VJ.

>17§03d Q"!

 

IIHIITIHWIEHIMIWWIWWIWflflfiiflfll

3 1293 00850 0237

 

MSU

LIBRARIES
-._——.

 

 

 

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

 

,WM
JUN 1 3 1994
2 3 7

 

 

 

ARABLE FARMING DEVELOPMENT PRIORITIES
IN THE CENTRAL AGRICULTURAL REGION, BOTSWANA:
A FARMING SYSTEMS ANALYSIS

By

Doyle Curtis Baker

A DISSERTATION

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

DOCTOR OF PHILOSOPHY

Department of Agricultural Economics

1987

Copyright by
DOYLE CURTIS BAKER

1987

'n-' .1;
...:.e L.

' ‘- mg...”
'e‘J‘: “\u3'i.
‘5
“6 ‘~a
V
'I- w.‘ .-
L:.l:.s 3‘ I

r- .
-oo.,.‘

“-=m-‘.g at.

UN".

9551‘s afilfir
H’

‘-I . .

k ‘b

ABSTRACT
ARABLE FARMING DEVELOPMENT PRIORITIES

IN THE CENTRAL AGRICULTURAL REGION, BOTSWANA:
A FARMING SYSTEMS ANALYSIS

By

Doyle Curtis Baker

There is an urgent need to identify ways to develop arable farming
in Botswana in order to reduce national dependence on food imports, save
foreign exchange, create rural employment, raise rural incomes, and
reduce household dependence on government subsidies.

The goal of this thesis is to provide a comprehensive, systems
analysis of arable farming in the Central Region, and the factors
affecting arable farming, in order to determine arable farming
deve10pment priorities. The analysis is based on village and on-farm
surveys and experiments carried out between October 1982 and June 1986.

The thesis uses the farming systems approach. The farming systems
approach: (a) is holistic not reductionist, (b) generally uses
households, farm systems or production subsystems as the unit of
analysis, and (c) follows a systems problem evaluation sequence. The
systems approach used in the thesis encompasses household circumstances,
traditional crop systems management, and local institutions, as well as
experimentation to identify improved practices.

The analysis of household circumstances describes the resource
constraints affecting arable farming development and reveals a more
pervasive pattern of inequality than is indicated in previous studies.
Recommendations are given on technical research priorities and targeting

strategies.

va‘fl

126331".th

an: can Di
The as;

servize, 1t;

Doyle Curtis Baker

The analysis of crOp systems management focuses on farmers'
perceptions and priorities, and the diversity in traditional practices.
Several potentially valuable technical investigations are identified.

An overview is given of the on-farm experiments carried out between
1982 and 1986. The trials program primarily focused on modifications in
the traditional broadcast, single plow system. A series of budget
analyses based on the on-farm trials shows there are modified practices
which can be profitably adepted even during drought conditions.

The analysis of agricultural support systems covers the extension
service, local traders, village groups, and two agricultural assistance
programs. Guidelines are given on appropriate institutional changes.

In the final chapter, a comprehensive strategy for arable farming
development is pr0posed and an assessment is given of the holistic

research approach used in the thesis.

 

Tze thesis
:he igrimitura
:a-s;c:scred a:
;::_‘e:: was in;

y
n

a ‘. .‘r ,-
Vj"~i" um ,HL.

nuts.

»n-n.. ' ’
.1:.L..tl-3n EC!

:f::A

.wjm to 186
tk.‘
“(L I earHEd.

‘-
Swag

State c

ACKNOWLEDGMENTS

The thesis is based on research carried out under the auspices of
the Agricultural Technology Improvement Project (ATIP). ATIP was
co-sponsored and funded by the Government of Botswana and USAID. The
project was implemented by the Mid-America International Agricultural
Consortium (MIAC). Kansas State University served as the lead
institution for MIAC. I gratefully acknowledge the financial and
logistical support of USAID, the Government of Botswana and Kansas State
University. I specially would like to thank Drs. Mark Johnson, John
Sjo, Vic Larsen and Jim Jorns at Kansas State University for their
support.

Several individuals associated with ATIP in Botswana made
invaluable contributions to the research. Dr. David Norman, ATIP Team
Leader, provided guidance and encouragement. Dr. Norman allowed me the
freedom to learn from my mistakes and showed more faith in my judgment
than I earned. I benefited greatly from discussions with all of my
Kansas State collegues, including Drs. Norman, J. Siebert, G. Heinrich,
F. Worman, D. Gray, C. Trent, B. Koch, J.A. Hobbs, and W. Miller.

I would particularly like to acknowledge my debt to Dr. Jay
Siebert, my Central Region team collegue. Dr. Siebert designed and
analyzed most of the on-farm trials carried out by the team. More
importanly, many of my ideas about arable farming develOpment resulted
from endless hours of discussion with Dr. Siebert.

ii

“he mean:
Sa’S'iéSl counter?

'-:‘; game, Jk 1

,av.

research. E. Hod
assume when n
1513: implezent
Sarah, and G.
satied by K. St

Prhap my 8

ts research. 2);-

:;::er, Glenn Jo:
i211: led to majg
:‘rlrzzupated on t

ilézts on the 1

33‘ crab'fOrd, E

Dr, Cravior:

to:

ake the t.

The research was possible only because of the efforts of my
Batswana counterparts, including at different times: Meschack Tjirongo,
Chada Tibone, John Lesotlho and Catherine Jonas. This thesis represents
their product as much as it does mine.

In addition to the above, many other ATIP staff contributed to the
research. E. Modiakgotla and J. Luzani, both team agronomists, provided
assistance when needed. The following field staff collected data and
helped implement trials: D. Dira, C. Mahilo, K. Okaile, R. Mosojane, R.
Serumola, and G. Mogotsi. Administrative assistance and data entry were
provided by K. Seleke, L. Seretse, and P. Monyane.

Perhaps my greatest debt is to the many farmers who participated in
the research. During four difficult seasons, they answered questions
and tried new practices withh little personal benefit.

At Michigan State, guidance committee members Eric Crawford, Carl
Eicher, Glenn Johnson and Carl Liedholm provided advice and criticism
which led to major improvements in the thesis. Rick Bernsten
participated on the dissertation committee and gave constructive
comments on the thesis organization. Special acknowledgments are due
Drs. Crawford, Eicher and Johnson.

Dr. Crawford assumed the responsibility of being major professor
and thesis supervisor when he could not afford the time. He worked with
‘me to make the thesis more focused, without losing its systems
orientation. I greatly appreciate his helpful comments.

Dr. Eicher was my major professor and mentor throughout my graduate
program at Michigan State. He took a personal interest in my career
deve10pment and provided me with innumerable Opportunities to learn and

grow.

iii

m2: pm
Eras a gra:
returzed to

I -'1.
F116;."

.no:..’, L_
flutsye tn:

Dr. Johnson is re3ponsible for the intellectual perspective of the
thesis (though not accountable for its weaknesses).

I also would like to thank Drs. Connor and Mandersheid for the
support provided by the Department of Agricultural Economics, both while
I was a graduate student and in the years since when I have periodically
returned to campus to finish the thesis.

Finally, I would like to thank my parents and wife, Kathy, for the
patience they have shown. Seven years is a long time to wait. I am

sure they wondered if the waiting would ever end.

iv

I. INTRJDICTIG

A. Backgroun
1. Profile
2. Role of

8. Problem 5

C. Research

3. Organiza:

. RESEARCH A?

5. Agricultu
B. Concepzua

L Definit
2. Systems
3. Farm 5;
5~ Problem
C- Field Res

1- Village
2- Surveys
3- EXperia
0' Data Anal
5. Distincrj
NOU‘Hoc
SUbjec‘
Farmer
SCOpe ‘

N P—‘
‘r r" . .
c

31- arrszaarn 1
A' Human Re
' HOUSeh.
° Main Ay
- Reside
' Land. Ca
' RESOur
' CORPOS
ab0r Al
HOUSeh
GendEr
SEaSOn
Fieidu

('7 (I!
. WM.—
)—4 N'—
hwy-
o.

TABLE OF CONTENTS

I. INTRODUCTION

A. Background
1. Profile of National DevelOpment
2. Role of Agriculture

B. Problem Statement

C. Research Objectives

D. Organization of Thesis

II. RESEARCH APPROACH

A. Agricultural Technology Improvement Project
B. Conceptual Framework
1. Definition of a "System"
2. Systems PersPective
3. Farm Systems Concept
4. Problem Evaluation Sequence
C. Field Research Activities
1. Village and Farmer Selection
2. Surveys and Studies
3. Experimentation
D. Data Analysis
E. Distinctive Features of the Research Approach
1. Non-Modernist
2. Subject Matter Orientation
3. Farmer First Planning Perspective
4. Sc0pe of Research

III. HOUSEHOLD CIRCUMSTANCES AFFECTING ARABLE FARMING

A. Human Resources
1. Household Composition
2. Main Activities
3. Residential Patterns
B. Land, Capital and Livestock Resources
1. Resource Profile, 1983
2. Composition and Value of Farm Assets, 1984
C. Labor Allocation
1. Household Labor Use
2. Gender Division of Labor
3. Seasonal Profile of Labor Use
4. Fieldwork Hours by Household Type

14
15
15
16
17
20
24
24
28
36
40
43
43
45
45
47

49

49
50
52
54
56
56
58
62
64
65
67
69

D. Cash Re
1. Casn
2. Casof

E. Food Co

Main

Sour:
F. Conclus

II. CROP SYSI

A. Croppir
B. CrOps :
1. Host
2. Desir
C. Draft 1
1. Rana!
2- Inte:
3- Iratti.
1. Ireo
2- Tizi'
E- CrOppi
1. Vari
2. Use
3- Per:
F- Post—5
1» Seed
Sto:
3- Cr0p
G. Perce;
Shos
2- Find
5' Hales
' coach
Cro;
Trac
Timi
Proc
P05:
Dec:

H

CRUPPIX;

A. Cropp:
19%;

1%;

195.

‘ Histo;
Expér:

PAGE

D. Cash Revenues and Expenditures 70
1. Cash Sources, 1983 70
2. Cashflow Analysis, 1984 71

E. Food Consumption 75
1. Main Dishes and Food Items 75
2. Food Consumption Frequencies 78
3. Sorghum and Maize Meal Consumption Levels 80
4. Sources of Food Items 81

F. Conclusions and Implications 82

IV. CROP SYSTEMS MANAGEMENT 87

A. Crapping Objectives 87

B. Craps and Varieties 88
1. Most Important Craps 88
2. Desired Varietal Characteristics for Sorghum 89

C. Draft Management 90
1. Management of Owned Traction 91
2. Inter-Household Draft Arrangements 92

D. Traction Use Patterns 95
1. Trends in Traction Use 95
2. Timing of Plowing 98

E. Crapping Practices 102
1. Variations in Traditional Practices 102
2. Use of Recommended Practices 106
3. Perceptions of Recommended Practices 108

F. Post-Harvest Practices 110
1. Seed Selection 110
2. Storage 111
3. Crap Residue Use 112

G. Perceptions of Problems 113
l. Shoshong and Makwate Farmers 113
2. Findings from Other Studies 115

H. Roles in Decision Making 118

I. Conclusions and Implications 121
l. Crops and Varieties 122
2. Traction Use and Access 123
3. Timing of Plowing 124
4. Production Practices 124
5. Post-Harvest Practices 126
6. Decision Behavior 127

V. CROPPING OUTCOMES AND THE RESULTS OF ON-FARM TRIALS 128

A. Crapping Outcomes 128
1. 1982-83 Season and "Normal" Seasons 128
2. 1983-84 Season 131
3. 1984-85 Season 133

B. Historical Perspective On Agricultural

Experimentation in Botswana 136

vi

C. Central
1. 1952-5
2. 1953-5

3. 1954-5

4. 1955-5

1. Budget A
L Dcubie
2. Early

3. Sole P

3. Hand P

D. Conclusi
1. Croppi
2. The Rt
3. Profit

”Y *“v-y-c q...
“' AUAALLBLCF

A. Agricult
1. Histori
2. Extensi

5. Trading
1. Profit]
2- Avail:

ltple:

3- Source
4- Busing
C. Village

PAGE

C. Central Region On-Farm Trials 141
1. 1982-83 Season 142
2. 1983-84 Season 143
3. 1984-85 Season 145
4. 1985-86 Season 147

D. Budget Analyses of Selected Practices 151
1. Double Plowing 151
2. Early Plowing Plus Row Planting 159
3. Sole Planting of Secondary Cr0ps 164
3. Hand Planting for Cap Filling 167

D. Conclusions and Recommendations 170
l. Crapping Outcomes 171
2. The Record of Experimentation 172
3. Profitable Production Practices for Drought 173

VI. AGRICULTURAL SUPPORT SYSTEMS AFFECTING ARABLE FARMING 175

A. Agricultural Extension 175
1. Historical Perspective of Extension in Botswana 176
2. Extension in the Central Region 178

B. Trading Network 187
1. Profile of Trading Establishments 188
2. Availability and Prices of Food Commodities,

Implements and Fencing Materials 191
3. Sources of Items Sold 193
4. Business Problems 195

C. Village Groups 196
1. Recent Research Findings 197
2. Groups in Shoshong and Makwate 198
3. Household and Individual Participation 199
4. Experiences in Farmer Groups 200

D. Participation in ALDEP and Drought Relief 202
1. Participation in ALDEP in 1983 203
2. Participation in Drought Relief and ALDEP in 1985 203
3. Farmer Assessment of Programs 204

E. Conclusions and Recommendations 206
1. Extension 206
2. Traders 208
3. Group Formation 210
4. ALDEP and Drought Relief 211

VII. SUMMARY AND IMPLICATIONS 213

A. Research Objectives and Approach 213

B. Summary of Findings 215

C. Implications for the Ministry of Agriculture 218
1. Shift to a Holistic, Systems Approach 218
2. Give Greater Emphasis to "Intermediate" Producers 219
3. Target Research, Extension and Planning Activities 221
4. Shift the Focus of Technical Research 222
5. Increase the Emphasis on Macro Policy Analysis 223

vii

D. Assess
Guidel

1. Mr

D. Assessment of Research Approach and
Guidelines for Future Research
1. Representative Villages
2. Small Samples
3. Holism

APPENDIX: PROFILE OF AGRICULTURE IN BOTSWANA

A. Technical Environment
1. Rainfall
2. Temperature and Evaporation
3. Soils
4. Vegetation

B. Agricultural Sector Overview
1. Sector Structure
3. Crapping Practices
4. Livestock Management and Husbandry

C. Agricultural Institutions
1. Ministry of Agriculture
2. Input Supply, Credit and Marketing Agencies
3. Village Government and Organizations

D. Summary

REFERENCES

viii

PAGE

224
224
225
227

230

230
230
231
232
232
233
233
234
237
238
238
240
242

243

Catalog of
‘ E
Catalog 0.

Household 1
HYPE Surve;
Principal 1
Cattle, Eq.
Value of P.-
Bousehold 1
Household 1
Household 1
Primary In.
Average Ho:
Farmer Rani
Parser Rani
Pood COhSu:

Irattion t
1113ng of '
159 Of Rec.
Farmers' 1:.
Storage Pr.
Use of Cow
“merS' a.
R0198 in 3.

Days of Pl.
{0‘ "Good

Normal Par:
Days of P1.
CIOP Enter
Partial Bu.
Partial Bu.
Partial Bu<
1983-84 311:
Investment
at PrESEn1
I358 Vail“
and SorghUr
SOle CIOpp

LIST OF TABLES

PAGE

2.1 Catalog of Surveys and Studies; October, 1982 to 1986 29
2.2 Catalog of On-Farm Experiments; October, 1982 to 1986 38
3.1 Household Structure, 1983-84 51
3.2 MVRU Survey Participants' Most Important Activities, 1984 53
3.3 Principal Residence by Period, 1984 55
3.4 Cattle, Equipment and Land Resources, 1983 57
3.5 Value of Farm Assets (Pula), 1983-84 61
3.6 Household Labor Use, 1983-84 64
3.7 Household Labor Use by Age-Gender Categories, 1983-84 66
3.8 Household Labor Use By Month, 1983-84 68
3.9 Primary Income Source, 1983 71
3.10 Average Monthly Cashflows, November, 1983 - June, 1984 (Pula) 72
3.11 Farmer Rankings of Main Dishes, 1983 76
3.12 Farmer Rankings of Other Main Foods, 1983 77
3.13 Food Consumption Frequencies, October 1984 to September 1985 79
4.1 Traction Used and Draft Access, 1982-83 and 1984-85 96
4.2 Timing of Plowing; 1982-83 and 1984-85 99
4.3 Use of Recommended CrOpping Practices, 1983 107
4.4 Farmers' Perceptions of Recommended Practices, 1983 109
4.5 Storage Practices, 1983 111
4.6 Use of Cowpea Residues, 1984 113
4.7 Farmers' Rankings of Resource Constraints, 1983 114
4.8 Roles in Decision Making, 1985 120
5.1 Days of Planting and Acres Planted; Recall Data

for "Good Seasons", 1982-83, and 1984-85 129
5.2 Normal Farm Production of Major Craps, 1983 130
5.3 Days of Plowing, Area Plowed, Sorghum Harvested, 1983-84 132
5.4 Crop Enterprise Budget, 1984-85 134
5.5 Partial Budget: RM, FI Double Plowing, 1983-84 and 1984-85 153
5.6 Partial Budget: 1985-86 FM,FI Double Plowing Trial 157
5.7 Partial Budget: Early Plow and Row Plant,

1983-84 and 1984-85 160
5.8 Investment in a Row Planter: Benefit-Cost Ratio,

Net Present Worth, Internal Rate of Return 163
5.9 Gross Values of Production for Secondary CrOps

and Sorghum, 1983-84 and 1984-85 166
5.10 Sole Crapping Trial: Enterprise Analysis, 1985-86 168

ix

fen

.-

p

b

.3
an

5

HJ

. nit

5

"J

I’d

VII.“

«.1 .
ob

_U
.5
Fl o

11¢

1.“

T 7..
1:83.

PAGE

Extension Area Characteristics by District, 1983 179
AD Supply and Use of Equipment, 1983 182
AD Assessments of the Most Valuable Crops; 1983 185
Changes Farmers Could Make to Increase Productivity, 1983 186
Number of Traders in the Central Region

By Location and Size of Village, 1985 188
Trading Activities, Central Region, 1985 190
Availability and Prices of Major Food Grains, 1985 192
Sales of Implements, Hand Tools and Fencing Materials, 1985 194
Business Problems, 1985 196
Farm Structure; Traditional and Freehold Sectors, 1983 234
Cropping Practices by Agricultural Region:

Traditional Sector, 1983 235

1.1 Pap of Bots

2.1 Harm Syst
22 Farming Sys

1.1

2
2

l
2

LIST OF FIGURES

Map of Botswana; Agricultural Regions and Districts

A Farm System Model
Farming Systems Problem Evaluation Sequence

xi

PAGE

19
21

‘1: Agricultural I
1...? , Arable 1.371
11., Animal Prod
£19,1ctelerate:
:15 Agricultural
P, lgri: ultura
51'}, Botswana {1;
5!: Botswana Plea
1.511, Internet 1
"C, Central Sta:
355, Department
3:, District Ag:
‘ De Jenartnent 0
31. “Livable Plow
-'-"1 3i“ision of
:v‘1513ryland Fa
T“£1510er "
“3. Evaluatio
-1 , Economic 51; r
13, Food and Ag:
‘1! Father Manage
1 Farmer Imple:
PL Emmi 5 st
:, Government c
.i " luteStated
1 Labor Base:
“'3' iliisuy o.
:31! “mistry 0‘
Vii-L) hitiple- -v.
:National De
1 Karim,“ F
’ 3351011211 Hi
Uplanaigg an
_ :;Rutal Dale 1-
’ Rum Into
)RESear Cher
1 when. ”rile
Ta
11,1, 33%
0V
w. laiteda at

1,1:

:73:

“f,"

u..." g

’l’

LIST OF ABBREVIATIONS

AD, Agricultural Demonstrator

ALDEP, Arable Lands Deve10pment Program

APRU, Animal Production Research Unit

ARAP, Accelerated Rainfed Arable Program

ARS, Agricultural Research Station

ATIP, Agricultural Technology Improvement Project
BAMB, Botswana Agricultural Marketing Board

BMC, Botswana Meat Commission

CIMMYT, International Maize and Wheat Improvement Center
CSO, Central Statistics Office

DAFS, Department of Agricultural Field Services
DAO, District Agriculture Officer

DAR, Department of Agricultural Research

DP, Double Plow

DPS, Division of Planning and Statistics

DLFRS, Dryland Farming Research Scheme

DUMI, Decision Unit Management Information
EFSAIP, Evaluation of Farming Systems and Implements Project
ESM, Economic Survey Mission

FAO, Food and Agriculture Organization

FM, Farmer Managed

FI, Farmer Implemented

FSR, Farming Systems Research

GOB, Government of Botswana

IFPP, Integrated Farming Pilot Project

LBDR, Labor-Based Drought Relief

MOA, Ministry of Agriculture

MFDP, Ministry of Finance and Deve10pment Planning
MVRU, Multiple-Visit Resource Use

NDP, National Deve10pment Plan

NFMS, National Farm Management Survey

NMS, National Migration Study

PSU, Planning and Statisics Unit

RDC, Rural Development Council

RIDS, Rural Income Distribution Study

RI, Researcher Implemented

RM, Researcher Managed

RSU, Rural Sociology Unit

SP, Single Plow

USAID, United States Agency for International Deve10pment
VDC, Village Development Committee

xii

l‘nis resear-
Certral Agricult:
largest agricult=
patter of its f1
the region is 10'
years with good ?
:ienselves. In 1
production subsi

I’ne resears‘:
Iazageaent, expe
EEPPOH Systems.
carried out betw-

lliis chapte
”12' the role of

.ne research ob i
J

BOtswann is

333: 3115!: nor
ii
333333r3- Host

1

tzallcefltrated in ‘

I. INTRODUCTION

This research provides information on farming systems in the
Central Agricultural Region, Botswana. The Central Region is the
largest agricultural region in Botswana and encompasses more than a
quarter of its farmers (Figure 1.1). The level of crop production in
the region is low and unstable, as it is throughout Botswana. Even in
years with good rainfall, many farmers do not produce enough to feed
themselves. In drought seasons, nearly all farmers rely on government
production subsidies and feeding programs, and on purchased food.

The research covers household circumstances, crop systems
management, experimentation on production practices, and agricultural
support systems. The research is based on village and on-farm research
carried out between October 1982 and June 1986.

This chapter gives background information on national development
and the role of agriculture, presents a problem statement, summarizes
the research objectives, and describes the organization of the thesis.

A. BACKGROUND

Botswana is a lightly populated, land-locked country in southern
Africa, just north of South Africa. Two-thirds of its 582,000 square
kilometers are covered by the desert and semi-desert sands of the
Kalahari. Most of the population of around one million people is

concentrated in villages located on the eastern border of the country.

 

4"
."

Po—.’

3

-..-- -..--J
r-— ‘q .\1 .
I 5 \

Ngomulofld We 1'

/
d

,-

WESTERN

./

/
. I?
9' I

| (hobo \

1 i

\

Ngom‘fa'nd (on

 

 

   

 

.\.
\ .r '
‘ /
l" . I
. .1
.‘ . _,
I f’
' I
l ,-- '

 

. i .
GA BORONE a
Kwoncng $00"! \(ng:
\ ,*

\
Novakonc Nor!

SOUTHERN

 

UNL ....... [__A‘
.

\

\
\
lulume

RANCISTOWN‘;

'\

P‘-

"(09‘
c

 

 

 

FIGURE 1.1:

MAP OF BOTSWANA;.AGRICULTURAL REGIONS AND DISTRICTS

 

;. PRJEILE 0? NE

mg Botswa:
eeustries in Afri
bee: little insti
mama and re:
were the major 5C

mating the 1
Four events had a
the country was t
is :opper, nickel

mes. Diamonds

with access to tt
Ki.
....i, the South

Locate Botswan.

.’\..
“4“;

it of the 1
albmfi a treme

The combine
319 30tsvana eco
iiltiuding rural

ne-

dCt STE? at

5%?
t

um“ iDahl ‘

l. PROFILE OF NATIONAL DEVELOPMENT

When Botswana became independent in 1966, it was one of the poorest
countries in Africa and was in a life threatening drought. There had
been little institutional, infrastructural or industrial deve10pment.
Livestock and remittances from mine laborers working in South Africa
were the major sources of rural household and national income.

During the 19708, circumstances in Botswana changed dramatically.
Four events had a significant impact on national development. First,
the country was found to possess mineral wealth on a large scale. There
is copper, nickel, coal, iron ore, manganese and two massive diamond
mines. Diamonds alone account for two—thirds of the value of exports
[MFDP, 1985]. Second, the Lome Agreement with the EEC provided Botswana
with access to the European beef market at no or reduced tariffs.

Third, the Southern African Customs Union Agreement was renegotiated to
allocate Botswana a larger share of customs union revenues. Fourth, the
drought of the 19608 was followed by several years of good rainfall,
allowing a tremendous expansion of the national cattle herd.

The combined effect of these events was a dramatic restructuring of
the Botswana economy. First, there was a large increase in earnings
(including rural earnings). Between 1966 and 1983, the gross domestic
product grew at an average annual rate of 13 percent [MFDP, 1985].
Second, there was a rapid expansion of demand, particularly demand for
imports [Dahl, 1981]. Imports rose by around 400 percent during the
decade following independence. Third, based on mine rents and customs
union revenues, a budgetary surplus was achieved in 1983, and has been

maintained since [Lewis, 1981]. Fourth, the contribution of agriculture

:: we fell from
the :mtributlon
mgfacturinie tt

Accompanying
age in rural inf
hare‘noie develop:
tints. Heanwfni
ievelopnent. AS
Viewed as one of

Despite its
elninate three I
Lhe economy to st
generation, and t
seztor More, 19,

The wuinera
li'is when Botsw
tttbreak of Food
recovered, the e
“Fifi swings in

The employ:

lepeideat der 1 .
K

4

to GDP fell from 39 percent in 1966 to 7.4 percent in 1983. By 1983,
the contribution of agriculture was outstripped by mining,
manufacturing, trade and hotels, services and government.

Accompanying the change in the economic picture, major strides were
made in rural infrastructure, particularly with respect to roads and
borehole deve10pment, and rural services such as primary education and
clinics. Meanwhile the political climate remained conducive to
deve10pment. As a result, by the end of the 19708, Botswana was being
viewed as one of the outstanding success stories in Africa.

Despite its notable successes, the GOB has not been able to
eliminate three persistent development problems: (a) vulnerability of
the economy to shifts in export demand, (b) insufficient employment
generation, and (c) poor and erratic performance of the crap farming
sector [MFDP, 1985].

The vulnerability of the economy was made clear during the late
19708 when Botswana beef was temporarily banned from the EEC due to an
outbreak of Food and Mouth Disease. Before the economy had fully
recovered, the export demand for diamonds dropped. The net effect was
rapid swings in the balance of payment between 1978 and 1983.

The employment problem resulted largely from the pattern of diamond
dependent deve10pment [Lewis, 1981]. Because there are few linkages
from diamond mining to the rest of the economy, the eXpansion of formal
sector employment was much less than the rate of growth of value added
[Dahl, 1981]. Moreover, the increase in employment opportunities was
largely offset by an increasing number of job seekers. The lack of

formal sector job Opportunities led to a problem of rural

underemployment, with nearly half the potential labor time not being

reei iliPme 19:
2, EQLE 0F AGRIfi

By the lld'i
sector but it has
:‘erelspneflt- 0“
cf :r'oan househoi
agriwltural sect
inerts, save for
at raise rural i

The Cowernne
agriculture. Cat
ezoraay and devel
priority during 1
sector remained t
Plan iDP V) but
3??3ftunities, t
Silesia given t

Lea Livestock d

Mat-mien would

a 1
uhEAOLds d0 DC

used (Lipton, 1978].
2. ROLE OF AGRICULTURE

By the mid-19708, agriculture was no longer the largest economic
sector but it has continued to play a vital rOle in national
development. Over eighty percent Of rural households and a large share
of urban households are involved in agricultural production. A viable
agricultural sector is necessary to reduce national dependence on food
imports, save foreign exchange, create rural employment Opportunities
and raise rural incomes.

The Government of Botswana recognizes the importance Of
agriculture. Cattle historically have been the backbone of the rural
economy and deve10pment Of a strong commercial livestock sector was a
priority during the first four national plans. The commercial livestock
sector remained the tOp priority in the 1979-85 National Development
Plan (NDP V) but, motivated by a concern with equity and employment
Opportunities, the Ministry of Agriculture was mandated to increase the
emphasis given to crOp production, limited resource farmers and communal
area livestock deve10pment. Continued emphasis on commercial livestock
production would have had severe equity implications since many
households do not have cattle [C80, 1976].

Although the Ministry Of Agriculture received an increased budget
allocation under the 1979-85 NDP, financial and trained personnel
resources available to the Ministry of Agriculture continued to be
limited [Litschauer, 1980]. Through 1981, there had been little
progress toward the goals envisioned for agricultural policy.

The performance of the arable farming sector became a critical

national problem when, in 1981, another drought started. Between the

377-75 season a:
perrent. Food 31
great of natio:
drag in the numb!
agioraent in cr-
aid-197'ds to 19C]
to stimulat-
eousehold food 5
agricultural ass
1954, store than
rations equivale
Food provided :11
596mm of th
ties, all farmer
35 Percent SUbsi
amides, Subs
raccinations. I
receive an 85 pe
and water catchy;

“931s.

me Sovermr

1:5 ,
Nultml ass

‘35. 1983]. In

J
ufi‘
men“ on to

the .
Programs aim

k‘leil
OWEN Plan

5m
. a falling in

6

1977-78 season and the 1982-83 season, the agricultural GDP fell 33.6
percent. Food grain production fell to 10,000 tons, less that ten
percent of national food grain requirements. There was a 20-30 percent
drOp in the number of households planting crops. As a result, estimated
employment in crOp production drOpped from around 250,000 in the
mid-19708 to 190,000 in the early 19808 [RDC, 1985].

TO stimulate farm employment, maintain rural assets and ensure
household food security, the government has set up several feeding and
agricultural assistance programs which subsidize rural households. In
1984, more than a third of the total pOpulation received supplementary
rations equivalent to 35 percent Of their basic food needs [RDC, 1985].
Food provided through the various feeding programs accounted for nearly
25 percent Of the total availability of staple grains. At the same
time, all farmers were eligible to receive 20 kilograms of free seed, an
85 percent subsidy on plowing (up to three hectares), payments for field
destumping, subsidized stockfeed, and free or subsidized livestock
vaccinations. In addition, farmers having fewer than forty cattle could
receive an 85 percent subsidy for the purchase of implements, fencing,
and water catchment tanks, and a 60 percent subsidy for traction
animals.

The government does not want to continue the various feeding and
agricultural assistance programs at their current level [MFDP, 1985;
RDC, 1985]. In addition to the political drawbacks from national
dependence on food imports and household dependence on the government,
the programs simply are not sustainable. The current National
Development Plan (NDP VI) projects that the growth of mine revenues will

begin falling in the next few years while the growth of government

experiitures wiI
Tierefore, the 1
food iaports an:
:3 rural househ.

37513? 3 Viable

performance of .

he pro'ole:
.‘iristry of Ag:
deelopuent whi
present, Minist'
nezagerial prac
iaproved crop p

Because of
different targe
1"? Segments
“6 aade to all

is “Ch or more

““3 and Cohen

re-r '
Janendatmns

Io COntrib
more informtic
do t

950 n
”“98 (h

35 ':
d1‘f‘irent te

he and cash t
tray“

is, ‘
“firs tnink a

7

expenditures will increase, leading to a government budget deficit.
Therefore, the government expects to have a declining ability to pay for
food imports and provide free or subsidized food and agricultural inputs
to rural households. Consequently, the Ministry Of Agriculture must
deveIOp a viable strategy for addressing the stagnation and erratic
performance of arable agriculture.

B. PROBLEM STATEMENT

The problem addressed in this thesis is the limited capacity of the
Ministry Of Agriculture to identify ways to contribute to arable farming
deve10pment which are not based on resource transfers and subsidies. At
present, Ministry officials do not know enough about farmers' resources,
managerial practices and decision processes to evaluate and target
improved crop production technologies and support systems.

Because of the Ministry's limited capacity to develOp programs for
different target groups, the agricultural assistance programs encompass
large segments of the pOpulation and the same extension recommendations
are made to all farmers. The result is that richer households benefit
as much or more from the assistance programs as do poorer households
[Holm and Cohen, 1986] and few households follow the main extension
recommendations.

To contribute to arable farming development, the Ministry needs
more information about several issues, including the following: (a) How
do resources (human, land, capital, and livestock) affect the relevance
of different technologies and programs; (b) DO rural households have the
time and cash to invest in new practices or implements; (c) Can
traditional practices be modified to increase production; (d) What do

farmers think about the practices recommended by the Ministry; (e) Are

“here any a”duc
ccrtizions; and
addressed to m
The lack of
essence of agric
istead, lack of
gas: agricultura
sirce the 197-0‘s.
Several an:
household dynana
Essen [1977], A
'n'aluahie, these
Zactites or agr
lotions for th
Shah of the
filtrated throng
’1 a Particular

Nukes, one fOC

.r‘at
' a

’ one On in
“Wu, 1975},
Dr 'u'lt ' e

' 3 lied mrorma

r». .
.45! Vere “wail;

8

there any production practices which perform well under drought
conditions; and (f) Are there institutional constraints which might be
addressed to facilitate crop deve10pment?

The lack of information on the above issues is not due to an
absence of agricultural and rural deve10pment research in Botswana.
Instead, lack Of information can be attributed to: (a) the pattern of
past agricultural research and (b) the change in national circumstances
since the 19708.

Several anthrOpological studies have provided information on
household dynamics and village institutions, such as Schapera [1943],
Kerven [1977], Alverson [1978] and Gulbrandsen [1980]. Although
valuable, these studies generally have not addressed production
practices or agricultural support systems, nor have they identified
Options for the Ministry of Agriculture.

Much of the information about Botswana agriculture has been
generated through national studies with predetermined mandates to focus
on a particular issue. During the 19703, there were at least four major
studies, one focused on the (now ended) food-for-work program [FAO,
1974], one on income distribution [C50, 1976], another on employment
[Lipton, 1978], and the fourth on migration [080, 1982]. These studies
provided information on agricultural households and the rural economy.
They were invaluable in describing the circumstances of poor rural
households, particularly households without cattle and female—headed
households.

The importance Of the various national studies is reflected in the
influence they have had on national planning. The studies were

unanimous in urging the government to shift its orientation toward the

pm. Host rec
generation. A
capital, both 1
in agricultural
of the studies '
productive capi
another subsidy

l'he hinist
audit knows, i'
what to do ahou
general. the v.
gractices, rela
development, or

A third so-
?ractices, is tj
35% Ministry, .
generauy has b.
lli‘l‘iios’ MOSt
Stations_ FEW
large gap bEtve

paling“, (b)

9
poor. Most recommended a focus on crop production and rural employment
generation. A common prescription was to create access to productive
capital, both livestock and implements, in order to increase the returns
in agricultural production. The main recommendations and prescriptions
of the studies were accepted in NDP VI. ALDEP was set up to provide
productive capital. Within a few years, however, ALDEP was changed into
another subsidy program.

The Ministry of Agriculture now knows it should focus on the poor
and it knows, in general, who the poor are. It does not, however, know
what to dO about the poor or about the stagnation of crop production in
general. The various national studies say little about production
practices, relationships between household circumstances and technology
development, or agricultural support systems.

A third source of information, which does address production
practices, is the agricultural research which has been carried out by
the Ministry. Unfortunately, agricultural research by the Ministry
generally has been equated with technical research and, until the
mid-19708, most technical research was carried out on experiment
stations. Few guidelines were developed for farmers because Of: (a) a
large gap between management Of the on-station trials and farmers'
practices, (b) a lack of significant yield benefits from many tested
practices, and (c) a failure to evaluate the economic benefits from
alternative practices (for examples, see ARS [1978] and DAR [1969]).

The weakness Of on-station technical research was recognized in the
19708 and two multidisciplinary on-farm research projects were
initiated. These projects, however, concentrated on testing

technological packages generated through on-station research [Gaosegelwe

9.110! 1953i-
;rrjects turnec
prrject focuses
livestock inter
with farm svste
such as patter:
rasrflows, fooc
Ecrctioning of
faraing develo;
Finally, I
P1313133 throw
“Neils carrie:
Statistics. 5,
data required
resources. The
Feetices, "as
Sezder roles,
:1953i and (:33
As a resu
target,

t

01d circ

iterates and

we. ;
iEIOre’ an

any“, avid p,

the: .

The goal

this“ of ar

10
et al., 1983]. After more than five years Of limited success, both
projects turned to a farming systems approach but, even then, one
project focused mainly on improved implements and the other focused on
livestock interventions [Baker and Hobbs, 1986]. Neither project dealt
with farm systems in the Central Region and neither addressed issues
such as patterns in household resources, whole farm labor use,
cashflows, food consumption, variations in traditional practices, or the
functioning of local institutions, and how these issues influence arable
farming development.

Finally, the Ministry does generate its own information for
planning through national farm management and agricultural statistical
surveys carried out each year by the Division of Planning and
Statistics. Both surveys, however, concentrate on a limited range of
data required for characterizing production trends and the returns to
resources. The data ignore mixed cropping, patterns in traditional
practices, wage employment, income sources, household labor allocation,
gender roles, resource sharing, and market participation (see Boykin
[1983] and C80 [1984]).

As a result, little is known about the relationships between
household circumstances, crOps systems management, modified production
practices and local agricultural institutions in the Central Region.
Therefore, an integrated, farming systems analysis is needed in order to
provide guidance to Ministry Of Agriculture personnel and to improve
their capacity to contribute to arable farming development.

C. RESEARCH OBJECTIVES
The goal Of this thesis is to provide a comprehensive, systems

analysis of arable farming in the Central Region, and the factors

u I ' .N
utectine a

niectives
1. To des:
identif

farming

l to gene

farmers

4' 3° anal
identii

PerfOrrr

11

affecting arable farming, in order to determine arable farming

development priorities. To accomplish this goal, the specific research

objectives are as follows:

1. TO describe patterns in household resource control and use, and to
identify the implications of household circumstances for arable
farming development.

2. To generate information on existing production practices, and
farmers' priorities and perceptions, in order to identify technology
deve10pment priorities.

3. TO characterize crOpping outcomes, and to identify production
practices which can improve cropping Outcomes even under drought
conditions.

4. To analyze the performance Of agricultural support systems, and to
identify policy and institutional Options to improve their
performance.

D. ORGANIZATION OF THESIS

Chapter II describes the research approach. The conceptual
framework is introduced. After that, an overview is given of the field
research activities and the data analysis procedures. The final section
identifies the distinctive features of the research approach.

Chapter III gives an overview Of the household circumstances
affecting arable farming. Information is presented on human, land,
labor, capital and livestock resources. Household resources are key
parameters affecting farm system performance and Options for improved
performance. Household labor use, revenue and expenditure, and food
consumption patterns are also described in order to place arable crOps

production into a household and farm systems context. At the end of the

 

, ‘ ‘
--. ' 9
war“

‘31

para;

8.8.. you
v6.58: 5
I

v ' e
1 a I.‘ a
iewlash

.fii‘hn 1
"P‘I‘L-o‘

but ge :

55“!

elm.-
f'tl:

.
9

W o'
“4, s.

3'.

12
chapter, the main implications of a household and farm systems
perspective for arable farming development are highlighted.

Chapter IV shifts from a household perspective to an analysis of
crop systems management. Information is presented on cropping
objectives, crops and varieties, draft management, traction use
patterns, production practices, post-harvesting practices, and roles in
decision-making. Attention is given to farmers' priorities and
perceptions, as well as to describing current practices. The final
section identifies implications of crops systems management for arable
farming deve10pment.

Chapter V gives an assessment of cropping outcomes and reviews the
efforts which have been made to identify improved arable production
technologies. The results of the on-farm trials are described and
budget analyses are presented for the most promising practices. The
budget analyses focus on minor changes since the harsh and uncertain
environment precludes as too risky interventions which require
substantially increased investments.

Chapter VI presents information on the performance of three
agricultural support systems. The first section gives an evaluation of
agricultural extension in the Central Region. The trading network in
the Central Region is addressed in the second section. The third
section examines village groups and a group formation "institutional
experiment." The final section gives recommendations on policy and
institutional changes.

The last chapter reviews the research approach and objectives,
summarizes the findings, points out implications for the Ministry Of

Agriculture, identifies limitations of the research, and suggests

3323!}

r ‘;
jade:

’31-‘03i

sssu

13
priorities future on-farm research in Botswana.

The Appendix gives an overview Of agriculture in Botswana, for
readers not familiar with Botswana. Information is given about the
technical environment, the agricultural sector structure, recent
production trends, production practices in the different agricultural

regions, and the main agricultural institutions.

«A
.34—

sari;
t0 3:

EETJI

II. RESEARCH APPROACH

The thesis is based on village and on-farm research carried out
under the auspices Of the Agricultural Technology Improvement Project
(ATIP). This chapter gives an overview of the research approach. The
first section introduces the ATIP project. The next three sections
describe: (a) the conceptual framework which guided the research, (b)
the field research activities, and (c) data analysis procedures. The
final section points out four distinctive features of the research
approach.

A. AGRICULTURAL TECHNOLOGY IMPROVEMENT PROJECT

ATIP was initiated in 1982 with a mandate to improve the capacity
of the Ministry Of Agriculture to develop and extend technologies
relevant to the needs of resource poor farmers. To accomplish the
mandate, there were four main components to the project: (a) an
agricultural economist was based at the main research station in order
to strengthen the capacity for multi-disciplinary research, (b) an
agronomist was appointed tO a newly created position Of
Research-Extension Liaison Officer (RELO), (c) long- and short-term
training were provided to Ministry Officers, and (d) two on-farm
research teams were established. The core of the project was the two
on-farm research teams. One had a mandate to cover the Central
Agricultural Region. The second had a mandate to cover Tutume District

in the Francistown Agricultural Region.

14

, AL-
.4,
In

2133
be "
.a‘
rcrtr
ll". :5

15

The Central Region team comprised twO expatriates—-an agronomist
and an agricultural economist--plus counterparts and field assistants
from the Ministry of Agriculture. The team had a mandate to focus on
crop production and the crop-livestock interface. The Objective was to
improve arable farming in a manner that contributed to equity and
national food independence. The target group of farmers was defined to
be those cultivating from one to ten hectares and having fewer than
forty head of cattle.

B. CONCEPTUAL FRAMEWORK

Arable farming deve10pment priorities in the Central Region were
analyzed using a farming systems approach. The farming systems approach
has three distinguishing characteristics: (a) reductionism is rejected
in favor of a systems perspective, (b) farm systems generally are taken
as the unit of analysis, and (c) a systems problem evaluation process is
followed.

This section defines the term "system," characterizes the systems
perspective, introduces the concept of a farm system, and describes the
problem evaluation sequence. The systems problem evaluation procedures
used in this research are comparable to those develOped by CIMMYT for
"on-farm research with a farming systems perspective" [Byerlee,
Collinson, et al., 1981] and by Norman for the "farming systems approach
to research" [Gilbert, Norman, Winch, 1980] but are presented using
terminology suggested by Manetsch and Park [1979].

1. DEFINITION OF A "SYSTEM"
A system is a set of interacting components which are organized

toward a goal or set of goals. A system reacts as a whole to external

a c 4 a .C a - e e w it HIM
"V .8 a 0L AG ~ A as
D * at,- ‘U. y» n .a I \J tutu.

D' a s h C . b‘h Is!‘ “'6 ‘2’-

.5

16

stimuli and each system has a specified boundary based on the inclusion
Of all significant feedback [Spedding, 1979; Dent and Anderson, 1971].

Following Manetsch and Park [1979], systems can be identified with
reference to input variables, output variables, and parameters which
define the system structure. The system structure is the set of
interacting components (or subsystems) and related variables (including
parameters) that link system outputs to system inputs. System
parameters are the fixed attributes of the system structure. System
input variables include environmental and controllable inputs.
Environmental inputs affect the system but are not significantly
influenced by it and, therefore, are beyond the boundary of the system.
Controllable inputs are provided by system managers. System outputs can
be either desired or undesired. System performance refers to the
success of a system in attaining desired outputs.
2. SYSTEMS PERSPECTIVE

The systems perspective is the antithesis of reductionism.
Following Dillon [1976], the reductionist approach entails reducing
phenomena to more basic part8, analyzing the parts as independent
entities, and then aggregating the results in order to explain the
phenomena under study. Dillon argued that reductionism has led
disciplinary researchers away from real world problems. Dillon
contrasted reductionism with the systems concepts of expansionism and
teleology. Expansionism is the view that all objects and events are
parts of larger wholes. Teleology is the belief that purpose and design
affect phenomena. The teleological approach involves setting a target
and assessing alternative actions for achieving the target.

The practical limitations Of reductionism have been identified by

17
many systems analysts. Two arguments are frequently mentioned. One
relates to global versus local Optimization. It is argued that
Optimization with respect to a part Of a system does not necessarily
mean that the performance of the entire system will be improved. The
second holds that the whole of a system is greater than the sum Of the
parts. Because of the synergistic interactions of systems components,
systems cannot be studied by considering their components in isolation
[Rountree, 1977].

Dillon [1976] said that the systems approach inherently leads to
better understanding since it is more consistent with real world
processes. With reference to specific contributions, Dillon argued that
the systems approach: (a) provides an indication of where research is
needed, (b) helps researchers to consider more than the material product
of agriculture, and (c) provides a workable procedure for research, and
for evaluating the likely ramifications of research.

Based on the expansionist and teleological concepts, the search for
solutions using the systems perspective cuts across the boundaries of
traditional disciplines and is philOSOphically eclectic, with an
emphasis on pragmatism. A distinguishing feature of pragmatism is the
belief that knowledge is valid only if it leads to "workable solutions"
tO identified problems [Johnson, 1986].

3. FARM SYSTEMS CONCEPT

A farm system is a biological, social and physical system in which
decision makers attempt tO control biological subsystems in an uncertain
environment in order to achieve multiple goals. "Farming systems" are
groups of farm systems which have nearly the same social, institutional

and technical environment and, a8 a result, pursue similar goals through

 

18

similar sets of system components. In a strict sense, farming systems
generally are not systems since: (a) no boundary is specified which
includes all significant feedback and (b) they do not react as a whole
to external stimuli.

A farm system encompasses most, but not all, the activities of an
agricultural household. Agricultural household activities also include
wage employment, schooling, child rearing, gathering, and attending
ceremonies. Consequently, research on farm systems may be sufficient
for addressing most of the factors affecting households, but not all.

