Front Cover
 Title Page
 Table of Contents
 List of Tables
 List of Illustrations
 Background of the study area
 Response to fertilizer applica...
 Adoption and fertilizer use...
 Farmers' access to fertilizers
 Determinants of fertilizer use
 Fertilizer use and transition to...
 Conclusions and policy implica...
 Appendix 1: Lima ladder for eastern...
 Back Cover

Group Title: Research report
Title: Fertilizer use on smallholder farms in Eastern Province, Zambia
Full Citation
Permanent Link: http://ufdc.ufl.edu/UF00085437/00001
 Material Information
Title: Fertilizer use on smallholder farms in Eastern Province, Zambia
Series Title: Research report - International Food Policy Research Institute ; 94
Physical Description: 76 p. : ill., map ; 26 cm.
Language: English
Creator: Jha, D ( Dayanatha )
Hojjati, Behjat, 1954-
International Food Policy Research Institute
Publisher: International Food Policy Research Institute
Place of Publication: Washington D. C.
Publication Date: 1993
Copyright Date: 1993
Subject: Fertilizers -- Zambia -- Eastern Province   ( lcsh )
Farms, Small -- Zambia -- Eastern Province   ( lcsh )
Genre: bibliography   ( marcgt )
non-fiction   ( marcgt )
Spatial Coverage: Zambia
Statement of Responsibility: Dayanatha Jha, Behjat Hojjati.
Bibliography: Includes bibliographical references (p. 73-76).
 Record Information
Bibliographic ID: UF00085437
Volume ID: VID00001
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: oclc - 28417709
lccn - 93025675
isbn - 0896290972

Table of Contents
    Front Cover
        Page i
        Page ii
    Title Page
        Page 1
        Page 2
    Table of Contents
        Page 3
    List of Tables
        Page 4
        Page 5
    List of Illustrations
        Page 6
        Page 7
        Page 8
        Page 9
        Page 10
        Page 11
        Page 12
        Page 13
        Page 14
        Page 15
    Background of the study area
        Page 16
        Page 17
        Page 18
        Page 19
        Page 20
        Page 21
        Page 22
    Response to fertilizer application
        Page 23
        Page 24
        Page 25
        Page 26
        Page 27
        Page 28
        Page 29
        Page 30
        Page 31
    Adoption and fertilizer use practices
        Page 32
        Page 33
        Page 34
        Page 35
        Page 36
        Page 37
        Page 38
        Page 39
        Page 40
        Page 41
        Page 42
        Page 43
    Farmers' access to fertilizers
        Page 44
        Page 45
        Page 46
        Page 47
        Page 48
        Page 49
        Page 50
    Determinants of fertilizer use
        Page 51
        Page 52
        Page 53
        Page 54
        Page 55
        Page 56
        Page 57
        Page 58
        Page 59
        Page 60
    Fertilizer use and transition to commercial agriculture
        Page 61
        Page 62
        Page 63
        Page 64
        Page 65
    Conclusions and policy implications
        Page 66
        Page 67
        Page 68
    Appendix 1: Lima ladder for eastern province hoe cultivators
        Page 69
        Page 70
        Page 71
        Page 72
        Page 73
        Page 74
        Page 75
        Page 76
    Back Cover
        Page 77
        Page 78
Full Text


Dayanatha Jha
Behjat Hojjati

RESE1ARH 1 ri el F, I

IFPRI Research Reports
Publications Review Committee and Procedures

Gunvani Desai, Chairman
Ousmane Badiane Christopher Delgado
Romeo Bautista Maarten Immink
Joachim von Braun Nurul Islam
Barbara Rose (ex officio)

All manuscripts submitted for publication as IFPRI Research Reports undergo extensive
review. Prior to submission to the Publications Review Committee, manuscripts are
circulated informally among the author's colleagues and presented in a fonnal seminar.
Following submission to the Review Committee, four or five reviewers are selected. At
least two reviewers are identified from outside IFPRI, one from inside IFPRI, and one
from the Review Committee. Reviewers are chosen for their expertise in the manuscript's
subject matter and methodology and, when applicable, their familiarity with the country
setting. The author responds in writing to the reviewers' comments and resubmits the
manuscript to the Re% iew Committee after making any necessary rev visions The Rev lew
Committee then makes its recommendation on publication of the manuscript to the
Director General of IFPRI. With the Director General's approval, the manuscript becomes
part of the IFPRI Research Report series.

IFPRI Board of Trustees

Gerry Helleiner
Chairman, Canada

Sjarifuddin Baharsjah

David E. Bell

Henri Carsalade

Anna Ferro-Luzzi

Ibrahim Saad Ahmed Hagrass

Yujiro Hayami

James Charles Ingram

Dhanna Kumar

Harris Mutio Mule

Abdoulaye Sawadogo
C6te d'lvoire

Nicholas H. Stem
United Kingdom

M. Sveduzzaman

Per Pinstrup-Andersen
Director General
Ex Officio, Denmark



Dayanatha Jha
Behjat Hojjati

Research Report 94
International Food Policy Research Institute
Washington, D.C.

Copyright 1993 International Food Policy
Research Institute.
All rights reserved. Sections of this report may be
reproduced without the express permission of but
with acknowledgment to the International Food
Policy Research Institute.

Library of Congress Cataloging-
in-Publication Data
Jha, D. (Dayanatha)
Fertilizer use on smallholder farms in Eastern
Province, Zambia / Dayanatha Jha and Behjat
p. cm. (Research report ; 94)
Includes bibliographical references.
ISBN 0-89629-097-2
1. Fertilizers-Zambia-Eastern Province.
2. Farms, Small-Zambia-Eastern Province.
I. Hojjati, Behjat, 1954- II. International Food
Policy Research Institute. m. Title. IV. Series: Re-
search report (International Food Policy Research
Institute); 94.

S633.5.Z33J43 1993



1. Summary 9
2. Introduction 12
3. Background of the Study Area 16
4. Response to Fertilizer Application 23
5. Adoption and Fertilizer Use Practices 32
6. Farmers' Access to Fertilizers 44
7. Determinants of Fertilizer Use 51
8. Fertilizer Use and Transition to Commercial Agriculture 61
9. Conclusions and Policy Implications 66
Appendix 1: Lima Ladder for Eastern Province Hoe Cultivators 69
Appendix 2: Supplementary Tables 70
Bibliography 73


1. Fertilizer sales in Eastern Province and share of total sales,
1981-89 17
2. Characteristics of agroecological zones in Eastern Province 20
3. Fertilizer use recommendations for major crops, 1982/83 and
1986 24
4. Weather conditions and maize yield levels at Msekera Re-
search Station, Zambia, 1982/83-1986/87 26
5. Average response of local and hybrid maize to fertilizer
application per hectare at Msekera Research Station,
1983/84-1985/86 26
6. Average response and value-cost ratios for recommended
levels of fertilizer application in the plateau region 27
7. Fertilizer use on smallholder farms in Eastern Province 32
8. Allocation of area and fertilizer among crops on smallholder
farms in the Eastern Province plateau region 34
9. Allocation of fertilizer among crops for fertilizer users, East-
ern Province 36
10. Probit regression explaining fertilizer use on local maize
(probit maximum likelihood estimates) 37
11. Percent of area fertilized and rate of application per fertilized
hectare for important crops in the plateau region 39
12. Rates of nutrient application per fertilized hectare for major
crops by agroecological zones in the plateau region 40
13. Use of plant nutrients, plateau region 41
14. Fertilizer use by farm size, plateau region 42
15. Use of different kinds of fertilizers, plateau region 45
16. Monthly fertilizer purchase patterns of farmers 46
17. Credit and fertilizer use, 1985/86 48
18. Average distance and mode of transport used for fertilizer
purchases, 1985/86 49
19. Definitions of variables in the fertilizer use regressions 53

20. Determinants of fertilizer use and intensity 56
21. Plot-level (tobit) analysis of determinants of fertilizer appli-
cation rates 58
22. Mean values of variables in each fertilizer user category 62
23. Results of an ordered probit regression explaining transition
to commercial agriculture 62
24. Location of sample households 70
25. Marketed production of maize, groundnuts, and cotton in
Eastern Province, 1978-87 70
26. Adaptive Research Planning Team (ARPT) trials, 1982/83-
1987/88 71
27. Fertilizer and maize prices in Zambia, 1980/81 to 1989/90 72


1. Fertilizer consumption in Zambia, 1969-88 17
2. Agroecological zones and agricultural districts, Eastern Prov-
ince, Zambia 19
3. Fertilizer-maize prices in Zambia and Malawi 29
4. Lima ladder 69


In Sub-Saharan Africa, which so far has benefited little from the green revolution,
the adoption of high-yielding maize has great potential for closing the gap between
food demand and supply. To bring about this transformation, fertilizer is essential for
realizing the yield potential of hybrid maize while sustaining the fertility of Africa's
fragile land. This study of Eastern Province, Zambia, shows that use of fertilizer on
traditional varieties can also be a catalyst for agricultural growth.
This work is part of an extensive body of research on adoption of new agricultural
technology carried out by IFPRI in Asia and Africa. The study was undertaken in
collaboration with several Zambian institutions, including the Rural Development
Studies Bureau (University of Zambia), the National Food and Nutrition Commis-
sion, and the Eastern Province Agricultural Development Project (both of the govern-
ment of the Republic of Zambia). It was funded by the Swiss Development Coopera-
The relationship between technological change and government policy has al-
ways been an important part of IFPRI's research program.IFPRI's ongoing research
on fertilizer use is part of an effort to devise workable policies for translating new
technology into rapid agricultural growth and sustainable development that benefit all
segments of society, but particularly the poor. It also relates to other IFPRI research
on input market reforms, which examines ways to improve access of the poor to
inputs such as fertilizer through efficient pricing and distribution policies.

Per Pinstrup-Andersen
Director General


We would like to thank our colleagues Raisuddin Ahmed, Gunvant Desai, Vasant
Gandhi, and John Mellor for their comments on earlier versions of this manuscript.
Careful and critical reviews by Ousmane Badiane, Bruce Stone, Carlos Baanante, and
an anonymous referee helped us in clarifying and improving the manuscript further.
We also owe a great deal to the members of the Zambia research team: Sudhir
Wanmali, Rafael Celis, Shubh Kumar, Neal Bliven, and the late Emmanuel Shula.



This study, conducted in Eastern Province, Zambia, looks at farmer's practices
regarding use of fertilizer and analyzes the role of fertilizers in the transition from
subsistence farming to a more commercialized agriculture. The findings here are also
relevant for similar agroecological zones in central and southern Africa.
The study is based on data collected from 330 households in Eastern Province by
the International Food Policy Research Institute (IFPRI) in collaboration with Zam-
bian research institutions during 1985/86 and on data provided by the Department of
Agriculture and the Adaptive Research Planning Team, which conducted field trials
for the Eastern Province Agricultural Development Project, funded by the World
Bank, during 1983-87.
Eastern Province has recorded impressive gains in agriculture since the late
1970s and has contributed increasingly to national food stocks, particularly of maize.
Most of this growth has come from the plateau region, which covers nearly half of the
province and has 80 percent of the population. The other region of the province, the
Luangwa Valley, is heavily infested with tsetse flies and is lacking in infrastructure
and support services. Since fertilizer use was negligible in the valley region-only 6
percent of the sample households used fertilizers-the analysis in this report is
confined to the plateau sample. About 96 percent of the farms in the province are less
than 10 hectares, although it is a land-surplus area. Maize is the dominant crop,
accounting for more than 80 percent of the cultivated area in the plateau and about 60
percent in the valley. Groundnuts are the other major crop in the plateau. Fertilizer
use in Eastern Province began in the late 1960s, mainly on hybrid maize. Belying the
national trend, fertilizer use continued to grow impressively through the mid-1980s
when the province accounted for nearly a quarter of national consumption of fertilizer.
Experiments conducted at the Msekera Agricultural Research Station and on
farmers' fields in the province during 1983-87 indicated that physical responses to
fertilizer application in the plateau region were high, ranging from 16 to 22 kilograms
of grain per kilogram of plant nutrients for hybrid maize and from 9 to 15 kilograms
for local maize at recommended levels. Cotton and beans also responded well to
fertilizer in tests, but farmers use fertilizer almost exclusively on maize. Interactions
between fertilizer and different agronomic practices such as time of planting, weed-
ing, crop rotation, and time of fertilizer application were established, findings that
have been explicitly incorporated in extension messages in the province.
The high physical response to fertilizer in the plateau was supported by a
favorable price environment during the 1980s. Fertilizer prices rose sharply after
1980, but maize prices kept pace, and relative fertilizer prices remained favorable
until 1989/90, when real prices increased by 81 percent.
Data collected in the IFPRI survey show that the use of fertilizer has been widely
adopted by smallholders in the plateau region. About 67 percent of the sample
households used fertilizers, and more than 55 percent of the cropped area was

fertilized at an average rate of 96 kilograms of total plant nutrients (N + P205 + K2O).
The average rate of application per hectare of cultivated land was about 53 kilo-
grams-more than three times the national average. These numbers conceal substan-
tial variation in the plateau, however, where use varied from 20 to 100 percent among
Hybrid maize accounted for 54 percent of the total fertilizer used; the rest was
allocated to local maize. About 48 percent of fertilizer was allocated to mixed stands
of local and hybrid maize varieties. However, fertilizer use recommendations are
based on pure stands, because little research has been done on responses of mixed
During the survey year, 49 percent of the fertilizer users applied fertilizer to local
maize only. This goes against the traditional wisdom regarding diffusion, which holds
that in a cross-section of households there will be some new fertilizer users who
fertilize only hybrid maize and others with more experience with fertilizers who will
use it on both hybrid and local varieties. A probit model designed to identify the
distinguishing characteristics of farmers who only fertilize local maize shows that
smaller farmers, households headed by females, older farmers, and farmers with
fewer dependents are more likely to apply it only to local maize. This implies that
even those farmers who could not (because of limited labor or capital) or did not (by
preference) grow hybrid maize found it advantageous to use fertilizer. Apparently,
knowledge regarding fertilizers is so widespread that farmers are aware that it pays to
fertilize local maize.
Rates of fertilizer application varied across different agro-ecological zones
within the plateau. In general, fertilizer users tended to use all three major nutrients,
although in some locations, inadequate supplies of mixed fertilizer forced farmers to
use only a top dressing (nitrogen) fertilizer such as urea. Substitutions led to diver-
gences from recommended levels and imbalances in the use of nitrogen and phospho-
rous in some cases.
Fertilizer use parameters-adoption (use), crop area fertilized, and rates of appli-
cation-did not vary significantly across farm size categories. Only on commercial-
ized farms of more than 5 hectares were the indicators sharply higher. However, the
pattern of fertilizer allocation between crops did change as farm area increased,
shifting from local maize to hybrid maize.
Constraints in the fertilizer distribution system were inferred from analysis of the
fertilizer purchase patterns of sample households. Less than 3 percent of the total
quantity of mixed fertilizers used was of the recommended X mixture (20-10-5 NPK).
In some locations where even the D mixture (10-20-10) was not available, farmers
were forced to use only nitrogenous fertilizers. Lack of the recommended fertilizer
mixture raised farmers' fertilizer costs by about 16 percent and created inefficiency
in timing of application. Availability of fertilizers at the right time also depended on
physical access; farmers had to travel from 0.7 to 9 kilometers to obtain fertilizers,
depending on location.
Only 15 percent of the farmers used institutional credit, although more than 66
percent used fertilizers. Among fertilizer users, more than 75 percent did not use
credit during the 1985/86 season.
Farmers' fertilizer use behavior was analyzed through an econometric model that
considers two simultaneous decisions-whether to use fertilizers and how much to
apply. The model postulates that these decisions are influenced by a common set of

personal variables (age, education, sex, and subsistence pressure), resource endow-
ment variables (family size, cultivated area, oxen use, crop sales, and hybrid use),
institutional access variables (cooperatives, extension, markets, and fertilizer sup-
ply), and location variables (agroecological).
The results on adoption indicate that younger farmers were more likely to use
fertilizers, as were those who had better access to capital either through farm sales or
through membership in cooperatives. Fertilizer supply conditions and access to
markets affected this decision significantly. The results did not support the view that
households headed by women lagged behind in use of fertilizer, but they were
reluctant to grow hybrid maize and therefore more likely to use fertilizers on local
maize only. The extension advice and education variables were not found to be
significant, probably because knowledge about fertilizer use was already widespread.
As farmers gained familiarity with fertilizers and began making more complex
decisions on fertilizer levels to be used, the study finds that education and extension
advice assumed greater importance.
Membership in a cooperative and access to markets-variables found to be
important in the decision to use fertilizers-were not significant in the decision on
how much fertilizer to use. The only market-related variable that mattered was the
quality of input distribution in the area.
Farmers' fertilizer allocation behavior was also examined to gain insight on the
process of transition from subsistence to commercial farming. It was hypothesized
that fertilizer use on local maize triggered the transition process in Eastern Province.
The resulting surplus was used to expand cultivated area and area under hybrid maize,
culminating in larger farms and commercial maize production. Empirical evidence
based on an ordered-probit model shows that education, credit, and improved access
to infrastructure facilitated this process and inefficiencies in fertilizer distribution
dampened it. Older farmers and farm households headed by women found it difficult
to make this transition. Availability of land, existence of a labor market, possibility
of using oxen for cultivation, and a favorable price and market environment all
contributed to the process.
Some important policy conclusions emerge from these findings. First, regions of
high physical response, like the Eastern Province plateau, can sustain rapid growth in
fertilizer use and agricultural production. It is, therefore, more efficient to target
scarce fertilizer and supporting infrastructure to these areas.
Second, Zambia's favorable incentive environment made it profitable to use
fertilizers on maize. A recent deterioration in relative fertilizer prices may have an
adverse effect, especially because fertilizer use is concentrated on only one crop. In
the long run, it will be necessary to diversify the crop base.
Third, public investments in schooling, infrastructure, and credit institutions were
important determinants of fertilizer use and commercialization. The research and
extension systems have also contributed. And, in Zambia, the state has procured all
the maize surpluses that farmers produced. These findings emphasize the critical role
of public investments in modernizing African agriculture.
Finally, the study shows that fertilizer use can trigger the process of transition of
subsistence agriculture to commercialized farming. Under favorable circumstances,
yield-increasing technologies offer viable opportunities even in land-surplus regions.
Development strategies for regions with poor endowments, like the Luangwa Valley
in Eastern Province, require more research on their agriculture and farming systems.



The crucial role of chemical fertilizers in Africa's agricultural future is well
recognized (FAO 1981, 1986; Mellor, Delgado, and Blackie 1987; Lele, Christiansen,
and Kadiresan 1989). Evolving population-food imbalances underscore the critical
role of new and improved agricultural technologies, and fertilizers constitute a key
component of this strategy. Actual performance regarding the use and diffusion of
chemical fertilizers in different countries of Sub-Saharan Africa has, however, been
generally disappointing. Levels of fertilizer use continue to be low. Growth in
fertilizer consumption has faltered in Africa, lagging behind that for developing
countries as a whole (Desai and Gandhi 1989).
In summarizing the fertilizer-related experiences of selected countries in Africa,
Lele, Christiansen, and Kadiresan (1989, 6) identify supply and demand constraints
that inhibit growth in fertilizer consumption:1 "The underutilization of fertilizer
makes fertilizer pricing, subsidy, and distribution policy, together with the alleviation
of other technological and institutional constraints, one of the most pressing issues in
the modernization of African smallholder agriculture."
From another perspective, researchers have emphasized the need to examine the
nature of farming systems before deciding on appropriate technological interven-
tions. Large parts of the continent, it is argued, are characterized by land-abundant
forest or bush-fallow systems, where yield-increasing technologies like fertilizers
have limited relevance (Binswanger and Pingali 1988). This perception helps explain
the observed low use of biochemical technologies for food crop production. This
system constraint notwithstanding, active public policy with regard to regional target-
ing, infrastructure, research and extension, prices, marketing, and so forth can, even
in such situations, promote these technologies and help achieve rapid production and
income growth (Binswanger and Pingali 1988; Lele and Stone 1989).
Micro studies in different agroeconomic settings are, therefore, essential to
understanding the variations that characterize the African fertilizer scene. Scores of
such studies have helped rationalize and formulate policies in Asia, for example,
where fertilizer use has grown dramatically over the last two decades. Similar efforts
have been lacking in the African context. In fact, many more studies are needed in
view of the large variability in farming systems, economic environment, and infra-
structure conditions over the continent.
This research reports the results of one such study conducted in the Eastern
Province of Zambia-a country endowed with abundant land resources. It is based on
data collected from 330 smallholder households during the 1985/86 agricultural year.
Fertilizer consumption in Zambia has remained stagnant over the last 10 years,

'Fertilizer supply includes fertilizer imports, aid, and distribution, and demand includes relative fertilizer
prices, crop response to fertilizer application, credit, and other facilitating institutions.

although Eastern Province has recorded significant gains in fertilizer consumption as
well as output growth. In addition to analyzing the pattern and determinants of
fertilizer use on farms, this study also tries to further understanding of the process of
the transition from subsistence to commercial agriculture by analyzing differences in
fertilizer allocation behavior among farmers. The major objectives of the study are (1)
to describe the use pattern of chemical fertilizers by smallholder farms in Eastern
Province; (2) to identify the major determinants of fertilizer adoption and uses, and
(3) to examine the role of fertilizers in promoting agricultural transformation of the area.
A brief background of the study area is provided in Chapter 3. Chapter 4 focuses
on the physical fertilizer response environment and the profitability of fertilizer use,
based on on-farm and experiment station trials conducted by the Adaptive Research
Planning Team (ARPT) of the Eastern Province Agricultural Development Project
(EPADP) during the period 1983-87.2 The relative price scenario is also briefly
reviewed. This is followed in Chapter 5 by an analytical description of fertilizer use
practices of smallholders, including adoption and levels of use. Chapter 6 examines
farmers' access to fertilizers and credit. An econometric analysis of determinants of
fertilizer use is undertaken in Chapter 7, and fertilizer allocation behavior is studied
in Chapter 8. The final chapter highlights the major conclusions and policy implica-
tions of the study.

