I
’.

luv :1,-_'-n

«7:4

II

 

1. Wu: Am

~me mm

 

1:93 15*? “uh."

I’

Lin—AL

3

“AA‘ \ 9: A1

' "‘ " 5‘: :74“;Td:1‘.lu“iayl

v

 

 

 

  
    
 

      

,... ).,, A
I\'V ,‘tmwnup

   

A A
.u.
unfit-u
Mun: «A
n
1

      
    

A I H."
‘n‘m,
4.34:. '.._ ,A“
...
{R \fii.’ “RA"V

    
  

      
    

     

“I" A».

      

“WIT. |..l\A—h‘ '4: u
l

     
   
 
 
 
 
 
 

  
   
  

       
  

\
3m
"1‘1" _ A
1-2:.u::v‘

:1

A
Av
40‘

sA ”an;

an
1:45

  

   

‘vh":vl:: \- m 1t

I“, wavy; 3"

       
      
 

 
 

”A.“
Ally-7“ AA.A.-,‘ A
*I m‘ I"
m -.A

\

~'A-kv.'.~:‘.<‘n

b A. ‘1 u.
\

   
  
  
 
  

    
  
  

 

1..»ft.»

an
AA. ..A
«A Jiffy} :x‘
1457 -

   

A
;U
'~AI‘|AA
pm...” A
4~ 7-1.: _
AA. 1-
'u-V 5:3”?
sq 1m. .,

 

    
  
  
  
   

  
  
 

vb
‘v-& A: ,1.‘ A"
drunmu

   
 
 

A 'l A
’v MK".

..A"A‘. LAAAAA ‘ .A A
o «m
..A , .

    
  

(AA

 

 
 
   
  
 
  

:5; A
.JIAI..AAAA.
AA. A
I

  
 
 

\5 "

A. if»; Lu

44-..“, T“;
5

    
 
 
 

 
 

“w h" A‘

.AA...

‘ . A :,.:L Am“; :23 £2:
w-AA l‘l‘

»A A ,:

      

,
QAL‘X'A‘A-A. .7

IRAA."

   
 

 
 
 
   
 
 
 
 
  
  
 
 
 

 
  

I
.«l AAV

um .4 up.

'u «u-

  
  
 

     
 
   
  

     

1,1,5

4
.A...

 

.Aw.‘

   

..
if”; iv £4. AM“

 

     

,

    

 
 

 

  
   
  

I

 

A

~ A.» . A

A l;.! 25....‘13‘
I

    
 

a.“

   

—.

   

In

 

.u-u~‘n~.

 

   
 

_ r
.A,A.,,,.AAPA1. A}.

        
 

 

 

   
 
 

   
 

 

‘ . V V

A . A . A '

‘ ., . , , .‘ .‘ t

A A A . A .

' . . . J

‘ V' ‘ . ~ A I . 5

, . . ‘ F

A .A . . r .

‘ ‘ I r‘ .l' . A .
A '. - | A J . .- w ‘

' A . g

  

WWWWWWW 9WE’WOCO

3 1293 00572 5530

 

 

 

 

 

 

 

 

 

 

 

 

LIBRARY
Michigan Stat.
University

 

 

 

This is to certify that the

thesis entitled
The Rate of return to Maize
Research in Kenya: 1955-88

presented by

Daniel David Karanja

has been accepted towards fulfillment
of the requirements for

 

 

flaster's Ofdegree in Agricultural
Science Economics

Cad—K ELL/W

Major professor

 

Date 9 October 1990

 

0-7639 MS U is an Affirmative Action/Equal Opportunity Institution

PLACE iN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or hetero one due.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU Io An Affirmdlve ActioNEquel Opportunity Institution

THE RATE OF RETURN TO MAIZE RESEARCH IN KENYA: 1955-88

BY

DANIEL DAVID KARANJA

A THESIS

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

MASTER OF SCIENCE
Department of Agricultural Economics

1990

ABSTRACT
THE RATE OF RETURN TO MAIZE RESEARCH IN KENYA: 1955-88

By

Daniel David Karanj a

Kenya's agricultural sector is facing several contemporary
challenges including the need to feed a rapidly growing population.
Maize is the staple food for over ninety percent of the population. In
1955, the Kenyan government initiated a hybrid maize research program
which resulted in the release of high-yielding maize varieties. The
hybrids contributed to a doubling of the national maize yield, near
tripling of the area under maize and a fivefold increase in national
maize output over the 1955-88 period.

Today, there is a growing interest in the assessment of
productivity of agricultural research and development of guidelines on
how much national governments and donors should invest in research in
Africa. This study pioneers the evaluation of returns to agricultural
research in Kenya and uses a production function approach to evaluate
the rate of return to investments in maize research from 1955 to 1988.
The results indicate that past maize research, extension. and seed
development programs contributed to increased maize production. The
average rate of return to investments in maize research over the 1955-88
period was found to be sixty-eight percent. In policy terms , the
results show that one Kenyan pound invested in maize research
contributed.sixty-eight pounds to the Kenyan society over the 1955-88

period.

To my mother Janet Wanjiru
and my brothers John Kamau and Martin Maina.

Their love and patience gave me strength.

ii

I wish to thank RJR-NABISCO of U.S.A. for funding my graduate
study at Michigan State University (U.S.A.) through the Kenya Canners
Scholarship Program, the International Maize and Wheat Program (CIMMYT)
for funding my field research in Kenya, and my graduate committee for
guiding me through the study.

My special gratitude go to my major professor, Prof. Carl K.
Eicher, for his inspiration and invaluable support; Dr. James Oehmke for
seeing me through the analysis; and Dr. Carl Liedholm for his comments
on an earlier draft of the thesis. I am indebted to Dr. Derek Byerlee of
CIMMYT (Mexico) for his support and comments on the study, Dr. Ruben
Echeverria of ISNAR (The Hague, Netherlands) for his suggestions and Dr.
P. Ananda of CIMMYT (Nairobi) for his valuable assistance during the
field research.

I also thank Dr. C. Ndiritu (Director, Kenya Agricultural
Research Institute (KARI)), Dr. J. K. Wanjama (Director, KARI-NPBRC,
Njoro), KARI Centre Directors of NARC-Kitele, NDFRC-Katumani, RRC-Embu
and RRC-Mtwapa, and all KARI researchers and personnel for their
participation and support in the .field research. Last but not least, I
thank Sherry Rich (Michigan State University) for her typing assistance
and the support staff at CIMMYT (Nairobi) for their efficient service

and warm hospitality .

iii

Page
List of Tables .................................................... (vi)
List of Figures ................................................... (vii)
List of Abbreviations ............................................. (viii)
1. CHAPTER ONE: INTRODUCTION ANDBACKGROUND .................... 1
1.1. Introduction ................................................ l
1.2. Location, Climate and Demography ............................ 3
1.2. Agriculture in the National Economy ......................... 7
2. CHAPTER TWO: OVERVIEW OF THE MAIZE SUB-SECTOR .............. 11
2.1. Origin and Introduction of Maize ........................... 11
2.2. Maize Research ............................................. 13
2.2.1. Maize Breeding ............................................. 13
2.2.2. Maize Agronomy ............................................. 18
2.3. Maize Seed Production and Distribution ..................... 22
2.3.1. Seed Development ........................................... 22
2.3.2. Seed Distribution .......................................... 23
2.4. Maize Production and Consumption ........................... 27
2.5. Maize Marketing ........... . ................................ 34
2.6 Financing Maize Research ................................... 37

iv

TABLE OF CONTENTngContinued)

Page
CHAPTER THREE: LITERATURE REVIEW ON RATE OF RETURN ......... 42
Technical Innovation and Resource Productivity ............. 42
Measurement of Research Productivity ....................... 44
Rate of Return Studies in Africa ........................... 52
CHAPTER POUR: RATE OF RETURN TO MAIZE IN KENYA ............. 55
The Sample ................................................. 56
Results and Discussion ..................................... 59
CHAPTER FIVE: SUMMARY ..................................... 67
APPENDICES ................................................. 73
APPENDIX A ................................................. 73
APPENDIX B ................................................. 82
BIBLIOGRAPHY .............................................. 100

Table

Table

Table

Table

Table

Table

Table

Table

Table

Table

Table

9:

10:

11:

Page
: Kenya: Agra-Ecological Zones and Land Potential ............. 5
: Kenya: Improved/Hybrid Maize Varieties Developed and
Released .................................................... 17
: Effects of Husbandry and Input Use on Maize Yields .......... l9
: Kenya: Area Under Improved Maize Seed, 1967-75 .............. 26

: Kenya: National Maize Production, Yield and Area, l955-88...28

: Western Kenya: Percentage Adoption of Maize Technology by

Farm Size, 1973 ............................................. 33

: western Kenya: Responses of Non-Hybrid Adapters, 1973 ....... 33

: Kenya: Distribution of Research Expenditure by Private

Sector, 1985-86 ............................................. 38
Kenya: NCPB Maize Purchase by Province, 1966-88 ............. 80
A summary of Returns from Agricultural Research and

Extension from 1958-90 ..................................... 82
Kenya: Calculations of Rate of Return to Maize Research,

1955-88 .................................................... 98

vi

Figure
Figure
Figure
Figure
Figure
Figure
Figure
Figure
Figure
Figure

Figure

Page
1: Administrative Map of Kenya ................................ 4
2: Maize "Diamond" in Kenya ................................... 21
3: Kenya: Hybrid Seed Sales ................................... 25
4(a): Kenya: Maize Area ....................................... 73
4(b): Kenya: Maize Yield ...................................... 74
4(c): Kenya: Maize Production ................................. 7S
5: Kenya: Hybrid Maize Adoption ............................... 76
6: Kenya: MoA Expenditure ..................................... 77
7: Kenya: Maize Expenditure ................................... 78
8: Kenya: Kenya Maize Price ................................... 79
9: Kenya: KARI: Organizational Structure ...................... 81

vii

CBS =

CGIAR

CIMMYT=

CPI =

FAO

GDP

IARC

IITA

ISNAR

KARI

KGGCU

KSC =
Ki =
HALO 8

MOA

MRST =

NARC 8

NARS 8

NCPB 8

NDFRC

NPBRC

NSQCS 8

LIST OF ABBREVIATIONS

 

Central Bureau of Statistics

Consultative Group of International Agricultural Research
International Maize and Wheat Improvement Centre

Consumer Price Index

Food and Agriculture Organization of the United Nations
Gross Domestic Product

International Agricultural Research Centre

International Institute of Tropical Agriculture
International Services for National Agricultural Research
Kenya Agricultural Research Institute

Kenya Green Growers Cooperative Union

Kenya Seed Company

Kenyan Pound=Kenyan Shillings 20 = approx. USS 0.87 (Jun,1990).
Ministry of Agriculture and Livestock Development
Ministry of Agriculture

Ministry of Research, Science and Technology
National Agricultural Research Centre

National Agricultural Research System

National Cereals and Produce Board

National Dryland Farming Research Centre

National Plant Breeding Research Centre

National Seed Quality Control Station

viii

I DO

LLJntmdsstien

Kenya's agricultural sector faces the challenge of feeding a
rapidly growing population on about 17 percent of the country's total
land area. Because almost all the arable land is under cultivation,
future increases in food production will have to rely primarily on yield
improvement rather than on area expansion. Maize is the staple food for
over 90 percent of the population and a cheap source of calories. To
meet future food demand, projections indicate that maize supplies will
have to double in the next fifteen to twenty years.

The development of maize hybrids since the early 19605 has led
to a doubling of maize yields, a near tripling of the area under maize
and a fivefold increase in maize production in what can be considered as
the Green Revolution success story of Kenyan agricultural research. But
the challenge of increasing food production requires continuous
investment in the generation, transfer and adoption of productive
agricultural technologies. Because such investments are costly and
compete for scarce public resources, it is neccessary to ensure that the
resources are allocated to priority research programs.

Several techniques have been developed to evaluate agricultural
research productivity. Ex;pg§; approaches have been used to evaluate the
payoff to past investments in research while ex-ante approaches have

been used to estimate future returns to investments in research. The

2
results of such analyses can be used to determine the productivity of
alternative research investments and assist in research prioritization
and resource allocation. Many such evaluations in Asia, Latin America
and in developed countries have revealed a high rate of return (ROR) to
investment in agricultural research. But only four of about 170 studies
world wide have been published for Africa.

This study evaluates the payoff to investment in improved/hybrid
maize research in Kenya from 1955-88. The study is based on data and
information collected from a field research conducted in Kenya from
October 1989 to March 1990 in order to compile the costs and benefits of
maize research from 1955 to 1988. The data and information were
assembled by extensive use of archival sources, personal interviews and
secondary data sources. The study uses a production function analysis
technique to estimate the ROR to maize research over the l955-88 period.

The outline of the thesis is as follows: Chapter one provides a
background to the country and the economy, and highlights the research
problem and objectives of the study. Chapter two presents an overview of
the maize sub-sector including maize research, production, marketing and
consumption. Chapter three reviews the literature on rate of return
assessment by highlighting past studies on agricultural research
productivity and different assessment techniques. Chapter four presents
the results of the rate of return evaluation for investments in maize
research in Kenya over the 1955-88 period. Chapter five summarizes the

study results and draws implications for research policy in Kenya.

Warns:

Kenya, located in East Africa astride the Equator, has a total
land area of 582,646 square kilometres (10112). It is bordered to the
north by Ethiopia and Sudan, Tanzania to the South, Uganda to the West
and the Indian Ocean to the East. The country is divided into seven
provinces, excluding the city of Nairobi. These are further divided into
41 districts which are, in turn, subdivided into divisions, locations
and. sub-locations. Figure 1 shows the administrative boundaries of
Kenya.

Largely influenced 'by its equatorial location, Kenya has a
diverse climate that varies greatly with topography. The result is a
climate ranging from hot to wet tropical climate on the coastal belt to
arid and semi-arid conditions in the north and north-east, and temperate
climate in the highlands. Annual rainfall and temperatures follow strong
seasonal patterns and vary with altitude. There are two distinct
rainfall distribution patterns: a bimodal pattern characteristic of the
Rift, Central highlands and the coastal belt and a unimodal pattern
characteristic of the regions west of the Rift Valley. The amount and
distribution of rainfall combined with soil characteristics are the
major factors that determine the agricultural potential of land. Based
on these factors, Pratt at al (1966) devised a comprehensive
classification of land potential. Six broad Agra-Ecological Zones (AE2)

are distinguished from his work and are presented in Table l.

Figure 1: Administrative Map of Kenya

’oVL iv:

\ TUIIAIA

1

 

 

   
 

.wo

\ .

 

   
   

 

 

 

 

4 0°!

 

u
‘1 r
. :0 h ‘
15-3 \ l
. ‘ an s A I tr \
\ ' ” “
x - . 1
\ 0' ¢
UGANDA 3",,\ ’l {5" u I
toner . CJ " ' ‘ J u I
5 x’ ‘ a " i
.t.‘=§¥:" Mia \ Santiago R .SOMALT
Krona—‘1‘ \rudo—nuurr . 3 . o L 9 z I
:IAMNCO 5-“, \ .
12m cur-035:.“ o : Lo't-nu u \ g. IREPUBLIC
/’ ,‘ a 1’ ‘uwii . a ! ,.- ‘\
0.0," a I :‘C'm .\ p. -O' \ “ O
I s" \i (LAIKIPIA "5 ‘\ ’1
“l‘.‘.‘ \\ ' ‘ z \- _" o
.J A ~.‘ "‘ l o-
mum~ ‘5‘ ‘3‘] a ‘ “'3 - ‘ o . .—
E ,‘iuxunu MYAMMRUA { z I
e- \: ~ 2 5“" Y Vimmviuea’hr 'Ali33 A '
annexe ,’ ~ .‘ Q3: tnou'
o I 8’“ o .A Q: ”i i
‘ o
L ”32'” l ,1, its”: an. "mil .
\" \ 0 "\ 9 \
. . 3" rune ,- v .
5 \ . I A I O 8 [Ir ‘2 «uncut \.\
\. ( ( t\ tutu: nu RIVER
\.\ (i ht ls
. L. #0300 MACMAKOS \
\J 0 i
\ nuance ‘g‘ “’ Ltd“!
0 9
\.\..r \\ Y I \ '5---
o\ \‘ I” “
K E Y \. \u o /
coolant" \. ' l’
8......"0... —e—_O i“'k"'
Provincial ———- g t I N D I A N
mum: —-——---., . ,0.
L. TAITA TAVITA ~""‘.‘ 0 C E A N
.\ o: ‘s 4"
'\‘,o" " ‘ I HOMASA
L 1° too too to o ‘ S ' "
a o L o u r 'r are '\

 

 

 

 

 

 

w

 

 

 

 

 

 

 

 

 

 

 

ource I ratt

 

 

_ J i -
ZONE AND POTENTIAL CHARACTERISTICS AND LAND .AREA 1 OF
USE ('000Ha) TOTAL
AREA
I. TROPICAL-ALPINE Above 3000m. Moorland and 60 0.11
ZONE barren. Land above forest
(Low Potential) line. Use limited to water
catchment and tourism.
II. UPPER HIGHLAND Between 2400-3000m. Mostly 5,087 8.94
ZONE high rainfall. Suitable for
(High Potential) forests, coffee, tea,
pyrethrum, intensive
livestock. Wheat, barley
and maize on lower
altitudes.
III. LOWER HIGHLAND Between l800-2400m. 4,670 8.21
ZONE Moderate rainfall.
(Medium Potential) Suitable for mixed farming,
maize, wheat, barley, oil
crops and livestock. Higher
altitudes good for tea,
coffee and pyrethrum.
IV. UPPER MIDLAND Between l300-l800m. Semi- 5,342 9.39
ZONE arid. Marginal
(Low Potential) agriculture. Subsistence
crop farming. Sunflower,
maize, sisal, livestock and
wild life.
V. LOWER MIDLAND Between 800-1300m. Arid. 30,192 53.06
ZONE Sub-marginal agriculture.
(Low Potential) Sisal, cotton. Moderate
rangeland potential. Live-
stock and wildlife.
VI. LOWLAND ZONE Between 0-800m. Very Arid. 11,550 20.29
(Low Potential) Mostly rangeland limited to
nomadic pastoralism. Higher
altitudes may be suitable
for sorghum, millet.
TOTAL 56,901 100.00
-—

 

 

6

Table 1 indicates that only about 17 percent of Kenya's total
land area is medium and high potential. The rest is a vast arid and
semi-arid land (ASAL) characterized by nomadic pastoralism, submarginal
agriculture, livestock and wildlife (Kenya, 1983a). With prevailing
technology, these dry low-potential lands cannot be used for productive
agriculture. But since the early 1980s the government is focusing
attention on the development of sustainable technologies for small-scale
dryland farming, irrigated agriculture and pastoral subsectors. These
ventures require substantial investments in physical and social
infrastructures.

Kenya's rapid population growth has put considerable pressure on
arable land for food production and settlement. This has also resulted
in an out-migration from the high and medium potential land to the ASAL.
In the last twenty years arable land per capita has declined by over 50
percent (Henin, 1981). Moreover, the high population growth poses a
threat to economic growth and national development. The population has
increased from 5.4 million people in 1948 to an estimated 22.7 million
in 1988; implying a quadrupling of the population in forty years.
Assuming no change in age-specific fertility and the current growth
rate, the population is estimated to reach 35 million by year 2000. The
combination of a high population growth rate, diminishing per capita
arable land and a rising demand for health, education and shelter are
exerting severe pressure on the government budget and the need to
develop improved and sustainable technologies to increase food

production .

Kenya's economic growth since independence has been very good
relative to most countries in Africa. The Gross Domestic Product (GDP)
grew at an average 6.5 percent per year between 1964-71, one of the
highest growth rates in Africa during the sixties. This growth was
fueled by the agricultural sector through a transfer of land from large-
to small-farm use, an increase in smallholder cultivation of high-value
crops such as tea and coffee, and a modest growth in industrialization
based on import-substitution (Kenya, 1983a). Between 1972-81 the growth
rate declined to an average 5.2 percent per year; the rate grew at an
average 3.8 percent per year between 1982-87.

