 

LIBRARY

Michigan State
Unlverslty

 

 

 

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

 

DATE DUE DATE DUE DATE DUE

MAY 200

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1/93 Woes-p14

 

 

THE ECONOM

THE ECONOMICS OF SOIL CONSERVATION INVESTMENTS IN THE
TIGRAY REGION OF ETHIOPIA
By

Berhanu Gebremedhin

A DISSERTATION

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

DOCTOR OF PHILOSOPHY

Department of Agricultural Economics

1998

mtcon'omc

Land degradatio:
problems of the low pro
campaign against land d
ﬁveymrs. However, fa
Mites have not been
We adoption and sus
Ethiopi; one of the reg

The Study anal}:
fan“ Proﬁtability, the le‘
mammom and the fa

“Marion Practices.

ABSTRACT

THE ECONOMICS OF SOIL CONSERVATION INVESTMENTS IN THE
TIGRAY REGION OF ETHIOPIA

By

Berhanu Gebremedhin

Land degradation, mainly in the form of soil erosion, constitutes the basis for the
problems Of the low productivity and sluggish growth of Ethiopian agriculture. A national
campaign against land degradation has been going on in the country for more than twenty-
ﬁve years. However, farmer adoption of conservation practices has been low and adopted
practices have not been sustained. This study seeks to understand the determinants of
farmer adoption and sustained use of conservation practices in the Tigray region of
Ethiopia, one Of the regions where soil erosion is most severe.

The Study analyzes the impact Of stone terraces on wheat and fava bean yields and
farm proﬁtability, the levels and determinants of farmer perception of soil erosion and
conservation, and the factors affecting farmer adoption and intensity of use of
conservation practices. Primary data were collected during 1995/96 crop year from a
sample Of 250 farm households in the Tigray region. Analysis of variance and investment
analysis are used to determine the impact of stone terraces on crop yields and farm
proﬁtability. Ordered probit and probit models are used to identify and estimate the
quantitative impact Of variables affecting farmer perceptions. A double hurdle model is
used to ﬁnd out the factors inﬂuencing adoption and use intensity of conservation

practices.

Stone terraces
proﬁtability of investr
discount rate of 50 pe
soil erosion on his plc
and plot physical char
invest riskiness of im
Charadenstics aﬂ’eae
mum to Stone tem
We. The facu
that ﬁected adOption
°iP0mtnity Cosr Ofla

bum" ImleCaIlOns f0

 

Stone terraces increased crop yields signiﬁcantly and enhanced yield stability. The
proﬁtability of investment in stone terraces dropped almost to break-even at a farmer
discount rate Of 50 percent. An average farmer had 58 percent probability of perceiving
soil erosion on his plots as severe and the perceptions were inﬂuenced chieﬂy by village
and plot physical characteristics. Village and ﬁeld physical factors, household capacity to
invest, riskiness Of investment, socio-institutional and household demographic
characteristics affected farmer adoption of stone terraces and soil bunds. However,
contrary tO stone terraces, investment in soil bunds was made from a low-cost short-term
perspective. The factors inﬂuencing intensity of use of stone terraces differed from those
that aﬂ‘ected adoption. Intensity of use was inﬂuenced by foregone land productivity and
Opportunity cost of labor. Farmers were likely to maintain stone terraces once they were

built. Implications for conservation policy, research and extension are presented.

 

 

 

Copyright by
Berhanu Gebremedhin
l 998

 

 

To my parents Gebremed
went to school but who ho

their children achieve the h

lomytife Eliu Gebrem:
mppon and patience playe

Program and

To my late ﬁiends Berhe
Gebrucmayat Woldu “i

untimeh' death will always

DEDICATED

TO my parents Gebremedhin Gebreegziabiher and Sesen Woldegebriel who never
went to school but who hold ﬁrm belief in modem education and do all they can to help

their children achieve the highest level of education,

To my wife Elfu Gebremariam and our son Samrawi Berhanu whose continuous

support and patience played a Signiﬁcant role in the successful completion Of my graduate

program, and

TO my late friends Berhe Kiros, Yemane Gebreegziabiher, Girmay Gebre and
Gebresemayat Woldu who were always eager to see me achieve my dream, and whose

untimely death will always pain me.

First and foremost.
Dr Scott Swinton my dis:
themmssﬁtl completion t
worked with my adiisor wl
style of writing in the right
members of my committee.

and Dr. Mitiltti Haile

lam thankrm to the

 

ﬁeld work would not have b

neitutional mpport oi \ie‘t

smith tomp‘tetion mi 1

\M. for his support an:

WC Especiahw to my
I l

sou more mdysis w
i was impress

Especially. I would l

ACKNOWLEDGMENTS

First and foremost, I would like to thank Dr. Carl Eicher, my major professor and
Dr. Scott Swinton, my dissertation advisor, whose support and guidance was crucial to
the successful completion Of my graduate program. I feel proud and privileged to have
worked with my advisor who was always available for me, and who effectively shaped my
style Of writing in the right direction. I would also like to express my gratitude to other
members of my committee, Dr. Eric Crawford, Dr. John Strauss, Dr. Thomas Reardon
and Dr. Mitiku Haile.

I am thankﬁrl to the Rockefeller Foundation without whose ﬁnancial support my
ﬁeld work would not have been possible. I would also like to acknowledge the
institutional support ofMekelle University College (MUC), which was crucial to the
successful completion my ﬁeld work. I am especially grateful to Dr. Mitiku Haile, dean of
MUC, for his support and encouragement. My recognition goes to the faculty and Staff of
MUC, especially tO my colleague Fassil, whose contribution in enumerator training and
soil texture analysis was signiﬁcant.

I was impressed by the cooperation I got from ofﬁcials at different levels in Tigray.
Especially, I would like to thank Berhane Haile, head of the Regional Bureau of
Agriculture, and Belete and Gebru Teka, from the then Regional Bureau of Natural
Resource Development and Environmental Protection. My recognition also goes to the
district level administrators, agricultural experts, development agents and farmers in the

study area, The diligence and hard work of the enumerators is greatly acknowledged.

vi

lam grateful to m}
dataentry. My recognitior

tieDevelopment Bank of E

 

Zenebe. expert at the Susta.

project Aderajew Haddusl:

 

ot'Agricultural Research .\l
forit.

My family and l wot;
Swintori Dr. Julie Howard,

crumbly at home durinr:

I am grateﬁil to my late ﬁiend Gebresemayat Woldu, who contributed a lot during
data entry. My recognition also goes to Fray, head Of the Tigray regional branch office of
the Development Bank Of Ethiopia; Yibabe Tilahun, expert at the Relief Society of Tigray;
Zenebe, expert at the Sustainable Agriculture and Environmental Rehabilitation Of Tigray
project; Aderajew Haddush, Amare Belay and Gebremedhin Woldewahid Of the Institute
of Agricultural Research, Mekelle center who readily provided support whenever I asked
for it.

My family and I would like to express our appreciation to Dr. Scott and Sylvia
Swinton, Dr. Julie Howard, and Gordon and Patricia Decker who made us feel

comfortably at home during our stay at Michigan State University.

vii

LIST OF TABLES

 

 

 

CHAPTER l

DRODUCTION . . . .
1.1 Background .
12 Problem Stater
1-3 Research Obie
1'4 organization c

TABLE OF CONTENTS

Page
LIST OF TABLES XII
CHAPTER 1
INTRODUCTION .................................................... 1
1. 1 Background ................................................. l
1.2 Problem Statement and Justiﬁcation of the Study ..................... 3
1.3 Research Objectives .......................................... S
1.4 Organization Of the Dissertation .................................. 5
CHAPTER 2
REVIEW OF THE LITERATURE ........................................ 7
2.1 Land Degradation and Conservation in Developing Countries ............ 7
2.1.1 Causes Of Land Degradation ............................. 8
2.1.2 Soil Erosion: The Major Form of Environmental Degradation in
Developing Countries ................................ 10
2.2 Conservation Policies and Programs: Lessons From Past Eﬁ‘orts ........ 12
2.2.1 Past Soil Conservation Strategies in Developing countries ...... 12
2.2.2 Towards a New Approach for Soil Conservation in Developing
Countries ......................................... 16
2.3 Farmer Perceptions of the Problem of Soil Erosion and Technology Attributes
of Conservation Practices ................................... 18
2.3.1 Farmer Perceptions of Soil Erosion ....................... 18
...................................................... 18
2.3.2 Farmer Perceptions of the Technology Attributes of Conservation
Practices .......................................... 20
2.4 Factors that Affect Conservation Behavior of Farmers ................ 21
2.4.1 Research on Factors Affecting Conservation Adoption ......... 21
2.4.2 Empirical Models in Adoption Research .................... 23
2.4.3 Dependent and Independent Variables in Conservation Adoption
Research .......................................... 25
2.4.4 Results of Past Research on Factors Affecting Conservation
Adoption ......................................... 27
CHAPTER 3
LAND DEGRADATION AND CONSERVATION EFFORTS IN ETHIOPIA AND THE
TIGRAY REGION ................................................... 30
3.1 Land Degradation and Conservation in Ethiopia ..................... 30

viii

 

 

3.2.3 Crop
3.2.4 Lives
3.2.5 Natu
3.2.6 Rura
3.3 Summary . _ _

CHAPTEIM

lhhhll‘iG SYSTEMS (
4-1 General Chm
4-2 Land Tenure

-3 Libor use . ,
4-4 cm and Lt
4'5 FWTS’ Pet

tervem

3.1.1 Extent and Form of Land Degradation in Ethiopia ............ 30

3.1.2 Physical Factors of Land Degradation ..................... 34
3.1.3 Agricultural Practices as Factors Of Land Degradation ......... 35
3.1.4 The Policy Factor .................................... 36
3.1.5 Conservation Efforts in Ethiopia ......................... 38
3.2 Society, Agriculture, and Natural Resources in the Tigray Region ....... 41
3.2.1 Topography and Climate ............................... 42
3.2.2 People and Society ................................... 44
3.2.3 Crop Production ..................................... 49
3.2.4 Livestock production .................................. 59
3.2.5 Natural Resource Degradation and Conservation ............. 61
3.2.6 Rural Institutions ..................................... 70
3.3 Summary .................................................. 77
CHAPTER 4
FARMING SYSTEMS CHARACTERISTICS OF THE STUDY AREA .......... 79
4.1 General Characteristics of the Surveyed Villages .................... 79
4.2 Land Tenure ................................................ 81
4.3 Labor use .................................................. 88
4.4 Crop and Livestock Production ................................. 93
4.5 Farmers’ .Perceptions of Development Constraints and Areas of Institutional
Intervention ............................................ 105
4.6 Farmer Involvement in Soil and Water Conservation ................ 106
4.7 Summary ................................................. 110
CHAPTER 5
THEORETICAL MODEL, EMPIRICAL MODELS AND RESEARCH DESIGN . . 112
5.1 Investment in Soil Conservation ................................ 112
5.2 Conceptual Model .......................................... 114
5.3 Empirical Models and Hypotheses .............................. 118
5.3.1 Effect of Stone Terraces on Crop Yields .................. 118
5.3.2 Adoption of Conservation Practices ...................... 118
5.4 Explanatory Variables ....................................... 123
5.5 Sampling and Data Collection ................................. 131
CHAPTER 6

EFFECTS OF STONE TERRACES ON CROP YIELDS AND FARM
PROFITABILITY: A CASE STUDY IN DEGUA TEMBEN, CENTRAL TIGRAY 133
6.1 An Overview of Research on the Effects of Conservation Practices on Crop

Yields ................................................. 133
6.2 The Study Area ............................................ 135
6.3 Data and Research Design .................................... 136
6.4 Analytical Methods ......................................... 137

ix

 

6.4.1 A:
6.4.2 RE
6.4.3 I-
6.4.4 Pr
6.5 Results and
6.51 E
6.5.2 Pr
6.6 Summary _

CHAPTER7
DETERMINANTS Ol
EROSION AND ATT
7.1. Speciﬁcap
7.1.1 B
7.1.2 0
7‘2 Empirical .‘
7.21 1»

7.2.2 F

7.23 E
7.3. Empirical

6.4.1 Analysis of variance (ANOVA) ......................... 137

6.4.2 Regression analysis .................................. 138
6.4.3 T-test for difference in means ........................... 139
6.4.4 Proﬁtability analysis .................................. 139
6.5 Results and Discussion ....................................... 140
6.5.1 Effects on Crop Yields ................................ 140
6.5.2 Proﬁtability ........................................ 146
6.6 Summary ................................................. 153
CHAPTER 7
DETERMINANTS OF FARMERS’ PERCEPTIONS OF THE PROBLEM OF SOIL
EROSION AND ATTITUDES TOWARDS SOIL CONSERVATION ........... 155
7.1. Speciﬁcation of Empirical Models ............................. 155
7.1.1 Binary Dependent Variable Models ...................... 155
7.1.2 Ordinal Dependent Variable Models ...................... 159
7.2 Empirical Models ........................................... 160
7.2.1 Models Of Farmer Perception of Soil Erosion and Conservation
................................................ 160
7.2.2 Farmer Preference Between Own and Public Campaign Conservation
Work Models .................................... 162
7.2.3 Econometric Model Tests ............................. 164
7.3. Empirical Results and Discussion ....... ' ....................... 165
7.3.1 Farmer Perceptions Of Severity of Erosion ................. 165
7.3.2 Farmer Perceptions of the Yield Effect of Soil Erosion ........ 169
7.3.3 Farmer Perception of the Effectiveness of Soil Conservation
Practices in Reducing Soil Erosion ..................... 173
7.3.4. Farmer Preference Between Own and Public Campaign Conservation
Work ........................................... 176
7.4 Summary ................................................. 180
CHAPTER 8
FACTORS AFFECTING FARMER ADOPTION AND INTENSITY OF USE OF SOIL
CONSERVATION PRACTICES ....................................... 181
8.1 Speciﬁcation of Empirical Models .............................. 181
8.1.1 The Double Hurdle Model ............................. 182
8.2 Empirical Models ........................................... 183
8.2.1 Adoption Models for Physical Conservation Practices ........ 183
8.2.2 Intensity of Use Models of physical Conservation Practices . . . . 184
8.2.3 Adoption Models for Biological Conservation Practices and
Maintenance of Stone Terraces ................... 185
8.2.4 Tests Of the Econometric Models ........................ 186
8.3 Empirical Results and Discussion ............................... 187
8.3.1 Adoption of Stone Terraces and Soil Bunds ............. 187

 

 

 

  

CHAPTER 9
Shh-DIARY, CONCLL'S
9.1 Introduction

8.3.2 Intensity of Use Of Stone Terraces and Soil Bunds ........... 195
8.3.3 Farmer Adoption of Biological Practices and Maintenance of Stone

Terraces .................................. . ....... 199
8.4 Summary ................................................. 204
CHAPTER 9
SUMMARY, CONCLUSIONS AND IMPLICATIONS ...................... 205
9.1 Introduction ............................................... 205
9.1.1 Problem Statement ................................... 205
9.1.2 Research Objectives .................................. 206
9.1.3 Conservation Adoption Studies: An Overview .............. 207
9.2 Summary of Results ......................................... 209
9.2.1 Farming Systems of the Study Area ...................... 209
9.2.2 Impact of Stone Terraces on CrOp Yields and Farm Proﬁtability 210
9.2.3 Farmer Perceptions of Soil Erosion and Conservation ........ 211
9.2.4 Determinants of Adoption and Intensity Of Use of Conservation
practices ......................................... 213
9.3 Policy Implications .......................................... 215
9.3.1 Implications for Conservation Policy ..................... 216
9.3.2 Implications for Agricultural Extension ................... 218
9.3.3 Implications for Conservation Research ................... 219
9.4 Limitations Of the Study ...................................... 221
APPENDIX I ...................................................... 223
REFERENCES ..................................................... 245

xi

 

 

D

 

 

Table 3.1:
Table 32:
Table 3.3:
Table 3.4:
Table 3.5:
Table 3,6;
Table 37
Table 4,1,
Table 42
Tillie 43
Table 4.4:
Table 4.5:
T1916 4.6:
Table 4.7:

Table 4.8:
Table 4.9;

Ttble 4, 10:
Table 4,1 1.

Table 4.12:

Table 4.13;
Iable 5.1.

We 5.2:
We 6.1:

We 6.2:

Estimated
National ar
Role of Ho
Distributior
(1993-95 a
Three Wide
(1993-1995
Yield (100‘,
(l993~95 3
Soil and \T‘
Campaign ‘
Rental Fon
Definition :
Probit Mot
(1991151 sta-
Perceived '
Availability
Household
PrObii MOT
(robust Sta
Improved
PTObTT Mo.
(IObUSI STE
PIObit MO
(IObust Sla
Pliobll MO
(I0bust 512
De‘V'CIOPm
ASSTSIanCe
alTner 111‘
Sir Aiiitt
eﬁmTTOn
eﬁnliion

 

Table 3.1:
Table 3.2:
Table 3.3:
Table 3.4:

Table 3.5:
Table 3.6:
Table 3.7:
Table 4.1.
Table 4.2:
Table 4.3:
Table 4.4:
Table 4.5:
Table 4.6:
Table 4.7:

Table 4.8:
Table 4.9:

Table 4.10:
Table 4.11:
Table 4.12:
Table 4.13:
Table 4.14:
Table 5.1:
Table 5.2:
Table 6.1:

Table 6.2:

LIST OF TABLES
Page

Estimated Soil Loss Rates in Ethiopia by Land Cover .............. 32
National and Regional Rainfall Averages and Variability (1961-87) . . . . 44
Role Of Household Members in Decision Making in Central Tigray . . . . 49
Distribution of Area Coverage of Cereal Crops in Tigray

(1993-95 average) ........................................ 51
Three Widely Grown Cereal Crops in Tigray, by Zone

(1993-1995 average) ...................................... 52
Yield (100kg/ha) of Cereal Crops, Oil Seeds and Pulses in Tigray

(1993-95 average) ........................................ 53
Soil and Water Conservation Accomplishments Through Public

Campaign Work in Tigray, 1992-1995 ......................... 68
Rental Form for Land Leasing and Share Crop Ratio .............. 82
Deﬁnition and Measurement of Explanatory Variables ............. 84
Probit Model Predicting Land Leasing Practice of Farmers

(robust standard errors) .................................... 86
Perceived Tenure Security Among 243 Interviewed Farmers, 1995-96 . 87
Availability Of Hired Labor as Rated by Respondents .............. 88
Households’ Oﬁlfarm Involvement by Type Of Work .............. 89
Probit Model Predicting Participation in Off-farm Employment

(robust standard errors) .................................... 92
Improved Seed: Sources and Reasons for Not Using .............. 94
Probit Model Predicting Household Self-Sufficiency in Local Seed

(robust standard errors) .................................... 97
Probit Model Predicting Household use of Commercial Fertilizer

(robust standard errors) ................................... 100
Probit Model Predicting Household Self-Sufﬁciency in Draﬁ Power
(robust standard errors) ................................... 104
Development Constraints and Priorities for Government

Assistance in Agriculture (farmers’ rating) ..................... 105
Farmer Investment in Conservation Practices ................... 107
Risk Attitude of Farmers Towards Conservation Practices ......... 109
Deﬁnition and Measurement of Dependent Variables ............. 126
Deﬁnition and Measurement of Explanatory Variables ............ 128
Mean Wheat Grain and Straw Yield Under the Differen Treatments

100 kg/ha (standard deviation) .............................. 141

Regression Coefﬁcients for Wheat Yield Differences at 5% Adjustment
(stdandard errors) ........................................ 142

xii

 

 

Table 6.3:
Table 6.4:
Table 6.5:
Table 6.6:

Table 7.1:
Table 7.2:

Table 7.3:
Table 7.4:

Table bl;
Table 8.2:

Table 3.3;

 

Mean Gr:
Treatmer.
Regressio
(standard
Summary
and Fave
Cash Flo
Terraces (
Regressiot
Regressior
Soil Erosi
Regressio
Conserva
RCgressic
I)l'Obll Rt
Probit ar
Of Stone
Probit R
Practice

Table 6.3:
Table 6.4:
Table 6.5:
Table 6.6:

Table 7.1:
Table 7.2:

Table 7.3:
Table 7.4:

Table 8.1:
Table 8.2:

Table 8.3:

Mean Grain and Straw Yields of Fava Bean Under Different

Treatments, 100 kg/ha (standard deviation) .................... 144
Regression Coefficients for Fava Bean Yield Differences at 5% Adjustment
(standard error) ......................................... 145
Summary of Incremental Cost and Beneﬁts of Stone Terraces on Wheat
and Fava Bean Fields (in Birr) .............................. 149
Cash Flow for Wheat-Wheat-Fava Bean Cropping System Under Stone
Terraces (in Birr) ........................................ 152

Regression Results for Farmer Perception of Severity of Soil Erosion . 168
Regression Results for Farmer Perceptions of Yield Reduction Effect of

Soil Erosion ............................................ 172
Regression Results for Farmer Perception of Effectiveness of

Conservation Practices .................................... 175
Regression Results for Farmer Preferences Between Own and Public
Campaign Conservation Work .............................. 179

Probit Regression Results for Adoption of Stone Terraces and Soil Bund94
Probit and Truncated Regression Results for Adoption and Intensity of Use
of Stone Terraces (std. err.) ................................ 198
Probit Regression Results for Adoption of Biological Conservation

Practices and Maintenance of Stone Terraces ................... 203

xiii

 

1.1 Background
Land degrade“
declining agriculture
of agricultural and fl
some time now. hit
from rapid populatit
Due to recur
productivity of the
considerably. lnsti
environmental prot

ASittilture (110A

CHAPTER 1

INTRODUCTION

1.1 Background

Land degradation, mainly in the form of soil erosion, has been a major cause of
declining agricultural productivity in Ethiopia. Despite its potential to be a net exporter
of agricultural and food commodities, Ethiopia has been dependent on food imports for
some time now. Moreover, this second most populous country in Africa is reeling
from rapid population growth.

Due to recurrent drought and famine, which is partly a result Of declining
productivity of the land due to soil erosion, environmental awareness has grown
considerably. Institutions have been modiﬁed to give greater attention to
environmental problems. The main focus Of the extension work of the Ministry of
Agriculture (MOA) has been on soil conservation practices. The Food-For-Work
(FFW) Program has been used in mobilizing rural labor to build terraces and soil
bunds, and afforest hillsides. A National Conservation Strategy (N CS) was adopted in
1989 to coordinate all natural resource related programs. In 1990, the Ministry of
Natural Resource Development and Environ mental Protection (MONRDEP) was created
which Operated until 1996 when it was again merged with the Ministry of Agriculture
(MOA).

However, the efforts Of the government and NGOs have been minuscule

compared with the magnitude of the problem. Many conservation practices that have

 

 

in

 

2

been adopted have not been sustained by farmers. In fact, there is growing evidence
that soil erosion is accelerating, particularly in the northern part of the country. The
physical conservation structures, stone terraces and soil bunds, have encountered
technical problems that have led to a reduction in cultivated area and increased
incidence Of pests. In Short, the situation can be described as being characterized by a
combination of inadequate resources for an extensive problem, faulty approaches, and
inadequate adoption and sustained use of conservation practices.

Soil conservation policy in Ethiopia needs to be re-examined. But new
approaches can only succeed if they are based on a thorough understanding Of existing
farming systems and how previous conservation policies have performed. Speciﬁcally,
four questions need tO be addressed:

a) What are the principal determinants of soil erosion in the Ethiopian farming

systems?

b) What soil erosion control techniques have been recommended by the

government? '

c) Why are recommended soil conservation techniques not being adopted or

sustained by farmers?

d) How should the current conservation interventions be modiﬁed to increase

the adoption and sustained use of conservation practices?

 

 

O

 

 

 

12 Problem Statement

The highlandsl it
88 percent of the popula
land area was taken out (

third of the highlands ha‘

 

estimated to need soil co:
erosion is estimated to re

by another 1 percent per
notations grasses which
livestock productivity.
Soil erosion is m
Mateo practices. Sr
ho shonn that crop Ian
mnhihutor to soil erosi

m.
Unrated that soil loss .

I 80 Percc

an .

 

3

1.2 Problem Statement and Justiﬁcation of the Study

The highlands‘ in Ethiopia occupy about 43 percent of the country and support
88 percent of the population. In the 19803 about 3.7 percent (2 million hectares) of the
land area was taken out of production because of soil erosion (Wood, 1989). Nearly a
third of the highlands have slopes exceeding 30 percent and about three quarters are
estimated to need soil conservation measures in order to support cultivation. Soil
erosion is estimated to reduce yields by 1 percent per year and biological degradation
by another 1 percent per year (Stahl, 1990). Moreover, eroded lands grow less
nutritious grasses which means poor grazing for livestock and consequently low
livestock productivity.

Soil erosion is more severe in cultivated lands because of single cropping and
associated practices. Small mded crops require ﬁne tilth seed beds. Dejene (1990)
has shown that crop land covers about 16.3 million hectares and is the largest
contributor to soil erosion in Ethiopia. Hans Hurni as quoted by Dejene (1990)
estimated that soil loss on cultivated land is four to ten times higher than on grazing
land and about 80 percent of the annual soil loss occurs during the months of plowing
' and in the ﬁrst month after planting.

Natural resources in Ethiopia are the foundation for increasing the rural standard
of living. However, land resources continue to be degraded. Wood (1989) argues that

unless land use changes are made so that erosion is slowed, 10 million hectares (10

 

1Highlands are areas which have altitude of 1500 meters above sea level or more.

‘I

 

percent of the crop land
lives of 10 million peep
The Tigray regic
Ethiopia. Tigray is locat
80,000 square km with a
severely affected by recu
30percent of the populat
arable land lies on the pl
W61 in ash. The n
W5, Which account f

”Elm grains are tefF, b

““3 Wthum. millet a

4

percent of the crop land) will be taken out of cultivation by 2010 which will affect the
lives of 10 million people.

The Tigray region is one of the most seriously degraded environments in
Ethiopia. Tigray is located in the northern part of the country and it covers more than
80,000 square km with a population of about 3.2 million. It is the region that had been
severely affected by recurrent drought and declining agricultural productivity. About
80 percent of the population lives in the highlands and depends on agriculture. Most
arable land lies on the plateau, on altitudes ranging from 1000 to 3400 meters above
sea level (m asl). The intensity of cropping is once a year. The predominant crops are
cereals, which account for 70-75 percent of the cultivated land. High on the plateau,
major grains are teffz, barley and wheat while on the lower plateau and in the lowlands
white sorghum, millet and maize predominate. Farming practices include plowing with
a pair of oxen and iron tipped wooden plow, sowing by broadcasting seeds, hand
weeding, harvesting by a short handled sickle, and threshing by oxen hooves. Crops
are typically rotated among grains with occasional crops of legumes.

Agriculture in the region is characterized by increasing land pressure,
continuous cultivation of arable land, inadequate manuring and cleared forest lands. As
a result, the land is left barren and highly susceptible to erosion. Although land
degradation is a pressing problem throughout Ethiopia, the areas that are in greatest

need of soil conservation are the northern and eastern highlands (Wood, 1989).

 

2 Tell (Eragmstis abyssinica) is a small seeded grain used as a staple food in Ti gray region and other parts of
Ethiopia. Its cultivation is almost limited to Ethiopia.

 

 

’3

However. there is a lac

collection.

13 Research Objectiv

The general objec

 

 

and economic factors inﬂ
intheTigray region of E:
The speciﬁc obj cc
1, Study t2
study area
2. Estimat
proﬁtabilit
3. Discov:
erosion
4. Identify
household
5. Determj

pracllCes b

5

However, there is a lack of research based on village level studies and primary data

collection.

1.3 Research Objectives
The general objective of this study is to determine the physical, social, institutional
and economic factors inﬂuencing fanners’ decisions to invest in soil conservation practices
in the Tigray region of Ethiopia.
The speciﬁc objectives are to:
1. Study the farm management and farming systems characteristics of the
study area,
2. Estimate the effects of stone terraces on crop yields and farm
proﬁtability.
3. Discover the levels and detemiinants of fanners' perceptions of soil
erosion,
4. Identify the social, economic and technical determinants of farm
household adoption of soil conservation practices,
5. Determine what factors explain the intensity of use of conservation

practices by farm households,

1.4 Organization of the Dissertation
The dissertation is organized in nine chapters. Chapter 2 reviews the literature on

farmers’ adoption of soil conservation practices, policy instruments used by governments,

 

and approaches and strat-
the conceptual model ant
Chapter 4 present

inEthiopia and the Tigraj

 

that have bearing on land
be are discussed. Chap
area south central Ti gray
inputs and constraints on
film research on the effec

Dcsoiption of the study 2

analysis results.

Chapter 7 deals \s

problem of soil erosion to

preferences of farmers be

Canpaign work are discu
and Intensity of use of cc

hows conclusions and p

init' ‘-

mention strategies It

hi '
tattoos of the study at

6

and approaches and strategies applied in conservation intervention. Chapter 3 describes
the conceptual model and research methods used for the study.

Chapter 4 presents the background information on the problem of land degradation
in Ethiopia and the Tigray region. It describes the social, political, and physical conditions
that have bearing on land degradation. Agricultural production and the natural resource
base are discussed. Chapter 5 discusses the farming systems characteristics of the study
area, south central Tigray. It presents descriptive statistics on land tenure, agricultural
inputs and constraints on agricultural production. Chapter 6 presents the results of an on-
farm research on the effect of stone terraces on crop yields and farm proﬁtability.
Description of the study area and method of study is followed by statistical and investment
analysis results.

Chapter 7 deals with the levels and determinants of farmer perceptions of the
problem of soil erosion and effects of conservation on crop yields. Determinants of
preferences of farmers between private investment on soil conservation and community
campaign work are discussed. Chapter 8 examines the detemiinants of farmer adoption
and intensity of use of conservation practices. Chapter 9 summarizes the results and
draws conclusions and policy implications for national and regional conservation
intervention strategies. Implications for research and extension are outlined. Finally the

limitations of the study and possible ﬁiture research topics are presented.

 

 

 

2.1 Land Degradation :1

Developing count
facing the triple problems
usetheir natural resource
that hiture growth in agr
than through area eXpan
one of the central object
“33' lint ﬁrture product
Mm“ (Delgado a:

olStlStainable develOpr

CHAPTER 2

REVIEW OF THE LITERATURE

2.1 Land Degradation and Conservation in Developing Countries

Developing countries in general and sub-Saharan Aﬁ'ican countries in particular are
facing the triple problems of how to increase agricultural production, reduce poverty, and
use their natural resources sustainably. High population growth rates in Africa necessitate
that future growth in agriculture will have to come increasingly from yield increases rather
than through area expansion (Eicher, 1994). Natural resource sustainability has become
one of the central objectives of development. Production will have to increase in such a
way that future production capacity of the natural resources is enhanced rather than
diminished (Delgado and Anderson, 1993). These concerns are embodied in the paradigm
of sustainable development which links together population, poverty and the environment.

Environmental degradation has attracted the attention of policy makers,
researchers, and development practitioners for some time now. However, understanding
of the problem of environmental degradation has been difﬁcult due to difﬁculty in
obtaining accurate measurements of the problem, and inseparability of the effects of
natural conditions ﬁ'om that of humans (Anderson and Thampapillai, 1990).
Environmental Economics, concerned with the impact of human activities on the
environment, has developed as a coherent body of knowledge since the 1960's. Its
applications to the conditions of developing countries began around the 1980's (Pearce

and Maler, 1991).

 

I

 

The forms of ens
categories: land degrada
process by which the pr
and weather conditions
dirinution of the produ
erosion deforestation t
alkalization or nutrient

llhile the prob
industrialized countries
developing countries. 7
p l
arranger in develtiptr

billet rate of use of cc

11.1 Causes of Land

 

Lind degradap‘

"hiltoptiate land Use I

and“ 1992). Star

mdllce emimnmemal I

 

8

The forms of environmental degradation can be classiﬁed into three general
categories: land degradation, pollution and loss of biodiversity. Land degradation is a
process by which the productivity of the land is reduced if production inputs, technology
and weather conditions are held constant. In other words, land degradation refers to the
diminution of the productive capacity of the land. This diminution can arise due to soil
erosion, deforestation, deterioration of soil structure, waterlogging, salinization,
alkalization, or nutrient depletion.

While the problems of pollution and loss of biodiversity are more important in the
industrialized countries, the consequences of land degradation are more prevalent in
developing countries. The impact of soil erosion on the productivity of land, for instance,
is stronger in developing countries than in developed countries partly because of the

higher rate of use of commercial inputs in the latter (Stocking, 1988).

2.1.1 Causes of Land Degradation

Land degradation results from inappropriate land use. The real causes of
inappropriate land use are social, economic and institutional rather than technical
(Sanders, 1992). State policies may fail to encourage conservation behavior or even
induce environmental degradation. Subsidies that encourage mechanization may induce
degradation while controlled farm product prices may hinder conservation investments by
lowering the proﬁtability of farming. Public subsidy of commercial fertilizer could boost
production in the short run but can also be a disincentive to farmers to adopt conservation

practices (Barbier, 1990; Anderson and Thampapillai, 1990).

Extraction of
economic base and ca
that in Haiti, rural em
nstematic rating the l
senices National gor
unemployment may de
from in Ethiopia in tht
to land degradation ain
hen a larger scale can
technologies, and lack .

Marginalization

most of the soil erosior.

ttsource poor farmers 1

huts (Anderson and Tr

 

.lhlans into less fertile
tesult, more severe soil .

Market failures
hthtftction of existing

liturance markets in de

 

loot ’ ‘ l
tanners lunder ins-e

Itote likely to be long te I

9

Extraction of economic surplus from the rural population weakens the rural
economic base and can result in degradation of the land. White et al. (1995) have argued
that in Haiti, rural environmental degradation is the consequence of the government’s
systematic taxing the rural population without providing agricultural development
services. National governments in their attempt to solve urban social problems such as
unemployment may devise erosive land development projects. The development of state
farms in Ethiopia in the 1980's is a case in point. Resettlement schemes can be a response
to land degradation aimed at reducing the pressure on land. However, land degradation of
even a larger scale can occur in the newly settled areas due to fragile lands, unproven
technologies, and lack of experience (Anderson and Thampapillai, 1990).

Marginalization of farmers can be another cause of land degradation. The fact that
most of the soil erosion that occurs in developing countries is on land areas operated by
resource poor farmers is an indication of the forced movement of people into marginal
lands (Anderson and Thampapillai, 1990). Colonial settlers in Aﬁica pushed the native
Africans into less fertile lands which put a lot of pressure on the marginal lands. As a
result, more severe soil erosion was seen in native agriculture (Stocking, 1985).

Market failures induce land degradation (Bojo, 1991). Market failures include
imperfection of existing markets, incomplete markets or nonexisting markets. Lack of
insurance markets in developing countries coupled with high discount rate of resource
poor farmers hinder investment in resource conservation measures whose pay offs are

more likely to be long-term. Poorly deﬁned property rights, common in developing

countries result in 6
resources have been

Environment
can also become an it
degradation for insta
major cisilizations in t
(1988) indicated that :

Lanztr99o)

argues th

tween government a

Agriculture is r

Impact on the ensirorirr

‘Hll lr

10

countries, result in extemalities. In these countries, open access and common property
resources have been major sources of environmental degradation.

Environmental degradation is not only the result of social and political factors. It
can also become an important factor that inﬂuences social and political conditions. Land
degradation, for instance, was among the major factors that contributed to the decline of
major civilizations in China, Mesopotamia, Egypt, and Greece (Sanders, 1992). Stocking
(1988) indicated that soil erosion was directly related to a disruption of Nepal’s economy.
Lanz (1996) argues that environmental degradation can weaken the structural linkages
between government and people.

Agriculture is one of the major sectors that causes environmental degradation. Its
impact on the environment has been manifested through four major problems: food
security due to land degradation, food safety due to chemical inputs, environmental quality
due to water and air pollution and loss of biodiversity. In Africa, land degradation has
been at the base of food supply problems (Dejene, 1990). Land degradation can also
incur costs on the national economy through increased food imports, increased

susceptibility to drought, and higher fertilizer imports (Stocking, 1988).

2.1.2 Soil Erosion: The Major Form of Environmental Degradation in Developing
Countries

The soil resource is a complex structure which provides a growing medium for
plants by providing rooting zone, soil moisture and air. It also serves as a source of

nutrients. While nutrient loss can be easily replenished by growing legume crops,

 

manuring or commer.
maintenance of the PI
fenhiw, for instance-
offset other effects 0f
the rate of soil loss ex
deposition), the net 10
resource. Reduced 0'
inﬁltration and increas

Among the for
entironmental problen
loss in land productisit
tobedelayed due to cr
thect crop yields. In th

iterate to other areas

Different metho

held including

(1) rem

g .
indent erosion levels

 

ll

manuring or commercial fertilizer, maintaining the other functions of the soil require
maintenance of the proper topsoil depth and structure. Application of commercial
fertilizer, for instance, can compensate for nutrient loss due to soil erosion but can not
oﬁ‘set other eﬂ‘ects of topsoil loss (Anderson and Thampapillai, 1990). In the event that
the rate of soil loss exceeds the tolerable limit (i.e the rate of soil formation plus
deposition), the net loss in topsoil represents deterioration of the quality of the soil
resource. Reduced organic matter and water retention capacity of the soil, reduced
inﬁltration and increased runoﬁ‘ are all results of soil loss.

Among the forms of land degradation, soil erosion stands out as the major
environmental problem in developing countries. The direct impact of soil erosion is the
loss in land productivity through topsoil loss. Soil erosion can also cause ﬁeld operations
to be delayed due to crusting and compaction of soils. Delayed ﬁeld operations in turn
aﬂ‘ect crop yields. In the extreme, soil erosion can force farmers abandon their land and
migrate to other areas, oﬁen to marginal lands inducing ﬁrrther land degradation.

Different methods have been used to estimate the effect of soil erosion on crop
yield including (1) removal and addition of top soil, (2) comparative study of soils at
different erosion levels, (3) analysis of yield data from plots under varying management
practices and landscape positions, (4) use of factor analysis and application of geostatistics
and (5) simulation techniques (Olson et al., 1994). Time series data on erosion rates and
crop yields have also been used to determine the effect of soil erosion on yield over time
(Weesies et al. 1994). The comparative analyses of crop yields from eroded and non-

eroded lands, or from lands under different management practices or landscape locations,

 

on the other hand, us
soil erosion on yield.
Calculator (EPIC) ar
identify agroecologic
Multivariate :
crop yields while get
Cmp parameters of a
with the relations
yield reducuon dEper

a non-linear relations

22 C°“5°Wation p.
L2.l Past Soil Cons

h‘iOdem 50“ C

12

on the other hand, uses a cross-sectional approach to determine the productivity effect of
soil erosion on yield. Simulation models such as the Erosion Productivity Impact
Calculator (EPIC) and Nitrogen-Tillage Residual Model (NTRM) have been used to
identify agroecological and crop factors that limit productivity under speciﬁc conditions.
Multivariate analysis have been used to identify soil and other variables that aﬂ‘ect
crop yields while geostatistics have been applied to analyze spatial variability of soil and
crop parameters of a given site. Econometric damage functions have been estimated to
specify the relationship between crop yield and soil depth. The effect of soil erosion on
yield reduction depends on the level of erosion that has already taken place. This results in

a non-linear relationship between soil loss and yield reduction (Stocking, 1988).

2.2 Conservation Policies and Programs: Lessons From Past Efforts
2.2.1 Past Soil Conservation Strategies in Developing countries

Modern soil conservation efforts have a history of only about 50 years (Sanders,
1992). In Africa, the necessity of soil conservation was ﬁrst recognized in the 1920's
(Stocking, 1985). State intervention to encourage soil conservation has been direct
through programs and projects or indirect through policy instruments. When conservation
practices are proﬁtable to society but not to individuals, public subsidies are justiﬁable on
economic grounds. Seitz et al. (1979) as quoted in Anderson and Thampapillai (1990)
showed, for the US corn belt, that private net returns with conservation were $50 per
hectare lower than those without conservation. McConnel (1983) applied a dynamic

model to the problem of soil conservation and concluded that it may be rational for

 

 

farmers to let 50” ”Od'
Indonesil need econo
practices. In Kenya di
to use on-farm consen‘
have been widely used
The strategies 3
include direa expenditt
regulatory measures or
management behasiors
encourage soil conserv

SUbSldleS, institutional t

13

farmers to let soil erode. Barbier (1990) showed that farmers in the uplands of Java,
Indonesia, need economic incentives in order change their land use and management
practices. In Kenya, direct cash payments were used as incentives for individual farmers
to use on-farm conservation practices (Admassie, 1992). In Ethiopia, FFW payments
have been widely used for conservation purposes.

The strategies available to governments to mitigate the effects of land degradation
include direct expenditure on public works, provision of education and information, and
regulatory measures or indirect through policy instruments to inﬂuence farmers land
management behaviors (Cary and Wilkinson, 1997; Ervin, 1982). Policy instruments to
encourage soil conservation in developing countries include appropriate production
subsidies, institutional credit for construction of conservation practices, price and tax
incentives, monetary policies that encourage investment in soil conservation, and land and
tree tenure security (Anderson and Thampapillai, 1990).

In addition to public subsidies when conservation practices are proﬁtable to society
but not to individuals , soil conservation practices need to combine the twin objectives of
conserving soil and providing short term real beneﬁts to farmers (Sanders, 1992). The
short term beneﬁts may be in the form of yield increase, or the reduction of risk or
drudgery. While soil erosion can induce suboptimal input use, conservation practices can
push the input mix to optimal (Barbier, 1990). Hence, another way of improving farm
level proﬁtability of conservation practices is by encouraging farmers to grow high value

crops.

 

Past soil conse
on two major assumP’
degradation is primaﬁ
willing and able to im
followed the three 518
dCﬁHlIlOl'l proceeded i
proposed were engine
Inappmpriate land use
involved in the design
to lack adaptability to
cOnsequence of coerci

e to the coercive n;

14

Past soil conservation programs and projects in developing countries were based
on two major assumptions (Norman and Douglas, 1994): (1) the solution to land
degradation is primarily technical, and (2) farmers are aware of land degradation and are
willing and able to invest on it. Hence the classical approach to soil conservation
followed the three steps of problem identiﬁcation, planning, and implementation. Problem
deﬁnition proceeded in terms of soil loss rather than productivity loss, and the solutions
proposed were engineering structures that would keep soil in place (Sanders, 1992).
Inappropriate land use and management were rarely considered, and farmers were not
involved in the design of conservation programs. Such top-down approaches have tended
to lack adaptability to local conditions (Stahl, 1990). An extreme example of the
consequence of coercive top-down approach to soil conservation can be found in Kenya.
Due to the coercive nature of soil conservation strategy during the colonial period, soil
conservation was totally rejected by farmers for nearly a decade after independence; it
took the establishment of a Permanent Presidential Commission on Soil Conservation in
1981 to reinstitute conservation intervention programs (Admassie 1992).

The soil conservation programs and projects promoted by governments and non-
govemment organizations (NGOs) have been expensive in terms of land conserved or soil
retained and government eﬂ‘ort required (Sanders, 1992). Hence these programs have
proven diﬂicult to expand. The mechanical soil conservation practices promoted by such
projects were labor intensive, provided little short term beneﬁts, and were not familiar to
farmers. The practices required external assistance for adoption, and were usually

abandoned after project termination (White et al., 1995).

 

   

Norman and

 

 

15

Norman and Douglas (1994) identiﬁed the reasons for the failure of past soil

conservation programs and projects as follows:

(1) they considered loss of t0p soil separately ﬁom loss of productivity,

(2) they were unable to provide short-term beneﬁts to farmers,

(3) they excluded farmers from the design process, and

(4) they ignored the factors that hinder adoption of conservation practices.

The failures of past conservation projects prompted researchers to investigate
factors associated with successful adoption of practices. Understanding why indigenous
conservation practices have often been abandoned might indicate what needs to be done to
facilitate the adoption of modern practices. After reviewing the performance of soil
conservation efforts in Aﬁ'ica during the colonial era Stocking (1985) concluded the
following:

(1) conservation designs and systems of land use planning need to be simple,

(2) familiar techniques with short term beneﬁts have higher chances of success,

(3) labor demanding techniques have lower chances of success,

(4) local people need to be involved in planning of conservation programs and projects,
and

(5) implementation should be primarily the responsibility of local people.

de la Briere (1996) found that FFW incentives would bring poor farmers into
conservation in the Dominican Republic. But once the FFW program is over, these
farmers do not follow up with the conservation practices due to high opportunity cost of

labor. In their study of Colombian farmers, Ashby et al. (1996) found that involving

 

 

farmers in the design
criteria for acceptabll
short term beneﬁts as
found that farmers in
because the ruin obje
instinrtional incentives
of research extensior

Similarly, mu:
consenation practices
required limited labor ;
a(billable, improved in

Itsources. However, it

 

and developing countri

 

 

16

farmers in the design of soil conservation technologies and consideration of fanners’
criteria for acceptability improved adoption of conservation practices. They identiﬁed
short term beneﬁts as the most important acceptability criterion. Sain and Barreto (1996)
found that farmers in one area of El Salvador successfully adopted conservation practices
because the twin objectives of conservation and productivity were linked by economic and
institutional incentives. The institutional incentive was provided through the coordination
of research, extension and credit institutions.

