family requirements:- 12 bags Grams
4 bags Legumes
10 bags Potatoes

Surplus 9 bags Maize @ 30/- 270/-
2—6 bags Grain 180/-
2 bags Legumes @ 40/— = 80/-
6 bags Potatoes @ lO/- 60/-
Shs.590/- 590/-

acres. Grass (including some fodder grass)
carrying 3-3k stock units
= 2 Adult cows & followers 2 - 2% years

2 Adult cows produce 200 gallons surplus/annun

= 400 gallons @ 1/50 600/-
Surplus stock (calculated on 5 yr.
basis with 20% loss) 250/-

Shs.850/- 850/-

acre. Cash crop. Either % ac.Coffee & % ac.

mulch for same. 8 cwt.Coffee @ £300 per ton 1,200/-
pp g ac. Pyrethrum: if rotated with

an extra % ac.fodder grass this will
enable another cow to be kept
Value of Pyrethrum 350/-)

Cattle Products 425/—) 775/-

pp 1 ac.Tea producing 3,000 lb.

green leaf per acre. 1,000/-

325

326

.pp 1 ac.Sugar Cane yielding @ 130
tons cane/cycle or 26 tons cane/year

@ 30/-/ton of cane, 780/-
This also enables another cow to be
kept on tops 425/- 1,205/-

(5) % acre Trees
Total 7 acres. producing an income of Shsp2Jl95 - Shs.2,745/-

Note: (1) Sugar cane yield above is based on recent experi-
mental results, Coffee on actual practice. Milk
surplus is based on a cow producing 500 gallons/lacta-
tion of which 100 is required for the calf & the
family.

(2) Wimbi has been included as a grain crop for beer
making etc; its superior feeding value in other re-
spects would not be necessary if the family got ade-
quate eggs, milk etc.

Farming System for Zone II. High Rainfall Savannah.

(l) 1% Ac. House. Vegetables, Bananas & Cassava
Surplus products eggs etc. Cassava

in some years 300/-

(2) 3% Ac. Arable, Producing as follows -
LR. 2 AC. Maize 1% Ac.Wimbi & Sorghum

20 bags 10 bags
SR. 2% Ac. Legumes 1 Ac. Sweet potatoes
10 bags 10 bags

(TOTAL 30 bags Grain: 10 bags Legumes and
10 bags roots)

Family Requirements 12 bags Grain
4 bags Legumes
6 bags roots

Surplus 18 bags Grain @ 30/- 540/-
6 bags Legumes @ 40/- 240/-
4 bags Roots @ lO/- 40(-

820/-

327

(3) 4% Ac.Grass of which 1 ac. will be fodder grass.
Will keep at least 6 stock units i.e. 4 Adult
animals & followers. Of the adults 2 will be
cows and 2 oxen.

2 Cows producing @ 150 gallons surplus/annum
300 gallons @ 1/50 per gallon 450/-
Surplus stock (on 5 year basis with

25% loss) 200/- 650/-

(4) l Ac.Cash Crop Either Sugar Cane @ 3/4 Zone I yield 800/-
pp 1 Ac. of cotton 400/-

1 Ac. Rice in a swamp 400/- 800/-

% acre extra to allow for more grass in
rotation with cash crops

(5) l Ac.Trees. (Planted on boundaries as shelter belts)
Value of surplus timber, token figure 100/-

Total 12 acres. Producing a Cash Income of Shs.2,470 - 2:870l-

Note: (1) In this zone it is envisaged that Cash Crops may have
to be grown outside the holding in areas of swamp etc.
which are particularly suitable to them.

(2) Practically no account is taken of the value of the
large area of bananas possible in the homestead plot;
this would almost certainly enable a greater propor-
tion of surplus grain etc. to be sold.

Farming System for Zone III. Less favourable dry land areas.

 

(l) 1 Ac. House and vegetable garden.