Analytical Framework

In almost all developing countries chemical fertilizers were introduced into the
main agricultural systems in the years following the Second World War. Prior to that,
the use of chemical fertilizer was confined to some plantation or cash crops. Research
on farmers' responses to this innovation and factors affecting those responses has
drawn heavily from the adoption-diffusion framework of sociology and anthropol-
ogy, on the one hand, and from factor demand theory on the other. Such studies range
from socio-cultural-economic determinants at the farm household level to more
aggregative, often time-series, analyses incorporating the role of prices, environ-
mental and other shifter variables, and policy factors.3 From these studies inferences
have been drawn regarding appropriate interventions to augment fertilizer use, such
as research, extension, credit, irrigation, and price policy.
A complementary, evolutionary framework focuses on the processes governing
growth in fertilizer use. As is evident from the following quotation, several of these
processes are still nascent in the developing-country context and hence act as con-
straints. This perspective emphasizes the institutional dimensions.
Forces behind growth in fertilizer use may be viewed as development of and
interactions among four sets of processes: (1) those that influence the
economic potential of fertilizer use through development of resources such

2EPADP was established with World Bank funding in 1982. Raising agricultural production through
rapid diffusion of improved technology was the main objective of the project. Major emphasis was placed
on reorganization of the extension system along training and visit (T&V) system lines.
3See Feder, Just, and Zilberman 1985 for a comprehensive review.

as irrigation and new technology that shift fertilizer response functions
upwards; (2) those that convert the potential into farmers' effective demand
for fertilizers by providing them with knowledge on fertilizer use, credit,
and assured markets for output; (3) those that determine the growth of
aggregate fertilizer supply through imports and domestic production; and
(4) those that help develop geographically dispersed fertilizer distribution
systems and determine how they operate. (Desai and Gandhi 1990).
This study, based on cross-sectional household data at a subnational level, essen-
tially uses the adoption-diffusion framework. Differences in fertilizer use between
households are examined in relation to socioeconomic, cultural, and specific farm
characteristics. Prices and other dynamic variables cannot be included in such mod-
els. A descriptive analysis of some of these processes is nevertheless provided, and
this insight is used to understand the observed differences. Also, locational differ-
ences in market access and fertilizer supply conditions observed in cross-sectional
data do allow incorporation of some of these variables into the analysis.
In a broader context, a cross-section of farming households in a growing, dy-
namic setting does, in a sense, capture individuals in different stages of transition as
they try to move from traditional to modern agriculture. Some persist with traditional
modes, others are experimenting with new ideas, while some have already achieved
high levels of modernization. Using a suitable criterion for classifying individuals in
different categories of transition, one can decipher forces or factors that govern this
process. In this study, an attempt is made at an exploratory analysis along these lines,
using fertilizer allocation behavior as a criterion.
Literature on fertilizer diffusion in developing countries suggests that fertilizer
use starts with high response crops and then spreads to other crops in a hierarchical
fashion as agriculture becomes more progressive. This process is used to depict the
transition from subsistence farming with no fertilizer use to greater commercializa-
tion and modernization of the smallholder farming sector.
A progressive transition of this kind has, in fact, been articulated by Zambian
development workers in the concept of a lima ladder. A lima is a unit of land (0.25
hectare). This step-by-step approach begins with a subsistence farmer and, as a first
step, introduces a small area (a lima) of a cash crop such as soybeans or sunflowers
(see Appendix 1, Figure 4). Sales from this activity generate cash to purchase
fertilizers to be used on local maize the next year. The increase in maize further
improves the farmer's cash position, and so in the third year local maize area is
reduced and hybrid maize is added. Up to this stage, area expansion is restricted
because of labor constraints. In the fourth year, the farmer has enough cash to hire
oxen and expand his hybrid maize area. Fertilizer plays a key role in this scheme.
Using the observed fertilizer use pattern on farms as an indicator of this transitional
pathway, the analysis seeks to identify factors that influence this process.

Data and Limitations

Eastern Province is divided into six administrative districts, namely, Chadiza,
Chama, Chipata, Katete, Lundazi, and Petauke. The provincial Department of Agri-
culture has demarcated 10 agricultural districts. This study is based on a repre-

sentative sample of 330 smallholder households spread over 9 agricultural districts in
the province. The survey was conducted as part of a collaborative study entitled
"Growth and Equity in Zambian Agriculture," funded by the Swiss Development
Cooperation. The International Food Policy Research Institute (IFPRI), the Rural
Development Studies Bureau, University of Zambia (RDSB), the National Food and
Nutrition Council (NFNC), and Eastern Province Agricultural Development Project
(EPADP) were the collaborators. Appendix 2, Table 24, shows the distribution of
sample households across districts. From each district, a cluster of villages (called a
branch) was selected randomly. In Lundazi District, two clusters were selected. From
each branch about 33 households were selected for detailed data collection. House-
hold activities were monitored and data collected by resident enumerators from
November 1985 to December 1986. The range of information gathered included area
measurements, monthly income, expenditures, labor use, crop cultivation practices,
inventory of assets, access to institutions and services, technical assistance, house-
hold decisionmaking, nutrition and consumption, and off-farm activities. Data were
also collected from reports of the provincial Department of Agriculture on general
agricultural characteristics of the province, and from the annual reports and other
bulletins of the Adaptive Research Planning Team of the EPADP on fertilizer
responses and recommendations.
Since 1985/86, the survey year, dramatic changes have taken place in the macro-
economic environment in Zambia. To the extent possible, the relevant changes have
been incorporated in the analysis. While no comment can be made on how farmers
have responded to these changes, the basic conclusions are nevertheless enduring and
remain relevant.



Zambia is endowed with abundant land resources. Of a total arable area of about
9 million hectares, only 1.4 million hectares are currently cultivated. Population
densities in the countryside are low and projections suggest that despite high popula-
tion growth rates, Zambia will remain a low-density region until about the second half
of the twenty-first century (Binswanger and Pingali 1988). In agricultural perform-
ance, however, the country follows the trend of declining per capital food availability
that is prevalent in most other countries in Sub-Saharan Africa (World Bank 1981).
Promoting fertilizer use has been an important element of Zambia's agricultural
strategy since the 1950s. Prior to that the focus was on good crop husbandry and land
conservation measures. Total fertilizer (nutrient) consumption grew from less than
20,000 metric tons in the late 1960s to more than 103,000 tons in 1981, but it has
faltered since then, fluctuating between 60,000 and 95,000 tons in recent years
(Figure 1).4 Fertilizer use intensities are low, declining from about 20 kilograms per
hectare of arable land in 1981 to 15-16 kilograms in recent years. The use is highly
concentrated-nearly 90 percent of Zambia's total fertilizer use is on maize. The
smallholder sector accounts for about 80 percent of the nation's aggregate consump-
tion. More than 80 percent of the current fertilizer supplies come from imports,
including concessional aid (Muleya 1990). It has been argued that the uncertainties of
donor assistance have contributed to the recent deceleration and instability in fertil-
izer consumption in the country.
These trends are partly reflected in the data for Eastern Province. The population
density, at 9.6 persons per square kilometer in 1985, is marginally higher than the
national average (7.5 persons per square kilometer). The annual cropped area repre-
sents only 19 percent of the total area and 35 percent of the arable land base of the
province, indicating easy access to land. The population growth rate is lower and the
proportion of rural population is much higher than the national average (NCDP
1989). The province has recorded significant gains in fertilizer consumption since
1981 (Table 1). By the mid-1980s it accounted for about 20 percent of the total
fertilizer consumption in the country in some years, although its share of the total area
is less than 10 percent. Even when national consumption of fertilizer faltered during
the early 1980s (Figure 1), use in Eastern Province continued to grow impressively.
The pace has slowed since.
In Eastern Province, agriculture is dominated by smallholders. About 96 percent
of the farms are less than 10 hectares in size; 72 percent are categorized as traditional
and 24 percent as small-scale commercial farms (Katongo 1988). Data from the
Eastern Province Department of Agriculture suggest that the cultivated area grew by
almost 10 percent per year between 1978/79 and 1985/86; most of this increment (93

4All tons referred to in this report are metric tons.

Figure 1-Fertilizer consumption in Zambia, 1969-88

Quantity offertilizer
(1,000 metric tons)












1969 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88

- NPK(1,000tons)

- NPK/hectare (kilograms)

Source: Food and Agriculture Organization of the United Nations, FAO Fertilizer Yearbook (Rome: FAO, various

Table 1-Fertilizer sales in Eastern Province and share of total sales, 1981-89

Sales in Share in
Year Eastern Province National Sales

(1,000 metric tons) (percent)

1981 28 13
1982 34 16
1983 40 25
1984 18 13
1985 55 26
1986 39 20
1987 51 21
1988 54 18
1989 60 20

Source: NAMBOARD (National Agricultural Marketing Board), Annual report and accounts (Lusaka, Zambia:
NAMBOARD, various years).
Note: Sales include total fertilizer material.

percent) is accounted for by maize (Eastern Province, Department of Agriculture
various years). This growth in cultivated area is the result of greater use of oxen.
Growth in marketed maize production in the province has been remarkable-the
quantities nearly trebled between the late 1970s and the late 1980s (Appendix 2,
Table 25): the share of Eastern Province in national supplies rose from less than 1
percent in 1970 to more than 27 percent in 1983 (Mumeka 1991). The growth pattern
in Eastern Province is particularly noteworthy because it has emerged largely from
the smallholder sector. Groundnuts, the traditional marketed crop of the province (the
commonly grown Chalimbana variety is well known as an export crop), stagnated
during most of this period, though it showed signs of recovery in the late 1980s. As
indicated earlier, fertilizer consumption rose steadily until the mid-1980s. Adoption
of hybrid maize and animal traction made significant headway in the plateau region
(Jha, Hojjati, and Vosti 1991). By all indicators, Eastern Province has been one of the
more dynamic regions of the country in agricultural performance.
The province has two distinct agroecological regions-the Eastern Plateau and
the Luangwa Valley. A semipermanent bush-fallow system of cultivation prevails
throughout the province, although some areas in the plateau region have begun to
experience local land scarcity, and fallow periods are being reduced (ARPT various
years). The Adaptive Research Planning Team has identified three major agro-
ecological zones in each of these two regions (Figure 2).5 Table 2 shows the major
characteristics of these zones. The annual rainfall ranges from 850 millimeters to
1,050 millimeters in the higher-altitude plateau, concentrated between the months of
November and April. Agricultural activities are confined to this period. Sowing starts
with the onset of rains and harvesting is completed by the end of May. The soils vary
from sandy to clay-loam in texture. Over 80 percent of the human and all of the cattle
population is concentrated in this region, and oxen cultivation is becoming quite
widespread.6 Consequently, farms are larger in the plateau region. Maize is the
dominant crop, accounting for more than 80 percent of the cultivated area. Both
traditional and hybrid varieties of maize are grown; the latter is cultivated exclusively
as a market crop.
The valley is thinly populated, receives lower rainfall, has higher temperatures,
and is heavily tsetse infested. Maize and groundnuts are important in this zone too,
but crops like sorghum, rice, millet, and cotton also occupy significant areas. Ground-
nuts and cotton are the only cash crops in this zone. Hybrid maize is practically
nonexistent. As will be shown in Chapter 5, fertilizer use is negligible. Thus, all three
major technological options-hybrid maize, fertilizer, and animal traction-are ab-
sent from the valley. Farms are smaller in this zone and hoe cultivation prevails.
The Msekera Agricultural Research Station near Chipata, the provincial head-
quarters, is the main research facility in the province. There is a small substation in
the valley (at Masumba), which is slated for upgrading. The Msekera Station is the
lead center for groundnut and grain legume research in Zambia. For maize, the lead
research center is at Mount Makulu in Lusaka. It directs the maize trials at Msekera.

5Within each plateau zone, at least two subsystems--hoe and oxen-are also identified.
6Note that the incidence of east coast fever makes cattle rearing risky in this zone also. Parts of Katete in
the southern zone of Eastern Plateau and the entire western zone of the plateau are identified as vulnerable
to tsetse (EPADP 1987).

Figure 2-Agroecological zones and agricultural districts, Eastern Province,



Zambia ( ,



O (North and
e South)

Sastr Chadiza
I Katcte
""' I'Petauke


Source: ARPT (Adaptive Research Planning Team), Eastern Province Agricultural Development Project, "Almual
Report, 1985-86" (Chipata, Zambia, 1986, mimeographed).
Note: The households sampled for this study were selected from the agricultural districts shown on the map.

Table 2- Characteristics of agroecological zones in Eastern Province

Eastern Plateau Luangwa Valley
Item Western North/South Central Chama Central South

Altitude (meters) 900-1,500 900-1,500 900-1,500 400-600 400-600 400-600
Mean annual rainfall (millimeters) 900-950 850-1,000a 850-1,050 800-900a 750-850a 700-800
Mean growing season (days) 125 135 140 145 140 125
Predominant soil Clay-loam Sandy Clay-loam Sedimentary Variable Variable
Population density 1980b (persons/kilometer) 9.3 11.7 17.1 ... 2.0
Fann families (1987 estimate) (percent of total) 17 41 24 7 6 4
Cultivated area/famnn (hectares) 3.39 2.99 3.20 .1.39d
Average family sizec (number) 6.67 5.77 5.49 5.93d
Cattle population/familyc (number) 1.7 3.0 1.4
Share of crop area sownc (percent)
Maizee 85.4 84.7 82.2 ... 58.3
Sorghum .. .... 15.6
Millet ... 5.0 .. .. 5.6
Rice 2.4
Groundnuts 9.9 4.7 13.4 ... 11.1
Soybeans 1.1 0.2 ... 0.4
Sunflower 1.6 3.7 1.9
Cotton 1.0 ... 0.4 ... 2.8
Percent of farmers adopting
Hybrid maize 31.1 42.7 34.4 ... 3.2
Animal traction 67.2 73.4 27.9

Source: Agricultural household survey conducted in 1985/86 by the International Food Policy Research Institute; the Rural Development Studies Bureau, University of
Zambia; the National Food and Nutrition Commission, Zambia; and the Eastern Province Agricultural Development Project, in Eastern Province, Zambia.
Notes: The Adaptive Research Planning Team identified hoe and oxen-based systems in each of the plateau zones. These have been aggregated because data are not
available at this level. The ellipses indicate a nil or negligible amount.
aShort drought periods occur.
bObtained by district approximation.
CBased on the survey for this study.
dBased on a survey for Chama and Central zones by the Adaptive Research Planning Team.
epure and mixed cropping prevail in the province.

The ARPT at Msekera has been conducting on-station as well as on-farm trials on
different crops since 1982. Since 1992, the provincial agricultural extension system
has been gradually reorganized along training and visit (T&V) system lines under the
World Bank-assisted Eastern Province Agricultural Development Project (EPADP).
The Research and Evaluation Unit and the ARPT have also been supported by the
project. The National Agricultural Marketing Board (NAMBOARD) and the Eastern
Cooperative Union (ECU) were the agencies responsible for input and credit supplies
and output purchases in the province.7

Evidence on Fertilizer Use in Eastern Province

The earliest systematic work on smallholder agricultural systems in the province
was conducted in Chipata District in 1971/72 (Harvey 1973). This study showed that
fertilizer use was largely confined to hybrid maize and tobacco-both cash crops-
though a small proportion (5 percent) of farmers used it on local maize. About 18
percent of the farmers grew hybrid maize, which covered nearly 10 percent of the
maize area. Hybrid maize was a new crop; no farmer had grown it for more than three
years. Before that time, fertilizer use was practically nonexistent in the smallholder
sector. The levels of fertilizer application on hybrid maize were close to those
recommended by the extension service.
Since the inception of the EPADP, lack of information on small farm systems has
been recognized as a serious handicap. Some small-farm surveys have been con-
ducted by the Research and Evaluation Unit of EPADP and ARPT. A survey in
Lundazi District, located in the northern zone of the plateau, which is the highest
fertilizer-consuming district in the province, showed that all sample farmers used
fertilizer on local as well as hybrid maize, often at higher than recommended levels.
No other crop was fertilized. In this high-adoption location, 74 percent of the farmers
were growing hybrid maize, and animal traction was common (EPADP 1984). A
similar survey in Eastern Plateau South revealed significant inefficiencies in fertilizer use
practices with regard to timing and method of application (ARPT various years). Severe
labor constraint around planting time was identified as an important reason for this.
Another small-farm survey, spread over the entire province, was conducted in
1984/85. It showed wide variability in resource endowments and agricultural prac-
tices within the province. Fertilizer use was found to be practically nonexistent in the
valley region and varied substantially between plateau districts. The districts of
Lundazi and Chipata South were found to be high fertilizer-consuming areas, fol-
lowed by Katete (EPADP 1986b).
Three major conclusions emerge from these surveys. First, very little fertilizer
was used in the province until the early 1970s. Fertilizer use picked up with the
emergence of hybrid maize as a cash crop, which shifted the fertilizer response curve
and profitability of fertilizer use dramatically. Even during these early years, farmers
used the recommended levels. Subsequently, farmers extended fertilizer use to local
varieties of maize also. In some areas, fertilizer use on hybrid and local maize has

7The marketing board, NAMBOARD, has since been dissolved and its function taken over by the
cooperative sector.

become universal. In others, the diffusion phase is still going on. In the province, no
other crops are fertilized. Second, agroecological conditions play an important role in
diffusion of fertilizer use. Fertilizer use in the valley is negligible, and even within the
plateau there is considerable variation. Finally, although fertilizer use has become
widespread, farmers do not always follow directions for optimal timing or method of
application, either because they lack sufficient knowledge or because of labor con-
straints during the planting and weeding period. Subsequent sections of this report
examine these and other related issues in greater detail, based on household data
collected during 1985/86.



Traditionally, fertilizer application was recommended only for hybrid maize and
tobacco, mainly in the plateau region of Eastern Province. These recommendations
were derived from trials conducted at the Mount Makulu and Msekera research
stations at Lusaka and Chipata, respectively. Since the inception of the ARPT in
1982, attempts have been made to evaluate crop responses to fertilizer on farmers'
fields in the province. This work tries to integrate agronomic and economic consid-
erations. Trials conducted by the ARPT during the period 1982/83-1987/88 are listed
in Appendix 2, Table 26, which indicates that by this time the research system was
conducting verification trials on farmers' fields for improved practices on crops such
as maize, cotton, soybeans, and groundnuts, whereas rice and finger millet trials were
still being conducted at the experiment station level. While fertilizer and maize
continued to be important in ARPT work, the focus was shifting to other crops and

Fertilizer Use Recommendations

Table 3 shows how fertilizer recommendations evolved from 1982/83 to 1986
based on this work. In the valley region, not enough experimental work had been
done to be able to make recommendations, so the same recommendations were given
for both the plateau and valley regions. The recommendations in 1982/83 were based
on a review of past work (mainly on experimental stations), as well as on current
recommendations based on on-station and on-farm trials.
The major focus of fertilizer work in the earlier period was on hybrid maize.
Though experimental work indicated a viable response, use of fertilizer on local
maize varieties was not pushed strongly by the extension system. Local maize was
viewed as a purely subsistence crop, on which the use of costly inputs could not be
justified. Groundnuts, the other major crop, did not respond to fertilizers. Other
traditional crops like sorghum and rice were not adequately investigated. Cotton
showed some response, particularly with plant protection and timely planting, but the
major cotton-promoting agency, the Lint Company of Zambia (LINTCO), did not
include fertilizer in its package of recommendations, perhaps because it would make
the package even more costly.
Early research in the province showed that the major nutrient deficiency in
plateau soils was nitrogen. The recommended phosphate and potash levels were
essentially for maintenance. Severe sulfur deficiency occurs in parts of the plateau, as
well as the valley, but it is adequately corrected by the 10 percent sulfur content in
most of the basal fertilizer formulations. Similarly, some locations have the problem
of high soil acidity, for which application of lime has been found effective. Research
in the valley region was largely neglected.