Similarly, White et al. (1995) in their Haitian study concluded that widely accepted
conservation practices shared the traits that they (1) provided multiple beneﬁts, (2)
required limited labor and low ﬁnancial expenditure, and (3) were simple, locally
adaptable, improved indigenous techniques that could be constructed using local
resources. However, it appears that research and development efforts in both developed
and developing countries have neglected the simultaneous pursuit of productivity and

environmental quality objectives (Ervin and Graffy, 1995).

2.2.2 Towards a New Approach for Soil Conservation in Developing Countries
Lessons from the failures of past conservation programs and projects and the few

success stories have led to the development of a new approach to soil conservation based

on the following principles (Norman and Hudson, 1994):

(1) Since productivity loss rather than soil loss is the main problem to be addressed, soil

conservation must be considered within the overall framework of agricultural

development; improved farming practices should be the basis for conserving soil.

(2) The focus should
(3) Soil conservation
implementation phase:
(4) Soil conservation
(5) Biological or cultu:
(6) The combat agains
emerging research find
This new appro

soil conservation progr
laid users. and deveIOp
del'elOPiUS a Soil conser

Inappropriate land use a

l7

(2) The focus should be on preventing land degradation, not on curing it.

(3) Soil conservation efforts need to involve farmers in the design, planning and
implementation phases.

(4) Soil conservation projects should incorporate short term tangible beneﬁts to farmers
(5) Biological or cultural practices can be more effective than mechanical ones.

(6) The combat against land degradation should be a long term effort supported by
emerging research ﬁndings.

This new approach to soil conservation implies that in order to develop effective
soil conservation programs, every country needs to aim at improving land use, involving
land users, and developing appropriate institutions (Sanders, 1992). The ﬁrst step in
developing a soil conservation strategy should, therefore, be to understand the causes of
inappropriate land use and management.

The choice of policy instruments needs to be based on information derived from
research on 1) the economic status of farmers in the target area; 2) effectiveness, technical
feasibility and proﬁtability of the practices, and 3) factors inﬂuencing their adoption.
Comparison of project-assisted vs control group to analyze the effect of policies and basic
inﬁ'astructure on the adoption of conservation practices may reveal the effectiveness of
particular policy instruments. Since different policies may have contrary effects on soil
conservation, coordination of policies that have bearing on soil conservation becomes
necessary. Moreover, the ﬂow of information from research and the reﬁnement of policy
instruments needs to be a continuous process. In sum, the solution to environmental

problems in developing countries lies in an approach which gives more power to farmers

 

and less to govern

2.3 Farmer Perctp
Conservation Prat
2.3.1 Farmer Perce
An understan
designing conservatic
threat due to soil eros
often overlooked in Si
soil erosion non~econ
conserve soil, and non

Subjective perception c

. (“R

18

and less to government and non-govemment agencies and institutions with the latter

serving as a guide, and source of support and technical knowledge.

2.3 Farmer Perceptions of the Problem of Soil Erosion and Technology Attributes of
Conservation Practices
2.3.1 Farmer Perceptions of Soil Erosion

An understanding of how farmers perceive soil erosion can be more important in
designing conservation programs than the physical quantiﬁcation of soil loss, once the
threat due to soil erosion is widely understood. Hefferman (1982) classiﬁed the factors
often overlooked in soil conservation programs into fanners’ perception of the problem of
soil erosion, non-economic motives for farming, community inﬂuence on farmers to
conserve soil, and norms of good farming. Pender and Kerr (1996) claim that farmers’
subjective perception of soil erosion are more relevant for their decision on conservation
investment than objective measures.

The way farmers perceive erosion problem on their lands versus on others lands
may be different. Ervin and Alexander (1981) found that the percentage of farmers in
Monroe county, Missouri, USA, who felt that soil conservation was a problem in
agricultural land in general was much higher than those who felt the same on their own
lands. The same study showed that farmers believed that there was more erosion on rented
than on owned lands. Moreover, Carlson et. al. (1976) as quoted in Hoover et. al. (1982)
found that farm operators in Palouse area of Northern Idaho and Eastern Washington

believed that erosion on their land was one-half that of their neighbors.

ResearC
perceptions of U
perceptions of If
level, farming ear]
steeper soils inﬂU
Hoover et. al. (1 9
l982, and found tl

explained the highe
al. (1994) investiga

conservation betwer

l9

Researchers have also been interested in knowing what factors inﬂuence fanners’
perceptions of the erosion problem. Gould et. al. (1989) used farmers’ area-wide
perceptions of the problem of soil erosion in a probit analysis. They found that education
level, farming experience, contact with soil conservation services, and proportion of
steeper soils inﬂuenced perceptions positively while land size had a negative inﬂuence.
Hoover et. al. (1982) compared Nebraska fanners’ perception of soil erosion in 1978 and
1982, and found that increased access to information and knowledge of soil erosion
explained the higher perception of the problem of soil erosion found in 1982. Carlson et.
al. (1994) investigated changes in fanners’ attitudes and behaviors towards soil
conservation between 1976 and 1990 in Palouse and Camaras prairies of north-central
Idaho, eastern Washington and northeastern Oregon, some of the areas with the highest
erosion rates in the United States. They found that farmer awareness of the erosion
problem increased and that while socioeconomic variables were important in explaining
adoption of conservation practices in 1976, attitudinal factors were more important in
1990.

Farmers’ estimates of yield loss due to soil erosion can be inaccurate due to the
difﬁculty in isolating the effect of erosion from that of climate and inputs. Nevertheless, it
is important to understand these perceptions because they drive fanners’ decisions on
conservation investment. Farmers’ perceptions need to be interpreted with caution. It
could be that farmers are unaware of the problem because they are so accustomed to it

that they may consider it as a natural outcome of farming. It could also be that they are

anare of the probler
do not consider it as

Studies aime
into adoption and us
as explanatory varia
and intensity of use
and use of conserva
Perceiving an erosic

e-‘fl~‘lains adoption p

Pfrictices

In addition 1

20

aware of the problem but do not want to admit it because it reﬂects on their ability or they
do not consider it as economically important (Hefferman, 1982).

Studies aimed at determining whether or not perceptions are actually translated
into adOption and use of conservation practices have used actual or predicted perception
as explanatory variables. Ervin and Ervin (1982) used actual perceptions in their adoption
and intensity of use equations and found that perception explains positively both adoption
and use of conservation practices. Gould eta] (1989) used predicted probability of
perceiving an erosion problem in their tobit model of adoption and found that perception

explains adoption positively.

2.3.2 Farmer Perceptions of the Technology Attributes of Conservation
Practices

In addition to the understanding of farmers’ perceptions of soil erosion,
perceptions of the technology attributes of conservation practices need to be considered in
conservation programs. Farmers’ criteria for the effectiveness of conservation practices
include factors other than reduction in soil loss, such as feasibility of the practices, ease of
farm operation, effect on yield and proﬁtability. Feasibility of the practices was assumed
given in most studies. Perceived rather than actual proﬁtability of conservation practices
may be important for adoption of conservation practices (Cary and Wilkinson, 1997; Ervin
and Alexander, 1981). This may not be true for intensity of use of the practices, because
once the practices are tested, observed proﬁtability is likely to be more important. New

technologies entail two types of risks: subjective (due to learning) and objective (due to

such factors a
important in r
Ervin
USA were c<
terraces and
and Zinnah (
and found th
111d intensity
vallables fro
However. Si:
inﬂuenced h;

f‘” Simultam

2‘4 FQCIOrs

2.4.1 Rum

ECOn

”Diem level
”allow Or in

influence fan]

is}
Len adopllr

Mi“? i“Strum

0 .

21

such factors as weather conditions etc.), of which the subjective risk may be more
important in determining adoption of innovations (Peder et al., 1985).

Ervin and Alexander (1981) reported that farmers in Monroe county, Missouri,
USA were concerned about the tillage, cultivation and drainage difﬁculties caused by
. terraces and efﬁciency loss was the primary concern regarding contour plowing. Adesina
and Zinnah (1993) studied farmers’ perceptions of improved rice varieties in Sierra Leone
and found that perceived technology attributes explained most of the variation in adoption
and intensity of use. These results indicate that omission of technology perception
variables from adoption studies can result in biased estimates of adoption variables.
However, since perceptions about the technology attributes of the practices can be
inﬂuenced by experience, their use in predicting conservation behavior needs to account

for simultaneity bias.

2.4 Factors that Affect Conservation Behavior of Farmers
2.4.1 Research on Factors Affecting Conservation Adoption

Economic analysis can play a role in combating environmental degradation through
project level cost-beneﬁt studies, macro policy analysis, cost assessment of damages at
national or international level, and identiﬁcation and quantiﬁcation of the factors that
inﬂuence farmer adoption of conservation practices. Understanding of the factors that
aﬂ‘ect adoption of conservation practices is important in order to choose the appropriate
policy instruments that would inﬂuence how farmers manage the land. Quantitative and

qualitative research geared towards this end has been going on since the 1930‘s.

‘I

However, most adopt
studies tended to use t

studies in developing C

 

1990).

Following Nov
decisions can be grout
iru‘msrnrcture or orgar
These paradigms deve
of these models have

developed, focuses or
driving adoption. The

eCommie factors and

f.
”“1595 0n the orgaru
Physic -

al 0’ 0 Ogrca
differs

22

However, most adoption studies have been limited to the developed countries. These
studies tended to use detailed technical information and much rigor. By contrast, the
studies in developing countries tend to be more qualitative (Anderson and Thampapillai,
1990)

Following Nowak (1993), the paradigms for explaining conservation adoption
decisions can be grouped into ﬁve: innovation-diﬁ‘usion, economic constraints,
infrastructure or organizational, physical or ecological, and macro or systems perspectives.
These paradigms developed in a historical way, one following the other. Hybrid versions
of these models have also been used. The innovation-diffusion paradigm, the ﬁrst to be
developed, focuses on information communication and social pressure as major forces
driving adoption. The economic constraints paradigm emphasizes the importance of
economic factors and proﬁtability. The infrastructure or organizational perspective
focuses on the organization and the system by which it promotes the innovation. The
physical or ecological perspective focuses on the applicability of the technologies in
diﬂ‘erent physical environments. Finally, the systems model gives more importance to
macro policies and institutional arrangements. A new paradigm focusing on farmer
perceptions seems to be emerging (Adesina and Zinnah, 1993). No model has proven to
be consistently superior to others because the complexity and site speciﬁcity of adoption

decisions involves factors from each paradigm to varying degrees.

2.4.2 Empirical Moi
The predomin
multiple regression an
the three basic statistic
(Besley and Case, 1 99:
aggregate diffusion pro
com in the United State
an aggregate measure 0.

the relative inﬂuence of '

,el

 

23

2.4.2 Empirical Models in Adoption Research

The predominant analytic method used in conservation adoption studies has been
multiple regression analysis. Time series, cross sectional and panel data models have been
the three basic statistical models used in the study of technology adoption in general
(Besley and Case, 1993). The main focus of the time series models has been to study the
aggregate diffusion process over time. Zvi Griliches’ (195 7) study of adoption of hybrid
corn in the United States is a classic example. The dependent variable has typically been
an aggregate measure of adoption. These models are of limited importance in explaining
the relative inﬂuence of those factors affecting the decision to adopt. The cross-sectional
studies either take a snapshot of use by farmers of a given technology or use recall data. In
the former case the dynamic nature of the technology adoption process is ignored. In case
the diﬂ‘usion process is incomplete, parameter estimates may be biased and results could
be misleading. Ifanalysts are careﬁrl and explicit in differentiating between adoption and
intensity of use of the technology, results from a cross-sectional studies could be indicative
of important causal factors. Panel data models, which are based on data pertaining to
farm and ﬁrm characteristics and adoption choices over time, can potentially address the
drawbacks of both time series and cross-sectional models. However, the high cost
involved in collecting the panel data have made use of these models somewhat restricted.
Their use has been expanding recently.

Irrespective of the type of model used, research on the adoption of agricultural
innovations shares some common shortcomings (F eder et al., 1985). These shortcomings

include the following: (1) empirical analysis not based on theoretical models, (2) lack of

 

 

 

consideration of endo~

 

statistical assumptions
proportion of explaine
(Lockeretz, 1992).

It is often diﬁic

 

 

with underlying theory
underlying theoretical c
involved. Researchers :
terms of the data needs
underStanding of the ch

Non-statistical r

“Nels, have also been

Statistical models have i

Patio" Practices,

(1 - .
etenmnrng the proﬁtat

24

consideration of endogeneity problem of explanatory variables, and (3) violation of
statistical assumptions required for valid hypothesis testing. Perhaps as a result, the
proportion of explained variation of the dependent variable has been consistently low

(Lockeretz, 1992).

It is often difﬁcult to interpret results of an empirical model which is inconsistent

with underlying theory. The difﬁculty with using empirical models consistent with an
underlying theoretical choice model arises from the data needs and the complexity
involved. Researchers need to weigh the costs of using theoretically consistent models in
terms of the data needs and complexity required and the beneﬁts in terms of increased
understanding of the choice problem.

Non-statistical models, including simulation and mathematical programming
models, have also been used in the study of adoption of conservation practices. While the
statistical models have been used to determine the factors that affect farmer adoption of
conservation practices, the non-statistical models have, for the most part, been aimed at
determining the proﬁtability of the practices. Lopez-Pereira et al. (1994) used discrete
stochastic a whole farm programming model to determine the proﬁtability of a technical
package of physical conservation practices, living tree barriers, seed and fertilizer
technologies on low input hillside farms of Southern Honduras. They concluded that a
combination of soil conservation, improved seed and fertilizer technologies could increase
farm income substantially and facilitate adoption. Gray et al. (1987) used simulation to

4evaluate the proﬁtability of zero-tillage and minimum tillage practices in central

Saskatchewan, Canada. Painter et al. (1993) showed, using mixed integer programming,

 

that adopt
policies) Cl

USA diver

2.4.3 Deper
Research
The
Soil erosion,
This last has
or terraces), .

maintenance r

25

that adopting a more ﬂexible agricultural policies (as opposed to mandatory set-aside
policies) could result in signiﬁcant economic and environmental gains in North Carolina,

USA, diversiﬁed farming systems.

2.4.3 Dependent and Independent Variables in Conservation Adoption
Research

The dependent variables used have included farmer perception of the problem of
soil erosion, use/non-use of conservation practices, and intensity of use of the practices.
This last has been measured either as physical length of conservation structures (eg. bunds
or terraces), outcome in soil loss reduction or ﬁnancial expenditure on construction and
maintenance of the practices. The binary dependent variable of use/non-use of
conservation practices has been employed to examine the factors affecting the probability
of adoption. These factors may not be the same as the ones that inﬂuence how intensely
those practices which are adopted are used. Different factors may be associated with the
adoption of different conservation practices (Ervin and Ervin, 1982). Moreover, adoption
and continued use may be inﬂuenced differently by different factors. Rikoon et al. (1996),
in their study of sixteen counties in Missouri, USA, found that the factors associated with
adoption and continued use of banded pesticide application were not the same.

The most common explanatory variables used in the studies of technology
adoption in developing countries included farm size, subjective and objective risk, human
capital, labor availability, credit and type of tenure (Feder et al., 1985). The explanatory

Variables used to adoption of conservation practices could, in general, be grouped into

 

 

personal characteristi.
economic and instituti
and education of the f:
characteristics include.
included slope and soil
farm income. risk farrr
senices, and other pub
rambles. the operation
Ming comparison of r

Researchers had

 

 

Into demographic charac
Pb’srcal incentives to int

Eninll982)‘ .

26

personal characteristics, farm-ﬁrm characteristics, physical characteristics of the land,
economic and institutional factors. Personal characteristics have typically included the age
and education of the farm operator or farm household head, while the farm-ﬁrm
characteristics included type and size of the farm. The physical characteristics of the land
included slope and soil type. Economic and institutional factors related to farm and 0&1
farrn income, risk, farmer planning horizon, land tenure, technical and educational
services, and other public policies. In addition to the use of different sets of explanatory
variables, the operational measurement of the variables has also been widely different,
making comparison of results difﬁcult.

Researchers had to select the appropriate mix of variables for a particular study
area. Clay and Reardon (1996) in their Rwanda study grouped the explanatory variables
into demographic characteristics of the farm household, ﬁnancial incentives to invest,
physical incentives to invest, capacity to invest and riskiness of the investment. Ervin and
Ervin (1982) in their Missouri, USA, study grouped the variables into personal,
institutional, economic and physical factors. Cary and Wilkinson’s (1997) study of
Australian farmers used ﬁve sets of variables which emphasized perceptions: farmer
. perception of the environmental problem and technical feasibility of the practices,
perceived proﬁtability of the practices, farm size and environmental concern. They
contended that non-economic attitudinal factors could play a more important role when
proﬁtability is low. Similarly, Hansen et al. (1987) argue that socioeconomic factors may

be relevant in explaining adoption of practices that are commercially proﬁtable but not

 

ones that enhan

trnironmental t

14.4 Results 0

The res:
the nicro~level
be 3 better cons
of factors were
(19%) found in
Practices Em,
Educational [m

PilCllCes, GOUl

27

ones that enhance environmental quality. However their distinction of commercial versus

environmental quality enhancing practices was not clear.

2.4.4 Results of Past Research on Factors Affecting Conservation Adoption

The results of conservation adoption studies appear to be inconsistent regarding
the micro-level variables that affect adoption of soil conservation practices; there seems to
be a better consensus on the effect of macro-level variables (Ervin, 1994). A wide variety
of factors were found to explain conservation behavior of farmers. Sureshwaran et al.
(1996) found in the Philippines that education positively inﬂuences use of conservation
practices. Ervin and Ervin (1982) found in Iowa, USA, that the farm operator’s
educational level favors perception of erosion problem, adoption and use of conservation
practices. Gould et al. (1989) found education explaining adoption of conservation tillage
positively in Wisconsin, USA. Operator age seems to work both ways (Rikoon et al,
1996; Sureshwaran, 1996; Hoover and Watilla, 1980).

Under imperfect capital markets, wealthier farmers are more likely to invest on
conservation. Hence, farm income tends to encourage conservation (Rikoon et al., 1996;
Gould et al., 1989). The effect of farm size varies by the type of practice considered. In
general, farm size tends to have a negative effect on use of conservation practices (de la
Briere, 1996; Sureshwan et al.,1996; Clay and Reardon, 1996). The effect of off-farm

employment appears to mixed (de la Briere, 1996; Clay and Reardon, 1996; Gould et al.,
1989; Blase, 1960) . Oﬁlfarm employment may compete for labor with conservation

investment while the income generated could ease liquidity constraint thus encouraging

I

investment. Farm
and Ervin 1982).

A farmers
period during whi
be quite long. Tl
(1982) but contra
H980) found ten
their Colombian
Wet e fewer eros‘.
W from homes
Reardon. 19%)

Faﬁner .
ofumsen'Eltion
0f loci] Comm'
Bultena and m
storm are.
consen'iilion ti

Studies.

“This all tu

28

investment. Farmer risk aversion tends to discourage use of conservation practices (Ervin
and Ervin, 1982).

A farmers’ planning horizon may not affect adoption signiﬁcantly because the time
period during which net return without conservation exceeds that with conservation can
be quite long. This statement seems to be in line with the ﬁndings of Ervin and Ervin
(1982) but contradicts the ﬁndings of Gould et al. (1989). Reinhardt (1987) and Lee
(1980) found tenure security to be positively associated with conservation practices in
their Colombian and US studies respectively. Ervin (1982) found in Missouri that there
were fewer erosion control practices on rented than on owner operated lands. Distance of
plot from homestead was negatively associated with use of practices in Rwanda (Clay and
Reardon, 1996).

Farmer perception of the erosion problem tends to enhance the adoption and use
of conservation practices (Ervin and Ervin, 1982; Blase, 1960). Perception of the attitude
of local community towards farmers who fail to use conservation practices was found by
Bultena and Hoiberg (1983) to be the only variable, in addition to age, that showed
Signiﬁcant difference between early adopters, late adopters and non-adopters of
Conservation tillage. This variable is excluded from almost all conservation adoption
Studies.

In sum, the fact that demographic, physical, economic, institutional and perceptual

variables all turn out signiﬁcant in studies of soil conservation adoption clearly indicates

the complexity involved in conservation decision making of farmers. As a result, the

tendency in con:

explanatory vari

29

tendency in conservation adoption research should be towards expanding the number of

explanatory variables considered.

rc-

 

 

Ian) DEGR

3.1 Land Degr

In a stuc
Practices and th.
”ll conServatior
ilistitutional and
diets of land dt
expended in com

me Co“Senati

EthiOpia i

mural, 1985)_ E

CHAPTER 3
LAND DEGRADATION AND CONSERVATION EFFORTS IN ETHIOPIA AND

THE TIGRAY REGION

3.1 Land Degradation and Conservation in Ethiopia

In a study of land degradation, the physical and ecological conditions, agricultural
practices and the policy environment need to be considered together. The feasibility of
soil conservation techniques should be evaluated in terms of technical, economic,
institutional and policy constraints. This section, therefore, presents the extent, form and
eﬂ‘ects of land degradation in Ethiopia and discusses the causal factors. The efforts

expended in combating land degradation will be evaluated critically to draw lessons for

ﬁrture conservation policies and programs.

3.1.1 Extent and Form of Land Degradation in Ethiopia
Ethiopia is considered the most environmentally troubled country in the Sahel belt

(Hurni, 1985). Environmental degradation in Ethiopia is almost synonymous with land
degradation. Land degradation is manifested as soil erosion, loss of soil fertility and
depth, reduced plant water availability, and deforestation. Of these, soil erosion is the
most important environmental problem in the country and its causes are primarily human
rather than natural factors. The most important human activities contributing to land
degradation in Ethiopia are over-cultivation, overgrazing and deforestation (Dejene,

1990). Grepperud (1996) analyzed the relationship between population pressure and land

30

 

degradation in l
activities are sit
loss occurs nit
(1994) estimatt
rainy period.
There 1
Ethiopian Hi gl
CSlimated that
“5 Slgfilican
Study Conclud
rehabilitation,
Cultiv
estitrated am
land was 35 ‘
cultivated 1a,

31

degradation in Ethiopia. He found that pressure variables that emanate from human
activities are signiﬁcantly positively associated with land degradation. Most of the soil
loss occurs with the intensive rain storms. Constable (1984) as quoted in Grepperud
(1994) estimated that 58-80 percent of soil loss in Ethiopia occurs during the intensive
rainy period.

There are varying estimates of soil loss and its effect on Ethiopian crop yields. The
Ethiopian Highland Reclamation Study (EHRS) as quoted in Bojo and Cassels (1995)
estimated that by the mid-1980's about 50 percent of the highlands (27 million hectares)
was signiﬁcantly eroded while more than one-fourth was seriously eroded. The same
study concluded that more than 2 million hectares of cultivated land was beyond
rehabilitation.

Cultivated land is the major contributor to soil loss. According to EHRS the
estimated annual soil loss ﬁ'om cultivated land was 130 tons/ha while the average for all
land was 35 tons/ha. On the other hand, Humi (1988) estimated annual soil loss from
cultivated land to be 42 tons/ha and an associated soil formation rate of 3-7 tons/ha/year
(Table 3.1). Hurni estimated total annual soil loss to be 1.5 billion metric tons. Campbell
(1991) reported that total annual soil loss could be as high as 2-3 billion metric tons.

In addition to the variations in the soil loss estimates, Bojo and Cassels (1995)
argue that all previous estimates did not differentiate between gross and net soil losses. In
particular, redeposition of eroded soil was not considered. According to them net soil loss
is limited only to cultivated and unproductive land and nutrient loss is more important than

soil erosion. They estimated that gross ﬁnancial loss due to soil erosion amounts to US $2

 

 

lllluiOl

for ftu

asqu

annur

Tabi

 

 

 

 

32

million per year while nutrient loss due to removal of crop residues and use of animal dung
for fuel results in a loss of US $100 million per year. In this vein, Wood and Stahl (1989)

argue that use of dung for fuelwood reduces crop yield by 10-20 percent. Belshaw (1989)
as quoted in Campbell (1991) indicated that soil erosion reduces cereal yields by 2 percent

annually.

Table 3.1: Estimated Soil Loss Rates in Ethiopia by Land Cover

 

 

 

 

 

Land Cover Percent area Estimated soil loss
covered metric tons/ha/year million metric
tons/year

Crop land 13. 1 42 672
Perennial crops 1.7 8 17
Grazing and 51.0 5 312
Browsing land
Unproductive 3.8 70 325
Uncultivable 18.7 5 114
Forests 3.6 1 4
Wood and bush 8.1 5 49
land
Total 100.0 12 1493

 

 

 

Source: Hurni, 1988, p. 127

 

 

in the 1.
resour
incapz

estim

Elfie:

Tree
shes
Wet

aror

Da

Fe

 

33

Various predictions have also been made regarding the impact of land degradation
in the highlands in the absence of changes in agricultural practices, population grth and
resource use. One scenario estimated that by the year 2010 the amount of total land
incapable of supporting cultivation will reach 10 million hectares (Stahl, 1990). Another
estimated that by the same period three-fourth of the highlands would be food deﬁcit
(Ministry of Agriculture and FAQ, 1984; FAO/UNDP, 1984). While speciﬁc numbers
diﬁ‘er, the evidence reviewed all suggest that soil erosion is severe in Ethiopia.

In addition to soil loss, the forest resources of the country have also been depleted.
Trees have been used for ﬁre wood, house construction, farm implements, animal yard and
sheds. The wood supply used to come from homesteads, cultivated land , natural
woodland and forests. Forty percent of Ethiopia’s land area was covered with forests
around the turn of the century. This ﬁgure reduced to 16 percent by the 1960's and to 3.1
percent by the 1990's (Wolde-Giorgis, 1993). A shortage of fuelwood has forced farm
households to use animal dung for fuel. Fuelwood shortage is directly related to altitude;
higher altitude zones use animal dung almost entirely as source of fuel.

The progression of land degradation proceeded with the population settlement
pattern from the north of Ethiopia to the center and then southwards and westwards.
Fertile soils, cool weather and abundant rainfall favored the beginning of mixed
agriculture on the northern Ethiopian highlands about three thousand years ago. The
Abyssinian kingdom dated back to the Axumite empire, which was based in central Tigray.

Population pressure started to intensify in the 20th century. By the 1980's most of Tigray

 

 

was $8

Shoat

3.1.21

Thisc

Canbt

Series

(J
H,

F]

34

was severely eroded and the other northern provinces of Wollo, Gondar and northern

Shoa exhibited rapid land degradation (Stahl, 1990).

3.1.2 Physical Factors of Land Degradation

About 43 percent of Ethiopia is highland, more than 1500 meters above sea level.
This constitutes more than half of the Aﬁican Highlands (Humi, 1988). The topography
can be characterized as mountainous plateau extending from north to south, surrounded by
series of lowlands. The southern highlands are more productive and stable with lower
population density than the northern highlands. In general, the Ethiopian highlands
provide a favorable living environment and they have become home to 80 percent of the
human population and 67 percent of the livestock. The population density in the rural
highlands can go as high as 50 per km2 (Lanz, 1996). Ninety percent of the country’s crop
land is also found in the highlands (Dejene, 1990).

More than 30 percent of the highlands, occupying 537 000 km2 of land area, have
slopes exceeding 30 percent gradient (Dejene, 1990; Hurni, 1988). Nearly 79 percent of
the highlands have slopes of more than 16 percent gradient (Campbell, 1991). The rainfall
pattern in Ethiopia is aﬂ‘ected by altitude with the highlands, in general, having higher
rainfall than the lowlands. Most of the rain falls in high intensity. Fifty to sixty percent of
the rainfall in Ethiopia is lost as runoff (Lanz, 1991). It is estimated that about 80 percent
of the water resource of the Blue Nile originates from Ethiopia.

Due to high population pressure the sloping areas in the highlands have been over-

cultivated, often with inappropriate land management practices. Jointly, steep slopes,

 

 

 

heavy
erosic

Tigra

3.1.3

intert
Ihresi

livest

“ith
as ne
Pitsg

CIQp

[edits

35

heavy rains, high population density and intensive cultivation have resulted in severe soil

erosion in the Ethiopian highlands. Conditions are most serious in the north, especially

Tigray.

3.1.3 Agricultural Practices as Factors of Land Degradation

Agriculture in Ethiopia is predominantly mixed farming. There is strong
interdependence between crop farming and livestock rearing. Land preparation and
threshing is done using oxen power. Straw from crop production is a major contributor to
livestock feed. Cows are reared mainly for oxen reproduction and milk production.

Increased demand for food crop production due to increased population is ﬁrst met
with expansion of cultivated land. The extensive cultivation strategy soon reaches its limit
as new land becomes scarce. In the next stage, land is cultivated more intensively. At
present the Ethiopian highlands are at this juncture. Fallowing is becoming obsolete.

Crop residue and animal dung are increasingly used for fuel, and the practice of manuring is
decreasing. This reduces the organic content of the soil. Lack of organic content in turn
reduces fertility which reduces vegetative cover, contributing to soil erosion.

Crop lands that cover about 16 million hectares are the major contributor to soil
erosion in Ethiopia (Dejene, 1990). Humi (1983) found that the ratio of soil loss to soil
formation is four to ten times higher for cultivated land than for grassland depending on
agroclimatic zone. The high soil losses on cropland originate in the perceived need of small
seeded crops for ﬁnely tilled seed bed. Plots are plowed from three to eight times after

harvest. The ensuing dry periods are long. This hinders the development of vegetative

ﬂ

 

 

 

 

unp

crop

esrirr

yields

cattlea

lll’estot

mainly

36

cover. The rains begin as torrential storms and washes away substantial amounts of
unprotected topsoil.

Soil erosion, in turn, reduces the water retention capacity of the soil. As a result
crop production has become increasingly dependent on the availability of rainfall during the
growing period (Stahl, 1990). Webb and Braun (1989) as quoted in Campbell (1991)
estimated that a 10 percent decline in rainfall would result in a drop of 8—9 percent in cereal
yields. The problem is compounded with rainfall being more erratically distributed.

Ethiopia has the largest livestock population in Aﬁica. It is comprised of 27 million
cattle, 24 million sheep, 18 million goats, 7 million equines and 1 million camels (Dejene,
1990). The number of working oxen is estimated to be 6 million. During the dry season
livestock graze freely, mostly on croplands. During the rainy season livestock are limited
mainly to mountainsides and pasture land.

The huge livestock population is another contributor to land degradation. Free
grazing on croplands during the dry season reduces the vegetative cover of the land
exposing the top soil to the torrential rainstorm. During the rainy season, mountainsides
support livestock population beyond their capacity. Daniel (1988) as cited in Wolde-
Giorgis (1993) estimated that the Ethiopian highlands support livestock population more

than three times their carrying capacity.

3.1.4 The Policy Factor
The Ethiopian government began to expand state control into the rural areas in the

1940's (Lanz, 1996). After the second world war, capitalist elements started to be

 

 

' tr
bull
wer

allor

forct

CUlti‘

37

introduced into the feudal system. The modernization effort of 1950-1974 concentrated on
building internal inﬁastructure, developing industry and commercial agriculture. Peasants
were left out of the modernization process. During 1963-74, the share of state budget
allocated to the development of agriculture was 4.2 percent of which the most part was
spent on the commercial sector (Lanz, 1996). The expansion of commodity production
forced peasants into marginal lands. Forest areas were cleared and mountainsides put into
cultivation. Soil erosion intensiﬁed.

The modernization process itself fell short of success, and grth was very slow.
Surplus extraction ﬁom the rural areas intensiﬁed and weakened the agricultural productive
base. The modernization effort had a direct effect on the environment by destroying a
substantial amount of forest without replacement.

In 1974, a revolution took place and a military dictatorship replaced the feudal
regime. The following year, a land reform nationalized all land, abolished tenancy,
prohibited land transfer and limitted land holdings tolO hectares. Land administration
became the responsibility of local peasant associations. Private ownership of forest areas
was also abolished. Peasant associations and government agencies took the responsibility
of managing forest areas.

The agricultural policies of the military regime were aimed at strengthening state
control over grain trade, expansion of state farms and promotion of agricultural
cooperatives. Smallholder agriculture was once again left out of the scene although the
productivity of individual farms was higher than that of the agricultural cooperatives (Stahl,

1990). To these were later added forced resettlement and villagization. Villagization was

he

 

,J

 

requiri
goven

Corpc

trader

SUbst.
the ft
19%
and 1

dual

The

38

requiring farmers to abandon their villages and build new residences on sites selected by the
government. The govemment’s grain marketing parastatal, the Agricultural Marketing
Corporation (AMC), was mandated to ﬁx grain prices and enforce quota delivery ﬁom
farmers. The ﬁxed prices were in most instances below the market prices. Private grain
traders were either forced out of business or made to operate as agents for AMC.

The natural outcomes of the agricultural policies of the military regime were
substantial land tenure insecurity of farmers and reduced agricultural proﬁtability. During
the ﬁrst six years of the military regime only, food production declined by 6 percent (Lanz,
1996). Moreover, civil war intensiﬁed and the government was forced to withdraw more
and more resources from development and allocate them into combat. The defence share
during the military regime rose from 18 percent at the beginning of the 1970's to 50 percent
in 1988 (Dejene, 1990). Forced enlistment of farmers became a routine practice. In effect,
the military government replaced feudal lords in the appropriation of agricultural output
and even peasant labor. Land degaradation intensiﬁed as land and forest resources were

literally left without an owner, and the incentives for farm productivity diminished.

3.1.5 Conservation Efforts in Ethiopia

Ethiopian farmers are reported to be aware that soil erosion is a problem and that
their agricultural practices are aggravating the problem (Stahl, 1990), although no study
has been conducted to date on the level and determinants of their awareness. Despite their
awareness, farmers continue to use erosive land management techniques unless provided

with incentives to undertake soil conservation. The reasons for this could range from

p

 

 

pra;
exec
corn

Lhel

39

short-term liquidity problems to structural factors such as tenure insecurity or lack of
farmer involvement in development administration. Although traditional conservation
practices have a long history in the Ethiopian highlands, their use is decreasing with time
except in few areas in the south (Tilahun, 1996). A successful example of traditional
conservation practices is found in southern Ethiopia, among the peoples of Konso. Here
the Konso eﬁ‘ectively integrate terracing with crop and livestock production and forestry.

Based on descriptive results, researchers have also reported that Ethiopian farmers
believe that conservation practices could increase crop yields. The detemrinants of these
perceptions are unknown, however. Amare (1988) as cited in Tilahun (1996) found that
out of 72 farmers interviewed in the southern part of the country, 80 percent believed that
soil bunds increase crop yield. Admassie (1988) conducted a survey of 600 households in
the north, south and east and concluded that farmers believe that conservation practices
stabilize production due to increase in soil moisture. Research ﬁndings seem to conﬁrm
farmers views. Gebre Michael (1980) as quoted in Stahl (1990) found that yield on ﬁelds
with soil bunds can be higher by 60 percent compared to ﬁelds without soil bunds.

The farmer awareness of soil erosion should create a favorable condition for
conservation intervention by government or nongovemment bodies. Conservation
intervention in Ethiopia started in early 1970 (Campbell, 1991). The interventions took
almost entirely food-for-work (FFW) approach. In the 1980's FFW programs in Ethiopia
became the second largest in the world, next to India (Campbell, 1991). The World Food

Program (WFP) and the European Economic Community (EEC) were the main suppliers of

 

‘Q

C0

ye

bu

'5?

pr:

4o

grain and edible oil for FFW while other donors provided tools and technical equipment
(Stahl, 1990).

By the late 1980's, more than 1 million km of soil and stone bunds have been
constructed on cultivated land, plus more than half a million km of hillside terraces had
been built, and more than 80 000 hectares of hillsides had been closed off (Hoben, 1995).
Not surprisingly, most of the conservation work was done in the northern Highlands.
Humi (1988) argues that soil accumulation behind bunds can form bench terraces in 5-20
years. Martin Bundez as cited in Dejene (1990) found that in hillside closures, grasses and
bushes could regenerate within 2-3 years, erosion signiﬁcantly reduced in 3-5 years, and
gully formation stopped in 10 years.

FFW was also instrumental in afforestation programs with support for the
programs coming mainly from WFP, Swedish International Development Agency (SIDA),
and the Federal Republic of Germany (FRG) (Dejene, 1990). Trees and forests help to
reduce soil erosion by reducing the erosivity of rainfall, improving soil structure by
increasing its organic content, keeping soil intact by the effect of roots, and reducing the
velocity of winds. In the 1980's, about 300,000 hectares were planted with trees, and
nurseries with a capacity of raising about 100 million tree seedlings annually were operating
(Stahl, 1990). Grass seedlings to stabilize terraces were also raised.

Impressive as these results may appear, the FFW approach was conﬁ'onted with
several problems. The environmental reclamation effort was not based on any on-farm
research on the production, economic or environmental impacts of the technical package,

much less on the diverse agroecological conditions (Hoben, 1995). The conservation

 

 

pract
3m.
CXlCll‘.
impler
works
emiror
Palmer
Prepara

OilOCaj

PTESSUre

llOl prov

s‘m'll’al r

41

practices had technical problems. Stone terraces, in addition to reducing cultivated land
area, were harboring pests. The land area covered was quite limited compared with the
extent of the problem. The lack of involvement of local people in the planning and
implementation of the FFW conservation projects made farmers view the conservation
works mainly as a means of getting the FFW payments rather than as necessary means for
environmental rehabilitation. As a result, sustainable use of the practices without the FFW
payments was very low. The conservation techniques were labor intensive. The project
preparation and implementation followed a pure ‘top-down’ approach where involvement
of local people was minimal. In particular, implementation was coercive.

As for afforestation, there were no management plans for the closed areas and
pressure from grazing animals increased on nearby areas. The afforestation programs did
not provide any guidelines regarding tree ownership rights. Perhaps as a result, the

survival rate of seedlings was low (Stahl, 1990).

3.2 Society, Agriculture, and Natural Resources in the Tigray Region

This section deals with the effect of agro-ecological, social and physical factors on
the problem of land degradation in Tigray. The mechanisms by which the rural population
adopt to changing agricultural and environmental conditions including drought, and input
constraints are discussed. Current practices of crop and livestock production are
presented. The level of natural resource degradation, farmers perception of the problem,
and institutional and private efforts to combat land degradation currently underway in the

region are assessed. Finally the role of rural institutions including land tenure, agricultural

 

 

credit

availal

3.2.1 T

longituc
approxjr
Plateau.
Center to
loin low];
salt minin
At
Went Ct
which is a
admlmslrat

The

Mum-0’ 197

42

credit and extension services is discussed. Since most of the secondary information was

available by administrative zones the discussion is, in most cases, accordingly structured.

3.2.1 Topography and Climate

Tigray is situated between 12°20’ and 14°40’ N latitude, and 36° and 41 °30’ E
longitude on the Sudano-Sahelian Aﬁican drylands zone. It covers an area of
approximately 80 000 km’. The topography of the region is characterized by mountain
plateau. A Plateau of height between 2500 and 3400 meters above sea level crosses the
center towards the west with associated mountains on both sides. In the west the plateau
join lowlands of less than 500 m a.s.l.. In the cast an escarpment sharply extends into the
salt mining depression of the Danakil desert.

Administratively, the region is classiﬁed into four administrative zones- southern,
eastern, central and western. The basic administrative unit in the region is called a Labia
which is a collection of villages. A group of Lamas make up the next higher level
administration referred to as a male. Several Was make up a zone.

The climate of the region can be characterized as tropical semi-arid (Virgo and
Munro, 1978). As such, Tigray is confronted with the challenges of development caused
by a dryland environment including erratic and unreliable rainfall, shallow and infertile soils,
and outbreak of crop pests and diseases. Agroecologically, the region can be classiﬁed into
four zones, viz., highland, intermediate highland, lowland and desert, with relative area
coverages respectively of about 5 percent, 41 percent, 49 percent and 5 percent (Regional

Bureau of Agriculture, 1995).

 

 

 

 

Atm
incentral, e;
most import.
lemPOral dis

Annu
moves from g
“”5 may sht

l‘fStem zone
nth small raj;
pellOd lUne to
may begin as e

The dis
the three montl
falls at a late 0:

correlation CXis

M avenge raj

higher lhan the t

43

Atmospheric temperature in the region ranges from 5 ° to 40° celsius. Most areas
in central, eastern and southern zones have temperatures between 15° and 199° C. The
most important factor of agricultural production in Tigray is the amount, and spatial and
temporal distribution of rainfall, particularly during the months of July and August.

Annual rainfall ranges from 450 to 980 mm. In general, rainfall increases as one
moves from east to west and from north to south. However, depending on altitude, nearby
areas may show wide differences in the amount of rainfall they receive. While most of the
western zone has a unimodal rainfall pattern, the rest of the region has bimodal rainfall,
with small rains falling during the period March to May, and the big rains falling in the
period June to September. In some areas in the eastern and southern zones, the small rains
may begin as early as January thus allowing double cropping.

The distinguishing characteristic of the precipitation is that most of it falls within
the three months and in high intensity. It was estimated that 75 per cent of the precipitation
falls at a rate of more than 25 min/hour (Virgo and munro, 1978). Moreover, no strong
correlation exists between number of rainy days and amount of rainfall (Tilahun, 1996).
Another major characteristic of rainfall in the region is its inter-annual variability. A 27-
year average rainfall indicates that rainfall variability is high in absolute terms and much

higher than the corresponding national ﬁgure (Table 3 .2).

1‘

 

44

Table 3.2: National and Regional Rainfall Averages and Variability (1961-87)

 

 

 

 

 

 

 

 

Average Annual % of Standard Coefﬁcient of
Rainfall (mm) National Deviation Variation
Average
Tigray 578 63 162 28
Eihiopig 921 100 71 8

Source: REST, 1995, P. 31

There are two major drainage systems in the region. The Mereb river, which is
seasonal drains northern Tigray in the west direction to the Sudan. The perennial Tekeze
river drains central and southern Tigray to the Nile (Belay, 1995). Despite a substantial
amount of potential irrigable land, estimated to range from 0.3 million to 0.5 million
hectares, little irrigation is practiced by Tigray farmers.

Very limited basic inﬁ'astnrcture is found in Tigray at the moment. Average walking
distance to nearest all-weather road is estimated to be 6 hours (Woldegiorgis, 1993). A
study conducted in central Tigray also indicated that average walking distance to nearest

market place in the area was 2.5 hours .

3.2.2 People and Society
Population and social organizations

The population of Tigray is estimated to be 3.2 million making crude population
density of 40/krn2 (Regional Bureau of Economic Planning and Development, 1995).

Ninety percent of the population depends on agriculture for its livelihood. Average life

ll

 

 

cxpmrrcy i
basic unit of
households (
makes decisit
conditions wl
Social organiz
rdllllous assoc
50dal function
famine conditic
The hot
do“ SCnrtiny if
relationships “,1:
eﬁmeb houn.
household decisi
hOUSCholds and ‘
Strong belief 3111c
aCCOIIIijSlmmmS
hither We as an ‘

irritates risk lalq'

Bailing“, n l5

(6 Th

areb

45

expectancy is 46 years (REST, 1995). The household, built mostly on nuclear family, is the
basic unit of society. The rural population is organized into more than half a million farm
households (Meles and Hagos, 1995). The household ﬁmctions as an economic unit. It
makes decisions on resource allocation dictated by the ecological and environmental
conditions which determine climate and operating costs. Besides the household, several
social organizations aimed at mutual help and cooperation exist, including neighborhoods,
religious associations, savings associations, and labor and/or oxen exchange groups. The
social function of these organizations is signiﬁcant, particularly during drought and/or
famine conditions.

The household in Tigray appears to be a stable unit of social structure. However, a
close scrutiny into the development of a typical household indicates that vertical
relationships within a family are easily liable to change. The household , although
eﬁ'ectively bound together as a unit, can easily change in size. The vertical ties between the
household decision maker and junior members can and do break at any time. Hence
households and villages can rapidly change in size and adapt to environmental changes. The
strong belief among the Tigrayans that status-honor should be achieved by own .
accomplishments and not by inheritance gives the motivation for every individual to ﬁnd
his/her role as an active member of the social system. This ﬂuidity of social organizations

facilitates risk taking, and would encourage innovation and rapid diffusion of innovations.
Bauer (1985, p. 151) describes this situation as follows:
“ The apparent stability and rigidity of social stratiﬁcation [in Tigray]

are belied by the facts. The appearance of moribundity falls to a view

F

 

 

of the pc
of non-u
Sim-ion:

S.
Munro, 1
due to on
Coarse ten
Nonnd M
eastern an

46

of social relations that encourage, almost force, innovation both
social and technical. The family as a transgenerational unit is weak;
emphasis is laid upon the leadership of individual household heads.”