A farm system can be viewed as having a management unit which is
responsible for the standard management functions, including problem
identification, observation, analysis, decision making, taking action
and assuming responsibility for outcomes. The goals Of the management
unit are accomplished by adjustments in the timing, level and
composition of controllable inputs applied to the various subsystems.
The management unit usually does not consist of a single individual.

Farm system managers have limited control over the nature and
output of the system. The structure and performance Of a farm system
are determined primarily by the system environment and by endogenous,
but uncontrollable in the short run, system parameters. Farm system
parameters include the number and composition Of components (or
subsystems), demographic composition of a household, behavioral
characteristics Of system managers and household assets.

A model farm system model is given in Figure 2.1. The first part
of the figure shows a general model which characterizes all farm
systems. The second part lists the key inputs, outputs, system

components, and resource parameters identified for farm systems in the

19

 

 

 

SYSTEM ENVIRONMENT

 

 

 

 

l

 

 

 

 

 

OVERT INPUTS ,7 FARM SYSTEM OUTPUTS

 

 

 

 

 

 

 

 

 

F'—

UTILITY

 

 

 

 

SYSTEM MANAGEMENT L

 

 

IDENTIFICATION OF BOTSWANA FARM SYSTEMS

SYSTEM ENVIRONMENT:
Technical: Soils, Rainfall, Climate, TOpography
Biological: Birds, Livestock, Pests and Vegetation; Varieties
Rural Economy: Prices, Wages, Credit, Input & Output Markets
Village Institutions: Groups, Inter-Household Exchanges
Government Programs: Drought Relief, ALDEP, Extension, ARAP
INPUTS: Labor, Cash, Traction, Seed, Implements, Land
FARM SYSTEM:
Components: Crop Production, Livestock Husbandry, Beer Brewing
Demographic: Size, Residence, Sex Of Head, Education
Assets: Livestock, Land, Implements, Buildings
Behavior: Perceptions, Priorities, Beliefs
OUTPUTS: Food, Beer, Consumer Goods, Livestock, Cash

 

 

FIGURE 2.1: A FARM SYSTEM MODEL

”-2,-
a.”

01;.

 

20
Central Region. Structural relations between system components are
implicit in the feedback lOOp (through management) in which outputs from
one component can contribute directly to utility or can have
instrumental value.
4. SYSTEMS PROBLEM EVALUATION SEQUENCE

The farm systems model does not indicate what should be
investigated, or how an investigation should proceed. To investigate
farm systems in the Central Region, a systems problem evaluation
sequence was used.

There are many characterizations Of the systems problem evaluation
sequence in the farming systems literature. The most common
representation was developed by D. Norman [1980]. Norman proposed four
stages of research: description and diagnosis, design, testing and
extension. Norman's terminology identifies processes which are
conceptually distinct, although they may be carried out at the same time
or even in a different order than is implied in use of the term
"stages."

A problem with most representations Of the systems problem
evaluation sequence is that they assume a focus on technology
evaluation. In this research, however, information is another important
output. Information can relate to human and institutional perspectives
as well as to technologies. Therefore, a slightly different
conceptualization of the systems problem evaluation sequence was
develOped and is presented in Figure 2.2.

Figure 2.2 is based on a merger between the stages concept Of
Norman and the feasibility evaluation process of the systems approach

presented in Manetsch and Park [1979]. The sequence moves through a

2i

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

r--- Needs
] Analysis r I
I
I
Ade uate NO
: l q l 3
: (Research)
: (TOpic) I~
: NO
r- System
i Identificationk_—‘ [Propose Backup ResearchL [Possible]
I
: [Adequate] N0 Check for
I Technical
: (System) Solutions --1
n (Components) as l
: NO i
k-Eescriptioni , 1
: LRecommend Policy Changes L [Main Problem] :
: [Adequate] No ,
I Check Policy -4
: (What is) ,t Environment :
' [Disseminate] —jT I
i—--[_Diagnos is I, No :
s ‘ '
: [Adeqlate] NO No [Viable] [ExperimentL [Soleble] ]
l E [On-Farm] '
1 (Problem) 7 :
i (Statement) , NO Prescreen I
i J, ‘ I (Set of) For "—1.
: Design [Complete]__(0ptions) Viabilit s
I
a J
..1‘ 7* r

 

I a l
1 L 41
LQather Information] [generate Information_] I Provide Information I

 

 

 

 

FIGURE 2.2: FARMING SYSTEMS PROBLEM EVALUATION SEQUENCE

22
series of processes (or stages). After each process, there is an
evaluation to determine whether the outcome of the process was completed
satisfactorily. The solid lines and arrows indicate the path through
the processes. Reverse arrows indicate iteration or repetition of prior
processes. The broken lines represent the accumulation and processing
of information at all phases Of the sequence. As represented in Figure
2.2, the collection, generation and provision of information are
separate processes, which are outside but complementary to the actual
problem evaluation sequence.

The sequence begins with a "needs analysis." The objective of the
needs analysis is to select a research topic which addresses an
identified need (or felt problem). The "need" can derive from a
research mandate, a review Of literature, intuition or an assessment of
farmer problems, and usually will be based on a combination. The second
process is to identify the system tO be investigated. This process
ordinarily starts with the farm system model. However, alternative
systems can be identified depending on the needs analysis.

Two of the most important processes are systems description and
diagnosis. Farming systems research procedures for systems description
and diagnosis include the following steps. First, select areas and
groups of farms systems with reasonably similar characteristics as
targets for research. The target groups are called recommendation
domains [Byerlee, Collinson et a1, 1980] or research domains [Hildebrand
and Poey, 1985]. Second, identify and describe systems components,
management behavior, environmental circumstances, resource and
demographic parameters, and performance measures. Third, examine

subsystems management and the determinants of system performance in

23
order to diagnose: (a) why practices are being used, (b) what problems
constrain productivity, and (c) the relative importance of constraints.

In the design process, a range of alternatives are generated to
address the identified constraints. The alternatives generally are
based on experiment station research results, farmers' indigenous
technical knowledge, and existing practices. The alternatives are
prescreened taking into account technical feasibility, economic
viability and social acceptability [Norman and Collinson, 1985].

The main design process output is a series Of hypotheses about the
technical and economic benefits from specific biological inputs,
husbandry practices, practices by environment effects, or policy and
institutional changes. The hypotheses are then tested through on-farm
experiments or evaluated through policy and institutional research.

The problem evaluation sequence ideally encompasses attempts to
test the validity of prOposed solutions by having the clients of
research react to them. The workability of solutions can then be judged
by seeing if the clients take advantage of the findings and if, once
adopted, the findings contribute to improved farm systems performance.

The problem evaluation sequence is not a one-time activity. There
is continual iteration with respect to the processes in any given
sequence. Also, new needs are identified over time. When a new need is
identified, it might take priority over an existing problem evaluation
sequence. In this case, the prior sequence is aborted, whatever
information has been produced is distributed, and attention is shifted

to the new needs analysis.

24
C. FIELD RESEARCH ACTIVITIES

This section describes: (a) the procedures used to select
representative villages and farmer cooperators, (b) the surveys and
studies used for systems description and diagnosis, and (c) the types of
trials used to test production practices.

1. VILLAGE AND FARMER SELECTION

During the first three seasons, field research activities were
focused on 52 farmer cOOperators in two representative villages.
Approximately half the cOOperators participated in the resource
monitoring survey (described below). Initially, only the other half
participated in on-farm trials but, eventually, most of the COOperators
hosted trials. All the cooperators also participated in several
single-visit surveys. This section describes the procedures used to
select the representative villages and farmer cooperators.

a. Village Selection In 1982

USAID and the GOB specified that the research coverage should be
limited to a representative village in each of two extension areas. In
conjunction with the District Extension Officers, three trips were made
to identify and select representative villages.

The villages selected were Shoshong, in Mahalapye West District,
and Makwate, in Mahalapye East District. Shoshong was selected to
represent the dominant pattern of cattle and tractor farming. Shoshong
is a large village with several active village groups, trading
establishments and schools. Makwate was selected to represent donkey
traction farming. In addition, Makwate has fewer services and less
infrastructural deve10pment than Shoshong. Since Shoshong encompasses

two extension areas, it was decided to work just in Shoshong East, one

c

F.

.1.
:4.-

r.
ml

c a
A:

25
of the two extension areas.
b. Farmer Selection

To select farmers, a three stage procedure was followed. First,
village exploratory surveys were carried out to establish a research
focus and define corre8ponding research domains. Second, a 16 question
census was administered in each village to generate a sample frame.
Third, nearly two months were spent contacting and interviewing the
households selected from the sample frame, to make sure COOperators
understood what would be expected of them.

The exploratory surveys were carried out during October 1982.
Eight to nine person-days were spent in each village. The interviews
were informal and unstructured. A checklist of information on practices
and problems was compiled during the interviews. Debriefing meetings
were held each evening.

The main arable production problem identified was inadequate and
uneven plant establishment (based on a literature review and
recommendations from on-station researchers, as well as the exploratory
surveys). Therefore, research domains were defined on the basis of
factors affecting the ability to implement timely planting or multiple
tillage Operations. Two RDs were identified. The domains were based on
whether a household controlled the use of traction resources (owned,
managed or borrowed) or was dependent on other households (hired or
shared draft). Lack of draft control had been identified in several
studies as the central constraint on timely planting and, consequently,
as a key problem affecting crOp production [FAO, 1974; Oland, Alverson,
Cummings, 1980; Livingstone and Srivastava, 1980; Vierich and Sheppard,

1980].

 

"‘

c‘»

b
H'

26

The census was administered during late October and early November,
1982. Following a brief training session, nine enumerators covered both
areas in two work-days. Eight additional person-days were spent
recontacting households. Approximately 90 percent of the households
were contacted in both villages.

The results of the censuses were hand tabulated so farmer selection
could begin immediately. Independent samples were selected for the
on-farm trials and resource monitoring so results of the resource
monitoring survey would not be affected by participation in researcher
managed trials. Based on available resources, it was possible to
include approximately 25 households in the on-farm trials program and
25-30 households in resource monitoring. The size of the resource
monitoring sample was based on an assessment that each of three
enumerators could interview five households a day (or ten households
total, each twice a week, with one day for recontacts).

Before selecting cooperators, each RD was divided into sub-domains.
The first sub-division was based on type of traction used--donkeys,
cattle or tractors. The type of traction was important to consider
because of differences in expense, and the speed and quality of field
operations (due to draft power). For the resource monitoring sample, an
additional sub-division was made on the basis Of cattle assets and
gender of the household head. Cattle assets were considered since the
research mandate was to focus on resource poor farmers. Also, cattle
asset inequality had been a major agricultural planning issue in
Botswana since the Rural Income Distribution Study (RIDS) [C80, 1976].
Gender of the household head was considered because female-headed

households had been identified as a particularly vulnerable segment of

 

r"
I.)

in

(J
P ‘9-

27
the rural pOpulation [Bond, 1974; FAO, 1974; C80, 1976; Kerven, 1977;
Livingstone and Srivastava, 1980; C80, 1982].

In order to make sure households representing each research domain
and sub-domain were included, the population was stratified before
randomly selecting COOperators within strata. The relative number of
households selected in each stratum was determined by the number of
households in each stratum in the p0pulation so results would not have
to be weighted when presenting findings for each or both villages.

Some adjustments were made in the course of selecting each sample.
For the trial cOOperators, cell proportions were adjusted slightly to
make sure there was minimal representation of each domain and traction
sub-domain. Also, all members of the trials sample had to have fenced
fields in order to protect trial plots. Two adjustments were made for
the resource monitoring sample. First, households with only one or two
members were not included, since a perspective Of household resource
allocation, not just for individuals, was wanted. Second, the sample in
Shoshong East was clustered, since enumerators had to bicycle to the
lands for interviews (a minimum of 10-15 kilometers). Once most farmers
had been selected through the stratified random sampling procedure,
clusters were completed by making contacts in the lands area. Farmers
were interviewed at randomly encountered compounds. A household was
included if it fitted the quotas based on draft access, traction, gender
of household head, and cattle ownership.

By the middle of December, 1982, 52 cOOperators had been selected,

and both trial and survey activities initiated.

28
c. Addition of Makoro Village in 1984

Beginning the 1984-85 season, Makoro in the Palapye Agricultural
District was added as a research village (following a GOB request to
work in an additional district). To select Makoro, representative
extension area characteristics were identified on the basis of a
regional survey (see Chapter VI). The objective was to find a village
which had less infrastructural deve10pment than Shoshong but, like
Shoshong, represented the dominant pattern of cattle-tractor farming.
The Ministry specified that the village had to be in Palapye District.
Logistical considerations required that only villages within two hours
of Mahalapye be considered.

Five potential villages were identified on the basis of survey
results. Each village was visited once or twice. Makoro had all the
specified criteria and provided a new settlement pattern. Although
having a borehole, health clinic, general trader, extension agent, and
various village groups, Makoro was a distant lands area for Serowe and
Palapye villages.

2. SURVEYS AND STUDIES

Four types of surveys and studies were used for systems description
and diagnosis: (8) resource use monitoring, (b) subject surveys, (c)
plot monitoring, and (d) technical studies. A catalog of the surveys
and studies, including the focus, reason, sample, dates and primary
investigator for each activity, is presented in Table 2.1.

a. Resource Use Monitoring

Resource use monitoring was a major activity during the first three

seasons. The primary resource monitoring instrument was the

Multiple-Visit Resource Use (MVRU) Survey, which was administered from

29

swan axon vooawuaoo

 

 

mo>uom aoummm wowuooHaISan
o< maumw mm mm motion: ucoammomm< woos cw waoHnoum woos muwucomH manouuma vooa
mo>uom mn< no muamouumdoo mono» Oawmouo mm Hmoomwou
ocumuumaoaon uoomoum ca mooaumaum> "maoHnoua nowadOuxo
2m ma< mm mm museum: HonouHOOfiuw4 unwowmwowwm mmwuoovH mm:0wuaoouom a mowum>fiuu<
mm can mvoum cofluoomoum mono pow maMOm mo mowuuomoum
u< mvaoww as mm now hufiafinmmum> Hwom mucfimuumaoo HMO» awwuaovH Hausaoso can amowmmnm
ho>uom cocoa wcwaoaa xaoawu wcwaoaa mo cowuouwcmwuo
2m as: «N mm maximum towowuu< among waOOOLMm ouOuomm xmwuoao Homoou "uaOfiuoumn NOMSOHm
Bounmm voowuomua aaaoaaoo
mwtqwmfi wcwuomHaISOHd vuwvcnun wofiuooamsowmaafiu wo
moOmoom qwsmwma Hommmoooom you vouwsvou mmoawHoawu cu moaoouso
o< sauna +00 mmsmwaa wcwuouflcoa uon mooooumaoouwo onwauouoa wcflnnouo mo qwnocowumaom
>o>n=m muoowowuwooo vow monsoon wowcmnuxo uoxuoasaoo van
was monoOmom omwuououoo mmumoo zuflaou assume "mowua>wuom oocoaou
ow uoo uwmw>sonfluaoz suoamo can muafimuumaoo segue vaonomso: a oaoocw
zm was um um own anon macs: oousomou xwmuoovH .Om: Hanna .kaHm sumo
.xmz w.monm unmade madam madame “mam mo mommaaw> ca «mm a
2m 8w an: mom mm >ozsuoo madam manamm ooamuuomaw humuo> mouusomou mo OOMuanwuuuwn
axo>usm .nnm oowmov mamanoum huwuowun
museumm mucuouonxm vow mowuwuofiuo "nucmmuumaoo consanom
m\2m on vooou< Nw uoo owoaaw> nouoonou uom “noomuonum huvamnasm

.mm>zH mqmz<m mma<a >9H>HHU< 20m<mm mpuom

 

omma OH Nwmfi

.mmmoaoo "mmHnaem az< mwm>msm no ooq<H<o "H.N

mqm<fi

3O

owed uxo: voaawuaoo

 

mowumwuouoouono
accumumwv now: mmaonomoon

someone muouoo>aw
xooumo>wa ”mowu«>fluom

 

mm mom >o>uom you mowuw>muom Show owcunoxo .oowuomuu

zm as: an ow uoo zuw>wuo< mo moooowoum ouomaoo .OOMOQEOaaOO .uonna

zo>u=m amuwoum monouumm

oawHowmm haocouwm mooaoo oaonom :Omuaaomaoo “somumuwfiwua

2m on: as am can momsoo ou cOaOmapowaw xumnsvoom uoovoum “anemones: momsou

o< mvHOMM ea om mmZIHQ< wcwuoomsawom mucwmuumaoo unfluaovH mauouuom oawmoum Hwom

2m mm nomsomh mo>u=m moow>uow moowmwoov m.uoanwm wowuw>wuoo msoum HOOOH

oumaxmz can a ououosuummumaH mewooOOHwaw maoouw can moow>uom unsecuo>ow

2m w woozmonm aw quIcmh .moowunuflumcH was moow>uom huwucovH “moowum a muovmuu owmaaw>

mowcono huouao>dw mom

on: an mm Hosanna m>o>uom mommauomuom saunas Ou moquuao>cfl xoouao>wa

vow muouco>aa dehumaou w auouoamumm mamuwnoo show monsooouun

Zn an: Nu «m umzsnom vaonomoom couscous mo mqmzamod omnaouwoaov vaonomoom

>o>uom mammuu now aowuooaoo mumawnuawm>o moon

madam uoapow “woman .muoaumm no “moowuooum was nowuw>wuou

Zn on: no mm >oz wowooouo unwsouv mo uoooam unomo¢ wawmmouo vonooam

mowuauowum noumomou «momma noooouomoua OOOM

>o>uom nouuon ou oOOwumoouom "mucwuuuuooo ouusouou mo

uaoaowmdmz van .oaoHnoua .moomuomum msom> .muoaumm “mundanmon

zm on: cad mm Gunshot mono no «use onwaooum vow nonwumuouoo mono
.mm>zH mqmz<m mma<a >9H>HHU< zom<um maven

 

Aeasaauaouv a.“ assay

31

.umwaooouwm a o< muonoumomou uaoaowmoma Boom n in

”noumwmumo>cw HongOMHm .o

 

 

mo>u=m mmmq< aw on vasonm
Em on: mm mm Homeosh om: uuwo muumo nonuons mmoum< own van ownmuocso uumo
mo>usm macaw: Houomnu mumou uwmnou can
2m muonao «m cm wo<lasm noose neuomua cu oow>uom Houuon ooamaouawma “mamumn
cm #86 mmvoum wowaoom m ovw>oum unwwa choose uowuuaoo cu monsoon
2m muouoouu mm mm uoo whooom pouomnH u0uomuu Bo: ouooam>m ”manouuoo on: wouomua
>o>uom mommumuOan xooumo>wa aw ocuouuma OOMOouwkus
couguomum monsoon wcmmmouocw now can .moomuomua wawvcou
2m an: mm mm boondom xooumo>flq mowufiasuuommo hmwuoovH .moquuao>om xoouwo>mg
mommaaw> as mo>uom ooooauowuom .uoovaoo mamanouo "moowum muxooum
2m “muovouu MQH mm cohsua< onwaommm povmuH .ousuosuum uoxuma «momm< ”vomnnouso a vHOn maouu
owvoasoax auowanoou
moowumvaoaaooou .muoaumw "noowuomua
Ammo>u=m thHocnooO ou momaoamou oowuonmamu: .wamumo>uwn
Zn on: um Hmahomv mucosawaw muonu so: .aomuoououm "moowuonum
mo owmuo>< mm monsoon mvoum was amenasoax Hmowanoou wamuaoaalaon "wawuonm
Amo>uam anumEOOMaH msooowwvo« .ouoanmw can causeway vaononoonluouaw
muououonxmv uooaowoomz mowaoomv use: nonwoov "coach vaonomoonluuuaw
2m mommaaw> N aw coaluoo a awn: OOMmMOOG cw usages“ nomoov cwmo “momnomuwoaov vaonomaom
no use mo>uom =m>z momum>wuoo aomuoovoum uuomuoo a ausmaw wawaoun
Zn on: Mg «m >oz vowvmun¢ anew hos Ou nonsuom noon was secuoo>wa .mouo

.mm>zH mqmz<m mmH<n NHH>HHU< 20m<mm maoom

 

Avooawuoouv ~.N OAAmH

32

December 1982 to September 1985. The MVRU Survey evolved during this
period in the degree of structuring, primary Objectives, and
comprehensiveness.

During the first year, the survey was not pre-coded. Twice-weekly
interviews were carried out with 27 farmers; 17 in Shoshong East and 10
in Makwate. The survey was designed to provide three categories of
information: (a) crop production activities, including fieldwork by
both household and non-household members, non-labor inputs, and traction
use; (b) resource acquisition and disposal, covering sales, purchases,
animal use, changes in livestock inventories, gifts and remittances,
loans, and miscellaneous expenditures and receipts; and (c) household
maintenance activities, including cooking, washing, gathering firewood,
fetching water, compound construction and repair, and livestock tending.
Initial attempts were made to collect information on travel, attending
ceremonies, gathering wild food, and hunting, but these were generally
not successful and, therefore, were abandoned.

Households were defined to include all individuals residing for
most of the season at either the lands or village compound Of the
household, whether family members or not. In addition, all conjugal
family members living Outside the village area were defined to be
non-resident household members if they had not yet established their own
household. The distinction between resident and non-resident members
was needed since transfers from non-resident members to resident members
were recorded while transfers among resident members were not.

Information was collected from individuals who participated in
activities. This generally meant that the respondents were women. In

nmny cases, multiple respondents were interviewed so they could help

each ott

Br
instrune
questior
ea:h mm
on hous«
flow da

Du
outputs
activit
househo
In gene

deSpite

33
each other recall activities and times.

By the beginning of the second year, a structured, pre-coded survey
instrument was developed. The survey instrument comprised a master
questionnaire and 14 followbup sheets which were collected and replaced
each month. The master questionnaire had a series of "yes-no" questions
on household activities. The follow~up sheets were used to record the
flow data corresponding each activity.

During the third season, data were collected only on inputs and
outputs related to arable production, beer brewing and livestock
activities in the land/village area. The survey was administered to 13
households which had participated in the MVRU Survey in prior seasons.
In general, these were farmers who had managed to produce some crOps
despite the drought conditions in 1982-83 and 1984-84.

To complement the 1984-85 MVRU Survey, an Activity Survey was
carried out from October 1984 to September 1985. The survey provided a
qualitative, seasonal profile of labor use, traction use, marketing and
consumption activities for households with different characteristics.
The sample for the survey was the cOOperators selected in 1982. Farmers
were asked on a monthly basis to recall the frequency with which an
event took place during the month.

b. Plot Monitoring

During the first two seasons, plot monitoring was carried on fields
of the trial participants and some fields of the MVRU Survey
participants. During the third season, the number of fields and plots
monitored was reduced. Eleven of the MVRU Survey participant's fields
were monitored that season.

A plot was defined to be any contiguous area planted on the same

34

day to the same crop or crop mixture. For each plot, data were

collected on date(s) of plowing and planting, area, plowing rate and

depth, soil moisture, seeding rate, seed quality, tillage and planting
method(s) used, and weed level. Rainfall data were collected on a field
by field basis. Production responses included plant emergence counts,
harvest plant counts, and grain yield. Yield 1088 estimates for bird
feeding, smut in sorghum, and cattle and wildlife feeding were made from
harvest samples.

c. Subject Surveys

The term "subject survey" was used to distinguish tOpical
single-visit surveys from the MVRU Survey. From 1983 to 1986, thirteen
formal subject surveys were designed and carried-out. The general
objective of these surveys was to broaden understanding of the
circumstances affecting farm systems decision making and performance.

The following surveys were administered:

1) 1983: Draft Arrangements; Crop Management; Agricultural
Demonstrator; and Cropping Plans (for 1983-84 season) surveys.

2) 1984: Village Institutions and Services; and Cowpea Baseline
surveys.

3) 1985: Trader Baseline; Management Information for PlowbPlanting;
Decision Unit Profile; Crop Protection, Harvesting and Utilization;
and Livestock Practices surveys

4) 1986: Tractor Ownership and Use; and Cart Use surveys.

The farmer cooperators were the core sample for the Draft

Arrangements, Crop Management, Cropping Plans, Cowpea Baseline and

Livestock Practices surveys. In the Draft Arrangements Survey, the

cooperators were supplemented with 11 randomly selected farmers in

35
Makwate (to make the sample from the donkey village comparable to the
cattle-tractor village) and 13 purposively selected households involved
in cooperative plowing arrangements. The last set was included in order
to have a sufficient sample size for evaluating cooperative
arrangements.

The most important subject survey was the 1983 CrOp Management
Survey. To analyze the representativeness of the cooperators, 66
randomly selected farmers were interviewed as well as the cooperators;
38 in Shoshong East and 28 in Makwate. As many as possible of the
cooperators and randomly selected households were interviewed again for
the 1985 Livestock Practices Survey.

Participants in the three surveys of the 1985 DUMI Study
(Management Information for Plow-Planting; Decision Unit Profile; CrOp
Protection, Harvesting and Utilization) comprised a quota sub-sample of
52 households selected from among the 66 randomly selected households
which participated in the 1983 CrOp Management Survey. The overlapping
sample was used to allow an evaluation of cropping trends. Cooperators
were not included in the DUMI Study since participation in trials may
have influenced their managerial practices and attitudes.

The Agricultural Demonstrator Survey was based on a questionnaire
distributed to all the ADs in the Central Region at their monthly
management meetings in March, 1983. ADs returned the questionnaire
though the mail. After a questionnaire was returned, each AD was
interviewed in order to eliminate mistakes. Eventually, questionnaires
were received from 52 of 54 ADs.

In the Trader Baseline Survey, 173 trading establishments were

enumerated in 49 villages, nearly all the villages in the Central Region

except for
the resear
villages,
selected f
In ad
appraisal?
following
6. Techn
Save
Particula
arable pr
observatj
Tech

Variabil

there ‘

36
except for Bobonong District. (Bobonong had by then been excluded from
the research mandate zone.) All traders were interviewed in 46
villages, while a subset of around 15 traders per village were randomly
selected from the three major villages (Mahalapye, Serowe, and Palapye).

In addition to the formal subject surveys, many informal (rapid
appraisal) and end-of-season trial assessment surveys were carried out
following the subject survey format.

d. Technical Subject Studies

Several technical studies were carried out in order to describe
particular characteristics of the physical environment pertinent to
arable production. These studies included direct measurements and
observations, as well as analyses of secondary data.

Technical subject studies addressed the following: (a) soil
variability, (b) weed burdens and composition in cultivated fields, (c)
soil profiles in relation to the depth of sorghum rooting, (d)
vegetation cover, soil types and topography in the project area, (e)
rainfall, temperature and evapo-transpiration during the four seasons of
field research, and (f) long term rainfall probabilities based on data
from the Department of Meteorological Services.

3. EXPERIMENTATION

Many trial designs were used for experimentation. The number of
treatments and replications varied depending on the trial objectives and
the degree of researcher versus farmer involvement. In some cases,
there were replications within farms and in other cases only between
farms.

During the first two seasons, most experiments were researcher

managed and farmer implemented, following standard farming systems

{2583“

involvi

techno}
season,
“sage:

A
researc
categox
3- Far

Sp]

ICC

811;

Bit

37

research guidelines. Beginning the third season, several approaches,

involving different combinations of researcher and farmer management,

were used. The relative importance of researcher managed and
implemented trials increased overtime, rather than decreased, because:

(a) pressure from the drought made it difficult for farmers to implement

trials and (b) the experimentation program increasingly focused on

technology design rather than final stage testing. After the third
season, essentially all farmer implemented trials were also farmer
managed.

A catalog of the trials carried out during the four seasons of
research is given in Table 2.2. The types of trials carried out can be
categorized as follows:

a. Factorial design trials. Factorial experimental designs, involving
split and strip plots, were used in order to evaluate component
technologies and technology by environment effects. These trials
were researcher managed and implemented and included within-site
replications.

b. Informal design studies. Informal investigations depended on
subjective farmer and researcher evaluation of interventions. These
usually were evaluations without comparison checks.

c. Superimposed trials on farmers' plots. Treatments generally were
researcher implemented on farmers plots. In cases where the primary
objective was to assess technical responses, two to four
replications per site were implemented. In cases where economic
analysis (and/or farmer assessment) was the main objective, fewer
within-site and more between-site replications were used.

d. Researcher managed tillage-planting (including crOp-variety)

 

 

Ego ..._...._._.wz.<m ww_s._.._<.: are... > ~ .~.:< 227.52: 35cm;

'3:

 

 

333d flak. N23~ oann—Avhcuva «mu.-.zul:~:nvda~xui ::<.Ll-za 5‘3 2v1~<sfi<0 nNsN. Hl~=<oNs

38

owoa uxoa vooowuooo

 

mmsemma oomwummaoo muwawnowum> Hmuowuo> mama» a uooanmaanwumo
o< maumw Am emsmmmfi wowmmouu spoons vow mouosuouoH mono mnemoooom
mauwm n owsmwmd wcwuowam Dom Bounmm memos a uooaooHo>ov
oucw owmaawu ouoa oumuon .uooanmmanmumo no moonuoa
manna m mwsowma oaonom Iuooow OO “womaoam manna wowuoon w wOwaoHn
u< moon m swimmmfi wcmuomamsowoaame now moonuoa wowuomfin vooz haumo mo nuoowwm
2m manmw 0H swimwma ucoaumouH room huwawpmfi> moon o>ounaH mucoauoouu voom
oaonom oonowm Henna wcwumom ooovou cu
u< moumm a qwlmwmfi mcmumom cyan mowwaov womswo wcfivoom owmaov coma mo aouunoo
swimwma woodwamsaoam moousomou nuuxo now now: unoaowovo
u< somehow a mwlmmafi mo oomumwmumo>aH unmannflanmuoo o>ounEH womuomamlowoaama
mwcwmmouuuouoH sumo» hue ow uoOMOfim
o< moan H mwlmmmfi Eonwuom stmcw one monomeou mono mosvwoou mono
Howw> woaavooo a uooamOHo>ov mono a
nanny mN owsmwma oowuoOMahow uwmooon O» unmanmflanmuuo nomad no
o< «Show N mwsmwoa wowaoam manooo OOMOo>HOnoOo nouns wowsoam hands mo ouoowwm
noOMOoOOA omsmwma Swagocnooa :H
ed snowman m nwlqwmfi «doom auflouosaoo ouoommo omuuflwuoahn uOOfiuomum
mnwuum was uOOwuoououom aouooa muonmmosm a unwuawan
64 mausw mH owlmwma oumnmmosm soouw>oo now mcmooouum .owuaawu movcoaaooom
.mu>zH mqmz<m mmw<n >HH>HHo< z0m<mm mboom

 

owaa OH Nwma .MMQOHUO mmHZMZHMmmxm 2M<hlzo mo GOA<H<U "N.N qu<fi

 

l.‘ll.lllii|.i||l|]llll ‘Vlils

«via/2 — s.— —._—2<U .. .4....-i<r. ...f . .....1. a [l s I - I ll 1 [lil Islll ii]- l.l I

fiuumsu.-.uUa-CUV N.N ONCE“.

39

 

 

HmHnH around an5 onuooHaHHo around uoHHHa a

2m manmm m omsmmmH wchoHnH mo mowmnnm>om onu momu< aoswnoa >>mon onooHLH

mucous oo>o onoa o>OHnom nouomHn was:

HmHnH On onHHHm new mom mouHcom :uHa OOHHHE

2m manmm NH emsmmmH onHHHm one nooaoHaaH onu nmomm< w aonwnom onucoHoom
muono vouooHom onucon

HmHnH OHOm mo naoawmommm amono mnmvoOOOm oon>

2m manna mm owtmmmH onncme oHom nosonwomon was noanmm :an mo onuaon OHom

HmHnH uchnnmooo nmea mH oaoumhm

coHuo>nomaoo unsumHoa HHom .moon80m onnoonlownHHHu

u< mouHm m omlmme nouns mnmvoooom aonw mHmmononm :OHno>nomooo nouo3
maoummo onuomHm

HmHnH unmovoonn now voom uoo oOHuOHno> womaou

muoHnw>smvonuoz one «OHOOHno> onoa no noow waHummovoonn

Sm manna m mwsqme onuomHm moosoo ooo nonnons ouosHm>m .m> onuomHm Don comm

u< oonHH> H mmlqmmH ucoaowmomz ummnn moonm BOHnonn %MHn:ovH nooEowmoma nwwnv hoxaon

:OHuao wowuoon mono onnw powwow

0¢ manmm m mwsowaH onsomnOOOD oumH .HHleamw now vooz owan waHBomnovo: Odom
H aOHuosvonm ooovona

o< nannm m mmaammH O>HnoouoH economsmwo now mooz ouOHn oovnmw monouos

N ooHnoooonm
u< canon m mwsamoH o>HooouoH OOHumvaoaaooon mmdn onoooa Honnoo «O on:
.mm>zH mHmz<m mma<a VHH>HHU< zom<um mboom

 

AvooaHnoOOV N.N oHnnH

cooper:

trial 1

and ma'

were I

tradit

e Farmer

were 1
assess
was: I

imple:

Most
differenc
exPiéCted
Sender ca
anticipat
inclUded
h0u5ehold

The
tract'mn
(e) Catt]
traction,
Mad “as
in the 1a
with rest
RCCQSS VIE

40

comparisons with farmer implementation. This was the most common

trial format. Tillage-planting practices were researcher specified

and managed. Farmers implemented the tillage-planting Operation and
were responsible for all non-treatment aspects of the trial. A
traditional check plot nearly always was included.

e. Farmer managed and implemented trials. All managerial decisions
were left up to farmers and few measurements were made to formally
assess outcomes. This format was used when the primary interest
was: (a) researcher assessment of problems encountered with new
implements or (b) farmer assessment of practices or varieties.

D. DATA ANALYSIS

Most survey data analysis was based on evaluating significant mean
differences (for quantitative variables) or deviations from cell
expected values (for categorical variables) across household type or
gender categories. The design of all survey forms took into account the
anticipated analysis based on household types. Index variables were
included on each questionnaire for the strata assignment of each sampled
household.

The main stratifying variables were: (a) village, (b) type of
traction used, (c) draft access, (d) gender of head of household, and
(e) cattle owned. Type of traction referred to the primary source of
traction, and distinguished among donkey, cattle and tractor. Gender of
head was based on the de facto head. (Some male heads were not resident
in the lands/village area and a female was the de facto household head
with respect to farm systems management.) The categories for draft
access were: owned, borrowed or managed, hired, or obtained through a

COOperative arrangement. Five levels were distinguished for cattle

med:

recate
calcul
For co

least

these

wariat

draft

41
owned: 0, 1-15, 16-35, 36-70, and more than 70.

During data analysis, the stratifying variables generally were
recategorized on the basis of statistical "rules-of-thumb." For
calculating means, the target was to have 20-30 observations per mean.
For contingency analysis, the target was to have an expected value of at
least five per cell. Aggregations were made as necessary to approach
these targets.

Given the small sample sizes for many subject surveys, the index
variables generally were categorized to take on only two values. For
draft access, draft owners and borrowers were treated as
draft-controlling households. Hiring households and households which
obtained traction through cooperative arrangements were treated as draft
dependent. Two approaches were taken for dividing households on the
basis Of cattle assets. For most subject surveys, 15 cattle was used as
the breaking point. This was the nearest break to the median number of
cattle owned. For analyzing the resource monitoring data, 36 cattle was
used at the breaking point since was close to the 40 head of cattle used
by the G03 to determine eligibility for ALDEP.

The collection of labor flow data anticipated gender and age
disaggregated analysis, as well as analysis based on the household
strata. Separate sheets were used to record the labor times for
household and non-household members.

To reduce the number of records, labor data were recorded only on
the basis of the number of men, women, boys and girls working on an
activity, and the length of time worked. The adult, non-adult division
was based on the year of independence, 1966, since this was easy to

recall. To give an example of the approach: if one man and two boys

exam;
durii

IECO‘.

Ptha

fiest
Soft,
crea,
had 1
t‘tEs
inter

and t

42
spent three hours plowing, a single line was used to record: "1" under
"days" for men, "3" under "hours per man," "2" under "day" for boys, and
"3" under "hours per boy." One advantage of the approach was that
multiple days of work could be recorded on a single line. Thus, for
example, if one woman spent an average of two hours per day cooking
during the three days since the last interview, the enumerator could
record "3" under "days" (one woman times three days) and "2" under
"hours per day" for women.

Revenue and Expenditure data also were combined in the field to
reduce the number of records. The following categories were used for
expenditures: production inputs, grain and meal, other food, animals
and animal products, household goods, gifts and loans, wages, transport,
and other. The categories used for revenues were: crop sales, livestock
sales, beer sales, other sales, gifts and loans, wages, and other. The
field aggregations were particularly useful for purchases since many
types of "other food" often were purchased at the same time and
respondents could more easily remember the total than they could the
prices of individual items.

Because of the data collection approach, it was possible to analyze
most of labor and expenditure data using only database management
software. This was valuable since no software was available for
creating aggregated data files, and the available statistical software
had limited memory capacity.

The trials data were analyzed using analysis of variance or paired
t-tests. Because of the drought, trials data analysis often focused on
intermediate cropping outcomes instead of yield. Emergence stand counts

and the percentage field emergence were the main intermediate cropping

outcome
energen
tillage
(neasur
depends

E:
This wa
changes
product

prices.

7“»-

non-nod
solving
Perspec
IQSEarc

1. sort

Operati
by tien
similar
exPansi
tends t
compata
rather
Fo

43

outcomes analyzed. (Percentage field emergence is based on the
emergence stand counts divided by the number of seeds planted.) For
tillage-planting trials, the effects of treatments on weed burdens
(measured as the percent field coverage by weeds) also was an important
dependent variable.

Economic assessments of trial outcomes relied on partial budgeting.
This was possible since the interventions analyzed required minor
changes in resource allocations and no changes in the end-use of
products. An attempt was made to value resources at their local market
prices.

E. DISTINCTIVE FEATURES OF THE RESEARCH APPROACH

The research approach has four distinctive features: (a) it is a
non-modernist approach, (b) has a subject matter rather than a problem
solving orientation, (c) is based on a farmer first planning
perspective, and (d) is broader than is typical for farming systems
research.
1. NON-MODERNIST

The research approach used in the thesis represents an attempt to
operationalize the concept of non-modernist research, as characterized
by Glenn Johnson [Johnson, 1983]. Non-modernist research is very
similar to systems analysis in that it is multidisciplinary,
expansionist, teleological and pragmatic. Non-modernist research also
tends to be eclectic with respect to research methods, relying on
comparative analysis, descriptive statistics and hypothesis testing
rather than quantitative analytical models.

Following Johnson, modernist research is based on the view that:

(a) explanation is derived from scientific laws, (b) knowledge not

express:
predicti
research
are not
respect
that res
performer
is more
are econ
Optima.

Joh

argued t

44

expressible in numbers is inadequate, and (c) evaluating the accuracy of
predictions based on theory is the main goal of research. Non-modernist
research is based on the view that the theories of any single discipline
are not a sufficient framework for addressing real world problems. With
respect to farm management research, the non-modernist perspective holds
that research focused on technologies and institutions, and human
performance in adjusting to new technologies and institutional change,
is more relevant to farm managers, extension agents and planners than
are econometric and mathematical programming models focused on static
Optima.

Johnson's concerns with modernism date to the 19508 when he first
argued that "management (including farm management) cannot be defined,
abstractly or in practice, as only a subfield of economics" [Johnson,
1957: 442]. In both 1952 and 1957, Johnson posited that managers are
concerned with many types of information not commonly regarded as
economic in nature, of which the most important are institutional
arrangements and technological change.

In 1983, Johnson cited several examples from the post-war history
of farm management research which illustrated the difficulties
encountered when attempting to use modernist methods to address
agricultural development problems. In contrast, Johnson cited other
examples in which non-modernist research did provide information of
relevance to farm managers, extension agents and planners.

Based on the American farm management research heritage, Johnson
[1982] argued at the 1981 Kansas State University Farming Systems
Research Symposium that the top priority for farming system researchers

was to carry out non-modernist research on: (a) farm-household

relatio
human c

promote

solving
researc
The put
0f prob
solving
Problem
I?
Practic
the get
constra
Problen
surveyS
h°listi
Problen
analYSe
Third,
decisic
anal)rsi
require
3. FAE
A

agricUl

45
relationships, (b) values and preferences, (d) induced institutional and
human change, (d) managerial processes, and (e) policy changes needed to
promote agricultural development.
2. SUBJECT MATTER ORIENTATION

Johnson [1983] also has clarified the distinction between problem
solving and subject matter research. The purpose of problem solving
research is to solve specific problems facing specific decision makers.
The purpose of subject matter research is to provide information on sets
of problems to sets of decision makers. As Johnson notes, problem
solving requires simultaneous attention to all variables defining a
problem and necessary for its solution.

This research follows a subject matter orientation for several
practical reasons. First, the top priority in Botswana is to improve
the general capacity of the Ministry to address a multitude of problems
constraining arable crOps deve10pment, not the solution of any single
problem. Second, the research is based on village and farmrlevel
surveys and experimentation and, therefore, does not encompass a
holistic enough perspective to actually solve arable farming development
problems. The solution of these problems would require macro level
analyses that go beyond the mandate of the Ministry of Agriculture.
Third, the research was carried out in isolation from the relevant
decision makers for agricultural policy in Botswana. Without a needs
analysis and diagnosis involving direct input from decision makers, the
requirements for problem solving research could not be met.

3. FARMER FIRST PLANNING PERSPECTIVE
A key postulate of the farming system approach is that planning for

agricultural development must start with an understanding of existing

farming pat!
based on an
shows that i
1950], exper
processes [C
technical an

Aside i
systems," tn
regarding [‘3
systems rese
economic via
h“ a passiv
Minn, Wine
30315 of res
r81ations an
ReSfearchers

In reac
Pronouns a
and 300th, 1
Edel, farme
PerspectiVQS

above' A fa

d - .
far

Hers are

46
farming patterns [Collinson, 1972; Norman, 1980]. This postulate is
based on an accumulation of experience in farmrlevel research which
shows that farmers have indigenous knowledge [Brokensha, Warren, Werner,
1980], experimentation procedures [A. Johnson, 1972] and adjustment
processes [Collinson, 1968] which they use to c0pe with an uncertain
technical and institutional environment.

Aside from a general consensus on ”the need to understand existing
systems," there are differences in farming systems research approaches
regarding the roles of farmers. Through the early 19803, most farming
systems research focused on evaluating the technical feasibility and
economic viability of modified practices and varieties. Farmers often
had a passive role. Farmers were the clients of research [Gilbert,
Norman, Winch, 1980] and generally implemented the trials, but the real
goals of research were agronomic and economic analyses of input-output
relations and profitability [Byerlee, Harrington, and Winkelmann, 1982].
Researchers assumed the role of decision maker.

In reaction, a number of FSR social scientists recently have been
promoting a "farmer first" approach to farming systems research [Rhoades
and Booth, 1982; Chambers and Ghildyal, 1985]. In the farmer first
model, farmers are involved in all research processes and their
perspectives dominate the problem evaluation sequence.

The research approach in this thesis is distinct from both of the
above. A farmer first viewpoint is adOpted with respect to the
importance of farmers' perspectives, since ultimately it is the farmers'
decisions that matter--not researchers' decisions. On the other hand,
farmers are only indirectly considered to be the clients of the

research. The direct clients of the research are the planning, research

47

and extension personnel of the Ministry of Agriculture.

The thesis emphasizes the Ministry of Agriculture for two reasons:
(a) research and development efforts focused on particular farmers in
particular areas often do not spread to other farmers and (b) arable
farming deve10pment will take a long time in Botswana. In the long run,
the capacity of the Ministry to facilitate deve10pment is more important
than short run improvements in farmer welfare (given the various
assistance programs which buffer farmers against the worst effects of
drought and low production [Holm and Cohen, 1986]).

4. SCOPE OF RESEARCH

Despite the broad conceptual framework of the farming systems
approach, farming systems research projects and programs often have
ignored local institutions (support systems) and household dynamics.

The guidelines for FSR emphasize cost- and time-efficient research
methods. For example, Gilbert, Norman and Winch [1980] recommended that
minimal investments should be made in information gathering; just enough
to develop the understanding needed to design and test production
technologies. Collinson [1982] recommended focusing on one or a few
activities or enterprises which most affect overall farm systems
perfonmance ("leverage points"). In general, more attention has been
given to how quickly diagnosis could be completed and testing started
than to the quality of results obtained.

After a great deal of early enthusiasm, questions started emerging,
in the mid-1980s, about the potential of "quick and efficient" farming
Systems research methods. C. Eicher [1986], among others, argued that
time scape of FSR was too narrow to have a significant impact on the

deteriorating agrarian situation in Africa. A second argument held,

insteac

unknown

of the
system:
system!
system:
abandor
time it

intra-t

48
instead, that the potential of the farming systems approach was still
unknown because farming systems research methods had deviated so much
from a true systems perspective.

Two approaches have been taken to address the perceived weaknesses
of the farming systems approach. One has been to reject the farming
systems approach in favor of other systems, including agro-ecological
systems [Gibbs, 1985], agro-forestry systems [Rochelieu, 1986], and food
systems [Riley and Staatz, 1981]. The second approach has been to
abandon the emphasis on production technologies and quick turnaround
time in favor of systems research methods which more adequately address
intra-household dynamics [McKee, 1984; Moock, 1986; Feldstein and Posts,
1985], inter-household linkages [Behnke and Kerven, 1983], consumption
perspectives [Frankenberger, 1986], and the requirements for national
development planning [Norman and Baker, 1984].

Since the problem addressed in this research relates to the
requirements of the Ministry of Agriculture, and since the Ministry has
been directed to use the farming systems research methodology [MFDP,
1985], this thesis has followed the second approach. An important
contribution of the thesis, therefore, is to illustrate a broader

approach to farming systems research.

III. HOUSEHOLD CIRCUMSTANCES AFFECTING ARABLE FARMING

This chapter gives an overview of household and farm system
structure and resource use in Shoshong and Makwate villages. The first
section is on household demographics. Section two characterizes
household wealth, including livestock inventories and farm fixed
capital. The next three sections describe household labor use, cash
revenues and expenditures, and food consumption. The final section
points out the implications of household and farm system circumstances
for arable farming development.

There are two main objectives to the chapter. The first is to give
a profile of household circumstances in order to place arable farming
into a broader household and farm systems context. The second is to
analyze differences in circumstances for: (a) households in Shoshong
versus households in Makwate, (b) male-headed versus female-headed
households, (c) cattle-rich versus cattle-poor households, and (d)
draft-dependent versus draft-controlling households.