Surplus poultry etc. 200/-
(2) 4 Ac. Arable Land producing as follows -
LR. 3 Ac. Sorghum 1 Ac. cotton
18 bags

SR. 2 Ac. Legumes; 1 Ac. Cotton, 1 Ac. Sweet Potatoes
6 bags 600 lbs 10 bags

328

Family requirements 12 bags Grain
4 bags Legumes
8 bags Sweet Potatoes

Surplus -
6 bags Grain @ 20/- 120/-
2 " Legumes @ 40/- 80/-
2 " Sweet Potatoes @ 10/- 20/-

(3) 6 Ac. Grass ley started with arable
1 Ac. Fodder rotated with Cassava
Will keep 6 cattle units (2 oxen & 2 cows)

annum each 300/-
Surplus stock (on 10 year basis
with 40% loss) 300/-

(4) 1 Ac. Cassava rotated with Fodder grass
In a good year produces 400/- extra in
cash; in a bad required for food. If one
year in two is good this gives a mean
income per annum of

(5) Cash Crop (a) Internal Sisal Hedges, % ton
fibre @ £80 per ton

(b) Cotton in Arable rotation

(6) 1 Ac. Trees, for fuel, windbreaks etc.

Total 14 acres producing an income of Shs.

N.B. Many families in this zone will derive additional
income from fishing, which appears to affect
people living up to 5 miles from the Lake Shore.

230/-

600/-

ZOO/-

800/-
300/-

2,330/-

BIBLIOGRAPHY

BIBLIOGRAPHY

Books

Allan, William. The African Husbandman. London: Oliver and
Boyd, 1965.

 

Studies pp African Land Usage 1p Northern Rhodesia.
Rhodes-Livingstone Papers Number 15, 1949.

 

 

Alexander, John W. Economic Geography. Englewood Cliffs, N. J.:
Prentice-Hall, Inc., 1963.

 

Clayton, Eric. Agrarian Development 1p Peasant Economies. New
York: The Macmillan Company, 1964.

 

 

Eden, T. Tea. London: Longmans, Green and Co., 1958.

Eicher, Carl, and Witt, Lawrence (eds.). Agpiculture 1p Economic
Develqpment. New York: McGraw-Hill Book Company, 1964.

 

 

Frank, Charles S. The Spgar Industry 3p East Africa. Nairobi:
The East African Institute of Social Research, 1965.

 

 

Friedmann, John R. P. The Spatial Structure pf Economic Devel-
opment ip the Tennessee Valley, Research Paper No. 39.
Chicago: Department of Geography, University of Chicago,
1955.

 

 

 

 

Fromm, Gary (ed.). Transport Investment and Economic Development.
Washington: The Brookings Institution, 1965.

 

Ghosh, Santikumar. The Financing pf Economic Development.
Calcutta: The World Press Private Ltd., 1962.

 

 

Ginsburg, Norton. Atlas pf Economic Development, Research Paper
No. 68. Chicago: Department of Geography, University of

Chicago, 1961.

 

(ed.). Essays pp Geography and Economic Develppment,'gp-
search Ppper No. 62. Chicago: Department of Geography,
University of Chicago, 1960.

 

330

331

Gould, Peter R. The Development pp the Transportation Pattern V/
.ip Ghana, Studies 1p Geography No. 5. Evanston:
Northwestern University, 1960.

 

Grist, D. H. Rice. 4th ed. London: Longmans, 1965.

Haarer, A. E. Coffee Growing. London: Oxford University Press,
1963.

 

Hance, William A. African Economic Development. New York: ‘¢
Harper and Brothers, 1958.

 

Hartshorne, Richard. Perspective pp the Nature pf Geography.
Chicago: Rand McNally & Company, 1959.

 

 

Hawkins, E. K. Roads and Road Transport 1p pp Underdeveloped \/'
Country: 'p Case Stupy pf Uganda. Colonial Research
Studies No. 32. London: H. M. Stationery Office, 1962.

 

 

Road Transport 1p Nigeria. London: Oxford University x/
Press, 1958.

 

Hay, William W. ‘Ap Introduction.pp Transport Engineering. New
York: John Wiley and Sons, Inc., 1961.

Hirschman, Albert 0. The Strategy pf Economic Development. New v”
Haven: Yale University Press, 1958.

 

International Bank for Reconstruction and Development. The Econom-
ip Development pf Kenya. Baltimore: The Johns Hopkins
Press, 1963.