Children of non-working age, less than 15 years of age, constitute nearly 50 percent
of the population. A study conducted in central Tigray indicates that proportion of children
of non-working age in the survey area was 46 percent while that of old age people, above
sixty-four years of age, was less than 4 percent (REST, 1995).

Settlement pattern in the region appears to be determined by soil type (Virgo and
Munro, 1978). While ﬁne textured and less easily drained soils favor nucleated settlement
due to difﬁculty of traversing ﬁelds during rainy season, dispersed settlement is found in
coarse textured and easily drained soils. The former is found in the southern zone and
around Mekelle, the capital town of the region. The latter settlement pattern is found in the
eastern and central zones. This difference in settlement pattern has a direct bearing on the

possibility of using manure for maintaining soil fertility.

Rural labor supply

The major occupation of rural labor is farming, with limited involvement in off-farm
work. Extent of participation in off-farm work, however, differs from zone to zone where
Participation is highest in the eastern zone followed by the central zone. Farmers in the
southern and western zones show less involvement in off-farm work. The supply of labor
in the region is regulated in part by the ﬁmctioning of an institution of celebrating saints

days Which prohibit agricultural work during those days. A survey in central Tigray

1‘

 

 

 

 

 

indicat

47

indicates that on average 11 1 days in a year are work prohibition days (REST, 1995).
During peak labor demand seasons, the most frequent source of additional labor is
exchange labor. Use of hired labor is very limited although the daily wage rate of an
agricultural laborer seems to increase ﬁ'om year to year.

Prior to 1975, seasonal migration of labor from Tigray to commercial farms in the
southern and to the coﬂ‘ee growing areas of the western parts of the country was common.
As many as 100 000 laborers used to migrate southwards from Tigray and the neighboring
former province of Wollo every year in the 1960's (Lanz, 1996). Such an outﬂow of labor
ﬁ'om the northern regions was an indicative of the fact that the north was converted into a
progressively weaker economy and a source of surplus labor to the growing commercial
sector which was part of the modernization strategy of Emperor Hailesselassie’s regime.
Although such opportunities for employment did not exist after 1975 due to the
nationalization policy of the military government and civil wars, the outﬂow of productive
labor from the region continued in the form of forced resettlement. In other words, the
indirect forces of neglected and weakened economy were replaced by the direct military
forces of resettlement. The outﬂow of labor from Tigray has been reversed since 1991.
There is now an inﬂux of returnees to Tigray, people who had migrated to other parts of
Ethiopia, were dislocated by the resettlement programs of the military government or after
Staying in the Sudan as refugees. Most of these returnees are farmers (Wolde-Giorgis,
1993),

On the other hand, shortage of trained manpower, workers with post-secondary

education, is still very critical. In 1994 the trained manpower in the regional bureaux of

 

 

 

 

 

 

 

agricult
B.Sc. at
in the re
hating c
One nee

need to t

The role
ll

mills Sur-

 

 

48

agriculture and natural resource development combined consisted of only 45 M.Sc., 9O
B.Sc. and 352 Diploma (two years of college education) holders. The manpower situation
in the regional agricultural research center was even worse with less than 10 researchers
having only M.Sc. or B.Sc. degrees and most of whom had limited research experience.
One nwds to see this acute shortage of trained manpower in conjunction with the urgent

need to combat land degradation and increase food production.

The role of women

Households headed by women are a permanent feature of social organization in
Tigray. Bauer (1985) indicates that in 1973, the proportion of female headed households
in his survey area around mekelle, the capital of the region, was 20 percent. A survey
conducted by REST in 1993-94 in central Tigray came up with a corresponding ﬁgure of
25 percent (REST, 1995). The latter ﬁgure is higher than the national average of 14
percent (Gebru, et al., 1995). Women’s role in the region in household decisions, and
political and administrative matters is increasing. Active participation of women in local
administration from village to woreda level is particularly signiﬁcant. Involvement of
Women in decision making at the household level is also high. A survey in central Tigray

showed that most decisions at the household level are also made jointly by the husband and

the wife (Table 3.3).

49

 

 

 

 

 

 

 

 

 

Table 3.3: Role of Household Members in Decision Making in Central Tigray

Decision Decision Making Role (percent)
Husband Wife Both Other

Planting 12.91 19.75 57.70 9.64

Selling/Buying 8.71 16.49 69.67 5.13

Crop

Selling/buying 11.35 20.68 62.74 5.23

livestock

Seed Selection 16.49 20.06 52.41 11.04

Social 10.58 24.42 60.33 4.69

menses

Source: REST, 1995, P.115

3.2.3 Crop Production
Land use and cropping pattern
Out of the total estimated regional cultivable land area of 1.5 million ha, about 1

million ha is currently under cultivation. Most of the uncultivated area, a relatively fertile
land mass, is found in the north western lowlands of the region (Meles and Hagos, 1995).
Forty percent of the total area is estimated to be used for grazing while forest cover
accounts for only 0.3 per cent (Regional Bureau of Agriculture, 1995). Cereal crops
production accounts for 84 percent of the cultivated area, while oil seeds and. pulses

account for 9 percent and 7 percent respectively.

 

 

0P
ﬁel
due
on

oft

Uta

oils

50

Cropping pattern is determined by a combination of factors including risk attitudes
of farmers, preference for food, yield considerations and the time required by each crop for
optimum planting. The most important factor perhaps is risk. Farmers prefer to plant
ﬁelds in dispersed locations and use a variety of crops in order to avoid total crop failure
due to erratic and unreliable rainfall or crop pest outbreak. Moreover, farmers seeding rate
practice shows an inverse relationship between seeding rate, and depth and fertility status
of the soil.

The three most important cereal crops in the region in terms of area coverage are
sorghum, barley and teff (Table 3.4). Sesame accounts for more than 80 percent of the

area covered by oil seeds while fava beans account for about 30 percent of area covered by

oil seeds.

 

p

 

51

Table 3.4: Distribution of Area Coverage of Cereal Crops in Tigray (1993-95

 

average)
Crop Percent of area covered
Sorghum 28.84
Barley 16.67
Teﬂ‘ 16.3 b
Millet 13.55
wheat 10.75
Maize 9.09
Barley/wheat mix 2.95
her 1.85

 

 

 

Source: Abraha et.al., 1995, p.4

There is also a difference in cropping pattern among zones. While barley is the
most important cereal crop in the southern and eastern zones, sorghum takes the lead place
in the central and eastern zones (Table 3.5). Among oil seeds, linseed is most important in
the southem, eastern and central zones while sesame takes the place in the western zone.
Chickpea dominates the area devoted to pulses in the western zone, while fava bean is

more important in the other zones.

 

52

Table 3.5: Three Widely Grown Cereal Crops in Tigray, by Zone (1993-1995
average)

 

 

 

 

 

 

 

Zone Rank
lst 2nd 3rd
Southern barley wheat teﬂ‘
Eastern barley wheat barley/wheat mix -
Central sorghum teff barley
Western sorghum ﬁnger millet teff
Reg’on sorghgm barley teff

Source: Own computation based on Abraha et. al.,(1996, p.4)

Under normal conditions, the ﬁrst crops to be planted are ﬁnger millet, maize and
sorghum, followed by barley, barley/wheat mix, wheat and teff. Oil seeds and pulses are
planted late in the season, generally only when the early crops fail. This planting pattern
may be disrupted if the small rains fail. Perennial crop production is very limited. Due to
the variety of agroecological conditions in the region, diﬁ‘erent perennial crops including
orange, lemon, guava, papaya, banana and grapes can be grown and the reason for the
limited production of these crops is not obvious. However, farmers’ interest in perennial

crops appears to be growing.

 

53
Crop yields
Due to moisture stress, crop pests, and low soil fertility, crop yields in the region
are low. The factors for low crop yield are direct consequences of the challenge that the
dryland environment poses for development. The regional average yield for cereals is 6.93

meteric quintals (100 kg) per hectare (qt/ha), while the corresponding ﬁgures for oil seeds

h

and pulses respectively are 5.66 and 4.59 (Table 3.6).

Table 3.6: Yield (lOOkg/ha) of Cereal Crops, Oil Seeds and Pulses in Tigray

 

 

 

(1993-95 average)
—
Cereals Oil Seeds Pulses
Crop Yield Crop Yield Crop Yield
Barley 6.82 Noug 3.12 Field pea 4.23
Wheat 5.40 Linseed 2.50 Chick pea 4.45
Barley/wheat 5.95 Sesame 6.26 Lentils 3.68
mix
Finger millet 5.88 Haricot
bean 4.50
Maize 8.69 Lathyrus 5.67
Sorghum 9. 15 Fava beans 5.06
Teﬂ‘ 4.06
Re ' n 6.93 5.66 4.59

 

 

 

 

 

 

 

Source: Abraha et. al., 1995

 

In t
sour
Fan.

mois

mall;

these

Whicl
better
Stlecr
When

diﬁ‘erg

mulllPl;

0:1:ch

ﬁo’n dis

54

Use of improved seeds in the region is very limited and farmers grow local seeds.
In addition, farmers prefer to use their own seeds rather than seeds obtained from other
sources. However, seed selection ﬁom among the local varieties is widely practiced.
Farmers select seed based on several criteria including time required for maturity; soil and
moisture requirements; weed, pests and disease tolerance; baking quality, suitability for
different food products, color of food product; grain and fodder yield; and size of grain.
Among these the most important factors considered by farmers are time required for
maturity, drought tolerance and grain yield (Haile, 1995). Plant breeders may need to take
these factors into consideration in designing research programs.

Seed selection is done both pre-harvest and post-harvest. In the former method,
which is more widespread, farmers mark and harvest separately those plants which showed
better performance ﬁ'om germination to fruition. In post-harvest selection, better seeds are
selected from harvested heads. Seed exchange among neighboring plot owners is practiced
when one of the plots has better crop stand or patches of crop stands. Farmers residing in
diﬂ‘erent villages also exchange seed. In addition to seed selection, intercropping is also
practiced by farmers in Tigray, the most common intercropping practices being cereals with
oil seeds or pulses. Seed selection is accompanied by seed multiplication. This is done in
two ways. One way is by sowing the selected seeds in part of a plot designated for
multiplication during planting season. In the second method, farmers plant the seeds on
catchments during dry periods.

Despite efforts by farmers to preserve their own seed, in practice, farmers use seeds

from diﬂ‘erent sources. This is a direct result of the inability of most farmers to be self-

“h

 

frequ

opera

Opera

Oren

hOUsel
disiilbr
social .

China]

55

suﬁcient in their production. A study in central Tigray indicates that seed sources other
than own production were mentioned more than sixty per cent of the time. The most
frequently used alternative sources being market and relatives.

Community seed banks aimed at providing seed to the needy farmers are now
operational in some mm of Tigray. They are organized at the woreda level and
operate at the village level. Their purpose is to store good quality seed varieties. Each
seed bank is administered by a committee comprised of agricultural experts, contact
farmers, and selected farmers. Seeds are purchased from local farmers who multiply better

perfomring varieties or have ﬁeld of better crop stand in the area.

Oxen ownership and use

The current land disribution system, assigning land to individuals rather than to
households, resulted in remarkable equity in land distribution. With this equity in land
distribution in the region, agricultural capital and inputs have become the major factors of
social differentiation. The most important factor, perhaps, is oxen ownership. A study in
central Tigray indicates that 38.5 per cent of farmers in the survey area were without any
0X, only 16.6% of them reported being self-sufﬁcient in draft power, and average oxen
Ownership was 0.83 per household (REST, 1995). Shortage of draft power is even more
acute to female headed households. In addition to oxen losses from recurrent droughts, the
availability of fodder in normal weather years is another reason for households’ inability to
keep the required amount of oxen. The problem of fodder is particularly pressing before

the beginning of the rainy season when most of the grazing pasture is depleted. It is

 

 

 

 

 

 

 

common

dwﬁp

lhe situa
growth. 1
of land pr
most lmpr
very recer
imillemenr

about 361

AEtitultur

Alt
ads‘ante f
aEllttulrura
hhck‘Smit'

black‘SIrn't

56

common to supplement this dry period feeding with straw and stovers. However, the yield
of such products does decline from year to year together with the decline of grain output.
The situation is compounded with the decreasing land holding size due to population
growth. The consequences of inadequate draft power include late and reduced frequency
of land preparation, and late planting with a major depressing effect on crop yield. The

most important source of additional draft power is oxen exchange. Oxen rent for cash has

h

very recently been started. In the eastern zone a pair of oxen including labor and
implements is rented for 25-30 Birr’lday. In the central and southern zones, the rate is

about 36 Birr/day.

Agricultural implements
Although less serious than the shortage of draft power, some households also lack
adequate farm implements. Agricultural villages in Tigray used to be self sufﬁcient in
agricultural tools and implements. Iron parts of the oxen-drawn plow were made by local
black-smith and wooden parts by the farmers themselves. There would be at least one
black-smith workshop in each village for producing and sharpening metallic tools.
The oxen-drawn traditional plow in use in Tigray, and indeed in many parts of
Ethiopia, only turns the soil without breaking it. It is possible that unploughed strips may
be left between consecutive passes. As a result, ploughing of three to four times‘ is

normally required before the soil is ready for sowing. At ﬁrst plowing, plow depth ranges

k

31:: 1995/96, US $1 =6.30 Bin:

4Teﬂ‘ may require up to six ploughing before sowing.

 

~

 

 

from
Cultir
POSl-l

purpo

diiiere;
and ear
the pan
Cultirar
Sol Th
lRTPD)
into Con:
moldboa,
dQPEndin!
reduce th,

is aimed a-

Milan

The
[Hamming ar
his become

her}.- limited.

57

from ﬁve to ten centimeters depending on the soil type and moisture content. At ﬁnal
cultivation, ploughing can be as deep as twenty centimeters (Salih and Morris, 1996).
Post-planting plowing is commonly done in ﬁelds sown with maize, sorghum and millet, the
purpose of which is to increase water percolation and reduce weeds.

The reasons for the widespread use of the traditional plow include its versatility for

diﬁ‘erent operations and soil types, simplicity in operation and maintenance, light weight

I

and ease of transportation, low draft power requirement, and the possibility that most of
the parts can be made by the farmer himself. Its drawbacks include its being primarily a
cultivation rather than a plowing implement, low pulverizing effect, and inability to turn the
soil. The efforts currently underway by the Rural Technology Promotion Department
(RTPD) of the region to introduce improved farm implements need to take these points
into consideration. The regional RTPD has been trying since since 1992 to introduce the
moldboard plow, chisel plow, broad bed maker, and tied-ridger to farmers in the region,
depending on the speciﬁc needs of a particular environment. Broad bed maker is meant to
reduce the problem of water logging in areas with predominantly clay soil, and tied-ridger

is aimed at conserving soil moisture.

Fertilization and pests

The most widely used fertility maintenance practices of farmers in the region are
manuring and crop rotation. Due to land scarcity and possibily tenure insecurity, fallowing
has become a disappearing practice. Use of chemical fertilizer by farmers in the region is

Very limited. The proporition of farmers using commercial fertilizer was 11 percent in 1993

   

58

and 16 percent in 1995. Among the few farmers who use commercial fertilizer, the
application rate is much lower than the rates recommended by the regional bureau of
agriculture. A study conducted in 1995 in central Tigray indicated that the average rate of
fertilizer application in the study area was 18 kg/ha (Haile, 1995). There is also zonal
disparity in the use of commercial fertilizer. The eastern zone has the highest proportion of
households that use commercial fertilizer while the southern zone has the least. The
proportion of households in the central zone is slightly higher than those in the western
zone. The most common reasons given by farmers for not using commercial fertilizer
include high price, and inadequate and/or unreliable rainfall. Some farmers believe that
application of fertilizer favors the development of the crop pest shoot ﬂy which aﬂ‘ects
mainly teff. Due to its high value, most fertilizer is applied on teff.

The dry land environment of farming in Tigray provides favorable conditions for
weeds, crop pests and diseases. However, the threat to agricultural production comes
more from weeds and pests than plant diseases (Fitiwy et al., 1996). Broad leaved weeds,
grasses and parasitic weeds are economically important. The parasitic weed, striga,
favored by the regional ecological conditions of erratic rainfall, low soil fertility, and mono-
cropping practices, causes substantial loss in sorghum, ﬁnger millet, and maize ﬁelds.
Important crop pests in the region include barley shootﬂy, temrites, stalk borer, weevils and

beetles. Rodents also cause signiﬁcant problem both pre- and post-harvest.

 

h

 

.F'

 

 

 

[it

59
3.2.4 Livestock production

Livestock population and use

Livestock play an important role in the rural economy of Tigray. Oxen are the only
sources of draft power for plowing. As such, oxen ownership is major indicator of wealth
and food security status of farm households. Livestock products such as milk, meat, and
eggs are important food items for rural households in addition to being a source of cash.
Hides and skins are used for storing agricultural products and for sale. Even though dung
cakes are increasingly becoming used for the], animal manure is still used as fertilizer.
Moreover, livestock are used as dowry and source of prestige for many rural families.
Some rural households consider investment in livestock as a risk spreading strategy- a
source of cash income from sale in case of crop failures due to drought.

Estimates of livestock population in the region vary considerably. The regional
bureau of agriculture’s 1992 livestock survey report estimated that Tigray had 2.1 million
cattle, 5.6 million sheep and goats, 0.5 million horses, mules and donkeys, 1 million
Pollltry, and 0.1 million beehives. As such, the region accounts for about eleven per cent of
the total livestock population in the country (Berhane, 1995). While cattle dominate in the

highlands, sheep and goats are more prevalent in the low altitude zones.

Livutock management and feed

The most important problems of livestock production in Tigray is shortage of feed,
Particularly in the southern, eastern and central zones. Livestock population is much higher

than the carrying capacity of the region as a whole. Animal feed in the region include

 

1'.

 

 

Cu

dc»
“6e

Ptoc

liar;

use,

60

straw, maize and sorghum stover, hay, pasture and browse trees. More than forty-six
percent of the feed comes from crop residues, the most important of which are teff, wheat
and barley straw, and pulse halms. During the dry season animals graze freely on cultivated
and communal pasture areas while during planting time they would be restricted to pasture
areas and areas bordering the cultivated plots. After the 1984/85 drought, prickly pear is
becoming popular as animal feed in the region. Although the crop is not a traditional
fodder crop, recurrent drought and coping efforts of farmers is making the crop an
important fodder crop. My discussion with farmers in the ﬁeld regarding the reason for the
growing popularity of prickly pear as a fodder crop indicates that its resistance to drought
and its moisture conserving ability make the crop suitable as a fodder crop during dry
season. Its cultivation which was limited to the eastern zone is now spreading to all over
the region. Many farmers are planting their gardens permanently with the crop. Farmers
Who live in areas with relatively more livestock feed take custodianship for raising cattle,
Sheep, and goats owned by those who live in areas of more serious feed shortage.
CuStodians get 25 to 50 percent of goat and sheep offsprings, and all the milk ﬁom cows.
Poor management is another major problem of livestock production. Most farmers
deVOte minimal managerial care only to oxen and cows. Sheep and goats are believed to
need no managerial care. Moreover, quantity of animals is preferred to quality and
productivity. As a result, livestock production in the region is characterized by late
maturity, late ﬁrst calving, low milk yield, and long calving interval. First calving in cattle
is at about ﬁve years and average calving interval is three years (Berhane, 1996). Animal

diseases are prevalent.

 

e-V
F

 

 

 

 

T r

xi

61

The extension service provides demonstrations and farmer training on improved
livestock production techniques. In addition to expanding the provision of veterinary
services, the regional bureau of agriculture is trying to introduce an improved feeding
system including restricted grazing, improved hay making, stall feeding, and forage
development. The shift towards stall feeding needs to be seen within the overall context of
agricultural production in the region. Stall feeding can increase availability of manure and
reduce the energy loss of livestock due to walking in search of fodder where there usually
is little. On the other hand, stall feeding requires more labor for watering, housing and
breeding. Moreover, oxen and pack animals need the physical exercise required for
plowing and transporting.

Overpopulation of livestock is one major cause of land degradation in the region.
Ilnprovement of the quality and productivity of livestock could facilitate destocking thus
tﬁducing the pressure on grazing land, and better opportunity for stall feeding. This would,
ho“fever, require farmers to change their attitude towards livestock ownership. Strategies
for natural resource conservation and development in the region need to explicitly consider
1iVBStock production. Failure to do so may mean ineffective use of human and capital

resources for natural resource conservation and even more land degradation.

3~2-.5 Natural Resource Degradation and Conservation

Soil erosion

Land degradation in Tigray, mostly in the form of soil erosion and deforestation is

V917 severe, particularly in the southern, eastern and central zones. The western zone, due

 

 

 

“in

ii

the r

62

to relative low population density and adequate grazing area for livestock, exhibits less
severe land degradation. The severity of land degradation in the region is manifested by
gullies cutting arable lands, exposure of stones and rocks in cultivated and grazing areas,
destruction of grazing land and declining crop yield. REST’s study (REST, 1995) in
central Tigray indicate that about 50 per cent of the cultivated land in the study area was
severely eroded. This result conﬁrms the ﬁndings of a study conducted in 197 5 (Virgo and
munro, 1978). Michael Stahl succinctly described the land degradation situation in Tigray
as follows: “ If any region [in Ethiopia] would need environmental ﬁrst aid, that would be
Tigray” (Ornas and Salih, 1989, p. 192).

The main causes of soil loss is in appropriate land use coupled with erosive rainfall,
Steep slopes. There has never been a policy that delineates arable from nonarable areas. As
a result, although most arable lands are found on the plateau and undulated terrain, steep
slopes are also widely cultivated. Except in the northwestern part of the region, almost all
land that is physically cultivable is currently under cultivation. Areas that are not cultivated
are in most cases drainage-line depressions used for grazing livestock and steep rocky lands
whiCh are the only places for browsing.

The seemingly paradoxical combination of lack of adequate soil moisture and
e"massive run-oﬁ‘ are among the major contributors to the severity of soil erosion. Due to
the concentration of rainfall in the months of June to August, shallow soils, removal of ﬁne

Soil particles by soil erosion and lack of soil organic matter, soil moisture is in the available
faItge for less than three months in a year. Lack of adequate soil moisture reduces the

WBgetative cover of the land exposing the soil for further erosion. Moreover, cr0p residue

 

 

s...

in"

v14

63

is grazed right after harvest after which the land is ploughed repeatedly to combat weeds.
When the rainy season begins, cultivated land remains without cover. According to Virgo
and munro (1978), even during the growing season, vegetative cover in the grazing and
browse areas does not exceed 30 percent, the rate below which accelerated soil erosion
could occur; and that the highest rate of soil erosion occurs in marginal cultivated lands of
a middle slope of up to 15 percent gradient.

Quantitative soil loss studies are rare in Tigray. The ﬁrst attempt in this regard was
the study conducted by Virgo and Munro in 1975. Based on a combination of empirical
and suspended sediment measurements, the study estimated soil loss for the central part of
the region to be 17-33 metric tons per hectare. The same study applied the Universal Soil
Loss Equation (U SLE) on cultivated lands and concluded that potential soil loss rates for
the region could range ﬁom 200 tons /ha to 400 tons/ha, the difference being principally
acmounted for by sowing date, which affects the level of vegetative cover. Moreover, a
co"lparison of photographs of 1965 and 1974 indicated that gullies expand at an average
rate of 5-10 meters per year and that gullies which were as recent phenomena as since 1943
are now a major erosion feature in most parts of the region.

The USLE could have been applied to assess soil loss and indicate the relative
it“Dor'tance of the causative factors of soil erosion in the region thereby providing guidance
for conservation programs. However, I have not come across any other quantitative soil

1038 study. The Soil Conservation Research Project (SCRP), funded by the government of
S\Witherland since 1981, and aimed at providing applied research ﬁndings to support the

Soil and water conservation efforts in Ethiopia did not have a research station in Tigray.

 

 

 

 

 

Ill

and
and ;

4;“.

Deforestation

Several areas in Tigray are said to have been covered with forests only about 40
years ago (W olde-Giorgis, 1993). There is virtually no forest area left at present except for
a few remnants at places such as Desa and Hugumburda, and church yards. Cutting trees
for fuelwood, timber and agricultural implements, and clearing forests to expand
agricultural land have exhausted the forest cover of the area. Use of oxen for plowing
particularly needed clear land. While cultivated land was operated under individual
households, forest areas were either communally utilized or open access resources. The
resulting deforestation has caused a shortage of fuelwood forcing households to use animal
dung and crop residues for ﬁle] thereby reducing the organic matter and 'nutrients returned
to the soil.

There has not been any formal policy that deﬁnes ownership or use rights of trees
that grow on cultivated land, communal forests, gullies, wasteland and closed areas. There
iS only an informal institution that allocates ownership and user rights to households when
trees are grown on plots adjacent to the homestead. The price of a load of ﬁrel wood
depends only on the consideration of transportation and labor costs with no value
accorded to the trees themselves. Currently, cutting trees from cultivated land or forest
area is prohibited without obtaining permission from the concerned administrative body.

The development of forest resources in the region requires assurance of ownership
and use rights of trees, an appropriate organizational and legal framework for tree planting
and forest management on communal lands, and/or allocation of communal lands to

individuals for tree planting and grazing purposes. Area closures may result in increased

 

 

 

PH

per

nor

65

pressure on productive land due to grazing requirements unless products from the closed
area are used systematically to feed animals in a cut-and-curry system. This requires an
appropriate policy for the utilization of products from the closed areas as well.

Livestock management and grazing systems need to be accorded serious
consideration in addressing the problem of land degradation. Solving the problem of land
degradation in a signiﬁcant way may be impossible without appropriate policies pertaining
to livestock population and utilization of pasture land. Expansion of the provision of
veterinary services which has been the main focus of livestock development all over

EthiOpia may actually have contributed to land degradation by emphasizing the quantity

rather than the quality of livestock.

Community involvement in soil and water conservation

Farmers in the region apperar to be aware of the problem of land degradation
(REST, 1995). However, pressed by short-term needs, they continue to use cultural
Practices that induce land degradation. Several local expressions indicate how farmers
perceive the problem. Expressions such as ‘ﬂat lands are converted into gullies that could
not even be crossed by small animals’ and ‘small streams we used to cross on foot have
Widened by erosion to become larger than a stone’s throw’ indicate the extent of land
degradation (W olde-Giorgis, 1993). Famers also praise households that protect their land
from soil erosion. Hence there is a favorable condition for mobilizing local communities to

Combat land degradation and develop natural resources.

' .

 

do.
Cor

hm

(ill

66

Farmers in the region have used terracing for soil and water conservation since as
back in history as the Axumite empire, around 700 AD (Oyhus, 1995). Check dams, grass
strips, agroforestry and contour plowing were also used traditionally by farmers in many
parts of the region. The conservation practices currently recommended for use by farmers
are in eﬁ‘ect improved versions of the traditional practices.

Soil and water conservation efforts supported by government or foreign aid donors
are very recent. The United States Agency for International Development (U SAID)
supported the ﬁrst afforestation and terracing program in Tigray during 1971-1974
(Hunting Technical Service, 1976, as cited in Tilahun, 1996). Under this program about
4400 km of bunds were constructed and 3.4 million seedlings planted. After 1974, the
environmental rehabilitation efforts in the region were supported by the UN/FAO FFW
program. During the period between 1983 to 1988, more than 94,000 km of hillside

terraces, 18,000 km bunds and 500 km check dams were constructed (Tilahun, 1996). On
the other hand, the Tigray People Liberation Front (TPLF) was organizing~ soil and water
Conservation programs since 1988 in areas it had controlled. The prominent features of the
TPLF—led programs were community participation in their planning and implementation,
and emphasis on achieving food security through conservation based agricultural
development. This was in contrast with the donor-led programs in the government
Controlled areas which did not involve local communities in their design and
iInplementation.
Due to the highly degraded and barren sloppy areas, all soil and water conservation

Efforts focused on constructing physical conservation structures whether sponsored by the

 

 

Plr

Ph}
90f.
Phy

lEm

 

 

67

government, donors or rebels. No scientiﬁc or systematic study has examined the
eﬁ‘ectiveness and sustainability of the structures. But the urgency of reducing run-off and
soil loss at least for the short term justiﬁes use of the engineering practices. Some zonal
experts in the regional Bureau of Natural Resource Development consider this movement
as the ﬁrst phase in the eﬁ‘ort to combat land degradation, the second phase being

supplementing the physical structures with biological measures.

Since the downfall of the military government in 1991, every able person in Tigray
has been required to provide free labor for soil and water conservation efforts, nearly 50
percent of which is contributed by women. Initially, every able citizen was required to
provide four months of ﬁ'ee labor annually. This was reduced to twenty days in 1995. Such
mass mobilization in campaign form has resulted in the construction of many different
physical conservation structures. During 1992-1995, soil and stone bunds were
Constructed on more than 0.37 million hectares of land (Table 3.7). Lying the emphasis on
Physical structures may result in an immediate reduction of soil loss. However, the long
term effectiveness and sustainability of the structures is uncertain. Hence there is a need to

sl-Ipplement and strengthen them with trees, shrubs, or grasses that have been selected

lOcally using adaptive research.

68

Table 3.7 : Soil and Water Conservation Accomplishments Through Public
Campaign Work in Tigray, 1992-1995

 

 

Type of Work Unit of Accomplishment
measurement
Soil and stone bund hectare 370,432
Terrace maintenance hectare 35,389
Afforestation hectare 3,555
Area closure hectare 125, 000
Diversion ditch kilometer 1,618
Grass strip kilometer 1,646
Check dam kilometer 1,938
P nd construction Unit 28

 

 

 

Source: Teka and Edwards, 1996, p.2

In addition to community involvement, an environmental rehabilitation project
known as Sustainable Agriculture and Environmental Rehabilitation in Tigray (SAERT) is
currently operational in the region. The main objective of the project is to develop small
Scale irrigation schemes. It incorporates the construction of small scale irrigation dams,
conservation of the associated watershed, and where necessary, afforestation. The project

emphasizes capacity building and participation of local beneﬁciaries.

 

69
Institutional aspect of soil and water conservation public campaign work
Since 1992, the soil and water conservation campaign works have been coordinated
at diﬁ‘erent levels by committees composed of experts, administrators and local people. At
the regional level, a soil and water conservation team composed of experts leads the effort
by developing annual plans, and conducting monitoring and evaluation activities. At zonal
and woreda levels, committees are constituted by a chairperson and other members of the
respective administrative body, head of the bureau of natural resources, and representatives
of farmer associations. The committees make detailed conservation plans, and facilitate
and monitor their implementation. The expert from the bureau of natural resources is
responsible for technical aspects. At tabja level, the local committee is made up of the
chairperson and other members of the administration, a development agent, contact farmers
and representatives of farmer associations. The Labia committee closely supervises the
implementation of the plans at village level and reports results. At village level, where the
soil and water conservation work is actually carried out, work is coordinated by a
committee chaired by a development agent and constituted by contact and selected farmers.
Although annual plans for soil and water conservation activities are prepared at the

regional level, they are approved at the village level before implementation. An annual plan
drafted at the regional level will be passed to zonal committee for evaluation and
mOdiﬂcation. The revised version of the plan will again be evaluated and possibly modiﬁed
by the med; and Labia committees before it is passed to the village level. The village

co“'Imittee evaluates the plan in terms of feasibility before it ﬁnally approves it for

““Dlementation.

70

Once a village approves its annual plan, implementation proceeds through several
steps. First, development agents and contact farmers decide on which part of the village
land should be conserved. Next, they estimate the number of person-days that able people
in the village can contribute for the campaign work. Slope measurement, technique
selection, and marking of contours will then follow. Three conservation techniques are
commonly applied, namely, stone terrace, soil bund, and check dams. On hillsides stone
terraces are used. On cultivated lands, either stone terraces or soil bunds are applied. The
choice between stone terraces and soil bunds is made based on the avaialability of stones
and slope degree. Generally, stone terraces are used as there is no shortage of stones for
construction and steep slopes. Checkdams are used to control gullies.

Villagers are grouped into work teams of 7-10 individuals. Since 1995, a man or
youth is required to construct 12 meters of soil bumd or 7 meters of stone terraces per day
while a woman is required to construct 6 meters of soil bund and 3.5 meters of stone
terraces. The village community imposes ﬁnes on those villagers who fail to contribute

labor during the campaign work. Most ﬁnes take the form of a double work load.

3.2.6 Rural Institutions
Land tenure

Prior to the land reform of 1975, there existed private, state, church and communal
based tenure systems. The private tenure system was based on inheritance through the
family who ﬁrst cleared the land or occupied it after the previous tenure rights lapsed.

U“der this system, which was locally called mm, a person who could prove his descent

 

r

‘F

 

71

from the original tenure right holder was entitled for a share of the land. Under the state

ownership system, locally called gum, state ofﬁcials at different administrative levels were

 

entitled to collection of taxes from the tenants who operated the land. Part of the collected
tax could be given to the regional or national government. The church-owned lands
operated in the same way as the state owned lands except that the taxes collected ﬁ'om the
land were used by the churches or else the land would be operated by church servicemen in
exchange for their service. The communal tenure system evolved over time as a
replacement for the other systems when the latter failed to ﬁt the changing social and
environmental conditions of villages. Under the communal system, land distribution was
based on residency in a particular village. It was not uncommon for an individual to own
holdings under different tenure systems.

Under the dsti system, it was the discretion of the father to give land to his children,
and parents had signiﬁcant power over their children due to land inheritance expectations.
The continuous subdivision of land resulted in considerable fragmentation of holdings
(Bruce, 1976). Moreover proving blood relationships would usually end up in court, and
Conﬂicts among relatives based on land claims were common (Wolde-Giorgis, 1993).

Neither of the different land tenure systems that existed prior to 1975 gave tenure
security to farmers. This may have discouraged investments in land development. The
land reform of 1975 by the military government proclaimed that land was the collective
Property of the Ethiopian people. Land ownership was effectively abolished, giving

far“\el's only useﬁ'uct rights. Coupled with the collectivization policy of the government,

 

72

the land reform rendered farmer land tenure even more insecure. The reform could not be
implemented in most parts of Tigray due to security reasons.

However, TPLF itself was an advocate of land reform in areas it controlled. The
TPLF-pioneered land distribution also gave farmers useﬁuct rights to land. The system
was based on two simple principles: (1) a farmer needs to reside as a farmer in a given Labia
to get land ﬁ'om them and (2) land will be allocated to individuals, not to households.
Adults get equal shares and four underage children are counted as one adult. In most
places, the minimum age for qualiﬁcation as an adult for land distribution is 22 years for
male and 15 years for a female. However, the minimum age requirement for entitlement to
land can vary from district to district. It is the district congress, locally called m
balm, that decides on the age limit. Since residence is one criterion for entitlement to land,
the current land tenure system precludes ‘ownership’ of plots in different mains.

Moreover, the system has given legal land tenure rights to women; upon divorce, the
husband-and wife retain their individual shares.

Land is distributed by a committee elected by Labia farmers. The share allotted to an
adult household member is referred to as one-half of a gihrj. The sum of the shares of two
adult household members, therefore, makes up one gibn. Two underage children receive
One-fourth of a gibri. The size of a share of an adult, however, depends on the availability
and fertility status of cultivated land in the Labia. For distribution purposes, cultivable land
in alabia is classiﬁed into three groups based on fertility status as fertile, medium and

clegl‘aded. Mitiku (1995) found that the local land classiﬁcation used for land distributuion

pl"Pesos was consistent with modern scientiﬁc classiﬁcation system.

 

 

73

Each household is allotted land from each category. A considerable variation may
exist in the size of a share between different tabias. Usually the size of land allocated from
the degraded category is larger in size. Land holding per household also varies between
administrative zones. In the western zone, the land holding per household is 2.4 ha, while
the corresponding ﬁgures in the central, eastern, and southern zones respectively are 1.06
ha, 0.9 ha, and 1.25 ha (Regional Bureau of Agriculture, 1995). The regional average of
land holding size is 1.2 ha.

Land was redistributed in order to accommodate young families and returnees from
settlement areas of the military government. This had resulted in further fragmentation of
holdings. Although there is no policy that formally establishes that there would be no
further land distribution, land distribution is temporarily halted in the region and several
new families and returnees have not been allocated land.

Land leasing is also widely practiced in the region. Households who are unable to
operate their own land may lease it for a share of the harvest. Lease agreements specify
input contributions, including labor and seed, as well as the division of the harvest. The
prevailing arrangement is to share output equally between the share cropper and the land
owner. In some places, a renter may pay a lumpsum payment.

Like its predecessors, the current land tenure system does not guarantee farmers
full security in their holding. In the absence of a law that prohibits further redistribution of
land, farmers expect that land can be redistributed at any time. This uncertainty seems to
have limited land development efforts mostly to plots that are adjacent to the homestead,

because of the relative security farmers have on the nearer plots. In conclusion, the

 

 

 

74

absence of tenure security for many generations in the region is likely to have contributed

signiﬁcantly to the land degradation process.

Agricultural credit

The predominant sources of agricultural credit in Tigray have been relatives, ﬁiends
and merchants. Although the regional branches of government banks such as Commercial
Bank of Ethiopia (CBE) and Development Bank of Ethiopia (DBE) are mandated to
provide credit to farmers, the operational procedures of these formal ﬁnancial institutions
have not been suited to the credit needs of farm households, particularly the small farmers
which constitute the bulk of the rural population. In most cases, collateral requirements
have put the ﬁnancial institutions out of farmers’ reach. A study conducted in central
Tigray (REST, 1995) indicated that relatives and friends made up for more than 65 per cent
of the sources of credit in the study area.

After many years without a formal ﬁnancial institution that meets the credit needs of
farmers, REST initiated a rural credit scheme in 1994 designed to meet these needs. The
s‘ih'i’the combines both credit and savings services. The organizational structure of the
credit scheme stretches from the head ofﬁce at Mekele to the mobile branch offices at the
tibial level. Each branch ofﬁce is staffed with a’branch manager, an accountant and a
caSl'lier. Loans are processed and savings accepted at the branch level.

Priority in loan provision is accorded to women and the relatively poorer farmers

when poverty is deﬁned based on oxen ownership or equivalent wealth. For this purpose,

be“efrciaries are selected by the whole community. The selected individuals will then make

 

 

 

 

sch
Per.
and

l8pa

A19m:
resear.
madn-
Collegé
Ag’lCul:

minim.

75

up credit and saving groups, and a collection of groups will make up credit and saving
centers. Group members approve loan requests by individuals. Groups and centers are
collectively responsible for loans taken by individuals. Although no collateral in the form of
physical capital is required, group pressure and group savings serve asindirect forces to
facilitate repayment. Ma and mm congresses work in close collaboration with the
branch ofﬁces and local community in the process of beneﬁciary selection, coordination of
inter— and intra-group activities, monitoring of credit use, and repayment enforcement in
case of potential default.

Although it is too early to evaluate the performance of this innovative credit
scheme, results of the ﬁrst few years of operation appear to indicate promising
performance. Until late 1995, 14000 households in twenty different districts were served
and 19 million Birr was disbursed. Out of this, 46 percent was received by women. A

repayment rate of 93-98 per cent was recorded (Gebremedhin et al., 1996).

Agricultural extension and farmer cooperatives

, Agricultural extension service in Ethiopa was ﬁrst provided in 1951 under the then
Alemaya College of Agriculture, where it was coordinated with the college’s teaching and
research activities. The extension organization has since gone through successive
modiﬁcations. In 1961, the Ministry of Agriculture took the responsibility from Alemaya
College. In order to generate technology for diffusion to farmers, the Institute of
Agricultural Research (IAR) was established in 1966, administered by a board chaired by

the minister of agriculture. Since this time, integrated rural development programs and/or

 

 

 

 

eCr

fol]
Il'ler

Sen,

i”Du

l9§r<

76

extension programs aimed at diffusing package of technologies have pre-dominated.
However, most of these agricultural extension efforts did not operate in Tigray for various
reasons.

During the 1980's, TPLF provided extension service in areas it liberated, placing
major emphasis on soil and water conservation. In the absence of modern technologies
tested for use in the region, the extension program focused sound indigenous practices such
as land preparation, timely operation of farm activities, seed selection and pest control. By
1991, on the eve of the downfall of the military government, the program was operating in
almost all parts of Tigray.

Currently, a modiﬁed version of the training and visit (T&V) extension model is
being used in the region. The whole region is classiﬁed into extension centers. Each
extension center, which is a collection of villages, has three development agents, one each
for agricultural production, natural resource development and home econorrrics. Each
village has seven contact farmers of which two each are for agricultural production, home
econonrics, and natural resources development, and one for plant protection. Development
agents receive two days training each month from woreda level agricultural experts. This is
followed by a meeting of the development agents and contact farmers. The contact farmers
then meet with the village community every two weeks. The focus of the extension
services in the region are agronomic and crop protection demonstrations, distribution of
improved seed, distribution of forage and horticultural seedlings, provision of modern
inputs and associated credit, formation and development of farmer cooperatives. By late

1995, about 300 extension centers were operating, in which 65 half-hectare demonstration

in

 

77

plots managed by the bureau of agriculture, 1 566 one-tenth hectare demonstration plots
for cereals on farmers ﬁelds, and 104 ﬁfty square meter farmer ﬁeld demonstration plots for
horticultural crops were used (Meles and Hagos, 1995).

Four types of farmer co-operatives function currently in the region viz.,
multipurpose cooperatives, handicrafts producers’ cooperatives, ﬁshers’ cooperatives and
water users’ associations. Multipurpose cooperatives provide members with diverse
services including agricultural inputs and consumer goods. Water users associations are
aimed at coordinating the use of irrigation water that is being developed as a result of the
SAERT project. In 1995, about 150 multipurpose cooperatives and 47 water users

associations were functioning with total membership of 98,700 and 3,280 respectively.

3.3 Summary

Soil erosion is the most serious environmental problem in Ethiopia and cultivated
land is the major contributor to soil loss. Steep slopes and erosive rainfall are the major
physical factors of soil erosion. To these is added intensive cultivation of land due to
increasing population pressure. The huge livestock population is another contributor to
soil erosion due to overgrazing.

Agricultural policies of the previous governments had left small holder agriculture
out of focus of development programs. Farmers lacked land tenure security. The
marketing policies of the military government depressed the proﬁtability of small holder
agriculture. As a result, small holder investment in land improvement practices had been

very low.

 

78

Public intervention in soil conservation started in the early 70's, mostly in the form
of FFW projects. However, the interventions used technical packages that were not tested
on-farm and which lacked adaptability to the diverse agroecological conditions.
Involvement of local people was also minimal.

The Tigray region, which receives about 30 percent less rainfall than the national
average, is one of the most severely degraded areas in Ethiopia. Moreover, the region was
not a signiﬁcant beneﬁciary of public interventions for natural resource conservation during
the past regimes. Conservation and development of the natural resources and rehabilitation
of the environment has been at the center of any development policy in the region since

1991, after the dawn fall of the military government.

CHAPTER 4

FARMING SYSTEMS CHARACTERISTICS OF THE STUDY AREA

This chapter characterizes the farming systems in the study area using the
descriptive statistics ﬁ'om the household survey and econometric models of availability and
use of farm inputs. Land tenure and crop production; availability and use of draft power,
local and improved seed, commercial fertilizer; and household involvement in oﬂ‘-farm
employment are described. Farmer ratings of development constraints and priorities for

government intervention are also discussed.

4.1 General Characteristics of the Surveyed Villages

The study area, in south central Tigray, ranges in altitude from 1560 to 2510 meters
above sea level. Areas between 1500 and 2500 meters above sea level are categorized as
intermediate highlands and those above 2500 meters are in the highland zone. Three of the
thirty villages included in the survey were in the highland zone. According to the villagers,
the predominant soil textures were sand and clay and the predominant topography of
twenty-three of the villages was hilly or intermediate hilly. The average walking time ﬁom
the surveyed villages to the nearest market place and all weather road, respectively, were
1.6 and 1.5 hours.

Land degradation has forced villagers to depend more on animal dung as a source

of cooking fuel. Seventy-ﬁve percent of the villages depended primarily on animal dung for

79

80

fuel while the rest used fuel wood from forests and bushes. To fetch fuel wood (without
considering the time spent to collect the wood), the average round trip walking time was
6.3 hours. In some villages, residents had to travel for as long as 12 hours. In the highland
zone, the average walking distance was 5 hours while the corresponding ﬁgure for the
intermediate highland zones was 7 hours. However, animal dung was reported as the only
primary source of ﬁre] in all the villages in the highland zone while twenty-eight percent of
the villages in the intermediate highland zone depended primarily on forests and bushes for
fuel. Eighty percent of the villages had closed areas (areas protected ﬁ'om human or
livestock intrusion) with the duration of area closure ranging ﬁ'om 1 to 17 years. The
beneﬁts villagers get from the closed areas included feed and construction wood.