A. HUMAN RESOURCES

Information on household demographics was collected in household
censuses administered yearly to the MVRU Survey participants. The
primary objective was to characterize the human resources available for
crop production activities. The 1984 census revealed around 12 members
per household, of which ten were in residence. In contrast, many of the

studies from the 19708 reported only 7-8 resident members (for example,

49

SOC

ge:

50
FAQ [1974], Bond [1974], C80 [1976]). No significant differences in
household size were associated with village, gender of household head,
draft access or cattle assets.

To assess consumption requirements, resident consumer units were
derived using adult caloric requirements to weight individuals of
different ages. Overall, households had to feed just over seven adult
consumer units. Makwate households had more consumer units but
otherwise there were relatively small differences associated with the
household categories.

Since human resources are a key farm system parameter, the census
data were used to examine three demographic issues: (a) age and gender
composition of the farming p0pulation, (b) main activities of the
population besides crOp production, and (c) residential patterns at
different periods of the year.

1. HOUSEHOLD COMPOSITION

There are age and gender specific production roles in most African
societies [Boserup, 1970]. Consequently, differences in the age and
gender compositions of households often are related to resource use
patterns and farm systems performance. Information on the gender and
age composition of households is presented in Table 3.1.

The age and gender compositions of the household categories were
similar. Approximately 45 percent of resident members were adults.
Girls (under 16) were particularly prevalent in cattle-poor and
female-headed households. A larger prOportion of household members were

boys in Makwate than Shoshong East, reflecting the fact that fewer

.—.~< .—.z...:_z..._1..._: 23n~ WW]: #1? rd-<!¢.un- ..l-<: tapr<~._\-<: .\:r~:r.uf\:'u

 

IIICU< 9.4.6.1: m...—.H._.f..n..< BJLLV<U 26.2: EC XIV. ..L..»<.~‘->

.i- I iixl l..]\l [I]; I. ll II I

 

 

cminmo_ .213P3:zhm Qqazxmzcz "~.n qu<h

51

W85 a case: sages mmN.o u mNuoN owe mom.o . man «we moN.o u «no «we

.oo.~ n cos uasvn
"ounwwoa pad: uoasmcoo .n

.vouwuoasco mamoom mam .mvaonomno: mo>usm Dm>z nu “unmade vaonomaom «wad ”ouusom

 

 

N.N m.o m.N N.N N.N a.o m.N N.m e.o uu\maHz= mmzsmzoo
N o m o e o m o m u>NueHmmuaoz
NN NN 0 NH NH wN m m «N m>NumNmm umeuo
mN NN mN mN NH Hm NH NN mN eNNeoeauuu
on me am as Hm Ne on as on eNNeo
o a N o N N m o o museum
m a w m m m m m m exam
“ANV q<mm oa ona<ANN wm onaNmomzoo
mm me Nm oe 0N Ne om Nm mm nauNo
mN NN em mN on «N Nm mm eN when
NN mN ON NN NN NN NN mN NN asses
mN NN «N «N 0N mN ea MN mN cox
"ANV mazmaHmmN onaHmomzoo amazmonuu<
Nm om NN mm mN mm NN 0N mm «NuNu
eN oN mN mN Hm NN mN Hm NN amen
NN mN ON NN NN mN oN ON NN nose:
mN NN «N NN mN NN «N NN «N can
“ANV Newman: 44¢ onaNmomzoo mmnzmoumu<
Nm Nw Nw mm NN mm NN mm aN saueNmmm uauonom
m.m m.m a.¢ o.oN w.w N.oN m.a a.o~ ¢.a nuaoeNaom Hogan
N.NN o.HN ¢.NN N.mN m.oN m.NN o.NH o.NN N.NN moaned: NN<
“mNHm agommmsom
aq< azmazmmma zso mmuo +om maazmm was: we<zu<= ozommomm
mmmooa edema mammm< ugaa<o n<mm mo xmm mo<ga~>
emunmmN .mmseosmem naommmaom "~.n mam<a

Makwate h

Diff
snen focr
15 per: 1
Inienal
amen on
These ho
:oaprise

miEratic

Popula
quartE
”age ‘
types
have

ample

52
Makwate households had distant cattle posts.

Differences in age and gender compositions were more pronounced
when focusing just on resident household members. Men only constituted
15 percent of the resident pOpulation compared to 22 percent for women.
In female-headed, cattle-poor and draft-dependent households, resident
women outnumbered resident men by an average of around 70 percent.
These households also had a relatively large of their populations
comprised of young grandchildren. Similar effects of temporary
migration on household composition were noted in FAQ [1974] and C80
[1982].

2. MAIN ACTIVITIES

In the Central Region, crOp production usually does not provide
labor returns which are comparable to wage employment. Therefore,
involvement in wage employment, or even schooling (since that is the
route to wage employment) must be considered a parameter for the crop
production subsystem. Thus, an important demographic issue is who
considers farming to be their main activity. Findings on this issue are
summarized in Table 3.2.

Crap production was the main activity of only 25 percent of the
papulation. A quarter of the papulation were going to school, another
quarter were too young or too old to be active, and 19 percent said that
wage employment was their main activity. Differences across household
types were relatively small. Female-headed and poorer households did
have a slightly higher prOportion of individuals involved in wage
employment, while male-headed and cattle-rich households had more who
considered herding to be their main activity.

More substantial differences were found when looking at main

MN

1\ \s

TABLE 3.2:

53

MVRU SURVEY PARTICIPANTS' MOST
IMPORTANT ACTIVITIES, 1984

 

 

 

EMPLOY- NOT
FARM HERD STUDENT MENT ACTIVE
(Percent of POpulation)
HOUSEHOLD TYPE
VILLAGE:
Shoshong 27 6 24 17 25
Makwate 22 3 27 20 27
SEX OF HEAD:
Male 27 6 26 16 25
Female 24 3 24 21 28
CATTLE ASSETS:
0-35 24 3 23 20 30
36+ 28 8 29 15 20
DRAFT ACCESS:
Control 26 6 26 17 25
Dependent 25 3 25 18 28
PERSONAL
ACE-SEX:
Men 39 4 3 53 1
Women 69 0 5 20 6
Boys 4 14 41 1 40
Girls 3 1 44 3 47
EDUCATION:
0-6 Years 23 5 26 11 35
7+ Years 25 5 25 19 26
egg, 25 3 25 19 26

 

Source: 1984 Household Census.

ctivitil
said cro
sen. P101
activity
herding.
3. R331:

Toe:
enigratic
second is
beginning
these mig
comPlated
1ignition
Parliament
sYStetns d
°P€ration
SignifiCa
collated
Findirigs

In 1
heme“ t
approxima
remainitial.
FIOPOrtiO
beCauSe t

mute c.

FeV

54
activities in relation to age and gender. Nearly 70 percent Of women
said crop farming was their main activity compared to only 39 percent Of
men. More than half the men said wage employment was their main
activity. The main activity for boys, aside from being students, was
herding.
3. RESIDENTIAL PATTERNS

There are two major migration patterns in Botswana. One is
emigration from the lands or village area to seek employment. The
second is seasonal migration between the village and lands at the
beginning and end Of the crOpping season. Much has been written about
these migration patterns; a major national study assessing migration was
completed in 1982 [080, 1982]. One major conclusion was that seasonal
migration is declining. A reduction in seasonal migration, in favor of
permanent lands settlement, would remove a major constraint on crOp
systems deve10pment by creating an Opportunity for Off-season field
Operations [Livingstone and Srivastava, 1980]. To assess the
significance Of seasonal migration in Shoshong and Makwate, data were
collected on the principal residence at different periods of the year.
Findings are summarized in Table 3.3.

In 1984, there was a substantial amount Of seasonal migration
between the village and lands area. During the crOpping season,
approximately half of the pOpulation moved to the lands. Most of those
remaining behind in the village were school age children. The
proportion Of individuals remaining in the village was higher in Makwate
because the lands areas are close to the village. Several households in
Makwate commute on a daily basis rather than make a seasonal migration.

Few individuals had taken up permanent residence in the lands area.

 

 

 

caeu .:C.zmm >= muzsaNmr: J<LNUZ~2L n.n £4=<s

55

.usaaoo vHonomnom qme “oouaom

 

ON

mm
NH

NH
ON
Om

wH
NN

NN
oH

HN
ON

mH
HN

.-s
‘¢

aq
Nw

N
x?

N Na
2
m 2
N .3

00050
O
....

Nu
an

....
NO‘

m a no
on

...-1
N

N mm
N m mm

N

o o Hm
N 5 mo

ON

.3
NH

NH
ON
we

mH
NN

NN
oH

HN
oH

wH
HN

H he

H mm
Hm Nm

o HO
0 on
0 mm
m mm

H me
H «m
H me

H mq
H mq

AcOHuOHsaom mo uaoouomv

Nm

ON
on

Nm
mm
mN
oH

H44

muwmw +h
mummy olo
"onH¢o=nm
nHHHU
whom
:oaO3
cox
"xmmlmu4

 

moHHmHmm90<m<mo A<zomMmm

mm
wN

NN
mm

«m
Hm

oe
wN

unovaonoa
Houuaoo
“mmmoo< Hm<ma
+om
mnlo
“memmm< MHHH<U
OHanm
OHmz
un<mm mo xmm
ouosxnz
waonmonm
"mu<HHH>
mmwy agommmoom

 

 

mQHmHDo Hmom.H<U mnz<4 MU<HHH>

mmysz UZHMDG

mnHmHDo Hmom.H<U mnz<A mU<AAH>

zom<mm qummOMo

«me .nonmm we mozmaHmmm AdeUszm

"m.m m4m<H

56

Also, few considered cattle posts to be their principal residence at any
time of the year. On the other hand, around 20 percent Of the
population resided outside the lands/village area during all periods of
the year. TO the extent seasonal migration had declined in Shoshong and
Makwate, it appears it had declined due tO permanent rural-urban
migration rather than permanent lands area settlement.

B. LAND, CAPITAL AND LIVESTOCK RESOURCES

This section presents an overview Of household land, capital and
livestock resources. The first part gives a profile of resource control
based on the 1983 CrOp Management Survey. The second part presents a
more detailed analysis of the composition and value Of the MVRU Survey
participants' assets.

1. RESOURCE PROFILE, 1983

In the 1983 Crop Management Survey, resource data were collected on
cattle assets, major implements, land availability and field
characteristics. Results are summarized in Table 3.4.

Substantial differences were identified in cattle assets. Fewer
Makwate, female-headed, and draft-dependent households had cattle than
did Shoshong, male-headed and draft-controlling households,
respectively. Those Makwate, female-headed and draft-dependent
households which did have cattle, tended to have fewer cattle. Although
substantial, the extent Of cattle asset inequality in the CrOp
Management Survey population was less than that reported in the FAO
[1974] and RIDS [C30, 1976] studies. In both those studies, 45 percent
Of the households had no cattle and, in the RIDS, the poorest 50 percent
Of households only had 0.5 percent Of the cattle.

The differences by household types noted with respect to cattle

 

 

 

 

. :zmuflebt IHCZ‘,NZOUIIIII+10..H- [IsnrHtawmzflw ...... <2 . 3 Miami. whip—fin 41:.
Mimxnuic< .Hinwlxlwwc 34.59639 C<vL= XHW HM...L<J1~H>

 

mum‘s n o mavzhuflvth Anz<ufl 32“ uflrzvizflflfl 33”.... Q M\~.~.vfl.<o n 0V 3 MM Mi‘u‘u‘fH.

57

.mvHonomson oHH mmO>usm uaoaowwcnz mono mme "Oousom

 

 

 

Ne HN ON NO He ON ON Na NN oaoz
NN NH NH NN NN OH HN HN HN swam
ON NN NN NN ON NO NN NN Ne mNHs
"Hum NO OHNHN NO quosz
NO NN NN NO ON ON NN NO HO HNHONNN
NN NO NO NN ON Os Ne NN NN uumHaaoo
“New NO OHNHN NO oszzpemmO
NH NH OH HH HH NH NH NH NH New NO azOH OONNON NN>N
N.N O.N N.q N.O H.N O.e N.N O.N N.H Oz<H HHO NNO OH zH NzONONN
HN Os ON NN NN Ne NN ON NO Hem NO Ozaa ONNOzO N><m
O.NN N.OH N.NN N.NH N.ON O.NN O.NH N.NN N.HN ONN<>HNHOO OHNHN Omen»
Oz<H
O NH NH N N «H a NH OH +N
Os «O NO NO OO NO NN ON NN N no H
ON HN OH NO NN NH OO NN NN O
"meHz mm NzNONNO
N.O ¢.H N.H O.O O.O H.H O.H N.H H.H NNNzOz NOONN><
azmzmHOmN
ON NN ON O NH NN NH NN ON +ON
OH NN as O NH ON OH NN HN NNuOH
eN NN O «O ON ON NH ON NN NH-H
HO ON O ON NO NH ON NH ON O
"quzzo a: NzNOemm
N.O O.N O.N N.O O.O 0.0 N.H N.N O.N NHaO Oswaaqz
O.N N.N N.N e.O O.N e.N O.H N.N O.N Newman: uoz .Ounao
N.NH O.NN O.NN H.N O.NH N.NN H.O N.NO O.NN Omwucez New Ooaao
“NNNzOz NONNN><
NHNN<O
OZNNNO HONNzOO +OH NHsO NHazNN Nqaz .3N4z .NOON HH<
Nmmoo< NN<NO NHNNNO O<mm NNN NOOHHH>

 

NNOH .mmumpommm Oz<a Oz< azmmeOom .NHNN<O "¢.N mHm<e

assets vs
female-hi
equipnen
equipnen

To
identifi
draft-cc
fields I
more Hi1
cultiva'
386 bet-

1n
SEasons
unused

Haiti 3 t e

58

assets were also found for equipment ownership. Fewer Makwate,
female-headed, cattle-poor and draft-dependent households owned
equipment than their counterparts, and the average numbers Of pieces Of
equipment owned were less.

To a slightly lesser extent, similar differences were again
identified for land resources. More male-headed, cattle-rich and
draft-controlling households had completely destumped and wired fenced
fields than had their comparison households. However, there had been
more wire fencing in Makwate than in Shoshong. Fields had been
cultivated longer in Shoshong East than in Makwate but differences field
age between households categories were insignificant.

In both villages, land allocations were completely used in five
seasons out Of ten. Less than half the households reported they had
unused land in good seasons. Few draft-controlling, cattle-rich or
Makwate households had unused land in good seasons.

2. COMPOSITION AND VALUE OF FARM ASSETS, 1984

In 1984, two inventory surveys were administered to the MVRU Survey
participants in order to develOp a more detailed picture Of household
resources. One survey covered livestock inventories. The other was on
farm fixed capital, including buildings, field characteristics,
implements, hand tOOls and receptacles. This part presents selected
findings on the composition and value of farm assets.

a. Farm Fixed Capital

All households had access to at least one field; 22 percent had
access tO a second field. On average, the current heads Of households
obtained their fields twenty years ago. More than half the fields were

Obtained via inheritance. However, nearly half the fields in Shoshong

East and more
and/or fenced
fields were bi
cannon in Halo
female-headed
households.
Most hon;
rcent had m
percent of thc
only half the
lost hou:
axes, hoes, :31
takes and auge
racePtacles, [
one or more d1
hOUSEholdS hac
Wheelbarro" 1'
households hax
for h°“3€hold,
HOuSQhoh
On in the 13m
households hac
Hikwate hOUSer

generally were

building.

59
East and more than a third overall were initially Opened (destumped
and/or fenced) by the current household head. A small prOportion Of
fields were borrowed or were Obtained as gifts. Borrowing land was more
common in Makwate than in Shoshong, and for cattle-poor and
female-headed households relative to cattle-rich and male-headed
households.

Most households had one single-furrow mouldboard plow. Nineteen
percent had no plow while 22 percent had two or more plows. Over 60
percent Of the plows owned were purchased more than 20 years ago and
only half the plows were purchased new.

Most households had a variety Of hand tools, usually including
axes, hoes, and spades. Thirty percent or more households had picks,
rakes and augers. Most households also had several types of
receptacles, usually including small and large buckets, large bowls, and
one or more drum and large plastic containers. Only ten percent Of
households had a wheelbarrow, despite the potential value of
wheelbarrows in reducing time spent fetching water. The proportions Of
households having the various hand tools and receptacles were comparable
for households in the two villages and across household categories.

Households had on average more than five buildings; just under two
on in the lands area and around four in the village. All Shoshong East
households had buildings in the village and the lands area. Several
Makwate households did not have lands area buildings. Village buildings
generally were better constructed and require a larger investment per

building.

60

b. Livestock Inventories

Just over 85 percent Of the households owned some cattle. Nearly
all owned chickens and 85 percent own goats. Less than half owned
donkeys, a quarter owned sheep and none owned either horses or pigs.

Patterns of cattle ownership followed those identified in the 1983
CrOp Management Survey. For goats, there were differences by village in
the proportions Of households owning and the average numbers owned, with
more Shoshong households owning. There were no significant differences
by household type. The FAO [1974] study also showed that differences in
smallstock ownership were less skewed than was cattle ownership. No
significant differences were found for chicken ownership or average
chicken inventories by village but male-headed, cattle-rich and
draft-controlling households had 50-100 more chickens than their
comparison households. Donkeys were more commonly owned by Makwate,
male-headed and draft-controlling households. Relatively more
male—headed households owned sheep.
c. Value Of Farm-Household Assets

Values of livestock inventories and farm fixed capital for MVRU
Survey participants are presented in Table 3.5. Overall, households had
around P10,000 in assets. Shoshong East, male-headed, cattle-rich and
draft-controlling households all had greater assets, largely due to
livestock assets. These same households consistently achieve higher
levels Of arable crops production (see Chapter V).

A substantial prOportion Of assets were tied up in buildings,
particularly village buildings. This was particularly true in Shoshong
East. Households with more cattle had buildings which were worth more

than double those of households with fewer cattle assets. Neither

61

one

.umOO ucoaoooHaou Soc um OOOHO> moHonunooou can OHOOu vsnm
.coHqucoo mama Onu OH uOOaOOoHnou O OHOunO Ou umoo uaouuso

do women OOOHO> xOOumo>HH

.qmmH OH mo.o m venous n OHsm
.OOOHNO OOHOO ownuo>o

mo mucoammommo m.uoauow no women mwaHvHHon van .ucoamHsvo Honuo .OBOHQ mo OOHuOOHO> .n

 

 

 

 

.mtherm kufiawo vmxwh Sham van mucuflmNrcH vHooummNNMA ..vwma "mOUH—Hom
NN NO NN ON ON OO NN HN ON NNOHOHHON
O O N OH N HH N OH N HNOHOOO .OHNNO
NO NO NO NO NO ON ON OO OO OOOONO>HH
"HNNNNNO NO NO NNONOONNN NNNNO
OOON HONN NNNN OONH OOOO ONHN ONNN ONOH OONN «NNHHH>
ONN ONN ONO NOH NOO NON ONO NOH OON «OOO ONOOH
"NozHOHHON
ONN NNO OON OOO NON NNN HHN ONN HOO NNNOH
NHH NO NNH OO HNH NOH NHH OO NNH NOHONOOOOON
NN ON NN HO NN OO HO ON NN NHOON New:
NNN OOO NNN OOO NNH ONN OON ONN NNN nemaONzON NaOuO
OO NN NNH OO OOH NN HNH OOH NO Naon
“HONHOOO HONONHOOHNOO
OOON ONNN OOON ONNN NNNOH HNNN NONN ONON HOON NNNOH
ON N OO NO HN O OO OO NNH umeuo
NNH ON NON NNH ONH NO NON OON OO ONONOOO
NN NO ON HO HN ON HO HN ON NNOHOON
NNN ONN ONN NON OHN NNN NON ONH NNO NONOO
OONN ONON ONNO NNNH ONOOH ONNN HOHN OOON NNNO «Hausa
“NOONNN>HH
OHNO NOHN NHNHH NONO ONOOH OONO NNHHH NNNN NONHH NNNNNNNO NNNOH
HHN azNOzNONO zzo NN-O +ON NHOzNN NNO: NNOONOz onNNONN
NNNOOe NNONO NNNNNO NHNNOO ONNN NO NNN NOOHHH>
ONuNNOH .HOHONO NNNNNO zm<m NO NOH<> "N.N NHNON

62
female-headed nor draft-dependent households had fewer building assets.
Female-headed households had as much invested in housing as male-headed
households--despite fewer cattle and lower incomes--due to inheritance,
family help, and remittances.

Agricultural capital was an insignificant proportion of household
assets. (Other equipment can only partially be treated as agricultural
capital since it included carts and, for a few households, vehicles
which had non-agricultural uses.) Households with fewer cattle did not
have fewer hand tools and receptacles but did have less invested in
plows (reflecting few plows). Plows are the single most important
implement in arable production and investments in plows are less
divisible than are those in hand tools or receptacles. A positive
correlation between cattle assets and plow ownership also was found in
the RIDS [Lucas, 1985].

C . LABOR ALLOCATION

This section presents an overview Of household labor use by
activity, month and age-gender category. Where applicable, findings are
compared to those from the RIDS, recently published in Chernichovsky,
Lucas and Mueller [1985], and a labor use study carried Out in Shoshong
in 1970 (reported in Bond [1974]). A comparison is useful since both
studies are now more than ten years Old. Also, the RIDS labor data were
collected from 1,000 households but were based on only five days per
year per household.

Household labor allocations were monitored through the MVRU Survey.
There were three main Objectives: (a) to assess labor requirements for
crOp production relative to household maintenance and other income

generating activities, (b) to characterize the gender division Of labor,

63

and (c) to examine seasonality in labor use.
1. HOUSEHOLD LABOR USE

An overview of household labor use is summarized in Table 3.6.
Note that the percentages in the table do not take into account leisure,
school and child care-~each Of which would account for a lot of time if
recorded [Mueller, 1985].

The single most time consuming activity was tending livestock,
accounting for over a quarter of the hours worked by household members.
However, taken together, household maintenance activities took more than
forty percent of the total time recorded. Nearly 1,100 hours per
household, or an average of three hours per day, were spent on cooking
alone. Beer brewing accounted for a surprisingly small share of time,
slightly more than half the amount of time spent in wage employment.

Fieldwork accounted for a only 15 percent of household members time.
The most time consuming field activity was birdscaring, but birdscaring
often was combined with other field activities (so separate hours were
not recorded for birdscaring). The most time consuming "active" task
was plowing and planting, followed by harvesting and fencing.

A short length of time usually was spent on any single activity on a
given day but several different activities Often were done each day.
This was particularly true Of household maintenance activities.
Fieldwork days averaged 3.75 hours, on days fieldwork was done. The
longest amount of time spent on a single activity during a day was for
wage employment, an average Of 7.3 hours per day on days when wage
employment was done. Thus, wage employment seemingly precludes doing
many other activities on the same day. This may in part explain why

women can fit crOp production activities into their heavy schedules,

64

TABLE 3.6: HOUSEHOLD LABOR USE, 1983-84

 

PERSON-DAYS PERCENT HOURS PERCENT
PER YEAR/a OF DAYS PER YEAR OF HOURS

 

(Per Household)

FIELDWORK: 199/b 8.9 747 14.9
Clear & Destump 8 38
Plow & Plant: 28 131
Weed & Thin/c 18 63
Birdscare/d 56 208
Gather Morogo 7 13
Other Harvest 35 116
Fix Threshing Floor 7 21
Thresh, Winnow & Bag 17 53
Fence 24 105
LIVESTOCK: 541 24.2 1362 27.2
Tend 443 1290
Milk 98 72
BEER: 93 4.2 261 5.2
Make 63 101
Sell 30 160
HOUSEHOLD MAINTENANCE: 1332 59.5 2125 42.4
Gather Firewood 154 189
Fetch Water 545 531
Cook 431 1097
Wash 159 232
Construct & Repair 43 76
OFF-FARM: 73 3.3 513 10.3
Other Fields 10 54
Wage Employment 63 459
ALL 2240 5009

 

Source: 1983-84 MVRU Survey; 27 households.

a. Each person working during a day counted as one
person-day.

b. Numbers and percentages do not total due to rounding.

c. Includes a small number of hours spent hilling jugo
beans.

d. Birdscaring time was not recorded. To account for
birdscaring, 3.75 hours (the average length of an active
fieldwork day) was added for each birdscaring person-day.

65
while men are less able to do so with their lighter schedules.
2. GENDER DIVISION OF LABOR

Table 3.7 shows the average hours per household spent on income
generating and household activities by men, women, boys and girls. The
table shows there was a strong gender division of labor. There was also
somewhat of an age division of labor.

Men and boys were responsible for most Of the livestock tending and
milking labor, plowing, and field maintenance (mainly destumping and
fencing). Boys actually spent more time tending livestock and plowing
than did men. Boys also helped with household maintenance activities
more than did men. Men, however, spent nearly twice as many hours in
wage employment as did any other worker category.

WOmen worked more hours on more activities than either men or boys.
On average, women were responsible for more than three-quarters of
household maintenance time, sixty percent of weeding time, two-thirds of
birdscaring time, eighty percent Of harvesting and threshing time, and
essentially all the time spent brewing and selling beer. Girls
primarily did the same activities as women, but accounted for less than
ten percent of the time recorded. Girls contributed significant amounts
of labor only to gathering firewood and water, and cooking and washing.
NO categories of labor spent much time on construction and repair
activities.

In total, females worked 37 percent more hours than did males.
Since most of the activities not recorded are primarily done by females,
the gap between males and females is much greater than this. On an
individual basis, women worked the greatest number of hours, nearly

1,200 hours per active woman. This was more than sixty percent more

66

TABLE 3.7: HOUSEHOLD LABOR USE BY
AGE-GENDER CATEGORIES, 1983-84

 

MEN WOMEN BOYS GIRLS

 

(Hours)

FIELDWORK:

Plow & Plant 48 20 61 l

Weed & Thin 12 40 6 6

Birdscare/a 23 143 11 34

Harvest (inc. Morogo) 11 103 4 11

Field Maintenance/b 103 32 26 2
LIVESTOCK:

Tend 468 72 732 18

Milk 18 7 46 l
BEER:

Make 1 98 c 1

Sell c 159 1 c
HOUSEHOLD MAINTENANCE:

Gather Firewood 37 77 45 29

Fetch Water 55 306 48 122

Cook 3 895 24 175

Wash 3 197 10 23

Construct & Repair 11 57 5 3
OFF-FARM:

Wage Employment 261 154 25 19

Other Field 14 28 8 4
TOTAL:

Per Household 1071 2435 1052 453

Per Active Person/d 729 1188 612 246

 

Source: 1983-84 MVRU Survey.

a. Birscaring hours calculated as in Table 3.6.

b. Includes clearing, destumping, fixing threshing
floor, and fencing.

c. Less than one.

d. Active people were defined to be the number Of
residents in each category who were not
inactive due to age or health (determined on
the basis Of the Household Census). The number
of individuals actually contributing labor
fluctuated as individuals moved in and out of
the household on a temporary basis.

67
hours than worked per active man. The low number Of hours recorded per
active girl mainly reflects a failure to record hours for two of their
main activities, childcare and bush gathering.

The findings on the gender division of labor generally support
those from the RIDS [Mueller, 1985], the early research in Shoshong
[Bond, 1974], and the NMS [C80, 1982]. In those studies, males
accounted for most of the livestock tending labor and were relatively
more engaged in wage employment than females. Females on the other
hand, accounted for nearly all of the household maintenance activities
and a majority of the fieldwork.

3. SEASONAL PROFILE OF LABOR USE

Table 3.8 presents a monthly profile of household labor use by
activity. Aside from the amount Of time devoted to construction and
repair, the largest monthly variation in hours worked was for fieldwork,
as would ordinarily be expected. Time spent on beer brewing and wage
employment also varied quite a bit over the year but generally in a
counter cyclical pattern to fieldwork. As a result, total labor inputs
varied less over the year than was the case for any single labor
activity.

With the exception Of high figures for a single month, January,
there was not much difference in household maintenance labor by month.
Excluding January the coefficient Of variation of monthly labor inputs
for household maintenance was only around fifteen percent. Mueller
[1985] also concluded there was a lack Of significant seasonality in
labor requirements for household maintenance activities.

The MVRU Survey data showed that household maintenance labor inputs

increased slightly during the cropping season. This was because the

68

TABLE 3.8: HOUSEHOLD LABOR USE BY MONTH, 1983-84

 

 

 

FIELD- LIVE— WOOD/ COOK/ CONST./

WORK STOCK BEER WAGE WATER WASH REPAIR ALL
NOVEMBER 50 102 12 44 57 94 2 361
DECEMBER 103 95 6 12 46 95 2 360
JANUARY 93 138 6 18 117 180 5 559
FEBRUARY 52/8 102 25 68 71 142 10 469
MARCH 65 159 20 62 55 114 3 478
APRIL 106 139 22 47 51 114 2 481
MAY 126 186 31 69 56 109 5 581
JUNE 108 110 19 18 52 86 5 397
JULY 38 92 24 18 43 89 2 306
AUGUST 31 87 3o 30 52 100 4 333
SEPTEMBER 24 76 31 31 70 104 9 345
OCTOBER 6 75 36 44 50 103 26 339
79131115; """" a; 114 22‘ 38 NA 111/c 6 .1;
S.D. 39 35 10 20 20 27 7 93
c.v. (Z) 59 31 45 53 33 24 105 22

 

Source: 1983-84 MVRU Survey.

a. February through June includes birdscaring days at 3.75 hours
per day.

b. Without January, the mean, s.d., and c.v. for gathering
firewood and fetching water are 5.5, 8.7 and 16%.

c. Without January, the mean, s.d., and c.v. for cooking and
washing were 104, 15.7 and 152.

69

time required for cooking, washing, gathering firewood and fetching
water increased as the households split (some members going to the lands
and some remaining in the village). In contrast, the RIDS showed that
the time spent on household activities decreased during the crOpping
season. The difference may stem from the seasons monitored. The RIDS
was carried out in a season with good rainfall, so crOpping labor
requirements may have forced a reduction in household maintenance labor.

The MVRU Survey also showed that livestock tending labor did not
change much on a monthly basis but did tend to increase at the same time
that fieldwork labor was the greatest. The 1985 Livestock Practices
Survey confirmed that this stemmed from a need increase the intensity Of
herding during the crOpping season in order to reduce liability for crop
damage.

When examined by age-gender categories, the monthly profile of
hours spent by women on household maintenance and non-cropping
activities showed the least variation (a c.v. of 21 percent) while hours
spent by boys showed the greatest variation (a c.v. Of 39 percent). NO
particular month or period could be identified in which labor
requirements were substantially higher than average monthly
requirements. Moreover, there were no discernible seasonal trends.

4. FIELDWORK HOURS BY HOUSEHOLD TYPE

Only fieldwork hours were analyzed by household type. The analysis
showed that Makwate, male-headed and draft-controlling households worked
substantially more hours on fieldwork than Shoshong, female—headed, and
draft-dependent households, respectively. The difference was due to the
area planted, since labor inputs per hectare were similar. Shoshong

households spent less time on plowing, due to use Of tractors, but

70

Makwate households spent more time on the other field activities. The
results on the relative hours of male-headed and female-headed
households differ from those of the RIDS, which showed that
female-headed households did more fieldwork [Lucas, 1985]. The
difference in results reflects the difficulties female-headed households
face in getting an adequate area plowed during drought seasons.

D. CASH REVENUES AND EXPENDITURES

This section gives an overview of cash revenue and expenditure
patterns. The primary Objectives are to describe: (a) the role of crOp
sales relative to other cash sources, (b) the importance Of food grain
purchases relative to other expenditures, and (c) differences in
cashflow patterns by household category.

1. CASH SOURCES, 1983

A preliminary profile of cash sources was generated through the
1983 Crop Management Survey. Findings are summarized in Table 3.9.

Most respondents said there were several sources of cash. A majority of
households had three or more sources, from both the lands/village area
and outside the village area.

CrOp production played little or no role in generating cash. Few
respondents said crop sales were a source Of cash at all. Selling
cattle was reported to be the main source Of cash. Following cattle
sales, selling beer, remittances, and village wage employment were
ranked (in that order) as the principal sources of cash.

There were several differences in the relative importance of cash
sources by village and household type. Cattle sales were particularly
important for male-headed and cattle-rich households. Female-headed and

cattle-poor households said they relied primarily on beer sales and

71

TABLE 3.9: PRIMARY INCOME SOURCE, 1983

 

SELL SELL REMIT- WAGE SELL SELL
CATTLE BEER TANCE LABOR CROPS OTHER

(Percent of Households)

ALL 36 23 22 17 1 3
VILLAGE:
Shoshong 46 25 18 7 0 3
Makwate 21 19 27 29 2 2
SEX OF HEAD:
Male 51 13 12 21 0 3
Female 13 36 36 10 2 2
CATTLE ASSETS:
0-15 8 34 28 23 2 5
16+ 67 9 15 10 0 0
DRAFT ACCESS:
Control 42 19 16 17 2 0
Depend 22 27 29 16 0 6

 

Source: 1983 Crop Management Survey.

remittances. Remittances were the primary source of cash for
draft-dependent households. Village wage labor was the primary source
Of cash for more than twenty percent of male-headed, Makwate and
cattle-poor households, compared to ten or less percent for their
comparison households.
2. CASHFLOW ANALYSIS, 1984

Data on cash revenues and expenditures were collected during the
first two years of the MVRU Survey. These data were used to: (a) verify
results from the CTOp Management Survey, (b) evaluate the importance Of
food expenditures relative to other expenditures, (c) assess
relationships between cashflow patterns and household characteristics,
and (d) identify seasonal cashflow patterns. Table 3.10 presents an

overview Of average monthly cash inflows and outflows for the MVRU

72

.OOHuoouu mcHuN: HON van NOON movsHocH .O

.NO>N=N ON>z ON-NNOH ”wousom

 

OH.Ou ON.OH- NN.OH ON.NN- OO.OH NO.N OO.N- OOHNNNOO NNN
NN.NOH NO.NN OO.HOH NO.NO OO.NOH OO.OO OH.ONH NNNONHONNNNN HONON
OH.NN NO.NN OO.ON NN.ON ON.NN NO.OH NO.NN Nesso
NN.N HN.H NO.O ON.N OO.O NO.H ON.O ONONNOONN
ON.NN NN.NH ON.NO OH.ON NN.NN NH.O N0.00 NNNONN:
ON.O NO.O NO.O NO.O NO.O OO.O NO.O OONOH OON NONHO
. ”NNNONHONNNNN .ONHz
NN.H ON.H OH.H NN.O NN.H ON.O OO.H Condo
OO.OH OO.HH ON.NN OO.O NN.OH NN.N NO.HN NOOOO OHOOooaoN
OO.N NH.O NN.O OO.O NN.N OO.O NH.N NuoOOO>HH
OH.OH NN.HH ON.NH NO.NH HO.OH OO.O NN.NH OOON NOO8O
NN.HH HH.OH OH.NH HO.NH OO.OH NN.O NH.OH Has: New OHNNO
NO.O ON.O OO.N ON.N OH.N NN.H OH.O NusaeH
"NNNONONON
NH.NOH NN.OO ON.NNH NH.NN NO.ONH OO.NN NN.ONH NNOzN>NN HONON
ON.N NN.H O0.0 OO.H NN.O HO.O NO.N NOOuO
NO.OH OH.NH NO.NN OO.N HO.NN NN.O NO.HN NONN:
ON.NN ON.NN ON.NH ON.ON NO.ON NN.NN NO.OH OONOH OOO NONNO
"NNOzN>NN .ONHz
OO.H NO.O ON.N NN.O NN.H ON.O OO.N NONOO
OO.NH OO.NH NN.ON ON.ON O0.0 ON.H NO.NN NOON
OO.ON OH.OH HH.HN OO.N NH.NN NN.NH ON.NN Nooomo>NH
ON.O ON.O HN.O OH.H OO.O NO.O OO.H OaONO
”NN.H«N
HHO NN-O +ON NHOzNN NNO: NNOONON ONONNONN
NNNNNO NHNNOO NONN NO NNN NOOHHH>

 

NOHONO ONOH .NzOO u NNOH .NNNzN>oz .N3OHNNN<O NHmNzoz NO<NN><

HOH.m mqm<a

73
Survey participants during the period November 1983 to June 1984-~the
cropping season.

Overall, approximately P60 was received each month via sales, and a
total of just over P100 was received from all sources. Shoshong,
male-headed, cattle-rich, and draft-dependent households all had
substantially greater cash inflows than did their comparison household
types. Female-headed households had lower cash incomes in Shoshong in
the early 19708, as well [Syson, 1972].

Livestock sales were the single largest source Of cash, averaged
across all households. Shoshong East, male-headed and cattle-rich
households all received more cash from livestock sales than their
comparison groups. CrOp sales provided very little cash, as would be
expected in a drought year. Beer sales were an important source of cash
for several households, particularly for female-headed households in
Shoshong. Cattle-rich households Obtained more cash from beer sales
than did cattle-poor households, because poorer households could not
afford to buy the inputs required to make beer.

Gifts and loans, primarily remittances from non-resident household
members, were a major source of cash for all categories Of households.
Gifts and loans accounted for nearly half of each month's cash inflow
for Makwate and cattle-poor households. The relative contribution of
remittances was less for Shoshong, male-headed and cattle-rich
households than their comparison groups. Village wage cash incomes were
important, particularly in Shoshong and for male-headed and richer
households.

Turning to cash outflows, purchases of grain and meal, other food

and households goods were substantial for all types of households.

74
Relatively little was spent on inputs and livestock. Inputs
expenditures were higher for female-headed households, reflecting
purchases of grain for making beer.

The largest single expenditure for Shoshong, male-headed,
female-headed, and cattle-rich households was for traction hire,
included in the wages category. Expenditures for traction hire were
particularly high for male-headed and cattle-rich households, averaging
P40-50 per month, but were relatively the greatest for female-headed
households. Traction fees were not the largest expenditure only for
cattle-poor and Makwate households, farmers who did not hire tractors.

The positive association between traction hire fees and cattle
assets differs from results of the RIDS, in which cattle-rich households
(who had the ability to hire) spent less on plowing services than did
poorer households [Lucas, 1985]. There are two likely explanations: (a)
during drought, richer households are better able to mobilize the cash
required for hiring and therefore spend more on plowing services, and
(b) there may be a greater preference for traction hiring in the 19808
than there was in the 19708. The second factor is at least partially
true because several traction owners chose to hire traction even after
they said their animals had recovered.

The total level of expenditure in Makwate was only a third of that
in Shoshong. Cattle-poor households spent just over half of that spent
by the cattle-rich households. There was surprisingly little difference
in expenditures by male-headed and female-headed households, despite a
large difference in revenues. The net cashflow per month was negative
for Shoshong, female-headed and cattle-poor households--and for all

households combined. However, there was a positive net cashflow of over

75
P25 when one excludes traction hire fees.

An analysis of cashflows by month showed that beer revenues and
expenditures on grain, other food and households goods were quite
regular. Cash revenues from livestock sales were concentrated in a
single month, April, when five households made large cattle sales.
Remittances were high in November and October, fell to around ten Pula
during most of the season and then increased to above thirty-five Pula
during May and June, when it had become apparent there was going to be
widespread crOp failure. There was less monthly variation in
expenditures than for revenues.

E . FOOD CONSUMPTION

Over the past few years, there has been a growing appreciation of
the importance of a food consumption perspective in on-farm research
[Frankenberger, 1985]. Since most farm production is consumed by the
producing household, consumption requirements and food preferences exert
a major influence on farmers' responses to suggested innovations.

Farmer welfare can be improved by targeting arable farming interventions
to problems areas in the diet, even if household cash income is not
increased.

This section presents selected findings on the main dishes and
other foods consumed in Shoshong and Makwate, the frequency of with
which the major food items were consumed in 1984-85, and the primary
sources of food items.

1. MAIN DISHES AND FOOD ITEMS

Farmers were asked to identify and rank their main dishes and other

foods in the 1983 Crop Management Survey. The resulting rankings of

main dishes and other main foods are presented in Tables 3.11 and 3.12.

TABLE 3.11: FARMER RANKINGS OF MAIN DISHES, 1983

76

 

DISH/a DESCRIPTION/b

SHOSHONG

MAKWATE

ALL

RANK POINTS/c RANK POINTS RANK POINTS

 

LEFATA
Sorghum or Maize &
Cowpeas or Jugo Beans
BOGOBE
Sorghum & Morogo or
Meat or Milk
THOPI
Sorghum & Melons
PHALECHE
Maizemeal & Morogo or
Meat or Milk
MOGODU
Cowpeas & Melons
MOKGANIYANE
Sorghum Cooked With
Milk
SETAMPA
Semp & Meat or Oil
or Milk

7

91

100

93

35

19

3

5

100

43

30

46

42

37

100

81

61

35

32

21

18

 

Source: 1983 CrOp Management Survey; 116 farmers.

a. Dish names vary within the Mahalapye area and across the
country. The most common names in Shoshong and Makwate

have been used.

b. There can be variations in ingredients for each dish.

For

example, some farmers use ground melon seeds in several
dishes; other use stamped maize rather than unstamped; some
use fresh milk and some sour milk, etc.
c. Farmers were asked to cite their four main dishes in order
of importance. Points were given in reverse order, with
four points for the most important dish, three points for

the second dish, etc.

differences due to sample sizes.

Points were then transformed to a
basis Of 100 for the top ranked dish, to eliminate

77

TABLE 3.12: FARMER RANKINGS OF OTHER MAIN FOODS, 1983

 

 

SHOSHONG MAKWATE ALL
RANK POINTS/a RANK POINTS RANK POINTS
WATERMELONS 1 100 4 4o 1 100
MELONS 3 83 3 52 2 96
MEAT 6 56 2 78 3 94
BEANS 8 26 1 84 4 76
MILK 2 88 7 12 5 72
SWEET REED 4 63 10 0 6 46
CUCUMBER 5 58 9 3 7 44
PUMPKINs 7 39 8 5 8 32
GREEN MAIZE 9 11 5 33 9 30
BREAD 10 7 6 32 10 26

 

Source: 1983 Crop Management Survey.

a. Farmers cited their four "other main foods" in order
of importance. Points were calculated as described
in Table 3.12.

Sorghum, maize and cowpeas were the primary constituents of most
dishes. Most dishes included leafy greens (referred to by the general
term "morogo"), milk or meat. Milk was consumed fresh or sour and was
in some cases cooked into the grain meal while in other cases was added
after cooking. Thinner porridge is generally consumed with milk for
breakfast while thick porridge with a relish was consumed later in the
day. The relative importance of dishes was different between Shoshong
East and Makwate even though the same ingredients were used in both
villages.

Watermelons and melons were viewed by most farmers as important
foods. Both were consumed in large quantities beginning in March
through the end of the season. Meat and beans, such as jugo beans, were
important food items aside from their role in the major dishes. Milk,

sweet reed (a sweet sorghum species eaten like sugar cane), cucumbers

78

and pumpkins were highly ranked foods in Shoshong but not Makwate. The
reverse held for green maize and bread.
2. FOOD CONSUMPTION FREQUENCIES

The source Of information on consumption frequencies was the
1984-85 Activity Survey. Respondents were asked on a monthly basis how
frequently they consumed 19 major food items. Categorical responses
were recorded, distinguishing between "not at all during the month,"
"one to three times during the month," "two to three times a week,"
"four to five times a week," and "everyday or nearly everyday." Table
3.13 summarizes findings for October 1984 to September 1985. The
figures refer to household-months. There were 554 Observations, an
average of just over 46 responding households per month.

The most frequently consumed food was maize meal, followed by
sorghum meal. Sorghum meal was not consumed at all in a surprisingly
high number of household-months. Bread and vegetables were eaten at
least occasionally in many household-months, but green maize, rice, and
cowpeas were infrequently eaten. Milk was the major source of animal
protein, although it was consumed in small quantities. Chicken meat and
eggs were consumed infrequently or not at all by nearly all the
households.

Seasonal differences in consumption were most pronounced for animal
products (except goat meat), lands morogo, and vegetables. Beef was
least frequently consumed during the January to March period, at which
time no beef was consumed in nearly half of the household-months.
Except for the January to March period, beef was eaten more than once a
week in nearly a third of the household—months. This is certainly high

according to African standards. Goat milk consumption was concentrated

79

TABLE 3.13: FOOD CONSUMPTION FREQUENCIES,
OCTOBER 1984 TO SEPTEMBER 1985

 

A7a B C D E F

 

(Percent of Household-Months/b)
GRAIN AND MEAL:

Sorghum Meal 22 10 6 4 4 53
Maize Meal 16 9 10 5 4 56
"Stamps" 47 35 10 6 1 0
Green Maize 91 8 c c 1 c
Bread 24 32 17 16 7 4
Rice 74 21 4 1 c 0
Cowpeas 62 25 8 3 2 c
OTHER CROPS:
Lands Morogo 50 18 7 8 8 9
Other Morogo 90 3 1 2 1 3
Watermelons 83 5 3 3 3 4
Starchy Melons 91 3 2 1 2 1
Vegetables 45 21 11 12 7 5
ANIMAL PRODUCTS:
Beef 29 37 9 9 9 8
Goat/Mutton 37 24 14 13 7 5
Chicken 67 30 2 l 0 1
Cow Milk 57 7 3 6 4 23
Goat Milk 44 2 1 2 1 49
Eggs 77 18 1 l c 1

 

Source: 1985 Activity Survey.

a. A I not at all during month; B = one to three times
during the month; C I once a week; D - two to three times
a week; E - four to five times a week; F a everday or
nearly every day.

b. Based on 554 responses, an average of 46 households per
month.

c. Less than one.

80
between October and March. Cow milk was relatively more important later
in the season. When cows were milked, three to four times as much milk
was consumed as compared to when goat milk was consumed.

Consumption of lands morogo (leafy greens) was concentrated in the
April to June period. Consumption was possible outside of that period
because farmers often boil and then dry the morogo so it will keep. .
Winter vegetables, available in October through December, were consumed
more frequently by more households than were summer vegetables. 5

Consumption of the major cereal grains increased only slightly as

 

the season's harvest became available. This was because maize and E
sorghum meal were purchased by most households. The frequency of maize
meal consumption dropped off beginning in April, just as consumption of
home produced sorghum increased.

Shoshong and cattle-rich households consumed most items
significantly more frequently than did Makwate and cattle-poor
households, respectively (as determined by Chi-statistics from a cross
tabulation analysis). There were few significant differences in
consumption frequencies associated with either gender of household head
or draft access.

3. SORGHUM AND MAIZE MEAL CONSUMPTION LEVELS

During the 1985 Activity Survey, respondents were asked to estimate
how many kilograms of sorghum and maize meal were eaten each month by
all the adults and children who ate with the household on a daily basis.
The numbers of adults and children eating daily were also recorded.