 

Isard, Walter. Location and Space Econopy. New York: John Wiley
and Sons, Inc., 1956.

Jackson, E. F. (ed.) Economic Develppment pp Africa. Oxford:
Basil Blackwell, 1965.

King, John A., Jr. Economic Development Projects and their 1,”
Appraisal. Baltimore: The Johns Hopkins Press, 1967.

Klein, Martin S. Economic Feasibility pf Improvipg Two Inter-
national Highway Links ip Kenya. Cambridge, Mass.: United
Research Incorporated, 1965.

 

Maas, Arthur, et al. Design pf Water Resource Systems. Cambridge,
Mass.: Harvard University Press, 1962.

332

McKean, Roland N. Efficiency Lp Government through Systems
Analysis. New York: John Wiley and Sons, Inc., 1958.

 

Mellor, John W. The Economics pf Agricultural Development.
Ithaca, N. Y.: Cornell University Press, 1966.

 

 

Morgan, W. T. W., and Shaffer, N. Manfred. Population pf Kenya;
Density and Distribution. Nairobi: Oxford University
Press, 1966.

 

 

Myrdal, Gunnar. Economic Theory and Underdeveloped Regions.
London: Duckworth, 1957.

Nurkse, Ragnar. Problems Lf Ca apital Formation in Underdeveloped
Countries. New York: Oxford University —Press, 1962.

 

Randall, Darrell. Factors pf Economic Develppment and the
Okavapgp Delta, Research Paper No. 47. Chicago: Depart-
ment of Geography, University of Chicago, 1957.

 

 

Robinson, E. A. G. (ed.) Economic Development for Africa South
pf the Sahara. London: Macmillan, 1965.

 

Rostow, W. W. The Stages pf Economic Growth. Cambridge: The
University Press, 1961.

 

 

Russell, E. W. (ed.) The Natural Resources pf_East Africa.
Nairobi: East African Literature Bureau, 1962.

 

 

Ruthenberg, Hans. African Agricultural Development Policy 3p
Kepya 1952-1965. Berlin-Heidelberg, New York: Springer-
Verlag, 1966.

 

//
‘/
Schultz, Theodore W. Transforming Traditional Agriculture. New

Haven: Yale University Press, 1964.

Steiner, Henry Malcolm. Criteria for Planning Rural Roads 12.2
Developing Counppy: Ipp Case pf Mexico. Project on
Engineering-Economic Planning, Report EEP-17. Stanford:
Stanford University Institute in Engineering-Economic
Systems, 1965.

 

/
~/

Wilson, George W., et al. The Impact Lf Highway Investment Ln
Development. Washington: The— Brookings Institution, 1966.

 

 

Winch, David M. The Economics Lf Highway Plannipg. Toronto:
University of Toronto Press, 1963.

 

333

Wood, Alan. The Groundnut Affair. London: The Bodley Head, V/
1950.

 

Articles and Periodicals

 

Baier, W. "The Interrelationship of Meteorological Factors, Soil
Moisture and Plant Growth," International Journal pf
Biometeorology, IX (1965), 5-20.

 

Bauer, P. T., and Yamey, B. S. ”Economic Progress and Occupational
Distribution," The Economic Journal, LXI (December, 1951),
741-55.

 

Berry, Brian J. L. "City Size Distributions and Economic Develop- _/
ment," Economic Development and Cultural Chapge, IX
(July, 1961), 573-88.

Brunner, C. T. "Developing Countries: Roads Economic Justifica- u/f
tion," British Road Federation, Bulletin, No. 310
(December, 1962), 235-42.

Garnier, B. J. ”Geography and National Development," The Nigerian
Geographical Journal, I (April, 1957), 3-6.

 

Ginsburg, Norton. "Natural Resources and Economic Development,"
Annals pf the Association pf American Geographers, XLVII
(September, 1957), 197-212.

 

 

Hastie, M. S. (abstract of research) Empire Cotton Growipg Review,
XLII (1965), 318.

Hill, Polly. "A Plea for Indigenous Economics: the West African
Example," Economic Development and Cultural Change, XV
(October, 1966), 10-20.