The number of households living in a village ranged from 59 to 610 with an average
of 242 households per village. Male-headed households accounted for 83 percent of the
total households. The average family size was 4.8, and the average female and male
members per household were 2.4 and 2.6 respectively. The number of working age
household members (IS-64 years old) averaged 3 per household. Only twenty three
percent of household heads were able to read and write. The age of household heads
ranged ﬁ'om 20 to 78 with an average of 47 years. Twenty-ﬁve percent of the households
reported sending a child to school. FFW projects were operational in seventeen of the

villages during 1992-95.

 

81

4.2 Land Tenure

The national policy on land tenure stipulates that farmers have usuﬁ'uct rights on
land, but may not buy or sell it. In principle, a farmer is entitled to get land as long as his
primary Occupation is farming and he dwells in a village. In Tigray, redistribution of land
has been halted for some time now. In the survey area, the time since last land distribution
up to 1995, ranged ﬁom 2 to 12 years and the time of operation of a plot of land by a given
household ranged ﬁom 1 to 38 years. During the 1995/96 crop season, the average area of
cultivated land per household was 1.7 hectares, made up of 3.5 plots with an average plot
size of half a hectare. Some farmers cultivated as many as 11 plots. The average cultivated
land per capita was 0.4 hectares. Since farmers are entitled to land in their Labia, most
plots cultivated were not far away from their villages. The average walking distance to a
plot from homestead was half an hour. Plots with slopes as high as 23 degrees were
cultivated by the surveyed households.

The average land share of a couple (husband and wife) was 1.3 hectares, and the
average share of a child 0.2 hectares. The average share of a couple was lower in the
highland zone (1.0 hectare) as compared to the intermediate highland zone (1.3 hectares).
The minimum age for entitlement to land ranged from 12 to 17 years for females and from
20 to 26. years for males.

Land leasing is commonly practiced in the area. Twenty-six percent of surveyed
farmers reported having leased land during the 1995/96 crop year. The average land size
leased out was 0.7 hectares. The most common rental form for land lease is share

cr0pping, the dominant ratios of share being 1:4 and 1:3 between the owner and the

 

82

operator (Table 4.1). The most important reasons for leasing out land were lack of draft

power, labor and seed in that order.

Table 4.1. Rental Form for Land Leasing and Share Crop Ratio

 

 

 

 

 

 

 

 

 

(n=6l) (n=64)

Share cropping 1:5 6.6
1:4 36.1 85.9

1:3 32.8

1:2 24.6
Share cropping and lumpsum payment 10.9
Fix mount in kin 3.1

 

 

 

In addition to farmers responses about why they lease land out, a probit model was
estimated to determine the factors associated with land lease practice after other factors
were controlled for. Land leasing was expected to be determined by agroecological
condition and village characteristics, land area owned, labor availability, and household
demographic characteristics.

Location in a highland zone (HIGHLAND) was expected to reduce the likelihood
of leasing out land due to the relative land scarcity compared to the intermediate highland
zone. Distance of village from market place (MKTDIST) and all weather road

(ROADDIST) were expected to encourage leasing out land due to their effect in reducing

83

the return to agriculture. Use of animal dung as primary source of fuel in a village
(FUELDUNG) was expected to encourage leasing out due to the relative severity of land
degradation which could reduce agricultural productivity. Land area owned (LANDOWN)
was expected to encourage land leasing. Number of working age family members
(WORKERS) should discourage leasing out land due to the availability of labor for farm
activities. Among the demographic variables, male household head (MALEHEAD) was
hypothesized to reduce leasing out as male headed households would be expected to have
higher access to resources and age of household head (AGEHSI-IED) to increase land lease
for the opposite reason. The expected effect of literacy of household head (LITERACY)
was ambiguous because literate households could have better access to resources or more
likely to be involved in non-farm activities. Having access to credit (CREDITAC) should
discourage leasing out since it eases liquidity problems. Explanatory variables are deﬁned

in Table 4.2 below.

Table 4.2:

84

Deﬁnition and Measurement of Explanatory Variables

 

 

 

 

 

Variable code Deﬁnition Measurement Level of
observation
Village=VILL
Household= HD
HIGHLAND Agroecological =intermediate VILL
Zone highland
l=highland
MKTDIST Distance of village walking hours VILL
ﬁ'om nearest market
ROADDIST Distance of village Walking hours VILL
ﬁom nearest all-
weather road
FUELDUNG Major source of 0=forest and bush VILL
cooking fuel l=animal dung
AGEHSHED Age of household years HD
head
MALEHEAD Sex of household 0=female HD
head 1=male
LITERACY Literacy of =illiterate HD
household head 1=literate
SIZEHSHED Size of household number HD
WORKERS Number of working Number of people HD
age household aged 15-64
members
LANDOWN Owned land Hectares HD
LANDSIZE Size of cultivated Hectares HD
land
CREDITAC Ifhousehold had =no HI)
access to formal or 1=yes
inform l a . Credit

 

85

The following probit speciﬁcation was estimated to determine the characteristics of
farmers who leased land out.
LEASEOUT = F( HIGHLAND, MKTDIST, ROADDIST, FUELDUNG, LANDOWN,
WORKERS, AGEHSHED, MALEHEAD, LITERACY, CREDITAC)
LEASEOUT refers to leasing land to others during the 1995-96
cropping year (0=no, 1=yes), and all explanatory variables are as deﬁned in

Table 4.2 (page 84).

Eight of the ten explanatory variables had the expected signs. Highland zone and
sex of household head signiﬁcantly explained the leasing out of land (Table 4.3). Farm
households residing in the highland zone had less probability of leasing land out as
compared to households who lived in the intermediate highland zone. This may be due to
the relative land scarcity that exists in the highland zone. Female headed households are
more likely to lease land out because they possess inadequate labor and/or draft power.
Male household head had the highest marginal effect (-0.39) on a the probability of leasing
out land by a household, followed by highland zone (-0.21). An average farmer had 14

percent probability of leasing out land.

86

Table 4,3: Probit Model Predicting Land Leasing Practice of Farmers
(robust standard errors)

 

 

 

 

 

 

Indep. Variable Coeﬁcient Marginal
effect (dF/dx)
HIGHLAND -1.034 (.422)" -.211
MRKTDIST -.225 (.173) -.061
ROADDIST .038 (.089) .010
FUELDUNG -.258 (.250) -.074
LANDOWN .003 (.110) .001
WORKERS -.099 (.106) -.026
MALEHEAD . -1.148 (.275)*** -.393
AGEHSHED .010 (.008) .003
LITERACY .190 (.267) .053
CREDITAC -.257 (.229) -.066
Constant .688 (.541) ---
Chi-square 38.26 ---
Prob. > chi-square 0.0000 «-
Pseudo R-square 0.1912 ---
Predicted prob. at x-bar --- .136
N 215 «-

***, ** signiﬁcant at 1% and 5% respectively.

Farmers in the survey area feel more insecure about their holdings in the long run
than in the medium term. Sixty percent of farmers responded that they felt certain that they
would cultivate their plots ﬁve years from 1995 while only 41.7 percent felt sure of being

able to leave their plots to their children (Table 4.4).

87

Table 4.4: Perceived Tenure Security Among 243 Interviewed Farmers, 1995-96

 

 

 

Security Type Percent of farmers
(N=243)
Will cultivate plot in ﬁve years Yes 60.5
No 6.2
Am not sure 33.3
Will leave plot to children Yes 41.7
No 24.4
Am not sure 33.9

 

 

 

Farmers’ estimates of land area did not show strong correlation with actual
measurements of ﬁeld sizes. Farmers were asked to report the size of each plot of land
they cultivated in the 1995/96 crop year in mm, a local measure of land area equivalent
to the area that could be ploughed using a pair of oxen in a day. Four mm is
approximately equivalent to one hectare. The farmers’ estimates were converted into
hectares using this equivalence. The size of each plot was again measured directly in
hectares. A comparison of fanners’ estimates and the direct measurement showed a
correlation coefﬁcient of 0.67. This result indicates that farmers’ estimate of land size

needs to be taken with caution.

88

4.3 Labor use

The primary activities involving adult males include land preparation, cultivation
and harvesting, while adult females are involved primarily in household and cultivation
activities. Fuelwood collection is done by all household members. Work prohibition on
religious days is widely observed in the Tigray region. In the study area, the number of
such days excluding weekends ranged from 0 to 11 per month. Weekends are not working
days.

Farmers in the area experience labor shortage for farm operations. Fifty eight
percent of surveyed farmers reported labor shortage for the 1995/96 cropping season of
which only 30 percent used hired labor while 52 percent used exchange labor or help ﬁ'om
ﬁiends and relatives to alleviate the problem. Casual hired labor is mainly male. While
permanent hired labor was reported to be scarce throughout the year, casual hired labor

was abundant except for the months of September to December (Table 4.5).

Table 4.5: Availability of Hired Labor as Rated by Respondents

 

 

 

 

 

 

 

Abundant Scarce Seasonally scarce
Permanent 12.7 82.4 4.9
1 51 0 29.6 19.3

 

 

"'Li‘a ‘ 1..
Witch Ad

@3153er

 

 

 

 

89

Casual hired labor is mainly used for cultivation and harvesting. The crops for
which labor shortage was more prevalent were teff, wheat and barley. Farmers reported
the need for additional labor especially for weeding, harvesting and threshing during the
months of July to November. In addition to agricultural activities, 29 percent of
households reported participation in off-farm work, the most important being construction,

trade and selling ﬁrewood (Table 4.6).

Table 4.6: Households’ Off-farm Involvement by Type of Work

 

 

Type of work Percent of households
involved (N=71)
Trade 25.4
Construction 43.7
Food processing 2.8
Selling fuelwood 12.7
Fetching salt I 1.4
Other“ 14.0

 

 

—
* includes, among others, crafting, quarrying, spinning and hair dressing

In order to identify the factors associated with participation in off-farm activities a
probit model was estimated. Off-farm work was expected to be determined by
agroecological condition and village characteristics, land area owned, family labor

availability, and household demographic characteristics.

Ar
08
(11

W01

9O

Farming in a highland zone (HIGHLAND) should reduce off-farm involvement due
to higher land productivity. Distance of village from market place (MKTDIST) and all-
weather road (ROADDIST) were expected to reduce participation due to possible lack of
oﬂ'-farm employment opportunities. Villages with more degraded soil proxied by use of
dried dung for cooking fuel should encourage off-farm work due to lower land
productivity. Age of household head (AGEHSHED) was expected to have an ambiguous
effect because such households could have more experience in off-farm activities or less
available family labor. Land area owned (LANDOWN) was hypothesized to reduce
participation as wealthier households should have less need for off-farm income. The
impacts of sex of household head (MALEHEAD) and availability of family labor
(WORKERS) were ambigous. Male-headed households might have less need for additional
income than female-headed households, but they are more likely to have off-farm skills.
Availability of fanrily labor could increase farm income but also encourage participation in
off-farm activities due to the availability of labor. Literacy of household head
(LITERACY) should increase involvement as literacy could result in skills for oﬂ‘-farm
work. The impact of access to credit (CREDITAC) was ambiguous as credit can
encourage investment in both farm and off-fann activities.

The following probit speciﬁcation was estimated to identify the factors inﬂuencing
household involvement in off-farm work:

OFF-FARM = F(HIGHLAND, MKTDIST, ROADDIST, FUELDUNG,

LANDOWN, WORKERS, AGEHSHED, MALEHEAD,

LITERACY, CREDITAC), where

 

91

OFF-FARM refers to household involvement in off-farm activities
during the 1995-96 cropping year (0=no, 1=yes). All explanatory variables
are as deﬁned in Table 4.2 (page 84).

Participation in oﬁlfarm work was explained negatively by distance of village from
market place and positively by major source of fuel (Table 4.7). Households living in
remote villages were less likely to participate in off-farm employment perhaps due to
limited employment opportunities. Those living in more degraded environments were more
likely to participate in off-farm employment. Relative degradation of land had the highest
marginal eﬂ’ect (0.17) on the probability of household participation in
oﬁlfarm employment. An average farmer had 27 percent probability of participating in off-

farm employment.

92

Table 4.7: Probit Model Predicting Participation in Off-farm Employment
(robust standard errors)

 

 

 

effect

(dF/dx)
HIGHLAND -.l33 (.304) -.042
MKTDIST -.357 (.161)" -.118
ROADDIST -.038 (.100) -013
FUELDUNG .604 (.313)* .172
LANDOWN .178 (.112) .058
WORKERS -.091 (.099) - .030
MALEHEAD .223 (.320) .069
AGEHSHED .008 (.007) .002
LITERACY . .238 (.220) .081
CREDITAC -.121 (.205) -.039
Constant -1.050 (.631) «-
Chi-square 21 .73 ---
Prob. > chi-square 0.016 ---
Pseudo R-square 0.088 «-
Predicted prob. at x- --- .266
bar
N 215 ---

 

 

** ," Signiﬁcant at 5% and 10% respectively.

4.4 Crop and l

Barley i
Pulses and oil 5
disappearing dr
“1°51 Commoru‘:
Sorghum. Dror
Varieties. Othe
and 1a"d search

 

 

The ma
use of some im
improved Seed,
most lmponam
regiOnal [mm
Consisrem With
plltchasing p0

(WW).

93

4.4 Crop and Livestock Production

' Barley is the most widely grown crop in the study area, followed by wheat and teﬁ’.
Pulses and oil seeds are also grown on limited areas. Some crop varieties have been
disappearing due to environmental and agricultural conditions. Wheat varieties are the
most commonly mentioned disappearing crop varieties followed by teﬂ’ , barley and
sorghum. Drought is the most important reason that induces abandonment of crop
varieties. Other important reasons include water logging, plant diseases, decreasing yield
and land scarcity. Farmers shift into growing different varieties notably ones that give high
yield, tolerate drought, and resist water logging.

The majority of farmers surveyed used local seed with only 22 percent reporting
use of some improved seed during the 1995/96 cropping year. Of those who used
improved seed, 81 percent used wheat varieties and 17 percent used Leif varieties. The
most important source of improved seed for farmers was purchase on credit from the
regional bureau of agriculture. Provision of improved seed on credit to farmers was
consistent with the reason reported by farmers for not using improved seed. Lack of

purchasing power was reported to be the primary reason for inability to use improved seed

(Table 4.8).

hOU 5‘

94

Table 4.8: Improved Seed: Sources and Reasons for Not Using

 

 

 

 

 

 

Source of and reason for not using improved seed Percent of
farmers
Source of seed Own production 14.0
Purchase on credit 72.0
Reason for not using Was not available 12.8
(N= 179) Lack of purchasing power 49.7
Did not want to use 22.0
Did not know its beneﬁt 15.6

The following probit model was estimated to study the factors associated with
households’ adoption of improved seed:
SEEDIMP = F(HIGHLAND, MKTDIST, ROADDIST, FUELDUNG,
AGEHSHED, MALEHEAD, LITERACY, WORKERS,
LANDSIZE, CREDITAC), where
SEEDIMP refers to household use of improved seed during the 1995-96
cropping year (0=no, 1=yes). All explanatory variables are as deﬁned in
Table 4.2 (page 84).

However, the regression was not signiﬁcant, so results are not reported.

95

Although farmers prefer to use local seed, not all surveyed farmers were self
suﬁcient in seed; 44 percent of farmers did not have adequate local seed of their own
during the 1995/96 cropping year. The shortage of local seed was equally spread among
the important crops including barley, wheat, Leﬁ‘ and sorghum. Farmers’ main sources of
additional seed were market and credit from local areas.

A probit model was estimated to identify the factors associated with farmer self-
sufﬁciency in local seed. Self-sufﬁciency in local seed was hypothesised to be a product of
agroecological and village characteristics, area of cultivated land, availability of family labor
and household demographic charateristics.

Location in a highland zone (HIGHLAND), male-headed households
(MALEHEAD), higher family labor (WORKER), and larger cultivated land area
(LANDSIZE) were expected to increase households’ likelihood of being self-sumcient in
local seed through their impact on increasing production. Villages that are more distant
from roads (ROADDIST) and markets (MKTDIST) should increase probability of self-
sumciency due lessintegration into market. Villages that are more degraded
(FUELDUNG) should decrease self-sufficiency through reduced production. Larger family
size (SIZEHSHD) was expected to reduce self-sufficiency as it would mean more mouths
to feed. The effect of age (AGEHSHED) and literacy (LITERACY) of household head
were ambiguous. Old age could mean more experience but also less family labor. Access
to agricultural credit (CREDITAC) was expected to reduce self-sufﬁciency as farmers

would rely on credit. for additional seed.

96

The following probit speciﬁcation was estimated:
LOCALSEED = F(HIGHLAND, MKTDIST, ROADDIST, FUELDUNG,
AGEHSHED, MALEHEAD, LITERACY, SIZEHSHD,
WORKERS, LANDSIZE, CREDITAC), where
LOCALSEED refers to farmer self-sufficiency in local seed during the
1992-95 cropping year (0=no, 1=yes). Explanatory variables are as deﬁned

in Table 4.2 (page 84).

Farming in a highland zone, male household head and more family labor all
increased a household’s likelihood of being self-sufficient in local seed as expected (Table
4.9). Distance to market also increased the likelihood, perhaps because less market access
forces households to reserve their own seed. Age of household head reduced self-
suﬁciency, perhaps due to less access to resources. Larger households were less likely to
be self-sufficient in local seed. Seed reserves may be consumed during planting time by
larger households who have more mouths to feed. An average farmer had 57 percent
probability of being self sufﬁcient in local seed. Highland zone had the highest marginal
effect (0.35) on the likelihood of a household’s self-sufﬁciency in local seed, followed by

male household head (.29), and distance of villages ﬁom market place (0.17).

Table 4.9:
(robust standard errors)

97

Probit Model Predicting Household Self-Sufficiency in Local Seed

 

 

 

 

 

Independent variable Coefficient Marginal effect
(dF/dx)
HIGHLAND 1.002 (.292)*** .350
MRKTDIST .443 (.149)*** .174
ROADDIST -.081 (.090) -.032
FUELDUNG .192 (.237) .076
AGEHSHED -.014 (.008)* -.006
MALEHEAD .747 (300)" .289
LITERACY -.169 (.223) -.067
SIZEHSHD -.211 (.069)*** -.083
WORKERS .246 (125)" .097
LANDSIZE .054 (.106) 0.021
CREDITAC -.008 (.208) -.003
constant -.543 (.554) ---
chi-square 36.06 «-
prob. > chi-square 0.0002 «-
pseudo R-square 0.1229 ---
predicted prob. at x-bar --- .568
N 214 «-

 

_
‘, ", "" signiﬁcant at 10%, 5% and 1% respectively.

98

Fertilizer use by farmers is limited. Only 28 percent of surveyed farmers reported
having used fertilizer during the 1995/96 crop year, and not all of them apply it at the
recommended rate. Most fertilizer was applied on Leﬂ’ ﬁelds due to the high cash value of
the crop. As' in the case of the use of improved seed, the most important source of
fertilizer for farmers was purchase on credit ﬁom the regional bureau of agriculture. The
most common reason for not using fertilizer was lack of purchasing power. Manuring was
reported to have been practiced by 40 percent of the farmers, although use of animal dung
for ﬁrel is rendering the practice less frequent.

A probit model was estimated to determine the factors associated with farmer use
of conunercial fertilizer. Use of commercial fertilizer was hypothesized to depend on
agroecological and village and village characteristics, area of cultivated land, availability of
family labor and household demographic charateristics.

Location in a highland zone (HIGHLAND), male household head (MALEHEAD),
more family labor (WORKERS), larger size of cultivated land (LANDSIZE), and access to
credit (CREDITAC) were expected to increase probability of fertilizer use through their
impact on increasing production. Literacy of household head (LITERACY) was expected
to increase fertilizer adoption due to its impact on access to information. Age of household
head (AGEHSHED) was expected to reduce the likelihood of fertilizer use. Villages that
are more degraded (FUELDUNG), and more distant from roads (ROADDIST) and market
places (MKTDIST) should reduce fertilizer adoption as they reduce productivity or the

return to agriculture.

99

FTZEUSE = F(HIGHLAND, MKTDIST, ROADDIST, FUELDUNG,
AGEHSHED, MALEHEAD, LITERACY, SIZEHSHD,
WORKERS, LANDSIZE, CREDITAC), where
FT ZEUSE refers to household’s use of commercial fertilizer during the
1992-95 cropping year (0=no, lacs). Explanatory variables are as deﬁned

in Table 4.2 (page 84).

Farming in a highland zone, distance of village from an all weather road, cultivated
land area were the signiﬁcant variables (Table 4.10). Highland zone reduces the likelihood
of fertilizer use, contrary to expectation. Farmers who cultivated more land were more
likely to use commercial fertilizer. Distance of villages from all weather road discouraged
fertizer use perhaps by reducing the return to agriculture or adding to fertilizer transport
costs and delays. The probability that an average farmer adopts commercial fertilizer was
27 percent. Highland zone had the highest marginal effect (-0. 16) followed by cultivated

land area (0.09).

100

Table 4.10: Probit Model Predicting Household use of Commercial Fertilizer
(robust standard errors)

 

 

 

 

 

 

Independent Coefficient Marginal effect
variable ‘ (dF/dx)
HIGHLAND -.544 (.298)* -. 161
MRKTDIST .194 (.162) .064
ROADDIST -.231 (.107)" -.076
FUELDUNG .159 (.261) .051
AGEHSHED .001 (.008) .000
MALEHEAD .219 (.342) .068
LITERACY .338 (.226) .117
WORKERS -.045 (.089) -.015
LANDSIZE .289 (.106)*** .095
CREDITAC .-055 (.219) -.018
Constant -1.50 (.586) m
Chi-square _ 28.40 m
Prob. > chi-square 0.0016 «-
Pseudo R-square 0.114 ---
Predicted prob. at --- . .269
x-bar

N 213 --

_
"*f‘g * signiﬁcant at 1%, 5% and 10% respectively.

The most important livestock reared in the area are cattle, followed by goats and

sheep. The'major reason for raising cattle is to obtain draft power for plowing. Goats and

101

sheep are primarily used as source of cash for the household. Most of the villages had a
delineated grazing area. However, 71 percent of them reported inadequacy of grazing area
during the rainy season while 58 percent reported shortage during dry season. The major
reasons mentioned by farmers for shortage of grazing land during rainy season were
extensive cultivation and reservation for the dry season, while the corresponding reasons
for the dry season were land degradation and area closure. Farmers supplement pasture
with straw and hay. Some farmers also send their livestock temporarily to places which
have more livestock feed.

With remarkable equity in land distribution in the study zone, the availability of
draft power and labor are becoming important factors of social differentiation. Forty-one
percent of the households had no ox and the percentage of households who had one ox and
a pair of oxen were 31 and 25 respectively. Only 3.5 percent of households had three or
more oxen. Oxen ownership appears to show different patterns between the highland and
intermediate highland zones. The proportions of households with no ox at all or a single ox
is higher in the highland zone than in the intermediate highland zone. The proportion of
households who have two or more oxen is higher in the intermediate highland zone.

Sixty-two percent of farmers reported a shortage of draft power for the 1995/96
cropping season. The most common solution used by farmers for inadequate draft power is
to arrange an exchange of ox among households; 64 percent reported using this system.

A probit model was estimated to determine the factors associated with household
self-suﬂiciency in draft power. Self-sufﬁency in draft power was hypothesized to depend

on agroecological and village characteristics, cultivated land area, availability of family

102

labor, and household demographic charateristics. Farm location in a highland zone
(HIGHLAND), male household head (MALEHEAD), availability of family labor
(WORKERS), area of cultivated land (LANDSIZE) and access to credit (CREDITAC)
were expected to increase the likelihood of self sufﬁciency in draft power due to their
impact on increasing production or access to ﬁnancial resources. On the other hand,
distance from market (MKTDIST) and roads (ROADDIST), relative degradation of
village lands (FUELDUNG) were expected to decrease the probability of a household’s
being self-sufﬁcient in draft power due to their impact on reducing agricultural productivity
or return to agriculture. Age of household head (AGEHSHED) was expected to reduce
self-sufﬁciency as older people could be expected to have less farm resources. The impact
of literacy of household head (LITERACY) was considered ambiguous.
The following probit speciﬁcation was estimated:
DRFTPOWER = F(HIGHLAND, MKTDIST, ROADDIST, FUELDUNG,
AGEHSHED, MALEHEAD, LITERACY, WORKERS,
LANDSIZE, CREDITAC), where
DRFTPOWER refers to households self-sufﬁciency in draft power during
the 1995-96 cropping year (0=no, lacs). Explanatory variables are
deﬁned in Table 4.2 (page 84).
Agroecological zone, age, sex and literacy of household head, availability of family
labor and area of cultivated land explained self-sufﬁciency in draft power signiﬁcantly
(Table 4.11). Households in the highland zone were more likely to be self-sufﬁcient in

draft power probably due to higher land productivity. Male-headed households had higher

chance of bein,
household hea.
land was posiri
draft power se
inversely relate
sumClent in dr;
SuﬁlCient in let

followed by hj S

 

103

chance of being self-suﬁcient due to better access to resources while older age of
household head reduced the likelihood perhaps for the opposite reason. More cultivated
land was positively associated with having adequate draft power. Family labor increased
draft power self-sumciency, perhaps due its impact on production. Literacy of beads was
inversely related to self-sufficiency. The probability of an average household being self
Meient in draft power (34 percent) was much less than the probability of being self-
suﬁicient in local seed. Male head of household had the highest marginal effect (0.37)

followed by highland zone (0.29).

104

Table 4.11: Probit Model Predicting Household Self-Sufﬁciency in Draft Power
(robust standard errors)

 

 

 

 

 

 

Indep. variable Coefﬁcient Marginal effect
(dF/dx)
HIGHLAND .756 (.289)*** 0.289
MRKTDIST .191 (.167) 0.070
ROADDIST .057 (.098) 0.021
FUELDUNG .251 (.256) 0.089
AGEHSHED -014 (.008)* -0005
MALEHEAD 1.553 (.506)*** 0.367
LITERACY -.418 (.220)* -0145
WORKERS .232 (092)" 0.084
LANDSIZE .435 (.125)*** 0.159
CREDITAC -.080 (.212) -0.029
Constant -3.22 (.705) ---
Chi-square 43.06 ---
Prob. > chi-square 0.0000 ---
Pseudo R-square 0.192 «-
Predicted prob. at --- .338
x-bar
N 214 ---

***, * signiﬁcant at 1% and 10% respectively

Although less serious than the shortage of draft power, a lack of adequate

agricultural implements is also experienced by farmers. In the study area, this was reported

105

by 30 percent of the farmers. Unlike the case with draft animals, farmers can easily borrow
additional implements from ﬁiends and relatives. Of the farmers who faced shortage of

agricultural implements, 39 percent reported that they did not have enough plow shares.

4.5 Farmers’ Perceptions of Development Constraints and Areas of Institutional
Intervention

Farmers in the study area perceive that the most important development constraints
for agriculture are, in order of importance, drought, pests and land scarcity (Table 4.12).
Despite these perceived constraints, farming was rated the highest among investment
priorities of farmers. Sixty-four percent of surveyed farmers reported that agriculture was

their highest priority for investment, followed by household investment (20 percent), off-

farm work (15 percent), and education of children (1 percent).

Table 4.12: Development Constraints and Priorities for Government
Assistance in Agriculture (farmers’ rating)

*

 

 

 

 

 

 

Development ' Percent of response Priority for Percent of response
constraint (N=401)* intervention (N=621)*
Drought 32.4 Credit for oxen 27.5
Pests 18.5 Credit for seed 22.9
Land scarcity 16.2 Supply of inputs 20.5
Water logging 10.2 Development of

irrigation 11.4

* Percentages do not add up to 100 because only the top four categories are included.

Int

 

power at tl
irrigation \l

haxing acct

4'6 Fﬁrmer

106

In their priorities for government support, the surveyed farmers put credit for draft
power at the top, followed by credit for seed, supply of farm inputs and development of
irrigation water supply (Table 4.12). Twenty-nine percent of surveyed farmers reported

having access to agricultural credit during the 1995/96 cropping season.

4.6 Farmer Involvement in Soil and Water Conservation

Currently the government and non-govemment organizations are implementing a
wide array of soil and water conservation efforts in the region, mainly in the form of
extension services, and assisting and organizing community work. Communities contribute
free labor for conservation activities in the form of campaign work. Government and
nongovemment organizations supply technical services and provide tools. Community
cammign works construct conservation structures on communal as well as private lands.
Out of the surveyed farmers, 70 percent had conservation work done on their plots through
community campaign labor by 1995. Alongside the campaign work, extension service is
assisting farmers to invest in soil conservation practices on their own. Fifty-seven percent
of surveyed farmers had contact with extension service related with soil conservation
during 1992-95. Moreover, 34 percent of surveyed farmers made private investment in
physical conservation practices and 60 percent reported having used biological soil
conservation practices during 1992-1995.

Of those who invested in physical practices, 95 percent used stone terraces, 19
percent used soil bunds and 9 percent invested in check dams. About 23 percent invested

in more than one practice. Among the biological practices, crop rotation was the most

107

widely used practice (80% of cases), followed by contour plowing (60.3% of cases) and
grass strips (31.5% of cases) (Table 4.13). About 82 percent of households used more

than one biological practice. ’

Table 4.13: Farmer Investment in Conservation Practices

—

 

 

 

 

 

Percent of cases
(N 1=98, N2=265)*
stone terrace 95
Physical practices 5°“ bund 13-3
check dam 8.8
grass strip 315
Biological practices contour plowing 60.3
crop rotation 80,1
tr lantin 9.6

 

* N1 and N2 refer to number of cases in physical and biological practices respectively.

More than 60 percent of farmers reported having maintained physical conservation
practices. The most important reason given by farmers for not maintaining conservation
structures is labor shortage. Most of the maintenance work was done on stone terraces.
Regarding preference between own and campaign work for soil conservation, 57 percent of
farmers preferred campaign to private work. When asked if they would have constructed

the conservation structures built on their plots through campaign work by themselves, 61

108

percent replied negatively. Labor shortage was given as the reason by more than 90
percent of them.

In some cases, farmers are involved in soil and water conservation activities
including aﬂ’orestation, area closure, dam construction, and soil conservation structures
through food-for-work programs. Forty-four percent of farmers surveyed participated in
such programs over the 1992—95 period. In addition to augmenting the conservation
efforts in the region, the food-for-work programs were fairly effective in isolating the
poorer from the better of households thus contributing to food security of the most needy
households (Gebremedhin et al., 1997).

Almost all participants were of the opinion that such projects would increase yields
beneﬁting the community as a whole, apart from the wage payment. The same opinion was
held regarding conservation work done privately or through community work. When
asked by how much conservation would increase crop yields, seventy-ﬁve percent of them
replied that yield would increase by one-ﬁﬁh to one-third of the yield level without
conservation. Surveyed farmers considered that the major causes of soil erosion were
deforestation (51%), over cultivation (22%), poor agricultural practices (17%), and
overgrazing (5%). Only ﬁve percent reported that intensity of rain and steep slopes were
the major causes of soil erosion.

Risk has been identiﬁed as one factor affecting fanners’ conservation decisions.
Farmers’ risk attitudes towards conservation practices were measured through fanners’
perception of the effect of conservation on stability of yield. More than 63 percent of

farmers surveyed thought that conservation practices increased the stability of crop yields

109

(Table 4.14). Moreover, more than 90 percent believed that investment in conservation
practices is proﬁtable. Such subjective evaluation of proﬁtability may determine investment
decisions more strongly than actual proﬁtability during the initial adoption of conservation
practices. Once the household adopted the practice, actual proﬁtability may be more

important.

Table 4.14: Risk Attitude of Farmers Towards Conservation Practices

 

 

 

 

 

 

Increases Decreases Has no effect Can not
stability stability determine
Stone terrace 66.7 17.4 15.1 1.3
Soil bund 63.3 16.3 18.9 1.5
Biolgg’cal practigg 67. 8 12.0 18 . 9 1i3

 

 

Despite fanners’ overall favorable evaluation of conservation practices, many
farmers have encountered operational problems with stone terraces. Fifty seven percent of
farmers surveyed reported that stone terraces harbor pests such as rats while 14 percent
and 8 percent complained of land reduction and difﬁculty in farm operation respectively.
About 60 percent of farmers received extension education related to soil and water

conservation during 1992—95.

110

4.7 Summary

Farmers have usuﬁuct rights on land and they feel more insecure about their
holdings in the long run than in the medium term. Population pressure has rendered
fallowing obsolete. Land lease, mainly on share cropping, is widely used in the area and
female-headed households are more likely to lease land out. Cultivated land during the
1995/96 crop year averaged 1.7 hectares. Farmers in the study area experience labor
shortage for farm operation and exchange labor is the major source of additional labor.
Twenty-nine percent of the surveyed farmers reported participation in off-farm activities,
mainly construction, petty trade and selling ﬁrewood.

Barley, wheat and teﬁ‘ are the most widely grown crops in the study area in that
order. Several crop varieties had been abandoned due to environmental and agricultural
conditions. Farmers could obtain improved seed on credit from the regional bureau of
agriculture, but use of improved seed is very limited and farmers use local seeds. However,
not all farmers are self-sufficient in local seed. Use of commercial fertilizer is very limited
and those who use it do not apply at the recommended rate. Most of the commercial
fertilizer is applied on Leﬂ‘ due to the high cash value of the crop. Farmers reported lack of
purchasing power as the major reason for not using fertilizer. However, farmers could
obtain fertilizer on credit from the regional bureau of agriculture. Sixty-two percent of
surveyed farmers experienced shortage of draft power during the 1995/96 crop season.
The most common solution for inadequate draft power is exchange oxen between

households.

111

The most important development constraints as perceived by farmers are, in order
of importance, drought, pests, and land scarcity. However, farmers rated agriculture as
their top priority for investment. Regarding priorities for public support of agriculture,
surveyed farmers put credit for draft power at the top.

Thirty-two percent of surveyed farmers invested in physical soil conservation
practices and another 60 percent in biological practices during 1992-95. Seventy percent
of farmers had conservation done on their plots through public campaign work. Farmers
showed islight preference for public campaign conservation work than private work.

Farmers perceive that soil conservation enhances crop yield stability.

CHAPTER 5

THEORETICAL MODEL, EMPIRICAL MODELS AND RESEARCH DESIGN

5.1 Investment in Soil Conservation

When soil erosion occurs, productivity of the land may decrease and less output
will be produced from a given level of inputs than would be obtained without soil erosion,
using the same technology and under similar climatic conditions. This results in a
downward shiﬁ of the production function. Soil conservation can be deﬁned as the effort
that goes into preventing the diminution of the productive capacity of land, while holding
inputs, technology, and climate constant. In other words, conservation refers to the effort
directed at maintaining a given production function over time. An input is a conservation
input if it prevents such a downward shift Of a production function. On the other hand, an
input that causes a movement on the same production function is not a conservation input.

Soil conservation need not be the same as soil erosion control although soil erosion
control may be the major aim of soil conservation (Ervin and Ervin, 1982). Measures
aimed at improving soil structure such as incorporation of organic matter into the soil are
soil conservation practices but may not necessarily reduce soil erosion.

Conservation investment can be analyzed using economic principles that deal with
allocation of scarce resources among competing alternatives over time. From an economic
emciency point of view, investment on conservation practices should proceed as long as
the marginal productivity of resources invested is not less than it would be in other

competitive uses of the same resources. The beneﬁts of soil conservation occur over time.

112

113

Under certainty and lack of capital rationing, investment in soil conservation practices is
justiﬁable as long as the discounted net beneﬁts are greater than the investment cost. Under
uncertainty, the expected discounted beneﬁts may need to be higher than the investment
cost by a risk premium.

Conservation practices can be proﬁtable to both individuals and society, but they
can also be proﬁtable to society but not to individuals (Heady, 1952; Cary and Wilkinson,
1977). The reasons why conservation practices can be proﬁtable to society but not to
individuals include tenure security problems, capital scarcity and risk, which make private
discount factors higher than social discount factors. Tenure insecurity shortens the
planning horizon of farmers while capital scarcity raises the opportunity cost of investment
in conservation practices. Net returns to conservation practices depend on the discount
factor used and time horizon considered. Net returns without conservation are likely to
exceed those with conservation for shorter time horizons and higher discount rates.
Moreover, market failure and the existence of extemalities make farmers’ conservation
investments differ from the socially optimal level (McConnel, 1983). Market failure refers
to inefﬁcient outcome of existing markets due to market imperfection or the absence of
markets. Controlled crop prices aimed at lowering food prices to urban consumers lowers
the opportunity cost of lost topsoil thus inducing soil erosion. An extemality exists when
the action of one agent affects the welfare of another agent. An extemality can be positive
or negative. A positive extemality occurs when the action of an agent increases the welfare
of the affected agent while a negative extemality occurs when the welfare of the aﬂ‘ected

agent is decreased. For example, soil erosion on poorly conserved farm land uphill may

114

damage crops downhill due to sediment deposition or higher runoff, causing a negative
extemality. The upland farmer does not consider this cost and the beneﬁt to the downhill
farmer from conservation uphill in his/her decision of conservation investment. This will
understate both the social costs of erosion and the beneﬁts from conservation making the

social and private proﬁtability of conservation investment different.

5.2 Conceptual Model

Soil erosion is an insidious process and its yield reduction impact can be
negligible when top soil is deep. The economic impact of soil erosion needs to be felt
by farmers before any consideration of conservation investment will be made. Hence,
farmer perception of the problem of soil erosion needs to be incorporated into the
analysis of conservation investment decision making.

Since investment in conservation practices is a medium- or long-term investment
with payoffs distributed over several years in the future, a farmer’s conservation
investment decision can be modeled as a present value maximization problem. The
choice of whether or not Ethiopian farmers should invest in soil conservation is modeled
as a problem of maximizing the present value of resulting beneﬁts. The choice model
developed below is adapted from the one developed by Barbier (1990).

The production of multiple crops by a farm household is assumed to be explained
by a single crop production function. Livestock production is assumed to be subsidiary
to crop production. This is because the analysis is concerned with farmer investment in

conservation practices on cultivated lands and the major impact of soil erosion on

115

cultivated lands is crop yield reduction rather than deterioration of grazing land. The
production function is

<10) = f (210), X (0)

fl > 0, f" < O,

19 > 0, 1’22 <0,

fl2 = t" 2 0,

where q(t) is output, x(t) is soil depth, and z,(t) is a vector of non-conservation
(traditional) input package comprised of production inputs, crop varieties, cropping
patterns and techniques. 21 increases crop output at a decreasing rate as does soil depth,
x. Increased soil depth raises the productivity of the traditional inputs and increased use
of traditional inputs enhances the effect of soil depth on crop productivity as indicated
by the positive cross partial derivatives.

Soil depth is assumed to be a function of management practices and agroclimatic
factors. Use of 2,, the traditional input, induces soil erosion at an accelerating rate.
The household has a choice of using conservation inputs, 2,2. Thus, algebraically,

x0) =, h (2.0). 2.0)). and

h‘s 0, h11 s 0,

h2 > 0, h22 s 0,

h‘2 = h” <0.

Let p(t) be the price of the crop output normalized by the price of any crop that might

be associated with 2.2, and c,(t) and 0,0) be the costs of 21 and L2 respectively. The

116

conservation behavior of farm household can be modeled by the impact of soil depth on
proﬁts through the following present value maximization problem over a given time
horizon:

max PV = fe’" [pf(zl , x) - c,z1 - c221,] dt

Z]: L)

1

s.t. x = h (21 , 2,2)

W) = x0.

Where PV is present value
r is the farmer’s discount factor
x‘ is rate of change of soil depth

x(0) is initial soil depth.

Since farmers in Tigray have only usufruct rights for land the terminal value of
land is not included in the PV model. Assuming that the marginal value of soil lost to

be a continuous function, u(t), the ﬁrst order conditions from the Hamiltonian (H) are as

follows:
dH/dz, == pf‘- c,- uhl = o, (1)
dH/dz,z=-cz+uh2 =0 (2)
u1 =ru-dH/dx =ru-pf2 (3)
X' = h(zr. 2;) (4)

x(0) = x0 (5)

117

Condition (1) implies that, for the traditional productive inputs, the marginal
value product, pf‘, must equal the sum of input costs, c1, and the opportunity cost of
soil erosion, uh'_ Condition (2) implies that the total cost of soil conservation must
equal the marginal value it generates by controlling soil erosion. The costate variable,
u(t), is the shadow price of soil. Hence, the implicit cost of soil loss grows by the
discount rate less the soil’s contribution to current proﬁts, p19 (condition 3).

The comparative static analysis of the temporal equilibrium of the optimization
problem provides the following results:

zl = z,(u, p, c1, c2, x)

1’2 = 7:201: P, 01.qu X),

where u and 01 decrease the probability of use of 21 but increase that of 2,2,
and

p, c; and it increase the probability of use of 21 but decrease that of 2.2.

As the implicit cost of soil loss increases, optimizing farmers would choose to
use conservation input. Also, when the cost of the traditional input increases,
conservation inputs would be preferred. On the other hand, an increase in the relative
price of the traditional crop and in the cost of conservation input, and increased soil
depth would make zl preferred over z), Hence, the farm household’s choice of z, over
21 is inﬂuenced by the factors that increase the proﬁtability of the conservation input

relative to that of the traditional input.

118

5.3 Fanirical Models and Hypotheses
5.3.1 Effect of Stone Terraces on Crop Yields

Conservation practices are expected to raise crop yields. The proﬁtability of
conservation investments is another important consideration. To determine the effect of
stone terraces (the most widely applied conservation practice in Tigray) on crop yields
and farm proﬁtability, an on-farm quasi-experimental study was conducted. Analysis of
variance (ANOVA) is applied on crop yields from conserved and non-conserved plots.
Investment proﬁtability analysis using Net Present Value (NPV) criteria is conducted to

determine the proﬁtability of investment on stone terraces.

5.3.2 Adoption of Conservation Practices

Farmers need to perceive the problem of soil erosion before they will invest in
preventing it. Hence, it is important to study the determinants and level of farmer
perception of soil erosion in an adoption study of soil conservation techniques. Models
of determinants of farmer perception of soil erosion and adoption of conservation

practices are presented below.

Erosion Perception Models

In the present study, we examine perceptions of farmers regarding (1) the
severity and productivity effect of soil erosion, (2) effectiveness of conservation in
reducing soil erosion and (3) farmers’ preference between own and campaign

conservation work. Farmer perceptions were hypothesized to be the product of the

119

topographic and ecological characteristics of the village in which a farm household lives,
physieal characteristics of the land holding of a household, socio-institutional factors,
and demographic characteristics of the household. Village level factors include
agroecological zone, predominant topography of village and indicators of land
degradation. Villages situated in the highland zone were hypothesized to be associated
with higher perception of soil erosion problem because of higher rainfall. Hilly villages
and villages more degraded land were expected to raise erosion perception.

Physical characteristics of the land include plot size, degree and shape of slope,
distance from homestead, soil type, location of plot and length of time operated by
household. Slope degree, and duration of time of operation were expected to raise
perception while distance from homestead was expected to lower it as more distant
plots are likely to be less frequently observed. More erodible soils, location at upper
slope, and convex-like shape of slope were expected to raise erosion perception due to
their tendency to increase erosion.

Institutional variables include extension service, availability of food-for-work
(FFW) projects for resource conservation, community campaign work on individual
plots, and land tenure. All institutional variables except land tenure were expected to
raise erosion perception either through information access or technical service. The
effect of land tenure (owned versus leased in) is ambiguous. On one hand, research
results indicate that farmers perceive more erosion on other people’s land than on their
own. On the other hand, owned land is typically operated longer than leased in land, as

lease terms in Tigray last only for a few years.

120

Demographic characteristics of the household include age, sex and literacy of
household head. Age was expected to raise erosion perception through experience,
literacy to raise it through access to information, and male sex to raise it through higher
involvement in the agricultural production process. The effect of the explanatory
variables on farmer perceptions of the productivity effects of soil erosion and
effectiveness of conservation in reducing soil erosion were expected to be the same as in
the perception of severity of erosion model.

Farmers’ preference between own and campaign work was hypothesized to be a
function of the factors affecting capacity to invest, in addition to the factors affecting
perceptions. Village location in highland zone and severity of land degradation were
expected to favor own work, as were plot characteristics that cause higher erosion,
contact with extension service and duration under owner’s operation. Distance of plots
from homestead should favor campaign work because of the travel time required to.
conserve distant plots. The effect of availability of FFW projects is ambiguous because
FFW projects can be sources of information and/or technical support but can also

compete for time with own conservation investments.

