132.31..
.7. 725,1. 1X3 .. 2...“).
a r..- 1! if
r. 15: 5h
$64...
31 .
.r

 

.. .
sﬁixﬂlt
2:, .35
I 1951 a “PIA-‘6'. uni
. .izlzl :2 .
5 543?;

3:12....
statix
L25! .1...”

a! but
ﬁt... I.

 

I)“.
,1}! o

I. I. .IE...’}' 3:!-
! . 5:3... . 3:2 59.13.... .
«“15. $5.53?!)

:1-.. 45..., It. 33:; x...

{.1 :2...

 

 

. C .15
3:” n. o

‘9 Illvthl
AC0! 9‘ vi
5......) 1‘:-

1‘ .l

> 5 1-,: s .A
I... 1 . ‘9.
. 3.5!”... ldt’».

. I‘ll.

.A.w.aa..1v . v).

 

f ‘1
7....v1 u; . 'II .- ..n|t-D!IDA.. :vl ODI Ev
..'\..1.:J?..v.:. (A V... v.15...” «bu! :.l\y

.11.. .i...
T... .I .n‘?

\t 7. ...‘ v.-

     

     
 
 

 

. l . u ‘ ‘. Ax ‘ V ‘5' l 4 .I ‘ . D . E a ! I 1"
. V , .u. .3. E .... . s. .5 w . a» i. . u. t. L 2.1 . , . . . a I t :blV
.. . » A 1 s x v. z .. 1 )rl
, . y...‘.. t l... ,..... 1. .3 3 .7. 3. t:!.f....;r‘.wuf. .......1..ﬂu.uyv.u uni-T... Music! :1 rti, . .

 

LIBRARIES

" Illlllll lllllllllllll\ll‘l\lll llll‘llallll

This is to certify that the

thesis entitled

FOOD SECURITY IN ZAMBIA: AN ECONOMETRIC ANALYSIS
OF THE MAIZE MARKET

presented by
Bethel Nakaponda

has been accepted towards fulﬁllment
of the requirements for

M.S. degree in A2. Economics

Xmas/gs;

Major professor

Date 17 November 1992

 

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

 

 

 

 

f Lunar
Michigan Cute

University

" A

—_..__

 

 

oemoom'lll' P. I

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.

“ DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU Is An Affirmative Action/Equal Opportunity Institution
6.

HEWwa-e. ‘

 

FOOD SECURITY IN ZAMBIA: AN ECONOMETRIC ANALYSIS
OF THE MAIZE MARKET

By

Bethe! Nahponda

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

MASTER or SCIENCE

Department of Agricultural Economics

1992

ABSTRACT

FOOD SECURITY IN ZAMBIA: AN ECONOMETRIC ANALYSIS
OF THE MAIZE MARKET

By

Bethe] Nnkaponda

This study has speciﬁed and estimated an econometric model of the maize market
using time series data to determine the current and possible future conditions of prices,
production and consumption of maize in Zambia. Maize is the main staple food crop in the
entire country, a source of income for over half of the population that produce it, and it is
the second largest contributor to GDP after copper.

Separate equations are speciﬁed for production, consumption and prices of maize
using the OLS estimator. Equations that pass statistical and economic tests are combined
to form a recursive simultaneous equation model.

The model predicts that domestic maize production and total demand for maize will
grow at an average annual rate of 7.77% and 3.29%, respectively, over the period 1991-2000.
However, domestic production will be insufficient to meet the demand. Maize will be

imported to meet excess demand if no corrective measures are taken.

ACKNOWLEDGEMENTS

Many people guided, inspired, assisted and supported me in the course of this study
and during my graduate program in general. I would like to express my sincere appreciation
to my thesis committee members: Dr. Eric W. Crawford, my major professor, for his
detailed reading and critique of the thesis drafts and for his administrative support during
the whole graduate program; Dr. John (Jake) Ferris, for providing me with FAO data and
for his constructive criticism and suggestions from a technical science perspective; and Dr.
Carl Liedholm for his editorial and substantive contributions.

I am grateful to USAID and the Zambian government for providing financial
support for my graduate program through the USAID/ZAMS Project. I am particularly
grateful and thankful to the Zambia Cooperative Federation (ZCF) Management for giving
me the opportunity and for all the support during my entire study program.

My sincere appreciation goes to Julie Howard, of Michigan State University and Elly
Sinyinza, of Central Statistics Ofﬁce in Zambia, for their combined effort in providing me
with most of the data from Zambia. Without their help, it was not going to be possible for
me to carry out this study.

I would also like to thank David Makanda for his contributions to this study, as well
as for his patience, support and encouragement during its preparation.

Finally, I dedicate this work to my wonderful family, especially my mother Violet,
my late sister Moudie, and Mr. A. and Mrs. R. Sichizya for all their years of love, support,

understanding, and faith in me.

TABLE OF CONTENTS

CHAPTER PAGE
List of Tables ...................................................... viii
List of Figures ....................................................... x
Abbreviations ....................................................... xii
1. Introduction ....................................................... 1

1.1 Background .................................................... 1
1.1.1 The Importance of Maize in Zambia ............................. 1
1.1.2 The Importance of Sorghum in Zambia ........................... 3

1.2 Problem Statement .............................................. 3

1.3 Objectives of the Study ........................................... 7

1.4 Data Collection ................................................. 8

1.5 Organization of the Study ......................................... 9

2. The Zambia Economy .............................................. 10

2.1 Colonial Period ................................................ 10

2.2 Post-Independence Period ........................................ 11
2.2.1 Transitional Development Plan: 1965-1966 ........................ 11
2.2.2 First National Development Plan: 1966-1970 ...................... 12
2.2.3 Second National Development Plan: 1972-1976 .................... 16
2.2.4 Third National Development Plan: 1979-1983 ...................... 21
2.2.5 Restructuring in the Midst of Crisis: 1983-1987 .................... 24
2.2.6 New Economic Recovery Program .............................. 26

2.2.6.1 Fourth National Development Plan: 1989-1993 ............... 27

2.3 Overall Economic Performance and Development ...................... 28

2.4 Speciﬁc Agricultural Policies ...................................... 33
2.4.1 Pricing Policy .............................................. 33
2.4.2 Marketing Policy ............................................ 35
2.4.3 Subsidy Policy .............................................. 36
2.4.4 Trade Policy ............................................... 37
2.4.5 Exchange Rate Policy ........................................ 39
2.4.6 Research and Extension Policy ................................. 41
2.4.7 Credit Policy .............................................. 41

2.5 Summary ..................................................... 43

CHAPTER PAGE

3. The Maize Subsector in Zambia ....................................... 45
3.1 Maize Supply Analysis ........................................... 45
3.1.1 Trend in Maize Production .................................... 45
3.1.1.1 Trend in Maize Area Harvested ........................... 47

3.1.1.2 Trend in Maize Yield ................................... 49

3.1.2 Net Imports of Maize ........................................ 50

3.2 Maize Marketing. ............................................... 51
3.3 Marketed and Retained Maize ..................................... 54
3.4 Transportation for Marketed Output ................................ 55
3.5 Storage for Maize .............................................. 56
3.6 Maize Processing ............................................... 58
3.7 Maize Retailing ................................................ 60
3.8 Maize Consumption Analysis ...................................... 61
3.8.1 Relationship Between Maize Consumption and Real GDP ............ 63
3.8.2 Maize Consumption and Real Maize Retail Price ................... 66
3.8.3 Per Capita Maize Consumption and Sorghum Production ............ 67

3.9 Maize Price Analysis ............................................. 68
3.9.1 Real Maize Producer Price . . . ................................. 68
3.9.2 Maize Marketing Margin Analysis .............................. 71
3.10 Maize Subsidy Analysis .......................................... 73
3.11 Summary ..................................................... 84
4. Research Methodology .............................................. 85
4.1 Supply and Demand of a Commodity ................................ 85
4.1.1 Supply of a Commodity ...................................... 85
4.1.2 Demand of a Commodity ..................................... 86

4.2 Modeling Commodity Markets ..................................... 87
4.3 Modeling Methodologies ......................................... 91
4.3.1 Time Series Models Versus Econometric Models ................... 91
4.3.2 Market Models ............................................. 92

4.4 Modeling the Maize Market in Zambia .............................. 93
4.4.1 Model Structure ............................................ 93
4.4.2 Model Speciﬁcation ......................................... 94
4.4.3 Choice of Estimator ......................................... 98
4.4.4 Model Validation .......................................... 100
4.4.5 Model Forecasting ......................................... 100

4.5 Summary ..................................................... 101
5. Presentation and Analysis of Results .................................. 102
5.1 Estimated Maize Supply Equations .............. ‘ .................. 102
5.1.1 Estimated Average Yield of Maize ............................. 102
5.1.2 Estimated Area Harvested of Maize ............................ 103
5.1.3 Estimated Net Imports of Maize ............................... 104

5.2 Evaluation of Supply Equations ................................... 105
5.2.1 Statistical Evaluation of Supply Equations ....................... 106

CHAPTER 5 Continued PAGE

5.2.1.1 The Maize Yield Equation .............................. 106

5.2.1.2 The Area Harvested of Maize Equation .................... 107

5.2.1.3 Maize Net Imports Equation ............................. 108

5.2.2 Economic Evaluation of Supply Equations ....................... 108
5.2.2.1 The Yield of Maize per Hectare .......................... 108

5.2.2.2 The Area Harvested of Maize ............................ 109

5.2.2.3 Net Maize Imports .................................... 109

5.3 Estimated Maize Demand Equations ............................... 109
5.3.1 Estimated Per Capita Domestic Maize Consumption ............... 110
5.3.2 Estimated Demand of Maize for Industrial Processing .............. 111
5.3.3 Estimated Per Capita Demand of Maize for Feed ................. 112
5.3.4 Estimated Demand of Maize for Seed .......................... 113

5.4 Evaluation of Maize Demand Equations ............................ 114
5.4.1 Statistical Evaluation of Maize Demand Equations ................. 114
5.4.1.1 Per Capita Domestic Maize Consumption Equation ........... 114

5.4.1.2 Per Capita Industrial Processed Maize Equation .............. 114

5.4.1.3 Per capita Maize Feed Equation .......................... 115

5.4.1.4 Maize Seed Equation .................................. 115

5.4.2 Economic Evaluation of Maize Demand Equations ................ 115
5.4.2.1 Per Capita Domestic Maize Consumption Equation ........... 115

5.4.2.2 Per Capita Maize Demand for Industrial Processing ........... 116

5.4.2.3 Per Capita Maize Demand for Feed ....................... 116

5.4.2.4 Demand of Maize for Seed .............................. 116

5.5 Estimated Price Equations ....................................... 117
5.5.1 Deﬂated Maize Producer Price Equation ........................ 117
5.5.2 Deﬂated Maize Meal Retail Price Equation ...................... 118

5.6 Evaluation of Price Equations .................................... 118
5.6.1 Statistical Evaluation of Price Equations ......................... 118
5.6.1.1 Deﬂated Maize Producer Price ........................... 118

5.6.1.2 Maize Marketing Spread ................................ 119

5.6.2 Economic Evaluation of Price Equations ........................ 119
5.6.2.1 Deﬂated Maize Producer Price ........................... 119

5.6.2.2 Maize Marketing Spread ................................ 120

5.7 Supply Elasticities ............................................. 120
5.8 Demand Elasticities ............................................ 122
5.9 Validation of the Model ......................................... 123
5.9.1 Statistical Validation ........................................ 123
5.9.2 Graphical Validation ........................................ 124
5.9.3 Validation Using Mean Squared Error .......................... 125
5.9.4 Validation Using the Root Mean Squared Percentage Error ......... 126
5.9.5 Validation Using Theil’s U-statistic ............................. 128
5.10 Forecasts of Exogenous Variables ................................. 128
5.11 Forecast Results .............................................. 129
5.11.1 Maize Supply Forecasts ..................................... 129
5.11.2 Maize Demand Forecasts ................................... 132
5.11.3 Projection of Required Maize Imports ......................... 135

vi

CHAPTER 5 Continued PAGE

5.11.4MaizePriceForecasts............................... ....... 139
5.12 Summary .................................................... 141
6. Summary, Conclusion, Policy Recommendations and Areas of Future
Research .................................................... 143
6.1 Summary ..................................................... 143
6.2 Conclusion ................................................... 145
6.3 Policy Recommendations ........................................ 147
6.4 Areas of Further Research ....................................... 143
Appendix A Figures For Chapter 5 .................................... 153
Appendix B Sources of Data Used in the Study ........................... 162
Appendix C Calculation of the Fertilizer Consumption Index and the Fertilizer
Cost per Hectare ........................................ 166
List of References .................................................. 167

vii

LIST OF TABLES

TABLE PAGE
1.1 Zambia: Food Production Indices, 1963-1988 ............................. 4
2.1 Zambia: Planned Public Investment, 1966-1970 .......................... 14
2.2 Zambia: SNDP Planned Investment by Economic Sector, 1972- 1976 .......... 19
2.3 Zambia: TNDP Planned Sectoral Investment, 1979-1983 ................... 22
2.4 Zambia: FNDP Planned Sectoral Investment, 1989-1993 ................... 29
2.5 Zambia: Macroeconomic Indicators, 1960-1988 .......................... 30

2.6 Zambia: Percentage Distribution of GDP by Kind of Economic Activity at
Constant 1977 Prices ............................................ 32

2.7 Zambia: Number of Short-term Loan Applicants and Approved Loans, 1979/80-
1983/84 ...................................................... 42

3.1 Zambia: 1988 Grain Storage Capacity. by Type and Provincial C00perative
Union ....................................................... 57

3.2 Zambia: Number of Stores Retailing Maize Mealie Meal by Province in 1990 . . . 61

3.3 Zambia: Percentage Changes in Producer Prices from the Previous year’s Price
for selected crops ............................................... 70

3.4 Zambia: Ratio of Maize Producer Price to the Retail Price of Maize Mealie
Meal, 1964-1991 ................................................ 74

3.5 Zambia: Nominal Maize and Fertilizer Subsidies as a Percentage of the
Government Budget ............................................. 75

3.6 Zambia: 1988/89 Season Intra-Provincial Maize Handling and Transportation
Costs and Subsidies by Provincial Cooperative Union ................... 79

viii

TABLE PAGE

3.7 Zambia: 1988/ 89 Percentage Composition of Total Handling and

Transportation Cost of a 90-kg bag of maize .......................... 80
3.8 Zambia: Government Gazetted Transport Rates for Maize Haulage, 1984-

1990 ......................................................... 81
5.1 Short-run and Long-run Supply and Demand Elasticities .................. 120
5.2 Quantitative Measures of the Model’s Ability to Forecast ................. 127
5.3 Zambia: Projection of Area Harvested, Yield, and Production of Maize, 1991-

2000 ........................................................ 131
5.4 Zambia: Projection of Demand of Maize for Domestic Consumption, for Feed,

for Industrial Processing, and for Seed, 1991-2000 ..................... 133
5.5 Zambia: Projection of Required Maize Imports, 1991-2000 ................ 136

5.6 Zambia: Projection of the Maize Producer price, Maize Mealie Meal Retail
Price and Maize Marketing Spread, 1991-2000 ....................... 140

LIST OF FIGURES

FIGURE PAGE
1.1 Maize Production in Zambia: 19644990 ................................. 5
3.1 Maize and Sorghum Production in Zambia: 1964-1990 ..................... 46

3.2 Zambia: Relationship Between Area Harvested of Maize and the Real Producer
Price of Maize, 1964-1990 ........................................ 48

3.3 Zambia: Relationship Between Fertilizer Consumption Index and Maize Yield
1964-1990 ..................................................... 50

3.4 Zambia: Relationship Between Per Capita Maize Consumption and Per Capita
Maize Production, 1964-1990 ...................................... 62

3.5 Zambia: Relationship Between Per Capita Maize Consumption and Per Capita
Real GDP, 1964-1990 ........................................... 64

3.6 Zambia: Comparison of Real Retail Prices of Maize, Rice and Bread, 1964-
1990 ......................................................... 65

3.7 Zambia: Relationship Between Per Capita Maize Consumption and Maize
Mealie Meal Real Retail Price, 1964-1990 ............................ 66

3.8 Zambia: Relationship Between Per Capita Maize Consumption and Per Capita
Sorghum Production, 1964-1990 .................................... 68

3.9 Zambia: Comparison of Maize and Sorghum Real Producer Prices, 1964-1990 . . . 69
3.10 Zambia: Comparison of Real Maize Producer and Retail Prices, 1964-1990 . . . . 72

3.11 Zambia: Comparison of Real Maize Marketing Margin and Real Maize
Subsidies, 1964-1990 ............................................ 77

5.1 National Average Yield of Maize in Zambia: 1966-1990 and Projected to the
Year 2000 ................................................... 153

LIST OF FIGURES

FIGURE PAGE
1.1 Maize Production in Zambia: 1964-1990 ................................. 5
3.1 Maize and Sorghum Production in Zambia: 1964-1990 ..................... 46

3.2 Zambia: Relationship Between Area Harvested of Maize and the Real Producer
Price of Maize, 1964-1990 ........................................ 48

3.3 Zambia: Relationship Between Fertilizer Consumption Index and Maize Yield
1964-1990 ..................................................... 50

3.4 Zambia: Relationship Between Per Capita Maize Consumption and Per Capita
Maize Production, 1964-1990 ...................................... 62

3.5 Zambia: Relationship Between Per Capita Maize Consumption and Per Capita
Real GDP, 1964-1990 ........................................... 64

3.6 Zambia: Comparison of Real Retail Prices of Maize, Rice and Bread, 1964-
1990 ......................................................... 65

3.7 Zambia: Relationship Between Per Capita Maize Consumption and Maize
Mealie Meal Real Retail Price, 1964-1990 ............................ 66

3.8 Zambia: Relationship Between Per Capita Maize Consumption and Per Capita
Sorghum Production, 1964-1990 .................................... 68

3.9 Zambia: Comparison of Maize and Sorghum Real Producer Prices, 1964-1990 . . . 69
3.10 Zambia: Comparison of Real Maize Producer and Retail Prices, 1964-1990 . . . . 72

3.11 Zambia: Comparison of Real Maize Marketing Margin and Real Maize
Subsidies, 1964-1990 ............................................ 77

5.1 National Average Yield of Maize in Zambia: 1966-1990 and Projected to the
Year 2000 ................................................... 153

FIGURE ' PAGE

5.2 Area Harvested of Maize in Zambia: 1966-1990 and Projected to the Year

2000 ........................................................ 154
5.3 'Maize production in Zambia: 1966-1990 and Projected to the Year 2000 ...... 154
5.4 Net Maize Imports in Zambia: 1966-1990 and Projected to the year 2000 ..... 155
5.5 Net maize supply in Zambia: 1966-1990 and Projected to the year 2000 ....... 155
5.6 Per Capita Domestic Consumption of Maize in Zambia: 1966-1990 and

Projected to the year 2000 ....................................... 156
5.7 Total Domestic Consumption of Maize in Zambia: 1966-1990 with Projections

to the year 2000 ............................................... 156
5.8 Per Capita Demand of Maize for Feed in Zambia: 1966-1990 and Projected to .

the year 2000 ................................................. 157
5.9 Total Demand of Maize for Feed in Zambia: 1966-1990 and Projected to the

year 2000 .................................................... 157

5.10 Per Capita Demand of Maize for Industrial Processing in Zambia: 1966-1990
and Projections to the Year 2000 .................................. 158

5.11 Total demand of Maize for Industrial Processing in Zambia: 1966-1990 and
Projection to the Year 2000 ...................................... 158

5.12 Maize Seed Demand in Zambia: 1966-1990 and Projected to the year 2000 . . . . 159

5.13 Total Demand of Maize in Zambia: 1966-1990 and Projected to the Year
2000 ........................................................ 159

5.14 Change in Maize Stocks in Zambia: 1966-1990 and Projected to the Year
2000 ........................................................ 160

5.15 Real Maize Producer Price in Zambia: 1966-1990 and Projected to the Year
2000 ........................................................ 160

5.16 Real Maize Retail Price in Zambia: 1966-1990 and Projected to the year
2000 ........................................................ 161

ARMB

CPCMU
ECU
EDP
FAO
FEMAC
FNDP
GDP

NAMBoard
NCDP
NCU
NERP
NWCU
OLS

SNDP
SPCMU
TDP
TNDP
WPCMU
ZCF
ZEGA

ABBREVIATIONS

Agricultural Rural Marketing Board
COOperative Credit Scheme

Copperbelt Province Cooperative Union
Central Province Cooperative Marketing Union
Eastern Province Cooperative Union
Emergency Development Plan

Food and Agricultural Organization

Foreign Exchange Management Committee
Fourth National Development Plan

Gross Domestic Product

Grain Marketing Board

International Labor Organization

International Monetary Fund

Kwacha (Zambian Currency)

Kilogram

Kilometer

Luapula Province Cooperative Union

Lusaka Province Cooperative Union

National Agricultural Marketing Board
National Commission for Development planning
Northern Province Cooperative Union

New Economic Recovery Program
North-Western Province C00perative Union
Ordinary Least Squares

Provincial Cooperative Union

Second National Development Plan

Southern Province Cooperative Marketing Union
Transitional Development Plan

Third National Development Plan

Western Province Cooperative Marketing Union
Zambia Cooperative Federation

Zambia Export Growers Association

1. INTRODUCTION

1-1 Wt!
1-1-1 BMW

Various crops are grown in Zambia, including maize, soya beans, sunﬂower, groundnuts,
cotton, tobacco, rice, beans, millet, wheat and cassava. The main staple food crop for the
whole country is maize. In the urban areas, maize is supplemented, to a lesser extent, with
wheat flour, bread and rice. In the rural areas, maize is supplemented with traditional food
crops - sorghum, millet and cassava. The whole agricultural sector is dominated by maize
in Zambia. Of all the food and non-food crops that are grown, maize accounts for about
70% of total area under cultivation and contributes about 90% to total food production in
Zambia. It is produced throughout the country by more than half of the population
(Chenoweth and McKenzie, 1992) and it is also the main source of income for most of the
households in the rural areas.

Maize is consumed in both rural and urban areas. In 1980, average cereal
consumption per capita was estimated at 172 kg of which 62% (106.64 kg) was derived from
maize while the rest was accounted for by cassava, millet, sorghum, wheat and rice (Republic
of Zambia, 1990). The same study revealed that with the decrease in the production of
traditional food crops over the years and the decontrol of prices for all foods other than
maize mealie meal in the 1980s, per capita consumption of maize is even higher now. The

decontrol of prices for other food staﬂs apart from maize has left maize mealie meal

2
relatively cheaper and more affordable than other cereals like wheat and rice. This has also

been inﬂuenced by the general economic position of the population which has continued to
worsen starting as far back as in the 1970s as reﬂected by the real per capita gross domestic
product (GDP).

The importance and strategic position of maize has also been recognized by the
government. Since the colonial period, the government has intervened in the maize
subsector mainly for socio-political reasons. The main aim of government controls of maize
marketing has been to ensure the supply of cheap food to the vocal urban population.
Uniform and pan-seasonal maize producer prices have been ﬁxed by the government
without considering regional differences. The marketing and processing of maize, as well
as maize meal distribution have been government controlled. The marketing spread has also
been determined by the government, which has been setting uniform prices paid and
received by maize millers, and uniform consumer prices for all maize processed products.
Actually, maize production and marketing have been promoted by the government to an
extent that maize is the second most important economic commodity after copper, in terms
of domestic revenue.

However, with the gradual removal of maize subsidies, hence maize support
production and marketing support programs, maize production and marketing may become
too expensive in some parts of the country. In such areas, production of alternative crops
such as sorghum and millet may have to be encouraged if they are suitable. The importance

of sorghum in the country is reviewed below.

3

1.12 W
Although the amount of sorghum produced in Zambia is very little relative to the

amount of maize produced (less than 5%). sorghum and other traditional food crops are still
important staple food crops for some rural households that grow maize mainly as a cash
crOp. Sorghum is the main food crop among traditional food crops. Because sorghum and
millet are not retailed and consumed in urban areas as a staple food, over 99% of sorghum
produced in the country is retained for consumption in rural areas. For households that
produce traditional food crops, maize is retained for consumption only as a supplement to
those traditional food crops. Moreover, in drought-prone areas where maize cannot be
produced, sorghum and millet, which are more drought resistant than maize, are the main
food crops produced.

The Republic of Zambia (1990) revealed that between 1982/83 and 1988 / 89,
sorghum took up the second largest area under cultivation after maize in Lusaka,
Copperbelt and North-Western Provinces while it took up the third largest area under
cultivation in Western Province. All these provinces are maize deficit areas. Therefore,
there is a possibility that if policies encouraged the production of suitable crops other than
maize and considering the decline in maize support programs, sorghum production might

stand a good chance in these provinces where it is already produced.

12 Wm

Generally, food production in Zambia has been increasing since independence in 1964,
as shown in Table 1.1, which shows Zambia’s food production indices for the period 1963-
1988. The food production increase since the late 1970s has been quite erratic. This can

partly be attributed to the weather conditions that characterized the period, particularly in

4

Table 1.1

Zambia: Food Production Indices, 1963-1988
(1979-1981= 100)

 

Year mm
1963 0.65
1964 0.65
1965 0.73
1966 0.84
1967 0.89
1968 0.80
1969 0.87
1970 0.86
1971 0.96
1972 0.98
1973 0.96
1974 1.09
1975 1.13
1976 1.21
1977 1. 17
1978 1.10
1979 0.97
1980 1.02
1981 1.00
1982 1.98
1983 1.03
1984 1.04
1985 1.15
1986 1.13
1987 1.14
1988 1.26

 

Source: FAO Production Yearbook 1988 Vol. 42 for data from 1963-1968; 1980
vol. 34 for data from 1969-76; and 1974 vol. 28.1 for data from 1977 to
1988.

Note: Data from 1963-1968 was given as indices with 1961-65 as base years
while data from 1969-76 was given as indices with 1969-1971 as base
years. The author converted them to indices with 1979-1981 as base
year to tally with data for the period 1977 to 1988.

Figure 1.1. Maize Production in Zambia: 1964-1990

 

2000
1800i
1600'
1400'
1200-

1000‘

THOUSANDS OF METHC TONS

800-

 

 

 

"'19'70' '19'75' .4 i936 ' ' 11985] ' 1 3950
YEAR

39235

the 1980s. The ﬂuctuating increase. in total food production is also reﬂected in the
production of maize during the same period, as shown in Figure 1.1.

However, this increase in food production has not been enough to meet the demand.
Food production has been increasing at a lower rate than population. Excess food demand
has been met by imports, mostly in the form of cereals. Imports of non-essential foodstuffs
have been restricted by government policy. Some of the problems in the maize subsector
that need to be looked into include excessive transaction costs, producer prices, high levels
of government subsidies, heavy transportation bottlenecks, lack of credit, insufﬁcient storage

facilities and other market infrastructure, high grain losses, late delivery of inputs, imperfect

syn

{on

pro
pro

dete

6
information, technological implementation problems, lack of spare parts for machinery and

trucks, and limited access to both domestic and international trade.

Therefore, there is need to increase food production further and have an eﬂicient
food system. To have adequate maize production and an efﬁcient food marketing and
distribution system, there is need to have proper future commodity pricing, production and
consumption information for planning purposes by all the groups involved in the food
system. Hence the importance of understanding how the maize market operates and to
forecast prices and quantities of maize.

The information from forecasting can assist several users: farmers in their
production and marketing decisions; marketing agencies and intermediaries as they assess
production prospects for the season for their resource allocation purposes; consumers in
determining their consumption patterns; lending institutions; and the government in setting
agricultural priorities and other agricultural programs like price support programs.

A well-coordinated marketing system for both inputs and outputs is of vital
importance if food self-sufficiency is to be obtained. It is important because about 50% of
the pOpulation resides in urban areas and depends on the food market system for the supply
of their daily food needs. This is in addition to the need for farmers to have efficient
produce marketing and timely distribution of farm inputs.

A well-coordinated food system is also required for the attainment of food security
which goes beyond increased production and food self-sufﬁciency. Valdes and Siamwala
(1981) define food security as

the ability of food deficit countries, or regions or households within these countries,
to meet target consumption levels on a year-to-year basis. (p.2).

7
Therefore, it is possible to have both national food self-sufficiency (adequate food

production) and hunger in some areas of the country.

Timmer 9331. (1983) argue that hunger is caused by restricted choices of basic food
quantities available to individual consumers, resulting mainly from inefficient and
uncoordinated food systems. They further argue that this hunger problem is embedded in
the set of processes of the food system starting from the point of commodity production
(farm), through the transformation of these commodities into food in the marketing sector,
to selling of food to consumers to satisfy nutritional and social needs. The actual
functioning of this food system frequently leaves many poor people inadequately fed because
of their employment and income status, the prices they pay for food and other goods and
services, and their ability to move in search of better opportunities. Hence, solving the
problem of hunger involves expanding the available choices, which are a function of incomes,

food prices, food supplies, and consumer knowledge.

1.3 W

The main objective of the study is to carry out an econometric analysis of the maize
subsector to determine the current and possible future maize demand and supply conditions.
This objective is three-fold. First, the study will give a descriptive and diagnostic analysis
Of the major institutional, policy, environmental and economic factors and relations that
characterize the agricultural sector and the maize subsector in particular. This will show
how the Zambian maize subsector’s performance is a function of the institutional and policy
environment within which it operates. This will also provide a background for the economic

formulation of the supply and demand behavioral equations for the maize econometric

model.

8
Second, the study will carry out a maize market analysis which will involve the

speciﬁcation and estimation of an econometric model of supply and demand for maize.
Then predictions of maize prices and quantities will be made to the year 2000. Lastly,

results of the regression model will be used to calculate supply and demand elasticities.

1.4 Won

The 1964-1990 time series data used in this study are secondary data. Some of the
data were collected from the documents published by the Central Statistical Office and other
government ministries of the Republic of Zambia. The rest of the data are from the Food
and Agricultural Organization (FAO) and from some books written about Zambia which
report data obtained from Zambian government documents. Specific sources of data are
given in Appendix B.

The rainfall ﬁgures are national averages which have been computed from rainfall
levels at stations from the various parts of the country as reported in the government
publications consulted. They may or may not be representative of the true population mean
value, even when the sample is large.

It is worth pointing out that various sources of country-data show big differences in
the same variables. For Zambia, the data on area harvested of maize, yield, production,
exports, imports and domestic consumption from the United States Department of
Agriculture (USDA) and from FAO showed big differences. However, both sources have
almost the same data as those from Zambia for the period 1980 to 1990 for which period
data on the latter could be found.

Although both FAO and USDA have almost the same data as those from Zambia

for the 1980s, FAO data, and not USDA data, were selected for use in this study for two

9
reasons. First, an inquiry into the sources of data from the USDA revealed that the USDA

obtains some of its country-data from FAO among other sources, while FAQ is said to be
collecting its data from the respective country governments. Second, FAO has its data
broken-down into various types of demand for maize and maize wasted in Zambia unlike

USDA which just reports domestic consumption of maize.