Just over 50 kgs. were consumed per household per month. This gave an
average of approximately 7.5 kgs. per consumer each month (or about

one-quarter kg. per day.) Shoshong, male-headed, draft-controlling, and

81
cattle-rich households all consumed greater amounts of meal, both in
total and per consumer (around one-third to one-half kg. more per
consumer per month).
4. SOURCES OF FOOD ITEMS

In the 1985 Activity Survey, respondents also were asked to
indicate which of the following was the primary source of each food item
during that month: a relative, another farmer, a local trader, a trader
in another village, a government feeding program, or household
production.

In more than 90 percent of the household-months, traders were the
primary source for maize meal, bread and rice. Green maize was obtained
primarily from other farmers. Despite the dominance of sorghum in farm
production, household production was the primary source of sorghum in
only 31 percent of the household-months. This of course reflected the
influence of the drought. With reference to other crops, melons and
watermelons were Obtained almost exclusively through home production or
from other farmers. Most vegetables, on the other hand, were purchased
from traders.

Animal products generally were obtained from home production or
from other farmers. In most cases, farmers sell some beef or goat meat
when they slaughter animals. Across a village, a few individuals might
slaughter animals on any given day. In this way, many peOple can eat
small quantities Of meat. Farmers also Often sell each other small
quantities of milk. Chicken and eggs were Obtained almost exclusively
from home production.

A number of differences were identified in the sources of food by

village and household type. Overall, Makwate households tended to rely

 

82
on home production or other farmers more than did Shoshong households.
Female-headed households relied less on home production of most
commodities, relative to male-headed households, as did draft-dependent
household relative to draft-controlling households, and cattle-rich
households relative to cattle-poor households.
F. CONCLUSIONS AND IMPLICATIONS

The chapter had two main Objectives. One was to provide a
comprehensive profile of household circumstances, to place arable
farming into a broader household and farm system context. The second
was to identify significant differences in household circumstances for
four distinct household categories: households in Shoshong versus
households in Makwate, and male-headed versus female-headed, cattle-rich
versus cattle-poor, and draft-controlling versus draft-dependent
households.

The review of household circumstances showed that household and
farm system resources are limited and maldistributed. A number of
resource constraints were identified, any combination of which might
impinge on a given household. Some households had too little labor,
while others lacked implements, and still Others had too little land-~at
least during seasons with good rainfall.

Female-headed, cattle-poor and draft-dependent households
consistently had less develOped land resources, fewer capital assets and
fewer livestock than their counterpart household categories. These
households also tended to have less cash to spend, to be more dependent
on remittances, to eat less grain and meal per consumer, and to be more
dependent on traders and other households for their food. Consequently,

there was a "pattern of vulnerability" which extended beyond the issues

83
Of access to cattle or male labor, as identified in the RIDS [080, 1976]
and the NMS [080, 1982], respectively.

Significant differences in resources also existed between
households in Makwate versus households in Shoshong. Makwate households
had fewer cattle and, because they were generally poorer, had less cash
to spend and ate less food. At the same time, households in Makwate had
as much invested in plows and hand tools, and had better developed land
resources. As shown in Chapter IV, households in Makwate were more
strongly oriented toward arable farming than were households in
Shoshong.

The chapter showed that the allocation of existing resources to
crop production was further constrained by household maintenance
requirements and the allocation of resources to other production
enterprises. Most of the available supply of labor was needed for
households activities and livestock tending. Cash was needed to buy
food and other household goods. The top priority for investment was
cattle.

Because Of the variety of demands on household resources, even the
better Off household categories (male-headed, cattle-rich, and
draft-controlling households) had severe problems. These households
generally had small amounts invested in agricultural implements. Most
had negative cashflows, largely due to hiring traction. They were
unable to eat most food items any more frequently than did the poorer
households.

Finally, the chapter revealed a potential conflict between national
and farmer goals for crop production. While the government would like

the country to be self-sufficient in grain production, the food security

 

84
of farming households depends on their ability to generate cash from
non—cropping sources (including remittances). If farmers increase their
investments in crOp production, the potential for crop failure means
that the risk Of food shortfalls could actually increase [Vierich and
Sheppard; 1980].

Farmers obviously are aware of the riskiness Of crOp production
and, as a result: (a) minimize their investments in fixed capital, (b)
try to rely on retained seed and household labor rather than purchased
inputs, (c) invest minimal amounts of labor in plowing and planting
(through broadcasting seed mixtures), (d) only invest labor in plant
protection after it can be seen that there is reasonably good stand
establishment and plant development, and (e) give higher priority to
wage employment and the health of their cattle than they do crop
production. With their current whole household risk management
strategy, farmers do not lose so much in drought years and can still
approach self-sufficiency in good years.

The analysis of household circumstances presented in this chapter
suggests the following priorities and targeting strategies for the
Ministry of Agriculture.

First, the most useful criterion for distinguishing household types
would appear to be the division based on draft access. The causal
relationships between draft access and cropping outcomes are clearer
than are those between gender of head, cattle assets and crop production
results. Moreover, the Ministry can be confident that any interventions
developed for draft-dependent households will also address the needs of
most poor and female-headed households.

To address the requirements of the draft access research domains,

85

two sets of modified practices should be developed; one set assuming
that control of timing and multiple tillage operations are possible, the
second set assuming they are not possible. For example, technical
research for the draft-dependent RD could focus on remedial stand
management interventions (such as hand gap filling) and crop protection
interventions, since timely and sufficient plowing cannot be assumed.

Second, because there were a variety of resource constraints
besides draft access, alternatives need to be developed for each
traction RD, taking into account different resource constraints. To
keep targeting from getting to be too complicated, researchers should
give top priority to labor constraints for both RDs. A second priority
for the draft control RD would be interventions to address land
shortages.

In addressing various labor and land constraints, attention should
be given to production practices which do not require additional
implements or cash investments. Many households cannot reasonably
divert the cash needed for food into crop production and any required
investment in even one new implement could easily double or more the
value of capital tied up in arable production.

Third, instead of focusing just on the poor (as many consultants
have recommended), the Ministry must develop technologies and programs
to help all household types. Essentially all the households are poor
and need help. The different household types just need different kinds
of help.

Fourth, units of analysis other than the household might be
appropriate for targeting some research and extension activities. For

example, the Ministry should take advantage of the fact that households

86
in Makwate use donkeys and appear to be more willing to invest in arable
farming, given their resources, than do Shoshong households. As an
alternative to "village targeting," the Ministry should consider
targeting particular worker categories when assessing technological
Options. In this context, steps to alleviate the existing work burden
of women should be a top priority.

Fifth, planning and policy analysis needs to pay increased
attention to household circumstances. For example, the Ministry should
have been able to predict that ARAP (which provided a one-time
Opportunity to have the government pay for up to ten hectares of
plowing) would have a biased impact in favor of male-headed and
cattle-rich households. Another example Obtains from the food
consumption analysis. Because maize, vegetables and wheat are consumed
regularly, the Ministry needs to evaluate the macro economics of
different approaches to increase their supply (import, Batswana

commercial farmers, diversified household production).

 

IV. CROP SYSTEMS MANAGEMENT

This chapter presents an overview of crop systems management.
Information is given on crOpping objectives, major crops and varieties,
draft management, traction use patterns, cropping practices,
post-harvest practices, and roles in decision making. The objectives
are to: (8) describe existing production and post-harvest practices, (b)
characterize farmers' priorities and perceptions, and (c) synthesize
implications for arable farming deve10pment priorities.

A. CROPPING OBJECTIVES

In the 1985 DUMI Study, respondents were asked to list their
primary cropping objectives. Every respondent said the primary
objective was to grow food for household consumption. Several
respondents did not mention any other objective. The most commonly
mentioned secondary Objectives were to: (a) secure a diverse diet, (b)
produce a surplus of sorghum to sell, (c) generate cash by growing and
selling cowpeas or jugo beans, and (d) minimize the risk of complete
crop failure (by growing a mixture Of crops).

To verify the listing of Objectives, farmers were asked to rank the
relative importance of prices, consumption requirements, and seed
availability in determining how much to grow of sorghum, maize, and
cowpeas. All the respondents but three said that household requirements
were the most important determinant of the amount of sorghum grown.

Seed availability was cited second by nearly all the respondents. Only

87

 

88
four percent said that price was one of the tOp two considerations.
Almost sixty percent said they pay no attention to prices when making
decisions about sorghum.

Household consumption requirements were the most important
determinants of maize and cowpeas production as well, but prices played
a relatively greater role than was the case with sorghum. For maize,
seed availability was the second consideration, but nearly twenty
percent of the respondents said price was the first or second
consideration in growing maize. Nearly half said price was the most
important determinant in growing cowpeas. Cowpeas are the closest to
being a cash crop in the Central Region.

B. CROPS AND VARIETIES

To develop guidelines for on-station technical researchers, data
were collected on: (a) the relative importance Of different crops and
varieties and (b) farmers' priorities with respect to varietal
characteristics.

1. MOST IMPORTANT CROPS

Information on the number of crops grown and farmers' views on the
relative importance of crops was collected in the 1983 Crop Management
Survey. Respondents said they grew an average of six crOps. Sorghum
was the most important, followed by cowpeas, melons, watermelons, maize
and jugo beans. Relative crop rankings were almost the same for farmers
in Shoshong East and Makwate, and for respondents from different types
Of farm systems (as identified in Chapter III). Few farmers mentioned
either millet or groundnuts.

Segaolane was the most important sorghum variety in 1982-83.

89
(Segaolane is an improved local variety of intermediate maturity.)
Goromente and Moterane sorghum varieties were nearly as important and
were preferred by many farmers for their taste in porridge and beer
brewing. (Both are long maturing improved, local varieties.)
Approximately an equal number of farmers grew Blackeye and Tswana
cowpeas. (Blackeye is a dual purpose-~grain and leaf--semi—erect and
semi-determinant variety with intermediate maturity. Tswana cowpeas are
a long maturing indeterminant, spreading variety.)

One problem encountered in asking farmers open-ended questions
about which crops are grown is their tendency to ignore minor crops. TO
generate a complete profile of crOps grown, farmers were asked in 1985
to indicate which of a specified list of crOps they had planted.
Sorghum, melons and cowpeas were grown by more than 90 percent of
households, followed by maize (75%) and sweet reed (71%). NO other
crops were grown by more than thirty percent of the households. Jugo
beans and tepary beans were grown by more than twenty percent but
millet, sunflower, mung beans and groundnuts were grown by only a few
farmers.

2. DESIRED VARIETAL CHARACTERISTICS FOR SORGHUM

Farmers were asked during the 1985 DUMI Study to identify the most
important sorghum varietal characteristics and to say what they like and
dislike about the current main varieties. The most commonly cited
characteristics were early maturity, large seed size, and heat and
drought tolerance. Lesser important criteria were palatability and
color, ease of threshing and stamping, disease and bird resistance, and
good storing qualities.

Nine major sorghum varieties were identified. Moterane, Goromente

 

III

90
and Marupantse were the main varieties grown in the area until the
government introduced Segaolane (from another part of the country).
Most respondents said the three traditional varieties taste good, thresh
and stamp well, store well, and are heat and drought resistant. Many
said they yield well and have big heads, particularly Goromente. The
biggest complaint against them was the length of time to maturity. The
main advantage of Segaolane is its relatively quick maturity. More
respondents than not said it is heat resistant and tastes good. Some
complained that Segaolane is susceptible to weevils when in store. A
few said they did not like its taste nor the way it cooks. Kromo,
Raigop, and Kanye were grown by only a small proportion of farmers.
Comments about these varieties generally were favorable, noting their
quick maturity, heat tolerance, and relatively big seeds.

Most criticisms were reserved for the Red Swazi variety, provided
by the government through the Drought Relief program. Many respondents
said it is not heat tolerant and is susceptible to pests. Several said
the porridge is tasteless. Others noted that the grain take a long time
tO ripen, leaving it susceptible to bird attacks. 8-D, another quick
maturing variety introduced by the government, was criticized for being
susceptible to heat.

C. DRAFT MANAGEMENT

A distinctive feature of crop production in Botswana is the
importance of animal and tractor traction. While the introduction of
animal traction is a major issue in many African countries, animal draft
has been used in the Botswana since the beginning of the century
[Pinga1i, Bigot, Binswanger, 1987]. Control of traction resources is

strongly associated with control of other resources (Chapter III), crOp

91
system performance, and the viability Of recommended practices (Chapter
V).

To understand factors affecting access to and use of traction
resources, an investigation was carried out on draft management in 1983.
Two main issues were investigated: (a) how are owned traction animals
managed, and (b) what are the rights and Obligations associated with
inter-household drafts arrangements.

1. MANAGEMENT OF OWNED TRACTION

Data were collected on: (a) how many animals are used, (b) how long
animals are used, (c) whether animals are fit to start plowing when the
rains begin, (d) the length of time spent training animals, and (e)
plowing for others. Each of these issues affects the timing and amount
of plowing.

Half of the households using cattle plowed with eight or ten
animals and the other half used six. Most of the households plowing
with donkeys used six animals but a third only used four animals. NO
households using cattle plowed with just four animals.

There was no consensus on how long animals should be used. The
most common response for cattle users was three to four years. Cattle
users usually took into consideration when animals were the proper age
to be sold for beef. Donkey users were not concerned about the issue
since selling donkeys was not anticipated.

As of the 1982-83 season, most respondents said their animals
usually were strong enough in to begin plowing in October. Half of the
respondents, however, qualified their answer by saying it depended on
whether rains had been good during the prior season. These views

directly affect the viability of forage crops.

92

Approximately a third of the respondents began plowing after no or
a single day of training. Another third said only two to three days
were required for training. In general, training did not delay the
start of plowing.

Just over half the households had plowed for relatives but only a
third had ever hired out. Respondents who had not plowed for relatives
usually said their relatives had their own draft or that they had never
been asked to plow. Of the households which had never plowed for hire,
most said they felt plowing for hire would not allow them to finish
their own fields. A quarter said they had never been asked and Others
expressed concern about overworking their animals. In general, farmers
did not like to plow for others but did so out Of a sense of Obligation.
2. INTER-HOUSEHOLD DRAFT ARRANGEMENTS

Control over traction resources affects the ability to do timely
plowing and multiple tillage operations [FAO, 1974; Oland, Alverson,
Cummings, 1980]. But control of traction is not a simple issue in
Botswana. There are several ways of getting access to traction, each
associated with different rights and Obligations [Curtis, 1972;
Alverson, 1978; Gulbrandsen; 1980]. The networks of draft arrangements
which link households, therefore, need to be considered when evaluating
cropping outcomes and production recommendations [Behnke and Kerven,
1983].

Despite the importance of draft arrangements, there have been few
systematic studies which address the rights and Obligations affecting
the control Of timing. Most of the available information is based on:
(a) ethnographic descriptions which discuss resource exchange patterns

in general terms or (b) case studies focused on the specific networks of

93
specific households (for example, Behnke and Kerven [1983]). To
complement the existing studies, information was collected in 1983 on
three key draft arrangements: (a) borrowing, (b) hiring, and (e)
cooperative plowing.
a. Use of Managed or Borrowed Traction

Of the entire sample, only seven households used borrowed or
managed traction. Out of this limited sample, traction animals were
Obtained from both relatives and neighbors, so loaning was not
restricted to family. In only one case was a gift given to or work done
for the household which provided traction. In four cases, no time was
specified as to when the animals were to be returned. In the remaining
cases, the loans were at least for the entire plowing period.

In six of seven cases, the person borrowing traction was free to
use that traction for hiring out. Respondents from five of those
households said they would not give any money to the traction owner if
they did hire out. In fact, in only one case did the traction owner
retain any say on how the traction was to be used. Even if an animal
died while under the control of a borrower, no compensation was due the
owner. In most cases, however, the carcass would either be given to or
shared with the owner.

b. Hiring Traction

Thirty-nine households had hired traction but only six households
had hired out traction. In twelve cases donkeys had been hired. Cattle
had been hired in another 12 cases. In the remaining cases tractors had
been hired. Tractors hire rates were more than double cattle hire rates
and three to four times higher than donkey hire rates.

In almost forty percent of the cases, the people involved in the

94
hiring arrangement were relatives. Nevertheless, the hiring of a
relative instead of a neighbor did not give the hirer any additional
rights. For example, in nearly all cases a complete payment was made at
the time of plowing, the person hiring out also provided the plow, and
the person hiring in provided food.

The time of plowing generally was scheduled by the person who
hired, usually in relation to the following set of rains. However,
plowing often was not done when scheduled. If there was too long of a
delay and soil moisture became insufficient, the person hiring did have

the right to say that the plowing should be delayed until another rain.

 

This was an important decision since replowing was rarely done if
germination failed after a first plowing. Replowing was more frequently
done if there was substantial early weed development, demonstrating that
the first plowing had been poorly done.

c. COOperative Plowing

Two-thirds of COOperative arrangements were with relatives. In
most cases, the motivation for the relationship was to gain access to
plowing resources, either labor or traction. In ten cases, however, the
motive was to help friends or relatives. Most of time, the provider of
labor initiated the COOperative arrangement. Nevertheless, most
respondents said that both cooperators benefited equally from the
arrangement.

Both providers of labor and providers of traction agreed that the
household providing traction had the major say over the timing of
plowing, although some households said this was a joint decision. Both
partners usually were involved in the decision as to whether soil

moisture was adequate on the day plowing was scheduled.

95

There generally was a fixed amount of time spent on each field
before shifting. Most frequently reported was three days or a week.

The main factors considered when determining how much plowing was done
each day were soil moisture, the condition of animals and the starting
time. All households reported that replowing would be done if there was
poor germination. Essentially all households also reported that if the
reciprocal plowing was not completed in one season due to a lack of
rain, the obligation would be carried out the following season.

D. TRACTION USE PATTERNS

Two traction use pattern issues were investigated: (a) trends in
traction use and (b) the timing of plowing during the season and
vis-a-vis rains.

I. TRENDS IN TRACTION USE

Data on traction use patterns were collected in the 1983 Crop
Management Survey. To evaluate trends in traction use, many Of the same
households were contacted again in 1985 DUMI Study. A summary of
findings is presented in Table 4.1.

Between 1980 and 1985, a majority of households in Makwate only

used donkeys and more than ninety percent did some plowing with donkeys.
In 1982-83 more than eighty percent considered donkeys to be their
primary traction but this fell to under seventy percent by 1984-85.
This was due to a forced substitution of tractors for donkeys, since
many donkey had died or were in too poor condition to plow. Tractors
were almost unknown in Makwate in 1982-83 but were the primary traction
of twenty percent of the interviewed households in 1984-85.

Cattle were the main type of traction in Shoshong East in 1982-83

but there was a dramatic increase in the use of tractors between the

 

96

TABLE 4.1: TRACTION USED AND DRAFT ACCESS,
1982-83 AND 1984-85

 

 

 

 

VILLAGE SEX HEAD CATTLE
SHOS. MAKW. MALE FEMALE 0-15 16+ ALL
(Percent of Households)

TYPES OF TRACTION; 1982-83:

Donkeys 7 77 30 45 52 18 36

Cattle 46 17 38 28 33 35 34

Tractors 34 O 25 13 11 29 20

Tractors & Cattle 13 0 3 15 3 13 16

Tractors & Donkeys 0 2 1 0 0 2 2

Cattle & Donkeys 0 4 3 0 O 4 4
TYPES OF TRACTION; 1982-85:

Donkeys 11 54 43 21 44 10 31

Cattle 14 8 4 17 6 20 12

Tractors 39 0 22 21 16 30 21

Tractors & Cattle 32 0 13 21 13 25 17

Tractors & Donkeys 4 8 4 7 9 0 6

Cattle & Donkeys 0 17 9 7 6 10 8

All Three Types 0 13 4 7 6 0 6
PRIMARY TRACTION; 1982-83:

Donkeys 4 83 37 38 51 22 37

Cattle 51 17 38 35 34 40 37

Tractors 44 O 25 26 15 38 26
PRIMARY TRACTION; 1984-85:

Donkeys 11 67 52 24 50 15 37

Cattle 29 13 13 28 13 35 21

Tractors 61 21 35 48 37 50 42
PRIMARY ACCESS; 1982-83:

Own, Borrow 51 78 72 48 55 70 62

Hire 39 16 17 47 31 26 29

COOp Arrange. 10 6 11 5 14 4 9
PRIMARY ACCESS; 1984-85:

Own, Borrow 29 54 65 21 28 60 40

Hire 68 33 26 72 59 40 52

Coop Arrange. 4 13 9 7 13 0 8

 

Soures: 1983 Crop Management Survey.
1985 DUMI Study.

97
1982-83 and 1984-85 seasons. In Shoshong East, less than thirty percent
of the households did not use tractors between 1980 and 1985 and by
1984-85 more than sixty percent of households considered tractors to be
their primary traction. The prOportion of Shoshong households which
considered donkeys to be their primary traction (while still small)
almost tripled between the 1982-83 and 1984-85 seasons. The use Of
donkeys as a supplementary source of traction increased even more
dramatically.

In the 1982-83 season, tractor use was equally common for
male-headed versus female-headed households. Fortmann [1981] also
reported no significant difference in the use of tractors by male-headed
versus female-headed households. By the 1984-85 season, however, nearly
half the female-headed households primarily relied on tractors, compared
to just over one-third of the male-headed households. In both the
1982-83 and 1984-85 seasons, a larger proportion of cattle-rich
households used tractors than did cattle-poor households. The share of
household using tractors increased for both household types between 1982
and 1985.

The shift in traction use was reflected in the primary means of
draft access. In 1982-83, two-thirds of the households primarily relied
on owned traction. By 1984-85, less than half the households used any
owned traction and nearly two-thirds hired at least some of their
traction. A majority of farmers considered hiring to be their primary
form of draft access.

The primary impetus for the shift in traction use and draft access
was the drought. Many farmers reported that their draft animals were

unfit for plowing. While draft dependency increased during the drought,

98
it is notable that so many farmers were able to continue plowing by
hiring traction. During prior droughts in Botswana, many farmers simply
stopped plowing. The ability of farmers to hire traction was
facilitated through government payments for plowing under two drought
relief programs.

The trend toward traction hire may not be completely reversed once
the rains recover. Many farmers now hiring tractors express little
interest in returning to animal traction. In addition, the continuing
migration of younger household members to larger villages and towns may
soon leave too little labor for many households to plow with animals.

It can also be expected that some households now hiring tractors will
shift to donkeys after the drought, rather than go back to cattle, since
several households have lost all or most of their cattle. Donkeys sell
for one-third to one-quarter the price of oxen, and fewer are used in
plowing.

2. TIMING OF PLOWING

Two notions of timing are important determinants of crOpping
outcomes: when plowing is done during the season and the timing of
plowing vis-a—vis any given planting rain (usually a rainfall of at
least 15-20 mm). November and December plantings have the highest
probability of success (although this may not hold in any given season).
Consequently, early planting has been a long-standing extension
recommendation. The Ministry also recommends planting as soon as
possible after it rains, to ensure there is adequate soil moisture for
germination (without having to rely on a post-planting rain).

A profile of the starting and ending months for plowing in the

1982-83 and 1984-85 seasons, and "usually" is given in Table 4.2.

99

TABLE 4.2: TIMING OF PLOWING; 1982-83 and 1984-85

 

 

 

 

START PLOWING END PLOWING
BEFORE AFTER BEFORE AFTER
DECEM. DECEM. DECEM. JAN. JAN. JAN.
(Percent of Households)
USUALLY
VILLAGE:/ab
Shoshong East 57 24 19 24 41 35
Makwate 90 7 3 21 69 10
DRAFT ACCESS:/a
Own, Manage 89 8 3 28 56 17
Hire, Coop 57 14 28 14 62 24
SEX HOUSEHOLD HEAD:
Male 77 15 8 2 46 26
Female 62 19 19 15 63 22
ALL 72 17 12 23 53 24
1982-83 SEASON
VILLAGE:/b
Shoshong East 37 23 40 37 40 23
Makwate 59 26 15 72 20 8
DRAFT ACCESS:/a
Own, Manage 61 22 17 57 31 12
Hire, Coop 24 29 47 45 30 25
SEX HOUSEHOLD HEAD:
Male 56 18 26 45 36 19
Female 35 35 30 64 23 17
ALL 47 25 28 53 31 16
1984-85 SEASON
VILLAGEz/a
Shoshong East 0 35 65 23 50 27
Makwate 29 42 29 38 42 20
DRAFT ACCESS:
Own, Manage 14 48 38 33 38 29
Hire, COOp 13 30 57 27 50 23
SEX HOUSEHOLD HEAD:
Male 17 39 44 35 39 26
Female 11 36 53 25 50 25
ALL 14 37 49 29 45 26

 

Sources: 1983 Crop Management Survey and 1985 DUMI Study.

a. Proportions of households beginning plowing in each period is
significantly different at .95 confidence level.

b. Proportions of households ending plowing in each period is
significantly different at .95 confidence level.

100
Makwate households tend to start and end their plowing before those in
Shoshong East. These is true regardless of whether one refers to
farmers' perceptions of usual starting times or to actual times in
1982-83 and 1984-85. Draft-dependent households consistently start
later than draft-controlling households. The same relationship between
draft control and the start of plowing was noted by the FAO [1974],
Sheppard and Clement-Jones [1979], Vierich and Sheppard [1980], and in
others studies. Since female-headed households are more dependent on
hired traction than male-headed households (Table 4.1), they too tend to
start their plowing later. Bond [1974] similarly found that unmarried
women and widows tend to start later than married couples.

Two managerial issues relating to the timing of plowing during the
season were investigated in the DUMI Study. The first was on which set
of rains does plowing normally start. Most respondents said they
generally wait until the third set of rains. The reasons for waiting
included the following: (a) to avoid the extreme heat and drought spells
which Often follow the first couple of rains, (b) draft animals Often
are in poor condition early in the season, (c) there is better weed
control if weed seeds are allowed to germinate before plowing, (d) to
avoid crop damage from livestock, (e) cash to hire traction or buy seeds
Often does not arrive through remittances until later, and (f) to there
tend to be multiple rains during late December and early January.

Some additional reasons for starting late were identified by Hertel
[1977] and Gulbrandsen [1980], including: (a) important village
celebrations take place in early spring, (b) school children and mine
laborers cannot help until Christmas vacation, (c) draft animals cannot

be hired until owners finish plowing for themselves, and (d) higher bird

101
damage on early planted crOps.

The second seasonal timing issue addressed was: How late in the
season will plowing be done? As noted in the Appendix, night
temperatures turn too cold for sorghum maturation by April or May. When
there are infrequent rains during the early part of the season, farmers
often have to decide whether to continue planting sorghum later into the
season than ordinarily desired. When asked in 1983, 55 of the
respondents said they would continue into February, including nearly
thirty percent who said they would continue into early March. Early
January is about the latest that technical researchers are willing to
plant their trials.

Data on the timing of plowing vis-a-vis planting rains were
collected in the MVRU Survey and through technical plot monitoring.

Plot monitoring data from both 1983-84 and 1984-85 showed that
households controlling access to draft plowed more area and a greater
proportion of their area on good to excellent soil moisture than did
hiring households. More than half Of the area planted by hiring
households in 1983-84 was planted on drying or insufficient soil
moisture. Data from the 1984-85 MVRU Survey (all households controlling
traction) showed, in contrast, that 62 percent of the area was plowed
within two days of when plowing could begin after a rain. Only ten
percent of the area was plowed more than four days after plowing started
on each rain.

One key managerial issue relating to the timing of plowing
vis-a-vis rains was investigated in the DUMI Study: Do farmers
consciously plow and plant on drying soil moisture conditions? Nearly

all experiment station trials are put-in immediately following rains, on

102

excellent soil moisture but farmers often continue plowing beyond the
point at which soil moisture is adequate for germination and emergence.

Nearly all the respondents reported that they do plant on drying
conditions. Most of those who plow and plant on drying conditions said
that they believe the seed can remain in the ground for long periods, if
it is dry, and will germinate when there is a following rain. Several
farmers said that whether planting is done on marginal soil moisture
depends on the month. Early in the season, they use only good moisture
days. Later in the season, when the number of potential plowing
opportunities has decreased, they plow on less Optimal days. The few
respondents who said they do not plow and plant on drying conditions
expressed concern over their draft animals, not concern over poor
emergence percentages.

E. CROPPING PRACTICES

National data show that most farmers in the Central Region
broadcast plant sorghum-dominated seed mixtures, incorporate the seed
using a mouldboard plow, do not apply fertilizer, and only do a single
weeding (see the Appendix). One objective in conducting research on
cropping practices was to verify that practices in Shoshong and Makwate
were typical for the Central Region. Three additional Objectives were
to: (a) characterize "traditional" farming practices, (b) determine
whether the use of recommended practices was associated with household
(or farm system) types, and (c) have farmers assess the major
recommended practices.
1. VARIATIONS IN TRADITIONAL PRACTICES

In the 1983 Crop Management Survey, farmers were asked about their

standard plowing and planting practices. NO single set of traditional

103
practices could be identified. For example, just over half the
households planted some crops in sole stands, even though nearly all
households planted mostly crop mixtures (as national data indicate).
Fifty percent planted the same crops in the same places every year but
the others changed crop locations. Over forty percent purposively
rotated crops. Just over half changed the direction of plowing each
year, while the Others always plowed in the same direction. Finally,
just under half planted particular parts of their field first each year,
while the others started at different places.

There were substantial differences in the plowing and planting
practices used in Makwate and Shoshong East. In general, Makwate
households used more "progressive" practices. For example, more Makwate
households rotated their crops and took into account field and
crop/variety characteristics when deciding what to plant, when and
where. The use of the different plowing and planting practices was not
associated with the sex of household head, draft control or cattle
assets.

To characterize the "progressive" traditional practices,
participants in the 1983 CrOp Management Survey were asked: (a) which
crops are sole planted, (b) which crOps are rotated and why, (c) which
parts Of the field are planted first each year, and (d) which crOp are
planted first and why.

Sorghum was sole planted by more households than any other crop but
the proportion of sorghum area which was sole planted was quite small.
Cowpeas, maize and jugo beans were frequently planted sole and for each
a relatively large proportion of the area planted was sole planted.

(This finding stimulated an interest in exploring the advantages of

104
planting secondary crOps in sole stands; see Chapter VI).

The most common crOp rotation was sorghum and cowpeas, followed by
sorghum and maize, maize and cowpeas, sorghum and jugo beans, and maize
and jugo beans, in that order. A total of 16 different crop rotations
were identified, but sorghum and cowpeas accounted for nearly 40 percent
of the responses.

The main reason for changing crop locations in the field was to
increase soil fertility (due to nitrogen fixation by cowpeas). Twenty
percent shifted their crOps in order to control striga on sorghum and
witchweed on cowpeas. Most of the others shifted their crops in order
find out where different crOps grow the best.

The most common strategy, when selecting where to start plowing,
was to plow on the sandier parts of the field. Several respondents said
that the sandier soils are easier to plow when their traction animals
are still weak. Other respondents rightly noted that sandier soils
moisten to plow depth sooner than do clayey soils. Four other types of
areas were plowed first during the season: heavily weeded areas (to
control weed deve10pment), well destumped parts of the field (since
plowing is easier), parts nearer the compound (easier to protect from
livestock), and parts nearer other fields (less bird damage).

Most farmers naturally said they plant long maturing crops first
each season, including jugo beans, Goromente and Moterane sorghum,
Tswana cowpeas, and watermelons. These crOps also tend to be tolerant
of the high December and January temperatures. Quick maturing varieties
of sorghum, cowpeas, and maize were planted later in the season. Late
plantings were made in order to escape mid-season droughts and rain

damage (either lodging or various grain molds). A small number Of

105
respondents said they plant some maize and quick maturing cowpeas early
in the season, to provide "green mealies" (corn on the cob) and morogo
(spinach) for consumption in March and April.

In the 1984 Cowpea Baseline Survey and the 1985 DUMI Study, further
attempts were made to characterize traditional practices. The Cowpea
Survey revealed that farmers do not have any strategy for dealing with
insect pests, although pests such as grasshOppers can quickly destroy
cowpeas plots. To control disease problems, they select seed only from
the healthy plants. Several strategies were followed in picking cowpea
leaves: some farmers completely defoliated plants, others removed
selected vines, and others periodically took a few leaves from each of
several plants. Only a few waited until after the grain harvest to
strip leaves. Reflecting the diversity of defoliation practices, there
was no consensus as to whether defoliation reduced grain yield.

In the 1985 DUMI Study, farmers were asked whether they take into
account soil moisture when deciding what to plant (because different
crOps and varieties have differing moisture requirements for prOper
germination). Essentially all the farmers were aware of relative soil
moisture requirements, and took them into account. For example, all the
farmers said they try to plant maize only when there is good moisture.
Jugo beans and groundnuts also were planted on good moisture, if
possible. Most farmers said that their traditional varieties of sorghum
and cowpeas could be planted into drying soils.

The DUMI Study showed there was diversity in post-planting
practices as well. While a majority did not thin or replant, a number
of respondents did do these operations when needed. The farmers who did

not thin gave a number of reasons for not doing so, most of which

106

implied they thought thinning was not beneficial. Most of the farmers
who thinned said they thin when the soil is moist, so the remaining
plants do not become desiccated. The farmers who replanted generally
said they wait one or two rains before replanting, since there can be
substantial delayed emergence if the initial planting was done on poor
soil moisture. Some farmers replanted by replowing, others used metal
harrows, and still Others dragged branches over the plot in order to
incorporate the seed. Hand replanting generally was done with a hand
hoe.

In summary, there was substantial heterogeneity in "traditional"
plowing and planting practices, and post-planting practices. Some
practices seem to be more progressive, based on basic agronomic
principles and observation. Few of the actual benefits associated with
the use of "progressive" traditional strategies and practices have been
investigated.

2. USE OF RECOMMENDED PRACTICES

There is not a single recommended sorghum package in Botswana.
Instead a number of extension recommendations have been developed and
promoted over the past forty years. A summary of the main current
recommendations and their use by households participating in the 1983
CrOp Management survey is presented in Table 4.3.

Early planting was the most commonly used recommendation, followed
by multiple weeding and hand planting. These practices actually are
"progressive traditional" practices which the extension service now
recommends to all farmers.

Few farmers used any of the "modern" practices recommended by DAFS,

including row planting, harrowing, third furrow hand planting,

107

.mo>u=m unoaowoomz mono mme ”oounom

 

 

 

N O m N N N H O O MMDZOZ H<mmoo Gum:
N N N N N O n O m msonmm memo nmmoem
O O O O m N O OH O Hz<Hm Qz<m 3ommsm can
O OH O m N O n O n qu3OHH<m Mm:
m N OH OH N O OH N NH BZOHA o\3 304m me<m
OH O OH HH O O NH mH O Hz<Hm 30M
OH O NN OH HH N NN NH NH 3omm<m mm:
5N ON mN ON mm NO ON mm OH HZde nz<m
Hm mN mm ON Om Om ON 50 O mUZHnmmz MHmHHHDZ
OO OH 00 NO Om HN Nm me ON HZdelonm MHm<m
Aanonooaom mo uaoouomv
HHO .nzmmmn Homazoo +oH mHIO MH<zmm MAO! .3M<z .momm
mmmoo< HHOMQ MHeado Q<mm Nmm MU<HHH>
mmmH .wmoHHUOdm oszmomo Omnzmzzoomm mo mm: NN.O mHm<H

108
fallowing, cutting crOp residues, or incorporating corral manure. Only
14 percent of the households had ever used a harrow and only ten percent
had ever tried row planting. Even fewer households had ever tried third
furrow hand planting, fallowing, cutting and storing crOp residues, or
incorporating corral manure. Aside from row planting and harrowing, few
respondents even knew a user of the "modern" practices and, therefore,
had little or no basis for evaluating the relevance of the extension
recommendations.
3. PERCEPTIONS OF RECOMMENDED PRACTICES

In order to determine why farmers were not using the extension
recommendations, participants in the 1983 survey were asked whether a
practice would increase production, if used. Respondents who did not
use recommendations were asked why. Summary findings are presented in
Table 4.4.

The respondents expressed the greatest confidence in row planting
and early planting. Very few farmers doubted that either practice would
increase their crOp production. Lack of resources, particularly control
over traction and implements, was the main reason why more farmers had
not tried row planting. Few farmers said they were deterred by the
extra expense. Some respondents said they could not early plant because
of the risk of cattle damage.

The next most positively perceived recommendations were multiple
weeding, incorporation of corral manure and harrowing. Half the
respondents said that multiple weeding would increase production but
most of them also said they did not have enough labor to do more than a
single weeding. Also, nearly a third said multiple weeding would not

help, more than was the case for any other practice. More than forty

109

.muowouno vocHnaoo OHnu OH vovsHocH
one on vow Osouuon OOO OuouomHo sou voxooH NHHOOO: OOHuonuu wconoH Nuoaunm .n
.OHH ou NO aouw vowcou oomooaoou mo nonaoa one .o
.mo>u:m uooaowoooz mono mmmH "oouoom

 

 

 

N ovooa who:
m ensue: 02
OH uuoaoaoua uoccno
m m OcHOm oHnOHHOHOD
m ownfina oHuuoo
H N O N Omcoaxm Ouuxm
m m «manum
HN HH NN OH N N N N OHNooom oNNONONoooH oz
ON NO ON OO NN ON OH NH omooHaooN NooH
N OH OO NN OH O NH OH NoOOH xooo
NN O ON ooooaoHoaH NONN
HH NN OO ON O\.aoHNsH O .oNNN NONN

u: NNNNNzONNNN NO NzNONNN 1-- ”NNO Noz NNO
HN NN OO ON NN NN NN NN zone Nooon
O N N NH O N N N oz
NN HO HN NN OH ON ON ON No»

"onHUDQomm mm<mmozH aHboz

 

mmmzmdm mo Hzmommm

 

 

onH<HOM 02H amm3 Hz<Hm Hz<Hm 30mm<m Hz<Hm Hz<Hm
zH Im=z<z MHmH mom 3omm=m NHM<M 30M
3OHH<m H<Mmoo IHHDX az¢m Oz<m

mme .mmoHHo<mm amnzmzzoomm ho monHmmummm .mmm2M4h HO.O MHm<H

110
percent of the farmers said they thought that incorporating corral
manure would increase production. Again, however, several resource
constraints were identified, including too little labor, lack of manure,
and no transport. Some farmers said that manuring causes more weeds.
More than a third of the respondents felt that harrowing (as a planting
method) would help. Lack of traction and a harrow were the main
constraints.

Farmers felt least certain about the benefits from hand hoe
planting, fallowing and hand furrow planting. The latter two were
essentially unknown in the area. Less than thirty percent of the
farmers said than hand planting would increase production and an even
larger share said they did not have the labor needed to do hand
planting. Many respondents said they just did not like the idea of hand
planting or the idea of fallowing. Fallowing was seen as waste of land.

F . POST-HARVE ST PRACTICES

It was only possible to investigate post-harvest issues by asking
farmers about their normal practices, since most farmers harvested
little or no grain. Two issues were addressed in the 1983 CrOp
Management survey: (a) where is seed selected for future plantings and
(b) how are the food grains stored. In 1984, participants in the the
Cowpea Baseline Survey were ask about their use of crop residues.

1. SEED SELECTION

More that two-thirds of the respondents in 1983 said that they
selected seed at the threshing floor before threshing takes place. Only
23 percent said that they selected seed in the field, while ten percent
said they selected seed after threshing. A greater proportion of

respondents from Makwate and draft—controlling households than Shoshong

111
and draft-dependent households, respectively, said they selected seed in
the field where plant characteristics could be considered.

Most seed selection was based on head characteristics. Half the
respondents said that they select fully mature heads without any smut.
More than forty percent said they pick the heads having large seeds. In
essence, the respondents merely looked for the best heads among the ones
they harvested for each of the varieties grown. Plant characteristics
were not taken into account, nor were interactions between
micro-environments and head deve10pment.

2. STORAGE

As can be seen in Table 4.5, nearly all households stored their
grain for both food and seed in bags. Traditional storage methods, such
as mud cribs and reed baskets, have disappeared throughout much of

Botswana [Hamilton, 1975].

TABLE 4.5: STORAGE PRACTICES, 1983

 

SORGHUM MAIZE COWPEAS
FOOD SEED FOOD SEED FOOD SEED
(Percent of Househods)

CONTAINER:
Bags 98 94 98 96 97 89
SEED PROTECTION:
Nothing 52 44 53 49 45 40
Ash 43 47 42 43 48 48
Chemicals 5 9 5 8 7 l2

 

Source: 1983 Crop Management Survey.

112

Research from around Africa generally has shown that bag storage is
as good or better than other methods [Eicher and Baker, 1982]. Hamilton
[1975], however, Observed high insect pest infestation in bag storage in
a study in the S.E. Kweneng District.

In 1983, nearly half the respondents used nothing to protect their
seed grain, while most of the rest used burnt manure or wood ash. Less
than ten percent of the households used seed protection chemicals.
Similar proportions were noted by Hamilton.

3. CROP RESIDUE USE

One of the main recommendations of the Ministry is to cut and store
crop residues, to be fed to draft animals (or cows) during the dry
winter. The 1983 CrOp Management Survey showed, however, that few
farmers cut, remove and store crOp residues - or feel they have the
labor resources to do so. In 1984, as a follow-on, farmers' field
residue management strategies were investigated. Results are summarized
in Table 4.6.

There was little or no management involved in the use of crOp
residues. Which animals grazed residues was determined by which animals
were in the vicinity of the field at the end of the season. As a
result, 80 percent Of the respondents reported that non-household
animals grazed their crOp residues. In few cases did the owners of
non-household animals pay when the residues belonging to other
households were grazed.

MOst residues were grazed immediately after the harvest and few
respondents reported that it would be possible to prevent grazing after
the harvest was complete. Only one household reported that residues had

ever been removed and stored for later feeding. At present, the

113

TABLE 4.6: USE OF COWPEA RESIDUES, 1984

 

SHOSHONG MAKWATE ALL
(Percent of Households)

 

OWN ANIMALS GRAZE RESIDUES:

Yes 60 74 65
No 33 5 22
DO Not Own Animals 7 21 12
IF ANIMALS GRAZE, wHICH ANIMALS:
Large and Small Stock 32 46 38
Cattle 47 0 28
Donkeys 5 38 19
Goats 16 0 9
Cattle and Donkeys 0 15 6
CONTROL OF ACCESS TO GRAZING:
Non-Household Animals Graze 87 68 80
Grazing Immediately Follows Harvest 100 37 76
Farmer Could Prevent Grazing 13 32 20
Owners of Non-Household Animals Pay 12 8 10
Farmer Selects Animals 5 0 3
EVER HAVE REMOVED STOVER 3 O 2

 

Source: 1984 Cowpea Baseline Survey.

dominant managerial strategy relating to use of crOp residues is to let
the livestock have one last, big feast before they have to cOpe with the
shortage of graze and browse during the dry season.
G. PERCEPTIONS OF PROBLEMS

To determine farmers' priorities, participants in the 1983 Crop
Management Survey were asked to identify their top three crOp production
problems. Followeup questions on problems were asked in 1985. This
section summarizes findings, and then notes the problems farmers have
identified in five other studies.
1. SHOSHONG AND MAKWATE FARMERS

In 1983, the most common problem was livestock damage (and

implicitly a problem of inadequate fencing), cited by more than a third

114

of the respondents. The other major problems cited were (in order of
frequency): lack of draft power, lack of cash (for hiring traction and
buying seed), lack of implements, labor shortages and a shortage of
water at the lands. Even though the question was Open-ended, all the
the main responses referred to resource constraints--not husbandry
problems.

The 1983 survey participants also were asked to rank the relative
importance of their resource constraints. Findings are summarized in

Table 4.7 for different household types.

TABLE 4.7: FARMERS' RANKINGS OF RESOURCE CONSTRAINTS, 1983

 

TRACTION
CASH/a LABOR FENCING EQUIP. ANIMALS/b LAND
(Percent of Responses/c)

 

VILLAGE:
Shoshong 31 24 31 3 6 4
Makwate 53 20 7 16 5 4
SEX OF HEAD:
Male 29 28 25 9 5 5
Female 52 15 17 6 6 2
CATTLE ASSETS:
0-15 55 15 12 10 7 2
16+ 20 31 33 6 4 6
DRAFT ACCESS:
Control 33 26 25 7 3 7
Dependent 46 18 18 10 8 0
ALL 39 23 22 8 5 4

 

Source: 1983 Crop Management Survey.

a. Lack of cash for seed, transport to traction hiring.

b. TOO few or weak traction animals.

c. Respondents identified their top two constraints out of the
list of six. Percentages are based on the number of
responses received.

115

Overall, lack of cash for obtaining needed inputs was the main
resource constraint. This was followed by too little labor at times and
inadequate fencing. Respondents from female-headed and cattle-poor
households emphasized the problem of no cash, while male-headed,
draft-controlling and cattle-rich households reported that labor
shortages and inadequate fencing were equally important.

In the 1985 DUMI Study, respondents were asked to identify from a
list which problems had reduced crop production that season. Nearly
all, of course, cited the drought. In addition, more than 60 percent of
the respondents had severe insect pest attacks, 55 percent had poor
plant establishment, and 43 percent lost crOps to livestock. TOO many
weeds, crop diseases and labor shortages were cited by relatively few
farmers, reflecting the impact of the drought. When respondents were
asked to identify their major insect pests, the most frequently
identified pests in Shoshong were the Corn Cricket, CMR Beetle and the
American Bollworm. In Makwate, the main pests were the Elegant
grasshOpper, American Bollworm, and Corn Cricket.

2. FINDINGS FROM OTHER STUDIES

Additional information on farmers' views of their problems is
available from the FAO [1974], Bond [1974], Alverson [1979], Fortmann
[1981] and Gaosegelwe, et al. [1983]. The FAO study covered the whole
country. Bond's results are from the Southern and Gaborone Agricultural
Regions. Alverson's survey was of a small number of farmers in the
Gaborone Region. Fortmann's survey covered 355 households in 12
‘villages from the Francistown, Central, Gaborone and Southern Regions.
Gaosegelwe, et a1. interviewed farmers in the Ngamiland West District of

the Maun Region .

116

In the FAO study, respondents were asked what factors kept them
from plowing more land than they did. The responses in order of
importance were: drought, lack of seed, draft power not available, lack
of implements, labor shortages, oxen too weak, and no food at the lands.

Bond asked about arable farming problems, and presented separate
results based on gender and marital status. For the pOpulation as a
whole, the main problems in order of importance were: difficulty in
hiring traction, labor shortages, weeds and birds (one category), and
animal damage. The main problem for single women (including widows) was
a lack Of labor. Both married couples and married women whose husbands
were absent said that difficulties in hiring traction was the main
problem.