Conservation Adoption Models

The adoption of soil conservation practices was hypothesized to be a function of
the ﬁnancial and physical incentives to invest, the capacity to invest, risk of investment,
and socio-institutional factors. Village location in a highland zone and indicators of the

severity of degradation were expected to encourage adoption due to higher need for

121

conservation. Plot characteristics that induce more erosion and thus possibly higher
returns to conservation were expected to encourage adoption. Duration of cultivation by
owner should encourage adoption. Family labor supply and area of cultivated land were
hypothesized to increase adoption since they increase the capacity to invest. Access to
credit and contact with extension service were hypothesized to encourage adoption by
augmenting the capacity to invest through easing liquidity problem and provision of
technical knowledge respectively. The effects of distance of village from market place
and all-weather road were ambiguous. These variables could reduce the return to
agriculture thus reduce the probability of adoption; on the other hand they could lower
the opportunity cost of labor thus increase adoption. Distance of plot from homestead
was expected to reduce the probability of adoption. Leased in land, and short- and
long-term land tenure insecurity were expected to lower the probability of adoption due
to their effect on the riskiness of the investment. The effect of FFW project availability
was considered ambiguous as FFW projects can serve as sources of technical
information but also compete for labor with conservation investments. The effect of
campaign conservation work on own land was also considered ambiguous because
campaign work can be a substitute for own investment but can also be a source of
technical knowledge. Several groups were expected to have a lower propensity to
adopt: older people due to shorter time horizon, female headed households due to less

access to resources, and illiterate people due to less access to information.

122

Intensity of Adoption Models

The decision to take an action and the degree of action taken need not be joint
decisions (Cragg, 1971). Unlike the Tobit model (Tobin, 1958), where the two
decisions are assumed to be made simultaneously and so inﬂuenced by the same set of
variables, technology adoption and intensity of adoption may be viewed as a two-step
process requiring examination of the two decisions separately. This can be done using a
double-hurdle model where the decision to take an action is modeled as a probit model
and the decision on the intensity of the action modeled as a truncated regression, the
distribution truncated at zero for decisions that have only positive action outcomes
(Cragg, 1971).

In this study, adoption and intensity of use of conservation practices are modeled
as separate decisions. A farm-household ﬁrst makes a decision on adopting a
conservation practice and then on how much to invest in it. While the decision on
adoption can be a one-period decision, investment on the intensity of conservation
practices is necessarily a multi-period decision as time will be required to implement
most practices, physical conservation practices in particular. Intensity of conservation
practiced was hypothesized to be a function of the same variables as adoption. Although
it is difﬁcult to predict how the effect of the explanatory variables would differ between
the adoption and intensity models, it was hypothesized that some of the effects would

differ either in direction or magnitude, or both.

123

5.4 Explanatory Variables

Production inputs, crop varieties, cropping patterns and techniques and cost of
conservation investment did not vary in the study area. Information on soil depth and
cost of soil loss were not available. Hence, the variables included in the empirical
models were selected following the literature on farm-level investment theory
(Christensen, 1989; Feder et al., 1992; Clay and Reardon, 1996). Following this
literature, farm investment can be modeled as a function of ﬁnancial and physieal
factors that affect the proﬁtability of investment, risk of investment, capacity to invest,
household demographic characteristics and socio-institutional variables.

Financial factors of investment refer to the price and cost variables that affect,
directly or indirectly, the relative proﬁtability of investment in conservation practices.
Ideally such factors would include crop prices, cost of labor and materials used for
conservation and the yield effect of conservation practices. Data on crop prices could
not be collected, nor were they available from secondary sources. Cost of conservation
did not vary across farms in the study area. Information on the effect of conservation
on yield was not available. Distance from market place and access to all weather road
was expected to affect farmers’ revenue from sale of crop output in addition to prices.
Each of these variables was measured as walking distance from village of residence.

Physical incentives to invest in conservation practices include the village level
ecological factors and physical characteristics of plots. The more environmental
degradation in a village, the more likely resident farmers would be to invest on

conservation practices. One such measure is what the major source of cooking and

124

heating fuel is. As degradation gets severe, farmers depend more on animal dung for
fuel than fuelwood. The distance villagers need to travel to fetch fuelwood is another
indieator of the extent of degradation. Hilly villages induce soil erosion thus should
encourage conservation. Highland zones have higher rainfall than the intermediate
highland zones and should induce greater soil erosion with more incentive for
conservation practices to reduce runoff. Physical characteristics of plots include slope
degree and slope shape, soil type, plot location, plot size, distance from homestead, plot
fragmentation and period in years since operated by current operator.

Three different measures were used to capture the risk of investment, one each
for the “immediate period”, the short-term and the long-term. In the immediate period,
risk was measured in terms of whether or not the land was owned or leased in. For the
short-term, farmer tenure security was measured in terms of whether or not farmers
believed that they would cultivate the same plots ﬁve years from the time of the survey.
Long-term tenure security was gauged by whether farmers believed they would bequeath
plots to their was solicited. At the village level, time elapsed since last land distribution
was used as measure of the stability of land tenure.

The factors that affect the capacity to invest include cash income, wealth, land
area and family labor. The cash income and wealth data were unusable due to under-
reporting. Land area was measured as hectares of cultivated and cultivated land per

capita. Family labor was measured as number of household members aged 15-64.

125

Household demographic variables include age, sex and literacy of household
head. Age was measured in years. Literacy was measured in terms of the head’s
ability to read and write.

Relevant socio-institutional variables include the pressure village communities
would have on households to conserve soil, access to formal or informal agricultural
credit, contact with the agricultural extension service, availability of FFW projects in a
village, and whether a household beneﬁted from soil conservation campaign works.
Pressure from community to conserve soil was measured in terms of whether a
households felt that the community expects him to conserve his land.

Deﬁnitions and measurements of dependent and independent variables are given

in Table 5.1 and Table 5.2 respectively.

126

Table 5.1: Definition and Measurement of Dependent Variables

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Variable code Deﬁnition measurement Level of
Observation:
Village 8 VILL,
Household=HD,
Plot=PT
1.Farmer Perceptions of Soil Erosion and Conservation
ERSEV Farmer perception of the 0=no problem PT
severity of soil erosion 1=slight
before conservation was 2 =moderate
applied 3 = severe
ERSEVI Farmer perception of the 0= no problem or PT
severity of soil erosion slight 1=moderate
before conservation was or severe
applied
EFFECTCN Farmer perception of the l=aggravated PT
effectiveness of 2 = same
conservation in reducing 3 = reduced
erosion
ERSPRD Farmer perception of the O=no reduction PT
effect of erosion on l=one-ﬁﬂh
productivity 2=one-fourth
3 = one-third
4=one~half
ERSPRDl Farmer perception of the 1=0 2=0.20 PT
effect of erosion on 3=0.25 4=0.33
productivity 4 = 0.5
OWNCGPRF Farmer preference 0= campaign
between own and l=own HD
campaign conservation
work
OWNFORCG Willingness to substitute O=no 1=yes HD
own work for campaign
conservation work by
farmers who already had
public campaign
conservation work on
private land

 

 

Table 5.1 (Cont’d)

127

 

 

 

 

 

 

 

 

 

 

 

2. Adoption and intensity of use of conservation practices
ADOPTTER Adoption of stone terrace 0 = no 1 = yes PT
ADOPTBUN Adoption of soil bund O=yes 1=no PT
ADOPTBIO Adoption of biological O=no 1=yes HD
conservation practices
MNTNTER If household maintained 0=no 1=yes HD
stone terraces
TERACEHA Terraces per hectare meters PT
S

 

128
Table 5.2: Definition and Measurement of Explanatory Variables’
_

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Variable Code Deﬁnition Measurement Level of
observation
Village=VILL
Household=}!
D, Plot=PT
1. Financial factors
MKTDIST Distance of village from nearest walking hours VILL
market
ROADDIST Distance of village from all walking hours VILL
weather road
2. Physical factors
FUELDIST Average round trip distance to walking hours VILL
fetch fuel wood
FUELDUNG . Major source of cooking fuel 0=forest and VILL
bush
1 =animal
dung
HIGHLAND Agroecological zone O=intermediat VH1.
e highland
1 =highland
HILLYVIL predominant topography of 0 = plain VILL
village 1 =hilly
PLOTCVT Fragmentation of plots number of HD
plots cultivated
SLOPEDGR Slope of plot degrees PT
SLOPBAVE Average slope of plots degrees HD
SOILSAND‘ Predominant soil type of plot l=sand PT
0=otherwise
SOILSILT Predominant soil type of plot 1=silt PT
0=otherwise
SOILLOAM Predominant soil type of plot l=loam PT
0=otherwise

 

 

 

 

 

 

 

3 Some variables could be categorized differently. This classiﬁcation is based on the adoption and intensity of
use models.

6 Clay soil was the base for comparison among all soil dummies

Table 5.2 (Cont’d)

129

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

SLOPCNVX" Plot slope shape 1=convex PT
0=otherwise
SLOPCNCV Plot slope shape 1 =concave PT
0=otherwise
SLOPMIX Plot slope shape l=convex- PT
concave
0 =otherwise
PLOTUPI Location of plot = upper PT
slope
0 =otherwise
PLOTMID Location of plot 1= middle PT
slope
0 = otherwise
PLOTLOW Location of plot 1= lower PT
s10pe
0 = otherwise
PLOTAREA Size of plot hectare PT
DISTHOME Distance of plot from home walking hours PT
DISTHMAV Average distance of plots from walking hours HD
home
PLOTAGE Duration plot operated by owner years PT
PLOTAGAV Average duration plots operated years HD
by owner
3. Capacity factors
WORKERS Number of working age number of HD
household members people aged
15-64
AREAPCAP Size of cultivated land per capita hectare HD
LANDSIZE Size of cultivated land hectare HD
4. Risk factors
OWNNOW If plot is owned or leased in 0=leased in PT
1 = owned

 

 

7 Rectilinear shape of plot was the base of comparison for all slope dummies. -

' Plain or Plateau was the base of comparison for all plot location dummies.

 

130

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Table 5.2 (Cont’d)
OWNSYRS If owner feels certain to cultivate 0= no/not HI)
the same plots after ﬁve years sure
1 =yes
OWNINHRT If owner feels certain to leave 0: nolnot I-ID
plots to children sure
1=yes
5. Soda-institutional factors
LANDDSTR Time in years since last land years VILL
distribution in village
PRESSURE If household feels pressure from =no HD
cumrnunity to conserve soil 1=yes
CREDITAC If household had access to =no HD
formal or informal ag. credit 1=yes
EXTCONS If household had contact with 0=no HD
extension conservation service 1=yes
FFWAVAIL If food-for-work was available in =no VILL
village 1 = yes
CAMPAIGN If household had conservation =no HD
work done on his plots by public 1=yes
campaign work
HOLIDAYS Number of holidays household days HD
observes per month excluding
week ends
6. Household demographic characteristics
AGEHSHED Age of household head years HD
MALEHEAD Sex of household head 0=female HD
l=male
LITERACY Literacy of household head O=illiterate HD
1 = literate

 

 

 

 

 

W

 

131

5.5 Sampling and Data Collection

The study area, approximately 400 kmz, covered 30 villages in six districts in
the Tigray region. Five districts were in southern Tigray and one district in central
Tigray. For sampling purposes, the area was classiﬁed into four topographic zones:
steep, moderately steep, moderately steep and hilly, and plain. Representative villages
were purposively selected in each topographical class. The number of villages selected
was proportional to the land area covered by each class. A sampling frame of household
heads in each village was then prepared in cooperation with extension agents and village
leaders. A random sample of 250 households was drawn. The number of households
sampled from each village was proportional to the number of households in the village.

Data were collected at village, household and plot levels. Observation,
measurement and interviews with farmers were used to collect the data. First, a rapid
rural appraisal (RRA) study was conducted using semi-structured questionnaires. ’
Discussions were made with experts at the Ministry of Natural Resource Development
and Environmental Protection (MoNRDEP) in Addis Ababa, and the regional branch
ofﬁces of the same ministry, Ministry of Agriculture (MOA), the Institute of
Agricultural Research (IAR), and several NGO’s (local and foreign) in Mekelle, Tigray.
Meetings were also held with subject matter specialists, extension workers, contact
farmers, and elderly farmers in Tigray. Available secondary data were reviewed.

The results of the RA study were used in developing the survey instrument
which was later tested at two representative locations in the study area, one in the

highland zone and another in the intermediate highland zone. Enumerators were

kn

 

 

 

132

selected and trained on interview techniques, slope measurement, determination of slope
shape and location of plot, identiﬁcation of soil texture using the ‘ﬁnger assessment’
method, and plot size measurement using the ‘rectangularization’ approximation’
method. A soil scientist from Mekelle University College participated in the training.

Most of the village level data were collected by observation and interview with
village leaders. Data related to household characteristics and agricultural activities were
collected by interview with household heads. Data on physical characteristics of plots
were collected by observation and measurement. Soil samples were taken from each
plot operated by a sample household and soil texture determined by the ‘ﬁnger
assessment’ method. The ﬁnger assessment method is a method of soil texture
determination by manipulating the soil sample to determine the predominant feel and
creating balls and threads of soil to determine its elasticity. Physical conservation
practices constructed on each plot operated by the sampled households were measured
using measuring tapes. Plot sizes were measured by ‘rectangularization’.

An on-farm quasi-experimental study was conducted to determine the effects of
stone terraces on crop yields. 8 m2 quadrats were marked on conserved and non-
conserved farmers’ ﬁelds sown with wheat and fava beans. Enumerators supervised the
maintenance of the plot markings while crop management was left to farmers. Crops on

the quadrats were harvested separately and grain and straw/hay yields recorded.

 

9Using this method, a plot was ﬁrst approximated to have rectangular shape with the opposite sides not
necessarily equal. Then the average of the two lengths and the two widths were taken as measures of the width

and length of the plot respectively.

CHAPTER 6
EFFECTS OF STONE TERRACES ON CROP YIELDS AND FARM

PROFITABILITY: A CASE STUDY IN DEGUA TEMBEN, CENTRAL TIGRAY

Stone terraces are the most common soil conservation structures used on cultivated
ﬁelds in Tigray. However, no published evidence exists that quantiﬁes the effect of
terraces on crop yields in the region. This study was conducted during the 1995-96
cropping season in order to evaluate the effect of stone terraces on crop yields and farm

proﬁtability. ’°

6.1 An Overview of Research on the Effects of Conservation Practices on Crop Yields
Soil erosion reduces crop yield through nutrient loss, shallower rooting zone,

reduced organic matter and water holding capacity, and changes in soil structure and clay
content (Weesies et al., 1994). These effects can be represented by soil loss or reduction in
the depth of top soil. Soil type and climatic factors, particularly rainfall, condition the
eﬂ‘ect of soil erosion on productivity. Gantzer and McCarty (1987) as cited in Weesies et
al. (1994) found that on Mexico silt loam soils in the United States yield reduction due to
topsoil loss was higher in drought years than in normal years. The major beneﬁts from soil

conservation are increases in crop and fodder yields (Tjemstrom, 1992).

 

10The study was conducted in collaboration with Mr. Yibabe Tilahun, a masters student in management of
neural resources and sustainable agriculture at the Agricultural University of Norway, who was doing ﬁeld
work for his thesis.

133

Very few studies have been conducted on the effects of soil erosion and soil
conservation on productivity in Ethiopia. Kejela (1992) studied the costs of soil erosion in
central Ethiopia and found a strong correlation between nutrient loss and soil loss. From
experimental plots, he estimated that the cost of the two major nutrients , nitrogen and
phosphorus, lost due to soil erosion, could be as high as 2,500 Birr/ha or US$ 1,208/ha at
1985-86 fertilizer pricesll (Kejela, 1992, p.227).

Gebremichael (1992) conducted a survey in central Ethiopia to analyze the eﬁ‘ects
of soil conservation on crop yields in the 1986/87 production year. Based on 603 harvest
samples ﬁ'om 4 m2 quadrats, he found that (1) grain and straw yields from graded Fanya

juu” terraced ﬁelds were 15—47 percent higher, and were more stable than yields ﬁ'om non-
conserved plots; (2) within the Fanyajuu terraced ﬁelds, grain and straw yields ﬁom above
the terraces were higher than yields from below or the middle of the terraces; (3) grain and
straw yields ﬁom ﬁelds under graded F anya juu were higher by 31 percent and 26 percent
respectively than those from level bunds.

Tegene (1992) studied the impacts of erosion on the properties and productivity of
soils in southern Ethiopia based on data obtained from 17 experimental plots. By
classifying the study area into three categories as severely, slightly, and moderately eroded,
he analyzed the effect of erosion on the grain and straw yields of maize and haricot bean,

the two most important cr0ps in the area. He found that maize yields on the slightly eroded

 

it The prices of urea and diammonium phosphate (DAP) considered in the study were 88 Birr and 58 Birr per
100 kg respectively. In 1985/86 US $l= 2.07 Bin.

12 Fanja Jun terraces are ernbankments along the countour, made of soil and/or stones, with a basin at its lower
side, graded sideways towards a waterway.

134

135

ﬁelds were 3.2-3.6 metric tons/ha compared with 06-1 .0 metric tons/ha on the severely
eroded ﬁelds. The average haricot bean yield on the slightly eroded ﬁelds was 0.4 metric
tons/ha while the yield ﬁom the severely eroded ﬁelds was 0.15 metric tons/ha. The
average straw yield was 1.1 and 0.4 metric tons /ha for the slightly and severely eroded
ﬁelds respectively.

The results of these studies cannot be extrapolated to the farming systems in T igray
region because of the differences in agro-climatic, socioeconomic and extent of land
degradation between the farming systems in central and southern Ethiopia and those in the
Tigray region. In fact, there may be a lot of variation in the characteristics of the farming
systems within central or southern Ethiopia alone due to differences in altitude, cultivation

practices and culture.

6.2 The Study Area

The study area was located in the highland zone of central Tigray with an altitude
of over 2500 meters above sea level. The major crops grown include wheat , teﬂ‘, barley
and fava beans. Average annual rainfall for the 1992/93 to 1995/96 cropping years was
848 mm. During the 1995/96 cropping year, when the study was conducted, annual rainfall
was particularly good at 1004 mm. Moreover, the rainfall distribution during the cropping
season was also relatively good, with 56 days of rain out of 90 during the months of June,
July and August. Atmospheric temperature can fall below freezing and frost incidences can

occur during the harvest season of October to January. Depending on the intensity of

136

rainfall, water logging can be a problem in some ﬁelds. Moreover, hail storms are

common occurrence in the area.

6.3 Data and Research Design

The selection of plots for the study was based on (1) similar soil and geological
formation, (2) slope gradient range of 20-3 0%, and (3) availability of terraces at least three
years of age. Based on these criteria, a watershed was selected and in it 70 terraced plots
divided equally between wheat and fava beans, and 70 non-conserved plots were also
divided equally between same two crops.

Two quadrats of 8 m2 (2m * 4m) were marked on each of the terraced plots sown
with each crop; one just above the terrace and another one parallel to this but below the
next upper terrace. Only one quadrat was marked on each of the non-conserved plots, for
a total of 210 quadrats included in the study. The 4m side lay laterally across the slope
(along the contour). The quadrats above the terraces were designated as Soil
Accumulation Zone (Accum Zone) and those under the upper terrace, Soil Loss Zone
(Loss Zone). The quadrats on the non-conserved plots were designated as Control Zone
(Control). Crop management was totally left to farmers except for monitoring to ensure
that the quadrat markings were preserved. F ava bean ﬁelds which were heavily aﬁ‘ected by
hail around mid-August were replaced by other ﬁelds in the same watershed. Crops from
the quadrats were harvested and threshed separately from the rest of the plots. Grain and

straw yields were then weighed and recorded.

137

The on—farm research was quai-experimental in the sense that farmers decided on
which plots to use stone terraces. If farmers used stone terraces only on plots of that had
some common characteristics, systematic bias might have been incorporated into the yield
data. However, the magnitude of the bias is likely to be very small because there is no wide

difference in the severity of erosion between ﬁelds in the study area.

6.4 Analytical Methods
6.4.1 Analysis of variance (ANOVA)

ANOVA is used to test the null hypothesis that several population means are equal
(Watson et. al., 1990). Comparing the within-group variabilities of observations and the
variability between group means, inference is drawn about the differences in the underlying
population means. When only one variable is used to classify the cases, one-way ANOVA
is used. The purpose of the on-farm experiment was to determine if terraces result in
signiﬁcant difference in crop yields. Hence, one-way ANOVA was primarily used for
analysis with the treatment category as the classifying variable.

Terrace construction takes some land out of cultivation. To determine the net
effect of stone terraces on yields and proﬁtability, the production loss due to land occupied
by terraces was considered. The minimum and maximum amount of land that could be lost
to terraces were computed based on the terrace construction guidelines used in extension

work in the region.

138
(1) Determining the minimum:

’ Terrace width averages 0.75 meter. Assuming that a terrace is constructed every
15 meters, there would be 6.67 terraces in a square one-hectare plot laid out across the
slope contour, occupying 6.67 * 0.75m * 100m = 500.25 m2 of land. This is 5 perCent of a
hectare.

(2) Determining the maximum:

If terraces are constructed every ﬁve meters and the terrace width remains the same
as above, 20 terraces will be constructed on a hectare of land, occupying 20 * 0.75m “
100m = 1500 m 2, or 15 percent of a hectare of land.

The yields ﬁom the two terraced treatments were adjusted to a per-hectare basis
under the two assumptions regarding land lost to terraces. Yields from the control plots
were not adjusted. One-way ANOVA was conducted on both sets of yields. This allowed

testing the sensitivity of yield results to assumptions regarding land loss.

6.4.2 Regression analysis

On-farm research ensures that technologies are evaluated under farmers’ own
management practices, and environmental and agricultural conditions (CIMMYT, 1980).
But the heterogenous management of farmer-managed on-farm trials may confound
experimental results. In this study, different frequency of tillage and weeding were
observed on the experimental plots. Hence it was necessary to determine whether the
management diﬂ'erences affected the results of the ANOVA analysis. Linear regression

analysis was used for this purpose. Differences in yield, ﬁequency of tillage, and frequency

139

of weeding were computed between treatment groups. The differences in yield were
regressed on a constant, differences in tillage and differences in weeding. The constant

term captures the treatment effect of stone terraces.

6.4.3 T-test for difference in means

Terrace ages varied among the fava bean ﬁelds. Seventeen plots had terraces 3
years old and eighteen had terraces 4 years old. This raised the question of whether or not
crop yield diﬁ‘erences could be observed due to differences in terrace age. T-test was
conducted to test whether or not mean fava bean yields were equal in Accum Zone and

Loss zone.

6.4.4 Proﬁtability analysis

Partial budgeting was used to evaluate the proﬁtability of investment in stone
terraces. Since the beneﬁts from soil conservation accrue over time, the net returns were
discounted and net present values (NPV) computed over time horizon of thirty years
(Gittinger, 1982). A wheat-wheat-fava bean rotation is considered to represent the

dominant cropping system in the area.

140

6.5 Results and Discussion
6.5.1 Effects on Cr0p Yields
Meat

The unadjusted mean grain yield of wheat from the Accum Zone was signiﬁcantly
higher than those of the Loss Zone and the Control respectively at 5 percent level as were
the adjusted yields at 5 and 15 percent land reduction levels (Table 6.1). After adjustment
for 5 percent land loss, the mean wheat yield from Accum Zone was more than twice the
mean yield from Control and about twice that from Loss Zone. Moreover, the yields from
Accum Zone were much more stable than the yields ﬁom Control with coefﬁcients of
variation respectively of 38 and 61 percent. The coefﬁcient of variation of the yields ﬁom
Loss Zone was 40 percent.

The mean straw yield from Accum Zone was also signiﬁcantly higher than the straw
yields ﬁ'om Loss Zone and Control respectively, with and without adjustment at the two
levels of land loss. At the 5 percent adjustment, straw yields from Accum Zone were twice
the yields from Loss Zone and Control. The straw yields from Accum Zone were more
stable than the yields from Control with coefﬁcients of variation, respectively, of 35 and 50
percent. The coefﬁcient of variation of straw yields from Loss Zone was 37 percent. On
the other hand, neither the mean wheat grain nor straw yield from Loss Zone was higher
than those ﬁ'om Control with and without adjustment. Mean wheat grain and straw yields

are given in Table 6.1.

Table 6.1:

141

100 kg/ba (standard deviation in parentheses)*

Mean Wheat Grain and Straw Yield Under the Dil'feren Treatments,

 

 

 

 

 

 

 

 

 

 

 

Output Treatment Unadjusted percent percent
Grain Accum Zone 16.10 (6.09)a 15.29 (5.79)a 13.68 (5.18)a
Loss Zone 8.49 (3.35)b 8.07 (3.18)b 7.21 (2.85)b
Control 6.64 (4.08)b 6.64 (4.08)b 6.64 (4.08)b
Straw Accum Zone 27.87 (9.84)a 26.48 (9.35)a 23.68 (8.36)a
Loss Zone 14.52 (5.42)b 13.79 (5.15)b 12.33 (4.6l)b
ntrol 12.03 6.05 b 12.03 6.05 b 12 03 6 05

 

* Figures followed by different letters were signiﬁcantly different at 5 percent level, using
Bonferroni multiple range test (Watson et al., 1990).

In order to control for the effect of the differences in management practices of
farmers on comparison of mean yields, yield differences at 5 percent adjustment were
regressed on differences in frequencies of weeding and tillage. Tillage and weeding
frequencies were the principal management differences exhibited by the farmers. Wheat
grain and straw yield differences between each treatment pair were regressed on a constant,
differences in weeding frequency (rounds of weeding), and differences in Tillage ﬁ'equency
(rounds of tillage). The sign and signiﬁcance of the coefﬁcients of the constant term in
each regression were evaluated to see if the differences in means were affected by the
management diﬁ‘erences.

Except for the wheat grain yield differences between Accum Zone and Loss Zone,

all other regressions were insigniﬁcant at the 5 % threshold level (Table 6.2). These results

show that the grain and straw yield differences between treatment pairs are generally not
signiﬁcantly related to differences in management practices. Since the constant term shows
the mean yield diﬁ‘erences adjusted for farmer tillage and weeding ﬁ'equency, results are
reported even though most of the regressions were insigniﬁcant. In the one instance where

wheat grain yield diﬁ‘erence was affected by farmer management (Accum Zone - Loss

142

Zone), the constant term is large (8 qt/ha) and over nine times its standard error.

Table 6.2:
(stdandard errors)

Regression Coefﬁcients for Wheat Yield Differences at 5% Adjustment

 

 

 

 

 

 

 

 

 

 

 

 

 

Wheat grain yield Wheat straw yield
differences (100 kg/ha)* differences (100 kg/ha)
Tl-T2 T1-T3 T2-T3 Tl-T2 Tl-T3 T2-T3
Constant 8.06 10.57 1.67 13.27 15.25 0.65
(0.85) (1.90) (1.72) (1.54) (3.06) (2.43)
Weeding frequency 3.29 1.28 -0.03 2.69 2.21 -l.96
diﬁ‘erences (1.41) (1.52) (1.18) (2.55) (2.45) (1.67)
Tillage frequency 3.10 -l .67 -0.24 -l .11 -0.25 1.09
differences (2.86) (1.43) (1.46) (5.17) (2.29) (2.06)
R2 0.15 0.05 0.01 0.05 0.03 0.04
Pr > F 0.08 0.44 0.92 0.47 0.66 0.51

* T1, T2 and T3 represent Accum Zone, Loss Zone and control respectively.

 

143
1‘3an

In the fava bean plots, the AN OVA’s found grain yields from Accum Zone were
signiﬁcantly higher than the corresponding yields from Loss Zone and Control without
adjustment and under 5 percent adjustment. When yields were adjusted for 15 percent land
loss, mean yields from Accum Zone were signiﬁcantly higher than mean yields ﬁom Loss
Zone but not ﬁom Control. At the 5 percent adjustment, mean yield from Accum Zone
. were more than 40 percent higher than the mean yields from Loss Zone and Control. The
grain yields ﬁ'om Accum Zone were also more stable than yields from Control with
coefﬁcients of variation, respectively, of 40 and 78 percent. The coefﬁcient of variation of
yields ﬁom Loss Zone was 43 percent.

Straw yields from Accum Zone were higher than yields from Loss Zone and
Control with and without adjustment. At the 5 percent adjustment, straw yields ﬁom
Accum Zone were 43 percent higher than the yields from Control. Straw yields ﬁ'om
Accum Zone were more stable than those from Control with coefﬁcients of variation
respectively of 34 and 80 percent. Mean grain and straw yields from Loss Zone were not
higher than mean yields from Control with or without adjustment. Mean grain and straw

fava bean yields are given in Table 6.3.

144

Table 6.3: Mean Grain and Straw Yields of Fava Bean Under Different
Treatments, 100 kg/ha (standard deviation)*

_

 

 

 

 

 

 

 

Out put Treatment Unadjusted Adjusted at 5 Adjusted at 15
percent percent

Grain Accum Zone 8.04 (3.13)a 7.63 (2.97)a 6.83 (2.66)a
Loss Zone 5.46 (2.37)b 5.19 (2.25)b 4.64 (2.01)b
Control 5.35 (4.19)b 5.35 (4.19)b 5.35 (4.19)ab

Straw Accum Zone 11.82 (4.07)a 11.23 (3.87)a 10.05 (3.46)a
Loss Zone 7.46 (3.22)b 7.09 (3.05)b 6.34 (2.73)b

n r 6.40 5.09 b 6.40 5 09 6 40 5 09

 

 

 

 

 

* Figures followed by different letters were signiﬁcantly different at 5 percent level using
Bonferroni multiple range test (Watson, et al., 1990).

As in wheat, regressions were run on the differences in fava bean grain and hay
yields between treatment pairs to determine whether differences in yields were caused by
diﬁ‘erences in crop management. The yield treatment differences with 5 percent land loss
adjustment were regressed on differences in frequency of weeding, and differences in
ﬁequency of tillage. All treatment pairs were compared except for Accum Zone and Loss
Zone where there were no management differences in fava beans.

The fava bean regressions show much less beneﬁt from stone terraces than wheat
did. Only for fava hay did the constant (yield difference) term show Accum Zone
outyielding Control (Table 6.4). In the other regressions, the constant term was
insigniﬁcant and weeding effort differences inﬂuenced yield more. The negative eﬁ‘ect of

diﬁ‘erences in weeding frequency indicate that weeding has more yield effect on

145

unconserved plots than on conserved plots. This is likely because the reduced availability

of top soil as a nutrient source and water-holding resource makes weed competition more

damaging to crop yield.

Table 6.4: Regression Coefficients for Fava Bean Yield Differences at 5%
Adjustment (standard error)

 

 

 

 

 

 

Fava bean gram yield Fava bean hay yield
difference (100 kg/ha)* difference (100 kg/ha)
T1-T3 T2-T3 T1-T3 T2-T3
constant 1.05 (0.89) -1.2 (1.01) 3.57 (1.39) -1.17 (1.23)
Weeding frequency
diﬂ'erence -2.71 (1.01) -244 (1.05) -229 (1.44) -347 (1.27)
Tillage ﬁ'equency
diﬁ'erence -l.69 (1.30) -2.01 (1.36) 0.10 (1.87) -0.17 (1.23)
R2 0.25 0.22 0.07 0.20
r > F i 0.01 0.01 0.29 0.03

 

 

 

 

 

 

* T1, T2 and T3 represent Accum Zone, Loss Zone and Control respectively.

The t-tests for the difference in mean grain and hay yields by terrace age showed
that fava bean grain and straw yields under Accum Zone were higher under 4-year old
terraces (mean, 0.73; std.dev., 0.21) than under the 3-year-old ones (mean, 0.55; std. dev.,
0.26). However, straw and grain yields under Loss Zone showed no signiﬁcant difference

by terrace age. Apparently, the soil accumulation in the fourth year after terrace

146

construction still enhanced yields, though the soil loss zone did not show a comparable

yield loss reduction in fava grain and hay yield. This suggests that the terraces had not ﬁilly

developed.

6.5.2 Proﬁtability

Expected proﬁtability can be a major consideration for farmers in deciding whether
to adopt new agricultural practices. Proﬁtability and income stability can be especially
important to smallholders who live on the edge of subsistence and Operate in an extremely
risky agricultural environment. This section presents the results of proﬁtability analysis of
investments in stone terraces. It ﬁrst develops cash ﬂows based on partial budgeting, and
then extends these into capital investment budgets.

The average length of terraces in the study area is estimated to be 700 m/ha. The
regional campaign work for soil and water conservation assigns a target of preparing 7-8 m
of stone terraces per adult man per day. At this rate 100 man days are required to
construct stone terraces on one hectare of land. The terraces are expected to require
annual maintenance equivalent of 5 man days per hectare. Construction and maintenance
of stone terraces is done during dry season when there is little cr0p work.

An important ancillary cost of soil conservation is that more labor is required for
harvest and threshing due to increased yield. The ANOVA results showed that wheat yields
are doubled and fava bean yields increased by about 50 percent due to stone terraces. It is
assumed that where conservation is practiced, higher crop yields will require additional 15

mandays per hectare in wheat ﬁelds and 5 mandays per hectare in fava bean ﬁelds as

147

compared to ﬁelds without conservation. The daily wage rate in the area during the dry
season ranged from 5 to 7 Birr, while during the agricultural season it ranged from 6 to 10
Birr. Hence the average wages of 6 and 8 Birr were used respectively for the two seasons.

Shovels, axes and spades are required to construct the terraces. It is assumed that
one implement of each type is required, and each depreciates over a period of 20 man days.
The market prices of the implements are 20 Birr, 15 Birr, and 15 Birr per unit respectively.

Yield differences between the soil accumulation and soil loss zones after adjustment
for 5 percent land reduction were used for the proﬁtability analysis. The 5 percent land
reduction corresponds with the use of 700 m of terraces per hectare. The grain yield
differences were 8.65 and 2.55 quintals/hectare respectively for wheat and fava bean. The
corresponding straw yield differences were 14.45 and 5.15 quintals/hectare. Yield
differences are conservatively assumed to be zero during the ﬁrst three years of
establishment because of the extent of degradation that has already occurred and the time
needed for the terraces to stabilize. The yield differences observed on the fourth year of
establishment are assumed to remain constant for the entire planning period.

The farm gate prices of 130 Birr/quintal for wheat and 100 Birr/quintal for fava
beans were used, calculated at 1995-96 market prices during harvest season less
transportation cost from farm to market. The prices of straw and hay were 25 Birr/quintal
for both crops. A thirty-year farm planning horizon was assumed.

Two discount factors were used. First, a discount factor of 15 percent, the
agricultural interest rate in the region was used. The 15 percent discount rate was used to

evaluate the proﬁtability of investment in stone terraces if farmers would have access to

148

targeted credit for terrace construction at the current agricultural interest rate. Secondly, a
discount factor of 50 percent, based on an estimated time preference rate (deﬁned as the
consumption rate of interest or intertemporal marginal rate of substitution) of Ethiopian
farmers (Holden, S.T. et al., 1997) was used. Farmers would not be expected to invest
unless the return to their investment is at least as high as their time preference rate.

Using the data and assumptions explained above, the incremental costs and beneﬁts
per hectare due to use of stone terraces were computed. The labor and implement cost of
initial terrace establishment was 825 Birr and the subsequent maintenance costs amounted
to 80 Birr annually for both crops. For wheat, the increased annual harvest and threshing
labor cost amounts to 120 Birr and the market value of increased grain and straw output is
1,486 Birr. For fava bean, the increased annual labor cost is 40 Birr/hectare and the market
value of increased grain and hay output is 384 Bin/hectare. The costs and beneﬁts for

wheat and fava beans are summarized in table 6.5

149

Table 6.5: Summary of Incremental Cost and Beneﬁts of Stone Terraces on

Wheat and Fava Bean Fields (in Birr ‘3

 

 

 

Wheat
Increased Cost
WWW
Labor (100 mandays @ 6 Birr/manday) 600.00
Materials (5 shovels @ 20Birr + 5 axes @
15 Birr + 5 spades @ 15 Birr) 250.00
Total establishment Cost 850.00
. Annualmst
Terrace maintenance
Labor (5 mandays @ 6 Birr/manday) 30.00
Materials (1 shovel @ 20 Birr + 1 axe @ 15 Birr +
l spade @ 15 Birr) 50.00
Increased labor cost for agricultural operation (after year 3)
wheat (15 mandays @ 8 Birr/man day) 120.00
fava beans (5 mandays @ 8 Birr/manday)
Total annual cost 200.00
Increased Revenue (after year 3)
Grain: wheat (8.65 quintals @ 130 Birr/quintal) 1125.50
fava beans (2.55 quintals @ 100 Birr/quintaI)
Straw/hay: wheat (14.45 quintals @ 25 Birr/quintal) 361.25
fava beans ( 5.15 quintals @ 25 Birr/quintal)
Total annual revenue gain 1485.75

 

 

 

Fava bean

600.00
250.00

850.00

30.00

50.00

40.00

120.00

255.00

128.75

383.75

 

 

 

'3 US $1.00 = 6.30 Birr in 1995/96

150

Under the 15 percent discount rate the NPV of investment on terraces for the
wheat-wheat-fava bean rotation over 30 years was 3, 907 Birr (Table 6.6). The discounted
value of the incremental beneﬁts exceeded the discounted sum of the incremental costs in
the ﬁfth year. The nominal pay back period was four years.

When the estimated time preference rate of 50 percent was used as the discount
factor, the NPV was 12 Birr. Although proﬁtability dropped to almost a break-even,
investment in stone terraces remained proﬁtable at the higher discount factor. The
discounted value of the incremental costs exceeded the discounted value of the incremental
beneﬁts in the fourteenth year. The internal rate of return (IRR) of 50 percent suggests the
severity of the impact of soil erosion on productivity and the high potential from investment
in soil conservation in the region. Moreover, since individuals discount ﬁrture income more
heavily than society does, investment in stone terraces would be more proﬁtable if a social
time preference rate were used as a discount factor.

Rural society in Tigray is composed of net sellers and net buyers of cereals. The
analysis above uses crop sale price, and is oriented toward net sellers. An alternative price
to use could be the mean purchase price or time-of-consumption weighted average price.
Since most crop are made right after harvest while rural households make their purchases
later in the season, purchase prices are generally higher than sale prices. Hence, if purchase
prices were used instead, the proﬁtability of investment in stone terraces would be higher
since the cost side of the analysis remains the same.

Although investment in stone terraces has been shown to be proﬁtable to individual

farmers, public intervention is still required to encourage wider use of stone terraces in the

151

region. Targeted credit for investment in stone terraces at the current agricultural interest
rate of 15 percent could ease the liquidity problem of subsistence farmers and make
investment in stone terraces more attractive. Alternatively, a public subsidy to cover part
of the investment cost could be instituted. This would be justiﬁed by the 50 percent of [RR
which makes sense for society despite being a break-even proposition at farmer rates of
time preference.

However, neither of these policy approaches can be a substitute for public technical
and educational intervention through extension service. More extension service is required
to strengthen the knowledge base and technical capability of farmers to construct stone

terraces.

Table 6.6:
Stone Terraces (in Birr)

152

Cash Flow for Wheat-Wheat-Fava Bean Cropping System Under

Y In .

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

ear cremental Incremental Cumulative Discounted Cumulative
cost beneﬁt incremental incremental net discounted
net beneﬁt beneﬁt incremental net
beneﬁt
(15%) (50%) (15%) (50%)

1 850 --- (850) (850) (850) (850) (850)
2 80 -- (930) (70) (53) (920) (903)
3 80 --- (1010) (61) (36) (981) (939)
4 200 1486 276 846 381 (135) (558)
5 200 1486 1562 736 255 601 (303)
6 120 3 84 1826 131 3 5 732 (268)
7 200 1486 3112 556 113 1288 (155)
8 200 1486 4398 484 76 1772 (79)
9 120 384 4662 86 10 1858 (69)
10 200 1486 5948 365 33 2223 (36)
11 200 1486 7234 318 22 2541 (14)
12 120 384 7498 57 3 2598 (11)
13 200 1486 8784 241 10 2839 (1)
14 200 1486 10070 210 6 . 3049 5
15 120 384 10334 37 l 3086 6
16 200 1486 11618 158 3 - 3244 9
17 200 1486 12906 138 3 3382 12

 

 

153

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Table 6.3 (Cont’d)
18 120 384 13170 25 0 3407 12
19 200 1486 14456 104 0 3511 12
20 200 1486 15742 90 0 3601 12
21 120 384 16006 16 0 3617 12
22 200 1486 17292 68 0 3685 12
23 200 1486 18578 59 0 3744 12
24 120 384 18842 11 0 3755 12
25 200 1486 20128 45 0 3800 12
26 200 1486 21414 39 O 3839 12
27 120 '384 21678 7 0 3846 12
28 , 200 1486 22964 30 0 3876 12
29 200 1486 24250 26 0 3902 12
30 120 384 24514 5 0 3907 12

6.6 Summary

Stone terraces improve crop yields under farmer management. Wheat grain and
straw yields from soil accumulation zone were more than twice the yields ﬁom the control
zone (unterraced plots). Fava bean grain and hay yields increased by more than 40 percent
due to stone terraces. Yield stability is also enhanced by stone terraces. Yield variance was
lower for terraced plots in each crop for both grain and straw. The investment analysis
results show that, under current farming practices, stone terraces can be proﬁtable to
farmers both at the current 15 percent agricultural interest rate and farmers’ time

preference rate of 50 percent. However, targeted credit at the current interest rate or a

 

-l-
.-[.

154

public subsidy to cover part of the investment cost are justiﬁable to encourage adoption
and capture the 50 percent IRR. Technical and institutional support need to be
strengthened in order to enhance wider use of stone terraces in the area. Future research
on the productivity eﬂ‘ects of stone terraces should incorporate other important crops
grown in the region, such as barley, sorghum and teff. Moreover, a similar proﬁtability
analysis of stone terraces with improved farming practices, such as the application of
commercial fertilizer or improved seeds, could shade light on the complementarity between

soil conservation and land use intensiﬁcation.

CHAPTER 7
DETERMINANTS OF FARMERS’ PERCEPTIONS OF THE PROBLEM OF SOIL

EROSION AND ATTITUDES TOWARDS SOH. CONSERVATION

This chapter deals with the levels and determinants of farmer perceptions regarding
soil erosion and conservation, and preferences between own and public campaign
conservation work. First, the determinants of farmer awareness of the severity of soil
erosion are analyzed followed by the factors affecting farmer perceptions of the yield loss
due to soil erosion. Determinants of farmer perceptions of the effectiveness of
conservation in reducing soil erosion are presented next. Finally, the factors affecting
farmer preferences between own and public campaign work are discussed. Before
empirical results are presented, the speciﬁcation of the econometric models is discussed,

and the nature of the dependent variables explained.

7.1. Specification of Empirical Models
7.1.1 Binary Dependent Variable Models

Farmer preferences can be analyzed using binary (qualitative) response statistical
models. Using these models, the outcome of interest is modeled to take on the values (0,1),
l for a choice of an action or preference, and 0 for non-preference. Based on the
assumption of utility maximization, a choice t2 is preferred to another choice tl when the
utility derived ﬁom t2 is higher than that derived from t,. The utility function involved in

making the decision between choices can be represented as UM C,), where utility U is a

155

156

ﬁrnction of a vector M, of distribution of net returns for choice t, and a vector C, of
technical characteristics of the choice (Rahm and Huffman, 1984). The variables
constituting the vectors M, and C, are usually unavailable or unobservable, and the utility
ﬁrnction (also unobservable) of the ith farmer is assumed to be a function of a vector of
observable farm/ﬁrm characteristics, X6, and a disturbance term eﬁ. Thus, the utility
function takes the form of :
U,..t(x,) + cs, t=1,2 and 1= 1,..., n.
The ith farmer chooses/prefers t2 over tl when U2,- > U“. Thus the choice can be modeled
as a binary response variable y, where
y= l, ifU2,>U,,, and
=0, if U2i < Um
The focus of the binary response models is on the factors that determine the probability of
choice. The probability that Yi takes on the value 1 can be represented as a ﬁrnction of
farm/ﬁrm characteristics. Thus,
Pi = P(y=l) = P(U2i >U“).
= P(X.a2+ 6:. >Xa1 +61.)
= P1032. - e. > X. (a .- as]
= P0. < Xi» = F 040),
Where P(.) is a probability function,
u, is a random disturbance term,

[3 is vector of coefﬁcients and

157

F(Xiﬂ) is a cumulative distribution function for u, The distribution of 11, determines
the distribution F.