Agriculture is the largest sector of the Kenya's economy
providing nearly all the country's food requirements, 28-38 percent of
the GDP, 60-70 percent of total export earnings, 75 percent of total
employment and livelihood for about 85 percent of the population (Kenya,
1983a; Ruigu, 1985). The sector remains the main foreign exchange earner
with coffee and tea jointly contributing nearly 70 percent of
agricultural export earnings and about 45 percent of the total export
earnings. Tourism and industrial exports are the other sources of
foreign currency.

Kenya's agriculture faces.two contemporary problems: a shortage
of arable land and the need to provide adequate food for a rapidly
growing population. In the 19603 and early 19708 Kenya produced
sufficient food but since the mid-19703, there has been periodic

imbalances between food production and demand (Kenya, 1988). The rising

8

population and vagaries of weather have resulted in insufficient
domestic food production, thus, undermining national food security.
Agriculture also needs to employ a growing potential laborforce and
generate foreign exchange and raw materials to support expansion and
diversification of the economy.

A drought in 1979 and 1980 resulted in major food shortages and
a US$ 265 million food-import bill (Kenya, 1982). Based on this
experience, the government re-evaluated its food policies and formulated
an agenda to stimulate agricultural production and ensure food security
(Kenya, 1981). The severe drought of 1984 again dramatized the

precarious food situation in Kenya and the need to accord greater

priority to food production. Ihg Sessigng], Pang; No.1 gt 1,986 93
WW outlined the challenges facing

the economy and the urgent need to stimulate agricultural and economic
growth (Kenya, 1986a). Since then the government has adopted policies to
revitalize food production through increased public investment in
agriculture, particularly subsistence crops, and improvement of the
efficiency of production, marketing and distribution of food crops.
Agricultural support services have been reorganized to provide a
favourable environment for increased and sustained agricultural
production. But the options for crop intensification are limited.
Because virtually all of the arable land is under cultivation, Kenya has
to rely on increasing yields rather than area expansion for increased
agricultural production. Over the long run, attempts will have to be
made to convert most of the ASAL region, covering 80 percent of the

total land area, into agriculturally productive land.

9

The development and adoption of improved maize varieties, one of
the great successes of Kenya's agricultural research, increased national
maize output fivefold. The maize yield has more than doubled and the
area planted to maize has nearly tripled since 1955. This success is
important because maize is the staple food for over 90 percent of
Kenya's population and accounts for over 40 percent of the total dietary
intake (Blackie, 1989). The average per capita consumption of maize in
Kenya is about 113 kilograms per year (kg/Yr) compared to 20 kg/Yr for
the whole of the developing world (CIMMYT, 1987; ISNAR, 1985a). In 1981
maize was ranked first in the total area harvested, total production
value and total annual employment of all crops in Kenya (Mwangi, 1980).

Increased maize supplies through increased production and/or
imports are required to meet future food demand. Projections on
population growth and food demand indicate that maize supplies will have
to double in the next 15-20 years to meet food demand. The challenge of
:increasing food production has been met by a government decision to give
,priority to investment in agricultural research and improve its
effectiveness in technology generation and transfers Yet, in order to
achieve this goal, research priorities must be focused on the key staple
foods.

Pressure on. budgets and growth of public expenditures have
increased the demand for economic assessment of research priorities and
investments. Schuh. and. Tollini (1979) highlights three benefits of
assessing a research activity: 1) to provide a basis for soliciting and
justifying budget support; 2) to provide for an efficient allocation of

scarce research resources; and 3) to enable financing of priority

10
research.

Much of Kenya's agricultural success is attributed to
agricultural institutions such .as the research, extension, credit and
input delivery systems, some of which were inherited from the colonial
era. Agricultural research in Kenya began with the establishment of a
multidisciplinary research centre in Nairobi in 1903. Today, there are
22 agricultural research centres located in. different agro-climatic
zones and managed by a newly formed Kenya Agricultural Research
Institute (KARI), a government parastatal set up to co-ordinate, execute
and manage all crop and livestock research activities. Figure 9 in the
Appendix presents the organization structure of' KARI. The research
centres are allocated national and/or regional mandated commodity-
research programs (Kenya, 1986b).

The general objective of this study is to estimate the rate of
return to maize research in Kenya from l955-88. The specific objectives
are: 1) to describe the evolution of maize and its importance in Kenya;
2) to review the development of maize research and the maize sub-sector;
3) to review studies on the measurement of agricultural oresearch
productivity; 4) to generate data and information on the costs and
benefits of maize research between l955-88; and 5) to use a production
function approach to evaluate the rate of return to investments in maize

research over the 1955-88 period..

WM

Maize was introduced. to Kenya following the importation. of
several types of maize from North and latin America via South Africa
around the turn of this century. By 1903 maize occupied about 20 percent
of the total foodcrop area and formed the staple diet of the Kikuyu and
Kamba (Meinertzhagen, 1957).1 By 1960 maize was already established as
an important food crop, occupying about 44 percent of the total crop
area (Kenya, 1966). The crop spread quickly because it is easy to grow,
has few serious pests and diseases, provides a good yield, is easy to
store and is more palatable in various forms than traditional cereals
such as millet and sorghum (Allan, 1971).

The most popular imported varieties were Hickory King, White
Horsetooth, Ladysmith White and Salisbury White. These and other
introductions became intercrossed in Kenya to such an extent that their
identity’ was lost, but European. settler-farmers established locally
adapted strains for different areas mainly by crib selection. In this
way, Kenya Flat White emerged as a recognized, variable but reasonably
stable, mixed population or complex. From the settler-farms, Kenya Flat
White was spread by African labourers to their home areas. During this

process, some admixtures with the local Caribbean Flint types must have

 

1 These two Kenyan tribes were the first agriculturists to have contact
with white settlers.

ll

l2
taken place, accounting for the relatively large proportion of yellow
and purple kernels observed in smallholder crops. The Kenya Flat White
Complex was adapted locally in Central Province as "Muratha" and in
Eastern Province as the "Machakos Local White."

The last distinct type of maize to be introduced to Kenya was a
high altitude race, "Cuzco" from Peru (Grobman M, 1961). It was
introduced by missionaries before 1914 and is traditionally known by the
Kikuyu as "Githigu." Today, it is the only type found over 2400m ASL in
the Central Highlands of’ Kenya. This variety has Shown remarkable
resistance to maize streak-virus disease which is common problem in the

cool highlands.

13

W

The initial hybrid maize breeding effort stemmed from a
suggestion in 1909 by an American scientist, G.H. Schull, that sought to
exploit the genetic explanation of hybrid vigor (Hallauer and Miranda,
1988). During the period 1910-30 the genetics of hybrid maize breeding
were unlocked and a number of synthetic and hybrid varieties developed.
By 19503 most of the American Corn Belt was already planted to hybrids
(Griliches, 1958).

In Kenya, maize breeding work started at Njoro in 1930, focusing
mainly on producing varieties for large-scale settler farmers. The
initial effort involved disease screening, inbreeding and hybridization.
The program made little progress and was abandoned in 1945. It was
resumed in 1948 but the breeding stock and records were destroyed by a
fire. However, a more systematic maize improvement program was started
at Kitale in 1955 when a full time breeder, Michael Harrison, was
appointed to develop late-maturity maize hybrids suitable for the maize-
growing regions receiving 750-2000 mm of rainfall annually in 6 to 8
months. From inbred lines of the Kenya Flat White Complex a synthetic
variety, Kitale Synthetic II (KS 11) was developed and released in 1961.
This open pollinated variety gave a seven percent yield advantage over
the local maize and was widely adopted by both large- and small-scale
farmers. But a narrow genetic base of the Flat White Complex posed a

problem to the breeders. To overcome this, accessions from Latin America

14

were imported to provide needed genetic diversity. These introductions
formed the foundation of a successful maize improvement program in
Kenya. After preliminary screening in a top-crossing block, to KS II as
a tester, 124 test crosses were grown in 1961. Most outstanding among
them was the cross of KS II and Ecuador 573, an unimproved stock of a
high altitude, late-maturing variety fom Ecuador. These two varieties
had enough genetic diversity to provide excellent heterosis. The first
successful cross, released as Hybrid 611 (H611) in 1964, yielded 40
percent more than KS II (Harrison, 1970). During the same year two other
hybrids, H621 and H631, having an average 26 percent yield-advantage
over KS 11, were released.

Between 1965 and 1989, eleven high-altitude maize hybrids were
developed and released to farmers. H625 and H626, released in 1981 and
1989 respectively, were jointly developed by the Kitale program and
Kenya Seed Company (KSC). H626 is currently the highest yielding hybrid
variety. On average, the late-maturity hybrids out-yielded the local
farmers' maize by 30-53 percent (NARC, 1990).

Meanwhile, the Kenya maize improvement program had been expanded
to develop varieties suited to different agro-climatic zones, from the
semi-arid east and sub-humid coastal belts to the mmist highlands and
cool, frost-ridden zones above an altitude of 2400 m. Maize research on
early-maturity varieties was started at Katumani in 1956 to develop
varieties for the semi-arid regions which receive low, erratic rainfall
of about 250-400 mm a year falling within 60 days, and barely enough for
a successful maize crop. The challenge was to develop drought-

resistant/tolerant maize varieties that would withstand long durations

15

of low moisture and still give good yield. The initial breeding effort
involved early-maturity ”Taboran" from Mexico and Machakos Local White,
a derivative of the Kenya Flat White Complex. The results of their
crosses led to the development and release of Katumani Composite A (KCA)
and Katumani Composite B (KCB) in 1966 and 1968 respectively. These
open-pollinated varieties tassled in 55 days and were fairly successful
in the semi-arid areas. However, because of their dismal performance in
regions where rainfall lasted only two months, a concerted effort was
needed to further reduce the flowering and maturity period. The effort
paid off in the development of Dryland Composite I (DC I) which was
released in 1989. This variety flowers 4 to 7 days earlier and is more
reliable in drier conditions than both KCA and KCB. Current programs at
Katumani focus on. improving existing composites and. evaluating 'new
introductions from CIMMYT and IITA. There is, however, great concern on
the poor yield performance of the composites, a reason that has made
them less attractive to farmers. Part of the reason is that farmers in
these regions use little fertilizer or complementary inputs (NDFRC,
1986).

Research on medium-maturity maize was started at the Embu
Regional Research Centre in 1965 for regions of the Central Highlands
receiving 350-750 mm of rainfall in two distinct seasons and requiring a
variety that takes 5 to 6 months to mature. A cross between Kitale late-
maturity hybrids and Katumani early-maturity maize led to the release of
H511 in 1968, the firstt medium-maturity hybrid, followed by another
medium-maturity variety, H512, in 1970. These varieties had a research

yield-advantage of 36 percent over the local maize, 'Muratha”. H511 and

l6
H512 have been successful in the lower and drier parts of the Central
Highlands. In the higher and cooler parts the farmers cultivate local
maize types which are more resistant to frost, maize leaf-rust and
streak-virus. Current maize programs at Embu focus on improvement of
yield and disease-resistance.

Maize improvement for the medium-rainfall, low-altitude tropical
region began in 1952 at the Coastal Research Station when short season
varieties were screened for resistance against maize leaf-rust (Eggginia
Mm). Ten varieties performed well and were incorporated into a
composite, Coast Composite (C.C.), developed from basic material
imported from the Latin American lowlands. The C.C. was not widely
adopted by farmers because of its low yield and yellow kernels, a
characteristic that is not acceptable to the marketing board.
Consequently, an ear to row selection program for clearing the yellow-
kernel coloring was initiated in 1983. The work was not successful until
its sixth cycle in 1987 when the yellow color disappeared. A notable
breakthrough by the coastal maize program was the successful development
and release of Pwani Hybrid I (PH 1), the first hybrid for the Coastal
region. PH 1, released in 1989, has a 5-15 percent yield advantage over
C.C. and matures ten days earlier.

Table 2 presents the remarkable output of the maize improvement
program in Kenya since its inception in 1955. Thousands of inbred lines
are currently being screened and tested in the maize program; KSC has
about 1400 lines on test (Ndegwa, 1990). It is expected that by the year
2000 many more superior varieties will have been released to provide

sufficient and suitable choice of maize types to farmers.

17

Table 2: Kenya: Improved/Hybrid Maize Varieties Developed and Released.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

   

 

Source: Ochieng,

1988

 

; Ndegwa and Ndambuki, 1988.

 

 

VARIETY YEAR OF DAYS TO YIELD GROWING DEVELOPED
RELEASE MATURITY (t/ha) ALTITUDE BY
KSII 1961 l80-240 3.37 l500-2100 NARC
H611 1964 l80-27O 4.50 1800-2400 NARC
H621 1964 l80-240 4.05 1000-1700 NARC
H631 1964 180-240 4.45 1000-1700 NARC
H622 1965 l80-210 5.22 1000-1700 NARC
H632 1965 180-210 4.45 1000-1700 NARC
H612C 1966 l80-270 5.88 1200-1800 NARC
H6110 1971 180-240 5.85 1800-2400 NARC
H6l3C 1972 180-270 5.96 1500-2100 NARC
H614C 1976 180-270 6.30 1500-2100 NARC
H625 1981 l80-240 6.75 1500-2100 NARC,KSC
H612D 1986 180-240 6.43 1500-1800 NARC
H613D 1986 l80-24O 6.03 1500-2100 NARC
H614D 1986 180-240 6.56 l500-2100 NARC
H626 1989 180-240 6.76 l500-2100 NARC,KSC
Kat.SII 1963 90-120 2.00 1000-1700 NDFRC
KCA 1966 90-120 2.25 1000-1700 NDFRC
KCB 1968 90-120 2.80 1000-1700 NDFRC
H511 1968 120-150 3.60 1000-1500 RRC-EMBU
H512 1970 125-155 4.05 1200-1900 RRC-EMBU
DC I 1989 80-110 2.89 1000-1900 NDFRC
CC 1974 120-150 3.30 0-1000 RRC-MTWAPA
1989 100-130 3.78 0-1300 KSC

 
 

 

 

Key: NDFRC = National Dryland Farming Research Centre, Katumani.
NARC - National Agricultural Research Centre, Kitale.

RRC
KSC

Regional Research Centre.
Kenya Seed Company.

18

LAW

Maize improvement in Kenya is based on plant breeding and
complementary research on disease and pest resistance, fertilizer and
soil characteristics. An extensive maize agronomy program has been in
existence since the breeding program started. Between 1950 and 1960,
many uncoordinated trials were carried out and observations made on
fertilizers, spacing and other treatments. But this pioneering work came
under heavy criticism because most of the limiting factors to yield
increases were examined singly or at best in pairs. Also maize varieties
with the genetic potential to give full response to improved conditions
were not available, planting time of the trials was often late, plant
populations were usually low and other cultural conditions were seldomly
optimum. Because of these shortcomings, there was confusion over the
many factors limiting maize production.

The most notable work on maize agronomy was performed by A.Y.
Allan with funds provided by the Rockefeller Foundation and the British
Government. Starting in 1963 and working closely with maize breeders at
Kitale, Allan developed a systematic agronomy program and evaluated new
hybrids from the breeding program over a wide range of agronomic
characteristics. He determined six agronomic requirements for the
hybrids using 33 and 26 factorial district maize variety and husbandry
trials. His results are shown on Table 3. Each of the six factors
considered was included at two levels, a "high" level representing the
recommended practice and a "low" level corresponding to the farmer's

practice. Time of planting and the genotype were found to be the most

19

scope u“ Knows"

mmmoonm on accompans can Hanan cum 0: Sense «rowan.

Hommtmd.

 

 

 

 

 

 

 

 

 

 

 

 

«mar—u >95.— wnEJ 5:32. 08. N
3.22. 4.48.32: =.a\>n.o m2.==m«\>03 m3.=an<>n8

is... on 132:“ 9»: ea a?“ 38 NB $5 :50
a $8.8 .22. 8.8

22.3 12. and 58¢ .38 :u a
«.8: 3.3

4.35 ca «02. :11,— sms Sn 5
hon.— uumo

>395. 2. tone?“ .930 :38. Saw. 38 Ba 8
0:8. _..:o ”88

3.8158 e2. 38 3 5. :3 8 u»
Zoao 88

256%: can and .3 =9 38 a .3
Zoao was

 

monsoon >. .5 2.2.. .938. $5925 .39.“ me £883 .933. 38 a... 33.. 9.22. m: 3.2. 3.38:. .330

573333. E 92218.? O. r. >. rag—"Q. 03—. 539338. m: m»: >38. 5.3. a. a.

>. 58 .18 0.. NM\ 62. ”8 .55... 95.

war...— o: 8a.” on mac...“ Sacra.— uac 8:332. 33.. 829

 

20
important factors explaining low farmers' yield, followed by weeding
standard and plant population. Fertilizer application was found to be
the least important factor influencing yield (Allan, 1971).

Maize researchers at Kitale put the husbandry trials to good use
by developing a demonstration "diamond” to serve as a reference for
agricultural extension staff. Figure 2 presents the interaction effects
of Allan's trials. The interactions between time of planting, plant
population and genotype with fertilizer response indicated that apparent
uneconomic application of fertilizer is profitable if all the other
limiting factors are removed. The lesson of the trials, therefore, was
that fertilizer 'use and.'hybrid. seed. are 'not substitutes for sound
farming practices such as timely planting, weeding and appropriate plant
population. Thus, it would be unwise to advise farmers, especially poor
smallholders, to spend large sums of money on hybrid seed and fertilizer
before they can raise their husbandry standards to levels that allow
hybrids to express their full yield potential (Harrison, 1970).

Currently, the agronomy program augments breeding research by
evaluating the new varieties in different agro-ecological zones, seasons
and farmer circumstances. Subsequent agronomic and on-farm research
have identified the following factors as influencing maize yields, in
order of importance: 1) Land preparation and time of planting; 2) Weed
control and plant population; 3) Genotype; a) Fertilizer; and 5) Pest

control and time of harvesting.

21_

Eigure 2: Kenya: Maize "Diamond”

4—vL—

 

ti BegeIAete
Pearle-hem, Local HebeSeed. Noledihet

 

 

 

 

 

[—14.0 Begum (01% been)? [ 21.! loge Here (was lneteeee)
Poet I'll-bench. Hybrid Seed. Good Husbandry, Local Seed.
'1 human I l No rum

 

 

 

 

as Bags/Ac" Miocene)
Goal-laundry. W30“. Feather

 

 

Sontce: Allan (1968)

22

Today, the production of certified maize seed is accomplished
through varietal development, evaluation and release, maintenance
multiplication, processing, storage and distribution. The development,
evaluation and release of the new seed is the responsibility of the
National Maize Programme under KARI. Seed production, processing and
distribution is the responsibility of the Kenya Seed Company (KSC) while
seed certification and testing are handled by the National Seed Quality
Control Services (NSQCS), a designated certifying agency within KARI.

Once a breeder has identified a high performance maize variety
in the initial screening trials, it is recommended for inclusion in the
National Performance Trials where it is tested for various agronomic
characteristics. After three to four seasons in the NPT, the co-
ordinator of the trials submits a detailed analysis of the performance
data to the National Maize Research Committee which examines the
findings and, in turn, submit its recommendations to the National
Varietal Release Committee. The latter committee, comprising of
participants from KARI, KSC, NSQCS, Universities, extension agents and
representative farmers, evaluates the release recommendations and
determines whether a variety should be released to farmers.

Usually, a small quantity of breeders' seed is supplied to the
KSC from the maize program for multiplication and maintenance. Most

certified maize seed is grown through an outgrowers' seed production

23
program by farmers contracted by the KSC. The farmers and farms are
selected based on their suitability to basic requirements set by the KSC
and NSQCS for seed production. The proximity of the farms to the KSC is
important for regular inspection of seed and husbandry practices used by

the outgrowe rs .