Alverson disaggregated the farmers' assessments based on relative
wealth. The wealthiest farmers said their main problems were pest and
cattle damage, lack of water at the lands, inadequate repair services
for implements and tractors, and labor shortages. The middle farmers'
problems were: a lack cash for purchasing inputs, labor shortages, pest
and cattle damage, and a lack of water at the lands. Poor farmers said
draft power was the main problem, followed by labor shortages, lack of
implements, not enough seed, and cattle damage. Alverson further
explained that the reasons for labor shortages differed depending on
wealth. The wealthiest farmers faced a shortage in harvesting and
threshing due to their large scale of production, while the poorest had
too little labor for plowing and weeding.

Fortmann asked respondents to identify the main problems which kept
them from growing more crOps. Separate results were presented for

male-headed versus female-headed households. Overall, the main problems

117
(as ide from lack of rain, of course) were poor land, lack of seed, lack
0f draft animals, labor shortages, lack of land, lack of implements and
“Kiéllc. draft animals (in order of importance). A significantly higher
Proportion of female-headed households than male-headed households
reported that seed shortages, labor shortages and a lack of draft were
major constraints.

Gaosegelwe, et a1. did not rank the relative importance of problems
for farmers in the Ngamiland area. The four most consistently mentioned
‘problems were: lack of draft, lack of labor, crOp damage, and lack of
seed.

In general, the problems reported by farmers elsewhere in Botswana
echo those mentioned by farmers in Shoshong and Makwate. Lack of draft
or money to hire draft is the main problem facing poorer farmers. Once
draft-dependent households can get the money to hire traction (or
exchange labor), they Often get poor germination. If good germination
is achieved, most households--but particularly those draft-controlling
households able to plant large areas--run into labor problems in getting
their fields weeded and harvested. Meanwhile, the potential harvest is
constantly under attack from birds, cattle, other stock, and wild
animals. In the end, many of the poorer households do not even harvest
enough that they have a secure supply of seed for the following season.

Given the above scenario, it is no wonder farmers perceive resource
constraints as dominating, and hesitate to consider many of the
extension recommendations which merely aggravate cash, traction, and

labor resource constraints.

118

H. ROLES IN DECISION MAKING

This section presents information on the relative roles of resident
adults in decisions about crop production activities. Roles in decision
making were investigated in order to make sure the correct individuals
could be contacted about proposed changes. The research was motivated
by earlier findings on the gender division of labor and the male bias of
the extension service [Bettles, 1984; Fortmann, 1985].

In one of the surveys of the 1985 DUMI Study, male household heads,
female household heads, females spouses and adult children adults were
asked to rank their relative roles in decision-making. The categorical
responses included the following: "decide alone," "discuss with Others
but have the main role in deciding," "decide jointly with others,"
"discuss alternatives but Others decide," and "not involved in the
decision."

The series of individual interviews showed that most decisions are
made jointly by adults after a discussion of alternatives. The relative
influence of adult children varied but, as a rule, children did not play
a major role. Three basic patterns of decision roles were identified
for different types of decisions.

The most common pattern of decision making was for joint decision
making in households with a male head and spouse, the head alone in
female-headed households, and with children having relatively little say
in both types of households. This pattern prevailed for decisions about
hiring or hiring out labor or traction, whether to cooperate with other
households, whether to try new crOpping practices, and whether to sell
crops. The same basic pattern prevailed, except that males within

male-headed households had a slightly stronger role, for the decision

119
whether to use non-household labor for weeding. The joint decision
making pattern was also observed with respect to decisions about helping
non-household family members, except that children played a greater role
in this decision.

In a second pattern, males dominated decision making, even in
female-headed households. This pattern was found for the decisions when
to start plowing during the year and whether traction animals were fit
to plow. Men in male-headed households also were dominant in the
decisions about when to plant vis-a-vis a given rain and about the
adequacy of soil moisture. Female heads of households assumed control
of these decision for their households.

The third pattern involved a dominant role for females in both
male-headed and female-headed households. This pattern prevailed for
the selection of seed to plant, and decisions about replanting or
weeding.

To further characterize the patterns of decision making, indices of
relative control were calculated for each decision. For each decision,
the index of relative control represents the perceived sense of control
for each respondent category relative to other household members. The
lower the index, the greater the perceived sense of control for that
group. The indices of relative control are presented in Table 4.8.

(The procedures for calculating the indices are described in the notes
to the table.)

The indices of relative control show that it primarily was a
person's position in the household, rather than just gender, which
determined roles in decision making. Female heads of households had the

greatest individual control over decisions because there often were no

120

TABLE 4.8: ROLES IN DECISION-MAKING, 1985

 

MALE FEMALE FEMALE SONS/—
HEADS HEADS SPOUSES DAUGHT.

 

(Indices of Relative Control/a)

WHETHER DRAFT ANIMALS FIT 56 107 112 121
WHEN TO PLANT VIA RAINS 68 88 115 121
WHEN To START PLOWING 72 93 112 118
WHETHER SOIL MOISTURE OKAY 82 77 111 120
To HIRE-IN LABOR/TRACTION 92 77 108 115
To HIRE-OUT LABOR/TRACTION 101 73 112 108
To SELL CROPS 102 61 106 119
TO USE NON-HH WEEDING LABOR 104 68 104 114
TO ENTER COOP ARRANGEMENT 105 79 108 104
WHETHER & WHEN TO REPLANT 108 60 92 124
To HELP NON-HH FAMILY 109 75 112 100
HOW MUCH SEED TO PLANT 119 54 82 128
WHEN To WEED 146 59 59 120
AVERAGE 96 75 104 116

 

Source: 1985 Decision Unit Profile Survey.

a. Each respondent ranked his/her own role in each decision
from 1 (Idecide alone) to 5 (- not even involved in the
decision). The numer of respondents were as follows: 23
male heads, 19 female heads, 21 female Spouses, and 30
daughters and sons. For each respondent category, the
index of control was calculated as the mean value for
that group relative to the average value for all groups
combined (multiplied by 100).
greater the perceived sense of control for that group.

The lower the index, the

121
other senior males or females resident. There were, however, gender
related decision roles in households having a male head and female
spouse. In general, males controlled decisions relating to the plowing
Operation and females controlled decisions about post-planting
practices. There was less gender bias in decisions about resource
acquisition and disposal (include crOp sales) than in production
decisions.

The findings on decision making patterns are, for the most part,
consistent with those reported by Bond [1974]. Due to her survey method
(each respondent was only asked who was responsible for each of four
crop production decisions), Bond: (a) understated the role of women in
seed selection, (b) did not identify the important role of women in
decisions about post—planting Operations, (c) seemingly overstated the
role of women in decisions about traction use and timing, and (d) did
not fully represent the extent to which decisions are made by consensus.
There not only is a free exchange of ideas, several individuals often
are involved in actually making decisions even if one is the principal
decision maker. Nevertheless, Bond made an invaluable contribution when
she documented the existence of distinct decision roles and showed that
the extension service must work with directly with women.

I. CONCLUSIONS AND IMPLICATIONS

This chapter has described crOp systems management, emphasizing:
(a) farmers' priorities and perceptions, and (b) the diversity in
existing practices. This section highlights key findings and

synthesizes implications for arable farming deve10pment priorities.

122

l. CROPS AND VARIETIES

Farmers were unanimous in saying their primary objective was to
grow food for home consumption. Secondary objectives were to produce a
surplus to sell, to have a diverse diet and to minimize the risk of
complete crOp failure. Based on farmers' Objectives, the Ministry
appears to be on the right track in emphasizing sorghum and GOWpea
research.

With reference to sorghum variety screening, length of growing time
and drought resistance were considered to be the most important
characteristics. Still, farmers were not ready to sacrifice the taste
of their traditional varieties to get the short growing periods of Red
Swazi and 8-D. Cowpea screening should, as it currently does, focus on
at least two distinct types of cowpea plants: (a) a quick growing,
determinant plant for seed production and, (b) a long season, spreading
plant for leaf production. Farmers also like plants such as Blackeye
and Sri Lanka which combine the characteristics. Therefore, an
intermediate plant would be a third screening goal.

There is enough interest in groundnuts and jugo beans to warrant
additional research, if only in the context of on-farm research. The
top priority is to identify drought resistant, quicker maturing
varieties. Farmers do not use groundnuts for oil, so Oil extraction
should not be a criterion for variety selection.

In general, the screening of all crOps and varieties needs to pay
more attention to farmers' standard production practices. For example,
most farmers plant their cowpeas and sorghum into drying soils. As part
Of the varietal screening process, establishment should be evaluated

when seeds are broadcasted into drying soil moisture conditions. The

123

tendency of farmers to plant late in the season also should be
considered. Quicker growing crOps/varieties and crops which produce
products even if there is not time to mature (such as cOWpea leaves) are
Options for farmers who plant late.

2. TRACTION USE AND ACCESS

One of the main challenges facing the Ministry is to help farmers
take advantage of rapidly changing patterns of traction use and draft
access. Although no one knows the extent to which the trends will
reverse themselves after the drought, the trends do have some positive
implications for arable farming deve10pment.

One positive implication of the trend toward tractor hire is the
Opportunity to impact on many farms by concentrating on appropriate
implements and practices for a much smaller number of farmers. Even
more important, the control, speed and draft power of tractors (even
with low horsepower) create several additional Options for field
Operations, including sub-soiling, stubble sweeping, and inter-row
cultivating.

Use of donkeys has several advantages, particularly for the poor.
Timing can be controlled and more area can be plowed than when one hires
traction. An entire donkey team costs only the equivalent Of one ox or
cow. Moreover, donkeys survive better in drought and generally require
less management. Row planting can be done with one or two donkeys,
while row planting with untrained oxen is impractical. Finally, donkeys
can be used throughout the year to pull carts, reducing household
maintenance labor requirements and freeing additional labor for arable
farming.

All things considered, the Ministry should make promotion Of donkey

124

traction one of its tOp priorities. Options to take advantage of the

speed and draft of tractors would be another priority. Efforts to

improve or maintain the traditional oxen-based plowing system should be

a low priority, at least in the Central Region.

3. TIMING OF PLOWING

Several timing issues were addressed in the research on traction
use and plowing. In each case, farmers cited several reasons why they
use their current timing strategy, rather than follow the
recommendations and practices of researchers. The Ministry should take
farmers' timing strategies as a given and concentrate on interventions
which are consistent with their strategies. Several examples can be
given:

a. Row planting into drying soils might be more successful if done with
an injection planter, rather than the current (shallow) planters.

b. The Option of dry planting should be examined. Under the right
conditions, dry plantings often establish better than do wet
plantings.

c. The advantages of seed soaking could again be examined (as they were
inconclusively during the 19708) as a Option for planting into
drying soils.

d. Fungicide seed treatment might be useful for dry plantings, to
protect the seed until the following rain.

4. PRODUCTION PRACTICES

There were two main findings in the research on production
practices. First, there was substantial heterogeneity in traditional
practices. Second, many farmers had favorable perceptions of several

recommended practices but were not using them due to resource

125

constraints.

To take advantage of the diversity of traditional practices, a
greater share of research resources needs to be shifted from tillage
systems deve10pment work to technical and economic assessments of
farmers' current practices. It should be quite feasible to address the
following issues with minor investments of existing research resources:
a. What criteria should farmers use for site selection? Right now,

farmers plant their groundnuts and jugo beans on sandier soils. Are
there other criteria which might be helpful? For example, the depth
and texture of the soil may reflect water holding capacity, which in
turn is related to the benefit from double plowing (see Chapter V).

b. Farmers currently plant secondary crOps in their crop mixtures at
only 1:25 or even 1:50 ratios relative to their sorghum. Can the
ratio increase without hurting sorghum production?

c. What are the best cowpea defoliation strategies for different types
of cowpea plants?

d. What is the threshhold level of plant stand, below which it would
make more sense to replow and plant the entire acre rather than gap
fill?

e. Does the extra production from hilling jugo beans and groundnuts (to
facilitate pegging) justify the labor investment.

f. Can women start their weeding earlier without a drop in yields (1-2
weeks after emergence rather than 4-6 weeks), thereby reducing the
labor constraint when weeding overlaps with birdscaring and early
harvesting?

Since most farmers do not have the resources to implement the

current recommendations, the recommendations need to be targeted to

126

particular environmental circumstances. Targeting will maximize the

expected benefit and minimize the resources required for implementation.

Some examples where this approach might be tried are as follows:

a. Hand planting might be feasible (and profitable) if targeted to high
value crops, planted on high potential sites.

b. Thinning might be a better use of labor than weeding if done only in
the few spots where stands exceed 100,000 to 125,000 plants per
hectare.

c. Harvesting only some residues for high priority animals such as
calves and kids might be profitable and would not take too much
labor.

d. Manuring could be concentrated on a small, apprOpriate site for
growing maize for green mealies (corn on the cob).

5. POST-HARVEST PRACTICES

Post-harvest research will not be a priority until the drought is
over. However, in the medium term horizon, there are options for
improvement. First, farmers have a minimal set of criteria for
selecting seed. If simple, but more comprehensive criteria can be
develOped, this would be a "non-resource using" intervention--which is
always useful.

Second, the management and harvesting Of crop residues is likely to
be of increasing importance over time. Residue management is not part
of traditional crop systems management because, until recently, few
fields were fenced.

Third, storage losses can be a problem for sorghum and appear to be
a major problem for cowpeas. The Ministry might reactivate its storage

extension program from the late 19608 and early 19708, in which mobile

127
teams circulated to provide storage advice and to help treat seed
[Hamilton, 1975].
6. DECISION BEHAVIOR

The research on decision behavior showed that: (a) farmers felt
that resource constraints are the main factors limiting increased
production, and (b) there were gender roles in production decisions
corresponding to the gender bias in fieldwork.

Given farmers' current concerns with resource constraints, along
with the government's commitment to equitable deve10pment, the
Ministry's emphasis on ALDEP and drought relief appears warranted. The
proposals presented in this thesis on Options to resource transfers are
not meant to imply that resource transfers are not useful or are
undesirable.

Finally, research activities clearly need to be targeted to
individuals within households, not only to particular household types.
Interventions related to the plowing Operation should be targeted toward
males and should relate to the problems perceived by males.
Interventions targeted toward crop/variety selection or
post-establishment Operations should be designed and evaluated in
relation to the problems and interests of females. The extension
service also should taken into account gender roles in decision making

and fieldwork.

V. CROPPING OUTCOMES AND THE RESULTS OF ON-FARM TRIALS

This chapter presents an overview of crOpping outcomes for the
Shoshong and Makwate households, gives a historical perspective on
experiment station and on-farm research in Botswana, and provides an
assessment of the production practices examined in on-farm trials. The
objectives are to: (a) characterize differences in outcomes by household
type, (b) determine the returns to the resources invested in arable
farming, (c) clarify the rationale for the Central Region
experimentation program, (d) distinguish promising interventions from
those that seem less promising, and (e) analyze whether the most
promising interventions are profitable enough to justify adOption.

A. CROPPING OUTCOMES

Several crOpping outcomes were examined, including days of
planting, area planted, percent of area and plots weeded and abandoned,
yields, and returns to labor, management and traction.

1. 1982-83 SEASON AND "NORMAL" SEASONS

In the 1983 Crop Management Survey, farmers were asked how many
"acres" they planted during the 1982-83 season, and how many they plant
in years with normal rainfall. ("Acres" are a traditional plowing unit
which, on average, are equal to about one-third hectare.) Farmers also
were asked about the number of days on which some planting was done.
The number of planting days and area planted reflect the ability of

different households to take advantage of planting rains. Results are

128

129
summarized in Table 5.1.

Due to the drought, the number of planting days and the area
planted were less in 1982-83 than in normal seasons for all household
categories. The percentage differences between good seasons and the
1982-83 drought season were greater for female-headed compared to
male-headed households, and for draft-dependent compared to
draft-controlling households. The relatively greater effects of drought
on draft-dependent households was also noted by Vierich and Sheppard

[1979].

TABLE 5.1: DAYS OF PLANTING AND ACRES PLANTED; RECALL DATA
FOR "GOOD SEASONS", 1982-83, AND 1984-85

 

GOOD SEASON 1982-83 SEASON 1984-85 SEASON
DAYS ACRES/a DAYS ACRES DAYS ACRES

VILLAGE:
Shoshong East 7 l4 5 9 2 8
Makwate 11 18 7 10 5 10
PRIMARY TRACTION:
Donkey 11 18 7 10 6 10
Cattle 8 15 8 10 3 7
Tractor 4 13 2 7 2 10
DRAFT ACCESS:
Control 11 18 8 11 6 ll
Dependent 5 14 2 6 2 8
HOUSEHOLD HEAD:
Male 9 16 7 11 5 11
Female 9 15 5 7 3 8
ALL 9 l6 6 9 4 9

 

Sources: 1983 CrOp Management Survey; 1983 DUMI Study.
a. Acres refers to traditional acres. Traditional acres
average just over 1/3 hectare. Donkey plowed acres
generally are smaller, around .25 ha., while tractor

plowed acres average around .75 hectares.

 

 

5636011.

 

Findings

The

 

130

Respondents also were asked to estimate their total annual

production of the major crops for the 1982-83 season and for a normal

season.

Little was produced in either Makwate or Shoshong East in 1983.

Findings for normal production levels are summarized in Table 5.2.

TABLE 5.2: NORMAL FARM PRODUCTION OF MAJOR CROPS, 1983

 

 

SORGHUM/a MAIZE GOWPEAS JUGO B. MELONS

VILLAGE:

Shoshong East 752/b 65 56 37 489

Makwate 343/b 108 39 45 586
HOUSEHOLD HEAD:

Male 701/b 125/b 68/b 53/b 560

Female 431/b 23/b 28/b 21/b 434
DRAFT ACCESS:

Control 636 118/b 70/b 54/b 539

Dependent 527 36/b 20/b 21/b 474
CATTLE OWNED:

15 or Less 353/b 32/b 33 26 503

16 or More 87l/b 136/b 68 54 517

 

Source:

1983 CrOp Management Survey.

a. Production refers to kgs. for all crops but the number

of melons.

b. Means for same category and crop followed by "b" are

significantly different at a .95 confidence level.

The estimated normal production levels for sorghum were greater for

Shoshong East, male-headed, draft-controlling, and cattle-rich

households.

For male-headed versus female-headed households and

draft-controlling versus draft-dependent households, there were

significant differences reported for nearly all the major crops. The

RIDS also showed significant differences in production associated with

gender of head and cattle assets [Lucas, 1985].

The relationships identified between household types and estimated

131
production levels were reconfirmed in additional analyses of crOpping
outcomes during the following two seasons.
2. 1983-84 Season

An overview of the days of plowing, area plowed and sorghum
harvested in 1983-84 by the 27 MVRU Survey households is presented in
Table 5.3. Makwate, male-headed and draft-controlling households plowed
many more days on average than did Shoshong East, female-headed, and
draft-dependent households, respectively. Cattle-rich households did
not plow on more days than cattle-poor households because many of them
hired tractors.

Male headed, cattle-rich and draft-controlling households all
plowed more area than their comparison groups--in total, per person
plowing, and per consumer equivalent. A positive association between
area plowed and cattle assets has been reported in several studies,
including the FAO [1974], C80 [1976] and Vierich and Sheppard [1979].

Male-headed, cattle-rich and draft-controlling households also
Obtained better results than their counterparts, in terms of the
proportion of households harvesting, total amount harvested, and amount
harvested per hectare, per consumer, and per hour. The differences were
particular dramatic depending on draft access.

The association between resource access (cattle and draft) and
yields deserves note. Several studies based on years with good rainfall
have not reported an association between yields and resource access (for
example, the RIDS [Lucas, 1985] and Alverson [1979]). Instead,
differences in production generally have been due to a larger amount of
area planted. During drought seasons, however, lack Of draft access and

control over timing become particularly crucial, and affect yields as

132

.oHnOHHo>O no: one mpowouoo mwooon ammuv an mason xuosvHOHm no mouame .o

.cHeuw wonmounu «0 woo woo .mwx ON «0 oonuo>ooo Ounvcoum O do women aonwuow .OwM .n
.NuooaouHsdou OHNOHOO vouoaHuOo OON do woman ONO muanos OOOLH OO.H n cos uHOOO “OO.O

n coaoz uHsvo mmN.O u mHIOH owe mOm.O u aim owe “ON.O n OIO owo "muanoB muHc: uoEOOcoo .n

.onHONHaoa uoHo vow >o>uam Dm>z OOIMOOH "oouoom

 

 

O.m o O N.H N.N O.H N.O O.N m.m ooooooucHez usonqu
O.N o o O.H H.N H.H H.N N.N N.N .ooooHOz OHOHN OOH:
"poo: pom
OO N ON OO ON ON NO NN OO .>HOON NanOooO NON
ON NN NO NN NN NO ON NN OO ONOOOON NON
ONN HN NNN NNN NNO ONH ONO ONO NNN HOooN
ONHONNNNONON NONONON .NOM
NO NN NO HN NN OO NN OO NN zONONON N24 ozHNNN>N<N
N2N<N NO NNNOmNN
0.0 O.H N.O N.O N.H N.O 0.0 0.0 N.H NON NON
0.0 N.O N.O 0.0 0.0 0.0 N.O 0.0 0.0 O\.>HOON NanNooO NON
N.H 0.0 N.H 0.0 N.H 0.0 N.H H.H O.H NoHsoHN ooONON NON
m.m N.H N.N O.N N.N N.N N.O m.O O.N HONOH
"ONOOHN <NN<
0.0 O.H N.O N.O H.N N.N N.N N.OH H.O ozHOOHN NO NOON
NN NN OOH HN OOH ON OO OOH NN ON3OHN NOHN: N2N<N N
HH< NzNOzNNNO zzo NNnO +ON NHOzNN NHOz NN<3N<z onNNONN

mmmoo< HHOMQ

mHmmm< MHHH<U

OOH: mo xmm

mU<HHH>

ON-NNOH .ONNNN>N<N zONONON .ONSOHN <NN< .ozH3OHN NO NN<O "N.N NHNON

133

well as area plowed.
3. 1984-85 SEASON

During the 1984-85 season, the analysis of cropping outcomes was
based only on the 13 households which participated in the abridged MVRU
Survey. All but one of these households controlled traction resources.
The Objective was to evaluate crOpping outcomes and enterprise returns
for draft-controlling households.

The season was again dominated by drought conditions. An average
area of 6.4 hectares was planted per household but only two-thirds of
the area planted was weeded or harvested. Only nine of the 13
households harvested any grain. Across all 13 households, grain
production averaged 221 kgs. of sorghum and 23 kgs. of cowpeas. Average
sorghum yields were only 36 kgs. per hectare.

A crop enterprise budget for the thirteen farms participating in
the income and plot monitoring program is presented in Table 5.4.
Technical data in the budget are area weighted. Local market prices
were used for seed, sorghum, maize and beans. The prices were Obtained
through a regional survey of trading establishments (see Chapter VI).
The official BAMB prices were not used since no farmers traded through
BAMB. It was difficult to determine an appropriate price for sweet
reed, morogo and melons, since these items usually are neither purchased
nor sold by farmers in Shoshong and Makwate. To value these items, the
most common sale price in Mahalapye village was used, minus an estimated
tranSport charge. This represents the Opportunity cost for home
consumption.

NO attempt was made to establish prices for farmers' land, labor,

traction, or management. There is no formal market for land or

TABLE 5.4:

134

CROP ENTERPRISE BUDGET,

1984-85

 

LEVEL OF PRODUCTION

ALL FARMERS
PER FARM PER HA.

IF HARVESTED
PER FARM PER HA.

 

Area Plowed (ha.) 6.4 -- 8.0 --
Area Weeded 4.0 -- 6.1 --
Area Harvested 4.2 -- 6.6 --
INPUTS
Seed Planted (kgs.) 39.5 6.2 44.9 6.2
Traction Hours 63.3 10.0 80.6 11.5
Person-Hours:
Plowing 183 29 248 35
Weeding 149 23 222 32
Birdscaring/a 297 46 467 67
Harvesting 113 18 177 25
Threshing 28 4 44 6
Maintenance 116 24 145 18
PRODUCTION
Sorghum (kgs.) 221 34.5 348 43.5
Cowpeas (kgs.) 23 3.6 33 4.1
Maize (kgs.) 1.3 0.2 2 0.3
Misc. Beans (kgs.) 3.2 0.5 5 0.6
Melons (number) 938 147 1475 184
Morogo (fresh sacks) 5 0.8 7.6 1.0
Sweet Reed (stalks) 30 4.7 47 5.9
VALUE OF PRODUCTION/b 249.64 39.01 386.40 48.30
COST OF SEED 20.15 3.15 22.90 2.86
GROSS MARGIN
NO Labor, Tract. Cost 229.49 35 86 363.50 45.44
Per Hour .26 --- .28 ---
FIXED COSTS
Equipment 53.90 8.42 65.38 8.17
Lands Buildings/i 33.67 5.26 35.17 4.40
NET MARGIN
No Labor, Trac. Cost 141.92 22.18 262.95 32.87
Per Hour .16 --- .20 ---

 

Source: 1984-85 MVRU and plot monitoring.

a. For this budget, 8 hours a day are assumed for birdscaring.

b. The following prices were used: sorghum - 39t/kg; maize -
60t/kg; cowpeas, other beans - P1.56/kg; morogo P5.00 per fresh
(70 kg.) sac; melons 10t each; sweet reed lot/stalk.

135
agricultural labor [Duggan, 1979] and the number of assumptions required
to value home use of owned traction animals would make any particular
figure meaningless. Consequently, the gross and net margins presented
in the table represent returns to farmers' land, labor, traction and
management.

The gross margin for crop production was approximately P386 per
farm and P48 per hectare for those farms which harvested. The gross
return to farmers' land, labor, traction, and management was
approximately P0.28 per hour (P0.26 for all farmers).

To calculate a net margin, depreciation charges were subtracted for
plows, hand tools, receptacles, and lands buildings. For plows, a
straight line depreciation of current value was used. The plows
averaged nine years of age and it was assumed that they could last, on
average, five more years with no salvage value. The replacement value
was used for hand tools and receptacles, depreciated over five years.
The five years was based on farmers' estimates of the normal useful life
of these items. The charge for buildings was based on the average value
of lands buildings for farmers who did not live at the lands except for
when active in cropping (all but two farmers), depreciated over ten
years with no salvage value. While buildings might have a much longer
life when maintained, it can reasonably be assumed that an investment
equal to ten percent of the value of the buildings would be required to
maintain them in reasonable condition.

Based on the above assumptions, the net return was approximately
P0.20 for the seven farmers which harvested and P0.16 for all farmers
combined. Even without deducting a charge for traction, these returns

to farmers labor, traction and management in the 1984-85 season were

136

below urban or even village wages.

Data from the 1981-82 National Farm Management Survey (also a
drought season) showed even lower returns to arable farming [Boykin,
1983]. Based on 165 households from 11 extension areas, the gross
margins were only P77 per farm, or P19 per hectare. However, the gross
margin per hour was about the same as the MVRU Survey data, P0.30 per
hour (because fewer hours were reported). The NFMS data also showed
that male-headed households and "draft-adequate" households plowed more
area and produced more than did female-headed households and households
without adequate draft resources, respectively.

B. HISTORICAL PERSPECTIVE ON
AGRICULTURAL EXPERIMENTATION IN BOTSWANA

Arable production research in Botswana was initiated in 1936 on a
small experimental field in Mahalapye [ARS, 1978]. During the next
thirty years, the field at Mahalapye remained the main station. There
was a restricted research program, focused on: (a) assessments of new
crOps and varieties, (b) reactions of the new varieties to different
fertilizer levels and selected husbandry practices (such as row
spacing), (c) constraints on the introduction of cash
crops--particular1y cotton, and (d) practices for maintaining soil
fertility--particularly crop rotations [ESM, 1965].

Throughout the pre-independence period, most technical researchers
assumed that Batswana farmers would soon adOpt the capital intensive,
row planting farming methods then being used in South Africa [ARS,
1978]. Their Objective was to identify an appropriate package of
"modern farming methods" for progressive farmers [ESM, 1965].

Substantial progress was made. The greatest success was in

137
identifying appropriate staple food grain varieties. Information was
also generated on the importance of winter plowing and apprOpriate
timing of field Operations. However, the cotton work was a failure
because no solution could be found for the American bollworm. Also,
plant spacing trials did not reveal that any particular pattern was most
advantageous, and fertilizer trials Often were inconclusive [ARS, 1978].

Over time a seemingly effective package of "modern farming"
techniques was develOped. The main elements of the package were as
follows: (a) winter plowing, (b) row planting on early spring rains, (c)
regular weeding, (d) use of quick maturing varieties, (e) crop
rotations, and (g) selective use of manure and fertilizer (mainly
phosphate) [ESM, 1965]. Beginning in 1962, this package was introduced
in a series of steps to a selected number of "pupil farmers." If
farmers eventually adopted all the components of the package, they
became "master farmers" (see Chapter VI).

Just prior to independence, there was a great deal of confidence in
the package. The Economic Survey Mission, which prepared the last
deve10pment plan prior to independence, said the the pupil farmers using
the package were "strikingly successful" in increasing yields, producing
even during drought, adding cash crOps to their cropping systems, and
increasing the area plowed [ESM, 1965: 41].

Soon after independence, the main experiment station was shifted to
Sebele and research continued on dryland farming practices, crop and
variety selection, and plant protection. However, the top priority of
the government was to increase the number of extension workers, in order
to expand their coverage and increase the benefits from the existing

package [MFDP, 1968]. In the 1970 NDP, the government declared that

138
production could be doubled or tripled by increasing the number of pupil
farmers [MFDP, 1970].

In the next National Deve10pment Plan, only three years later,
there was a substantial change in tone [MFDP, 1973]. When rainfall
recovered during the 19703, the relative advantage of the techniques
being used by the pupil farmers was less than expected. Equally
important, national agricultural statistics revealed how atypical the
pupil farmers were. In addition, there was a growing concern with the
effects of existing farming techniques on the environment [Gibbon,
Harvey, Hubbard, 1974]. Consequently, there was a renewed emphasis on
research, and national research policy shifted to focus on risk reducing
and soil conserving practices.

During the 19708, research at Sebele--and at sub-stations located
at Goodhope, Mahalapye, MotOpi and Moshu--included trials on dryland
crop varieties, fertilizers, crOp rotations, tillage systems, and
horticulture crOps. Much of the research at Sebele, and off-station,
was conducted through donor funded projects, including a FAQ fertilizer
project, a British bird pest research project, UNDP/FAO research in the
Okavango, and a U.S. crOp improvement research project.

The backbone of arable production research between 1971 and 1983
was the British sponsored Dryland Farming Research Scheme (DLFRS). The
Objectives of DLFRS were to study factors limiting arable production and
to investigate means of making crop production more reliable. During
Phase I, which lasted from 1971 to 1974, DLFRS took an "applied
research" approach--as directed by the Ministry. By 1974, a minimum
tillage crOpping system was develOped, derived from techniques used in

the U.S.

139

The DLFRS system was based on the Makgonatsotlhe two-wheeled tool
bar and included the following: (a) crop rotation including bare
fallowing before sorghum, (b) maintaining a weed free environment using
blade sweeps, (c) sub-soil plowing (with a chisel) during autumn, (d)
row planting after the first spring rains, (e) inter-row weeding
immediately after emergence and whenever necessary afterwards, (f)
thinning, (g) tOp dressing N in wet years, (h) breaking the surface to
improve rainfall infiltration, (i) harvesting early in the season, and
(j) sweeping crOp stubble immediately following harvest [Gibbon, Harvey,
Hubbard, 1974]. To prevent runoff, all Operations were carried out on
the contour. With this system, yields on-station were more than 600
percent higher than the national average [DLFRS, 1977].

In 1975, two companion British projects were established to begin
testing the recommendations develOped by DLFRS on farmers' fields. The
Evaluation of Farming Systems and Implements Project (EFSAIP) had both
on-station and on-farm research activities, and took over responsibility
for implements development and testing. The Integrated Farming Pilot
Project (IFPP) was established in the Department of Agricultural Field
Services in order to provide an extension situation under which
technologies developed by DLFRS and screened by EFSAIP could be tested
under farmers' conditions with farmers' management.

It was quickly found that the Makgonatsotlhe system did not work
when moved off-station. In EFSAIP trials in 1976-77 and 1977-78, the
Makgonatsotlhe system gave higher yields than the traditional broadcast
system but the gross margin per hectare (not including capital charges)
was negative [EFSAIP, 1978; 1979; Gulbrandsen, 1980]. Also, the amount

of grain produced per hour invested was lower than for the traditional

140

check plots [Alverson and Srivastava, 1980]. However, an intermediate
system involving autumn plowing and spring planting (using a mechanized
method of furrow plow-planting) slightly outperformed the traditional
system in 1976-77 and did significantly better in the 1977-78 season.
Similarly in IFPP trials, the Makgonatsotlhe system did worse than
traditional broadcasting but a "modified traditional" mechanical row
planting system increased the gross margin per hectare by 33 percent
[Gulbrandsen, 1980].

During the late 19708, the Makgonatsotlhe system was abandoned. In
addition to modest yields benefits, the equipment was expensive and did
not reduce draft requirements as expected. Many of the associated
crOpping systems recommendations could not be implemented by farmers.
As Gibbons had said, "to raise output significantly the system needs to
be applied as a whole and not as a series of isolated steps." [Gibbons,
n.d,: 7]. Unfortunately, farmers generally cannot and do not adopt
entire packages.

In 1982, when on-farm research began in the Central Region, there
was a great deal of uncertainty about appropriate crOp production
practices. There were disagreements among technical researchers over
the advantages of row planting versus broadcasting, sole versus
intercrOpping, and the profitability of using fertilizer [Lightfoot,
1983]. The possibility of pesticides or herbicides was not seriously
considered, given the low yield base. Even the prevailing plant
population recommendation for sorghum did not provide much guidance;
45,000 plants plus or minus 20,000 plants was recommended for much Of
the country.

The extension service continued to recommend many of the practices

141
originally develOped for the pupil farmers, but there was little
research supporting the recommendations. Moreover, as the national
statistical data showed, few farmers were paying any attention to the
recommendations.
C. CENTRAL REGION ON-FARM TRIALS

Between 1982 and 1986, many interventions were tested in a variety
of experimentation formats. The trials program primarily focused on
modifications in the traditional broadcast, single plow system.
Emphasis was given to minor modifications, generally involving an
additional tillage and/or the introduction of more precise planting
methods. More substantial modifications were investigated beginning the
third season, but only in RI trials.

Two other tOpics were systematically investigated: (a)
production of secondary crOps and (b) post—establishment plot management
practices. Research on secondary crops mainly comprised a series of
crop-variety comparison trials, plus attempts to identify specialized
planting methods for crOps or varieties which established poorly in the
traditional broadcast system. Post-establishment plot management
practices emerged as an important theme during the third and fourth
seasons, primarily in relation to stand adjustment techniques (thinning
and replanting). Several planned post-establishment trials were
abandoned due to the drought.

In addition to the primary investigations, many issues were
addressed on an ad hoc basis. For example, garden plots, manuring and
ridging were evaluated--primarily to assess their potential. Some
investigations were added in the interest of establishing collaboration

with Sebele researchers and researchers from the other on-farm teams.

142

This section gives an overview of the trials program and a summary
of trial results, on a season-by-season basis. The summary of trial
results is primarily based on analyses carried out by J. Siebert and
presented in ATIP [1986].

1. 1982-83 SEASON

During the first season, the experimentation program consisted of
three small trials, all based on issues already under investigation
elsewhere in Botswana. The main experiment was a comparison of five
planting methods-~each representing a different degree of seed placement
control. The objective was to evaluate the importance of improved seed
placement for sorghum stand establishment and yield. In addition to a
broadcast control plot, the methods tested were third furrow hand
placement, the EFSAIP plow-planter (a mechanized version of furrow
placement), the EFSAIP single row planter, and use of a harrow to cover
broadcast seed. The last method was suggested by farmers during the
exploratory surveys.

In the second trial, double plowing--once early and once at
planting--was assessed at two sites. Double plowing was examined
because EFSAIP had found in a "time of planting" trial that double
plowing seemed to improve water infiltration and soil tilth, leading to
better stand establishment [EFSAIP, 1983].

In addition to the tillage trials, one RI trial was carried out in
which three special purpose fodder crOps (teff grass, Dolichos lab lab,
Bambala millet), as well as the traditional (Tswana) cowpea variety,
were undersown following the establishment of a sorghum crop. The
Objectives were to test crop varieties suitable to undersowing and to

evaluate a hand jab planter.

143

Results for the planting methods trial failed to show significant
differences for plant stand establishment. A majority of the
comparisons failed to give any grain yield and no yield analysis was
carried out. In the two double plowing comparisons, double plowing
yielded substantial benefits for plant establishment and grain yield.
In the undersowing trial, Tswana cowpeas produced as much vegetative
material as any of the special purpose fodders. Farmers said they would
prefer cowpeas, because there could be a grain yield.

2. 1983-84 SEASON

In the second season, early plowing was added to the planting
methods trial. Sixteen farmers sole plowed (plowed without planting)
and later superimposed the different planting methods. Super single
phosphate was applied in strips across the sole plowing and planting
methods trial. There were statistically significant plant stand and
grain yield benefits from double plowing and early plowing plus row
planting when compared with the traditional and other tillage-planting
systems. However, the differences were not substantial. There were no
measurable benefits from the phosphate applications.

The double plowing trial from the first season was modified to
include three early tillage options: deep plowing, shallow plowing and
harrowing. This was done in order to separate the benefits of early
double plowing into effects due to time of the Operation versus the
depth of plowing. The early plowing strips (implemented at three
depths) produced a benefit for stand establishment but the effects were
confounded with pocket variability of soil types within fields. Due to
drought, grain yield effects due to treatments could not be measured.

There was no visible effect due to the depth at which early plowing was

144
done. Harrowing had a negative effect on soil structure and plant
growth.

Because of the continued enthusiasm of EFSAIP, the plow planter
unit was compared to the traditional broadcast system in a separate FM,
FI trial. Few implementations were obtained. The unit was difficult to
use when mounted on a single furrow plow (since farmers normally drOp
the plow when turning at the top of acres). There was little farmer
interest in the plow-planter unit. -

In response to an observed problem of weak draft animals, six
farmers were provided with a supplemental ration of sorghum and cowpea
residues for feeding their draft teams. The objective was to maintain
animals during the plowing period by providing energy supplements as a
complement to existing grazing. (Prior feeding schemes had concentrated
on the unrealistic goal of feeding draft teams throughout the winter.)

When provided with residues, the farmers were willing to feed their
draft animals prior to plowing but did not want to feed residues at
breaks in the plowing Operations. Cattle Often would not eat the rough
residues even when provided (although donkeys did). Overall, it did
appeared that neither farmers nor the draft animals were reaponsive to
the concept of a residue harvesting and supplemental feeding scheme.

A secondary crOps comparison trial was initiated in order to
evaluate stand establishments and relative yields for selected secondary
crOps, particularly GOWpeas, under the traditional broadcast system.
Most replications arranged were never implemented due to the drought.

Of the comparisons planted, only seven replications of the comparison of
cowpea varieties established measurable stands; of these only five

produced grain yield. Both Blackeye and Tswana cowpeas emerged better

145

than ER-7 (a determinant variety promoted by the Sebele-based Cowpea
CRSP). Tswana and Blackeye cowpeas also produced significantly higher
grain yields.

A relatively minor birdscaring scheme, based on a polyethylene "hum

' was added late in the season. The hum line was installed on some

line,‘
sorghum plots in four farmers' fields. Bird damage was reduced to an
extent, but doves seemingly were unaffected. To be installed at a heavy
enough concentration to be effective, the cost of the "hum lines" would
be prohibitive.

3. 1984-85 SEASON

During the third season, an RM,RI "steps in technology" factorial
trial was implemented in order to assess combinations of
tillage-planting methods. The experimental treatments were sub-soiling,
early plowing, phosphate fertilizer, and several planting methods. The
objective of the factorial trial was to evaluate the technical
performance of various inputs. Sub-soiling and phosphate fertilizer
gave significant, but not substantial, benefits for sorghum grain yield.
The only substantial treatment main effect was from row planting. There
were significant interactions with location.

Two "design" activities were not based on a formal trials. One
involved an RM,RI technical evaluation of a ridging plow and a ridging
planter. Neither piece of ridging equipment performed well on
previously untilled soil. When used on previously tilled soils, ridge
construction was poor by both units and the ridges disappeared following
a heavy rainfall. Emergence of planted seed was poor.

The second design activity was an examination of intensive

production, involving early planting and the addition of corral manure

146
at sites selected for high production potential (usually having deeper
soils). The intensive production plots also failed. Stand
establishment was adequate on only three of five sites and there was no
measurable difference in grain yield due to the corral manure
applications.

The F1 tillage-planting methods trial was modified to focus on
early plowing plus row planting and double plowing, versus the
traditional system. Sole plowing was targeted to drying soil moisture,
to reduce the Opportunity cost of the sole plowing. The planting
methods strips were implemented on two days, one with good soil moisture
and one with drying soil moisture. The benefit of early plowing for
grain yield was statistically significant, but not large, as in the
previous season. Most farmers were not able to follow the plowing
moisture guideline. They complained that the condition of their animals
would not permit the harder work on drying soils and the abrasive action
of the dry soil would quickly ruin a plow share.

A modified version of the secondary crops comparison trial was
again tested during the third season. Sri Lanka cowpeas (purchased and
imported by the government due to a seed shortage) and mung been
established and yielded well. Blackeye and Tswana cowpeas established
well but gave an erratic yield performance. ER-7 cowpea was erratic in
establishment but yielded well on a per plant basis. Groundnuts
established and yielded poorly. Tepary bean gave adequate establishment
but limited grain yield.

As a complement to the crOpping comparison trial, a trial comparing
hand planting versus broadcast planting of four cowpea varieties was

implemented. The Objective was to test differences in establishment

147
between hand planting (with control over seed depth placement) and
broadcast planting after having controlled for soil moisture (by making
all plantings on good moisture). Good emergence plant stands were
Obtained for all varieties and both planting methods (ranging from
36,000 to 53,000 plants per hectare). Neither planting methods nor
plant methods by variety effects were significant. Yields were low but,
except for Sri Lanka, were much higher than those in the crOpping
comparison trial. The results showed that adequate stands of the new
cowpeas, ER-7 and Sri Lanka, are obtainable using the standard broadcast
system.
4. 1985-86 SEASON

During the fourth year, two RM,RI tillage-planting trials were
implemented. One was the second year of the "steps in technology"
factorial trial. Grain yield results for double plowing and row
planting did not differ substantially from prior years. Most of the
significant responses to early plowing and phosphate were obtained at
environments with the highest total water holding capacity and when
drought conditions immediately followed planting. Row planting gave
significant benefits for grain yield regardless of the soil type or
rainfall pattern.

The second RM,RI tillage trial focused on several "water
harvesting" tillage—planting systems, including deep ripping, minimum
tillage and strip cultivation, as well as combinations of some of the
components tested in the other RI trial. The strip tillage systems were
difficult to implement, produced smaller and fewer plants, had
u“Hitacceptable weed burdens along the untilled strips, and resulted in

very low yields.

148

In the FI program, separate experiments were set-up for row
planting and double plowing. Double plowing was compared to the
traditional system at more than thirty sites in order to evaluate the
relative benefits across soil and weed environments. A yield increase
of 94 percent was obtained across 31 sites. Yield levels, however, were
quite low for both the traditional and double plowing systems. Weed
control was significantly improved through double plowing.

Three tillage-planting systems were compared for row planting: (a)
plowing and planting on the same day, (b) early plowing followed by row
planting on a subsequent rain, and (c) early plowing followed by a weed
control tillage (cultivating or harrowing) and row planting after a
subsequent rain. The goal was to evaluate the benefits of early plowing
(and the associated problem of weed development before planting) versus
the standard row plant system based on same day plowing. The trial was
severely affected by poor moisture conditions. Due to the absence of
rainfall during the early planting period, little rainfall was conserved
through early plowing and weed control was not a serious issue in any of
the plots.

Following two years of the RM,FI cropping comparison trial, sole
cropping of secondary crOps (in small plots) was examined in an FM,FI
framework in the 1985-86 season. Small plots were recommended because:
(a) potential labor conflicts between sorghum and secondary crops could
be minimized, and (b) small plots would be sufficient to meet the
limited demand for secondary crOps. The objectives of the trial were
to: (a) facilitate farmer assessments of the advantages and
disadvantages of sole crOpping and (b) collect data for crOp specific

enterprise analysis.

149

The farmers who tried to implement the sole planting trial
encountered many problems. A late start to the rains prevented many
farmers from planting jugo beans or groundnuts. Plant establishment for
most crOps was less than had been hOped, in part because the recommended
soil moisture guidelines were not well followed. Even when reasonable
establishment was achieved, insect pests, heat, drought, disease and
wild animal damage destroyed most plots.

Hand furrow planting was again examined, this time as an Option for
secondary crops that do not establish consistently in the traditional
broadcast system. Few implementations were achieved. Based on limited
observations, there did not appear to be any advantage in hand furrow
planting for sunflower, groundnuts or jugo beans. The first two
established poorly regardless of the planting system, while jugo beans
established well under broadcast. There was somewhat better
establishment with hand furrow planting for ER-7 COWpeas and maize, but
farmers expressed little or no interest in pursuing hand furrow planting
on their own.

Until the fourth season, the main strategy for improving plant
stands was to modify the tillage-planting system. However, this
strategy was practical only for draft-controlling households--as low
trials implementation rates demonstrated. Research on a strategy for
draft-dependent households was initiated during the fourth season. The
strategy was to invest labor in post-establishment stand management,
including gap filling and thinning.

Hand planting for gap filling was examined for sorghum and millet.
Millet (cv. Serere 6-A) and sorghum (cv. 65-D) were planted in paired

plots with two replications. On some fields, the main plots were split

150
into dry versus wet planting sub-plots. The average planting time was
only 10.2 hours per hectare. Acceptable stands were achieved: 45,663
plants per hectare for sorghum and 42,428 for millet. Dry plantings
established significantly better than wet plantings. Unfortunately,
after the plots were planted, there was little rain. Plant development
was poor and yields were low; 134 kgs. per hectare for millet and only
10.2 kgs. per hectare for sorghum.

Thinning was evaluated in two formats. As part of the double
plowing trial, thinned strips were superimposed across the single and
double plowed plots in order to standardize the post-establishment plant
populations. This thinning resulted in a significant (though not large)
yield benefit even though stands in the thinned plots were not
significantly reduced. The result was largely attributable to the
elimination of non—viable plants rather than a reduced overall
pOpulation.