Table 3-Fertilizer use recommendations for major crops, 1982/83 and 1986

Crop 1982/83 1986

Local 200 kilograms per hectare of X mixture at 100 kilograms per hectare of X + 100
planting or 10 days after emergence kilograms per hectare urea 10-15 days
after emergence
Hybrid 200 kilograms per hectare of X + 150 200 kilograms per hectare of X + 150
kilograms per hectare urea at planting or kilograms per hectare urea or 200
10 days after emergence for early-planted kilograms per hectare urea 10-15 days
hybrids after emergence
Groundnuts Not recommended Lime at 400 kilograms per hectare on
highly acidic soils
Cotton 200 kilograms per hectare of X for early- 200-300 kilograms per hectare of X
planted and sprayed crops on light soils, for early-planted and sprayed crops
nil otherwise
Sorghum (Valley) No trial Not recommended
Soybeans 200 kilograms per hectare of D on phos- 150 kilograms per hectare of X
phate-deficient soils
Sunflower Not economical Not economical
Rice No trial 200 kilograms per hectare of X on
poor soils
Tobacco burleyy) Not available 200 kilograms per hectare of C + 200
kilograms per hectare of ammonium
nitrate as top dressing
Carioca beans Not available 150 kilograms per hectare of X

Sources: 1982/83 recommendations are based on survey of earlier experimental work reported in Adaptive Research
and Planning Team, Eastern Province Agricultural Development Project, "Annual Report, 1982/83,"
Appendix 13 (Chipata, Zambia, 1983). 1986 recommendations are from Eastern Province Department of
Agriculture "Guidelines on Fertilizer Use in Eastern Province Farming Systems: A Memo for Agricultural
Extension Workers," Chipata, Zambia, October 1986 (mimeo).
Notes: X, D, and C are fertilizer mixtures, where X = 20-10-5, D = 10-20-10, and C = 6-18-12 combinations of N,
P205, K20, respectively, and urea = 46-0-0.
aTop dressing can be applied up to four weeks after emergence.

Most of these recommendations have been confirmed and refined by the efforts
of the ARPT since 1982. Trials on sorghum, soybeans, sunflower, and rice showed
poor or uneconomic responses to fertilizer, while carioca beans showed promising
results with fertilizer. For the major crop, maize, modifications have been made in the
recommendations for local and improved local (MMV600) varieties. In addition,
trials have established the nature and magnitude of specific interactions of fertilizer
with planting time, weeding, impact of the preceding crop, plant protection, and so
forth (ARPT various years). These are now explicitly included in a qualitative manner
in the recommendations (Eastern Province, Department of Agriculture 1986b). For
example, the following cautionary points are emphasized in training materials for
extension workers:

No fertilizer should be used for cotton if planted late (December).
Nitrogenous fertilizers should never be used without basal fertilizer.
Phosphorous must be applied within two weeks of emergence.

No fertilizer should be applied to local maize if planted after December 20.
For late-planted hybrids (December), reduce the fertilizer level by half.
For hybrid maize preceded by beans, soybeans, or groundnuts, reduce the
quantity of top dressing fertilizer (urea) by half.
Benefits from fertilizers are dramatically reduced if weeding is not done properly.

Informal and formal surveys of farmers' practices conducted by the ARPT at
different locations conclude that these interactions and farmers' inability to follow
correct practices (mainly due to labor constraints at planting and weeding time) are
the main factors responsible for less-than-expected yields, particularly for hybrid
maize, despite high application rates of fertilizer. Surveys conducted in the Eastern
Plateau North and South zones, for example, revealed that 50-60 percent of fertilizer
users applied it on the surface or in uncovered holes, and 40-50 percent applied it very
late. Similarly, the quantity of fertilizer applied per stand was found to vary consid-
erably from the recommended practice (ARPT 1986). Apart from emphasizing these
technical efficiency parameters in the fertilizer recommendations, the ARPT has been
trying to address labor constraints by testing varieties and hybrids suitable for late
planting and alternative weed management techniques.
Two comments need to be made about the on-farm trials on which fertilizer
recommendations are based. First, the main thrust of ARPT work in this area has been
on variety and fertilizer response evaluation at different sites within the province.
Scrutiny of data for 1982/83-1985/86 indicates that, while some meaningful results
have been obtained on variety, fertilizer response trials did not generate statistically
valid results in either of the two years (1982/83 and 1983/84) for which details are
provided in the annual reports (ARPT various years). Whereas the ARPT has used
data on agroclimatic and farming systems characteristics to delineate agroclimatic
zones within the province, it has not been able to establish zone-specific responses.
This has constrained the capacity of the research system to develop location-specific
recommendations. Therefore, the same recommendations are made for the entire
province, with modifications suggested to suit individual farm situations.
Second, stability of responses over seasons and the resulting implications for
risks have not received analytical attention. In general, the perception is that yields in
the plateau region have been fairly stable over years. Table 4 indicates the variations
in weather and yields of hybrid and local maize in on-farm trials in the province over
a five-year period. The yield levels are remarkably stable over a wide range of
weather conditions, represented in the table by the length of the rainy season and the
number of drought spells during the growth period.
In order to investigate whether response to fertilizer application also shows
similar stability, data from a maize management trial, conducted over three years
(1983/84-1985/86) at Msekera Research Station, were examined (Table 5) (ARPT
various years). Comparable data were not generated in other trials. These years cover
the extremes of weather conditions experienced at this location. The effects of
fertilizer on both local and hybrid maize were statistically significant in all three
years. Average responses per kilogram of plant nutrient were found to fluctuate by
30-40 percent between years (Table 5). Yet even the lowest figures (9.5 kilograms for
local maize and 15.7 kilograms for hybrid maize) are high by tropical field crop
standards. This suggests that the response is also favorable in the plateau region from
the stability point of view.

Table 4-Weather conditions and maize yield levels at Msekera Research
Station, Zambia, 1982/83-1986/87

Number of Mean Yield of Maize
Length 10-Day Droughts, in On-Farm Trials
Start End of Rainy November 1-
Year of Rains of Rains Season March 31 Hybrida Localb
(days) (kilograms/hectare)
20-year average November 17 April 10 144 2.3 ...
1982/83 November 21 April 7 138 4.0 n.a. 2,517
1983/84 November 30 April 1 122 7.0 5,798 n.a.
1984/85 November 5 April 9 155 3.0 5,626 2,294
1985/86 November 5 April 21 167 0.0 5,073 2,317
1986/87 November 1 March 31 150 4.0 n.a. 2,799
Mean 5,499 2,480

Source: EDADP (Eastern Province Agricultural Development Project), "Eastern Province Agricultural Develop-
ment Project: Project Completion Report" (Chipata, Zambia, November 1987, mimeographed).
Note: n.a. is "not available."
aTrials were conducted at 27 sites.
bTrials were conducted at 46 sites.

Average Responses and Value-Cost Ratios
The ARPT has worked out average responses and value-cost ratios at recom-
mended levels of fertilizer application to different crops. These are conservative
estimates depicting expected responses on farmers' fields in 1983/84 and 1986/87
and are based on results of on-farm trials and researchers' considered judgments
(Table 6).
These data indicate that the plateau region enjoyed a favorable physical response
environment, particularly for hybrid maize, local maize, cotton, and beans. Among
the major crops, hybrid maize had the highest fertilizer use potential. The value-cost
ratios indicate that at 1986/87 prices, fertilizer use was profitable on maize, beans,
cotton, and tobacco. The other crops in Table 6-groundnuts, soybeans, and sun-

Table 5-Average response of local and hybrid maize to fertilizer application
per hectare at Msekera Research Station, 1983/84-1985/86

Average Response per Kilogram of Nutrients
Year Local Hybrid

1983/84 9.5 18.9
1984/85 15.3 21.8
1985/86 11.7 15.7

Source: Derived from data on Maize Management Trials in ARPT (Adaptive Research Planning Tcam), Eastern
Province Agricultural Development Project, "Annual Report," (Chipata, Zambia, various years, mimeo-
Notes: The fertilizer effect was statistically significant for both local and hybrid maize in all three years. The
fertilizer application was 200 kilograms of 20-10-5 NPK and 200 kilograms of urea.

Table 6-Average response and value-cost ratios for recommended levels of
fertilizer application in the plateau region

Value-Cost Ratio
Average 1983/84 1986/87
Crop Response Pricesa Pricesb
(kilograms of grain/
kilogram of nutrient)

Hybrid maize 18.1 4.0 4.3
Local maize
Traditional 11.1 2.6 2.7
MMV600 variety 13.3 3.2 3.3
Cotton 7.0 3.1 2.5
Soybeans 4.5 2.8 1.6
Groundnuts ... 2.8d n.a.
Sunflower 0.8 1.5 1.1
Tobacco 1.4 n.a. 2.3
Carioca beans 6.1 5.1 6.5

Source: ARPT (Adaptive Research Planning Team), Eastern Province Agricultural Development Project, "Annual
Report" (Chipata, Zambia, various years, mimeographed).
Notes: X is a 20-10-5 NPK fertilizer mixture. Prices are in kwacha (K).
aMaize = K0.31/kilogram, X = K26.75/bag (50 kilograms), urea = K26.75/bag (50 kilograms).
bMaize = K0.87/kilogram, X = KO0/bag, urea = K63/bag. Later period prices for commodities other than maize are
not available.
Responses for cotton and tobacco are in terms of kilograms of lint and dry leaves, respectively.
dResponse to lime application at the rate of 400 kilograms per hectare in areas where more than 20 percent popping
(splitting of immature shells) occurs.

flower-either did not respond or response levels were too low for use to be viable.
Rice and sorghum, not shown in the table, also fell in the same category.
Two other studies have reported average responses of maize to fertilizer applica-
tion, based on farm-level input-output data from the province. A study by Harvey
(1973) showed an extremely high response.8 Using the yield data collected in the
survey on which this report is based, Jha (1991) estimated farm-level response at
about 6 kilograms of maize per kilogram of plant nutrient for local maize and 19
kilograms for hybrid maize. The latter figure is close to the ARPT estimate but the
former is significantly lower. Judging by the enthusiastic adoption of fertilizer use by
local maize growers (discussed in Chapter 5), the former appears to be unrealistic.

Fertilizer Prices
Evaluating the impact of price changes on fertilizer consumption and related
issues such as fertilizer subsidies is beyond the scope of this study. However, the

8Harvey's 1973 study estimated a linear regression relating yield to fertilizer and other factors. The
fertilizer variable (defined in kilograms of fertilizer material) was highly significant and indicated an
average response of 12.7 kilograms of maize per kilogram of fertilizer. Assuming 40 percent nutrient
content in fertilizer, this implies an average response of about 32 kilograms. Considering that the
observation included local as well as hybrid maize plots, this number is very high.

price situation is briefly reviewed in order to provide an indication of the economic
viability of fertilizer use.
Fertilizer prices have been rising since 1980 (see Appendix 2, Table 27). Big
jumps occurred in 1983/84, 1986/87, 1988/89, and 1989/90. The last one was most
dramatic-the price of urea increased by 441 percent and D compound (10-20-10) by
302 percent, compared with 1988/89.9 The average price per kilogram of plant
nutrient rose by 315 percent; the maize price also increased about 100 percent in
1986/87 and 127 percent in 1989/90, compared with the preceding years.
Figure 3 shows the movements of the nutrient-maize price ratio in Zambia since
1980/81.10 Despite large fertilizer price increases, the nutrient-output (maize) price
ratio did not rise very high until after 1988/89, largely remaining between 3 and 4
kilograms. In the adjoining country of Malawi, where growth in fertilizer use on
smallholder farms has been sluggish despite a somewhat better physical response
(Lele, Christiansen, and Kadiresan 1989), very high relative fertilizer prices have
obviously dampened growth in fertilizer consumption (Figure 3). During most of the
1980s, Zambia was able to maintain a favorable relative price environment. In
1989/90, real fertilizer prices increased by 81 percent. The average value-cost ratio
declined to 3.13 for hybrid and 1.92 for local maize.
Comparing the relative price scenario in Figure 3 with fertilizer consumption data
in Figure 1 shows that there was some correlation between the two. Fertilizer prices
were rising from 1980/81 to 1983/84; fertilizer consumption was falling. When prices
fell between 1983/84 and 1985/86, consumption rose. Since then, the correlation
seems to have weakened. However, the post-1987 stagnation in fertilizer consump-
tion in Eastern Province could be due to a lack of growth in national supplies. By
1985, the share of Eastern Province in national consumption had become so large that
it could not be insulated from fluctuations in aggregate supply.
Because of changes in prices, the profitability of fertilizer use on maize has gone
down; on local maize it is now marginal, a matter of serious concern because fertilizer
use on local maize has been an important growth strategy for smallholder farms (see
Chapter 8). For other crops, the situation is worse. Because of price changes, returns
from fertilizer use dropped for cotton and soybeans (see Table 6). These crops, which
offer some opportunities for diversification of the crop base of fertilizer use, do not
receive the same kind of price support.
Increased efficiency assumes critical significance in the context of the new price
regime. Apart from concerns about technical efficiency articulated earlier, there are
two price-related factors that offer scope for improvement. First, in fixing prices for
different fertilizer mixtures, the nutrient composition does not seem to be taken into
account. For example, there is very little difference in the price of X (20-10-5) and D
(10-20-10) compounds, although the latter has a significantly higher nutrient content.
Moreover, D mixture has twice as much phosphate (P2O,)-a more expensive nutri-

9Three combinations of nitrogen (N), phosphorus (P205), and potassium (KO) (potash) are commonly
used in Eastern Province. X compound is mixed in proportions of 20-10-5, D compound is 10-20-10, and
C compound is 6-18-12.
'OThe average price per kilogram of plant nutrient was obtained by computing the price of nitrogen from
the price of urea and the price per kilogram of nutrient for compound 10-20-10-the most commonly used
fertilizers in Eastern Province-and averaging the two.

Figure 3-Fertilizer-maize prices in Zambia and Malawi

Kilograms ofmaize



Average response (maize/kilogram nutrient)

17- Hybrid

12" Average response (maize/kilogram ofnutrient)



Fertilizer price in maize equivalent (kilograms)

)/81 1981/82 1982/83 1983/84 1984/85 1985/86 1986/87 1987/88

Kilograms ofmaize

1988/89 1989/90


1980/81 1981/82 1982/83 1983/84 1984/85 1985/86 1986/87 1987/88



Sources: L. Mumeka, "Policy Constraints on the Development of Agriculture and Small-Scale Enterprises in Rural
Zambia," in Adopting Improved Farm Technology: A Study of Smallholder Farmers in Eastern
Province, Zambia, ed. R. Celis, J. T. Milimo, and S. Wanmali, 62-92 (Washington, D.C.: International
Food Policy Research Institute, 1991); and U. Lele, R. E. Clristiansen, and K. Kadiresan, Issues in
Fertilizer Policy in Africa: Lessons from Development Programs and Adjustment Lending. Managing
Agricultural Development in Africa (MADIA) Discussion Paper 5 (Washington, D.C.: World Bank, 1989).

Average response (maize/kilogram ofnutrient)

Average response (maize/kilogram ofnutrient)

Fertilizerprice in
maize equivalent (kilograms)

7 '

ent in the international market. Rationalizing prices by basing them on real prices of
nutrients would improve pricing efficiency.
Second, these differences have implications for fertilizer recommendations. Re-
search results indicate that X compound (20-10-5) is the proper basal fertilizer to use
for maize. Availability of this fertilizer in the province is restricted, D compound
(10-20-10) is made available instead. In response to this supply constraint, the
recommendation is modified to add more urea to make up for the lower nitrogen
content of the D compound. At 1989/90 prices, these two recommendations are
evaluated as follows:

200 kilograms X + 150 kilograms urea = (109N+20P20,+10K20), (1)

which costs 2,684 kwacha (K) per hectare," and

200 kilograms D + 200 kilograms urea = (112N+40 P205+20K20), (2)

which costs K3,120 per hectare. (Prices for X, D, and urea were K0.383, K0.396, and
KO.384 for each 50-kilogram bag, respectively).
Nonavailability of X implies that the fertilizer bill for the farmer is raised by 16
percent. Moreover, use of D compound leads to waste of P205 and K20 because the
proportions of those compounds are higher than needed. A recombination involving
100 D + 200 urea would provide 102N + 20P205 + 10K2O and would cost K2,328 per
hectare, thus lowering the fertilizer bill by 14 percent, compared with the X-urea
combination, and by more than 25 percent compared with the D-urea combination.
More research is needed to find the best way of combining the two fertilizers, but this
example clearly illustrates how greater economic efficiency can be generated. These
changes are purely technical in nature. Addressing them would significantly improve
the economic viability of fertilizer use and, to some extent, compensate for higher
fertilizer prices.
Five major findings emerge from this description of the fertilizer response envi-
ronment in Eastern Province.
Information is still deficient on fertilizer responses in the valley region. Yet
similar recommendations for fertilizer application on local maize, hybrid
maize, and cotton are made for both regions.
Even for the plateau zone, experiments with nutrients have been confined to
some early work at Msekera. Researchers have mainly focused on nitrogen, the
most deficient nutrient in Eastern Province soils, and have not investigated or
analyzed interactions with phosphorous or potash (KO2). Recommended levels
of the latter two are perceived as maintenance levels. In the context of the new
price regime confronting farmers, research on nutrient balance, which was not
as crucial in the subsidized price environment that prevailed earlier, has to be
given high priority. Researchers must focus on efficient nutrient combinations.
On-farm trials on fertilizers have not really been successful in establishing site-
and location-specific responses and recommendations in the plateau zone.
Major conclusions are still drawn from trials at the Msekera Research Station.

"US$1.00=K35 in 1989/90.

Limited resources available to the ARPT for testing, such as on-farm fertilizer
trials and soil testing, constrain this work.
All recommendations pertain to pure-stand (sole) cropping situations. As long
as farmers feel strongly compelled to use a mixed cropping system, there will
be a need to evaluate fertilizer use strategies for this system.12
The physical as well as the economic environment has contributed to high
fertilizer use in the plateau region of the province. Despite recent increases in
fertilizer prices, fertilizer use on hybrid maize remains profitable, although on
local maize, its viability is threatened. For other crops, no attempt has been
made to evaluate output prices in relation to fertilizer prices, perhaps because
no fertilizer is used on these crops. Consequently, the profitability of fertilizer
use is declining. Prices and marketing of crops other than maize will emerge as
relevant issues in the future.

12Area under mixed cropping was found to vary from 11.1 percent of the total cultivated area in the
Western Zone of the plateau to 49.1 percent in the Central Zone, 62.5 percent in the North/South Zone,
and 47.4 percent in the valley, according to the survey on which this report is based. Higher security of
output, symbiotic crop interactions, and the need for variety are the important rationales for mixed
cropping in subsistence agricultural systems. As will be shown subsequently, fertilizer use is common
under mixed cropping situations.