Prior to 1963, little certified maize seed was available to
farmers. Better farmers selected seed from their own crop, processed it
on-farm and sold it to other farmers, retaining a portion for the
following season. The sale of seed was mainly on a farm-to-farm basis
with little or no control by any authority. KSC was formed by large-
scale farmers in 1956 in. Trans-Nzoia, mainly to multiply improved
pasture seed. As a result of a reduction in the demand for pasture
seed, KSC almost collapsed in 1961. The National Maize Programme came to
its rescue by proposing that it undertake the production of' seed for
maize varieties released by Kitale. In 1963 KSC entered into an
agreement with the Kenyan government to produce and distribute improved
maize seed.

Since the KSC was initially geared to serve only large scale
farmers in the highlands, its distribution ‘network ‘was limited to
terminals at the main centres along the railway line. To respond to the
growing demand for improved maize seed, the KSC initiated a seed growing

programme and expanded its distribution network by recruiting seed

24

stockists in almost every trading centre. From 104 in 1966, the number
of seed stockists increased to 2541 in 1976 and an estimated 6000 by
1980, approximately one for every eighty hectares of improved maize
planted. This rapid expansion of coverage followed the existing road
network. Although some disparity in the stockists distribution between
districts was noted, there was a general improvement in the availability
of seed to small-scale farmers (Rundquist, 1989).

Figure 3 shows the rapid increase in hybrid seed sales by the
KSC over a period of 26 years. Starting with a mere 4 metric tons (t) in
1962/63, the company's maize seed output increased to 10,600 t in
1975/76 and 21,800 t in 1987/88. This indicates a rapid adoption of the
improved seed.2 Table 4 shows the area planted to improved seed by
small- and large-scale farmers.3 The figures reveal that the greatest
increase in area planted to improved and hybrid varieties was due to
adoption of hybrid seed by small-scale farmers. Currently, the KSC
produces about ten different hybrids and £29 open-pollinated varieties
suitable for four agroclimatic areas of Kenya. In 1984, the company
also produced seven new experimental hybrids (Ndegwa g§_§1, 1985). In
essence, the KSC and KARI researchers at Kitale, Katumani, Embu and
Mtwapa are close partners in the quest for excellence in maize seed

production.

 

2See Figure 5 in the Appendix A.

3 The area is estimated from a recommended seed rate of 22.45Kg/Ha and
assuming no loss of seed at planting.

METRIC TONS

25

Figure 3: Kenya: Hybrid Seed Sales

 

 

 

 

25

20- \
{”3 15‘
C2
(0
05'
:3
2
b 10-

5- +\F*/(/

19‘5'5‘ 'ie‘eb' 'iéeis' 'i9‘7b‘ ' i9'7‘sf'19‘ab' #92357 '

‘YEAG!
Source: KSC. Unpublished data.

 

26

 

 

 

 

 

-7
Year Small-scale Large-scale Total Area
Farms Farms
( Area in Hectares)

1967/68 51320 36516 87836
1968/69 64333 39516 103849
1969/70 96971 45915 142886
1970/71 149971 63811 213782
1971/72 206947 73975 280922
1972/73 264871 53392 318263
1973/74 292501 39232 331733
1974/75 352276 50717 402993
1975(26, 421553 50607 472160
Source: KSC. Unpublished data.

27

Worm

Maize is considered the national food staple because it accounts
for over 50 percent of both calories and protein intake and is a major
source of fats (ISNAR, 1985b). In Kenya maize can be consumed as: (1)
roasted or boiled grain on cob, a favorite snack; (2) whole grain meal,
boiled in mixture with beans or peas to make a popular meal called
"Githeri"; (3) ground maize flour which is boiled in water to make
another popular meal called "Ugali", a thick semi-solid mass, or a thin
porridge, ”U11"; and (5) an alcoholic drink brewed from a fermented
mixture of maize and sorghum or millet in water. A survey of 349
families in Nairobi in 1958 revealed that maize accounted for 80 percent
of starchy-staple calories while data from a 1969-70 Nairobi Urban
Survey indicated maize as the cheapest source of calories and was second
only to cowpeas as a source of inexpensive protein (Miracle, 1966;
Gerhart, 1975).

About 90 percent of smallholders account for 80 percent of the
total maize area and produce between 70 and 90 percent of the total
annual maize production (Ruigu, 1985; Ackello-Ogutu and Odhiambo, 1986;
Odhiambo, 1988). These smallholders consume most of their maize on the
farm, selling only 22 percent to the market compared to their large-
scale counterparts who market 75 percent (Paliwal M, 1984). Table 5
presents a summary of estimates of maize output, area and yield in Kenya
for the period 1955-88. Figures 4(a), 4(b) and 4(c) in Appendix A show

the trends of maize area, yield and production, respectively.

28

 

 

 

! 5 ., . . ., , .3
YEAR OUTPUT YIELD AREA IMRBQEEDZHXEBID_MALZE
(t) (tlhg) (ha) ABEA‘hg)‘ 1 92 IOIAL
1955-59 542224 0.868 625240 - -
'1960-64 805650 1.058 760900 13623 1.79
1965-69 1089678 1.292 843195 95487 11.32
1970-74 1344400 1.510 891078 309378 34.72
1975-79 1772000 1.562 1143817 498989 43.62
1980-84 1936880 1.622 1204823 648855 53.85
1985-88 2670400 1.845 1445085 917576 63.50
SOURCE: Kenya colony, Crop Production Review. Various issues.

Kenya, MOA/DPD. Various Reports.
FAO Production Yearbook, various issues.
a Estimated from the recommended seed rate of 22.45kg/ha.

During the decade from 1955-1964. maize production and. area
increased steadily while average national maize yield remained
relatively constant. The total maize area averaged 684,000 ha,
increasing from about 509,000 ha in 1955 to 830,000 ha in 1964, the time
of the first hybrid release. This increase in area was largely
attributed to increased accessibility of agricultural land to the local
people and greater incentives to farmers following the initiation of the
Swynnerton Plan in the mid-19503 (Ruthenberg, 1969).

From 1965-1974 seven hybrids and three composite varieties were
released from the maize program. The average yield increased from 0.937

t/ha in 1955-64 to 1.401 t/ha in 1965-74, an increase of about 49

29

percent. In the same period, maize area increased by 27 percent. These
area and yield effects led to an 83 percent increase in total maize
output. Between 1975 and 1989, seven more hybrids and one composite were
made available to farmers; the average yield improved by about 19
percent during this period accompanied by an area and output increase of
44 and 72 percent respectively. The area planted to improved maize seed
increased from an average 1.8 percent in 1962-64 to an average 63.5
percent in 1985-88. This compares to an increase in 'hybrid. maize
coverage in the US Corn Belt from 0.2 to 90 percent in about the same
number of years, from 1933-58 (Jugenheimer, 1958). In summary, the
national average maize yield increased from 0.87 t/ha in 1955-59 to 1.85
t/ha in 1985-88; the area more than doubled, from 625,000 ha to 144,500
ha; and output increased fivefold, from 540,000 t to 2,670,000 t during
the same period.

Disaggregation of maize yield, area and output by high- and
marginal-potential regions revealed that between 1955-88: 1) a larger
proportion of maize area was in the marginal-potential (M-P) maize-
growing region; 2) farmers in the M-P region achieved lower maize yield
than those in the high-potential (H-P) maize region; and 3) the rate of
growth of maize yield in the M-P region was slower than that in the H-P
region.4 Four reasons account for these trends. First, maize varieties
for the H-P areas are higher-yielding than those available for the M-P

areas. Second, the climate of the H-P region is more favorable for maize

 

4 High-Potential regions represent maize-growing in the highlands
traditionally occupied by settler farmers. Marginal-Potential regions refer to
maize-growing area in the medium- and low-potential zones which are largely
occupied by smallholders.

30

production, allowing greater exploitation of the genotype, than that of
the M-P region. Third, farmers in the H-P region adopted maize
technologies relatively faster than the farmers in the M-P region
(Ongaro, 1988). Finally, a greater proportion of agricultural land is in
the M-P regions. The trend indicates that maize area, yield and output
increased in both the H-P and M-P regions between 1955-88. Area
increases in the H-P region are attributed to a shift from pasture to
maize cultivation by large-scale farmers and the sub-division of
formerly large wheat farms into smallholder farms where mixed-farming is
pursued. In essence, large wheat farms have been transformed into maize
farms (NPBS, 1987). On the other hand, the increase in area in the M-P
regions was largely due to an increase in cultivated land.

But the percentage changes in maize area, yield and output
reveal a slowing of maize production increases. Yield increased by about
49 percent between 1955-59 and 1965-69, but only by 18 percent between
1975-79 and 1985-88; area increases slowed from 35 to 26 percent for the
same periods. Efforts are required to increase maize production and
sustain the increases in the wake of rising food demand in Kenya.
Various medium term options have been suggested, such as increased
adoption of existing maize technologies, and appropriate pricing,
marketing and credit incentives. In the long term, improved
technologies to enhance land ,and labour productivity in maize
production, especially among the smallholders, are required.

Previous research identified the time of planting followed by
genotype as the most important factors affecting maize yield (Allan,

1969,1980; Ngugi, 1982). ‘ Based on Allan's work at Kitale, the KSC

31

estimated, as a rule of thumb, that there is a loss of 90 kg/ha per day
for maize planted late, so that with approximately one million hectares
of maize planted 10 days late on average, the loss from planting is
approximately 900,000 t of maize per year (Schluter, 1984), about 30
percent of the average national maize output. Late planting is often
caused by delays in payments to farmers for the previous year's crop
delivery to the NCPB, credit disbursement and input delivery. In many
countries it has been shown that maize output can be increased through
the adoption of hybrid seed and complementary inputs such as fertilizer.
But on-farm research in Kenya reveals low input-use on maize among
smallholders, especially in marginal areas (NPBRC, 1987). Ackello-Ogutu
and Odhiambo (1986) reported that between 1969-85, phosphorus and
nitrogen application on maize in Kenya increased by one and four percent
per year respectively, and that an increase in fertilizer expenditure by
10 percent led to a 15 percent increase in maize output between 1976-
1978.

Gerhart noted that despite the fact that "both large- and small-
scale farmers in the high rainfall portions of western Kenya adopted
hybrid maize at a rate faster than American farmers in Iowa in the 19205
and 19308" (1975, p.26), there were differences in adoption rates
between small- and large-scale farmers, and between agro-ecological
zones. Table 6 shows the adoption of hybrid maize technology in western
Kenya by farm size and agro-ecological zones. The results indicate that
large farms consistently led small farms in adopting maize technologies
except insecticides in zone 3. Bridging the gap in maize technology

adoption between the two farm categories and between different agro-

32
climatic zones will increase maize production. The World Bank (1984)
contends that maize production in Kenya could increase by up to 70
percent if the current maize technology is improved and that, in many
cases, smallholder yield could be doubled through the expanded use of
improved inputs and husbandry. But intensification of maize production
will require a concerted effort of agricultural support services,
particularlly' extension, and. credit, and. improvement of ‘policies to
stimulate production. Table 7 shows the response of some farmers in
western Kenya who had not planted hybrid seed in 1973. All the farmers
interviewed in Zones 1 and 2 gave "cost" as the only reason for non-
adoption while in Zone 3 identified "cost” as the major constraint. Even
today, the cost of hybrid seed and fertilizer make up a large share of a
smallholder's farm-inputs budget and constantly feature as a constraint

to adoption of improved maize technology.

1! Q '911' xeee e10 1‘. ,‘ ; . LCL!‘ e-'

 

Zone 1 Zone 2 Zone 3

 

1. HYBRID MAIZE:

Small 95.7 83.7 14.9

Large 95.8 95.1 17 4
2. FERTILIZER:

Small 47.8 71.4 2.1

Large 75.0 92.5 6.4
3. INSECTICIDES:

Small 2.2 21.3 7.1

Large 10.6 43.9 2.2

 

Source: Gerhart, 1975, p. 24.

 

 

 

(Percentage)
Zone 1 Zone 2 Zone 3

No. of responses 9 14 102
1. Cost: Too expensive/No money. 100 100 52
2. Not available/Distance to stocklist/

Must buy every year. 0 0 6
3. Never heard of it/ Does not know how

to use it/No experience with it. 0 0 l4
4. Performance: Yields less/Less Certain/

Does not do well. 0 0 20
S. Congruence: Too much work/Planted

late/Misc. 0 0 7

 

Source: Gerhart, 1975, p. 23.
Note: Total responses exceed the number of non-adopters since some
farmers gave more than one answer.

34

24.811121113951112

Maize marketing is controlled by the Government through the
operation of a parastatal marketing board, the National Cereals and
Produce Board (NCPB), established in July 1979 as a result of a merger
of the former Maize and Produce Board and the Wheat Board of Kenya. NCPB
handles the purchase, storage and distribution of all major food grains
throughout the country. Maize is its most important commodity.

A dual marketing structure was developed for maize as a result
of pre—independence policy and practice. Under colonial rule, African
producers had no access to a marketing board, thus, forcing them to
create a parallel or informal market. Most smallholders sell their maize
in the parallel market. Because of the small size and dispersion of such
maize sales, small traders tend to be more efficient and cost effective
in handling smallholder sales than NCPB, whose activities are largely
concentrated in maize-surplus, large farm areas (World Bank, 1986). The
NCPB controls the official marketing channel.

For most of the period after independence, the grain marketing
board has had instructions to purchase all the marketed maize that is
not sold directly from a producer to a consumer at a controlled price,
and to sell the maize at a higher, also controlled, price. The official
producer price is announced each. year ‘before planting and is not
expected, by law, to fluctuate through the year. But, in reality, both
producers and consumers have faced large price fluctuations because of
an increasing dependency of both groups on the parallel market,

especially when surpluses have exceeded NCPB's storage capacity

35

(Pinckney, 1988). This often results in maize flooding the parallel
market and depressing prices.

The Sixth National Development Plan outlines the intentions of
the government to reduce the mandate of the NCPB to a custodian of a
proposed national strategic food reserve and a buyer of the last resort.
Under this plan, NCPB would leave up to 75 percent of maize marketing to
private traders, millers and farmer co-operative societies. This reform
will be accompanied by the removal of the inter-district maize movement
restriction imposed to reduce consumer price parity beWeen maize-
surplus and maize-deficient regions (Kenya, 1988). These changes are
expected to take place during the current five-year development period.
Meanwhile, the government is taking steps to reorganise and prepare NCPB
for its new role and to improve the existing market infrastructure.
Already, studies have been undertaken to work out a reform program for
NCPB in the field of general management, rationalizing of buying
centres, financial restructuring and monitoring of market conditions
(Odhiambo and Wilcock, 1989). This restructuring will, however, require
substantial financial costs because the NCPB has incurred huge financial
losses in the past.

Kenya urgently requires more grain storage capacity. The current
NCPB's storage capacity is 896,000 tons. This is supplemented by a
886,000-ton capacity provided by.the private sector. Maize occupies 84
percent of the total storage capacity. However, it is estimated that a
capacity of 1,346,000 tons is needed for maize. Because of poor on-farm
storage conditions and inadequate NCPB storage capacity, a considerable

amount of grain is lost. Most NCPB's buying centres have no storage

36

facilities. During peak buying periods, bags of grain from farmers are
piled up or stacked on the yards of buying centres until trucks are
available to transport them to depots. If transport is delayed and
rainfall comes, the grain losses are considerable. The construction of
additional storage facilities is underway. And when it is completed, the
storage capacity is envisaged to increase from 40 to 70 percent of the
total marketable maize production (Kenya, 1984).

Table 9 in Appendix A presents the amount of maize purchased by
NCPB by province from 1966/67 to 1988/89. The data show an increasing
trend in maize purchases from about 225,000 t in 1966/67 to about
570,000 t in 1976/77 and 833,700 t in 1985/86. Large fluctuations in the
amount of purchased maize exist by province and year. These fluctuations
are attributed to fluctuations in maize production and NCPB's purchasing
ability. Rift Valley and Western Provinces combined account for about 70
percent of the total maize sales to NCPB. Future projections indicate
increases in marketable maize production and hence the need for storage

expansion.

37

2.6 Financing Maize Research

The 4th National Development Plan (1979-83) identified the need
for a substantial increase in resources to agricultural research to
overcome technological constraints (n1 agricultural production (Kenya,
1979). A target level of two percent of Kenya's agricultural GDP has
been suggested as an appropriate level of funding for research (ISNAR,
1985a). Funding for agricultural research in Kenya is provided almost
entirely by the public sector through the Ministry of Research, Science
and Technology (MRST). Until 1987/88, funds were channelled through the
Ministry of Agricultural and Livestock Development (MALD), except in
1982/83 when HRST provided the appropriation for KARI. Kenya's research
also 'benefits from donations by' external governments and agencies.
Besides KARI and independent institutions such as the befee Research
Foundation, Tea Research Foundation and the National Irrigation Board,
agricultural research is carried out by universities (Nairobi, Moi and
Egerton) and several international research centres and regional
programs based in Kenya. Private sector involvement in agricultural
research has been minimal, mainly adaptive in nature and the results are
seldomly made public.5

In a survey carried out in 1985/86 on private sector research
and resource allocation, it was 'found that 69 out of 364 responding

firms were undertaking some research and development (R & ‘0) work

 

5 Examples of private firms engaged in agricultural research include East
African Industries (Oilcrops), Kenya Breweries Ltd. (Barley), British Tobacco
C°|Pany (Tobacco). Kenya Canners Ltd. (Pineapples), East African Tanning
COlpany (Wattle) and Wellcome Kenya Ltd. (Livestock Drugs). Several of these
work very closely with the public research network.

38
(Makau, 1988). These firms spent :1 total of Kenyan pounds (K 19 2.38
million on R & D in 1985/86, averaging about K £ 34,500 per firm. By
comparison, the public sector spent K £ 25.06 million on R & D at an
average of K £ 1.14 million per institution.6 Therefore, the country
spent nearly K .f 27.5 million on R & D in 1985-86, of which only 8.7
percent came from the private sector. The total national expenditure on
R & D was only 0.51 percent of the national GDP. Table 8 shows the

distribution of research expenditure by the private sector.

Table 8: Kenya: Digtributiog of ResearchiExpenditure by the Private
Sector. 1985-86.

 

 

Science Group Number of Total Average
Establishments Expenditure Expenditure
(K!) (Ki)
Agricultural 17 1,000,000 58,824
Medical 4 802,250 20,063
Industrial 44 1,273,750 28,950
Natural & Physical 2 18,750 9,375
Social 2 5,250 2,625

 

Source: Makau, B.F. (1988). p.7.

The level of gross recurrent and development budget funding to
the Ministry of Agriculture (MOA) increased considerably between 1955
and 1988. In real terms, the recurrent expenditure increased from an
average of 1054.066 million per year in 1955-59 to K£12.242 million per

year between 1985-88 while the development expenditure rose from K{

 

6Excludes salaries of teaching staff at the national universities.

39

5.018 million to K£ 16.279 million over the same periods. Thus, the
total agriculture expenditure increased from an average K£ 9.8 million
per year for the period 1955-59 to about Kf 28.5 million per year in
1985-88 in real terms. 'This represents a threefold. increase in real
terms; in nominal terms, the total agricultural expenditure was
increased. by ¢over twenty times between 1955 and 1988. Figure 6 in
Appendix A shows the trend of the MOA expenditure in nominal and real
terms.

The government‘s expenditure on crop improvement averaged 43
percent of the total agriculture budget between 1955 and 1988. Whereas
the crop development expenditure rose, in real terms, from Kf 4.29
million per year between 1955-59 to nearly K! 12 million per year in
1985-88, the crop research allocation increased only from Kf 1.03
million to about K£'1.2 million annually in the same periods. Estimated
maize research expenditures increased in real terms by about 182 percent
between 1955-59 and 1974-79, and then declined by 55 percent between
1974-79 and 1985-88.’ The decline in ‘budget-funding in real terms
constrained the maize research program from the late 19705 until 1987-88
fiscal year when KARI received aid from several donors.

Meanwhile, there has been a continuous build up of resources,
facilities and manpower in Kenya's agricultural research system, making
it one of the largest research establishment in Africa. Kenya is second
to Nigeria in per capita research expenditure. The number of Kenyan

agricultural researchers increased from 18 in 1963 to 566 in 1982.

 

(Figure 7 in Appendix A compares the estimated maize research and
extension expenditures in real terms.