1-5 W

This study is organized into six chapters. Following this introduction (Chapter 1),
Chapter 2 gives an overview of Zambia’s general and speciﬁc agricultural policies and the
post-independence sectoral and economic performance of the economy. In Chapter 3, the
maize subsector is discussed with regard to production, marketing, consumption and pricing.
Chapter 4 discusses the methodology used in estimating the maize supply, demand and price
behavioral equations. Chapter 5 presents empirical ﬁndings from the econometric model
of the study. Finally, in Chapter 6, a summary, conclusions, and policy implications of the

study, as well as further research recommendations, are given.

2. THE ZAMBIA ECONOMY

This chapter gives an overview of Zambia’s economy with particular emphasis on the
agricultural sector. The ﬁrst part gives an overview of the pre-colonial period. The second
part provides a descriptive and diagnostic analysis of the post-independence development
plans undertaken by the Zambian government and the resulting sectoral and overall
performance of the economy. Performance is measured by the achievement of the targets
set, and the relative performance of the sectors of the economy. This is followed by the

discussion on speciﬁc agricultural policies.

2-1 W
The British colonial rule in Zambia was such that the economy was overwhelmingly

geared towards the export of copper with agriculture and manufacturing left severely
underdeveloped (Clark and Allison, .1989; Bhagavan, 1978). The Federal government had
a policy of confining the manufacturing and industrial base to Zimbabwe (then Southern
Rhodesia). Large sums of revenue that accrued from the copper exports in Zambia (then
Northern Rhodesia) were drained out of the territory and used to underpin industrial
development in Zimbabwe (Ceary, 1989; Kaunga em, 1983). The few industrial activities
and commercial agriculture were concentrated along the railway line which ran through the
Southern, Central and Copperbelt provinces of Zambia.

Agricultural production, particularly peasant production of traditional food crops
such as sorghum, millet and cassava and other simple commodities, was irreparably
damaged, and peasant society totally destabilized by the phenomenon of labor migration

from rural to urban areas, induced by the colonial regime. As a result, great regional

10

11
imbalances arose between the rural and urban areas. These regional imbalances and the

heavy dependence of Zambia’s economy on the copper industry, whose export earnings
have ﬂuctuated and declined so much since early 1970s as a result of world copper price
instability, became a great concern of the government upon independence in 1964. Upon
independence, the Zambian government’s policies emphasized the diversiﬁcation of the
economy, with particular emphasis on the development of the agricultural sector on which
the rural areas heavily rely for their income and other social and economic needs.

22 WW

At Independence in 1964, Zambia inherited a highly dualistic economy. The modern
sector, mainly owned by foreigners, was dominated by the copper industry and by large-scale
commercial agriculture. The rest of the economy consisted mainly of rural small-holder
farmers with very few productive links to the modern sector. The difference in the two
sectors of the economy were also reﬂected in their high income disparities in favor of the
modern sector. Therefore, since independence, the Zambian government has undertaken
objectives and policies that would help develop the rural sector so as to raise the standard
of living of the rural population and to narrow the rural-urban gap and to reduce the
economy’s dependence on copper. The major policies and objectives that have been
undertaken by the government and the performance of the economy in general and the

agricultural sector in particular are discussed below.

2.2-1W
The first post-independence plan was the Emergency Development Plan (EDP).

It was a prelude to the 18 months Transitional Development plan (TDP) which ran from

12
January 1965 to June 1966. The TDP was meant for bridging the period of transition before

a main development plan could be formulated. It covered a period of change from the
rudimentary planning presented by the colonial public expenditure programs to the initiation
of full resource programming, which the government wanted to introduce (Chileshe, 1986).
It had two categories of projects: those carried over for completion and new ones whose
starting period was dependent on the progress made in those carried over. Low priority was
allocated to agriculture and the rural sector (Chileshe, 1986). The agricultural sector was
neglected because Zambia was 'born" at a time when revenues from the copper industry

were plentiful enough to comfortably cushion the economy.

22.2 W
In 1966, the First National Development plan was launched and it covered the period

1966-1970. Its emphasis was to establish a broad social and economic infrastructure as
quickly as possible, before any direct agricultural production could be emphasized. The
Plan’s main objectives were to diversify the economy so as to eliminate over-dependence on
the copper industry; to increase employment; to improve per capita monetary output; to
maintain reasonable price stability; to minimize the inherited economic imbalance between
the urban and rural sectors; to raise the general levels of education and develop a wide
range of speciﬁc technical, administrative, executive, professional and management skills of
the population; to provide more and better living accommodation and other social welfare
needs; and to develop new communications, sources of energy, transport and other economic
infrastructure which constitute the basic framework for real economic growth and

development.

13

During this period, agriculture was not taken seriously and was not in practice
considered to be complementary to the development of the industrial sector in Zambia.
This is depicted in the sectoral concentration of the public investment funds, as shown in
Table 2.1. A large proportion of public expenditure was allocated to the transport and
infrastructure sectors which accounted for 38.9% of the total public sector investment of
K5637 million. Agriculture, industry and mining, and social infrastructure absorbed about
12.2%, 21.4% and 18.1%, respectively. Forestry, game, ﬁsheries and research together were
allocated the least (3.2%) despite the importance of research to a country’s development
process. These public sector investment funds (563.7 million) accounted for about 65% of
the total national investment funds (K8586 million), indicating the lower importance
attached to the private sector.

The justification for the lower priority given to agriculture relative to the other major
sectors of the economy was mainly the “time lags and bottlenecks, absent from other sectors"
but present in the agricultural sector (First National Development Plan, p. 22). Another
factor cited that prohibited a rapid expansion of the agricultural sector was the need to
spend considerable sums of money on programs such as economic infrastructure and
training which would be fruitful only after 1970. However, the agricultural sector also
benefitted from the planned investment that was channelled into provision of economic
infrastructure in the rural areas, including provision of basic water supplies, access roads and
major conservation works.

The report on the "Second National Convention on Rural Development, Incomes,
Wages and Prices in Zambia" (1970) provides the ﬁrst major assessment by the government
of the policies that were described in the First National Development Plan. This convention

was held in December 1969. President Kaunda pointed out that the past government

14
Table 2.1

Zambia: Planned Public Investment, 1966-1970

 

Elﬁn ‘ Erin——
mm
1. Infrastructure and Transport

Housing and Construction 86.3
Roads 69.8
Railways 19.5
Aerodromes 15.1
Telecommunications 12.9
Meteorology 0.7
Supporting Services 113
215.7 383
II Agriculture and Lands
CrOps and Livestock 69.0 12.2
Forestry 2.7
Game and Fisheries 6.6
Research . 5.6
Other 32
87.0 15.4
III Industry and Mining
Industrial development 50.1
Mines 12.1
Tourism 5.5
Electricity 51.5
121.0 21.4
IV Social infrastructure
Health 18.5
Education 79.5
Labor and Social Welfare ;4_._1
102.1 18.1
V Other
Oﬂice of the President 29.5
Secretary to the Cabinet 1.1
Ofﬁce of National Development and 1.2
Planning
Foreign Aﬂairs 0.6
Justice 0.8
National Assembly 0.5
Home Affairs 3.1
Finance L9
37.9 6.7
Total m1 m9

 

Source: Zambia, Oﬂice of National Development and Planning W

W July. 1966. p. 12, converted from pounds to
Kwacha.

15

policies were inconsistent with the philosophy of Humanism, which is based on the equality
of all men and the dignity of even the most impoverished individual. He exprewed concern
and dissatisfaction with the widening rural/ urban gap and with the fact that government
policies as at that date had benefitted the ’urbanites’ while they affected only a small
minority of the population in the rural areas. He also criticized the poor performance of
the cooperative movement and the bias toward capital-intensive methods of production at
the expense of the labor-intensive production methods.

Overall, the performance of the economy during the First National Development
Plan was quite satisfactory, mostly due to higher copper prices that prevailed at that time.
The introductory part of the Second National Development Plan (SNDP) reports that
between 1965 and 1970, nominal GDP increased by 83% with an average annual growth rate
of 10.6% (Second National Development Plan (SNDP), 1971). This growth rate compared
favorably with the planned increase of 11.7% per annum in the First National Development
Plan. Over the period 1964-1970, gross fixed capital formation grew at an average annual
rate of 21% while the planned rate was 20% per annum. As a percentage of GDP,
investment rose from about 16% in 1964 to about 28% in 1970. Resources (external and
internal) made available during the period exceeded the amounts planned by 43%; recurrent
expenditure exceeded the target by 21%; and real capital expenditure was close to the
amount budgeted for. Therefore, on the whole, the resource and expenditure targets of
First National Development Plan were more than met (SNDP, pp. 2-4).

Although the overall performance of the economy was good between 1965 and 1970,
that of the agricultural sector was not satisfactory. The poor performance is reﬂected in the
amount by which the Plan targets had been missed and by the relative performance of the

other sectors. While the target annual growth rate was 9%, the actual rate of growth was

16
only 3.3% per annum, and out of the K884 million planned capital expenditure for the rural

sector, only 76% was actually spent (SNDP, p. 4). The 1970 targets for marketed production
for individual commodities in the First National Development Plan were not met except for
raw sugar, dressed poultry, and eggs (SNDP, p. 14). Even this sugar was produced on an
industrial estate, not in the rural areas. The SNDP also pointed out that during the First
National Development Plan, greater problems were experienced with communal agricultural
cooperatives than with family farm cooperatives and that in the future, emphasis should be

put on the latter.

223 W

The Second National Development Plan (SNDP) was designed to carry further the
process of economic and social growth started during the First National Development Plan
period. It laid major emphasis on rural development - ”the expansion of agricultural
production as a top priority” (SNDP, p. 33) - both to reduce the level of income disparity
between urban and rural areas and to diversify the economic base. The SNDP sought to
fulﬁll a number of objectives which included the creation of new employment and income
opportunities; the improvement of infrastructural services to complement rural employment
creation and counteract rural-urban migration; the increase in the contribution of the
agricultural sector to the GDP; the reduction in the growing dependence on imported
commodities; the achievement of staple food self-sufficiency; and the emphasis to improve
rural nutrition by increasing the production of protein foods in the rural areas for local
consumption (SNDP p. 61).

To achieve these objectives, the government was required to coordinate and

emphasize policies that favored balanced and accelerated rural development and encouraged

17

self-reliance. For instance, it was required to ensure and maintain fair prices for agricultural
products, inputs, and consumer goods. In addition, it had to improve input supply, credit
facilities, extension services, storage facilities, marketing and processing services to
agriculture, forestry, and ﬁsheries. There was also policy redirection through the promotion
and development of long-term family farms, and use of more labor-intensive techniques and
the shift of emphasis from large-scale state production to private sector production.

Two new major policies were introduced in the SNDP: the setting up of regional
production priorities, and the formation of Intensive Development Zones (IDZ’s).
However, both policies were not implemented. The policy of regional production priorities
was hindered by the movement toward uniform regional crop pricing which began in
1971/72 such that by 1974/75 season, all regional price differentials were eliminated. This
was because it was supposed to be based on considerations of local demand and supply and
comparative costs. On the other hand, in 1975, a decision was taken against the
implementation of the IDZ’s because it was discovered they were not going to be as feasible
as they were planned for.

Despite the proposal that the agricultural sector would be given top priority, the
planned allocation of investment gave several other objectives just as high a priority in the
SNDP. The other objectives given almost the same priority included expansion and
diversiﬁcation of industry and improvement of social and economic infrastructure. As in
the First National Development Plan period (1966-1970), the agricultural sector was not
expected to grow at a higher rate, relative to the growth rates of the other major sectors.
During the SNDP period, the projected average annual growth rate of the agricultural sector

(5.3% for subsistence agriculture and 5.9% for commercial agriculture) at constant 1969

18

prices was lower than that of the manufacturing and mining sectors (14.7% and 6.1%,
respectively) (SNDP, p.40).

Table 2.2 shows the allocation of the investment funds between the public sector and
the private sector, and partly reﬂects the importance attached to each economic sector.
Data in Table 2.2 show that the public sector again utilized nearly 65% of total sectoral
investment funds in the SNDP. On a sector basis, public investment in transport and
communications infrastructure continued to receive top priority. Allocations to most sectors,
including the agricultural (rural) sector declined.

Both the overall performance of the economy and the performance of the
agricultural sector was not satisfactory during the SNDP period. Between 1965 and 1970
(approximate period of the First National Development Plan) and at constant 1965 prices,
GDP grew at an annual average rate of 2% while during the SNDP, the average annual
growth rate of GDP was 3.4% at constant 1965 prices - as against the target growth rate of
7.4% (Third National Development Plan (TNDP), 1979, p.1). Although the GDP annual
average growth rate was higher during the latter plan than during the former, the latter
growth rate was still lower than the target.

Despite the major objective of the SNDP being that of diversifying the economy,
hardly any development took place in this direction. At constant 1965 prices, the
contribution of the mining sector to the total GDP decreased from 24% in 1971 to 23.2%
in 1976. The agricultural sector’s contribution of 14.2% to GDP in 1976 was almost the same
as that of 1971. However, there was a signiﬁcant development~ in the crop subsector but its
growth impact was canceled by the shortfalls in the livestock subsector. The average annual
contribution of the manufacturing sector to GDP in 1976 was marginally above that of 1971;

the share of construction in the GDP almost remained constant; while the shares of

Zambia: SNDP Planned Investment by Economic Sector, 1972-1976

19
Table 2.2

 

 

Million Kwacha
Private

Aetiziﬂ SEW % E Z Intel
Rural Sector 122.5 9. 6 30. 9‘ 4. 4 152. 5
Mining 41.0 3.2 361.0 52.7 402.0
Manufacturing 107.0 8 .4 146.0 21.8 253.0
Construction - 0. O 63.0 9.8 63.0
Power 198.8 15. 6 - 0.0 198.8
Trade 45.0 3 .4 10.0 1.5 55.0
Transport and 369.3 29.1 5.4 0.8 374.7
Communications

Tourism 15.5 1.2 10.0 1.5 25.0
Housing, etc. 146.0 11.5 54.0 7.9 200.0
Education 114.5 9.0 3.0 0.4 40.0
Health 37.0 2.0 3. 0 0. 4 40.0
Others 74.9 5.9 0. 0 74.9
Total 1211.0 ms 653 ms 19564
Source: Compiled from data given in Tables 1. 1 to 3 .2 of Total Investment

Outlay 1n the
Ministry of Development Planning and National Guidance, Lusaka,

Zambia.

t

9

" Mainly investments of the subsidiaries of MINDECO which are 51% owned by

the government.

20
transport, communication, storage, and wholesale and retail sectors in the total GDP

declined, particularly in the last years of the plan. Therefore, overall, “the SNDP failed to
bring about any diversiﬁcation in the country’s economic structure.” (TNDP, pp. 3-4).

The uneven and inadequate growth performance of the economy during the period
of the SNDP was attributed to a number of both endogenous and exogenous factors that
characterized the period. The endogenous factors included the pattern of investment, the
import of highly capital-intensive technology, and other factors. For the pattern of
investment, (i) about half of the total expenditures was on maintenance and replacement
of already existing ﬁxed investment which did not result in any net addition to the capital
stock thereby reducing ﬁxed assets; (ii) the proportion of investment absorbed by the
economic and social infrastructure and services (that is, sectors that are not directly
productive) was also high as shown in Table 2.2; (iii) the investment was also biased towards
the adoption of highly capital-intensive technology which was against the need to utilize the
country’s domestic raw materials and unsuited to Zambia’s resource endowment which
warranted less capital-intensive techniques; and (iv) the weaknesses in the implementation
of the plan at both the central and provincial levels, lack of properly appraised projects for
absorbing capital funds, shortages of skilled man-power and managerial capacity, and policy
shortcomings such as pricing of products unrelated to economic costs (TNDP, pp. 2-3).

The domestic economic difficulties were compounded by exogenous factors such as
the collapse of the copper prices on the world market, the negative effect of weather
conditions on crop harvests, the impact of world inﬂation and recession, the fuel crisis, the
break-up of the international monetary system, and the disruption of supply routes caused
by the Rhodesian border closure and the Angolan civil war and the consequential re-routing

of Zambian distribution channels (TNDP, foreword and p. 1).

21

2.2.4 W
The TNDP was also formulated based on the lessons learnt from the past

performance of the economy. It called for a major shift in the investment pattern in favor
of three productive sectors namely agriculture, industry and mining. Again, rural
development was emphasized through the need to expand the agicultural sector to attain
both self-sufﬁciency in staple foods and to promote exports, through the promotion of rural
reconstruction centers, school production units, ranches and state farms. In addition, the
TNDP recognized the importance of maintaining the prominence of village regouping
schemes and model-village projects in the rural areas.

The Plan supportive policies and strategies for the achievement of rural development
included the provision of complementary social and economic services and amenities such
as water supply, electriﬁcation, feeder roads, housing and educational and health facilities.
These were in addition to ensuring that producer prices were based on economic costs;
ensuring timely collection of the agricultural produce; and investment in production and
marketing progams that would ensure provision of adequate and timely agicultural inputs
and services such as credit, fertilizers, implements, seeds, and storage and extension.

Table 2.3 shows the planned expenditure of the investment funds for the TNDP
period whose total was K3,354 million. According to the TNDP, the private sector was only
supposed to utilize about 9% of the total investment funds. The agicultural sector had the
third largest allocation of the total investment funds (about 15%), with the second largest
allocation given to the transport and communications sector (about 19%) and the ﬁrst
largest allocation accruing to the mining sector (about 20%). Out of K673 million allocated

to the mining sector, K580 million (over 86%) was allocated to the copper industry.

22
Table 2.3

Zambia: TNDP Planned Sectoral Investment, 1979-1983

 

Million Kwacha

Parastatals
Govt Tnternal External Private Total

 

 

 

National Commission for Development Planning, Lusaka, p. 37.

m 39.48:! Km Emcee: m
Wmm 440. 0 55 502.0—
Mining 13.0 450 180 30 673.0
Manufacturing 60.0 240 120 30 450.0
Power 50.0 140 - - 190.0
Transport and 220.0 170 230 20 640.0
Comm.(excluding roads)

Public Works (including 155.0 - - - 155.0
roads)

Commerce 10.0 28 - 20 58.0
Housing & real estate 95.0 23 - 125 243.0
Education 106.0 - - - 106.0
Health 43.0 - - - 43.0
Tourism, National Parks 10.0 24 - 20 54.0
& Wild life

Information and 30.0 - - - 30.0
broadcasting

Youth and Sports 4.2 - - - 4.2
Scientific Research 6.0 - - - 6. 0
General Admin. 46.8 - - - 46. 8
Provinces 150.0 - - - 150.0
Total 1439.0 1085 530 300 3354.0
Source: Republic of Zambia, t' o

23
The allocation to the agicultural sector did not increase much from the percentage

allocation in the SNDP (12%).

The performance of the economy during the TNDP period was not satisfactory, as
reﬂected in all macro-economic indicators such as GDP; Government Budget; balance of
payments; general price and employment levels. Real GDP gew by 1% throughout the Plan
period or by an average of 0.2% per annum, compared to the projected gowth in real GDP
of 4.8% per annum (Fourth National Development percent Plan (FNDP), 1989, p. 2). The
only positive gowth rate in real GDP occurred in 1981 when there was 6.2% gowth. But
this was followed by negative GDP gowth rates of 2.8%, 2.0% and 0.4% in 1982, 1983 and
1984, respectively. Moreover, the share of goss ﬁxed capital formation in real GDP
expenditure during the TNDP period averaged 14.3% as against a target of 29%. In
contrast, the share of private and government consumption taken together was 79.7% as
against the planned target of 71%, while domestic savings remained short of planned targets
since the government did not raise savings for the whole period under review (FNDP, 1989,
p. 2). Failure to achieve the projected real GDP gowth rate resulted in the failure to
achieve the projected gowth in per capita incomes of 1.5% over the Plan period.
Aggavated by a higher gowth rate of population, per capita income declined by 12.1%
during the period 1980 to 1984 (FNDP, 1989, p. 2). Inﬂation and terms of trade were also
negatively affected during the Plan period. Inﬂation increased from 12% in 1980 to 19.4%
in 1983. Using 1979 as thebase year, the terms of trade index declined from 59.74 in 1980
to 26.77 in 1983.

With this poor performance of the TNDP, the government embarked on another
strategy for the period 1983 to 1987 which tried to incorporate the mistakes learned from

the past.

24

225B . . II I“ [CH'1233-12EZ
The planned reforms for the period 1983-1987 are laid out in a government

document entitled ”Restructuring in the Midst of Crisis”. It also provides measures of how
to restructure the economy, reduce Zambia’s heavy dependence on the mining sector and
accelerate the country’s social and economic development. The document discusses the
poor performance of the agicultural sector since independence. At the time of writing this
document, there had been no substantial improvement in livestock and crop production; per
capita crop production had been declining since mid-1970s; the agicultural sector had failed
to make a signiﬁcant contribution towards raising incomes in the rural areas. Hence,
regional-income disparities continued between the rural and urban areas.

The document attributes the poor performance of the agricultural sector to several
factors, of which inappropriate or ineffective government policies are a part. These
inappropriate policies include the pan-territorial and pan-seasonal (uniform) pricing system,
which discouraged the development of crops according to regional comparative advantage
and thus discouraged the production of traditional crops like sorghum, cassava and millet;
the changing terms of trade against the agicultural sector that reduced the real producer
prices; the inadequate producer incentives in the late 1960s and the 1970s; the limiting of
marketing functions to government parastatals only that bred inefﬁciencies in the marketing
system due to lack of competition and resulted in increasing budgetary subsidies, which in
turn weakened the capital expenditure progam; the inadequate research and technical
services, especially in the traditional sector; the inefficient and inadequate supply of farming
inputs (like fertilizer, seeds, credit, and farm implements) because of the shortage of foreign
exchange; and insufficient budgetary allocations to the agicultural sector - for example,

while the operating base and personnel of the Ministry of Agiculture and Water

25
Development (MAWD) had gown substantially, its operating expenses for materials and

services did not increase in relative proportions. The result was only a marginal increase
in output.

In the “Restructuring in the Midst of Crisis” document, the government identiﬁed
long-, medium- and short-term objectives which would promote small-holder agriculture.
As a long run strategy, institutional changes and proper and adequate investment allocation
for agiculture were sought. To provide incentives for economically eﬁicient farming at all
levels, it was prescribed that the government should improve producer price setting. The
medium-term strategy called for the government to eliminate subsidies, streamline and
improve marketing and input delivery, and to increase credit availability. In the short-run,
the government would expand commercial agriculture by providing more of the imported
inputs that were scarce, like fertilizer, farm machinery, spare parts and ago-chemicals, to
increase production. The government also undertook some adjustment progams prescribed
by the IMF and World Bank.

However, the economic problems continued in the economy. The gowth in the real
GDP was close to nil and little diversiﬁcation took place. As a result, real GDP per capita
declined by 0.1%, 0.6% and 3.7% in 1985, 1986 and 1987, respectively. The share of goss
ﬁxed capital formation in real GDP expenditure between 1985 and 1987 dropped from
14.3% to 8.6% per annum. In 1987, it further dropped by 7.8% (FNDP, 1989, p.2). This
indicates that there was a contraction in the general performance of the economy. Inﬂation
continued to be on the rise and terms of trade continued to decline. Inﬂation increased A
from 19.4% in 1983 to 53% in 1986 and dropped to 45% in 1987 with 1975 as the base year.

Agiculture, which was supposed to have the geatest potential for gowth, continued to lag

26

behind its potential due to lack of effective and sufficient incentives as well as bad weather

conditions during the 1980s.

226 W

In response to the worsening economic crisis and in need to restore stability and
create a conducive environment for investment and planning, the government decided to
abandon the IMF/ World Bank sponsored progams. It launched its own New Economic
Recovery Progarn (NERP) in May, 1987. The NERP had three interrelated goals:

(a) to reactivate the economy on the basis of maximizing the use of own resources;

(b) to change the consumption pattern from ir‘nport dependence to reliance on

domestically produced goods, and services; and
(c) to rationalize the use of foreign exchange as a strategic resource. (FNDP, 1989,

p- 6)

On the basis of the NERP, the Interim National Development Plan (INDP) was
drawn-up covering a period of eighteen months, starting July 1987. The progam had
notable sectoral and overall performance achievement (FNDP, 1989, p. 22-3). GDP gew
by 6.3% at constant 1977 prices and per capita income gowth rate was positive for the ﬁrst
time since 1981. The agicultural sector gew by 6.4% over the previous year’s level.
Although the mining sector’s gowth rate was below the target, the sector gew by 2.7% in
1988 and contributed about 9% to GDP. The manufacturing sector recorded good
performance in the pulp and paper products subsectors, textiles and clothing and food,
beverages and tobacco subsectors. The index of industrial production of the pulp and paper
products subsector increased from 204% in 1987 to 322.2% in 1988 while that of the textiles

and clothing subsector increased from 115.7% in 1987 to 200% in 1988. The food,

beverages and tobacco sub-sector’s index increased from 95.1% in 1987 to 115.7% in 1988.

27

To continue with the implementation of the NERP, the government launched the

Fourth National Development Plan (FNDP) in 1989 covering the period 1989-1993.

22.61 W

The present policy orientations for economic development are outlined in the Fourth
National Development Plan (FNDP). The FNDP is the ﬁrst broad medium-term progam
in the implementation of the country’s New Economic Recovery Progam. The theme of
the progam is "gowth from own resources”, which incorporates the idea of self-reliance in
all aspects. The major agicultural-related objectives of the FNDP are to improve
agicultural productivity and increase production for export; to achieve food self-sufﬁciency;
to promote import-substitution; to diversify the economy so as to reduce economic
dependence on revenues from copper; to streamline the marketing system so as to improve
input and output marketing; and to improve the standard of living of the rural population
through increased employment creation and income generation. Increased production at
minimum costs is envisaged in the short-run while providing medium- and long-term
facilities to consolidate economic gowth and development.

However, minimum costs of production can only be achieved through efﬁciency in
agricultural production and marketing. The FNDP is meant to bring about this efficiency
through various strategies. The overall strategy is to provide price and non-price incentives
as well as provide the necessary support services. Seasonal and regional pricing is envisaged

in order to encourage comparative advantage and on-farm storage. The existing price
incentives, foreign exchange retention scheme and tax incentives are expected to increase
production, food availability and employment opportunities. The need to provide sufficient

credit is recognized, although this is in conﬂict with a credit squeeze aimed at controlling

28
money supply in the economy. The plan also emphasizes the need to mobilize funds for the

Irrigation Revolving Fund, the need to maximize the use of electricity in rural areas and to
strengthen the irrigation branch in the Ministry of Agiculture. This is to hedge against bad
weather which characterized the country in the last decade, and has contributed to the poor
performance of the agicultural sector particularly in the 1980s.

The plan also incorporates nutrition and population issues into the agicultural
development strategy. With regard to nutrition, the aim is to provide enough staple food-
stuffs, fruit and vegetables and protein and energy rich products. Improvement of staple
food accessibility by all consumers has also been emphasized through the improvement and
expansion of the local production, processing, preservation, storage, and distribution system.

A break-down of resource allocation in the FNDP is given in Table 2.4. Out of a
total of K26,102.6 million, the mining sector has the largest share in investment allocation
(K7492.9 million), compared to manufacturing and trade and agriculture and water
development sectors which have been allocated K4,946.6 million and K3,152.7 million,
respectively. The new emphasis in the mining sector is on the expansion of the mining
sector to some minerals other than copper, while the manufacturing and trade sector is
trying to promote import substitution and shift consumption from imported to domestically
produced goods and services. Therefore, the agicultural sector is considered the least
among the major economic sectors of the economy, although most of the Zambian

population still depends on it for a living.

23 W
Table 2.5 gives a summary of some macroeconorrnic indicators for the period 1960-

1988. It shows that from 1960 to 1963 (years before independence), the average annual

29
Table 2.4

Zambia: FNDP Planned Sectoral Investment, 1989-1993

 

 

 

MWwacha
m Govt Parastatal Private Iota!
Bum Qrsanizations Sector
W Water Devt. 505.4 1639.4 1W

Land & Natural Resources 141.6 490.1 283.1 884.8
Mining 515.3 5061.0 1916.6 7492.9
Manufacturing & Trade 791.5 2572.2 1582.9 4946.6
Tourism 118.5 385.1 237.0 740.6
Energy 179.6 583.7 359.2 1122.5
Transport & Comm. 348.1 1131.4 696.2 2175.7
Construction & Housing 203.6 661.6 407.2 1272.4
Education 736.8 138.2 46.1 921.1
Health 689.4 129.3 43.1 861.8
Mass Communications 320.3 60.1 20.0 400.4
Employment & Social Devt. 611.2 114.6 38.2 764.0
Govt. Admin. 1336.1 - - 1336.1
Total 6497.4 12966.7 6638.5 26102.6

 

Source: Republic of Zambia. WWW
Wanda National
Commission for Development Planning, Lusaka, p.38.

30

Table 2.5

Zambia: Macroeconomic Indicators, 19604988

 

 

(Mn—k) MI!) W of
m 331.3121? W W‘ hm am
1900 3775 1.193 2303 32.94 40.10 5300
1901 3759 1.105 2433 40.04 43.10 53.13
1902 3743 1.131 22.09 2319 4539 5310
1903 3313 1094 1730 21.90 4024 53.10
1904 4323 1.23s 1135 32.47 3103 05.14
1905 5593 1547 24.47 43.00 3099 52.40
1900 5307 1.442 2339 3321 3950 53.77
1907 5029 1.403 3033 3030 43.47 49.03
1903 5m 1401 32.39 27.97 4420 51.41
1909 5940 1.400 1311 2154 32.42 05.03
1970 0157 1.470 2000 2091 3304 5019
1971 0271 1.455 2332 2321 39.12 5039
1972 0042 1.493 31.10 2951 39.40 3753
1973 0057 1.459 2923 3023 2131 4004
1974 7093 1512 3005 29.01 2035 37.93
1975 0913 . 1.423 4050 17.09 37.73 32.91
1970 7209 1.443 2331 1340 24.74 39.00
1977 0372 1.332 24.07 1123 2009 3505
1973 0905 1294 2330 1035 21.90 3052
1979 0700 1212 14.14 34.00 2233 40.93
1930 0901 1203 22.33 24.07 45.40 2133
1931 7321 1227 19.94 15.93 41.15 23.04
1932 7119 1.140 10.77 1725 3001 27.02
1933 0930 1.079 13.75 20.00 31.79 30.04
1934 0955 1.033 14.03 22.07 3233 30.05
1935 7072 1.009 14.90 1332 3322 3374
1930 7111 0.977 2331 15.91 45.04 4559
1937 7093 0.939 13.74 13.00 33.73 4135
1933 0394 0337 1137 23.03 2037 3420
MW
023% 231% 430% -1534% 025% 0.04%
1131 974
4.79% 2.40% 033% 023% 033% 320%
MW
027% 423% 45.05% 4.00% 4.0% 435%

 

 

 

 

 

Notes:

' Real GDP is derived by deﬂating GDP in current prices dy the deflator (1935-100)

bInvestmentisdeﬁnedasGrossFrxedCapitalFormationplusCliangeinStocks

c Saving '8 derived as a residual (GDP-consumption)

' Source: Republic of Zambia, Wm various issues.