This chapter presents an overview of different dimensions of fertilizer use in
Eastern Province, based on data from the sample households. Data on adoption,
extent of area fertilized, and average rates of fertilizer application are presented in
Table 7. Information is provided for 10 locations to reflect variability.
First, there is a sharp distinction between the plateau and valley regions. Sixty-
seven percent of farmers in the plateau used fertilizers, against only 6 percent of those
in the valley. The distinction is equally sharp for area fertilized and rates of applica-
tion. The perception is that poor infrastructure and an inadequate fertilizer distribu-
tion network are the main factors responsible for low fertilizer use in the valley
(ARPT 1986). Recent work based on the same data set on access to infrastructure
reveals that infrastructure deficiencies exist in the valley zone (Wanmali and He

Table 7-Fertilizer use on smallholder farms in Eastern Province

Percent Percent Average Rate of Applicationb Percent
of Farmers of Cultivated Per Per of Farmers
Region/Zone Using Area Cultivated Fertilized Using
Location Fertilizersa Fertilized Hectare Hectare Manure
Plateau Region 67.1 55.6 53.6 96.4 1.2
North/South Zone
Chipili 100.0 90.9 87.8 96.6
Sinda 20.0 10.7 14.0 130.7 8.0
Chaweya 90.9 60.8 60.9 100.1
Kasendeka 87.9 68.4 78.7 115.0
Central Zone
Makangila 93.9 80.8 83.3 103.1
Mtenthela 53.6 45.6 36.8 80.8
Western Zone
Chiwizi 25.8 18.0 13.4 74.4 3.2
Kamwala 45.1 49.2 31.0 63.0
Luangwa Valley Region 6.3 5.2 3.5 66.9
Nkhoka 10.0 9.5 7.5 79.4
Mphata 3.0 2.8 1.2 42.2

Source: Agricultural household survey conducted in 1985/86 by the International Food Policy Research Institute; the
Rural Development Studies Bureau, University of Zambia; the National Food and Nutrition Commission,
Zambia; and the Eastern Province Agricultural Development Project, in Eastern Province, Zambia
Note: The ellipses (. .) indicate a nil amount.
aThis is a measure of the extent of adoption.
bThe average rate of application is in kilograms of total plant nutrients (N + P205 + K20). Total quantity divided by
the total cultivated area gives the average rate per unit of cultivated area. The rate per fertilized hectare is obtained by
dividing the total quantity by the area receiving fertilizer.

1989). However, as mentioned in the preceding chapter, because there is so little
experimental evidence on fertilizer response, low or uneconomic response cannot be
ruled out as a relevant factor explaining low use in the valley zone. Hybrid maize, the
most fertilizer-responsive crop, is seldom grown in the valley. Therefore, the valley
subsample has been excluded from subsequent analysis and discussion.

Average Fertilizer Use Indicators

The average rate of fertilizer applied per hectare of cultivated area in the plateau
region-more than 53 kilograms of plant nutrients-is high by developing-country
standards, and the average rate per fertilized hectare is even more impressive-96
kilograms. These high rates of application suggest that the extension service con-
vinced the adopters that it was not worthwhile to use small quantities. Experimental
work suggests a threshold level of 50-60 kilograms, below which no response is
obtained (ARPT various years). The average rate of application per fertilized hectare
was not lower than that at any location in the plateau region. This is an important
finding, because most studies on fertilizer use in Asia, for example, indicate that rates
of applications were not high in the early years of use. In Eastern Province, however,
farmers have used the recommended amounts right from the beginning (Harvey
Data by locations (branches) reveal substantial variation. Among the plateau
locations, adoption ranges from 20 to 100 percent, and the area fertilized, from 11 to
91 percent. The variation in rate of application per fertilized hectare is smaller: in four
out of eight locations, the rate is over 100 kilograms of plant nutrients per fertilized
Looking at agroecological zones within the plateau, the indicators are low in the
Western Zone, but even in the North/South Zone there are pockets of low fertilizer
use. (North and South are combined for purposes of enumeration because results are
similar.) In subsequent analysis, reasons for this spatial pattern are explored, and this
evidence has obvious implications for fertilizer promotion (mainly research and
extension) and distribution strategies within the province.
Finally, the last column of Table 7 shows that use of organic (cattle) manure has
not been integrated as a fertility-enhancing strategy in the farming systems in Eastern
Province. In only 2 out of 10 branches did a few farmers report this practice, although
households do maintain cattle in all three major agroecological zones in the plateau
region (see Table 2). Even where organic manure use was reported, the practice was
to "kraal" (corral) the animals on the plot rather than to use decomposed cattle
manure. Surveys carried out by the ARPT indicate that farmers perceived that
increased weed infestation from the use of cattle manure was a major disadvantage
(EPADP 1984). There was some use on intensively cultivated dimba (wetland)
gardens. Manuring is a labor-intensive activity, which does not mesh with the
relatively limited labor endowments of the region (Binswanger and Pingali 1988).
The practice of burning weeds and other vegetative matter before planting was,
however, common. Fallowing was practiced as a long-term fertility enhancement
strategy under a bush-fallow system all over the province.
Although fertilizer use was practically nonexistent in the province before 1970, it
reached about 54 kilograms per hectare in 15 years. This rate of growth is comparable

to any recorded in high fertilizer-using pockets in developing countries (see, for
example, Desai 1988). This rate and the fact that more than 55 percent of the cropped
area receives fertilizer indicate that fertilizer use has become a well-established
practice in the plateau. Use levels in this region of the province far exceed the national

Allocation of Fertilizers

Experience in different developing countries indicates that, during the process of
diffusion, fertilizer use spreads from high and more profitable responses and more
profitable environments and crops to those with lower responses (Desai 1988; Desai
and Gandhi 1990; Lele, Christiansen, and Kadiresan 1989). This seems to have been
the pattern in Eastern Province also. A study by Harvey (1973) finds that fertilizer use
was almost totally confined to the most fertilizer-responsive crop-hybrid maize-
in the early 1970s.
Table 8, which provides data on the current pattern of fertilizer allocation in
sample households and area shares of different crops in the plateau, clearly shows that
fertilizer use is still largely confined to maize. Other crops are seldom fertilized. In
contrast to the early 1970s however, the maize fertilization profile has changed.
Fertilizer use has diffused to local varieties of maize and to mixed cropping situ-

Table 8-Allocation of area and fertilizer among crops on smallholder farms in
the Eastern Province plateau region

Local Hybrid Local Hybrid
Maize Maize Maize Maize Other
Zone/Location (Sole) (Sole) (Mixed) (Mixed) Crops
Area allocation
North/South Zone 7.6 15.0 51.7 10.4 15.3
Central Zone 14.9 19.4 42.1 5.8 17.8
Western Zone 55.6 18.8 6.7 4.3 14.6
Fertilizer allocation
Plateau 12.6 39.3 33.0 14.8 0.3
North/South Zone
Chipili 4.6 6.3 51.7 37.4 0.0
Sinda 0.0 49.5 50.0 0.0 0.5
Chaweya 28.0 35.1 30.3 6.6 0.0
Kasendeka 6.4 46.2 33.0 14.4 0.0
Central Zone
Makangila 18.6 42.8 30.4 8.2 0.0
Mtenthela 8.8 43.5 47.6 0.0 0.1
Western Zone
Chiwizi 33.4 57.1 0.0 0.0 9.5
Kamwala 14.2 69.2 3.9 12.7 0.0

Source: Agricultural household survey conducted in 1985/86 by the International Food Policy Research Institute; the
Rural Development Studies Bureau, University of Zambia; the National Food and Nutrition Commission,
Zambia; and the Eastern Province Agricultural Development Project, in Eastern Province, Zambia.
Note: Fertilizer is in total plant nutrients (N+P205+K20).

nations, but the response in these situations is inferior to that for hybrids and single
cropping. For the plateau region as a whole, about 46 percent of the fertilizer use is
allocated to local maize, ranging in different locations from 18 percent to more than
58 percent. Similarly, nearly 48 percent of fertilizer is used on mixed crops ranging
from 0 to 89 percent in different locations. Qualitatively, however, the fertilizer scene
is still dominated by hybrid maize. Only about 25 percent of the total maize area in
the plateau is devoted to this crop (against nearly 60 percent to local maize), yet its
share in total fertilizer use is more than 54 percent.
The distribution of fertilizer between pure (sole) and mixed cropping situations
also depends on the dominance of one or the other type of cropping. For example,
mixed cropping is dominant in the North/South and Central zones (accounting for
48-62 percent of the cultivated area), and maize mixtures account for a substantial
share of the total fertilizer used in these zones. In the Western Zone, where pure
cropping is dominant, the opposite is true. It is important to note here the complete
neglect of mixed cropping in agronomic research in general and fertilizer research in
It is tempting to speculate about a diffusion path by relating the allocation
behavior in Table 8 to the adoption data in Table 7. In the Western Zone, where
fertilizer adoption is low, fertilizer use is concentrated on "better" response alterna-
tives-57-82 percent of the fertilizer is allocated to hybrid maize and 83-90 percent
goes to pure crops. At high adoption locations, on the other hand, use is more
diffused. This supports the inference that in the initial stages of adoption, fertilizer is
concentrated on crops where its use is most profitable.
Caution is, however, needed in equating adoption with use in any given season.
The former concept represents the first overt step in acceptance of an innovation.
Respondents to the survey were not canvassed to find out when they actually adopted
fertilizer. Subsequent use or nonuse, as well as the pattern of allocation in any specific
year or season, would depend on availability of funds to buy fertilizers and availabil-
ity of fertilizer supplies. For example, even for a progressive farmer, nonavailability
could result in concentration of limited supplies on hybrid maize only, or even nonuse
in a specific season. It is, in fact, argued later that fertilizer use is so widespread in the
plateau region that individual farmers are no longer constrained by lack of knowl-
edge. Even those who do not grow hybrids now use fertilizers.
Adoption of fertilizer use is more pervasive than adoption of hybrid maize. The
proportion of farmers growing hybrids varies from 31 percent to 43 percent between
the three plateau zones (Table 2), whereas about 67 percent of farmers use fertilizers
(Table 7). Less than 25 percent of the cultivated area is devoted to hybrid maize
(Table 8), whereas fertilizer is used on about 56 percent of total area. Obviously, there
are fertilizer users who do not grow hybrid maize. An examination of the distribution
of maize fertilizer users in three mutually exclusive categories shows that 49 percent
of those plateau farmers who used fertilizers on maize confined their use to local
maize only. Only 8 percent used it exclusively for hybrid maize, and the rest fertilized
both hybrid and local maize.
This result is surprising. A priori, one would expect that some farmers (early
adopters as well as highly commercialized farmers) would use fertilizers only on
hybrid maize, which is most fertilizer responsive, and others would use it on both
hybrid and local varieties. Diffusion literature would suggest a "no-use hybrid
maize hybrid + local hybrid" continuum, depicting the transition to commercial

agriculture. This pathway follows from the concept of a hierarchy of responses (Desai
1969), implying that the proportion of farmers using fertilizers on hybrid maize
would be larger. Data show that 92 percent of the fertilizer-using households use it on
local maize, as against 51 percent on hybrids (Table 9).
That such a large proportion of farmers fertilize only local maize calls for further
examination of the characteristics of those farmers who currently use fertilizer but not
on the most responsive crop-hybrid maize. A probit regression based on data for
fertilizer-using households"1 shows a dependent fertilizer-use variable as dichoto-
mous, having a value of unity for those households who fertilize only local maize and
zero for those belonging to other categories (Table 10). The independent variable set
includes the following:
Personal attributes of the farmer: age, education, sex;
Response characteristics: cultivated area, family size, oxen use, dependency
ratio, cash sales, extension advice, credit; and
Locational factors: agroecological zones.
Results indicate that farm size has a strong influence; farmers cultivating small
areas are likely to be the ones who fertilize local maize. Households headed by
females and older people and families having fewer dependents are also more likely
to fall in this category. Those who fertilize local maize only are more likely to be
located in the North/South Zone where fertilizer adoption is generally high (Table 7).
Variables like education, capital, oxen use, and extension advice do not discriminate
between the two groups of fertilizer-using households. These findings lend support to
the point made earlier that knowledge regarding fertilizers is now widespread, par-
ticularly in areas where adoption is high, like the North/South Zone, and even
traditionally disadvantaged farmers (small-scale farmers, women, and older farmers)
use fertilizer on local maize. This group of fertilizer users is likely to be most similar
to the subsistence-oriented nonuser. These are important findings and are pursued
further in a broader context in Chapter 7.

Table 9-Allocation of fertilizer among crops for fertilizer users, Eastern

Number of Percent
Crop Fertilized Households of Iouseholds

Hybrid maize only 10 8
Hybrid maize and local maizea 57 43
Local maize only 64 49
Total 131 100

Source: Agricultural household survey conducted in 1985/86 by the International Food Policy Research Institute; the
Rural Development Studies Bureau, University of Zambia; the National Food and Nutrition Commission,
Zambia; and the Eastern Province Agricultural Development Project, in Eastem Province, Zambia.
aIncludes pure as well as mixed crops.

13This can lead to sample selectivity bias (see Chapter 7).

Table 10-Probit regression explaining fertilizer use on local maize (probit
maximum likelihood estimates)

Variable Coefficient t-Statistic Mean

Constant 0.821 1.248
Age (years) 0.017* 1.572 41.33
Primary education (0,1, yes is 1) -0.286 0.753 0.26
Secondary education (0,1, yes is 1) -0.074 0.194 0.37
Female household head (1 is female) 0.649** 1.715 0.27
Dependency ratio (dependents/worker) -0.526*** 2.137 0.73
Family size (number) 0.058 0.959 5.84
Area cultivated (hectares) -0.708*** 4.649 3.13
Total farm sales (100 kwacha) -0.001 0.123 12.73
Cultivation method (0,1, 1 is oxen) -0.354 0.967 0.70
Extension advice (0,1, yes is 1) 0.311 0.949 0.30
Cooperative membership (0,1, yes is 1) -0.016 0.045 0.25
North/South Zone (0,1, yes is 1) 0.645* 1.517 0.64
Western Zone (0,1, yes is 1)b 0.225 0.390 0.13
X2 74.850
Number of observations 131

Notes: The Central Zone is omitted.
*Significant at the 20 percent level.
**Significant at the 10 percent level.
***Significant at the 5 percent level.

Maize is the only crop that receives fertilizer in significant amounts. This implies
that if relative fertilizer-maize prices and marketing conditions deteriorate,14 fertilizer
consumption could be adversely affected. Price manipulations of fertilizer relevant to
maize in the mid-1980s have been careful (Figure 3): it was only in 1989/90 that price
changes adversely affected the economic incentive to use fertilizer. Other cash crops
being promoted to augment and diversify smallholder incomes in the province (for
example, soybeans and sunflower) do not respond well to fertilizer application (Table
6) and, therefore, cannot help diversify the crop base of fertilizer use. Cotton is an
exception. LINTCO, the parastatal agency responsible for all aspects of cotton
promotion in the province, including extension, input supplies, and marketing, does
not yet include fertilizer as part of the recommended package of practices for cotton,
despite evidence from on-farm trials that fertilizer can result in significant responses
and profits (ARPT 1986). Beans also offer some opportunities for diversification.
This crop has not received adequate attention in development programs, including
extension and marketing. Maize has been the center of attention for policymakers,
and maize prices have been adjusted upward in recent years, often distorting relative
output prices. For example, decline in production of groundnuts in Eastern Province
has been attributed to this factor (Zambia, Ministry of Agriculture and Water Devel-
opment, Planning Division 1983). As the maize production and supply situation
improves at the national level, these issues will emerge as major constraints. It is

14Prices are officially fixed and maize purchases are handled by a parastatal company.

important to focus attention on crops like cotton and beans and to initiate fertilizer
promotion programs for them.

Fertilizer Use on Crops

Disaggregated information on levels of fertilizer use by crops confirms the trends
noted earlier (Table 11). Hybrid maize is almost always fertilized, regardless of
whether it is grown in pure or mixed stands. This holds for all locations within the
plateau region. Fertilizer use on local maize, on the other hand, is highly variable
across locations, although, on average, 43 percent of the pure and 58 percent of the
mixed stand local maize crops are fertilized.15 Again, fertilizer use indicators in the
Western Zone locations and at Sinda in the North/South Zone, are poor. These are
relatively low adoption zones (Table 7). In these situations, almost the entire hybrid
maize area is fertilized, but fertilizer use is constrained on local maize (Table 11),
which is what one would expect during initial phases of adoption or when availability
is limited. Another factor may be lower or uneconomic response to fertilizer use on
local maize at these locations, but no data are available to test this possibility.
Rates of application are, as expected, significantly higher for hybrids than for
local varieties and generally higher for pure crops than for mixed crops. The recom-
mended rate of fertilizer application for hybrid maize in the plateau region is 140-170
kilograms of plant nutrients per hectare, depending on availability of different kinds
of fertilizers. In most cases, actual application rates for hybrid maize planted alone
are in this range. For local maize, on the other hand, actual rates are generally higher
than the recommended 80 kilograms of plant nutrients, except in the Western Zone.
The Research and Evaluation Unit Survey of Lundazi District (North Zone) also
found that in 1982/83 farmers were using higher than recommended levels, and this
was attributed to inefficient fertilizer application techniques (EPADP 1984). Data for
1985/86 show that at comparable locations (Chaweya and Kasendeka), rates for
hybrid maize were not noticeably higher, but rates for local maize were, even though
1985/86 was a subnormal year in terms of fertilizer supplies to the province (Table
11) (EPADP 1987).
In order to provide firmer estimates of rates of fertilizer application (a number of
estimates in Table 11 are based on too few observations), the data are aggregated at
the level of agroecological zones (Table 12). The application rates, presented in terms
of individual nutrients, are highest in the North/South Zone and lowest in the Western
Zone. This is in line with the observed pattern of fertilizer consumption in the
province (EPADP 1984). The major nutrient shortfall (in relation to recommended
levels) is in nitrogen, particularly in the Central and Western zones. The actual levels
of phosphate and potash are generally higher than recommended, again with the
exception of the Western Zone. This is, as discussed in Chapter 4, related to the
availability of correct fertilizer formulations in the province. The recommended
mixture is X fertilizer (20-10-5), which contains a higher proportion of nitrogen, but

15The latter figure is misleading because location data show that the proportion of mixed crop area
fertilized is higher than sole crop area at only one location (Makangila). At all other locations, the reverse
is true.

Table 11- Percent of area fertilized and rate of application per fertilized hectare for important crops in the plateau region

Percent of Crop Area Fertilized Rate of Applicationa
Local Hybrid Local Hybrid Local Hybrid Local Hybrid
Maize Maize Maize Maize Maize Maize Maize Maize
Zone/Location (Sole) (Sole) (Mixed) (Mixed) (Sole) (Sole) (Mixed) (Mixed)

Plateau Region 43 97 58 98 74 119 87 106
North/South Zone
Chipili 100 100 97 100 102 143 86 112
Sinda 94 7 174b 99b
Chaweya 100 100 61 93 97 130 77 152b
Kasendeka 91 96 72 100 124 148 88 115
Central Zone
Makangila 88 100 98 100 83b 96 109 125b
Mtenthela 74 91 49 ... 47b 183 59
Western Zone
Chiwizi 19 81 .. 47 198b
Kamwala 28 99 20 100 31 77 66b 56b

Source: Agricultural household survey conducted in 1985/86 by the International Food Policy Research Institute; the Rural Development Studies Bureau, University of
Zambia; the National Food and Nutrition Commission, Zambia; and the Eastern Province Agricultural Development Project, in Eastern Province, Zambia.
aln kilograms of total plant nutrients (N+P205K20) per fertilized hectare.
bBased on less than five observations.

Table 12-Rates of nutrient application per fertilized hectare for major crops
by agroecological zones in the plateau region

North/South Zone Central Zone Western Zone
Crop N P205 K20 N P205 K20 N P205 K20
Hybrid maize (sole) 96 33 17 74 27 13 61 16 8
(109) (20) (10) (109) (20) (10) (109) (20) (10)
Local maize (sole) 68 25 12 49 17 9 32 4 2
(66) (10) (5) (66) (10) (5) (66) (10) (5)
Hybrid maize (mixed) 75 26 13 n.a. n.a. n.a. n.a. n.a. n.a.
Local maize (mixed) 56 20 10 61 20 10 n.a. n.a. n.a.

Source: Agricultural household survey conducted in 1985/86 by the International Food Policy Research Institute; the
Rural Development Studies Bureau, University of Zambia; the National Food and Nutrition Commission,
Zambia; and the Eastern Province Agricultural Development Project, in Eastern Province, Zambia.
Notes: Figures in parentheses indicate recommended levels. These are based on the most efficient recommendation
using X (20-10-5) rather than D (10-20-10) fertilizer mixture (Table 3). Recommendations are not available
for mixed crops.
n.a. is "not available."
N is nitrogen, P205 is phosphate, and K20 is potash.

D mixture (10-20-10) is commonly supplied (EPADP 1986a). Use of this D mixture
results in relatively higher than recommended application rates for phosphate and
These tables suggest two major findings. First, there appears to be some scope for
raising application rates for hybrid maize, particularly in the Western and Central
zones. Availability and use of the correct fertilizer formulation may be the simple
answer to this. Lack of the correct type of fertilizer not only reduces nitrogen levels,
it also creates an imbalance in nutrient use. The extension service has obviously not
been able to make farmers aware of the adjustments that need to be made when
shifting from X to D fertilizer mixture. Second, the Western Zone, where all fertilizer
use indicators are low, deserves special attention. More on-farm research on fertilizer
is clearly needed in this zone.