Hunter, John M. "Cocoa Migration and Patterns of Land Ownership \/
in the Densu Valley near Suhum, Ghana," Transactions and
Papers, The Institute pf British Geographers, No. 33
(December, 1963), 61-87.

 

 

"Population Pressure in a Part of the West African
Savanna: A Study of Nangodi, Northeast Ghana," Annals pf
the Association pf American Geographers, LVII (March, 1967),
101-14.

 

334

 

Isard, Walter. "Distance Inputs and the Space Economy," The
QuarterLy Journal pf Economics, LXV (1951), 181-98 and
373-99.

Johnston, Bruce F., and Mellor, John W. "The Nature of Agricul-

ture's Contributions to Economic Development," Food
Research Institute Studies, Stanford University, I
(November, 1960), 335-56.

 

, and Mellor, John W. "The Role of Agriculture in Econom-
ic Development,” American Economic Review, LI (September,
1961), 566-93.

 

Kuznets, Simon. "Economic Growth and the Contribution of Agri-
culture: Notes on Measurements," International Journal
.pf Agrarian Affairs, III (April, 1961), 59-75.

 

 

Lewis, W. Arthur. "Economic Development with Unlimited Supplies
of Labour," Manchester School pf Economic and Social
Studies (May, 1954), pp. 139-92.

 

 

Moss, R. P. "Soils, Slopes and Land Use in a Part of South-
Western Nigeria: Some Implications for the Planning of
Agricultural Development in Inter-Tropical Africa,"
Transactions and Papers, The Institute p: British
Geographers, No. 32 (June, 1963), 143-68.

 

 

 

 

Penman, H. L. "Natural Evaporation from Open Water, Bare Soil and
Grass," Proceedings pf the Royal Society, CXCIII (1948),
120-45.

Prest, A. R., and Turvey, R. "Cost-Benefit Analysis: A Survey,"

The Economic Journal, LXXV (December, 1965), 683-735.

 

"Program Presented at the 49th Annual Meeting by the AAG Committee
on Economic Development," Professional Gepgrapher, V
(July, 1953), 16-18.

Nurkse, Ragnar. "Trade Fluctuations and Buffer Policies of Low-
Income Countries," Kyklos, XI (1958), 141-54.

Reiner, Thomas A. "Sub-national and National Planning: Decision
Criteria," Papers, Regional Science Association, XIV
(1965), 107-36.

"Revenue and Cost of Production of African Grown Crops in
Nyasaland," Nyasaland Farmer and Forester, VI (1962), 8,
abstract in Eppire Cotton Growing Review, XL (1963), 145-46.

 

 

335

Smith, R., and Brown, R. A. "Improved Methods of Cotton Growing V//
in Eastern Region, Tanganyika," The Empire CottOn Growipg
Review, XL (1963), 268-77.

Thornthwaite, C. W. "An Approach toward a Rational Classification
of Climate," Geographical Review, XXXVIII (1948), 85-94.

, and Mather, J. R. "The Water Balance," Publications 3p
Climatology, VIII (1955).

 

 

, Mather, J. R., and Carter, D. B. "Three Water Balance
Maps of Southwest Asia," Publications 1p_Climatology,
XI (1958).

Vinge, Clarence L. "The Language of Geography," Professional
Geographer, V (May, 1953), 8-9.

Vining, Rutledge. "Delimitations of Economic Areas: Statistical
Conceptions of the Spatial Structure of an Economic
System," Journal pf the American Statistical Association,
XLVIII (March, 1953), 44-64.

Witt, Lawrence W. "Role of Agriculture in Economic Development -- \//
A Review," Journal pf Farm Economics, XLVII (February,
1965), 120-31.

Wolpert, Julian. "The Decision Process in Spatial Context,"
Annals pf the Association pf American Geographers, LIV
(December, 1964), 537-58.

Public Documents

Federation of Rhodesia and Nyasaland. Ap Agricultural Survey_pf
Southern Rhodesia. Parts I and II. 1959.

 

East African Common Services Organization. East African
Meteorological Department. Collected Climatological
Statistics for East African Stations. Part 1. Kenya ppp
Seyphelles. June, 1964.