40

During the period, the proportion of expatriates in the research
institution dropped from 86.9 percent to 11.3 percent (Jamieson, 1981;
ISNAR, 1981,1985a). A review of Kenya's agricultural research system
indicates that there were 504 Kenyan agricultural researchers (275 Bsc.,
212 Msc., and 17 Phd.) in 1986 (ISNAR, 1990). In 1980, Kenya had a
larger number of agricultural research scientists per million people
(24.3) than Sub-Saharan Africa (15.1), Asia (15.2) and Latin America
(22.7) (Oram and Bindlish, 1981).

Agricultural research receives 70 percent of Kenya's national
government budget for research, reflecting the importance and dominant
role «of agriculture in the economy (Ruigu, 1985). Research on the
traditional export crops, coffee and tea, receives about 32 percent of
the total agricultural research funding as compared to 25 and 23 percent
for food crops and livestock research respectively. In 1979/80, tea and
coffee received nearly one-third of the allocation while livestock
research got about one-fifth (ISNAR, 1985b).

The allocation of research funds varies by region. In 1983/84,
more than 90 percent of funds and scientific manpower were concentrated
on high- and medium-potential areas. But in recent years there has been
greater effort by the government to shift resources to medium-potential
regions as they become increasingly important sources of agricultural
growth following rapid papulation growth and migration. Jamieson(l98l)
studied resource allocation in agricultural research during the first
fifteen years of independence and found that: (l) the government played
an important and direct role in agricultural research through the former

Scientific Research Division (SRD); (2) research was predominantly

41

applied in nature; (3) a greater amount of research effort was spent on
biological and chemical research rather than mechanical, other applied
or basic research; (4) nearly all the research was directed at the needs
and ‘problems of large-scale, high-income farmers but it *was mostly
scale-neutral; (5) the largest proportion of the resources was allocated
to livestock followed by cash-crops, food-crops, pasture and fodder
crops; (6) coffee, maize, wheat and tea received the largest allocation
of crop research in that order; (7) relatively more research was done on
medium-potential than high- or low-potential areas, with low-potential
zones receiving the least attention; and (8) most research focussed on
increasing the yield level rather than increasing the reliability of
yield or improvement of crop storage. The study also found that the
resource allocation process was influenced by the availability of farm
inputs whereas changes in the relative product prices or levels of
output seemed to have little impact. Lastly, the allocation process was
found to be only slightly responsive to economic, social and political
inducements for change, and the political and institutional factors
retarded changes in the orientation of research policy.

Several studies have identified the following constraints to
Kenya's agricultural research: 1) a persistently high turnover of
trained and experienced scientists; 2) lack of adequate manpower and/or
research-oriented training; 3) inadequate and erratic budget funding for
research; and 4) a growing deterioration of research infrastructure.
Unless these problems are offset, they could seriously reduce the
capacity of KARI to generate agricultural technologies needed for

agricultural and economic growth (ISNAR, 1981, 1990).

3. CHAPTER THREE

RATE OF RETURN: LITERATURE REVIEW
3.1 Technical Innovation and Reggurce Productivity

”The capacity to develop and to manage technology in a manner
consistent with. a nation's physical and cultural endowments is the
single most important variable accounting for differences in
agricultural productivity among nations" (Ruttan, 1982, p.17).
Exploitating such capacity generates appropriate technologies which
facilitate the substitution of inexpensive, abundant resources for
scarce and. expensive resources, thus releasing constraints to
agricultural growth imposed by inelastic supply of resources. The
resulting benefits accrue to consumers from increases in productivity
and supplies of agricultural goods at relatively lower prices while
producers experience lower production costs as efficiency is enhanced by
the new technologies. For these reasons, it has become widely accepted
that agricultural research is an important means of raising agricultural
productivity.

Investments in research are usually costly and in most cases the
results are not immediate. This raises the issue of how scarce research
resources should be allocated among alternative uses. Rationally, such
resources should be allocated to those investments that ensure a high
rate of return. Moreover, pressure on public and private investments in
agricultural research has heightened the need to justify such

investments vis a vis alternative public investments such as extension,

42

43
irrigation and non-agricultural investments. Tighter research budgets
make it important for national agricultural research systems (NARS) to
improve their priority-setting and resource allocation among competing
research programs.

Most NARS managers make allocative decisions based on their past
experience, an understanding of research goals and objectives, research
problems) and a sense of what is achievable through research. Then
allocative decisions are made on priority commodity research programs,
regions, factors of production, long-term versus short-term, basic
versus applied research, and the {distributional effects of research
(Norton and Pardey, 1987). In Kenya, resource allocation decisions are
heavily influenced by the previous year's budget. Changes in research
programs and budgets often results from requests by scientists, which
are evaluated relatively informally and aggregated into an overall plan,
and from the introduction of donor-assisted research projects.

Resource allocation decisions based on past experience are
important. In addition, increased use of analytical techniques may be
neccessary to improve priority-setting procedure for NARS decision-
makers and aid funding decisions. Such techniques can minimize large and
costly changes in research priorities, especially in systems
characterized by rapid turnover of research administrators (Schuh and
Tollini, 1979). There is a great need to improve the relevance and
productivity of agricultural research by focusing on priority research

and adopting appropriate funding decisions.

44

3.2 Measurement of Research Productivity

Numerous studies have been carried out to assess the impact of
research on agricultural productivity growth, particularly in the US,
Asia and Latin America. Table 10 in Appendix 8 presents a summary of the
results of these studies. Almost all of the studies indicated a high
rate of return to investment in agricultural research. Empirical studies
of social profitability in national agricultural research, particularly
in developing countries, have also shown consistently significant
productivity growth as a result of these kinds of investments (Ruttan,
1982).

There are several methodologies for assessment of the
contributions of research to agricultural growth. They can be grouped
into two major categories: ex post and ex ante approaches. The ex post
approach has been used to determine the past impact of investments in
research while the ex ante approach has been a useful guide to future
allocation of research resources in order to maximize their social
return. These methods help to evaluate research proposals and programs
with respect to their funding requirements in order to establish
priorities and justify budget requests.

The procedures used to make ex post evaluations can be grouped
into five different approaches: 1) the inputs-saved approach; 2) the
impact on national economy; 3) the production function approach; 4) the
economic surplus approach; and 5) the nutritional impact approach (Schuh
and Tollini, 1979). The inputs-saved approach attempts to estimate the

resources saved by the adoption of a new technology. The benefits of

45

research are estimated as the resources saved when a new technology is
used to produce an output compared to a base period production
technology. The costs are computed from publicly- and privately-financed
research and extension expenditures. A benefit-cost ratio or rate of
return can then be computed from the data. Schultz (1953) used this
method in what is believed to be the first attempt to quantify the
returns to investment in 0.8. agriculture. He approximated social
benefits by comparing the costs of producing 1950 outputs using 1910
technologies with the actual cost of producing the 1950 outputs. He was
interested in the whole sector and, thus, he made no attempt to consider
individual research programs or particular innovations. This approach
could be extended to individual research programs or to narrowly defined
technological innovations, especially’ those that are more resource-
saving than output-increasing. The procedure requires only modest data
but provides a crude estimation of the social benefits.

The impact-on-national-income approach is almost similar to (1)
and it provides a crude estimation of agricultural productivity by
considering the benefits of a new innovation as the resources it
releases to the non-farm sector. Tweeten and Hines (1965) calculated how
much lower the national income would have been if the farm population
never changed and the additional farmers had the income of today's
farmers rather than today’s non-farmers. This provided a crude measure
of benefits from technical change as a results of farmers adopting new
production technologies. The costs of’ public and private research,
education. and federal support programs were estimated and used to

compute a benefit/cost ratio.

46

The production function methodology involves the estimation of
input-output relationships for a commodity or sector. Griliches (1958),
Peterson (1967) and Evenson (1967) used this approach in the analysis of
hybrid corn, poultry and U.S. agricultural research productivity
respectively. Evenson (1967) estimated the rate of return (ROR) to
research from1949-59 to be 47 percent. Griliches (1964) had estimated
35-40 percent ROR to aggregate U.S. agricutural research for the same
time period; education of the farm labour force was also found to be a
significant variable in explaining increased agricultural productivity.

The production function approach involves aggregating various
outputs from the same commodity, using e.g. price weights, and inputs
(such as agricultural research, extension, etc.), and estimating the
contribution of each input to changes in output. One advantage of this
method is that marginal productivity of an input can be calculated.
Regression analysis is used to aportion the increase in productivity to
the various farm and non-farm inputs, and to test statistically the
impacts. Thus, the benefits of research can be imputed to particular
research programs, forming a basis for resource allocation. Brendahl and
Peterson (1976) estimated the marginal rate of return to investments in
research on cash grains, poultry, dairy and livestock by state for the
U.S., thereby providing a guide to allocation of research resources
among the commodities and geographic areas. Evenson and Kislev (1975)
separated research into indigenous research and research done in other
countries. Using cross-country data on maize and wheat, they determined
the contribution of each type of research to improvements in yields.

Evenson and Binswanger (1979) used publications as a proxy of research

47

output and, after separating research into scientific and applied
research, they calculated the impact of research on cereal grain
production across countries in nine major geoclimatic zones and forty
five regions in the world. Their results reported a strong
complementarity between the two types of research. The study determined
the contribution of regional and zonal research to country production.
Later, Evenson, Flores and Hayami (1976) also used a basic production
function approach to assess the difference in technology transfer
between an international and a national research program on rice.

The third method of estimating research productivity is the
economic surplus approach. This method measures the benefits and losses
from the adoption. of a new technology by assessing its impact on
producers and consumers. The new technology is assumed to shift the
supply curve to the right and create benefits to consumers through an
increased supply of product at reduced prices; producers benefit from
reduced unit production costs. This) method is flexible and can be
modified to assess the distributional impact of price and trade policies
of different structure of the economy (open or closed), among other
things. Akino and Hayami (1975) used this approach to estimate the
distribution of returns to investment in Japan's rice breeding program
among producers and consumers in different economic scenerios. Evenson
g£_gl (1976) evaluated the distribution of benefits from rice research
in Phillipines, while Echeverria (1990) studied the benefits of
generation and transfer of technology for rice in Uruguay. Griliches
(1958) study of the benefits of hybrid corn in the U.S. and the range of

examples cited here indicate the potential of using the economic surplus

48
model in different problem settings.

Although the economic surplus approach does identify the
contribution of research to an. overall increase in productivity, the
specification of an appropriate supply curve and supply-curve shift are
crucial to the results. Misspecification of the supply shift can lead to
inaccurate quantification of research benefits and their distribution to
producers and consumers (Daniels M, 1990). Several studies have
focused on the effects of inappropriate supply function formulation.
Lindner and Jarrett (1978) used an alternative formula for measuring
social surplus in corn production and found that Griliches (1958) had
overestimated the ROR to hybrid corn research by at least 50 to 100
percent. Similar comparisons with Petersons'(l967) study revealed that
he had overestimated social surplus by more than 150 percent.

Some studies have evaluated the effects of policy interventions
on research productivity measurement. 0ehmke (1988) found a divergence
of up to 100 percent in the rate of return (ROR) by comparing benefits
without intervention with benefits which account for market and

government budget effects. Alston et a1 (1988) examined the effects of

 

production quotas, subsidies and target prices on the benefits from
research to producers, consumers and the government. They concluded that
government intervention modified the pattern of distribution of benefits
from research relative to free-trade and, therefore, proposed that such
interventions should be accounted for in ROR measurements.

The nutritional impact approach does not in itself provide a
rate of return estimate but instead provides valuable information in

establishing research priority when improved nutrition is the research

49
goal. Pinstrup-Andersen, Londono and Hoover (1976) developed a such a
procedure to estimate the distribution of supply increases among
consumer groups, the related adjustment in total food consumption and
implications of these to calorific and protein nutrition. This procedure
derived the impact of increases in agricultural output on nutrition, by
income groups, and equity.

In addition to these five ex post evaluation procedures, there
are several ex ante procedures that are used to improve the decision-
making process in research resource allocation. The models range from
approaches which provide a systematic means of utilizing informed
judgement to approaches which attempt to provide empirical knowledge on
the consequences of alternative causes of action. These models can be
classified on the basis of time-frame (i.e. static or dynamic). degree
of uncertainty (i.e. deterministic or probabilistic) and by the
"environment” of the decision-maker (one-, two- or n-decision-makers).
In turn, the degree of complexity of the models will depend on 1) the
scope of the agenda; 2) the set of control variables; and 3) the degree
of programming required.

The degree of methodological sophistication of ex ante models
ranges from the simpler scoring models to more complex mathematical
programming and simulation models. Scoring models utilize weighted
multiple criteria for ranking .priorities. A panel of specialists
evaluate and rank various research programs using a predetermined set of
criteria. The programs are then funded according to their ranks until
the budget is exhausted. This procedure is commonly used to allocate

limited research funds to research, projects in the order of their

50
priorities. Other ex ante models that have been used in priority-setting
include: Minnesota Model, Pinstrup-Andersen and Franklin Model,
Cartwright Model, Castro and Schuh Model, Easter and Norton Model and
Atkinson-Bobis Model.8 In addition, benefit-cost analysis has been
used in various forms to select research priorities; Araji g£_gl (1978).

Fishel (1971), Davis et a1 (1986) and. Norton. et a1 (1987) present

 

examples. Most ex ante analyses employ the economic-surplus concept and
incorporate expert opinion to determine 'projected research impacts,
adoption rates and probabilities of success and provide estimates of the
economic efficiency and distributional implications of agricultural
research resource allocation.

Mathematical programming models rely on mathematical
Optimization to choose a research portfolio through maximizing a
multiple-goal objective function, given the research resource
constraints. This procedure may require more detailed information than
the weighted-criteria method, and usually selects an ”optimal" research
option rather than simply ranking them. Russell (1969) used this method
to maximize the contributions of a research program in the United
Kingdom to several goals, given budget and human resource constraints,
and different policy scenerios. Simulation models vary in their
construction and require extensive amounts of data and estimation of
mathematiCal relationships. Pinstrup-Andersen and Franklin (1977)
provides an example of the use of a simulation model as an appoach to

agricultural research resource allocation in developing countries.

 

8 See Schuh and Tollini (1979) for a discussion of these models.

51
The advantage of ex ante models compared to ex post is that they
provide a basis for decision-making with a focus on the future rather
than in the past. Hence, they provide a means of relating research
efforts explicitly to a set of goals. The disadvantage is that they are
based more on predictions about the future, can be both time-consuming

and costly, and rely on subjective judgement.

52

3.3 Rate of Returg Studies in Africa

Various studies in developed countries, Latin America and Asia
have indicated high rate of return to investment in agricultural
research. Many feasibility teams have drawn on these studies to justify
projects for donor assistance to NARS in Africa. But there is a need for
caution in interpreting these studies because most agricultural
institutions, including NARS, in Africa are at a relatively early stage
of institutional development (Eicher, 1990). There is now growing
evidence that a combination of misplaced and inappropriate projects,
weak management and financial accountability, and shifts of the research
agenda have resulted in a poor performance by the African NARS.

Surprisingly3 only four published studies on the payoff to
investments in agriculture have been documented for Africa: Norgaard
(1988). Abidogun (1982), Evenson (1987) and Schwartz g;_gl_(1989). By
contrast there are at least 25 for Asia and 60 for Latin America.

Schwartz et a1 (1989) estimated an average ROR of 60-80 percent to the

 

combined research and promotion programs for cowpea in Senegal. Noorgard
(1988) estimated the benefits and costs of research on biological
control of cassava mealybug in Ghana, deriving a benefit-cost ratio of
149:1 for the program. The average rate of return to four cotton
development projects in Burkina. Faso, Cote d'Ivoire and Togo were
estimated at were 31, 37, 11 and 41 percent (Daniels g£_gl, 1990).
Evenson (1987) used data from 24 countries in Africa, Asia and
Latin America to estimate the ROR to national research investments by

region and commodity groups. The ROR ranged from 30-40 percent for maize

53

and staple crops in Africa and maize in Latin America, to 60-70 percent
for maize and cereals in Latin America, cereals in Asia and staple crops
in Asia. He also studied the effects of research and extension
investments on productivity of various commodities in West and
East/Southern Africa. He found that national research investments had
significant production impacts cu: wheat, rice, cotton, sugar, cassava,
irish, and sweet potatoes; East/Southern .Africa showed larger wheat,
maize and groundnuts impacts than West Africa, but lower impacts for
potatoes, cotton, sugar, cassava, soybeans, beans and rice; the
International Agricultural Research Centre's (IARC) investments had an
impact on wheat, irish and sweet potatoes production in West and
East/Southern Africa, beans in East/Southern Africa and rice in West
Africa; and impacts of national research and, to a lesser extent, IARC
research on productivity are lower in West Africa than in other
developing countries, suggesting that many of the smaller NARS such as
those found in West Africa have little or no impact on productivity.

Studies using a descriptive non-ROR. approach to assess the
impact of agricultural research include an evaluation of the impact of
the thirteen Consultative Group of International Agricultural Research
(CGIAR) centres on world agriculture (Anderson, Herdt and Scobie, 1988).
The authors found that the CG-centres have had a large impact on NARS
through wheat and rice research but less impact in maize research. An
evaluation of USAID involvement in Kenyan maize research (Johnson gt_gl,
1979) indicated that great success is possible when public and private
sectors cooperate in technology development and diffusion. Research

funded by the government of Kenya, USAID and other donors led to the

54

development of a high-yielding hybrid maize -H6ll- in 1964 which had a
yield advantage of 40 ‘percent over the local synthetic. Subsequent
breeding generated sixteen hybrids and three composites for almost all
maize-growing regions of Kenya. Hybrid maize research in Zimbabwe began
in 1932 and took seventeen years to release its first hybrid, SR-l. But
the major breakthrough came eleven years later in 1960 when SR-52, a
high yielding long-season variety, was released. Smallholder area under
hybrids increased from 30 percent in 1979 to nearly 80 percent in 1986
(Rohrbach, 1988).

Despite Kenya's widely acclaimed success in agriculture since
independence, no rate of return study has been documented for agricul-
tural research in Kenya. Many evaluations of agricultural projects have
relied upon quantitative and qualitative indicators such as rates of
adoption of particular technologies, the extent of farmer participation
in a project, nutrition improvement, potential income generation, etc.
Indicator matrix-type of analysis (M81, 1990) and simple benefit-cost
analysis have been commonly used by donor analysts. In the Kenyan NARS,
most researchers use indicators such as technology diffusion rates, crop

area and yield increases as measures of research success.

4. CHAPTER FOUR

RATE OF RETURN: MAIZE IN KENYA

This study represents the first assessment of the impact of
agricultural research within KARI. Whereas the study aimed at estimating
the payoff to past investment in maize research, its broader objective
is to lay the groundwork for systematic evaluation of agricultural
research by Kenyan scientists. Agricultural institutions need to develop
the capacity to internalize such evaluations and to generate information
on the payoff to research to justify budget support to policy makers.
The overall aim is to develop a set of tools that can be used to aid
decision-making on resource allocation and research priorities.

The maize program in Kenya has been considered successful by
many' agriculturalists. But no one has ever attempted to quantify,
economically, this success. To date, the following indicators have been
used to measure the success: rate of adoption of improved maize
technology, area under hybrid seed, increase in maize output over the
years, and improved nutrition and cash-flow for the rural poor.

This study uses a production function approach to isolate the
impact of breeding and agronomic research on maize production from the
effects of maize extension and seed development, and to identify the
relationship between the production and expenditure variables.9 The

study focuses on the research component.

9The success of maize in Kenya is attributed to the maize breeding and '
agronomic research, extension and seed development programs.

55

56

4.1 The Sample

The data and information for this study were collected from
secondary sources in Kenya between October 1989 and March 1990. The
sources included the Ministry of Agriculture, Ministry of Research,
Science and Technology, Ministry of Planning and National Development,
Kenya Agricultural Research Institute, Kenya Seed Company, Kenya Green
Growers Cooperative Union, University of Nairobi, Egerton University,
Kenya National Farmers Union, Central Bank of Kenya and Central Bureau
of Statistics. International data sources included: the Food and
Agriculture Organization of the United Nations (FAO), the International
Maize and Wheat Improvement Centre (CIMMYT) and the International
Services for National Agricultural Research (ISNAR).