Republic of Zambia, National Accounts Statistics Bulletin, No. 2, Central Statistics Ofﬁce, Lusaka, 1988.

Republic of Zambia, Zambia in Figures, 1991, Central Statistics Office, Lusaka, 1991.

T00

pep

thar
me
4m
9ng

and 1

31
gowth rate of real GDP was 0.28% while the real per capita GDP had a decreasirng gowth

rate of -2.31%. After independence in 1964 and before the world copper prices started to
ﬂuctuate in 1974, average real GDP gowth rate increased to 4.79% while that of average
real per capita GDP improved tremendously to 2.4%. However, both the real GDP and the
real GDP per capita average gowth rates became negative during the period between 1975
and 1988, although the nominal gowth rate was positive.

The percentages of sectoral contributions to total GDP for selected years are given
in Table 2.6, which shows that the contribution of the mining sector to total GDP has been
decreasing over time. In 1970, the mining sector contributed about 34% to total GDP while
in 1989 it only contributed about 7.8% at constant (1977) prices. On the other hand, the
GDP shares of the other sectors of the economy have been increasing. The contribution
of the agicultural sector to GDP rose from 10.8% in 1970 to 19.1% in 1989 at constant
(1977) prices.

However, this increase in earnings, particularly in the agicultural sector, has not
been at a rate fast enough to beat the rate at which population has been increasing. For
instance, the agicultural sector currently contributes less than 20% to the GDP and has
recorded an annual gowth rate of just about 2.4% in recent years on the average while the
population gowth rate has been over 3% (Central Statistical Office, 1990).

The contribution of the agicultural products to total exports has never been more
than 5% and has declined to less than 2% since 1978. Agicultural imports have far
exceeded agicultural exports over the years. For instance, according to the Quarterly
Agicultural Statistics Bulletin, Vol. IV, March /J une, 1975, agicultural exports were only
worth K9.63 million and K 5.5 million while agicultural imports were worth K29.3 million

and K607 million for the years 1965 and 1972 respectively.

32

Table 2.6

Zambia: Percentage Distribution of GDP by Kind of Econonnic Activity at Constant
1977 Prices

 

Total GDP (%)

Total GDP value (1977 prices)
Agic., Forestry & Fishing
Mirning & Quarrying
Manufacturing

Electricity, Gas & Water
Construction

Wholesale & Retail Trade
Hotels & Restaurants
Transport, Comm. & Storage
Financial Institutions & Insurance
Real Estate & Business Service

Community, Social & Personal Services

Import Duties
Less imputed Banking Service Charges

1210M128912§512§2

100 100 100 100
304 344 425

11 16 15 17 19

34 12 10 9 8

10 18 19 21 25

2 3 3 4 2

7 6 5 4 3

9 10 10 9 8

1 2 2 3 2

4 7 6 5 5

2 4 3 3 2

3 6 7 9 9

14 17 17 17 17

3 3 2 1 1

1 1 1 1 1

 

* Preliminary estimate likely to undergo revisions

Sources:

Republic of Zambia, Central Statistical Ofﬁce, Mgrlthly Mg!

Statistics, (Lusaka: Central Statistical Ofﬁce), various issues.

Republic of Zambia, W (Lusaka: Ministry of Finance
and Development Planning and National Commission for Development

Planning), various issues.

Note: Figures are rounded off to the nearest whole number.

33
The poor performance of the Zambian economy in general and the agicultural

sector in particular can be attributed to a number of factors. A major turning point in the
Zambian economy came in 1975 as copper prices on the world market started to decline
while domestic copper production was declining too.

After the 1965 Unilateral Declaration of Independence (UDI) of Southern Rhodesia,
Zambia introduced sanctions (tlnrougln import controls) against Zimbabwe (then Southern
Rhodesia). Since Zambia is a land-locked country, the import controls meant the need for
more export earnings. However, since export earnings were declining, import controls had
to adjust to the former.

The economic difficulties resulting from cutbacks in exports, imports, and revenues
were aggavated by the effects of droughts on agicultural production. Moreover, the poor
plan performance is partly attributed to irnconsistencies in objectives and lack of effective
coordination between the various organs of the Party and Government responsible for
policy formulation and implementation (ILO,1977, p.265; Chileshe, 1986, pp.74-77). The
combination of these problems, plus such problems as shortage of technical and managerial
skills, world and national irnﬂation, the oil crisis, the collapse of the International Monetary

System, etc, resulted in a huge ﬁnancial and economic crisis.

2.45'05'1 12]..

24.1mm
Up to 1971, there were regional producer price differences that mainly reﬂected

transport cost differentials. However, during the 1972/73 marketing season, the government
introduced the policy of uniform pricing for all controlled products. At that time, controlled

products included maize, maize meal, goundnuts, different types of beans, cowPeas,

34
sunnhemp, sunﬂower seeds, sorghum, soya beans and seed cotton. Uniform prices were set

annually on the basis of a number of often conﬂicting criteria, namely cost of production,
fair 3 return to producers, fairness to consumers, import-export parity, food security, and
political acceptability. Fairness to consumers and political acceptability had been the
overriding considerations (Jansen, 1988), since the pricing system was inﬂuenced by the
desire to provide cheap food for the beneﬁt of the vocal urban population.

The pricing policy was modiﬁed slightly in 1982. Some consumer items were
decontrolled. Consumer items such as mealie meal, wheat products, sugar, etc. were still
controlled. The government continued to control all producer prices of major commodities,
fertilizer selling prices as well as tractor hire charges. In 1985, producer prices of most
agicultural cr0ps were decontrolled and in June of 1989, all producer prices were
decontrolled apart from that of maize. However, the government still sets ﬂoor prices of
major agicultural crops as a way of giving farmers a strong negotiating base for their cr0ps.

The policy of uniform pricing has been the subject of considerable criticism from
both local and international organizations. The policy has resulted in inefﬁcient allocation
of resources. It has encouraged the production of maize at the expense of traditional crops
such as cassava, millet and sorghum. The neglect of traditional crops has contributed to food
insecurity, especially in areas that are prone to droughts, since it provides no incentives for
on-farm storage. The policy has made it unproﬁtable to store crops on farms and has
encouraged farmers to sell crops immediately after harvest. Thus, the policy has overtaxed
the buying, transport and storage facilities of government marketing agencies.

Finally, in 1991, the maize producer price was decontrolled. Now, the government
just sets the ﬂoor price for maize. This move was meant to attract private traders and other

parastatal organizations into maize marketing to make it more efﬁcient and to relieve the

35

government of the high subsidy expenses it was incurring in maize marketing. However,
while the producer price of maize is decontrolled, the government continues to set a
uniform retail price of maize mealie meal. This has been unattractive to private traders
since the whole business is unproﬁtable. Maize mealie meal retail prices are set at too low
a level to allow private traders to recover their handling and transportation costs and make
a proﬁt. As a result, Zambia Cooperative Federation (ZCF) and its designated afﬁliates
have continued to monopolize maize marketing.

. The government has not encouraged the production of traditional staple food crops
(sorghum, cassava and millet) to the same extent as it has for maize. This is partly reﬂected
in the fact that the government never used to set prices or market some of the traditional

crops until the 1970s for millet and the late 1980s for cassava.

2.4.2 21310311032009:

Parastatal marketing organizations have always existed in Zambia, and, until recently,
they monopolized commodity marketing and input distribution. Since the pre-colonial
period, maize marketing has been the sole responsibility of the gain marketing boards and
their agents. Details about this will be discussed in chapter three. Since the decontrol of
some commodities, some commodity speciﬁc organizations have been formed to render
output marketing and input distribution services for the respective commodities. For
instance, the Lint Company (Lintco) is responsible for the cotton subsector while the
Zambia Coffee Company (ZCC) is responsible for coffee. In the case of seed and fruit and
vegetable, there is the Zambia Seed Company (ZamSeed) and the Zambia Horticultural
Company (ZamHort), respectively. Other commodity-speciﬁc marketing organizations

include the Zambia Pork Products (ZAPP) for pork and pork products; the Zambia Cold

36
Storage Board (ZCSB) for beef and beef products, the Dairy Produce Board (DPB) for nnilk

and milk products; and the National Tobacco Company/Tobacco Board of Zambia
(NATCO/TBZ) for tobacco. Private traders have also been involved in the marketing and
processing of many commodities such as beef, milk and milk products, fruit and vegetables,
wheat, soybeans, etc.

2-43 W

In general, subsidies are meant to integate an economy and to foster a balance
among various sectors (Timmer M. 1983). In Zambia, priority in the allocation of
subsidies was meant to be given to the rural sector and to industries which increase their
use of local resources in the production process.

Between 1965 and 1975, investigations into the actual beneﬁciaries of subsidies
revealed little evidence to support this purpose of subsidies (Chileshe, 1986). Most subsidies
have been and still are used to support certain public institutions which could not survive
commercially. Many subsidies were used to compensate for price increases and for the
losses incurred by public enterprises, resulting into a fast rise in subsidies. Subsidies
increased from K 4 million in 1965 to K 16 million in 1966, to K24.1 million in 1967 and to
K 34 million in 1974. In 1967, the government adrrnitted in the SNDP that subsidies were no
longer acting as anticipated and that the continuous losses made by the subsidized public
enterprises showed that

subsidies of any type, whether on maize or any other commodity, not only result in

a misdirection of resources, but are also apt to put a premium on inefﬁciency unless

they are given after careful scrutiny of the various factors going irnto the cost of

production of commodities concerned. (SNDP, p.7).

Subsidies were also used to establish and to prop up many of the country’s

cooperative societies. Many cooperatives were led by goups of misguided ofﬁce bearers

37

who under normal circumstances would not have been permitted to run such enterprises.
There was also lack of supervision by the central government authorities. As a result, many
cooperative societies wasted resources. Therefore, the fact that a geater proportion of the
Zambian population lives in rural areas and derives its subsistence therefrom was not a
good economic reason for subsidizing cooperative societies on a country-wide basis.
There have also been commodity subsidies. Before the introduction of uniform
producer prices irn 1972/73 season, the government subsidized crop marketing in areas
where it wanted to increase crop production, and when the marketing board could not cover
its costs. But with the introduction of uniform prices, the subsidy situation worsened. The
marketing agencies involved in handling commodities (particularly maize) and inputs have
continuously and increasingly been subsidized to cover their handling, transportation, storage

and management overhead costs. Maize subsidies will be discussed in chapter 3.

2.4.4 11400110113
Trade policy has also had a geat impact on the food and agicultural sector. One of

the main objectives in the agicultural sector has been export diversiﬁcation. Under the
50% foreign exchange retention scheme introduced in 1984, exporters of non-traditional
products were allowed to retain 50% of their export earnings in foreign currency for use in
acquiring necessary inputs for their operations, thereby providing incentives for export
production. Currently, some non-traditional exporters are allowed to retain 100% of their
earnings, in special cases.

Other non-price incentives for expanding production, processing and marketing are
provided in the Investment Act of 1986 and various government statutory instruments.

These include concessional tax rates for agicultural income; 100 percent write-off of capital

38

cost of farm implements and machinery within two years; tax rebates; foreign exchange
retention allowances for producers marketing at least 5,000 bags of maize; and removal of
duties on agicultural maclninery and most agicultural inputs. For example, tax rates on
farming incomes were reduced from 80% to 25% in 1981 and to 15% in 1982 (Jansen, 1988,
p. 48). However, the beneﬁciaries of these incentives are mostly emergent (medium-scale)
and commercial farmers who constitute only about 9% of total farmers in Zambia.
Moreover, these strategies are not intended to promote traditional crops like sorghum.

Export activities are also supported under the Export Development Act of 1985.
This Act led to the establishment of the Export Council of Zambia, whose functions are to
formulate and approve national emort promotion policies, and the establishment of the
Export Board of Zambia, whose main responsibilities are to develop and promote exports
and give assistance to exporters. In addition to the two institutions, the government
established the Zambia Export and Import (EXIM) Bank in 1987 to provide ﬁnancial
resources for trade to the non-traditional export sector.

Apart from government institutions, a goup of commercial farmers have formed the
Zambia Export Growers Association (ZEGA). The ZEGA was formed in 1984 to act as
an agent in the export of fruits and vegetables. ZEGA coordinates, identiﬁes markets,
ensures quality control, etc. for exports on behalf of its members. The association is also
used as a negotiating forum on issues such as air freight space and charges as well as any
other issues affecting the export gowers as a result of government policies.

Besides activities towards export promotion and expansion, the government has also
maintained its food security concerns. It is government policy that there should be no

exports of major food stuffs until self-sufﬁciency has been achieved. Other food stuffs can

39
be exported but only with government approval. International trade in maize has always

been and continues to be done by marketing boards (ZCF).

Zambia is ideally placed to export maize to Zaire, Malawi and Mozambique.
However, in 1989/90 and in 1990/91 when there was supposed to be relatively good maize
exports, these markets were not utilized to the extent they could have been. This could have
been because ZCF has not yet fully understood the intricacies of international trade.
Moreover, Subrarnanyan (1990) shows that the system itself has defects. Although only ZCF
is allowed to export maize, maize is generally the property of the vaincial Cooperative
Unions (PCUs). Therefore, ZCF is supposed to pay the PCUs ﬁrst and then export maize.
But what usually happens is that the ZCF just exports and promises to reimburse the PCUs
later. However, even after receiving the export payment, ZCF does not pay the PCUs right
away. Hence, the PCUs have to bear the interest charges on money held in maize stocks,

thus worsening their already poor ﬁnancial position.

245 W

Like marketing and pricing policies, the exchange rate policy has gone through
changes over the last few years. Prior to October 1985, the government followed a system
of ﬁxed exchange rates. The system of weekly auctions was then introduced and lasted until
May, 1987. During the ﬁxed exchange rate system, the country’s currency was overvalued.
This meant that imported goods were cheaper than those produced locally, suppressing
demand for locally produced goods. The overvalued currency also made exports more
expensive on the international market, hence less competitive.

During the auction system, the kwacha was devalued signiﬁcantly within a short

period of time. The immediate effect was a general increase in prices, especially for

40
imported inputs. The agicultural sector depends on imported inputs such as fertilizers,

machinery and chemicals. The increase in prices of these imported inputs as a result of
devaluation made the auction system counter productive. Exports became cheaper and
more competitive on the international market. This opportunity was not fully exploited,
however, due to government export restrictions on major agicultural commodities for self-
sufﬁciency reasons. In addition, the agicultural sector did not beneﬁt much from the
auction system partly due to low liquidity levels and the irnability of most farmers to compete
in the auction. Furthermore, the general increase in prices due to the auction system had
geat adverse effects on the poorer sections of the population. The end result was that the
auction system failed to beneﬁt either the sectors considered as priority sectors by the
government, and the poor, whose standard of living the government had wanted to improve.

For the above reasons, the government decided to discontinue the auction system,
and reverted to a ﬁxed exchange rate system. A Foreign Exchange Management Committee
(FEMAC) was established to prioritize the country’s scarce foreign exchange resource
allocation. FEMAC allocates foreign exchange every two weeks. Several government
departments are represented in FEMAC. These include Ministries of Finance and NCDP,
Agiculture, Commerce and Industry, and Cooperatives. Some government departments and
agencies like the Department of Customs, Bank of Zambia, Prices and Incomes Commission
and representatives from commercial banks are also represented. Currently, the central
bank has a two-tier exchange rate system under which there is one exchange rate for its
dealings with the government and another exchange rate for' its day-to-day dealings with

customers other than the government.

41
24.6 W

Research and extension policies are intended to provide suitable technological
packages and extension services to farmers. However, available technological packages
remain largely unsuited to the needs of small-holders in general and subsistence farmers in
particular. Very little attention has been given to traditional subsistence crops like cassava,
sorghum and millet. The extension services have not also provided advice on practical ways
of increasing productivity within the constraints imposed by labor, managerial skills, and
capital resources at the disposal of farmers. Because extension services have been geared
to commercial farmers mainly along the line of rail, there is little, if any, technical advice
that extension staff can convey with conﬁdence to small-holders. Since maize is produced
mostly by small-scale farmers, if maize production has to increase, there is need to improve
extension services to these farmers so as to narrow or even eliminate the gap between them

and commercial farmers.

2.4.7 9330110113

Although almost all farmers, including small-holder farmers, use their own resources
to purchase some of the inputs for agicultural production, credit provided for crop inputs
is the major type of farm credit in Zambia. For instance, the maize subsector study of 1990
done by the Zambian government reported that in 1983/84, 93% of all funds loaned by
lenders were short-term credit, of which 83% was for crop inputs (Republic of Zambia,
1990).

Commercial farmers get credit for agicultural inputs from commercial banks since
they have collateral. For small-scale farmers who lack the collateral needed to obtain credit

from banks, the two main sources of seasonal loans are the Cooperative Credit Scheme

42
(CCS) and Lima Bank. The CCS is operated by the ZCF Finance Services Division, which
assumed responsibility for the Rural CooperativeAgicultural Credit Scheme (RCACS) in
1986. For the RCACS to take off, the government provided ZCF with a gant of K60
nnillion as a revolving fund for credit provision. This enabled the cooperative movement to
acquire a substantial credit organization at no cost to its members.

Despite the sources of credit for commercial as well as small-scale farmers, their
loanable amounts have not been large enough to meet the demand. Table 2.7 below shows
the number of applicants and the approved loans for the period 1979 / 80 to 1983 / 84. The
Table shows that during these seasons, the number of short-term loans approved far
exceeded the number of applications. Only 41.5% of the short-term loan applications were
approved in the 1981 / 82 season. According to the same study, the number of applicants was
only a small fraction of the total number of maize producers.

The credit system faces several problems: high nominal interest rates and negative

real interest rates due to the high inﬂation rate; delays in the repayment of loans for seeds

Table 2.7

Zambia: Number of Short-term Loan Applicants and Approved Loans,
1979/80-1983/84

MWWW

 

 

1979/80 31773 21428 07.4
1980/81 50127 22302 44.5
1981/82 51315 21281 41.5
1982/83 45039 22856 50.1
1983/84 34931 20278 58.1
Source: Republic of Zambia. W

W 1990. p- 65

43
and fertilizers due to delays in paying farmers for their maize; inadequate credit facilities

for small scale farmers; low loan recovery rates; and inadequate training for staff of lending

institutions.

25 Summary
As a backgound to the analysis of the maize subsector in Zambia, sectoral and

overall economic performance of the national economy is reviewed, with particular emphasis
on agiculture. Each national development plan that has been employed by the Zambian
government since independence in 1964 is discussed with speciﬁc concentration on the
allocation of public funds. This is done to partly show the priorities of the government in
its development strategies over time, especially as the government has been concerned with
reducing the dependence on copper and expanding agicultural production.

The general feature of Zambian development planning has been the gap between
policies, plans and actions. Sectoral and overall performance has fallen short of the set
targets. Although the development plans have emphasized prioritizing agicultural
development, in reality, not much attention has been paid to the development of the
agicultural sector. The mining industry has continued to be the highest priority where fund
allocation is concerned. As a result, the real GDP per capita has been declining since the
late 19703. However, some of the objectives of the agicultural sector in the FNDP, such
as the removal of the subsidies, the h'beralization of the marketing system for all crops, etc.,
are moving on the right track, although the liberalization of maize marketing is not yet
operational because the government still controls the maize mealie meal retail prices.

Speciﬁc agicultural policies that affect the maize subsector are also discussed.

These reﬂect how most of the policies have been beneﬁtting commercial farmers at the

44
expense of small-scale farmers. Sirnce maize is mostly produced by small-scale farmers while

sorghum is solely produced by small-scale farmers, these farmers have not beneﬁtted much
from these agicultural policies.

3. THE MAIZE SUBSECI‘OR IN ZAMBIA

To be able to specify behavioral relationships required to analyze a subsector, there
is a need to understand the operations of the subsector in question and how it is affected
by the environment and institutions within which it operates. This chapter gives a

descriptive and diagnostic analysis of the production and marketing activities in the maize

subsector.
alums
3.1-1 hammer;

There are three categories of farmers in Zambia: small-scale, medium-scale and
commercial farmers. Maize is produced throughout the country by all categories of farmers.
Commercial farmers represent only about 1% of the total farmers yet they contribute about
30% to total maize production. Medium-scale farmers constitute about 8% of total farmers
and they contribute about 35% to total maize production. The remaining 35% of maize
production accounted for by small-scale farmers who constitute about 91% of total farmers
in Zambia.

Maize production and marketing has been heavily supported by the government at
the expense of the traditional food crops. This is partly reﬂected in the quantities of maize
and sorghum produced over time. Figure 3.1 shows that from 1964 to about 1975, the
quantity of sorghum produced was roughly 5% that of maize. Since 1975, sorghum
production has dropped to about 1% of total maize production. However, it should be
noted that the sorghum production ﬁgures are likely to be underestimated considering the

low proﬁle sorghum is given and that there is no concrete time series data available on it

45

46

Figure 3.1 Maize and Sorghum Production in Zambia: 1964-1990

 

2000
1300-
A ’5‘
ggtsoo- 2
Q
.—
§ 1400- 8
g o
g 1200- g
0 3
[3|
5 1000- 5
‘ 5
I In

or

O

O
l

 

 

 

600

 

I I I I I I I I I I I I I I I I I I I 10
1975 1980 1985 1990
YEARS

i303 " 197d

 

[:4F-MAQE -*—-SORGHUM |

 

within the country. The only data that is reported on sorghum is the marketed output
which would also not be representative of production since most of it is consumed within
the areas where it is produced.

There have been ﬂuctuations in the quantity of maize produced throughout the
period under study. In the 1960s, maize production was relatively low because the
government had not yet started promoting its production and marketing. During that
period, the policy was more oriented toward the building of infrastructure in readiness for
agricultural take-off than toward direct agicultural production. In the early 1970s, maize
production increased substantially. This increase can partly be attributed to an increase in

the maize producer price. However, in the late 1970s, there was a drastic decrease in maize

47

production. Although maize production started to pick-up in the 19803, for the most part
of the 19803 this increase in maize production was erratic, partly due to bad weather
conditions.

Over the years, especially since the 19703, the government has taken notable steps
towards encouraging maize production by revising nominal producer prices upwards.
However, the producer prices have not been signiﬁcant enougln to stimulate increased
production. The real producer prices have been low particularly in the 1980s resulting in
a decrease in maize production by commercial farmers who switched to the production of
the other cash crops whose prices were more attractive. Small-scale farmers, who have
lesser opportunity to switch to other commodities continue to increase maize production
while decreasing the production of traditional food crops like sorglnum, millet and cassava.
This has been so because the government, through the Provincial Cooperative Unions
(PCUs), gives some farmers maize seed and fertilizer on credit relatively easily under its
maize promotion strategy. However, the amount of these inputs issued on credit is limited

by the availability of government funds and are not enough to meet the demand.

3.1.1-1W
The increase in maize production in the mid-1970s and in the 1980s can partly be

attributed to an increase in the areas harvested of maize, as shown in Figure 3.2. Although
area harvested of maize has also been ﬂuctuating like maize production, the cycles are less
pronounced in the case of the former. However, the drastic decrease in maize production
in the late 19703 and early 19805 is very much reﬂected in the decrease in the area

harvested of maize during that same period. These ﬂuctuations in maize area can partly

48

Figure 3.2 Zambia: Relationship between Area Harvested of Maize and the Real Producer
Price of Maize, 1964-1990

 

 

 

 

1200 5

c
"1100- 2
E ~4.5 >
81000- If
8 9.4 E
o 900- a.
E a
I; 300- -5.5 §
5 8
I 700- 0
E "3 ”
2 000- 5'
kl ~25 z
2 500- ' 2’
i E

400

I I I I I I I I I I I I I I I I I I I j I I I I I I I I 2
1965 1970 1975 1980 1985 1990
YEARS

 

+ MAIZE AREA y +REAL PRODUCER PRICE

 

be attributed to the weather that characterized the period. The rains came later than usual
so that farmers could not plant much of the land with maize considering its gestation period.

Figure 3.2 shows that the area harvested of maize increased from 0.89 million
hectares in 1964 to 1.1 million hectares in 1976. It then dropped to 0.95 million hectares
in 1977 to about half a million hectares in 1982. However, it increased again from 0.547
million hectares in 1983 to about 0.905 million hectares in 1989, but dropped again in 1990
to 0.764 million hectares. Figure 3.2 shows a big unexplairned drop in the area harvested of
maize in the late 1970s. On the overall, area harvested of maize constitutes about 70% of

the total area under cultivation in the country.

49

Figure 3.2 also shows some responsiveness of farmers to real producer prices of
maize in some years particularly between 1968 and 1973 and between 1976 and the early
1980s. Although there is a positive relationship between the area harvested of maize and
the real producer price of maize, a distinct trend is absent. Thus the real maize producer
price would not be expected to be a signiﬁcant determinant of the area harvested of maize

in the econometric model, althougin the sign is expected to conform to economic theory.

3.1-1.2 Wield

Average maize yield has generally been increasing over time, as shown in Figure 3.3.
This increase in maize yield can partly be explained by the increase in the use of commercial
fertilizer and hybrid maize seed which is believed to have a higher yield rate than the
traditional maize seed. Currently, there are about ten hybrid maize seed varieties utilized
in Zambia. Moreover, about 70% of the total fertilizer sold in the country is used in maize
production (Republic of Zambia, 1990, p.54).

The maize yield response to the fertilizer consumption index (see its calculation in
Appendix C) is shown in Figure 3.3, which shows that there is a causal relationship between
maize yield and fertilizer consumption. Although hybrid seed and commercial fertilizer are
used in maize production, there have been problems of accessibility and delivery of these
inputs to farmers in terms of timing, quantities and suitability of different varieties to the
ecological conditions of various places. For instance, there have been times when the maize
area has increased proportionately more than the increase in fertilizer use in maize
production as was the case in the 1980s (Republic of Zambia, 1990). This implied that
average fertilizer application rates decreased thereby reducing the chances of having better

maize yields. This was mainly due to the increased proportion of maize produced by small-

3.1

50
Figure 3.3 Zambia: Relationship Between Fertilizer Consumption Index and Maize Yield,
1964-1990

 

bl

MAIZE YIELD (MT/HA) AND FCI
0.

 

 

 

0 ‘I I I I I I I I I I I I I I I I I I T IﬁT I I I I
1965 1970 1975 1980 1985 1990
YEARS

 

+ YIELD —"— FERTIJZER "DEX

 

scale farmers who have little access to fertilizer relative to commercial farmers. The same
study reported that commercial farmers who are in a better position to buy and store large
quantities of fertilizers base their decisions regarding fertilizer type and application levels
mairnly on the analysis of potential proﬁts. This is not the case with most of the small-scale
farmers who have little cash available for fertilizer purchases even if the yield response

potential is good.

3.1.2 Net Imports 91 Mar: e

Although maize production can generally be said to have been increasing, the

population gowth rate has been faster than that of maize production. This can partly be

5 I
explained by the fact that there has been nnigation from the rural areas, where most of the

maize is produced, to the urban areas. Domestic maize production has not been sufficient
to meet the increased demand for maize, particularly in the late 1970s and in the 1980s. As

a result, Zambia has been a net importer of maize over the years.

32 Marketing.
Parastatal marketing organizations have always existed in Zambia, and, until recently,

they monopolized commodity marketing and input distribution due to various government
controls and policies. Food-gains marketing has been one area that has been subjected to
extensive controls by the government for a long time. The controls were instituted by the
colonial administration in the 1930s and were inherited by the post-colonial administration
at independence in 1964. Since 1964, the government has experimented with a number of
marketing arrangements for controlled agicultural products.

Upon independence in 1964, two statutory marketing agencies, namely the Grain
Marketing Board (GMB) and the Agicultural Rural Marketing Board (ARMB), were
established and given the responsibility of marketing agicultural produce. These two
marketing boards lasted up to 1969. The GMB was initially responsible for marketing
certain controlled agicultural crop products and residual crop products along the railway
line. Later on, its operations expanded to include the distribution of farm inputs such as
seeds and fertilizer. Controlled products were products which producers were obliged to sell
only to the GMB or its appointed agents. Residual crop products were products which were
not necessarily supposed to be sold to the Board, but could be delivered to the Board if the
farmer so wished. Being one of the controlled products, maize produced along the railway

line was marketed by GMB which lasted from 1964 to 1969.

52

The ARMB was responsible for marketing crops in ”nonviable” areas. A nonviable
area was an area which had agicultural potential but in which the value of the surplus
agicultural produce was insuﬁicient to cover the transportation and handling costs and
where markets were unduly depressed (Dodge, 1977 p. 82). It was the Minister of
Agiculture’s responsibility to declare an area nonviable. The ARMB had the monopoly
over marketing of certain commodities and the provision of general marketing services in
all the areas that were declared nonviable. Maize and sorghum were among other crops
that were marketed by ARMB. The ARMB also used to distribute seeds, fertilizer, and
other agicultural requisites. Some subsidies were given to the ARMB to allow it to provide
producer price incentives irn the nonviable areas, although this was envisaged as a temporary
measure.

In September 1969, the GMB and the ARMB were merged to form the National
Agicultural Marketing Board (NAMBoard). NAMBoard was established primarily to
function as the government’s marketing agency; to implement the country’s major food
policy; and to ensure a guaranteed market for maize, sorghum, and other controlled crops.
Therefore, upon formation, NAMBoard had a monopoly in the purchase, sale, import,
export, and storage of maize; in ensuring availability of marketing outlets for the maize
surplus regions; in ensuring adequate and equitable maize supply in deﬁcit regions; in the
marketing of other controlled crops including sorghum; and in the procurement and
distribution of fertilizers. However, in 1972, the government decontrolled the marketing of
some crops including sorghum, but not maize. That is, while maize continued to be
marketed solely by NAMBoard, sorghum became a ”residual” crop that could be marketed
by private traders too. The government continued to set ﬂoor prices for decontrolled crops,

thereby assuring farmers a minimum price they could receive for their crops.

 

 

 

53
Later on, NAMBoard’s functions increased to include the marketing of some

additional crops and distributing other inputs like seeds, chemicals and implements,
throughout the country. Increased responsibilities made NAMBoard inefficient. Additional
inefficiencies resulted from insufﬁcient marketing funds for controlled crops, particularly
maize, since both the NAMBoard buying and selling prices of various commodities were
determined by the government. These prices were inadequate to cover handling and
transportation costs and allowed for no proﬁt margin at all. As a result, NAMBoard
became heavily dependent on subsidies from the government for its maize marketing
operations. Delays in the disbursement of these subsidies to NAMBoard impaired its ability
to meet ﬁnancial obligations to third parties, and to procure and distribute inputs and
outputs in a timely manner.

The increased and insupportable responsibilities given to NAMBoard necessitated
the formation of the Provincial Cooperative Unions (PCUs) in 1981 as one of the
government’s measures to improve maize marketing. During the co-existence of NAMBoard
and PCUs, the government kept on shifting the marketing responsibilities between the two
institutions. Immediately after the PCUs were formed, they took over the responsibilities
of rural fertilizer procurement and distribution, and intra-provincial maize marketing from
NAMBoard. Inter-provincial maize marketing still remained the responsibility of
NAMBoard. However, in 1985, all marketing responsibilities were returned to NAMBoard
due to increased inefficiency and ﬁnancial mismanagement by PCUs. Instead, PCUs only
acted as agents of NAMBoard in the crop and input marketing process and the former
received a commission for rendering their marketing services.