East African Common Services Organization. East African
Meteorological Department. Summary pf Rainfall ileenya
for the Year 1965. 1966.

336

East African Common Services Organization. East African Meteoro-
logical Department. Temperature Data for Stations pp East
Africa. Part I. Kenya. 1959.

 

Kenya. Atlas 9; Kenya. 1959.

Kenya. Development Plan, 1966-1970. 1966.

 

Kenya. Economic Survey, 1966. 1966.

 

Kenya. Ministry of Agriculture, Animal Husbandry, and Water Re-
sources. African Land Development pp Kepya, 1946-1955.
1956.

 

 

Kenya. Ministry of Economic Planning and Development. Farm
Economics Survey Unit. Some Economic Aspects pf Agricul-
tural Development ipyNyeri District, 1963. Report No. 24.
August, 1966.

 

 

 

Kenya. Ministry of Finance and Economic Planning. Economics and
Statistics Division. Farm Economics Survey Unit. Some
Economic Appects pf Agricultural Develppment ip_Nyeri
District, 1962. Report No. 21. August, 1964.

 

 

 

Kenya. Ministry of Finance and Economic Planning. Economics
and Statistics Division. Farm Economics Survey Unit.
The System 9; Recording and Analysis used pp Production
Cost Studies pf Large Scale Farmipg 1p Western Kepya.
Report No. 17. December, 1963.

 

 

 

 

 

 

Kenya. Ministry of Works. Loan Application for Trunk Road Devel-
ppment 1967-1970. August, 1966.

 

Kenya. Office of the Minister of State for Constitutional Affairs
and Economic Planning. Economics and Statistics Division.
Kepyp African Agricultural Sample Census, 1960/61. Parts
I and II. 1962.

 

Organization of American States. General Secretariat. Department
of Economic Affairs of the Pan American Union. Farm pp \J/
Market Roads 2p Latin America; their administration and

financing. 1964.

 

United Nations. Tranpport Problems 3p Relation.pp Ecpnomic Devel-
opment Lp West Africa. E/CN.l4/63. 1962. \:

 

 

 

337

United Nations Conference on Trade and Development. Stabili-
zation pf International Commodity Markets. E/CONF.46/8.

1964.

United Nations Economic and Social Council. Economic Commission
for Africa. East African Transport Problems in Relation
to the Promotion Lf Economic Development (Progress Report).
E/CN. 14/148. January 24,1962.

 

Reports

Brown, L. H. A National Cash Cropp Poli_y for Kenya. Parts I
and II. Report for Government of Kenya, Ministry of Agri-
culture and Animal Husbandry. Nairobi: The Government
Printer, 1963.

Central Nyanza District. Annual Reports pf the District Agricul-
tural Officer. 1960-1965.

 

Deutsches Institut fur Entwicklungspolitik. Vorgutachten zur
Raumplanupg 1p der Region Kisumu (Kenia). Berlin, 1967.

 

Edwards and Burrow. Preliminapy Report pp the Permanent Secretapy,
Ministry pf Works Communications and Power, Government pf
Kenya, pp Tea Roads. Nairobi, l964(?).

Gibb, Sir Alexander, and Partners (East Africa). Report pp Agr’-
cultural Roads Miwani : Chemelil : Muhoroni Areas. Report

for the Government of Kenya. Nairobi, 1965.

 

Halcrow, Sir William, and Partners. Report pp the Feasibility pf
Constructing Roads pp Serve the Buyagp Cotton Zone and /
Preliminary Route Location for a Road from Busisi via \/
Buyagu_ to Kharumwa and for a Link to Kasamwa. Report for
the Ministry of Communications and Works, Government of
Tanzania. August, 1966.

Kenya. Report pp the Committee pp Carry Qpp pp Economic Survey pf
South Nypnza and Kericho District - with 3 view pp advising
whether the economic potential would jpstify Rail Develop-
ment. Nairobi: The Government Printer, 1957.

Kenya Cotton Lint and Seed Marketing Board. Tenth Annual Report
and Accounts for the Year ended 3lst October, 1964.