Balanced Use of Plant Nutrients

Another way of looking at the question of using another fertilizer mixture when
the correct one is unavailable is to examine whether farmers show a preference for a
particular nutrient. There is, for example, evidence from Asia suggesting a neglect of
nutrients other than nitrogen in an environment where nitrogen deficiency is the most
critical problem (Desai and Gandhi 1989). The resulting imbalance threatens long-
term fertility of the soil. In the plateau region, nitrogen is also the major nutrient
deficiency; how the sample households address this problem is shown in Table 13.
In general, and in contrast to the Asian experience, farmers apply all three major
nutrients. It should be noted that use of only nitrogenous fertilizers is not recom-
mended (EPADP 1987). Generally, farmers follow a balanced pattern: only 8.5
percent of the households use only nitrogen and 5.7 percent of the fertilized area
receives only nitrogen. Variations do exist from one location to another, however. For

Table 13-Use of plant nutrients, plateau region

Percentage of Farmers Percent of Fertilized Area
Using Using Receiving Receiving
Zone/Location N Only NPK N Only NPK

Plateau Region 8.5 91.5 5.7 94.3
North/South Zone
Chipili ... 100.0 .. 100.0
Sinda 25.0 75.0 10.0 90.0
Chaweya 3.3 96.7 0.7 99.3
Kasendeka ... 100.0 ... 100.0
Central Zone
Makangila ... 100.0 ... 100.0
Mtenthela 46.7 53.3 34.8 65.2
Western Zone
Chiwizi 50.0 50.0 92.1 7.9
Kamwala 7.1 92.9 4.1 95.9

Source: Agricultural household survey conducted in 1985/86 by the International Food Policy Research Institute; the
Rural Development Studies Bureau, University of Zambia; the National Food and Nutrition Commission,
Zambia; and the Eastern Province Agricultural Development Project, in Eastern Province, Zambia.
Notes: The ellipses indicate a nil or negligible amount; N is nitrogen, P is phosphorous, and K is potassium.

example, in Chiwizi and Sinda, both low-adoption areas (Table 7), a significant
proportion of fertilizer users apply only nitrogen. The same is true in Mtenthela. At
the other five locations, almost the entire fertilized area receives all three nutrients.
These locational differences could be due to unavailability of basal fertilizer types.
No data are available on this, although use and purchase patterns, and by inference
availability of different kinds of fertilizers, are examined in the next chapter.

Farm Size and Fertilizer Use

The preceding sections have demonstrated that fertilizers are widely used in the
plateau region of Eastern Province. But how uniformly are fertilizer use practices
diffused within the smallholder sector? The interest here is in farm size. Literature on
distributional consequences of new technology have focused on whether these tech-
nologies have any size bias.
Farm size distribution data are not available for Eastern Province as a whole. A
small farm is customarily defined as less than 10 hectares in size, and about 96
percent of the farms in the province belong in this category. Table 14 shows the
distribution of cultivated area of sample households in the plateau region and the
corresponding fertilizer use in each size category within the smallholder sector. This
section provides a descriptive summary and Chapter 7 addresses the question more
Contrary to popular perceptions regarding land ownership in traditional agricul-
tural systems in Africa, land distribution in the plateau region appears to be highly
skewed. About 24 percent of the sample households cultivate less than 1 hectare of
land, and their share in the total cultivated area is only about 6 percent (Table 14). At

Table 14-Fertilizer use by farm size, plateau region

Farm Size Classes (in Hectares)
Less More
Item than 1 1.1 2.0 2.1- 3.0 3.1 5.0 than 5

Percent of households 23.9 31.6 15.4 18.6 10.5
Percent of total area 5.8 18.6 15.0 28.2 32.4
Percent of maize area under hybrids 3.1 15.1 17.6 24.2 49.2
Percent of farmers using fertilizers 51.7 67.9 65.8 67.3 100.0
Percent of area fertilized 45.5 47.5 51.2 41.3 70.7
Kilograms of nutrientsa/fertilized hectare 102.5 92.2 94.6 104.5 93.9
Fertilizer allocation (percent)
Local maize (sole) 28.1 14.2 7.1 11.6 11.9
Local maize (mixed) 59.4 47.5 54.3 26.7 21.1
Hybrid maize (sole) 5.5 27.2 20.9 41.8 52.6
Hybrid maize (mixed) 5.5 9.0 17.7 19.9 14.4
Other crops 1.5 2.1 ... ..

Source: Agricultural household survey conducted in 1985/86 by the International Food Policy Research Institute; the
Rural Development Studies Bureau, University of Zambia; the National Food and Nutrition Commission,
Zambia; and the Eastern Province Agricultural Development Project, in Eastern Province, Zambia.
aTotal plant nutrients (N+P205+K20).

the other extreme, only 10.5 percent of the households cultivate more than 5 hectares,
but they control 32 percent of the total area. Fertilizer use information shows that
there is not much difference between size categories in rate of adoption, percent of
area fertilized, and rate of application on farms up to 5 hectares. Farms in the smallest
category (less than 1 hectare) have marginally lower rates for adoption and area
fertilized. The average levels on smaller farm-size categories suggest fairly high
fertilizer use--one-half to two-thirds of the households use fertilizer and about half
the cultivated area is fertilized at an average rate of about 100 kilograms of total plant
nutrients per hectare.
On farms of more than 5 hectares, adoption is universal. About 71 percent of the
cultivated area receives fertilizer at a rate of 94 kilograms per hectare. These farms
are clearly more commercialized. For rate of application, however, no trend is
discernible between the different farm categories.
The pattern of fertilizer allocation to crops, however, changes with farm size. In
the smallest category, more than 87 percent of fertilizer is allocated to local maize; on
farms of more than 5 hectares, hybrid maize accounts for 67 percent of the total
fertilizer used. As farm size increases, there is also a tendency to allocate more to sole
crops than to mixed crops of maize. These patterns could be related to relative land
scarcity and food requirements at the household level. Farmers concentrate on local
maize for their food needs; hybrid maize is mainly grown as a cash crop. To meet
their food requirements from a meager land base, farmers in the smaller size classes
allocate larger shares of their cultivated area to local maize varieties and use fertilizer
to obtain higher yields. It has been empirically demonstrated in Table 10 that smaller
farmers dominate the category that uses fertilizer on local maize. A small amount of
cultivated area also implies more mixed cropping to meet the need for variety in food.
These constraints are not binding on larger holdings, where adoption of hybrids is
much higher (Jha, Hojjati, and Vosti 1991).

This is an important finding. It has been argued that yield-increasing technologies
do not mesh well with a land-surplus farming system unless infrastructure and market
conditions make it profitable to produce surpluses that can be sold (Binswanger and
Pingali 1988). Hybrid maize in Eastern Province illustrates this case. However, this
report shows that even very small farmers who are primarily subsistence-oriented
find it worthwhile to use fertilizers (Table 10). When fertilizers generate high re-
sponses, they represent a viable option for small farmers. By raising the possibility of
surplus production (and command over more labor resources), fertilizer use facilitates
a medium- to longer-term transition to larger farms and greater commercialization.
An attempt is made in Chapter 8 to elaborate on this theme and to test this proposition


Farmers' fertilizer use decisions and practices are also influenced by fertilizer
supply and distribution conditions. Availability of the right kinds of fertilizers at the
right time and at accessible locations is crucial. Access is also constrained by
inadequacy of working capital in subsistence-dominated agricultural systems. Pro-
cesses such as distribution systems, credit, and procurement are still evolving in most
developing countries. In the Zambian context, fertilizer procurement and distribution
are handled by the state through parastatals or cooperatives.16 The latter are also the
main source of institutional credit for smallholders.17 There is no private-sector
involvement. Unfortunately, no data were collected on the existence, scale, or opera-
tion of these institutions in the study locations. Therefore, inferences are drawn from
observations of the extent to which the sample households interact with institutions.
For example, availability of fertilizer is judged by farmers' purchases of different
kinds of fertilizer and the timing of these purchases. Access is evaluated on the basis
of use of credit, distances, and modes of travel.

Use of Major Fertilizer Materials

Several fertilizer mixtures are available in Zambia. These are meant to be used
for basal application at or soon after planting. Research done in Eastern Province
found X compound (20-10-5) to be the most effective basal fertilizer. Urea is
recommended as the nitrogenous fertilizer for top dressing. The fertilizer materials
actually used by the sample farmers during 1985/86 are presented in Table 15.
Almost the entire quantity of basal fertilizer used by sample farmers consisted of
D compound (10-20-10), and only urea was used for top dressing. The Eastern
Province Agricultural Development Project (EPADP) staff and scientists of the
Adaptive Research Planning Team (ARPT) have consistently emphasized the need
for X compound, but the distribution system has not been able to respond, primarily
because assessment of demand for different components at the national procurement
level has been faulty (Eastern Province, Department of Agriculture 1986). Though
unit prices for both these formulations are about the same, substitution of D com-
pound for X necessitates revision of the recommendations (increasing the quantity of

6In 1985/86, the National Agricultural Marketing Board (NAMBOARD) was responsible for
procurement of fertilizers and participated in distribution along with cooperatives. Subsequently, this
institution was abolished and cooperatives assumed most of NAMBOARD's functions.
17Commercial banks and the Lima Bank are other institutions that provide credit. The Lima Bank is a
public agricultural bank, established in 1987 through the amalgamation of the Agricultural Finance
Company and the Zambia Agricultural Development Bank.

Table 15-Use of different kinds of fertilizers, plateau region

Mixed Fertilizers
D Compound X Compound R Compound
(10-20-10 (20-10-5 (20-20-0 Urea
Zone/Location NPK) NPK) NPK) (46N)
(50-kilogram bag)
Plateau Region 978.0 27 4 1,028.5
North/South Zone
Chipili 170.5 4 ... 176.0
Sinda 13.0 11 2 22.0
Chaweya 122.0 4 ... 118.0
Kasendeka 340.0 3 .. 342.0
Central Zone
Makangila 226.0 2 2 212.0
Mtenthela 21.5 3 ... 31.5
Western Zone
Chiwizi 9.0 ... ... 48.0
Kamwala 76.0 ... ... 79.0

Source: Agricultural household survey conducted in 1985/86 by the International Food Policy Research Institute; the
Rural Development Studies Bureau, University of Zambia; the National Food and Nutrition Commission,
Zambia; and the Eastern Province Agricultural Development Project, in Eastern Province, Zambia.
Note: The ellipses (. .) indicate a nil or negative amount.

urea) and raises the effective cost for the farmer (as mentioned in Chapter 4). In
addition, use of D compound wastes phosphates.
Generally, a one-to-one relationship is maintained between materials meant for
basal and top dressing (Table 15). This is what one would expect because the
recommended combination is usually in this ratio, and at most locations, farmers are
able to maintain this level of application. This suggests that both these types of
fertilizers are available. The exceptions are Chiwizi and Mtenthela, where availability
of basal (mixed) fertilizer was limited. It may be noted that at these locations farmers
were fertilizing a substantial area with a single nutrient, nitrogen, coming from urea,
which was the major fertilizer available (Table 13). This imbalance should be attrib-
uted to inefficient supply conditions rather than inadequate knowledge. Clearly,
unavailability of the right kind of fertilizer is a constraint resulting from poor
coordination between the Department of Agriculture and NAMBOARD (leading to
procurement and supply of an inefficient fertilizer formula to the province), and
logistic failures in the distribution system (leading to supplies that do not reach all

Time of Fertilizer Purchase

The issue of timely availability is examined by looking at the fertilizer purchase
patterns of farmers at different locations. Monthly expenditures of the sample house-
holds were monitored from November 1985 to December 1986. From these records,
expenditures on fertilizer purchase were retrieved for each household. Table 16
shows the monthly distribution of total expenditure on fertilizers. The planting season
starts in November and extends to January. This is the period for basal application;
top dressing extends to March.

Table 16-Monthly fertilizer purchase patterns of farmers

Percent of Total Fertilizer Expenditure
Zone/Location December January February March

North/South Zone
Chipili 92.6 ... 7.4
Sinda 38.8 61.2 ...
Chaweya 84.6 10.9 4.5
Kasendeka 47.4 52.6 ...
Central Zone
Makangila 100.0 .. .
Mtenthela 11.0 33.3 55.7
Western Zone
Chiwizi 72.0 16.0 8.6 4.4
Kamwala 100.0 ......

Source: Agricultural household survey conducted in 1985/86 by the International Food Policy Research Institute; the
Rural Development Studies Bureau, University of Zambia; the National Food and Nutrition Council,
Zambia; and the Eastern Province Agricultural Development Project, in Eastern Province, Zambia.
Note: The ellipses (. .) indicate a nil or negative amount.

All fertilizer purchases were confined to the November-March period.18 At four
out of eight locations, more than 80 percent of the purchases were made by Decem-
ber. At these locations, fertilizer was obviously available at the right time. At two
locations, Mtenthela and Chiwizi, significant purchases were made in February or
March. By then, it is too late for basal application and late even for top dressing.
Supplies were apparently not timely in those locations. These late purchases were
obviously for the top-dressing fertilizer-urea. As was shown earlier, the rate of
adoption is low at these locations (Table 7) and nutrient use is not well balanced
(Table 13). These problems were also noted at Sinda, where very little fertilizer (39
percent) was purchased by the end of December, the most critical period for fertilizer
These data suggest that at some locations fertilizer use indicators could be low
because of the unavailability of the right kind of fertilizer or untimely delivery of
supplies. The latter is more critical for basal fertilizers because the time constraint is
more binding. Top dressing with urea can be done up to 5-6 weeks after planting, but
basal application cannot be delayed beyond 2-3 weeks after planting. To the extent
that the delays in fertilizer purchases are not voluntary, they reflect problems in the
fertilizer delivery system, which arise from institutional constraints. For example,
there may be delays in processing credit applications, or supplies from the main
depots may arrive late. These cannot be categorically addressed with data from this
It is important to mention that delayed fertilizer application (and the consequent
inefficiency in fertilizer use) does not occur due to delayed availability alone. Even

18The survey did not cover the period September-October 1985, but this period was monitored in 1986.
In that year as well, fertilizer purchases started in November and not earlier. Thus, information presented
in Table 16 covers the entire fertilizer purchase season.

in areas where availability and timing problems do not exist, fertilizer application
could be delayed due to labor constraints (ARPT various years). Highest priority is
accorded to planting operations during November and December, leaving farmers
very little time for operations like fertilizer application and weeding, which could be
delayed or inefficiently done or both. The recommended practice of basal fertilizer
application (making a hole 7.5 centimeters deep, 7-8 centimeters away from the plant,
applying a measured quantity of fertilizer, and covering the hole) is a fairly labor-
intensive operation, especially for hoe farmers.
It was not possible to empirically demonstrate this on the basis of the survey
because of data problems.19 Nevertheless, the following results are indicative. It was
assumed that basal fertilizer application about a month after planting would reflect
correct timing. The proportion of fertilized plots on which the timing of basal
application fell within this time band was as follows: Chipili, 6 percent; Sinda, 71
percent; Chaweya, 71 percent; Kasendeka, 11 percent; Makangila, 98 percent; Mten-
thela, 0 percent; Kamwala, 10 percent; and the average for the plateau region, 37
percent. Thus, in four out of seven locations, there was ample indication that the
timeliness of this operation was suboptimal.20 Two of these, Chipili and Kasendeka,
were locations where the rate of adoption and other fertilizer use indicators were high
(Table 7).
On the agricultural research side, the labor-constraint-induced delays have been
addressed in three ways (ARPT various years). First, instead of two applications
(basal and top dressing), only one application of basal and top fertilizers has been
tested and recommended. Second, experiments have been conducted to show that top
dressing with urea can be deferred until 7-8 weeks after planting without detrimental
yield effects. Neither of these findings, however, helps alleviate the labor constraint
at the critical period during November-December. In fact, the first aggravates the
early-season labor bottleneck. Finally, attempts have been made to identify maize
varieties and hybrids that require shorter periods to reach maturity and therefore can
be planted late. This would enable farmers to spread their operations over time and
reduce peak-period labor requirements.
Technical efficiency in fertilizer use practices has been identified as a major
extension theme. Surveys conducted by ARPT have shown that farmers' practices are
inefficient in terms of quantity applied per stand (many farmers use a "handful"
measure that results in uneven application over the field), method of fertilizer appli-
cation (application on surface or uncovered holes), nutrient balance (only top dress-
ing with urea), and timing of application. These have been identified as the reasons
why, despite high absolute levels of fertilizer use, hybrid maize yields remain below
the potential (EPADP 1984). Now that profitability of fertilizer use has been signifi-
cantly eroded by removal of subsidies, the major thrust for extension services should
be to remove these sources of technical inefficiency.

19Dates of different operations on each cultivated plot were obtained in a one-shot crop management
survey toward the end of the season. While critical operations like planting were easily recalled, dates of
other operations reported by farmers were likely to be less reliable.
20These results are also reported in ARPT 1986. The survey, conducted in the South Zone of the plateau,
showed that 40-60 percent of the farmers did not apply basal fertilizer at the right time. Note also that data
on timing of operations were not available for Chiwizi.

Access to Fertilizer

Traditional farming is characterized by very low capital intensities. To ease the
transition to modem agriculture, almost all agricultural development projects, there-
fore, emphasize developing and strengthening credit institutions to provide working
and investment capital for purchase of modem inputs. The cooperative sector and
commercial banks are extending credit to smallholders in the province for purchase
of short-term production inputs (Banda 1991). Based on data from the study survey,
Table 17 provides information on the spread of institutional credit and the extent to
which fertilizer users in the plateau region depended upon borrowed funds for
fertilizer purchases.
Only 15 percent of the sample households in the plateau region used institutional
short-term credit in 1985/86, a figure not very different from the estimate of 17
percent for 1984/85 projected by an earlier survey in the province (EPADP 1986b).
Data for different locations indicate that, at best, about one-third of the households
were covered (Chipili and Mtenthela). On the other extreme, in Sinda and Chiwizi
none of the farmers used any institutional credit. The coverage is not only thin, it is
also highly uneven.
More significantly, Table 17 shows that most of the fertilizer users do not use
institutional credit. Less than one-fourth used credit, on average, although at some
locations (notably Mtenthela), the figures are high. This indicates that the majority of
fertilizer users depend on their own funds or informal credit sources for fertilizer
purchases. Comparing these figures with those on the extent of adoption of fertilizers
in Table 7, it appears that locations where adoption of fertilizer was high did not
necessarily have a high proportion of farmers depending on credit. On the other hand,
at locations where adoption was poorest (Sinda and Chiwizi), access to credit was
also poorest. At locations in the intermediate adoption range (Kamwala and Mten-

Table 17-Credit and fertilizer use, 1985/86

Percent of
Percent of Farmers Fertilizer Users
Obtaining Credit from Obtaining Credit from
Zone/Location Institutional Sources Institutional Sources

Plateau Region 15.2 23.8
North/South Zone
Chaweya 9.1 10.0
Kasendeka 6.1 6.9
Chipili 36.4 36.4
Sinda 0.0 0.0
Central Zone
Mtenthela 37.5 80.0
Makangila 12.1 12.9
Western Zone
Chiwizi 0.0 0.0
Kamwala 18.8 42.8

Source: Agricultural household survey conducted in 1985/86 by the International Food Policy Research Institute; the
Rural Development Studies Bureau, University of Zambia; the National Food and Nutrition Commission,
Zambia; and the Eastern Province Agricultural Development Project, in Eastern Province, Zambia.

thela), however, dependence on credit was high. At these locations 43 and 80 percent
of the fertilizer users, respectively, were borrowers. This suggests that credit support
is perhaps crucial in the early stages of fertilizer diffusion. This initial support helps
to generate surpluses that can then be reinvested by farmers for more fertilizer
purchases or area expansion or cultivation of hybrid maize. This is the pathway
encouraged by EPADP. These inferences are tentative, however. It has been shown
earlier that at some locations (Sinda and Chiwizi) there were also fertilizer distribu-
tion problems. Obviously, low use at these locations cannot be attributed to credit
constraints alone. This theme is pursued more rigorously in the succeeding chapters.
Credit eases the financial constraint to access, but in poorly developed regions,
there are physical barriers to access as well. Roads and transport facilities near the
settlements may be inadequate, making market towns inaccessible. Inadequate facili-
ties also raise the real cost of fertilizers if time and effort dimensions are considered.
The modal distance traveled to purchase fertilizers varies from less than 1 kilometer
at Sinda to nearly 9 kilometers at Chiwizi (Table 18). The apparent relationship
between distance and fertilizer adoption at these locations appears to be weak (Table
7), but in three out of four low-adoption locations-Chiwizi, Mtenthela, and Kam-
wala-farmers have to travel long distances to purchase fertilizers. The fourth loca-
tion, Sinda, is an exception. It has the lowest adoption level, yet the fertilizer sales
point is less than a kilometer away. Nevertheless, a study on access to and use of
different services, based on the same data set, showed that distance was an important
determinant of fertilizer purchases (Wanmali and He 1989).
Table 18 also shows that motorized transport is used in only two out of eight
locations. This reflects the nature of the road network. Ox carts, or in Lundazi, boats
or canoes, are generally used to transport bulky inputs. Oxen and cart ownership is

Table 18-Average distance and mode of transport used for fertilizer
purchases, 1985/86

Percent of Trips Made
Average Ox Cart,
Distance Boat, or Motorized
Zone/Branch Traveled Foot Canoe Transport Bicycle
North/South Zone
Chipili 2.15 11 84 5
Sinda 0.72 56 22 .. 22
Chaweya 1.78 21 74 ... 5
Kasendeka 3.76 7 90 ... 3
Central Zone
Makangila 2.30 11 78 11
Mtenthela 4.33 12 70 18
Western Zone
Chiwizi 8.67 67 ... ... 33
Kamwala 2.63 ... 100

Source: Agricultural household survey conducted in 1985/86 by the International Food Policy Research Institute; the
Rural Development Studies Bureau, University of Zambia; the National Food and Nutrition Commission,
Zambia; and the Eastern Province Agricultural Development Project, in Eastern Province, Zambia.
Notes: Figures pertain to the most frequently used purchase location in each branch. The ellipses (. ..) indicate a nil

limited and, in spite of a rental market, farmers' access to this mode of haulage is
restricted. In some areas where ox carts are not available (for example, Chiwizi),
farmers have to walk or use bicycles. This obviously limits the volume of fertilizer
that can be transported.
Data presented in this chapter indicate wide variations in financial conditions and
physical access to fertilizers within the province. There are indications that these have
an effect on fertilizer use. In the econometric analysis attempted in the next chapter,
the impact of these variables is examined in greater detail.