In 1987, marketing arrangements changed again. The NAMBoard was given the sole

responsibility to buy maize only along the line of rail, to coordinate inter-provincial maize

 

 

 

Ni

7.

nc

54

marketing, to carry out maize international trade, and to continue with the distribution of
the agicultural inputs. The PCUs concentrated on intra-provincial crop marketing in the
rural areas. This arrangement did not last long either. In 1988, NAMBoard was directed
to buy maize only for national reserves at designated places where there were silos for
storage. That is, it was again the responsibility of the PCUs to do all intra-provincial maize
and other crop marketing

As inefficiencies continued even under the different marketing arrangements, the
government dissolved NAMBoard in June, 1989. The food-gain marketing function of
NAMBoard at that time was transferred to the Zambia Cooperative Federation (ZCF).
The PCUs continued with their intra-provincial maize marketing, acting as buyers of the last
resort for the other decontrolled crops, and seed and fertilizer distribution. ZCF is
responsible for coordinating inter-provincial maize marketing and the importing and
exporting of maize.

In 1991, the government h'beralized maize marketing to allow participation by private
traders. Unfortunately, private traders have not started marketing maize yet because they
have found it to be unproﬁtable since the government still provides a ceiling on how much

private traders can charge when selling maize by continuing to set the maize mealie meal

retail price.
33 WW

Out of the total maize produced, only about half is marketed through the ofﬁcial
marketing organizations. Small-scale maize producers based in rural areas where there are
no large-scale maize processing plants and where maize meal processed in urban areas is

not readily available, retain some maize for own consumption. They then sell the remaining

55

maize output to official marketing organizations. Therefore, the marketed output depends
on how much maize the rural people decide to retain, which in turn depends on the
availability of other food crops produced and consumed in the rural areas such as sorghum,
millet and cassava. The amount of maize retained has been increasing over time. The
inefﬁciencies in the distribution system for maize meal, the decreasing production of some
traditional food crops and the increase in rural population over time are some of the
explanatory factors for the increase in the maize retention rate for food security reasons at
household level. Maize bought by official marketing organizations is resold to milling
companies in the urban areas where it is processed into mealie meal for domestic

consumption and for other uses.

34 IranmcrtatrloachMaizeMarketeinuput
Zambia has a fairly good trunk road system although it is in need of repairs in many

places. However, the rural road system, which is the responsibility of District Councils, is
still poor. Repair work is hampered by lack of earth moving equipment and lack of cash.
No system has yet been develOped to tax farmers to generate revenue for the construction
and maintenance of rural roads. This is a big problem in crap marketing since private
truckers shun areas where roads are bad.

Maize is transported from farms to PCUs depots on ox-carts, tractor trailers or hired
trucks by individual farmers. From depots to storage facilities, maize is transported by
transporters on contract with the cooperative unions. The introduction of uniform producer
prices (1972) for the whole country and throughout the marketing season, and the
determination of the transport rates for maize marketing have made it difficult for PCUs

to contract enough transporters. These policies have made maize haulage by private

56

transporters unproﬁtable. Moreover, most of the trucks available in Zambia are 30-ton
trucks owned by private traders and engaged in long distance maize haulage in areas with
good roads. What PCUs need for maize haulage in the interior parts of rural areas where
roads are bad are lightweight trucks that can cover short distances. The government and
individual PCUs have only a few of these trucks. However, in January, 1990, ZCF bought
110 lightweight trucks speciﬁcally for maize haulage.

The shortage of spare parts for the existing ﬂeet of vehicles is yet another problem
which has resulted from the foreign exchange problem tied to the overall economic situation.
There is also heavy reliance on road use in maize haulage. The short supply of locomotives
and wagons has given way to a minimum use of the railway system. This transport problem
delays getting stocks ﬁrst from rural buying depots to safe storage, and then from safe

storage to consumption areas. It also results in some maize loss since maize gets soaked

by rain.

35 W

The maize that is retained is stored by individual households in their own small
ganaries. The marketed output is stored by marketing organizations. For marketed output,
there is currently a national gain storage capacity of approximately 12.5 million 90-kg bags
as shown in Table 3.1. Out of this capacity, only about 5 million bags worth of storage is
of reasonable quality (sheds and silos). The rest is in the form of ”hardstanding" which is
supposed to consist of a concrete platform on which maize is put and then covered with

tarpaulins.

57
Table 3.1

Zambia: 1988 Grain Storage Capacity by Type and Provincial Cooperative Union

 

 

 

 

 

Minions of 90.101 Bags
Existing Additional Requirement
Coop. Silos Sheds Hard Total Silos Sheds Hard Total
Union Standings Standings
‘CPCU 0.5 0.444 1.8 2.744 0.2 0.55 - 0.75—
CCU 0.41 0.331 1.179 1.919 - 0.05 - 0.05
ECU - 0.71 0.413 1.123 0.2 0.75 0.55 1.5
LCU - 0.057 0.174 0.232 - 0.1 - 0.1
LPCU 0.16 0.555 0.824 1.539 - 0.05 - 0.05
NCU - 0.278 0.398 0.675 0.2 0.35 - 0.55
NWCU - 0.106 0.072 0.178 - 0.04 - 0.04
SPCMU 0.16 0.888 2.641 3.689 0.2 0.4 - 0.6
WPCMU - 0.255 0.236 0.491 - 0.1 - 0.1
Total 1.23 3.624 7.735 12.589 0.8 2.39 0.55 3.74
Source: Republic of Zambia, Ministry of Cooperatives, Department of

Marketing, Logistics and Marketing Information Center, 1990.

The ”hardstanding' storage ‘does not lend itself to effective fumigation and if
tarpaulins are of poor quality there is further damage to crops. Despite the disadvantages
of 'hardstanding' storage, it is used in many places due to transport problems and lack of
money. In fact, some of it has a base built of wooden poles, instead of concrete, which
easily rot. Yet in some places, maize stocks have been kept in these wooden-base
"hardsta'nding' storage facilities for even two years or more although it is only good enough
for a maximum of three months. As of 1988, there was a short-fall of about 4 million 90-kg
bags of gain storage capacity.

This storage problem has led to geat losses resulting from rain damage, rodents,

other storage pests, and improper use of gairn bags. Furthermore, a shortage of storage

facilities has partly contributed to giving up the gading of maize by marketing organizations

58
upon buying it since gading entails storing according to gades, thus calling for extra storage
facilities. However, the government set aside some K750 million of which K400 million is
foreign exchange, for the construction of 75 storage sheds in 1990 (Ministry of Agiculture,

1990).

3.6 11112021003510;
Most of the food crops are processed before reachirng the ﬁnal consumer. Bagachwa

(1991) discusses the various types of gain (including maize) processing in Tanzania, which
are almost the same as those in Zambia. The two types of maize mealie meal processed
by commercial milling plants in urban areas in Zambia are breakfast and roller mealie meal.
Breakfast is the ﬁnest of the two since all the maize husks are removed before milling. The
Republic of Zambia (1990) study indicates that from 100 kg of maize, roughly 63 kg of
breakfast meal or 92 kg of roller meal can be obtained. The rrnilling of maize is currently
carried out mostly by parastatal millers-following the nationalization of almost all the major
milling companies after the 1986 mealie meal riots on the Copperbelt.

Most of these milling parastatals are concentrated along the old line of rail provinces
especially Lusaka and the Copperbelt. About 71.5% of total urban population is
concentrated in these areas. These provinces are also the major maize deﬁcit provinces and
most of the maize consumed is brought in from the surplus provinces.

The milling plants have a number of problems. Most of them were installed a long
time ago and therefore require rehabilitation. This is compounded by lack of foreig1
exchange to bring in spare parts, which is also partly the reason why the milling plants do
not operate at full capacity. For instance, during the period April to June 1987, the milling

companies operated at about 57% capacity (FNDP). The capacity under-utilization of maize

59
millirng plants is also attributed to insufﬁcient supply of maize to milling plants at times.

This is aggavated by inadequate storage facilities at milling plants. A number of major
mills have storage facilities for only 2 to 3 weeks supply of maize. The rigid pricing. system
does not give millers incentives to invest in storage facilities.

In rural areas, most households rely on laborious, time consuming and often back-
breaking staple food processing facilities available at household level. Usually, hard maize
varieties are wet-processed. This method entails removing husks on maize gains usually
by pounding and soaking the gains for a varying period of 2 to 5 days. Soaking allows for
gains to be softened before pounding and sieving. The major disadvantage of this method
of processing maize is the signiﬁcant loss of nutrients through leaching when the gains are
soaked. However, fermentation leaves some ﬂavors which contribute to characteristic taste
and palatability of the processed product (Siandwazi, 1990).

Mechanized processing of maize in rural areas, usually hybrid maize whose gains
are soft, is done using hammermills and handmills. The availability of hammermills
increases maize retention by farmers and reduces maize mealie meal importation from
urban areas. Maize retention has also increased with the rise in maize meal prices over
time and the erratic maize meal distribution system.

The hammermills are also not without problems. As of 1989, about 30% of all
hammermills in the country were non-operational due to lack of spare parts and skilled
manpower at the local level to maintain them (Ministry of Cooperatives, Department of
Marketing, 1990). Moreover, mechanized maize processing facilities (hammermills) are not
well distributed and women are still subjected to walking long distances to a nearby
hammermill. The levy that women have to pay for processing maize through these

mechanized processing technologies is not uniform from one region to another or witlnin the

60

same area. However, the total number of hammermills that are reported to be in the
country may be underestimated in that usually only those that are distributed through the

PCUs are considered, not those owned by private individuals.

3.7 14313233121033

The principal retailers of maize meal in urban areas are state-owned retail outlets.
There is a government rule that nnillers should deliver 80% of the total maize meal to state-
owned shops which are based in Lusaka and Copperbelt provinces. Private traders just
retail the remaining 20% and they have to obtain a special permit to be able to do that.
The government’s intention of using these measures has been to guard against malpractice
like smugglirng and exploitative pricing.

Table 3.2 shows the number of stores that were retailing maize mealie meal in
Zambia in 1990. It shows that the number of private traders retailing maize meal far
exceeded the number of state-owned shOps, despite the former’s small share in the total
amount of maize meal retailed.

In rural areas, because households retain maize for own consumption, they rarely
depend on the market for maize mealie meal supplies. If some households run out of their
retained maize stocks and traditional food crops, which is mairnly around
November / December, that is when they buy maize meal from retail shops, if it is available.
This seasonality of the trade in maize mealie meal in the rural areas and the diversity in the
allocation of licenses for maize retailing to private traders at district level make it diﬂ'icult
to keep track of this trade. Usually, distrrbutors shun the distribution of maize mealie meal

to rural areas because it is too expensive and unproﬁtable given the uniform

61
Table 3.2

Zambia: Number of Stores Retailing Maize Mealie Meal by Province in 1990

 

 

219210131 Whoa: 1211231280003
Lusaka 18 63
Copperbelt 18 103

Southern - 42

Western - 16

Eastern - 35

Luapula - 52

Northern - 26

Central - 23

Total 36 360

Source: Republic of Zambia. WWII/[aize

5911521301. 1990. p- 151-

national government ﬁxed and controlled prices and margins. Therefore, maize mealie meal
sales in rural areas are neglig'ble and they are part of the 20% stocks of maize meal retailed

by private traders.

3.8 W

The staple food crops in Zambia are maize, rice, wheat, sorglnum, millet, and
cassava. While the ﬁrst three crops are gown both as cash crops as well as food crops, the
other crops are gown only for subsistence purposes. Out of these, maize contributes about

95% to total staple food consumption in the country.

62
Figure 3.4 Zambia: Relationship Between Per Capita Maize Consumption and Per Capita
Maize Production, 1964-1990

 

 

 

 

0.35 0.2
A ’1':
3 —o.19 3,
z 0.3- z
.9 -o.1s S
8 3
80'25_ '0'I7 In
E 5
g -o.1s 3
§ 0.2- 'i'
< ”0.15 I
t <
% -o.14 E
oo.ns~ 6
E -0.13 a:
it

0.1 12

1305' ' ' i976 ﬁia'n' ' ' {3'30 ' ' 1935' ' ' 1990‘“
YEARS

 

[—4-— PRODUCTION —..— CONSUMPTION

 

Figure 3.4 shows the relationship between per capita maize consumption and per
capita maize production. It depicts that generally per capita maize production was higher
than per capita maize consumption in the 1960s and 1970s such that maize exports were
geater than maize imports. Moreover, during that same period, the production of
traditional food crops was still high relative to that of maize. This implies that more of the
traditional food crops were being consumed in rural areas thereby reducing the overall per
capita maize consumption relative to per capita maize production. However, with the
decrease in per capita maize production starting in the late 19703, per capita maize
consumption has since been higher than per capita maize production, generally, resulting

in substantial maize imports in the 1980s.

 

The relationship between per capita maize consumption and per capita real GDP
is given in Figure 3.5. Awarding to economic theory, consumption of a normal commodity
is positively related to income. However, Figure 3.5 shows that while per capita real GDP
has been decreasing over time, per capita maize consumption has been increasing. A
possible conclusion from this result, based on economic literature, would be that maize is
an inferior commodity in Zambia. Rubey (1992) cited two studies done by Cadiz (1984)
and Van Zyl (1986), both using country-wide data in South Africa, which found white maize
meal had a negative income elasticity of demand -0.38 and -0.30, respectively, and therefore
was an ”inferior" good. However, Rubey (1992) also reported the findings of the study done
by Elliott (1991) using consumption data that only included black consumers in South Africa
which has called the two earlier results into question. He reports that Elliott (1991) is
reported to have found that white maize meal was a I‘normal" good since it had an income
elasticity of demand of 0.06 for the urban population and 0.20 for the rural population.
Since about 94% of total maize consumption is accounted for by the black population in
South Africa (Rubey, 1992), it is possible that the negative income elasticities of demand
found by Cadiz (1984) and Van Zyl (1986) were inﬂuenced by the income differences
between whites and blacks in that country. The difference could also be due to the data sets
used in their analyses.

Another study done by Jayne and Chisvo (1991) in Zimbabwe showed a positive
relationship between the maize demanded by millers for processing it into commercial meal
from the Grain Marketing Board (GMB). Therefore, maize is not an ”inferior" good in
Zimbabwe. The study reports that the demand for maize by millers from the GMB is

essentially a derived demand for maize meal by consumers.

64

Figure 3.5 Zambia: Relationship Between Per Capita Maize Consumption and Per Capita
Real GDP, 1964-1990

 

 

 

 

0.2 0.024
A A
" §
50.19- ~o.022 o
(
z 3
80.18- -o.02 x
§ 0
O
0.17- ~o.018 3
5
‘émr fame §
I _l
£0.15- ~o.014 Q
‘ <
50.”.- -o.012 g
0
gms- ~o.01 E
0.12 I I Iffl I [j I I I I I I I I I I I I I I I I T I 0.008
1955 1970 1975 1980 1985 1990

YEARS

 

+MA|ZECONSUWTION +REALGDP

 

 

 

Based on the ﬁndings of these studies from Zimbabwe and South Africa where
maize is also the main staple food crop for the majority of the population, it is possible that
maize may not be an inferior good in Zambia. Although the negative relationship between
per capita real GDP and per capita maize consumption was found in this study, with a
relatively large urbanized population relative to the other Sub-Saharan countries, Zambia’s
per capita GDP may not be representative of the incomes of people across the country.
Therefore, further investigations are required to determine the income levels of different
categories of people both between and within rural and urban areas before a conclusion can

be made whether maize is an inferior good or not in Zambia.

65

Figure 3.6 Zambia: Comparison of Real Retail Prices of Maize, Rice and Bread, 19644990

 

 

 

 

 

I I l U I I I U T j

isisTTT ishd’ 3935 1990
YEARS

39%5 3936

 

+ MAIZE '+- RICE + BREAD

 

 

 

The relationship between per capita maize consumption and per capita real GDP
may be inﬂuenced by the pricing system in the Zambian situation. The other major staple
food stuffs are wheat (bread) and rice. The prices for these commodities are not uniform
for the whole country. They are more variable from time to time within a given year and
they are decontrolled. However, the price of maize meal is uniform throughout the country.
It is ﬁxed and controlled by the government and it does not change as often as the changes
in the prices of rice and bread.

Therefore, maize meal has always been cheaper than the other staple food crops, as shown
in Figure 3.6. As such, as the economic situation worsens from time to time, as rural

people’s food consumption patterns change from consumption of traditional food crops to

66

Figure 3.7 Zambia: Relationship Between Per Capita Maize Consumption and Maize Mealie
Meal Retail Price, 1964-1990

 

 

 

 

0.2 9
50.19% '8 E
éora— 9
g ' ’7 V

0.17- E
g '6 _,
00.16- <
Ems- .5 E
“ ' E
O

<
50.“.- i‘ 2
IL _I
3013- -3 g
I!
0.12 I I I I I I I I I I I I I I I I I I I I I I I I I I I 2
1955 1970 1975 1930 1935 1990
YEARS

 

|+ OONSLMTION + PRICE

 

maize consumption (Republic of Zambia, 1990), and as rice and bread become more
unaffordable by the urban population, most people have generally tended to reduce the
consumption of other staple food stuffs like rice, bread and sorghum and increase maize

consumption.

3.8.2 Maize gnsumptign and Real Mag e Retaﬂ Brice

Economic theory stipulates that consumption of a commodity is negatively related
to its own price. The relationship between per capita domestic maize consumption and real
maize retail price is given in Figure 3.7 above, which shows that there have been more

ﬂuctuations in the real maize retail price than in per capita domestic maize consumption.

67

This indicates that although there is some negative relationship between these two variables,
their correlation is not very strong. Thus, one would expect the real maize meal retail price
not to be a signiﬁcant determinant of domestic maize consumption. One of the explanations
for this could be the lack of readin aﬂordable substitutes for maize meal as explained in the

previous section such that most people have no choice but to consume maize meal.

 

The relationship between per capita domestic maize consumption and per capita
sorghum production is given in Figure 3.8, which depicts a negative relationship between the
two variables. While per capita sorghum production has been decreasing, per capita
domestic maize consumption has been increasing since people have been substituting
sorghum for maize.

This decrease in the production and hence in the consumption of sorghum in rural
areas which implies an increase in maize consumption can partly be attributed to changes
in the food consumption pattern (Republic of Zambia, 1990). The same study reported that
nutritional surveys for the country suggest that there has been a gradual change taking place
in the pattern of food consumption in rural areas with maize being more widely consumed
throughout the country. The study also mentioned that the decontrol of prices of all foods
except for maize mealie meal in the 19808 has resulted in increased dependence on maize
in urban areas where cassava, millet and sorghum are not consumed. Hence, in both urban
and rural areas, there has been an increase in maize consumption which is reﬂected in
Figure 3.8. Therefore, it would be interesting to disaggregate rural and urban maize

consumption to see the changes in the pattern, although this is beyond the scope of this

study.

68

Figure 3.8 Relationship between Per Capita Maize Consumption and Per Capita Sorghum
Production, 1964-1990

 

 

 

 

 

 

0.2 0.02 A

13 5
50.19- >o.o1a v
.50 18- -o.015 g
g -0... §

0 17- g
5 ~o.o12 ‘1
0 o 15 ~ 3
E -o.o1 E

o 15-
: -o.ooa m
E o 14- E
( -o.oos E
0
5 0-13‘ -o.004 :
‘L it!

0.12 I T r I I I I I I I I I I I I I I r T 1 I I ‘I—Iﬁﬁ W 0.002

1965 1970 1975 1980 1985 1990
YEARS
|+ MAIZE consuwnon —~— SORGHUM PRODUCTION J
3.9 W
391m

The decreasing world copper prices since the mid-1970s made the government
realize the need for promoting agricultural production. However, in promoting agricultural
production, particularly that of staple food crops, the government has biased its policies
towards maize production and marketing. Some of this bias in policies against traditional
food crops and in favor of maize is reﬂected in Figure 3.9, which shows the trend in real
maize and sorghum producer prices. It shows that the sorghum real producer price was

higher than that of maize only between 1964 and 1975 before the government started

69
Figure 3.9 Zambia: Comparison of Maize and Sorghum Real Producer Prices, 1964-1990

 

,. 5*
0| 0| 0'
l l

‘
1

“I

5"
0|
1

 

01
/

REAL PRODUCER PRiCES (K/MT)
N
in

N
i

 

 

 

—a
(ll

I I j j IIIIIIIIIII

i930 1975 1980 i935 ‘ ' i9§6
YEARS

iahé

 

+MA|ZE +SORGWM

 

 

 

promoting maize production. Since 1975, the government has been ﬁxing the maize
producer price at almost the same level as that of sorghum, in addition to rendering a
readily available market for maize but not for sorghum. This increase in maize price at the
expense of that of sorghum partly explains the decrease in sorghum production over time.
‘One of the reasons for maize production and marketing promotion by the government is
that in government circles, food self-sufﬁciency is considered to be equivalent to increased
maize production.

Table 3.3 gives the annual percentage change in the producer price from the
previous year for major crops grown in Zambia. It is observed that out of the staple food
crops (maize, wheat, rice, sorghum, millet and cassava), millet and sorghum almost never

had any price revisions from 1976 to 1982. The wheat and rice producer prices also did not

70

Table 3.3

Zalbia: Percentage Changes in Producer Prices from the Previous Year's Price for Selected Crops

 

YEAR
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990

CSVA
27
60
47
28

-16
-14
22
0
33
50
100
17
43
40
129

HLET
0
0
O
0
0
0
O

383
2
29
48
64
74
24
120

SBEAN UHEAT
29 O
0 0
26 25
16 0
28 0
13 30
16 23
7 12
16 19
16 6
84 91
32 28
47 71
29 19
106 116

RICE

2O
0

0
ll
13
3
Si
43
0

0
39
49
34
233
58

SFLVER
19
0
35
1
20
7
18
4
0
30
50
67
29
80
85

GNUT
47
0
14
12
9
32
4
15

. 30
28
43
23
79
16
103

CTON
33
O
15

11
12
16
25
90
88
20
169

SGHU!
0

000°

50

78
17
44
59
73

36
161

18660
15
13
32

4
4
5
45
13
4
23
48
22
124

317

MAIZE

26
O
8
32
30
15
19
14
34
16
94
42
3
56
127

 

Source: Calculated from data obtained from

Republic of Zambia. Ministry of Agriculture and Hater Development; Ministry of
Cooperatives. and Central Statistics Office

FAO "Agrostat" data on the diskette

71

change as much as the maize and cassava prices did. However, while the government used
to set ﬂoor producer prices for most of the staple food crops, it just started setting those of
millet and cassava in 1979 and 1986, respectively. This, plus the lack of assured market
for the two crops as with sorghum did not encourage cassava and millet production.
Therefore, out of the staple food crops, even the price incentive for the early part of the

period was only given to maize.

3.92 MmMﬂkcﬁnaMstﬂnAnﬂxsis

A marketing margin (or marketing spread) is the difference between the farm price
and the consumer price of a commodity. Timmer 9131. (1983) argue that large marketing
margins can occur for two reasons: either high real marketing costs cause consumer prices
to be much higher than farm prices or monopolistic elements in the marketing system earn
excess proﬁts. They also give two approaches (direct and indirect) that can be used to
determine whether there are excess profits and serious inefﬁciency in food crop marketing
or whether wide margins are due to high real costs. The direct approach analyses marketing
costs by looking at marketing functions (transportation, storage, processing, etc) whose
combined costs constitute the marketing margin.

The indirect approach entails price analysis to determine market efficiency. The
latter approach involves statistical comparisons of pairs of price series that should be
connected by marketing. However, due to lack of costs and returns to marketing agents on
the various marketing functions, and due to lack of prices at different levels of the
marketing system, only the farm and retail prices of maize will be used to analyze the maize

marketing margin in this study.

72

Figure 3.10 Zambia: Comparison of Real Maize Producer and Retail prices, 1964-1990

 

 

o H
1965' ' ' '19'70' ' 1 '19'75' ' ' '19'80' ' ' '19335' ' '
YEARS

 

 

 

REAL PRODUCER PRICES AND SPREAD (K/MT)
p

'19'90'

 

+ PRODUCER PRICE + RETAIL PRICE + MARKETNG SPREAD l

 

The real maize producer price, the real maize mealie meal retail price, and the
marketing spread are given in Figure 3.10. A high marketing spread implies high marketing
(transportation, storage, processing, etc.) costs. Zambia has a wide marketing spread for
maize because of high maize transportation and handling costs. Figure 3.10 shows that in
the late 1970s, the real maize marketing margin (marketing'spread) was even higher than
the real maize producer price.

It can also be inferred from the same Figure that generally the higher the real maize
producer price, the higher the real maize mealie meal retail price. However, the ratio of
the two prices varies from time to time with the highest ratios reﬂected in the 1980s. This
indicates that for the most part of the 1980s, the marketing spread was relatively small, such

that it was even negative in 1987 and 1988.

73
The relatively lower values of the marketing spread for the most part of the 1980s

are not necessarily a reﬂection of improved efficiency in maize marketing in the Zambian
situation. This is because the government was spending a lot of money subsidizing maize
marketing such that actual maize transportation and handling costs were not reﬂected in the
retail price of maize. This is depicted in Figure 3.11 in the next section which shows how
the maize marketing spread has been moving together with the maize subsidies. The higher
the maize subsidy the lower the maize marketing spread. The maize marketing spread
increased tremendously in 1989 and 1990 because the government embarked on a maize
subsidy-reduction program. The details of this will be discussed in the next section.
Table 3.4 shows the ratio of the maize producer price to the maize meal retail price.
The higher the ratio, the smaller the maize marketing spread. It also confirms that the
smallest maize marketing margin was in the 19808 when the maize subsidies reached the
highest level. However, from 1989 onwards, the ratio started to decline tremendously

because of the gradual removal of maize subsidies since that year.

3.10 MSW
Maize-related subsidies have been of various forms, the main ones being:
(a) a marketing subsidy on commercial fertilizer;
(b) a price differential subsidy on imported fertilizers;
(c) a marketing subsidy to Provincial Cooperative Unions (PCUs) to cover
transportation and handling costs incurred in intra-provincial maize marketing;
(d) a marketing subsidy on inter-provincial maize marketing; and

(e) provision of maize meal coupons to selected groups of urban consumers.

74

Table 3.4

ZambhzkaﬁodNomMMaberderbewtthMuﬁeMuiRenﬂPﬁOeJW-lm

 

m l: I E“ B 'l 2' E' B .
1964 41.44 67.60 0.61
1965 4133 63.20 0.65
1966 36.39 63.20 0.53
1967 _ 34.44 62.40 055
1963 3222 62.40 052
1969 3656 6330 052
1970 3339 7920 0.49
1971 4444 36 40 051
1972 47.73 3640 055
1973 47.73 3640 055
1974 47.73 36.40 055
1975 , 55.56 11300 0.47
1976 70.00 155.00 0.45
1971 70.00 154.00 0.45
1973 . 7556 192.00 039
1979 100.00 130.00 0.56
1930 130.00 130.00 0.72
1931 150.00 26520 057
1932 177.73 294.40 0.60
1933 20333 40720 050
1934 272.22 459.00 059
1935 314.67 436.00 0.65
1936 611.11 76600 030

l 1937 366.67 766.00 1.13
1933 33339 766.00 . 1.16
1939 1333.39 3206.60 0.43
1990 3157.73 10760.00 029
1991 555556 21520.00 026

 

Source: Republic of Zambia, Ministry of Agriculture and Water Development and Ministry of Cooperatives, Prices
and Incomes Commission and Central Statistics Ofﬁce.

75

Maize marketing subsidies began increasing at a faster rate after the introduction
of the uniform prices for maize in 1972. Table 3.5 shows the maize and fertilizer subsidies
in nominal terms and as a percentage of the total government budget for the period 1980
to 1990. It shows that the share of maize subsidies in the total government budget rose
from 9.3% in 1980 to 16.9% of the total government budget in 1988 in nominal terms. The
fertilizer subsidy was 3.6% of the budget in 1980, 5.7% in 1982, and 3.9% in 1990.

Figure 3.11 gives the relationship between real maize subsidies and the real maize
marketing margin. These huge subsidies resulted partly from the inefficiencies of the
marketing organizations and partly from the pricing policy. The pricing policy did not allow
marketing agencies to have any proﬁt margins. For instance, from 1985 to 1988, the
nominal maize into-mill price of a 90-kg bag of maize remained at K35 while the producer
price of the same bag was changing from year to year. In some years, both the producer
price and the into-mill price of maize were set at the same level. Such into-mill prices never
reﬂected the transportation and handling costs of maize. Worse-still, the maize producer
price was even higher than the maize into-mill price in 1986, 1987 and 1988. The
transportation and handling costs of maize and the other parts of the price differential were
paid for by the government as part of maize subsidies. In 1987 and 1988, the marketing
margin was even negative.

All these huge maize marketing subsidies, particularly in the 19803 were meant to
ensure provision of cheap, affordable maize mealie meal to the vocal urban population as
the economic situation was increasingly worsening. These maiz'e subsidies have been utilized
mostly by the urban population at the expense of the rural population which rarely consumes

maize meal processed in urban areas to which maize marketing subsidies apply. In turn, this

76

Table 3.5

Government Budget

Zambia: Nominal Maize and Fertilizer Subsidies as 3 Percentage of the

 

 

 

 

 

 

Million Kwacha
Maize Subsidies Fertilizer Subsidies
m: Govt. Govt. Alum % of Amount % of
m Budget Govt Govt
Deﬁcit Mm 39.9891
1980 1657.6 568 154.00 9.3 60.0 3.6
1981 1388.6 450 87.10 6.3 48.5 3.5
1982 1643.2 668 138.00 8.4 84.6 5.7
1983 1475.9 327 124.70 8.4 19.9 1.3
1984 1484.6 414 81.60 5.5 20.1 1.4
1985 2184.3 1073 134.00 6.1 n/a n/a
1986 5383.6 1026 565.00 10.5 n/a n/a
1987 5837.5 2147 638.40 10.9 205.9 3.5
1988 8359.3 1531 1413.00 16.9 205.9 2.5
1989 9838.0 3699 1585.60 16.1 357.0 3.6
1990 24503.3 2801 2193.73 13.7 950.0 3.9
Source: s

Republic of Zambia, v '
W Tables 11.2 and 115, 1990, Lusaka.

77

Figure 3.11 Zambia: Comparison of Real Maize Marketing Margin and Real Maize
Subsidies, 1964-1990

 

3‘
:5 U!
1

3.5 -

9
DUI

 

 

 

 

REAL MARKETNG SPREAD 3: SLBSDY (K/MT)
- N
.2 u: N Ln 0:

'19'35' ' ' '19'70' ' ' '19'75' ' ' 1930' ' '19'35' ' ' 1930'
s

 

L+ swsuozs —o— MARKETNG SPREAD

 

kind of marketing arrangement and pricing policy has contributed and continues to
contribute to the poor performance of marketing agencies.