Digging up a reliable time-series data set is probably the
hardest task of carrying out research in many developing countries.
Collecting data for this study was not an exception. A lot of time was
spent extracting data from the archives and from loose, dusty files.

This study covered the 1955-88 period, which coincides with the
inception of a maize improvement program in Kenya in 1955. Data were
collected on maize production and research and development (R 8 D)
expenditure, as well as other micro- and macro-economic variables that
explain observed changes in production and consumption of maize over
time. Most of the data were aggregate because of unavailable or inade-
quate micro-level data.

Maize area, yield and output data were collected from the

Ministry of Agriculture (MOA) and KARI, and supplemented with FAO

57

estimates. Both recurrent and development estimates of expenditure on
agriculture were obtained from the MOA and the Treasury Department.
Using a time-series ratio of the number of maize breeders and
agronomists (with at least one university degree) to that of the total
crop researchers in the Kenyan research organization, an estimate of
Maize Research Expenditure (MRE) was derived from the crop research
expenditure.10 Maize Extension Expenditure (MEE) data was obtained from
gross non-research crop development expenditure using the same ratio of
maize researchers to the total crop researchers, assuming that
allocation of resources to maize extension activities follow the same
pattern as research resource allocation.

The ‘value of maize production was computed from the total
national output of maize and the official domeStic producer-price for
maize. This price is announced by the government at the beginning of
each crop season.11 Kenya is normally self-sufficient in maize in good-
weather years. During extensive drought-years like 1984, large ammounts
of basic food commodities, including maize, are imported. Because such
incidences are infrequent and Kenya's participation in the world maize-
grain market is minimal, the domestic producer price was used for this
study.

Data on several other variables related to maize production were
collected. The amount of annual sales of improved and hybrid seed were

obtained from the KSC office at Kitale. The price of seed was obtained

 

w Some studies use number of publications or proportion of crop area as
a proxy of the proportion of resources allocated.

11Although a parallel market exists, prices vary greatly with location
and time, making it difficult to measure and to obtain them for this study.

58

from seed stockists and the Kenya Green Growers Co-operative Union
(KGGCU). the largest wholesaler of maize seed. The volume of fertilizer
imported. was extracted from trade reports of the Central Bureau. of
Statistics (CBS). Prices of fertilizers were obtained from the FAO
production statistics. The CBS was also a valuable source of information
on the economy such as on demography, food supply and demand,
agricultural and economic growth, etc., extracted from various
government publications and reports.

The MRE included both the recurrent and development expenditures
on breeding and agronomic activities.12 The MEE estimated the recurrent
and. development expenditure (on extension activities, including soil
conservation and farm management, on maize crop. The cost of the
multiplication, maintenance research and distribution of improved/hybrid
seed by the KSC was estimated by using the market price of seed. This
assumed that the KSC, in which the government has the controlling
shares, does not extract abnormal or economic profits from farmers.”
Additional private funding for maize R 6 D was also included.“ All
expenditure and price variables were deflated by the Consumer Price

Index (CPI, 1971.100),

 

n This does not include personal emoluments and development expenditures
on the work done by Dr. Harrison and Dr. Allen in the 19505 and 19605.

13Griliches (1958) used a similar proxy to estimate the expenditure on
maize seed production in the United States.

1“Most funding from donor sources are channelled through the Department
of Treasury in the Ministry of Finance and included in the annual
appropriations.

59

4.2 Results and Discussion

Before calculating a rate of return (ROR), regression analyses
were carried out to derive the relationship between the production and
expenditure variables. Instead. of regressing on output, logarithmic
regressions on area and yield were done to gain greater insight into the
specific effects of explanatory variables on maize yield and area, and
thus on output:

OUTPUT = YIELD * AREA
Ln(0UTPUT) = Ln(YIELD) + Ln(AREA)
Where: OUTPUTsMaize output in tons;
YIELD=Average national maize yield in kg/ha;
AREA=Area under maize in Hectares; and
anNatural logarithm.
The endogenous variables were area and yield while the instrumental
variables included research and extension expenditure variables, seed
development variable, producer price for maize, an estimate of the
quantitity of fertilizer used on maize, and area and yield. The
inclusion of area and yield on the right-hand side of the equations
enabled determination of the secondary effects of area on yield. A seed
development variable, the volume of hybrid seed sales by the KSC, was
included to isolate the effect of seed development on maize production.

A two-stage least square (ZSLS) regression technique was used
because of a high likelihood that there was lack of statistical
independence between the random variables representing the equations'
errors and the explanatory variables. This is due to the interdependent

nature of the data-generation procedure. A residual plot of the

60

regression results indicated that some residuals were scattered. These
residuals were identified to correspond to unusual climatic and economic
seasons. Thus, a dummy variable (D) was introduced to capture the
exceptional agro-climatic and economic effects on maize production.

Actual data on fertilizer use on maize were not available.
Therefore, an estimate of fertilizer use was derived by assuming that
farmers were 1) using fertilizer only on hybrids; 2) using the blanket
recommendation rate of 123.5 kg/ha; and 3) applying only phosphatic
fertilizer to maize. The world. price (c.i.f. Mombasa) of Di-Ammonium
Phosphate (DAP) fertilizer was used ‘to estimate fertilizer cost to
farmers in this study.15 Since Kenya imports almost all of its
manufactured fertilizer requirements, it was more appropriate to use the
world rather than the local fertilizer prices which are mostly
subsidized.

Research usually has a long gestation period. It took a decade
(1955 to 1964) to develop and release the first maize hybrid in Kenya.
Subsequent releases took approximately eight years.16 Most of the maize
varieties released afterwards came from similar or close parentage of
the initial inbreds. It takes the Kenya Seed Company two to three years
after the release of a variety to make it available to farmers in
sufficient quantities. Therefore, a lag of ten years between the cost

and benefits of maize research was used in the study.

 

15Various on-farm investigations have identified that DAP is the most
common fertilizer used on hybrid maize.

M More time is required when the source of germplasm is a new
introduction.

61

The maize extension variable was unlagged because it was assumed
that extension messages delivered by extension agents to farmers is
usually most valuable in the current crop season. It was also assumed
that the farmers get the extension messages prior to cultivation so that
they can benefit from it during the same season. In later years, the
impact would include both extension information land the farmers'
experience.

The cost of seed to the society is computed as the volume of
improved/hybrid seed sales by KSC multiplied by the market price of seed
(based on assumptions discussed earlier in this chapter). The cost of
fertilizer is obtained by multiplying the volume of fertilizer used on
maize by the world market price of fertilizer. Both seed and fertilizer
are the common and major cost inputs in maize cultivation.

Ideally, the cost to a farmer of using improved/hybrid maize
should include the cost of seed, fertilizer, herbicide (if applied).
incremental labour cost, other capital costs such as credit and all
extra costs that are incurred as a result of using the new maize
technology. But since data on some of these inputs were lacking, it was
assumed that seed and fertilizer are the only extra costs of using the
improved maize technology. This is a reasonable assumption, especially
for small-scale farmers who usually adopt only these two inputs.

The calculation of ROR involved an assessment of the benefits
and the costs of breeding and agronomic research from 1955 to 1988. The
results of the regressions were used to identify the impact of the
research on maize production. A marginal rate of return was derived in

the following way:

62
Suppose V = Value of Maize Output; A = Maize Area; Y = Maize Yield;
R a Research Expenditure; P = Maize Price; and Q = Quantity

of Maize output, where V a Q * P.
Then, Ln V = P (Ln A + Ln Y), assuming P is constant. (i)
By differentiation,

d(LnV)/d(LnR) = P ((d(LnA)/d(LnR)) + (d(LnY)/d(LnR)), (ii)
assuming that maize research does not affect the maize price. This would
be the case for research in a tradable commodity by a country that is
too ”small” to affect the world market price of that commodity or a
country where there is strict government price control. Also, the same
case would apply to a scenerio where maize forms a small proportion of
the overall food demand by the people and/or substitutes are available.
Kenya fits into the price-control scenerio since maize prices are

controlled by the government.

Let LnV . v'; LnA . A'; LnY . Y‘; and LnR=R*.

Thus, dv’/dn* . p HdA‘IdR‘) + (dY'ldR’)! (iii)
But arm" . (awn/(am) . (R/Y)*(dY/dR) ' (iv)
and dA’/dn' s (dA/A)/(dR/R) =- (R/A)*(dA/dR) (v)

It follows that:
dY/dR . ( (Y/R) * (dY‘IdR') )

and dA/dR - ((A/R) * (ax/am)

But, dYfi/dRit is given by coefficient of research in the yield regression
equation (0.2529), and dA'ldR' by the coefficient of research in the

area regression equation (0.4922).

63

The change in maize output due to research, given by the sum of
dA‘ldR' and dY’/dR‘, is adjusted by the secondary effects of yield and
area, dY'ldA' and dAi/dR. respectively. These secondary negative effects
of yield and area were obtained from the coefficients of the regression
results. The resulting overall equation measures the change in the value
of maize production due to research. In order to get the marginal ROR,
benefits were lagged ten years and the following formula used:

ans (1n)10 = 8 (vi)
where MRE a Deflated maize research expenditure; 8 = Benefits due to
research; and r a marginal ROR.

In essence, the marginal ROR gives an estimate of how much
benefits are obtained by the society when research expenditures are
increased by one Kenyan Pound (K19.

The .study also computed an average ROR which measures the
averager benefits that accrue to all previous expenditure on maize
research. Because of the nature of the data and the fact that estimation
of an ROR is more accurate around the means of the variables, this study
used the mean values of research (R‘), area (A.) and yield (Yi) to derive
an average ROR. From the regression equations and using sample means of
all the explanatory variables, before (1955-64) and after (1965-88)
research impact on maize production, the changes in maize area and maize
yield attributed to research were.computed. The benefits due to research
were estimated as the difference between the value of maize before the
research impact and the value after research impact. Equation (vi) is
then applied on the average research expenditure and the benefits from

research to obtain the average rate of return. The average ROR is a

O

64
measure of the average benefit to the society achieved by investing all
the previous streams of research expenditure. The calculations of the
marginal ROR and average ROR are presented in Table 11(a) and 11(b) in
Appendix B.
The regression analysis results of the logarithmic functions of

yield and area are presented below:

1) YIELD 8 9.0478 - 0.5705AREA + 0.2529MRE10 + 0.0732MEE + 0.0763FERT
(2.545) (0.293) (0.187) (0.068) (0.071)

+ 0.1216HYBSALE + 0.2406DUM
(0.058) (0.055)

(F VALUE=7.455; PR08>F=0.0006; DF822; ADJ R-SQIO.6377)

2) AREA 8 8.4670 - 0.2623YIELD + 0.4922MRE10 + 0.0673MEE + 0.1013MPP
(1.467) (0.175) (0.084) (0.049) (0.110)

+ 0.0177HYBSALE + 0.1262DUM
(0.053) (0.057)

(F VALUE823.468; PROD>F-0.0001; DFIZZ; ADJ R-SQ=0.8597)

where: AREA 8 Maize area in he;
YIELD - Average national maize yield in kg/ha;
MEE I Deflated Maize Extension Expenditure in K£ ;
MRElO - Deflated Maize Research Expenditure in K15 lagged ten
years;
FERT I Fertilizer Imports in t;
MPP - Deflated annual Official Maize Producer Price in K£ per t;
HYBSALE - Annual Maize Seed Sales by KSC in t; and
DUM I Dummy variable.

The results indicate that expenditures on maize research, extension and
seed development increased maize yield, area and production. From the
coefficients of the regression equations, the impact of research on both

maize area and yield, and consequently on maize production, was greater

than that of extension and seed development. The impact of extension on

65
maize area was greater than that of seed development but the converse
was true for maize yield.

The use of improved/hybrid maize varieties and fertilizer has
contributed to an improvement in maize yield throughout the country. The
results of this study show that the availability of hybrid seed and
fertilizer led to increases in the maize yield, with the former having a
larger impact on yield. The producer price of maize also induced a maize
area expansion.

The effects of weather and the economy on maize production were
crudely captured by the dummy variable. The results indicate a large and
positive effect of favourable weather and economic conditions on both
area and yield, and hence on maize production. This suggests that during
adverse weather or economic problems, maize production falls whereas
maize production increases in periods of good ‘weather and economic
prosperity. Looking at the maize production data over the past three
decades, it is clear that good maize harvests coincide with good-
weather years and/or periods when the economy is performing well.

The regression analysis also indicated that an increase in maize
research expenditure (MRE) by one percent lead to an increase in yield
by 0.2529 percent and an area increase of 0.4922 percent after a decade.
But the change in area by 0.4922 percent resulted in a change in yield
by -0.2808 percent (i.e., 0.4922-* -0.5705); and the yield change by
0.2529 percent underestimates the area by 0.0663 percent (i.e., 0.2529 *
-0.2623). Thus, the secondary effect of area on yield was -0.2808 and
yield on area was -0.0663. Therefore, when adjusted, an increase in

research expenditure by one percent raised maize output by 0.3988

66
percent ten years later.

The marginal rate of return (ROR) on investment in maize
research from 1955 to 1988 was found to b2532.59w2122353225) This is a
high rate of return and in policy terms it means that an increase in
research expenditure by one Kenyan pound increased benefits to the

Kenyan society by between thirty-three and forty-seven Kenyan pounds.

The mean marginal rate of return was found to be 41 percent.

‘ 'W"WW“
r W‘

. ,- WW-_\\.
The average rate of return was estimated té be 68mpercent,)1.e.,
overall, each Kenyan pound invested in maize research between 1955 and

1988 yielded sixty-eight Kenyan pounds.

s. gamma FIVE

SUMMARY

Today, the agricultural sector provides nearly all of Kenya's
food requirements, more than a quarter of the GDP, over three-fifths of
the export earnings, and three-quarters of total employment. But the
agricultural sector is facing several contemporary challenges such as
feeding a rapidly growing population on less than twenty percent of the
country's land area. Whereas the country produced sufficient food
staples in the sixties and early seventies, there have been marked
imbalances between food production and demand since the mid-seventies.
These imbalances are basically caused by rapid population growth, the
growing demand for food and the vagaries of weather. Because of limited
supplies of arable land, Kenya has to rely primarily on yield increases
rather than area expansion for increased maize production in the years
ahead.

The development and adoption of improved maize varieties, one of
the great successes of Kenya's agricultural research and extension
program, led to a fivefold increase in maize output, doubling of yield
and near tripling of area in three decades since the inception of a
maize improvement program in 1955. This success is of strategic
importance to food policy because maize is the staple food for over 90
percent of the population and it.accounts for over 40 percent of the
total dietary intake. But projections indicate that maize supplies from
domestic production and/or imports will have to double in the next 15-20

years to meet future food demand.

67

68

The government of Kenya introduced a systematic maize breeding
program in 1955 and it took ten years to produce the first Kenyan
hybrid, H611, which had a 40 percent yield advantage over the best local
variety, KS II. The succeSs of H611 hybrid maize was based on excellent
cooperation between maize breeders and agronomists, an agressive seed
development program and an extensive promotion campaign by the extension
services. To date, the maize research program has released seventeen
hybrids, four composites and two synthetic maize varieties. These
varieties have a 30-55 percent on-station yield-advantage over the
traditional varieties and they are suited to a wide range of agro-
climatic conditions. More than half of these are still being supplied by
the Kenya Seed Company (KSC).

The KSC played a crucial role in the diffusion of maize hybrids.
Responding to a government request to undertake the production and
distribution of hybrid maize seed, the KSC expanded its service to both
the large? and small-scale farmers through an extensive seed development
and distribution network. This network improved smallholders‘ access to
seed and as a result, the KSC seed sales increased from a mere 4 tons in
1962/63 to 10,600 tons in 1975/76 and 21,800 tons in 1987/88, indicating
a rapid adoption of the hybrids. Data indicate that the greatest
increase in the area planted to hybrid seed was due to smallholder
adoption of seed.

Despite the availability of the new seed and complementary
inputs, the average maize'yield is low compared to the potential. In
1981, the average national yield was 25-50 percent of the on-station

yield. This is attributed to low adoption of the new maize technology,

69

particularly among smallholders. In many countries, maize output has
been increased through improved adoption of hybrid seed and
complementary inputs such as fertilizer and insecticides. The World Bank
(1984) contends that maize production could be increased in Kenya by 70-
100 percent through the expanded use of improved inputs and husbandry.
But intensification of maize production, especially among smallholders,
will require a concerted effort from the various agricultural support
services.

The government is deeply committed to invest in agricultural
research in order to overcome technological constraints on agricultural
production. Currently, agricultural research receives the largest share
of the total national budget for research. The budget on crop research
and development (R & D) accounted for an average of 43 percent of the
total agriculture expenditure during 1955-88. In real terms, the
government's crop R 6 D expenditure rose from an average of K£ 4.29
million in 1955-59 to about Kf 12 million in 1985-88. During the same
period, maize research expenditures increased about threefold in real
terms and nearly tenfold in nominal terms. Over the past 30 years,
government investment and foreign assistance to agricultural research
have made KARI one of the largest research establishments in Africa.

Agricultural research is an important means of raising
(agricultural productivity. But this requires costly investments with
long gestation periods. The growing pressure on scarce public resources
. has heightened the need to justify such investments vis a vis
-alternative public investments such as agricultural extension and

irrigation as well as non-agricultural investments. Tight research

70
budgets require NARS managers to improve their research priority-setting
and resource-allocation procedures.

Many NARS managers make their research-priority decisions on the
basis of past experience and intimate knowledge of national research
policies and goals. These sources are important but the identification
of the payoff to past research can provide supplementary information to
guide future research funding decisions. Several analytical techniques
are now available for the NARS managers to analyze the costs and returns
to various research programs and generate information which can be used
for priority-setting.

Numerous studies have documented a high rate of return to
investments in agricultural research throughout the world. Several
methods are available for assessing the payoff to research. They are
grouped into two major categories: ex post and ex ante approaches. The
ex post approach is used to determine the impact of past investments in
research while the ex ante approach estimates the impact of future
funding. The results of these studies have been used by the managers of
NARS to justify the need for continued political support for
agricultural research and improve the decision-making process with
respect to research resource allocation.

This study evaluated the payoff to investment in maize research
over the period 1955-88. The data and information used in the study were
collected from archival sources, personal interviews and secondary
sources in Kenya between October 1989 and March 1990. The study found
that the maize research program, in conjuction with an active

agricultural extension program and a seed delivery system contributed to

71
a doubling of the national maize yield, near tripling of the area under
maize and a fivefold increase in maize production over the 1955-88
period.

The results from regression analyses show that an increase in
Kenya's maize research expenditure by one percent increased maize
production by 0.40 percent a decade later. This may seem like a small
effect but considering the size of national maize output of over 2
million tons, then 0.40 percent is a substantial increase in production
attributed to research. The marginal rate of return (ROR) on investment
in maize research from 1955 to 1988 was found to be 33 to 47 percent.
This is a high rate of return and in policy terms it means that an
increase in research expenditure by one Kenyan pound increased benefits
to the Kenyan society by thirty-three to forty-seven Kenyan pounds. The
average rate of return was found to be sixty-eight percent; this means
that one Kenyan pound invested in maize research yielded sixty-eight
Kenyan pounds over the 1955-88 period. Whereas this study cannot
determine the appropriate level of future funding for maize research,
the high ROR suggests that Kenya's maize research program had been
underfunded in the past.

This study points up the crucial complementary role of the seed
industry, extension service and complementary government policies in
increasing maize production and contributing to the overall success of
the maize research program. Because of these institutional linkages,
further research should be carried out to identify the contribution of
the seed program, extension and government policy to maize production.

The focus of rate of return studies should move from this ex-post study

72
to an ex-ante study to evaluate the potential future benefits of maize
research in Kenya.