As articulated in Chapter 2, the process of fertilizer diffusion is governed by
four sets of forces: the nature of the physical response environment, factors respon-
sible for creating effective demand for fertilizers at the farm level, the status of the
fertilizer distribution system, and conditions of aggregate supply of fertilizer at the
national level. Some of these lie in public domain (for example, prices and market-
ing, fertilizer production, imports and distribution, research, extension, education,
and credit) and others relate to agroclimatic conditions and characteristics of the
farm or the farmer (such as education, age, experience, subsistence priorities, and
farm resource endowments).
In a cross-sectional, household-level data set, only a part of these processes can
be captured. Effects on fertilizer use decisions of household-level variables like
education, farm size, size and composition of family, age, gender, and household
participation in credit and extension programs can be analyzed. Also, since the
households are spread throughout the province and the locations vary on access to
institutions and markets and agroecological conditions, these factors can also be
brought in to explain differences in fertilizer use decisions. Other factors, such as
the incentive environment, aggregate supply situation, and availability of new
technologies that shift the fertilizer response curve, are fixed in the short run, and
their effects cannot be deciphered from this kind of data set.
Some information on these fixed factors has been provided in the preceding
chapters. For example, it has been shown that in the plateau region, the dominant
crop of the farming system-maize-is highly responsive to fertilizer. Availability
of suitable hybrids further improves the agronomic potential of fertilizer use. It has
also been shown that the relative fertilizer-output price ratio has historically been
conducive to growth in fertilizer use (Chapter 4). Thus, regarding incentives and
supply, the plateau region was favorably placed in 1985/86 for widespread adop-
tion of fertilizer. Regarding credit and infrastructure, however, some constraints
were noted at certain locations in the plateau, as well as in the valley. These
obviously exert a dampening influence. No information was available from the
survey on the fertilizer distribution system, though deficiencies in timing of sup-
plies and types of fertilizer materials available were observed that may be attrib-
uted to inadequacies in the distribution system (see Chapter 6). Finally, aggregate
supplies at the national level seem to have constrained fertilizer use in Eastern
Province in the second half of the 1980s. Problems with assessment of demand for
different kinds of fertilizer were also noted. These findings provide the context
within which results of the econometric exercise reported in this chapter need to be
interpreted. These econometric analyses aim at rigorous assessment of the effects
of sociocultural, institutional, and environmental factors on fertilizer use decisions
of individual smallholders.

Theoretical Framework and Model

First, farmers decide whether they will use fertilizer (USEFERT). Second, they
make a decision regarding intensity of use, represented here by the rate of fertilizer
application per hectare (NPKFHA). Factors explaining both these decisions are
identical, since both are generated from the same underlying decisionmaking process.
An econometric problem that arises in estimation of these relationships is sample
selectivity bias. This is relevant for the intensity of use variable, but it is not observed
for the sample as a whole. By excluding individuals who do not use fertilizers, the
dependent variable is censored and the residuals do not satisfy the condition that the
sum of the residuals must equal zero (Maddala 1977). This problem is handled in this
study by using the Heckman method, which consists of two steps (Heckman 1974;
1980). First, USEFERT is estimated by using a likelihood function on both fertilizer
users and nonusers. Then probit is applied to maximize this function.21 From the
probit equation the inverse of the Mill's ratio, LAMBDA (X), which is the ratio of the
ordinate of a standard normal to the tail area of the distribution, can be computed
(Heckman 1980). The Mill's ratio reflects the probability that an observation belongs
to the selected sample. In the second step, X is included as an additional variable in
the ordinary least squares estimation of the intensity equation (NPKFHA) for fertil-
izer-using households. This technique eliminates the potential sample selection bias.
If X is not statistically significant, then sample selection bias is not a problem
(Heckman 1979; 1980).
The estimated models in this study are defined as follows:

Probability of fertilizer use (USEFERT) is
Prob (USEFERT=1) = 1 F(-yX), (3)
and intensity of fertilizer use (NPKFHA) is

NPKFHA = Po + P,X + ,2X + E, (4)
for USEFERT > 0, where X is the vector of the regressor and k is the inverse of the
Mill's ratio. Pi and y are vectors of unknown coefficients, and E is the new residual
with the property that E(s) =0 (Maddala 1983).
It is hypothesized that the two decisions (USEFERT and NPKFHA) are influ-
enced by the following set of variables:
1. Personal attributes of the farmer: age, level of education, sex, family subsis-
tence pressure;
2. Farming system and resource characteristics: cultivated area, family size, use
of oxen, hybrid maize cultivation, and liquidity position of the farmer;
3. Institutions and infrastructural factors: access to credit, extension advice,
market and fertilizer supply situations; and
4. Environmental factors: agroecological zones.
The variables used in this analysis are defined in Table 19. Although the terms
"fertilizer use" and "adoption" have largely been used interchangeably in this report,

21For a description and discussion of the probit model, see, for example, Maddala 1983, 22-27.

Table 19-Definitions of variables in the fertilizer use regressions

Variable Definition Meana Deviation

Dependent variables
USEFERT Use = 1; nonuse = 0 0.67 0.47
NPKFHA Total nutrient used per fertilized hectare 68.79 63.72
Personal attributes
AGE Age of the household head (years) 43.24 14.97
HEADHH Sex of household head (male = 0, female = 1) 0.32 0.47
EDPb Up to 5 years of schooling (0, 1, yes = 1) 0.25 0.43
EDSb More than 5 years of schooling (0, 1, yes = 1) 0.33 0.47
DRATIO Number of dependentsC/number of adults 0.73 0.60
Resource characteristics
AREAHA Cultivated area in hectares 2.69 2.48
CULTMETH Oxen use (0,1, yes = 1) 0.64 0.48
FAMILYSZ Number of persons in the family 5.91 3.08
USEHYM Adoption of hybrid maize (0,1, yes = 1) 0.36 0.48
ANCR Total farm sales (100 kwacha) 9.21 17.40
Institutional access
CDD Distance to fertilizer market (kilometers) 15.60 9.37
COOPMEM Membership in a cooperative (0, 1, yes = 1) 0.19 0.39
ADVICE Receipt of extension advice (0, 1, yes = 1) 0.27 0.44
SDUM Fertilizer supply (0, 1, 1 for Sinda,
Kasendeka, Mtenthela, and Chiwizi) 0.46 0.50
Environmental factors
Z2, Z4 Dummy variables (0,1) for North/South and
Western zones

aBased on 199 plateau households.
b"Illiterate" is excluded from the education category, and the Central Zone is excluded from the zone categories.
cDependents are defined as persons below 14 years and those over 60 years.

USEFERT does not measure adoption in a strict sense. The education variables, EDP
and EDS, give better access to information about the new technology. Thus, the
education of the head of the household is expected to have a positive impact on the
decision to use fertilizer. Older farmers are less likely to adopt a new technology, and
this is expected to have a negative effect.
The status of women-headed households has received considerable attention, and
such farms are generally perceived to face more constraints than others. They also are
more subsistence-oriented (see, for example, von Braun, Puetz, and Webb 1989; von
Braun and Pandya-Lorch 1991). Nearly one-third of the sample households are
headed by women, as indicated by the variable, HEADHH.
The more dependents in a family the greater the pressure to provide subsistence
crops; the variable DRATIO was included to capture such effects. On the one hand,
a large family creates more consumption pressure, and, on the other, a larger supply
of labor is expected to affect adoption rates favorably. Thus, it is hypothesized that
the more intense the subsistence pressure, the greater will be the tendency to adopt
the new technology.
Adoption of hybrid maize (USEHYM) is expected to affect the use of fertilizer
positively. As was shown earlier, fertilizer and hybrid maize go together (Table 11).
The economic status of farmers (usually measured as wealth) exercises a positive

influence. In the absence of wealth data, and in the context of African smallholder
farming systems, the area (AREAHA) and oxen use (CULTMETH) variables capture
this effect.22 The shortage of working capital can also be a major constraint to
adoption of the new technology. In this analysis, value of sales of crop and livestock
products during the year (ANCR) has been used as a proxy for farmer's liquidity,
which is expected to have a positive influence on adoption.
Influence of credit is measured in terms of membership in cooperatives (COOP-
MEM). Fertilizer is promoted by cooperatives. If the farmer is a member of a
cooperative, credit and fertilizer are provided as a package. When the farmer buys
fertilizer, the loan for the purchase is granted simultaneously. Thus, membership in a
cooperative is very important in the adoption of fertilizer. Having access to other
sources of credit may not have much effect on purchase of fertilizer, because a farmer
may not know where to buy fertilizer. Of course, membership in a cooperative also
indicates higher socioeconomic status.
Extension contact (ADVICE) is measured by the household's contact with an
extension agent during the year. It is expected to have a positive effect on fertilizer
use. Market access is defined as a location-specific variable and is defined as the
mean distance that a farmer travels to purchase fertilizer (CDD). As the distance to
market increases, the purchase of fertilizer is anticipated to decrease. Fertilizer supply
conditions are also measured in location-specific terms. In Chapter 6, four locations
(Sinda, Kasendeka, Mtenthela, and Chiwizi) were identified as having problems with
fertilizer supplies (SDUM). Thus, unavailability of fertilizer is anticipated to have a
negative influence on adoption. Finally, agro-ecological differences are accounted
for by using dummy variables for the three low zones (Z2 and Z4) (see Chapter 3).
Those zones that consume little fertilizer are expected to have a negative influence.
Full and consistent information on this set of variables is available only for 199 (out
of 262) sample households in the plateau region. Table 19 and all other subsequent
analyses are based on these 199 observations.
Using equations (3) and (4), the empirical model may be specified as

COOPMEM, ADVICE, SDUM, Z,, Z4), and (5)

COOPMEM, ADVICE, SDUM, Z2, Z4, 1). (6)

Thus specified, the model still has a problem. None of the right-hand side
variables used in equations (5) and (6) are truly independent or exogenous. For
example, it has been shown (in Table 11) that hybrid maize and fertilizer go together:
farmers seldom adopt hybrid maize without adopting fertilizer use. Also, in the land-
surplus context of Eastern Province, depending upon the farmers' ability to hire more
labor or oxen during planting time, cultivated area can be expanded. Both options for

22Number of cattle was also tried but it did not prove satisfactory.

increasing output (expansion of cultivated area and use of yield-increasing inputs like
hybrid maize and fertilizers) are available to farmers, and both should be treated as
In subsequent analysis, an approach combining an instrumental variable and the
Heckman approach is used to address these problems. First, AREAHA and USE-
HYM equations are estimated and the predicted area and hybrid maize adoption
variables are employed as instruments in the fertilizer use and intensity equations.
The NPKFHA equation is only estimated for fertilizer-using households.
Thus, AREAHA, USEHYM, USEFERT, and NPKFHA form the endogenous
variable set, and the others (described in Table 19) are treated as exogenous. An
additional variable, SETFMYRS, which determines how long the household has
resided at the present location, has also been included in the instrumental variable
equations. A longer stay provides time to acquire more land and higher socioeco-
nomic status (von Braun, Puetz, and Webb 1989). Since labor availability is crucial
for area expansion, another variable, LOFFEMP, measuring hours of off-farm work
per household (in 100-hour units), was also considered in these equations. The system
is estimated simultaneously and the results are presented in Table 20.
The area and hybrid adoption equations (used as instruments in the fertilizer
equation) provide some interesting insights. None of the personal attributes have any
influence on area. All farmers share the desire to acquire more land and produce
more. Larger labor supply, indicated by larger family size, access to oxen, and higher
liquidity level (which allows access to hired labor and oxen) are important in realizing
this goal. Longer duration of stay at one site is also land-augmenting in the sense that
some land can be cleared every year and land stock can be cumulatively augmented.23
Among institutional factors, better access to markets and credit also encourages area
expansion. The odd result here is the negative coefficient for DRATIO. Adoption of
hybrid maize, however, is influenced by personal attributes-older farmers and
female farmers are reluctant to adopt. Larger family size, higher liquidity level, and
access to oxen affect adoption favorably, as do access to credit and markets. Decline
in family labor supply due to off-farm work inhibits adoption of hybrid maize.
For fertilizer use, the variables age, liquidity, fertilizer supply, credit, and market
access emerge as significant. None of the others (including personal and institutional
factors) affect this decision. Current fertilizer use and adoption of hybrid maize
decisions are independent of farm size. Consistent with earlier results, adoption in the
Western Zone is poor. These findings lend support to the view articulated in Chapter
5 that farmers in the plateau region are now well aware of the benefits of fertilizer use.
Farmers' liquidity, access to credit, and market infrastructure, as well as a functioning
distribution system, are the major factors determining use or nonuse of fertilizer in
any particular year.
The predictive power of the equation explaining intensity of use was very low.
Farmers in the Western Zone used lower levels of fertilizer. A higher level of
education and access to oxen led to higher fertilization rates. The education result is
important when the intensity of fertilizer application is considered, but not in the
decision of whether to use fertilizer. A decision on how much fertilizer to use requires

23This works even in a bush-fallow system where land is cultivated for six-to-seven seasons and then

Table 20-Determinants of fertilizer use and intensity

Estimation of Instruments Fertilizer Use Decisions
Variable Coefficient t-Statistic Coefficient t-Statistic Coefficient t-Statistic Coefficient t-Statistic

Constant 0.060 0.082 -0.355 0.645 2.486*** 3.563 83.459 3.840
AGE -0.007 0.638 -0.022*** 2.446 -0.028*** 2.705 -0.437 1.229
HEADHH -0.188 0.566 -0.382* 1.413 -0.309 1.064 -5.589 0.542
EDP -0.167 0.447 0.304 0.012 0.141 0.407 9.712 0.830
EDS -0.095 0.239 0.045 0.151 -0.421 1.134 17.809* 1.441
DRATIO -0.299* 1.307 0.937 0.504 -0.006 0.027 3.552 0.465
CULTMETH 0.744*** 2.019 0.461* 1.561 -0.153 0.317 37.123*** 2.119
FAMILYSZ 0.221*** 4.303 0.080*** 2.064 0.067 0.715 1.696 0.533
ANCR 0.050*** 5.559 0.027*** 3.252 0.101*** 2.931 0.479 0.708
SETFMYRS 0.389*** 2.359 0.005 0.397 ... ...
AHAT ... ... ... ... -0.186 0.506 -4.501 0.358
MZHATC .. .. ... ... -0.136 0.296 -5.196 0.396
CDD -0.055*** 3.054 -0.255** 1.791 -0.047** 1.860 0.421 0.413
LOFFEMP -0.022 0.856 -0.034** 1.698 ......
COOPMEM 0.826*** 2.153 0.749*** 2.633 1.149** 2.151 14.651 0.858
ADVICE -0.337 1.003 -0.216 0.851 -0.053 0.157 4.521 0.417
SDUM 1.031*** 3.146 0.050 0.198 -0.731* 1.550 -8.176 0.481
Z2 1.280*** 2.844 0.409 1.127 0.813* 1.319 -15.718 0.806
Z4 0.855** 1.746 -0.316 0.774 -0.957** 1.778 -55.547*** 2.507
a ... .. .. 33.798** 1.777
R2 0.391 . ... ... ... ... 0.110
X2 ... ... 76.655 . 111.150 ....
N 199 ... 199 ... 199 ... 133

Notes: The variables are defined in Table 19, with the exception of SETFMYRS, which is the number of years the
family has resided at its present location, and LOFFEMP, which is the hours of off-farm work per household.
The estimation results reported in Table 10 are for fertilizer-using households. Here the estimations are based
on all 199 plateau households.
aEstimated by ordinary least squares.
bEstimated by a probit maximum likelihood procedure.
Predicted values of AREAHA and USEHYM.
dX is the inverse of Mill's ratio obtained from the USEFERT equation.
*Significant at the 20 percent level.
**Significant at the 10 percent level.
***Significant at the 5 percent level.

more knowledge. The significance of X in the intensity equation indicates that
important differences exist between users and nonusers of fertilizers and that these
differences need to be taken into consideration in estimating the intensity equation.

Effects of Crop Management Practices

Crop management practices, like choice of crops, timeliness of operations, effi-
cient weeding, and residual fertilizers, affect the response of crops to fertilizer
application. These variables, therefore, affect the quantity of fertilizer used by farm-
ers, assuming they know the nature of these interactions. The effects of these factors
cannot be captured by the household analysis because these data are plot-specific,
whereas in Table 20 the average rate is derived from the sum of all plots belonging to

the household. The results of a plot-specific tobit regression that seeks to explain the
variation in fertilizer application rates (NPKFHA) using plot-level data from farms in
the plateau are presented in Table 21. Only maize plots are considered. In addition to
the variables included in Table 21, the following plot-specific variables are also

PGHY = Previous crop on the plot (hybrid maize or
groundnuts = 1; others = 0),24
DPLANTIM = Date of planting (weeks),25
NWEED = Number of weedings done on the plot,
DISTHOUS = Distance of the plot from the house (kilometers),
HYV = Maize variety grown (hybrid maize = 1; local =
0), and
PM = Cropping practice followed (sole cropping = 1;
mixed = 0).

Moreover, NPKFHA is now defined as the quantity of fertilizer (kilograms/hec-
tare) used on the plot. The results need to be carefully interpreted because of the
partial nature of the model.
Among the personal attribute variables, in addition to age, the knowledge vari-
ables emerge as significant determinants of actual rates of fertilizer application.
Schooling is important, and a higher level of schooling affects rates positively. The
extension variable also has a significant and positive coefficient. Decisions regarding
rates of application demand more knowledge. As mentioned in Chapter 4, the amount
of fertilizer applied needs to be adjusted depending on specific soil, agronomic, and
management situations. The results here support this. Farmers with a better knowl-
edge base use larger quantities of fertilizer. This result also implies that the extension
service has been effective in overcoming the farmers' tendency to use low rates of
fertilizer application.
As expected, use of fertilizer-responsive hybrid maize results in a sharp increase
in the fertilizer application rate. Delayed planting affects rates adversely. Better-
weeded plots receive more fertilizer. These results imply that farmers are aware of
some of the critical interactions and adjust their fertilizer application rates accord-
ingly. Distant plots, which are more difficult to supervise, are fertilized at lower
levels. These plots could also be more recently cleared fields with relatively high
native fertility, thus requiring less nutrient replenishment. The nature of the cropping
(sole or mixed) does not make any difference in the rate of fertilizer application.26
This finding has implications for maize researchers. It was stated in Chapter 4 that,
despite the widespread prevalence of mixed cropping, little research has been done

24These crops leave residual nutrients that are available to the succeeding crop.
25For each location, the week of first planting was given the number 1. Actual date of planting for each
plot was scaled in relation to this date. Thus, this variable measures the extent of delay in planting and is
standardized over branches (locations) even though the actual dates of first planting are different across
26Rates of application on mixed crops have been found to be lower in a number of agronomic experiments.