However, according to the current national development plan (FNDP), one of the
objectives of the government is to get rid of all maize and fertilizer subsidies by 1993. To
begin the process of removing maize marketing subsidies, the government raised the into-
mill price of maize in 1989 so that the system could pay for part of the intra-provincial
transport and handling costs. In the 1988 / 89 marketing season, farmers sold maize at K125
per 90-kg bag and millers bought the same bag at K160, allowing a marketing margin of
K3500.

78
Table 3.6 shows a 1988 / 89 marketing season’s break-down of the intra-provincial

maize marketing subsidies paid by the government for maize marketed by each PCU. Out
of the total transportation and handling cost incurred by each cooperative union per bag,
the marketing system paid K35 per bag while the government paid the remaining amount.
That is, the government was able to reduce its intra-provincial maize marketing subsidies
by K35 per 90-kg bag of maize. The intra-provincial maize marketing subsidies amounted
to about K510 million in 1989. This reﬂected the decrease in maize subsidies as a
percentage of the total government budget in 1989, as shown in Table 3.5 above.

A closer scrutiny of the composition of the handling and transportation cost per 90-
kg bag of maize reveals that transport costs accounted for about 74% of the total cost (see
Table 3.7). Since transportation cost is the major component of total maize marketing costs,
any strategies aimed at reducing marketing costs in the medium-term should prepose ways
and means of reducing such costs. One way to do this is to reduce the movement of grain
from producing areas to either centrally located safe storage or consumption areas by
locating storage and milling plants in maize producing areas. This could substantially reduce
the burden on government treasury as far as maize handling and transportation cost
subsidies are concerned.

Another move in the direction of removing maize marketing subsidies came in the
1990 / 91 marketing season when the government got rid of all intra-provincial maize
marketing subsidies. The minimum transportation and handling cost per 90-kg bag of maize
during the same season was K158 since the buying price paid to farmers was K284 while the

selling price to millers was K442 per bag.

79

Table 3.6

Zambia: 1988 / 89 Season Intra-Provincial Maize Handling and Transportation Costs
and Subsidies by Provincial Cooperative Union

 

 

 

(Kwacha per 90-Kg Bag)
Coop. H & T Producer Into-Mill m Quantity Total
Ilniszn Quits Brice Brice Mmlmed Subsidy Paid
(90-ks bass»)

CPCMU 68 125 160 33 4494680 148324440

CCU 68 125 160 33 483947 15970251

ECU 68 125 160 33 2470675 81532275

LCU 97 125 160 62 386601 23969262

LPCU 68 125 160 33 465596 15364668

NCU 85 125 160 50 1451029 72551450

NWCU 117 125 160 82 159561 13084002

SPCMU 68 125 160 33 3352818 110642994

WPCMU 140 125 160 105 285255 29951775
13550162 509951376

Total

Source: Republic of Zambia, Ministry of Cooperatives, Department of

Marketing, Logistics and Information Center, Lusaka, 1990.

80

Table 3.7

Zambia: 1988 / 89 Percentage Composition of Total Handling and Transportation Cost

 

 

of a 90-kg Bag of Maize
m H ii' 11: . E
Transport 74.0%
Salaries 7.0%
Bank Interest 9.0%
Bank Charges 0.7%
Weight Loss 4.0%
Proﬁt Margin _ 5.0%
F umigation 0.3%
Total 100.0%
Source: Republic of Zambia, Ministry of Cooperatives, Department of

Marketing, Logistics and Information Center, Lusaka, 1990.

The pinch of the removal of these subsidies has really been felt by the population
through dramatic increases in maize mealie meal prices. The price of a 25-kg bag of
breakfast maize meal has increased from K2000 in December 1988 to K114.00 in 1989 to
K269.00 in June 1990 and to about K990 in December 1991.

The other aspect which merits attention is the inter-provincial transfer of maize and
other crops from surplus regions to deficit regions. The government has continued to
subsidize inter-provincial maize transportation costs. The subsidy is based on the formula:
distance x transport rate x total quantity moved. The transport rates, ﬁxed by the
government for crop marketing, are given in Table 3.8 for selected years.

These transport rates are ﬁxed in consultation with transporters and are the same

throughout the country. As a result, some rural areas are shunned by transporters because

81

Table 3.8

Zambia: Government Gazetted Transport Rates for Maize Haulage, 1984-1990

 

 

(Kwacha per ton per Kilometer)
W .1282 1285 1222 1231 12.82 1222 1229
1 - 50 km 0.25 0.70 0.80 0.91 1.45 4.00 7.20
51 - 100 km 0.20 0.65 0.70 0.80 1.20 3.30 5.94
101 - 200 km 0.15 0.41 0.55 0.63 1.00 2.80 5.04
200 - Above 0.10 0.30 0.45 0.51 0.80 2.60 4.85
Source: Republic of Zambia, Ministry of Cooperatives, Department of

Marketing, Logistics and Information Center, Lusaka, 1990.

University of Zambia, Institute of African Studies, Draft Report of

"Rice Production, Processing and Distribution in Zambia: Analysis of

Problems and Prospects”, Lusaka, 1990.
they have bad roads. Hence, some crops remain uncollected and get soaked by the rains.
Given the structure of these rates, the inter-provincial maize transportation subsidies have
been very enormous too. For instance, during the 1988/ 89 season, it is estimated that about
6.5 million 90-kg bags of maize were moved from the surplus provinces of Southern,
Eastern, Central and Northern to the major deﬁcit provinces of Lusaka and Copperbelt.
Some maize was also moved to Luapula, North-Western and Western provinces which are
also deﬁcit provinces. During the same season, the cost of a ton of maize transported from
Kasama to Ndola, over a distance of 759 km, was about K1,973.40. Thus the transport cost
was about K177.62 per 90okg bag of maize, which was higher than the producer price of
K125.00 per 90-kg bag for the same season.

To reduce the inter-provincial transportation costs, it is prudent to encourage

production near consumption areas. This can be achieved through regionally differentiated

82
pricing intended to regionalize crop production based on economic and agro-climatic

conditions of each region. The guiding principle in crop promotion and production should
be regional comparative advantage, especially for maize.

The other maize-related subsidy was the maize meal consumption subsidy. Any type
of food subsidy is aimed at providing food to economically weaker groups of the population
at a lower price in the absence of a more direct intervention. But what the general subsidy
did instead was to create opportunities whereby a substantial proportion of the beneﬁts
accrued to households that were economically better off.

In the 1980s, the government embarked on measures to modify and improve the
general consumer subsidy. In 1986, the government introduced a form of self-targeting
subsidy by subsidizing “roller mealie meal", which it believed was the type of maize meal
mostly consumed by the poor people. The other type of maize meal, "breakfast mealie
meal”, believed to be consumed by the relatively better-off people, was left to be sold at a
higher price determined by the government. This was a wrong assumption on the part of
the government because it sparked food riots which indicated that actually most people
consume breakfast mealie meal including some in the low income.

In January 1989, the goverﬁent introduced the coupon system to facilitate the
gradual removal of the subsidy from the maize meal price. This coincided with a general
decontrol of other commodities. The coupon system was introduced to subsidize urban
maize mealie meal consumption through the distribution of coupons initially to the whole
urban population, and later to only those urban households that earned less than K20,500.00
per annum (Siandwazi, 1990). The coupon, with a given kwacha value based on the estimate

of maize meal requirement of 14 kg per person per month, could be exchanged for any type

83
of maize meal (Siandwazi, 1988). However, this coupon system was eliminated at the

beginning of 1992 after the coming of the new government in October 1991.

There have also been maize fertilizer-related subsidies. Nitrogen Chemicals of
Zambia (NCZ), the only fertilizer manufacturing organization, meets only about 40% of the
national fertilizer requirements (Ndalamei, 1990). The rest of the fertilizer is imported.
Since imported fertilizer usually costs more than the domestic selling price set by the
government, the government subsidizes the importation of fertilizer too. In fact, the selling
price of fertilizer also does not cover the production costs incurred by the NCZ, implying
that fertilizer production as well as distribution is subsidized.

The fertilizer subsidies are shown in Table 3.5 above. The government made the
ﬁrst move in reducing fertilizer subsidies in 1983 and as a result fertilizer prices increased
by over 50%. In the study conducted by the Zambian government on the performance of
maize subsector in 1990, it was reported that maize uses about 70% of the total fertilizer
sold in the country. So, fertilizer subsidies have been part of the process of providing cheap
maize mealie meal to the vocal urban population since the bulk of fertilizer is used in maize
production.

All these maize-related subsidies have left fertilizer, maize producer prices and maize
mealie meal retail prices lower in Zambia than those in neighboring countries, leading to
smuggling of the commodities, especially into Zaire in the case of maize mealie meal.
Fertilizer smuggling is estimated between 15,000 to 25,000 tons per year (Ndalamei, 1990).
Currently, the government is in the process of removing all maize marketing and fertilizer

subsidies, which are expected to be completely eliminated by 1993.

3.11 Summer
This chapter describes the maize subsector and shows how its performance has been

inﬂuenced by the institution and environmental policies within which it operates. It
discusses maize yields, area harvested and production. With the use of hybrid maize seed
and commercial fertilizer, maize yield has been increasing over time. However, the area
harvested of maize has been ﬂuctuating resulting in ﬂuctuations in maize production. It
shows that the government policy has had an impact on the direction of sorghum production
over time due to its bias toward maize production at the expense of the other staple food
crops.

Maize consumption is also analyzed in relation to its own retail price, real GDP, and
the production of sorghum. Maize retail prices and sorghum production showed their
expected relationship to maize consumption. Per capita real GDP was negatively related
to per capita maize consumption which implies that maize is an inferior good. However,
based on the literature reviewed on South Africa and Zimbabwe where maize also the main
staple food, it is possible that maize could be a normal good in Zambia. Hence this study
could not conclude that maize is an inferior good. It is possible that the per capita GDP
does not reﬂect the incomes which maize consumers have considering that the country is
highly urbanized. The negative relationship between per capita GDP and per capita maize
consumption could also partly be due to the retail pricing policy for staple food crops whose
aim has been to provide cheap maize meal to the vocal urban population for political

reasons.

4. RESEARCH METHODOLOGY

To be able to achieve the forecasting objective outlined in the introductory chapter, an
analytical tool was required. A quantitative model of the maize market was selected to
meet this objective. This chapter first provides the economic theory of supply and demand
for a commodity. It then discusses the general steps involved in commodity modeling and

applies them to the maize subsector in Zambia.

4.1' Wm
4-1-1 We

Traditionally, supply analysis is embedded in the theory of production. According
to production theory, output produced is a function of factors of production available to the
producer. However, supply analysis is based on economic theory which goes beyond
production theory. Apart from the factors of production, it incorporates other various
supply shifters which affect the supply of a commodity. The various factors that affect the
supply of a commodity include own price of a commodity under consideration; weather and
pests; changes in technology that inﬂuence both yields and costs of production; changes in
input (or factor) prices; changes in prices (hence returns) of commodities that compete for
the same resources; changes in the prices of the joint products (that is, commodities that
are produced together such as wool and mutton); changes in the level of price and/or yield
risks faced by the producer; and institutional constraints such as government acreage-control
programs. Any supply function should be formulated based on this economic theory.

The relationship between the commodity under analysis and the factors that affect

it is important for economic analysis. According to economic theory, supply of a commodity

85

86
is negatively related to prices of inputs used in its production, but positively related to its
own price and the prices of commodities with which it has a joint relationship. Other things
remaining the same, a rise in the price of commodity B can be expected to lead to a
decrease in the area planted to commodity A. Supply of a commodity is positively related

to improved technology and weather which lead to improvement in yield.

4.12 Dcmansuzfammcdm

Asuming-Brempong (1991) cites three common aspects of consumption behavior
used in economic policy analysis and forecasting models, namely consumer demand analysis,
Engel curve analysis and consumption functions in more aggregated contexts. Engel curves
are employed to estimate income elasticities; aggregate consumption functions relate
microeconornic variables, such as household consumption, to income and socioeconomic
characteristics; while demand analysis provides price and income elasticities that are
important policy parameters. Demand analysis incorporates demand equations which are
formulated according to the purpose for which a commodity is demanded and the factors
affecting demand. Of these three aspects of consumption behavior, this study will focus on
demand analysis.

Demand analysis is traditionally embedded in the theory of consumer behavior. A
commodity can be demanded for various reasons (Ferris, 1991). It can be demanded for
domestic consumption (the basis of consumer demand theory), for export, and for storage
and speculation. According to consumer demand theory, the quantity of a commodity
buyers will be willing to purchase from the market over a given period of time depends on
the price of the commodity itself; the prices of substitutes and complements; the consumers’

incomes and income distribution; the level of general inﬂation; the consumers’ tastes and

87

preferences; living patterns; erééwmodity; and the population size and its
distribution by age, sex, ethnic group and region.

Demand for storage and speculation of a given commodity depends on current price
versus expected prices; storage costs; interest rates; and storage availability. The demand
of a commodity for export depends on the commodity’s own price in the exporting country;
its price in competing exporting countries; transportation costs; shortage possibilities in the
importing country(ies); and various other factors in the importing country(ies) such as price
of the commodity, its production, consumer income, livestock numbers (for feed demand),

and farm programs and trade policies.

42 WW
There are various steps involved in building a model for any given commodity. First,

the model builder must determine the purpose for which the model is being developed and
the information being sought from it. Palm and Smit (1991) argue that a model can be used
for an analytical problem and for a policy problem. On one hand, an analytical problem
involves solving the endogenous variables for the ﬁxed values of the exogenous variables.
That is, it involves forecasting short- and long-term prices, production and consumption.
On the other hand, a policy problem entails giving some endogenous and exogenous
variables a certain target value and the model is used to find the matching values of the
other exogenous and endogenous variables. In the latter case, the analysis tries to answer
"what if' questions about market responses to speciﬁc economic, policy and / or institutional
changes.

The second step in commodity modelling is the selection of the appropriate model

structure to use in the analysis. The selection of the structure of a commodity model should

88

reﬂect the attributes of a commodity market or a particular commodity problem to be
analyzed, in addition to the formal methodology employed which includes model
speciﬁcation, estimation and forecasting. The stochastic structure of prices and quantities
must be speciﬁed to justify the use of a particular estimator. In general, it is important to
know whether or not price or quantity of a commodity whose supply and demand is being
analyzed is predetermined. Thurman (1987) argues that the text-book competitive market
determines price and quantity simultaneously and requires instrumental variables-type
estimators of demand. This is based on the assumption that both the price and quantity of
the commodity are determined by market forces. Thurman (1987) further argues that if one
is dealing with a market structure in which prices or quantities are predetermined, then
regression estimates of equations should be such that the predetermined variable is an
independent variable which should only appear on the right-hand-side of the equation.
Therefore, even the regression estimator to be used should be one which allows for the
predetermined variables in question to be on the right-hand-side.

The third step in commodity modelling is the specification of relationships of a
particular commodity market. Equations of a model should be speciﬁed in such a way that
they are able to capture economic relationships between model variables while maintaining
consistency with the model structure. While this process includes determination of
behavioral relationships based on economic theory, model speciﬁcation goes beyond
economic theory. Economic theory guides a commodity model builder to select variables
that should appear in each equation but it does not tell you which specific prices to include,
whether to lag variables or not and by how much time period a lag should be, what
functional form the equation should take, etc. Answers to all these questions are also part

of model speciﬁcation.

89

The fourth step in commodity model building is the estimation of parameters for the
variables in the behavioral equations of the model. This process involves selecting the
appropriate technique to use.

The ﬁfth step in commodity model building is the validation of the model. The ﬁnal
validity of a model depenck on the conclusions reached from the application of the different
techniques aimed at the analysis of the information supplied by a model once it has been
speciﬁed and estimated. This process involves performing tests and diagnostic checks with
the model, most of which are statistical in nature, to select the model with the most
appropriate speciﬁcation. The acceptance of the model as satisfactory depends on the
intended purpose of the model. Sowey (1973) points out:

By validation, on the other hand, is understood simply determining whether the

model fulﬁlls well the demand made of it. It is not a question of whether the model

embodies strictly causal mechanisms, but rather whether the estimated model, with
all its inherent imperfections, does an adequate job of prediction, both within and

beyond the estimation period. (p.196).

Dhrymes M. (1972) affirm that validation of a given model differs from case to case as
the proposed use of the model under consideration changes. However, they still argue that
although a particular model may be validated for one purpose and not for another, in each
case the process should be such that the model fulﬁlls the stated purpose.

Grasa (1989) argues that a model is selected if it adequately characterizes the data
by being more accurate and more precise than the other models. To be able to adequately
characterize the data, Grasa argues that a model should resist repeated refutation with
success and should provide more information than other models. This information should

be corroborated by new facts which are not themselves used in the speciﬁcation of the

model.

90

Generally, models built for forecasting purposes are subjected to more rigorous tests
than those developed for policy analysis purposes. Some of the most commonly used
statistical tests discussed by Asuming-Brempong (1991) which are also used in this study are:

(a) consistency to a priori relationship - the estimated parameters are evaluated on
the basis of their magnitude and signs to see if they conform to their expected values and
signs as stipulated by economic theory.

(b) Level of signiﬁcance of regression coefﬁcients - this will be tested using the t-test
(or student t-statistic). The rule of thumb is that if the, t-statistic is at least equal to 2, then
the regession coeﬁ‘icients are signiﬁcant at the 5% level of signiﬁcance.

(c) Validity of the model - This is tested using the F-test which shows the
signiﬁcance of the entire regession equation and its appropriateness for the particular
analysis being conducted.

(d) Coefficient of determination (R2) - This measures the degee of the relationship
between the dependent and independent variables, and shows how much of the variation in
the dependent variable is explained by the independent variables. The higher the magnitude
of R2 the better the model. However, the magnitude of R2 increases with the increase in
the number of the independent variables while the degees of freedom decrease. Therefore,
the corrected (or adjusted) R2 is a better statistic for empirical analysis.

(e) Standard error of Regession (SER) - SER measures the extent to which the
estimated (or observed) value approaches the true value of the variable. Since the
deviations of the observed value from the true value must be minimum, the smaller the
SER, the geater the precision and the more reliable the estimates of the parameters of the

regression equation.

91
(f) Durbin Watson (D-W) test - the DW statistic is used to measure the presence

of serial correlation. Its value is expected to be around 2 when there is no auto-correlation
in the model.

The ﬁnal step in commodity model building is the performing of a forecast, a
prediction or a simulation. To do a forecast, one has to ﬁrst forecast the future values of
exogenous variables and then use them to forecast the future values of the endogenous

variables.

4.3 Mﬂehn' g Methgdglggjes

A commodity model can be built for various purposes. It is not possible to have all
of the purposes met with the use of one methodology only. Therefore, a model may
sometimes require the use of different kinds of quantitative methodologies. Labys (1988)
discusses some categories of models which are based on econometric methods (like market

models) as well as linear programming methods.

4.3.1 MW

Time series models are mainly used for forecasting in commodity market analysis.
Harvey (1990) and Kennedy (1985) argue that while time series models attempt to explain
the behavior of a variable in terms of its own past, an econometric model is set up on the
basis of a behavioral relationship suggested by economic theory. Although time series
models will be used to forecast exogenous variables in this study, an econometric model will
be used to specify, estimate and forecast the endogenous variables. One major category of

econometric models is the market models.

432 Music]:
These are models usually based on econometric methods concerned with the
deterrrnination of prices and with the behavior of participants in a competitive commodity

market. The price is determined by the market clearing conditions. The following is the

standard commodity market model:
Q? ' “1’19in T1) v
0i ' «Pt-P we Pi Gt)
Pt s f(St)
St = st-l “3i ’ oil
Where:
0‘: == quantity demanded in time t;
0: -- quantity supplied in time t;
Pt = commodity price in time t;
St = level of stock-holding of the commodity in time t;
P? = price of other commodities (substitutes or complements) that can be
demanded in time t;
Yt = income level in time t;
Tt = consumer tastes and preferences in time t;
Wt = weather in time t;
P: = prices of substitutes that can be supplied in time t; and
Gt = government policy in time t.

In other cases where prices are not determined by competitive adjustment, other modeling

approaches have to be used.

93

4.4““. l 11° lll . Z l.
The description of the maize subsector given in Chapter 3 is essential in choosing

the correct model structure, and especially for the selection of variables to include in the

price, supply and demand equations.

4.4.1 MW

From the discussion on the maize subsector in Zambia, it is clear that the
government sets all the maize subsector-related prices (maize producer price, fertilizer
prices, maize seed prices, maize marketing transportation rates, maize buying and selling
prices by milling companies, and wholesale and retail prices). On this note, one would
consider prices to be exogenous in the model and use Thurman’s argument (1987) since
maize prices would be considered to be predetermined or controlled. However, in setting
the maize producer price, the Zambian government considers the previous year’s maize
stocks and production levels which are endogenous in the maize subsector, along with
production costs and other policies that might be exogenous to the model. Therefore, since
maize prices are partially endogenously determined, they will be formulated as endogenous
variables so that the government policy can be captured.

Although maize prices will be considered endogenous, it does not imply that an
equilibrium maize price exists which is determined by market clearing conditions. This
implies that the market clearing conditions (equating maize supply and demand) cannot be
applied to determine an equilibrium price for maize given the fact that prices are not fully

determined by the interaction of supply and demand.

442 MW
Total maize supply in Zambia includes domestic production and net imports and

stocks at any given time. As argued by Tomek and Robinson (1990), separate equations are
required to explain each component of supply of a commodity. Each of these maize supply
components will therefore have a separate equation. For production, it is simpler to
formulate only one equation in which total production is a dependent variable in analyzing
producers’ supply response to various factors. However, Tomek and Robinson (1990) argue
that it is usually more appropriate to consider yields and units separately since some detail
is lost in the process. Therefore, this study will formulate separate equations for the area
planted to maize and for the average maize yield. Maize production will be derived as a
product of the behavioral equations of yield and area harvested of maize.

The area planted to maize will initially be formulated as a function of area planted
to maize in the previous year, the producer price of maize, the producer prices of the other
crops a farmer can produce, the production cost, and the amount of rainfall. Because of
lack of data on other inputs, the cost of fertilizer will be used as a proxy for the total cost
of producing maize. ‘

Producers usually have more control over the number of units planted than over the
yield. Yields are largely inﬂuenced by factors over which a farmer has no control (like
moisture, temperatures, pests, etc.). These factors render the speciﬁcation of the yield
equations difﬁcult; the latter frequently exhibit strong underlying trends. Moreover, it is
often difficult to identify or measure precisely how much of a given change in output is due
to technical improvements and how much is due to changes in factor or product prices.
Because of the deﬁnitional and measurement problems involved, there has been a tendency

in empirical analysis of supply to use time or some simple trend variable as a measure of

95
technological improvements without speciﬁcally identifying and measuring those factors
responsible for shifts in supply (T omek and Robinson, 1990). Therefore, yield will be
formulated as a function of time trend, fertilizer consumption index and rain.

Net maize imports are deternnined by the change in stocks which are derived as the
residual of domestic production plus net imports less total maize demand and waste. The
level of change in stocks is used as a proxy for the level of stocks since the there are no data
on maize stocks.

Maize is demanded for domestic consumption, for feed, for seed, and for industrial

ﬂ
processing into maize products other than mealie. meal. Domestic consumption is
determined by commodity price, prices of substitutes, income, change in tastes and
preferences captured by time trend, and population which is incorporated by estimating per
capita domestic maize consumption. Demand of maize for feed will be determined as a
derived demand due to lack of data on livestock and feed price. That is, demand for feed
is derived from the demand for livestock, which in turn is determined by income and time
trend. The same argument follows for the demand of maize for industrial processing.

Maize demand for seed is determined by the area planted to maize.

The regression equations will initially be speciﬁed as follows:

(1) T‘MZY, - «ca/1119071146139
(37312411154, - f(C,TMZA.REAt-1, M2313), 3919?, 173mm,, amt)
(3) 11629301) - mzv - TM
t t
(4) NMZIMP, - r(c, Mzsmcx,_1, 11123300,)
(5) mzo: - TMZPROD, + NMZIMP, + MZSTOCKH
\(yﬁwzcrc, - 1(c, MZMRPDI, met, GDPPCDt, men—217.313,, scpaonrct)
(7) 'I'Mzr:t - T'MZCPC, * nor,

(3) PROCPC, - r(c, current, TIMI-1,)

 

(9) uzraoc,
(10) 1-‘33171’ct
(11) 13213331)t
(12) M25539:
(13) TMZQ‘,’
(14) uzsrocx,
(15) M2931),
(16) MZMS,
(17) 31mm,
Where:

“1'36ant
11462141134,
We
TMZPROD:
1123171)t

3313‘:

FERTPDt

scpaonpc,

MZMRPD,
1?.1c3111’1)t
MZPROC,

1123331)t

96
- PROCPC, ' mp,
- ac, current, 11163,)
- FEBDPCt ‘ T'POP,
- {(C, W)
- TMZC, + MZPROC, + MZFEBD, + MZSEEDt
- mo: - mzo‘,’
- {(c, 33111717,, mock”, Paoncosrt)
- {(C, CPI” MZSUBSPQ
- (M318, 0 mpg/c121,

- average maize yield in time t (in mt./lna.)

- areaharvestedofmaizeintimet(in000ha.)

- area harvested ofmaize lagged one period

- total maize production in time t (000 mt.)

- deﬂated producer price of maize in time t (K/mt.)

- deflated producer prices of other crops in time t (K/mt.)
- deflated fertilizer price in time t (K/mt.)

- Per capita sorghum production irn time t (000 mt.) v
- total domestic maize consumption in time t (000 mt.)

- domestic maize consumption per capita in time t (mt.)

- deﬂatedpercapitagossdomesticproductintimet(000K)
- Netmaizeirnportsintimet(i11000mt.)

- maizestocksintimet(000mt.)

- maize stocks lagged one period (000 mt.)

- deﬂated maize mealie meal retail price in time t (K/mt.)
- deflated retail price of rice in time t (K/mt.) v

- maize demanded for processing in time t (000 mt.)

- maize demanded for feed in time t (000 mt.)

V

f

97

PROCPC, - per capita maize demanded for processing in time t (mt.)
FEEDPC, - per capita maize demanded for feed irn time t (mt.)
MZSEED, -maizedemandedforseedintimet(m0mt.)
mzo: - total quantity of maize supplied in time t (000 mt.)
'I'MZQ‘,i - total quantity of maize demanded in time t (ill) mt.)
TPOP, - total population 131 period t (’000 persons)

RAIN, - average rainfallinpariodt(mm.)

TIME, - time trend «\I

PRODCOSI‘, - other costs in maize production in time t

CPI, - consumer price index in time t (1985 -100)
MZSUBSP, -. maize marketing subsidy per bag in time t (K/mt)
MZMS, - nominal maize marketing spread in time t (K/mt)
MZPP, - nominal producer price of maize in time t (K/mt)

C - constant

Equations (1) to (5) are supply equations while equations (6) to (13) are demand
equations. Equations (5), (13) and (14) are identities. Equations (15) to (17) are price
equations. Although the equations are originally speciﬁed as above, other speciﬁcations will
be tried and only those that pass the statistical and economic tests will be selected for use
in forecasting.

Generally, it is argued that if area planted in the current time period is the
dependent variable, then the independent variables for prices and costs are those for the
previous year (Tomek and Robinson, 1990). This is based on‘the fact that at the time of
planting, farmers do not know the prices at which they will sell their produce. However, in
Zambia, the government announces both the producer prices and the input prices (for seed

and fertilizer) at the beginning of each marketing/cropping season (May), which is ﬁve

98

months before planting time. Since these prices remain in effect until the next crop and

marketing season, no time lags are included in the model for prices.

4-43 W

In this study, the ordinary least squares (OLS) estimator will be employed to
generate parameter estimates. To obtain unbiased and consistent results from the OLS
technique, the following conditions are assumed to hold (Kennedy, 1985):

(1) The dependent variable should have a linear functional relationship with both
the independent variables and the error term. The unknown coefﬁcients of this linear
function are assumed to be constant.

(2) The conditional expected value of the error term (e) given any exogenous

variable is assumed to be zero. That is, E(e:|X:)=0
(3) E(e,,es) s 0 fort 1‘ s; t,s = 1, 2, ..., n; and E(e,)2 a Var(€:l&)=af- The ﬁrst

equation states that the error terms are not correlated with each other since they are
assumed to be distributed randomly. The second equation states that all the error terms
have the same variance, thus assuming homoscedasticity in the model.

(4) The fourth assumption is that observations of independent variables are ﬁxed in
repeated samples such that the explanatory variables are not correlated with the error term.
That is, E(e2,X,) = 0.

(5) The last assumption is that the number of observations is geater than the
number of independent variables and that there are no linear relationships between the

independent variables. That is, there is no perfect multicollinearity among the regressands.

99

(5) The last assumption is that the number of observations is geater than the
number of independent variables and that there are no linear relationships between the
independent variables. That is, there is no perfect multicollinearity among the regessands.

When lagged endogenous variables are used in the equation as explanatory (and
hence exogenous) variables, the fourth assumption is violated since the lagged dependent
variable is correlated with all of the past disturbances (Kennedy, 1985). In this case, the
model is said to have auto-regession. However, Kennedy (1985) argues that although the
lagged dependent variable is not independent of the disturbance vector, it is
contemporaneously independent of the disturbance since it is not correlated with the current
and future disturbances. This implies that although OLS is a biased estimator of the
parameters, it is consistent and it is on these gounds that the OLS is usually adopted as the
most appropriate estimator. Moreover, Potlury and Miller (1970) point out that when
lagged dependent variables are on the right-hand-side of the equation, no estimation
procedure has been shown to be better in small samples. Therefore, based on the above
justiﬁcations, the 01.8 estimator will be used even though some equations will violate the
fourth assumption.

When a model has endogenous variables among independent variables, it is said to
be simultaneous. When a model is simultaneous, Kennedy (1985) says the OLS is both
biased and inconsistent. In such a situation, a two-stage least squares (ZSLS) estimator is
preferred to the OLS since it is consistent, and if it there are no lagged endogenous
variables appearing among the exogenous variables, it is also unbiased. However, Kennedy
(1985) further argues that not all simultaneous systems of equations are biased. If a
simultaneous system of equations is recursive, there is no contemporaneous correlation

between the disturbance and the regressors. Hence the OLS is consistent, and if no lagged

100
endogenous variables appear among the exogenous variables, it is unbiased. Therefore,

although the model used in this study is simultaneous, the use of the OLS estimator is

justiﬁed by the fact that the system of simultaneous equations is recursive.

4.4.4 mm
A number of validation statistics will be applied to the estimated models to come up

with one to use to forecast maize prises, supply and demand. These will include the ones
mentioned in section 4.2, the graphical method, the Mean Squared Error (MSE) method,

the Root Mean Squared Percentage Error (RMSPE) and Theil’s U-Statistic.

4.45 W

Once estimates of the parameters of an econometric model are available, the model
can be employed to forecast the dependent variables if the associated values of independent
variables are given. Both ex ante forecasting and ex post forecasting will be utilized in this
study. The former is the forecasting of unknown values while the latter is the forecasting
of krnown values.