It follows, therefore, that in general, the probability of choice/preference of an
action can be represented as Pi = P,(Y= 1) = G(Xi ,B), for I= 1,2, ..., n) where Xiis a
vector of explanatory variables and B is a vector of unknown parameters (to be estimated).
In order to estimate the parameter values, one needs to specify the functional form of G.
Three alternative functional forms of G have been suggested: the linear probability, probit

and logit models (Greene, 1990).

The linear probability model:

The linear probability model is a linear regression of the binary outcome variable on
a set of explanatory variables:

Y = B’X + e where Y is the binary dependent variable, X is a vector of explanatory
variables and e is the disturbance term. Ordinary Least Square (OLS) is used to estimate
the relationship. The linear probability model suffers from two major drawbacks (Greene,
1990): (1) e is hetroskedastic (and depends on B), and (2) the model can produce predicted
probabilities less than zero or greater than one. The need to produce probabilities
consistent with theory has led to the development of alternative speciﬁcations of which the

probit and logit models are the most popular.

158

The probit and logit models:

The difference between the probit and logit models is in the assumed cumulative
distribution ﬁrnction of the associated disturbance term. While the probit model assumes
the cumulative normal distribution ﬁrnction, the logit model is based on the logistic

cumulative distribution function. In the probit model, the probability of a choice/preference

B'x

P[Y=1]= f ¢(t)dt.

is deﬁned in terms of an index which is converted into probability value through the
assumed cumulative normal distribution ﬁmction. The probit model takes the following
form:
P[Y=l]=<I>(B’x),where
P is probability of choice/preference and

(P(.) is the normal cumulative distribution ﬁrnction

In the logit model, the probability of choice/preference is given as follows:

epl"
P[Y=l]=——l—
1 + epx

= A0354).

159

where P is probability and

A(.) is the logistic cumulative distribution function.
Whatever distribution is used, the probability model is a regression P(Y) = F(B’x) where x
is a vector of explanatory variables. Unlike in the linear probability model, the relationships
are estimated by maximum likelihood estimation technique and the parameter coefﬁcients
are not necessarily the marginal effects. However, the marginals can be computed using the
estimated parameter coefﬁcients.

The choice between the probit and logit models is hard to make on theoretical

grounds. However, unlike in the multiple outcome cases, choice of one over the other does
not make much difference in binary outcome situations (Green, 1990). In this study, the

probit model is used.

7.1.2 Ordinal Dependent Variable Models
Some choice problems involve multiple choices that are inherently ordered.
Ordered probit and logit models have been widely used to analyze such choice outcomes.
As in the binary choice models, these models are based on a latent regression:
y* = ﬂ’x + e
where y“ is unobserved. What is observed is:
y = 0 if y“ s 0,
= 1 if0 s y*<u,,
=2 if»: s y" <02,

= J ifu,,l s y*.

160

The p’s are unknown parameters to be estimated along with B. As in the binary outcome
case, the assumption of normal distribution of the disturbance term yields the ordered
probit model, while the logistically distributed disturbance term yields the ordered logit
model. The choice between the two distributions (hence between the two models) is
practically immaterial (Greene, 1990). The normal distribution results in the following
probabilities:

P(Y=0) = ¢(-B’X),

P(y=1) = 9011-1330 - M-B’X)

P(y=3) = 4’01:- B’X) - (Pf-BX)

P(Y=J) = 1‘ (p(llr-r ' WK)-
Again as in the binary outcome models, the parameter coeﬂicients in these models
are not the marginal eﬂ’ects and the marginals need to be computed using the estimated

coefﬁcients. In this study the ordered probit model is used.

7.2 Empirical Models
7.2.1 Models of Farmer Perception of Soil Erosion and Conservation

Farmers need to perceive the problem of soil erosion before they will consider
investing in its prevention. An understanding of the levels and determinants of farmer
perceptions of soil erosion and conservation can help design appropriate conservation

intervention strategies. Perceptions of farmers regarding (1) the severity of soil erosion,

161

(2) the productivity effect of soil erosion, and (3) the effectiveness of conservation in
reducing soil erosion were examined.

Farmers were asked to rate the severity of soil erosion on each of their plots before
any conservation was done“. Perceptions were solicited in four ordinal categories (no
problem, slight, moderate, severe). These four ordinal categories were then converted into
binary categories by classifying farmers into those who perceive erosion as a problem and
those who do not. Ordered probit was used for the ordinal dependent variable while
ordinary probit was used with the binary dependent variable. Contrasting the results of the
two models helped in testing the robustness of the results to model speciﬁcation.

In order to capture farmer perception of the productivity eﬁ‘ect of soil erosion,
farmers were asked to estimate the rate of yield reduction on each of their plots due to soil
erosion during a normal cropping year before any conservation had been done on each plot.
These responses were used both as ordinal and continuous values. The continuous values
are the perceived rates of yield reduction. Ordinary linear regression (OLS) was used for
the continuous variable and ordered probit was used for the ordinal variable. As in the
erosion perception models, a contrast between the results of the two models showed the
robustness of the results.

Conservation practices are expected to reduce soil erosion. What factors determine
farmer perception of the effectiveness of conservation practices in reducing soil erosion?

To answer this question, farmers were asked if they perceived that conservation practices

 

14 By the time the survey was conducted, some farmers already had conservation practices on their plots. This
may have caused some bias in the reported perceptions.

162

had aggravated erosion, left it unchanged or reduced erosion in comparison to the level of
erosion before conservation each plot. These responses were used as ordinal values (0, 1,
2) and ordered probit was used to analyze the determinants.
Farmer perception of the severity of soil erosion, its productivity impact, and the
effectiveness of soil conservation were all modeled in two steps as follows:
Perception = F ( true information, socio-institutional factors, household
demographic characteristics)
True Information = F (topographic and ecological characteristics of village,
physical characteristics of land holding)
By substituting the factors that explain true information into the perception equation, the
following speciﬁcation was developed:
Farmer Perception = F( l. topographic and ecological characteristics of the village,
2. physical characteristics of the land holding, 3. socio-institutional

factors and 4. Household demographic characteristics).

7.2.2 Farmer Preference Between Own and Public Campaign Conservation Work
Models

Three general approaches are being used in Tigray to conserve land and rehabilitate
the environment: (1) communities are organized in mandatory public conservation
campaigns to work on communal or private lands (2) agricultural extension services assist
farmers to apply conservation techniques by themselves and (3) FFW projects use paid

labor to construct terraces, afforest hillsides and protect closed areas. The ﬁrst two

163

approaches are widely used in the region. F arrners are divided in their attitude towards
own versus campaign conservation work. Hence, it was important to understand what
factors determine these preferences. An understanding of the detemrinants can be used to
better organize the public campaign work and/or strengthen the extension soil conservation
eﬁ‘orts.

All farmers interviewed were asked to reveal their preference between public
campaign and own conservation work. Farmers who already had conservation work done
on their plot(s) through public campaign work were then asked if they would have done the
conservation work by themselves had there been no campaign work available. This was
done to see if having campaign work would change farmers’ preferences between own and
campaign work.

Since the dependent variables were binary, probit models were used to analyze the

factors aﬁ‘ecting the preferences. The following speciﬁcation was used:

Partner preferences between own and campaign work = F ( l. topographic and ecological
characteristics of the village, 2. physical characteristics of the land
holding, 3. socio-institutional factors 4. demographic characteristics
of the household and 5. factors affecting capacity to invest).

The same set of explanatory variables was used as for the perception models,
except for the additional factors affecting capacity to invest, which may affect private

investments in soil conservation.

_4_rL.-_‘

 

164

7.2.3 Econometric Model Tests

A signiﬁcant multicollinearity problem could make it difﬁcult to isolate the effects
of the collinear explanatory variables. Multicollinearity was assessed following the method
suggested by Belsley et al. (1980). For the plot level models, only one of the principal
components (eigenvalues) of the X’X matrix had a condition index of above 30 (40.17) and
the only variables whose coefﬁcient variances accounted for more than 30 percent were the
constant, topography of village (HILLYVIL), and literacy of household head
(LITERACY). Both variables were retained in the models because neither had its
coeﬁcient variance accounted for more than 50 percent. In the models estimated at the
household level, no principal component had a condition index greater than 30, so no
variable was dropped.

The different levels at which the data were collected (village, household and plot)
could result in heteroskedasticity of the error term. Models were estimated using
heteroskedasticity robust (White-corrected) standard errors and they resulted in very
similar results to the models estimated with ordinary standard errors suggesting that
heteroskedasticity was not a problem in the data set.

A speciﬁcation test for functional form was performed on each of the models
estimated, based on the idea that if a model were correctly speciﬁed, there should be no
other signiﬁcant variable other than by chance. Each model was reﬁt with the predicted
and squared predicted value as the only right-hand side variables. A signiﬁcant squared

predicted value would indicate speciﬁcation error. All models passed the test.

165

7.3. Empirical Results and Discussion
7.3.1 Farmer Perceptions of Severity of Erosion

Farmer perceptions of the severity of soil erosion (ERSEV) was measured at plot
level as an ordinal variable (0=no problem, 1=slight, 2=moderate and 3=severe). This was
later converted into binary variable (ERSEVI) taking on values 0 or 1 (0= no or slight
problem, l= moderate or severe problem).

The determinants of farmer perceptions of the severity of soil erosion were analyzed
using ordered probit for the ordinal dependent variable and ordinary probit for the binary
dependent variable. The following regressions were estimated:

(1)ERSEV =F[ 1. (HIGHLAND, HILLYVIL, FUELDUNG, FUELDIST),

2. ( PLOTAGE, SOILSAND, SOILSILT, SOILOAM,
SLOPEDGR, SLOPCNVX, SLOPCNCV, SLOPMIX,
DISTHOME, PLOTAREA, PLOTUP, PLOTMID, PLOTLOW),
3. (OWNNOW, FFWAVAIL, CAMPAIGN, EXTCONS'), 4.
(AGEHSHED, MALEHEAD, LITERACY)]

(2) ERSEV1= F( the same variables as in (1)), where

The explanatory variables are as deﬁned in Table 5.2 (page 128).

The results from the two models are similar (Table 7.1). Farmers perceive soil
erosion as a serious problem. An average farmer had a 58 percent probability of perceiving
soil erosion as a severe problem. Farmer perceptions of severity of soil erosion are based

Primarily on village and plot physical characteristics.

166

Among the village level physical factors, farming in the highland zone
(HIGHLAND) reduced erosion perception. This was contrary to expectation as the
highland zone has higher rainfall and so more runoff. Villages with more degraded lands
(FUELDUNG) raised erosion perception, as expected. Living in a predominantly hilly
village (HILLYVIL) raised erosion perception, since hilly villages have steeper slopes.

Among the physical characteristics of plots, steeper slope (SLOPEDGR), convex-
(SLOPCNVX) and concave-like (SLOPCNCV) shape of plots were associated with higher
erosion perception since these topographic factors aggravate soil erosion. Erosion
perception was not inﬂuenced by plot location factors.

Farmers are more likely to perceive erosion on’plots they cultivated longer
(PLOTAGE). Loam soil of plots reduced perception relative to clay soil contrary to
expectation. Larger sized plots (PLOTAREA) raised erosion perception of farmers,
Perhaps because erosion features on cultivated land are more recognizable on wider plots.
Plots more distant from the homestead (DISTHOME) were associated with lower
Perception since such plots are less frequently observed.

Among the household demographic characteristics, age of household head
(AGEHSHED) reduced erosion perception indicating that older household heads are less
likely to perceive erosion as being serious problem. Older people might be so used to soil
erosion that they consider it as a normal process.

Contact with the extension service for natural resource conservation (EXTCONS)
Was associated with lower erosion perception. This result contradicted expectation and

was puzzling. Perhaps, extension services focused on the construction of conservation

167

practices and advantages from conservation to the extent that farmers felt that soil erosion
was a manageable problem, thus reducing their perception of its severity. However, an
adequate explanation of this counterintuitive result requires a close examination of the
messages transferred to farmers via the extension service. Both models were estimated
without the extension service variable and the same variables were signiﬁcant with the same
signs as in the models that included the extension variable.

Farmer perception of erosion was not inﬂuenced by ownership status of plot.
Neither was availability of FFW projects in a village or campaign conservation work on

private land important in explaining erosion perception.

168

TABLE 7.1: Regression Results for Farmer Perception of Severity of Soil Erosion

 

   

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

ERSEV (std.err) ERSEV] (probit)
(ordered probit) coeﬂicient (robust std. 7 marginal eﬁ‘ects"
7 err.)'

1. Village physical factors
HIGHLAND -.702 (.204)”‘ -.472 (.253)* -.186
HILLYVIL .355 (.157)" .438 (.193)" .173
FUELDUNG .399 (.l45)“”‘"'I .386 (.172)“ .151
FUELDIST -.005 (.023) .009 (.028) .003
2. Plot physical factors
PLOTAGE .024 (.011)“ .019 (.013) .007
SOILSAND .109 (.141) .017(.l75) .007
SOILSILT -.l62 (.348) -.569 (.398) -.224
SOILLOAM -.l91 (.115)* -.289 (.139)“ -.114
SLOPEDGR .049 (.01 l)*" .045 (.014)*** .018
SLOPCNVX .467 (.229)" .592 (.287)"”'I .209
SLOPCNCV .686 (.210)"* .601 (.245)" .213
SLOPMIX .140 (. 188) .239 (.230) .091
DISTHOME -.43l (.l4l)*** -.505 (.156)*** -.l97
PLOTAREA .389 (.160)“ .456 (.188)“ .178
PLOTUP .017 (.188) -.069 (.223) -.027
PLOTMID .005 (.171) -.030 (.210) -.012
PLOTLOW .144 (.128) .230 (.157) .089
3. Socio-institutional factors
OWNNOW .039 (.145) .138 (.170) .054
FFWAVAIL. .042 (.112) .052 (.132) .020
EXTCONS -.452 (.107)*** -.520 (. l 30)""‘”" -. 198
CAMPAIGN .073 (.128) .196 (.151) .077

 

 

169

Table 7.1 (Cont’d)

 

 

 

 

 

 

 

 

 

 

 

4. Demographic characteristics

AGEHSHED -.016(.004)*“ -.018 (.005)“* -.007
MALEHEAD -.161 (.179) -.248 (.210) -.094
LITERACY -.195 (.129) .092 (.157) -.036
constant ---- .213 (.454) ---
Chi-square 128.27 89.78 ‘ «-
Prob >chi-square 0.0000 0.0000 ---
Pseudo R-square 0.084 0.135 ---
Predicted prob. at --- --- .583
x-bar

N 565 565

' robust standard errors are white-corrected standard errors
" all marginal effects are computed at mean values of variables
"‘3 **, * signiﬁcant at 1%, 5% and 10% respectively
7.3.2 Farmer Perceptions of the Yield Effect of Soil Erosion

Farmers were asked to estimate the yield reduction that would occur on each of
their plots due to soil erosion, assuming a normal cropping year and without any
conservation. Responses were solicited in one of ﬁve rates (1= 0 or no reduction, 2= one-
ﬁfth, 3=one-fourth, 4= one-third, 5= one-half). These responses were used both as ordinal
categories (ERSPRD) and as continuous values (ERSPRDI). The continuous values were

the perceived rates of yield reduction. Ordered probit was used for the ordinal variable,

while OLS was used for the continuous variable.

 

170

The following two models were estimated:

(1) ERSPRD = F[ 1. (HIGHLAND, HILLYVIL, FUELDUNG, FUELDIST),
2. (PLOTAGE, SOILSAND, SOILSILT, SOILOAM,
SLOPEDGR, SLOPCNVX, SLOPCNCV, SLOPMDf,
DISTHOME, PLOTAREA, PLOTUP, PLOTMID, PLOTLOW),
3. (OWNNOW, FFWAVAIL, CAMPAIGN, EXTCONS), 4.

(AGEHSHED, MALEHEAD, LITERACY)]

(2)ERSPRD1 = F ( same variables as in (1)), where

The explanatory variables are as deﬁned in Table 5.2 (page 128)

Again the results from the two models were similar (Table 7 .2). In general, the
reSlrlts ﬁom this analysis are consistent with the results obtained from the analysis of the
determinants of farmer perceptions of severity of soil erosion. Factors associated with
higher farmer perception of the severity of soil erosion were also generally associated with
higher perception of yield impact of soil erosion. Farmer perceptions of the yield effect of
s<>il erosion were based on village and plot physical characteristics, and socio-institutional
factors. I

Living in villages with more degraded land (FUELDUNG) raised farmer perception
Of the productivityloss due to soil erosion, as expected. Steeper plots (SLOPEDGR), and
Convex- (SLOPCNVX) and concave-like (SLOPCNCV) shape of plots increased farmer
perception of the yield reduction effect of soil erosion, perhaps through their effect on

increasing erosion. Perception of yield impact of erosion was not inﬂuenced by soil type or

171

location of plot. More distant plots (DISTHOME) reduced farmer perception of yield
impacts as expected since more distant plots are less frequently observed or operated.

Extension service contact (EXTCONS) reduced farmer perceptions of the yield
impact of soil erosion. This result contradicted expectation but was consistent with the
result ﬁom the analysis of the determinants of erosion perception. The models were
estimated without the extension variable and the same set of explanatory variables were
signiﬁcant with the same signs as in the models that included the extension variable.
Farmer beneﬁt from campaign conservation work (CAMPAIGN) also reduced perception
of the yield impact of erosion.

Among the demographic factors, age of household head (AGEHSHED) explained
yield loss perception signiﬁcantly but negatively. This result was consistent with the result

that the same variable reduced farmer perception of soil erosion.

TABLE 7.2: Regression Results for Farmer Perceptions of Yield Reduction Effect of

Soil Erosion

Variable

172

ERSPRD ( std. err)

ERSPRD] (robust stderr)’

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

(Ordered probit) (OLS)
1. Village physical factors
HIGHLAND -.049 (.204) -.003 (.030)
HILLYVIL -.226 (.198) -.034 (.027)
FUELDUNG .318 (.153)" .031 (.022)
FUELDIST -.066 (.206) -.009 (.013)
2. Plot physical factors
PLOTAGE -.006(.011) -.001 (.002)
SOILSAND -. 160 (. 144) -.024 (.020)
SOILSILT .178 (.324) .029 (.027)
SOILLOAM -.l75 (.122) -.022 (.017)
SLOPEDGR .031 (.010)“* .004 (.001)”
SLOPCNVX .362 (.230) .061 (.028)"
SLOPCNCV .416 (.197)“ .058 (.030)
SLOPMIX .174 (.185) .023 (.028)
DISTHOME -.349 (.144)" -.052 (.017)*“
PLOTAREA .136 (.171) .019 (.024)
PLOTUP -.264 (.187) -.027 (.026)
PLOTMID .027 (.176) .008 (.026)
PLOTLOW .186 (.133) .027 (.019)
3. Socio-institutional factors
OWNNOW .149 (.154) .021 (.022)
FFWAVAIL -.015 (.119) -.005 (.016)
EXTCONS -.344 (.l l l)*** —.050 (.016)"*
CAMPAIGN -.286 (.141)" -.036 (.021)”*

 

 

173

 

 

 

 

 

 

 

 

 

Table 7.2 (Cont’d)
4. Demographic characteristics
AGEHSHED -.01 1 (.004)"”""I -.002 (.001)”"
MALEHEAD -.055 (.198) ‘ -.017 (.027) I
LITERACY -.089 (.131) -.010 (.018)
constant ----- .393 (.060)
Chi-square IF 72.44 4.52
prob >chi-square IF 0.0000 0.0000
pseudo R-square / R-square 0.0479 0.1487
H :87 387

 

'robust standard errors are white-corrected standard errors
*,**, "* signiﬁcant at 10%, 5% and 1% respectively.
7.3.3 Farmer Perception of the Effectiveness of Soil Conservation Practices in
Reducing Soil Erosion

Farmers were asked about the effectiveness of conservation practices in reducing
soil erosion on each of their plots which had physical conservation practices. Farmer
perception of the effectiveness of conservation (EFFECTCN) was measured as an ordinal
variable (1=aggravated erosion, 2=same erosion, 3=reduced erosion), relative to the level
of erosion before conservation was used on the plot. Ordered probit was used to analyze
the determinants of perceptions.
The following relationship was estimated:
EFFECTCN = F [ l. (HIGHLAND, HILLYVIL, FUELDUNG, FUELDIST), 2.

(PLOTAGE, SOILSAND, SOILSILT, SOILOAM, SLOPEDGK

SLOPCNVX, SLOPCNCV, SLOPMIX, DISTHOME, PLOTAREA,

174

PLOTUP, PLOTMID, PLOTLOW), 3. (OWNNOW, FFWAVAIL,
CAMPAIGN, EXTCONS) 4. (AGEHSHED, MALEHEAD, LITERACY)]

The explanatory variables are as deﬁned in Table 5.2 (page 128).

Farmer perceptions of the effectiveness of conservation in reducing soil erosion
were explained by village and plot physical factors, and contact with extension service
(Table 7.3). Farming in highland villages (HIGHLAND) reduced the likelihood that
farmers would consider conservation effective. Perhaps, the conservation practices in the
highlands are not effective enough to mitigate the consequences of high runoff.

Farmers considered conservation more effective on plots with predominantly loam
soils (SOILLOAM) than on plots of clay soil. Conservation practices used on concave-
(SLOPCNCV) or convex-like (SLOPCNVX) shaped plots were more likely to be
considered eﬁ‘ective by farmers. Perhaps conservation practices reduce erosion more
effectively on plots more prone to erode. More distant plots (DISTHOME) make farmers
consider conservation less effective; the same variable reduced farmer perception of the
severity of soil erosion. Location of plot did not explain perception.

Contact with extension service (EXTCONS) raised the likelihood that farmers
would consider conservation effective in reducing soil erosion. This result supports the
explanation given above that extension services might be focusing on conservation
practices and the advantages on conservation rather than on educating farmers regarding
the severity and productivity impact of soil erosion.

The other institutional factors (ownership status of plots, availability of FFW

projects and farmer beneﬁt ﬁ'om campaign conservation work) did not explain signiﬁcantly

175

the perceived effectiveness of conservation practices. Household demographic

characteristics also did not inﬂuence farmer perception.

Table 7.3: Regression Results for Farmer Perception of Effectiveness of

Conservation Practices

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Variable EFFECTCN (std. Err)
(ordered probit)
1. Village physical factors
HIGHLAND -.956 (.278)""‘
HILLYVIL .218 (.298)
FUELDUNG -.292 (.240)
FUELDIST .057 (.038)
2. Plot physical factors
PLOTAGE -.008 (.016)
SOILSAND -.082 (.215)
SOILSILT .196 (.462)
SOILOAM .332 (.185)‘
SLOPEDGR -.016 (.014)
SLOPCNVX .823 (.476)*
SLOPCNCV .325 (.293)
SLOPMD( .356 (.257)
DISTHOME' -.431 (.255)‘
PLOTAREA -.421 (.242)*
PLOTUP .006 (.263)
PLOTMID -.018 (.248)
PLOTLOW .022 (.195)

 

 

 

 

176

Table 7.3 (Cont’d)

 

3. Socio-institutional factors

 

 

 

 

OWNNOW .086 (.233)
FFWAVAIL .161 (.191)
EXTCONS .438 (.158)
CAMPAIGN -.026 (.238)

 

4. Demographic characteristics

 

 

 

 

 

AGEHSHED .002 (.006)
MALEHEAD -.301 (.297)
LITERACY . .230 (. 179)
Chi-square 98.46

Prob. > chi-square ' 0.0000
Pseudo r-square ' 0.168

 

 

W

*, *** signiﬁcant at 10% and 1% respectively.

7.3.4. Farmer Preference Between Own and Public Campaign Conservation Work

Farmers in the study area differed in their preferences between mandatory public
campaign and private conservation work. In this section, we investigate factors that
determine farmer preferences between public campaign and own conservation work. First,
the preferences between private and public work are analyzed. Farmer preferences
between own and campaign conservation work (OWNCGPRF) was measured at the
household level as a binary variable (l=own 0=campaign). For those farmers who already
had campaign conservation work on their land, we examine the factors affecting their

propensity to do the work by themselves had there not been any public conservation

177

campaign. This variable too (OWNFORCG) was measured at the household level as a
binary variable (1=yes, 0=no).

The following probit models were estimated:

(1)OWNCGPRF = [ l. (HIGHLAND, HILLYVIL, FUELDUNG, FUELDIST),

2. (PLOTCVT, PLOTAGAV, SLOPEAVE,

DISTHMAV) 3. (FFWAVAIL, CAMPAIGN,

EXTCONS), 4. (AGEHSHED, MALEHEAD, LITERACY)
5. (WORKERS, CREDITAC, AREAPCAP)]

(2) OWNFORCG = F (same as in (1) with CAMPAIGN omitted). CAMPAIGN
equals one for households who beneﬁt from public campaign conservation work. The
explanatory variables are as deﬁned in Table 5.2 (page 128).

The results indicate that, farmers prefer public campaign to private conservation
work. For the average household the probability of preferring public work was 59 percent
(Table 7.4). Among those who already had public conservation work on their land, the
probability of willingness to substitute own work for the public work was even lower (36
percent).

Living in highland areas (l-HGHLAND) and exposure to extension service contacts
(EXTCONS) were the only variables that affect preferences signiﬁcantly. Living in
highland areas favored preference for public conservation work. Living in highland areas
had a very high marginal effect (46 percent) on a household’s preference of public
campaign work over own work. That highland areas favor campaign work may be

indicating that campaigns have been carried out most often in the highland areas. On the

 

178

other hand, contact with extension services favored private work. These results indicate
that the extension service has been effective in teaching farmers of the importance of
conservation work and how to construct/use the practices.

Among those who already had campaign conservation work on their land, extension
service contact (EXTCONS) and average years of plot cultivation (PLOTAGAV) were the
only signiﬁcant variables. Both variables favored private work. The link between longer
period of cultivation and willingness to conserve it by himself suggests that greater land
tenure security may encourage farmers to use conservation practices. The positive
coeﬂicient of the extension contact indicates that extension service is effective in extending

conservation technologies to farmers.

Table 7.4:

179

Campaign Conservation Work

Regression Results for Farmer Preferences Between Own and Public

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Variable OWNCGPRF(probit) OWNFORCG (probit)
coefficient (robust marginal I coefﬁcient (robust marginal
std. errors) effects std. error) eﬁ‘ects

1. Village physical factors

HIGHLAND -l .418 (.475)*“ —.461 -.555 (.493) -.l98

HILLYVIL .353 (.364) .132 .711 (.455) .230

FUELDUNG .044 (.307) .017 .254 (.371) .092

FUELDIST .01 1 (.049) .004 .036 (.061) .014

2. Plot physical factors

PLOTCVT .031 (.078) .012 .042 (.077) .016

PLOTAGAV -.001 (.039) -.000 .071 (.042)’ .027

SLOPEAVE .022 (.030) .009 -.021 (.031) -.010

DISTHMAV .307 (.374) .120 -.126 (.453) -.047

3. Socio-institutional factors

FFWAVAIL .328 (.257) .126 .413 (.288) .151

CAMPAIGN .056 (.270) .022 --- ---

EXTCONS .774 (.236)“* .290 .623 (.241)"* .225

4. Demographic characteristics

AGEHSHED -.005 (.009) -.002 -.010 (.013) -.004

MALEHEAD .328 (.348) .123 .564 (.470) .187

LITERACY .225 (.277) .088 .182 (.282) .069

 

 

 

 

 

180

 

 

 

 

 

 

 

 

 

 

Table 7.4 (Cont’d)
5. Capacity factors
WORKERS ' -.039 (.096) -.015 -.005 (.111) -.002
AREAPCAP’ .267 (.718) .104 .503 (.834) .188
CREDITAC .264 (.250) .104 .313 (.264) .119
Constant -l.553 (.837) --- -2.630 (.941) --
Chi-square 38.84 m 30.45 «-
Prob. > chi-square 0.0019 --- 0.016
pseudo R-square 0.1805 --- 0.1580
predicted prob. at x-bar --- 0.414 --- 0.359
N 171 --- 143

 

 

_
‘robust standard errors are white-corrected standard errors

*,",*" signiﬁcant at 10%, 5%,and 1% respectively.
7.4 Summary

The analysis in this chapter showed that farmer perceptions regarding soil erosion
and conservation are chieﬂy a function of physical factors. In general, farmer perceptions
of the different aspects of soil erosion and conservation are consistent with each other
implying the need for and importance of farmer involvement in the design and
implementation of conservation programs and projects. Extension services appear to focus
on technical service rather than educational intervention to raise farmer awareness of soil
erosion. Farmers showed slightly higher preference for campaign work, suggesting the
need to proceed with both approaches in the region. However, emphasis needs to be given
to strengthening the extension service as farmers will eventually be individually responsible

for implementing conservation practices on their land.

CHAPTER 8
FACTORS AFFECTING FARMER ADOPTION AND INTENSITY OF USE OF

SOIL CONSERVATION PRACTICES

This chapter deals with the determinants of farmer adoption and intensity of use of
conservation practices. Factors aﬁ‘ecting plot level adoption of stone terrace and soil bunds

are discussed separately. The determinants of household level adoption of biological

 

conservation practices as a group (grass strips, contour plowing, crop rotation and tree
planting) are discussed. Factors affecting farmer maintenance of stone terraces are
presented and discussed based on household level data collected from a sub-sample of
households who claimed need for terrace maintenance.

This chapter is presented in three parts. First, the empirical models are speciﬁed.
Next, discussion of the nature of dependent variables and categories of explanatory
variables included in each model, and econometric tests of the models are presented.

Finally, the empirical results are presented and discussed.

8.1 Specification of Empirical Models

The technology adoption behavior of farmers can be studied in terms of the factors
affecting probability of adoption and the factors affecting the intensity of use. Adoption
and intensity of use decisions can be analyzed jointly or separately. When the decisions are
joint, the Tobit model is appropriate for analyzing the factors that affect the joint decision

(Greene, 1990).

181

182

However, adoption and intensity of use decisions are not necessarily made jointly.
The decision to adopt may precede the decision on the intensity of use, and the factors
aﬁ‘ecting each decision may be different. Such decision situations can be analyzed using the
Heckrnan two-step procedure or the two-part double hurdle model. To use the Heckrnan
sample selection model, one needs strong theoretical justiﬁcation for the selection of
variables that affect the probability of adoption but not the intensity of use or make a
functional form assumption for identiﬁcation of the model. The literature on the economics
of soil conservation investment does not identify variables that affect adoption but not
intensity of use. Moreover, the results are sensitive to the assumed functional form
(StataCorp, 1997).

g In this instance, the “double hurdle” model can be used to model adoption as a
probabilistic choice (probit model) and intensity of use as a regression truncated at zero
(Cragg, 1971). In this study, adoption and intensity of use of conservation practices are
considered separate decisions, and the double hurdle model is used to analyze the factors

affecting them.

8.1.1 The Double Hurdle Model

The basic situation handled by the double hurdle model is when there is an event
which may or may not take place, and if it takes place, takes on continuous positive values.
Unlike the Tobit model (Tobin, 1958), the double hurdle model allows the determinants of
the size of an event when it is non-zero to differ from those that determine the probability

of occurrence of the event.

183

In the case of farmer adoption and intensity of use decisions of conservation
practices, a decision on adopting the practices is made ﬁrst, and the decision on the
intensity of use of the practices follows. Following Cragg (1971), the decision on adoption
can be modeled as a probit regression:

f(y= 1| X, , X2) = C(X’,0)
where C(.) is the normal cumulative distribution ﬁrnction, and X, and X2 are vectors of
independent variables, not necessarily distinct; and the decision on the intensity of use can
be modeled as a standard regression:

f(yl X,, X2) = (27:)"” 0’1 exp {-(y - X’,y)2/202} C(X’,[3) for y ,1 0.

To incorporate the non-negativity of y in the model, we need to truncate the distribution at
zero:

f(yl X, , X2) = (2n)"” 0'1 exp {-(y - X’,y)2/202} C(X’,B) / C(X’2 y/ o) for y > 0.

In this study, farmer adoption decisions are analyzed using probit models while the

decisions on the intensity of use are modeled as truncated regressions.

8.2 Empirical Models
8.2.1 Adoption Models for Physical Conservation Practices

Stone terraces and soil bunds are the two most widely used physical conservation
practices in Tigray. Understanding the factors associated with the adoption of each of
these practices can be useﬁrl in guiding efforts aimed at encouraging wider adoption.

Farmers were asked if and what type of physical conservation practice they applied on each

 

184

of their plots during 1992-95. Adoption was measured at plot level as a binary variable and
probit models were estimated to analyze the determinants.
The following speciﬁcation was used:
Adoption = F( 1. Financial incentives to invest, 2. Physical incentives to invest, 3.
Capacity to invest, 4. Risk of investment, 5. Socio-institutional

factors, 6. Household demographic characteristics)

8.2.2 Intensity of Use Models of physical Conservation Practices

Intensity of use of a conservation practice can be a better indicator of the eﬂ‘orts
that go into conserving soil and the effectiveness of the practices than just adoption.
Understanding the factors inﬂuencing intensity can help guide efforts aimed at reducing soil
erosion effectively. The length of stone terraces and soil bunds constructed by farmers on
their plots during 1992-95 were measured in meters. Intensity of use was measured in
meters/hectare and truncated regression was used to analyze the factors inﬂuencing it.

The following speciﬁcation was used:

Intensity = F (same category of variables as in the physical practices adoption

model above)

185

8.2.3 Adoption Models for Biological Conservation Practices and Maintenance of
Stone Terraces

In addition to physical soil conservation practices, farmers in Tigray also use
biological practices including grass strips, contour plowing, crop rotation and tree planting.
Farmers were asked about their use or non-use of any of these practices during 1992-95.
As it was not possible to observe most of the biological practices at plot level during the
survey, data were collected at household level. Adoption was measured as a binary
variable.

Stone terraces can be damaged or destroyed by agricultural operations, grazing
animals or high run-oﬂ‘. Sustained use of the practices requires maintenance. Data on
terrace maintenance were collected from a sub-sample of farmers who had need of terrace
maintenance. Farmers were asked to indicate if they maintained terraces during 1992-95.
As it was not possible to identify maintenance practices at plot level, data were collected at
household level. Maintenance was measured as a binary variable. Probit models were
estimated to analyze the determinants of adoption and maintenance.

The following speciﬁcation was used:

Adoption of Biological Practices = F( same category of variables as in adoption

model above that could be measured at household level)

Maintenance of Stone Terraces = F(same variables as in adoption model above

that could be measured at household level)

 

186

8.2.4 Tests of the Econometric Models

The explanatory variables used in the adoption and intensity of use models included
new variables in addition to those used in the perception and preference models in the
preceding chapter. Hence it was necessary to conduct another test of multicollinearity.
The method developed by Belsley et al. (1980) was applied. In the plot level models, one
principal component (eigenvalue) of the X’X matrix had a condition index above 30. This
component accounted for more than 30 percent of the variance of the ﬁve coefﬁcients. All
variables were retained in the models as no variable had its coefﬁcient variance accounted
for more than 50 percent by the principal component. Multicollinearity was not a problem
in the household level models.

As in the perception models, the different levels at which the data were collected
(village, household and plot) could result in heteroskedasticity of the error term. Models
were estimated using heteroskedasticity robust (White-corrected) standard errors and
resulted in very similar results to the models estimated with ordinary standard errors
suggesting that heteroskedasticity was not a problem in the data set.

A speciﬁcation test for functional form was performed on each of the models
estimated, based on the idea that if a model were correctly speciﬁed, no other explanatory
variable should turn out signiﬁcant except by chance. Each model was reﬁt with the
predicted and squared predicted value as the only right hand side variables. Signiﬁcant
squared predicted value would indicate speciﬁcation error. Speciﬁcation errors were not

detected.

 

 

187

A test was also conducted to determine whether a double hurdle model would be
more suitable than a tobit model. Following Greene (1995), a likelihood ratio test was
conducted for the adoption and intensity of use of stone terraces (no test was conducted
for soil bunds as the truncated regression was not signiﬁcant). First the log-likelihood
values ﬁ'om the three models were computed:

LLprobit = log-likelihood of probit model = -239

LLtruncated= log-likelihood of truncated model = -913

LLtobit = log-likelihood of tobit model = -3 56 values. Then, comparing the test statistic, A
= 2(LLprobit + LLtruncated - LLtobit) = 796, with the chi-square value x235.” = 55.76, the

double hurdle model was found to be more suitable than the tobit model.

8.3 Empirical Results and Discussion
8.3.1 Adoption of Stone Terraces and Soil Bunds

Farmer adoption of physical conservation practices was disaggregated into adoption
of stone terraces and soil bunds. Since the two types of practices have different
characteristics, it was important to analyze the factors associated with the adoption of each
practice separately.

Household adoption of stone terraces (ADOPTTER) and soil bunds (ADOPTBUN)
during 1992-95 were measured at plot level as binary variables (1=adoption, 0=non-
adoption). To analyze the factors affecting farmer adoption of physical conservation

practices, the following probit models were estimated:

188

(1) ADOPTTER= 1=[ 1. (MKTDIST, ROADDDIST) 2. (HIGHLAND,
HILLYVIL, FUELDIST, PLOTCVT, PLOTAGE, SOILSAND,
SOILSILT, SOILOAM, SLOPEDGIL SLOPESQ, SLOPCNVX,
SLOPCNCV, SLOPMIX, DISTHOME, PLOTAREA, PLOTUP,
PLOTMID, PLOTLOW) 3. (WORKERS, LANDSIZE) 4.
(OWNNOW, OWNSYRS, OWNINHRT) 5. (LANDDSTR,
PRESSURE, EXTCONS, FFWAVAIL, CAMPAIGN) 6.

(AGEHSHED, MALEHEAD, LITERACY)].

(2) ADOPTBUN = F(same variables as in (1) above), where
ADPOTTER and ADOPTBUN refer to farmer adoption of stone terraces and soil

bunds respectively and the explanatory variables are as deﬁned in Table 5.2 (page 128).

Adoption of stone terraces

The regression results (Table 8.1) showed that household investment in stone
terraces is inﬂuenced by a wide range of factors. Physical incentives to invest, household
capacity to invest, risk associated with the investment and socio-institutional factors were
important in explaining household adoption of stone terraces. Overall, the likelihood of
adoptionof stone terraces was low. An average farmer had 14 percent probability of

adopting the practices.

 

189

Among the physical factors of investment, farming in the highland zone
(HIGHLAND) reduced probability of adoption, perhaps because public campaign work
was more widely used in the highland zone. Loam soil texture (SOILLOAM) decreased
adoption relative to clay soils, contrary to expectation. Farmers are less likely to use stone
terraces on plots more distant from the homestead (DISTHOME), perhaps due to the
travel time required to conserve those plots. Hilly topography of villages (HILLYVIL),
slope of plot (SLOPEDGR) and concave shape of plot (SLOPCNCV) all raised the
likelihood of terrace adoption due to their effect on increasing soil erosion. However,
farmer adoption of stone terraces decreased as plot slope gets steeper (SLOPESQ). Area
of plot (PLOTAREA) increased farmer likelihood of adoption of stone terraces. Farmers
tend to favor using terraces on plots at the lower part of a slope (PLOTLOW) relative to
ﬂat plots.

Among the factors affecting farmers’ capacity to invest, availability of family labor
(WORKERS) increased the likelihood of adoption. This could be due to the high labor
requirement of terrace construction. Among investment risk factors, the perceived
medium-term (OWNSYRS) and long-term (OWNINHRT) tenure security were signiﬁcant,
the former negatively and the latter positively. As investment in stone terraces is
necessarily a long-term investment, farmers were more likely to invest in it when they felt
secure enough in their land tenure to bequeath land to their children. The investment
analysis in stone terraces (chapter 6) showed that at farmers’ presumed discount rate of 50
percent, the investment breaks even in seventeen years. Ownership status of plot

(OWNNOW) was marginally signiﬁcant with a positive coefﬁcient indicating that farmers

 

190

are more likely to conserve their own land as compared to leased land because the lease
term in Tigray is very short.

Among the socio-institutional variables, the time elapsed since last land distribution
(LANDDSTR), availability of FFW (FFW AVAIL) and farmer beneﬁt from public
campaign conservation work (CAMPAIGN) were signiﬁcant. The longer the period since
last land distribution, the more likely were farmers to adopt stone terraces, again
suggesting the importance of tenure stability for long term investment. Longer duration of
plot operation by a household (PLOTAGE) also increased probability of terrace adoption
providing additional justiﬁcation for the role of tenure stability. Availability of FFW
projects increased likelihood of adoption perhaps due to the information and technical
service provided to farmers. Beneﬁt from public campaign conservation work reduced

own investment, indicating that public works are substituting private investments.

Adoption of soil bunds

Results of the probit regression (Table 8.1) showed that, like stone terraces,
adoption of soil bunds is inﬂuenced by a wide range of factors. Financial and physical
factors of investment, risk of investment, socio-institutional factors and household
demographic factors were important in explaining the adoption of soil bunds. However,
most of the effects were contrary to those on stone terraces. Overall, the probability of
adoption of soil bunds was very low. An average farmer had only 1.3 percent probability

of adopting soil bunds.

 

 

191

Distance of villages from market place (MKTDIST) reduced adoption as expected.
Among the physical factors of investment, loam soil texture (SOILLOAM) increased
adoption relative to clay soils. In addition to the relative severity of soil erosion on ﬁelds
with loam soils relative to clay soils, this soil texture might be more suitable for the

construction of soil bunds. Fragmentation of plots (PLOTCVT) raised the likelihood of

 

adoption of soil bunds perhaps because farmers ﬁnd it easier to use soil bunds when the
transaction cost is higher. The longer a plot was cultivated by a household (PLOTAGE),

the more likely farmers would apply soil bunds, indicating the importance of land tenure

 

stability. Slope (SLOPEDGR) and convex shape of plots (SLOPCNVX) raised likelihood
of adoption of soil bunds due to their effect on increasing soil erosion. Convex-concave
(SLOPMIX) shape of plots reduced use of soil bunds relative to rectilinear shaped slopes.
This might be because irregular shaped plots make it difficult to constnrct soil bunds. All
location of plot variables, upper slope (PLOTUP), middle slope (PLOTlvﬂD) and lower
slope (PLOTLOW), reduced the likelihood of soil bund adoption relative to plain or
plateau. This is due to the fact that soil bunds are more effective on relatively ﬂat areas
than on steeper slopes.

Farmers preferred to use soil bunds on own plots (OWNNOW) as compared to
leased in land. However, long-term tenure security (OWNINHRT) detracted from adoption
of the same practice. This indicates that long-term tenure security favors more durable
stone terraces. Availability of FFW (FFW AVAIL) explained adoption of soil bunds
negatively since FFW projects dealt more with stone terraces. Beneﬁt from public

campaign work (CAMPAIGN) detracted from own investment on soil bunds indicating

192

that public conservation works substitute private investments. Among the demographic

factors, age (AGEHEAD) and literacy (LITERACY) of household head explained adoption

 

of soil bunds negatively. Perhaps farming experience and access to information show soil

bunds less effective in reducing soil erosion.

Contrast between stone terraces and soil bunds

The differences between factors that favored adoption of stone terraces versus soil
bunds offer useﬁrl insights about soil conservation in Tigray. Farmers were more likely to
adopt stone terraces than soil bunds.

Farmers preferred to construct soil bunds on heavier soil textures that are easier to
work with and stay compact. Degree of slope increased the use of both stone terraces and
soil bunds. On the other hand, the inﬂuence of location of plots was much more important
for soil bunds than for stone terraces. Farmers prefer to use soil bunds on ﬂat lands. The
fact that hilly topography of villages was an important determinant of the adoption stone
terraces but did not matter for soil bunds suggests that Tigrayan farmers believe that stone
terraces are more eﬁ‘ective when soil erosion is more severe. Distance of plots from home
detracted from the adoption of stone terraces but did not matter for soil bunds perhaps
because bunds do not require transport of their construction materials.

Household capacity to invest, as measured by family labor supply, favored stone
terraces due to the more labour intensive nature of stone terraces. However, it did not
matter for soil bunds. The effect of long-term tenure security perception of farmers was

strongly positive for stone terraces but strongly negative for soil bunds, indicating the

193

importance of land tenure security for investments that have a long-term payoﬂ‘. In
contrast, insecurity of land tenure detracted from adoption of stone terraces but increased
the adoption of soil bunds with marginal signiﬁcance. At the village level, land tenure
stability favored stone terraces but detracted from soil bunds. Availability of FFW
increased adoption of stone terraces but decreased that of soil bunds. This could be due to
the fact that FFW projects emphasized the rehabilitation of hillsides, and thus focused
more on stone terraces. Household that had beneﬁted from public conservation campaigns
were less prone to adopt either soil bunds or stone terraces, indicating that farmers
consider public investments to be substitute for private conservation work.