Further studies of factors influencing research productivity
could provide information (”I the constraints (”1 improving the
productivity of KARI. Several constraints require further research: lack
of adequate budget support for KARI, high personnel turnover, inadequate
and/or inadequately-trained manpower and the optimal size of scientists
and support staff. It is important that the managers of KARI and
government policy makers realize the need to train and maintain an
experienced cadre of scientists and provide them with working
incentives, competitive remunerations and the means to carry out their
research programs efficiently. The re-organization of KARI is expected
to help overcome some of these problems.

The urgency of increasing maize production in the wake of a
rising food demand compels KARI researchers to search for germplasm
throughout the world. After all, the initial success of the maize
hybrids in the mid-sixties started with the importation of maize seed
(Ec 573) from Ecuador. Therefore, KARI should develop and hone the
capacity to generate improved maize varieties through conventional
breeding, "intelligent“ borrowing of germplasm from other countries, and
biotechnology. But it takes vision, continuity of research staff and
stability of domestic financial support to develop an efficient capacity

to borrow, screen and adapt technology to micro-environments.

APPENDICES

APPENDIX A

Appendix A

Hectares

.(Millions)

 

73‘

Figure 4(a): Kenya: Maize Area

.8-

 

 

1960 ‘ie‘sé-s' ”1976‘ 1975 1980
Year

OOOOO

1 985

 

74

Appendix A [Continued]

Fioure 4(b): Kenya: Maize Yield
22

 

Tons/Ha

1.4- ./ ..

1;?va '

 

 

 

0.8TToIirrr; UUUUUU Uniforfffi f;
1955 1960 1965 1970 1975 1980 1985

Year

75

Appendix A [Continued]

Fiaure 4(c): Kenya: Maize Production

 

Metric Tons
(Millions)
-* 93
<9 I? 9'

.—I-
I

05-

 

 

 

. U I F

OFFOOO‘KTTU‘I—‘OIICT—OIIIE66“ 80'
1955 ' 1960 1965 1970 1975 1980 1985

Year

76

Appendix A [Continued]

Ficure 5: Kenya: Hybrid Maize Adoption

 

 

 

 

90
80- §
70- ‘
60- V
2 \ .
g 50 \Q \ ‘ \
g 40- §_ \\ §
30- .l
g ‘ i l
20‘ i \ §
\ .
. A §
0 ...... \

 

'197'0' '197'5

Year

1980

 

 

77

Appendix A [Continued]

Kenyan Pounds
(Millions)

160

Figgre 6: Kenya: MoA.Exnenditure

 

140-
120-
100'

m
C?

 

 

 
      

ls

 

O8.6“8';:U;I‘jFTCfiIUTfITI0869508600880
1 955 1 960 1965 1970 1 975 1 980 1 985

 

Year

 

->*- in Nominal Terms -Ar- in Real Terms

 

 

 

78

Appendix A [Continued]

Eigure 7: Kenya: Maize Expenditure
Estimated in Real Terms

 

2.5

 

"P

c—l.
0.‘

Kenyan Pounds
(Millions)
.1

.0
‘1‘

    

         

 

 

 

075:8; 8.5;;0888‘T18TT8008788'88'888
1955 1965 1970 1975 1980 1985

Year

° ieeé‘

 

—*<— Maize Research —i- Maize Extension

 

 

 

 

79

Appendix A [Continued]

Fioure 8: Kenya: Maize Price

 

120

 

110‘
100-

501

Kenyan Pound/Ton

 

 

 

195$ . 1960 7985'“ 19'76‘ ‘19‘73' '- ‘ 1980 ‘ ' 19235
Year

 

[—2- Nominal Price -!- Real Price

 

 

Appendix A [Continued]

80

Table 9: Kenya: NCPB Maize PgrcMe (in '000 Tons) by Province, 1966-88.

.7

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Paovmcc

YEAR R/Valley‘ Western Nyenza Eastern Central Coast TOTAL
1966/67 148.5 58.1 9.44 4.87 9.2 0.5 225.6
1967/68 106.7 92.2 18.98 5.51 9.7 0.1 228.2
1968.169 167.8 72.7 24.04 22.70 5.2 0.01 292.0
1969/70 178.8 75.2 15.51 - 7.68 8.5 - 275.1
1970/71 178.8 75.2 1.78 1.20 0.5 - 252.0
1971/72 217.1 57.8 15.50 15.95 11.9 - 818.8
1972/78' 290.8 100.2 86.07 85.14 19.8 - 481 0
1978/74 214.8 126.1 12.58 2.05 12.8 - 868.8
1974/75 284.8 150.6 21.09 8.89 84.0 - 448.9
1975/76 888.8 173.1 35.51 0.64 12.9 0.01 555.5
1976/77 270.0 171.8 60.58 48.82 21.6 2.4 570.2
1977/78 140.6 82.6 14.96 8.14 2.8 0.1 244.2
1978/79 154.8 51.3 4.21 8.89 8.5 - 226.7
1979/80 95.7 28.2 8.24 4.91 0.1 - 182.2
1980/81 269.7 60.8 31 40 0.84 0.4 - 382.6
1981/82 469.8 123.2 54.47 41.84 7.6 0.03 696.4
1982/83 487.6 96.0 50.95 88.29 9.8 - 627.1
1988/84 874.8 74.9 45.85 0.04 1.9 - 497.5
1984/85 288.8 115.8 14.71 10.03 0.9 0.05 879.8
1985/85 560.7 175.7 51.00 16.19 10.1 0.04 888.7
1986/87 ’ 544.9 118.1 45.06 1.93 5.8 0.08 718.8
1987/88 889.6 81.6 55.41 0.66 0.2 0.02 477.5
1988./89 467.8 69.1 - 22.28 85.55 28.8 0.5 624.0
Source: NCPB. Unpublished data. '

 

 

Note:

"-" means 'negliglble'.

 

81

 

 

wvvmamwx y ”nosnwacmma

_. mazcnacmm

on

 

  
  

I‘C.—OJ
fillolbr
.332S:o..:

  
   

 

      

2.3.-:0
ungnn E
rX—I—n—uF-us

   

     

v.0.C.-:nu v BA.- » 2.5:. p 2:16.
2.5.3.: 3:129-

3354:.-
. 52:5..- . 5!! o

 

 

 

.-
n
n. I '25.:-
. . 7 :55?“
5.. .5282:

O
1 ’5.
a E3:
.2919-
E

    

  

P—nnndu
guns—n:

 

  

ac:_._:

 

 

 

 

“E

:- - _9-hh

   

 

 

 

 

 

III.|I| 21:03: '0 i=3 3:96.
uI-IIIII man—E adds—:0- ?anoh v. .5235 :5-

qu t—vhas- 17:91:: hq;h
Ilnluliol. wig—E udncdnu-g rig ~uozn—o- U-h U—P—aep 22g 83:

E ya—SF Dal—05::- E29
EF. .53.: 55:91.32. Elf—oz...
5.! .55: EE 5.: E. E... Ecn. :52.

III-'1' Anni—E us—r‘u23 bung gn-

APPENDIX B

>VVmsapx m
Hman .o. > massmn< on wmnanam no >mnwocwnanmw xmmomnnz man axemamwoa. Homm no Home.

 

 

82

9i. 9...: . conga... .38.... :3». 283...}...2alg
.2992 :i asisnsati zfifiw 98<
an; Ex. r9388 .34. em) «:85 «aloe. mm 33*
033.3 .8. cm) .4285 $8.89... mm 2:3:
E8fl~1§o
9.3.55: 3.. .3 096..» >88».- .83» mm 8i 1.99.8. 238 :3 I... 383.8.
vaifie 6153 .
03.. 3.. 2.83 .3 p! 33* .898 MG .378: I .8! 8g .2 33.8. 2...! a. .893
v.23- !x. .45.. .3 03!.» an; .3élaggaufigioaggu9g
Efififiafiifig
m§B=IEP§E= .25 aflfi iii .mah. mm mi
28823 2:
954 a?
>5!» .3.
19:50 .3 a? 2.5. 33.8 mm .3..-8
2.53%. .34. E328- 28 88.8 Sift
25$ 38 3.2!! >83... .83.. mo 9.. 1.358 988.8
$832.! .a 0389 2.5.3252 mm m.8l§§21§§3$.§§.§2g
mo .9. lat-lilies;
_!£t .28 .w. 2111. .QUH. m6 nfl
0:5926a3883 .33 59 2831. .afiflw no 3%
a:52al&m5!fl..nu SF. Ilia mm Bi
fizxfiasfil
035: .25 59 “985. .6 oxiahlalizaz.
wastna. .
«ex 38 SS 538- 83.8 no .Salgggul ..3..2§%
389.9.
n33 .2§}18!!§c232w38!5I828882RBIE<Zt€fl
agigr

mocnom. moso<wnnwm .Hmmo..

83

>vvoaaux m Anoanhacmav

emvwm .o .noanwacma.

 

 

g. 9.5: 03-82 3.8 a. «as m :35.» en. a. .35 l 8392.»

ISQE :fi edisas&£4 217w 98‘

1.313: a. 9.. 38 5) .4935 .3»... mm ..$.8$§8s8§§la§1§3¢&¢3

Ei§333232£. £9a81§!280
.2 88..
E§:&&)28: .25 VBKB. )sfifit .sfie ma .Rilénaaatfit.
.3332. .nu .ms mfifiénffifis .w aflfixlsfiiutxtanu
Siiaflltfi. .

€3.85..- .3 ah. >1» 3... as! .3 2 mm 2%

2363: .g 5) «33.: «its. .8988 mm ui....$lgitu.=5

gillsasalfi.

<53 33 ex 398* 38.3 no 3.3.3

mg 32 5E >383. .83“ no .8:

:REosnp .Ra afflr .fiv .n$2 mw .QflzaiIéssazafifiégustafitflfiaunsafifilu
388filiiflfiZiilxiizizaiafifiksnaglifi
111x383?

r333 is. .022 .8. Imam-.8 5.... .839 m0 nuts: I... 5.3. .3356 883.8

9.83 a. a. .8~ 331.8 «Sun-8 .82”. m0 «33$

22.331. .8. $3...sz >333- .8..~8o mm .33 gigsaig .8782

3.8. Sift Ill-3:2; .8~.§
2E6 8¥2$
<93 .38.»
PEER R955.
cola 28.2:

«333mg .8. 2938.3. 3983. .896. m0 git-Biésggguggug
.matflflfiufinfisaggg

MEE! :8~ SS 22!: 3869 mu .niLSfiu.3‘32983Q8R326Iifiuluu383503.

.P?§&3385 QfifigtfiiaatoZ§§lnu838529€§rCtaoai

g§§§83§a§t§.§
388.35%...

84

>vvmsowx w .noanwacma.

emuwo .o .nonnwssmm.

 

 

”8% n§la 0858! 3%31 .Eid aaiflfitaalilaiiga
2.993 <3 38...... .32.. 2.8.... wt...
mice: 3.. .8. cm) «23... mo 8.8.8.. .I Inns: 3.. a 8......
,Sta!. 311233295
«can!» .8. 2...... >383. .82: no 8%.... I 3...: a... .2 225 38.3.. 8.. 8.. ”3...: .358
.15832y995¢821333288§:
.5... 3.. m. .2 .8. 0...... >33... .8... mo .8.
an... .8. S. .gniuanou 3.. .8. mm 8a
is .

3iingiu .8¢R§. Hi
M93233... .8. 956. as. .8... mm Bilgguatgao
@532! .nl mist: can .28 no 33.8TI55313813ta2u§a3315.§595
1.23 .8. S. 333. 1.8.3 mm 0.2. 8 3* I. 59.... 8.23.8.
2.8. a. I. .8 03... 2.. .8... mm 3*
$.33 .. I. .8 3......“ :3... .398 no 5......
r... 9908. .23 9... 25.2 53!. 8.3. mm 88
2.63:. .3 >38 08.9. 8...? an > 8.1.8.. it... .3. .

2286: 83..
VQSRnizfiunx .nl ex .aiaraya 1381 mm >Z§§¢R:stasr.
mama-lag? .8. S. 28...... .83. m0 >§§ga§§
4...... I... 038...... .8. S. 58!. 33.. mo .3
$.63! a. I. .8 all... a... .8... mm 8*
5.2852... .8. 93.318. 9.8 ..$...al§i§uto§§§g
033:2. 898§xu§3£l§rElnitaiifibiai:flfitu
lafii.

Nona—:3. a. E. .28 as... 68.8.. .8. 0. mm a:

85

>vvmaawx m Anoanwssmav

emUHm .o Aooanwscmov

 

 

98.. 9...... nine... 3...... «as». aggiaiilg
zbfiz :fi. asizasafia 221w mEQ
mifieeeu .na 93m 28 mm aizlaafltignfifiawsnba
@398:!& .93 «89:958 :31 .298 mm :9:.tutagutinaoasssaaaaISSgsEEia
.Enyh QSQHEJ «82in .mi.SSEBBBBEPPlbhfiihfifiwgafifiuacsll
SE. .Sa
«9. a. 9.. 38 038.. 9.5 .892 mm 8*
:335328 .28 .m> >339: .353_ .saiéxnfiaaifi:azx=
m$§23 Sflissgfiatésgntau
Din aiéSRnfiilifirazxs
«3.1.2.. «on... 19.8.. 8.38
9.889 .568... 8a.... 833 .883:
_B8iéxxfiaxifirafizs
no.3": a. a. .8: mafia: .963 09.38 .8..2 mm 8.3.88
no.3... 3.. 88 can...» «Seal: 823 mm 5.8a
9593:.» gqfilsug
.8 m 9.6: 38 F.- 538. 38.83 mm 3.3% I {9.1 i3 2 3.8... v. a It 2.1.5.8 a. 8.2%....» a.
_ B‘ggfliaiugglanntg.
IRS-.8 «3.3 It: .3 Fl 3 393
@333 I... 38 s...» 3.8 .922 no 38 l 98.“ 38......
,kxyfll
:93: a. 9.. 38 «5w.» «3.. vale .038 mm 28
6.203 I... mil... 38 3.3.1.8 g .83. no Ewan:
u3<l3)32& .38 .ESFSB 2331: .2é2 m0 :8:
mac 3.. g; 38 $338 $8.58 .832: mm m. It at} g u :82. g .38: .3... 88:3 3

...§38§.r§3€§1§3§§3
T*%§1.g§.§£§§§sfifl.

.1... E.

86

>vvosmwx m .noawwazmm.

enUHm .o .ooanwasom.

 

 

9.3.. 9.3 03-23 3.8 a. .38 a. g. .i. 1.3-3 loo-Iain
>53... <9... 33.82.59... 3...... m5.
.Weaau .3. SB Erma! .26$ mm aiii.
5...... 395 .292 8.x
«26 .Ru his zfifli- .8988 .8 Hfi.
9:873 .8. cm) >333. .938 m0 93.8!
r13... .8. Fa.) . >833. .298 no as p.958-
03335 38 9. tag .983 mm 33.2.8833? §§%IEQ~..§
H... :3... 2.8 Bio 23.8.2
$193: .8. :9 $85 353 m0 3* nut
mSSRE .38 SB >§H28 .2¢$ mo ti
minus .89 ”2:8 means .23... mm algzga‘ggiggfigp

87

>Uvmsawx m .noanwacmm.

emcwm .o Anoanwacmm.

98% .855

«as». 288% .68 a. 38.3 Incl-sin

 

2.992 «3. 3.8 a. .5353
9.5.3 .93 cm)

2... manta

mm?» .8. 29.1.3»

~.§m33§ 38. figzfiis

0.38: 83 >83.» 33.. €3.33:
:58 8% an... o
@333 .8“. r...»

9.... .2...

Vain: .8“. ca:-

Rx. .8336 Haas...

25.38 3: cm) .9. 8..

0.50 33 :9

9.... £3 .93 >62...»

a 933 3 0913.. .82 Eq 48...»...

28...! 3.. mg: .8... .3»

2630.3 .Sm ca:

«.393: 83 cm)

.5: 99.»...

IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII

OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO

OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO

$.38

filggglassggggua
.838...

Eggs-agage‘gag;
gigi.

OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO

OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO

..$.§lg§§§.9§§3§§§.3§§
ggaiagfin.

8%...fl I g 3.38:.

wanton-«8.1.3....Buasgw .38?
.33:

3*
a;
$8 .
0.!

88

>vvmaawx m .Ooanpscoa.

amawm Ho Anoanwscaa.

98.. as...

.33». Enoch-2.83! 83......
m!&

 

mo 8.: £§§1.§.§8§§3

9.x Big-ggaufiafigg

8:

..$
mgagagégagwu..§a
assane$ .

8.12%.... .3.5 133533.88...

..................................................................................................................................................................

2.9.x... <3. «£258....
9.3%: i... 33 CM)
3.9!...
.635... 8... cm) .9... z. nun-B.
9.533
«2.83 33 PB-
mzfil... 20...... 3.. :9
manta... 3.. cm;
12.... .3. 3.1.1.8
93....» 39.93.
23.5.. I... 3.. .8...
2.5..
:38... o. l. 3.. 9.8....
£5.33 8.. Cm) .9553.
.92.... a. a. 3.. 5....

mm

no
no

Enigaelgfifiigasgg
31.3.1618...

i.§..us1..9.
Baggagggiggéig
gonna-81.3.

99.83.895.18".

8: gaggsgiiigig
3.. §§t§§§£§§§
9.8
.3
.3
5
8n

Engsggigfip

88.3!ggagogmi9ni
.ufl..mfll.§%uflal§lm5§g%
nagging-88.

82...»:

3.8a

..$..~fl

3 .

£128§2~8§.§1§1.3§

as
..u§anmuaa.>..xapafi.§i§p§§

89

vvvoampx w .noanwacma.

amvpm Ho .noanwacma.

 

 

9.9. 9...! 08.3.... 3.81 and gneiaiilg
2.992 < o. 38.8 o. .335... >321. g
E 8.. an... .8. 8» >88... .82» .6 ~33...
v3 3.. slain :58. .8»... mm 2*
3.8.8.. 3.... 3.. 8) >38:- 39...» mm 2?“; I 83.8.. .8. 9...! 3.8.. .82.. .8! 3...... u...“
353..» 323. 3.! In. .32.». 2.x. .32...“ 8938.. la 1...! 26.
2.5.5.9.... fl... 3.. on!» 98.. .893. mm a;
3:586. 2......- .. 8*
@388 .8. :9 >83... .29.. mo zealarlngsg
mg I... 3.. Br 38 3.83 m0 3.88
20.3 3.83 .83.. 38...:
3...... 2.....5. .

20.8 o. I. 3.- 3.11.8 38 .89.... mo .38 In .32.; a. 9. 31B.
33.9. I... 3.. E g .893 mm .88..§
19.5.3 0.. .836 2?. .3

.ixaau .arau
9:. 9.2.3. 3.. r... 3.- 5) >333 .83... no nail-.938;

38.... _ 3.. .54. 2....- 3.. m0 malt-1.3.5.9383... in»: tin 8.5.?
eat;
:09 R... «can: 3.- 93..» 388.8 333 mm 88...:
$3. 3.. vii: >333. mm
3.5.3; .82... gtulgg
115E. .gsn. nabarusaiunsar.
m o...- I... .633 3.- 983.... 2....- .83. mm 3.8:
‘

02w 3... 5.. >33... 83.3 no 8....8... 5...... 3.9.3.... .82....

90

rmeaapx m .noanwacoa.