In a separate thinning trial, farmer group members (see Chapter VI)
were asked to identify parts of their field where they felt the stands
were too high. Thinning was targeted to areas where the plant response
could be expected to be the greatest in order to maximize the potential
returns to thinning labor and because women do not have the time to thin
their entire field. Eight comparisons were implemented. The average
time required for thinning was 39.5 hours per hectare. Measurements
were made around six weeks after thinning on: live plants, heads, plants
with less than six surviving leaves, and dead plants.

Unfortunately, most plots were not thinned enough or early enough.
As a result, there was not sufficient compensation due to the lower

Population on the thinned plots (with respect to heads per plant or

151

average head weight) to provide a reasonable return to labor invested in

thinning. There were, however, some signs that proper thinning could

increase the production potential of areas with extremely high initial
establishments (and provide reasonable returns). A regression analysis

of the trial results showed, for example, that for each additional 1,000

plants, the percentage of viable plants decreased by nearly two percent

(r = -0.89) and the heads per plant decreased by 0.025 (r = -0.56).

D. BUDGET ANALYSES OF SELECTED PRACTICES
This section presents budget analyses Of the most promising
modified practices, including: (a) double plowing, (b) early plowing
plus row planting, (c) sole planting of secondary crops, and (d) gap
filling.

1 . DOUBLE PLOW ING

During the first three seasons Of RM,FI trials, the average yield
increase for double plowing was just greater than forty percent. In the

1985-86 FM,FI trials, there was a 94 percent yield increase due to

double plowing. Double plowing also gave statistically significant

benefits for grain yield in the RM,RI factorial trial.

a. RM,FI Trial Results

After the 1984-85 season, a partial budget analysis was carried out
using the combined RM,FI trial results. Two specific issues were
addressed:

1) Should a farmer double plow or single plow a given hectare? This
issue is relevant to farmers who have a limited amount of land to
plow or who are otherwise constrained in reaching production goals
via planting greater area. It is also relevant to farmers who are

able to do an early plowing at a time when they would otherwise be

152

unable to plant.

2) Should a farmer double plow a given hectare or single plow two
hectares? This issue is relevant to farmers who are not land
constrained, but may have limited weeding labor. The key trade-off
to these farmers is the Opportunity cost of the yield from the plot
which might have been planted when early plowing was done. If
farmers lack the labor resources to manage two plots, they might
instead increase the inputs to a single plot, thereby by generating
higher returns to the labor which is available, even if farm grain
production is reduced somewhat.

The issues were analyzed through partial budgeting. Results are
presented in Table 5.5. Each budget presents the expected changes in
costs and benefits associated with the prOposed shift to double plowing.
The main costs and benefits are grain yields, labor time, and seed. The
second part of each table presents a sensitive analysis of the basic
budget taking into account different yields, resource use levels and/or
resource valuation.

The analysis in the basic budget was oriented to traction owners
since these farmers are the ones most likely to be able to implement
double plowing. To the extent possible, resources were valued at their
Opportunity costs. In the case of labor, the government's Labor Based
Drought Relief (LBDR) wage rate was used. There is no formal
agricultural labor market in Botswana [Duggan, 1979] and the LBDR
program provides the main Opportunity for short term rural wage
employment. Moreover, the LBDR wage is very close to the few reported
observations on agricultural wage employment (for example, EFSAIP

[1983]). For yields, local market prices were used. Local prices

153

TABLE 5.5:

1983-84 AND 1984-85

PARTIAL BUDGET: RM,FI DOUBLE PLOWING,

 

 

 

 

DP 1 HA VS DP 1 HA VS
SP 1 HA SP 2 HA
BASIC BUDGET:

Reduced Cost

Seed Saving ----- 3.12/a

Weeding Time Saving 3.57/b 12.46/c

SP Harvesting Time/d 8.90 18.47
Added Benefit

DP Yield (198 kgs) 77.22 77.22
Added Cost

DP Second Plow/e 9.04 -----

DP Harvesting Time 15.05 15.05
Reduced Benefit

SP1 Yield (126 kgs) ----- 49.14

SP2 Yield (117 kgs) 45.63 45.63
Net Gain 19.97 1.45

SENSITIVITY ANALYSIS:

Yields from 1983-84 Trial Only/f 23.73 28.14
Yields from 1984-85 Trial Only/g 16.83 -27.43
No Weeding Time Saving 16.39 -ll.01
Weeding Time Saving Plus 50% 21.75 7.69
Labor at Urban Minimum Wage (53t) 18.95 7.72
Grain Market Board Price (27t/kg) 10.25 5.89
Hired Traction (P50/ha) -20.99 1.45

 

a.
b.

Seeding rate I 8 kg/ha.
Weeding times were not recorded.

All grain valued at 39 t/kg.
An imputed weeding time was

calculated on the basis of plot monitoring data and data from
EFSAIP. EFSAIP data show a 602 weeding time saving on double
plow plots. The average weeding time by cooperators on plots
weeded was 37.4 hours. Only 12 of the 19 trial plots were
harvested. Assuming no gain in weeding time on plots not
harvested, the calculated weeding time savings would be .63 x
.4 x 37.4 I 9.4 hrs. All labor valued at 38 t/hr, the LBDR
wage.

Weed times saving of 9.4 hours plus the 23.4 hours required
for the second traditional plot. 23.4 hours is the time
spent weeding, averaged across all plots.

Based on a standardized harvesting and threshing rate Of 5
kg/hr. This is the average Obtained from monitoring data
[ATIP, 1986b].

The average plowing time for trial participants was 11.9
hrs/ha. Plowing labor was standardized to two peOple.

Yields were low: 53 kgs for SP1, 24 for SP2, and 117 for DP.
Higher yields: 271 for DP, 208 for SP1, and 200 for SP2.

154
represented a competitive equilibrium price, since trade in farm
products is not regulated in Botswana.

Based on the 1983-85 trials, double plowing would have increased
the value of production by more than 31 Pula and would have provided a
net gain of nearly twenty Pula per hectare. The double plowing system
required 30.6 more hours, so the return per extra hour invested would
have been over one Pula per hour. Thus, double plowing provided an
Opportunity to work more hours during drought at an imputed wage which
exceeded the urban minimum wage rate.

The sensitivity analysis showed that there would have been a net
gain per hectare in both the 1983-84 and 1984-85 seasons, the first of
which had very low yields while the second had relatively high yields
for drought conditions. Valuation of labor using the urban wage rate
had little effect on the profitability of double plowing. Use of the
(low) BAMB price for sorghum reduced the net gain. An increased weeding
savings would have increased the net gain.

The net gain from double plowing would have been lower for farmers
substituting one hectare of double plowing for two hectares of single
plowing. Due to the foregone yield from the second hectare, the net
gain would have been only Pl.45. However, forty fewer hours were
required for the double plowing system, mostly due to fewer weeding
hours. Thus, households facing a weeding labor constraint would have
benefited from this system change. While farm grain production would
have been reduced, farm system profitability would have been increased.

One drawback with the decision to double plow is a risk that the
relative benefit might not hold up with better rains. A decision to

substitute one double plowed hectare for two single plowed hectares

155

would have resulted in a loss in the 1984-85 season, when higher yields
were Obtained for both single and double plowing.

In the two hectares versus one sensitivity analysis, valuing labor
at the urban minimum wage rate increased the net gain for double
plowing. Lack of a weeding saving resulted in a loss from double
plowing, highlighting the importance of weeding benefits when assessing
the value of double plowing.

The sensitivity analysis on traction hiring points out several
crucial issues for draft-dependent households. The negative return from
the first decision (1 ha DP instead of 1 ha SP) reflects the fact that
neither system resulted in yields which were sufficient to justify an
investment in tractor hire. Draft-dependent households would have been
better off not plowing at all. Given the fact that crOp failure, or
near failure, is always a possibility in Botswana, the analysis suggests
that hiring households should minimize their risk of loss by only
plowing once.

However, the recommendation gets more complicated when one
considers the second decision as well (1 ha DP instead of 2 ha DP).
With this substitution, there would have been a net gain, and a
reduction of labor requirements. The complications derives from the
source of funds used for plowing. If the government were to give money
only for plowing, or relatives sent remittances only for plowing, it
would have been more profitable to invest that money in double plowing
half as much land as could have been single plowed.

As with tractor owners, the second decision would have increased
profitability and reduced labor requirements, but also reduced farm

production. Which Option is best for a hiring household, therefore,

156

depends on a combination of labor availability, any restrictions the use
Of funds, and relative interest in profitability versus total farm
production. On balance, it would not seem advisable to recommend double
plowing to hiring households except on a Special case basis.
b. FM,FI Trial Results

The same two questions were addressed in 1986 using results from
the 1985-86 FM,FI double plowing trial. The partial budget analyses
again showed there would have been a net gain from double plowing, as
seen in Table 5.6. The net gain for double plowing any given hectare
would have been slightly lower than for the 1983-85 RM,FI trials, but
the net gain for substituting one hectare of double plowing for two
hectares of single plowing would have been greater. When analyzed on a
individual site basis, there would have been a positive gain for double
plowing a given hectare at two-thirds of the sites. The decision to
substitute one hectare of double plowing for two single plow hectares
would have resulted in a net gain at 86 percent Of the sites. Total
production was almost as great from the one double plowed plot as two
single plowed plots, but the total time invested was substantially less.

As before, it would not have been profitable for households hiring
traction to invest in a (first or) second plowing on any amount of land.
However, for any given amount spent on plowing, it would have been more
profitable to double plow half as much land as could have been single
plowed. Again, farm system profitability might not be the main issue in
such a decision, since total farm grain production would decrease.
c. RM,RI Sensitivity Analysis

Table 5.6 also presents a sensitivity analysis based on the RM,RI

trial results, so two additional issues can be addressed: application

157

TABLE 5.6: PARTIAL BUDGET: 1985-86 FM,FI
DOUBLE PLOWING TRIAL

 

DP 1 HA DP 1 HA
VS 1 HA SP VS SP 2 HA

 

BASIC BUDGET:

Reduced Cost

Seed Saving/a ----- 2.34
Weeding Time Saving/b 1.41 8.93
SP Harvesting Time /c 6.80 13.40

Added Benefit

DP Yield 67.74 67.74

Added Cost

DP Second Plow/d 11.71 _____
DP Harvesting Time 13.20 13.20

Reduced Benefit

SP Yield 34.91 69.81

Net Gain 16.13 9.40

 

 

SENSITIVITY ANALYSIS:

Hired Traction/e -22.16 9.40
RI Yields: All Sites

NO Phosphate 20.60 -85.99
With Phosphate/f 0.43 -97.66

RI Yields: Environment A/g

No Phosphate 105.94 -34.56
With PhOSphate 145.25 9.70

 

Seeding rate = 6 kg/ha. All grain valued at 39 t/kg.
Average weeding times were 19.8 hrs/ha for SP and 16.1
for DP. All labor valued at 38 t/hr (the LBDR wage
rate).

Based on harvesting and threshing rate of 5 kg/hr.
Standardized to donkey plowing with two workers. Average
donkey plowing time was 15.4 hr/ha.

. Substitutes the standard plowing hire rate in Central

Region for the plowing labor charge.

Phosphate budget includes: (a) labor charge for two hours
to haul bags and broadcast, and (b) a P62.89 charge for
3.8 bags Of SSP at P16.55 per bag. The without P or DP
yields are compared to the with P and DP yields.
Environment A has deep to moderately deep soils with high
total water holding capacity and with a rainfall pattern
favorable for response to early tillage (i.e., good
pre-planting rain with dry conditions immediately
following).

158
of phosphate and site selection. The analysis assumes that the relative
weeding times would have been the same for the RM,RI trial as were
observed for the FM, FI trial. The reasonableness of this assumption
cannot be proven. Harvest times for all the partial budgets were based
on a standardized ratio of time per unit harvest, since harvesting times
are only affected by tillage treatments via the effects of the
treatments on yields.

Based on yields achieved in the RM,RI trial, there would have been
a net gain (for traction owners) from substituting double plowing for
single plowing, but only if land was a constraint. Otherwise, it would
have been better to keep single plowing additional hectares. This
result was due to the relatively better yields achieved with single
plowing when single plowing was researcher managed and implemented.

The analysis using RM,RI trial shows that the profitability of
double plowing is quite sensitive to the soil and rainfall environment
in which double plowing is implemented. In Environment A, that most
conducive to a double plowing benefit, the net gain per hectare would
have exceeded P105. Even in Environment A though, it would have been
more profitable to single plow two hectares than it would have been to
double plow a single hectare, if land and weeding labor were available.

Averaged across all sites, the combination of phosphate and double
plowing would have been even less profitable than double plowing alone.
However, in the proper environment, there was a synergistic effect
between double plowing and phosphate, which would have made the
application of phosphate profitable. With phosphate application, the
decision to double plow would have been profitable in Environment A even

if traction had been hired.

159

2. EARLY PLOWING PLUS ROW PLANTING

The yield increases Obtained from row planting in the RM,FI
tillage-planting systems trials during the first three seasons were
almost identical to those from double plowing. Row planting yields

Often were higher in the RM,RI trials than with double plowing, and row

planting has a smaller draft requirement than does double plowing.

a. The Decision to Row Plant

Following the 1985 season, two issues were addressed for row
planting, which are analogous to those raised for double plowing:

1) Should a farmer early plow and row plant rather than broadcast and
plow a hectare? This issue is relevant to farmers who own row
planters but do not use them even though they have a limited amount
of land or can plow when there is too little moisture to plant. The
issue could be of some importance in the Central Region where many
farmers who own row planters use them infrequently.

2) Should a farmer early plow and row plant one hectare rather than
broadcast and plow two hectares? This issue is most relevant to
farmers who have the land, but not the labor resources to support an
extensive orientation toward farm production (but do own a row
planter).

The results of partial budget analyses Of these questions are
summarized in Table 5.7. The results are quite similar to those for
double plowing. For each hectare farmers early plowed and row planted
they would have gained over P21, relative to the traditional
broadcast-single plow approach. The return per extra hour invested
would have been above one Pula. The net gain and returns to labor would

have increased even more if farmers would have weeded using a

160

PARTIAL BUDGET: EARLY PLOW AND
ROW PLANT, 1983-84 AND 1984-85

TABLE 5.7:

 

 

 

 

RP 1 HA VS RP 1 HA VS
BP 1 HA BP 2 HA
BASIC BUDGET

Reduced Cost

Seed Saving/a 0.78 3.90

Plow 2nd BP plot ---- 9.04/b

Weeding 2nd BP Plot ---- 8.89

BP Harvesting Time/c 8.90 18.47
Added Benefit

RP Yield (201 kgs) 78.39 78.39
Added Cost

Hours Row Planting/d 6.08 6.08

RP Harvesting Time 15.28 15.28
Reduced Benefit

SP1 Yield (126 kgs) ----- 49.14

SP2 Yield (117 kgs) 45.63 45.63
Net Gain 21.08 2.56

SENSITIVITY ANALYSIS

Yields from 1983-84 Trial Only 19.67 24.23
Yields from 1984-85 Trial Only 22.49 -21.92
Saving from Cultivator Weeding/e 28.03 9.51
Labor at Urban Minimum Wage (53t) 16.16 8.50
Grain Market Board Price (27t/kg) 10.76 6.40
Hired Traction (P50 plow; P25 plant) 2.16 24.60

 

Row plant seeding rate = 6 kg/ha. Broadcast seeding rate
= 8 kg/ha. All grain valued at 39 t/kg.
Plowing rate of 11.9 hr/ha (timed rate, some used double

furrow plows). Standardized to two peOple plowing. All
labor valued at P0.38/hr.

Based on harvesting and threshing rate of 5 kg/hr.

Row plant rate of 8 hr/ha with two people planting.
Saving from cultivator weeding of 18.3 hrs/ha. This is

the difference between an average weeding time of 37.4
hrs/ha on hand weeded plots (obtained from plot
monitoring) and the average cultivating time of 8.4
hrs/ha (measured trial data). The difference, 29 hrs. is
multiplied by .63, since only 12 of 19 plots were
harvested. Farmers usually do not weed plots which are
abandoned before harvesting.

Low yields: 104 for RP, 53 for BPl, and 24 for BP2.
Higher yields: 288 for RP, 208 for BPl, and 200 for BP2.

161
cultivator. Both a higher Opportunity cost for labor and a lower price
for output reduced the net gain in the one hectare versus one hectare
decision. Unlike double plowing, however, early plowing and row
planting would have been a better Option than broadcasting even if
traction had to have been hired for both the plowing and row planting
Operations.

There would have been a slight net gain for the decision to
substitute one row planted hectare for two double plowed hectares. The
value of production would have decreased by over P16, but the value of
the labor saved would have been even greater, increasing farm system
profitability. However, a net gain would have been achieved in only one
of the two seasons, the one with worse rains and lower yields. This
suggests that the relative advantages of intensification are less in
good seasons. This could be a major reason farmers have not been more
responsive to the extension recommendation to row plant.

In the sensitivity analysis for the one row planted hectare versus
two broadcasted hectares, a higher Opportunity cost for labor or lower
value for output increased profitability relative to the basic budget.
Row planting should be particularly attractive for hiring households,
assuming Of course they can find peOple who will row plant at the rate
of P25 per hectare.

b. The Decision to Buy a Planter

Because row planting involves a new implement for most farmers, an
additional question was addressed: should a farmer who does not own a
row planter buy one, given the row planting trial results (obtained
during drought seasons)?

The decision to invest in a row planter was addressed using capital

162
budgeting and the net gain from Table 5.7. The basic budget assumed a
ten year life for the planter, that the planter could be used for two
hectares each year without a reduction in the average area planted, and
a discount rate of ten percent. Sensitivity analysis took into account
different discount rates, numbers of hectares row planted each year, and
acquisition costs.

Results on the benefit-cost ratio, the net present worth, the
internal rate of return, and the "break—even net gain" are summarized in
Table 5.8. One would ordinarily make an investment if the benefit-cost
ratio is above one or the net present worth of the investment is greater
than zero. The internal rate of return represents the average annual
rate of return to the capital invested in the row planter. This rate of
return can be compared to the Opportunity cost of alternative capital
investments. The "break-even net gain" is the threshold level of net
gain required for the benefit-cost ratio to equal one (or the net
present worth to be zero).

The basic capital budget analysis indicated that farmers who did
not own a row planter should not have bought one. However, the
sensitivity analysis showed that several factors could have made the
investment profitable. For example, any farmer Obtaining a higher than
average net gain per hectare (for example, through site selection or
learning how to row plant better) or row planting more hectares each
season, could have profited from investing in a row planter.

Most significantly, a farmer would have had to obtain a net gain of
only P6.15 per hectare in order to have profitably invested in a row
planter—-if the row planter was purchased through ALDEP. Under the 85

percent ALDEP subsidy, the benefit-cost ratio for a row planter

163

TABLE 5.8: INVESTMENT IN A ROW PLANTER:
BENEFIT-COST RATIO, NET PRESENT WORTH,
INTERNAL RATE OF RETURN

 

 

 

ASSUMPTIONS:
Discount rate: 0.10 Net gain per hectare: 21.08/b
Initial cost: 350.00/a Hectares RP per year: 2
Years: 10 Maintenance per year: 9.85/c
BENEFIT NET PRESENT BREAK-
-COST WORTH I.R.R. EVEN GAIN
BASIC BUDGET/a 0.69 -115.42 -18.8 30.48

SENSITIVITY ANALYSIS:
Discount Rate:

OZ 0.94 -26.90 ---- 22.43
5% 0.80 -80.67 ---- 26.31
152 0.61 -137.48 ---- 34.83
Hectares RP:
3 1.04 14.11 11.3 20.32
5 1.73 273.17 35.4 12.19
Cost:
85% Subsidy/d 3.43 187.57 119.0 6.15
502 Subsidy 1.30 60.46 20.4 16.17

 

8. Average cost of single row planter in Mahalapye area.

b. Based on Table 5.7.

c. Maintenance charges assume P25.00 in set-up cash costs plus
cash maintenance costs equal to 1% of the original cost per
year beginning in year 5 and increasing each year following
by 12 per year. Higher maintenance costs would be
inconsistent with observed practices, particularly under the
assumed usage rate of only 2 hectares per year.

d. The current ALDEP subsidy.

164

investment would have been 3.4. Even with only a 50 percent subsidy the
benefit-cost ratio would have been 1.36.
3. SOLE PLANTING OF SECONDARY CROPS

During three seasons in which broadcasted secondary crOps were
compared in both an RM,FI and FM,FI format, yields were low. However,
because of their high value during drought, it may still be advantageous
to sole plant secondary crOps.

Following the third season, yield data from the crOpping comparison
and the cowpea planting methods-variety trials were used to make a
preliminary assessment whether farmers who do not currently plant small
plots of secondary crops in sole stands might benefit from doing so.

The analysis was based on a comparison of the net value of production
per hectare (yield minus the amount of seed planted) for sorghum and the
main secondary crOps. (Because of the original trial objectives and the
small plot sizes, no attempt was made to record labor data, so returns
to labor could not be calculated.) Sorghum yields were based on
findings from the MVRU Survey.

The value of minor crops (eg. sweetreed and melons) normally
present in sorghum mixtures was not included in the analysis since it
was assumed that all requirements for these crops could still be met if
only a small portion of land was diverted to sole planted cowpeas,
groundnuts, mung bean or tepary bean. Similarly, the value of the
morogo (spinach) from the sole planted cowpea plots was not included
since it was assumed that the normal crop mixtures of farmers satisfy
the demand for morogo. Two sets of prices were used to value the grain
produced. The first set was based on results from the 1985 Trader

Baseline Survey. These prices had a large gap between the value of

165
sorghum and the various secondary crops. The second set was the BAMB
prices for Mahalapye. BAMB price differentials were much smaller.

The results of the analysis are presented in Table 5.9. Although
widely varying yield results were obtained during the two seasons,
essentially all the secondary crOps gave a higher net value of
production than did sorghum. The value of production was less only for
ER-7 cowpeas in the 1983-84 season when using BAMB prices. In general,
groundnuts and ER-7 cowpeas returned the lowest value of production,
despite the high prices received for groundnuts. In two of the three
trials, Tswana cowpeas gave the highest value of production per hectare.
Tswana cowpea plants also provided the most (unvalued) morogo.

It is not possible to draw definite conclusions based on a "value
of production" analysis. Despite the higher values of production for
the secondary crops relative to sorghum, the net margin may not have
been higher since labor requirements tend to be higher on secondary
crops. For example, harvesting cowpeas and mung beans often takes
longer than does harvesting sorghum. Groundnuts and jugo beans
traditionally are "billed" to facilitate pegging. Nevertheless, the
results presented in Table 5.9 suggest that sole plantings of secondary
crops would have generated returns to both land and labor which would
have exceeded those from sorghum. This is reflected in the fact that
the gross values of secondary crOps produced per unit area were several
hundred percent higher than for sorghum (when using local market
prices). Labor requirements for secondary crops are not that much
higher than for sorghum.

In the 1985-86 season, sole cropping of secondary crOps was

evaluated in an FM,FI framework, and labor data were collected. An

TABLE 5.9:

166

GROSS VALUES OF PRODUCTION FOR SECONDARY
CROPS AND SORGHUM, 1983-84 AND 1984-85

 

 

 

 

PRICES GROSS VALUE/c
YIELD/a Pl/b P2 P1 P2
1983-84 SEASON
Sorghum/d 74 .39 .26 26.44 16.64
Cowpeas:
Tswana 144 1.56 .35 209.04 46.90
Blackeye 141 1.56 .47 204.36 61.57
ER-7 27 1.56 .39 26.52 6.63
1984-85 SEASON
Sorghum 35 .39 .26 11.23 6.50
Cowpeas (1)/e:
Tswana 52 1.56 .35 65.52 14.70
Blackeye 76 1.56 .47 102.96 31.02
ER-7 58 1.56 .39 74.88 18.72
Sri Lanka 168 1.56 .39 246.48 61.62
Cowpeas (2):
Tswana 175 1.56 .35 257.40 57.75
Blackeye 116 1.56 .47 165.36 49.82
ER-7 112 1.56 .39 159.12 39.78
Sri Lanka 127 1.56 .39 182.52 45.63
Groundnuts 21 2.00 .89 26.00 11.57
Mung Bean 114 1.56 .41 170.04 44.69
Tepary Bean 79 1.56 .41 115.44 30.34

 

P1 - prices from 1984 Trader Survey.
prices for Mahalapye.

. All yields are threshed kilograms per hectare.

P2 ' 1986 BAMB

The value of grain produced minus the value of seed.
Average yields for cooperators.

Cowpeas (1) . CrOpping Comparison Trial.

Planting Methods - Variety Trial.

Cowpeas (2) =

167
enterprise analysis of the sole planted plots was carried out in 1986.
Results from the analysis are presented in Table 5.10.

For farmers who were able to plow using their own animals, the
returns to labor exceeded the LBDR wage for four of the six crOps. For
purposes of comparison, the approximate value of sorghum production per
hectare for the farmers implementing the trial (based on total farm
production divided by the hectares planted) was only P6.05, less on a
per hectare basis than all secondary crOps except maize. Jugo beans was
the most profitable crop, giving a value of production per hectare of
more than P100, and a net margin of over P54, nearly twice as high as
the next most profitable secondary cr0p. The biggest constraint on
expanded jugo bean production is seed availability.

If hired traction had been used, the net margin per hectare would
have been negative for all crops except jugo beans, even when using
local prices (high for output and low for labor). However, the loss
would have been even greater if the entire field would have been planted
to sorghum.

At the end of the season, farmers were asked to compare their
experiences in growing the crops sole versus in mixed planting. Every
farmer said that the problems encountered in growing crOps sole were no
more severe or were less severe than when crOps are mixed. Most farmers
further noted some advantages in growing sole plantings, particularly
with respect to the timing of planting, site selection, and harvesting
labor.

4. HAND PLANTING FOR GAP FILLING
Hand planting was first examined during the 1984-85 season, in

relation to secondary crops establishment. Hand planting did not

168

 

Amman uxo: voaamucofiw

 

~.¢H

man
mafia

ea

¢.om N.¢n
m.m o.mN
s o¢
5N do
¢.m o.m~
o.~ m.m
~.o m.~
o.o~ a.m¢
«.0 m.o~
~.m 0.x
o.~ o.m
m.o~ m.-
om «H
mm nu
ofi ma
on mm
N.HH N.O
«aw own
oooH com
0N nu

w.m

new
con

ON

«.ma

mas
mmm

oH

o.wm cacao wouwo>umm
m.m< macam A~<
"Aan\mmxv essay
no vmumo>umn muoHa .uom

mu mucous muon .uom
s.sa AoooH x us\ssamaav
unmanmfianwumo vcmum
“mmzoowao
m.m muoam woumo>uwm
a.e muofia HH<
n\uAm£\unv oamu wcwnmouny
m.mm ouoad vmumo>uwm
m.h~ muofia Haa
"Awn\u:v mafia mawumo>uwm
~.as suoHa cosmos
o.mm ”Sofia 33¢

"non\unv oawu wcmwooz
m.n «\Aon\unv mafia wagsoam

ma Amaov sumov wowsofim
mm mucuowus
om oeuumo
nu mmoxaoa
"Amuoaa NV vow: domuouua
H.m Amn\wxv mush wawcmmm.
cmm A8 way won<
on“ Amawv vmom
"wHszH
ma myogm mo Mmmxbz

 

zo mHAAmm MNH<Z mum nlmm

cwlmwma

<MZ<A Ham 24mm czaz z<mm OUDH

.mHmVA<z< mmHmmMmHzm HA<HMH UZHmmomo MAOm "o~.m mam<9

169

 

.un\wm.om .oww3 menu on» us vmsau> nonma HH< .m
.cowwmm kuuaoo a“ moon on“: nouomuu vumvawum .o
.muoxuoz osu cu wouflvumvawum Honda wcwaoam .v

.oo.u .muscvasouw moo.

.mmm mom.H .nlmm vow oxcoq wum .cmon mesa mmN.N .mamon ownm "who; vow: mooflum .o
.onwwa vac .ucmon own“ .muacvaaouw
wow un\wx n can soon mass was mmmaaoo now us\wx «a «a mommy magnomunu co comma .n
.vmvuooou mums mmawu wcwsoam coxo 3mm .muouowuu you u£\u:

w.m can mmmxaou you «s\us e.na mums

.c30m mono

mo «moavuwwou .mmuwu wcwsoam .u

 

 

NH.O- No.5H: mm.on ao.o w~.ma oo.o sausage scuba
ms.o «E.ou sm.o ma.o om.ou Na.o mauxaoo use
"noon umm
o~.a~- ow.mnu NN.Hm- m~.-- mo.oou Na.sa passage emuwm
o~.m om.kau wo.k am.o~ mm.w~: ~o.mn mawxaoa use
”wuwuomfi Hmm
"zHomaz mmomo
oo.- w~.a ka.a ¢~.mH ma.~ ma.m~ m\ponaa scaaausaom
oo.om O0.0m oo.om oo.on oo.om oo.om u\uouuaua can“:
ok.sa ok.- ON.HH ok.aa ca.aa e\m»mxaon aao
”MGMSon
o¢.w~ «5.0 ne.oH mo.m oa.o~ m~.ma swam
"Ame\mv mamoo mam<~¢<>
ON.H“ OH.N oo.am om.~o wa.k we.wo~ “o\e=ma=o mo maq<>
zu mHgamm mNHaz mm“ kumm axzaq Ham z<mm ozaz z<mm cosh

 

 

Avoaawuaoov o~.m mqm<s

170
outperform broadcast planting but, in the context of a gap filling
strategy, this is not the relevant analysis. When considering gap
filling, plowing and weeding inputs must be treated as unvalued sunk
costs. The key issue is the return to hand planting and harvesting
labor, after deducting a charge for the seed.

In the 1984-85 cowpeas planting methods-variety trial, yields
averaged from 112 to 175 kgs. per hectare. At local prices, this would
have provided a return to the hand planting and harvesting labor of
around two Pulas per hour (more than three times the urban minimum
wage). Even if one were to include a charge for weeding at a rate of 30
hours per hectare, the returns to gap filling labor would have been more
than double the urban wage rate.

In the 1985-86 season, hand planting for gap filling was examined
for sorghum and millet. Due to late planting and post-planting drought,
yields were low. Despite the low yields, the average returns to labor
across all sites were P2.l8 per hour for millet and P0.27 for sorghum
(taking into account planting and harvesting labor and the value of
seed). Even if an additional 30 hours per hectare of weeding had been
required, the return per labor hour for gap filling, weeding and
harvesting the millet still would have been P0.96.

E. CONCLUSIONS AND RECOMMENDATIONS

This chapter has: (a) given an overview of cropping outcomes during
drought, (b) reviewed efforts to identify improved production practices,
and (c) presented budget analyses of a few promising interventions for
production during drought. This section presents conclusions and

recommendations on promising production practices.

171

1. CROPPING OUTCOMES

The research revealed a "normal" pattern of poor cropping outcomes,
which was aggravated due to drought conditions. In each season, a large
prOportion of the area planted was neither weeded nor harvested.
1 Despite the drought, some real income was generated by draft
controlling households. For example, the value of craps production
above variable costs in the 1984-85 season was equivalent to the cost of

450 kgs. of store bought maize. On a per hour basis, the gross margin

 

was only P0.26. This is less than half the urban minimum wage rate and
only 70 percent of the LBDR village wage. The returns were insufficient
to cover traction hire fees. Nevertheless, for traction owners, the
returns were equivalent to those women earn in beer brewing and selling
[ATIP, 1986].

Two elements of the traditional system were responsible for what
little success was achieved. First, the non-grain component of
production was substantial. Second, farmers wisely cut back on their
labor inputs as expected crop outcomes were revised downward in the face
of the unfolding drought year.

Significant and consistent differences were identified in crOpping
outcomes by household type, both with reference to "normal" outcomes and
the outcomes observed during the drought seasons. Male headed, cattle
rich and draft controlling households plowed more days, achieved higher
production levels and (at least during drought) had higher yields.

The findings on crOpping outcomes reinforce several conclusions
from Chapter III and IV, including: (a) the value of secondary and minor
crops, (b) the need for interventions for all households not just poor

households, (c) the significance of the trend toward tractor hire and

172
ownership, and (d) the need to keep capital investments to a minimum
2. THE RECORD OF TECHNICAL EXPERIMENTATION
In 1982, the bottom line on technical experimentation in Botswana
was that few answers had been developed relative to the resources

' on-farm research in

invested. Building on this "state of knowledge,‘

the Central Region initially focused on a comparison of planting

methods. Before long, the scOpe of investigations expanded to include
several types of tillage-planting systems, secondary cr0p and variety
selection, sole crOpping, and post-establishment stand management
practices.

After four seasons of research, the results of the various
investigations could be divided into four categories.

a. No actual or potential benefits were identified for: plow-planters
(particularly when mounted on single furrow plows), harrowing
(particularly as a weed control tillage), ridge plows, ridge
planters, strip tillage systems, manuring or draft animal feeding.

b. Not enough trials were successfully conducted to reliably assess the
potential benefits of (or work out problems with): seed treatments,
hand furrow planting, the advantages of replanting versus relying on
an existing poor stand, undersowing, the relative advantages of
different tillage-weed control systems for row planting, early
weeding versus standard weeding, or compound garden plots.

c. There were promising indications, but no firm conclusions, for:
thinning, sub-soiling, multiple purpose uses of cultivators
(pre-plant weed control or to incorporate broadcasted seed),
phosphate applications on certain soil types, and site selection for

double plowing.

 

173
d. There was some confidence that farmers could obtain at least small
improvements by double plowing and/or early plowing followed by row
planting, production of sole planted secondary crops, and gap
filling.
3. PROFITABLE PRODUCTION PRACTICES FOR DROUGHT CONDITIONS
The partial budget analyses showed that traction owners could
benefit from a shift to double plowing, and could even benefit slightly
if half as much area had to be plowed in total in order to double plow.
Similar results were obtained for early plowing plus row planting. The

benefits from either system, however, were not great and may not

 

stimulate much farmer interest. Both systems were viable alternatives
for farmers facing either land or labor constraints.

The benefits of a second plowing were not great enough to pay for a
second hiring of a tractor, so double plowing cannot be recommended as a
strategy for draft-dependent households. This result stems from the
general unprofitability of arable farming during drought.

When comparing row planting and double plowing, it is hard to
choose based on the budget analyses. Row planting, of course, requires
an additional implement but most farmers can obtain a planter at 15
percent of its cost. Several less tangible factors support a relatively
greater emphasis on row planting as a long run strategy, including: (a)
eventual elimination of the weeding labor constraint through inter-row
cultivation, (b) better control over seed depth placement and plant
spacing, (c) greater stability over a range of soil environments [ATIP,
1986], (d) smaller draft requirements if plowing is hired and planting
is done with two donkeys, and (e) flexibility in replanting. Therefore,

the challenge of solving the problems constraining the Spread of row

174

planting should be the top priority of both research and extension in
the Central Region (particularly with reference to the draft-controlling
research domain). Such a priority would be consistent with farmers'
interest in row planting.

The budget analyses of secondary crOps provided evidence that the
returns were as high or higher than for sorghum. On balance, it would
appear that small plots of sole planted secondary crops should be
recommended. However, maize should be actively discouraged. Jugo beans
and dual purpose cowpeas (intermediate plants such as Sri Lanka and
Blackeye) would seem to be the best secondary crOps to promote. There
would not appear to be any need to recommend special planting methods
for the secondary crops.

Based on two drought seasons in which poor yields were obtained,
hand planting for gap filling provided returns to the additional labor
invested that exceeded urban minimum wages and were four times the
average return to crop production labor. Hand planting for gap filling
would appear to be a viable option to promote, particularly to farmers

lacking control of traction resources.

VI. AGRICULTURAL SUPPORT SYSTEMS AFFECTING ARABLE FARMING

This chapter presents findings from research on: agricultural
extension, the trading network, village groups, and participation in
ALDEP and Drought Relief. The objectives are to: (a) analyze the
performance of selected agricultural support systems, and (b) identify
policy and institutional options to improve their performance.

A. AGRICULTURAL EXTENSION

Agricultural Demonstrators (ADs) are the link between farmers and
the Ministry of Agriculture. All messages about modified production
practices and assistance programs are channeled through ADS. The
Department of Agricultural Field Services uses the ADs' annual reports
to determine extension service priorities. Therefore, any institutional
changes which improve the effectiveness of ADs in communicating
messages, demonstrating innovations, and providing feedback to the
Ministry should eventually contribute to arable farming development.

To identify Opportunities for improvement, the activities and
problems of ADs in the Central Region were investigated in a 1983
regional survey. The survey had two objectives. The first was to
determine the extent to which farming systems differ across the Central
Region. The second was to find out what ADs do, what they believe to be
important, and what problems constrain their effectiveness.

This section begins with a brief historical perspective on

extension in Botswana. The second part summarizes the results of the

175

176
1983 regional survey.
1. HISTORICAL PERSPECTIVE OF EXTENSION IN BOTSWANA

The primary mandate of the extension service through the 19505 was
to provide advice on methods. There was no mandate to offer services to
farmers. The main activity of extension workers was to carry out
demonstrations on farmers' fields. The relative emphasis on providing
advice versus services has completely changed.

In 1962, the Pupil Farmer Scheme was established as the primary
mode of agricultural extension. In this approach, each AD concentrated
on improving the practices of about 25 farmers. In order to qualify for
the scheme, a farmer had to own a plow and draft oxen, and needed
destumped land. As their farming methods improved, pupil farmers were
promoted through "progressive" and "improved" stages, until they became
a "master farmer." In the end, master farmers autumn plowed, row
planted in early spring using appropriate spacing, used quick-maturing
varieties, did regular weeding, applied fertilizer as needed, followed a
defined crop rotation, and used contour bunds (if needed) [ESM, 1965].

During the early 19708, three major consultancies were commissioned
to examine the Pupil Farmer Scheme and the organization of the extension
service [Willett, 1981]. Each of the consultancies prOposed abandonment
of the Pupil Farmer Scheme. The consultants argued that, even though
the pupil farmers did become more productive, the recommended
technologies rarely spread to other farmers. Most other farmers lacked
the necessary resources [Lever, 1970; Willett, 1981] and ADs often were
too busy with their pupils to spend time with other farmers. Lever

[1970] reported that ADs spent approximately 90 percent of their time

 

177
with the pupil farmers.

The consultants identified several major problems in the
administration and management of the extension service. For example,
agricultural extension activities were being carried out by several
sections of the Ministry of Agriculture, leading to overlap and
confusion. In addition, field staff were inadequately supervised,
poorly housed, lacked necessary extension equipment, and were short of
suitable transport.

Both the extension approach and the organization of extension
activities were changed in the mid-19703 to their current status. In
1973, a principle of pursuing both group and individual extension was
adopted. In 1976, the Pupil Farmer Scheme was ended. ADs now are
required to work with groups of farmers whenever possible. Different
approaches are used for communicating extension messages, including
demonstrations, speaking at village meetings and participating in
village groups.

Following the consultants' advice, the Ministry of Agriculture was
reorganized during 1974 and 1975. A regional administrative structure
was established for the extension service. To improve field staff
management, several institutional changes were introduced, including:
(a) farmer record cards, (b) monthly management meetings involving all
the ADs for each extension district, (c) monthly and annual plans for
each AD, and (d) a daily activities planning and record system.

One of the key issues facing the Ministry in the 19808 is whether
the reorganization of the extension service and the adoption of a group

extension approach have accomplished the envisioned aims.

178
2. EXTENSION IN THE CENTRAL REGION

Fifty-two out of 54 ADs then stationed in the Central Region
participated in the 1983 Agricultural Demonstrator Survey. Of the 52
respondents, approximately ten came from each of the five extension
districts. (A sixth district, Letlhakane, was later added to the
region). All but four ADs were male, with no district having more than
one female AD. On average, the ADs had completed their formal education
ten years prior to the survey but each had attended an in-service
training course within the previous two years. In terms of formal
education, only a third had attended the equivalent of high school,
approximately 40 percent stOpped with the equivalent of a junior high
school degree, and 30 had less than a junior high school degree. All
ADs had technical agricultural certificates, mostly from the BAC.

a. Extension Area Characteristics

An overview of extension area characteristics is presented in Table
6.1. On average, extension areas encompassed two or more villages and
more than 400 farming families. Lands areas were on average between six
and twenty-four kilometers from the villages where ADs were stationed.
Sorghum, maize and cowpeas were the most important crOps in every
district.

Donkey traction was used by 25 percent of households, higher than
the national average. Donkey traction was particularly important in
Mahalapye East and Bobonong. Tractors or tractors in conjunction with
animal traction were used by 36 percent of the households.

In all areas, there was little use of progressive practices. For
example, an average of less than 30 households per area did more than

one weeding, less than 20 households used row planting, and even fewer

179

.ma< >3 mawmwmuamm m>wuommnsm mo wcwunwwms mmum>cw Eoum vm>wuov mwawxcmm .n
.mwwawawm

wcaauwm mo nonasc onu oumeumo no: vaaoo Q< umnuo may .mn< OH «0 a do vommm .m

.mm>u=m heumuumcoaon Hmuauasowuw< mwma "mousom

 

mm om mm om mm «m onHo<mH MMH: mm .Hom
mu ma wH on «N qm mamawq< w mHOuowHH
ma 0H mm NH 5 ma mucuowufi
am NH as no mm us mauuwu
2 8 2 2 cm 3 3938
"Am: NV ZOHHU<MH wm<szm
mm 3 8 on 3 em 9:82. 02
mm 3 mm om Z .3 385m swam
om wq ma qa «m NN mcmocmm muw3
ANV":uM3 mvaonomaom
ma ma NH NH ca NH A:mouom:v swam
"mQAMHm
umaafiz umHHMZ umaafiz cmom owsh comm owsw comm owsh nunsom
mommaou mmoaaoo mamasoo mmomsoo mommsoo mamasoo wHLSH
onwmz oufimz mafia: swam: swam: mafia: vcooom
asnwuom annwuom Escwuom asnwuom Banwuom aanwuom n\umuwm
"mmomo Hz<HmomzH Hmoz
«N wN NM mm mm 0H mvcmq ummsuumm
o m n m o o Amaxv mecca umoumoz
and mom mmm «\oqo mmm mmm mowawamm wcfiaumm
m.~ m.~ m.m m o.N H.H mmmeHw>

nm¢mm¢ onmzmexm mo MNHm

 

AA< ozozomom mzommm mwm<q<m Hm<m Hmmz
m>m<4<m<z mwm<q<m<2

HOHMHmHQ Wm moHHmHMMHO<M<EU <mm< ZOHmzmHNm u~.© mqm<fi

180
households applied fertilizer, did winter or spring plowing, use a metal
harrow, used fallow in rotation, or had cut and stored fodder.

In total, more 50 characteristics were recorded for each extension
area. When extension area characteristics were analyzed by district,
few significant differences were identified. The most important
difference was the importance donkey traction in the Bobonong and
Mahalapye East Districts, versus the importance of cattle and tractor
traction in the other three districts. Otherwise, the variability among
areas within districts was greater than differences across districts.
Also, there was a general homogeneity of farming problems and practices
in the Central Region.

The survey did reveal important differences between extension areas
when compared on an individual basis. For example, row planting was
dominant in one extension areas, but in no other extension area.
Similarly, there were differences in the amount of wire fencing, the
distance to the lands, and the percent of households owning cattle.

b. Extension Activities

ADs provided advice and services to farmers, acting as field agents
for several government programs. Advice was provided through direct
contacts and at village meetings. Village meetings were the primary
forum for contacting farmers. On average, village meetings were
addressed ten times during the crOpping season, more frequently in
Mahalapye East and Serowe Districts than elsewhere. The emphasis of AD
messages at village meetings was on government programs, particularly
ALDEP, rather than promotion of improved technologies. Group formation
and drift fencing projects were priority topics in every district.

According to the ADs own estimates, only one-fifth of the farming

181
families were contacted even once a season. Only one in fifteen farmers
was contacted once per week, and these farmers often were the same ones
week after week. When ADs were asked what distinguished the farmers
contacted regularly, 85 percent of the responses related to attitudes,
productivity, and resources that would be associated with the best, most
progressive farmers. This raises the issue as to whether abandonment of
the Pupil Farmer Scheme increased the amount of extension advice being
received by poorer, less progressive farmers.

In providing advice to farmers, ADs stressed the same recommended
practices throughout the Central Region, almost in the exact same order
of importance. Row planting was the main recommendation in each
district, followed by early plowing, crop rotation and early weeding.
There clearly was more unanimity among ADs on what farmers should be
doing than there is among DAR personnel.

In addition to providing advice, ADs were responsible for
demonstration plots and for supplementing supplies of equipment, seed,
and fertilizer. During the 1982-83 season, few demonstrations were
attempted and very few were successfully implemented. The number of
demonstrations ranged from just over one per area in Bobonong District
to three per area in Serowe District, averaging two per area across the
region. Eighty-seven percent of the demonstrations dealt with row
planting and/or fertilization (including manuring). The main types of
equipment provided were planters, cultivators, harrows, burdizzos (an
implement for castrating) and dehorning irons. A summary of equipment
supplies is given in Table 6.2. As can be seen, few farmers made use of
AD planters, cultivators or harrows.

Just over sixty percent of the ADs reported sales of sorghum and

182

TABLE 6.2: AD SUPPLY AND USE OF EQUIPMENT, 1983

 

 

NUMBER NUMBER OF FARMERS USING
AVAILABLE TOTAL PER PIECE

PLANTERS 1.7 6.0 3.5
(1.5) a (9.1)

CULTIVATORS 0.9 3.1 3.4
(1.4) (8.2)

HARROWS 0.8 5.0 6.2
(0.9) (7.9)

BURDIZZOS 0.4 67.3 156.5
(0.6) (196.1)

DEHORNING IRONS 0.8 55.0 70.5
(1.7) (192.8)

 

Source: 1983 Agricultural Demonstrator Survey.
a. Standard deviations in parentheses.

maize seed. Nearly all the ADs reported they could have sold more seed
had it been available. Forty-two percent provided fertilizer but the
number of farmers receiving fertilizer was quite small (3.6 per area
where any fertilizer was provided). The remaining major activity of ADs
was helping with ALDEP applications.

An analysis of correlations among farmer contact activities showed
there were positive and significant (above .9 confidence level)
correlations of .35 to .40 between times addressing a village meeting,
families met individually during a season, and the number of
demonstrations. There also was a positive (.34) and significant
correlation between the number of record cards kept and the number of
families met during a season. Each of these activities were negatively
related to the number of visits to the lands area per week (but not
significantly so). It would appear that certain ADs are more active and
this is reflected in all their farmer contact activities, except for

visits to the lands areas. Visits to the lands areas may be a

183
substitute for other types of farmer contact activities.
c. Institutional Support

Monthly management meetings are the main format for giving
instructions to ADs and helping them with problems they encounter.
Eighty-eight percent of ADs said they had reported problems to a monthly
management meeting during the six months preceding the survey.
Eighty-two percent said they felt monthly management meetings were
useful. The main value of the meetings was seen as the Opportunity to
interact and get ideas. Among the eighteen percent who were
dissatisfied, the primary concern was that complaints and problems were
never solved.