Household demographic characteristics were more important in inﬂuencing soil
bunds than stone terraces. The fact that greater age and literacy did not matter for stone
terraces but inﬂuenced adoption of soil bunds negatively might suggest that, through
experience and access to information, farmers ﬁnd soil bunds less effective in reducing soil

erosion.

 

 

Table 8.1:
Bunds

Variable

194

ADOPTTER

Probit Regression Results for Adoption of Stone Terraces and Soil

ADOPTBUN

 

coefficient (robust std.

, err.)

 

marginal effect I

std. err)

coefﬁcient (robust .

marginal

effect

 

1. Financial incentives to invest

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MKTDIST I .047 (.114) .011 I -.338 (.194)‘ I -.011
ROADDIST I .018 (.077) .004 I -.034 (.125) I -.001
2. Physical incentives to invest

HIGHLAND -1.087 (.377)*” -.172 -.3 16 (.469) -.009
FUELDIST .035 (.030) .008 .066 (.056) .002
HlLLYVIL .923 (.260)"* .139 .271 (461) .007
PLOTCVT .059 (.057) .013 .101 (.061)‘ .003
PLOTAGE .024 (.015) .005 .038 (.015)“ .001
SOILSAND -.112 (.181) -.024 .412 (.266) .019
SOILSILT -.008 (.467) -.002 .704 (.622) .050
SOILOAM -.273 (.154)‘ -.058 .578 (.243)" .027
SLOPEDGR .116 (.040)"* .025 .156 (.053)”" .005
SLOPESQ -.004 (.001)"* -.001 -.004 (.002)* -.000
SLOPCNVX .398 (.259) .106 .721 (.355) .050
SLOPCNCV .550 (.256)" .153 .038 (.414) .001
SLOPMIX .305 (.214) .077 -.773 (.437)* -.014
DISTHOME -.884 (.251)"" -.l96 -.262 (.31 l) -.009
PLOTAREA .751 (.233)“" .167 .278 (.335) .009
PLOTUP .1 11 (.106) .025 -.917 (.366)“ -.031
PLOTMID .312 (.222) .078 -.697 (.328)" -.014
PLOTLOW .551 (.181)"* .139 -.404l(.275) -.011
3. Capacity to invest factors 7
WORKERS .111 (.066)* .025 I .020 (.091) I .001
LANDSIZE -.109 (.089) -.024 I -.089 (.102) I -.003

 

 

 

Table 8.1 (Cont’d)

195

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

4.Risk of investment
OWNNOW .375 (.233) .073 .713 (.328)" .016
OWNSYRS -.327 (.185)‘ -.075 .346 (.257) .011
OWNINHRT .332 (.176)* .077 -1.l41 (.268)“* -.036
5. Socio—institutional factors
LANDDSTR .075 (.034)" .017 -.O68 (.047) -.002
PRESSURE .225 (.172) .052 -.382 (.244) -.012
EXTCONS -.120 (.155) -.027 -.095 (.223) -.003
FFWAVAIL .423 (.169)“ .091 -.898 (.292)"""l -.039
CAMPAIGN -.431 (.177)” -.108 -.426 (.263)‘ -.019
6. Demographic characteristics
AGEHSHED -.0001 (.005) -.000 -.020 (.010)" -.001
MALEHEAD -.002 (.246) -.000 .322 (.535) .008
LITERACY .058 (.172) .013 -.509 (.252)“ -.013
constant -3.740 (.690)"* m -l.400 (1.041) --
Chi-square 141.79 --- 101.22 «-
Prob. > chi-square 0.0000 0.0000
Pseudo R-square 0.2839 --- 0.2762 «-
Predicted prob. at x-bar .219 .013
--- 33 ...

 

 

*, ”, *** signiﬁcant at 10%, 5% and 1% respectively.

8.3.2 Intensity of Use of Stone Terraces and Soil Bunds

The density of stone terraces (TERACEHA) and soil bunds (BUNDHA) during
1992-95 was measured in meters per hectare. Density was treated as a measure of intensity
of use. Adoption and intensity of use of the conservation practices were modeled as

separate decisions. A farmer ﬁrst decides on adoption of a given practice and then decides

196

on how much to invest in it. This two—stage conservation investment decision of farmers
was analyzed using a double hurdle model where the decision on adoption was analyzed
using a probit model, and the intensity of use using a truncated regression model.

Results of the adoption decisions (probit models) were presented and discussed in
sections 8.3.1 above and are reproduced here for comparison. This section presents results
of the truncated regression for stone terraces. The truncated regression for soil bunds was
not signiﬁcant and is not reported.

To determine the factors associated with the intensity of use of stone terraces, the
following truncated regression was estimated:

1. TERACEHA = F (same variables as in the adoption of stone terraces model),
where TERACEHA refers to length of terrace in meters per hectare. Explanatory variables
are as deﬁned in Table 5.2 (page 123).

Results of the truncated regression showed that the factors that inﬂuence adoption
and intensity of use of stone terraces are different (Table 8.2). Capacity to invest, risk of
investment and socio-institutional factors which were important in determining adoption,
had no inﬂuence on intensity of use.

The two ﬁnancial incentives of investment, distance of villages ﬁom market place
(MKTDIST) and all weather road (ROADDIST), both favored intensity of terrace
construction despite being insigniﬁcant in their adoption. Perhaps once they adopted,
farmers living in remote villages were more likely to invest in stone terraces due to lower
opportunity cost of labor. These variables explained participation in off-farm employment

negatively (Chapter 4). Silt soil texture (SOILSILT), which did not matter for adoption,

 

197
explained intensity of use positively. Size of plot (PLOTAREA), which explained adoption
positively, explained intensity negatively. On the other hand, duration of operation of a
plot by a farm household (PLOTAGE) explained adoption and intensity of use positively;
but was marginally signiﬁcant in explaining the former. Highland zone (HIGHLAND)
which explained adoption negatively, explained intensity of use signiﬁcantly positively.
This might be because once farmers in the highland zone adopt stone terraces, the yield

advantage they realize induces them to invest in stone terraces ﬁirther.

U:

 

 

 

198

Table 8.2: Probit and Truncated Regression Results for Adoption and Intensity of
Use of Stone Terraces (std. err.)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Variable ADOPTTER (probit) TERACEHA
(truncated)
1. Financial incentives to invest
MKTDIST .047 (.114) 208.80 (126.6)“
ROADDIST .018 (.077) 164.20 (63.88)“
' 2. Physical incentives to invest
HIGHLAND -1.087 (.377)*" 659.47 (296.2)"
FUELDIST .035 (.030) -16.74 (32.29)
HILLYVIL .923 (.260)*" 174.54 (245.6)
PLOTCVT .059 (.057) -56.13 (51.88)
PLOTAGE .024 (.015) 18.42 (10.59)*
SOILSAND -.112(.l8l) 207.04(l61.7)
SOILSILT -.008 (.467) 1383.3 (387.4)*"
SOILLOAM .273 (.154)* 102.33 (214.8)
SLOPEDGR .l 16 (.040)“* 52.25 (42.64)*"
SLOPESQ -.004 (.001)*" -2.12 (1.73)*
SLOPCNVX .398 (.259) 191.88 (227.8)
SLOPCNCV .550 (.256)" 65.41 (208.6)
SLOPMIX .305 (.214) 145.72 (183.5)
DISTHOME -.884 (.251)*“ -300.97 (222.4)
PLOTAREA .751 (.233)"* -749.9 (254.8)""
PLOTUP .111 (.106) 248.92 (232.4)
PLOTMID .312 (.222) 194.65 (239.8)
PLOTLOW .551 (.181)*" 61.71 (184.2)
3. Capacity to invest factors
WORKERS .1 11 (.066)‘ 22.22 (66.18)
LANDSIZE -.109 (.089) 1.19 (87.71)

 

 

 

 

 

 

 

Tal

er _T ﬁt __r # ﬁr ﬁr .1i '7

 

“ﬂ.

199

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Table 8.2 (Cont’d)
4. Risk of investment
OWNNOW .375 (.233) -264.58 (201.3)
OWNSYRS -.327 (.185)‘ 113.87 (205.3)
OWNINHRT .332 (.176)‘I -127.88 (185.9)
5. Socio—institutional factors
LANDDSTR .075 (.034)" -39.94 (33.16)
PRESSURE .225 (.172) -1 16.43 (159.2)
EXTCONS -.120 (.155) ~201.42 (141.1)
FFWAVAIL .423 (.169)" 241.81 (160.8)
CAMPAIGN -.431 (.177) -7089 (182.6)
6. Household demographic characteristics
AGEHSHED -.0001 (.005) -l.27 (6.86)
MALEHEAD -.002 (.246) -l62.64 (226.3)
LITERACY .058 (.172) -l66.24 (161.7)"
constant -3.740 (.690)"‘" m
Chi-square 141.79 ---
prob. > chi-square 0.0000 ---
pseudo R-square 0.2839 ---

 

 

 

 

%

*, ", *** signiﬁcant at 10%, 5% and 1% respectively

8.3.3 Farmer Adoption of Biological Practices and Maintenance of Stone Terraces
Biological conservation practices used in the study area included grass strips,

contour plowing, crop rotation, and tree planting. It was not possible to collect data on

adoption of each of the biological practices at the plot level as most of them were not

observable during the ﬁeld work. Hence the biological practices were considered as a

 

 

gr 0UP

pract.
vane!

Pratti

 

farmer

are as

200

group. Farmer adoption of any individual or combination of biological conservation
practices (ADOPTBIO) during 1992—95 was measured at the household level as a binary
variable (0=non-adoption, l=adoption).

Data on farmer maintenance of stone terraces were collected for those households
who claimeda need for maintenance. Again it was not possible to observe the maintenance
practices on the plots. Hence, farmer maintenance of stone terraces (MNTNTER) during
1992-95 was also measured at household level as a binary variable (0=did not maintain,
lanaintained).

To determine the factors associated with adoption of biological conservation
practices, and maintenance of stone terraces, the following probit models were estimated:

1. ADOPTBIO = F[ l. (MKTDIST, ROADDIST), 2. (HIGHLAND, HILLYVIL,

FUELDIST, PLOTCVT, PLOTAGAV, SLOPEAVE,
DISTHMAV) 4. (WORKERS, LANDSIZE), 5. (OWNSYRS,
OWNINHRT), 6. (PRESSURE, EXTCONS,

FFWAVAIL, CAMPAIGN) 7. (AGEHSHED, SEXHSHED,

LITERACY)]

2. MNTNTER = F( same variables as in adoption of biological practices model
above), where
ADOPTBIO and MNTNTER refer to farmer adoption of biological practices and
farmer maintenance of stone terraces during 1992-95 respectively. All explanatory variables

are as deﬁned in Table 5.2 (page 128).

201

Adoption of biological practices"

Household capacity to invest, risk of investment and socio-institutional factors were
important in explaining adoption of biological practices (Table 8.3). Household
demographic factors did not matter for the adoption of biological practices. Overall,
farmers were highly likely to adopt the practices. An average farmer had 73 percent
probability of adopting biological practices. 1

The severity of land degradation in a village as measured by travel time required to
fetch ﬁlCl wood (FUELDIST), increased likelihood of adoption as expected. Availability of
family labor (WORKERS) detracted from adoption, perhaps because households with more
family labor prefer to use physical practices, particularly stone terraces. Medium-term
tenure security (OWNSYRS) reduced adoption. This could be because farmers who felt
secure of their land holding in the medium-term prefer to invest in physical structures,
particularly soil bunds. Time elapsed since last land distribution (LANDDSTR) explained
adoption of soil bunds negatively but was marginally signiﬁcant. Perhaps, physical
practices are preferred with stable land tenure.

Extension service contact (EXTCONS) increased the likelihood of adoption of
biological practices, as expected. Availability of FFW (FFWAVAIL) explained adoption
positively, perhaps due to provision of tree seedlings and related services. Farmers who
felt community pressure to conserve soil (PRESSURE) were more likely to adopt

biological practices than those who did not.

 

1’ Since the biological practices considered here have different characteristics, the results may not apply to
individual practices. Consideration of the practices separately will be required to verify if the results are , in
fact, generalizable to biological practices as a whole.

 

 

hf

inc
anc
1ve
\vh

101'1

pos
labc
(PL
long

"Eh

exn:

eXpl.

exnn
enCOl
land ‘
main,

ﬁoml

202

Maintenance of stone terraces

Farmer maintenance of stone terraces was inﬂuenced by ﬁnancial and physical
incentives to invest and socio-institutional factors. Capacity to invest, risk of investment
and household demographic characteristics did not inﬂuence maintenance. Overall, farmers
were highly likely to maintain stone terraces. On the an average, a farmer with terraces
who perceived a need to maintain them, had 88 percent probability of maintaining stone
terraCes. '

Distance of village from market place (MKTDIST) inﬂuenced maintenance
positively. This suggests that once terraces are constructed, a low opportunity cost of
labor encouraged their maintenance. The longer a plot was cultivated by current owner
(PLOTAGAV), the more likely stone terraces would be maintained. This might be because
longer period of cultivation increases the likelihood of terrace destruction and thus need for
maintenance.

Among the socio-institutional factors, community pressure (PRESSURE),
extension service contact (EXTCONS), and public campaign work (CAMPAIGN) all
explained maintenance positively. The marginal effects of each of these variables were also
high. These results suggest that households respond to community expectations, and the
extension service was eﬁ‘ective not only in facilitating adoption of stone terraces but also
encouraging their sustained use. Having public campaign conservation work done on own
land encouraged farmers to maintain the terraces, perhaps by increasing the need for
maintenance. Time elapsed since last land distribution in a village signiﬁcantly detracted

ﬁ'om maintenance, contrary to expectation. However, its marginal effect was very low.

 

 

 

Table 8.3:

203

Practices and Maintenance of Stone Terraces

Probit Regression Results for Adoption of Biological Conservation

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Variable ADOPTBIO MNTNTER
Coefficient (robust Marginal Coeﬁicient (robust Marginal
std err.) effects std. err.) eﬁ‘ects
1. Financial incentives to invest
MKTDIST I .181 (.192) .060 .518 (.225)" .101
ROADDIST I -.060 (.140) -.020 .053 (.144) .010
2. Physical factors to invest
HIGHLAND .859 (.666) .246 -.630 (.854) -.l44
FUELDIST .129 (.052)“ .043 -.048 (.064) -.009
HILLYVIL .445 (386) .160 .367 (.461) .083
PLOTCVT .024 (1 17) .008 -.027 (l 17) -.005
PLOTAGAV .024 (.040) .008 .174 (.064)"" .034
SLOPEAVE .029 (.037) .010 .078 (.059) .015
DISTHMAV -.310 (.400) -.103 .606 (.660) .l 18
3. Capacity to invest factors
WORKERS I -.168 (.099)* -.056 .089 (.142) .017
LANDSIZE I .096 (.147) .032 .328 (.224) .064
4. Risk of investment
OWNSYRS I -.673 (.298)” -.214 -.127 (.373) -.025
OWNINHRT I .442 (.306) .141 -.518 (.436) -.109
5. Socio-institutional factors
LANDDSTR -.091 (.058) -.030 -.280 (.077)"* -.055
PRESSURE .657 (.322)" .226 1.245 (.436)*“ .310
EXTCONS .633 (.251)" .214 1.035 (.327)*" .228
FFWAVAIL .505 (.274)* .172 -.240 (.377) -.045
CAMPAIGN -.356 (.301) -.l 1 l 1.07 (.403)"* .276

 

 

 

 

 

*..4 -.

bu
fax
the
of
ant
ten

Wer

204

 

 

 

 

 

 

 

 

 

 

 

Table 8.3 (Cont’d)

6. Household demographic characteristics

AGEHSHED .012 (.011) .004 .013 (.014) .002

MALEHEAD .017 (.397) .006 .680 (.480) .177

LITERACY .309 (.319) .096 -.326 (.381) -.069

Constant -l.603 (1.001) --- -3.61 (1.38)“‘ --

chi-square 45.33 --- 46.03 -

prob. > chi-square 0.0016 --- 0.0013 «-

pseudo R-square 0.2527 --- 0.412 «-

predicted prob. at x-bar --- .73 --- .88
74 --- --

 

 

*, **, ""‘ signiﬁcant at 10%, 5% and 1% respectively.

8.4 Summary

The likelihood of farmer adoption of stone terraces was higher than that of soil
bunds. Long-term land tenure security of farmers and household capacity to invest, which
favored adoption of stone terraces detracted from the adoption of soil bunds. The factors
that inﬂuenced adoption and intensity of use of stone terraces were different. The degree
of effort farmers put into conserving soil was inﬂuenced by the opportunity costs of labor
and foregone land productivity. Farmers had a high probability of maintaining stone
terraces once they were built. Terrace maintenance and adoption of biological practices

were chieﬂy inﬂuenced by socio-institutional factors.

CHAPTER 9

SUMMARY, CONCLUSIONS AND IMPLICATIONS

9.1 Introduction
9.1.1 Problem Statement

Land degradation, particularly soil erosion, constitutes the basis for the problems of
the low productivity and sluggish grth of Ethiopian agriculture. The principal
determinants of soil erosion in Ethiopia include hilly and mountainous topography, erosive
rainfall, inappropriate agricultural practices, and agricultural policies that were biased
against small holders. Massive efforts have been put into combating land degradation in
the country since early 1970, mostly in the form of food-for-work (FFW) projects.
However, these eﬁ‘orts have been minuscule compared with the magnitude of the problem
and conservation practices have not been sustained by farmers.

The northern region of Tigray is one of the most seriously degraded areas in
Ethiopia. Increasing land pressure, continuous cultivation, inadequate manuring,
overgrazing and cleared forests have left the land barren and susceptible to erosion.
Intense tropical rain storms wash away topsoil from the virtually unprotected lands in the
region. Lack of appropriate dryland agricultural development policies and civil wars have
crippled the economic base of the region.

Since the downfall of the dictatorial military government in 1991, environmental
rehabilitation and the conservation and development of natural resources have been at the

center of the rural development policies pursued by the regional government of Tigray.

205

 

 

206
Three general approaches have been used for this purpose: (1) rural communities have been
organized in mandatory public campaign work for resource conservation on public or
private lands, (2) extension services have been active in transferring conservation
technologies to farmers, and (3) FFW projects have been used to rehabilitate hillsides and
construct small scale irrigation dams. This study was designed to ﬁll a gap in knowledge
regarding the determinants of farmer ad0ption and intensity of use of soil conservation '-
practices in Tigray, Ethiopia. An understanding of the factors inﬂuencing the adoption and

use of conservation practices could be used in formulating conservation intervention

policies aimed at facilitating wider adoption and sustained use of conservation practices.

9.1.2 Research Objectives

In Tigray, stone terraces and soil bunds are the most widely used physical
conservation practices. Biological conservation practices including grass strips, contour
plowing, crop rotation and tree planting are also used. The general objective of the study
was to determine the physical, social, institutional and economic factors that affect farmer
adoption of soil conservation practices in the region.

The speciﬁc objectives were to:
(1) Characterize the farming systems of the study area,
(2) Estimate the crop yield and farm proﬁtability impact of stone terraces,
(3) Analyze the levels of and factors that affect farmer perceptions of soil erosion and
conservation,

(4) Analyze the determinants of farmer adoption of conservation practices,

 

207

(5) Analyze the factors that explain intensity of use of conservation practices

9.1.3 Conservation Adoption Studies: An Overview

Quantitative and qualitative research aimed at understanding the factors inﬂuencing
farmer adoption of conservation practices has been going on since the 193 0's. Most of
these studies were conducted in the developed world using detailed technical information
and considerable rigor. Studies of soil conservation adoption in the developing world have
been fewer and more qualitative. This implies that, in the developing countries, a more
concerted and sustained investigation of the causes of land degradation and the factors
affecting farmer conservation practice is required.

Statistical and non-statistical methods have been employed in the study of adoption
of conservation practices. The predominant statistical method has been multiple regression
analysis. The non-statistical methods used in conservation adoption research include
simulation and mathematical programming models. While the statistical methods have been
more appropriate for studying the determinants of farmer adoption of conservation
practices, the non-statistical methods have been used to assess the proﬁtability and yield
impact of the practices.

Because farmers need to perceive the problem of soil erosion before they will invest
in preventing it, a study of farmer perception of the problem becomes an important
component of conservation adoption studies. Farmer adoption and intensity of use of
conservation practices are likely to be separate decisions. A farm household ﬁrst decides

on using a conservation practice and subsequently decides how much of the practice to use.

 

208

Most studies have ignored this distinction. Moreover, the determinants of adoption and
intensity of use of conservation practices may differ by type of practice.

Farmer conservation adoption decisions are complex. Research results to date have
shown that demographic, physical, economic, institutional and perceptual variables all
contribute to explaining the conservation behavior of farmers. Hence, conservation
adoption studies need to expand the number of explanatory variables considered. Cross
sectional studies that explicitly differentiate between types of conservation practices and
between adoption and intensity of use of practices, and that incorporate the analysis of
farmer perceptions of the problem of soil erosion can provide results indicative of
important causal factors.

This study analyzed the determinants of adoption by type of practice, and
diﬂ‘erentiated between the determinants of adoption and intensity of use. Adoption results
were supplemented by analysis of the determinants of farmer perception of soil erosion.
Data were collected from a randomly selected sample of 250 farm households in Tigray,
northern Ethiopia in 1995/96. Data were collected on conservation practices of farmers for
the 1992-95 period, as well as agroecological and physical characteristics of villages,
physical characteristics of plots, socio-institutional factors, and demographic characteristics
of farm households. Probit, ordered probit and double hurdle models were used to analyze
the survey data. In addition, a quasi-experimental ﬁeld plot study was conducted to
evaluate the yield and proﬁtability impact of stone terraces. Analysis of variance and

investment analysis were used to analyze the ﬁeld plot data.

 

209

9.2 Summary of Results
9.2.1 Farming Systems of the Study Area

The study covered an area of about 400 km’, consisting of thirty villages in six
districts in south central Tigray. Seventy-ﬁve percent of the surveyed villages depended
primarily on animal dung for ﬁrel. Food-for-work projects were operational in seventeen of
the villages during 1992-95. Agriculture is predominantly mixed subsistence farming, and
land preparation is entirely based on oxen power. Eighty-three percent of households were
headed by men. The average family size was 4.8 with three economically active persons per
household.

Land has been distributed to farmers based on family size and farmers have only
usuﬁuct rights to land. Farmers feel more insecure about their land holding in the long-
terrn than in the short-term. During 1995-96, cultivated land per household in the study
area averaged 1.7 hectares divided among 3.5 plots. About 26 percent of farmers reported
having leased out land during 1995-96. Share cropping was the most common rental form.
Households headed by women were more likely to lease land out.

Farmers commonly experience seasonal labor shortages. Use of hired labor was
limited, and farmers depended on exchange labor to alleviate the problem. Twenty-nine
percent of surveyed farmers participated in off-farm activities. The most important off-
farrn work was construction, followed by trade and selling ﬁrewood. Farmers living in
villages with more degraded soils and forests were more likely to participate in off-farm

work. By contrast, farmers residing in remote villages were less likely to participate.

 

210

Interestingly, surveyed farmers believed that the major causes of land degradation
were caused by humans. In order of importance, they were deforestation, overcultivation,
poor agricultural practices and overgrazing. By comparison, intensity of rain and steep
slopes were considered less important factors of land degradation.

Seventy percent of surveyed farmers had conservation work done on their plots by
community campaign labor during 1992-95. Thirty-two percent of surveyed farmers
invested in stone terraces, while 60 percent applied biological practices during 1992-95.
During the same period, 60 percent of surveyed farmers reported having maintained
physical conservation structures. Farmers in the study area were divided regarding their
preference between mandatory public campaign and private conservation work, with 57
percent preferring campaign work. Surveyed farmers held the opinion that conservation
increases crop yield and enhances yield stability. However, farmers also reported
operational problems with stone terraces, including reduction of land area and harboring

pests such as rats.

9.2.2 Impact of Stone Terraces on Crop Yields and Farm Proﬁtability

A watershed in central Tigray was selected for a on-farm quasi-experimental study
to ﬁll the information gap regarding the effect of stone terraces on crop yields and farm
proﬁtability. Seventy terraced and 70 non-conserved plots were equally divided between
wheat and fava beans. Two quadrates of 8 m2 were marked on each of the terraced plots;
one just above the terrace (soil accumulation zone) and another one parallel to this but

below the next upper terrace (soil loss zone). Only one quadrate (control zone) was

 

211

marked on each of the non-conserved plots. Yields from the conserved plots were adjusted
for 5 and 15 percent land loss due to terrace construction.

After yields were adjusted for 5 percent land loss, four major results stood out: (1)
grain and straw yields for both crops were signiﬁcantly higher in the soil accumulation zone
than in the soil loss zone or in the non-terraced control zone, (2) crop and straw yields
ﬁom the soil accumulation zone were more stable than those from the control zone, (3)
fava bean beneﬁted less than wheat from stone terraces, and (4) investment in stone
terraces was proﬁtable both at the bank loan agricultural rate of 15 percent and at a higher
presumed farmer discount rate of 50 percent. However, proﬁtability nearly dropped to the
break-even level at the 50 percent discount rate, indicating the need for public intervention

to encourage wider adoption of stone terraces.

9.2.3 Farmer Perceptions of' Soil Erosion and Conservation

Farmers in the study area perceived soil erosion to be severe. Their perception
could be explained statistically primarily by village and ﬁeld physical characteristics.
Location of plots did not affect for farmer perception of soil erosion. In general, the
perception was greatest where erosion potential was greatest. Farmers perceived soil
erosion to be less severe on plots more distant ﬁ'om the homestead. The perception of soil
erosion was less among those households with more experience in farming (older heads),
and those having contact with the extension service for natural resource conservation. The
latter result, although apparently counter-intuitive, suggests that knowledge of

conservation practices reduces farmer rating of the severity of soil erosion.

 

 

212

The results from the analysis of farmer perception of the yield impact of soil erosion
were, in general, consistent with the results from the analysis of the severity of soil erosion.
Most of the physical factors that explained perception of yield impact of soil erosion
signiﬁcantly also explained the perception of erosion severity, with the same direction of
eﬁ‘ect. Contact with extension service, which reduced perception of soil erosion, also
reduced perception of the yield impact of soil erosion. Farmers who beneﬁted from public
campaign conservation work perceived lower yield impact of soil erosion. Older household
heads, who perceived soil erosion to be less severe, also perceived lower yield impact of
soil erosion.

Farmer perceptions of the effectiveness of conservation practices in reducing soil
erosion were primarily determined by village and plot physical characteristics. In general,
the factors associated with higher erosion perception were also associated with higher
perceived effectiveness of conservation practices. These results supported the ﬁndings
from the models of farmer perceptions of soil erosion and its yield impact.

Farmers in the study area had different preferences between public campaign and
private conservation work, the two most widely used approaches for soil conservation in
the Tigray region. Analysis of the determinants of the preferences showed that, on
average, farmers had a slight preference for public campaign work. The propensity of
those who already had public campaign conservation work done on their plots to substitute
their own labor for the campaign work was lower. Farmers who had contact with the
extension service preferred to do their own work. Highlanders preferred public campaign

work.

 

 

213

9.2.4 Determinants of Adoption and Intensity of Use of Conservation practices
Physical conservation practices

Probit regressions showed that the adoption of stone terraces and soil bunds was
aﬂ‘ected by a wide range of factors including physical factors, household capacity to invest,
risk of investment, socio-institutional and household demographic factors. On average, the
probability of adoption of stone terraces and soil bunds was low (22.0 and 1.3 percent,
respectively).

Farmers tended to use stone terraces when soil erosion was likely to be more
severe. Soil textures that were easier to work with were positively associated with soil
bunds. Irregular slope shapes of plots detracted from adoption of soil bunds. Location of
plots was important in explaining the adoption of both stone terraces and soil bunds.
Farmers used stone terraces on lower sloping land, but they used soil bunds on ﬂat lands.

Family labor availability favored the adoption of stone terraces but did not matter
for soil bunds. Long-term land tenure security contributed to the adoption of stone
terraces but detracted from the adoption of soil bunds. On the other hand, insecurity of
land tenure detracted from adoption of stone terraces but increased the adoption of soil
bunds. Land tenure stability at the village level also favored stone terraces. These results
indicate that land tenure security is an important determinant of investments that have long-
term payoff.

Availability of food-for-work projects at the village level favored adoption of stone

terraces but detracted from the adoption of soil bunds, suggesting that stone terraces had

 

214

been the focus of such projects. Public campaign conservation work substituted for own
investment for both stone terraces and soil bunds.

Age and literacy of household head tended to reduce adoption of soil bunds but did
not matter for stone terraces. Experience and access to information might have proven that
soil bunds were less effective in reducing soil erosion.

The factors that affected the length of stone terraces constructed were different
ﬁom the factors that affected their adoption. Households located in the highland zone,
although less prone to adopt stone terraces, tended to build more terraces if they did adopt.
This result indicates that foregone land productivity might be more important in explaining
the level of use of conservation practices since land productivity is higher in the highland
zone. Distance of villages from the nearest market place and all-weather road (ﬁnancial
incentives to invest factors) were insigniﬁcant in explaining adoption but explained
intensity of use positively, indicating that once adoption took place, low opportunity cost
of labor encourages investment in stone terraces. Tenure security, institutional factors and
household demographic factors did not inﬂuence effort.

Farmer maintenance of stone terraces were determined by a wide range of factors.
Probit regression showed that on average, farmers were highly likely to maintain stone
terraces (with 88 percent probability). Low opportunity cost of labor encouraged
maintenance. Institutional factors that favored maintenance activities included extension

contact, community pressure and beneﬁt from campaign conservation work.

215

Biological conservation practices

Biological conservation practices including grass strips, crop rotation, contour
plowing and tree planting were widely adopted. The likelihood of their adoption was
explained by relative severity of land degradation in a village, household capacity to invest,
risk of investment and socio-institutional factors. Higher family labor availability and
tenure security detracted from adoption of biological practices, perhaps through their effect
in favoring stone terraces. Villagers who resided in villages with more degraded lands were
more likely to adopt biological practices. Institutional factors encouraged adoption,
notably contact with the extension service and availability of FFW, perhaps through
provision of technical information and tree seedlings. Farmer perception of community

pressure to conserve soil encouraged use of biological practices.

9.3 Policy Implications

No previous study has combined village, household and plot level data for the
analysis of the determinants of farmer conservation behavior in Ethiopia. This research
attempted to analyze results of four years of conservation efforts (1992-1995) in Tigray
region. Since the study had a limited area coverage and the analysis was based on farmer
recall data and farm management data collected from one production year, caution must be
exercised in drawing speciﬁc policy recommendations and actions for conservation
intervention for the entire region of Tigray. However, some general recommendations for

policy, research and extension are warranted.

 

216

9.3.1 Implications for Conservation Policy

Agricultural scientists tend to disregard indigenous farmer knowledge of soil
conservation. Natural resource conservation and development programmes and projects
are usually designed and implemented without farmer involvement. This study has shown
that Tigrayan farmers perceived soil erosion as a severe problem for agricultural
production. Their perceptions of soil erosion and conservation were found to be consistent
and reasoned. This indicates the need for and importance of farmer involvement in the
design of conservation programmes and projects and the choice of speciﬁc techniques.
Farmers need to take primary responsibility for implementation of conservation
programmes and projects.

Soil erosion on upland areas results in negative extemalities downstream, and the
effectiveness of conservation efforts on lower Slopes can be undermined in the absence of
similar efforts on upper slope areas. This linkage demands a watershed approach to soil
conservation. The public campaign soil conservation work being pursued in Tigray seems
to be in line with the watershed approach. However, this study has shown that the public
campaign work and the extension-assisted private conservation efforts are substitutes for
each other. The public campaign approach may have alleviated the problem of labor
shortage and eased the dmdgery that farm households would have faced to construct
conservation practices. Once all rural lands, private or communal, that suffer erosion have
been conserved, public campaign work needs to be restricted to communal lands and

emphasis needs to be given to strengthening the extension service to facilitate private

217

investment in soil conservation. Appropriate public policies should accompany the private
investments.

The analysis of adoption determinants highlighted the importance of secure land
tenure for encouraging farmers to adopt conservation practices such as stone terraces that
have a long-term payoff. It would be difﬁcult to expect farmers to invest in land
development unless they enjoy secure land tenure for a long enough period to appropriate
the returns. This research found that perceived land tenure security inﬂuences the choice of
conservation practice, and it may compromise the suitability of the practices chosen. With
secure land tenure, the choice of techniques is likely to be dictated by suitability. Hence, a
legal assurance of farmer land and tree tenure rights remains crucial to facilitate wider use
of soil conservation practices in Tigray. This legal assurance of land tenure need not,
however, take the form of private ownership of land.

Most Ethiopian farmers are risk averse and tend to discount ﬁiture income heavily.
The stone terrace investment analysis showed that the proﬁtability of investment in stone
terraces dropped almost to break-even at a farmer discount rate of 50 percent. Moreover,
farmers may not be able to pay for the initial investment cost calling for targeted credit for
terrace construction. The 50 percent internal rate of return of investment in stone terraces
also justiﬁes public subsidy. The provision of credit or public subsidy should be
coordinated with extension services, the latter serving as sources of technical and
educational support.

Soil conservation in the region has focussed on curing soil erosion rather than on

preventing it. This approach could be justiﬁed by the severity of soil erosion that has

218

already taken place in the region and the time it takes to design land use systems to prevent
soil erosion. However, a long-term soil conservation strategy needs to focus on changing
the farmer land management practices that induce soil erosion, a focus on preventing soil
erosion rather than on curing it. One way of doing this is by developing a conservation
package that combines physical, biological and agronomic practices. At present,
conservation and intensiﬁcation appear to be considered as separate objectives in the
region. In fact, conservation and intensiﬁcation can and do complement each other.
Hence, long-term strategies of agricultural development in the region need to combine

conservation and intensiﬁcation technologies.

9.3.2 Implications for Agricultural Extension

F ifty-seven percent of surveyed farmers reported having contact with extension
service related to soil conservation during 1992-95. However, contact with extension
service explained neither adoption of stone terraces and soil bunds, nor their intensity of
use. Mobilization of the community labor for public campaign conservation work may
have taken too much time from extension agents. The ﬁiture emphasis of extension service
for resource conservation needs to be on encouraging private investment in soil
conservation.

Adoption of different conservation practices are affected by different factors. The
extension services should recognize this fact in their programs. They should devote more
effort to the highland zone to promote use of physical conservation structures and their

maintenance. The distance of plots from homestead tend to detract from adoption, intensity

 

219

and maintenance of physical conservation practices. Extension services need to take this
fact into account during the program planning stage. The ﬁnding that age and literacy of
household heads detract from adoption of soil bunds suggests that extension service may
nwd to target such households in areas where soil bunds are more suitable to use. Farmers
are highly likely to maintain stone terraces once they are built, suggesting that extension
services may not need to deal with maintenance separately.

The factors that affect adoption and intensity of use of conservation practices seem
to be different. This suggests the need for different intervention strategies for adoption and
sustained use of conservation practices. To encourage effort among those who adopt stone
terraces, the extension services need to focus on households with higher opportunity cost

of labor and areas of low land productivity.

9.3.3 Implications for Conservation Research

Very limited research has been conducted in Tigray regarding soil erosion and
conservation. In particular, very little research has been done on the traditional
conservation practices of Tigrayan farmers. An inventory of the traditional conservation
practices needs to be done in order to identify the practices that are still in use from those
that have been abandoned. An understanding of the reasons for the abandonment of some
of the practices and the continuing use of the others could be a valuable information for the
design of modern conservation techniques. Moreover, case studies of a few villages which
have successﬁrlly instituted conservation practices would provide important insights about

the sustainable use of conservation practices at the community level.

 

220

The same conservation techniques are recommended through out the region. Such
a blanket recommendation may result from the lack of research results regarding the
feasibility and effectiveness Of the conservation practices in different parts of the region. It
is likely that different areas may have need for different conservation practices. Wind
erosion may be more important than water erosion in some parts of Tigray. Hence, an
interdisciplinary study on the feasibility and effectiveness of conservation practices for the
diﬁ‘erent parts of Tigray would be required.

Proﬁtability is an important determinant Of investments. Very little research has
been conducted on the private or social proﬁtability of conservation practices in Tigray.
The on-farm experiment conducted in this study to determine the impact Of stone terraces
on wheat and fava bean yields and farm proﬁtability needs to be replicated to include other
important crops in the region such as sorghum, barley and teff. The study also needs to be
conducted on different agroecological zones over several years in order to account for
changes in climatic conditions. A more systematic treatment Of the social proﬁtability
analysis is also required.

The battle against land degradation will continue for many years. Stistained
interdisciplinary conservation research plays a key role in this battle. Due to the
aggressive rural-centered development strategy being pursued in Ethiopia, the socio-
institutional, economic and agroecological environment under which Tigrayan farm
households Operate is likely to change. As a result, conservation policies need to be
updated based on current and emerging research ﬁndings on the economic status Of

farmers, feasibility, effectiveness and proﬁtability of conservation practices, and

 

221

determinants of farmer adoption of the practices. This requires the institutionalization of

conservation research in Tigray.

9.4 Limitations of the Study

Farmer adoption Of conservation practices is a complex decision involving
agroecological, physical, economic, and socio-institutional factors. The conservation
behavior Of farmers is partly site-speciﬁc. The limitations of the study emerge ﬁ'om these
facts.

The most important limitation of the study is its limited time and area coverage, and

small sample size. Although an attempt was made to represent the two most important
agroecological zones in the region, time and logistics did not allow for a wide enough
coverage or a large enough sample size to permit drawing deﬁnitive policy
recommendations for the entire region of Tigray. Replicating the study in different areas of
the region, particularly in the eastern and western zones could help draw some general
conclusions.

Biological conservation practices were treated here as a group, but future research
should disaggregate them by type of practice at the plot level. As the practices have
different characteristics, separate analysis Of the determinants Of adoption of each practice
could provide more useful information.

This research has shown that Tigrayan farmers perceive the problem Of soil erosion
as a severe problem for agricultural production. Investment in conservation practices may

increase and enhance the stability of crop yields but public intervention may be required to

 

 

222

make the investment attractive to farmers. Extension service need to treat different
conservation practices differently. Farmers make decisions on adoption and intensity of use
Of conservation practices separately suggesting different intervention strategies for

adoption and use intensity.

APPENDIX

 

 

223

APPENDIX I

The Economics of Soil Conservation Investments in the Tigray Region of Ethiopia
Village, Household and Plot Survey Instrument

Name Of Enumerator

 

Name of Woreda

 

Name of Tabia

 

Name Of Village

 

Name of Respondent

Village Level

 

Household and Plot Level

 

Date Of Interview

 

A. Village Characteristics
1. Agroecology Agroeco

1. Highland 2. Intermediate Highland 3. Lowland

 

 

 

2. Altitude Altitud

3. Predominant Topography Topogra
1. Hilly 2. Plain 3. Intermediate

4. Predominant Vegetation Vegetat
1. Bush 2. Grass 3. Wood

4. Forest 5. Barren

224

5. Three major crops grown in normal year in order of importance
by area occupied Majcrop

1 st 2nd 3 rd

 

 

 

6. How many times per year do farmers in the village produce? Intcrop

 

 

 

 

 

7. Year land distributed last in village Landistr
8. Number Of land classiﬁcation categories Landcls

9. Land classiﬁcation categories Landctg

1. 2. 3.

10. Size of one gibri Landsze
11. Size of land allotted to children Landchd
12. Size Of land allotted to an adult Landadt
13. What is the size of an average family? Fmszeav

 

14. What is the area Of land holding of this average family

 

in Timads? Avehold
15. Is there a variation of duration of land lease in the village? Varlease
1. Yes 0. NO

16. What is the average duration in years of leasing out land

 

in village? leease
17. Is there a lumpsum payment for leasing in land? Lplease
1. Yes 0. No
18. The rental arrangement for leasing land is in the form of: Rtlease
1. Cash 2. Share cropping 3. Other (specify)—

225

19. Rental Rate for leasing land:

 

 

lumpsum payment: (Per timad or plot)
if cash: (per timad)
if share: (ratio of share)

 

20. Type Of Two major livestock reared in the village:

 

 

21. Distance of village from nearest market place in walking hours:

 

22. Distance of village from all weather road

 

23. Major off-farm involvement of village people by number
of people involved:

1. Arho 2. Construction 3. Handcraft 4. Weaving

Lumpay

Cashpay

Shratio

Mjlvstk

Mktdist

Roaddist

Offarm _

‘5. Food processing 6. Transport 7. Trade 8. Other (specify) __

24. Total Number of Households in village:

 

25. Number Of male-headed households in village:

 

26. Number OfFemale-headed households in village:
27. Major source of cooking and heating fuel in village:
1. Forest and bush 2. Animal dung 3. Crop residue

4. Other (specify)

 

28. Average round trip distance traveled for collection of

fUClWOOd in walking hours:

 

Totlhd
Malehsd
Femlhsd

Fuelsrc

Fueldist

 

 

226

29. Predominant soil texture in village: Soiltyp __
1. Sandy 2. Clayey 3. Silt 4. Loam

30. Is there shortage Of ﬁrelwood in village? Fuelsrt __
1. Yes 0. No

31. Are there crops that are disappearing (decreasing in area coverage)

 

in the last ﬁve years? Crpdisp

crop Major Reason

 

 

 

 

 

 

32. Are there crops that are gaining popularity in the village

 

in the last ﬁve years? Croplr __
33. What is the number of farm households with the following
ox/0xen ownership?

Size of holding Number of households

1. Without any ox

 

2. Only one ox

 

3. A pair Of oxen

 

4. Three oxen

 

5. Two pairs of oxen

 

6. More than two pairs of oxen

 

34. Is there delineated grazing area in the village? Grzeare

 

227

35. What is the size of the grazing area in timad?

 

36. Is grazing land during rainy season adequate in the village?
1. Yes 0. NO
37. Is grazing land during dry season adequate in the village?
1. Yes 0. NO
38. What is the major reason for the inadequacy Of grazing land during
the rainy season?
1. Extensive cultivation 2. Highly degraded land

3. Other (specify)

 

39. What is the reason for the inadequacy of grazing land during
the dry season?
1. Highly degraded land 2. Area closure

3. Other (specify)

 

40. How do villagers try to alleviate the problem of shortage of
grazing land?

1. Using straw and weeds 2. Using tella residue

3. By Cut and carry 4. Other (specify)
41. Is there closed area in the village?

1. Yes 0. NO

42. What is the size of the closed area in Timad?

 

43. When was the area closed?

Gnesiz

Rngrzad

Drygzad

Rngzind

Drgzind

strtsl __

Areacls

Szeclos

Yrclose

 

228

44. What beneﬁts do villagers get from the closed area?

 

 

 

45. Is there a natural forest(s) in the village? Natfrst __
1. Yes 0. No

46. Is there an man-made forest(s) in the village? Artfrst
1. Yes 0. No

47. What are the major problems of agricultural production

in the village in order Of importance?

 

 

 

B. Farm Operations (1995/96 Cropping Year)

1. How many plots did you cultivate last cropping season? Plotcvt

 

Note T o Enumerator: Please number the plots by size: 1= largest etc.. In case of equal

size, use distance from homestead. The nearest will be numbered ﬁrst.

2. Have you leased out land? Ldlseut

1. Yes 0. NO

Note to Enumerator : If the answer to question B(2) above is NO, go to question 7; If

YES skip question 7.

 

1"-

 

229

3. What is the size Of the plots you leased out?

 

4. What was the rental arrangement?

1. Cash 2. Share cropping

3. Share cropping and lumpsum payment 4. Other (specify)

5. What was the rental rate?

Cash (Birr/timad)

 

Ratio of share

 

Lumpsum payment

 

6. What was your reason for leasing land out?

1. Lack of oxen 2. Lack of labor 3. Lack Of seed

4. Other (specify)

 

7. If you have not leased out land, what was the reason?
1. Could not ﬁnd leasee
2. Low crop share and lumpsum payment

3. Oxen and labor available

4. Other (specify)

 

8. Have you fallowed any of your plots last cropping season?
1. Yes 0. No

9. Have you fallowed any of your plots a year ago?
1. Yes 0. NO

10. Have you fallowed any of your plots ﬁve years ago?

1. Yes 0. No

Szlseut

Lserent

Lsereas

 

Ntlsers

Flwlstyr

Flwyrago __

Flwfvyr

230

11. Indicate the number Of cattle and other animals you own:

 

 

 

 

 

 

 

 

 

oxen Cows Heifer
Calves Goat Sheep
Donkey Mule Camel
Poultry Beehives

 

 

12. What did you use for plowing last cropping season?

(Choose all that apply) leuse __
l. Oxen 2. Cow 3. Donkey 4. Mule
5. Other (specify)

13. Did you have enough draft power of your own last cropping season? Enghdrft _
1. Yes 0. NO

14. Ifthe answer to question B(13) above is no, What was your source of additional

 

plowing power? (Choose all that apply) Sorcdrft
1. Oxen exchange 2. Cash rent 3. Share cropping
4. Help ﬁ'om relatives and ﬁiends 5. None

6. Other (specify)

 

15. Did you use improved seed last cropping season? Seedimp

1. Yes 0. No

Note to Enumerator: If the answer to question B(15) above is NO, Go to question 18; if
yes skip question 18.

l6.What was the improved seed(s)?