For independent variables whose actual future values are known, these values will
be incorporated a priori in forecasting future values of dependent variables. For the period
where actual future values of the independent variables are unknown, the values will be
forecasted using the method that seems appropriate and then used to forecast dependent

variables.

101

45 Summary

This chapter reviews economic theory behind supply and demand of a commodity
and discusses various steps involved in commodity market modeling. The steps in
commodity modeling include model structure, model specification, parameter estimation,
model validation, and model forecasting. This information is used in specifying the
econometric model suited for the maize market in Zambia. Based on the structure of the
model, the 01.8 estimator is selected to be used to generate parameter estimates in the
model. Since a number of speciﬁcations will be tried, the model will be validated using

various methods to come up with the ’best’ one that will be used for forecasting.

 

5. PRESENTATION AND ANALYSIS OF RESULTS

As pointed out earlier, a good model for forecasting should be both economically
and statistically sound. This chapter discusses the results of the model that is selected for
analysis based on the economic and statistical tests and forecasts of prices and quantities

of maize. The equations are estimated using annual data from 1964 to 1990. Prices and

A.

income are deﬂated using the consumer price index (CPI) to incorporate the general
inﬂation rate rather than use the CPI as a separate variable. Total population is also not

used as a separate variable since demand equations are estimated on a per capita basis.

5-1 W

The aim of estimating supply equations is to measure the farmers’ response, which
is indicated by the acres planted in the case of crop production. Due to the need to
separate farmer response from weather and disease effects which affect yield, equations for
the area planted to maize and average maize yield are estimated separately. Domestic
production of maize is obtained by multiplying maize yield by maize area harvested. The
estimated equations on the supply side of the model are presented below together with the

selected diagnostics. The values in brackets are the t-statisties for the respective coefﬁcients.

5.1-1 W
TMZY - 0.498 + 0.038 TTME + 0.440 FCI + 0.0002 RAIN +0.720 DV88 - 0.609 DV90 (5.1)
(1.735) (2.957) (1.530) (0.721) (3.743) (.2994)

ADJ. ReSQUARED: 0.88 SER: 0.175 D-W STAT:1.19 F-STAT: 38.647
PERIOD OF FIT: 1964-1990

Where:
TMZY - national average maize yield
TIME - time trend variable
FCI - fertilizer consumption index

102

103

RAIN -nationalannualaveragerainfall
DV88 - d1rnmyvariab1eequal1irn1988and00therwise
DV90 -dnunmyvariableequallinl990and00therwise

As mentioned in chapter four, the yield of maize is dependent on the use of
improved technology such as the use of high-yielding seeds and fertilizer. In section 3.1.12,
it was observed that there is a positive relationship between the use of fertilizer and maize
yield. In the model, the effect of commercial fertilizer on maize yield is captured by the
fertilizer consumption index variable, while other technological eﬁects, diseases and secular
shifts are captured by the time trend variable.

The yield of maize also depends on weather conditions which are captured by the
rainfall variable. It is due to a priori information on the importance of weather on yield
that the rainfall variable has been maintained in the maize yield equation even though it is
not signiﬁcant. The explanation for the non-signiﬁcance of rainfall could be that the stations
on which data are available and from where the national average rainfall is calculated are
not representative of the locations of maize production. Zero-one dummy variables for 1988

and 1990 are used to capture the abrupt shifts in average maize yield for the respective

years.
5.12 W
T'MZAREA - 556.800 + 37334 MZGMD + 0.439 73124113361) - 16.371 cmrmn
(3.005) (1.990) (3241) (4.435)
433.406 DV780N (52)
(.3342)

ADJ. R-SOUARED: 0.84 SER: 86.439 D-W STAT:2.510 F-STAT: 31282
PERIOD OF FIT 1967-1990

Where:
TMZAREA - total area harvested of maize
MZGMD - ma'ae goss margin deﬂated

104

TMZAREAGI) -totalareaharveatedofmaizelaggedoneperiod

GN'UI'PPD -groundnutprorhcerpricedeﬂated
DV780N-dunnmyvariableequalto1from19780nwardsandOotherwise

The response of farmers represented by the area harvested of maize is explained by
the deﬂated maize grow margin which captures the producer price of maize and the price
of fertilizer which is a major input; the area harvested of maize in the previous period; the
deﬂated producer price of goundnuts as a substitute; and a dummy variable from 1978
onwards. Due to lack of data on total cost of maize production, only the cost of fertilizer
is taken into account when calculating the maize gross margin. The separate use of maize
producer price and fertilizer price generated coefﬁcients that were not signiﬁcant and some
variable combinations even generated wrong signs on coefficients of these variables, hence
the use of the maize goss margin.

A dummy variable of ones from 1978 onwards and zeros otherwise is used to capture
the unexplained drop in the area harvested of maize starting from the late 1970s of which
1978 seems to be in the middle.

Domestic production of maize was an identity obtained as:

TMZPROD - TMZAREA ‘ TMZY (5.3)

Where:
TMZPROD - total maize production
TMZAREA - total area harvested of maize

TMZY - total maize yield
5.13 W
1mm - 4.143 - 0.061 731er01) - 0.063 uzsrocm-1)+ 6.063 711.13 . 166.857 DV67
(0.093) (-1441) (.1267) (3.154) (.2662)
+ 253.229 DV71 + 256.141 DV80 (5.4)
(4.130) (4213)

ADJ. R-SQUARED: 0.72 SER: 58.697 D-W STAT: 1.73 F-STAT: 11.466
PERIOD OF FIT: 1965-1990

105

NMZIMP -netmaizeirnports
TMZPROD -totalma’neprorh1aion
MZSPOCK(-1) -clnangeinma'zeatockalaggedoneperiod

TIME -tirnetrendvariable

DV67 -¢nmmyvariableerpnaltolin1967,00therwiae
DV71 -4annnmyvariableequaltolinl971,00therwise
DV80 -dumrnyvariableequaltolin1980,00therwise

Variations in net maize imports are explained by domestic production, the previous
year’s change in stocks of maize, the time variable, and three zero-one dummy variables for
1967, 1971 and 1980, respectively, which have captured the abrupt shifts (outliers) in the
dependent variable. Consumption of maize in the previous year also affects the current
year’s net maize imports but this is captured in the lagged change in maize stock variable
because of the way the latter variable was calculated.

The sum of equations 5.3 and 5.4 less maize wasted gives an identity which is the
total net maize supply.

mzo‘ - 131231101) + 19312131? - MZWASTE (55)
Where:

T‘MZDs - total quantity of maize supplied

TMZPROD - total maize production

NMZIMP - net maize irrnports

MZWASI'E - maize wasted
52 Watts

The statistical diagnostics selected for analysis are the t-statistic, the adjusted R-
squared, the standard error of regession (SER), the F -statistic and the Durbin-Watson (D-
W) statistic. The adjusted R2 will be used in the analysis rather than an unadjusted R2
because the latter is biased if many independent variables are used to explain the variation
in the dependent variable. Each of these statistics is discussed in chapter four section 4.2.

In addition to statistical viability, the estimated supply and demand equations are also

selected based on the correctness of the relationship between the dependent and

106
independent variables as indicated by the sign of the coeﬁ'icient and the signiﬁcance and

importance of that relationship.

52.1 W

5.2.1.1 W
In general, the higher the R2 the closer is the functional relationship between the

dependent and the explanatory variables. The yield equation has an adjusted R2 of 0.88
which means that 88 percent of the variation in the dependent variable is explained by the
independent variables after correcting for the eﬁ'ect of the number of explanatory variables
included in the equation.

The yield equation is also tested with regard to the t-statistics corresponding to each
independent variable used in it. A t-statistic tests whether a coefficient of an independent
variable explains nothing in the regression. It is the ratio of the coefﬁcient to its standard
error. In general, a t-statistic of geater than one means there is at least a two-thirds
likelihood that the true value of the coeﬁ'icient is not zero, while if the t-statistic exceeds two
in magnitude it is at least 95% likely that the coefﬁcient is not zero (Hall, e331, 1990).

Using this rule of thumb, it is found that in the yield equation, all the regression
coefﬁcients, except for the ones corresponding to the fertilizer consumption index and
rainfall, are signiﬁcant at the 5% level of signiﬁcance. However, the coefﬁcient on fertilizer
consumption index has a t-statistic geater than one indicating that at least there is a two-
thirds chance that it has inﬂuence on the dependent variable. Due to a priori information
on the effect of fertilizer and rain on maize yield, the fertilizer consumption index and
rainfall variables are maintained in the equation even though they are not signiﬁcant at the

5% level of signiﬁcance.

107

The F-statistic is a test of the hypothesis that all the coefﬁcients in a regession,
except the intercept or constant, are zero. If the F-statistic is geater than the critical level,
it means that at least one of the coeﬁicients is probably non-zero. In the yield equation, the
F-statistic of 49.28 exceeds the critical values at both the 5% (FD-956121) s 2.69) and the
1% (FD-”(5 '21) - 4.04) levels of signiﬁcance, respectively. This implies that there is a
regession relation between the yield of maize per hectare and at least one of the
independent variables.

The Durbin-Watson (D-W) statistic is a statistic for serial correlation. As a "rule of
thumb”, a D-W statistic of much less than 2 implies there is positive serial correlation and
vice versa if it is much above 2. A D-W statistic of around 2 indicates that there is no serial
correlation. When there is serial correlation, it is better to re-specify the equation, although
it can also be corrected. In the yield equation, D-W statistic of 1.19 fell in the inconclusive
region as to whether serial correlation is present or not. Hence, there was no need to re-

specify the equation or correct for auto-correlation.

5.2.1.2 W620
The adjusted R2 of 0.84 shows that the independent variables used in the area

harvested of maize equation explain 84% of the variations in the dependent variable. T-
statistics geater than 2 for coefﬁcients of the lagged area harvested of maize and the
dummy variable for 1978 onwards show that these variables are signiﬁcant at the 5% level
of signiﬁcance. The coefﬁcients on maize goss margin and the deﬂated retail price of
goundnuts are not signiﬁcant at the 5% level of signiﬁcance but they still provided at least
two-thirds certainty that they affect the dependent variable since they have t-statistics that

are geater than one.

108

The equation for the area harvested of maize has an F-statistic of 31.282 which
exceeds the critical values of 2.9 and 4.50 at the 5% and the 1% levels of signiﬁcance,

respectively. This implies that not all the independent variables are equal to zero.

5.2.13 MW

The equation for the net imports of maize has an adjusted R2 of 0.72. The t-
statistics of the coefficients corresponding to the time variable and all dummy variables are
signiﬁcant at the 5% level of signiﬁcance. The t-statistics corresponding to lagged change
in maize stocks and total maize production are not signiﬁcant at the 5% level of signiﬁcance.
However, with their coefﬁcients being geater than one, they are maintained in the model.

The F-statistic of 11.466 is signiﬁcant since it exceeds the critical values at both the
5% (F0.95(6,19) = 2.63) and the 1% (F0.99(6,19) = 3.94) levels of signiﬁcance. A D-W

statistic of 1.73 in the equation of net maize imports indicates no presence of auto-

correlation.
522 W
5.2-2-1 We!)

Economic evaluation of the estimated equations refers mairnly to the signs and
magnitudes of the estimated parameters. Both the TIME and FCI variables have expected
signs according to econonnic theory. Positive coefﬁcients on the TTME, PC] and RAIN
variables in the yield equation imply that holding other things constant, the yield of maize
in the current year would increase with the use of improved technology like hybrid seed
(captured by TIME), the use of more fertilizer in maize production and more rainfall (at

least within a normal range).

109

5222 W
The coeﬁicients of all independent variables in the equation for area harvested of

maize have the riglnt signs according to a priori information. The goss margin of maize has
a positive sign indicating that the more proﬁt farmers expect from the sales of maize, the
more land they will plant to maize. A negative sign on the producer price of goundnuts
indicates that the higher the producer price of goundnuts in relation to that of maize, the
less land that will be utilized for maize production. This indicates that goundnuts are a

substitute for maize.

5.2-23 Winn

The net maize imports equation reﬂects correct signs on the coefficients of all the
independent variables. The negative sign on the coefficient of domestic maize production
implies that holding other things constant, net imports of maize would decline with an
increase in domestic maize production. The same interpretation follows for the negative
sign on the change in maize stocks lagged one period. However, the positive sign on the
time trend depicts the increase in net imports of maize over time. This is true since for the
most part of the period under analysis, Zambia has not been self-sufficient in maize

production and has been a net importer of maize.

5.3 W

The selected maize demand equations are presented in equations 5.6 through 5.13,

together with their respective diagnostics.

110
53.1 W

TMZCPC - 0.114 + 0.291 rum-1) - 0.023 MZMRPDR - 0.688 SGPRODPC

(5323) (2132) (.1352) (.1333)
+ 0.001 TIME - 0.023 DV70 (5.6)
(2501) (.4346)

ADJ. R-SQUARED: 0.91 SER: 0.908 D—W STAT:2.27 F-STAT: 50.643
PERIOD OF FIT: 1965-1990

Where:
TMZCPC -totalmaizeconsumptionpercapita 0,0)«315‘3
TMZCPC(-1) -totalmaizeconsunnptiornpercapitalaggedoneperiod «Mn.
MZMRPDR -ntioofdeﬂatedmaizemealretaﬂpricetodeﬂatedpoﬁshedriceretaﬂpﬁce
SGPRODPC - per capita production of sorgnum
TIME -timetrendvariable
DV70 -dummyvariableequaltolin1970,00therwise

Per capita domestic maize consumption is explained by domestic maize consumption
lagged one period, the ratio of the deﬂated retail price of maizemealie meal to the deﬂated
retail price of polished rice, the per capita production of sorghum, the time variable, and
a zero-one dummy variable for 1970 which captures the abrupt decrease in maize
consumption in 1970. The ratio of the retail price of maize to that of rice captures th
effects of own price and the price of a substitute (rice). The ratio of the two retail pri
is used instead of separating them because where the right signs were obtained upon using
separate prices for maize and rice, the generated coefficients were insigniﬁcant. In othe
cases, the use of separate prices gave wrong signs for the coefficients of these variables.

Generally, if quantity is the dependent variable, then the prices are supposed to
independent variables. However, per capita produ ' n sorghum was used as an .
independent variable (substitute) in this model because r um is basically produced for i, \i
own consumption with no readily available market such at theiretail price of sorghum has
never been published.

According to economic theory, income is one of the major determinants of demand.

For a normal commodity, there is a positive relationship between income and consumption

11 1

of a commodity while there is a negative relationship between income and consumption of
an inferior good. The deﬂated per capita goss domestic product had a negative
relationship with per capita domestic consumption of maize indicating that it is an inferior
goodinZambia. Thiswasexpectedinthemodelasisobservedanddiscussedinsection
3.8.1 of chapter three. However based on the discussion in section 3.8. 1, this study could
not conclude that maize is an inferior good in Zambia and per capita real GDP was
dropped from the model.

The total demand of maize for domestic consumption is given in the identity below:

TMZC - TMZCPC ‘ TPOP (5.7)
Where:
T'MZC - total maize consunnption

TMZCPC - total maize consurnnption per capita
T'POP - total population

532 EsiimatsdﬂemandsfMaizefcdndnatrialImina

Pnocrc - - 0.0002 + 0.0565 GDPPCD + 0344 mom-1) +0001 DV77 (5.3)
(.0436) (1500) (3260) (2.012)

ADJ R-SQUARED: 0.94 SER' 0.0004 D—W STAT: 1.74 F-ST AT: 132.304
PERIOD OF FIT: 1965-1990

Where:

PROCPC a per capita maize for industrial processing

GDPPCD - per capita goss domestic product deﬂated

DV77 - dummy variable equal to one in 1977, 0 otherwise

Variations in per capita demand of maize for industrial processing are explained by
the deﬂated per capita goss domestic product, the per capita demand of maize for
industrial processing lagged one period, and a zero-one dummy variable for 1977 in which
year there was an abrupt shift in the dependent variable. Due to lack of variables to

estimate demand of maize for industrial processing as a direct demand, demand of maize

for industrial processing is estimated as a derived demand whereby it is assumed maize is

112
demanded for processing based on the demand for the ﬁnal products in which processed

maize is used as an input. Therefore per capita goss domestic product has to be included
in the equation as a major variable. It is for this reason and because it at least has a t-

statistic greater than one that it is maintained in the equation even thougln it is not

signiﬁcant at the 5% level of signiﬁcance.
Total demand of maize for industrial processing is obtained as:

MZPROC - PROCPC ’ T'POP (5.9)
Where:
MZPROC - total processed maize for industrial use

PROCPC - percapitaprowssedma'nefor industrialu
T'POP - total population .

533 WWW

FEEDPC - - 0.0012 + 0264 GDPPCD + 0.772 FEEDPq-n)
(.0735) (3227) (7.7186)

- 0.005 DV79 + 0.006 DV83 (5.10)
(.2371) (3.137)

ADJ. R-SQUARED: 0.78 SER: 0.0018 D-W ST AT:2.2 F-SI'AT: 23.06
PERIOD OF FIT: 1965-1990

Where:
FEEDPC - maize feed per capita
GDPPCD - per capita goss domestic product deﬂated
FEEDPC(-1) - maize feed per capita lagged one period
TTME - time trend variable
DV79 - dumnny variable equal to 1 in 1979, 0 otherwise
DV83 - dummy variable equal to 1 in 1983, 0 otherwise

The per capita demand of maize for feed is explained by itself lagged one period,
the deﬂated per capita goss domestic product, and zero-one dummy variables for 1979 and
1983, respectively. In the absence of variables like the livestock numbers which could allow
for the direct estimation of the demand of maize for feed, the same logic that is used to
estimate demand of maize for industrial processing is used to estimate the demand of maize

for feed. The total demand of maize for feed is obtained as:

113
MZFEED - FEEDPC ‘ TPOP (5.11)

Where:
MZFEED - totalmahefor feed
FEEDPC - maizefeedpercapita
TPOP - totalpopulation

53.4 W

MZSEED - 2120 + 0.0273 1112411314 - 3.193 DV78 (5.12)
(0.969) (10.425) (.2925)

ADJ. R-SQUARED: 0.82 SER: 2.749 D-W STAT: 1.68 F-STAT: 58.655
PERIOD OF FIT: 1964-1990

Where:
MZSEED - maize seed
TMZAREA - total area harvested of maize
DV78 - thnmmy variable equal to zero in 1978, 0 otherwise

The variation in the demand of maize for seed is explained by total area harvested
of maize and a zerooone dummy variable for 1978.

The total demand for maize is represented as an identity as follows:

'11»erd - mzc + mzraoc + MZFEED + MZSEED (5.13)

Where:
T'MZQ‘l - total quantity of maize demanded
TMZC - total ma'ne consumption
MZPROC - total maize demanded for industrial prowssing
MZFEED - total maize demanded for feed
MZSEED - maize seed

Equation 5.14 is an identity representing changes in maize stocks in each year and

it is used in the model as a proxy for maize stocks whose data are not available.

uzsrocx - T'MZQ‘ - “mmd (5.14)

where:
MZS'I‘OCK - annual change irn maize stocks
TMZQ‘ - total net maize quantity supplied
“11020d - total quantity of ma'ne demanded

 

The independent variables in the per capita domestic maize consumption equation
explain 91% of the variation in the dependent variable. The lagged per capita domestic
maize consumption, the time variable and the dummy variable for 1970 have signiﬁcant t-
statistics at the 5% level of signiﬁcance. Althougln per capita production of sorglnum and
the ratio of the retail prices of maize to that of rice are not signiﬁcant at the 5% level of
signiﬁcance, their t-statistics are greater than unit. Therefore, they are maintained in the
equation.

The per capita domestic maize consumption equation has an F-statistic of 38.46.
This is greater than the critical values of 2.71 and 4.10 at the 5% and the 1% levels of

significance, respectively, thereby verifying its signiﬁcance. A D-W statistic of 2.27 shows

that there is no serial correlation irn the equation.

 

The equation of per capita demand of maize for industrial processing has an
adjusted R2 of 0.94. The t-statistic of the coefﬁcient corresponding to deﬂated per capita
gross domestic product is not signiﬁcant at the 5% level of signiﬁcance but has a t-statistic
which is geater than one. It is maintained in the equation for that reason and because it
is taken to be an important variable based on the assumption that demand of maize for
industrial processing is a derived one in this model. The t-statistics of the other coefﬁcients

of the independent variables are signiﬁcant at the 5% level of signiﬁcance.

1 15
An F-statistic of 132.304 in the per capita demand of maize for industrial processing

equation is signiﬁcant at both the 5% and the 1% levels of signiﬁcance since it exceeds the
critical values of 3.05 and 4.82 at those levels of signiﬁcance, respectively. The equation has

3 DW statistic of 1.74 which shows no presence of serial correlation.

5.4.1.3 W
The independent variables in the equation for per capita demand of maize for feed

explain 78% of the variation in the dependent variable. The t-statistics of all the coefﬁcients
are signiﬁcant at the 5% level of signiﬁcance. An F -statistic of 23.06 exceeds the critical
values of 2.84 and 4.37 at the 5% and the 1% levels of signiﬁcance, respectively, and is
hence signiﬁcant. A D-W statistic of 2.2 in the per capita demand of maize for feed

equation shows no presence of auto-correlation.

5.4.1.4 W

The adjusted R2 in the demand equation for maize seed is of 0.82. The t-statistic
of the coefﬁcients of both independent variables in the regession is signiﬁcant at the 5%
level of signiﬁcance. An F-statistic is also signiﬁcant at both the 5% and the 1% levels of

signiﬁcance while 3 DW statistic of 1.68 falls in the inconclusive region.

5.4.2 EmmaﬁzaluaticufMaizeﬂemaniﬁauaﬁcns

5.4.2.1 Wm
All the coefﬁcients of the independent variables have expected signs. The negative

coefficient on per capita sorghum production shows that holding all other things constant,

the higher the per capita production of sorghum the lower the per capita domestic maize

116

consumption. The negative sign on the ratio of retail price of maize to that of rice indicates
that per capita domestic consumption increases as the ratio decreases. This is true
according to a priori information since rice is a substitute of maize. A low retail price ratio
implies that the retail price of maize is relatively lower than that of rice, and vice versa for
the reverse situation. The positive coeﬂ'ncients on the time variable and on the lagged per
capita domestic maize consumption indicate that per capita domestic maize consumption
has been increasing with time and with per capita domestic maize consumption in the

previous year.

5.4.2.2 WW3

In the equation of per capita demand of maize for industrial processing, both the
deﬂated per capita goss domestic product and lagged per capita demand of maize for
industrial processing have positive coeﬁicients as expected. This means that holding other
things constant, an increase in either of them leads to an increase in the current per capita

demand of maize for industrial processing.

5.423 W

As in the per capita demand of maize for industrial processing equation, all the
coefficients of independent variables in the per capita maize feed demand equation are

positive. The same interpretation as above applies here too.

5.4.2.4 DcmanicLMaizefccﬁes-d

The positive sign on the area harvested of maize coefﬁcient implies that the demand

for maize seed increases with the increase in area harvested of maize in the current year.

117
This result is expected since any additional area will require more seeds for planting,

assuming the same planting density.

55 E . , E . E .

55.1 W020

MZPPD - 0.831 + 0396 MZPPD(-1) + 0.754 FERTAPD -0.003 TMZPROD + 1.430 DV86 (5.15)
(1.435) (3.197) (3.409) (-1320) (3.915)

ADJ. RoSQUARED: 0.66 SER: 2.433 D-W STAT: 2.11 F-STAT: 13.381
PERIOD OF FIT: 1965—1990

Where:
MZPPD - deﬂated producer price of maize
FERTAPD - deﬂated retail price of fertilizer
TMZPROD - domestic maize production
DV86 - zero-one dummy variable with 1986 equal to 1, 0 otherwise

This equation attempts to capture the factors and policies that inﬂuence the
government in setting the maize producer price. Some of the factors found to be taken into
account by the government when setting the maize producer price in the current period
include the producer price of maize lagged one period, the deﬂated retail price of fertilizer,
the previous year’s maize production level, and a zero-one dummy variable for 1986. The
dummy variable for 1986 is irncluded to capture the introduction of fertilizer subsidies meant
to decrease production costs thereby keeping the maize producer price relatively higher.
Fertilizer price explains more than 50% of the variations in the maize producer price.

In the absence of the maize stocks, the variable of change in maize stocks was tried
in the equation as a proxy for the former. However, it was dropped from the equation
because it had a t-statistic of less than one. This does not necessarily mean that maize
stocks do not inﬂuence producer price setting. To partly capture the effect of the maize
stocks on the producer price of maize, the previous year’s maize production is used and

maintained in the equation since it has a t-statistic geater than one.

 

MZMS - - 34322 + 5.774 CPI - 1.016 uzsuasr - 1157.435 DV88
(.1313) (22359) (.7157) (.5071)

ADJ. R-SQUARED: 0.98 SER: 219.213 D-W STAT: 1.66 F-STAT: 386.762
PERIOD OF FIT: 1964-1990

Where:
MZMS - maizemarketingspread
CPI - consumer priceirndea (1988-100) forMarcln
MZSUBSP - maizemarketing subsidypermetric torn
DV88 -dumnryvariablewith1988equaltol,0otherwise

MZMRPD - (MZPP + MZMS)/CPI

Where:
MZMRPD -deﬂatedbreakfastmaizemealretailprice
MZPP - nominal maize producer price
MZMS - maize marketing spread
CPI - consumer price index (1988-100)

(5.16)

(5.17)

For the deﬂated retail price of maize, the equation for maize marketing margin

(marketing spread) is estimated (equation 5.16) and then the former is obtained as an

identity by adding the maize marketing spread to the producer price of maize and deﬂating

the sum (equation 5.17). The maize marketing spread is explained by the consumer price

index, the maize marketing subsidy per metric ton, and a dummy variable for 1988. The

dummy variable for 1988 captures the beginning of the gadual removal of maize marketing

subsidies when the government reduced into-mill price subsidies by K3000 per 90-kg bag

of maize in the 1988/89 marketing season.

5.6 W
56.1 W
56.1.1 W

Taking into account the number of variables used in the equation, the independent

variables included in the deﬂated maize producer price equation explain 67% of the

1 19
variations in the dependent variable. All the independent variables’ coefficients, except the

one for lagged maize production, have t-statistics geater than 2, implying that they are
signiﬁcant at the 5% level of signiﬁcance. However, the t-statistic of the lagged maize
production is geater than one in absolute terms which provides at least two-thirds certainty
that the previous year’s maize production level aﬁects the current producer price of maize.

AntF-statistic of 13.381 is signiﬁcant at both the 5% and the 1% levels of
signiﬁcance, while a D-W statistic of 2.12 does not indicate any presence of serial

correlation.

5-6-1-2 W

The adjusted R2 in the maize marketing spread equation is 0.98. All the coefﬁcients
of the independent variables are highly signiﬁcant at the 5% level of signiﬁcance. The F-
statistic of 386.762 is signiﬁcant at both the 5% and the 1% levels of signiﬁcance. A D“!

statistic of 1.66 falls in the inconclusive region.

562 £220210i2E22123020212022E2232i203
5621 W

All the independent variables are consistent with economic a priori information. A
positive sign on the coefficients of fertilizer price and lagged maize producer price imply
that holding other things constant, the higher the fertilizer price in the current year or the
higher the maize producer price in the previous year, respectively, the higher will be the
producer price of maize in the current year. However, as expected, the higher the previous

year’s production of maize, the lower will be the current year’s maize producer price.

120

5622 W
The signs on all the irndependent variables are as expected. Holding all other things

constant, the maize marketing spread increases with a rise in inﬂation but decreases with
an increase in the maize marketing subsidy. The positive sign on the coeﬁ'icient of the
dummy variable indicates that the maize marketing spread increased with the removal of
some into-mill price maize subsidy, resulting in the reduction of the overall marketing

subsidy for maize.

5.7 5.112201233021025

The short-run and long-run price and income elasticities of supply and demand are
given in Table 5.1. Short-run elasticities are calculated at the mean level of the respective
N___ _...

variables involved. As expected, the long-run elasticity of supply and demand are geater
W

Table 5.1

Short-run and Long-run Supply and Demand Elasticities

 

5020:1110 1203:1110
Gross margin elasticity of supply 0.19 0.37
Price elasticity of supply 0.22 0.51
Cross-price elasticity of supply -0.21 -0.42
Demand price elasticity (Consumption) -0.03 -0.04
Income elasticity (feed) 0.32 1.39

Income elasticity (processed maize 0.15 0.96

 

121

than short-run ones since in the long-run, farmers have more time to shift their resources
around than in the short-run while consumers’ expenditure on a commodity has a multiplier
effect over time.

The short-run (crop year-to-crop year) and long-run elasticity of supply with respect
to deﬂated maize goss margin is 0.19 and 0.37, respectively. This means that if farmers
expect the maize goss margin to increase by 1%, they would increase the area harvested
of maize by 0.19% in the short-run and by 0.37% in the long-run. The short-run and long-
run price elasticities of supply are 0.22 and 0.51, respectively. These results are quite
comparable to that estimated by Timmer am. (1983) who claim that short-run gain supply
elasticities are generally between 0.2 and 0.3 althougln there is a noticeable tendency for the
supply elasticity to be larger in more developed countries and in regions with longer
commercial histories.

The results in Table 5.1 show that farmers are more responsive to maize producer
price than to the maize goss margin. This could be because farmers may not be sure
whether the variable costs of maize production (like labor costs) would change from the
time they are planting to the time they are harvesting, fertilizer and seed prices are ﬁxed
by the prices set by the government. Therefore, if the government wanted to increase maize
production, it could choose between policies that increase the maize producer price and
those that reduce the cost of production thereby increasing the maize goss margin.

The short-run and long-run cross-price elasticity of supply with respect to the
deﬂated producer price of goundnuts is -0.21 and -0.42, respectively. This implies that
holding other things constant, a 1% increase in the producer price of goundnuts relative

to that of maize would result in a 0.21% shift in the area harvested of maize to the

122
production of goundnuts or other crops in the short-run while there would be a 0.42%

reduction in the area harvested of maize in the long-run.

This shows that goundnuts compete with maize for land resources, hence holding
other things constant, increasing the producer price of goundnuts would lead to farmers
allocating part of the land that was previously under maize production to goundnuts
production, and vice versa for the reverse situation.

This could be true in that goundnuts are produced throughout the country and some
parts of the country (Eastern, Northern and Luapula Provinces) have goundnuts accounting
for the second largest area under cultivation after maize (Republic of Zambia, 1990). In
fact, Northern and Eastern Provinces are even among the four maize surplus producing

provinces in the country.

58 12202025330002

In the equation for per capita domestic consumption of maize, only the ratio of
prices is included and not the separate price or income variables. Therefore, only the price
not income elasticity of demand is calculated. The short-run and long-run price elasticities
of demand of maize for domestic consumption are -0.03 and -0.04, respectively, as in Table
5.1. These are very inelastic price elasticities of demand, indicating that even if the maize
mealie meal price increased by 1%, people would only reduce their maize meal consumption
by 0.03% in the short-run and by 0.04% in the long-run. This result could be due to a
combination of the price structure of food stuffs that has left maize mealie meal the
cheapest staple food and the worsening economic situation of the people as reﬂected in per

capita real GDP that has continued to decline since the 1970s.