Table 21- Plot-level (tobit) analysis of determinants of fertilizer application

Variable Coefficient t-Statistic

Constant 102.346*** 3.523
AGE -0.924*** 2.417
HEADHH -0.488 0.039
EDP 25.315** 1.926
EDS 24.162** 1.830
DRATIO 9.985 1.178
FAMILYSZ -4.185*** 2.181
AREAHA 6.178*** 2.979
ANCR -0.030 0.329
CDD -0.308 0.505
COOPMEM 1.177 0.090
ADVICE 20.517** 1.817
SDUM -37.076*** 3.026
CULTMETH -8.478 0.703
PGHY 26.541*** 2.445
DPLANTIM -11.724*** 4.912
NWEED 16.341* 1.544
DISTHOUS -2.785* 1.402
HYV 95.644*** 7.720
PM -5.563 0.493
X2 209.000
N 431

Notes: The variables are defined in Table 19, except for the following: PGHY is previous crop on plot (hybrid maize
or groundnuts = 1; others = 0), DPLANTIM is date of planting (weeks); NWEED is the number of weeding
done on the plot; DISTHOUS is the distance of the plot from the house (kilometers); HYV is the maize variety
grown (hybrid maize = 1; local = 0); and PM is the cropping practice followed (sole cropping = 1; mixed = 0).
*Significant at the 20 percent level.
**Significant at the 10 percent level.
***Significant at the 5 percent level.

on fertilizer use in this situation. It seems, however, that based on their practices,
farmers don't consider mixed cropping an inferior response.
Some unexpected results are also obtained. The previous crop variable (PGHY)
had a positive and significant coefficient. It was hypothesized that these crops (hybrid
maize and groundnuts) leave some residual nutrients in the soil and, in response,
farmers should use less fertilizer on the succeeding crop grown on the same plot of
land. Table 21 indicates that such plots received more fertilizer in the plots surveyed
here. These crops (hybrid maize and groundnuts) are extremely important from a
market point of view, and plots on which they are grown obviously receive more
attention. These plots may also have been cultivated for a longer period of time;
therefore, farmers may be consciously trying to maintain soil fertility by using more
fertilizer. This is an important hypothesis, since population pressure will continue to
erode the traditional fertility-maintenance practice of fallowing. Fertilizers must
assume greater significance in the future. If this result does indeed reflect this
phenomenon, it provides an indication of how farmers are responding to intensifica-
tion pressures. Even though the region is characterized as land-abundant, fertility and
intensification-related issues are relevant in plot-specific circumstances. This hy-
pothesis has a bearing on the sustainability versus mining arguments often put

forward in literature pertaining to intensification of forest-fallow and bush-fallow
In contrast to the result in Table 20, cultivation of the plot with the help of oxen
did not lead to higher rates of application. This could be due to explicit inclusion of
the farm size variable (AREAHA), which was highly significant and positive. This
could also have affected the family size variable, which emerges as a significant
negative factor. The relationships between cultivated area, method of cultivation, and
labor are complex, and Table 21, which is based on a single-equation framework,
obviously fails to account for these.
These analyses reveal that among the personal attributes, only old age constrains
the decision to use fertilizers. Factors like education, gender, or extension contact and
farmer's resource endowments of land, labor, and oxen do not affect this decision
significantly. Agroecological differences within the plateau region are important, as
illustrated by the Western Zone. This is a challenge for agricultural research. Other
variables that matter in this regard are access to funds (ANCR and COOPMEM),
availability of supplies (SDUM), and access to market (CDD).
These findings must be put in proper perspective. Farmers in the study area have
been experimenting with fertilizers for a long time. A favorable incentive environ-
ment has encouraged this process. Knowledge regarding fertilizers is well diffused
and is accessible to all categories of farmers. Many of the sociocultural and economic
factors usually associated with adoption of innovations are no longer relevant. The
only constraints are those of liquidity, in order to buy fertilizers, and accessibility to
markets and supplies. However, over a large part of Africa (indeed, even within
Eastern Province in the valley and the Western Zone of the plateau), most of these
favorable conditions do not obtain. In these situations, factors such as research,
extension, and education may assume greater significance.
The roles of education and extension are clearly revealed in analyses pertaining
to decisions on rates of fertilizer application. These depend on specific agronomic
and management conditions, which vary from plot to plot. Farmers need more
knowledge to efficiently respond to these situations. Thus, knowledge-related vari-
ables assume greater significance. The analyses indicate that farmers, in fact, do
adjust their rates of application in response to factors such as planting time, variety,
and weeding. These results imply that in the plateau region, the extension services
should concentrate more on educating farmers about fine-tuning their fertilizer use
practices to further improve the efficiency of fertilizer use. Farmers in the plateau
region are generally well beyond the initial phase, where convincing them to use
fertilizers is the prime extension task.
In terms of the role of state policy and public investments, strong results were
obtained. Education, credit, infrastructure, and fertilizer distribution all affect deci-
sions regarding fertilizer, but current levels of these investments are low (NCDP
1989). Extension plays an important role in improving fertilizer use efficiency. In
Eastern Province, extension and research that feeds into the extension system have
received some support in recent years through a World Bank-supported agricultural
development project. Underlying all of these results has been the government support
of maize prices and marketing policy. Without these-and, of course, the favorable
fertilizer response environment-fertilizer use would not have taken off in the
province. Focusing exclusively on household-level regressions would completely
bypass these important preconditions.

To summarize, both area expansion and yield improvements appear to be occur-
ring simultaneously over most of the plateau region. This supports the view that even
in land-surplus regions, yield-increasing inputs can play an important role (Bin-
swanger and Pingali 1988; Lele and Stone 1989). There are two key provisos,
however. First, a high natural potential for the use of such technologies must exist,
and second, committed support for marketing and infrastructure and remunerative
prices must prevail. In the initial phases, there is no substitute for the role of the state
in providing these and other services such as extension, veterinary services, credit,
and even input distribution.



It was hypothesized in Chapter 2 that the process of transformation from subsis-
tence to commercial agriculture in the plateau region is triggered by the use of
fertilizer on local maize. The resulting surplus is then used to cultivate more land and
put more area under a purely commercial hybrid maize crop. Since plenty of arable
land is available and animal traction is feasible, the process culminates in larger
holdings and substantial commercial production. This perception of the transition
process is reinforced by (1) the fertilizer allocation behavior of the sample house-
holds-some farmers were using fertilizers for local maize, some for both local and
hybrid maize, and some for hybrid maize alone (Table 9), and (2) the increase in the
importance of hybrid maize with farm size (Table 14).
Analyzing these distinct patterns of fertilizer allocation behavior should provide
insight into the process of transition. By classifying farmers in categories according
to this criterion (fertilizer use) and trying to understand the factors responsible for
differences between the groups, it should be possible to identify some elements of this
dynamic transition process. It should be noted, however, that fertilizers are only a part
of this process; therefore, the perspective derived from this analysis is partial.
The transition process is analyzed with the help of an ordered probit model,
which attempts to explain a household's progressive movement from no fertilizer use
to fertilizer use on hybrid maize only, representing the progression from subsistence
to commercial agriculture. The dependent variable is defined in an ordered form as
Number of
Category Households

No fertilizer use = 0 68
Fertilizer use on local maize only = 1 64
Fertilizer use on local and hybrid maize = 2 57
Fertilizer use on hybrid maize only = 3 10

The set of independent variables includes personal (farmer-related) and resource
endowment characteristics. Institutional and agroecological factors are also consid-
ered. Variables like cultivated area, adoption of hybrid maize, oxen, and farmers
income and liquidity are all part of the transition phenomenon being explained and
are not included as explanatory factors.
Data on mean levels of the relevant variables in each of the four categories are
presented in Table 22, and results of the ordered probit regression are provided in
Table 23. A trend can be discerned from some of the variables in Table 22. For
example, all indicators of transition--cash income, cash sales, cultivated area, and

Table 22-Mean values of variables in each fertilizer user category

Fertilizer Use
Local Local and Hybrid
Maize Hybrid Maize
None Only Maize Only
Variable (0) (1) (2) (3)

CASH INCOME/CAPITA 40.35 133.47 305.89 157.49
ANCR 2.43 6.64 19.54 12.91
USEHYM 0.01 0.06 1.00 1.00
AREAHA 1.84 1.66 4.21 6.36
CULTMETH 0.53 0.53 0.86 0.90
AGE 46.91 42.15 39.98 44.00
HEADHH 0.43 0.33 0.21 0.20
EDP 0.20 0.25 0.28 0.30
EDS 0.23 0.31 0.47 0.20
DRATIO 0.72 0.66 0.81 0.70
FAMILYSZ 6.04 4.98 6.39 8.20
COOPMEM 0.07 0.19 0.26 0.60
ADVICE 0.20 0.34 0.30 0.10
CDD 19.60 13.92 13.63 10.47
Number of observations 68 64 57 10

Notes: The variables are defined in Table 19, with the exception of CASH INCOME/CAPITA, which is defined as
total cash sales from all sources in kwacha.

Table 23-Results of an ordered probit regression explaining transition to
commercial agriculture

Four Categories (0, 1, 2, 3)
Variable Coefficient I-Statistic

Constant 1.721*** 3.116
AGE -0.019*** 2.561
HEADHH -0.397** 1.734
EDP 0.149 0.649
EDS -0.180 0.714
DRATIO 0.004 0.022
FAMILYSZ 0.078*** 2.718
COOPMEM 0.684*** 2.946
ADVICE -0.158 0.680
CDD -0.038*** 3.358
SDUM -0.465*** 2.000
Z2 0.308 1.091
Z4 -0.636** 1.877
X2 76.727
N 199

Note: Variables are defined in Table 19.
*Significant at the 20 percent level.
**Significant at the 10 percent level.
***Significant at the 5 percent level.

use of hybrid maize-rise with ascending categories. This corroborates the proposed
conceptual framework: the categories do seem to reflect a transition pathway. The
proportion of female-headed households declines, whereas the share of households
with primary schooling, use of oxen, and membership in a cooperative increases as
agriculture becomes more commercialized, as expected. The income indicators of the
most progressive group (category 3) are lower than the trend, but other variables
generally conform to expectations. The regressions reported in Table 23 identify the
main forces more clearly.
Older farmers and households headed by females find it difficult to ascend the
commercialization hierarchy. Although neither fertilizer adoption nor intensity of use
is influenced by gender, as shown in Chapter 7, women farmers are more likely to
fertilize local maize only (Table 10). And women farmers find it difficult to graduate
to higher levels of fertilizer use (Table 23), possibly because they are reluctant to
grow hybrid maize (Table 20). It has been argued that poor processing and storage
qualities make hybrids less preferable for domestic consumption (Jha, Hojjati, and
Vosti 1991; Kydd 1989), and female-headed households place greater emphasis on
family food production than on sales. Women farmers may also be at a disadvantage
in other respects. Since households headed by women are generally short of adult
labor, particularly male labor, women farmers may not have time to be involved in
cash transactions that require travel, interactions with institutions primarily oriented
toward dealing with men, and time-consuming cash generating and purchasing activi-
ties. In Eastern Province, cash transactions are primarily associated with men; farm-
ing and household chores are the domain of women.
Education does not emerge as a significant determinant of fertilizer use. This may
be due to the partial nature of the model, which focuses only on fertilizers. A large
family, which denotes higher consumption needs and a better labor supply, exerts a
strong positive influence. Better infrastructure and availability of credit also play
important contributory roles. Obviously, inefficiencies in input distribution dampen
the process of commercialization.
The effects of another significant state intervention-animal disease control-
should be mentioned. The entire province is vulnerable to tsetse infestation, and
constant vigilance has to be maintained to contain incidence of the disease in the
plateau region. The valley cannot sustain any cattle population because of the disease.
Government-sponsored animal disease control measures such as quarantine, sprays,
dips, and veterinary clinics have contributed to greater use of animal traction in the
farming system. Despite such efforts, cattle rearing remains risky, particularly in the
Western Zone. In Petauke and Nyimba districts, which account for most of the
Western Zone area, for example, the cattle population declined by 37 percent between
1982 and 1984 because of cattle mortality (Eastern Province, Department of Agricul-
ture 1986a). Although the data from the survey do not show lower levels of oxen use
in the Western Zone (Table 2), higher risk levels could adversely affect cash input use
(for example, for fertilizer). These relationships need to be explored in future studies.
Three significant inferences arise from these results. First, the prospects for
progress for women farmers are constrained. Because nearly a third of the small-
holder farms in the study area are headed by women, this issue clearly requires more
attention in research and in public programs. Second, factors external to the farming
system are crucial. Credit, infrastructure, input distribution, animal disease eradica-
tion, and of course, marketing, play decisive roles in the transition process. Many of

these are outside the normal jurisdiction of the Ministry of Agriculture. A strategy for
agricultural transformation that ignores these investments, however, is unlikely to go
far. Finally, agroecological characteristics play a significant role. It is important to
identify these system constraints.
This pattern also lends credibility to the concept of the "lima ladder" put forth by
Zambian researchers (see Chapter 2 and Appendix 1). The initial step is to get
subsistence farmers to produce a small area of a cash crop-soybeans or sunflowers,
and to use the proceeds to buy fertilizers. In the next step (season), this fertilizer is
applied to local maize. One argument proposes that the farmer who uses fertilizer can
plant a smaller area in local maize because, with fertilizer, he can produce enough on
a smaller area for subsistence. He can use the labor, thus released, to grow hybrid
maize. Subsequently, sales of hybrid maize would give him enough cash to hire more
labor or oxen to increase the area under hybrid maize further. The actual progression
appears to be similar, only the initial cash crop component seems to be weak. Very
few of the sample households grow soybeans, sunflowers, or cotton, perhaps because
the seed supply and marketing arrangements for these new crops are weak. At least
some farmers manage to skip this step and somehow (perhaps with credit) arrange to
acquire fertilizer, which is put on local maize. Subsequent steps of increased hybrid
maize production and expansion of cultivated area follow.
The plateau region of Eastern Province illustrates how policy-led intensification
works (Lele, Christiansen, and Kadiresan 1989; Binswanger and Pingali 1988). By
assuring that prices are remunerative and backing them up with effective procure-
ment, the state has induced intensification in a land-surplus environment. Use of
yield-increasing inputs (hybrid seeds and fertilizers), as well as area expansion, are
proceeding side by side. This is probably not happening in the valley region because
of a number of constraints-both agroclimatic (uncertain responses and prevalence
of tsetse) and infrastructural (lack of adequate roads and services). The lesson is to
target investments and interventions only after a careful analysis of farming system
characteristics. Animal disease control is a critical step in this process.
Some important conclusions have emerged from the analyses in the last two
chapters. First, among the personal attributes, age has the greatest negative influence
on fertilizer use. And, although much concern has been voiced on the constraints
faced by women farmers, this analysis does not show them to be lagging behind in
fertilizer use, though other factors such as their reluctance to grow hybrids do inhibit
their progression toward increased commercialization. Education has a positive effect
in more sophisticated fertilizer use decisions (application rates, for example). In an
area where farmers have long been exposed to fertilizers, extension advice does not
seem to influence the decision to use fertilizers, but there is some indication that
extension efforts have an effect on decisions that require more technical knowledge,
such as rates of application. Farmers also seem to be aware of the need to adjust
fertilizer use levels in response to deviations from optimal crop management prac-
tices. Extension has obviously contributed to the improved technical efficiency of
fertilizer use by communicating these messages to farmers.
Among the institutional factors, access to credit, infrastructure, and input distri-
bution are found to be most important, and most of these determinants are in the
public domain. Credit, infrastructure, extension, and education are all provided by the
state. In Zambia, fertilizer distribution is also a state activity. State support has also
enabled animal disease eradication measures and the resulting growth in use of

animal traction. Clearly, more public investment in the rural sector is crucial for
promoting further growth in fertilizer use and output in the Eastern Province.
The state also plays a vital role in pricing and marketing areas. Through most of
the 1980s, it was successful in maintaining a favorable incentive environment. In the
current context of higher fertilizer prices and pressures to minimize the role of the
state, some of these critical functions and roles need to be provided for.



Results presented in the preceding chapters establish quite clearly that the plateau
region of Eastern Province is an area with high potential for fertilizer use, and the
process of diffusion of fertilizer use is well on its way. The other region, the valley,
faces agroecological and infrastructural constraints that inhibit fertilizer use. Over
two-thirds of the sample households in the plateau region were fertilizer users in
1985/86, and more than half the cultivated area was fertilized. At 53 kilograms of
total plant nutrients per hectare of cultivated land, the rate of application in this region
is way ahead of that in the country as a whole, which averages about 20 kilograms.
Fertilizer use, which started in the late 1960s, has grown rapidly in Eastern Province
ever since. A high physical response environment, favorable relative prices, and the
ability to procure farmers' surpluses have all contributed to this development. The
experience in the plateau region demonstrates that even in a highly land-surplus
situation, yield-increasing inputs are a viable option, if the above conditions are met.
The process has been accompanied and probably fueled by concomitant growth in
cultivated area (due to use of animal traction) and growth in area under highly
fertilizer-responsive hybrid maize.
These trends highlight two important policy issues. First, the gap between the
plateau and the valley strengthens the argument for regional targeting of scarce
fertilizers to areas of high use potential. Without significant new investments in
agricultural research, infrastructure, and the marketing and distribution network,
fertilizer promotion efforts in the valley would be fruitless. Second, the importance
of incentives and market infrastructure in promoting fertilizer use should be stressed.
Farmers have responded to these opportunities, when available, and have achieved
significant gains in surpluses and incomes. More will be said on this later.
A significant finding of this study is the total domination of the fertilizer scene by
one crop, maize. Fertilizer use has diffused to varieties other than hybrids, but it is
still almost exclusively based on one crop. This has important implications for future
growth in fertilizer use and output. Since the gap between recommended and actual
levels of fertilizer application on maize is not very large, that leaves three major
sources of future growth in fertilizer use: increase in fertilized local maize area,
increase in hybrid maize area, and absolute increase in cultivated area itself. Avail-
ability of surplus land ensures that if the economic environment remains favorable,
these growth sources will remain important. Changes in relative fertilizer prices
beginning in 1989/90 threaten growth in fertilizer use and output of maize. Thus,
from the longer-term point of view, diversification of fertilizer use to other crops is
crucial. Cotton and carioca beans have been highly responsive to fertilizer, and use
on these crops needs to be encouraged. Research on other crops and regions also
needs to be strengthened. In particular, adequate attention should be paid to relative
fertilizer-output prices.

The evidence obtained in this study on the roles of education, credit, input
distribution, and infrastructure in promoting fertilizer use and transformation of
subsistence agriculture is strong. All these fall in the public domain, as does maize
marketing. The state has also played a role in expansion of animal traction in the
plateau through animal disease control. These social, institutional, and physical
infrastructure investments are extremely important and should be accorded high
Agricultural extension is another area of public intervention that has been ac-
corded high priority in recent years. Substantial efforts have been made to reorganize
the provincial extension system along the lines of the training and visit system
supported by the World Bank. Perhaps because fertilizer use is already well diffused
in the plateau region, the cross-sectional analysis failed to capture any effect of
extension on adoption. However, there was evidence that decisions on rates of
application were influenced by extension advice. As the experimental evidence
shows, response to fertilizer is significantly influenced by other crop management
practices such as time of planting, weeding, and residual fertilizer effects. The
analysis here shows that farmers were aware of some of these interactions and
adjusted their fertilizer use practices accordingly. This also indicates that the exten-
sion services were able to convey these messages. What is needed now is more site-
and system-specific advice on fertilizer use. The focus in extension should shift from
promoting fertilizers to fine-tuning fertilizer use practices, at least in the plateau
region and for crops other than maize. In other areas (like the valley) where fertilizer
experience is deficient, the on-farm experimentation and demonstration roles of
extension could be important. Older farmers and women who operate farms continue
to be somewhat rigid in their responses to innovations. Over time and with education
this will change, but in the short term these groups should receive special attention in
extension programs.
Turning to the issue of agricultural research, the capacity to conduct on-farm
research is severely constrained. There is practically no capacity at the Masumba
substation in the valley region. The current needs of farmers include location-specific
recommendations and information on fine-tuning of fertilizer use practices, nutrient
balance, efficient forms of fertilizer, and use of fertilizer on crops other than maize,
all of which require more on-station and on-farm research. Prices of different fertil-
izer mixtures, for example, should reflect the cost of their nutrient content. Even
within the plateau, the Western Zone had low fertilizer use indicators, which should
be investigated in depth.
Analysis of the patterns of fertilizer allocation decisions by farmers has provided
useful insights into the process of transformation going on in the plateau region of the
province. The transition from subsistence to commercial agriculture was triggered by
the use of fertilizer on local maize. The resulting surplus fueled expansion of area and
hybrid maize production. Availability of surplus land and oxen facilitated the pro-
cess. To abet this process, the crucial first step of finding cash to buy fertilizers could
be aided by credit, or, as proposed under the lima ladder concept, by enabling farmers
to grow small areas of cash crops. Both concepts require strong support. Education,
credit, infrastructure, and input supplies, and potential for oxen use are identified as
some of the facilitating forces behind the transition process. It is useful to bear in
mind, however, that this recipe depends on a set of initial conditions-a high physical
response environment, abundant land and the means to cultivate it, and a favorable

incentive structure. In this setting, yield-increasing technologies go hand-in-hand
with area expansion.
Finally, a word about women farmers who constitute about one-third of the
farming community. This analysis clearly shows that they do not lag behind in terms
of adoption of fertilizers. However, their ability to move up to higher levels of
commercialization is constrained by their reluctance to grow hybrid maize. Policies
to overcome this reluctance need to be found.
Growth in fertilizer use in Eastern Province has some useful lessons for areas
with similar farming systems in Zambia and other parts of central and southern
Africa. A number of factors contribute to the diffusion of fertilizers and greater
commercialization of the farming sector, and these imply investments. The following
play important roles:
favorable price and physical response environments;
investments in market and transport infrastructure and animal disease control;
agricultural research to enhance fertilizer use potential;
investments in education, extension, and credit; and
development of economic institutions for surplus mobilization and input deliv-
ery systems.
These suggest several actions that are important in policy design. However, the
results of the study are limited inasmuch as the recommendations cannot be priori-
tized. A policy of intervention to improve the efficiency of public expenditures is a
major concern of the Zambian government's structural adjustment program. In set-
ting priorities, the government also must take into account the expenditure require-
ments of different actions.