231

 

 

 

17. What was the source of the improved seed? Srcimpsd
1. Market 2. Credit 3. Aid 4. Other (specify)
18. Ifyou did not use improved seeds last cropping season, what was

the reason(s)? Rsnntsd

 

1. Was not available 2. Did not have purchasing power
3. Did not want to use them 4. Other (specify)

19. Did you have enough local seed of your own last cropping season? Loclsdad _
1. Yes 0. NO

Note to Enumerator: If the answer to question B(l9) above is yes, go to question 22.

20. What was the crop (s) for which you faced seed shortage?

 

 

 

21. What was your source of additional seed? Lclsdsrc
1. Credit 2. Market 3. Relatives and friends 4. Aid

5. Other (specify)

 

22. Did you have enough agricultural implements of your own? Agimpad

1. Yes 0. No

Note to Enumerator: If the answer to Question B(22) above is yes, go to question 26.

23. What was the implement(s) for which you faced shortage?

 

 

 

 

 

 

232

24. What was the cause(s) of the shortage? Impsrtrs
1. Was not available 2. Did not have purchasing power

3. Other (specify)

 

25. What did you do to alleviate the shortage? Srceimp
1. Borrowed ﬁom friends 2. Shared with others 3. Bought on credit

3. Other (specify)

 

26. Did you use modern fertilizer last cropping season? theuse

1. Yes 0. No

Note to Enumerator: If the answer to question B(26) above is No, go tO question 28; if

YES, skip question 28.

27. What was the source of the fertilizer? thesrc
1. Yes 0. No

28. What was the reason for your not using fertilizer? Ntftzrs

1. Was not available 2. Did not have purchasing power
3. Did not want to use it 4. Other (specify)
29. Did you have labor shortage for farm operation last cropping season? Lbrshrt___
1. Yes 0. No
Note to Enumerator: If the answer to question B(29) above is NO, skip questions 30-31.
30. For what crops and Operations did you face labor shortage?

Crop Operation Month

 

 

 

233

 

 

 

31. What was the source of additional labor?
1. Hired labor 2. Exchange labor 3. Help from ﬁiends and relatives
4. Other (specify)

32. Plot Description:

 

 

Plot 1 Plot 2 Plot 3 Plot 4

 

Cultivated by owner since when

 

Dominant soil texture

 

slope degree

 

Shape of slope

 

Distance ﬁom home stead

 

Size in timad (farmer estimate)

 

Size in hectare (from measurement)

 

owned or leased in

 

 

Water logged or not

 

Irrigated or not

 

Rental arrangement if leased

 

Conserved or not

 

Type of Conservation practice

 

Year conservation practice was built

 

who constructed the conservation practice

 

Conservation practice maintained or not

 

Who did the maintenance

 

 

 

 

 

 

 

Location of plot

 

234

C. Household information

 

 

1. Age of household head Agehead _
2. Sex of household head Sexhshed __
3. Household size Szehshd __ '
4. Number of male household members Mlehshd __
5. Number of female household members Flehshd __
6. Literacy of household head Literacy _
1. Literate 0. Illiterate

7. Number of household members by age and sex

 

 

 

 

 

 

 

 

 

 

 

 

 

Age category Male Female Total
0-5
6.10
1 1-17
18-64
above 64
8. Are you sending any of your children to school? Schlchd __
1. Yes 0. No
9. Number of children being sent to school N umschcd _

10. Major occupations of family members by age and sex
1. Herding 2. Field work 3. Cultivation 4. Household

5. Other (specify)

 

235

 

 

 

 

 

 

 

 

 

 

 

 

 

Age Sex Three major occupations by time spent
1st 2nd 3rd
6-10 female
male
11-17 female
male
18-64 female
male

 

 

 

11. Have you or members of your family participated in food-for-work
program for soil and water conservation in the last three years? Fdwkswc

1. Yes 0. No

 

Note to Enumerator: If the answer to question C(11) above is NO, go to question 16; If

YES, skip questions 16-17.

12. About how many total days have you and your family

members worked on this projects? Tldyswk __

13. DO you think that the conservation work projects have or will

contribute to improve crop yields? chrpyd _
1. Yes 0. No

14. If the conservation project was in your village, do you think

that your community has beneﬁted from the project? antcom__

1. Yes 0. No

236

15. Ifthe answer to question C(11) above is NO, Was there food-for-work

project in your community? Fdwkava _
1. Yes 0. NO

16. Ifthe answer to question C(15) above is YES, why did you or members

of your community not participate in the food-for-work project? N tfdwkrs __
1. Did not want to 2. Did not get the chance 3. Old age
4. Illness 5. Other (specify)

17. Did you hire labor for agricultural work last cropping season? Lbrhire _

1. Yes 0. No

Note to Enumerator: If the answer to C(17) above is NO, go to question 22.

18. What type of labor did you employ? (Choose all that apply) Lbrtype__
1. Permanent male 2. Permanent female 3. Casual male
4. Casual female

19.What was the duration of employment in days for:

Permanent male permanent Female

 

 

Casual Male Casual Male

 

20. How do you characterize the availability of permanent hired

labor in your community? Lprpmava___
1. Abundant 2. Scarce 3. Seasonally Scare (specify season)

21. How do characterize the availability of casual hired

labor in your community? Lbrcsav

237

1. Abundant 2. Scare 3. Seasonally scarce (specify Season)
22. What was the major work done by permanent hired labor? M jpmlb _
. 1
(choose all that apply)

 

1. Land preparation 2. Cultivation 3. Harvesting 4. Other (specify)

 

23. What was the work done by casual hired labor? Mjcslb
(choose all that apply)

1. Land preparation 2. Cultivation 3. Harvesting 4. Other (specify)

 

24. Did you or members of your family participate in Oﬁlfarm
work last cropping season? Ofarmwk

1. Yes 0.No

Note to Enumerator: Ifthe answer to question C(24) above is NO, go to question 29.

25. How many family members including you were involved? Pernum

26. What was the off-farm work? (Choose all that apply) M jofmwk _
1. Trade 2. Construction 3. Weaving 4. Food processing
5. Tailoring 6. Crafts 7. Transport 8. Arho

9. Other (specify)

 

27. How many total days were you and members of your family
involved in off-farm work? Dyswkd
28. How much income did your family get from Off-farm work? Incomgt

29. Were you and members Of your family involved in SWC campaign

238

work last cropping season? chgnwk _
30. How many family members including you were involved? Persinv

31. How many total days were you and members of your family

involved in SWC campaign work? Ttldywk _
32. What is the starting time for agricultural work in a day

during the peak period? Agrstart _
33. What is the ﬁnishing time for agricultural work in a day

during the peak period? Agrf'rnsh__
34. How many times in a day do you take rest from agricultural

work during the peak period? Timerest _
35. What is the duration of each rest in hours? Durest

36. How many religious holidays do you strictly observe (not involved in any agricultural
work at all) in a month? Religdys __

37. What are the three most important constraints of agricultural production in your area?

 

 

38. How much do you pay for taxes per year? Taxpay

39. What type of taxes do you pay?

 

 

 

40. Do you have access to agricultural credit? Crditac
41. In the last two crop years, what have you done with the intent of increasing crop yield
on your plots? (Choose that apply) Yldincf'r

1. Used fertilizer 2. Used manure 3. Used improved seed

239

4. Used insecticides and herbicides 5. Invested in soil and water conservation

42. What community work were you involved in during last cropping year?

 

 

 

 

 

Type Of community work Number of days

43. Ifyou had additional income, how would you spend it? Priorivt
1. Invest in agriculture 2. Invest in off-farm work
3. Home consumption 4. House construction

5. Education of children 6. Other (specify)

 

44. What are your three priorities (in order of importance) for

government support of agriculture in your area? thsptpr_
1. Credit for buying draft power 2. Credit for seed
3. Credit for fertilizer 4. Agricultural extension education

5. Supply of farm inputs 6. Development of irrigation structure

7. Other (specify)

 

D. Soil and Water Conservation Activities
1. How do you perceive the problem of soil erosion in your plots before conservation
practice was built?

1. Severe 2. Moderate 3. Slight 4. NO problem

Plot 1 Plot 2

 

 

Plot 3 Plot 4

 

 

240

2. How do you compare the problem of soil erosion in your plots after conservation

practice was built with that before the conservation practice?

 

 

 

 

1. Aggravated 2. Same 3. Reduced
Plot 1 Plot 2
Plot 3 Plot 4
3. What do you think is the major cause of soil erosion? Mjercaus _
1. Deforestation 2. Overgrazing 3. Over cultivation
4. Poor agricultural practices 5. Steep sIOpes
6. Intensity Of rain 7. Other (specify)
4. What do you think is the consequence of soil erosion? Eroscons _

1. Land preparation becomes more difficult
2. Change in types of crops grown is induced
3. Land productivity is reduced

4. Other (specify)

 

5. DO you think that time needed for land preparation is reduced as a result Of conservation

practice? Lndprcns _
1. Yes 0. No

6. If the answer to question D(5) above is YES, by how much? Lndprhow _

i 1. One-half 2. One-third 3. One-fourth 4. One-ﬁfth

6. Can not estimate

241

7. By how much do you think soil erosion reduced land productivity in your plots

before conservation practice was built assuming a normal cropping year? Ersprd __
1. One-ﬁfth 2. One-fourth 3. One-third 4. One-half
5. No reduction 7. Other (specify)

8. Have you built physical conservation practice on your plots by your own

 

in the last three years? Ownphcn _
1. Yes 0. No

9. If the answer to question D(8) above is YES, what is the

physical conservation practice? Phyconpr_
1. Stone terrace 2. Soil bund 3. Check dam
4. Other (specify)

10. Have you used agronomic (biological) conservation practices on your plots by your

own in the last three years? Ownbio __
1. Yes 0. NO

11. If the answer to question D(lO) above is YES, What is the practice? Biocnpr _
1. Grass strip 2. Tree planting 3. Crop rotation

3. Contour plowing 4. Other (specify)

 

12. Have you done maintenance work of the physical conservation practices

in your plots by your own in the last three years? Ownmtphy_
1. Yes 0. No
13. What was the physical conservation practice you maintained? Mtconpr

1. Stone terrace 2. Soil bund 3. Check dam

242

4. Other (specify)

14. Ifthe answer to question D(l3) above is NO, what is the reason for not doing any
maintenance work? Ntmtres

1. There was no need for it 2. Shortage of labor

 

3. Did not want the practice to be maintained 4. Other (specify)

 

15. Ifthe physical conservation practices on your plots were built by campaign work,

would you have built them by yourself had there been no campaign work? Owncagn _
1. Yes 0. NO

16. Ifthe answer to question D(15) above is NO, what is the reason? Ntownrs __
1. Do not want the practice 2. DO not have enough labor

3. Land tenure insecurity 4. Other (specify)

 

17. DO you think that the conservation practices have resulted in increased yield as
compared to the yield without conservation practice in normal year? Yldcons
1. Yes 0. No

18. Are there any problems that stone terraces are causing on your

farming operation? (Choose all that apply) Stnterpr__
1. Land reduction 2. Harboring pests
3. Difﬁculty in plowing 4. Water logging
5.0ther (specify)-

19. Are there any problems that soil bunds are causing on your
farm operation? (Choose all that apply) Soilbdpr

1. Land reduction 2. Harboring pests

243

3. Diﬁiculty in plowing 4. Water logging 5. Other (specify)
20. How do you think stone terraces affect the stability of crop yield? Stntrstb
1. Increase stability 2. Decrease stability 3. NO difference

21. How do you think soil bunds affect crop yield stability? Slbdstb

 

1. Increased stability 2. Reduced stability 3. No difference

 

22. How do you think biological conservation practices affect

the stability of crop yield? Bioprstb

 

1. Increased stability 2. Decreased stability 3. NO difference

23. DO you believe that you will cultivate the same plots ﬁve years

 

from now? OwnSyrs __
l.Yes 2. No 3. Am not sure

24. DO you expect you will inherit your plots to your children? Owninhrt _
1. Yes 2. NO 3. Am not sure

25. How do you think your community would feel about you if you do not use

conservation practices on your plots? Pressure _
1. Positively 2. Negatively 3. Neutral

26. Would you prefer to construct conservation practices on your plots by yourself rather

than being involved in campaign work? Owncgprf _
1. Yes 0. No

27. Did you have any extension education on soil and water conservation? Extcons

1. Yes 0. No

244

28. DO you think investment in conservation practice is proﬁtable? Cnsproft
1. Yes 0. No

29. What traditional soil conservation practices do (did) you use?

 

 

 

30. What change would you suggest to the current public campaign work approach to soil

 

conservation?

 

 

 

31. Private investment in soil conservation

 

Plot 1 Plot 2 Plot 3 Plot 4

 

Type of physical practice

 

Length in meters of physical practice

 

Persondays used in building the practice

 

Persondays used in maintaining the

 

practice

 

Type of biological practice

 

Size of the agronomic practice

 

Persondays used in setting up the

 

 

biological practice

 

 

 

 

 

REFERENCES

245

REFERENCES

Abraha, E., B. Dimtsu., and T. Woldu. “Land Races and Improved Varieties in Tigray.”
Paper presented at Regional Workshop on Agricultural Extension in Tigray,
Mekelle. February, 1996.

Adesina AA, and M.M. Zinnah. “Technology Characteristics, Farmers’ Perceptions and
Adoption Decisions: A Tobit Model Application in Sierra Leone.” Agricultural
Economics, 9(1993):297-3 l 1.

Admassie, Y.. “ The Catchment Approach to Soil and Water Conservation in Kenya.”
Regional Soil Conservation Unit/SIDA Report NO. 6. 1992.

Admassie, Y., and S. Gebre. “Food-For-Work in Ethiopia: A Socioeconomic Survey.”
Unpublished report to UN/FAO WFP, Addis Ababa. 1985.

Amare, M. “Physical Geography of the Central Zone of Tigray” in Proceedings of the
Workshop on Agro-Ecological and Health Studies in the Central Zone of Tigray
(unpublished). Mekelle, March 1995.

Anderson, D., and R. Grove. (eds) Conservation in A ﬁica: People, Policies and
Practice. Cambridge: Cambridge University Press. 1987.

Anderson, J.R., and J. Thampapillai. “Soil Conservation in Developing Countries--Project
and Policy Intervention.” The World Bank Policy and Research Series 8.
Washington DC: The World Bank. 1990.

Asgedom, Y., and M. Yohannes. “Vegetation Resources of Central Tigray” in
Proceedings of the Workshop on Agro-Ecological and Health Studies in the Central
Zone Of Tigray (unpublished). Mekelle, March 1995.

Ashby, J.A., J.A. Beltran, M.P. Guerrero and HF. Ramos. “Improving the Acceptability
to Farmers of Soil Conservation Practices.” Journal of Soil And Water
Conservation, 51(4):309-3 12. 1996.

Barbier, E.B.. “The Farm Level Economics of Soil Conservation: The Uplands of Java.”
LandEconomics, 66(2): 199-21 1. 1990.

246

Barbier, E.B., Burgess, I .C. and Markandya, A.. “The Economics of Tropical
Deforestation.” Ambio, 20(2): 55-88. 1991.

Bauer, D.F.. Household and Society in Ethiopia: An Economic and Social Analysis of
Tigray Social Principles and Household Organization. Monograph No.6, second
edition, occassional paper series, Northeast African Studies Committee. East
Lansing: African Studies Center, Michigan State University. 1985.

Bauer, D.Z. “Land, Leadership and Legitimacy Among the Inderta Tigray of Ethiopia.”
Ph.D. Dissertation, University of Wisconsin. 1972.

Belay, A.. “General Aspects of Some of the Major Cultural Practices in Tigray.” Paper
presented at Regional Workshop on Agricultural Extension in Tigray. Mekelle.
February, 1996.

Belay, A.. “Climate, Major Agro-Ecological Zones, and F arming Systems in Tigray.”
Paper presented at Regional Workshop on Agricultural Extension in Tigray.
Mekelle. February, 1996.

Bennett H...H Elements of Soil Conservation. New York: McGraw-Hill Book
Company, Inc.. 1955.

Berhane, G. “Livestock Production in Central Tigray” in Proceedings of the Workshop
on Agro-Ecological and Health Studies in the Central Zone of Tigray. Mekelle,
March 1995.

Belsley, D.A., E. Kuh, and RE. Welsch. Regression Diagnostics Identijjring Inﬂuential
Data and Sources of Collinearity. New York: John Wiley and Sons. 1980.

Belsely, T. and Case, A.. “Modeling Technology Adoption in Developing Countries.”
American Economic Review, 83(2):396-402. 1993.

Bojo, 1.. “Cost-Beneﬁt Analysis of Soil and Water Conservation Projects: A Review of
20 Empirical Studies.” in Tato, K. and H. Humi. (eds) Soil Conservation For
Survival. Ankeny: Soil and Water Conservation Society. 1992.

Bojo, J.P.. “Economics and Land Degradation.” Ambio, 20(2):75-79. 1991.

Bojo, J., and Cassells, D.. “Land Degradation and Rehabilitation in Ethiopia: A
Reassessment.” AFTES Working Paper No. 17. Washington DC: The World
Bank.

Bosch, D.V.. “From the Swiss Alps to the Simien Mountains: A Proﬁle of Hans Humi.”
Journal of Soil and Water Conservation, 49(3):228-230. 1994.

 

 

 

247

Bruce, J.W.. “Land Reform Planning and Indigenous Communal Tenures.” Ph.D.
Dissertation. University of Wisconsin. 1976.

Bultena, G.L. and E0. Hoiberg. “Factors Affecting Farmers Adoption of Conservation
Tillage.” Journal of Soil and Water Conservation, 38(3):281-284. 1983.

Burt, O.R. “Farm Level Economics of Soil Conservation in the Palouse Area of the
Northwest.” American Journal of A gricultural Economics, (February 1981):83-
92.

Campbell, 1.. “Land or Peasants?: The Dilemma Confronting Ethiopian Resource
Conservation.” Aﬁican A ﬂairs, 90(358):S-21. 1991.

Carlson, J.E., B. Schnabel, C.E. Beus, and DA. Dillman. “Changes in the Soil
Conservation Attitudes and Behaviors of Farmers in the Palouse and Camaras
Prairies: 1976-1990.” Journal of Soil and Water Conservation, 49(5):493-500.
1994.

Cary, J.W. and R.L. Wilkinson. “Perceived Proﬁtability and Farmers’ Conservation
Behavior.” Journal of Agricultural Economics, 48(1): 13-21. 1997.

Christensen, G.N.. “Determinants of Private Investment in Rural Burkina F aso.” Ph.D.

Dissertation, Department of agricultural Economics, Cornell University, Ithaca,
NY.

CIMMYT. From Agronomic Data to Farmer Recommendations: An Economics
Training Manual. Completely revised edition. Mexico, DP. 1988.

Clay, D.C., T. Reardon and J. Kangasniemi. “Sustainable intensiﬁcation in the Highland
Tropics: Rwandan Farmers’ Investments in Land Conservation and Soil Fertility.”
Paper forthcoming in Economic Development and Cultural Change (accepted
October, 1996).

Convery, F.J.. “Applying Environmental Economics in Africa.” World Bank Technical
Paper No.27. Washington DC: The World Bank.

Cragg, J.G. “Some Statistical Models For Limited Dependent Variables with
Application to the Demand for Durable Goods.” Econometrica, 39(5):829-844.

Crosson, P. and JR. Anderson. “Achieving a Sustainable Agricultural System in Sub-
Saharan Aﬁica.” Environmentally Sustainable Development Division, Africa
Technical Department Paper No.2. Washington, DC: The World Bank. 1995.

 

 

248

Dejene, A. Environment, Famine, and Politics in Ethiopia: A View From the Village.
Lynne Reinner Publishers, Inc.. 1990.

de la Briere, B. L.. “Household Behavior Towards Soil Conservation and Remittances
in the Dominican Sierra.” Ph.D. Dissertation, University of California, Berkeley.
1996.

Delgado, CL, and P. Anderson. “Agricultural Production in the Third World: Patterns
and Strategic Issues.” Keynote address to the AAEA/IFPRI Pre—conference
Workshop, Orlando, Florida. 1993.

Demeke, M., W. Amha, S. Ehui, T. Zegeye. (eds) Food Security, Nutrition and Poverty
Alleviation in Ethiopia: Problems and Prospects. Proceedings of the inaugural and
ﬁrst annual conference of the Agricultural Economics Society of Ethiopia (AESE).
Addis Ababa. 1995.

Dixon, J.A. and Sherman, P.B.. “Economics of Protected Areas.” Ambio, 20(2):68-74.
1991.

Eicher, C.. “Zimbabwe’s Green Revolution: Preconditions for Replication in Africa.” staﬂ‘
paper 94—1, East Lansing, Department of Agricultural Economics, Michigan State
University. 1994.

Ervin, CA, and DE. Ervin. “Factors Affecting the Use of Soil Conservation Practices:
Hypotheses, Evidence and Policy Implications.” Land Economics, S8(3):272-292.
1982.

Ervin, DE. and A. Graffy. “Technology for Production and Environmental Quality: Are
We Missing Opportunities for Complementarity?” Journal of Soil and Water
Conservation, 50(4):352-3 53. 1995.

Ervin, D.E.. “Soil and Water Conservation Down on the Farm: A Changing Economic
Landscape.” Journal of Soil and Water Conservation, May-June 1994: 232-234.
1994.

Ervin, D.E.. “Soil Erosion Control on Owner Operated and Rented Cropland.” Journal
of Soil and Water Conservation, 37(5):285-288. 1982.

Ervin, D.E.. “Getting Conservation Practice Adopted: An Agricultural Economist’s
Viewpoint .” in Soil and Water Conservation: The Principle and the Practice.
Special Report 290, Ag. Experiment Station, University of Missiouri, Colombia.
1982.

 

249

Ervin, DE. and CT. Alexander. “Soil Erosion and Conservation in Monroe County,
Missouri: Farmer Perceptions, Attitudes and Performances.” Department of
Agricultural Economics, University of Missouri-Columbia Report No. 81-10. 1981.

Feder, G., L]. Lau, J .Y. Lin, and X. Luo. “The Determinants of Farm Investment and
Residential Construction in Post-Reform China.” Economic Development and
Cultural Change, 41(1): 1-26. 1992.

Feder, G., R. Just and D. Zilberrnan. “Adoption of Agricultural Innovations in
Developing Countries: A Survey.” Economic Development and Cultural Change,
33(January 1985):255-298. '

Firebrace, 1., and G. Smith. The Hidden Revolution: An Analysis of Social Change in
T igray Based on Eyewitness Accounts. London: War on Want. 1982.

Fitiwi, I., S. Oukbarebi and T. Belay. “Plant Protection Problems in Tigray.” Paper
presented at Regional Workshop on Agricultural Extension in Tigray. Mekelle.
February, 1996.

Foster A.B.. Approved Practices in Soil Conservation. Danville: Interstate Printers
and Publishers Inc.. 1964.

Gantzer,.C.J. and TR. McCarty. “Predicting Corn Yields on a Clay Pan Soil Using a
Soil Productivity Index.” Trans. ASAE 30(1987):1347-1352.

Gebremedhin, B., B. Assefa, H. Zelelew, Bahta, M. “Agriculural Credit and Marketing
in Tigray: Assessment of Current Conditions and Future Directions.” Paper
presented at Regional Workshop on Agricultural Extension in Tigray. Mekelle.
February, 1996.

Gebremedhin, B., S. Swinton, and Y. Tilahun. “Effects of Stone Terraces on Crop
Yields: Results of On-farm Research in Central Tigray, Ethiopia.” Poster
presented at the 52nd annual conference of the Soil and Water Conservation
Society (SWCS), Toronto, Canada, July 1997.

Gebremedhin, B., S. Swinton, and T. Jayne. “Effectiveness of Food-For-Work
Targeting in Tigray, Ethiopia.” Poster presented at the 23rd International
Conference of Agricultural Economists, Sacramento, California, August, 1997.

Gebremichael, Y.. “The Effect of Conservation on Production in the Andit-Tid Area.”
in Tato, K. And H. Humi (eds) Soil Conservation For Survival. Ankeny: Soil and
Water Conservation Society. 1992.

 

250

Gebru, G., M. Mesﬁn, and T. Assfaw. “Health and Environmental Sanitation in the
Central Zone of Tigray.” in Proceedings of the Workshop on Agro-ecological and
Health Studies in the Central Zone of Tigray. Mekelle, March 1995.

Gittinger, J.P.. Economic Analysis of Agricultural Projects. Washington DC: The
World Bank. 1982.

Ghose, A.J.. “Transforming Feudal Agriculture: Agrarian Change in Ethiopia Since
.1974.” Journal of Development Studies, 22(1): 127-149. 1985.

Gould, B.W., WE. Saupe and RM. Klemme. “Conservation Tillage: The Role of Farm
Operator Characteristics and the Perception of Soil Erosion.” Land Economics,
65(2): 167-182.

Gray, R.S., J.S. Tylor and W.J. Brown. “Economic Factors Contributing to the
Adoption of Reduced Tillage Technologies in Central Saskatchewan.” Canadian
Journal of Plant Science, 76(1996): 661-668.

Greene, W.H.. LIMDEP Version 7.0 User ’s Manual. New York: Econometric Software
Inc.. 1995

Greene, W.H.. Econometric Analysis. Second edition. New York: Macmillan Publishing
Company. 1990.

Grepperud, 8.. “Population Pressure and Land Degradation: The Case of Ethiopia.”
Journal of Environmental Economics and Management, 30 (1996):18-33.

Griliches, Zvi. “Hybrid Corn: An Exploration in the Economics of Technological
Change.” Econometrica, 25(1957):501-522.

Haile, M. “Indigenous Knowledge and Agricultural Practices in Central Tigray” in
Proceedings of the Workshop on Agro-Ecological and Health Studies in the Central
Zone of Tigray (unpublished). Mekelle, March 1995.

Haile, M. “Soil Resources of Central Tigray.” in Proceedings of the Workshop on Agro-
ecological and Health Studies in the Central Zone of Tigray (unpublished). Mekelle,
March 1995.

Hansen, D.O., J.M. Erbaugh and TL. Napier. “Factors Related to Adoption of Soil
Conservation Practices in The Dominican Republic.” Journal of Soil and Water
Conservation, 42(5):367-369. 1987.

Heady, E.O.. Economics of A gricultural Production and Resource Use. Englewood Cliffs:
Prentice-Hall, Inc.. 1964.

 

251

Heady, 13.0. and GP. Vocke. Economic Models of A gricultural Land Conservation and
Environmental Improvement. Ames: Iowa State University. 1992.

Heﬂ‘erman, W.D. “Getting Conservation Practice Adopted: A Rural Sociologists
Viewpoint.” in Soil and Water Conservation: The Principles and The Practice,
Special Report 290. Agricultural Experiment Station, University of Missouri,
Columbia. 1981.

Helms, G.L., Bailey, D. and TE Glover. “Government Programs and Adoption of
Conservation Tillage Practices on Nonirrigated Wheat Farms.” American Journal
of Agricultural Economics, 69(4):787-795. 1987.

Hoben A., “Paradigms and Politics: The Cultural Construction of Environmental
Policies in Ethiopia.” Working Papers in African Studies No. 193. Boston
University: Aﬁican Studies Center. 1995.

Hold, 1., and Lawrence, M., Making Ends Meet: A Survey of the Food Economy of the
Ethiopian North-East Highlands. London: Save the Children. 1993.

Holden, S.T., B. Shiferaw, and M. Wik. “Poverty, Market Imperfections, and Time
Preference: Of relevance for Environmental Policy?” Contributed paper presented
at the 23rd International Conference of Agricultural Economists, Sacramento,
California, August, 1997.

Hoover, H, R. Schultz, R. and K. Reed. “Farmer Attitudes Towards Soil Conservation
and Water Quality: Has Government Made A Difference?” Paper prepared for
presentation at the 37th annual meeting of the Soil Conservation Society of
America, August 8-11, New Orleans, Louisiana. 1982.

Hudson, N.. Soil Conservation. Ames: Iowa State University Press. 1995.

Humi, H. “Mountain Research for Conservation and Development in Simien, Ethiopia.”
Mountain Research and Development, 1(1):49-54. 1981.

Humi, H. “Soil Erosion and Soil Formation in Agricultural Ecosystems in Ethiopia and
Northern Thailand.” Mountain Research and Development, 3(2): 131-142. 1983.

Hurni, H. “Erosion-Productivity-Conservation Systems in Ethiopia.” Paper presented at
the 4th International Conference on Soil Conservation, Venezuela, November,
1985.

Humi, H. “Degradation and Conservation of the Resources in the Ethiopian Highlands.”
Mountain Research and Development, 8(2/3): 123-130. 1988.

 

 

 

252

Hurni, H. and K. Tato.(eds). Erosion, Conservation and Small-Scale Farming.
Marceline: Walsworth Pub. Com. Inc. 1992.

Hussien A. M.. (ed) Drought and Famine in Ethiopia. Aﬁican Environment Special
Report 2. London: International African Institute. 1976.

ICRISAT, “Farming Systems Principles for Improved Food Production and the
Control of Soil Degradation in the Arid, Semi-Arid, and Humid Tropics.”

Proceedings of an experts’ meeting sponsored by the United Nations Environment
Programme. Patancheru: ICRISAT. 1983.

Keddeman, W.. “An Economic Analysis of Soil Conservation Projects in Ethiopia.” in

Tato, K. and H. Humi (eds) Soil Conservation for Survival (pp. 239-250). Ankeny:

Soil and Water Conservation Society. 1992.

Kejela, K.. “The Costs of Soil Erosion in Anjeni, Ethiopia.” in Tato, K. and H. Hurni
(eds) Soil Conservation for Survival (pp. 218-230). Ankeny: Soil and Water
Conservation Society. 1992.

Kuru, A.. “Soil Conservation and Strategic State Policy: The Case of Ethiopia.”
Ph.D. Dissertation, Faculty of Agriculture and Forestry, University of Helsinki.
1986.

Lal, H. and Wyant, J .G.. “Wholistic Interpretation of Conservation Needed.” Journal of
Soil and Water Conservation, May-June 1994:23 1.

Lamb, Rand 8le. “Soil Erosion , Real Causes of Ethiopian Famine.” Environmental
Conservation, 10(2): 157-159. 1983.

Lanz, T.J.. “Environmental Degradation and Social Conﬂict in the Northern Highlands
of Ethiopia: The Case of Tigray and Wollo Provinces.” Africa Today, 43 (2): 157-
182, 1996.

Larney, F.J., R.C. Izaurralde, HH. Janzen, B.M. Olson, E.D. Solberg, C.W. Lindwall,
and M. Nyborg. “Soil Erosion-Crop Productivity Relationships for Six Alberta
Soils.” Journal of Soil and Water Conservation, 50(1):87-91. 1995.

Lee, L.K.. “The Impact of Land Ownership Factors on Soil Conservation.” American
Journal of Agricultural Economics. December(1980): 1070- 1 075.

Legesse, G. “The Livestock Health Situation in Central Tigray.” in Proceedings of the
Workshop on Agro-ecological and Health Studies in the Central Zone of Tigray
(unpublished). Mekelle, March 1995.

j _.
_‘.I

253

Lin, CT]. and J. W. Milon. “Attributes and Safety Perceptions in a Double-Hurdle
Model of Shellﬁsh Consumption.” American Journal of Agricultural Economics,
75(August 1993): 724-729.

Lipschutz, R.D.. “What Resources Will Matter?” National Geographic Research and
Exploration, 9(1):4-6. 1993.

Lockeretz, W.. “What Have We Learnt about Who Conserves Soil.” Journal of Soil
and Water Conservation, 45(5):517-523. 1992.

Lopez-Preira, M.A., 'J.H. Sanders, T.G. Baker, P.V. Preckel. “ Economics of Erosion-
Control and Seed-Fertilizer Technologies for Hillside Farming in Honduras.”
Agricultural Economics, 1 1(1994):27 1-288.

Lovejoy, SB. and TL. Napier. (eds). Conserving Soil: Insights ﬁom Socioeconomic
Research. Ankeny: Soil Conservation Society of America. 1986.

Lungern, L..”Twenty Years of Soil Conservation in Eastern Aﬁica.” SIDA Regional Soil
Conservation Unit Report No.9. Nairobi. 1993.

Mackenzie, D.. “Can Ethiopia Be Saved?” New Scientist, September 1987(1579):54-58.

Mackenzie, D.. “Ethiopia’s Hand to the Plough.” New Scientist, October
1987(1580):52-55.

Magrath, W.B.. “Economic Analysis of Soil Conservation Technologies.” The World
Bank Environment Department Division Working Paper No. 2. Washington
DC: The World Bank. 1989.

McConnel, K...E “An Economic Model of Soil Conservation.” American Journal of
Agricultural Economics, 65(February 1983): 83-89.

Mch 1.. From Poverty to Famine in Northeastern Ethiopia: A Rural History 1900-
1935. Philadelphia: University of Pennsylvania Press. 1987.

Meles, G., and Hagos, T., “Strengthening Agricultural Extension Intervention in
Tigray.” Paper presented at Regional Workshop on Agricultural Extension in
Tigray. Mekelle. February, 1996.

Millington, A. “Local Farmer Perceptions of Soil Erosion Hazards and Indigenous Soil
Conservation Strategies in Sierra Leone, West Africa.” Paper presented at the 4th
International Conference on Soil Conservation, Maracay, Venezuela, 1985.

254

Ministry of Agriculture. Guidelines for Development Agents on Soil Conservation in
Ethiopia. Addis Ababa: Community Forestry and Soil Conservation Development
Department. 1986.

Morgan,'R.P.C. (ed). Soil Conservation: Problems and Prospects. New York: John Wiley
and Sons Ltd.. 1981.

Meulen, H.A.B. van der, G.R. de Snoo and G.A.A. Wossink. “F arrners’ Perception of
Unsprayed Crop Edges in the Netherlands.” Journal of Environmental
Management, 47(1996): 241-25 5.

New Scientist. “Saving Ethiopia From Itself.” World Press Review, 34(11): 52

 

Norman, D. and M. Douglas. “Farming Systems Development and Soil Conservation.”
FAO Farm Systems management Series 7. Rome: FAO. 1994.

Nowak, P. “The Adoption of Agricultural Natural Resource Management Practices: A .
Critical Overview.” Discussion paper prepared for the Soil and Water Conservation
Society Task Force on Socio-Economic Research, Minneapolis, April 29-30, 1993.

Nowak, P.. “Why Farmers Adopt Production Technologies.” Journal of Soil and Water
Conservation, 47(1): 14-16. 1992.

Olson, K.R., R. Lal, and L.D. Norton. “Evaluation of Methods to Study Soil Erosion-
Productivity Relationships.” Journal of Soil and Water Conservation, 49(6):586-
590. 1994.

Omas, A. H., and M. Salih. (eds) Ecologi and Politics: Environmental Stress and
Security in Africa. Scandinavian Institute of African Studies. 1989.

Omas, AH. and M. Salih. “Research and Development Issues for African Drylands.”
Ambio, 20(2):388-394. 1991 .

Oyhus, A.O. “Introduction to Proceedings” in Proceedings of the Workshop on Agro-
ecological and Health Studies in the Central Zone of Tigray (unpublished). Mekelle,
March 1995.

Oyhus, A0. and G. Gebru. “Development Yardistics.” in Proceedings of the
Workshop on Agro-Ecological and Health Studies in the Central Zone of Tigray
(unpublished). Mekelle, March 1995.

Pagiola, 8.. “Estimating the Yield Effect of Soil Degradation.” Journal of Soil and
Water Conservation, 47(6):486-490. 1992.

255

Painter, K.M., D.C. Young, D.M. Granastein and DJ. Mulla. “Combining Alternative
and Conventional Systems for Environmental Gains.” American Journal of
Alternative Agriculture, 10(2):88-96. 1995.

Painter, KM. and D.L.Young. “Environmental and Economic Impacts of Agricultural
Policy Reform: An Interregional Comparison.” Journal of Agricultural and Applied
Economics, 26(2):45 1-462. 1994.

Pankhurst, R.. A Social History of Ethiopia. Addis Ababa: Addis Ababa
University, Institute of Ethiopian Studies. 1990.

Pankhurst, R.. State and Land in Ethiopian History. Addis Ababa: Hailesselassie I
University, Institute of Development Studies. 1996.

Pausewang, S., F. Cheru, S. Brune, and E. Chole. (eds) Rural Development Options.
London: Zed Books Ltd. 1990.

Pearce, D. and Maler, K.G.. “Environmental Economics and the Developing World.” _
Ambio, 20(2):52-54. 1991.

Pender, J. and J. Kerr. “Determinants of Farmers’ Indigenous Soil and Water
Conservation Investments in India’s Semi-Arid tropics.” IFPRI Environmental and
Production Technology Division Discussion Paper, No. 17. Washington DC.
1996.

Rahm, MR, and Huffman, W.B.. “The Adoption of Reduced Tillage: The Role of Human
Capital and Other Variables.” American Journal of A gricultural Economics,
November(1984):405-413.

Reardon, T., V. Kelly, E. Crawford, Jayne, T., K. Savadogo and D. Clay. “Determinants
of Farm Productivity in Africa: A Synthesis of Four Case Studies.” MSU
International Development Working Paper No. 22, Department of Agricultural
Economics. East Lansing. 1996.

Regional Bureau of Agriculture (Tigray). “Agricultue in Tigray.” Paper presented at
Symposium on Agricultural Development in Humera, Tigray. Mekelle. 1995.

Regional Bureau of Economic Development and Planning (Tigray). Five-year (1996-2000)
Develpment Plan of Tigray Regional Administration. Mekelle, Tigray. 1995.

Relief Society of Tigray (REST). Farming Systems, Resource Management and Household
Coping Strategies in Northern Ethiopia: Report of a Social and A gro-ecological
Baseline Study in Central T igray. Oslo: NORAGRIC. 1995

 

256

Relief Society of Tigray (REST). Report of the Symposium on Environmental
Degradation, 15-20 April, 1992, Mekelle.

Rikoon, J .S., D.H. Constance and S. Geletta. “Factors Affecting Farmers’ Use and
Rejection of Banded Pesticide Application.” Journal of Soil and Water
Conservation, 51(4):322-329. 1996.

Ryder, R. “Farmers Perceptions of Soils in the Mountains of The Dominican Republic.”
Mountain Research And Development, 14(3):261-266. 1994.

Sain, GE. and HJ. Baretto. “The Adoption of Soil Conservation Technologies in El
Salvador: Linking Productivity and Conservation.” Journal of Soil and Water
Conservation, 51(4):313-321. 1996.

Salih, J.M., and V.L. Monis. “Farm Implements in Tigray.” Paper presented at Regional
Workshop on Agricultural Extension in Tigray. Mekelle. February, 1996.

Sanders, D.W.. “New Strategies for Soil and Water Conservation.” Journal of Soil and
Water Conservation, 45(5):511-516. 1992.

Seiz, W.D., C.R. Tylor, R.G.F Spitze, C.Osteen and MC. Nelson. “Economic Impacts
of Soil Erosion Control.” Land Economics, 55(1):28-42. 1979.

Shaxson, T.F., N.W. Hudson, D.W. Sanders, E. Roose, and WC. Moldenhauer. Land
Husbandry A Framework for Soil and Water Conservation. Ankeny: Soil and
Water Conservation Society. 1989.

Stahl, M.. “Land Degradation in East Africa.” Ambio, 22(8):505-508. 1993

Stahl, M.. “Constraints to Environmental Rehabilitation Through People’s Participation
in the Northern Ethiopian Highlands.” United Nations Research Institute For Social
Development (UNRISD) Discussion Paper 13. Switzerland. 1990.

StataCorp. Stata Statistical Software: Release 5.0 College Station, TX: Stata Corporation.
1997

Stocking, M.. “Socioeconornics of Soil and Water Conservation in Developing
Countries.” Journal of Soil and Water Conservation, September-October
(1988):381-385.

Stocking, M. and L. Peake. “Erosion-Induced Loss in Soil Productivity: Trends in
Research and International Cooperation.” Keynote presentation, 4th International
Conference on Soil Conservation, Maracay, Venezuela. 1985.

257

Stocking, M.. “Soil Conservation Policy in Colonial Aﬁica.” Agricultural History,
59(2):]48-161. 1985. -

Sureshwaran, S., S.R. Londhe and P. Frazier. “Factors Inﬂuencing Soil Conservation
Decisions in Developing Countries: A Case Study of Upland Farmers in the
Philippines.” Journal of Agribusiness, 14(1):83-94. 1996.

Tegegne, 8.. “Effects of Erosion on Properties and Productivity of Euritic Nitosols in
Gununo Area, Southern Ethiopia.” in Hurni, H. and K. Tato (eds) Erosion,
Conservation and Small Scale Farming. Ankeny: Soil and Water Conservation
Society. 1992.

Teka, G., and F. Edwards. “Possible Soil and Water Conservation Recommendations in
Tigray.” Paper presented at Regional Workshop on Agricultural Extension in
Tigray. Mekelle. Febrauary, 1996.

Tietenberg, T.. Environmental and Natural Resource Economics. New York:
HarperCollins Publishers Inc. 1992.

Tigray Bureau of Agriculture. “Agriculture in Tigray: An Overview.” Paper presented
at Symposium on Agricultural Development in Humera, Tigray. Mekelle, March
1996. '

Tjemstrorn, R.. “Yields from Terraced and Non-Terraced Fields in the Machakos District
of Kenya.” in Tato, K. and H. Humi (eds) Soil Conservation For Survival.

Ankeny: Soil and Water conservation Society. 1992.

Tilahun, Y. “Impact of Conservation Bunds on crop yields in Degua Temben, Northern
Ethiopia.” Msc Thesis, Agricultural University of Norway, Oslo. 1996.

Tsegaye, M. “Soil Fertility Management in Tigray” Paper presented at Regional
Workshop on Agricultural Extension in Tigray. Mekelle. February, 1996.

Tobin, J. “Estimation of Relationships for Limited Dependent Variables.”
Econometrica, 26(195 8):24-36.

Turner, K.. “Economics and Wetland Management.” Ambio, 20(2):59-63. 1991.
UNDP, ECA, TDA and Regional Government of Tigray. “Statistical Master Book of
Tigray.” Sustainable Agriculture and Environmental Rehabilitation in Tigray

(SAERT) project document vol. 13. Mekelle, 1994.

Varnis, S.L. “Policy and Performance in Ethiopia.” Society, 25(6):38-41. 1988.

258

Virgo, KI. and RN. Munro. “Soil and Erosion Features of the Central Plateau Region
of Tigray, Ethiopia.” Geoderma, 20(1978): 13 1-157.

Watson, C.J., P. Billingsley, D.J. Croft, D.V. Huntsberger. Statistics for Management
and Economics. Massachusets: Allyn and Bacon. 1990.

Weesies, G.A., S.J. Livingston, W.D. Hosteter, and D.L. Schertz. “Effects of Soil Erosion
on Crop Yield in Indiana: Results of a 10 Year Study.” Journal of Soil and
Water Conservation, 49(6):597-600. 1994.

White, T.A., and IL Jickling. “Peasants, Experts, and Land Use in Haiti: Lessons ﬁ'om
Indigenous and Project Technology.” Journal of Soil and Water Conservation,
50(1):7-14. 1995.

Witter, S.G., MP. Robotham and DA. Carrasco. “Sustainable Adoption of
Conservation Practices by Upland Farmers in the Dominican Republic.” Journal of
Soil and Water Conservation, 51(3):249-254. 1996.

Wolde-Georgis, T.. “Survival Strategies and Coping Mechanism of Peasants to Famine:
The Case of Tigray, Ethiopia.” Ph.D. Dissertation, Faculty of the Graduate School
of International Studies, University of Denver. 1993.

Wondimagegn, F.. “Crop Production in the Central Zone of Tigray” in Proceedings of
the Workshop on Agro-ecological and Health Studies in the Central Zone of
Tigray. Mekelle, March 1995.

World Bank. Ecologically Sensitive Sites in Africa. Volume 11: East Africa. Washington
DC: The World Bank. 1993.

Yen, S.T.. “Working Wives and Food Away From Home: The Box-Cox Double Hurdle
Model.” American Journal of Agricultural Economics, 75(Nov. 1993):884-
895.

 

HICHIGRN STRTE UNIV. LIBRARIES
Ill11111llllHIWWIlll111111llllllllIIWIHHHI
31293017120233