In addition to the monthly management meetings, sixty percent of
the ADs had requested special assistance at some time in the two years
preceding the survey. The prOportion of ADs requesting assistance
ranged from a low of 30 percent in Mahalapye East to 75 percent in
Bobonong. On average, just over four requests were made during that two
year period by those ADs who made any requests. (The median number of
requests was lower since a few ADs made more than 20 requests.)

The most frequently contacted person was the District Agricultural
Officer. The DAO and the Regional Crap Production Officer were the main
individuals providing information directly to ADs. Less than half the
A08 requesting help felt they received timely and effective help.
Twenty-five percent reported they received no help and 25 percent said
they received help, but it arrived too late to be of use.

There was a clear consensus among ADs that closer ties were needed
between researchers, extension agents and farmers. For example, ninety

percent said there should be more village research and eighty-one

184
percent agreed when asked if researchers and ADs should try to learn
from farmers. Only 31 percent felt they knew what was going on in
Sebele research.
d. Perceptions and Priorities

ADs were asked about the value of different crOps and practices in
order to assess perceptions affecting AD behavior. A summary of
responses is given in Table 6.3. Nearly all ADs said that sorghum was
the most valuable crop because it is harvested in large quantities, and
has multiple uses and a relatively stable yield. Multiple uses also was
mentioned as an attractive feature of cowpeas (harvesting leaves in bad
years) and melons. High demand and/or prices was mentioned for five of
six of the most valuable crOps. ADs said early maturity (availability
of product) was a valuable feature of cowpeas, maize and melons.

Table 6.4 summarizes ADs' views on changes farmers could make to
increase productivity, why each change would increase productivity, and
why farmers do not make the change. Row planting was by far the highest
ranked change, followed by fertilizer use, fencing, winter or spring
plowing and early planting. In general, the reasons ADs cited for
productivity increases were obvious and somewhat superficial. It was
not clear from the survey that the ADs had a good understanding of the
advantages and problems of different production practices. Lack of a
proven benefit, or other technical questions, were only cited as major
problems for fertilizer use, better weed control and early planting.
Otherwise, the ADs tended to cite resource constraints and lack of

knowledge as the main reasons farmers do not make changes.

TABLE 6.3:

185

AD ASSESSMENTS OF THE MOST
VALUABLE CROPS; 1983

 

RANK CROP

MOST FREQUENTLY CITED REASONS

 

l SORGHUM

2 GOWPEAS

3 MAIZE

4 MELONS

5 JUGO BEANS
6 SUNFLOWER

Harvested in Large Quantities
High Demand

Multiple Uses

Stable Yield

High Demand

High Prices

Matures Early

Can Harvest Leaves in Bad Years

Good Prices
High Demand Relative to Supply
Can Sell Green or Dry

Easy Management
Ripens Early
Multiple Uses

Low Labor Relative to Price
High Demand

High Price
High Demand

 

Source: 1983 AD Survey.

 

 

 

 

TABLE 6.4:

186

CHANGES FARMERS COULD MAKE TO

INCREASE PRODUCTIVITY, 1983

 

RANK/a CHANGE

WHY CHANGE WOULD
INCREASE PRODUCTIVITY/b

WHY FARMERS
DO NOT USE/b

 

 

l - Row Plant PrOper Plant Spacing Lack Implements
Helps Weeding TOO Much Work
Better Seed Depth
Less Seed Used
2 - Fertilizer Increase Soil Fertility Lack Funds
Improve Soil Structure Lack Knowledge
Conserve Soil Moisture Fear More Weeds
3 - Fencing Reduce Livestock Damage Lack Funds
NO COOperation
4 - Winter or Improve Soil Moisture Lack Draft
Spring Plow Not at Lands
Lack Labor
5 - Early Plant Leaves Long Time to Mature Erratic Rains
Use First Rains Lack Draft
Reduces Pest Attacks
6 - Certified Quick Maturing Lack Funds
Seed Better Germination
7 - Better Weed Reduce Weed Competition Lack Labor
Control Moisture Conservation NO Benefit
8 - Crop Increase Soil Fertility Lack Knowledge
Rotation Reduce Diseases and Pests Lack Seeds
Promotes Different CrOps
9 - Dams Water for Draft Animals Lack Funds
10 - Destump Less Damage to Plows Lack Funds
Use Planters & Cultivators Lack Labor
Source: 1983 Agricultural Demonstrator Survey.

a.

b. Most frequently received reasons are cited.

Ranking based on the number Of times a change was cited.

Each AD was allowed tO cite three promising changes.

Among cited
reasons, no strict ranking is implied. Two "attitude"
reasons were cited frequently-farmers lazy and no
interest--but have not been included since they do not relate
to particular changes.

187

B. TRADING NETWORK

In Botswana, local trade historically has been dominated by
informal exchanges between households. Informal trades are made for
labor, traction, gathered items and household products on both a cash
and barter basis. In many cases, exchanges are made without specified
terms [Duggan, 1979]. Although, as Duggan points out, there is an
internalized sense of relative value and it is expected that trades will
balance out over time.

During the drought Of the 19803, the importance of informal trade
decreased. Households had few products to exchange and, as noted in
Chapter IV, cooperative plowing gave way to formal hire draft
arrangements. Formal traders certainly were the main source of food and
production inputs between 1982 and 1985, if not before.

A key marketing issue affecting crop production is the impact of
the trading network on production incentives. One particular issue is
the prices farmers receive for their products relative to the prices Of
commodities available through the traders. Inexpensive food grain
imports can discourage attempts to increase production. A second issue
is the availability of food. The incentive to invest in crOp production
is greater if commodities are not regularly available. A third issue is
the availability Of inputs. It does little good to recommend use of new
implements or chemicals if they are not available.

In order to assess the impact Of the trading network on crOp
production incentives in the Central Region, a regional survey was
carried out in 1985. There were three main research Objectives: (a) to
see if there were differences in trading activities by village size, (b)

to determine the availability and prices of agricultural commodities and

188

production inputs, (c) to identify constraints on the performance of the
trading network, with a focus on supply sources and business problems.
1. PROFILE OF TRADING ESTABLISHMENTS

There was an average Of 2.7 traders per village, as seen in Table
6.5. (All results exclude the major villages unless otherwise stated.)
The number of traders per village was about the same in the different
parts of the region. As would be expected, there were more traders in
the large villages than in the small villages. However, the average

number of households per trader was actually the lowest in the small

 

 

 

 

villages.
TABLE 6.5: NUMBER OF TRADERS IN THE CENTRAL REGION
BY LOCATION AND SIZE OF VILLAGE, 1985
NO. OF No. OF TRADERS/ HHS/ HHS/
VILLAGES TRADERS VILLAGE VILLAGE TRADER
LOCATION:
Mahalapye West 7 l9 2 7 244 90
Mahalapye East 14 43 3.1 202 65
Palapye South/a 11 28 2.5 240 96
Palapye North/a 8 21 2 6 228 88
Serowe 6 14 2.3 207 90
POPULATION:
Small Villages 21 41 2.0 94 48
Medium Villages 15 41 2.7 239 89
Large Villages 10 43 4.3 468 109
SUB-TOTAL 46 125 2.7 223 83
MAJOR VILLAGES 3 48 --- 3786 --
ALL 49 173 --- 441 --

 

Source: 1985 Trader Baseline Survey.

a. Palapye South includes villages south of the Tswapong Hills
while Palapye North includes villages north of the Tswapong
Hills and on the road to Bobonong.

189

Ninety percent Of the trading establishments were general traders.
There were only 15 cooperatives in the 46 villages. Most of the
COOperatives were in Serowe and Mahalapye West Districts. In the small
villages, there were only three COOperatives.

Most trading establishments were owner managed (63%). The
establishments had been in operation for an average of eight years.
Around ten percent of the establishments had been in business for more
than twenty years. The establishments employed an average of 2.6
full-time employees but few part-time employees. Traders in the small
villages only employed an average of two employees, compared to about
three in the large villages. COOperatives employed two to three times
as many employees as did general traders. Traders in the major villages
had an average Of 6.2 employees.

As would be expected, most customers were from the village where
the establishment was located. The prOportion of traders who considered
lands dwellers to be a large share of their customers was greater in the
small villages than in large villages (56% of establishments versus 23%
in the large villages).

A summary of trading activities in presented in Table 6.6.
Essentially all establishments sold agricultural commodities (including
canned and processed foods). Many establishments sold some types of
implements, but less than a quarter sold fencing materials, livestock
requisites or even seed. There were few establishments which had bought
either crOps or livestock from farmers. The findings on trader
activities can be contrasted to those from a trader survey carried out
in the Southern Region in 1970 (a year with good rainfall) [DPS, 1971].

While few traders in that study said they sold fertilizer, pesticides or

190

.mucoanmwanmumo mud “>O>uam mafiaommm Hummus mama "mouaom

 

 

 

 

 

 

N m mN NN on m oN m as 444
o o NN ad Hm oN NN s om mmuaaNN> None:
N 0 EN NN mm H EN 0 ma A<Noeumom
m N NN ON sq N NN n no mmmeNN> mwuma
m N NN ON as N NN n ma mummN~N> asNumz
o m NH NH NN o NN N Na mowaNNN> NNmam
"zoNN<N=Noa
o N NN 0 Ne N mN o ooH «soumm
o m NN mN mm o SN O Na nuuoz «NamNmN
s o NN NN HN c EA as ea nuaom mNamNmN
m N NN NN Ne o MN m mo swam «NamNmemz
o 0H Nm EN mm o 0N NH om saws mNamNmemz
"cheauoa
Amocoanmwanwumm mo unmouomv
qum amaze .Naomm .mmyaz mazmz mq<oN ammm mmamomsm Namm
Imam .mN>NN muzmm -NNNzN -szo mmHeNoozzoo
Nooemm>NN mmaam NeazN Nassaaponua

 

mwma

.zoNomm N<Nezmo .mmHaN>HHo< uanama

no.0 m4m<H

191
livestock requisites, most did purchase sorghum, maize and cowpeas from
local farmers. This suggests that the lack of purchases from farmers in
the Central Region may have been due to the drought.

Most traders (79%) made sales on credit but who got credit was
generally limited to those that were well-known or who had permanent
wage employment. No trading establishment gave credit to all its
customers.

2. AVAILABILITY AND PRICES OF FOOD COMMODITIES,
IMPLEMENTS AND FENCING MATERIALS

Because of the effects of food access on production incentives,
additional data were collected on the availability and prices of the
major food commodities in relation to the size of village. Findings on
the percent of traders selling the different food grain items, the
percent with the items in-stock on the day of the survey, and average
prices are summarized in Table 6.7.

The percent of traders selling most items was somewhat less in the
small villages compared to the medium and large villages. The
differences by village size were greater when comparing how many traders
actually had items in stock. Nevertheless, even in the small villages,
most items not only were sold, but were in stock. The items which were
most consistently in stock were those Obtained from wholesalers.
Cowpeas, sorghum grain, sorghum meal and bread (Obtained from millers or
local growers and bakers) were available on a more ad hoc basis.

Turning to prices, imported maize meal was the least expensive,
followed by sorghum meal, wheat flour and rice. Cowpeas and other beans
were two to three times as expensive. There was a relatively narrow

range of prices for any given agricultural commodity. The coefficients

192

TABLE 6.7: AVAILABILITY AND PRICES OF
MAJOR FOOD GRAINS, 1985

 

BY SIZE OF VILLAGE MAJOR
SMALL MEDIUM LARGE VILLAGES ALL

 

PERCENT SELLING:

Maize Meal 97 89 93 96 93
Maize Grain 79 98 85 96 90
Sorghum Meal 44 71 68 73 65
Sorghum Grain 44 58 35 46 46
Cowpeas 12 31 38 48 33
Bread Flour 85 87 85 93 88
Bread 44 38 50 75 52
Rice 82 84 95 89 88

 

PCT. WITH IN-STOCK:

Maize Meal 74 87 90 91 86
Maize Grain 56 84 68 82 74
Sorghum Meal 32 51 43 66 49
Sorghum Grain 24 40 15 30 28
Cowpeas 6 18 15 39 20
Bread Flour 65 71 80 93 78
Bread 21 24 33 75 39
Rice 56 73 88 73 73

 

AVERAGE PRICE (PULA):

Maize Meal (kg) .56 .51 .53 .49 .52
Maize Grain (kg) .65 .59 .61 .57 .60
Sorghum Meal (kg) .71 .67 .70 .65 .68
Sorghum Grain (kg) .43 .38 .38 .38 .39
Cowpeas (kg) 1.25 1.49 1.55 1.58 1.53
Bread Flour (kg) .87 .79 .82 .75 .80
Bread (loaf) .77 .76 .80 .74 .76
Rice (500 gms) .86 .81 .86 .78 .83

 

Source: 1985 Trader Baseline Survey.

193
Of variations of most prices were under 15 percent. The narrow range Of
prices can be attributed to the trading margins set by the Price Control
Unit of the Ministry of Commerce & Industry. Since the Control of Goods
Act was passed in 1973, maximum trading margins have been set for all
goods sold by all traders - both wholesalers and retailers.

Prices tended to be higher in the small villages, and lowest in the
major villages. The largest and most significant differences were for
the items Obtained through wholesalers in the major villages.

The availability and prices of implements and fencing materials
were also examined, since these are the main purchased inputs in the
region. Results are presented in Table 6.8. Only a small share of the
traders who reported they sold implements, sold anything but hand tools
and plow shares. Less than ten percent of the traders sold plows,
planters, or carts. Even a smaller prOportion sold harrows or
cultivators. There were large ranges of prices Observed, mostly due to
differences in the features of the items being sold.

3. SOURCES OF ITEMS SOLD

Essentially all traders in the Central Region relied on wholesale
depots in Mahalapye, Palapye or Serowe for the items sold. In addition,
nearly half obtained some supplies from businesses in towns in Botswana,
a quarter from businesses in other countries, and a quarter from
parastatals (primarily BAMB). Neither farmers nor businesses in the
same village were sources Of supply for more than three percent of the
traders.

Traders in the major villages relied primarily on wholesalers in
the same village. In addition, a significantly greater proportion Of

traders in the major villages (.95 confidence level) obtained supplies

 

194

TABLE 6.8: SALES OF IMPLEMENTS, HAND TOOLS

AND FENCING MATERIALS, 1985

 

 

 

 

 

 

PCT. PCT. WITH -------- PRICES/a -------
SELL IN-STOCK LOW HIGH MEAN S.D.
IMPLEMENTS:
IS Single Plow 10 7 67 270 141 50
IV Single Plow 12 9 40 195 109 32
Double Plow 4 4 334 482 380 46
Row Planter 4 4 203 303 256 32
Plow Shares (8-12") 21 14 3 l3 6 3
Donkey Cart 4 l 315 610 456 108
HAND TOOLS:
Wheel Barrow 14 9 35 72 48 11
Spade 23 15 7 15 10 2
Hand Hoe 23 17 1 6 4 1
Pick 18 15 6 16 11 2
Axes - Small 27 21 4 17 10 3
Axes - Large b b 18 29 21 3
FENCING MATERIALS:
Galvanized Wire 15 13 47 96 65 11
Anchor Wire 8 7 35 90 48 16
Binding Wire 16 12 43 75 63 11
Barbed Wire 10 10 49 80 64 9
Gates 9 6 22 80 51 21
Posts 9 8 4 15 11 2

 

a. Rounded to nearest Pula.

b. The question did not distinguish by axe size.

195
in Botswana towns and from other countries than did traders from the
other villages.

A majority of traders paid for all their supplies with cash (56%).
The remaining traders were able to Obtain some of their supplies on
credit. A significantly greater prOportion of traders in the major
villages were able to obtain supplies on credit compared to traders in
the other 46 villages (67% compared to 44%; significant at .95
confidence level).

4. BUSINESS PROBLEMS

Business problems were investigated by asking traders which out of
a set of ten potential problems were problems faced by that
establishment. Findings are presented in Table 6.9. The most frequent
business problems were lack of supplies, too much competition,
unaccounted losses, and lack of cash for supplies. Unaccounted losses
and competition were more frequently cited by traders in the major
villages while lack Of cash and supplies were cited more frequently by
rural traders. When the traders were asked to rank their most important
problem, too much competition was ranked overall as the greatest
problem.

For comparison, traders in the Southern Region reported in 1970
that their main problems were: (a) lack of transport and the high cost
of transport, (b) delays in securing supplies caused by import and
export permits, and (c) unpredictable levels of local crop production

[DPS, 1971].

196

TABLE 6.9: BUSINESS PROBLEMS BY SIZE OF VILLAGE, 1985

 

SMALL MEDIUM LARGE MAJOR ALL
(Percent of Establishments)

SHORTAGES OF:

 

 

 

Supplies 66 66 67 58 64
Cash 68 46 51 4O 51
Buildings 34 49 3O 35 37
Skilled Labor 42 24 30 27 31
Water 42 34 30 21 31
MANAGEMENT:
TOO Much Competition 46 54 58 65 56
Unaccounted Losses 49 44 6O 60 54
Poor Transport 44 32 28 29 33
Prices Too Low 37 20 21 27 25
Poor Record Keeping 27 20 23 15 21

 

Source: 1985 Trader Baseline Survey.

C. VILLAGE GROUPS

Most ministries, including the Ministry of Agriculture, encourage
their field agents to form and work with village groups [Willett, 1981].
Village "self-help” activities, including drift fencing and dam
construction, nearly always are channeled through groups. The 1983 AD
Survey showed that group formation was the second most common tOpic
raised.by ADS at village meetings. Consequently, improved guidelines on
group formation and management should increase the capacity of the
Ministry to promote modified production practices and programs.

This section summarizes recent national research findings on
village groups and presents selected findings from Shoshong and Makwate
on: (a) the activities and problems of groups, and (b) patterns of
participation in groups. The last part of the section reviews

experiences with a group formation "institutional experiment" initiated

197
in the 1985-86 season.
1. RECENT RESEARCH FINDINGS

Research on village groups and voluntary organizations has been
sponsored by both the Ministry Of Local Government and Lands and the
Rural Sociology Unit of the Ministry of Agriculture. The objective in
most investigations has been to identify group activities and evaluate
their contributions to deve10pment. Between 1980 and 1985, local
institutions were examined by Wynne [1981], Brown [1982], Manzardo
[1982] and Zufferey [1983a, 1983b].

Wynne carried out a study of two villages in each of the Gaborone,
Central and Southern Regions. She found confusion over the roles of the
various village groups. Few villagers felt the voluntary organizations
were making an important contribution.

Brown studied local institutions in the Gaborone Region. He found
that the Village Development Committees rarely met and were involved in
no major activities. Few villages had Farmers Committees. Brown
concluded that most village organizations made few contributions to
rural deve10pment. In what they did accomplish, they were oriented
toward infrastructure.

Manzardo's research had a national focus. He reported that
organizations which were the product of efforts by ADS usually had small
memberships and rarely met. Groups initiated by villagers, on the other
hand, generally were committed to some Special purpose and were
supported by several active farmers.

Zufferey carried out a series of studies of local institutions in
the Central Region. He identified a diverse set of problems affecting

group organization and operations, including: (a) many members did not

198
understand the purpose of the groups, (b) lack Of coordination among
groups, and (c) lack of group organization skills.

In summary, local institutions research in Botswana has
consistently led to two observations: (a) most group organizations are
making few contributions to village deve10pment, mainly due to
attendance problems and a lack of focus and (b) activities which
represent organized responses to locally felt needs consistently are
more successful than those formed by extension workers or those which
are part Of national organization.

2. GROUPS IN SHOSHONG AND MAKWATE

An overview of the major groups in Shoshong and Makwate was
initially compiled during the 1984 Village Institutions and Services
Survey. Follow-up information on group activities and problems was
collected from most of the groups during February and March, 1985.

The studies confirmed that groups in Shoshong and Makwate had the
same problems identified elsewhere in Botswana. The groups were an
important element of village life but were not Operating very
effectively. Most Of the general purpose groups were oriented toward
infrastructure projects. There was a lack of coordination both within
and between groups, leading to an overlap in projects and confusion Over
the roles of different groups. Most groups suffered from poor
attendance and a lack of commitment.

The VDCS are supposed to be the focal point of village deve10pment
efforts, but were not playing this role in either Shoshong or Makwate.
The VDCS emphasized projects with few beneficiaries, rather than their
role of setting and coordinating a deve10pment agenda. The Farmers

Committees were completely inactive in both villages.

199

As found in other studies, the special purpose agricultural groups
were among the most active. However, the groups requiring large fees,
such as the Shoshong spray race group, tended to serve the interests of
a few wealthy villagers. The existence of projects was an important
focal point of the more successful groups. Without Specific projects,
most groups became inactive. Moreover, the groups with better
attendance and more committed memberships tended to be those which had
personal incentives.

3. HOUSEHOLD AND INDIVIDUAL PARTICIPATION

Household and individual participation in village groups was
examined in the 1985 DUMI Study. With reference to permanent groups,
more households had participated in 4-B projects than those of any other
group. Members from female headed households were more active in
village groups, as were those from richer households. In general, more
households participated in group activities sponsored by the village
extension agents and leaders than they did the activities of permanent
groups.

On an individual basis, only seven percent of the enumerated
population had participated in village groups. The proportion of
individuals participating in groups did not change greatly, regardless
of whether a household was in Shoshong or Makwate or was male versus
female headed. A slightly greater share of household members in
households with more than 15 cattle participated in groups than did
those from poorer households.

There were significant differences in individual participation
based on gender, role in a household, and age. Ten percent of the

females enumerated participated in village groups, compared to only

200
three percent of the males. Most of the participants in groups were
either the most senior male or female in a household. Nearly a third of
the most senior females and a quarter of the most senior males
participated in groups. Group participation was primarily limited to
those above age 40.
4. EXPERIENCES WITH FARMER GROUPS

After the above institutional studies had been completed, farmer
groups were formed in Shoshong, Makwate and Makoro villages. Formation
of the groups represented an institutional experiment. Prior on-farm
experiments carried out through individual farmers had failed to
Stimulate community interest. In addition, the focus of most trials had
been on tillage-planting systems and had, as a result, involved a
disprOportionate number of male headed and richer households. It was
hOped that formation of farmer groups would create additional community
interest and would broaden the base of farmers involved on-farm
research.

As part Of the institutional experiment, different types of groups
were formed in the three villages. In Makwate, a large, heterogeneous
group was formed. In Shoshong, the group was based on active and
interested farmers who had adequate resources for farming, but were not
particularly wealthy. Most participating households comprised small
conjugal units and both spouses were encouraged to attend meetings. In
Makoro, the group involved just females and most were from female headed
households. In addition, nearly half of the members served on the local
VDC. Two of the groups had ten members, while the one in Makwate had 21
members. Most of the individuals attending meetings were females.

Farmer group meetings were held once a month. Meetings generally

1‘...

 

201

consisted of three sets of discussion. The first was a review of the
farmers' circumstances and problems since the prior meeting. The second
was a discussion of trials' implementation (during the early part of the
season) and intermediate trial outcomes (during the later part of the
season). The final part of each meeting consisted of a discussion of
existing government programs and how farmers might take advantage of the
program.

Several procedures were followed to increase the likelihood of
success: (a) the groups were focused on well-identified problems which
were recognized by the farmers themselves as being problems; (b) seed
and advice were given to farmers to provide personal incentives for
group membership; (c) open discussions of problems and trials
implementation created peer pressure to be more active; (d) regular
meeting dates were set to reduce the chance that farmers would forget
about meetings; (a) meeting agendas were prepared, to make sure meetings
appeared to be accomplishing something; (f) each farmer had one or more
trials which served as a focal point for farmer involvement in the
groups; and (g) farmers were taken on field visits so they could compare
their relative progress.

The farmer group experiment was generally successful, although not
all problems were resolved. The group discussions were particularly
helpful in clarifying the instructions for trials and creating pressure
to implement trials. The discussions of general farming problems were
less satisfying because most problems had no identifiable solution. For
example, there were severe insect pest attacks but there was no
possibility of a spraying intervention (neither technically nor

economically feasible).

202

The two smaller and more homogeneous groups in Shoshong and Makoro
worked out much better. There was a higher rate of successful trial
implementation and there were more vigorous discussions. The most
animated discussions took place when farmers interacted on the basis of
differing personal experiences. Females attended the group meetings
most regularly and were most active in the trials. Males tended to
dominate discussions in which they were involved, but Often were
satisfied to sit back and wait for a topic on which they felt they had
relatively greater expertise. The females were more prone to talk about
their problems and to seek advice.

After the first season, group members were asked to assess the
group meetings. All group members said that they wanted to be a member
of the group during the following season. Also, all members said they
understood the instructions for the trials which were presented through
the group meetings. All but one member said it was useful to hear about
the problems of other farmers.

D. PARTICIPATION IN ALDEP AND DROUGHT RELIEF

Despite the amount of Ministry resources funnelled into ALDEP,
Drought Relief, and ARAP (beginning the 1985-86 season), there is no
systematic monitoring of participation patterns or farmers' assessments
of the programs. Resources were not available to carry out the type of
monitoring which is needed, but some data were collected on
participation in ALDEP in 1983, participation in both ALDEP and Drought
Relief in 1985, and farmers' assessments of the programs. The
objectives were: (a) to see who is benefiting from the programs, and (b)
to identify priorities for a monitoring program (if and when one is

established).

203

1. PARTICIPATION IN ALDEP IN 1983

In the 1983 Agricultural Demonstrator survey, ADS reported that
they had processed an average of 13 to 15 ALDEP loans during the year
preceding the survey. This can be compared to an average of more than
400 farming families per extension area. The main package requested was
fencing, followed by water catchment tanks, implements (primarily row
planters) and then the draft power packages.

Meeting asset requirements for the implements package (for which
draft power and a destumped field were required), was cited as the main
problem farmers' faced in getting approval of loans, deSpite the focus
of ALDEP on resource poor farmers. Supply problems and late approval of
loans (relative to the planting season) were also seen as major problems
in the ALDEP program. (Since 1983, ALDEP has been changed from a
subsidized loan program, to a down payment plus grant program.)

An overview of household participation in ALDEP in Shoshong and
Makwate was obtained in 1983 through the Crop Management Survey. Of 116
households in the survey, only 62 percent had even heard of ALDEP.
Slightly greater prOportions of male headed and poor households had
heard of the program, but the differences in prOportions were not
statistically significant. Only 28 percent of the households which had
heard of ALDEP had actually ever applied for the program. The
prOportions of households which both knew of the program and had applied
for subsidized loans did not differ significantly depending on village
location, sex of head of household, or cattle assets.

2. PARTICIPATION IN DROUGHT RELIEF AND ALDEP IN 1985
During the 1985 DUMI Study, reSpondents were asked to indicate

which Drought Relief programs they had participated in during the year

204
preceding the survey, and which ALDEP packages they had ever received.

In the 1984-85 season, nearly all households received seed through
Drought Relief. In addition, more than a third participated in the
traction hire program and another third in the destumping program. A
greater proportion of female headed and cattle rich households
participated in the traction hire scheme than did male headed and poorer
households, respectively. Participation in the food aid and destumping
programs differed significantly by village.

As of 1985, fencing continued to be the most pOpular ALDEP package.
According to the regional ALDEP officer, the fencing package was by far
the most popular package throughout the Central Region, followed by the
row planter package. Neither water catchment tanks nor the traction
packages were in much demand.

3. FARMER ASSESSMENT OF PROGRAMS

During the 1984-85 season, the three main assistance programs in
Shoshong and Makwate were the seed distribution, destumping and traction
hire schemes, all under the Drought Relief. In the DUMI Study,
respondents were asked to assess these programs, plus any ALDEP package
they had received.

A majority of reSpondents felt the seed program most directly
benefited females, while males were seen as the main beneficiaries of
the destumping scheme. An equal number of respondents said females as
Opposed to males were the main beneficiaries of the traction hire
scheme.

The main source of information for the Drought Relief schemes was
village meetings addressed by the ADS. A few respondents said they had

heard about the programs through personal contacts with the ADS or from

205
the radio. There was little agreement among the reSpondents about the
qualifications for the various programs, indicating these were not
clearly explained to farmers. Many farmers had the impression that
qualifications were more stringent than they actually were.

Most respondents approved strongly of Drought Relief (as could be
expected). All participants in the destumping scheme said it helped
their household, due to additional cash income and the clearing of
additional land. However, several respondents said they were not able
to take full advantage of the traction hire and seed distribution
schemes. For example, more than a third of the participants in the
traction hire scheme were not able to arrange for anyone to plow for
them. For the same reason, and due to a lack of rain, more than a third
of the households did plant the seed they were given.

With reference to ALDEP, an equal number of respondents reported
that males versus females were the main beneficiaries, essentially all
the respondents heard about the programs at village meetings called by
ADS, all felt the programs had helped their household, and all said they
would recommend the program to other households. Again, there appeared
to be some confusion over requirements, with a tendency to feel
requirements were greater than they are.

MOst farmers reported that they saw little need to change the
programs. The main exceptions were with respect to seed distribution
and the traction hire scheme. Nearly half of the farmers receiving seed
complained about the Red Swazi sorghum seed distributed in the
government program (see Chapter IV as well). Unfortunately, Red Swazi
seed was all that the government could obtain but farmers were not made

aware of this. Several farmers said that more cowpeas Should be

206
distributed through Drought Relief, since a shortage of cowpea seeds was
one of their main problems.

With reference to the traction hire scheme, respondents said that
there were too few tractors to plow for everyone, even though the
government was willing to pay for plowing. Some respondents suggested
that the government Should regulate the program by setting up plowing
schedules. Others suggested that the government should buy tractors and
have government drivers supplement private contract services. Some even
prOposed that the government should subsidize the purchase of tractors.
Unfortunately, the suggestions of the farmers conflict with a large body
of experience in Africa which shows that government tractor schemes
(primarily public hire services and tractor subsidies) have been costly
failures [Pingali, Bigot, Binswanger, 1987].

E. CONCLUSIONS AND RECOMMENDATIONS

The Objectives of this chapter were to analyze the performance of
the agricultural support systems affecting arable farming, and to
identify policy and institutional changes to improve their performance.
Four support systems were investigated: the extension service, the
network of formal traders, village groups, and two agricultural
assistance programs. This section presents conclusions and
recommendations for the Ministry of Agriculture.

1. EXTENSION

The 1983 extension survey showed that there continue to be major
extension problems in the Central Region which go beyond the extension
message. The administrative changes introduced during the 19708 have
helped but the lack of a coherent extension approach is a major

constraint. Five issues should be addressed in order to improve the

207
effectiveness of ADS.

First, most ADS gave the same advice and used the same
demonstrations, regardless of extension area circumstances. Extension
priorities should be established for each extension area, based on areas
characteristics. The key characteristics would appear to be: type of
traction used, distance to nearest lands, amount of wire fencing, and
percent of farmers using early plowing and row planting. These
variables tend to distinguish progressive arable farming areas from
those which are less progressive.

Second, most ADS worked primarily with progressive farmers, but did
so less effectively than they did under the Pupil Farmer Scheme. An
extension approach should be adopted which is in between the Pupil
Farmer Scheme and the "community at large" approach. Following the
farming systems approach, recommendations should be develOped for a few
clearly identified RDs. Then, extension workers could work with
representatives ("pupils") from each domain. In this way, improvements
could be initiated and monitored for a range of farming households, but
ADS could be allowed to concentrate on a limited number of households.

Third, the ADS were fully occupied with ALDEP, Drought Relief, and
ARAP. Unless the administration of the programs is significantly
changed, the Ministry should set-up one or two mobile extension teams in
each District to implement demonstrations and hold periodic extension
"clinics."

Fourth, the ADS were not adequately prepared for their job. In BAC
and in-service training, the emphasis given to technical training should
be reduced relative to training in farm management analysis (or farming

systems diagnosis). At present, ADs receive essentially no training

 

208
which prepares them to deal with human and institutional problems. A
recent national extension assessment confirmed that ADS throughout the
country do not feel competent in farm management analysis or inquiry
methods [Trent, Styles and Ramolemana, 1986].

Fifth, there were essentially no linkages between research and
extension personnel. On-farm researchers need to participate regularly
in the ALDEP in-service training courses and to periodically attend
monthly management meetings. Extension agents assigned to villages
where there is on-farm research should be seconded to the on-farm
research team. This will improve communication to the other ADS
(through reports at the monthly management meetings) and ensure that
farmers do not received mixed messages from the research and extension
Officers.

2. TRADERS

The research on the regional trading network revealed an effective
and relatively efficient system for providing food and other consumption
items to rural households. The government should be extremely cautious
when considering any policy or institutional change which might
adversely affect the effectiveness of the local traders.

The main intervention point with reference to the trading network
would appear to be input supply. Few traders provided agricultural
implements, fencing materials, seed, fertilizer, or livestock
requisites. Input availability particularly was a problem in the small
villages. The problem of input availability might be addressed as
follows.

First, the Ministry's Department of COOperative Development should

encourage member cOOperatives to establish schedules for visiting

209
near-by small villages. Farmers could be picked-up, taken to ShOp, and
then returned with their purchased items. This is a practice general
traders use to increase their coverage.

Second, since few shOps carried the major implements, farmers had
little Opportunity for comparing prices and features at different
traders. To increase competition among suppliers and reduce the
information costs to farmers, the Department of Agricultural Field
Services should begin monitoring prices of inputs at the large traders
and cooperatives, compile a list, and have ADS post it in each Of their
extension areas.

Third, the Ministry should encourage BAMB to provide seed to
traders on credit, overcoming their cashflow problem, and saving BAMB
the expense of setting up mobile seed distribution teams. If the model
works, it could eventually be extended to agricultural implements.

The severity of the problem of product evacuation was difficult to
assess due to the lack of production. A tOp priority for the Ministry
will be to monitor production utilization patterns and trader behavior
during good seasons. If product evacuation problems are identified, the
Ministry could set-up grain selling days in each village, analogous to
the current cattle auction days. BAMB could then develOp mobile
purchasing units to circulate to the village selling days.

Finally, while the local trading network was functioning reasonably
well, there were business problems which affected the service traders
provided to farmers. The Ministry of Commerce and Industries should
provide training courses in business management at the rural training

centers .

210
3. GROUP FORMATION

The research on village groups was not encouraging. Despite their
importance in village life, most groups are not functioning effectively.
Since on-farm research and extension activities are supposed to use a
group format, improved group formation and management guidelines are
needed. Based on the village groups research and the farmer group
experiment, group formation should take into account the following
issues.

First, groups must have a very specific set of objectives and,
preferably, specific projects which serve to focus the group. The
failure of many VDCS and Farmers Committees can be attributed to their
lack of focus.

Second, most groups need to have a dual membership Structure, and
an associated dual meeting structure. If a group is small, it will not
impact on the village as a whole. But large groups often fall into
inaction due divergent interests and lack of organization. Most of the
successful groups had small management committees which met regularly
and made most of the groups decisions. Other members met only as
needed, so meetings were seen as more useful.

Third, some mechanism is needed to ensure attendance. The groups
which personally benefited group members had fewer attendance problems.
In this context, on-farm trials and demonstrations could be located
primarily on the fields of group members, and the ADS could make sure
the ALDEP, Drought Relief, and ARAP applications of group members are
processed rapidly.

Fourth, even though there should be personal benefits to group

membership, the possibility of contributing to the general welfare of

211

the community often was an important motive for the most active group
members. One approach would be to have ADS encourage groups to hold
field days and to participate, as groups, in the agricultural shows held
each year.

Fifth, ADS might need to give Special attention to younger farmers.
At present, group activities are dominated by senior household members
but it is often the younger farmers who tend to innovate. One Option
might be to set up separate "young farmers" groups. The ADS also should
pay more attention to the 4-B clubs, to ensure 4-B activities focus on
farming projects.

To develop additional guidelines, the Ministry Should turn to the
on-farm research teams. By 1985, all the on-farm research teams had
started to work with village groups. However, each team used a
different approach than the one tried in the Central Region. An effort
should be made to review and synthesize the other experiences with
farmer groups.

4. ALDEP AND DROUGHT RELIEF

An adequate assessment of ALDEP and Drought Relief could not be
carried out, due to limited resources. Because of the money invested in
these programs, along with the recently introduced ARAP program, a large
scale national assessment of the three assistance programs should be one
of the top priorities for the Ministry.

The research carried out on ALDEP and Drought Relief showed that
farmers liked the programs and had few suggestions about alternative
programs. There did not seem to be any major biases in participation
patterns. However, four major problems were identified which, if

verified nationally, should be addressed to improve the effectiveness of

 

 

 

 

 

 

   

 

212
the programs.

The most Obvious problem was widespread confusion about what was
available and what requirements had to be met. The confusion was
partially due to the inability of ADS to explain the requirements of the
programs. Equally important, the requirements for each of the programs
had changed several times. The continual iteration confused both
farmers and extension agents. Also, public announcements about the
programs often conflicted with the instructions given to ADS by their
District Officers.

Second, the programs did not address constraints on input access.
For example, many farmers could not plow because of a lack of access to
hired traction. A lack of fencing poles often led to delays in
installing ALDEP fencing.

Third, many households were unable to meet the minimal down payment
requirements for ALDEP. The Ministry might consider setting up village
work programs for ALDEP applicants, following the format Of the
"food-for-work" program of the 19603, so no cash down payments would be
required.

Fourth, there were too many packages for too many farmers. The
number of packages not only was a financial drain, it overwhelmed the
ADS and District Agricultural Officers. It clearly is necessary to set
priorities. This will require macro economic analysis, as well as a

national farm level survey.

VII. SUMMARY AND IMPLICATIONS

This chapter: (a) reviews the research objectives and approach, (b)
summarizes major findings, (c) identifies implications for the Ministry
of Agriculture, (d) gives an assessment of the research approach, and
(e) Suggests guidelines for future farming systems research.

A. RESEARCH OBJECTIVES AND APPROACH

The goal Of the thesis has been to provide a comprehensive, systems
analysis of arable farming in the Central Region, and of the factors
affecting arable farming, in order to determine arable farming
deve10pment priorities. To accomplish this goal, the specific research
Objectives were to: (a) to describe household circumstances and identify
implications for technology deve10pment; (b) to describe farmers'
production practices, priorities and perceptions, in order to determine
priorities for technology research; (c) to characterize cropping
outcomes and identify production practices which improve crOpping
outcomes under drought conditions; and (d) to analyze the performance of
agricultural support systems in order to identify policy and
institutional Options to improve their performance.

Guidelines on arable deve10pment priorities are needed because the
level of crop production is low and unstable in the Central Region, as
it is throughout Botswana. To stimulate farm employment and reduce food
imports, the Government of Botswana has established several programs
which subsidize rural households. The programs are not financially

213

 

214
sustainable. Consequently, there is an urgent need to identify ways to
develop arable farming which are not based on resource transfers.

The research used the farming systems approach. The farming
Systems approach: (a) takes a holistic rather than reductionist
perSpective, (b) generally uses households, farm systems or production
subsystems as the unit of analysis, and (c) follows a systems problem
evaluation sequence.

The systems approach used in this research took into account three
key methodological themes in the farming systems literature: (a) the
need to understand how household circumstances affect farm systems
performance and Options for improvement [McKee, 1984; Moock, 1986]; (b)
the value of a "farmer first" perspective [Rhoades and Booth, 1982;
Chambers and Ghildyal, 1985]; and (c) the need to analyze agricultural
support systems [Norman, 1980].

The research approach, including the data collection procedures,
was described in Chapter II. Most of the data were collected in a
series of surveys and on-farm trials carried out in Shoshong and Makwate
villages between October 1982 and June 1986. Chapter III presented an
analysis of household circumstances. Chapter V shifted from a household
perspective to a production subsystem focus. CrOp systems management
was characterized, emphasizing farmers' priorities and perceptions as
well as existing practices. Chapter VI continued with a focus on the
crop production subsystem, but turned to an assessment of cropping
outcomes and modified production practices. Chapter VI returned to a
broader systems perspective. Three key support systems were examined,

to identify problems and priority institutional changes.

215

B. SUMMARY OF FINDINGS

Household circumstances were investigated in order to characterize
and stratify the relevant farming pOpulation. The analysis made it
clear that, at least in drought, arable production plays a limited role
in meeting household objectives. Labor requirements for competing
income activities and for household maintenance resulted in labor
Shortages for many households. Land resources were inadequate, at least
for some households in normal rainfall seasons. Many households lacked
the basic complement of draft animals and a plow needed to control the
timing of plowing.

It was shown that the pattern of inequality was more pervasive than
indicated by previous studies. Households with fewer cattle also tended
to have less-develOped land resources, fewer implements, cashflow
problems, and more limited food consumption. The pattern of inequality
was associated with draft—dependent households, not just with
cattle-poor and female-headed households. Therefore, from a technology
deve10pment perspective, the recommendation was made to concentrate on
draft access, followed by labor shortages and land Shortages. Through
this Strategy, the Ministry can be confident of addressing the problems
of female-headed and cattle-poor households.

The research on crops systems management confirmed that there was
little use of recommended practices but also showed that there were
important variations in the use Of "traditional" practices. There were
small differences in practices across the household types, reinforcing
the importance Of resource control and timely field operations. Based
on the review of crOp systems management, several priority technical

investigations were identified. None of the prOposed investigations

216
would require substantial changes in practices or resource requirements.

The chapters on household circumstances and crop systems management
both revealed the importance of gender roles in labor use and decision
making. The chapters also showed that resources and practices differed
significantly across villages. Therefore, the argument was made that
the targeting of research and extension activities must in some cases
focus on individuals or villages, not just on households.

Several technological Options for enhancing the production
potential of Batswana farmers were examined. Although outcomes of the
research program were dominated by a severe drought, some profitable
Options were identified.

The intervention with the greatest potential for increasing farm
productivity is row planting. Additional farm production and employment
could be generated, and the system has the potential to be expanded
without a weeding labor constraint. In El trials during the 1983-84 and
1984-85 seasons, early plowing plus row planting increased yields by
more than 40 percent. Farmers could have made an extra Pula per hour by
shifting to row planting (or an extra P20 per hectare with labor valued
at P0.38 per hour). Yield increases with row planting were even greater
in RI trials.

Double plowing was equally profitable and represents an apprOpriate
intermediate technology, since no planter is required. The main
disadvantages of double plowing are: (a) a full draft team is needed
twice as Often and (b) the relative gain is sensitive to the soil
environment and post-planting rains. Nevertheless, on soils with high
water holding capacity, a shift to double plowing increased net income

per hectare by more than one hundred Pula (based on RI trials in

217
1985-86).

The main Options identified for draft-dependent households were:
(a) sole plantings of secondary crops and (b) hand planting for gap
filling. Small plots of secondary crops (a) would not interfere with
household food objectives, (b) would not require much additional labor,
and (c) would increase farm profitability. Hand planting for gap
filling provided returns exceeding the urban wage rate, despite drought
and low yields.

The research on the extension service showed that there are major
extension problems in the Central Region. The lack of a coherent
extension approach was identified as a major constraint. At present,
the extension approach is based on village meetings, groups and
demonstrations. However, only 1-2 demonstrations per year are
implemented in each area, the same messages are promoted at village
meetings regardless of extension area characteristics, and few ADS work
actively with village groups. Moreover, extension contacts are biased
toward male and progressive farmers. TO improve the effectiveness of
the extension service, guidelines were presented for a modified
extension approach which builds on the FSR concept of recommendation
domains.

The trading network was examined because of the importance of food
purchases and purchased production inputs. It was shown that
essentially all farmers have access to grain and meal at low prices.
Few traders, however, sold seed, farm implements, livestock medicine or
fencing materials. Alternatives for increasing access to inputs were
suggested.

The research on village groups showed that most groups are not

218

functioning effectively. Nevertheless, an "institutional group
formation experiment" was conducted successfully, resulting in higher
trials implementation rates than had been achieved using an individual
contact approach. With prOper group formation guidelines, village
groups could be used to expand the coverage of on-farm research.

C. IMPLICATIONS FOR THE MINISTRY OF AGRICULTURE

This section identifies five major implications of the research for
the Ministry of Agriculture. Taken as a whole, the recommendations
outline a strategy through which the Ministry can make a substantial
contribution to arable farming deve10pment over a ten to twenty year
horizon.

1. SHIFT TO A HOLISTIC, SYSTEMS APPROACH

Perhaps the single greatest implication of the research is the need
to approach arable farming deve10pment in a holistic, systematic manner.
This recommendation applies to the way the Ministry functions, as well
as to the scope of the deve10pment agenda.

The review of technical research showed that no major solutions
exist for dryland arable farming. Even if solutions are found, they
will be hard to implement due to the number of potentially binding
resource constraints. On the other hand, there do appear to be some
options for contributing to arable farming development through
institutional and policy changes. Even so, it is not likely that
substantial improvements will result from any Single change. Therefore,
a wide range of technological, institutional and policy interventions
must be develOped.

A holistic approach is both possible and affordable. Although the

Ministry as a whole must use a holistic approach, not everyone working

219
in the Ministry must. For example, the RSU could focus on village
institutions and farmer decision making. BAMB could be assigned
responsibility for working on input supply. A holistic approach would
encourage an evaluation of the activities of each department and
division in terms of their contribution to a well-defined development
strategy, which encompasses institutional improvement as well as
technological change.

The biggest constraint preventing the Ministry from addressing
arable farming deve10pment in a holistic manner is the lack of
communication between the various departments. DPS develops plans based
on technical assumptions which are not supported by research. DAFS
promotes technologies which DAR officers feel are no better than
existing practices. BAC does not provide the type of training ADS need,
and DAR researchers provide little or no input into the BAC curriculum.

To improve communication, the Ministry Should pursue the concept of
decentralization, as prOposed in the NDP. Regional officers should be
appointed for the DAR and DPS, corresponding to the RAOs in DAFS. These
officers could work together to design and implement region-specific
strategies for the Ministry. It would not be necessary to add new
Officers to fill the positions. Rather, as officers are trained over
time, an increasing share of trained officers should be allocated to
regional positions.

2. GIVE GREATER EMPHASIS TO "INTERMEDIATE" PRODUCERS

A major policy issue in Botswana is whether the government should
be devoting so much of its scarce resources to the traditional arable
farming sector. The last NDP, for example, mandated the Ministry to

give more attention to non-traditional producers, and to irrigated

220
farming [MFDP, 1985]. The National Food Strategy pays little attention
traditional producers relative to non-traditional producers and food
imports [RDC, 1985].