Mount Makulu National Research Station suggested the name "lima ladder" for
a practical series of steps a smallholder can take to improve his position from that of
a subsistence cultivator barely growing enough maize for his family to a hybrid maize
grower with a good cash income, and without waiting for credit that may never come
(see Figure 4).
The plan is to grow more late season crops like soybeans or sunflowers or beans
to raise cash to buy fertilizer for local maize, then to grow less area of local maize (but
with fertilizer) so that the yield is greater than before, which enables the farmer to
have more time (previously given to a large local maize area) to grow more cash crops
to sell or to grow hybrid maize.

Figure 4 -Lima ladder

Oxen (larger area)
Season 4
Farmer hires oxen; grows 4 limas of
hybrid maize

Season 3
Farmer halves local maize area, adds 2 limas of hybrid
maize (fertilizer or cash from surplus maize, and
sunflower or soybean crop sold)

Season 2
Farmer sells (or barters) 4 bags of sunflower seeds for 4 bags of
fertilizer for his local maize (or buys fertilizer with cash from 2 bags of
Season 1
Farmer grows 1 lima of soybeans or 1-2 limas of sunflowers (the farmer is more likely
to have labor for this in late December/early January rather than trying to grow more
groundnuts to sell, which need labor when the farmer is busy planting and weeding local

Last Season
Subsistence farmer with no cash crop and no fertilizer for his
local maize

Source: Based on Eastern Province, Department of Agriculture, "Plateau Camps with Cotton," Bulletin No. 2,
Chipata, Zambia, December 1986.
Note: A lima is a small plot of land (0.25 hectare).


Table 24-Location of sample households

/ B h Number of Households
Region/ Branch
Agricultural District (Location) Branch Sample

Eastern Plateau Region
Chadiza Chipili 221 33
Chipata South Makangila 405 31
Chipata North Mtenthela 214 33
Katete Sinda 247 33
Lundazi Chaweya 138 33
Kasendeka 148 33
Nyimba Kamwala 306 33
Petauke Chiwizi 257 33
Luangwa Valley Region
Chama Nkhoka 89 33
Mambwe Mphata 159 33
Total 2,184 328

Source: Agricultural household survey conducted in 1985/86 by the International Food Policy Research Institute; the
Rural Development Studies Bureau, University of Zambia; the National Food and Nutrition Conunission,
Zambia; and the Eastern Province Agricultural Development Project, in Eastern Province, Zambia.
aA branch is a cluster of villages selected from a district in Eastern Province for the survey above.

Table 25-Marketed production of maize, groundnuts, and cotton in Eastern
Province, 1978-87

Marketed Production
Year Maize Groundnuts Cotton
(1,000 metric tons)

1978 69.4 1.7 2.8
1979 46.6 2.2 3.4
1980 66.6 1.4 3.7
1981 106.6 0.9 1.8
1982 114.6 0.7 1.5
1983 143.8 0.9 2.6
1984 166.7 1.0 4.8
1985 176.8 4.7 3.7
1986 224.3 6.2 2.9
1987 189.8 8.2 n.a

Source: NCDP (National Commission for Development Planning), New Economic Recovery Program, Fourth
National Development Plan: 1989-1993 (Lusaka, Zambia: Office of the President, 1989).

Table 26-Adaptive Research Planning Team (ARPT) trials, 1982/83-1987/88

Title of Trial Period of Trial

Maize/groundnut intercropping
Maize fertilizer
Sunflower variety fertilizer
Sorghum variety
Sorghum/sunflower/legume intercropping
Sorghum planting method
Shorter-duration maize hybrids
Maize open-pollinated variety trials
Sunflower planting, date/thinning
Sunflower thinning
Cotton fertilizer
Finger millet variety test
Maize management factorial
Maize weeding (on station)
Maize weeding (on farm)
Maize field factor (pop/fert)
Live mulch (Mwase CDA)
Rice fertilizer
Rice variety
Rice planting method
Sorghum/maize/soya intercropping
Bean exploratory (nonclimbing)
Bean exploratory (climbing)
Residual N/rotation
Striga control
Striga suppression
Soybean fertilizer
Finger millet variety trial
Relay cropping trial
Cotton/groundnut intercropping
Finger millet management
Soya seed rate
Pigeon pea variety
Groundnut factorial (with CRT)
Soil conservation (obser. trials)
Sweet potato variety trial
Cowpea variety trial
Alternative late season crops trial

1982/83, 1983/84
1982/83, 1983/84, 1986/87, 1987/88
1982/83, 1983/84
1983/84 (failed)
1982/83 to 1985/86
1982/83 to 1986/87, 1987/88
1984/85 to 1985/86
1983/84 to 1984/85
1982/83 to 1984/85
1982/83 to 1985/86
1982/83 to 1984/85
1985/86 to 1986/87
1982/83 to 1986/87, 1987/88
1982/83 to 1986/87, 1987/88
1982/83 to 1986/87, 1987/88
1982/83 to 1986/87, 1987/88
1983/84 to 1985/86
1984/85 (failed)
1983/84 to 1986/87, 1987/88
1985/86 to 1986/87, 1987/88
1983/84 to 1986/87
1984/85 to 1986/87
1985/86 to 1986/87
1984/85 to 1986/87
1985/86 to 1986/87, 1987/88
1986/87, 1987/88
1986/87, 1987/88
1986/87, 1987/88
1986/87, 1987/88
1986/87, 1987/88
1986/87, 1987/88

Source: EPADP (Eastern Province Agricultural Development Project), "Eastern Province Agricultural Development
Project: Project Completion Report" (Chipata, Zambia, November 1987, mimeographed).
aF indicates on-farm and ST on-station trials.

Table 27-Fertilizer and maize prices in Zambia, 1980/81 to 1989/90

Price of Average
Price 10-20-10 Nutrient Maize
Year of Urea Compound Pricea Price

1980/81 0.19 0.19 0.45 0.13
1981/82 0.22 0.23 0.53 0.15
1982/83 0.30 0.30 0.70 0.18
1983/84 0.48 0.48 1.12 0.20
1984/85 0.52 0.52 1.18 0.27
1985/86 0.54 0.53 1.25 0.61
1986/87 1.30 1.60 3.42 0.87
1987/88 1.30 1.60 3.41 0.89
1988/89 1.42 1.97 4.41 1.39
1989/90 7.68 7.92 18.25 3.16

Source: Fertilizer prices from FAO (Food and Agriculture Organization of the United Nations), FAO Fertilizer
Yearbook (Rome: FAO, various years). Maize prices from L. Mumeka, "Policy Constraints on the Devel-
opment of Agriculture and Small-Scale Enterprises in Rural Zambia," in Adopting Improved Farm
Technology: A Study of Smallholder Farmers in Eastern Province, Zambia, ed. R. Celis, J. T. Milimo
and S. Wanmali, 62-92 (Washington, D.C.: International Food Policy Research Institute, 1991).
aDerived as the average of the nitrogen price of urea (46N) and the price of a kilogram of nutrient mix of 10-20-10
compound. These are the two major fertilizers used in Eastern Province, and they are applied equally.


ARPT (Adaptive Research Planning Team), Eastern Province Agricultural Develop-
ment Project. 1986. Annual report, 1985-86. Chipata, Zambia. Mimeo.
Various years. Annual report. Eastern Province Agricultural Development
Project, Chipata, Zambia. Mimeo.

Banda, M. 1991. Credit policies and agricultural credit in Eastern Province. In Adopting
improved farm technology: A study of smallholder farmers in the Eastern
Province of Zambia, ed. R. Celis, J. T. Milimo, and S. Wanmali, 93-128.
Washington, D.C.: International Food Policy Research Institute.
Binswanger, H., and P. Pingali. 1988. Technological priorities for farming in Sub-
Saharan Africa. Research Observer 3 (January): 81-98.

Braun, J. von, and R. Pandya-Lorch, eds. 1991. Income sources of malnourished people
in rural areas: Microlevel information and policy implications. Working Paper
on Commercialization of Agriculture and Nutrition 5. Washington, D.C.: Inter-
national Food Policy Research Institute.

Braun, J. von, D. Puetz, and P. Webb. 1989. Irrigation technology and commercializa-
tion of rice in The Gambia: Effects on income and nutrition. Research Report 75.
Washington, D.C.: International Food Policy Research Institute.

Celis, R., J. T. Milimo, and S. Wanmali, eds. 1991. Adopting improvedfarm technology: A
study of smallholder farmers in Eastern Province, Zambia. An IFPRI occasional
paper. Washington, D.C.: International Food Policy Research Institute.
Desai, G. M. 1969. Growth in fertilizer use in Indian agriculture: Past trends and future
demands. International Agricultural Development Bulletin 28. Ithaca, N.Y.:
Cornell University Press.
1988. Development of fertilizer markets in South Asia: A comparative over-
view. Paper prepared for the International Fertilizer Association Joint Middle
East/South Asia Fertilizer Conference, Lahore, Pakistan, December 3-6. Mimeo.

Desai, G. M., and V. Gandhi. 1989. Phosphorus for sustainable agricultural growth in
Asia: An assessment of alternative sources and management. Paper prepared for
the Symposium on Phosphorus Requirements for Sustainable Agriculture in
Asia, International Rice Research Institute, Los Bafios, Philippines, March 6-10.
1990. Fertilizer consumption in Sub-Saharan Africa: An analysis of growth
and profile of use. In Fertilizer policy in tropical Africa: Workshop proceedings,
Lomn, Togo, April 1988, ed. T. B. Tshibaka and C. A. Baanante, 118-135.
Muscle Shoals, Ala., U.S.A.: International Fertilizer Development Center.

Desai, G. M., and B. Stone. 1987. Rural development strategies in China and India: A
comparative perspective on fertilizer policy requirements for long-term growth
and transitional needs. Paper prepared for the International Symposium for Rural
Development Strategies, Beijing, China, October 25-29.

EPADP (Eastern Province Agricultural Development Project). 1984. Agricultural pro-
duction systems of the Lundazi District with special reference to the Mwase
cattle development area. Research and Evaluation Unit, Chipata, Zambia.

1986a. A report on Eastern Province ARPT work (Project Implementation
Workshop). Petauke, Zambia. Mimeo.

1986b. Summary report on small farm surveys 1984-85. Research and Evalu-
ation Unit. Chipata, Zambia. Mimeo.

1987. Eastern Province Agricultural Development Project: Project completion
report. Chipata, Zambia. Mimeo.

Eastern Province, Department of Agriculture. 1986a. Plateau camps with cotton. Bulle-
tin No. 2. Chipata, Zambia. December.
1986b. Guidelines on fertilizer use in Eastern Province farming systems.
Chipata, Zambia. Mimeo.

Various years. Annual report. Chipata, Zambia. Mimeo.

Feder, G., R. E. Just, and D. Zilberman. 1985. Adoption of agricultural innovations in
developing countries: A survey. Economic Development and Cultural Change
33 (2): 255-298.

FAO (Food and Agriculture Organization of the United Nations). 1981. Agriculture:
Toward 2000. Rome.

1986. Atlas of African agriculture. In African agriculture in the next 25 years.

Various years. FAOfertilizer yearbook. Rome.

Harvey, R. H. 1973. Some determinants of the agricultural productivity of rural house-
holds: Report of a survey in Kalichero District, Eastern Province. Msekera
Regional Research Station, Chipata, Zambia. Mimeo.

Heckman, J. J. 1974. Shadow prices, market wages and labor supply. Econometrica 42
(4): 679-694.

1979. Sample selection bias as specification error. Econometrica 47 (Janu-
ary): 153-161.

1980. Sample selection bias as a specification error. In Female labor supply:
Theory and estimations, ed. J. P. Smith, 206-249. Princeton, N.J.: Princeton
University Press.

International Food Policy Research Institute, Rural Development Studies Bureau, East-
ern Province Agricultural Development Project, and National Food and Nutri-
tion Commission. 1986. Agricultural household survey, Eastern Province, Zam-
bia, 1985-86. International Food Policy Research Institute, Washington, D.C.
Computer printout.

Jha, D. 1991. Input use and productivity. In Adopting improved farm technology: A
study of smallholder farmers in Eastern Province, Zambia, ed. R. Celis, J. T.
Milimo, and S. Wanmali, 202-212. Washington, D.C.: International Food Policy
Research Institute.

Jha, D., B. Hojjati, and S. Vosti. 1991. The use of improved agricultural technology in
Eastern Province. In Adopting improved farm technology: A study ofsmallholder
farmers in Eastern Province, Zambia, ed. R. Celis, J. T. Milimo, and S. Wan-
mali, 173-201. Washington, D.C.: International Food Policy Research Institute.

Katongo, K. 1988. Agricultural sector policies in Zambia: 1975-1985. Paper prepared
for the Seminar on Growth and Equity in Zambian Agriculture: An Eastern
Province Study, Livingstone, Zambia, December 8-11.

Kydd, J. 1989. Maize research in Malawi: Lessons from failure. Journal of Interna-
tional Development 1 (1): 112-144.

Lele, U., and S. W. Stone. 1989. Population pressure, the environment, and agricul-
tural intensification: Variations on the Boserup hypothesis. Managing Agricul-
tural Development in Africa (MADIA) Discussion Paper 4. Washington, D.C.:
World Bank.

Lele, U., R. E. Christiansen, and K. Kadiresan. 1989. Issues in fertilizer policy in Africa:
Lessons from development programs and adjustment lending. Managing Agri-
cultural Development in Africa (MADIA) Discussion Paper 5. Washington,
D.C.: World Bank.

Maddala, G. S. 1977. Econometrics. Economic handbook series. New York: McGraw-

1983. Limited dependent and quantitative variables in economics. New York:
Cambridge University Press.

Maida, J. H. A., and E. S. Malinda. 1990. Fertilizer supply and demand in Malawi. In
Fertilizer policy in tropical Africa: Workshop proceedings, Lomd, Togo, April
1988, ed. T. B. Tshibaka and C. A. Baanante, 52-61. Muscle Shoals, Ala.,
U.S.A.: International Fertilizer Development Center.

Mellor, J. W., C. L. Delgado, and M. J. Blackie, eds. 1987. Accelerating food produc-
tion in Sub-Saharan Africa. Baltimore, Md., U.S.A.: Johns Hopkins University
Press for the International Food Policy Research Institute.

Muleya, K. 1990. Fertilizer policy in Zambia. In Fertilizer policy in tropical Africa:
Workshop proceedings, Lome, Togo, April 1988, ed. T. B. Tshibaka and C. A.
Baanante, 93-106. Muscle Shoals, Ala., U.S.A.: International Fertilizer Develop-
ment Center.

Mumeka, L. 1991. Policy constraints on the development of agriculture and small-scale
enterprises in rural Zambia. In Adopting improved farm technology: A study of
smallholderfarmers in Eastern Province, Zambia, ed. R. Celis, J. T. Milimo, and

S. Wanmali, 62-92. Washington, D.C.: International Food Policy Research

NAMBOARD (National Agricultural Marketing Board). Various years. Annual report
and accounts. Lusaka, Zambia.
NCDP (National Commission for Development Planning). 1989. New economic recov-
ery program, fourth national development plan: 1989-1993. Lusaka, Zambia:
Office of the President.

Tshibaka, T. B., and C. A. Baanante, eds. 1990. Fertilizer policy in tropical Africa:
Workshop proceedings, Lomd, Togo, April 1988. Muscle Shoals, Ala., U.S.A.:
International Fertilizer Development Center.
Wanmali, S., and J. He. 1989. Determinants of rural service use amongst households in
Eastern Province of Zambia. Paper prepared for the Seminar on Growth and
Equity in Zambian Agriculture, Livingstone, Zambia, December 8-11 (revised).
World Bank. 1981. Accelerated development in Sub-Sarahan Africa: An agenda for
action. Washington, D.C.: World Bank.

Zambia, Ministry of Agriculture and Water Development, Planning Division. 1983.
Groundnut production and marketing in Eastern Province. Lusaka, Zambia:
Ministry of Agriculture and Water Development.

Dayanatha Jha was a research fellow at IFPRI from August 1982 to March 1992.
Behjat Hojjati is a research analyst at IFPRI.


Domingo Cavallo, and Roberto Domenech. 1989
EFFECTS ON INCOME AND NUTRITION, by Joachim von Braun, Detlev Puetz, and Patrick
Webb, 1989
BASIN, by Tshikala B. Tshibaka, 1989
COME, AND NUTRITION, by Joachim von Braun. David Hotchkiss. and Maarten Immink. 1989
DESH, by Raisuddin Ahmed and Andrew Bernard, 1989
KENYA, by Thomas C. Pinckney, 1988
FORM, b% Joachim Zzetz and Alberto Valdes. 1988
and David Hotchkiss. 1988
COLOMBIAN EXPERIENCE, by Jorge Garcia Garcia and Gabriel Montes Llamas. 1988
Hossain, 1988
SUMPTION, b) Geraldo Ni Calegar and G. Edward Schuh, 1988
DESH, b Mahabub Hossain, 1988

The International Food Policy Research Institute was established in 1975 to
identify and analyze alternative national and international strategies and policies
for meeting food needs in the world, with particular emphasis on low-income
countries and on the poorer groups in those countries. While the research effort is
geared to the precise objective of contributing to the reduction of hunger and
malnutrition, the factors involved are many and wide-ranging, requiring analysis
of underlying processes and extending beyond a narrowly defined food sector.
The Institute's research program reflects worldwide interaction with policymak-
ers, administrators, and others concerned with increasing food production and
with improving the equity of its distribution. Research results are published and
distributed to officials and others concerned with national and international food
and agricultural policy.
The Institute receives support as a constituent of the Consultative Group on
International Agricultural Research from a number of donors including Australia,
Belgium, Canada, the People's Republic of China. Denmark, the Ford Foundation,
France, Germany, India, Italy, Japan, the Netherlands, Norway, the Philippines,
Spain, Switzerland, the United Kingdom, the United States, and the World Bank.
In addition, a number of other governments and institutions contribute funding to
special research projects.


CROP PRODUCTION, by Leonardo A. Gonzales, Faisal Kasryno, Nicostrato D. Perez, and
Mark W. Rosegrant, 1993
HOUSEHOLD LEVELS, Patrick Webb, Joachim von Braun, and Yisehac Yohannes, 1992
Tshibaka, 1992
Teklu, Joachim von Braun, and Elsayed Zaki, 1991
TURE, by Anya McGuirk and Yair Mundlak, 1991
DEVELOPMENT IN RURAL EGYPT, by Richard Adams, Jr., 1991
Hartwig de Haen, and Juergen Blanken, 1991
by Paul Dorosh and Alberto Vald6s, 1990
Ahmed and Mahabub Hossain, 1990
AN INPUT-OUTPUT ANALYSIS, by G. S. Bhalla, B. K. Chadha, S. P. Kashyap, and R. K.
Sharma, 1990
Vasant P. Gandhi, 1990
and Lawrence J. Haddad, 1990
(continued on inside back cover)

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

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