 

 

soap» .0 .noanpacma.
98.. 9:3 9!!! 3.8.. :3... gas-natal;
>523 .3. 38.8 a .39... 2.3... 93..
93:3: a. I. .2. cm» >383. .8289 no at I I sag-u 3.8.9
.835 8:...3. I83. our-u... a. gig 3.8.9 32.68. .2
8.88.9618 .325"...
.93. .3 I 8.33.3.8. .835. a... .83 .9 .II 3383.... .833:
In 8.2.6». 883.8
@3353 .8! I 80.39.33 .383:
29.8... B: Isa-8896.55. .839
38a... 8.. I Edge. .38...
.938: a... 33 8., >33... .23» mo 31.... 3.838.... 3.8 3...... SE... a 3......
«18.3 Sontag-sitaaaafiliiflfisifi
.¢1icp
r... 9.5. I... 33 :9 >88!- .83~ .8 8x333 IE8 883.8
025.8
5....- .. F a: 8.. Scan. .83. mm 3.8:.
39895. .28 uaanfziis 036 .898. mm Sfifii
.. 3. 38 9328.. 53.qu .83. mm 83$
2.. .8. :9 18:33.. .8938 mm tuggsamfiaifi.§§iaa
I583! ggnfl.a.$§§.&§i8o§.§§
93:83.
.5... .8. a... .59. 5.5.... .83. mm 5533218
.93.. ._ 3......
298.. A 3.3...
.998. 5.33.
fizzy. .:a.=a
...fl.. .3
0.2... a... .8. :9 5.8... 38.: 8 93.8... lg . 2.9... 389.93 3.8. .ififigs
$298.. 315522.333...an .§I.§Pas.§.8fl 38.
. a... a... .2 32. as. 3.. a...
:3.an .8. >55.» >383. .868 no §.§§§s§§. .

91

>vvmsnwx m .noanwacma.

amcwm .o .nosnwacmav

 

 

95- 08.3 98.95 38...; 3.2». gueuoaglg
>352 <3. 38...: 2 59.3.3 >35.“ «85
29.3 .8. cm) 89!: .89 no 932ml. l... 58... .8 .92. #93. 3.838 .8 33. .883:
3.5 Bibs. lasagivfiug
on: 3.8: I... ”.33 .2 .2.
$380 3?...fl3g8flénflt. .3.
0.3 H... .8. :9 08. 88.3 no .5323
19in; «:83 5.8!
38.3 .3
mint. .8. cm) :13 8: ma :3 gugcoaofisfnslanfiigu
138 2a finalize-82.55.3129
«£8! 3: gun. gm».
so 0:5 a. I. .8» n.8- iniu .33. mm 8:
«at; .898 g
M338 .8» 9.: >833. .83. mo Ba
3.90.3 .8 an: >838. .83. m not
2. Gain. 088 8*
is 8*
5...: 3.. 53.3.. .8». cm) 393.3 .96.: mo 3.8: I 39.8. 983.8
< 3391 a. t. .8» D... .23: 83.3 mm 2%.»:
g. .08.: gt
Na; 83 03.8 <93 .293 8 gala-.8 :88...-
>$o a. I. 88 can I)! But . 32.8 mm 5.8:
“:23 /II \\.
395 In)...» .3 93. 331.. .33. mm 8..
3.3.62 gEIm—Egglas 33.5
2.95.2 2... my... .8“. 33.: g .93 8 mm .888»!
2.0363 .3853» 33
.23.. .3 3...! E: .82. mm .Slgg

ill 5

92

Jbvasmwx w .0oanM0000.

emcwm Ho 2000000000.

0 2990059§999899§.§..9999990050080.8930096999399280.9>00
.00395980900909.3..9.flnn1 .1 Ln LP .09900000800009....0.)¢0900P)50920099999990090.3099..0..

0 {00.000699 009099089000 00.009.900.00.

 

 

 

 

 

 

a 09909990980930.0900.99000900093999.89899099999930939.200009.900

93909980009090.999999009999 99999092909999.3039;

99300009099000.0990.

9999999993999999050‘9

.914.90<.2.9.9...89..1. L. L1.11.-1 PP}. .. L P!90.9..0..990§308.

«0.09:0. ..000 .m0 $900009 .999901969903 020109000... «Pi-09.0800 r L1 L030.0.20099900.990?9009.
a». .000.- ..r . Lang-995999095090.

2.0.8.0090 .99....010. .3.96980999090960.95303399990?99999069039899.9939.
0.9-0099.09.

>090! 90m. ... 0...... ..00. .9900. .99990309903 00.902009009900099;— :§ .05.

5).). .00.. 500095083000. 99-. 0990099.! .. nu 10.00.99rt:;:999.900..9
0920099099.? 00. 90990999990939.0099

>00. >20..00..marl-9.0083099009999890....0090083099990009890900820159009000099900909
009.90.99.99

)0...) q 0.. .. a.) .99.. .. 0. 0.0.0.0. 0 0.00.0. 300.090.008.897 0 . ”04.000990... 903mg). 29);..520009
9099);.992.9.9>1)..8

)9??? ......2299000 ..< <90. :2 0900;090:0009 . . . 139908303229ln§§ 9.0.9.9).30.

 

>093... 0:99.09900000 :999 00090999369999.8999.
3.000.200.38.290959990999998990903392699909999
90.00.900.009.000.>999.999.9999.9999H§.§90§§9§09.099.00.030
$090.. 00280. 29... 09090: 660.9 20.9..
099.92.)..0309909990-909899393699999008

. 0000-3 .08.092909.099.0989.99§.>999908009<099989999.

09.9.9.0. 2.00.. «09.00.80.600.- 0300:9909. 0890999098090 .680 0.

00:99:. 2.809090090323059 :10009098 ....90 .9999! ...:9990190099039900900
09990099319099 .90890009..20 7:099:20 :9948 290 .099 90299209..

00.9.8.) .090 ..00n .99. .902... 0.09.9 gcqgéiasga.m898 99.000909000080009 0.09.9.0... .9902)!

92.90....0093000906911 - vis-P.”- . - 1PL1:.00.90.9 09:89:99.0...
«go—9930290990.

0.0090590050199530 ...0 :93900000008990.:09 990.99.... P .. L9...1..P Lu-L Linings.

 

93

>00030px w .0oanwacm0.
emupm Ho. 2000003000 .003002000.

 

 

0.00.30.90.900095000. 33:99.0 09009931893010.0358 9.999.099.8089?! 009.93.000.00-h P LL
3:59.90 9.0909. 00.093 .9829 090M002;

0.950..) 990002.99. .392980998090000 !085800090r9¢0390..02...03.03.09.990§
00.7039 .309 2.000039399131L . 11 P111101: 1 P 9990 0980090009 90930808003)« 00.05.80.329 9:00.035.

 

29.9....
00.91... 3:00.009 909809009930 2.0.002 99098992598999.9990;

92093.. ...00000990. 8003:099999!>99 :09; 098.090.. 0980300000.. 0.0.8.0990 2.00-930.009
.))m 00.0.9092? 0.029.090. 09.39.300.09.

000.9. a. ... 90). 0. 0. >09 .03. 09998.9)89500999880039 0.909 0200);.3: «silica! .8099. .8.

00 0..... m. ... <. ..930 0.0 ). 0. 0. >..-... .000. .99 00 8.93009 09.03909 00 0200);. 20923909 .0290 09.9980. 33);.82. 00093039. ... 0.909
0:03.563.

000?... :90 000.999.32.908- 0500.00.82.00 9:99;!09890009898830 2983.0 :89. 95:00. 0.83099.
9990:.00. 95.099999099909990.

000‘...” .03. g9:.99§3§.09§958§93§ .29-009085.929.-

 

 

 

 

000902.133 --.. .. .. rr inn-800.081.11.01 L1 -.1r.1.rh011.-L LL P 1.3.99. .000. 0..)98 00993899929090.
00.0000 3.00 0929080008 80939989 3:098:98 80.809.238.080 0993880000.. 930.. 9018.9.
9509.... ...0 03... :93989 9909.999031 P L. I. 3.26.00.

09.0.. 0. 0.90000 .2993: 09999993090900.9990 99999 99.9.9 0209990000890 200. .0. ..3
09.90.00 0:29.909... 998009.330 .9390 80909509900600. . 00.11. IL 11 1 .9932090019 29.80.000.05)!
«0.83”: 0..... 0..-03000 9.09.000 .909 50910109990989.9909899. 0... 9:99:39. ..293

.89.. 20.80.8009.-- . 11 . 99980099000019.6890 0.29.0820. 0. 90000200900 ...-09309200350

 

2.02.9090... ..0000 290.090.8200 99180390993995!299369999.90989999
«89090.0..000.90999099999.9990999.80989990000.9r0.598999909038000.

0.. 589... m. .000. 005.099.9939.... 80.50800! 93909890989201.8090 090.900.900.09.

9. 0.990. m. .03. 09995999990998.2393! 58995999999989 93099303889990

00.909.009.99, £909... .098.

9.99.. ...000 99990959099809.8299 888888
«89090.0 n. ..000 280.909.099.99 003.90 .29-30959939999098.9999; 90300800039293 009.0903.

000002898099.

2.09090. 0. .3 09998099989939.93909099909089899. 0099100815059. 0.902.009
«(0..-990.29.00.80)... .000280003938-30002 08:00390993.)30292§. «099380.090. 05.0.. 0903000200.
.90....900109.03..9999098993899893.809899909809!090.08399.“?L. L. L 3.0.0.03.

 

70.009 ... s. 900.299. .000. 0095.8 09099800930080 803.90.000.88. .9009. 0.....
. .9 ... a. ... 0. 0.0.3030. 900. .2. 0.2.9.. .03. «09.938 9.2.8.3 8890.0 2.99.9029: 00.89.00. 00830003....“ 20......)
.9009 ... m. 0.0... £02.80. 33. 00890... 9.0.0.0? 39.38.90 90 0.00.9200 99.000.099.890. ! 0080.0? 909.099.00.058! 90985.8. 0...

0900.. 9.0 0 .n. ...00. 0.8 9.009 2. 00.! 09....- .0930 0.0?

00000003 0 .000000000.

H0000 .ou ammonmanmm .000000000.

 

49.0.. n. 95 ..2-00. .009 5095339909 00900550990993.0903" .3050. 09.. 9050505043 09500. 05909030890058. .0..
00.0

400. 0. .000. 959010.039... 3595095098505955095530555955.5595; 9

400. 0. 0. 5350...; .000. 550505.5555I .0005 300.440.00.058; 009.998.058.555
90009..

20.9w 04.95). .08... .000. 55500059995090 55555.555555 .0. 0.510”
0599.. 05020.5. 98500950.

0.0058 .000. 00009035950009.0003“55550559590500.0302.
0.005 ~. .80. 055 55589009055500.0003. 50.0558 0080.40.08.03.
.55.. >. .03. .5559 5555553.5 55005495005385.

.5000. >. 0.0.. 0.000950.?0.005. 30355955085309.000I.805.00..£900650~=0. 09.950099380955554
90009..

39:04. 38:99.5 09.90.00.504 00.8.0009 5:20..
539.. ...8 0.0030099 58859035599545.9904 39880900 0991008.... ..800

.5934 ..095 :30. .0.:595 00.8 :54 0985.09 .5 53555533000591; .5599
59029350000105; 0.008.909.5550. 5309.

.5905... #0509520 ..0095950 .0000 .000 :55955550550589 0.8.
55>. 05.000.000.09 509899..- 902.8053 5:505 >989? 0559.». 8?;

5.90.0 ...>..080.> 00405950 .380. 5:040. 09.0000. 99.300099. 05.0.0.3 35900059550540. 09995055099098.0009
5.2.. :99. 00004358953 0990:5950... 5:09.

 

38;».0 5505995200890. a0!0£!61:192!:E98saxtiafaalsaa:ai{EISle36000883092§5:iiiz.

298:0- 9.:85 .:£&02:£S390.300850 0088508895..0 0:1:5880512 tiaklfitthzsli st!9§8:6&i€03:: .58
5339a(.:....82803939.060909096:>3!09E§38:£§§Ea.239120..92?

3539»£..0000 00:390150 08:501.:03193§0031230!31 31005000838592925

3039p 05!. $015.30atoltailsack0985008800:5012208250958;38:8019sailillieal.21001051383

.52994 5055.5..550505.5I955 .580 5.5558? .009 559955584 0009..

09.05.20 0009.0. 209.0. 5950....05 :05 05000939090335: 3000900 050555804 3505953550911. :rh L -
>0..5.....0 0.595... (109.0295 :354 50.08.

 

.00:495 45.530995950598504 :50. 0.0009590. 040000893905. 03.30?

03.3.5.0. 95:0.5.03.550900990 953 98.809909805590395 9.950.955.9904 09880900 0.590.... 0.

95

>000=QHX w .noannacma.

600H0 .o. 20n0000000 .nonnpacaov

 

 

.90:9.....8...9..0009.0....00.000..0. I. L L r39§38589§§§.
.999>....9.03....0.9..80..0393099389099.§9..309:350§.9§.09910>§0.
.999 >. ... I. .. 00.0. 9.0 2. m. 30.9.0. .80. 3.3.0.3.. .035. 3.0.0903 3.3.3.0... 3 80909.03 ...0 89.30.. .0.-9.0» .0959 002.00.
.999 >. ... 2. m. 39.09.02. 0. 009000. 38. 9.3.0.30 .023... .0 3.090.. 3.3030... 33960350 1300030.. .0.-9.0.. .0939 002.00.
.999 >....9.02.m.n3.9.0.309 m§.99.0.9090:3<9.39..3090.<330§.93>3.0999333.

(...0 .0998 930.....3. 00:28:88.0 0:980... 953000.908 3.000.. >..-0.8 9.9.390. >:§..9 89.910.830.03.
3.30.3000 9.0...9960. 0:30.000 3:90.30...- 9309... 500.0 ...-0 9:0. 09.. 2.00.20 .:£0..3930050>30

320900.93: :9830993 :90. «3.38. 099.09.00.90: 00963.0 89.000 05.02.000.300» .83)....53990003 08999300993393
5.309398 .0300( 00.098390 {00990900798033 .89.“:8500369913900933.

29.1.30. 0.09... 0.33.. ..03 .:009.9..8.0.§ 3.020.993.8903 989999.883 ggngnggg
(9.3008928. 3030005.. 08.2.

£03.50 .969». ass-9.203. 8§é§2§3§>999§999§¥xrz .. .3..o..
29.99.). .00.. 091359.800: . . L 1 50090901930. 30.93.0000 00009891998098. 30...... .390...
33.0.0. u.><0.0000009.o...600000..9L09.099000-0000009. §§§§§ 330930.93...
£000? “8:908:39; 98390989999308;
9.000.... ..80 000.050 90.939000303999090 3809.80.99.90. 30.330.380.299 .80.
2.00.5? ..03>.0 00:38.8 .0900.
2000... ..80. ..3.0.30 .:(3.0009¢000929§99.0300 889893.338998889938.
2005... ....0 0.33.:00..0.0..0.00..023.3095.2393.090..9...9.0.0.0..3.3-.08.... Egglgm 0983...... 0.8.
a. . L .

:8:.09&( ...: .9... 8:39.03 00.... u » .....JHHHH.-. 1 1. 5.300.000.8909; err)
3......

20139.0 22833899398888.8898 8888.9503989! 0.90.9.3
290099.. -.8.=I390909.990§309(83 >309 L . LU“..- L1.» L- I. 028.3900.
29.9r0 ..9 8. 50909.0 .:9.00300..9.0.0.09 .5989...- 909380. 29:93.00 m889§§0.:29.9r£ 053).).098

 

 

 

 

93‘.- .E 09200098903393... 8.. .9...-00.0993999095929399399098090.
29.9r0 0...a(( (..00098393 0.0039009 ..80 8990983829.:999938399 L. . Ln..-1?F Lu Iraqi. ..0...
9.093.989. 32.30.3339; a. n. .8. ...-.9. 09999309939909.0308. 9993333 50.0.93
00.30». 0.0.9.0.. (3290099.03.09.93.1930.808388983888598939893“29.3.
000. ..(. .000. 89.333989933929003. 3.089.030... 8209.20.00.90?
00.9.0... (>...88098309219090330‘9500293999333991933803.

 

 

099.999.98.03393 .0 28909999.“...9 930...... 0... ... :09-920. 03:90:33.0... 35320020893 .. L. “1909.033 09309
9.0203599098593390?! 3. >..-9.0 0.833913. (0. 29850500099930.8090);
33> >..0 03389.0»!0990 .8. 93899898938999 000.039.9308.; 09.99.088.093...

. OmOnUlriuinsg. :028’80 :a 3.

96

0000009x m .noanwacam.

606H0.o.00n81000:806.

0951.).036

210.10. .08 81.108318013008113 0108010105101 .0.0.0... 01308013363319? 3:011:88

.01.:18129 9.161..

0.90.020.-

0.....000110801 018311.519! !...._..11601610.01.§.308 1.6181396301110616.

 

1.00....0...:...5001.108. 018.311.0511.!11001Pn ML}. Li>183~1181111~f8.

1.00. 091.080.00.09 .8..101960113190988010t§!.110000r!§1380§016§1!>3008390 90.01.160.150.
3...001fi050.9.20...01.3

10013...

.1. 38. ...0 1.01.1 018060!88.I.01 9.0180010 21.1 5.1830131185181138. 806.0.

1630.... .. .001. 00.130.133.10... 011601000116. .9...-508030181901.

003.9>

...00. .10330031301363001300160059 601001.36..1.1!P..10.8>1>.O§.

«136;.8..:0¢11801..01310 018.011.0511.! ...6116..8?0 03.11.615.351! 13508.0.
«1.3... ?16.”..>«0n..0..3:!0001.0601831116801310.001010510100131..>30.81...1I11!?F ...« I. 09806....

8.81.0...)

 

«0.0.3 382686.383; :381030601120510006010511038313 91.6011 >18.

m0110..h.hh.61.«.21.090¢130.
8010.0 801101.!511005093 38.318905031818360 .9000 0006.19.
«000.0 ......169 ... 11000.3 0:05.098 :51081116083033.00010180 !00106.>0:1..10011>08§01 1.1.8.:88.
800.0. O. .s. 16 0.. .s. 9.0.003 .80. ...0 5...... 5.3.0.0... ! 01800115511. ! 201.011.138.10 .0 38.0.. ..>. n01.038- 00511. 0001. ..0....10..1r 20...
3116.361831000000.
«3a.... .80.>010001.100.01!015.10030..105511.:000519800 0.160.161.3000 «1.050.811.8132. 0.8.

3.:166 ..2031. 8. 08131135018519!513060¢6.38>11§116«1189

«16.n..16.«

.100. :85g::6:1.001n1.500.0111.090¢61 . '1rrfli....1:5.03.>318.5..111.!!0051:.61365§0810

 

01.30.... 16150.6...5 218.100.8301 :18 01.06.036.00 350.....01830151830. 5000:: 59600.30.

«000.9..—

....1309 01.018600001160000:.186011010 LII L L.0110006100010...0«60116.O1.10|0..1.0.000m018060«010868

 

00.3.0261. 5.38.03. 8.00.

«.0.... 6.0..
800108.

 

 

:21316960808:08.30001985010000013.EUI 1. .. LINN 20581605010500.3119). F . u
Gain-8.00.. .0.

«00.0.3... .8. 0:536:30... {Littlr1..!.01608..8..01l10.>18.318380!901609808.

«1613900

 

0600.61.49 (.218... .8. 801.365.136.510: 03:13363!135601018315510?a00 21.1w...

6.1.??.>..v1501600 98938015108039.8r 3051011830. 901.00... «1393130130110.
.16.> :.08..051n01601811. 0001501 81.0.1 000080300... :000108939130 5.160100

...-36.1491
98??!

.987
61.16.!

(0.0.0.10... ..

. s. 69.7.60.

00:60.0.

501.230....

01.10.00. .08.. 93109600018191. .00-011.1 30006101008191 1.111: ..00.....
6:016 60.0000. ...8060095800163500 1.0036 .«100813801613961801-110160.

>.. 3...; 3.3.1). «8.3.0.. 38.9836 .00.. .0.-1821:3205 9:10.. 86.51 15.003910186010035 «0118:8386...

8361.1133018193101nr0103
P16561503..00~.«001310.00«5108105.180!.021.01830. .50101050016651.>30.81.01311>1831181.1.8u

30.6.0.0..00816 0.11:1 . .rlci1!.r059§00«.13!nipl.n -w.-r!!: .:. 33:61.0. 51.081
0.10103? ?c«6>.. 113.1080. 1210.01.10. 0!.1. 901.300.2001.