Efforts to develOp non-traditional producers certainly could help,
and should be pursued as part of a holistic deve10pment strategy. But
there is a limit to the potential contribution from the non-traditional
producers, both with reference to food production and employment
Opportunities.

The Ministry can make a larger contribution to national deve10pment
objectives by devoting more resources to improving the productivity and
capacity of the "intermediate" traditional farmers. The intermediate
farmers are those who consistently cultivate more than ten hectares,
often own (or at least hire) tractors, intentionally try produce a
surplus for sale, have above-median cattle wealth and well-develOped
land resources, and row plant (or at least do "progressive traditional"
practices such as early planting).

With appropriate access to inputs and extension advice, the
intermediate farmers should be able to make a substantial contribution
to national food requirements. Research on appropriate techniques for
the intermediate farmers could, for the most part, be carried out in
conjunction with research on techniques for the Barlong, Tuli Block and
Pandamentenga farmers.

Policy and institutional support is needed for the intermediate
farmers, as well as apprOpriate production techniques. For example, it
would be desirable to have the landboards allocate substantial blocks of
land. Also, steps to create a rural wage labor market for agriculture

would be beneficial, Since many traditional farmers face labor shortages

221

due to their higher levels of production. Many intermediate (and
non-traditional) producers also need to have better access to short term
loans.

The recommendation to give greater attention to intermediate
farmers does not imply that the Ministry Should abandon efforts to help
poor farmers. Instead, the recommendation is to adOpt a dual strategy
toward the traditional arable farming sector. For the majority of
poorer farmers, the focus would continued to be on resource transfers
(mainly through ALDEP) and modest technical interventions. However, the
Ministry of Agriculture cannot expect to solve the problems of the
resource poor ‘and make substantial progress in arable farming
development. Most of the needs of the poor will have to be addressed by
other ministries.

3. TARGET RESEARCH, EXTENSION AND PLANNING ACTIVITIES,

There is a need to develOp a comprehensive strategy for targeting
the Ministry's planning, research, and extension activities. Targeting
serves two interrelated functions. First, it clarifies the relevant
population for each technology and program. This should help field
personnel when providing advice, explaining Options, and processing
applications. Second, targeting can be used to establish priorities for
technology and program development. An example is the attention given
by the Central Region research program to post-establishment management
practices--because some interventions were needed for draft-dependent
households.

The research suggests that different targeting strategies could and
should be used for planning, research and extension activities. The

primary criterion for planners (focused on assistance programs) should

222
be cattle assets. Cattle assets are an excellent proxy for overall
wealth, cashflows and food consumption. Gender of household head could
serve nearly as well, but several researchers have rightly noted that
there is great diversity within the category of female-headed households
(for example, Kervin [1977] and Gulbrandsen [1980]).

Technical researchers should use draft access as a stratifying
variable. Within each draft access research domain, a range of options
should be develOped to address labor constraints and, as a second
priority, land constraints. Reducing the work burden of women should
receive increased emphasis relative to tillage—planting interventions.
To the extent possible, all interventions should require minimal cash
investments.

Targeting for the extension service should first focus on extension
area characteristics, so "themes of deve10pment" can be develOped for
each area. Within each area, ADS should try to set up groups
representing the draft access domains. Special groups should also be
set up for young farmers and for women.

4. SHIFT THE FOCUS OF TECHNICAL RESEARCH

In the past, technical research has emphasized the development and
evaluation of new tillage-planting systems, crOps and varieties, and
implements which perform new functions. Most new systems and implements
have failed when tried on-farm, and few farmers have adopted the
Ministry's extension recommendations. The failure of the Makgonatsotlhe
tool-carrier is a classic example of what happens when researchers try
to radically transform existing practices through package approaches and
"single best" solutions.

In future technical research, there should be less emphasis on the

223
development and introduction of new systems. Two alternative, but
complementary, types of investigations should instead be emphasized.

First, technical research should focus on the relative advantages
of alternative practices, varieties and implements under different
environmental (and household) circumstances. From the Central Region
research, the example of double plowing can be cited. Overall, the
benefits of double plowing are marginal but they can be substantial when
double plowing is done on soils with high water holding capacity. The
objective of this type of research would be to quantify expected
benefits.

Second, technical investigations should focus on identifying ways
to modify traditional practices (to increase farm productivity) and
existing extension recommendations (to make them more attractive to
farmers). Many farmers feel that the existing recommendations,
particularly row planting, could increase their production, but feel
they have constraints which prevent adOption. Researchers need to help
farmers figure out how to "fit" the new practices into their existing
production system.

5. INCREASE THE EMPHASIS ON MACRO POLICY ANALYSIS

Most Ministry economists are involved in generating data for
planning purposes or in analyzing various programs and projects. The
task of addressing problems in arable farming deve10pment has been
turned over to the on-farm research teams. However, the contribution of
on-farm research to policy and institutional analysis is limited.
Micro—economic research can identify problems and determine guidelines
for potential solutions, but the range of factors considered in

micro-research is too narrow to solve national development problems.

224

Therefore, a top priority for the Ministry is to reallocate some of
its scarce social scientist resources to the macro economic analysis of
arable farming development policy. Urgent analyses are needed on
tractor importation and/or subsidies, food grain pricing and food
imports, and the relative employment and income effects of the ALDEP
versus Drought Relief versus ARAP.

One goal should be to make sure government revenues are used
efficiently. A longer-term goal should be to develOp a realistic
sectoral deve10pment strategy, so that when new projects are solicited
or prOposed, the Ministry can make sure the prOposals contribute to
sector deve10pment goals. The NDP prepared by the MFDP is not
sufficient for this purpose.

E. ASSESSMENT OF RESEARCH APPROACH
AND GUIDELINES FOR FUTURE RESEARCH

The research presented in the thesis had a broad focus,
synthesizing findings from several surveys and experiments conducted
over several years. The surveys and experiments were primarily based on
a small sample of farmers from two representative villages. This
section: (a) briefly reviews the reasons why a small sample,
representative village, holistic approach was used; (b) points out the
major limitations of the approach, and (c) suggests priorities for
future farming systems research.

1. REPRESENTATIVE VILLAGES

The representative villages approach was based on the view that
differences among farmers within a village often are more important for
assessing agricultural programs and technologies than are differences

across villages. On balance, the thesis confirms this viewpoint--at

225
least for Botswana. The two regional surveys (the Agricultural
Demonstrator and Trader Baseline surveys) both showed that arable
farming practices and circumstances were relatively similar throughout
the Central Region. The analyses of circumstances for households in
Shoshong versus those in Makwate did reveal substantial differences, but
this only gives an argument against a single-village approach--not a
representative villages approach.

Future farming systems research should concentrate on a small
number of villages, rather than covering many villages or using large
national or regional surveys. It would be desirable, however, to make
two adjustments in the representative village approach. First, it would
be better to cover at least two villages for each "type-of—village," to
help distinguish unique village features from the "representative"
village features. For example, are all donkey villages poor, just
because Makwate is a poor village? The first adjustment implies
working in 6-8 villages rather than three, as done by the Central Region
team. Second, periodic regional verification surveys should be
administered. As long as the surveys are focused on specific issues
(ie., are not large baseline surveys), it should not be too expensive to
carry out a regional verification survey once a season, or every second
season.

2. SMALL SAMPLES

The main reasons for using small samples of farmers were: (a) to
minimize measurement errors in resource monitoring and enumerator errors
in subject surveys, and (b) to reduce the amount of researcher time
required for checking, cleaning, and analyzing the resource flow data.

From a practical standpoint as well, it was not necessary to generate

226
"large sample" baseline data because regional and national data were
already available, even if insufficiently detailed and disaggregated for
farming systems diagnosis.

In retrOSpect, the small sample approach worked extremely well for
the subject surveys, but not so well for the resource monitoring. The
small samples for the subject surveys enabled existing staff to conduct
all the surveys at non-peak periods. Therefore, both the real and
opportunity costs of the approach were almost nil. Data entry and
analysis for several of the surveys took less than ten person-days.

To refine the small sample-subject survey approach, three steps
should be taken. First, conduct only one large (baseline) subject
survey each year. The first season was over-loaded when three large
surveys were administered. Second, a core of "COOperator" farmers
should be included in the large surveys, to eventually generate a
complete household profile and to enable an analysis of trends over
time. Third, new farmers should be selected for each survey, to
supplement the cOOperators. This will reduce the burden on cOOperating
farmers and enable verification of key issues with new samples.

The small sample strategy for resource monitoring was not so
successful because, as it turned out, it was an impossible task to
eliminate data errors. TOO much time would have been required, and the
opportunity cost was too high.

A major question is whether farming systems teams should be.
investing in resource monitoring. In Botswana, the answer is "no." Due
to the difficulty and cost of resource monitoring it would be a poor use
of resources relative to the information gained. In this study and

previous studies, many of the resource flow patterns of households have

227
been well outlined. To the extent on-going monitoring is required, this
should be the responsibility of the DPS, not the farming systems teams.
Therefore, as an alternative to their own resource monitoring, the
farming systems teams should meet with the Farm Management Unit to
revise the National Farm Management Survey, to make that survey more
useful.
3. HOLISM

Perhaps the most distinctive feature of the thesis is the number of
topics covered. The holistic approach, in part, obtained from the
particular circumstances of the research. It was possible to cover many
topics because the research did extend over four years. The topics
covered during the first year were, in many respects, the most crucial
(crap management, draft arrangements, and the extension service). It
was possible to cover other areas because the drought reduced the time
required for analyzing trial outcomes.

A holistic approach was followed for three methodological reasons
as well. First, investment of research resources in any single
information gathering or data analysis activity results in rapidly
diminishing returns. Many of the subject surveys--and related analyses
of household circumstances, crop systems management and agricultural
support systems--only identified major issues. But they did so at low
cost. A valid question can be raised as to the utility of more in-depth
analysis until such time as a strategy is develOped about the major
issues.

Second, the major function of this thesis was to provide guidelines
on arable deve10pment priorities, in order to contribute to the

formulation of a viable development strategy. Strategic-level analyses

228
inevitably require a broader range of information than is required for
making decisions about particular issues or problems.

Third, to have focused on one or two problems in Botswana would
have been an obvious violation of two of the distinguishing principles
of the systems approach: (a) Optimization with respect to a part of a
system does not necessarily mean that the performance Of the entire
system will be improved, and (b) because of the synergistic interactions
of systems components, systems cannot be studied by considering their
components in isolation. While it is possible to study particular
components using a systems perspective, this was not feasible in
Botswana because: (a) the harsh environment precludes substantial
improvement from any single systems change, and (b) arable farming is
some reSpects a minor contributor to household objectives, and a minor
contributor (at present) to national objectives.

Although there are many advantages to a holistic approach, there
are some costs. Three of the most important are: (a) there is less
depth in the analysis of any given issue, (b) it generally is necessary
to make forays into areas where other researchers have greater
expertise, and (c) the range of issues addressed can be
overwhelming--leading to inaction due to uncertainty about priorities.
The first "cost" is not so severe in a multi-year program. In this
research, tillage-planting interventions and farmers' practices and
priorities were investigated in—depth because of their high priority
while other topics were addressed with a minimal research investment.

The "expertise" issue is a real problem, since systems research is
inevitably multi-disciplinary in scope. It clearly is necessary to have

back-stOpping from several disciplines if resources are not available

229
for a large research team. The research approach was feasible in the
Central Region because the team members had formal disciplinary training
in economics, anthrOpology, political science, production agronomy, and
plant breeding. This situation will not always Obtain on a small team.
If it does not, it certainly would be better to focus investigations on
areas which can be competently addressed, rather than to adOpt a
holistic approach as a methodological principle.

Finally, the issue of policy—maker "overload" should not be the
determining criterion of whether a holistic approach is used or not.
There is always a bureaucratic bias in favor of a focus on one or two
problems, one or two priorities, and one or two solutions. The
determining criterion should be a comprehensive "needs analysis" related
to systems performance. In the case of arable farming in the Central
Region, there was a greater need to identify priorities in several areas
than to solve a Specific problem. Once Ministry Officials have a viable
deve10pment strategy, future farming systems research can be more

focused on specific problems and issues.

APPEND IX

APPENDIX

PROFILE OF AGRICULTURE IN BOTSWANA

The appendix presents an overview of agriculture in Botswana. The
objectives are to characterize the research setting and place the
research into a national agricultural deve10pment context. The first
section describes the technical environment, including rainfall
patterns, temperature, soils, and vegetation. The second section
presents an overview of the agricultural sector. The final section
reviews the important national and local institutions affecting
agriculture.

A. TECHNICAL ENVIRONMENT
1. RAINFALL

Rainfall is the single most important factor affecting agriculture
in Botswana. In eastern Botswana, where most arable farming takes
place, annual rainfall declines in a diagonal transect from 600 mm. in
the northeast to 250 mm. in the southwest. Most of the annual rainfall
is received during the summer rainy period, generally beginning in late
October or November and continuing to March. However, the beginning of
the rains is often late, there is a high frequency of prolonged dry
Spells during the growing season, and rainfall tends to be localized.

Most rainfall analyses have focused on inter- and intra- annual
patterns. NO significant correlations have been found between years or
between periods within seasons [Pike, 1971; Bhalotra, 1984; 1985]. Dyer
and Tyson [1977] did identify a twenty year cycle, ten years of wet and

ten dry, for the South African sub—continent but it has too much annual

230

 

231
variability to be of use in prediction for any given year. The 19703
were accurately predicted to be above average rainfall years while the
19803 were predicted to have below average rainfall.

With reference to the Mahalapye area, Siebert [ATIP, 1986]
calculated an average annual rainfall of 464 mm. over a 73 year period,
with more than 80 percent of the total expected from November through
March. Rainfall was substantially below the long term average during
the years when the research was conducted. With the exception of a
freak month (March, 1983, when 257 mm. fell), rainfall averaged only 270
mm. per year during the first three season, with poor distribution.

2. TEMPERATURE AND EVAPORATION

There are two crop production problems caused by temperatures.
First, during December and January, the period with the highest
probability for rain, it is hot. Daily maximum air temperatures average
33 degrees celsius, and can reach 40 degrees or more. This can lead to
soil surface temperatures of as much as 60 degrees but the cardinal
temperature for sorghum seedling viability is 40 to 47 degrees. Thus,
temperatures alone can have a detrimental effect on seedling
establishment.

The second problem is the onset of cool nights at the end of the
season. The cardinal temperature for sorghum maturation is around 15
degrees celsius. This temperature is commonly reached at night
beginning in April. As a result, February plantings often will not
reach harvest stage before the maturation process is Slowed. If there
are even short drought periods, during which sorghum stOps growing, even

January plantings may not mature.

232

High temperatures, along with radiation and wind, lead to another
problem--high evaporation and evapotranspiration rates. Evaporation
rates have been estimated to be around 1.7 to 1.9 meters per year.
Evapotranspiration rates are around 1.4 meters per year, with daily
rates exceeding 5 mm. In every month of the growing season, these rates
exceed expected rainfall.

3. SOILS

Siderius [1972] reported that the arable part of Botswana is
covered by ferruginous tropical soils. Many soils are Shallow and have
low water holding capacity. Most soils are weakly develOped, consisting
of medium to coarse grained sands and sandy loams. In many places, the
soils are subject to crusting following heavy rains. Most soils are
deficient in phOSphorus, have low (but variable) levels of mineral
nitrogen, and are very low in organic matter.

Siebert [ATIP, 1986] made a detailed assessment of soils in the
Mahalapye area based on an analysis of soil samples from 41 fields in
the villages where the Mahalapye team has conducted research. Most
soils are high in coarse grained sands (sand being more than 50 percent
of the soil mass) with a sandy loam to sandy texture. These soils are
more easily plowed than are the sandy clay loam soils in the area but
their pH values are low for sorghum production, phOSphorus is very low,
organic matter is almost not present and the total base saturation is
low. There is tremendous within-field variability, some of which is
significantly associated with micro-topography.

4. VEGETATION
Tree savanna zones cover most of the arable parts of the country.

These zones are dominated by Acacia. Several species of Acacia are

 

 

III 1

 

233
bushes which encroach in areas with sandy loam soils, particularly where
there has been overgrazing or heavy cultivation.
Siebert [ATIP, 1986] noted that arable activity in the southern
part of the Central Agricultural Region is mainly carried out in

woodlands (or tree savanna) dominated by Combretum apiculatum and

 

 

Acacia nigrescens . These woodlands give away in the northern part of

 

 

the Region to ColoPhospermum mopane bushveld. The transition in

 

vegetation from south to north is accompanied by an increase in soil
clay content and alkalinity. There is variability of vegetation types
within localities due to soil texture, soil reaction, tOpography and
land use.

B. AGRICULTURAL SECTOR OVERVIEW
1. SECTOR STRUCTURE

Two subdivisions usually are made when characterizing the
agricultural sector in Botswana: livestock and arable farming, or
traditional and freehold. The sub-division of the sector into
traditional versus freehold is based on a special tenure arrangement
made for the white minority. The small freehold sector, which in 1983
comprised 360 farms, accounts for a disproportionate share of area
cultivated, total food crop production and cattle offtake. Selected
characteristics of the freehold and traditional sectors, as of 1983, are
presented in Table A.1.

The Sixth National Deve10pment Plan [MFDP, 1985] identifies (for
the first time) two sub-groups within the traditional sector. One group
comprises "intermediate" farmers who own above average cattle herds and
aim to produce a surplus to market. These farmers generally cultivate

over ten hectares. In 1983, the farmers cultivating over ten hectares

 

234

TABLE A.l: FARM STRUCTURE: TRADITIONAL AND FREEHOLD SECTORS, 1983

 

 

TRADITIONAL FREEHOLD ALL
TOTAL FARMS 82,000 360 82,360
FARMS WITH LAND 60,900 150 61,050
AREA PER FARM WITH LAND 5.0 166.7 5.4
FARMS PLANTING CROPS 48,100 100 48,200
CEREAL PRODUCTION PER CROP FARM 0.2 53.5 0.3
FARMS WITH CATTLE 58,300 345 58,645
CATTLE PER CATTLE FARM 41.3 1,190.4 48.1
CATTLE OFFTAKE PER CATTLE FARM 4.0 496.2 6.8
SMALLSTOCK 901,800 45,700 947,500
SMALLSTOCK PER FARM 11.0 126.9 11.5
CHICKENS 660,800 300,000 960,800
CHICKENS PER FARM 8.1 833.3 11.7

 

Source: Ministry of Agriculture, Division of Planning & Statistics,
1984.

accounted for 73 percent of total cereal production even though they
represented just 6.3 percent of the traditional farms. The second group
have few or no cattle and Often are not able to grow enough sorghum to
cover subsistence requirements.

Dryland arable farming, whether traditional or freehold tenure, now
is also being distinguished from two systems not so reliant on rainfall.
One is irrigated farming. There are only about 1,000 hectares under
irrigation in Botswana, nearly all of it on the freehold farms. The
second is the traditional system (in the northwest) of "melapo" farming.
Melapo are flood recession fields in which crOp establishment and
deve10pment takes place mostly on residual soil moisture.

2. CROPPING PRACTICES

CrOpping practices are quite similar throughout Botswana. Table

A.2 shows that sorghum is the most important crop in four of five of the

major agricultural regions, with Maun being the exception. The Maun

235

 

mwwa uxoa posawucoo

 

w 0

Ha ooH
am mm
mm mN
om mm
0H mm
«N mH
«a no
em mm
Hm n\o

me am
om om

oo~.w¢ ooq

ma
mm

ma
ON
mm

NA
6
Nu

om
as
me
mm

ooH.m

N
mm

on
HH
mo

m
ma
mo

co
on
mm
«a

oo~.o

m
mm

ma
mm
mm

ow
mm
ms

Ho
m

mm
mm

oom.¢~

N mN NNNSnuo
mm mm ammopwoum
"Amzm¢m mo NV nomHmz wzHHz<Am
nu ma m\uo:uo
cm «H cwlouwm
as we :30
"Amzm<m mo NV mmmoo< Hm<mo
N o mmoxaon
om om O\mu0uomue
NN MN mfiuumo
"AO\mzm<m mo NV mmzom Hm<mn
ac «m mason
Na N uoaaflz
oN co ONNO:
aa mm Eanwuow

"Amzm¢m mo NV oszomw m2m<m

ooN.mH oom.N «\mze<e Nome

 

AA< zmmHmm3

ZD<S ZBOHmHoz<mm A<MHzmo mzomom<w 2MmmHDOm

 

mwma

.moeomm A<zoNaHn<ma

"ZOHUMM A<MDHADUHMU< Vm mmoHHo<Mm qummomo "N.< mam<8

236

 

.uOOaa OO maouuwn om: maumm 380m mwcfiuamaa Sou haemawum .m
.mucoEoouww O>Numuoaooo nwsounu wcwcfimuno mam wamsouuon mOOOHOOH .O
.uwmup wouomuu ou ucoamamasm O we manages mafia: maumm movsaocH .v
.mxoxcov OOO mauumo om: maumw wcwcwmaou “OOH Ou Bum uoc moon .O

.ucoouom m. awn» mama How vououco aw ONON .n

.mmma aw mmouo woodman nowna manmm .m

 

 

.Tuwmd .OmUH moflumwumum w MGMCGme mo dewmfiNrMQ .mhzuafiowuwaoq mo Nahumwdwz ”wouflom
w o om N H m N mHaNusaz
No mm NS NS NN NN Hm coco
NN no N as 8N MN 83 aaoz
Amzmam mo NV mszammz
N NE E SE H N «H Assam:
m o N o c N NH 4 NsoNameo
“Aoszqmm4 mzm<m mo NV mu<mD «MNHAHemmm
Nw mm NS NS mN Na SN mammN
NN .. cm as SN mm as smHSNz
so mm on ma mN SN we muNmz
oN mm mm as EN EN ms aaswuom
"NamezaeN ammm No Ne NNNDNtz Nome zN amezaqm ammm
444 zmmemmz 23¢: zonmHoz<mm A<mezmo mzomom<u zmmmHoom

 

ApoacquOOv N.< oHan

237
Region, in the northwest of the country, has greater rainfall and is
ecologically similar to southern Zimbabwe and Zambia. Maize and millet
are of above average importance in the Francistown Region. Cowpeas are
an important intercrop throughout the country.

Nearly all land is mechanically plowed using mouldboard plows.
Most draft power is provided by cattle, usually teams of six to eight
oxen. Tractor and donkey traction are rapidly becoming major sources of
draft power, particularly in the Central Region. Private contract
services account for most of the tractor traction used. A majority of
farms relied on traction they owned in 1983.

Nearly all crOps are broadcasted and most seed is planted in
mixtures. Row planting has made the greatest inroads in the Southern
Agricultural Region. Very few farmers use fertilizers, either manure or
chemical. Few weed more than once and many do not do a single weeding
unless a sufficiently promising plant stand is achieved.

3. LIVESTOCK MANAGEMENT AND HUSBANDRY

Livestock management practices differ greatly between the freehold
and traditional sectors but are similar throughout the communal areas.
In communal areas, cattle are allowed to graze freely near the cattle
posts or lands of their owners [Carl Bro International; 1982]. However,
it is a standard practice to have at least one corral and associated
calf pen. Lactating cows are often penned part of the time for milking.
Calves generally are penned for the first several months of their lives.
General herding is done only occasionally, most commonly when there are
crOps being grown or there is a long distance to water. The standard
practice for avoiding crop damage is to maintain a separate residence

where cattle can be kept (a "cattle post") which can be quite a distant

238
from both the village and the lands areas.

Since there are few permanent water sources in Botswana, livestock
are usually watered at boreholes or Open wells during the dry season.
When there are rains, dams and pans become the main source of water.

The most common husbandry practices are vaccination, deticking,
castration, and dehorning. Cattle generally are vaccinated against
anthrax, quarter evil and brucellosis. External parasites usually are
controlled on an ad hoc basis through the hand application of "tick
grease." Castration is generally carried out at three to four months
using a burdizzo (a pinching implement). Approximately half the cattle
are dehorned.

Few farmers regulate the breeding of their animals. Nevertheless,
there is a distinct calving season, with most calves being born during
the rainy season. Most bulls are selected from within the herd.
Periodically, owners obtain new bull through purchases or exchanges with
neighbors. In few cases do herd owners seek out improved breeds to
upgrade their herds.

C. AGRICULTURAL INSTITUTIONS

Primary responsibility for deve10pment of the agricultural sector
rests with the Ministry of Agriculture. A number of parastatals have
been created to supplement private channels for credit, input supply,
marketing and trade. At the local level, several organizations
coordinate and initiate deve10pment activities.

1. MINISTRY OF AGRICULTURE

The Ministry of Agriculture has three major departments:

Agricultural Research, Agricultural Field Services, and Animal

Veterinary Services. The major units responsible for research and

 

239
extension activities are the Departments of Agricultural Research (DAR)
and Agricultural Field Services (DAFS), respectively. DAFS is the
largest department, followed by Veterinary Services.

The function of the Department of Agricultural Research (DAR) is to
provide information and technologies so agricultural production can be
increased and made more reliable. The DAR consists of a Division of
Arable Research, an Animal Production Research Unit, an Estate
Management Unit, and a Laboratory Services Unit. The Division of Arable
Research is responsible for a Seed Multiplication Unit.

Livestock research is conducted under auspices of the Animal
Production Research Unit (APRU). APRU conducts research on 16 MOA
ranches Spread throughout the country. APRU'S first priority has been
deve10pment of a technical basis for a commercial beef industry. Range
management, range improvement, beef cattle nutrition, and beef cattle
reproductive physiology have been the major areas of investigation.

The Department of Agricultural Field Services contains four major
Divisions: Animal Production, Crop Production, Land Utilization, and
Agricultural Management Associations. An Agricultural Information
Section, servicing the entire MOA, and a Field Section are also
included. The latter is responsible for all direct farmer contact by
the MOA, except that done by the Department of Veterinary Services.

There are four levels of staff in Agricultural Field Services: (a)
administrators and subject matter specialists at headquarters, (b)
agricultural officers and specialist officers in the six regions
(specialists are reSponsible administratively to the regional
agricultural officers, but are reSponsible technically to their division

chiefs), (c) agricultural officers and supervisors in the 22

240

agricultural districts, and (d) agricultural demonstrators in the 225
agricultural extension areas. Extension work is funnelled through the
agricultural demonstrators (ADS).

The Department of Veterinary Services is reSponsible for disease
prevention and control, meat inSpection, tsetse fly control,
administration of livestock advisory services (mainly input
distribution), trek route supervision, operation of a diagnostic
laboratory, and veterinary research.

The Ministry also includes a small Department of COOperative
Development, the Division of Planning and Statistics and the Botswana
Agricultural College. The Division of Planning and Statistics (DPS) is
responsible for agricultural sector planning, an annual national
agricultural survey, an on-going farm management survey, and rural
sociology research. The DPS is administratively attached to the office
of the Deputy Permanent Secretary. Botswana Agricultural College
provides training in technical agriculture through the diploma level.
BAC has trained most of the agricultural demonstrators in Botswana.

2 . INPUT SUPPLY, CREDIT AND MARKETING AGENCIES

Several governmental and parastatal agencies have been created to
promote development of the agricultural sector. Most have focused on
commercial production, with a particular emphasis on nurturing the
important export beef market.

The most important agency is the Botswana Meat Commission (BMC).
BMC has a statutory monOpoly over exports of meat and meat products, and
accounts for four-fifths of all cattle slaughtered in Botswana. BMC has
two abattoirs, one of which is the largest single manufacturing

enterprise in the country. BMC provides government revenue through a

241

turnover tax. Prices are set for different classes and change over the
year to encourage sales outside of the peak April to June sales period.

Two additional agencies directly serve the livestock sector. The
Botswana Livestock Development Corporation (BLDC) buys cattle which are
not yet ready for slaughter and has attempted to develOp finishing
ranches. BLDC is Specifically targeted to serve remote area purchases.
The Botswana Vaccine Institute (BVI) was set-up to produce an effective
foot and mouth disease vaccine. BVI is capable of an annual production
of more than 20 million doses [MFDP, 1985].

The most important agency affecting crOp production is the Botswana
Agricultural Marketing Board (BAMB). BAMB has three basic functions:
(a) to provide guaranteed, equitable prices, (b) to purchase enough
domestic produce to meet emergency demand requirements, and (c) to
provide necessary seed and chemical inputs to farmers. BAMB has
marketing facilities at 26 locations and a storage capacity of 55,100
tons. BAMB played a major role during the drought in purchasing seed
internationally.

The National Deve10pment Bank (NDB) and the Botswana Deve10pment
Corporation (BDC) provide credit and support services to both
traditional and large scale commercial operations. The NDB is the major
source of credit for farmers, currently providing more than half of all
lending to the agricultural sector [MFDP, 1985]. The BDC promotes and
assists commercial Operations and, in agriculture, has concentrated on
irrigation and dairy farming. BDC freehold farms account for the

largest irrigated area in Botswana.

242
3. VILLAGE GOVERNMENT AND ORGANIZATIONS

Since independence, traditional village authority, vested in a
headman and village meetings (kgotlas), has been subsumed into local
governmental structure. Village leaders are responsible to District
Councils. District Councils are under jurisdiction of the Ministry of
Local Government and Lands (MLGL). Land Boards under authority of the
District Councils have been given authority to allocate tribal lands.

Village Development Committees, created under the authority of
District Councils, are the executive authority in charge of deve10pment
at the village level. A village may also have a village extension team
(VET) comprised of the AD, the assistant community deve10pment officer
(village representative of the MLGL), a family welfare educator (FWE)
and the head teachers of the local schools. Most villages have one or
more Farmers' Committees which work together with the ABS to implement
Ministry projects.

Local branches of various types of COOperatives can be found in
many of the larger villages. Consumer cOOperatives create retail
outlets for consumer goods while various marketing and multi-purpose
cooperatives provide farm inputs, sell consumer goods, and purchase
livestock and farm produce. Local COOperatives are supported through
two secondary societies, the Botswana Cooperative Bank and the Botswana
Cooperative Union, and by the Ministry of Agriculture's Department of

COOperative Deve10pment.

LIST OF REFERENCES

REFERENCES

Agricultural Research Station. Review of Crop Experiments ,
Bechuanaland Protectorate 1947-1959 (Published 1959). Gaborone:
MOA (Republished), 1978.

 

 

Agricultural Technology Improvement Project. Farming System Research
Activities at Mahalapye: Summary of Activities, 1982-85;, ATIP
Research Paper Number 1. Gaborone: MOA, 1986.

 

 

Alverson, H. The Social and Economic Context of Agriculture in
Botswana: Some Indicators; Research Paper No. 6. Gaborone:
Institute of Deve10pment Management, 1979.

 

 

Alverson, H. Mind in the Heart of Darkness: Value and Self Identify
Among the Tswana of Southern Africa. New Haven: Yale University
Press, 1978.

 

 

Baker, D., and J. Hobbs. "Institutionalization of FSR&E in Botswana:
Current Programs and Issues." In Butler-Flora and Tomecek, 1986.

Behnke, R., and C. Kerven. "FSR and the Attempt to Understand the Goals
and Motivations of Farmers." Culture and Agriculture 19: 9-16, 1983.

 

Bettles, F. Ten Years After: A Look at Female Agricultural
Demonstrators in the Department of Agricultural Field Services.

Gaborone: MOA, 1984.

 

 

Bhalotra, Y. Climate of Botswana Part I: Climatic Controls. Gaborone:
Ministry of Works and Communications, 1984.

 

Bhalotra, Y. Drought in Botswana. Gaborone: MWC, 1985.

 

Bond, C. Women's Involvement in Agriculture in Botswana. Gaborone:
Government Printer, 1974.

 

Boserup, E. Woman's Role in Economic Deve10pment. New York: St.
Martins Press, 1970.

 

Boykin, C. Farm Management Survey Results, 1982 Survey, 1978-82
Survey Comparisons. Gaborone: MOA, 1983.

 

 

Brokensha, D., D. Warren, and 0. Werner, eds. Indiggnous Knowledgg
Systems and Development. Landham, Maryland: University Press of
America, 1980.

 

 

243

 

244

Brown, C. A Study of Local Institutions in Kgatleng District, Botswana.
Gaborone: MLGL, 1982.

 

Butler-Flora, C., and M. Tomecek, eds. Selected Proceedings of Kansas
State University's 1984 Farming_Systems Research Symposium,
Farming Systems Research & Extension: Implementation and
Monitoring. Manhattan, Kansas: Kansas State University, 1986.

 

 

 

 

Byerlee, D., and M. Collinson et a1. Planning Technologies Appropriate
to Farmers' Conggpts and Procedures. Mexico: CIMMYT, 1980.

 

 

Byerlee, D., L. Harrington, and D. Winkelmann. "Farming Systems
Research: Issues in Research Strategy and Technology Design."
American Journal of Agricultural Economics 64: 897-904, 1982.

 

Carl Bro International. An Evaluation of Livestock Management and
Production in Boswana with Special Reference to Communal Areas
(3 volumes). Gaborone: MOA, 1982.

 

 

Central Statistics Office. The Rural Income Distribution Survey in
Botswana 1974/75. Gaborone: MFDP, 1976.

 

 

Central Statistics Office. Migration in Botswana: Patterns, Causes
and Consequences (Three Volumes). Gaborone: MFDP, 1982.

 

 

Central Statistics Office. 1983 Agricultural Survey. Gaborone:
MFDP, 1984.

 

Chambers, R., and B. Ghildyal. "Agricultural Research for Resource-Poor
Farmers: The Farmer-First-and—Last Model." Agricultural
Administration 20: 1-30, 1985.

 

 

Chernichovsky, D., R. Lucas, E. Mueller. The Household Economy of Rural

 

Botswana: An African Caseig World Bank Staff Working Papers No. 715.
Washington, D.C.: Thw World Bank, 1985.

 

Collinson, M. "The Evaluation of Innovations for Peasant Farming."
Eastern Africa Journal of Rural Development 1: 50-59, 1968.

 

Collinson, M. Farm Management in Peasant Agriculture: A Handbook for
Rural Development Planningfiin Africa. New York: Praeger, 1972.

 

 

Collinson, M. Farmi g Systems Research in Eastern Africa: The
Experience of CIMMYT and Some National Agricultural Research
Services, 1976-8l;, MSU International Deve10pment Paper No. 1.
East Lansing: Michigan State University, 1982.

 

 

 

Curtis, D. "The Social Organisation of Ploughing." Botswana Notes
and Records 4: 67-80, 1972.

 

 

Dahl, H. "Economic and Social Deve10pment in Botswana 1966-78." In
C. Harvey, ed., pp. 1-13, 1981.

245

Dent, J., and J. Anderson. "Systems, Management and Agriculture." In
Systems Analysis in Agriculture Management, J. Dent and J.
Anderson, eds. New York: John Wiley, pp. 3-13, 1971.

 

Department of Agricultural Research. Agricultural Research in Botswana
1960-68. Gaborone: MOA, 1969.

 

Dillon, J. "The Economics of Systems Research." Agricultural Systems
1: 5-21, 1976.

 

Division of Planning and Statistics. Preliminary Investigation into
the Marketing Of Crops and Livestock in Botswana. Gaborone:
MOA, 1971.

 

 

Dryland Farming Research Scheme. Initial Scientific Report. Phase II.
Gaborone: MOA, 1977.

 

Duggan, W. Informal Markets, Technology and Employment on Arable Land
in Botswana; ALDEP Employment Study Paper 2. Gaborone: MOA.

 

 

Dyer, T., and P. Tyson. "Estimating Above and Below Normal Rainfall
Periods over South Africa 1972-2000." Journal of Applied
Meteorology, 16: 145-147, 1977.

 

 

Economic Survey Mission. The Development Of the Bechuanaland Economyp
Report of the Ministry of Overseas Development Economic Survey
Mission. Gaborone: Government Printer, 1965.

 

 

Eicher, C. "Food Security Research Priorities in Sub-Saharan Africa."
Paper presented at OAU/STRC/SAFGRAD International Drought Symposium,
Kenyatta Interantional Center, Nairobi, Kenya, May, 1986.

Eicher, C., and D. Baker. Research on Agricultural Development in
Sub-Saharan Africa: A Critical Survey; MSU International
Deve10pment Paper No. 1. East Lansing: Michigan State University,
1982.

 

 

Evaluation of Farming Systems and Agricultural Implements Project.
Report No. 1, 1976-77. Gaborone: MOA, 1978.

 

Evaluation of Farming Systems and Agricultural Implements Project.
Report No. 2, 1977-78. Gaborone: MOA, 1979.

 

Evaluation of Farming Systems and Agricultural Implements Project.
Report No. 7, 1982-83. Gaborone: MOA, 1983.

 

Feldstein, H., and S. Poats. "Case Studies for FSR/E Training: Concepts
and Format." Paper presented at Symposium on Farming Systems
Research and Extension: Management and Methodology, Kansas State
University, Manhattan, Kansas, October, 1985.

246

Food and Agriculture Organization. A Study of Constraints on
Agricultural Deve10pment in the Republic of Botswana (Including an
Assessment of the Role of Food Aid). Gaborone: MOA, 1974.

 

 

 

Fortmann, L. Women's Agriculture in a Cattle Economy; Gaborone:
MOA, 1981.

 

Fortmann, L. "Factors Affecting Agricultural and Other Rural Extension
Services in Botswana." Agricultural Administration 18: 13-23, 1985.

 

Frankenberger, T. "Adding a Food Consumption Perspective to Farming
Systems Research." In C. Butler-Flora and M. Tomecek, eds., 1986.

Gaosegelwe, P., G. Khumalo, S. Maphanyane, S. Masikara, Y. Merafe, {-
H. Sigwele. Agricultural Deve10pment Ngamiland Phase I Report
(1981/82). Gaborone: MOA, 1983.

 

 

Gibbon, D. Dryland Crop Production Systems in Semi-Arid Botswana: Their
Limitations and Potential for Improvement. Insitute of]
Deve10pment Studies Reprint. Norwich, U.K.: University of East
Anglia, no date. C

 

 

 

Gibbon, D., J. Harvey, and K. Hubbard. "A Minimum Tillage System for
Botswana." World Crops 26: 229-34, 1974.

 

Gibbs, C. "Agricultural Systems Research in Asia: A Comparative
Discussion of Human Ecology Agroecosystems Research, Farming
Systems Research, and CrOpping Systems Research." Paper presented
at 2nd SUAN-EAPI Regional Research Symposium, Baguio City,
Philippines, March, 1985.

Gilbert, E., D. Norman, F. Winch. Farming Systems Research: A Critical
Appraisal; MSU Rural Deve10pment Paper NO. 6. East Lansing:
Michigan State University, 1980.

 

 

Gulbrandsen. O. Agro-Pastoral Production and Communal Land Use: A
Socio-Economic Study of the Bangwaketse . Gaborone: MOA, 1980

 

 

Hamilton, A. Review of Post Harvest Technologies: Botswana .
Gaborone: MOA, 1975.

 

Harvey, C., editor. Papers on the Economy of Botswana . London:
Heinemann, 1981.

 

Hertel, T. The System of Mafisa and the Highly Dependent Agricultural
Sector; Rural Sociology Report Series Number 11. Gaborone:
MOA, 1977.

 

Hildebrand, P., and F. Poey. On-Farm Agronomic Trials in Farmiog
Syetems Research and Extension. Boulder: Lynne Rienner
Publishers, 1985.

 

 

247

Hill, P. "The Myth of the Amorphous Peasantry: A Northern Nigerian Case
Study." Nigerian Journal of Economic and Social Studies 10:
239-261, 1968.

 

Holm, J., and M. Cohen. "Enhancing Equity in the Midst of Drought: The
Botswana Approach." Ceres 19: 20-24, 1986.

Johnson, A. "Individuality and Experimentation in Traditional
Agriculture." Human Ecology 1: 149-159, 1972.

 

Johnson, G. "Handling Problems of Risk and Uncertainty in Farm
Management Analysis." Journal of Farm Economics 34: 807-817, 1952.~

 

Johnson, G. "Agricultural Economics, Production Economics and the Field
of Farm Management." Journal of Farm Economics 39(2): 441-450,
1957.

 

Johnson, G. "Small Farms in a Changing World." In Proceedings of
Kansas State University's 1981 Farming‘Systems Research Symposium;
Small Farms in a Changing World: Proepects for the Eighties;
Farming Systems Research paper No. 2, W. Sheppard, editor.
Manhattan: Kansas State University, 1982.

 

 

Johnson, G. "Holistic Modeling of Multidisciplinary Subject Matter and
Problematic Domains." Unpublished manuscript, 1983.

Johnson, G. Research Methodology for Economists; Philosophy and
Practice. New York: Macmillan, 1986.

 

Lever, B. Agricultural Extension in Botswana;_ Development Study No.
7. Reading: University of Reading, 1970.

 

Lewis Jr., S. "The Potential Problems of Diamond-Dependent
Development." In C. Harvey, ed., pp. 14-28, 1981.

Lightfoot, C., editor. "A Collection of Papers Relevant to Farming
Systems Research in Botswana," Gaborone: MOA (mimeo), 1983.

Lipton, M. Employment and Labour Use in Botswana, Final Report
(2 volumes). Gaborone: MFDP, 1978.

 

Litschauer, J. Five Year Training Plan 1980-1985 . Gaborone: MOA, 1980.

 

Livingstone, I., and R. Srivastava. Poverty in the Midst of Plenty:
Problem of Creating Incomes and Employment in Botswana;, Occasional
Paper No. 8. Gaborone: Institute of Deve10pment Management, 1980.

 

 

Lucas, R. "The Distribution and Efficiency of Crop Production in Tribal
Areas of Botswana." In Chernichovsky, Lucas and Mueller, 1985.

McKee, C. "Methodological Challenges in Analyzing the Household in
Farming Systems Research: Intra-Household Resource Allocation."
In Butler-Flora, ed., 1984.

 

248

Manetsch, T., and G. Park. Systems Analysis and Simulation with
Applications to Economic and Social Systems, Part I (Third
Edition). East Lansing: Michigan State University, 1979.

 

 

Manzardo, A. Planning for Local Institutions Development in the
CFDA's of Botswana. Gaborone: MLGL, 1982.

 

 

Ministry of Finance and Deve10pment Planning. National Development
Plan 1968-73. Gaborone, MFDP, 1968.

 

 

Ministry of Finance and Deve10pment Planning. National Development
Plan 1970-75. Gaborone, MFDP, 1970.

 

 

Ministry of Finance and Development Planning. National Development
Plan 1973-78. Gaborone, MFDP, 1973.

 

 

Ministry of Finance and Deve10pment Planning. National Development
Plan 1985-91. Gaborone, MFDP, 1985.

 

 

Moock, J., ed. Understanding Africa's Rural Households and Farmipg
Systems. Boulder: Westview Press, 1986.

 

Mueller, E. "The Value of Allocation of Time in Rural Botswana."
In Chernichovsky, Lucas and Mueller, 1985.

Norman, D. The Farming System's Approach: Relevancy for the Small

Farmer; MSU Rural Deve10pment Paper NO. 5. East Lansing: Michigan

State University, 1980.

 

Norman, D., and D. Baker. "Components of Farming Systems Research: FSR
Credibility and Experiences in Botswana." Paper presented at ADC
conference Intra-Household Processes and Farming Systems Analysis,
Bellagio, Italy, March, 1984.

Norman, D., and M. Collinson. "Farming Systems Approach to Research in
Theory and Practice." Paper presented at WorkshOp on Farming
Systems Research (FSR), Hawkesbury Agricultural College, Sydney,
Australia, May, 1985.

Oland, K., H. Alverson, R. Cummings. Targets for Agricultural
Development in Botswana. Gaborone: MOA, 1980.

 

 

Pike, J. Rainfall and Evaporation in Botswanei, Surveys and Training
for the Development of Water Resources and Agricultural Production,
Technical Document No. l. Gaborone: UNDP/FAO, 1971.

 

 

Pingali, P., Y. Bigot, H. Binswanger. ‘Agricultural Mechanization and
the Evolution of Farming Systems in Sub-Saharan Africa.
Baltimore: Johns H0pkins University Press, 1987.

 

Rhoades, R., and R. Booth. "Farmer-Back-to Farmer: A Model for
Generating Acceptable Agricultural Technology." ,Agricultural

 

249

Administration 11: 127-137, 1982.

 

Riley, J., and J. Staatz. Food Systems Organization and Management
Problems in Developing Countries; ADC Report No. 23. New York:
ADC, 1981.

 

 

Rocheleau, D. "Criteria for Re-appraisal and Re-design: Intra-Household
and Between-Household Aspects of FSR&E in Three Kenyan Agroforestry
Projects." In C. Butler-Flora and M. Tomecek, eds., 1986.

Rountree, J. "Systems Thinking - Some Fundamental Aspects."
Agricultural Systems 2: 247-254, 1977.

 

Rural Deve10pment Council. Report on the National Food Strategy.
Gaborone: MFDP, 1985.

 

Schapera, I. Native Land Tenure in the Bechuanaland Protectorate.
Alice, South Africa: Lovedale Press, 1943.

 

Sheppard, C., and D. Clement-Jones. Coping with Drought in Botswana;
Household Strategy and Government Poliey . Gaborone: MOA, 1979.

 

 

Siderius, W. Soils and Eastern and Northern Botswana;g Surveys and
Training for the Deve10pment of Water Resources and Agricultural
Production, Technical Document No. 3. Gaborone: UNDP/FAO, 1972.

 

Spedding, C. An Introduction to Agricultural Systems. London: Applied
Science Publishers, 1979.

 

Syson, L. Income, Expenditure, and Wealth in The Shoshong Area;
Surveys and Training for the Deve10pment of Water Resources and
Agricultural Production, Technical Note No. 31. Gaborone:
UNDP/FAO, 1972.

 

Trent, C., D. Styles and G. Ramolemana. An Assessment of Subject Matter
Competencies of Agricultural Demonstrators in Botswane;§ ATIP
Working Paper No. 6. Gaborone: MOA, 1986.

 

 

Vierich, H., and C. Sheppard. Drought in Rural Botswana: Socio-Economic
Impact and Government Poliey . Gaborone: MOA, 1980.

 

 

Willett, A. Agricultural Group Development in Botswana (3 volumes).
Gaborone: MOA, 1981.

 

Wynne, 8. Local Institutions and Development in Botswane;g Rural
Sociology Report Series No. 18. Gaborone: MOA, 1981.

 

Zufferey, F. A Study of Local Institutions and Resource Management
Inquiry in Eastern Central District. Gaborone: MLGL, 1983a.

 

 

Zufferey, F. A Study of Local Institutions in Mopipi Communal First
Development Area, Central District. Gaborone: MLGL, 1983b.

 

 

"INNER