123

The short-run and long-run income elasticity of per capita demand of maize for feed
is 0.32 and 1.39 while tlnose of the maize demanded for industrial processing are 0.15 and
0.96, respectively. A comparison of the income elasticities of demand of maize for feed and
for industrial processing shows that the income elasticity of demand of maize for feed is
higher than that for demand of maize for industrial processing.

5-9 WW

5-9-1 W20
The values of the adjusted R2 vary from 0.67 in equation (5.15) to 0.98 in equation

(5.16). In general, the higlner the R2, the closer the functional linear relationship between
the dependent and the explanatory variables. How high the R2 should be to be accepted
is not known. Since in all cases, apart from one, more than 70% of the total variation in
each equation was explained by the respective independent variables, the model can be said
to be satisfactory. ,

With regard to the t-statistic, most of the coefficients had t-statistics above 2, hence
these were signiﬁcant at the 5% level of signiﬁcance. The rest of the variables had t-
statistics at least equal to one and thus offered at least two-thirds certainty that they
inﬂuence the respective dependent variables. The level of signiﬁcance one chooses to work
with irn the social sciences is based on a value judgement by the analyst. While in the
natural sciences, where there are control experiments, a 1% loss in the level of signiﬁcance
could make quite a substantial difference, in the social sciences this may not be the case.
Problems may arise with the way the data were collected, the assumptions made, and so on.
Therefore, based on the fact that this study used secondary data whose methods of

collection were not known and considering a priori information on the importance of each

124

variable, it was decided that variables with a t-statistic of more than one be. maintained in
the model.

In this study, the F-statistics range from 11.466 to 386.762 and are all signiﬁcant at
both the 5% and the 1% levels of signiﬁcance. The DW statistics, which range from 1.19
to 2.51, either show no presence of serial correlation between variables or fall irn regions
where one cannot conﬁdently conﬁrm the presence or absence of serial correlation.

Therefore, it is concluded that the model is statistically viable.

5.9.2 23002132003020
To validate the model gaphically, each of the endogenous variables has its actual

and projected values plotted against time. Figures 5.1 to 5.16 in Appendix A provide visual
evidence of how well the model tracked by comparing the distance between the actual and
projected values. One may be misled by the gaphical validation of the model in the sense
that the size of the difference between the actual and projected values of a variable depicted
in the gaph depends on the scale used irn the gaph. The validation is more exact the
smaller the scale.

It is observed that in some periods of the forecast, the model tracks the actual values
better than in the other periods. In some years, the model generates forecast values that
are geater or less than the actual values. However, the visinal observation shows that the

model is satisfactory.

125
593 W

The mean squared error (MSE) measures the mean of the squared difference
between actual and forecast variables. When deﬁned in terms of the differences in the
levels of the variables, the MSE is formulated as follows:

1 I

HS --2 (I’,-21,)2
n M

where:
P, -= Predicted values of the dependent variable
A, 8 Actual values of the dependent variable

0 =- number of observations

With this kind of definition, the result of the MSE varies with the units of variables. To
provide comparisons of forecasting accuracy for variables used in this model which have

different units, the MSE is measured in tennis of percentage changes (MSEP) as:

MSEP = 1/62 (p, - 6,)2
where:

Pt " (P t ' An-IVAt-l
at = (A1 'A1-1)/A1-1

The results of this method are presented in Table 5.2, together with the results from
the other validation methods discussed below. The closer the MSEP is to zero, the better
the ability of the model to forecast. Thirteen variables (TMZPC, TMZC, FEEDPC,
MZFEED, PROCPC, MZSEED, TMZQd, TMZY, TMZAREA, TMZPROD, TMZQS,

MZPPD, and MZMRPD) have MSEP statistics that are less than 5%. Only one variable

126
(NMZIMP) has a MSEP geater than 10%. The other two variables (MZPROC and

MZSTOCK) have MSEPs equal to 6% and 9.4%, respectively. Overall, since almost all the

MSEPs of the endogenous variables are very close to zero, the model is said to be

satisfactory.

 

The root mean squared percentage error (RMSPE) is a statistic that shows how well
the forecast values of the endogenous variables match with their actual values. RMSPE is

deﬁned as:

RMSPE = (I/n): «P, - Ans/4,9)”2

where:

A, = actual value of an endogenous variable

Pt = predicted value of an endogenous variable

n = number of observations

The smaller the value of the RMSPE, the better the ﬁt and the more accurate the
model is for forecasting. However, this statistic has two drawbacks. First, the RMSPE is
a measure of central tendency. Therefore, sometimes a few very large errors can raise the
value of the RMSPE of an endogenous variable that otherwise tracks very well. Second,
small errors in absolute terms give rise to large errors in percentage terms when actual
values of an endogenous variable are small.

The RMSPE for the endogenous variables are presented in Table 5.2 above, which
shows that all variables have RMSPE statistics less than 5%, except the ones for change in

maize stocks (12.7%) and for net maize imports (9.9%). Thus, the net maize imports and

127

Table 5 .2

Quantitative Measures of the Model’s Ability to Forecast

 

TMZY 0.044 0.020 0.102
TMZAREA 0.031 0.017 0.090
TMZPROD 0.046 0.036 0.174
NMZIMP 0.667 0.099 0.490
T'MZQS 0.035 0.039 0.191
TMZCPC 0.031 0.005 0.025
TMZC 0.036 0.005 0.024
FEEDPC 0.047 0.029 0.146
MZFEED 0.049 0.031 0.160
PROCPC 0.039 0.016 0.079
MZPROC 0.060 0.017 0.035
MZSEED 0.046 0.023 0.112
TMZQd 0.036 0.005 0.024
MZSTOCK 0.094 0.127 0.622
MZPPD 0.034 0.016 0.081
MZMRPD 0.040 0.019 0.090

 

128

the change in maize stocks, which are the least accurate using the MSEP, turn out to be
the least accurate even using the RMSPE. However, the net maize imports variable, which
has a MSEP of 66.7%, has a RMSPE of 9.9%. Overall, the model has performed well in

tracking the actual values of the endogenous variables.

5.95:, H. 1,. Ii 1,1,. ..
Theil’s inequality coeﬁ'icient is related to both the MSE and the RMSPE. It is

deﬁned as:

U . 0:3

(ll—9’
where MSE, A, and n are as deﬁned above. The closer the U-statistic is to zero, the better
the forecast ability of the model.

Table 5.2 above gives the Theil’s U-statistics for the endogenous variables. Unlike
the results from the MSEP and the RMSPE, the Theil’s U-statistic is less than 10% for only
eight variables. Six variables have U-statistics between 10% and 20% while two of the
variables have U-statistics geater than 20%. Net maize imports and change in maize stocks
with U-statistics of 49% and 62 %, respectively, show consistent results with those obtained
using MSEP and RMSPE in that they are the least accurate variables in tracking the actual
values even with the U-statistic. However, even this test shows that the model has

performed well enough to be used for forecasting.

5.10 W323

Once the modeler is reasonably comfortable with the model’s performance against

actual values, the next challenge is to generate projections. Before endogenous variables can

129

be forecast, exogenous variables have to be obtained for the forecast period. Where
exogenous variables are not known, they have to be projected to the end of the forecast
period ﬁrst and then used to forecast the endogenous variables.

Exogenous variables that do not show a time trend have their future values
forecasted by ﬁrst regessing the respective variable against time and then the forecast is
made based on that result. This method of forecasting assumes that the same conditions
that prevailed during the observation period would continue. The exogenous variables that
are forecasted using this method include rainfall, deﬂated retail price of goundnuts, maize
wasted, the maize marketing subsidy, and the ratio of the retail price of maize to that of
rice.

The exogenous variables that indicate a trend (fertilizer consumption index, deﬂated
goss margin of maize, total population, and consumer price index), are forecasted using
double exponential smoothing This method of forecasting does not make use of
information from series other than the variable being forecasted. Double exponential
smoothing, which makes a forecast which grows along a trend, gives high weight to the most
recent value and little weight to a longer-run average. It carries out the recursive
computation within the sample and then extends it for the rest of the observations in the
range which includes the forecast period.

For the period between 1991 and 1995, deﬂated per capita gross domestic product
is forecasted at a constant declining rate of 4.08 percent (average decline rate between 1980
and 1990). It is assumed to gow at an average rate of 5% between 1996 and 2000. The
declining trend in the ﬁrst part of the forecast is assumed as a period of implementing
policy changes that the new government has proposed and is still proposing. It is assumed

that from 1996, the positive impact of the policies will start to be felt.

130

For per capita production of sorghum, an average decline rate of 0.04% for the
period 1964 to 1990 is calculated at which rate the future level of per capita sorghum

production is expected to decline.

5.11 W
5.11.1M0mﬁnnclxr2mm

The model’s forecasted values for the area harvested of maize, average yield and
production of maize are given in Table 5.3. Total area harvested of maize is projected to
be 732.74 thousand hectares in 1991 and to g'ow at an average annual rate of 0.94% per
year between 1991 and 1995 and 0.21% per year between 1996 and 2000, reaching 806.96
thousand hectares in the year 2000.

In 1991, the average yield of maize is projected to reach 2.07 metric tons per
hectare. It would then increase at an annual average gowth rate of 10.4% between 1991
and 1995 and 1.68% between 1996 and 2000. By the year 2000, the yield is projected to
have reached 2.41 metric tons per hectare. This growth rate in maize yield is justiﬁed in
that in the late 19805, the average maize yield was about 2.2 metric tons per hectare.
However, it dropped to about 1.4 metric tons per hectare in 1990 because of the drought,
which in turn resulted in relatively lower maize production than usual. Hence, large
percentage gowth rates for the forecast period for both maize yield and maize production
reﬂect the move from the unusual low levels of 1990.

Domestic maize production is projected to be 1512.74 thousand metric tons in 1991
but would increase to 1923.23 thousand metric tons by the year 2000. This would entail an

average annual growth rate of 13.65% and 1.89% between 1991 and 1995 and between 1996

131

Table 5 .3

Zambia: Projections of Maize Area Harvested, Yield, and Production, 1991-2000'

 

 

Year TMZAREAF Growth TMZYF Growth TMZPRODF Growth
(000 ha) Rate % (mt/ha) Rate % (000 mt) Rate %
1990 764.00 ' 1.43 994.35
1991 732.74 -4.09 2.07 44.84 1512.74 52.13
1992 776.39 5.96 2.11 2.27 1639.24 8.36
1993 779.13 035 2.15 1.66 1672.24 2.01
1994 793.50 1.72 2.18 1.63 1728.63 3.37
1995 798.59 0.77 2.22 1.60 1769.81 2.38
1996 796.94 -0.21 2.27 2.27 1806.24 2.06
1997 800.07 0.39 2.30 1.56 1841.66 1.96
1998 799.27 -0.10 2.34 1.54 1868.11 1.44
1999 803.99 0.59 2.37 1.52 1907.60 2.11
2000 806.96 0.37 2.41 1.49 1943.23 1.87
MW
1991-1995 0.94 10.40 13.65
1996-2000 0.21 1.68 1.89
1991-2000 0.58 6.04 7.77

 

. Variables are as deﬁned earlier in the Chapter, the only difference is the "F‘ at the
end of each variable name which indicates that it is a forecast value.

132
and 2000, respectively. The gowth rate in the ﬁrst period seems high because the

calculation included the movement from 1990 which had very low maize production level
because of the drouglnt. The increase in domestic maize production is expected to result
more from an increase in maize yield than from the increase in the area harvested of maize

over the forecast period (see Table 5.3).

5.112 MW

The forecasts for the various types of demand for maize are given in Table 5.4. In
1991, maize for domestic consumption is projected to be 1509.11 tlnousand metric tons. It
would grow at an average annual rate of 2.87% and 3.21% between 1991-1995 and 1996-
2000, respectively. For the whole ten-year forecast period, domestic maize consumption is
projected to grow at an average annual rate of 3.04%. By the year 2000, human maize
consumption would have increased to 2009.548 thousand metric tons according to the
projection.

The increase in domestic maize consumption is projected to result from both the
increase in the total population and in the increase in per capita domestic maize
consumption over the forecast period. On one hand, total population is assumed to increase
from 7818.45 thousand persons in 1990 (according to the 1990/91 census) to 9957.08
thousand persons in the year 2000. On the other hand, the projections on domestic maize
consumption reveal an average per capita domestic maize consumption of 0.188 metric tons

'72,? in 1991 and 0.202 in the year 2000. mWﬁumgum
‘9 is quite reasonable considering that starting from the 19805, per capita domestic maize

mption has been around 0.190 metric tons.

133

Table 5 .4

Zambia: Projections of Demand of Maize for Domestic Consumption, for Feed,
for Industrial Processing and for Seed, 1991-2000

 

Thousand Metric Tons

 

1221

1990 1490.00 62.75
1991 1509.11 72.07
1992 1553.20 72.51
1993 1605.24 32.34
1994 1660.16 35.30
1995 1715.33 33.74
1996 1772.33 112.25
1997 1330.37 132.53
1993 1889.66 129.16
1999 194923 140.54
2000 200955 153.24
WW
1991-1995 2.37 7.34
1996-2000 3.21 12.69
1991-2000 3.04 10.01

26.40
25.32
24.58
24.21
23.95
23.78
23.70
23.69
23.75
24.86
24.02

-2.06
0.23
-0.91

21.13
23.83
24.07
24.31
24.55
24.80
25.05
25.20
25.33
25.50
25.62

3.36
0.65
2.01

IMZQEMZEEEDEMZEBEEMZSEEDEIMZQEF

1600.28
1630.33
1672.26
1733.43
1790.24
1849.11
1929.64
2008.01
2063.64
2135.84
2213.19

2.93
3.66
3.29

 

1‘ Variables are as deﬁned earlier in the Chapter, the only difference is the 'F" at

the end of each variable name which indicates that it is a forecast value.

134
Maize demanded for feed is projected to be 72.07 thousand metric tons in 1991 and

would rise to 158.238 thousand metric tons irn the year 2000. There would be an average
gowth rate of 7.34% during the period 1991-1995 and this would increase to a rate of
12.69% during the period 1996-2000.

The maize demanded for industrial processing is projected to decrease from 25.32
thousand metric tons in 1991 to 24.02 thousand metric tons in the year 2000. This would
result from a projected declirne in the gowth of processed maize at an average rate of
2.06% between 1991 and 1995 and an increase in the gowth rate of 0.23% between 1996
and 2000. This result assumes that since demand of maize for industrial processing has
been declining since the 19703, this trend would continue in the ﬁrst part of the forecast
period but that the trend would change in the second half of the forecast period.

An explanation for the decrease in demand of maize for industrial processing may
lie in data errors. Maize demanded. for industrial processing is mairnly used in local beer
brewing. It is most probable that the data only reﬂect the maize that is used by the krnown
commercial local beer brewing companies, while the maize used in local beer brewing by
individual households is geatly under-estimated.

The maize seed demand is projected to be 23.83 thousand metric tons in 1991 and
is expected to increase at an average rates of 3.36% and 0.65% between 1991 and 1995 and
between 1996 and 2000, respectively. By the year 2000, maize seed is projected to have
reached 25.62 thousand metric tons. This increase in the demand for maize seed would
result from the increase in the area harvested of maize during that period.

The total demand for maize is projected to reach 1630.33 thousand metric tons in
1991, increase at annual average gowth rates of 2.93% and 3.66% for the periods 1991-1995

and 1996-2000 respectively, to reach 2108.21 thousand metric tons in the year 2000. The

135
projected gowth rate in total demand for maize is lower than the one projected for

domestic maize production. However, the projected domestic maize production is not
sufﬁcient to meet total demand for maize implying that maize will have to be imported to

meet that excess demand throughout the forecast period.

5.113 W
Table 5.5 shows the projected imports of maize that will be required to supplement

domestic production if demand for maize has to be met. Maize imports are projected to
increase from 188.59 thousand metric tons in 1991 to 342.73 thousand metric tons by the
year 2000. Maize import are projected to increase at an average annual gowth rate of
1.76% between 1991 and the year 2000.

If no appropriate measures are taken, the lowest maize imports are projected to be
101.85 thousand metric tons in 1992 while the largest maize imports (188.59 thousand metric
tons) will be in the year 2000. Importing such huge amounts of maize would entail a serious
drain of the already scarce Zambian foreign currencies and hence a big loss on the country’s
capacity to import the most needed technologies. In addition, considering that the aim of
the government has been to achieve self-sufﬁciency in maize and food production in general,
the question that arises is whether the projected imports of maize can be avoided or

There are two possible ways of at least reducing, if not avoiding, maize imports in
Zambia. First, maize imports can be reduced by increasing maize production. Increased
maize production at a level even higher than the projected one can result from an expansion
in the area harvested of maize and/ or an improvement in the maize yield. Increasing maize

production by increasing the area under maize is possible since there is a lot of arable land

136

Table 5.5

Zambia: Projection of Required Maize Imports, 1991-2000

 

 

Thousand Metric Tons
Year Production Waste Net Domestic Total Required
Supply Demand Imports

1990 994.35 65.00 929.35 1600.28 670.93
1991 1512.74 71.00 1441.74 1630.33 188.59
1992 1639.24 68.83 1570.41 1672.26 101.85
1993 1672.24 66.08 1606.17 1733.43 127.26
1994 1728.63 63.32 1665.31 1790.24 124.92
1995 1769.81 60.56 1709.25 1849.11 139.86
1996 1806.24 72.77 1733.47 1929.64 196.18
1997 1841.66 72.77 1768.89 2008.01 239.13
1998 1868.11 72.77 1795.34 2063.64 268.30
1999 1907.60 72.77 1834.83 2135.84 301.01
2000 1943.23 72.77 1870.46 2213.19 342.73
MW

1991-1995 13.65 2.93 46.56
1996-2000 1.89 3.66 20.08
1991-2000 7.77 3.30 1.76

 

137
that is unutilized and under-utilized. For instance, Jansen (1988) reported that

reasonably potential arable land was estimated at 9 million hectares while the study done
by the Republic of Zambia (1990) showed that only about 1.27 million hectares of land was
planted to major crops in the whole country in the 1988/89 season.

Moreover, this study has shown that farmers are quite responsive to both the
producer price and the goss margin of maize. Therefore, if more area has to be planted
with maize to increase maize production, policies may be required that would increase the
maize producer price and / or decrease the cost of maize production so as to motivate
farmers. Moreover, other constraints that are prevalent in the maize subsector should be
addressed such as lack of credit, inadequate infrastrustural facilities such as roads, storage
facilities, etc. There is also need to introduce competition in input and maize marketing and
in the processing and distribution subsector.

The second way in which maize production can be increased is through the
improvement in maize yield. Even in this area, there is a lot of potential for improvement.
For instance, the average maize yield on commercial farms is about 3.5 metric tons per
hectare while it is only about 1.5 metric tons per hectare on small-scale farms. The big
difference in maize yield between small-scale and commercial farmers is mainly as a result
of limited of accessibility to inputs such as commercial fertilizer and hybrid maize seed and
the inappropriate application rates of the same inputs by the former.

The study by the Republic of Zambia (1990) reveals that most of the agricultural
extension services are available to commercial farmers who reside along the railway line
while small-scale farmers receive almost no extension services at all. Since about 70% of
maize is produced by small-scale and emergent (medium-scale) farmers, it is possible that

if more and adequate credit facilities were made available to ease accessibility to inputs and

138

extension services provided to small-scale farmers, then maize yield would improve even
further, thereby increasing maize production. For instance, the Republic of Zambia (1990),
indicated that if national average fertilizer application rates were to increase to those levels
employed by large-scale commercial farmers, then there would be

- a three-fold increase in the total quantity of commercial fertilizer used on maize

- an increase in the national average yield of maize to about 55 bags per hectare

(assuming appropriate seed and cultural practices are used), and

1: (Sigubling in the national production of maize from the same land area cultivated.
Although higher real fertilizer prices may reduce this potential, they would not eliminate it.
It should be pointed out that although it is possible to increase maize production both by
increasing the area harvested of maize and by improving the maize yield, policies that would
be introduced for such a purpose would require detailed analysis of their impact on all
parties involved such as farmers, marketing agencies, processing agencies and consumers.

The second possible method of reducing or getting rid of maize imports would be
to decrease the total demand of maize, although this may not be as easy as increasing maize
production. The most effective way to decrease total maize demand would be to decrease
the domestic consumption of maize which accounts for about 90% of total maize demand.

Decreasing domestic maize consumption would entail a reduction in per capita
domestic consumption of maize. This may partly come about if per capita consumption of
other cereal crops is increased. This would further entail putting in place those policies that
would encourage that. For instance, the increase in demand of maize for domestic
consumption has partly resulted from the price structure that exists for food products
(Republic of Zambia, 1990). While all food products had their prices decontrolled in the

19805, those of maize were not. This has left maize to be the cheapest and most affordable

staple food crop as the real per capita GDP continued to decrease. Therefore, if imports

139
of maize have to be reduced by decreasing maize demand, consumption of staple food crops

other than maize has to be encouraged partly by removing price controls on maize and
encouraging the production of such crops as sorghunn, millet and cassava.

However, it should be pointed out that even if maize imports were to be got rid of
by increasing maize production and/or decreasing the total demand of maize, that would
only ensure adequate availability of maize at national level. Availability alone is not a
sufﬁcient criteria upon which to assess the nutritional state of individual households. It is
possible to have a coexistence of food self-suﬁ'iciency and food insecurity. Therefore,
policies should go beyond targeting only food self-sufﬁciency if each household is to be food

secure.

5.11.4 01mm

The projected real producer price of maize, the real retail price of maize, and the
real maize marketing spread, together with their gowth rates, are presented in Table 5.6.
Over the whole forecast period, the real producer price of maize is predicted to increase at
an annual average rate of 1.07%.

The overall average annual gowth rate of the real maize retail price is predicted at
2.06% over the whole forecast period. This would entail an increase in the real maize retail
price from K6.42 per metric ton in 1991 to K7.66 per metric ton in 2000. It is predicted that
a larger part of the increase in the real retail price of maize would take place in the ﬁrst
ﬁve years (average of 3.39%). This is compared to an average annual gowth rate of only

0.73% in the real retail price of maize in the last ﬁve years of the forecast period.

140

Table 5.6

Zambia: Projections of the Maize Producer Price, Maize Mealie Meal Retail Price

and Maize Marketing Spread, 1991-2000

 

 

 

Producer Price Retail Price Marketing Spread

Year K/MT % Growth K/MT % Growth K/MT % Growth
Rate Rate Rate

1990 2.68 6.26 3.58
1991 3.20 19.36 6.42 2.68 3.22 -9.83
1992 2.80 -12.40 6.67 3.88 3.87 20.03
1993 2.87 2.36 6.81 2.01 3.94 1.75
1994 2.90 1.05 7.21 5.92 4.31 9.48
1995 3.36 15.78 7.39 2.45 4.03 -6.52
1996 3.28 -2.39 7.50 1.50 4.22 4.74
1997 3.26 -0.56 7.57 0.92 4.31 2.06
1998 3.08 «5.53 7.63 0.85 4.55 5.67
1999 2.87 -6.69 7.65 0.23 4.78 4.92
2000 2.86 -0.31 7.66 0.13 4.79 0.40
Annual Average Percentage Growth Rates .
1991-1995 5.23 3.39 2.98
1996-2000 -3.09 0.73 3.56
1991-2000 1.07 2.06 3.27

 

141

This pattern of gowth rates in prices of maize is partly expected taking into account
the policies that the new government is trying to implement such as the gadual removal of
the maize-related subsidies, decreasing the inﬂation rate, and increasing the maize producer
price. This direction has already been evidenced by the real retail price of maize in 1990
and 1991 when some maize marketing subsidies and all the maize consumption subsidy in
form of coupons were removed. When all the maize subsidies have been removed, it is
predicted that the real retail price of maize would tend to approach the econonnic price and
somehow level off. Moreover, with another policy in effect to curb the inﬂation rate
(estimated at about 200% in 1991), it is further predicted that the gowth rate in the
producer price and retail price of maize would be relatively stable.

The maize marketing spread is projected to grow at a faster rate during the last half
of the forecast period. This is in line with the argument presented above since it is
predicted that the maize marketing subsidies that reduce the marketing spread would have

been eliminated or tremendously reduced by the second half of the forecast period.

5.12 Summary

The chapter presents the results of the estimated model of the maize market in
Zambia. Each regessed equation is subjected to statistical and economic tests before it is
selected for analysis. The statistically and economically sound model is then used to ﬁrst
forecast prices, production and consumption of maize over the period under observation.
Various methods, including the MSEP, RMSPE and Theil’s "U" statistic and the graphical

method, are used to test the model’s ability to forecast.

142
When satisﬁed with the model’s performance over the period of observation, the

model is used to forecast future values of the endogenous variables. This entails ﬁrst
forecasting the unknown values of the exogenous variables into the future and second using
these values to forecast the future values of the endogenous variables. Whenever the actual
value of the exogenous variable is known it is incorporated a priori in forecasting the
endogenous variables. Average gowth rates of the predicted values of the endogenous
variables are calculated and analyzed. These are broken down into three periods: the ﬁrst

ﬁve years, the last ﬁve years, and the whole ten-year period.

6. SUMMARY, CONCLUSION, POLICY RECOMMENDATIONS AND
AREAS OF FUTURE RESEARCH

6.1§.ummanz

The main purpose of the study is to specify and estimate an econometric model of
the maize subsector in Zambia, calculate supply and demand elasticities, and then forecast
prices, production and consumption of maize after testing and validating the model.

In chapter tw0, the Zambia’s economy and sectoral performance is reviewed since
independence in 1964, along with the priorities embodied in the different policies that have
been employed by the government over the years and how these have affected the different
sectors of the economy, particularly the agicultural sector.

A descriptive, diagnostic and gaphical analysis of the major institutional, policy,
environmental and economic factors and relations that characterize the agicultural sector,
and the maize subsector in particular, is presented in chapter three. This descriptive,
diagnostic and graphical analysis of the maize subsector is done with regard to pricing,
production and marketing policies and consumption patterns. It shows how the Zambian
maize subsector’s performance is a function of the institutional and policy environment
within which it operates. Information in chapters two and three provides a backgound for
the economic formulation of the supply, demand and price behavioral equations for the
maize econometric model in the later sections.

In chapter four, the supply and demand theories and the stages involved in
commodity model building are discussed. In addition, the initial model is speciﬁed and the
estimator selected for model analysis.

The results of the econometric model are presented and the model is tested and

validated in chapter ﬁve. On the supply side of the model, three equations are estimated.

143

144
The ﬁrst equation is that of the average maize yield which is speciﬁed to. be related to

fertilizer consumption index, average rainfall and the time variable.

The second equation is that of area harvested of maize whose variations are
explained by the maize goss margin, the producer price of goundnuts, and lagged area
harvested of maize. This estimation captures the price sensitivity of producers to changes
in relative prices of substitute crops and maize goss margin sensitivity of producers. In the
third equation, net imports of maize are estimated as a function of maize production, lagged
change in stocks of maize and time trend. The econometric results are satisfactory and
provide short-run and long-run cross-price elasticities of supply of -0.21 and -0.42,
respectively, with respect to the producer price of goundnuts. The short-run and long-run
elasticities of supply with respect to the maize goss margin are 0.19 and 0.37 while those
with respect to the price are 0.22 and 0.51, respectively.

On the demand side, separate equations are estimated for per capita demand of
maize for domestic consumption, per capita demand of maize for feed, per capita demand
of maize for industrial processing and demand for maize seed. Per capita demand of maize
for domestic consumption is estimated as a function of lagged domestic maize consumption,
per capita sorghum production, the ratio of the retail price of maize to that of rice, and the
time trend. In the per capita demand of maize for feed and for industrial processing
equations, the regressors are the respective lagged dependent variable and per capita gross
domestic product. The demand for maize seed is explained by the area harvested of maize.

The short-run income elasticities of demand of maize for feed and per capita
demand of maize for industrial processing are 0.32 and 0.15, respectively, while their

respective long-run income elasticities of demand are 1.39 and 0.96. The short-run and

145
long-run price elasticity of demand fwgpﬁammﬁﬁnianmmd -

0.4, respectively.
/
The remaining part of the model is the formulation of the producer price and retail

-.-

price equations. The producer price of maize is speciﬁed as a function of lagged maize
producer price, fertilizer price, and lagged maize production. The lagged maize production
partly captures the effect of maize stocks on the producer price since the maize stocks were
not available, and when the change irn maize stocks was included in the equation it had a
t-statistic less than one. The price ﬂexibility with respect to lagged maize production is -
9.71. The retail price of maize is obtained as an identity by adding the maize marketing
spread to the maize producer price. The nominal maize marketing spread is estimated as
a function of inﬂation rate (using consumer price index as a proxy) and maize marketing
subsidy per metric ton. ~

The model equations are subjected to statistical and econonnic tests, and the model’s
ability to track actual data is validated using various methods: the gaphical method, Mean
Squared Error (MSE), Root Mean Squared Percentage Error (RMSPE) and Theil’s U-
statistic. In general, the model is good enough to be used for forecasting under all the tests.
In the last part of chapter ﬁve, the forecast results are presented and analyzed, with gowth

rates calculated for endogenous variables.

6-2 S‘mnclusicn
The study has generally met the research objectives outlined in chapter one. An
econometric model of the maize market is developed and highlights some important insights

into how the maize subsector operates in Zambia; how its performance is determined by the

146
institutions, policies, the environment and economic factors within which it operates; and
where the subsector nniglnt be headed to in the future.

The Zambian agricultural sector, and the food system in particular, has not had an
impressive performance despite the emphasis placed on it by the government starting as far
back as the late 19703. The country only achieved maize self-suﬁciency in the last two to
three years of the 19803. The country has been a net importer of maize and still imports
about 75% and 50% of its wheat and rice requirements respectively (Ndalamei, 1990).

The model predicts that between 1991 and 2000, the producer price and retail price
of maize would increase at average annual gowth rates of 1.07% and 2.06%, respectively.
The retail price of maize would increase at a faster rate than that of producer price partly
because of the gadual removal of the maize marketing subsidies that is in effect. The
predictions of the model also show that the area harvested of maize would increase at an
average annual rate of 0.58% compared to the average annual gowth rate of 6.04% in
maize yield over the whole forecast period. This would result in an average annual gowth
rate of 7.77% in maize production. The low gowth rate in the area harvested of maize
would partly be due to the low real producer price of maize which has been decreasing since
the late 19803. This led to a reduction in maize production by commercial farmers during
that same period as they switched to the production of other crops.