338

Kenya Tea Development Authority. The Operations and Development
Plans pi the Kenya Tea Development Authority. Nairobi:
K. T. D. A., 3lst December, 1964.

 

Kericho District. Annual Reports pf the Distric£_Agricultural
Officer. 1960-1965.

 

 

Kisii District. Annual Reports pp the District Agpicultural
Officer. 1962-1965.

 

 

South Nyanza District. Annual Repprts pp the District Agricul-
tural Officer. 1960-1965.

 

 

 

Swynnerton, R. J. M. Plan for the Intensification pf African
Agriculture Lp_Kenya. Nairobi: The Government of Kenya,
1954.

 

 

Unpublished Material

 

Bonney, R. S. P., Daniel, M. N., and Noorani, A. T. "Traffic and
Road Characteristics in South Western Tanzania." Research
Note No. 5. British Road Research Laboratory, East
African Transport Planning Research Unit, Tanzania Branch,
May, 1966.

Bonney, R. S. P., and Stevens, N. F. "Report on the Investigation
into Vehicle Operating Costs in East and Central Africa.
Part III. Kenya and Uganda." Laboratory Note No. LN/606/
RSPB.NSF. British Road Research Laboratory, June, 1964.

Bonney, R. S. P., and Stevens, N. F. "Report on the Investigation
into Vehicle Operating Costs in East and Central Africa.
Part I. Zambia." Laboratory Note No. LN/344/RSPB.NFS
(revised). British Road Research Laboratory, February,
1965.

Brown, L. H. "Draft Agricultural Policy : Nyanza Province."
Government of Kenya, Ministry of Agriculture, 1957(?).
(Mimeographed.)

East African Railways and Harbours. "Temporary Distance Tables."
(being sections 371 to 377 of Tariff Book No. 3) Nairobi,
1965.

"Nyanza Province." A report for the 1966-1970 Development Plan.
(Mimeographed.)

339

"Preliminary Note on the Origin and Destination Survey of Road
Traffic in Kenya 1964/65." British Road Research
Laboratory, East African Transport Planning Research Unit
(Kenya), KN/4. (Mimeographed.)

"Report on the Bukumbi Survey, Completed February 1962." Western
Research Centre, Ministry of Agriculture, Tanganyika.
(Mimeographed.)

Roth, G. J. "The Economic Benefits to be Obtained by Road
Improvements with special reference to Vehicle Operating \/‘
Costs." Research Note No. RN/3426/GJR. British Road
Research Laboratory, March, 1959.

/

Smith, N. D. S. "A Pilot Study in Uganda of the Effects upon
Economic Development of the Construction of Feeder Roads."(/'
Research Note No. RN/3408. British Road Research Labora-
tory, February, 1959.

"A Study of Vehicle Journey Times on Roads in Kenya." British

Road Research Laboratory, East African Transport Planning
Research Unit (Kenya), KN/3. (Mimeographed.)

Other Sources

 

East African Railways and Harbours. Memorandum, from Chief Opera-
ting Superintendent to the General Manager. February 22,
1960.

East African Railways and Harbours. Memorandum CD/97. March 2,
1962.

East African Railways and Harbours. Monthly records of freight
carried by station from, station to, and commodity class.
1961-1965.

East African Railways and Harbours. Office Communication from
Chief Operating Superintendent to the General Manager.
October 19, 1965.

Kenya Cotton Lint and Seed Marketing Board. Personal interview
with Mr. J. Adero, Nairobi. November 25, 1966.

Kenya Tea Development Authority. Personal interview with Mr.
R. Moore, Acting Senior Tea Officer, West of Rift,
Kericho, Kenya. November 12, 1966.

340

Kericho District. Personal interview with Mr. J. Birir, District
Agricultural Officer, Kericho District. November 12, 1966.

Kisii District. Personal interview with Mr. Ogoda, Cash Crops
Officer, Kisii District. November 14, 1966.
7
Michigan State University. Personal communication from Professor '
John M. Hunter, Department of Geography, Michigan State
University. July 25, 1967.

Pyrethrum Marketing Board. Personal interview with Mr. R. J.
Munro, Nakuru, Kenya. November 11, 1966.