97

>vvoaawx m .noanwacmav

amcwo Ho" wommnmaow Anoanwacmmv

 

$2.9 ... a. ... .553- .r. can 0. 0:0. 33.323932 finishing 55.38%..5830312601 Rug-Q mg 5.33.5. 0968
$9.5. 33 .

sale. ... P 91:. .8180! 33. 013.1383. 85.351.800.996: 36.8330. 561.536. iggiigqngkuzcu 20..

£651. 7 a. «9.02:. 9556:. 38. ~3531§.o§l.§5¢ao§ .:oogficotanogsnfig. nggnfiig
335.92.33.90.

£95 .03. 33. aggigsgzuggofigag >£§E§n§u§.§l§ 0.9 altar-i.
4.9.993— ... a. 293.3... 965 <§ .3». gcéfiigfigsgigs r60<31~30l9 5~§<§§98

33383998883. {.mggmkéggi.

30.5.9. 0. 9:. 95329.55 .3. ailrIIEQQ-igginigl .898 .§.£g3§g.§.1>98.m§§§ in.

68:}.

~02}... a.) .3». gaggiiiggasnfigggggig.

98

 

1 9° _. ' ’! ° 1 ° 1 '1! ° x

Egagsxshi_12§§:§§1
191_nezginel_nene_9£_39turn (MRR)

HBE_LK£1 BEEEEII§_LK£1 M38131

1955 79550
1955 85301
1957 88832
1958 99100
1959 89402
1950 91057
1951 95050
1952 98505
1953 94105
1954 119315
1955 123739 3359445 45.54
1955 135834 4205820 47.48
1957 143253 3452924 44.20
1958 158445 2824547 39.79
1959 194579 2597552 40.05
1970 250815 2354378 38.44
1971 203815 3534710 43.55
1972 213885 3528495 43.41
1973 204195 3080544 41.74
1974 185358 3539347 40.74
1975 215454 5453819 45.05
1975 234524 5573138 45.13
1977 239197 5849479 47.22
1978 298090 5721437 42.27
1979 255437 4825375 37.85
1980 215448 5851908 35.51
1981 185570 4858457 37.35
1982 143057 3824449 33.43
1983 105550 5784539 39.71
1984 92293 4542570 37.70
1985 108772 5035377 39.49
1985 74087 5355757 39.11
1987 172045 5159541 38.40
1233. 195514 55812251, 35.49

 

 

MRR(%) - [1-(8/uaz)1/1°)]*1oo
where B - Benetita: MRS - maize research expenditure

The mean (average) Marginal Rate of Return - 40.90

99

Table 11 (Continued)

1hl_Axsrags_Bats_9f_Bsturn

Using the regression equations on maize yield and area,
and mean values of the explanatory variables, the following
values are obtained:

Yo - 737.90 kg/ha: YR = 1385.78 kg/ha: A0 = 579752.34 ha:

AR 8 1066280.10 ha:
vo - P*(Ao * Y0) and vR = P*(AR * YR); Benefit (8): WR -
v0):
where P a 29.022 Kl'per ton.
Average ROR - [ 1-{(8/MRE)1/1°}]*100

= [ 1-{(28325544/157094.54)1/1°}]*100

I 68.11%
where Y0=Average Maize yield before research impact;
YR-Average Maize yield after research impact: Ito-Average
area under maize before research impact; ARahverage area
under maize after research impact: VO-Average value of maize
before research impact: VR-Average value of maize after
research impact: B-benefits of research or increase in the
value of maize production- due to research; ROR-rate of
return: MEE-Average maize research expenditure; and P is

average official producer price for maize.

 

BIBLIOGRAPHY

BIBLIOGRAPHY

Abidogun, A. 1982. Cocoa Research in Nigeria: An Ex-post Investment
Analysis. The Nigerian Journal of Econggic and Social Studies;

Ackello-Ogutu, C.A. and M.0. Odhiambo. 1986. Maize Production in Kenya:
A Study of Costs and Technological Constraints Associated with
Output Expansion. Report to the USAID. Nairobi.

Akino, M. and Y. Hayani. 1975. Efficiency and Equity in Public Research:
Rice Breeding in Japan's Economic Development. Agericgg Journal of
Agricultural Economics 57(1):1-10.

Allan, A.Y. 1969. A Review of Maize Agronony Research in Western Kenya.
Contribution to the Third Eastern African Cereals Research
Conference. Kenya.

Allan, A.Y. 1971. The Influence of Agronomic Factors on Maize Yields in
Western Kenya with Special Reference to the Time of Planting.
Ph.D. Dissertation, University of East Africa.

Allan, A.Y. 1980. Early Planting of Maize-Essential for High Yields.
Kenya Earner (March 1980).

Alston, J.M., G.H. Edwards. and J.W. Freebairn. 1988. Market
Distortions and Benefits from Research . AJerichn Journal of
Agriculturgl Economics 70: 281-88.

Anderson, J.R., R.W. Herdt, and G. Scobie. 1988. Science and Food: The
CGIAR and Its partners. The World Bank. Washington. D.C.

Araji, A.A., R.J. Sin, and R.L. Gardner. 1978. Return to Agricultural
Research and Extension Prograns: An Ex-Ante Approach. American
Journal of Agricglturgl Econoecis 60:964-68.

 

Blackie, M.J. 1989. ”Maize in Eastern and Southern Africa”. CIMMYT.
(Unpublished).

Bredahl, M. and W.L. Peterson. 1976. The Productivity and Allocation of

Research: U.S. Agricultural Experiment Stations. Agericgn Journal
of Agricultural Economics 58(4):684-92.

CIMMYT. 1987. 1986 CIMMYT World Maize Facts and Trends: The Economics of
Coneercial Maize Seed Production in Developing Countries. CIMMYT.
Mexico, D.F.

100

101

Daniels, L., J. Howard, M. Maredia, J. Oehmke, and R. Bernsten. 1990.
"Assessment of Agricultural Research: Ex-Post, Ex-Ante and Needed
Methodologies". Michigan State University. (Unpublished).

Davis, J.S., P. A. Oram, and J.G. Ryan. 1987. Assessment of Aggregate
Agricultural Research Priorities: An International Perspective.
ACIAR in Collaboration with IFPRI. Canberra. Australia.

Echeverria, R.G., G. Ferreira, and M. Dabezies. 1989. Returns to
Investment in the Generation and Transfer of Rice Technology in
Uruguay; Working Paper No. 30. International Services for
National Agricultural Research (ISNAR). The Hague, Netherlands.

Echeverria, R.G. Assessing the Impact of Agricultural Research. In
Methods for Diagnosing Resegrch System Constraints and Assessing
the Impact of Agricultural Research. V01.II. Edited by R.G.
Echeverria. ISNAR. The Hague, Netherlands.

Eicher, C.K. 1990. Building African Scientific Capacity for
Agricultural Development. Agricultural Economics 4:117-43.

Evenson, R.E. 1967. The Contribution of Agricultural Research to
Production. Journgl of Farm Economics 49: 1415-25.

 

Evenson, R.E. and Y. Kislev. 1975. Agrigltugl Research and
Prodgctivity. New Haven and London. Yale University Press.

Evenson, R.E., P.M. Flores, and Y. Hayami. 1976. Costs and Return to
Rice Research. Resource Paper No. 11. Conference on Economics of
New Rice Technology. International Rice Research Institute. Los
Banos, Phillipines. Decenber 1976. (Mimeographed).

Evenson, R.E. 1987. The International Agricultural Research .Centres:
Their Impacts on Spending for National Agricultural Research and
Extension. CGIAR, Washington, DC.

Fishel, W.L. 1971. The Minnesota Agricultural Resource Allocation
Information System and Experiment in Rgsource Allocation in
Agricultural Research. Edited by W.L. Fishel, University of
Minnesota Press, Minneapolis.

Food and Agriculture Organization. Production Yearbook. Various
issues.

Gerhart, J. 1975. The Diffusion of Hybrid Maize in Western Kenya.
Abridged version. CIMMYT, Mexico, D.F.

Griliches, Z. 1958. Research Costs and Social Returns: Hybrid Corn and
Related Innovations. Journal of Politicgl Economics 66:419-31.

102

Griliches, Z. 1964. Research Expenditure, Education and Aggregate
Agricultural Production Function. American Economic Review 54:961-
74.

Grobman, A., W. Salhuana, R. Sevilla, and P.C. Mangeldorf. 1961. Races
of Maize in Peru: Their Origin, Evolution and Classification.
Publication No. 915. Nat. Acad. of Sci., Washington.

Hallauer, A.R. and J. B. Miranda. 1988. Quantitative Genetigs in Maize
Breeding. 2nd edition.

Harrison, M.N. 1970. ”Maize Improvement in East Africa" in Crop
Improvement in East Africa edited by C.L.A. Leakey. Commonwealth
Agric. Bureaux. Farnham Royal.

Henin. R.A. 1981. ”The Characteristics and Development Implications of
Fast Growing Populations”. In Papers on the Kenygn, Econogy;
PerformanceL Problems and Policias. ed. Tony Killick. Heinemman
Education Books Ltd., Nairobi.

International Services for National Agricultural Research (ISNAR). 1981.
Kenya. National Agricultural Research System: A Report to the
Government of Kenya. The Hague, Netherlands.

International Services for National Agricultural Research (ISNAR).
1985a. Kenya Agricultural Research Strategy and Plan. V01. 1:
Organization and Structure. A report to the Government of Kenya.
The Hague, Netherlands.

International Services for National Agricultural Research (ISNAR).
1985b. Kenya Agricultural Research Strategy and Plan. Vol. 2:
Priorities and Programs. A report to the Governnent of Kenya. The
Hague, Netherlands.

International Services for National Agricultural Research (ISNAR). 1990.
Kenya: National Agricultural Research Project. Mid-Term Review. A
Report by the Joint Government of Kenya/ Donor Team. Produced by
ISNAR for the Government of Kenya. (August, 1990). The Hague,
Netherlands.

Jamieson, B.M. 1981. Resource Allocation to Agricultural Research in
Kenya, 1963-78. Ph.D. Dissertation. Toronto University.

Johnson, C.W., K.M. Byergo, P. Fleuret, E. Simmons, and G. Wasserman.
1979. KJitale Maize: The Limits of Success. Project Impact
Evaluation Report No. 2. Agency for International Development.
Washington, D.C.

Jugenheimer, R.W. 1958. Hybrid Maize Breeding and Seed Proggctiog; Food
and Agriculture Organization. Rome.

103
Kenya Seed Company. Annual Report. Various issues.
Kenya Seed Company. Unpublished Data. Kitale.

Kenya, Colony of the Protectorate of -. Crop Production Review.
Department of Agriculture. Various issues.

Kenya, The Government of -. 1966. Statistical Digest. Central Bureau of
Statistics. Ministry of Planning and National Development.
Government Printer. Nairobi.

Kenya, The Government of -. 1978. Development Plan 1979-83. Government
Printer. Nairobi.

Kenya, The Government of -. 1981. Sessional Paper No. 4 of 1981 on
National Food Policy. Government Printer, Nairobi.

Kenya, The Government of -. 1982. Statistical Abstracts. Central Bureau
of Statistics. Ministry’ of Planning and National Development.
Government Printer. Nairobi.

Kenya, The Government of -. 1983a. The Official Handbook, 1963-73.
Government Printer, Nairobi.

Kenya, The Government of -. 1983b. Development Plan 1984-88. Government
Printer, Nairobi.

Kenya, The Government of -. 1984. Ministry of Agriculture and Livestock
Development. Consultant Report on the Grain Silo Construction
Project. Definite Development Plan. Government Printer, Nairobi.

Kenya, The Government of-. 1986a. Sessional Paper No. 1 of 1986 on
Economic Management for Renewed Growth. Government Printer,
Nairobi.

Kenya, The Government of -. 1986b. Kenya National Agricultural Research
Project. Project Preparation Report. Ministry of Agriculture and
Livestock Development. Government Printer, Nairobi.

Kenya, The Government of -. 1988. Development Plan, 1989-93. Government
Printer, Nairobi.

Kenya, The Government of -. 1989. Economic Survey. Central Bureau of
Statistics. Ministry of Planning and National Development.
Government Printer, Nairobi.

Kenya, The Government of -. Customs and Excise Department, Ministry of
Finance. Annual Trade Reports, 1963-81. Government Printer,
Nairobi.

Kenya“ The Government of -. Development Estimates. Various issues.
Government Printer, Nairobi.

104

Kenya, The Government of -. Development Planning Division. Ministry of
Agriculture and Livestock Development. Various reports.

Kenya, The Government of -. Estimates of Recurrent Expenditures. Various
issues. Government Printer, Nairobi.

Kenya, The Government of -. Ministry of Agriculture. Annual Report.
Various issues.

Management Systems International (M81). 1990. The Agricultural Research
Intermediate Impact Indicator Matrix: A Synthesis for Managers. A
report presented to USAID. Washington, DC.

Makau, B.F. 1988. Survey on Private Sector Research and Development,
Resources and Activities in Kenya. National Council of Science and
Technology. Publication No. 26. Nairobi.

Meinertzhagen, R. 1957. Kenyg Diary 1902-1906. Oliver and Boyd.
Edinberg. London.

Miracle, M.P. 1966. Maize in Tropical Africa. University of Wisconsin
Press. Wisconsin, Madison.

Mwangi, M.W. 1980. Alternatives for Improving Production, Employment and
Income Distribution in Kenyan Agriculture. Discussion Paper No.
273. Institute of Development Studies. University of Nairobi.

National..Agricultural Research Centre (NARC), Kitale. 1990. National
Yield Performance Trials, 1966-88. Unpublished data. Kitale.

National Cereals and Produce Board. Annual Report. Various issues.
Nairobi.

National Cereals and Produce Board. Unpublished Data. Nairobi.

National Dryland Farming Research Centre (NDFRC), Katumani. 1986. Annual
Report.

National Plant Breeding Research Centre (NPBRC), Njoro. 1987. Annual
Report.

Ndegwa, C.M., and F. Ndambuki. 1988. Seed Production and Certification.
Kenya Seed Company. Unpublished Paper.

Ndegwa, C.M., N.K. Arap Tum, and F. Ndambuki. 1985. ”Kenya Seed Company:
Growing For The Future”. In To Feed Ourselveg: Proceedings of the

First EasternI Central gnd Southern Africa Regional Maize
Workshop. Lusaka, Zambia. March 10-17, 1985.

105

Ndegwa, C.M., and F. Ndambuki. 1989. Seed Production, Certification and
Testing in Kenya. A Paper Presented at the Third Eastern and
Southern Africa Regional Maize Workshop at PanAfric Hotel. 17-23rd
September. Nairobi.

Ndegwa, C.M. 1990. Personal Communication. Elgon Downs Farm, Endebess.
February 12th, 1990.

Ngugi, D.N. 1982. Agricultural Research to Increase Food Production.
Paper Presented at the Food Research Priorities Conference.
National Council of Science and ‘Technologyu June 2nd, 1982.
(Mimeographed).

Noorgard, R.B. 1988. The Biological Control of Cassava Mealybug in
Africa. American Journal of Agricultural Economics 70(2):366-71.

Norton, G.W. and P.G. Pardey. 1987. Priority-Setting; Mechanisms for
National Agricultural Research Systems: Present Experience and
Future Needs. Working Paper No. 7. ISNAR, The Hague, Netherlands.

Norton, G.W., V.G. Ganoza, and C. Pomareda. 1987. Potential Benefits to
Agricultural Research and Extension in Peru. Agericgn Journal of
Agricultural Economics 69:247-57.

Ochieng, J.A.W. 1988. "Three Decades of Maize Improvement in Kenya: What
Dividends?” NARC, Kitale. (Unpublished).

Odhiambo, M.O. 1988. Grain Market Development: Agricultural Policy
Assessment Study. Agriconsult Report to the USAID. Nairobi.

Odhiambo, M.O. and D.C. Wilcock. 1989. ”Reform of Maize Marketing in
Kenya” in Food Security Policies in the SADCC Region. Edited by M.
Rukuni, G. Mudimu and T.S. Jayne. Proceedings of the Fifth Annual

Conference on Food Security Research in Southern Africa. October
16-18, 1989.

Oehmke, J.F. 1988. The Calculation of Returns to Research in Distorted
Markets. Agricultural Economics 2:291-302.

Ongaro, W.A. 1988. Adoption of New Farming Technology: A Case Study of
Maize Production in Western Kenya. Economiska Studier, Utgivna Av
Nationalekonomiska Institutionen Handelshogskolan vid Gdteborgs
Universitet. No. 22. Sweden-

Paliwal, W.L., B. Gelaw, C. Nissly, A.M. Marimi, P.M. Ndambuki. and P.
Motanya. 1984. Project Proposal for an Expanded Maize Improvement
Program in Kenya. Submitted to the Ministry of Agriculture and
Livestock Development. Nairobi, Kenya.

Peterson, W.L. 1967. Return to Poultry Research in the United States.
Journal of Farm Economics 49:656-69.

106

Pinckney, T.C. 1988. Storage, Trade and Fhice Policy Under Production
Instability: Maize in Kenya. Report No. 71. IFPRI. Washington,
D.C.

Pinstrup-Andersen, P., N.R. de Londono, and E. Hoover. 1976. The Impact
of Increasing Food Supply on Human Nutrition: The Implications of
Commodity Priorities in Agricultural Research and Policy.
American Journal of Agricultural Economics 58(2):131-42.

Pinstrup-Andersen, P., D. Franklin. 1977. "A Systems Approach to
Agricultural Research Resource Allocation in Developing
Countries". In Resource Allocation and Productivity in National
and International Agricultural Research. Edited by T.M. Arndt,
D.G. Darylmple and V.W. Ruttan. University of Minnesota Press.
Minneapolis, Minnesota.

Pratt, D.J., P.J. Greenway, and M.D. Gwyne. 1966. A Classification of
the East African Rangeland. Journgl of Applied Ecology 3:369-82.

Rohrbach, D.D., 1988. The Growth of Smallholder Maize Production in
Zimbabwe:Causes and Implications for Food Security. Ph.D.
Dissertation, Michigan State University. (Unpublished)

Ruigu, G.H. 1985. The Impact of Collaboration the Between International
Agricultural Research System and the National Agricultural
Research System in Kenya. Report Prepared for the CGIAR Impact
Study.

Ruigu, C.M. 1988. The Kenya Seed Industry Revolution: Current Status
and Future Prospects. Paper Presented at a Seminar on Technology
Development and Changing Seed Supply Systems. Tilburg,
Netherlands, June 22nd, 1988.

Rundquist, F. 1989. Hybrid Maize Diffusion in Kenya: Policies, Diffusion
Patterns and Consequences. Case Studies for Central and South
Nyanza Provinces. CWK Gleerup.

Russell, E.J. 1969. Research and. Development Evaluation: The Total
Program Approach. Financial Executive N0. 234. London.

Ruthenberg, H. 1969. African Agricultural Production Development Policy
in Kenya, 1952-55. Afrika-Studien No. 10. Ifo-Institut fur
Wirtschaftsforschung. Munchen.

Ruttan, V.W. 1982. Agricultgral Research Policy. University of Minnesota
Press. Minneapolis.

Schluter, M. 1984. Constraints on Kenya's Food and Beverage Exports.
Research Report No. 44. IFPRI in Collaboration with IDS,
University of Nairobi.

107

Schuh, G.E. and H. Tollini, 1979. Costs and Benefits of Agricultural
Research: The State of Art. World Bank Staff Working Paper No.
360. The World Bank, Washington, D.C.

Schultz, T.W. 1953. The Economic Organization of Agriculture. McGraw-
Hill Publishers. New York.

Science. 1990. Editorial, No. 4971(249). p. 837. August 24th, 1990.

Schwartz, L.A., J. A. Sterns, J.F. Oehmke and R.D. Freed, 1989. Impact
Study of the Bean/Cowpea CRSP for Senegal. Michigan State
University. (Unpublished).

Tweeten, L.G. and F.K. Hines. 1965. Contribution of Agricultural
Productivity' to National Economic Growth. Agricultural Science
Review 3(2):40-5.

World Bank. 1984. Kenya Agricultural Sector Report. Report No. 4629-KE,
Washington, D.C.

"WHIWTIEHMETTMTET