This study has projected a continuation of the national maize deﬁcit up to the year
2000, if no preventative measures are taken. It predicts that total maize demand will grow
at an average rate of 3.29%. This would mairnly result from an increase in domestic maize
consumption which accounts for about 90% of total maize demand. Domestic maize
consumption is projected to increase at an annual rate of 3.04% over the forecast period

due to both an increase in total population and in per capita maize consumption. Although

147

demand of maize for feed is projected to increase at an annual rate of 10.01%, it will not
inﬂuence total maize demand mucln since it only accounts for about 4% of the latter.
Demand of maize for seed and for industrial processing is projected to grow at average
annual rates of 2.01% and -0.91%, respectively, over the whole forecast period.

The conclusion emerging from this study is that although maize production is
projected to increase at a faster rate than that of total maize demand, the resulting increase
in maize production will be insufficient to meet maize demand. As such, it is predicted that
maize imports will gow at an average annual rate of 1.76% to meet the excess maize
demand if no preventative measures are taken to hedge against these projected maize
imports.

Considering the need for the government to efﬁciently use the already scarce foreign
currency and its interest to achieve food self-sufﬁciency as well as food security, there is a
need to make the food system more efﬁcient. There is a need to reduce maize imports, if
not elimirnate them. It is shown in the previous chapter that with supportive policies in
place, there is a good potential to reduce or eliminate maize imports by increasing maize
production (by expanding area under maize and/or by improving maize yield) and/or
reducing total maize demand by encouraging the production of other staple food crops such
as sorglnum or by decontrolling the maize retail price too as is the case with other staple

food crops like rice and wheat.

63 22026322200020230203
To ensure availability of adequate and accessible supplies of food products to the
population at large, production, marketing and distribution of food should be efficient.

There is need to increase labor productivity in agiculture and to promote and encourage

148

the production of crops well suited to the different ago-ecological zones. This means that
some resources should be shifted from maize to more suitable crops in regions that are not
suitable for maize. Increased production of food mops other than maize would even reduce
maize demand and hence the projected maize imports. At the same time, more maize
production should be encouraged in those areas where it is suitable.

However, increased maize production, increased production of other food crops and
the reduction in maize demand meant to reduce or eliminate maize import whilst aiming
to achieve food self-sufﬁciency and food security in the country, can only come about if
policies that would support such moves and motivate all parties concerned are put in place.

The problems constraining production, marketing and distribution in the food system
which need to be addressed include:

- shortage of labor at peak farnning periods coupled with non-use of labor-saving

technologies among small-scale farmers;

- inadequate extension services for small-scale farmers in maize production and
inadequate extension services and research for traditional crops gown by the .
majority of the poor rural households;

- insufficient seasonal and linnited availability of medium and long-term credit;

- high cost of farm inputs and implements;

- linnited use of irrigation in crop production to hedge against drought;

- late payment for crops purchased by the cooperative unions;

- late delivery of agricultural inputs;

- scarcity of foreign exchange for the importation of ago-chemicals, agicultural
implements, machinery and spare parts;

- inadequate storage facilities for agricultural inputs and output;

149
- limited availability of funds for crop purchases;

- inadequate serviceable vehicles, locomotives, wagons and spare parts compounded
by an inappropriate pricing regime for freigint transport;

- poor conditions of the basic irnfrastructure such as roads (feeder roads) and rail
lines in areas with potential land for agicultural development;

- lack of skilled personnel;

- undue government involvement in agicultural marketing and processing;

- capacity under-utilization of milling plants;

 

- declining real income; and ‘r
- non-availability of retail outlets in some rural areas.
Furthermore, the agicultural distribution system needs to be restructured. For

instance, more competition is required in fertilizer, seed, and maize marketing and in the

processing and retailing sectors.

6.4 W

Some areas have been identiﬁed for improving the model. First, future work should
collect and incorporate direct variables that inﬂuence demand of maize for feed and demand
of maize for industrial processing so as to formulate their direct demand functions, rather
than formulating their derived demand is the case in this study. For instance, the number
of livestock that use maize feed and the price of feed can be used in the maize feed
equation. The quantity produced of the ﬁnal product, its price, and the price of the maize
for industrial processing can be incorporated in the equation of maize demand for industrial

processing.

150

Second, collection of data on and inclusion of maize stocks would improve the
prediction, particularly on the maize imports. If in actual fact there were huge maize stocks
at the end of the previous year, then the current total maize supply is under-estimated since
that amount of maize stocks is not accounted for as beginning stocks if maize stocks are
missing.

Third, only fertilizer cost was considered in calculating the maize goss margin used
in this study. Therefore, for the maize gross margin to be more representative of the
expected proﬁt accrued to farmers per hectare, there is need to incorporate other maize
costs of production such as labor costs, maize seed costs and costs of herbicides and
insecticides.

Fourth, this study was carried out using annual time series secondary data
aggegated at the national level from 1964 to 1990. The information 1033 due to the
aggegation bias is obvious and implies that one should be cautious in interpreting some of
the results. Therefore, there is a need to investigate factors aﬁecting supply and demand
of maize and other crops at a more disaggegated level. For instance, separate equations
can be estimated for commercial farmers and for small-scale farmers to determine their
different supply responsiveness. This would help come up with more speciﬁc, targeting and
appropriate recommendations for the different categories of farmers.

Disaggregation can also be done at various levels, such as provincial level, regional
level, rural versus urban level, and at different income levels within and between rural and
urban areas. This would help determine the consumption and production patterns at each
of these levels separately. This would in turn help implement appropriate policies at

different levels and in different regions.

15 1
Fifth, research is needed to come up with possible institutional changes that should-

be made to improve the maize subsector in terms of improved production, transportation,
storage for inputs and output, marketing, production credit, input procurement and delivery,
and pricing policies. This is crucial in the case of the maize subsector in Zambia where
currently all the marketing of maize is done by the cooperatives which are inefﬁcient and
result in heavy subsidies by the government. For instance, even with the realization and
authorization. in 1990 that the private sector should be participating in maize marketing,
private organizations have not shown any interest to do so yet. Therefore, investigations
into reasons why the private sector is not attracted into that venture should be done and
corrections made.

Sixth, another area of research would be establishing how accurate information can
be collected and made available to different users at the time it is needed in the maize
subsector to improve its performance. This is crucial for policy purposes as well as for
planning by different organizations affected by the subsector. For instance, with the
decontrol of producer and retail prices for food crops other than maize in the 19703 and in
the 19803 and the maize producer price in 1991, different prices charged in different
markets should be collected and communicated to the users to promote competition.

Seven, one of the major concerns of the Zambian government is to achieve food self-
sufﬁciency and food security. Therefore, an area of research would be to deterrrnine the
least-cost, size and location of gain warehouses and transportation which would enhance
the efﬁciency of gain handling and distribution of maize (and other foods) within and
between surplus and deﬁcit regions.

Last but not the least, with the removal of government support of maize production

and marketing in form of subsidies and considering the drought that has been prevalent

152

since the 19803, there is a need for research to consider the feasibility of strategies designed
to promote production of other staple food crops that are more drought resistant than
maize, like sorglnum and millet, for sale as well as for increased consumption. Basic
constraints to irnnproving sorghum, millet and cassava production require socio-economic and
agonomic investigation. These drouglnt resistant crops are not perceived as competitive
with maize in production. They are grown for their drought resistance while consumers
strongly prefer maize. As a result, maize receives the geatest proportion of inputs and
attention. There is need to examine the elasticity of rural and urban consumer preferences

for alternative coarse gains.

APPENDICES

APPENDIX A

Figures for Chapter 5

Figure 5.1 NationalAverageYleldofMaizein Zambia: 1966-1990andPrqjectedtotheYear
2000

3.0

 

lT/HA

 

ACTUAL

 

 

 

 

0'5 Tss'srﬁrﬁﬁvv'svi

1970 1975 19130 1583 '
was

I 1 U 1 U U U I I

1990 1595 2000

153

 

154

Figure 5.2 Area Harvested ofMaize in Zambia: 1966-1990 and Projected to the Year 2000

10!” HT

10!!) HT

 

1200

6004

500-

 

—
,o—v
a

 

400

 

io'vt') '

srﬁrsss
1975

1990 1995

 

 

éooo

 

2000

1750-

1500-

1250~

1000-

750-

  

l

  

rrrr

 

 

III'I'

1970

I

'10'75 ' ' '

[1711]

199

 

2000

155

FigureSANetMaizeImportsinZa-bla: 1966-1990andProjectedtotbeYear2000
400

 

 

 

 

 

 

-300
I I T I I I I I I ii I T Y T I I I I I I I I U I“ I I I I I I I

197 19 5 1980 1985 1990 1995 éOOO
YEARS

Figure 5.5 Net Maize Supply in Zambia: 1966-1990 and Projected to the Year 2000

 

2000«

1750-

1500~

 

101” HT

1250-

10004

 

750-

l
.lroascas'r
surfstfrrrﬁfffsrrstsss

197 1975 1980 19 5 1990 1995 2000

 

 

 

 

III‘I'T'

YEARS

156

Figure 5.6 PerCapltaDomesticConsumptionofMaizeinZambia: 19664990de
to theYear2000

0.21

 

0.20 -
0.19 -
0.18 4
0.17 ..
0.10 -

METRIC TONS

0.15 4
0.144
0.13 .

 

 

0.12

 

 

 

s s‘wr I r s s s 1 r If T 7r 1' s s 1 s g ‘1’ 1’“r’ﬁr I if s ﬁr I s v ‘s‘1

197 197 1980 1985 1990 1995 2000
YEARS

Figure 5.7 Total Domestic Consumption of Maize in Zambia: 1966-1990 and projected to
the Year 2000

 

2250
’
II’I'
"
1750 '1 ”’
’I
a”
I

,500 q ,,,,, I'roaecas'r

1250-

1000 HT

1000 ~

750‘

 

 

 

 

 

I I I I I 1 I I I I I I I I I I I I I I I I I I I I I I I I I

1670 1975 1.980 19 5 1990 1995 2000
YEARS

157

FigumSSPerCapihDemaudofMaiseforFeedanambia: 1966-1990andProjectedto
theYear2000

0.0225

 

0.02004

0.0175 2

0.0150-

METRIC TONS

000185 "

0.0100-

 

 

 

 

0.0075 .......,
1070

 

1676‘ '1330' ' ' '1sas"' ‘ 1990' ' ”loser" 'Eooo
YEARS

Figure 5.9TotalDemandofMaizeforFeedinZambia: l966-l990andProjectedtotheYear
2000

 

175

1
too - - .
125 .

100~ , .

1000 111'
P
2
C:
>
I"

75 . ~~~~~~
‘\
‘

    

50:

a ’-J
l/Ponecas'r

    

 

 

 

25 I I I I I j I I I I I I I I I I I I I I I I I I I I I I I I I I

19 0 1975 1980 1985 1990 1995 2000
YEARS

 

157

FImreSSPerCapltaDe-andofMaheforFeedananbh: l966-l990andProjectedto
theYearzooo

0.0225

 

0.0200:

0.0175 -

0.0150-

II'I'RIC 70118

000125 "

0.0100-

 

 

00075 f1.,......1.w..vr.,vv.. Ytrﬁv.v.
1970 1975 1980 1985 1990 1995 2000

YEARS

 

 

 

Figure 5.9 Total Demand of Maine for Feed in Zambia: 1966-1990 and Projected to the Year
2000

 

175 J
150 4 4 ’0
1.00 . - i'

8 ,:

O a

d

75-

50-
y

 

 

 

 

25 I I T I I I 1 I I I I I I I I II I I I I T I I I I I I I I I I

1570 197 1980 1985 1990 1995 2000
YEARS

 

158

Figure 5.10 Per Capita Demand o! Maine [or Industrial Processing in Zambia: 1966-1990
andProjectedtotheYearzooo

Figure 5.11 Total Demand of Maine for Industrial Processing in Zambia: 19615-1990 and
Projected to the Year 2000

 

1000 HT

 

 

 

 

22.5 ITT I IjI I I I I

 

1606 1633' ' '1' H '1693‘ 'éooo
YEARS

16??) T 1575

159

FigureSJZMaizeSeedDemandinZambia: l966-l990andProiectedtotheYear2000
35

 

        
 

30 .
“ FORECAST

O

251

0.0-'0'“...

1000 HT

15-

 

 

10

 

I I I I

1970 19'79' 1990 1992's ' ' 1990

 

VIII

19'95 2000

2250

 

zoooq '

1750 -

1500 4

1250 q

1000 HT

1000 .

750.

 

 

I I I

 

IIIIIIIIITIIIIIIIII

1570 1975 1990 19 5 1996 ' ' '19'9‘ ' ' éooo
YEARS

 

 

160

Figure 5.14 Change in Maize Stocks In Zambia: 19664990 and Projected to theYearZOOO

750

“Am/HT

 

 

 

II vTv—r1v1

 

 

 

'19'71')"

1975"

19.801 ' '19'85 ' 1990 20

 

 

PORE‘CA ST

I

I
\

' |

1

   

ACT UAL

     

 

 

19'71') '

' '19'75 '

 

' '19'81') ' ' '19'95 ' ' 19931 11595 Y ' éooo
YEARS

 

     

161
Figure 5.16 Real Maize Retail Price in Zambia: 1966-1990, and Projected to the Year 2000
8
71 XI",
ACTUAL
a- .
5"
\
é 5-
a4
5

 

-av
'-
"__,...o
.9..-

'19'75"19'96"

 

 

1685' '1996' ' 1995' 'Tjjzo
YEARS

 

WW9!
XABIABLE 2281912
TMZY 1964-1979

1980-1990
TMZARFA 1964-1979
1930.1990
TMZPROD 1964-1979
19801990
NMZIMP 1964-1979
1980-1990
TMZC 1964-1990
MZFEED 1964-1990
MZPROC 1964-1990
MZSEED 1964-1990
MZWASTE 1964-1990
MZPP 1964-1969
1970-1989
1990
MZMRP 1964-1980

APPENDIX B

SQLZBCE

FAO "Agostat" data
Zambia, Agicultural Statistics Bulletin, various issues.

FAO "Agostat" data
Zambia, Agicultural Statisties Bulletin, various issues.

FAO "Agostat" data
Zambia, Agicultural Statistics Bulletin, various issues.

FAO "Agostat" data
Zambia, Whigs
MW Lusaka. 1990.

FAO "Agostat" data

. FAO "Agostat" data

FAO "Agostat" data

IFAO "Agostat" data

FAO "Agostat" data

Zambia, Ministry of Cooperative, Department of
Marketing, Market Logisties and Information section,
Lusaka.

Zambia, Ministry of Agiculture, Agicultural Statistics
Bulletin, Lusaka, 1988.

Zambia, Central Statisties Office, W
1991, Lusaka.

Zambia, 0 ' tatis '
Central Statisties Office, Lusaka.

various issues,

162

163

W BEBIQD scum:
1981-1989 Zambia. Wm 2010121019.
mm Dmka. 1990
1990 21111thme December. 1990.
Central Statistics Office, Lusaka.
GDP 1964-1979 Zambia, MW various issues,
Central Statistics Ofﬁce, Lusaka.
1980-1983 Zambia, 11W 3011911111.;
1987, Central Statistics Ofﬁce, Lusaka.
1984-1988 Zambia, J anuary-August,
1990, Central Statistics Office, Lusaka.
1989-1991 Zambia, W Central Statistics Office,
Lusaka, 1991.
RICERP 1964-1970 Zambia, W various issues,
Central Statistics Office, Lusaka.
1971-1985 Jansen D W101
W10. 1988
BRP 1964-1970 Zambia, MW various issues,
and Central Statistics Ofﬁce, Lusaka.
1986-1990
1971-1985 Jansen, D W
W 1988
GNUTPP 1964 Zambia, Ministry of Cooperative, Department of
Marketing, Market Logistics and Information section,
Lusaka.
1965-1969 Jansen, D.,
W 1988
1970-1989 Zambia, Ministry of Agiculture, Agicultural Statistics
Bulletin, Lusaka, 1988.
1990-1991 Zambia, W Central Statistics Office,

Lusaka, 1991.

164

YABIABLE 2281912 3911893
FERTAP 1964-1970 Zambia, Ministry of Cooperative, Department of
and Marketing, Market Logistics and Information section,
1990-1991 Lusaka.
1971-1989 Zambia, Ministry ongiculture. Agricultural Statistics
Bulletin, Lusaka, 1988.
CPI 1964-1989 Zambia, W various issues,
Central Statistics Oﬁice, Lusaka.
1990-1991 Zambia. W Fourth Quarter.
Central Statistics Oﬂ'ice, Lusaka, 1991.
TPOP 1964-1990 Zambia, MW various issues,
Central Statistics Office, Lusaka.
RAIN 1964-1990 Zambia, Wm various issues, and
Mount Makulu Research Station, Lusaka.
FCI 1964-1977 Calculated from the FAO Fertilizer Yearbook, various
and issues.
1988-1990
1978-1987 Calculated from data on fertilizer consumption in the
Agicultural Statistics Bulletin, Lusaka, 1988.
SGY 1964-1990 FAO "Agostat" data
SGAREA 1964-1990 FAO "Agostat" data
SGPROD 1964-1990 FAO "Agostat" data
SGPP 1964-1974 FAO "Agostat" data
1975-1990 Zambia, Ministry of Cooperative, Department of
Marketing, Market Logistics and Information section,
Lusaka.
MZSUBS 1964-1968 Zambia, MW December, 1971,
Central Statistics Ofﬁce, Lusaka.
1969-1979 Zambia, W various issues, Ministry of

Finance and National Commission for Development

Flaming Lusaka.

where:
TMZAREA

TMZY

TMZPROD
NMZIMP
MZFEED
MZSEED
MZWASTE
MZPROC

TMZC
TPOP
RAIN

. MZPP
MZMRP
FERTAP
FCI
RICERP
BRP
GDP
SGAREA
SGPROD
SGY
SGPP
CPI
GNUTPP
MZSUBS

165

I area harvested of maize

I average maize yield per hectare

I domestic maize production

I net import of maize

I maize feed

I maize seed

I maize wasted

I maize processed for industrial use
I human consumption of maize

I total population

I national average rainfall

I maize producer price

I maize breakfast meal retail price
I fertilizer average price for Urea and "D" Compound
I fertilizer consumption index

I Polished rice retail price

I bread retail price

I goss domestic product

I area harvested of sorghum

I sorghum production

I sorghum average yield per hectare
I sorghum Producer price

I Annual consumer price index

I chalimbana goundnut producer price
I maize subsidy

APPENDIX C

Calcuhtion of Fertilizer Consumption Index and
Fertiliza- cost per Hectare

Fertilizer consumption index (F CI) is calculated using FAO data on fertilizer
consumption in Zambia with 1978 I 1. That is, all fertilizer consumption series are divided
by the value for 1978 to come up with the fertilizer consumption index with 1978 as base
year. The FAO data are broken down into potash, nitrogen and phosphate fertilizers and
is available for the period 1964 to 1990.

1978 was selected as the base year because the same index is used in calculating
fertilizer cost per hectare for maize production which is in turn used in calculating the maize
goss margin, using the data from Zambia rather than using FAO data directly. Unlike
FAO data that is only broken down into potash, nitrogenous and phosphate fertilizers,
Zambian data is broken down into specific separate compounds and other types of
fertilizers.

The basal fertilizers used in maize production in Zambia are basically compounds "X",
"D" and "R" while the top dressing fertilizers usually used are Urea, Ammonium Nitrate and
Ammonium Sulphate. These are the fertilizers that are used in calculating the fertilizer cost
per hectare for maize production. Since Zambian data are only available for the period
1978 to 1987, the FAO data are adjusted to the Zambian data for 1978 to come up with
FAO Zambian-equivalent data for the period earlier than 1978. The fertilizer cost per
hectare is calculated using the following steps:

1 Fertilizer consumption index per hectare (FCI / ha) is calculated by dividing the FCI

(calculated as above with 1978 I 1) by total area harvested of maize for each year.

166

167
The fertilizer price index is calculated as the average of the price .of Urea (top

dressing) and the price of Compound "D" (basal dressing) divided by the 1978 ﬁgure.
Urea and "D" Compound prices are selected because these are the two commonly used
types of fertilizers among those mentioned above and their prices are quite
representative of the prices of other top dressing and basal fertilizers, respectively.
Fertilizer cost index per hectare is the product of the results in steps 1 and 2.

From 1978 to 1990, the total cost of fertilizer in maize production is calculated as a
sum of the cost of each type of fertilizer used in maize production. The cost of each
type of fertilizer used in maize production is calculated as the amount used multiplied
by its own price. To get the fertilizer cost per hectare of maize produced for the
period 1978-1987, the total cost of fertilizer is divided by the area harvested of maize.
To get the cost of fertilizer per hectare of maize for the periods 1964-1977 and 1988-
1990, the fertilizer cost index per hectare for these years (calculated in step 3) is

multiplied by the cost of fertilizer per hectare for 1978 calculated in step 4.

LIST OF REFERENCES

Asuming-Brempong, S., (1991), "Guidelines for the Analysis of Demand for Agicultural
Products", Training Manual Prepared for the Food and Agicultural Organization
(FAO) Regional Oﬂ'ice for Africa, Accra, Ghana.

Basachwa. MW (1991) W
W International Development Research Center, Ottawa, Ontario,
Canada.

Bhagavan. M R (1978) WWW
W Research Report No.44, Scandinavian Institute of African Studies,
Uppsala.

Bond, M., (1983), "Agicultural Responses to Prices in Sub-Saharan Africa", International
W No.30.

“PP" 1‘ ° and Semen 9.1111. (1986). W
9.2118111112119111 Department of Agicultural Economics, Virginia Polytechnic Institute

and State University, Blackburg, Virginia.

Chenoweth. F. and McKenzie. 1.. (1992). MW
Wm Lusaka. Zambia, 1992

Chileshe, J. H. (1986), I1 ' =_ - -
Publishing House Pvt Ltd., Bhibabad, U. P., India.

Clark, J., and Allison, C., (1989), W Oxford.

Cleary, Seamus, (1989), W A Catholic Fund for Overseas
Development (CAP OD) Report, London.

 

Dhrymes, PJ., Howrey, E. P., Hymans, S. H, Kmenta, J., Learner, E. E., Quandt, R.E.,
Ramsey, J. B, Shapiro, H. T. and Zarnowitz, V., (1972), "Criteria for Evaluation of

Econometric ModeIS" 101311001091 WWW Vol 1 PP 291-
324.

Dodge, Doris,J., (1977), . ~ ’ '9 2 ° ° ' '
311W Institute of International Studies, University of California,
Berkeley.

 

168

 

(1977), - '11 i ~ ' ' '
—1nggmes_in_nghia, A Dissertation submitted 1n partial fulﬁllment of the requirements
of Ph. D. in Economics, University of California, Berkeley.

Elliot, Charles, (ed) (1971). W Zambia
Nairobi: Oxford university Press.

 

Ghosh, SK, (1991), .
Cliffs, New Jersey.

 

~. Prentice Hall, Eaglewood

Grasa Aznar A.. (1989). MMWW Kluwer
Academic Publishers, Dordrecht, The Netherlands.

Hall, R.T., Johnston, J. and Lilien, D.M., (1990), WEI, Quantitative
Micro Software, Irvine, California.

Harvey. AC. (1990). W Second Edition, London.

11.0,,(1977) 101111 1' :o’.=111.ti. 'v‘vn :Histtu‘ 110011111
Zambia. A Report to the Government of Zambia by Jobs and Skills Progam for
Africa (JASPA) Advisory Mission, Addis Ababa 1977.

(1981). W Findings and
Recommendations of an ILO/JASPA Basic Needs Mission to Zambia, Addis Ababa,
Ethiopia.

Integated Rural Development Progam, (1987), 'The Integated Rural Development
Progam (IRDP) Serenje- Mpika-Chinsali: The Progam and its Effects after 5 Years",
No.10,Mpika, Zambia.

International Food Policy Research Institute, (1985), W
WW3. Lusaka: IP'PPJ/I‘TFI‘IC/RDSB/Ul‘izA

International Monetary Fund, (1991), W
Washington, DC.

Jansen, Doris, D., (1988), ,2 ‘
Zambia, Washington, DC.

J anvry, A., (1985), "Dilemmas and Options in the Formulation of Agicultural Policies in
Africa", A Paper prepared for the Economic Development Institute Seminar in
conjunction with German Foundation International for Development on Agicultural
Policy and its Relationship to Food Policy in Sub-Saharan Africa, Berlin, March 18-22.

 

Jayne, T.S., and Chisvo, M., (1991) "Zimbabwe’s Grain Marketing Policy Changes in the '
19908: Short Run Versus Long Run Options". A Paper presented at the Seventh Annual
Conference on Food Security in Southern Africa, October 28-30, 1991, Makasa Sun
Hotel, Victoria Falls, Zimbabwe.

170

Katongo, K., (1988), "Agicultural Sectoral Policies in Zambia, 1975 - 1985", A Paper
presented at the Seminar on Growth and Equity in Zambian Agiculture: an Eastern
Province Study, Livingstone, Zambia, December 8-11.

Kaunda. Kenneth. (1968). WWW
Lusaka: Government Printer.

_(1974).Humanism.in_Zam12u_and_LQuide_19_lts lmnlsmentaticnmn. Lusaka:
Government Printers.

(1968), MW Lusaka: Government Printer.

Kaunga, E..,C Kalyalya, Denny H., Mutaba Mwali, Hayashi Koji, (1983), W
W Joint Research Program Series.
No.34, Institute of Developing Economies, Tokyo, Japan, 1983.

Kennedy, P., (1985), W, Second Edition, MIT Press, Cambridge.

Kydd, J. G, (1986), "Changes in Zambian Agricultural Policy since 1983: Problems of
Liberalization and Organization" in W No.4: pp. 243-69.

(1987), "Coﬁ'ee after Copper? Structural Adjustment, Liberalization and Agriculture
in Zambia", a paper presented at ODI Conference on the Design and Impact of
Adjustment Progams on Agiculture and Agicultural Institutions held at Overseas
Development Institute Regent College, London, September 10-11.

._ .. (1988). ”umbia.’ in mm edited by Charles
Harvey, London: Macmillan Press.

and Scarborough, V., (1988), "Food Market Liberalization in Sub—Saharan Africa: A
Survey of Issues".

Labys, Walter C., (1988). "Recent Developments in Commodity Modelling: A World Bank
Focus". The World Bank PPR Working Papers Series # 119.

Maddala, 6.8., (1977), W McGraw Hill Inc., New York.
Malambo, L.M., (1987), "Rural Food Security in Zambia", a Dissertation presented to

Michigan State University for the completion of the Ph.D Progam in Agricultural
Economics Department, East Lansing.

Ndalamei, Likolo, (1989), "Pricing Policy of Agicultural produce in Zambia" ﬂocking
2mg No. 110, Uppsala.

(1990), "National Food System in Zambia", A Draft Chapter of the FAO Research
Project on Food Security in Zambia, Lusaka.

NiChOlSOII, W., (1989), u 1‘: '1'“ .
Edition, The Dryden Press, USA.

Palm. Franz 0. and Smit. Hidde 1’. (1991) Emncnu'cMmgllinunﬂolioLAnabsia Gower
Publishing Company, Vermont, USA.

Potluri, Rao and Miller, Roger LeRoy, (1971), W Belmont, California,
Woodwork Publishing Company, Inc.

RePublio of Zambia. (1988). W233. Statistics Section. Planning
Division, Ministry of Agiculture and Waster Development, Lusaka.

 

(1991), W Fourth Quarter, Vol. 36, Central Statistics Ofﬁce,
Lusaka.

' ' ' ° 9- Ministry of Finance,
and National Commission for Development Planning, Lusaka.

_W various issues, Ministry of Finance and National Commission for
Development Planning, Lusaka.

Lusaka.

 

__ (1990),
__ EjmngjaLBepmt, various issues, Ministry of Finance, Lusaka.

_ MW various issues, Meteorological Department, Lusaka.
__ (1988), W2, Central Statistics Ofﬁce, Lusaka.

_..(1989) '4 "'1'". s'v' ninth” t 11 :11 91-11111'01 1:1-

—1223_Qlumgsj_and_n, National Commission for Development Flaming Lusaka.

._.._ (1990). W310 Ministry of
Finance and NCDP., Lusaka.

_Qmamammsmmaum various issues, Statistics Section. Flaming
Division, Ministry of Agiculture and Waster Development, Lusaka.

(1970), Report of the Second National Convention on Rural Development, Incomes,
Wages, and Prices in Zambia: Policy and Machinery, held in Kitwe, 12-16 December,
1969, Lusaka: Government Printer.

(1984), BMW Consultative Group for Zambia, May 22-
24, 1984, Lusaka.

(1968) 1'11 " 1‘ .1” 1'! t‘ ' 1' 91111.1“ 11

—W Ministry of giculture and Water Development,
Lusaka.

(1971), , - ' '
of Development Planmng and National Guidance, Lusaka.

. (1990). W221. Ministry of Finance and NCDP., Lusaka.

.__. (1979). WW1. National Commission for
Development Planning, Lusaka.

 

(1989), I"Traditional Crops Promotion Study“, Ministry of Finance, Lusaka.

(1991), W Central Statistical Office, Lusaka.

Rubey, Lawrence, (1992), "The Maize Subsector in South Africa: Emerging Policy Issues”.
A report prepared for the Michigan State University Food Security Project, East
Lansing, Michigan.

Shonkwiler, J. S. and Taylor, T. G., (1984), The Implications of Estimating Market Demand

Curves by Least Squares Regression” in W
No. 11, pp. 107-18.

Siandwazi, C., (1990), ”Food Security at National Level and Household Level in Zambia",
A draft Research Paper for FAQ on Food Security in Zambia, Lusaka, Zambia.

(1988), "Nutrition and Agicultural Policy Issues. Current Evidence and Research

Agenda”, in Bemsten, H. and Mudimu, G. D. (ed.), WW
Wham. Harare Zimbabwe

Sowey, E. R. ,,(1973) "Stochastic Simulations 1n Macroeconometric Models: Methodology and
Interpretation" in NW- Eds by
AA. Powell and RA. Williams. Amsterdam: North-Holland.

Timmer, P.C., Falcon, W.P., and Pearson, SR, (1983) mm m Johns Hopkins
University Press, Baltimore.

Tomek, W.G, and Robinson, KL, (1990), W Third Edition,
Cornell University Press, USA.

Thurman, W. N., (1986), "Endogeneity Testing 1n a Supply and Demand Framework”,1n°
BMW No 68

'I'he Poultry Market: Demand Stability and Industry Structure", in Amgican
_Wm PP 30-37

Valdes, A. and Siamwala, A. (1981). ”Introduction,” in Valdes, A. (eds.,) M
Wes, Westview Press, Boulder, Colorado.

Waugh, F..,V (1964), c ‘ ° - ' .-
Economic and Statistical Analysis D1V1s1on /Eoonom1c Research Service, U. S.
Department of Agriculture, Technical Bulletin No. 13 16.

 

Wood, .00., (1984),

 

‘ ' .‘ ° - . 1 '5 Prepared for the Meeting of
Inter-African Research Working Groups on State Policies on Agiculture and Food
Production 1n Africa, 25-28 September, Addis Ababa, Ethiopia.

 

WOOd, A.P., eds., (1990), I1 .‘- .
State university Press/Ames, USA.

 

World Bank, (1984),
No. 4764-ZA.

 

"‘111111111111111“